Professional Documents
Culture Documents
FF
FF
Causes and
Recommendations
April 2004
U.S.-Canada Power System Outage Task Force
Causes and
Recommendations
April 2004
U.S.-Canada Power System Outage Task Force
We are pleased to submit the Final Report of the U.S.-Canada Power System Outage
Task Force. As directed by you, the Task Force has completed a thorough investigation
of the causes of the August 14, 2003 blackout and has recommended actions to minimize
the likelihood and scope of similar events in the future.
The report makes clear that this blackout could have been prevented and that immediate
actions must be taken in both the United States and Canada to ensure that our electric
system is more reliable. First and foremost, compliance with reliability rules must be
made mandatory with substantial penalties for non-compliance.
We expect continued collaboration between our two countries to implement this report’s
recommendations. Failure to implement the recommendations would threaten the
reliability of the electricity supply that is critical to the economic, energy and national
security of our countries.
The work of the Task Force has been an outstanding example of close and effective
cooperation between the U.S. and Canadian governments. Such work will continue as we
strive to implement the Final Report’s recommendations. We resolve to work in
cooperation with Congress, Parliament, states, provinces and stakeholders to ensure that
North America’s electric grid is robust and reliable.
We would like to specifically thank the members of the Task Force and its Working
Groups for their efforts and support as we investigated the blackout and moved toward
completion of the Final Report. All involved have made valuable contributions. We
submit this report with optimism that its recommendations will result in better electric
service for the people of both our nations.
Sincerely,
ii G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Contents
Page
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2. Overview of the North American Electric Power System and Its Reliability Organizations . . . . . . . 5
The North American Power Grid Is One Large, Interconnected Machine . . . . . . . . . . . . . . . . . . . . 5
Planning and Reliable Operation of the Power Grid Are Technically Demanding . . . . . . . . . . . . . . 6
Reliability Organizations Oversee Grid Reliability in North America . . . . . . . . . . . . . . . . . . . . . . . 10
Key Parties in the Pre-Cascade Phase of the August 14 Blackout . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3. Causes of the Blackout and Violations of NERC Standards. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
The Causes of the Blackout in Ohio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Linking Causes to Specific Weaknesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Institutional Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4. Context and Preconditions for the Blackout: The Northeastern Power Grid
Before the Blackout Began . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Electric Demands on August 14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Generation Facilities Unavailable on August 14. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Unanticipated Outages of Transmission and Generation on August 14 . . . . . . . . . . . . . . . . . . . . . . 27
Key Parameters for the Cleveland-Akron Area at 15:05 EDT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Power Flow Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Voltages and Voltage Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Past System Events and Adequacy of System Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Model-Based Analysis of the State of the Regional Power System at 15:05 EDT,
Before the Loss of FE’s Harding-Chamberlin 345-kV Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
System Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5. How and Why the Blackout Began in Ohio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Chapter Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Phase 1: A Normal Afternoon Degrades: 12:15 EDT to 14:14 EDT . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Phase 2: FE’s Computer Failures: 14:14 EDT to 15:59 EDT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Phase 3: Three FE 345-kV Transmission Line Failures and Many Phone Calls:
15:05 EDT to 15:57 EDT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Phase 4: 138-kV Transmission System Collapse in Northern Ohio: 15:39 to 16:08 EDT . . . . . . . . . 68
6. The Cascade Stage of the Blackout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Phase 5: 345-kV Transmission System Cascade in Northern Ohio and South-Central Michigan . . 77
Phase 6: The Full Cascade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Phase 7: Several Electrical Islands Formed in Northeast U.S. and Canada: 16:10:46 EDT
to 16:12 EDT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Why the Blackout Stopped Where It Did . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Under-Frequency and Under-Voltage Load-Shedding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Why the Generators Tripped Off . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
7. The August 14 Blackout Compared With Previous Major North American Outages . . . . . . . . . . . . . 103
Incidence and Characteristics of Power System Outages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Outage Descriptions and Major Causal Factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Common or Similar Factors Among Major Outages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Comparisons With the August 14 Blackout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G iii
Contents (Continued) Page
8. Performance of Nuclear Power Plants Affected by the Blackout. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Findings of the U.S. Nuclear Working Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
Findings of the Canadian Nuclear Working Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Perspective of Nuclear Regulatory Agencies on Potential Changes to the Grid . . . . . . . . . . . . . . . . 129
9. Physical and Cyber Security Aspects of the Blackout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Summary and Primary Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
SWG Mandate and Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Cyber Security in the Electricity Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Information Collection and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Cyber Timeline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
10. Recommendations to Prevent or Minimize the Scope of Future Blackouts . . . . . . . . . . . . . . . . . . . 139
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Recommendations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
Appendixes
A. Members of the U.S.-Canada Power System Outage Task Force and Its Three Working Groups . . . 175
B. Description of Outage Investigation and Process for Development of Recommendations . . . . . . . . 179
C. List of Commenters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
D. NERC Actions to Prevent and Mitigate the Impacts of Future Cascading Blackouts . . . . . . . . . . . . . 193
E. List of Electricity Acronyms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
F. Electricity Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
G. Transmittal Letters from the Three Working Groups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Tables
4.1. Generators Not Available on August 14. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.2. Benchmarking Model Results to Actual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.3. Comparison of Voltage Criteria (Percent) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
7.1. Changing Conditions That Affect System Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
8.1. U.S. Nuclear Plant Trip Times . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
8.2. Summary of Events for U. S. Nuclear Power Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
8.3. Summary of Shutdown Events for Canadian Nuclear Power Plants . . . . . . . . . . . . . . . . . . . . . . . . 128
Figures
2.1. Basic Structure of the Electric System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2. North American Interconnections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3. PJM Load Curve, August 18-24, 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4. Normal and Abnormal Frequency Ranges. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.5. NERC Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.6. NERC Regions and Control Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.7. NERC Reliability Coordinators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.8. Reliability Coordinators and Control Areas in Ohio and Surrounding States. . . . . . . . . . . . . . . . . 13
4.1. August 2003 Temperatures in the U.S. Northeast and Eastern Canada. . . . . . . . . . . . . . . . . . . . . . 25
4.2. Load Forecasts Below Actuals, August 11 through 14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.3. MW and MVAr Output from Eastlake Unit 5 on August 14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.4. Generation, Demand, and Interregional Power Flows on August 14, 2003, at 15:05 EDT. . . . . . . . 29
iv G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Figures (Continued) Page
4.5. Scheduled Imports and Exports for the Northeast Central Region, June 1 through
August 13, 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.6. Impacts of Transactions Flows on Critical Line Loadings, August 14, 2003 . . . . . . . . . . . . . . . . . . . 30
4.7. Actual Voltages Across the Ohio Area Before and On August 14, 2003. . . . . . . . . . . . . . . . . . . . . . . 34
4.8. Reactive Reserves Around Ohio on August 14, 2003, for Representative Generators in the Area. . . 37
4.9. Loss of the Perry Unit Hurts Critical Voltages and Reactive Reserves: V-Q Analyses . . . . . . . . . . . . 42
4.10. Impact of Perry Unit Outage on Cleveland-Akron Area Voltage Stability . . . . . . . . . . . . . . . . . . . . . 43
4.11. Frequency on August 14, 2003, up to 16:09 EDT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.12. Hourly Deviations in Eastern Interconnection Frequency for the Month of August 2003 . . . . . . . . 44
5.1. Timeline: Start of the Blackout in Ohio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.2. Timeline Phase 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5.3. Eastlake Unit 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
5.4. Timeline Phase 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.5. FirstEnergy 345-kV Line Flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.6. Voltages on FirstEnergy’s 345-kV Lines: Impacts of Line Trips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.7. Timeline Phase 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.8. Harding-Chamberlin 345-kV Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.9. Hanna-Juniper 345-kV Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.10. Cause of the Hanna-Juniper Line Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.11. Star-South Canton 345-kV Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.12. Cumulative Effects of Sequential Outages on Remaining 345-kV Lines . . . . . . . . . . . . . . . . . . . . . . 64
5.13. Timeline Phase 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.14. Voltages on FirstEnergy’s 138-kV Lines: Impact of Line Trips. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.15. Simulated Effect of Prior Outages on 138-kV Line Loadings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.1. Rate of Line and Generator Trips During the Cascade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
6.2. Area Affected by the Blackout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6.3. Sammis-Star 345-kV Line Trip, 16:05:57 EDT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
6.4. Sammis-Star 345-kV Line Trip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.5. Line Flows Into Michigan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.6. Ohio 345-kV Lines Trip, 16:08:59 to 16:09:07 EDT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.7. New York-Ontario Line Flows at Niagara . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.8. First Power Swing Has Varying Impacts Across the Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6.9. Map of Zone 3 (and Zone 2s Operating Like Zone 3s) Relay Operations on August 14, 2003 . . . . . . 81
6.10. Michigan and Ohio Power Plants Trip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.11. Transmission and Generation Trips in Michigan, 16:10:36 to 16:10:37 EDT . . . . . . . . . . . . . . . . . . 83
6.12. Flows on Keith-Waterman 230-kV Ontario-Michigan Tie Line. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6.13. Simulated 345-kV Line Loadings from 16:05:57 through 16:10:38.4 EDT . . . . . . . . . . . . . . . . . . . . 84
6.14. Simulated Regional Interface Loadings from 16:05:57 through 16:10:38.4 EDT. . . . . . . . . . . . . . . . 84
6.15. Michigan Lines Trip and Ohio Separates from Pennsylvania, 16:10:36 to 16:10:38.6 EDT . . . . . . . 85
6.16. Active and Reactive Power and Voltage from Ontario into Detroit . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.17. Measured Power Flows and Frequency Across Regional Interfaces, 16:10:30 to 16:11:00 EDT,
with Key Events in the Cascade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.18. Cleveland and Toledo Islanded, 16:10:39 to 16:10:46 EDT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
6.19. Generators Under Stress in Detroit, as Seen from Keith PSDR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
6.20. Western Pennsylvania Separates from New York, 16:10:39 EDT to 16:10:44 EDT . . . . . . . . . . . . . . 89
6.21. Northeast Separates from Eastern Interconnection, 16:10:45 EDT . . . . . . . . . . . . . . . . . . . . . . . . . . 89
6.22. PJM to New York Interties Disconnect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6.23. New York and New England Separate, Multiple Islands Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6.24. Generator Trips by Time and Cause. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
6.25. Events at One Large Generator During the Cascade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G v
Figures (Continued) Page
6.26. Measured Power Flows and Frequency Across Regional Interfaces, 16:10:45 to 16:11:30 EDT,
with Key Events in the Cascade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
6.27. Frequency Separation Between Ontario and Western New York . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.28. Electric Islands Reflected in Frequency Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
6.29. Area Affected by the Blackout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
6.30. Cascade Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
7.1. North American Power System Outages, 1984-1997. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
vi G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
1. Introduction
On August 14, 2003, large portions of the Midwest The Task Force divided its work into two phases:
and Northeast United States and Ontario, Canada,
u Phase I: Investigate the outage to determine its
experienced an electric power blackout. The out-
causes and why it was not contained.
age affected an area with an estimated 50 million
people and 61,800 megawatts (MW) of electric u Phase II: Develop recommendations to reduce
load in the states of Ohio, Michigan, Pennsylva- the possibility of future outages and reduce the
nia, New York, Vermont, Massachusetts, Connect- scope of any that occur.
icut, New Jersey and the Canadian province of
Ontario. The blackout began a few minutes after The Task Force created three Working Groups to
4:00 pm Eastern Daylight Time (16:00 EDT), and assist in both phases of its work—an Electric Sys-
power was not restored for 4 days in some parts of tem Working Group (ESWG), a Nuclear Working
the United States. Parts of Ontario suffered rolling Group (NWG), and a Security Working Group
blackouts for more than a week before full power (SWG). The Working Groups were made up of
was restored. Estimates of total costs in the United state and provincial representatives, federal
States range between $4 billion and $10 billion employees, and contractors working for the U.S.
(U.S. dollars).1 In Canada, gross domestic product and Canadian government agencies represented
was down 0.7% in August, there was a net loss of on the Task Force.
18.9 million work hours, and manufacturing ship- The Task Force published an Interim Report on
ments in Ontario were down $2.3 billion (Cana- November 19, 2003, summarizing the facts that
dian dollars).2 the bi-national investigation found regarding the
On August 15, President George W. Bush and causes of the blackout on August 14, 2003. After
then-Prime Minister Jean Chrétien directed that a November 19, the Task Force’s technical investi-
joint U.S.-Canada Power System Outage Task gation teams pursued certain analyses that were
Force be established to investigate the causes of not complete in time for publication in the Interim
the blackout and ways to reduce the possibility of Report. The Working Groups focused on the draft-
future outages. They named U.S. Secretary of ing of recommendations for the consideration of
Energy Spencer Abraham and Herb Dhaliwal, the Task Force to prevent future blackouts and
Minister of Natural Resources, Canada, to chair reduce the scope of any that nonetheless occur. In
the joint Task Force. (Mr. Dhaliwal was later suc- drafting these recommendations, the Working
ceeded by Mr. John Efford as Minister of Natural Groups drew substantially on information and
Resources and as co-chair of the Task Force.) insights from the investigation teams’ additional
Three other U.S. representatives and three other analyses, and on inputs received at three public
Canadian representatives were named to the meetings (in Cleveland, New York City, and
Task Force. The U.S. members were Tom Ridge, Toronto) and two technical conferences (in Phila-
Secretary of Homeland Security; Pat Wood III, delphia and Toronto). They also drew on com-
Chairman of the Federal Energy Regulatory Com- ments filed electronically by interested parties on
mission; and Nils Diaz, Chairman of the Nuclear websites established for this purpose by the
Regulatory Commission. The Canadian members U.S. Department of Energy and Natural Resources
were Deputy Prime Minister John Manley, later Canada.
succeeded by Deputy Prime Minister Anne Although this Final Report presents some new
McLellan; Kenneth Vollman, Chairman of the information about the events and circumstances
National Energy Board; and Linda J. Keen, Presi- before the start of the blackout and additional
dent and CEO of the Canadian Nuclear Safety detail concerning the cascade stage of the black-
Commission. out, none of the comments received or additional
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 1
analyses performed by the Task Force’s investiga- u Chapter 10 presents the Task Force’s recom-
tors have changed the validity of the conclusions mendations for preventing future blackouts and
published in the Interim Report. This report, reducing the scope of any that occur.
however, presents findings concerning additional
Chapter 10 includes a total of 46 recommenda-
violations of reliability requirements and institu-
tions, but the single most important of them is that
tional and performance deficiencies beyond those
the U.S. Congress should enact the reliability pro-
identified in the Interim Report.
visions in H.R. 6 and S. 2095 to make compliance
The organization of this Final Report is similar to with reliability standards mandatory and enforce-
that of the Interim Report, and it is intended to able. If that could be done, many of the other rec-
update and supersede the Interim Report. It is ommended actions could be accomplished readily
divided into ten chapters, including this introduc- in the course of implementing the legislation. An
tory chapter: overview of the recommendations (by titles only)
is provided on pages 3 and 4.
u Chapter 2 provides an overview of the institu-
tional framework for maintaining and ensuring Chapter 2 is very little changed from the version
the reliability of the bulk power system in North published in the Interim Report. Chapter 3 is new
America, with particular attention to the roles to this Final Report. Chapters 4, 5, and 6 have been
and responsibilities of several types of reliabil- revised and expanded from the corresponding
ity-related organizations. chapters (3, 4, and 5) of the Interim Report. Chap-
u Chapter 3 identifies the causes of the blackout ters 7, 8, and 9 are only slightly changed from
and identifies failures to perform effectively Chapters 6, 7, and 8 of the Interim Report. The
relative to the reliability policies, guidelines, Interim Report had no counterpart to Chapter 10.
and standards of the North American Electric This report also includes seven appendixes:
Reliability Council (NERC) and, in some cases,
deficiencies in the standards themselves. u Appendix A lists the members of the Task Force
and the three working groups.
u Chapter 4 discusses conditions on the regional
power system on and before August 14 and u Appendix B describes the Task Force’s investi-
identifies conditions and failures that did and gative process for developing the Task Force’s
did not contribute to the blackout. recommendations.
u Chapter 5 describes the afternoon of August 14, u Appendix C lists the parties who either com-
starting from normal operating conditions, then mented on the Interim Report, provided sugges-
going into a period of abnormal but still poten- tions for recommendations, or both.
tially manageable conditions, and finally into
u Appendix D reproduces a document released on
an uncontrollable blackout in northern Ohio.
February 10, 2004 by NERC, describing its
u Chapter 6 provides details on the cascade phase actions to prevent and mitigate the impacts of
of the blackout as it spread in Ohio and then future cascading blackouts.
across the Northeast, and explains why the sys-
tem performed as it did. u Appendix E is a list of electricity acronyms.
u Chapter 7 compares the August 14, 2003, black- u Appendix F provides a glossary of electricity
out with previous major North American power terms.
outages. u Appendix G contains transmittal letters perti-
u Chapter 8 examines the performance of the nent to this report from the three Working
nuclear power plants affected by the August 14 Groups.
outage.
u Chapter 9 addresses issues related to physical
and cyber security associated with the outage.
2 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Overview of Task Force Recommendations: Titles Only
Group I. Institutional Issues Related to Reliability
1. Make reliability standards mandatory and enforceable, with penalties for noncompliance.
2. Develop a regulator-approved funding mechanism for NERC and the regional reliability councils,
to ensure their independence from the parties they oversee.
3. Strengthen the institutional framework for reliability management in North America.
4. Clarify that prudent expenditures and investments for bulk system reliability (including invest-
ments in new technologies) will be recoverable through transmission rates.
5. Track implementation of recommended actions to improve reliability.
6. FERC should not approve the operation of new RTOs or ISOs until they have met minimum
functional requirements.
7. Require any entity operating as part of the bulk power system to be a member of a regional reli-
ability council if it operates within the council’s footprint.
8. Shield operators who initiate load shedding pursuant to approved guidelines from liability or
retaliation.
9. Integrate a “reliability impact” consideration into the regulatory decision-making process.
10. Establish an independent source of reliability performance information.
11. Establish requirements for collection and reporting of data needed for post-blackout analyses.
12. Commission an independent study of the relationships among industry restructuring, competi-
tion, and reliability.
13. DOE should expand its research programs on reliability-related tools and technologies.
14. Establish a standing framework for the conduct of future blackout and disturbance
investigations.
Group II. Support and Strengthen NERC’s Actions of February 10, 2004
15. Correct the direct causes of the August 14, 2003 blackout.
16. Establish enforceable standards for maintenance of electrical clearances in right-of-way areas.
17. Strengthen the NERC Compliance Enforcement Program.
18. Support and strengthen NERC’s Reliability Readiness Audit Program.
19. Improve near-term and long-term training and certification requirements for operators, reliability
coordinators, and operator support staff.
20. Establish clear definitions for normal, alert and emergency operational system conditions. Clarify
roles, responsibilities, and authorities of reliability coordinators and control areas under each
condition.
21. Make more effective and wider use of system protection measures.
22. Evaluate and adopt better real-time tools for operators and reliability coordinators.
23. Strengthen reactive power and voltage control practices in all NERC regions.
24. Improve quality of system modeling data and data exchange practices.
25. NERC should reevaluate its existing reliability standards development process and accelerate the
adoption of enforceable standards.
26. Tighten communications protocols, especially for communications during alerts and emergen-
cies. Upgrade communication system hardware where appropriate.
27. Develop enforceable standards for transmission line ratings.
28. Require use of time-synchronized data recorders.
29. Evaluate and disseminate lessons learned during system restoration.
30. Clarify criteria for identification of operationally critical facilities, and improve dissemination of
updated information on unplanned outages.
31. Clarify that the transmission loading relief (TLR) process should not be used in situations involv-
ing an actual violation of an Operating Security Limit. Streamline the TLR process.
(continued on page 142)
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 3
Overview of Task Force Recommendations: Titles Only (Continued)
Group III. Physical and Cyber Security of North American Bulk Power Systems
32. Implement NERC IT standards.
33. Develop and deploy IT management procedures.
34. Develop corporate-level IT security governance and strategies.
35. Implement controls to manage system health, network monitoring, and incident management.
36. Initiate U.S.-Canada risk management study.
37. Improve IT forensic and diagnostic capabilities.
38. Assess IT risk and vulnerability at scheduled intervals.
39. Develop capability to detect wireless and remote wireline intrusion and surveillance.
40. Control access to operationally sensitive equipment.
41. NERC should provide guidance on employee background checks.
42. Confirm NERC ES-ISAC as the central point for sharing security information and analysis.
43. Establish clear authority for physical and cyber security.
44. Develop procedures to prevent or mitigate inappropriate disclosure of information.
Group IV. Canadian Nuclear Power Sector
45. The Task Force recommends that the Canadian Nuclear Safety Commission request Ontario
Power Generation and Bruce Power to review operating procedures and operator training associ-
ated with the use of adjuster rods.
46. The Task Force recommends that the Canadian Nuclear Safety Commission purchase and install
backup generation equipment.
4 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
2. Overview of the North American Electric Power
System and Its Reliability Organizations
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 5
transmission lines. Operating the transmission independent from each other except for a few
lines at high voltage (i.e., 230,000 to 765,000 volts) small direct current (DC) ties that link them.
reduces the losses of electricity from conductor Within each interconnection, electricity is pro-
heating and allows power to be shipped economi- duced the instant it is used, and flows over virtu-
cally over long distances. Transmission lines are ally all transmission lines from generators to
interconnected at switching stations and substa- loads.
tions to form a network of lines and stations called
a power “grid.” Electricity flows through the inter- The northeastern portion of the Eastern Intercon-
connected network of transmission lines from the nection (about 10 percent of the interconnection’s
generators to the loads in accordance with the total load) was affected by the August 14 blackout.
laws of physics—along “paths of least resistance,” The other two interconnections were not
in much the same way that water flows through a affected.1
network of canals. When the power arrives near a
load center, it is “stepped down” to lower voltages Planning and Reliable Operation
for distribution to customers. The bulk power sys-
tem is predominantly an alternating current (AC)
of the Power Grid Are Technically
system, as opposed to a direct current (DC) sys- Demanding
tem, because of the ease and low cost with which
voltages in AC systems can be converted from one Reliable operation of the power grid is complex
level to another. Some larger industrial and com- and demanding for two fundamental reasons:
mercial customers take service at intermediate u First, electricity flows at close to the speed of
voltage levels (12,000 to 115,000 volts), but most light (186,000 miles per second or 297,600
residential customers take their electrical service km/sec) and is not economically storable in
at 120 and 240 volts. large quantities. Therefore electricity must be
While the power system in North America is com- produced the instant it is used.
monly referred to as “the grid,” there are actually u Second, without the use of control devices too
three distinct power grids or “interconnections” expensive for general use, the flow of alternat-
(Figure 2.2). The Eastern Interconnection includes ing current (AC) electricity cannot be controlled
the eastern two-thirds of the continental United like a liquid or gas by opening or closing a valve
States and Canada from Saskatchewan east to the in a pipe, or switched like calls over a long-
Maritime Provinces. The Western Interconnection distance telephone network.2 Electricity flows
includes the western third of the continental freely along all available paths from the genera-
United States (excluding Alaska), the Canadian tors to the loads in accordance with the laws of
provinces of Alberta and British Columbia, and a physics—dividing among all connected flow
portion of Baja California Norte, Mexico. The third paths in the network, in inverse proportion to
interconnection comprises most of the state of the impedance (resistance plus reactance) on
Texas. The three interconnections are electrically each path.
Maintaining reliability is a complex enterprise
Figure 2.2. North American Interconnections that requires trained and skilled operators, sophis-
ticated computers and communications, and care-
ful planning and design. The North American
Electric Reliability Council (NERC) and its ten
Regional Reliability Councils have developed sys-
tem operating and planning standards for ensur-
ing the reliability of a transmission grid that are
based on seven key concepts:
u Balance power generation and demand
continuously.
u Balance reactive power supply and demand to
maintain scheduled voltages.
u Monitor flows over transmission lines and other
facilities to ensure that thermal (heating) limits
are not exceeded.
6 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
u Keep the system in a stable condition. automatic under-frequency “load shedding,”
which takes blocks of customers off-line in
u Operate the system so that it remains in a reli-
order to prevent a total collapse of the electric
able condition even if a contingency occurs,
system. As will be seen later in this report, such
such as the loss of a key generator or transmis-
an imbalance of generation and demand can
sion facility (the “N-1 criterion”).
also occur when the system responds to major
u Plan, design, and maintain the system to oper- disturbances by breaking into separate
ate reliably. “islands”; any such island may have an excess
or a shortage of generation, compared to
u Prepare for emergencies. demand within the island.
These seven concepts are explained in more detail 2. Balance reactive power supply and demand to
below. maintain scheduled voltages. Reactive power
1. Balance power generation and demand contin- sources, such as capacitor banks and genera-
uously. To enable customers to use as much tors, must be adjusted during the day to main-
electricity as they wish at any moment, produc- tain voltages within a secure range pertaining to
tion by the generators must be scheduled or all system electrical equipment (stations, trans-
“dispatched” to meet constantly changing mission lines, and customer equipment). Most
demands, typically on an hourly basis, and then generators have automatic voltage regulators
fine-tuned throughout the hour, sometimes that cause the reactive power output of genera-
through the use of automatic generation con- tors to increase or decrease to control voltages to
trols to continuously match generation to actual scheduled levels. Low voltage can cause electric
demand. Demand is somewhat predictable, system instability or collapse and, at distribu-
appearing as a daily demand curve—in the tion voltages, can cause damage to motors and
summer, highest during the afternoon and eve- the failure of electronic equipment. High volt-
ning and lowest in the middle of the night, and ages can exceed the insulation capabilities of
higher on weekdays when most businesses are equipment and cause dangerous electric arcs
open (Figure 2.3). (“flashovers”).
Failure to match generation to demand causes 3. Monitor flows over transmission lines and
the frequency of an AC power system (nomi- other facilities to ensure that thermal (heating)
nally 60 cycles per second or 60 Hertz) to limits are not exceeded. The dynamic interac-
increase (when generation exceeds demand) or tions between generators and loads, combined
decrease (when generation is less than demand) with the fact that electricity flows freely across
(Figure 2.4). Random, small variations in fre- all interconnected circuits, mean that power
quency are normal, as loads come on and off flow is ever-changing on transmission and dis-
and generators modify their output to follow the tribution lines. All lines, transformers, and
demand changes. However, large deviations in other equipment carrying electricity are heated
frequency can cause the rotational speed of gen- by the flow of electricity through them. The
erators to fluctuate, leading to vibrations that
can damage generator turbine blades and other
Figure 2.4. Normal and Abnormal Frequency
equipment. Extreme low frequencies can trigger
Ranges
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 7
Local Supplies of Reactive Power Are Essential to Maintaining Voltage Stability
A generator typically produces some mixture of lines both consume and produce reactive power.
“real” and “reactive” power, and the balance At light loads they are net producers, and at
between them can be adjusted at short notice to heavy loads, they are heavy consumers. Reactive
meet changing conditions. Real power, measured power consumption by these facilities or devices
in watts, is the form of electricity that powers tends to depress transmission voltage, while its
equipment. Reactive power, a characteristic of production (by generators) or injection (from
AC systems, is measured in volt-amperes reac- storage devices such as capacitors) tends to sup-
tive (VAr), and is the energy supplied to create or port voltage. Reactive power can be transmitted
be stored in electric or magnetic fields in and only over relatively short distances during heavy
around electrical equipment. Reactive power is load conditions. If reactive power cannot be sup-
particularly important for equipment that relies plied promptly and in sufficient quantity, volt-
on magnetic fields for the production of induced ages decay, and in extreme cases a “voltage
electric currents (e.g., motors, transformers, collapse” may result.
pumps, and air conditioning.) Transmission
flow must be limited to avoid overheating and limits, like thermal limits, are expressed as a
damaging the equipment. In the case of over- maximum amount of electricity that can be
head power lines, heating also causes the metal safely transferred over transmission lines.
conductor to stretch or expand and sag closer to
There are two types of stability limits: (1) Volt-
ground level. Conductor heating is also affected
age stability limits are set to ensure that the
by ambient temperature, wind, and other fac-
unplanned loss of a line or generator (which
tors. Flow on overhead lines must be limited to
may have been providing locally critical reac-
ensure that the line does not sag into obstruc-
tive power support, as described previously)
tions below such as trees or telephone lines, or
will not cause voltages to fall to dangerously
violate the minimum safety clearances between
low levels. If voltage falls too low, it begins to
the energized lines and other objects. (A short
collapse uncontrollably, at which point auto-
circuit or “flashover”—which can start fires or
matic relays either shed load or trip generators
damage equipment—can occur if an energized
to avoid damage. (2) Power (angle) stability lim-
line gets too close to another object). Most trans-
its are set to ensure that a short circuit or an
mission lines, transformers and other current-
unplanned loss of a line, transformer, or genera-
carrying devices are monitored continuously to
tor will not cause the remaining generators and
ensure that they do not become overloaded or
loads being served to lose synchronism with
violate other operating constraints. Multiple
one another. (Recall that all generators and
ratings are typically used, one for normal condi-
loads within an interconnection must operate at
tions and a higher rating for emergencies. The
or very near a common 60 Hz frequency.) Loss
primary means of limiting the flow of power on
of synchronism with the common frequency
transmission lines is to adjust selectively the
means generators are operating out-of-step with
output of generators.
one another. Even modest losses of synchro-
4. Keep the system in a stable condition. Because nism can result in damage to generation equip-
the electric system is interconnected and ment. Under extreme losses of synchronism,
dynamic, electrical stability limits must be the grid may break apart into separate electrical
observed. Stability problems can develop very islands; each island would begin to maintain its
quickly—in just a few cycles (a cycle is 1/60th of own frequency, determined by the load/genera-
a second)—or more slowly, over seconds or tion balance within the island.
minutes. The main concern is to ensure that 5. Operate the system so that it remains in a reli-
generation dispatch and the resulting power able condition even if a contingency occurs,
flows and voltages are such that the system is such as the loss of a key generator or transmis-
stable at all times. (As will be described later in sion facility (the “N minus 1 criterion”). The
this report, part of the Eastern Interconnection central organizing principle of electricity reli-
became unstable on August 14, resulting in a ability management is to plan for the unex-
cascading outage over a wide area.) Stability pected. The unique characteristics of electricity
8 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
mean that problems, when they arise, can does not jeopardize the remaining facilities in
spread and escalate very quickly if proper safe- the system by causing them to exceed their
guards are not in place. Accordingly, through emergency ratings or stability limits, which
years of experience, the industry has developed could lead to a cascading outage.
a network of defensive strategies for maintain-
ing reliability based on the assumption that Further, when a contingency does occur, the
equipment can and will fail unexpectedly upon operators are required to identify and assess
occasion. immediately the new worst contingencies,
given the changed conditions, and promptly
This principle is expressed by the requirement make any adjustments needed to ensure that if
that the system must be operated at all times to one of them were to occur, the system would
ensure that it will remain in a secure condition still remain operational and safe. NERC operat-
(generally within emergency ratings for current ing policy requires that the system be restored
and voltage and within established stability as soon as practical but within no more than 30
limits) following the loss of the most important minutes to compliance with normal limits, and
generator or transmission facility (a “worst sin- to a condition where it can once again with-
gle contingency”). This is called the “N-1 crite- stand the next-worst single contingency with-
rion.” In other words, because a generator or out violating thermal, voltage, or stability
line trip can occur at any time from random fail- limits. A few areas of the grid are operated to
ure, the power system must be operated in a withstand the concurrent loss of two or more
preventive mode so that the loss of the most facilities (i.e., “N-2”). This may be done, for
important generator or transmission facility example, as an added safety measure to protect
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 9
a densely populated metropolitan area or when system as quickly as possible to a normal and
lines share a common structure and could be reliable condition.
affected by a common failure mode, e.g., a sin-
gle lightning strike. Reliability Organizations Oversee
6. Plan, design, and maintain the system to oper- Grid Reliability in North America
ate reliably. Reliable power system operation
requires far more than monitoring and control- NERC is a non-governmental entity whose mis-
ling the system in real-time. Thorough plan- sion is to ensure that the bulk electric system in
ning, design, maintenance, and analysis are North America is reliable, adequate and secure.
required to ensure that the system can be oper- The organization was established in 1968, as a
ated reliably and within safe limits. Short-term result of the Northeast blackout in 1965. Since its
planning addresses day-ahead and week-ahead inception, NERC has operated as a voluntary orga-
operations planning; long-term planning nization, relying on reciprocity, peer pressure and
focuses on providing adequate generation the mutual self-interest of all those involved to
resources and transmission capacity to ensure ensure compliance with reliability requirements.
that in the future the system will be able to An independent board governs NERC.
withstand severe contingencies without experi-
To fulfill its mission, NERC:
encing widespread, uncontrolled cascading
outages. u Sets standards for the reliable operation and
planning of the bulk electric system.
A utility that serves retail customers must esti-
mate future loads and, in some cases, arrange u Monitors and assesses compliance with stan-
for adequate sources of supplies and plan ade- dards for bulk electric system reliability.
quate transmission and distribution infrastruc-
ture. NERC planning standards identify a range u Provides education and training resources to
of possible contingencies and set corresponding promote bulk electric system reliability.
expectations for system performance under sev- u Assesses, analyzes and reports on bulk electric
eral categories of possible events, ranging from system adequacy and performance.
everyday “probable” events to “extreme” events
that may involve substantial loss of customer u Coordinates with regional reliability councils
load and generation in a widespread area. NERC and other organizations.
planning standards also address requirements u Coordinates the provision of applications
for voltage support and reactive power, distur- (tools), data and services necessary to support
bance monitoring, facility ratings, system mod- the reliable operation and planning of the bulk
eling and data requirements, system protection electric system.
and control, and system restoration.
u Certifies reliability service organizations and
7. Prepare for emergencies. System operators are personnel.
required to take the steps described above to
plan and operate a reliable power system, but u Coordinates critical infrastructure protection of
emergencies can still occur because of external the bulk electric system.
factors such as severe weather, operator error,
u Enables the reliable operation of the intercon-
or equipment failures that exceed planning,
nected bulk electric system by facilitating infor-
design, or operating criteria. For these rare
mation exchange and coordination among
events, the operating entity is required to have
reliability service organizations.
emergency procedures covering a credible
range of emergency scenarios. Operators must Recent changes in the electricity industry have
be trained to recognize and take effective action altered many of the traditional mechanisms,
in response to these emergencies. To deal with a incentives and responsibilities of the entities
system emergency that results in a blackout, involved in ensuring reliability, to the point that
such as the one that occurred on August 14, the voluntary system of compliance with reliabil-
2003, there must be procedures and capabilities ity standards is generally recognized as not ade-
to use “black start” generators (capable of quate to current needs.3 NERC and many other
restarting with no external power source) and to electricity organizations support the development
coordinate operations in order to restore the of a new mandatory system of reliability standards
10 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
and compliance, backstopped in the United States that owned and operated generation, transmis-
by the Federal Energy Regulatory Commission. sion, and distribution. While that is still true in
This will require federal legislation in the United some areas, there has been significant restructur-
States to provide for the creation of a new electric ing of operating functions and some consolidation
reliability organization with the statutory author- of control areas into regional operating entities.
ity to enforce compliance with reliability stan- Utility industry restructuring has led to an
dards among all market participants. Appropriate unbundling of generation, transmission and dis-
government entities in Canada and Mexico are tribution activities such that the ownership and
prepared to take similar action, and some have operation of these assets have been separated
already done so. In the meantime, NERC encour- either functionally or through the formation of
ages compliance with its reliability standards independent entities called Independent System
through an agreement with its members. Operators (ISOs) and Regional Transmission
Organizations (RTOs).
NERC’s members are ten regional reliability
councils. (See Figure 2.5 for a map showing the u ISOs and RTOs in the United States have been
locations and boundaries of the regional councils.) authorized by FERC to implement aspects of the
In turn, the regional councils have broadened Energy Policy Act of 1992 and subsequent FERC
their membership to include all segments of the policy directives.
electric industry: investor-owned utilities; federal u The primary functions of ISOs and RTOs are to
power agencies; rural electric cooperatives; state, manage in real time and on a day-ahead basis
municipal and provincial utilities; independent the reliability of the bulk power system and the
power producers; power marketers; and end-use operation of wholesale electricity markets
customers. Collectively, the members of the NERC within their footprint.
regions account for virtually all the electricity sup-
u ISOs and RTOs do not own transmission assets;
plied in the United States, Canada, and a portion
they operate or direct the operation of assets
of Baja California Norte, Mexico. The ten regional
owned by their members.
councils jointly fund NERC and adapt NERC
standards to meet the needs of their regions. The u ISOs and RTOs may be control areas them-
August 14 blackout affected three NERC regional selves, or they may encompass more than one
reliability councils—East Central Area Reliability control area.
Coordination Agreement (ECAR), Mid-Atlantic u ISOs and RTOs may also be NERC Reliability
Area Council (MAAC), and Northeast Power Coor- Coordinators, as described below.
dinating Council (NPCC).
Five RTOs/ISOs are within the area directly
“Control areas” are the primary operational enti- affected by the August 14 blackout. They are:
ties that are subject to NERC and regional council
standards for reliability. A control area is a geo- u Midwest Independent System Operator (MISO)
graphic area within which a single entity, Inde- u PJM Interconnection (PJM)
pendent System Operator (ISO), or Regional
Transmission Organization (RTO) balances gener- Figure 2.5. NERC Regions
ation and loads in real time to maintain reliable
operation. Control areas are linked with each
other through transmission interconnection tie
lines. Control area operators control generation
directly to maintain their electricity interchange
schedules with other control areas. They also
operate collectively to support the reliability of
their interconnection. As shown in Figure 2.6,
there are approximately 140 control areas in North
America. The control area dispatch centers have
sophisticated monitoring and control systems and
are staffed 24 hours per day, 365 days per year.
Traditionally, control areas were defined by utility
service area boundaries and operations were
largely managed by vertically integrated utilities
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 11
u New York Independent System Operator Electric Power (AEP)—and their respective reli-
(NYISO) ability coordinators, MISO and PJM (see Figures
u New England Independent System Operator
2.7 and 2.8). These organizations and their reli-
(ISO-NE) ability responsibilities are described briefly in this
final subsection.
u Ontario Independent Market Operator (IMO)
1. FirstEnergy operates a control area in north-
Reliability coordinators provide reliability over- ern Ohio. FirstEnergy (FE) consists of seven
sight over a wide region. They prepare reliability electric utility operating companies. Four of
assessments, provide a wide-area view of reliabil- these companies, Ohio Edison, Toledo Edison,
ity, and coordinate emergency operations in real The Illuminating Company, and Penn Power,
time for one or more control areas. They may oper- operate in the NERC ECAR region, with MISO
ate, but do not participate in, wholesale or retail serving as their reliability coordinator. These
market functions. There are currently 18 reliabil- four companies now operate as one integrated
ity coordinators in North America. Figure 2.7 control area managed by FE.4
shows the locations and boundaries of their
respective areas. 2. American Electric Power (AEP) operates a con-
trol area in Ohio just south of FE. AEP is both a
transmission operator and a control area
Key Parties in the Pre-Cascade operator.
Phase of the August 14 Blackout 3. Midwest Independent System Operator
The initiating events of the blackout involved two (MISO) is the reliability coordinator for
control areas—FirstEnergy (FE) and American FirstEnergy. The Midwest Independent System
12 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Operator (MISO) is the reliability coordinator Reliability Responsibilities of Control
for a region of more than 1 million square miles Area Operators and Reliability
(2.6 million square kilometers), stretching from
Coordinators
Manitoba, Canada in the north to Kentucky in
the south, from Montana in the west to western 1. Control area operators have primary responsi-
Pennsylvania in the east. Reliability coordina- bility for reliability. Their most important
tion is provided by two offices, one in Minne- responsibilities, in the context of this report,
sota, and the other at the MISO headquarters in are:
Indiana. Overall, MISO provides reliability
coordination for 37 control areas, most of which N-1 criterion. NERC Operating Policy 2.A—
are members of MISO. Transmission Operations:
4. PJM is AEP’s reliability coordinator. PJM is one “All CONTROL AREAS shall operate so that
of the original ISOs formed after FERC orders instability, uncontrolled separation, or cas-
888 and 889, but was established as a regional cading outages will not occur as a result of
power pool in 1935. PJM recently expanded its the most severe single contingency.”
footprint to include control areas and transmis- Emergency preparedness and emergency
sion operators within MAIN and ECAR (PJM- response. NERC Operating Policy 5—Emer-
West). It performs its duties as a reliability coor- gency Operations, General Criteria:
dinator in different ways, depending on the
control areas involved. For PJM-East, it is “Each system and CONTROL AREA shall
both the control area and reliability coordinator promptly take appropriate action to relieve
for ten utilities, whose transmission systems any abnormal conditions, which jeopardize
span the Mid-Atlantic region of New Jersey, reliable Interconnection operation.”
most of Pennsylvania, Delaware, Maryland, “Each system, CONTROL AREA, and Region
West Virginia, Ohio, Virginia, and the District of shall establish a program of manual and auto-
Columbia. The PJM-West facility has the reli- matic load shedding which is designed to
ability coordinator desk for five control areas arrest frequency or voltage decays that could
(AEP, Commonwealth Edison, Duquesne Light, result in an uncontrolled failure of compo-
Dayton Power and Light, and Ohio Valley Elec- nents of the interconnection.”
tric Cooperative) and three generation-only
control areas (Duke Energy’s Washington NERC Operating Policy 5.A—Coordination
County (Ohio) facility, Duke’s Lawrence with Other Systems:
County/Hanging Rock (Ohio) facility, and Alle- “A system, CONTROL AREA, or pool that is
gheny Energy’s Buchanan (West Virginia) experiencing or anticipating an operating
facility. emergency shall communicate its current
and future status to neighboring systems,
CONTROL AREAS, or pools and throughout the
interconnection . . . . A system shall inform
Figure 2.7. NERC Reliability Coordinators
Figure 2.8. Reliability Coordinators and Control
Areas in Ohio and Surrounding States
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 13
other systems . . . whenever . . . the system’s Operating personnel and training: NERC Oper-
condition is burdening other systems or ating Policy 8.B—Training:
reducing the reliability of the Interconnec-
tion . . . [or whenever] the system’s line load- “Each OPERATING AUTHORITY should period-
ically practice simulated emergencies. The
ings and voltage/reactive levels are such that
scenarios included in practice situations
a single contingency could threaten the reli-
should represent a variety of operating condi-
ability of the Interconnection.”
tions and emergencies.”
NERC Operating Policy 5.C—Transmission
2. Reliability Coordinators such as MISO and
System Relief:
PJM are expected to comply with all aspects of
“Action to correct an OPERATING SECURITY NERC Operating Policies, especially Policy 9,
LIMIT violation shall not impose unaccept- Reliability Coordinator Procedures, and its
able stress on internal generation or transmis- appendices. Key requirements include:
sion equipment, reduce system reliability
beyond acceptable limits, or unduly impose NERC Operating Policy 9, Criteria for Reliabil-
ity Coordinators, 5.2:
voltage or reactive burdens on neighboring
systems. If all other means fail, corrective Have “detailed monitoring capability of the
action may require load reduction.” RELIABILITY AREA and sufficient monitoring
14 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
capability of the surrounding RELIABILITY NERC Operating Policy 9, Functions of Reliabil-
AREAS to ensure potential security violations ity Coordinators, 6:
are identified.”
“Conduct security assessment and monitor-
NERC Operating Policy 9, Functions of Reliabil- ing programs to assess contingency situa-
ity Coordinators, 1.7: tions. Assessments shall be made in real time
“Monitor the parameters that may have sig- and for the operations planning horizon at
nificant impacts within the RELIABILITY AREA the CONTROL AREA level with any identified
and with neighboring RELIABILITY AREAS problems reported to the RELIABILITY CO-
ORDINATOR. The RELIABILITY COORDINATOR
with respect to . . . sharing with other
RELIABILITY COORDINATORS any information is to ensure that CONTROL AREA, RELIABILITY
regarding potential, expected, or actual criti- AREA, and regional boundaries are suffi-
cal operating conditions that could nega- ciently modeled to capture any problems
tively impact other RELIABILITY AREAS. The crossing such boundaries.”
RELIABILITY COORDINATOR will coordinate Endnotes
with other RELIABILITY COORDINATORS and 1 The province of Québec, although considered a part of the
CONTROL AREAS as needed to develop appro- Eastern Interconnection, is connected to the rest of the East-
priate plans to mitigate negative impacts of ern Interconnection only by DC ties. In this instance, the DC
potential, expected, or actual critical operat- ties acted as buffers between portions of the Eastern Intercon-
ing conditions . . . .” nection; transient disturbances propagate through them less
readily. Therefore, the electricity system in Québec was not
affected by the outage, except for a small portion of the prov-
What Constitutes an Operating Emergency? ince’s load that is directly connected to Ontario by AC trans-
mission lines. (Although DC ties can act as a buffer between
An operating emergency is an unsustainable systems, the tradeoff is that they do not allow instantaneous
condition that cannot be resolved using the generation support following the unanticipated loss of a gen-
erating unit.)
resources normally available. The NERC Oper- 2 In some locations, bulk power flows are controlled through
ating Manual defines a “capacity emergency” as specialized devices or systems, such as phase angle regula-
when a system’s or pool’s operating generation tors, “flexible AC transmission systems” (FACTS), and high-
capacity, plus firm purchases from other sys- voltage DC converters (and reconverters) spliced into the AC
tems, to the extent available or limited by trans- system. These devices are still too expensive for general
fer capability, is inadequate to meet its demand application.
3 See, for example, Maintaining Reliability in a Competitive
plus its regulating requirements. It defines an
Electric Industry (1998), a report to the U.S. Secretary of
“energy emergency” as when a load-serving Energy by the Task Force on Electric Systems Reliability;
entity has exhausted all other options and can National Energy Policy (2001), a report to the President of the
no longer provide its customers’ expected United States by the National Energy Policy Development
energy requirements. A transmission emer- Group, p. 7-6; and National Transmission Grid Study (2002),
gency exists when “the system’s line loadings U.S. Dept. of Energy, pp. 46-48.
4 The remaining three FE companies, Penelec, Met-Ed, and
and voltage/ reactive levels are such that a single
Jersey Central Power & Light, are in the NERC MAAC region
contingency could threaten the reliability of the and have PJM as their reliability coordinator. The focus of this
Interconnection.” Control room operators and report is on the portion of FE in the ECAR reliability region
dispatchers are given substantial latitude to and within the MISO reliability coordinator footprint.
determine when to declare an emergency. (See
pages 66-67 in Chapter 5 for more detail.)
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 15
3. Causes of the Blackout
and Violations of NERC Standards
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 17
Causes of the Blackout’s Initiation
The Ohio phase of the August 14, 2003, blackout Group 2: Inadequate situational awareness
was caused by deficiencies in specific practices, at FirstEnergy. FE did not recognize or
equipment, and human decisions by various understand the deteriorating condition of
organizations that affected conditions and out- its system.
comes that afternoon—for example, insufficient As discussed in Chapter 5:
reactive power was an issue in the blackout, but
it was not a cause in itself. Rather, deficiencies in A) FE failed to ensure the security of its transmis-
corporate policies, lack of adherence to industry sion system after significant unforeseen con-
policies, and inadequate management of reactive tingencies because it did not use an effective
power and voltage caused the blackout, rather contingency analysis capability on a routine
than the lack of reactive power. There are four basis. (See pages 49-50 and 64.)
groups of causes for the blackout: B) FE lacked procedures to ensure that its opera-
tors were continually aware of the functional
Group 1: FirstEnergy and ECAR failed to state of their critical monitoring tools. (See
assess and understand the inadequacies of pages 51-53, 56.)
FE’s system, particularly with respect to C) FE control center computer support staff and
voltage instability and the vulnerability of operations staff did not have effective internal
the Cleveland-Akron area, and FE did not communications procedures. (See pages 54,
operate its system with appropriate voltage 56, and 65-67.)
criteria. (Note: This cause was not identified in
the Task Force’s Interim Report. It is based on D) FE lacked procedures to test effectively the
analysis completed by the investigative team functional state of its monitoring tools after
after the publication of the Interim Report.) repairs were made. (See page 54.)
As detailed in Chapter 4: E) FE did not have additional or back-up moni-
A) FE failed to conduct rigorous long-term plan- toring tools to understand or visualize the sta-
ning studies of its system, and neglected to tus of their transmission system to facilitate
conduct appropriate multiple contingency or its operators’ understanding of transmission
extreme condition assessments. (See pages system conditions after the failure of their pri-
37-39 and 41-43.) mary monitoring/alarming systems. (See
pages 53, 56, and 65.)
B) FE did not conduct sufficient voltage analyses
for its Ohio control area and used operational Group 3: FE failed to manage adequately tree
voltage criteria that did not reflect actual volt- growth in its transmission rights-of-way.
age stability conditions and needs. (See pages
This failure was the common cause of the outage
31-37.)
of three FE 345-kV transmission lines and one
C) ECAR (FE’s reliability council) did not con- 138-kV line. (See pages 57-64.)
duct an independent review or analysis of
FE’s voltage criteria and operating needs, Group 4: Failure of the interconnected grid’s
thereby allowing FE to use inadequate prac- reliability organizations to provide effective
tices without correction. (See page 39.) real-time diagnostic support.
D)Some of NERC’s planning and operational As discussed in Chapter 5:
requirements and standards were sufficiently A) MISO did not have real-time data from
ambiguous that FE could interpret them to Dayton Power and Light’s Stuart-Atlanta
include practices that were inadequate for reli- 345-kV line incorporated into its state estima-
able system operation. (See pages 31-33.) tor (a system monitoring tool). This precluded
(continued on page 19)
18 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Causes of the Blackout’s Initiation (Continued)
MISO from becoming aware of FE’s system D) PJM and MISO lacked joint procedures or
problems earlier and providing diagnostic guidelines on when and how to coordinate a
assistance or direction to FE. (See pages security limit violation observed by one of
49-50.) them in the other’s area due to a contingency
near their common boundary. (See pages
B) MISO’s reliability coordinators were using
62-63 and 65-66.)
non-real-time data to support real-time
“flowgate” monitoring. This prevented MISO In the chapters that follow, sections that relate to
from detecting an N-1 security violation in particular causes are denoted with the following
FE’s system and from assisting FE in neces- symbols:
sary relief actions. (See pages 48 and 63.) Cause 1 Cause 2
Inadequate Inadequate
C) MISO lacked an effective way to identify the System Situational
Understanding Awareness
location and significance of transmission line
breaker operations reported by their Energy
Management System (EMS). Such informa- Cause 3 Cause 4
Inadequate Inadequate
tion would have enabled MISO operators to
Tree RC Diagnostic
become aware earlier of important line out- Trimming Support
ages. (See page 48.)
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 19
Significant additional training is needed to MISO Transmission Owners Agreement and
qualify an individual to perform system opera- was not under the MISO tariff, so MISO did not
tion and management functions. have the necessary authority as FE’s Reliability
Coordinator as required by NERC Policy 9, Sec-
tion B, Requirement 2.
Group 3: FE failed to manage adequately tree
growth in its transmission rights-of-way. This u Although lacking authority under a signed
failure was the common cause of the outage of agreement, MISO as reliability coordinator nev-
three FE 345-kV transmission lines and ertheless should have issued directives to FE to
affected several 138-kV lines. return system operation to a safe and reliable
level as required by NERC Policy 9, Section B,
u FE failed to maintain equipment ratings
Requirement 2, before the cascading outages
through a vegetation management program. A
occurred.
vegetation management program is necessary to
fulfill NERC Policy 2, Section A, Requirement 1 u American Electric Power (AEP) and PJM
(Control areas shall develop, maintain, and attempted to use the transmission loading relief
implement formal policies and procedures to (TLR) process to address transmission power
provide for transmission security . . . including flows without recognizing that a TLR would not
equipment ratings.) solve the problem.
u Vegetation management requirements are not u NERC Policy 9 does not contain a requirement
defined in NERC Standards and Policies. for reliability coordinators equivalent to the
NERC Policy 2 statement that monitoring
equipment is to be used in a manner that would
Group 4: Failure of the interconnected grid’s
bring to the reliability coordinator’s attention
reliability organizations to provide effective
any important deviations in operating
diagnostic support.
conditions.
Violations (Identified by NERC):
u NERC Policy 9 lacks criteria for determining the
u Violation 4: MISO did not notify other reliabil- critical facilities lists in each reliability coordi-
ity coordinators of potential system problems as nator area.
required by NERC Policy 9, Section C, Require-
u NERC Policy 9 lacks specifics on coordinated
ment 2.
procedures and training for reliability coordina-
u Violation 5: MISO was using non-real-time data tors regarding “operating to the most conserva-
to support real-time operations, in violation of tive limit” in situations when operating
NERC Policy 9, Appendix D, Section A, Criteria conditions are not fully understood.
5.2.
u Violation 6: PJM and MISO as reliability coordi- Failures to act by FirstEnergy or others to solve
nators lacked procedures or guidelines between the growing problem, due to the other causes.
their respective organizations regarding the
Violations (Identified by NERC):
coordination of actions to address an operating
security limit violation observed by one of them u Violation 1: Following the outage of the Cham-
in the other’s area due to a contingency near berlin-Harding 345-kV line, FE operating per-
their common boundary, as required by Policy sonnel did not take the necessary action to
9, Appendix C. Note: Policy 9 lacks specifics on return the system to a safe operating state as
what constitutes coordinated procedures and required by NERC Policy 2, Section A, Standard
training. 1.
Other Problems: u Violation 2: FE operations personnel did not
u MISO did not have adequate monitoring capa- adequately communicate its emergency operat-
bility to fulfill its reliability coordinator respon- ing conditions to neighboring systems as
sibilities as required by NERC Policy 9, required by NERC Policy 5, Section A.
Appendix D, Section A. Other Problems:
u Although MISO is the reliability coordinator for u FE operations personnel did not promptly take
FE, on August 14 FE was not a signatory to the action as required by NERC Policy 5, General
20 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Criteria, to relieve the abnormal conditions 1. Although NERC’s provisions address many of
resulting from the outage of the Harding- the factors and practices which contributed to
Chamberlin 345-kV line. the blackout, some of the policies or guidelines
are inexact, non-specific, or lacking in detail,
u FE operations personnel did not implement
allowing divergent interpretations among reli-
measures to return system operation to within
ability councils, control areas, and reliability
security limits in the prescribed time frame
coordinators. NERC standards are minimum
of NERC Policy 2, Section A, Standard 2, follow-
requirements that may be made more stringent
ing the outage of the Harding-Chamberlin
if appropriate by regional or subregional bodies,
345-kV line.
but the regions have varied in their willingness
u FE operations personnel did not exercise the to implement exacting reliability standards.
authority to alleviate the operating security
limit violation as required by NERC Policy 5, 2. NERC and the industry’s reliability community
Section C, Requirement 2. were aware of the lack of specificity and detail
in some standards, including definitions of
u FE did not exercise a load reduction program to Operating Security Limits, definition of
relieve the critical system operating conditions planned outages, and delegation of Reliability
as required by NERC Policy 2, Section A, Coordinator functions to control areas, but they
Requirement 1.2. moved slowly to address these problems
u FE did not demonstrate the application of effectively.
effective emergency operating procedures as 3. Some standards relating to the blackout’s
required by NERC Policy 6, Section B, Emer- causes lack specificity and measurable compli-
gency Operations Criteria. ance criteria, including those pertaining to
u FE operations personnel did not demonstrate operator training, back-up control facilities,
that FE has an effective manual load shedding procedures to operate when part or all of the
program designed to address voltage decays EMS fails, emergency procedure training,
that result in uncontrolled failure of compo- system restoration plans, reactive reserve
nents of the interconnection as required by requirements, line ratings, and vegetation
NERC Policy 5, General Criteria. management.
u NERC Policy 5 lacks specifics for Control Areas 4. The NERC compliance program and region-
on procedures for coordinating with other sys- based auditing process has not been compre-
tems and training regarding “operating to the hensive or aggressive enough to assess the capa-
most conservative limit” in situations when bility of all control areas to direct the operation
operating conditions are not fully understood. of their portions of the bulk power system. The
effectiveness and thoroughness of regional
councils’ efforts to audit for compliance with
Institutional Issues reliability requirements have varied signifi-
cantly from region to region. Equally important,
As indicated above, the investigation team identi-
absent mandatory compliance and penalty
fied a number of institutional issues with respect
authority, there is no requirement that an entity
to NERC’s reliability standards. Many of the insti-
found to be deficient in an audit must remedy
tutional problems arise not because NERC is an
the deficiency.
inadequate or ineffective organization, but rather
because it has no structural independence from 5. NERC standards are frequently administrative
the industry it represents and has no authority to and technical rather than results-oriented.
develop strong reliability standards and to enforce
compliance with those standards. While many in 6. A recently-adopted NERC process for develop-
the industry and at NERC support such measures, ment of standards is lengthy and not yet fully
legislative action by the U.S. Congress is needed to understood or applied by many industry partic-
make this happen. ipants. Whether this process can be adapted to
support an expedited development of clear and
These institutional issues can be summed up auditable standards for key topics remains to be
generally: seen.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 21
7. NERC has not had an effective process to ensure (While Policy 5 on Emergency Operations does not address
that recommendations made in various reports the issue of “operating to the most conservative limit” when
coordinating with other systems and operating conditions are
and disturbance analyses are tracked for not understood, other NERC policies do address this matter:
accountability. On their own initiative, some Policy 2, Section A, Standard 1, on basic reliability for single
regional councils have developed effective contingencies; Policy 2, Section A, Standard 2, to return a sys-
tracking procedures for their geographic areas. tem to within operating security limits within 30 minutes;
Policy 2, Section A, Requirement 1, for formal policies and
Control areas and reliability coordinators operate procedures to provide for transmission security; Policy 5,
the grid every day under guidelines, policies, and General Criteria, to relieve any abnormal conditions that jeop-
ardize reliable operation; Policy 5, Section C, Requirement 1,
requirements established by the industry’s reli- to relieve security limit violations; and Policy 5, Section 2,
ability community under NERC’s coordination. If Requirement 2, which gives system operators responsibility
those policies are strong, clear, and unambiguous, and authority to alleviate operating security limit violations
then everyone will plan and operate the system at using timely and appropriate actions.)
Violation 2: FE did not notify other systems of an impend-
a high level of performance and reliability will be ing system emergency. (Policy 5, Section A, Requirement 1,
high. But if those policies are ambiguous and do directs a system to inform other systems if it is burdening oth-
not make entities’ roles and responsibilities clear ers, reducing system reliability, or if its lack of single contin-
and certain, they allow companies to perform at gency coverage could threaten Interconnection reliability.
varying levels and system reliability is likely to be Policy 5, Section A, Criteria, has similar provisions.)
Violation 3: FE’s state estimation/contingency analysis
compromised. tools were not used to assess the system conditions. (This is
addressed in Operating Policy 5, Section C, Requirement 3,
Given that NERC has been a voluntary organiza- concerning assessment of Operating Security Limit viola-
tion that makes decisions based on member votes, tions, and Policy 4, Section A, Requirement 5, which
if NERC’s standards have been unclear, non- addresses using monitoring equipment to inform the system
specific, lacking in scope, or insufficiently strict, operator of important conditions and the potential need for
that reflects at least as much on the industry com- corrective action.)
Violation 4: MISO did not notify other reliability coordina-
munity that drafts and votes on the standards as it tors of potential problems. (Policy 9, Section C, Requirement
does on NERC. Similarly, NERC’s ability to obtain 2, directing the reliability coordinator to alert all control areas
compliance with its requirements through its and reliability coordinators of a potential transmission prob-
audit process has been limited by the extent to lem.)
Violation 5: MISO was using non-real-time data to support
which the industry has been willing to support the real-time operations. (Policy 9, Appendix D, Section A, Crite-
audit program. ria For Reliability Coordinators 5.2, regarding adequate facili-
ties to perform their responsibilities, including detailed
Endnotes monitoring capability to identify potential security viola-
1 Webster’s II New Riverside University Dictionary, Riverside tions.)
Violation 6: PJM and MISO as Reliability Coordinators
Publishing Co., 1984. lacked procedures or guidelines between themselves on when
2 A NERC team looked at whether and how violations of
and how to coordinate an operating security limit violation
NERC’s reliability requirements may have occurred in the observed by one of them in the other’s area due to a contin-
events leading up to the blackout. They also looked at gency near their common boundary (Policy 9, Appendix 9C,
whether deficiencies in the requirements, practices and pro- Emergency Procedures). Note: Since Policy 9 lacks specifics
cedures of NERC and the regional reliability organizations on coordinated procedures and training, it was not possible
may have contributed to the blackout. They found seven spe- for the bi-national team to identify the exact violation that
cific violations of NERC operating policies (although some are occurred.
qualified by a lack of specificity in the NERC requirements). Violation 7: The monitoring equipment provided to FE
The Standards, Procedures and Compliance Investigation operators was not sufficient to bring the operators’ attention
Team reviewed the NERC Policies for violations, building on to the deviation on the system. (Policy 4, Section A, System
work and going beyond work done by the Root Cause Analy- Monitoring Requirements regarding resource availability and
sis Team. Based on that review the Standards team identified the use of monitoring equipment to alert operators to the need
a number of violations related to policies 2, 4, 5, and 9. for corrective action.)
Violation 1: Following the outage of the Chamberlin- 3 NERC has not yet completed its review of planning stan-
Harding 345-kV line, FE did not take the necessary actions to dards and violations.
return the system to a safe operating state within 30 minutes.
22 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
4. Context and Preconditions for the Blackout:
The Northeastern Power Grid
Before the Blackout Began
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 23
unpredictable times. Sound reliability manage- It is a basic principle of reliability management
ment is designed to ensure that operators can con- that “operators must operate the system they have
tinue to operate the system within appropriate in front of them”—unconditionally. The system
thermal, voltage, and stability limits following the must be operated at all times to withstand any sin-
unexpected loss of any key element (such as a gle contingency and yet be ready within 30 min-
major generator or key transmission facility). utes for the next contingency. If a facility is lost
These practices have been designed to maintain a unexpectedly, the system operators must deter-
functional and reliable grid, regardless of whether mine whether to make operational changes,
actual operating conditions are normal. including adjusting generator outputs, curtailing
Geography Lesson
In analyzing the August 14 blackout, it is crucial This geographic distinction matters because
to understand the geography of the FirstEnergy the Cleveland-Akron area is a transmission-
area. FirstEnergy has seven subsidiary distribu- constrained load pocket with relatively limited
tion utilities: Toledo Edison, Ohio Edison, and generation. While some analyses of the blackout
The Illuminating Company in Ohio and four refer to voltages and other indicators measured at
more in Pennsylvania and New Jersey. Its Ohio the boundaries of FE’s Ohio control area, those
control area spans the three Ohio distribution indicators have limited relevance to the black-
utility footprints and that of Cleveland Public out—the indicators of conditions at the edges of
Power, a municipal utility serving the city of and within the Cleveland-Akron area are the
Cleveland. Within FE’s Ohio control area is the ones that matter.
Cleveland-Akron area, shown in red cross-hatch.
Area All-Time Peak Load (MW) Load on August 14, 2003 (MW)
Cleveland-Akron Area
(including Cleveland Public Power) 7,340 6,715
FirstEnergy Control Area, Ohio 13,299 12,165
FirstEnergy Retail Area, including PJM 24,267 22,631
NA = not applicable.
24 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
electricity transactions, taking transmission ele- August 11 and 14. Figure 4.2 shows forecast and
ments out of service or restoring them, and if nec- actual power demands for AEP, Michigan Electri-
essary, shedding interruptible and firm customer cal Coordinated Systems (MECS), and FE from
load—i.e., cutting some customers off tempo- August 11 through August 14. Variances between
rarily, and in the right locations, to reduce elec- actual and forecast loads are not unusual, but
tricity demand to a level that matches what the because those forecasts are used for day-ahead
system is then able to deliver safely. planning for generation, purchases, and reactive
This chapter discusses system conditions in and power management, they can affect equipment
around northeast Ohio on August 14 and their rel- availability and schedules for the following day.
evance to the blackout. It reviews electric loads The existence of high air conditioning loads across
(real and reactive), system topology (transmission the Midwest on August 14 is relevant because air
and generation equipment availability and capa- conditioning loads (like other induction motors)
bilities), power flows, voltage profiles and reactive have lower power factors than other customer
power reserves. The discussion examines actual electricity uses, and consume more reactive
system data, investigation team modeling results, power. Because it had been hot for several days in
and past FE and AEP experiences in the Cleve- the Cleveland-Akron area, more air conditioners
land-Akron area. The detailed analyses will be were running to overcome the persistent heat, and
presented in a NERC technical report. consuming relatively high levels of reactive
power—further straining the area’s limited reac-
Electric Demands on August 14 tive generation capabilities.
Figure 4.1. August 2003 Temperatures in the U.S. Figure 4.2. Load Forecasts Below Actuals,
Northeast and Eastern Canada August 11 through 14
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 25
transmission facilities known to be out of service determined that the regional system could be
in the day-ahead planning studies they perform to operated safely. The unavailability of these gener-
ensure a secure system for the next day. Knowing ation units did not cause the blackout.
the status of key facilities also helps operators
determine in advance the safe electricity transfer On August 14 four or five capacitor banks within
levels for the coming day. the Cleveland-Akron area had been removed from
service for routine inspection, including capacitor
MISO’s day-ahead planning studies for August 14 banks at Fox and Avon 138-kV substations.1
took the above generator outages and transmission These static reactive power sources are important
outages reported to MISO into account and for voltage support, but were not restored to
26 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
service that afternoon despite the system opera- team has shown that this outage did not cause
tors’ need for more reactive power in the area.2 the subsequent events in northern Ohio that led
Normal utility practice is to inspect and maintain to the blackout. However, since the line was not
reactive resources in off-peak seasons so the facili- in MISO’s footprint, MISO operators did not
ties will be fully available to meet peak loads. monitor the status of this line and did not know
The unavailability of the critical it had gone out of service. This led to a data mis-
Cause 1 match that prevented MISO’s state estimator (a
Inadequate reactive resources was not known
System to those outside of FirstEnergy. key monitoring tool) from producing usable
Understanding NERC policy requires that critical results later in the day at a time when system
facilities be identified and that conditions in FE’s control area were deteriorat-
neighboring control areas and reliability coordina- ing (see details below, Recommendation
tors be made aware of the status of those facilities pages 46 and 48-49). 30, page 163
to identify the impact of those conditions on their u Eastlake Unit 5 is a 597 MW (net) generating
own facilities. However, FE never identified these unit located west of Cleveland on Lake Erie. It is
capacitor banks as critical a major source of reactive power support for the
Recommendations
and so did not pass on sta- 23, page 160; 30, page 163 Cleveland area. It tripped at 13:31 EDT. The
tus information to others. cause of the trip was that as the Eastlake 5 oper-
ator sought to increase the unit’s reactive power
Unanticipated Outages of output (Figure 4.3), the unit’s protection system
detected that VAr output exceeded the unit’s
Transmission and Generation VAr capability and tripped the unit off-line. The
on August 14 loss of the Eastlake 5 unit did not put the grid
into an unreliable state—i.e., it was still able to
Three notable unplanned outages occurred in withstand safely another contingency. How-
Ohio and Indiana on August 14 before 15:05 EDT. ever, the loss of the unit required FE to import
Around noon, several Cinergy transmission lines additional power to make up for the loss of the
in south-central Indiana tripped; at 13:31 EDT, unit’s output (612 MW), made voltage manage-
FE’s Eastlake 5 generating unit along the south- ment in northern Ohio more challenging, and
western shore of Lake Erie tripped; at 14:02 EDT, a gave FE operators less flexibility in operating
line within the Dayton Power and Light (DPL) con- their system (see details on pages 45-46 and
trol area, the Stuart-Atlanta 345-kV line in south- 49-50).
ern Ohio, tripped. Only the Eastlake 5 trip was
electrically significant to the FirstEnergy system. Key Parameters for the
u Transmission lines on the Cinergy 345-, 230-, Cleveland-Akron Area
and 138-kV systems experienced a series of out-
ages starting at 12:08 EDT and remained out of at 15:05 EDT
service during the entire blackout. The loss of The investigation team benchmarked their power
these lines caused significant voltage and load- flow models against measured data provided by
ing problems in the Cinergy area. Cinergy made
generation changes, and MISO operators Figure 4.3. MW and MVAr Output from Eastlake
Unit 5 on August 14
responded by implementing transmission load-
ing relief (TLR) procedures to control flows on
the transmission system in south-central Indi-
ana. System modeling by the investigation team
(see details below, pages 41-43) showed that the
loss of these lines was not electrically related to
subsequent events in northern Ohio that led to
the blackout.
u The Stuart-Atlanta 345-kV line, operated by
DPL, and monitored by the PJM reliability coor-
dinator, tripped at 14:02 EDT. This was the
result of a tree contact, and the line remained
out of service the entire afternoon. As explained
below, system modeling by the investigation
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 27
FirstEnergy for the Cleveland-Akron area at 15:05 reactive power (330 MVAr) from shunt capacitors
EDT (just before the first of FirstEnergy’s key and line charging and a 10% increase (250 MVAr)
transmission lines failed), as shown in Table 4.2. in reactive losses from transmission lines, these
Although the modeled figures do not match actual parameters indicate that the Cleveland-Akron area
system conditions perfectly, overall this model would be precariously short of reactive power if
shows a very high correspondence to the actual the Perry plant were lost.
occurrences and thus its results merit a high
degree of confidence. Although Table 4.2 shows
only a few key lines within the Cleveland-Akron
Power Flow Patterns
area, the model was successfully benchmarked to Several commentators have suggested that the
match actual flows, line-by-line, very closely voltage problems in northeast Ohio and the subse-
across the entire area for the afternoon of August quent blackout occurred due to unprecedented
14, 2003. high levels of inter-regional power transfers occur-
The power flow model assumes the following sys- ring on August 14. Investigation team analysis
tem conditions for the Cleveland-Akron area at indicates that in fact, power transfer levels were
15:05 EDT on August 14: high but were within established limits and previ-
ously experienced levels. Analysis of actual and
u Cleveland-Akron area load = 6,715 MW and test case power flows demonstrates that inter-
2,402 MVAr regional power transfers had a minimal effect on
u Transmission losses = 189 MW and 2,514 the transmission corridor containing the Har-
MVAr ding-Chamberlin, Hanna-Juniper, and Star-South
Canton 345-kV lines on August 14. It was the
u Reactive power from fixed shunt capacitors (all
increasing native load relative to the limited
voltage levels) = 2,585 MVAr amount of reactive power available in the Cleve-
u Reactive power from line charging (all voltage land-Akron area that caused the depletion of reac-
levels) = 739 MVAr tive power reserves and declining voltages.
u Network configuration = after the loss of On August 14, the flow of power through the
Eastlake 5, before the loss of Harding- ECAR region as a whole (lower Michigan, Indiana,
Chamberlin 345-kV line Ohio, Kentucky, West Virginia, and western Penn-
sylvania) was heavy as a result of transfers of
u Area generation combined output: 3,000 MW
power from the south (Tennessee, etc.) and west
and 1,200 MVAr.
(Wisconsin, Minnesota, Illinois, Missouri, etc.) to
Cause 1 Given these conditions, the power the north (Ohio, Michigan, and Ontario) and east
Inadequate flow model indicates that about (New York, Pennsylvania). The destinations for
System 3,900 MW and 400 MVAr of real much of the power were northern Ohio, Michigan,
Understanding power and reactive power flow PJM, and Ontario. This is shown in Figure 4.4,
into the Cleveland-Akron area which shows the flows between control areas on
was needed to meet the sum of customer load August 14 based on power flow simulations just
demanded plus line losses. There was about 688 before the Harding-Chamberlin line tripped at
MVAr of reactive reserve from generation in the 15:05 EDT. FE’s total load peaked at 12,165MW at
area, which is slightly more than the 660 MVAr 16:00 EDT. Actual system data indicate that
reactive capability of the Perry nuclear unit. Com- between 15:00 and 16:00 EDT, actual line flows
bined with the fact that a 5% reduction in operat- into FE’s control area were 2,695 MW for both
ing voltage would cause a 10% reduction in transactions and native load.
28 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Figure 4.4. Generation, Demand, and Interregional Power Flows on August 14, 2003, at 15:05 EDT
Figure 4.5 shows total scheduled imports for the Figure 4.5. Scheduled Imports and Exports for
entire northeast region for June through August the Northeast Central Region, June 1 through
14, 2003. These transfers were well within the August 13, 2003
range of previous levels, as shown in Figure 4.5,
and well within all established limits. In particu-
lar, on August 14 increasing amounts of the grow-
ing imports into the area were being delivered to
FirstEnergy’s Ohio territory to meet its increasing
demand and to replace the generation lost with the
trip of Eastlake 5. The level of imports into Ontario
from the U.S. on August 14 was high (e.g., 1,334
MW at 16:00 EDT through the New York and
Michigan ties) but not unusual, and well within
IMO’s import capability. Ontario is a frequent
importer and exporter of power, and had imported
similar and higher amounts of power several times
during the summers of 2002 and 2003. PJM and Note: These flows from within the Northeast Central Area
Michigan also routinely import and export power include ECAR, PJM, IMO, NYISO, and exclude transfers from
Québec, the Maritimes and New England, since the latter areas
across ECAR.
had minimal flows across the region of interest.
Some have suggested that the level of power flows transactions to determine whether and how the
into and across the Midwest was a direct cause of transactions affected line loadings within the
the blackout on August 14. Investigation team Cleveland-Akron area. The MUST (Managing Uti-
modeling proves that these flows were neither a lization of System Transmission) analytical tool
cause nor a contributing factor to the blackout. uses the transactions data from TagNet along with
The team used detailed modeling and simulation a power flow program to determine the impact of
incorporating the NERC TagNet data on actual transactions on the loading of transmission
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 29
flowgates or specific facilities, calculating transfer generation, and as such excluded from TLR cur-
distribution factors across the various flowgates. tailments. An additional small increment of line
The MUST analysis shows that for actual flows at loading served transactions for which FE was
15:05 EDT, only 10% of the loading on Cleve- either the importer or exporter, and the remaining
land-Akron lines was for through flows for which line loading was due to through-flows initiated
FE was neither the importer nor exporter. and received by other entities. The Star-South
According to real-time TagNet records, at 15:05 Canton line experienced the greatest impact from
EDT the incremental flows due to transactions through-flows—148 MW, or 18% of the total line
were approximately 2,800 MW flowing into the loading at 15:05 EDT, was due to through-flows
FirstEnergy control area and approximately 800 resulting from non-FE transactions. By 15:41 EDT,
MW out of FE to Duquesne Light Company right before Star-South Canton tripped—without
(DLCO). Among the flows into or out of the FE being overloaded—the Sammis-Star line was serv-
control area, the bulk of the flows were for transac- ing almost entirely native load, with loading from
tions where FE was the recipient or the source—at through-flows down to only 4.5%.
15:05 EDT the incremental flows due to transac- Cause 1 The central point of this analysis
tions into FE were 1,300 MW from interconnec- Inadequate is that because the critical lines
tions with PJM, AEP, DPL and MECS, and System were loaded primarily to serve
approximately 800 MW from interconnections Understanding native load and FE-related flows,
with DLCO. But not all of that energy moved attempts to reduce flows through
through the Cleveland-Akron area and across the transaction curtailments in and around the Cleve-
lines which failed on August 14, as Figure 4.6 land-Akron area would have had minimal impact
shows. on line loadings and the declining voltage situa-
Figure 4.6 shows how all of the transactions flow- tion within that area. Rising load in the Cleve-
ing across the Cleveland-Akron area on the after- land-Akron area that afternoon was depleting the
noon of August 14 affected line loadings at key FE remaining reactive power reserves. Since there
facilities, organized by time and types of transac- was no additional in-area generation, only in-area
tions. It shows that before the first transmission load cuts could have reduced local line loadings
line failed, the bulk of the loading on the four criti- and improved voltage security. This is confirmed
cal FirstEnergy circuits—Harding-Chamberlin, by the loadings on the
Recommendations
Hanna-Juniper, Star-South Canton and Sammis- Sammis-Star at 15:42 EDT, 3, page 143; 23, page 160
Star—was to serve Cleveland-Akron area native after the loss of Star-South
load. Flows to serve native load included transfers Canton—fully 96% of the current on that line was
from FE’s 1,640 MW Beaver Valley nuclear power to serve FE load and FE-related transactions, and a
plant and its Seneca plant, both in Pennsylvania, cut of every non-FE through transaction flowing
which have been traditionally counted by across northeast Ohio would have obtained only
FirstEnergy not as imports but rather as in-area 59 MW (4%) of relief for this specific line. This
means that redispatch of generation beyond north-
Figure 4.6. Impacts of Transactions Flows on east Ohio would have had almost no impact upon
Critical Line Loadings, August 14, 2003 conditions within the Cleveland-Akron area
(which after 13:31 EDT had no remaining genera-
tion reserves). Equally important, cutting flows on
the Star-South Canton line might not have
changed subsequent events—because the line
opened three times that afternoon due to tree con-
tacts, reducing its loading would not have assured
its continued operation.
Power flow patterns on August 14 did not cause
the blackout in the Cleveland-Akron area. But
once the first four FirstEnergy lines went down,
the magnitude and pattern of flows on the overall
system did affect the ultimate path, location and
speed of the cascade after 16:05:57 EDT.3
30 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Voltages and Voltage Criteria were declining over the afternoon of August 14.
Actual measured voltage levels at the Star bus and
During the days before August 14 and throughout others on FE’s transmission system on August 14
the morning and mid-day on August 14, voltages were below 100% starting early in the day. At
were depressed across parts of northern Ohio 11:00 EDT, voltage at the Star bus equaled 98.5%,
because of high air conditioning demand and declined to 97.3% after the loss of Eastlake 5 at
other loads, and power transfers into and to a 13:31 EDT, and dropped to 95.9% at 15:05 EDT
lesser extent across the region. Voltage varies by after the loss of the Harding-Chamberlin line.
location across an electrical region, and operators FirstEnergy system operators reported this voltage
monitor voltages continuously at key locations performance to be typical for a warm summer day
across their systems. on the FirstEnergy system. The gradual decline of
voltage over the early afternoon was consistent
Entities manage voltage using long-term planning with the increase of load over the same time
and day-ahead planning for adequate reactive period, particularly given that FirstEnergy had no
supply, and real-time adjustments to operating additional generation within the Cleveland-Akron
equipment. On August 14, for example, PJM area load pocket to provide additional reactive
implemented routine voltage management proce- support.
dures developed for heavy load conditions.
Within Ohio, FE began preparations early in the Cause 1 NERC and regional reliability
afternoon of August 14, requesting capacitors to Inadequate councils’ planning criteria and
be restored to service4 and additional voltage sup- System operating policies (such as NERC
Understanding I.A and I.D, NPCC A-2, and ECAR
port from generators.5 As the day progressed,
operators across the region took additional Document 1) specify voltage crite-
actions, such as increasing plants’ reactive power ria in such generic terms as: acceptable voltages
output, plant redispatch, and transformer tap under normal and emergency conditions shall be
changes to respond to changing voltage maintained within normal limits and applicable
conditions. emergency limits respectively, with due recogni-
tion to avoiding voltage instability and wide-
Voltages at key FirstEnergy buses (points at which spread system collapse in the event of certain
lines, generators, transformers, etc., converge) contingencies. Each system then defines its own
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 31
acceptable voltage criteria based on its own sys- declines from the pre-contingency voltage. Good
tem design and equipment characteristics, detail- utility practice requires that these determinations
ing quantified measures including acceptable be based on a full set of V-Q (voltage performance
minimum and maximum voltages in percentages V relative to reactive power supply Q) and P-V
of nominal voltage and acceptable voltage (real power transfer P relative to voltage V)
32 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
analyses for a wide range of system conditions. Cause 1 Figure 4.7 shows the profile of
Table 4.3 compares the voltage criteria used by Inadequate voltage levels at key buses from
FirstEnergy and other relevant transmission oper- System southeast Michigan across Ohio
ators in the region. As this table shows, FE uses Understanding into western Pennsylvania from
minimum acceptable normal voltages which are August 11 through 14 and for sev-
lower than and incompati- eral hours on August 14. These transects show
Recommendation
ble with those used by its 23, page 160
that across the area, voltage levels were consis-
interconnected neighbors. tently lower at the 345-kV buses in the Cleve-
land-Akron area (from Beaver to Hanna on the
Cause 1 The investigation team probed
west to east plot and from Avon Lake to Star on the
Inadequate deeply into voltage management
north to south plot) for the three days and the
System issues within the Cleveland-
Understanding 13:00 to 15:00 EDT period preceding the blackout.
Akron area. As noted previously,
Voltage was consistently and considerably higher
a power system with higher oper-
at the outer ends of each transect, where it never
ating voltage and larger reactive power reserves is
dropped below 96% even on August 14. These
more resilient or robust in the face of load
profiles also show clearly the decline of voltage
increases and operational contingencies. Higher
over the afternoon of August 14, with voltage at
transmission voltages enable higher power trans-
the Harding bus at 15:00 EDT just below 96%
fer capabilities and reduce transmission line
before the Harding-Chamberlin line tripped at
losses (both real and reactive). For the Cleve-
15:05 EDT, and dropping down to around 93% at
land-Akron area, FE has been operating the system
16:00 EDT after the loss of lines and load in the
with the minimum voltage level at 90% of nominal
immediate area.
rating, with alarms set at 92%.6 The criteria allow
for a single contingency to occur if voltage remains Cause 1 Using actual data provided by FE,
above 90%. The team conducted extensive voltage Inadequate ITC, AEP and PJM, Figure 4.8
stability studies (discussed below), concluding System shows the availability of reactive
that FE’s 90% minimum voltage level was not only Understanding reserves (the difference between
far less stringent than nearby interconnected sys- reactive power generated and the
tems (most of which set the pre-contingency mini- maximum reactive capability) within the Cleve-
mum voltage criteria at 95%), but was not land-Akron area and four regions surrounding it,
adequate for secure system operations. from ITC to PJM. On the afternoon of August 14,
the graph shows that reactive power generation
Examination of the Form 715 filings made by Ohio
was heavily taxed in the Cleveland-Akron area but
Edison, FE’s predecessor company, for 1994
that extensive MVAr reserves were available in
through 1997 indicate that Ohio Edison used a
the neighboring areas. As the afternoon pro-
pre-contingency bus voltage criteria of 95 to 105 %
gressed, reactive reserves diminished for all five
and 90% emergency post-contingency voltage,
regions as load grew. But reactive reserves were
with acceptable change in voltage no greater than
fully depleted within the Cleveland-Akron area by
5%. These historic criteria were compatible with
16:00 EDT without drawing down the reserves in
neighboring transmission operator practices.
neighboring areas, which remained at scheduled
A look at voltage levels across the region illus- voltages. The region as a whole had sufficient
trates the difference between FE’s voltage reactive reserves, but because reactive power can-
situation on August 14 and that of its neighbors. not be transported far but must be supplied from
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 33
Figure 4.7. Actual Voltages Across the Ohio Area Before and On August 14, 2003
34 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Voltage Stability Analysis
Voltage instability or voltage collapse occurs on a falls the reactive requirement also falls. But when
power system when voltages progressively voltage is lowered at the bus and the reactive
decline until stable operating voltages can no requirement at that bus begins to increase (rather
longer be maintained. This is precipitated by an than continuing to decrease), the system
imbalance of reactive power supply and demand, becomes unstable. The voltage point correspond-
resulting from one or more changes in system ing to the transition from stable to unstable con-
conditions including increased real or reactive ditions is known as the “critical voltage,” and the
loads, high power transfers, or the loss of genera- reactive power level at that point is the “reactive
tion or transmission facilities. Unlike the phe- margin.” The desired operating voltage level
nomenon of transient instability, where should be well above the critical voltage with a
generators swing out of synchronism with the large buffer for changes in prevailing system con-
rest of the power system within a few seconds or ditions and contingencies. Similarly, reactive
less after a critical fault, voltage instability can margins should be large to assure robust voltage
occur gradually within tens of seconds or levels and secure, stable system performance.
minutes.
The illustration below shows a series of V-Q
Voltage instability is best studied using V-Q curves. The lowest curve, A, reflects baseline
(voltage relative to reactive power) and P-V (real conditions for the grid with all facilities avail-
power relative to voltage) analysis. V-Q analysis able. Each higher curve represents the same
evaluates the reactive power required at a bus to loads and transfers for the region modeled, but
maintain stable voltage at that bus. A simulated with another contingency event (a circuit loss)
reactive power source is added to the bus, the occurring to make the system less stable. With
voltage schedule at the bus is adjusted in small each additional contingency, the critical voltage
steps from an initial operating point, and power rises (the point on the horizontal axis corre-
flows are solved to determine the change in reac- sponding to the lowest point on the curve) and
tive power demand resulting from the change the reactive margin decreases (the difference
in voltage. Under stable operating conditions, between the reactive power at the critical voltage
when voltage increases the reactive power and the zero point on the vertical axis). This
requirement also increases, and when voltage means the system is closer to instability.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 35
Voltage Stability Analysis (Continued)
V-Q analyses and experience with heavily loaded the curve. After a contingency occurs, such as a
power systems confirm that critical voltage levels transmission circuit or generator trip, the new
can rise above the 95% level traditionally consid- condition set is represented by curve B, with
ered as normal. Thus voltage magnitude alone is lower voltages (relative to curve A) for any load
a poor indicator of voltage stability and V-Q anal- on curve B. As the operator’s charge is to keep the
ysis must be carried out for several critical buses system stable against the next worst contingency,
in a local area, covering a range of load and gener- the system must be operated to stay well inside
ation conditions and known contingencies that the load level for the nose of curve B. If the B con-
affect voltages at these buses. tingency occurs, there is a next worst contin-
P-V analysis (real power relative to voltage) is a gency curve inside curve B, and the operator
companion tool which determines the real power must adjust the system to pull back operations to
transfer capability across a transmission inter- within the safe, buffered space represented by
face for load supply or a power transfer. Starting curve C.
from a base case system state, a series of load The investigation team conducted extensive V-Q
flows with increasing power transfers are solved and P-V analyses for the area around Cleve-
while monitoring voltages at critical buses. land-Akron for the conditions in effect on August
When power transfers reach a high enough level 14, 2003. Team members examined over fifty
a stable voltage cannot be sustained and the 345-kV and 138-kV buses across the systems of
power flow model fails to solve. The point where FirstEnergy, AEP, International Transmission
the power flow last solved corresponds to the Company, Duquesne Light Company, Alleghany
critical voltage level found in the V-Q curve for Power Systems and Dayton Power & Light. The
those conditions. On a P-V curve (see below), this V-Q analysis alone involved over 10,000 power
point is called the “nose” of the curve. flow simulations using a system model with
This set of P-V curves illustrates that for baseline more than 43,000 buses and 57,000 lines and
conditions shown in curve A, voltage remains transformers. The P-V analyses used the same
relatively steady (change along the vertical axis) model and data sets. Both examined conditions
as load increases within the region (moving out and combinations of contingencies for critical
along the horizontal axis). System conditions are times before and after key events on the
secure and stable in the area above the “nose” of FirstEnergy system on the day of the blackout.
P-V (Power-Voltage) Curves
36 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
local sources, these healthy reserves nearby could Past System Events
not support the Cleveland-Akron area’s reactive
power deficiency and growing voltage problems. and Adequacy of System Studies
Even FE’s own generation in the Ohio Valley had
reactive reserves that could not support the sag- Cause 1 In June 1994, with three genera-
ging voltages inside the Cleveland-Akron area. Inadequate tors in the Cleveland area out on
System maintenance, inadequate reactive
An important consideration in Understanding reserves and falling voltages in
Cause 1
Inadequate reactive power planning is to the Cleveland area forced Cleve-
System ensure an appropriate balance land Electric Illuminating (CEI, a predecessor
Understanding between static and dynamic reac- company to FirstEnergy) to shed load within
tive power resources across the Cleveland (a municipal utility and wholesale
interconnected system (as specified in NERC transmission and purchase customers within
Planning Standard 1D.S1). With so little genera- CEI’s control area) to avoid voltage collapse.7 The
tion left in the Cleveland-Akron area on August Cleveland-Akron area’s voltage problems were
14, the area’s dynamic reactive reserves were well-known and reflected in the stringent voltage
depleted and the area relied heavily on static com- criteria used by control area operators until 1998.8
pensation to respond to changing system condi-
tions and support voltages. But a system relying Cause 1 In the summer of 2002, AEP’s
on static compensation can experience a gradual Inadequate South Canton 765 kV to 345 kV
voltage degradation followed by a sudden drop in System transformer (which connects to
Understanding FirstEnergy’s Star 345-kV line)
voltage stability—the P-V curve for such a system
has a very steep slope close to the nose, where experienced eleven days of severe
voltage collapses. On August 14, the lack of ade- overloading when actual loadings exceeded nor-
quate dynamic reactive reserves, coupled with not mal rating and contingency loadings were at or
knowing the critical voltages and maximum above summer emergency ratings. In each
import capability to serve instance, AEP took all available actions short of
native load, left the Cleve- Recommendation load shedding to return the system to a secure
23, page 160 state, including TLRs, switching, and dispatch
land-Akron area in a very
vulnerable state. adjustments. These excessive loadings were
Figure 4.8. Reactive Reserves Around Ohio on August 14, 2003, for Representative Generators in the Area
ITC –
SW area
Fermi
Ashtabula
Monroe Perry
Eastlake
Lakeshore
PJM –
FE – Western area
Cleveland area
Keystone
FE –
Ohio Valley
Sammis
Conemaugh
Cardinal
AEP – Kammer
SSE of Akron
Note: These reactive reserve MVAr margins were calculated for the five regions for the following plants: (1) Cleveland area of
FirstEnergy—Ashtabula 5, Perry 1, Eastlake 1, Eastlake 3, Lakeshore 18; (2) Northern central portion of AEP near FirstEnergy
(South-Southeast of Akron)—Cardinal 1, Cardinal 2, Cardinal 3, Kammer 2, Kammer 3; (3) Southwest area of MECS (ITC)—Fermi
1, Monroe 2, Monroe 3, Monroe 4; (4) Ohio Valley portion of FirstEnergy—Sammis 4, Sammis 5, Sammis 6, Sammis 7; (5) Western
portion of PJM—Keystone 1, Conemaugh 1, Conemaugh 2.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 37
calculated to have diminished the remaining life especially if it occurred in combination with
of the transformer by 30%. AEP replaced this sin- outages of other critical facilities, would cause sig-
gle phase transformer in the winter of 2002-03, nificant low voltages and overloads on both the
marginally increasing the capacity of the South AEP and FirstEnergy systems. AEP shared these
Canton transformer bank. findings with FirstEnergy in a meeting on January
10, 2003.9
Following these events, AEP conducted extensive
modeling to understand the impact of a potential AEP subsequently completed a set of system stud-
outage of this transformer. That modeling re- ies, including long range studies for 2007, which
vealed that loss of the South Canton transformer, included both single contingency and extreme
38 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
disturbance possibilities. These studies showed Eastlake #62 345/138 kV transformer and the Fox
that with heavy transfers to the north, expected #1 138-kV capacitor, with other capacitors down
overloading of the South Canton transformer and in real time. Last, the study assumed a single set of
depressed voltages would occur following the loss import and export conditions, rather than testing a
of the Perry unit and the loss of the Tidd-Canton wider range of generation dispatch, import-export,
Central 345-kV line, and probable cascading into and inter-regional transfer conditions. Overall, the
voltage collapse across northeast Ohio would summer study posited less stressful system condi-
occur for nine different double contingency com- tions than actually occurred August 14, 2003
binations of generation and transmission or trans- (when load was well below historic peak demand).
mission and transmission outages.10 AEP shared It did not examine system sensitivity to key
these findings with FirstEnergy in a meeting on parameters to determine system operating limits
May 21, 2003. Meeting notes indicate that “neither within the constraints of transient stability, volt-
AEP or FE were able to identify any changes in age stability, and thermal Recommendation
transmission configuration or operating proce- capability. 23, page 160
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 39
ECAR and Organizational Independence
ECAR was established in 1967 as a regional reli- ECAR has a staff of 18 full-time employees, head-
ability council, to “augment the reliability of the quartered in Akron, Ohio. The staff provides
members’ electricity supply systems through engineering analysis and support to the various
coordination of the planning and operation of the committees and working groups. Ohio Edison, a
members’ generation and transmission facili- FirstEnergy subsidiary, administers salary, bene-
ties.”a ECAR’s membership includes 29 major fits, and accounting services for ECAR. ECAR
electricity suppliers serving more than 36 mil- employees automatically become part of Ohio
lion people. Edison’s (FirstEnergy’s) 401(k) retirement plan;
they receive FE stock as a matching share to
ECAR’s annual budget for 2003 was $5.15 mil- employee 401(k) investments and can purchase
lion (U.S.), including $1.775 million (U.S.) paid FE stock as well. Neither ECAR staff nor board
to fund NERC.b These costs are funded by its members are required to divest stock holdings in
members in a formula that reflects megawatts ECAR member companies.e Despite the close
generated, megawatt load served, and miles of link between FirstEnergy’s financial health and
high voltage lines. AEP, ECAR’s largest member, the interest of ECAR’s staff and management, the
pays about 15% of total ECAR expenses; investigation team has found no evidence to sug-
FirstEnergy pays approximately 8 to 10%.c gest that ECAR staff favor FirstEnergy’s interests
Utilities “whose generation and transmission relative to other members.
have an impact on the reliability of the intercon-
ECAR decisions appear to be dominated by the
nected electric systems” of the region are full
member control areas, which have consistently
ECAR members, while small utilities, independ-
allowed the continuation of past practices within
ent power producers, and marketers can be asso-
each control area to meet NERC requirements,
ciate members.d Its Executive Board has 22 seats,
rather than insisting on more stringent, consis-
one for each full member utility or major supplier
tent requirements for such matters as operating
(including every control area operator in ECAR).
voltage criteria or planning studies. ECAR mem-
Associate members do not have voting rights,
ber representatives also staff the reliability coun-
either on the Board or on the technical commit-
cil’s audit program, measuring individual control
tees which do all the work and policy-setting for
area compliance against local standards and
the ECAR region.
interpretations. It is difficult for an entity domi-
All of the policy and technical decisions for nated by its members to find that the members’
ECAR, including all interpretations of NERC standards and practices are inadequate. But it
guidelines, policies, and standards within ECAR, should also be recognized that NERC’s broadly
are developed by committees (called “panels”), worded and ambiguous standards have enabled
staffed by representatives from the ECAR mem- and facilitated the lax inter-
ber companies. Work allocation and leadership pretation of reliability re- Recommendations
2, page 143; 3, page 143
within ECAR are provided by the Board, the quirements within ECAR
Coordination Review Committee, and the Market over the years.
Interface Committee.
40 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Model-Based Analysis brought back into a secure operating condition
within 30 minutes.
of the State of the Regional Power
System at 15:05 EDT, Before the Perry Nuclear Plant as a
First Contingency
Loss of FE’s Harding-Chamberlin
345-kV Line Investigation team modeling demonstrates that
the Perry nuclear unit (1,255 MW near Lake Erie)
As the first step in modeling the August 14 black- is critical to the voltage stability of the Cleve-
out, the investigation team established a base case land-Akron area in general and particularly on
by creating a power flow simulation for the entire August 14. The modeling reveals that had Perry
Eastern Interconnection and benchmarking it to tripped before 15:05 EDT, voltage levels at key
recorded system conditions at 15:05 EDT on FirstEnergy buses would have fallen close to 93%
August 14. The team started with a projected sum- with only a 150 MW of area load margin (2% of the
mer 2003 power flow case for the Eastern Inter- Cleveland-Akron area load); but had Perry been
connection developed in the spring of 2003 by the lost after the Harding-Chamberlin line went down
Regional Reliability Councils to establish guide- at 15:05 EDT, the Cleveland-Akron area would
lines for safe operations for the coming summer. have been close to voltage collapse.
The level of detail involved in this region-wide
power flow case far exceeds that normally consid- Cause 1 Perry and Eastlake 5 together have
ered by individual control areas and reliability Inadequate a combined real power capability
coordinators. It consists of a detailed representa-
System of 1,852 MW and reactive capabil-
Understanding ity of 930 MVAr. If one of these
tion of more than 43,000 buses, 57,600 transmis-
sion lines, and all major generating stations across units is lost, it is necessary to
the northern U.S. and eastern Canada. The team immediately replace the lost generation with MW
revised the summer power flow case to match and MVAr imports (although reactive power does
recorded generation, demand, and power inter- not travel far under heavy loading); without
change levels among control areas at 15:05 EDT on quick-start generation or spinning reserves or
August 14. The benchmarking consisted of match- dynamic reactive reserves inside the Cleveland-
ing the calculated voltages and line flows to Akron area, system security
Recommendation
recorded observations at more than 1,500 loca- may be jeopardized. On 23, page 160
tions within the grid. Thousands of hours of effort August 14, as noted previ-
were required to benchmark the model satisfacto- ously, there were no significant spinning reserves
rily to observed conditions at 15:05 EDT. remaining within the Cleveland-Akron area fol-
lowing the loss of Eastlake 5 at 13:31 EDT. If Perry
Once the base case was benchmarked, the team had been lost FE would have been unable to meet
ran a contingency analysis that considered more the 30-minute security adjustment requirement of
than 800 possible events—including the loss of NERC’s Operating Policy 2, without the ability to
the Harding-Chamberlin 345-kV line—as points of shed load quickly. The loss of Eastlake 5 followed
departure from the 15:05 EDT case. None of these by the loss of Perry are contingencies that should
contingencies resulted in a violation of a transmis- be assessed in the operations planning timeframe,
sion line loading or bus voltage limit prior to the to develop measures to readjust the system
trip of FE’s Harding-Chamberlin 345-kV line. That between contingencies. Since FirstEnergy did not
is, according to these simulations, the system at conduct such contingency analysis planning and
15:05 EDT was capable of safe operation following develop these advance measures, it was in viola-
the occurrence of any of the tested contingencies. tion of NERC Planning Standard 1A, Category C3.
From an electrical standpoint, therefore, before
15:05 EDT the Eastern Interconnection was being This operating condition is not news. Historically,
operated within all established limits and in full the loss of Perry at full output has been recognized
compliance with NERC’s operating policies. How- as FE’s most critical single contingency for the
ever, after loss of the Harding-Chamberlin 345-kV Cleveland Electric Illuminating area, as docu-
line, the system would have exceeded emergency mented by FE’s 1998 Summer Import Capability
ratings immediately on several lines for two of the study. Perry’s MW and MVAr total output capabil-
contingencies studied—in other words, it would ity exceeded the import capability of any of the
no longer be operating in compliance with NERC critical 345-kV circuits into the Cleveland-Akron
Operating Policy A.2 because it could not be area after the loss of Eastlake 5 at 13:31 EDT. This
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 41
means that if the Perry plant had been lost on the Avon bus would have risen to 90.5%—uncom-
August 14 after Eastlake 5 went down—or on fortably close to the limits which FE considered as
many other days with similar loads and out- an acceptable operating range. But had the Perry
ages—it would have been difficult or impossible unit gone off-line after Harding-Chamberlin, reac-
for FE operators to adjust the system within 30 tive margins at all these buses would have been
minutes to prepare for the next critical contin- even tighter (with only 60 MVAr at the Harding
gency, as required by NERC Operating Policy A.2. bus), and critical voltage at Avon would have risen
In real-time operations, operators would have to to 92.5%, worse than FE’s 90% minimum accept-
calculate operating limits and prepare to use the able voltage. The system at this point would be
last resort of manually shedding large blocks of very close to voltage instability. If the first line out-
load before the second contingency, or immedi- age on August 14, 2003, had been at Hanna-
ately after it if automatic load-shedding is Juniper rather than at Harding-Chamberlin, the
available. FirstEnergy system could not have withstood the
loss of the Perry plant.
Cause 1 The investigation team could not
Inadequate find FirstEnergy contingency Cause 1 The above analysis assumed load
System plans or operational procedures Inadequate levels consistent with August 14.
Understanding for operators to manage the System But temperatures were not partic-
FirstEnergy control area and pro- Understanding ularly high that day and loads
tect the Cleveland-Akron area from the unex- were nowhere near FE’s historic
pected loss of the Perry plant. load level of 13,229 MW for the control area (in
August 2002). Therefore the investigation team
To examine the impact of this worst contingency
looked at what might have happened in the Cleve-
on the Cleveland-Akron area on August 14, Figure
land-Akron area had loads neared the historic
4.9 shows the V-Q curves for key buses in the
peak—approximately 625 MW higher than the
Cleveland-Akron area at 15:05 EDT, before and
6,715 MW peak load in the Cleveland-Akron area
after the loss of the Har-
in 2003. Figure 4.10 uses P-V analysis to show the
ding-Chamberlin line. The Recommendation
23, page 160 impact of increased load levels on voltages at the
curves on the left look at the
Star bus with and without the Perry unit before
impact of the loss of Perry
the loss of the Harding-Chamberlin line at 15:05
before the Harding-Chamberlin trip, while the
EDT. The top line shows that with the Perry plant
curves on the right show the impact had the
available, local load could have increased by 625
nuclear plant been lost after Harding-Chamberlin
MW and voltage at Star would have remained
went out of service. Had Perry gone down before
above 95%. But the bottom line, simulating the
the Harding-Chamberlin outage, reactive margins
loss of Perry, indicates that load could only have
at key FE buses would have been minimal (with
increased by about 150 MW before voltage at Star
the tightest margin at the Harding bus, read along
would have become unsolvable, indicating no
the Y-axis) and the critical voltage (the point
voltage stability margin and depending on load
before voltage collapse, read along the X-axis) at
dynamics, possible voltage collapse.
Figure 4.9. Loss of the Perry Unit Hurts Critical The above analyses indicate that the Cleveland-
Voltages and Reactive Reserves: V-Q Analyses Akron area was highly vulnerable on the after-
noon of August 14. Although the system was com-
500 Loss of Perry
Before Harding-Chamberlin
Loss of Perry
After Harding-Chamberlin
pliant with NERC Operating Policy 2A.1 for single
400 contingency reliability before the loss of the Har-
300 ding-Chamberlin line at 15:05 EDT, had FE lost
200 the Perry plant its system would have neared volt-
Reactive Power (MVAr)
Highest Critical
100 Voltage at Avon
Highest Critical
Voltage at Avon age instability or could have gone into a full volt-
(0.905 pu)
0
(0.925pu)
age collapse immediately if the Cleveland-Akron
-100
area load were 150 MW higher. It is worth noting
that this could have happened on August 14—at
-200
Avon-345 13:43 EDT that afternoon, the Perry plant operator
-300 Harding-345
Lowest Reactive Ashtabula-345 called the control area operator to warn about low
-400 Lowest Reactive Margin at Margin at Harding
Harding (165 MVAR) (80 MVAR)
Star-345
Hanna-345
voltages. At 15:36:51 EDT the Perry plant operator
-500
0.1 85%
0.15 90%
0.2 95%
0.25 100%
0.3 85%
0.35 90%
0.4 95%
0.45 100%
0.5 called FirstEnergy’s system control center to
Voltage (percent)
ask about voltage spikes at the plant’s main
42 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
transformer.14 At 15:42:49 EDT the Perry operator change of about ±0.031 Hz). Significant frequency
called the FirstEnergy operator to say, “I’m still excursions reflect large changes in load relative to
getting a lot of voltage spikes and swings on the generation and could cause unscheduled flows
generator . . . . I’m taking field volts pretty close to between control areas and even, in the extreme,
where I’ll trip the turbine off.”15 cause automatic under-frequency load-shedding
or automatic generator trips.
System Frequency The investigation team examined Eastern Inter-
connection frequency and Area Control Error
Assuming stable conditions, the system frequency (ACE) for August 14, 2003 and the entire month of
is the same across an interconnected grid at any August, looking for patterns and anomalies.
particular moment. System frequency will vary Extensive analysis using Fast Fourier Transforms
from moment to moment, however, depending on (described in the NERC Technical Report)
the second-to-second balance between aggregate revealed no unusual variations. Rather, trans-
generation and aggregate demand across the inter- forms using various time samples of average fre-
connection. System frequency is monitored on a quency (from 1 hour to 6 seconds in length)
continuous basis. indicate instead that the Eastern Interconnection
There were no significant or unusual frequency exhibits regular deviations.16
oscillations in the Eastern Interconnection on The largest deviations in frequency occur at regu-
August 14 prior to 16:09 EDT compared to prior lar intervals. These intervals reflect interchange
days, and frequency was well within the bounds
of safe operating practices. System frequency vari-
ation was not a cause or precursor of the initiation Figure 4.10. Impact of Perry Unit Outage on
of the blackout. But once the cascade began, the Cleveland-Akron Area Voltage Stability
large frequency swings that occurred early on
became a principal means by which the blackout
spread across a wide area.
Figure 4.11 shows Eastern Interconnection fre-
quency on August 14, 2003. Frequency declines or
increases from a mismatch between generation
and load on the order of about 3,200 MW per
0.1 Hertz (alternatively, a change in load or gener-
ation of 1,000 MW would cause a frequency
Frequency Management
Each control area is responsible for maintaining
a balance between its generation and demand. If
persistent under-frequency occurs, at least one
Figure 4.11. Frequency on August 14, 2003,
control area somewhere is “leaning on the grid,” up to 16:09 EDT
meaning that it is taking unscheduled electric-
ity from the grid, which both depresses system
frequency and creates unscheduled power
flows. In practice, minor deviations at the con-
trol area level are routine; it is very difficult to
maintain an exact balance between generation
and demand. Accordingly, NERC has estab-
lished operating rules that specify maximum
permissible deviations, and focus on prohibit-
ing persistent deviations, but not instantaneous
ones. NERC monitors the performance of con-
trol areas through specific measures of control
performance that gauge how accurately each
control area matches its load and generation.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 43
schedule changes at the peak to off-peak schedule Figure 4.12. Hourly Deviations in Eastern
changes (06:00 to 07:00 and 21:00 to 22:00, as Interconnection Frequency for the Month of
shown in Figure 4.12) and on regular hourly and August 2003
half-hour schedule changes as power plants ramp
up and down to serve scheduled purchases and
interchanges. Frequency tends to run high in the
early part of the day because extra generation
capacity is committed and waiting to be dis-
patched for the afternoon peak, and then runs
lower in the afternoon as load rises relative to
available generation and spinning reserve. The
investigation team concluded that frequency data
collection and frequency management in the East-
ern Interconnection should be improved, but that
frequency oscillations before 16:09 EDT on
August 14 had no effect on the blackout.
Conclusion Endnotes
1 FE transcripts, Channel 14, 13:33:44.
Determining that the system was in a reliable 2 FE transcripts, Channel 14 at 13:21:05; channel 3 at
operational state at 15:05 EDT is extremely signifi- 13:41:54; 15:30:36.
cant for understanding the causes of the blackout. 3 “ECAR Investigation of August 14, 2003 Blackout by Major
It means that none of the electrical conditions on System Disturbance Analysis Task Force, Recommendations
Report,” page 6.
the system before 15:05 EDT was a cause of the
4 Transmission operator at FE requested the restoration of
blackout. This eliminates low voltages earlier in
the Avon Substation capacitor bank #2. Example at Channel
the day or on prior days, the unavailability of indi- 3, 13:33:40. However, no additional capacitors were
vidual generators or transmission lines (either available.
individually or in combination with one another), 5 From 13:13 through 13:28, reliability operator at FE called
high power flows to Canada, unusual system fre- nine plant operators to request additional voltage support.
quencies, and many other issues as direct, princi- Examples at Channel 16, 13:13:18, 13:15:49, 13:16:44,
pal or sole causes of the blackout. 13:20:44, 13:22:07, 13:23:24, 13:24:38, 13:26:04, 13:28:40.
6 DOE/NERC fact-finding meeting, September 2003, state-
Although FirstEnergy’s system was technically in ment by Mr. Steve Morgan (FE), PR0890803, lines 5-23.
7 See 72 FERC 61,040, the order issued for FERC dockets EL
secure electrical condition before 15:05 EDT, it
94-75-000 and EL 94-80-000, for details of this incident.
was still highly vulnerable, because some of its 8 Testimony by Stanley Szwed, Vice President of Engi-
assumptions and limits were not accurate for safe
neering and Planning, Centerior Service Company (Cleveland
operating criteria. Analysis of Cleveland-Akron Electric Illuminating Company and Toledo Edison), FERC
area voltages and reactive margins shows that docket EL 94-75-000, February 22, 1996.
FirstEnergy was operating that system on the very 9 Presentation notes for January 10, 2003 meeting between
edge of NERC operational reliability standards, AEP and FirstEnergy, and meeting summary notes by Paul
and that it could have been compromised by a Johnson, AEP Manager, East Bulk Transmission Planning,
January 10, 2003.
number of potentially disruptive scenarios that
10 “Talking Points” for May 21, 2003 meeting between AEP
were foreseeable by thorough planning and opera-
and FirstEnergy, prepared by AEP.
tions studies. A system with this little reactive 11 Memo, “Summary of AEP/FE Meeting on 5/21/03,” by Scott
margin would leave little room for adjustment, P. Lockwood, AEP, May 29, 2003.
with few relief actions available to operators in the 12 Ibid.
face of single or multiple contingencies. As the 13 Testimony by Stanley Szwed, Vice President of Engi-
next chapter will show, the vulnerability created neering and Planning, Centerior Service Company (Cleveland
by inadequate system planning and understand- Electric Illuminating Company and Toledo Edison), FERC
ing was exacerbated because the FirstEnergy oper- docket EL 94-75-000, February 22, 1996.
ators were not adequately trained or prepared to 14 FE transcript, Channel 8.
recognize and deal with emergency situations. 15 FE transcript, Channel 8.
16 See NERC Blackout Investigation Technical Reports, to be
released in 2004.
44 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
5. How and Why the Blackout Began in Ohio
and generating units automatically tripped by pro-
Summary tective relay action to avoid physical damage.
This chapter explains the major events—electri-
cal, computer, and human—that occurred as the Chapter Organization
blackout evolved on August 14, 2003, and identi-
fies the causes of the initiation of the blackout. This chapter is divided into several phases that
The period covered in this chapter begins at 12:15 correlate to major changes within the FirstEnergy
Eastern Daylight Time (EDT) on August 14, 2003 system and the surrounding area in the hours
when inaccurate input data rendered MISO’s state leading up to the cascade:
estimator (a system monitoring tool) ineffective.
u Phase 1: A normal afternoon degrades
At 13:31 EDT, FE’s Eastlake 5 generation unit trip-
ped and shut down automatically. Shortly after u Phase 2: FE’s computer failures
14:14 EDT, the alarm and logging system in FE’s u Phase 3: Three FE 345-kV transmission line fail-
control room failed and was not restored until ures and many phone calls
after the blackout. After 15:05 EDT, some of FE’s
345-kV transmission lines began tripping out u Phase 4: The collapse of the FE 138-kV system
because the lines were contacting overgrown trees and the loss of the Sammis-Star line.
within the lines’ right-of-way areas. Key events within each phase are summarized in
Figure 5.1, a timeline of major events in the origin
By around 15:46 EDT when FE, MISO and neigh-
of the blackout in Ohio. The discussion that fol-
boring utilities had begun to realize that the FE
lows highlights and explains these significant
system was in jeopardy, the only way that the
events within each phase and explains how the
blackout might have been averted would have
events were related to one another and to the cas-
been to drop at least 1,500 MW of load around
cade. Specific causes of the blackout and associ-
Cleveland and Akron. No such effort was made,
ated recommendations are identified by icons.
however, and by 15:46 EDT it may already have
been too late for a large load-shed to make any dif-
ference. After 15:46 EDT, the loss of some of FE’s Phase 1:
key 345-kV lines in northern Ohio caused its
underlying network of 138-kV lines to begin to A Normal Afternoon Degrades:
fail, leading in turn to the loss of FE’s Sammis-Star 12:15 EDT to 14:14 EDT
345-kV line at 16:06 EDT. The chapter concludes
with the loss of FE’s Sammis-Star line, the event Overview of This Phase
that triggered the uncontrollable 345 kV cascade
portion of the blackout sequence. Northern Ohio was experiencing an ordinary
August afternoon, with loads moderately high to
The loss of the Sammis-Star line triggered the cas- serve air conditioning demand, consuming high
cade because it shut down the 345-kV path into levels of reactive power. With two of Cleveland’s
northern Ohio from eastern Ohio. Although the active and reactive power production anchors
area around Akron, Ohio was already blacked out already shut down (Davis-Besse and Eastlake 4),
due to earlier events, most of northern Ohio the loss of the Eastlake 5 unit at 13:31 EDT further
remained interconnected and electricity demand depleted critical voltage support for the Cleve-
was high. This meant that the loss of the heavily land-Akron area. Detailed simulation modeling
overloaded Sammis-Star line instantly created reveals that the loss of Eastlake 5 was a significant
major and unsustainable burdens on lines in adja- factor in the outage later that afternoon—
cent areas, and the cascade spread rapidly as lines with Eastlake 5 out of service, transmission line
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 45
Figure 5.1. Timeline: Start of the Blackout in Ohio
loadings were notably higher but well within nor- 12:15 and 15:34 EDT, due to a combination of
mal ratings. After the loss of FE’s Har- human error and the effect of the loss of DPL’s Stu-
ding-Chamberlin line at 15:05 EDT, the system art-Atlanta line on other MISO lines as reflected in
eventually became unable to sustain additional the state estimator’s calculations. Without an
contingencies, even though key 345 kV line load- effective state estimator, MISO was unable to per-
ings did not exceed their normal ratings. Had form contingency analyses of generation and line
Eastlake 5 remained in service, subsequent line losses within its reliability zone. Therefore,
loadings would have been lower. Loss of Eastlake through 15:34 EDT MISO could not determine
5, however, did not initiate the blackout. Rather, that with Eastlake 5 down, other transmission
subsequent computer failures leading to the loss lines would overload if FE lost a major transmis-
of situational awareness in FE’s control room and sion line, and could not issue appropriate warn-
the loss of key FE transmission lines due to con- ings and operational instructions.
tacts with trees were the most important causes.
In the investigation interviews, all utilities, con-
At 14:02 EDT, Dayton Power & Light’s (DPL) Stu- trol area operators, and reliability coordinators
art-Atlanta 345-kV line tripped off-line due to a indicated that the morning of August 14 was a rea-
tree contact. This line had no direct electrical sonably typical day.1 FE managers referred to it as
effect on FE’s system—but it did affect MISO’s per- peak load conditions on a less than peak load day.
formance as reliability coordinator, even though Dispatchers consistently said that while voltages
PJM is the reliability coordinator for the DPL line. were low, they were consistent with historical
One of MISO’s primary system condition evalua- voltages.2 Throughout the morning and early
tion tools, its state estimator, was unable to assess afternoon of August 14, FE reported a growing
system conditions for most of the period between need for voltage support in the upper Midwest.
46 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
The FE reliability operator was concerned about asked to reduce their real power output to be able
low voltage conditions on the FE system as early to produce more reactive output. He called the
as 13:13 EDT. He asked for voltage support (i.e., Sammis plant at 13:13 EDT, West Lorain at 13:15
increased reactive power output) from FE’s inter- EDT, Eastlake at 13:16 EDT, made three calls to
connected generators. Plants were operating in unidentified plants between 13:20 EDT and 13:23
automatic voltage control mode (reacting to sys- EDT, a “Unit 9” at 13:24 EDT, and two more at
tem voltage conditions and needs rather than con- 13:26 EDT and 13:28 EDT.4 The operators worked
stant reactive power output). As directed in FE’s to get shunt capacitors at Avon that were out of
Manual of Operations,3 the FE reliability operator service restored to support voltage,5 but those
began to call plant operators to ask for additional capacitors could not be restored to service.
voltage support from their units. He noted to most
of them that system voltages were sagging “all Following the loss of Eastlake 5 at 13:31 EDT, FE’s
over.” Several mentioned that they were already at operators’ concern about voltage levels increased.
or near their reactive output limits. None were They called Bay Shore at 13:41 EDT and Perry at
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 47
Figure 5.2. Timeline Phase 1
13:43 EDT to ask the plants for more voltage sup- every 5 minutes, and the RTCA less frequently. If
port. Again, while there was substantial effort to the model does not have accurate and timely infor-
support voltages in the Ohio area, FirstEnergy per- mation about key pieces of system equipment or if
sonnel characterized the conditions as not being key input data are wrong, the state estimator may
unusual for a peak load day, although this was not be unable to reach a solution or it will reach a solu-
an all-time (or record) peak load day.6 tion that is labeled as having a high degree of error.
In August, MISO considered its SE and RTCA
Key Phase 1 Events tools to be still under development and not fully
mature; those systems have since been completed
1A) 12:15 EDT to 16:04 EDT: MISO’s state estima- and placed into full operation.
tor software solution was compromised, and
MISO’s single contingency reliability assess- On August 14 at about 12:15 EDT, MISO’s state
ment became unavailable. estimator produced a solution with a high mis-
match (outside the bounds of acceptable error).
1B) 13:31:34 EDT: Eastlake Unit 5 generation trip- This was traced to an outage of Cinergy’s
ped in northern Ohio. Bloomington-Denois Creek 230-kV line—
1C) 14:02 EDT: Stuart-Atlanta 345-kV transmis- although it was out of service, its status was not
sion line tripped in southern Ohio. updated in MISO’s state estimator. Line status
information within MISO’s reliability coordina-
1A) MISO’s State Estimator Was Turned Off:
tion area is transmitted to MISO by the ECAR data
12:15 EDT to 16:04 EDT
network or direct links and is intended to be auto-
It is common for reliability coordinators and con- matically linked to the SE. This requires coordi-
trol areas to use a state estimator (SE) to improve nated data naming as well as instructions that link
the accuracy of the raw sampled data they have for the data to the tools. For this line, the automatic
the electric system by mathematically processing linkage of line status to the state estimator had not
raw data to make it consistent with the electrical yet been established. The line status was corrected
system model. The resulting information on and MISO’s analyst obtained a good SE solution at
equipment voltages and loadings is used in soft- 13:00 EDT and an RTCA solution at 13:07 EDT.
ware tools such as real time contingency analysis However, to troubleshoot this problem the analyst
(RTCA) to simulate various conditions and out- had turned off the automatic trigger that runs the
ages to evaluate the reliability of the power sys- state estimator every five minutes. After fixing the
tem. The RTCA tool is used to alert operators if the problem he forgot to re-enable it, so although he
system is operating insecurely; it can be run either had successfully run the SE and RTCA manually
on a regular schedule (e.g., every 5 minutes), when to reach a set of correct system analyses, the tools
triggered by some system event (e.g., the loss of a were not returned to normal automatic operation.
power plant or transmission line), or when initi- Thinking the system had been successfully
ated by an operator. MISO usually runs the SE restored, the analyst went to lunch.
48 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Cause 4 The fact that the state estimator Eastlake 4 and Davis-Besse meant that FE had to
Inadequate was not running automatically on import more energy into the Cleveland-Akron area
RC Diagnostic its regular 5-minute schedule was to support its load.
Support discovered about 14:40 EDT. The
automatic trigger was re-enabled When Eastlake 5 dropped off-line, replacement
but again the state estimator failed to solve suc- power transfers and the associated reactive power
cessfully. This time investigation identified the to support the imports to the local area contrib-
Stuart-Atlanta 345-kV line outage (which uted to the additional line loadings in the region.
occurred at 14:02 EDT) to be the likely cause. This At 15:00 EDT on August 14, FE’s load was approxi-
line is within the Dayton Power and Light control mately 12,080 MW, and they were importing
area in southern Ohio and is under PJM’s reliabil- about 2,575 MW, 21% of their total. FE’s system
ity umbrella rather than MISO’s. Even though it reactive power needs rose further.
affects electrical flows within MISO, its status had Cause 1 The investigation team’s system
not been automatically linked to MISO’s state Inadequate simulations indicate that the loss
estimator. System of Eastlake 5 was a critical step in
Understanding the sequence of events. Contin-
The discrepancy between actual measured system
flows (with Stuart-Atlanta off-line) and the MISO gency analysis simulation of the
model (which assumed Stuart-Atlanta on-line) conditions following the loss of the Har-
prevented the state estimator from solving cor- ding-Chamberlin 345-kV circuit at 15:05 EDT
rectly. At 15:09 EDT, when informed by the sys- showed that the system would be unable to sus-
tem engineer that the Stuart-Atlanta line appeared tain some contingencies without line overloads
to be the problem, the MISO operator said (mistak- above emergency ratings. However, when Eastlake
enly) that this line was in service. The system 5 was modeled as in service and fully available in
engineer then tried unsuccessfully to reach a solu- those simulations, all overloads above emergency
tion with the Stuart-Atlanta line modeled as in limits were eliminated, even
with the loss of Harding- Recommendation
service until approximately 15:29 EDT, when the 23, page 160
MISO operator called PJM to verify the correct sta- Chamberlin.
tus. After they determined that Stuart-Atlanta had Cause 2 FE did not perform a contingency
tripped, they updated the state estimator and it Inadequate analysis after the loss of Eastlake
solved successfully. The RTCA was then run man- Situational 5 at 13:31 EDT to determine
ually and solved successfully at 15:41 EDT. Awareness
whether the loss of further lines
MISO’s state estimator and contingency analysis or plants would put their system
were back under full automatic operation and at risk. FE also did not perform a contingency anal-
solving effectively by 16:04 EDT, about two min- ysis after the loss of Harding-Chamberlin at 15:05
utes before the start of the cascade. EDT (in part because they did not know that it had
In summary, the MISO state estimator and real tripped out of service), nor does the utility rou-
time contingency analysis tools were effectively tinely conduct such studies.7 Thus FE did not dis-
cover that their system was no longer in an N-1
out of service between 12:15 EDT and 16:04 EDT.
This prevented MISO from promptly performing
Figure 5.3. Eastlake Unit 5
precontingency “early warning” assessments of
power system reliability Recommendations
over the afternoon of August 3, page 143; 6, page 147;
30, page 163
14.
1B) Eastlake Unit 5 Tripped: 13:31 EDT
Eastlake Unit 5 (rated at 597 MW) is in northern
Ohio along the southern shore of Lake Erie, con-
nected to FE’s 345-kV transmission system (Figure
5.3). The Cleveland and Akron loads are generally
supported by generation from a combination of
the Eastlake, Perry and Davis-Besse units, along
with significant imports, particularly from
9,100 MW of generation located along the Ohio
and Pennsylvania border. The unavailability of
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 49
secure state at 15:05 EDT, Recommendations effect on power flows and voltages in the FE area.
and that operator action was 3, page 143, 22, page 159 The team examined the security of FE’s system,
needed to remedy the testing power flows and voltage levels with the
situation. combination of plant and line outages that evolved
on the afternoon of August 14. This analysis
1C) Stuart-Atlanta 345-kV Line Tripped:
shows that the availability or unavailability of the
14:02 EDT
Stuart-Atlanta 345-kV line did not change the
The Stuart-Atlanta 345-kV trans- capability or performance of FE’s system or affect
Cause 1
Inadequate mission line is in the control area any line loadings within the FE system, either
System of Dayton Power and Light. At immediately after its trip or later that afternoon.
Understanding 14:02 EDT the line tripped due to The only reason why Stuart-Atlanta matters to the
contact with a tree, causing a blackout is because it contributed to the failure of
short circuit to ground, and locked out. Investiga- MISO’s state estimator to operate effectively, so
tion team modeling reveals that the loss of DPL’s MISO could not fully identify FE’s precarious sys-
Stuart-Atlanta line had no significant electrical tem conditions until 16:04 EDT.8
50 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Phase 2: that was hosting the alarm function, and then the
backup server such that all functions that were
FE’s Computer Failures: being supported on these servers were stopped at
14:14 EDT to 15:59 EDT 14:54 EDT. However, for over an hour no one in
FE’s control room grasped that their computer sys-
Overview of This Phase tems were not operating properly, even though
FE’s Information Technology support staff knew
Starting around 14:14 EDT, FE’s control room of the problems and were working to solve them,
operators lost the alarm function that provided and the absence of alarms and other symptoms
audible and visual indications when a significant offered many clues to the operators of the EMS
piece of equipment changed from an acceptable to system’s impaired state. Thus, without a function-
a problematic condition. Shortly thereafter, the ing EMS or the knowledge that it had failed, FE’s
EMS system lost a number of its remote control system operators remained unaware that their
consoles. Next it lost the primary server computer electrical system condition was beginning to
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 51
Figure 5.4. Timeline Phase 2
degrade. Unknowingly, they used the outdated after voltages had begun deteriorating but well
system condition information they did have to dis- before any of FE’s lines began to contact trees and
count information from others about growing sys- trip out. After that time, the FE control room con-
tem problems. soles did not receive any further alarms, nor were
there any alarms being printed or posted on the
Key Events in This Phase EMS’s alarm logging facilities. Power system oper-
2A) 14:14 EDT: FE alarm and logging software ators rely heavily on audible and on-screen
failed. Neither FE’s control room operators alarms, plus alarm logs, to reveal any significant
nor FE’s IT EMS support personnel were changes in their system’s conditions. After 14:14
aware of the alarm failure. EDT on August 14, FE’s operators were working
under a significant handicap without these tools.
2B) 14:20 EDT: Several FE remote EMS consoles However, they were in further jeopardy because
failed. FE’s Information Technology (IT) engi- they did not know that they were operating with-
neer was computer auto-paged. out alarms, so that they did not realize that system
2C) 14:27:16 EDT: Star-South Canton 345-kV conditions were changing.
transmission line tripped and successfully Alarms are a critical function of an EMS, and
reclosed. EMS-generated alarms are the fundamental means
2D) 14:32 EDT: AEP called FE control room about by which system operators identify events on the
AEP indication of Star-South Canton 345-kV power system that need their attention. Without
line trip and reclosure. FE had no alarm or log alarms, events indicating one or more significant
of this line trip. system changes can occur but remain undetected
by the operator. If an EMS’s alarms are absent, but
2E) 14:41 EDT: The primary FE control system operators are aware of the situation and the
server hosting the alarm function failed. Its remainder of the EMS’s functions are intact, the
applications and functions were passed over operators can potentially continue to use the EMS
to a backup computer. FE’s IT engineer was to monitor and exercise control of their power sys-
auto-paged. tem. In such circumstances, the operators would
have to do so via repetitive, continuous manual
2F) 14:54 EDT: The FE back-up computer failed
scanning of numerous data and status points
and all functions that were running on it
located within the multitude of individual dis-
stopped. FE’s IT engineer was auto-paged.
plays available within their EMS. Further, it
Failure of FE’s Alarm System would be difficult for the operator to identify
quickly the most relevant of the many screens
Cause 2 FE’s computer SCADA alarm and available.
Inadequate logging software failed sometime
Situational shortly after 14:14 EDT (the last In the same way that an alarm system can inform
Awareness
time that a valid alarm came in), operators about the failure of key grid facilities, it
52 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
can also be set up to alarm them if the alarm sys- receiving. Continuing on in the belief that their
tem itself fails to perform properly. FE’s EMS did system was satisfactory, lacking any alarms from
not have such a notification system. their EMS to the contrary, and without visualiza-
tion aids such as a dynamic map board or a projec-
Although the alarm processing function of FE’s
tion of system topology, FE control room operators
EMS failed, the remainder of that system generally
were subsequently surprised when they began
continued to collect valid real-time status infor-
receiving telephone calls from other locations and
mation and measurements about FE’s power sys-
information sources—MISO, AEP, PJM, and FE
tem, and continued to have supervisory control
field operations staff—who offered information on
over the FE system. The EMS also continued to
the status of FE’s transmission facilities that con-
send its normal and expected collection of infor-
flicted with FE’s system Recommendations
mation on to other monitoring points and authori-
operators’ understanding of 3, page 143, 22, page 159
ties, including MISO and AEP. Thus these entities
the situation.
continued to receive accurate information about
the status and condition of FE’s power system after Analysis of the alarm problem performed by FE
the time when FE’s EMS alarms failed. FE’s opera- suggests that the alarm process essentially
tors were unaware that in this situation they “stalled” while processing an alarm event, such
needed to manually and more closely monitor and that the process began to run in a manner that
interpret the SCADA information they were failed to complete the processing of that alarm or
Alarms
System operators must keep a close and constant “closed” (off or on) when they should be other-
watch on the multitude of things occurring wise, or whether they have changed condition
simultaneously on their power system. These since the last scan. These alarms should provide
include the system’s load, the generation and clear indication and notification to system opera-
supply resources to meet that load, available tors of whether a given device is doing what they
reserves, and measurements of critical power think it is, or what they want it to do—for
system states, such as the voltage levels on the instance, whether a given power line is con-
lines. Because it is not humanly possible to nected to the system and moving power at a par-
watch and understand all these events and con- ticular moment.
ditions simultaneously, Energy Management
Systems use alarms to bring relevant information EMS limit alarms are designed to provide an
to operators’ attention. The alarms draw on the indication to system operators when something
information collected by the SCADA real-time important that is measured on a power system
monitoring system. device—such as the voltage on a line or the
amount of power flowing across it—is below or
Alarms are designed to quickly and appropri-
above pre-specified limits for using that device
ately attract the power system operators’ atten-
safely and efficiently. When a limit alarm acti-
tion to events or developments of interest on the
vates, it provides an important early warning to
system. They do so using combinations of audi-
the power system operator that elements of the
ble and visual signals, such as sounds at opera-
system may need some adjustment to prevent
tors’ control desks and symbol or color changes
damage to the system or to customer loads—like
or animations on system monitors, displays, or
the “low fuel” or “high engine temperature”
map boards. EMS alarms for power systems are
warnings in a car.
similar to the indicator lights or warning bell
tones that a modern automobile uses to signal its When FE’s alarm system failed on August 14, its
driver, like the “door open” bell, an image of a operators were running a complex power system
headlight high beam, a “parking brake on” indi- without adequate indicators of when key ele-
cator, and the visual and audible alert when a gas ments of that system were reaching and passing
tank is almost empty. the limits of safe operation—and without aware-
Power systems, like cars, use “status” alarms and ness that they were running the system without
“limit” alarms. A status alarm indicates the state these alarms and should no longer assume that
of a monitored device. In power systems these not getting alarms meant that system conditions
are commonly used to indicate whether such were still safe and unchanging.
items as switches or breakers are “open” or
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 53
produce any other valid output (alarms). In the operators’ computer consoles. Thus at times on
meantime, new inputs—system condition data August 14th, operators’ screen refresh rates—the
that needed to be reviewed for possible rate at which new information and displays are
alarms—built up in and then overflowed the pro- painted onto the computer screen, normally 1 to 3
cess’ input buffers.9,10 seconds—slowed to as long as 59 seconds per
screen. Since FE operators have numerous infor-
Loss of Remote EMS Terminals. Between 14:20
mation screen options, and one or more screens
EDT and 14:25 EDT, some of FE’s remote EMS ter-
are commonly “nested” as sub-screens to one or
minals in substations ceased operation. FE has
more top level screens, operators’ ability to view,
advised the investigation team that it believes this
understand and operate their system through the
occurred because the data feeding into those ter-
EMS would have slowed to a frustrating crawl.13
minals started “queuing” and overloading the ter- This situation may have occurred between 14:54
minals’ buffers. FE’s system operators did not EDT and 15:08 EDT when both servers failed, and
learn about this failure until 14:36 EDT, when a again between 15:46 EDT and 15:59 EDT while
technician at one of the sites noticed the terminal FE’s IT personnel attempted to reboot both servers
was not working after he came in on the 15:00 to remedy the alarm problem.
shift, and called the main control room to report
the problem. As remote terminals failed, each trig- Loss of the first server caused an auto-page to be
gered an automatic page to FE’s Information Tech- issued to alert FE’s EMS IT support personnel to
nology (IT) staff.11 The investigation team has not the problem. When the back-up server failed, it
determined why some terminals failed whereas too sent an auto-page to FE’s IT staff. They did not
others did not. Transcripts indicate that data links notify control room operators of the problem. At
to the remote sites were down as well.12 15:08 EDT, IT staffers completed a “warm reboot”
(restart) of the primary server. Startup diagnostics
EMS Server Failures. FE’s EMS system includes
monitored during that reboot verified that the
several server nodes that perform the higher func-
computer and all expected processes were run-
tions of the EMS. Although any one of them can
ning; accordingly, FE’s IT staff believed that they
host all of the functions, FE’s normal system con-
had successfully restarted the node and all the
figuration is to have a number of host subsets of
processes it was hosting. However, although the
the applications, with one server remaining in a
server and its applications were again running, the
“hot-standby” mode as a backup to the others
alarm system remained frozen and non-func-
should any fail. At 14:41 EDT, the primary server
tional, even on the restarted computer. The IT staff
hosting the EMS alarm processing application
did not confirm that the
failed, due either to the stalling of the alarm appli-
alarm system was again Recommendation
cation, “queuing” to the remote EMS terminals, 19, page 156
working properly with the
or some combination of the two. Following pre-
control room operators.
programmed instructions, the alarm system appli-
cation and all other EMS software running on the Another casualty of the loss of both servers was
first server automatically transferred (“failed- the Automatic Generation Control (AGC) function
over”) onto the back-up server. However, because hosted on those computers. Loss of AGC meant
the alarm application moved intact onto the that FE’s operators could not run affiliated
backup while still stalled and ineffective, the power plants on pre-set programs to respond auto-
backup server failed 13 minutes later, at 14:54 matically to meet FE’s system load and inter-
EDT. Accordingly, all of the EMS applications on change obligations. Although the AGC did not
these two servers stopped Recommendation work from 14:54 EDT to 15:08 EDT and 15:46 EDT
running. 22, page 159 to 15:59 EDT (periods when both servers were
down), this loss of function
Cause 2 The concurrent loss of both EMS Recommendation
does not appear to have had 22, page 159
Inadequate servers apparently caused several
an effect on the blackout.
Situational new problems for FE’s EMS and
Awareness the operators who used it. Tests The concurrent loss of the EMS
Cause 2
run during FE’s after-the-fact Inadequate servers also caused the failure of
analysis of the alarm failure event indicate that a Situational FE’s strip chart function. There
concurrent absence of these servers can signifi- Awareness are many strip charts in the FE
cantly slow down the rate at which the EMS sys- Reliability Operator control room
tem puts new—or refreshes existing—displays on driven by the EMS computers, showing a variety
54 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
of system conditions, including raw ACE (Area EDT and 15:59 EDT, when the two servers were
Control Error), FE system load, and Sammis-South down. This meant that generators were not con-
Canton and South Canton-Star loading. These trolled during these periods to meet FE’s load and
charts are visible in the reliability operator control interchange obligations (except from 15:00 EDT to
room. The chart printers continued to scroll but 15:09 EDT when control was switched to a backup
because the underlying computer system was controller). There were no apparent negative con-
locked up the chart pens showed only the last sequences from this failure. It has not been estab-
valid measurement recorded, without any varia- lished how loss of the primary generation control
tion from that measurement as time progressed signal was identified or if any discussions
(i.e., the charts “flat-lined”). There is no indication occurred with respect to the computer system’s
that any operators noticed or reported the failed operational status.15
operation of the charts.14 The few charts fed by
EMS System History. The EMS in service at FE’s
direct analog telemetry, rather than the EMS sys-
Ohio control center is a GE Harris (now GE Net-
tem, showed primarily frequency data, and
work Systems) XA21 system. It was initially
remained available throughout the afternoon of
brought into service in 1995. Other than the appli-
August 14. These yield little useful system infor-
cation of minor software fixes or patches typically
mation for operational purposes.
encountered in the ongoing maintenance and sup-
FE’s Area Control Error (ACE), the primary control port of such a system, the last major updates or
signal used to adjust generators and imports to revisions to this EMS were implemented in 1998.
match load obligations, did not function between On August 14 the system was not running the
14:54 EDT and 15:08 EDT and later between 15:46 most current release of the XA21 software. FE had
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 55
decided well before August Recommendation
room staff either when they began work on the
14 to replace it with one 33, page 164 servers at 14:54 EDT, or when they completed the
from another vendor. primary server restart at 15:08 EDT. At 15:42 EDT,
the IT staff were first told of the alarm problem by
FE personnel told the investigation team that the
a control room operator; FE has stated to investiga-
alarm processing application had failed on occa-
tors that their IT staff had been unaware before
sions prior to August 14, leading to loss of the
then that the alarm processing sub-system of the
alarming of system conditions and events for FE’s
EMS was not working.
operators.16 However, FE said that the mode and
behavior of this particular failure event were both Without the EMS systems, the only remaining
first time occurrences and ones which, at the time, ways to monitor system conditions would have
FE’s IT personnel neither recognized nor knew been through telephone calls and direct analog
how to correct. FE staff told investigators that it telemetry. FE control room personnel did not real-
was only during a post-outage support call with ize that alarm processing on their EMS was not
GE late on 14 August that FE and GE determined working and, subsequently, did not monitor other
that the only available course of action to correct available telemetry.
the alarm problem was a “cold reboot”17 of FE’s
overall XA21 system. In interviews immediately Cause 2 During the afternoon of August
after the blackout, FE IT personnel indicated that Inadequate 14, FE operators talked to their
they discussed a cold reboot of the XA21 system
Situational field personnel, MISO, PJM (con-
Awareness cerning an adjoining system in
with control room operators after they were told of
the alarm problem at 15:42 EDT, but decided not PJM’s reliability coordination
to take such action because operators considered region), adjoining systems (such as AEP), and cus-
power system conditions precarious, were con- tomers. The FE operators received pertinent infor-
cerned about the length of time that the reboot mation from all these sources, but did not
might take to complete, and understood that a cold recognize the emerging problems from the clues
reboot would leave them with even less EMS func- offered. This pertinent information included calls
tionality until it was completed.18 such as that from FE’s eastern control center ask-
ing about possible line trips, FE Perry nuclear
Clues to the EMS Problems. There is an entry in plant calls regarding what looked like nearby line
FE’s western desk operator’s log at 14:14 EDT trips, AEP calling about their end of the Star-South
referring to the loss of alarms, but it is not clear Canton line tripping, and Recommendations
whether that entry was made at that time or subse- MISO and PJM calling about 19, page 156; 26, page 161
quently, referring back to the last known alarm. possible line overloads.
There is no indication that the operator mentioned
the problem to other control Without a functioning alarm system, the FE con-
room staff and supervisors
Recommendation trol area operators failed to detect the tripping of
26, page 161
or to FE’s IT staff. electrical facilities essential to maintain the secu-
rity of their control area. Unaware of the loss of
The first clear hint to FE control room staff of any alarms and a limited EMS, they made no alternate
computer problems occurred at 14:19 EDT when a arrangements to monitor the system. When AEP
caller and an FE control room operator discussed identified the 14:27 EDT circuit trip and reclosure
the fact that three sub-transmission center of the Star 345 kV line circuit breakers at AEP’s
dial-ups had failed.19 At 14:25 EDT, a control South Canton substation, the FE operator dis-
room operator talked with a caller about the fail- missed the information as either not accurate or
ure of these three remote EMS consoles.20 The not relevant to his system, without following up
next hint came at 14:32 EDT, when FE scheduling on the discrepancy between the AEP event and the
staff spoke about having made schedule changes information from his own tools. There was no sub-
to update the EMS pages, but that the totals did sequent verification of conditions with the MISO
not update.21 reliability coordinator.
Cause 2 Although FE’s IT staff would have Only after AEP notified FE that a 345-kV circuit
Inadequate been aware that concurrent loss had tripped and locked out did the FE control
Situational of its servers would mean the loss area operator compare this information to
Awareness of alarm processing on the EMS, actual breaker conditions. FE failed to inform its
the investigation team has found reliability coordinator and adjacent control areas
no indication that the IT staff informed the control when they became aware that system conditions
56 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
had changed due to un- Recommendations 3D) 15:35 EDT: AEP asked PJM to begin work on a
scheduled equipment out- 26, page 161; 30, page 163 350-MW TLR to relieve overloading on the
ages that might affect other Star-South Canton line, not knowing the
control areas. Hanna-Juniper 345-kV line had already trip-
ped at 15:32 EDT.
Phase 3: 3E) 15:36 EDT: MISO called FE regarding
Three FE 345-kV post-contingency overload on Star-Juniper
345-kV line for the contingency loss of the
Transmission Line Failures Hanna-Juniper 345-kV line, unaware at the
and Many Phone Calls: start of the call that Hanna-Juniper had
15:05 EDT to 15:57 EDT already tripped.
3F) 15:41:33-41 EDT: Star-South Canton 345-kV
Overview of This Phase tripped, reclosed, tripped again at 15:41:35
From 15:05:41 EDT to 15:41:35 EDT, three 345-kV EDT and remained out of service, all while
lines failed with power flows at or below each AEP and PJM were discussing TLR relief
transmission line’s emergency rating. These line options (event 3D).
trips were not random. Rather, each was the result Transmission lines are designed with the expecta-
of a contact between a line and a tree that tion that they will sag lower when they become
had grown so tall that, over a period of years, it hotter. The transmission line gets hotter with
encroached into the required clearance height for heavier line loading and under higher ambient
the line. As each line failed, its outage increased temperatures, so towers and conductors are
the loading on the remaining lines (Figure 5.5). As designed to be tall enough and conductors pulled
each of the transmission lines failed, and power tightly enough to accommodate expected sagging
flows shifted to other transmission paths, voltages and still meet safety requirements. On a summer
on the rest of FE’s system degraded further (Figure day, conductor temperatures can rise from 60°C
5.6). on mornings with average wind to 100°C with hot
air temperatures and low wind conditions.
Key Phase 3 Events
A short-circuit occurred on the Harding-
3A) 15:05:41 EDT: Harding-Chamberlin 345-kV Chamberlin 345-kV line due to a contact between
line tripped. the line conductor and a tree. This line failed with
3B) 15:31-33 EDT: MISO called PJM to determine power flow at only 44% of its normal and emer-
if PJM had seen the Stuart-Atlanta 345-kV gency line rating. Incremental line current and
line outage. PJM confirmed Stuart-Atlanta temperature increases, escalated by the loss of
was out. Harding-Chamberlin, caused more sag on the
Hanna-Juniper line, which contacted a tree and
3C) 15:32:03 EDT: Hanna-Juniper 345-kV line failed with power flow at 88% of its normal
tripped. and emergency line rating. Star-South Canton
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 57
Figure 5.7. Timeline Phase 3
contacted a tree three times between 14:27:15 EDT Like most transmission owners, FE patrols its lines
and 15:41:33 EDT, opening and reclosing each regularly, flying over each transmission line twice
time before finally locking out while loaded at a year to check on the condition of the
93% of its emergency rating at 15:41:35 EDT. Each rights-of-way. Notes from fly-overs in 2001 and
of these three lines tripped not because of exces- 2002 indicate that the examiners saw a significant
sive sag due to overloading or high conductor tem- number of trees and brush that needed clearing or
perature, but because it hit an overgrown,
untrimmed tree.22 Line Ratings
Cause 3 Overgrown trees, as opposed to A conductor’s normal rating reflects how
Inadequate excessive conductor sag, caused heavily the line can be loaded under routine
Tree each of these faults. While sag operation and keep its internal temperature
Trimming may have contributed to these below a certain temperature (such as 90°C). A
events, these incidents occurred conductor’s emergency rating is often set to
because the trees grew too tall and encroached allow higher-than-normal power flows, but to
into the space below the line which is intended limit its internal temperature to a maximum
to be clear of any objects, not because the lines temperature (such as 100°C) for no longer than a
sagged into short trees. Because the trees were so specified period, so that it does not sag too low
tall (as discussed below), each of these lines or cause excessive damage to the conductor.
faulted under system conditions well within spec-
ified operating parameters. The investigation team For three of the four 345-kV lines that failed,
found field evidence of tree contact at all three FE set the normal and emergency ratings at the
locations, including human observation of the same level. Many of FE’s lines are limited by the
Hanna-Juniper contact. Evidence outlined below maximum temperature capability of its termi-
confirms that contact with trees caused the short nal equipment, rather than by the maximum
circuits to ground that caused each line to trip out safe temperature for its conductors. In calculat-
on August 14. ing summer emergency ampacity ratings for
many of its lines, FE assumed 90°F (32°C) ambi-
To be sure that the evidence of tree/line contacts ent air temperatures and 6.3 ft/sec (1.9 m/sec)
and tree remains found at each site was linked to wind speed,a which is a relatively high wind
the events of August 14, the team looked at speed assumption for favorable wind cooling.
whether these lines had any prior history of out- Actual temperature on August 14 was 87°F
ages in preceding months or years that might have (31°C) but wind speed at certain locations in the
resulted in the burn marks, debarking, and other Akron area was somewhere between 0 and 2
vegetative evidence of line contacts. The record ft/sec (0.6 m/sec) after 15:00 EDT that afternoon.
establishes that there were no prior sustained out-
aFirstEnergy Transmission Planning Criteria (Revision 8),
ages known to be caused by trees for these lines in
page 3.
2001, 2002, and 2003.23
58 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
trimming along many FE transmission lines. Notes 3A) FE’s Harding-Chamberlin 345-kV Line
from fly-overs in the spring of 2003 found fewer Tripped: 15:05 EDT
problems, suggesting that fly-overs do not allow At 15:05:41 EDT, FE’s Harding-Chamberlin line
effective identification of the distance between a (Figure 5.8) tripped and locked out while loaded at
tree and the line above it, 44% of its normal and emergency rating. At this
and need to be supple- Recommendations
16, page 154; 27, page 162 low loading, the line temperature would not
mented with ground patrols. exceed safe levels—even if still air meant there
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 59
was no wind cooling of the con- Figure 5.8. Harding-Chamberlin 345-kV Line
Cause 3
Inadequate ductor—and the line would not
Tree sag excessively. The investigation
Trimming team examined the relay data for
this trip, identified the geo-
graphic location of the fault, and determined that
the relay data match the classic “signature” pat-
tern for a tree/line short circuit to ground fault.
The field team found the remains of trees and
brush at the fault location determined from the
relay data. At this location, conductor height mea-
sured 46 feet 7 inches (14.20 meters), while the
height of the felled tree measured 42 feet (12.80
meters); however, portions of the tree had been
removed from the site. This means that while it is
difficult to determine the exact height of the line
contact, the measured height is a minimum and SCADA input of the line’s status change, this was
the actual contact was likely 3 to 4 feet (0.9 to 1.2 presented to MISO operators as breaker status
meters) higher than estimated here. Burn marks changes rather than a line failure. Because their
were observed 35 feet 8 inches (10.87 meters) up EMS system topology processor had not yet been
the tree, and the crown of this tree was at least 6 linked to recognize line failures, it did not connect
feet (1.83 meters) taller than the observed burn the breaker information to the loss of a transmis-
marks. The tree showed evi- sion line. Thus, MISO’s operators did not recog-
Recommendations
dence of fault current dam- 16, page 154; 27, page 162 nize the Harding-Chamberlin trip as a significant
age.24 contingency event and could not advise FE regard-
When the Harding-Chamberlin line locked out, ing the event or its consequences. Further, with-
the loss of this 345-kV path caused the remaining out its state estimator and associated contingency
three southern 345-kV lines into Cleveland to pick analyses, MISO was unable to identify potential
up more load, with Hanna-Juniper picking up overloads that would occur due to various line or
the most. The Harding-Chamberlin outage also equipment outages. Accordingly, when the Har-
caused more power to flow through the underly- ding-Chamberlin 345-kV line tripped at 15:05
ing 138-kV system. EDT, the state estimator did not produce results
and could not predict an
Cause 2 MISO did not discover that Har- overload if the Hanna- Recommendation
Inadequate ding-Chamberlin had tripped Juniper 345-kV line were to 30, page 163
60 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
height (not near lines). Other sites at this location Figure 5.9. Hanna-Juniper 345-kV Line
had numerous (at least 20) trees in this right-
of-way.
Hanna-Juniper was loaded at 88% of its normal
and emergency rating when it tripped. With this
line open, over 1,200 MVA of power flow had to
find a new path to reach its load in Cleveland.
Loading on the remaining two 345-kV lines
increased, with Star-Juniper taking the bulk of the
power. This caused Star-South Canton’s loading
to rise above its normal but within its emergency
rating and pushed more power onto the 138-kV
system. Flows west into Michigan decreased
slightly and voltages declined somewhat in the
Cleveland area.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 61
3D) AEP and PJM Begin Arranging a TLR for developing a 350 MW TLR to mitigate it. The TLR
Star-South Canton: 15:35 EDT was to relieve the actual overload above normal
rating then occurring on Star-South Canton, and
Cause 4 Because its alarm system was not
prevent an overload above emergency rating on
Inadequate working, FE was not aware of the
RC Diagnostic Harding-Chamberlin or Hanna-
Support Juniper line trips. However, once Figure 5.10. Cause of the Hanna-Juniper Line Loss
MISO manually updated the state
estimator model for the Stuart-Atlanta 345-kV line
outage, the software successfully completed a
state estimation and contingency analysis at 15:41
EDT. But this left a 36 minute period, from 15:05
EDT to 15:41 EDT, during which MISO did not
recognize the consequences of the Hanna-Juniper
loss, and FE operators knew neither of the line’s
loss nor its consequences. PJM and AEP recog-
nized the overload on Star-South Canton, but had
not expected it because their earlier contingency
analysis did not examine enough lines within the
FE system to foresee this result of the Hanna-
Juniper contingency on top of the Harding-
Chamberlin outage. This August 14 photo shows the tree that caused the loss of
the Hanna-Juniper line (tallest tree in photo). Other 345-kV
After AEP recognized the Star-South Canton over- conductors and shield wires can be seen in the background.
load, at 15:35 EDT AEP asked PJM to begin Photo by Nelson Tree.
62 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
that line if the Sammis-Star line were to fail. But were not working properly,
when they began working on the TLR, neither AEP MISO’s ability to recognize Recommendations
22, page 159; 30, page 163
nor PJM realized that the Hanna-Juniper 345-kV FE’s evolving contingency
line had already tripped at 15:32 EDT, further situation was impaired.
degrading system conditions. Since the great
majority of TLRs are for cuts of 25 to 50 MW, a 350 3F) Loss of the Star-South Canton 345-kV Line:
MW TLR request was highly unusual and opera- 15:41 EDT
tors were attempting to confirm why so much The Star-South Canton line (Figure 5.11) crosses
relief was suddenly required before implementing the boundary between FE and AEP—each com-
the requested TLR. Less than ten minutes elapsed pany owns the portion of the line and manages the
between the loss of Hanna-Juniper, the overload right-of-way within its respective territory. The
above the normal limits of Star-South Canton line tripped and reclosed three
Recommendations
Star-South Canton, and the 6, page 147; 22, page 159; times on the afternoon of August 14, first at
Star-South Canton trip and 30, page 163; 31, page 163 14:27:15 EDT while carrying less than 55% of its
lock-out. emergency rating (reclosing at both ends), then at
15:38:48 and again at 15:41:33 EDT. These multi-
Cause 2 Unfortunately, neither AEP nor
ple contacts had the effect of “electric
Inadequate PJM recognized that even a 350
tree-trimming,” burning back the contacting limbs
Situational MW TLR on the Star-South Can-
Awareness temporarily and allowing the line to carry more
ton line would have had little
current until further sag in the still air caused the
impact on the overload. Investi-
final contact and lock-out. At 15:41:35 EDT the
gation team analysis using the Interchange Distri-
line tripped and locked out at the Star substation,
bution Calculator (which was fully available on
with power flow at 93% of its emergency rating. A
the afternoon of August 14) indicates that tagged
short-circuit to ground occurred in each case.
transactions for the 15:00 EDT hour across Ohio
had minimal impact on the overloaded lines. As Cause 3 The investigation’s field team
discussed in Chapter 4, this analysis showed that Inadequate inspected the right of way in the
after the loss of the Hanna-Juniper 345 kV line, Tree location indicated by the relay
Star-South Canton was loaded primarily with Trimming digital fault recorders, in the FE
flows to serve native and network loads, deliver- portion of the line. They found
ing makeup energy for the loss of Eastlake 5, pur- debris from trees and vegetation that had been
chased from PJM (342 MW) and Ameren (126 felled. At this location the conductor height
MW). The only way that these high loadings could was 44 feet 9 inches (13.6 meters). The identifiable
have been relieved would not have been from the tree remains measured 30 feet (9.1 meters) in
redispatch that AEP requested, but rather from sig- height, although the team could not verify the
nificant load-shedding by FE in the Cleveland location of the stump, nor find all sections of the
area. tree. A nearby cluster of trees showed significant
fault damage, including charred limbs and
Cause 4 The primary tool MISO uses for
de-barking from fault current. Further, topsoil in
Inadequate assessing reliability on key
RC Diagnostic flowgates (specified groupings of
Support transmission lines or equipment Figure 5.11. Star-South Canton 345-kV Line
that sometimes have less transfer
capability than desired) is the flowgate monitoring
tool. After the Harding-Chamberlin 345-kV line
outage at 15:05 EDT, the flowgate monitoring tool
produced incorrect (obsolete) results, because the
outage was not reflected in the model. As a result,
the tool assumed that Harding-Chamberlin was
still available and did not predict an overload for
loss of the Hanna-Juniper 345-kV line. When
Hanna-Juniper tripped at 15:32 EDT, the resulting
overload was detected by MISO’s SCADA and set
off alarms to MISO’s system operators, who then
phoned FE about it.27 Because both MISO’s
state estimator and its flowgate monitoring tool
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 63
the area of the tree trunk was disturbed, discolored Cause 2 System impacts of the 345-kV
and broken up, a common indication of a higher Inadequate failures. According to extensive
magnitude fault or multiple faults. Analysis of Situational investigation team modeling,
another stump showed that a fourteen year-old Awareness there were no contingency limit
tree had recently been violations as of 15:05 EDT before
Recommendations
removed from the middle of 16, page 154; 27, page 162
the loss of the Harding-Chamberlin 345-kV line.
the right-of-way.28 Figure 5.12 shows the line loadings estimated by
investigation team modeling as the 345-kV lines in
After the Star-South Canton line was lost, flows
northeast Ohio began to trip. Showing line load-
increased greatly on the 138-kV system toward
ings on the 345-kV lines as a percent of normal rat-
Cleveland and area voltage levels began to degrade
ing, it tracks how the loading on each line
on the 138-kV and 69-kV system. At the same
increased as each subsequent 345-kV and 138-kV
time, power flows increased on the Sammis-Star
line tripped out of service between 15:05 EDT
345-kV line due to the 138-kV line trips—the only
(Harding-Chamberlin, the first line above to
remaining paths into Cleveland from the south.
stair-step down) and 16:06 EDT (Dale-West Can-
Cause 2 FE’s operators were not aware that ton). As the graph shows, none of the 345- or
Inadequate the system was operating outside 138-kV lines exceeded their normal ratings until
Situational first contingency limits after the after the combined trips of Harding-Chamberlin
Awareness and Hanna-Juniper. But immediately after the sec-
Harding-Chamberlin trip (for the
possible loss of Hanna-Juniper or ond line was lost, Star-South Canton’s loading
the Perry unit), because they did not conduct jumped from an estimated 82% of normal to 120%
a contingency analysis.29 The investigation team of normal (which was still below its emergency
has not determined whether the system status rating) and remained at the 120% level for 10 min-
information used by FE’s utes before tripping out. To the right, the graph
state estimator and contin- Recommendation
22, page159
shows the effects of the 138-kV line failures
gency analysis model was (discussed in the next phase) upon the
being accurately updated. two remaining 345-kV lines—i.e., Sammis-Star’s
loading increased steadily above 100% with each
Cause 1 Load-Shed Analysis. The investi- succeeding 138-kV line lost.
Inadequate gation team looked at whether it
System would have been possible to pre- Following the loss of the Harding-Chamberlin
Understanding vent the blackout by shedding 345-kV line at 15:05 EDT, contingency limit viola-
load within the Cleveland-Akron tions existed for:
area before the Star-South Canton 345 kV line trip- u The Star-Juniper 345-kV line, whose loadings
ped at 15:41 EDT. The team modeled the system would exceed emergency limits if the Hanna-
assuming 500 MW of load shed within the Cleve- Juniper 345-kV line were lost; and
land-Akron area before 15:41 EDT and found that
this would have improved voltage at the Star bus
from 91.7% up to 95.6%, pulling the line loading
from 91 to 87% of its emergency ampere rating; an Figure 5.12. Cumulative Effects of Sequential
Outages on Remaining 345-kV Lines
additional 500 MW of load would have had to be
dropped to improve Star voltage to 96.6% and the
line loading to 81% of its emergency ampere rat-
ing. But since the Star-South Canton line had
already been compromised by the tree below it
(which caused the first two trips and reclosures),
and was about to trip from tree contact a third
time, it is not clear that had such load shedding
occurred, it would have prevented the ultimate
trip and lock-out of the line. However, modeling
indicates that this load shed
would have prevented the Recommendations
8, page 147; 21, page 158
subsequent tripping of the
Sammis-Star line (see page
70).
64 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
u The Hanna-Juniper and Harding-Juniper call at 14:51 EDT, expressing concern that they
345-kV lines, whose loadings would exceed had not seen any indication of an operation, but
emergency limits if the Perry generation unit lacking evidence within their control room, the FE
(1,255 MW) were lost. operators did not pursue the issue.
Operationally, once FE’s system entered an N-1 At 15:19 EDT, AEP called FE back to confirm that
contingency violation state, any facility loss the Star-South Canton trip had occurred and that
beyond that pushed them farther into violation AEP had a confirmed relay operation from the site.
and into a more unreliable state. After loss of the FE’s operator restated that because they had
Harding-Chamberlin line, to avoid violating NERC received no trouble or alarms, they saw no prob-
criteria, FE needed to reduce loading on these lem. An AEP technician at the South Canton sub-
three lines within 30 minutes such that no single station verified the trip. At 15:20 EDT, AEP
contingency would violate an emergency limit; decided to treat the South Canton digital fault
that is, to restore the system to a reliable operating recorder and relay target information as a “fluke,”
mode. and checked the carrier relays to determine what
the problem might be.31
Phone Calls into the FE Control Room
At 15:35 EDT the FE control center received a call
Cause 2 Beginning at 14:14 EDT when
from the Mansfield 2 plant operator concerned
Inadequate their EMS alarms failed, and until
about generator fault recorder triggers and excita-
Situational at least 15:42 EDT when they
Awareness tion voltage spikes with an alarm for
began to recognize their situation,
over-excitation, and a dispatcher called reporting
FE operators did not understand
a “bump” on their system. Soon after this call, FE’s
how much of their system was being lost, and did
Reading, Pennsylvania control center called
not realize the degree to which their perception of
reporting that fault recorders in the Erie west and
their system was in error versus true system con-
south areas had activated, wondering if something
ditions, despite receiving clues via phone calls
had happened in the Ashtabula-Perry area. The
from AEP, PJM and MISO, and customers. The FE
Perry nuclear plant operator called to report a
operators were not aware of line outages that
“spike” on the unit’s main transformer. When he
occurred after the trip of Eastlake 5 at 13:31 EDT
went to look at the metering it was “still bouncing
until approximately 15:45 EDT, although they
around pretty good. I’ve got it relay tripped up
were beginning to get external input describing
here . . . so I know something ain’t right.”32
aspects of the system’s weakening condition.
Since FE’s operators were not aware and did not Beginning at this time, the FE operators began to
recognize events as they think that something was wrong, but did not rec-
were occurring, they took Recommendations ognize that it was on their system. “It’s got to be in
19, page 156; 26, page 161
no actions to return the sys- distribution, or something like that, or somebody
tem to a reliable state. else’s problem . . . but I’m not showing any-
thing.”33 Unlike many other transmission grid
A brief description follows of some of the calls FE
control rooms, FE’s control center did not have a
operators received concerning system problems
map board (which shows schematically all major
and their failure to recognize that the problem was
lines and plants in the control area on the wall in
on their system. For ease of presentation, this set
front of the operators), which might have shown
of calls extends past the time of the 345-kV line
the location of significant
trips into the time covered in the next phase, when Recommendation
line and facility outages 22, page 159
the 138-kV system collapsed.
within the control area.
Following the first trip of the Star-South Canton
At 15:36 EDT, MISO contacted FE regarding the
345-kV line at 14:27 EDT, AEP called FE at 14:32
post-contingency overload on Star-Juniper for the
EDT to discuss the trip and reclose of the line. AEP
loss of the Hanna-Juniper 345-kV line.34
was aware of breaker operations at their end
(South Canton) and asked about operations at FE’s At 15:42 EDT, FE’s western transmission operator
Star end. FE indicated they had seen nothing at informed FE’s IT staff that the EMS system func-
their end of the line, but AEP reiterated that the tionality was compromised. “Nothing seems to be
trip occurred at 14:27 EDT and that the South Can- updating on the computers . . . . We’ve had people
ton breakers had reclosed successfully.30 There calling and reporting trips and nothing seems to be
was an internal FE conversation about the AEP updating in the event summary . . . I think we’ve
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 65
got something seriously sick.” This is the first evi- Emergency Action
dence that a member of FE’s control room staff rec-
For FirstEnergy, as with many utilities, emergency
ognized any aspect of their degraded EMS system.
awareness is often focused on energy shortages.
There is no indication that he informed any of the
Utilities have plans to reduce loads under these
other operators at this moment. However, FE’s IT
circumstances to increasingly greater degrees.
staff discussed the subsequent EMS alarm correc-
Tools include calling for contracted customer load
tive action with some control room staff shortly
reductions, then public appeals, voltage reduc-
thereafter.
tions, and finally shedding system load by cutting
Also at 15:42 EDT, the Perry plant operator called off interruptible and firm customers. FE has a plan
back with more evidence of problems. “I’m still for this that is updated yearly. While they can trip
getting a lot of voltage spikes and swings on the loads quickly where there is SCADA control of
generator . . . . I don’t know how much longer load breakers (although FE has few of these), from
we’re going to survive.”35 an energy point of view, the intent is to be able to
At 15:45 EDT, the tree trimming crew reported regularly rotate what loads are not being served,
that they had witnessed a tree-caused fault on the which requires calling personnel out to switch the
Eastlake-Juniper 345-kV line; however, the actual various groupings in and out. This event was not,
fault was on the Hanna-Juniper 345-kV line in the however, a capacity or energy emergency or sys-
same vicinity. This information added to the con- tem instability, but an emergency due to transmis-
fusion in the FE control room, because the opera- sion line overloads.
tor had indication of flow on the Eastlake-Juniper To handle an emergency effectively a dispatcher
line.36 must first identify the emergency situation and
After the Star-South Canton 345-kV line tripped a then determine effective action. AEP identified
third time and locked out at 15:41:35 EDT, AEP potential contingency overloads at 15:36 EDT and
called FE at 15:45 EDT to discuss and inform them called PJM even as Star-South Canton, one of the
that they had additional lines that showed over- AEP/FE lines they were discussing, tripped and
load. FE recognized then that the Star breakers pushed FE’s Sammis-Star 345-kV line to its emer-
had tripped and remained open.37 gency rating. Since they had been focused on the
At 15:46 EDT the Perry plant operator called the impact of a Sammis-Star loss overloading Star-
FE control room a third time to say that the unit South Canton, they recognized that a serious prob-
was close to tripping off: “It’s not looking good . . . . lem had arisen on the system for which they did
We ain’t going to be here much longer and you’re not have a ready solution. Later, around 15:50
going to have a bigger problem.”38 EDT, their conversation reflected emergency con-
ditions (138-kV lines were tripping and several
At 15:48 EDT, an FE transmission operator sent other lines overloaded) but they still found no
staff to man the Star substation, and then at 15:50 practical way to mitigate
EDT, requested staffing at the regions, beginning these overloads across util- Recommendation
20, page 158
with Beaver, then East Springfield.39 ity and reliability coordina-
At 15:48 EDT, PJM called MISO to report the tor boundaries.
Star-South Canton trip, but the two reliability
Cause 2 At the control area level, FE
coordinators’ measures of the resulting line flows
Inadequate remained unaware of the precari-
on FE’s Sammis-Star 345-kV line did not match, Situational ous condition its system was in,
causing them to wonder whether the Star-South Awareness with key lines out of service,
Canton 345-kV line had returned to service.40
degrading voltages, and severe
At 15:56 EDT, because PJM was still concerned overloads on their remaining lines. Transcripts
about the impact of the Star-South Canton trip, show that FE operators were aware of falling volt-
PJM called FE to report that Star-South Canton ages and customer problems after loss of the
had tripped and that PJM thought FE’s Hanna-Juniper 345-kV line (at 15:32 EDT). They
Sammis-Star line was in actual emergency limit called out personnel to staff substations because
overload.41 FE could not confirm this overload. FE they did not think they could see them with their
informed PJM that Hanna-Juniper was also out data gathering tools. They were also talking to cus-
service. FE believed that the problems existed tomers. But there is no indication that FE’s opera-
beyond their system. “AEP must have lost some tors clearly identified their situation as a possible
major stuff.”42 emergency until around 15:45 EDT when the shift
66 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
supervisor informed his manager that it looked as instructions to adjust generators and trip firm
if they were losing the system; even then, although loads. After the loss of the Star-South Canton line,
FE had grasped that its system was in trouble, it voltages were below limits, and there were severe
never officially declared that it was an emergency line overloads. But FE did not follow any of these
condition and that emergency or extraordinary procedures on August 14, because FE did not
action was needed. know for most of that time that its system might
need such treatment.
FE’s internal control room procedures and proto-
cols did not prepare it adequately to identify and What training did the operators and reliability
react to the August 14 emergency. Throughout the coordinators have for recognizing and responding
afternoon of August 14 there were many clues that to emergencies? FE relied upon on-the-job experi-
FE had lost both its critical monitoring alarm func- ence as training for its operators in handling the
tionality and that its transmission system’s reli- routine business of a normal day, but had never
ability was becoming progressively more experienced a major disturbance and had no simu-
compromised. However, FE did not fully piece lator training or formal preparation for recogniz-
these clues together until after it had already lost ing and responding to emergencies. Although all
critical elements of its transmission system and affected FE and MISO operators were NERC-
only minutes before subsequent trips triggered the certified, NERC certification of operators
cascade phase of the blackout. The clues to a com- addresses basic operational considerations but
promised EMS alarm system and transmission offers little insight into emergency operations
system came into the FE control room from FE issues. Neither group of operators had significant
customers, generators, AEP, MISO, and PJM. In training, documentation, or actual experience for
spite of these clues, because how to handle an emer- Recommendation
of a number of related fac- Recommendations gency of this type and 20, page 158
20, page 158; 22, page 159;
tors, FE failed to identify 26, page 161 magnitude.
the emergency that it faced. MISO was hindered because it
Cause 4
The most critical factor delaying Inadequate lacked clear visibility, responsi-
Cause 2
Inadequate the assessment and synthesis of RC Diagnostic bility, authority, and ability to
Support take the actions needed in this cir-
Situational the clues was a lack of informa-
Awareness tion sharing between the FE sys- cumstance. MISO had interpre-
tem operators. In interviews with tive and operational tools and a large amount of
the FE operators and analysis of phone transcripts, system data, but had a limited view of FE’s system.
it is evident that rarely were any of the critical In MISO’s function as FE’s reliability coordinator,
clues shared with fellow operators. This lack of its primary task was to initiate and implement
information sharing can be TLRs, recognize and solve congestion problems in
Recommendation
attributed to: 26, page 161 less dramatic reliability circumstances with
longer solution time periods than those which
1. Physical separation of operators (the reliability existed on August 14, and provide assistance as
operator responsible for voltage schedules was requested.
across the hall from the transmission
Throughout August 14, most major elements of
operators).
FE’s EMS were working properly. The system was
2. The lack of a shared electronic log (visible to automatically transferring accurate real-time
all), as compared to FE’s practice of separate information about FE’s system conditions to com-
hand-written logs.43 puters at AEP, MISO, and PJM. FE’s operators did
not believe the transmission line failures reported
3. Lack of systematic procedures to brief incoming by AEP and MISO were real until 15:42 EDT, after
staff at shift change times. FE conversations with the AEP and MISO control
4. Infrequent training of operators in emergency rooms and calls from FE IT staff to report the fail-
scenarios, identification and resolution of bad ure of their alarms. At that point in time, FE opera-
data, and the importance of sharing key infor- tors began to think that their system might be in
mation throughout the control room. jeopardy—but they did not act to restore any of the
lost transmission lines, clearly alert their reliabil-
FE has specific written procedures and plans for ity coordinator or neighbors about their situation,
dealing with resource deficiencies, voltage or take other possible remedial measures (such as
depressions, and overloads, and these include load- shedding) to stabilize their system.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 67
Figure 5.13. Timeline Phase 4
68 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
which then opened the line to FE’s Cloverdale 4A-H) Pleasant Valley to Chamberlin-West
station. Akron Line Outages
4G) 15:51:41 EDT: East Lima-N. Findlay 138-kV From 15:39 EDT to 15:58:47 EDT, seven 138-kV
line tripped, likely due to sagging line, and lines in northern Ohio tripped and locked out. At
reclosed at East Lima end only. 15:45:41 EDT, Canton Central-Tidd 345-kV line
tripped and reclosed at 15:46:29 EDT because
4H) 15:58:47 EDT: Chamberlin-West Akron 138- Canton Central 345/138-kV CB “A1” operated
kV line tripped. multiple times, causing a low air pressure problem
Note: 15:51:41 EDT: Fostoria Central-N. that inhibited circuit breaker tripping. This event
Findlay 138-kV line tripped and reclosed, but forced the Canton Central 345/138-kV transform-
never locked out. ers to disconnect and remain out of service, fur-
ther weakening the Canton-Akron area 138-kV
At 15:59:00 EDT, the loss of the West Akron bus transmission system. At 15:58:47 EDT the
tripped due to breaker failure, causing another Chamberlin-West Akron 138-kV line tripped.
five 138-kV lines to trip:
4I-N) West Akron Transformer Circuit Breaker
4I) 15:59:00 EDT: West Akron 138-kV bus trip- Failure and Line Outages
ped, and cleared bus section circuit breakers
at West Akron 138 kV. At 15:59 EDT FE’s West Akron 138-kV bus tripped
due to a circuit breaker failure on West Akron
4J) 15:59:00 EDT: West Akron-Aetna 138-kV line transformer #1. This caused the five remaining
opened. 138-kV lines connected to the West Akron substa-
4K) 15:59:00 EDT: Barberton 138-kV line opened tion to open. The West Akron 138/12-kV trans-
at West Akron end only. West Akron-B18 formers remained connected to the Barberton-
138-kV tie breaker opened, affecting West West Akron 138-kV line, but power flow to West
Akron 138/12-kV transformers #3, 4 and 5 fed Akron 138/69-kV transformer #1 was interrupted.
from Barberton. 4O-P) Dale-West Canton 138-kV and
4L) 15:59:00 EDT: West Akron-Granger-Stoney- Sammis-Star 345-kV Lines Tripped
Brunswick-West Medina opened. After the Cloverdale-Torrey line failed at 15:42
EDT, Dale-West Canton was the most heavily
4M) 15:59:00 EDT: West Akron-Pleasant Valley loaded line on FE’s system. It held on, although
138-kV East line (Q-22) opened. heavily overloaded to 160 and 180% of normal rat-
4N) 15:59:00 EDT: West Akron-Rosemont-Pine- ings, until tripping at 16:05:55 EDT. The loss of
Wadsworth 138-kV line opened. this line had a significant effect on the area, and
voltages dropped significantly. More power
From 16:00 EDT to 16:08:59 EDT, four 138-kV shifted back to the remaining 345-kV network,
lines tripped, and the Sammis-Star 345-kV line pushing Sammis-Star’s loading above 120% of rat-
tripped due to high current and low voltage: ing. Two seconds later, at 16:05:57 EDT, Sammis-
4O) 16:05:55 EDT: Dale-West Canton 138-kV line Star tripped out. Unlike the previous three 345-kV
tripped due to sag into a tree, reclosed at West lines, which tripped on short circuits to ground
Canton only due to tree contacts, Sammis-Star tripped because
its protective relays saw low apparent impedance
4P) 16:05:57 EDT: Sammis-Star 345-kV line (depressed voltage divided by abnormally high
tripped line current)—i.e., the relay reacted as if the high
4Q) 16:06:02 EDT: Star-Urban 138-kV line tripped flow was due to a short circuit. Although three
more 138-kV lines dropped quickly in Ohio fol-
4R) 16:06:09 EDT: Richland-Ridgeville-Napo- lowing the Sammis-Star trip, loss of the Sammis-
leon-Stryker 138-kV line tripped on overload Star line marked the turning point at which sys-
and locked out at all terminals tem problems in northeast Ohio initiated a cascad-
ing blackout across the northeast United States
4S) 16:08:58 EDT: Ohio Central-Wooster 138-kV
and Ontario.
line tripped
Note: 16:08:55 EDT: East Wooster-South Can- Losing the 138-kV Transmission Lines
ton 138-kV line tripped, but successful auto- The tripping of 138-kV transmission lines that
matic reclosing restored this line. began at 15:39 EDT occurred because the loss
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 69
Cause 1
of the combination of the Har- dropped within the Cleveland-Akron area to
Inadequate ding-Chamberlin, Hanna-Juniper restore voltage at the Star bus from 90.8% (at 120%
System and Star-South Canton 345-kV of normal and emergency ampere rating) up to
Understanding lines overloaded the 138-kV sys- 95.9% (at 101% of normal and emergency ampere
tem with electricity flowing north rating).44 The P-V and V-Q analysis reviewed in
toward the Akron and Cleveland loads. Modeling Chapter 4 indicated that 95% is the minimum
indicates that the return of either the Hanna- operating voltage appropriate for 345-kV buses in
Juniper or Chamberlin-Harding 345-kV lines the Cleveland-Akron area. The investigation team
would have diminished, but not alleviated, all of concluded that since the Sammis-Star 345 kV out-
the 138-kV overloads. In theory, the return of both age was the critical event leading to widespread
lines would have restored all the 138-kV lines to cascading in Ohio and beyond, if manual or auto-
within their emergency ratings. matic load-shedding of 1,500 MW had occurred
within the Cleveland-Akron
Cause 2 However, all three 345-kV lines Recommendations
area before that outage, the 8, page 147; 21, page 158;
Inadequate had already been compromised 23, page 160
blackout could have been
Situational due to tree contacts so it is
Awareness averted.
unlikely that FE would have suc-
cessfully restored either line had Loss of the Sammis-Star 345-kV Line
they known it had tripped out, and since
Figure 5.15, derived from investigation team mod-
Star-South Canton had already tripped and
eling, shows how the power flows shifted across
reclosed three times it is also unlikely that an
FE’s 345- and key 138-kV northeast Ohio lines as
operator knowing this would have trusted it to
the line failures progressed. All lines were
operate securely under emergency conditions.
loaded within normal limits after the Har-
While generation redispatch scenarios alone
ding-Chamberlin lock-out, but after the
would not have solved the overload problem,
Hanna-Juniper trip at 15:32 EDT, the Star-South
modeling indicates that shedding load in the
Canton 345-kV line and three 138-kV lines
Cleveland and Akron areas may have reduced
jumped above normal loadings. After Star-South
most line loadings to within emergency range and
Canton locked out at 15:41 EDT within its emer-
helped stabilize the system. However, the amount
gency rating, five 138-kV and the Sammis-Star
of load shedding required grew rapidly as FE’s sys-
345-kV lines were overloaded. From that point, as
tem unraveled.
the graph shows, each subsequent line loss
Preventing the Blackout with Load-Shedding increased loadings on other lines, some loading to
well over 150% of normal ratings before they
Cause 1
The investigation team examined
failed. The Sammis-Star 345-kV line stayed in ser-
Inadequate whether load shedding before the
vice until it tripped at 16:05:57 EDT.
System loss of the Sammis-Star 345-kV
Understanding line at 16:05:57 EDT could have FirstEnergy had no automatic load-shedding
prevented this line loss. The team schemes in place, and did not attempt to begin
found that 1,500 MW of load would have had to be manual load-shedding. As Chapters 4 and 5 have
established, once Sammis-Star tripped, the possi-
Figure 5.15. Simulated Effect of Prior Outages on bility of averting the coming cascade by shedding
138-kV Line Loadings load ended. Within 6 minutes of these overloads,
extremely low voltages, big power swings and
accelerated line tripping would cause separations
and blackout within the Recommendation
21, page 158
Eastern Interconnection.
Endnotes
1 Investigation team field visit to FE 10/8/2003: Steve
Morgan.
2 Investigation team field visit to FE, September 3, 2003,
Hough interview: “When asked whether the voltages seemed
unusual, he said that some sagging would be expected on a
hot day, but on August 14th the voltages did seem unusually
low.” Spidle interview: “The voltages for the day were not
particularly bad.”
70 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
3 Manual of Operations, valid as of March 3, 2003, Process 15 FE Report, “Investigation of FirstEnergy’s Energy Manage-
flowcharts: Voltage Control and Reactive Support – Plant and ment System Status on August 14, 2003.”
System Voltage Monitoring Under Normal Conditions. 16 Investigation team field visit to FE, October 8-9, 2003,
4 14:13:18. Channel 16 - Sammis 1. 13:15:49 / Channel 16 – Sanicky Interview: “From his experience, it is not unusual for
West Lorain (FE Reliability Operator (RO) says, “Thanks. alarms to fail. Often times, they may be slow to update or they
We’re starting to sag all over the system.”) / 13:16:44. Channel may die completely. From his experience as a real-time opera-
16 – Eastlake (talked to two operators) (RO says, “We got a tor, the fact that the alarms failed did not surprise him.” Also
way bigger load than we thought we would have.” And “…So from same document, Mike McDonald interview, “FE has pre-
we’re starting to sag all over the system.”) / 13:20:22. Channel viously had [servers] down at the same time. The big issue for
16 – RO to “Berger” / 13:22:07. Channel 16 – “control room” them was that they were not receiving new alarms.”
RO says, “We’re sagging all over the system. I need some 17 A “cold” reboot of the XA21 system is one in which all
help.” / 13:23:24. Channel 16 – “Control room, Tom” / nodes (computers, consoles, etc.) of the system are shut down
13:24:38. Channel 16 – “Unit 9” / 13:26:04. Channel 16 – and then restarted. Alternatively, a given XA21 node can be
“Dave” / 13:28:40. Channel 16 “Troy Control.” Also general “warm” rebooted wherein only that node is shut down and
note in RO Dispatch Log. restarted, or restarted from a shutdown state. A cold reboot
5 Example at 13:33:40, Channel 3, FE transcripts. will take significantly longer to perform than a warm one.
6 Investigation team field visit to MISO, Walsh and Seidu Also during a cold reboot much more of the system is unavail-
interviews. able for use by the control room operators for visibility or con-
7 FE had and ran a state estimator every 30 minutes. This trol over the power system. Warm reboots are not uncommon,
whereas cold reboots are rare. All reboots undertaken by FE’s
served as a base from which to perform contingency analyses. IT EMSS support personnel on August 14 were warm reboots.
FE’s contingency analysis tool used SCADA and EMS inputs 18 The cold reboot was done in the early morning of 15
to identify any potential overloads that could result from vari-
ous line or equipment outages. FE indicated that it has experi- August and corrected the alarm problem as hoped.
enced problems with the automatic contingency analysis 19 Example at 14:19, Channel l4, FE transcripts.
operation since the system was installed in 1995. As a result, 20 Example at 14:25, Channel 8, FE transcripts.
FE operators or engineers ran contingency analysis manually 21 Example at 14:32, Channel 15, FE transcripts.
rather than automatically, and were expected to do so when
22 “Interim Report, Utility Vegetation Management,”
there were questions about the state of the system. Investiga-
tion team interviews of FE personnel indicate that the contin- U.S.-Canada Joint Outage Investigation Task Force, Vegeta-
gency analysis model was likely running but not consulted at tion Management Program Review, October 2003, page 7.
any point in the afternoon of August 14. 23 Investigation team transcript, meeting on September 9,
8 After the Stuart-Atlanta line tripped, Dayton Power & Light 2003, comments by Mr. Steve Morgan, Vice President Electric
did not immediately provide an update of a change in equip- Operations:
ment availability using a standard form that posts the status Mr. Morgan: The sustained outage history for these lines,
change in the SDX (System Data Exchange, the NERC data- 2001, 2002, 2003, up until the event, Chamberlin-Harding
base which maintains real-time information on grid equip- had zero operations for those two-and-a-half years. And
ment status), which relays that notice to reliability Hanna-Juniper had six operations in 2001, ranging from four
coordinators and control areas. After its state estimator failed minutes to maximum of 34 minutes. Two were unknown, one
to solve properly, MISO checked the SDX to make sure that was lightning, one was a relay failure, and two were really
they had properly identified all available equipment and out- relay scheme mis-operations. They’re category other. And
ages, but found no posting there regarding Stuart-Atlanta’s typically, that—I don’t know what this is particular to opera-
outage. tions, that typically occurs when there is a mis-operation.
9 Investigation team field visit, interviews with FE personnel Star-South Canton had no operations in that same period of
on October 8-9, 2003. time, two-and-a-half years. No sustained outages. And
10 DOE Site Visit to First Energy, September 3, 2003, Inter-
Sammis-Star, the line we haven’t talked about, also no sus-
tained outages during that two-and-a-half year period. So is it
view with David M. Elliott. normal? No. But 345 lines do operate, so it’s not unknown.
11 FE Report, “Investigation of FirstEnergy’s Energy Manage- 24 “Utility Vegetation Management Final Report,” CN Utility
ment System Status on August 14, 2003,” Bullet 1, Section Consulting, March 2004, page 32.
4.2.11. 25 “FE MISO Findings,” page 11.
12 Investigation team interviews with FE, October 8-9, 2003.
26 FE was conducting right-of-way vegetation maintenance
13 Investigation team field visit to FE, October 8-9, 2003: team
on a 5-year cycle, and the tree crew at Hanna-Juniper was
was advised that FE had discovered this effect during three spans away, clearing vegetation near the line, when the
post-event investigation and testing of the EMS. FE’s report contact occurred on August 14. Investigation team 9/9/03
“Investigation of FirstEnergy’s Energy Management System meeting transcript, and investigation field team discussion
Status on August 14, 2003” also indicates that this finding with the tree-trimming crew foreman.
was “verified using the strip charts from 8-14-03” (page 23), 27 Based on “FE MISO Findings” document, page 11.
not that the investigation of this item was instigated by opera-
28 “Interim Report, Utility Vegetation Management,”
tor reports of such a failure.
14 There is a conversation between a Phil and a Tom that US-Canada Joint Outage Task Force, Vegetation Management
speaks of “flatlining” 15:01:33. Channel 15. There is no men- Program Review, October 2003, page 6.
29 Investigation team September 9, 2003 meeting transcripts,
tion of AGC or generation control in the DOE Site Visit inter-
views with the reliability coordinator. Mr. Steve Morgan, First Energy Vice President, Electric Sys-
tem Operations:
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 71
Mr. Benjamin: Steve, just to make sure that I’m understand- 34 Example synchronized at 15:36 (from 14:43) Channel 20,
ing it correctly, you had indicated that once after MISO transcripts.
Hanna-Juniper relayed out, there wasn’t really a problem with 35 Example at 15:42:49, Channel 8, FE transcripts.
voltage on the system until Star-S. Canton operated. But were 36 Example at 15:46:00, Channel 8 FE transcripts.
the system operators aware that when Hanna-Juniper was
37 Example at 15:45:18, Channel 4, FE transcripts.
out, that if Star-S. Canton did trip, they would be outside of
operating limits? 38 Example at 15:46:00, Channel 8 FE transcripts.
Mr. Morgan: I think the answer to that question would have 39 Example at 15:50:15, Channel 12 FE transcripts.
required a contingency analysis to be done probably on
40 Example synchronized at 15:48 (from 14:55), channel 22,
demand for that operation. It doesn’t appear to me that a con-
tingency analysis, and certainly not a demand contingency MISO transcripts.
analysis, could have been run in that period of time. Other 41 Example at 15:56:00, Channel 31, FE transcripts.
than experience, I don’t know that they would have been able 42 FE Transcripts 15:45:18 on Channel 4 and 15:56:49 on
to answer that question. And what I know of the record right Channel 31.
now is that it doesn’t appear that they ran contingency analy- 43 The operator logs from FE’s Ohio control center indicate
sis on demand.
Mr. Benjamin: Could they have done that? that the west desk operator knew of the alarm system failure
Mr. Morgan: Yeah, presumably they could have. at 14:14, but that the east desk operator first knew of this
Mr. Benjamin: You have all the tools to do that? development at 15:45. These entries may have been entered
Mr. Morgan: They have all the tools and all the information is after the times noted, however.
there. And if the State Estimator is successful in solving, and 44 The investigation team determined that FE was using a dif-
all the data is updated, yeah, they could have. I would say in ferent set of line ratings for Sammis-Star than those being
addition to those tools, they also have access to the planning used in the MISO and PJM reliability coordinator calculations
load flow model that can actually run the same—full load of or by its neighbor AEP. Specifically, FE was operating
the model if they want to. Sammis-Star assuming that the 345-kV line was rated for
30 Example synchronized at 14:32 (from 13:32) #18 041 summer normal use at 1,310 MVA, with a summer emergency
TDC-E2 283.wav, AEP transcripts. limit rating of 1,310 MVA. In contrast, MISO, PJM and AEP
31 Example synchronized at 14:19 #2 020 TDC-E1 266.wav, were using a more conservative rating of 950 MVA normal
and 1,076 MVA emergency for this line. The facility owner (in
AEP transcripts. this case FE) is the entity which provides the line rating; when
32 Example at 15:36 Channel 8, FE transcripts. and why the ratings were changed and not communicated to
33 Example at 15:41:30 Channel 3, FE transcripts. all concerned parties has not been determined.
72 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
6. The Cascade Stage of the Blackout
Chapter 5 described how uncorrected problems in complex that it may be impossible to ever com-
northern Ohio developed to 16:05:57 EDT, the last pletely prove these or other theories about the
point at which a cascade of line trips could have fast-moving events of August 14. Final modeling
been averted. However, the Task Force’s investiga- results will be published by NERC as a technical
tion also sought to understand how and why the report in several months.
cascade spread and stopped as it did. As detailed
below, the investigation determined the sequence Why Does a Blackout Cascade?
of events in the cascade, and how and why it
Major blackouts are rare, and no two blackout sce-
spread, and how it stopped in each general geo-
narios are the same. The initiating events will
graphic area.
vary, including human actions or inactions, sys-
Based on the investigation to date, the investiga- tem topology, and load/generation balances. Other
tion team concludes that the cascade spread factors that will vary include the distance between
beyond Ohio and caused such a widespread black- generating stations and major load centers, voltage
out for three principal reasons. First, the loss of the profiles across the grid, and the types and settings
Sammis-Star 345-kV line in Ohio, following the of protective relays in use.
loss of other transmission lines and weak voltages Some wide-area blackouts start with short circuits
within Ohio, triggered many subsequent line trips. (faults) on several transmission lines in short suc-
Second, many of the key lines which tripped cession—sometimes resulting from natural causes
between 16:05:57 and 16:10:38 EDT operated on such as lightning or wind or, as on August 14,
zone 3 impedance relays (or zone 2 relays set to resulting from inadequate tree management in
operate like zone 3s) which responded to over- right-of-way areas. A fault causes a high current
loads rather than true faults on the grid. The speed and low voltage on the line containing the fault. A
at which they tripped spread the reach and accel- protective relay for that line detects the high cur-
erated the spread of the cascade beyond the Cleve- rent and low voltage and quickly trips the circuit
land-Akron area. Third, the evidence collected breakers to isolate that line from the rest of the
indicates that the relay protection settings for the power system.
transmission lines, generators and under-fre-
quency load-shedding in the northeast may not be A cascade is a dynamic phenomenon that cannot
entirely appropriate and are certainly not coordi- be stopped by human intervention once started. It
nated and integrated to reduce the likelihood and occurs when there is a sequential tripping of
consequences of a cascade—nor were they numerous transmission lines and generators in a
intended to do so. These issues are discussed in widening geographic area. A cascade can be trig-
depth below. gered by just a few initiating events, as was seen
on August 14. Power swings and voltage fluctua-
This analysis is based on close examination of the tions caused by these initial events can cause
events in the cascade, supplemented by complex, other lines to detect high currents and low volt-
detailed mathematical modeling of the electrical ages that appear to be faults, even if faults do not
phenomena that occurred. At the completion of actually exist on those other lines. Generators are
this report, the modeling had progressed through tripped off during a cascade to protect them from
16:10:40 EDT, and was continuing. Thus this severe power and voltage swings. Protective relay
chapter is informed and validated by modeling systems work well to protect lines and generators
(explained below) up until that time. Explanations from damage and to isolate them from the system
after that time reflect the investigation team’s best under normal and abnormal system conditions.
hypotheses given the available data, and may be
confirmed or modified when the modeling is com- But when power system operating and design cri-
plete. However, simulation of these events is so teria are violated because several outages occur
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 73
simultaneously, commonly used protective relays Chapter 5 described the four phases that led to the
that measure low voltage and high current cannot initiation of the cascade at about 16:06 EDT. After
distinguish between the currents and voltages 16:06 EDT, the cascade evolved in three distinct
seen in a system cascade from those caused by a phases:
fault. This leads to more and more lines and gener-
u Phase 5. The collapse of FE’s transmission sys-
ators being tripped, widening the blackout area.
tem induced unplanned shifts of power across
How Did the Cascade Evolve on the region. Shortly before the collapse, large
(but normal) electricity flows were moving
August 14?
across FE’s system from generators in the south
A series of line outages in northeast Ohio starting (Tennessee and Kentucky) and west (Illinois
at 15:05 EDT caused heavy loadings on parallel and Missouri) to load centers in northern Ohio,
circuits, leading to the trip and lock-out of FE’s eastern Michigan, and Ontario. A series of lines
Sammis-Star 345-kV line at 16:05:57 Eastern Day- within northern Ohio tripped under the high
light Time. This was the event that triggered a cas-
cade of interruptions on the high voltage system, Figure 6.1. Rate of Line and Generator Trips During
causing electrical fluctuations and facility trips the Cascade
such that within seven minutes the blackout rip-
pled from the Cleveland-Akron area across much
of the northeast United States and Canada. By
16:13 EDT, more than 508 generating units at 265
power plants had been lost, and tens of millions of
people in the United States and Canada were with-
out electric power.
The events in the cascade started relatively
slowly. Figure 6.1 illustrates how the number of
lines and generation lost stayed relatively low dur-
ing the Ohio phase of the blackout, but then
picked up speed after 16:08:59 EDT. The cascade
was complete only three minutes later.
Impedance Relays
The most common protective device for trans- just inside the line or just beyond the line. In
mission lines is the impedance (Z) relay (also this case, the zone 2 relay on the far end trips
known as a distance relay). It detects changes in the breakers after a short delay, after the zone 1
currents (I) and voltages (V) to determine the relay near the fault opens the line on that end
apparent impedance (Z=V/I) of the line. A relay first.
is installed at each end of a transmission line.
u The third zone is slower acting and looks for
Each relay is actually three relays within one,
line faults and faults well beyond the length of
with each element looking at a particular “zone”
the line. It can be thought of as a remote relay
or length of the line being protected.
or breaker backup, but should not trip the
u The first zone looks for faults over 80% of the breakers under typical emergency conditions.
line next to the relay, with no time delay before
the trip. An impedance relay operates when the apparent
impedance, as measured by the current and volt-
u The second zone is set to look at the entire line age seen by the relay, falls within any one of the
and slightly beyond the end of the line with a operating zones for the appropriate amount of
slight time delay. The slight delay on the zone time for that zone. The relay will trip and cause
2 relay is useful when a fault occurs near one circuit breakers to operate and isolate the line.
end of the line. The zone 1 relay near that end All three relay zone operations protect lines from
operates quickly to trip the circuit breakers on faults and may trip from apparent faults caused
that end. However, the zone 1 relay on the by large swings in voltages and currents.
other end may not be able to tell if the fault is
74 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
loads, hastened by the impact of Zone 3 imped- Although much of the disturbance area was
ance relays. This caused a series of shifts in fully blacked out in this process, some islands
power flows and loadings, but the grid stabi- were able to reach equilibrium without total
lized after each. loss of service. For example, the island consist-
ing of most of New England and the Maritime
u Phase 6. After 16:10:36 EDT, the power surges
Provinces stabilized and generation and load
resulting from the FE system failures caused
returned to balance. Another island consisted of
lines in neighboring areas to see overloads that
load in western New York and a small portion of
caused impedance relays to operate. The result
Ontario, supported by some New York genera-
was a wave of line trips through western Ohio
tion, the large Beck and Saunders plants in
that separated AEP from FE. Then the line trips
Ontario, and the 765-kV interconnection to
progressed northward into Michigan separating
Québec. This island survived but some other
western and eastern Michigan, causing a power
areas with large load centers within the island
flow reversal within Michigan toward Cleve-
collapsed into a blackout condition (Figure 6.2).
land. Many of these line trips were from Zone 3
impedance relay actions that accelerated the What Stopped the August 14 Blackout
speed of the line trips and reduced the potential from Cascading Further?
time in which grid operators might have identi-
The investigation concluded that a combination of
fied the growing problem and acted construc-
the following factors determined where and when
tively to contain it.
the cascade stopped spreading:
With paths cut from the west, a massive power u The effects of a disturbance travel over power
surge flowed from PJM into New York and lines and become damped the further they are
Ontario in a counter-clockwise flow around from the initial point, much like the ripple from
Lake Erie to serve the load still connected in a stone thrown in a pond. Thus, the voltage and
eastern Michigan and northern Ohio. Relays on current swings seen by relays on lines farther
the lines between PJM and New York saw this away from the initial disturbance are not as
massive power surge as faults and tripped those severe, and at some point they are no longer suf-
lines. Ontario’s east-west tie line also became ficient to cause lines to trip.
overloaded and tripped, leaving northwest u Higher voltage lines and more densely net-
Ontario connected to Manitoba and Minnesota. worked lines, such as the 500-kV system in PJM
The entire northeastern United States and east- and the 765-kV system in AEP, are better able to
ern Ontario then became a large electrical absorb voltage and current swings and thus
island separated from the rest of the Eastern serve as a barrier to the spread of a cascade. As
Interconnection. This large area, which had seen in Phase 6, the cascade progressed into
been importing power prior to the cascade, western Ohio and then northward through
quickly became unstable after 16:10:38 as there Michigan through the areas that had the fewest
was not sufficient generation on-line within the transmission lines. Because there were fewer
island to meet electricity demand. Systems to
the south and west of the split, such as PJM,
Figure 6.2. Area Affected by the Blackout
AEP and others further away, remained intact
and were mostly unaffected by the outage. Once
the northeast split from the rest of the Eastern
Interconnection, the cascade was isolated.
u Phase 7. In the final phase, after 16:10:46 EDT,
the large electrical island in the northeast had
less generation than load, and was unstable
with large power surges and swings in fre-
quency and voltage. As a result, many lines and
generators across the disturbance area tripped,
breaking the area into several electrical islands.
Generation and load within these smaller
islands was often unbalanced, leading to fur-
ther tripping of lines and generating units until
equilibrium was established in each island.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 75
System Oscillations, Stable, Transient, and Dynamic Conditions
The electric power system constantly experi- rare cases when a disturbance occurs, the system
ences small power oscillations that do not lead to may appear to rebalance quickly, but it then
system instability. They occur as generator rotors over-shoots and the oscillations can grow, caus-
accelerate or slow down while rebalancing elec- ing widespread instability that spreads in terms
trical output power to mechanical input power, of both the magnitude of the oscillations and in
to respond to changes in load or network condi- geographic scope. This can occur in a system that
tions. These oscillations are observable in the is heavily loaded, causing the electrical distance
power flow on transmission lines that link gener- (apparent impedance) between generators to be
ation to load or in the tie lines that link different longer, making it more difficult to keep the
regions of the system together. But with a distur- machine angles and speeds synchronized. In a
bance to the network, the oscillations can system that is well damped, the oscillations will
become more severe, even to the point where settle out quickly and return to a steady balance.
flows become progressively so great that protec- If the oscillation continues over time, neither
tive relays trip the connecting lines. If the lines growing nor subsiding, it is a poorly damped
connecting different electrical regions separate, system.
each region will find its own frequency, depend-
ing on the load to generation balance at the time The illustration below, of a weight hung on a
of separation. spring balance, illustrates a system which oscil-
lates over several cycles to return to balance. A
Oscillations that grow in amplitude are called critical point to observe is that in the process of
unstable oscillations. Such oscillations, once ini- hunting for its balance point, the spring over-
tiated, cause power to flow back and forth across shoots the true weight and balance point of the
the system like water sloshing in a rocking tub. spring and weight combined, and must cycle
through a series of exaggerated overshoots and
In a stable electric system, if a disturbance such
underweight rebounds before settling down to
as a fault occurs, the system will readjust and
rest at its true balance point. The same process
rebalance within a few seconds after the fault
occurs on an electric system, as can be observed
clears. If a fault occurs, protective relays can trip
in this chapter.
in less than 0.1 second. If the system recovers
and rebalances within less than 3 seconds, with If a system is in transient instability, the oscilla-
the possible loss of only the faulted element and tions following a disturbance will grow in magni-
a few generators in the area around the fault, then tude rather than settle out, and it will be unable
that condition is termed “transiently stable.” If to readjust to a stable, steady state. This is what
the system takes from 3 to 30 seconds to recover happened to the area that blacked out on August
and stabilize, it is “dynamically stable.” But in 14, 2003.
76 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
lines, each line absorbed more of the power and transmission paths west and northwest into Mich-
voltage surges and was more vulnerable to trip- igan, causing a sequential loss of lines and power
ping. A similar effect was seen toward the east plants.
as the lines between New York and Pennsylva-
nia, and eventually northern New Jersey trip-
Key Events in This Phase
ped. The cascade of transmission line outages 5A) 16:05:57 EDT: Sammis-Star 345-kV tripped
became contained after the northeast United by zone 3 relay.
States and Ontario were completely separated 5B) 16:08:59 EDT: Galion-Ohio Central-Mus-
from the rest of the Eastern Interconnection and kingum 345-kV line tripped on zone 3 relay.
no more power flows were possible into the 5C) 16:09:06 EDT: East Lima-Fostoria Central
northeast (except the DC ties from Québec, 345-kV line tripped on zone 3 relay, causing
which continued to supply power to western major power swings through New York and
New York and New England). Ontario into Michigan.
u Line trips isolated some areas from the portion 5D) 16:09:08 EDT to 16:10:27 EDT: Several power
of the grid that was experiencing instability. plants lost, totaling 937 MW.
Many of these areas retained sufficient on-line 5A) Sammis-Star 345-kV Tripped: 16:05:57 EDT
generation or the capacity to import power from
Sammis-Star did not trip due to a short circuit to
other parts of the grid, unaffected by the surges
ground (as did the prior 345-kV lines that tripped).
or instability, to meet demand. As the cascade
Sammis-Star tripped due to protective zone 3
progressed, and more generators and lines trip-
relay action that measured low apparent imped-
ped off to protect themselves from severe dam-
ance (depressed voltage divided by abnormally
age, some areas completely separated from the
high line current) (Figure 6.4). There was no fault
unstable part of the Eastern Interconnection. In
and no major power swing at the time of the
many of these areas there was sufficient genera-
trip—rather, high flows above the line’s emer-
tion to match load and stabilize the system.
gency rating together with depressed voltages
After the large island was formed in the north-
caused the overload to appear to the protective
east, symptoms of frequency and voltage decay
relays as a remote fault on the system. In effect, the
emerged. In some parts of the northeast, the sys-
relay could no longer differentiate between a
tem became too unstable and shut itself down.
remote three-phase fault and an exceptionally
In other parts, there was sufficient generation,
high line-load condition. Moreover, the reactive
coupled with fast-acting automatic load shed-
flows (VAr) on the line were almost ten times
ding, to stabilize frequency and voltage. In this
higher than they had been earlier in the day
manner, most of New England and the Maritime
because of the current overload. The relay oper-
Provinces remained energized. Approximately
ated as it was designed to do.
half of the generation and load remained on in
western New York, aided by generation in
southern Ontario that split and stayed with
Figure 6.3. Sammis-Star 345-kV Line Trip,
western New York. There were other smaller
16:05:57 EDT
isolated pockets of load and generation that
were able to achieve equilibrium and remain
energized.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 77
The Sammis-Star 345-kV line trip completely sev- overloading and tripping within Ohio, this was
ered the 345-kV path into northern Ohio from happening slowly enough under relatively stable
southeast Ohio, triggering a new, fast-paced conditions that the system could readjust—after
sequence of 345-kV transmission line trips in each line loss, power flows would redistribute
which each line trip placed a greater flow burden across the remaining lines. This is illustrated in
on those lines remaining in service. These line Figure 6.5, which shows the MW flows on the
outages left only three paths for power to flow into Michigan Electrical Coordinated Systems (MECS)
western Ohio: (1) from northwest Pennsylvania to interfaces with AEP (Ohio), FirstEnergy (Ohio)
northern Ohio around the south shore of Lake and Ontario. The graph shows a shift from 150
Erie, (2) from southwest Ohio toward northeast MW imports to 200 MW exports from the MECS
Ohio, and (3) from eastern Michigan and Ontario. system into FirstEnergy at 16:05:57 EDT after the
The line interruptions substantially weakened loss of Sammis-Star, after which this held steady
northeast Ohio as a source of power to eastern until 16:08:59, when the loss of East Lima-Fostoria
Michigan, making the Detroit area more reliant on Central cut the main energy path from the south
345-kV lines west and northwest of Detroit, and and west into Cleveland and Toledo. Loss of this
from northwestern Ohio to eastern Michigan. The path was significant, causing flow from MECS into
impact of this trip was felt across the grid—it FE to jump from 200 MW up to 2,300 MW, where
caused a 100 MW increase in flow from PJM into it bounced somewhat before stabilizing, roughly,
New York and through to Ontario.1 Frequency in until the path across Michigan was cut at 16:10:38
the Eastern Interconnection increased momen- EDT.
tarily by 0.02 Hz.
Transmission Lines into Northwestern Ohio
Soon after the Sammis-Star trip, four of the five 48 Tripped, and Generation Tripped in South
MW Handsome Lake combustion turbines in Central Michigan and Northern Ohio: 16:08:59
western Pennsylvania tripped off-line. These EDT to 16:10:27 EDT
units are connected to the 345-kV system by the
Homer City-Wayne 345-kV line, and were operat- 5B) 16:08:59 EDT: Galion-Ohio Central-Mus-
ing that day as synchronous condensers to partici- kingum 345-kV line tripped
pate in PJM’s spinning reserve market (not to 5C) 16:09:06 EDT: East Lima-Fostoria Central
provide voltage support). When Sammis-Star trip- 345-kV line tripped, causing a large power
ped and increased loadings on the local transmis- swing from Pennsylvania and New York
sion system, the Handsome Lake units were close through Ontario to Michigan
enough electrically to sense the impact and trip-
ped off-line at 16:07:00 EDT on under-voltage. The tripping of the Galion-Ohio Central-
Muskingum and East Lima-Fostoria Central
During the period between the Sammis-Star trip
and the trip of East Lima-Fostoria at 16:09:06.3
Figure 6.5. Line Flows Into Michigan
EDT, the system was still in a steady-state condi-
tion. Although one line after another was
78 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
345-kV transmission lines removed the transmis- shows the impact of the East Lima-Fostoria Cen-
sion paths from southern and western Ohio into tral and the New York to Ontario power swing,
northern Ohio and eastern Michigan. Northern which continued to oscillate for over 10 seconds.
Ohio was connected to eastern Michigan by only Looking at the MW flow line, it is clear that when
three 345-kV transmission lines near the south- Sammis-Star tripped, the system experienced
western bend of Lake Erie. Thus, the combined oscillations that quickly damped out and
northern Ohio and eastern Michigan load centers rebalanced. But East Lima-Fostoria triggered sig-
were left connected to the rest of the grid only by: nificantly greater oscillations that worsened in
(1) transmission lines eastward from northeast magnitude for several cycles, and returned to sta-
Ohio to northwest Pennsylvania along the south- bility but continued to flutter until the
ern shore of Lake Erie, and (2) westward by lines Argenta-Battle Creek trip 90 seconds later. Volt-
west and northwest of Detroit, Michigan and from ages also began declining at this time.
Michigan into Ontario (Figure 6.6).
After the East Lima-Fostoria Central trip, power
The Galion-Ohio Central-Muskingum 345-kV line flows increased dramatically and quickly on the
tripped first at Muskingum at 16:08:58.5 EDT on a lines into and across southern Michigan.
phase-to-ground fault, reclosed and tripped again Although power had initially been flowing north-
at 16:08:58.6 at Ohio Central, reclosed and tripped east out of Michigan into Ontario, that flow sud-
again at Muskingum on a Zone 3 relay, and finally denly reversed and approximately 500 to 700 MW
tripped at Galion on a ground fault. of power (measured at the Michigan-Ontario bor-
der, and 437 MW at the Ontario-New York border
After the Galion-Ohio Central-Muskingum line
at Niagara) flowed southwest out of Ontario
outage and numerous 138-kV line trips in central
through Michigan to serve the load of Cleveland
Ohio, the East Lima-Fostoria Central 345-kV line
and Toledo. This flow was fed by 700 MW pulled
tripped at 16:09:06 EDT on Zone 3 relay operation
out of PJM through New York on its 345-kV net-
due to high current and extremely low voltage
work.2 This was the first of several inter-area
(80%). Investigation team modeling indicates that power and frequency events that occurred over
if automatic under-voltage load-shedding had the next two minutes. This was the system’s
been in place in northeast Ohio, it might have response to the loss of the northwest Ohio trans-
been triggered at or before this point, and dropped mission paths (above), and the stress that the
enough load to reduce or still-high Cleveland, Toledo, and Detroit loads put
eliminate the subsequent Recommendation s
onto the surviving lines and local generators.
8, page 147; 21, page 158
line overloads that spread
the cascade. Figure 6.7 also shows the magnitude of subse-
quent flows and voltages at the New York-Ontario
Figure 6.7, a high-speed recording of 345-kV flows Niagara border, triggered by the trips of the
past Niagara Falls from the Hydro One recorders, Argenta-Battle Creek, Argenta-Tompkins, Hamp-
ton-Pontiac and Thetford-Jewell 345-kV lines in
Figure 6.6. Ohio 345-kV Lines Trip, 16:08:59 to Michigan, and the Erie West-Ashtabula-Perry
16:09:07 EDT
Figure 6.7. New York-Ontario Line Flows at Niagara
ONTARIO
5C
5B
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 79
345-kV line linking the Cleveland area to Pennsyl- reach of the line and its zone 1 and zone 2 settings,
vania. Farther south, the very low voltages on the a zone 3 relay waits through a 1 to 2 second time
northern Ohio transmission system made it very delay to allow the primary line protection to act
difficult for the generation in the Cleveland and first. A few lines have zone 3 settings designed
Lake Erie area to maintain synchronism with the with overload margins close to the long-term
Eastern Interconnection. Over the next two min- emergency limit of the line, because the length
utes, generators in this area shut down after reach- and configuration of the line dictate a higher
ing a point of no recovery as the stress level across apparent impedance setting. Thus it is possible for
the remaining ties became excessive. a zone 3 relay to operate on line load or overload in
extreme contingency conditions even in the
Figure 6.8, of metered power flows along the New
absence of a fault (which is why many regions in
York interfaces, documents how the flows head-
the United States and Canada have eliminated the
ing north and west toward Detroit and Cleveland
use of zone 3 relays on 230-kV and greater lines).
varied at different points on the grid. Beginning at
Some transmission operators set zone 2 relays to
16:09:05 EDT, power flows jumped simulta-
serve the same purpose as zone 3s—i.e., to reach
neously across all three interfaces—but when the
well beyond the length of the line it is protecting
first power surge peaked at 16:09:09, the change in
and protect against a distant fault on the outer
flow was highest on the PJM interface and lowest
lines.
on the New England interface. Power flows
increased significantly on the PJM-NY and NY- The Sammis-Star line tripped at 16:05:57 EDT on
Ontario interfaces because of the redistribution of a zone 3 impedance relay although there were no
flow around Lake Erie. The New England and Mar- faults occurring at the time, because increased real
itime systems maintained the same generation to and reactive power flow caused the apparent
load balance and did not carry the redistributed impedance to be within the impedance circle
flows because they were not in the direct path of (reach) of the relay. Between 16:06:01 and
the flows, so that interface with New York showed 16:10:38.6 EDT, thirteen more important 345 and
little response. 138-kV lines tripped on zone 3 operations that
afternoon at the start of the cascade, including
Before this first major power swing on the Michi-
Galion-Ohio Central-Muskingum, East Lima-
gan/Ontario interface, power flows in the NPCC
Fostoria Central, Argenta-Battle Creek, Argenta-
Region (Québec, Ontario and the Maritimes, New
Tompkins, Battle Creek-Oneida, and Perry-
England and New York) were typical for the sum-
Ashtabula (Figure 6.9). These included several
mer period, and well within acceptable limits.
zone 2 relays in Michigan that had been set to
Transmission and generation facilities were then
operate like zone 3s, overreaching the line by more
in a secure state across the NPCC region.
than 200% with no intentional time delay for
Zone 3 Relays and the Start of the Cascade remote breaker failure protection.3 All of these
relays operated according to their settings. How-
Zone 3 relays are set to provide breaker failure and ever, the zone 3 relays (and zone 2 relays acting
relay backup for remote distance faults on a trans- like zone 3s) acted so quickly that they impeded
mission line. If it senses a fault past the immediate the natural ability of the electric system to hold
together, and did not allow for any operator inter-
Figure 6.8. First Power Swing Has Varying Impacts vention to attempt to stop the spread of the cas-
Across the Grid cade. The investigation team concluded that
because these zone 2 and 3 relays tripped after
each line overloaded, these relays were the com-
mon mode of failure that accelerated the geo-
graphic spread of the cascade. Given grid
conditions and loads and the limited operator
tools available, the speed of the zone 2 and 3 oper-
ations across Ohio and Michigan eliminated any
possibility after 16:05:57 EDT that either operator
action or automatic intervention could have lim-
ited or mitigated the growing cascade.
What might have happened on August 14 if these
lines had not tripped on zone 2 and 3 relays? Each
80 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Figure 6.9. Map of Zone 3 (and Zone 2s Operating Like Zone 3s) Relay Operations on August 14, 2003
Voltage Collapse
Although the blackout of August 14 has been workable bounds before individual transmission
labeled by some as a voltage collapse, it was not a trips began, and those trips occurred within nor-
voltage collapse as that term has been tradition- mal line ratings rather than in overloads. With
ally used by power system engineers. Voltage fewer lines operational, current flowing over the
collapse occurs when an increase in load or loss remaining lines increased and voltage decreased
of generation or transmission facilities causes (current increases in inverse proportion to the
dropping voltage, which causes a further reduc- decrease in voltage for a given amount of power
tion in reactive power from capacitors and line flow)—but it stabilized after each line trip until
charging, and still further voltage reductions. If the next circuit trip. Soon northern Ohio lines
the declines continue, these voltage reductions began to trip out automatically on protection
cause additional elements to trip, leading to fur- from overloads, not from insufficient reactive
ther reduction in voltage and loss of load. The power. Once several lines tripped in the Cleve-
result is a progressive and uncontrollable decline land-Akron area, the power flow was rerouted to
in voltage, all because the power system is other heavily loaded lines in northern Ohio,
unable to provide the reactive power required to causing depressed voltages which led to auto-
supply the reactive power demand. This did not matic tripping on protection from overloads.
occur on August 14. While the Cleveland-Akron Voltage collapse therefore was not a cause of the
area was short of reactive power reserves they cascade.
were just sufficient to supply the reactive power
demand in the area and maintain stable albeit As the cascade progressed beyond Ohio, it spread
depressed voltages for the outage conditions due not to insufficient reactive power and a volt-
experienced. age collapse, but because of dynamic power
swings and the resulting system instability.
But the lines in the Cleveland-Akron area tripped Figure 6.7 shows voltage levels recorded at the
as a result of tree contacts well below the nomi- Niagara area. It shows clearly that voltage levels
nal rating of the lines and not due to low volt- remained stable until 16:10:30 EDT, despite sig-
ages, which is a precursor for voltage collapse. nificant power fluctuations. In the cascade that
The initial trips within FirstEnergy began followed, the voltage instability was a compan-
because of ground faults with untrimmed ion to, not a driver of, the angle instability that
trees, not because of a shortage of reactive power tripped generators and lines.
and low voltages. Voltage levels were within
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 81
was operating with high load, and loads on each and limited transmission and generation
line grew as each preceding line tripped out of ser- capability.
vice. But if these lines had not tripped quickly on
zone 2s and 3s, each might have remained heavily Phase 6: The Full Cascade
loaded, with conductor temperatures increasing,
for as long as 20 to 30 minutes before the line Between 16:10:36 EDT and 16:13 EDT, thousands
sagged into something and experienced a ground of events occurred on the grid, driven by physics
fault. For instance, the Dale-West Canton line took and automatic equipment operations. When it was
20 minutes to trip under 160 to 180% of its normal over, much of the northeastern United States and
rated load. Even with sophisticated modeling it is the province of Ontario were in the dark.
impossible to predict just how long this delay Key Phase 6 Events
might have occurred (affected by wind speeds,
line loadings, and line length, tension and ground Transmission Lines Disconnected Across
clearance along every span), because the system Michigan and Northern Ohio, Generation Shut
did not become dynamically unstable until at least Down in Central Michigan and Northern Ohio,
after the Thetford-Jewell trip at 16:10:38 EDT. and Northern Ohio Separated from
During this period the system would likely have Pennsylvania: 16:10:36 to 16:10:39 EDT
remained stable and been able to readjust after 6A) Transmission and more generation tripped
each line trip on ground fault. If this period of within Michigan: 16:10:36 to 16:10:37 EDT:
deterioration and overloading under stable condi- 16:10:36.2 EDT: Argenta-Battle Creek 345-kV
tions had lasted for as little as 15 minutes or as line tripped
long as an hour, it is possible that the growing
problems could have been recognized and action 16:10:36.3 EDT: Argenta-Tompkins 345-kV
taken, such as automatic under-voltage load- line tripped
shedding, manual load-shedding in Ohio or other 16:10:36.8 EDT: Battle Creek-Oneida 345-kV
measures. So although the operation of zone 2 and line tripped
3 relays in Ohio and Michigan did not cause the 16:10:37 EDT: Sumpter Units 1, 2, 3, and 4
blackout, it is certain that units tripped on under-voltage (300 MW near
they greatly expanded and Recommendation
Detroit)
21, page 158
accelerated the spread of
the cascade. 16:10:37.5 EDT: MCV Plant output dropped
from 963 MW to 109 MW on over-current
5D) Multiple Power Plants Tripped, Totaling protection.
946 MW: 16:09:08 to 16:10:27 EDT
16:09:08 EDT: Michigan Cogeneration Venture Together, the above line outages interrupted the
plant reduction of 300 MW (from 1,263 MW to west-to-east transmission paths into the Detroit
963 MW) area from south-central Michigan. The Sumpter
generation units tripped in response to
16:09:17 EDT: Avon Lake 7 unit trips (82 MW)
16:09:17 EDT: Burger 3, 4, and 5 units trip (355 Figure 6.10. Michigan and Ohio Power Plants Trip
MW total)
16:09:30 EDT: Kinder Morgan units 3, 6 and 7 ONTARIO
82 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
under-voltage on the system. Michigan lines west is no evidence that the reclose caused harm to the
of Detroit then began to trip, as shown in Figure local generators.
6.11.
6B) Western and Eastern Michigan separation
The Argenta-Battle Creek relay first opened the started: 16:10:37 EDT to 16:10:38 EDT
line at 16:10:36.230 EDT, reclosed it at 16:10:37, 16:10:38.2 EDT: Hampton-Pontiac 345-kV
then tripped again. This line connects major gen- line tripped
erators—including the Cook and Palisades
nuclear plants and the Campbell fossil plant—to 16:10:38.4 EDT: Thetford-Jewell 345-kV line
the MECS system. This line is designed with tripped
auto-reclose breakers at each end of the line, After the Argenta lines tripped, the phase angle
which do an automatic high-speed reclose as soon between eastern and western Michigan began to
as they open to restore the line to service with no increase. The Hampton-Pontiac and Thetford-
interruptions. Since the majority of faults on the Jewell 345-kV lines were the only lines remaining
North American grid are temporary, automatic connecting Detroit to power sources and the rest of
reclosing can enhance stability and system reli- the grid to the north and west. When these lines
ability. However, situations can occur when the tripped out of service, it left the loads in Detroit,
power systems behind the two ends of the line Toledo, Cleveland, and their surrounding areas
could go out of phase during the high-speed served only by local generation and the lines north
reclose period (typically less than 30 cycles, or one of Lake Erie connecting Detroit east to Ontario and
half second, to allow the air to de-ionize after the the lines south of Lake Erie from Cleveland east to
trip to prevent arc re-ignition). To address this and northwest Pennsylvania. These trips completed
protect generators from the harm that an the extra-high voltage network separation
out-of-synchronism reconnect could cause, it is between eastern and western Michigan.
worth studying whether a synchro-check relay is
needed, to reclose the second breaker only when The Power System Disturbance Recorders at Keith
the two ends are within a certain voltage and and Lambton, Ontario, captured these events in
phase angle tolerance. No such protection was the flows across the Ontario-Michigan interface,
installed at Argenta-Battle Creek; when the line as shown in Figure 6.12 and Figure 6.16. It shows
reclosed, there was a 70o difference in phase clearly that the west to east Michigan separation
across the circuit breaker reclosing the line. There (the Thetford-Jewell trip) was the start and Erie
West-Ashtabula-Perry was the trigger for the 3,700
MW surge from Ontario into Michigan. When
Figure 6.11. Transmission and Generation Trips in Thetford-Jewell tripped, power that had been
Michigan, 16:10:36 to 16:10:37 EDT flowing into Michigan and Ohio from western
Michigan, western Ohio and Indiana was cut off.
The nearby Ontario recorders saw a pronounced
impact as flows into Detroit readjusted to draw
power from the northeast instead. To the south,
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 83
Erie West-Ashtabula-Perry was the last 345-kV 6C) Cleveland separated from Pennsylvania,
eastern link for northern Ohio loads. When that flows reversed and a huge power surge
line severed, all the power that moments before flowed counter-clockwise around Lake Erie:
had flowed across Michigan and Ohio paths was 16:10:38.6 EDT
now diverted in a counter-clockwise direction 16:10:38.6 EDT: Erie West-Ashtabula-Perry
around Lake Erie through the single path left in 345-kV line tripped at Perry
eastern Michigan, pulling power out of Ontario, 16:10:38.6 EDT: Large power surge to serve
New York and PJM. loads in eastern Michigan and northern Ohio
Figures 6.13 and 6.14 show the results of investi- swept across Pennsylvania, New Jersey, and
gation team modeling of the line loadings on the New York through Ontario into Michigan.
Ohio, Michigan, and other regional interfaces for Perry-Ashtabula was the last 345-kV line connect-
the period between 16:05:57 until the Thetford- ing northern Ohio to the east south of Lake Erie.
Jewell trip, to understand how power flows shifted This line’s trip at the Perry substation on a zone 3
during this period. The team simulated evolving relay operation separated the northern Ohio
system conditions on August 14, 2003, based on 345-kV transmission system from Pennsylvania
the 16:05:50 power flow case developed by the and all eastern 345-kV connections. After this trip,
MAAC-ECAR-NPCC Operations Studies Working the load centers in eastern Michigan and northern
Group. Each horizontal line in the graph indicates Ohio (Detroit, Cleveland, and Akron) remained
a single or set of 345-kV lines and its loading as a connected to the rest of the Eastern Interconnec-
function of normal ratings over time as first one, tion only to the north at the interface between the
then another, set of circuits tripped out of service. Michigan and Ontario systems (Figure 6.15). East-
In general, each subsequent line trip causes the ern Michigan and northern Ohio now had little
remaining line loadings to rise; where a line drops internal generation left and voltage was declining.
(as Erie West-Ashtabula-Perry in Figure 6.13 after The frequency in the Cleveland area dropped rap-
the Hanna-Juniper trip), that indicates that line idly, and between 16:10:39 and 16:10:50 EDT
loading lightened, most likely due to customers under-frequency load shedding in the Cleveland
dropped from service. Note that Muskingum and area interrupted about 1,750 MW of load. How-
East Lima-Fostoria Central were overloaded before ever, the load shedding did not drop enough load
they tripped, but the Michigan west and north relative to local generation to rebalance and arrest
interfaces were not overloaded before they trip- the frequency decline. Since the electrical system
ped. Erie West-Ashtabula-Perry was loaded to always seeks to balance load and generation, the
130% after the Hampton-Pontiac and Thetford- high loads in Detroit and Cleveland drew power
Jewell trips. over the only major transmission path remain-
The Regional Interface Loadings graph (Figure ing—the lines from eastern Michigan into Ontario.
6.14) shows that loadings at the interfaces Mismatches between generation and load are
between PJM-NY, NY-Ontario and NY-New Eng- reflected in changes in frequency, so with more
land were well within normal ratings before the generation than load frequency rises and with less
east-west Michigan separation. generation than load, frequency falls.
Figure 6.13. Simulated 345-kV Line Loadings from Figure 6.14. Simulated Regional Interface Loadings
16:05:57 through 16:10:38.4 EDT from 16:05:57 through 16:10:38.4 EDT
84 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
At 16:10:38.6 EDT, after the above transmission and black out (as evidenced by the rapid power
paths into Michigan and Ohio failed, the power oscillations decaying after 16:10:43 EDT).
that had been flowing at modest levels into Michi-
gan from Ontario suddenly jumped in magnitude.
While flows from Ontario into Michigan had been Figure 6.15. Michigan Lines Trip and Ohio
Separates from Pennsylvania, 16:10:36 to
in the 250 to 350 MW range since 16:10:09.06
16:10:38.6 EDT
EDT, with this new surge they peaked at 3,700
MW at 16:10:39 EDT (Figure 6.16). Electricity
moved along a giant loop through Pennsylvania
and into New York and Ontario and then into
Michigan via the remaining transmission path to
6B
serve the combined loads of Cleveland, Toledo,
and Detroit. This sudden large change in power
flows drastically lowered voltage and increased
current levels on the transmission lines along the
Pennsylvania-New York transmission interface. 6C
This was a power surge of large magnitude, so fre-
quency was not the same across the Eastern Inter-
connection. As Figure 6.16 shows, the power
swing resulted in a rapid rate of voltage decay.
Flows into Detroit exceeded 3,700 MW and 1,500
MVAr—the power surge was draining real power
out of the northeast, causing voltages in Ontario
and New York to drop. At the same time, local
voltages in the Detroit area were plummeting
because Detroit had already lost 500 MW of local
generation. Detroit would soon lose synchronism
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 85
Just before the Argenta-Battle Creek trip, when cases, this adjustment was unsuccessful and the
Michigan separated west to east at 16:10:37 EDT, generators tripped out from milliseconds to sev-
almost all of the generators in the eastern intercon- eral seconds thereafter.
nection were moving in synchronism with the
The Perry-Ashtabula-Erie West 345-kV line trip at
overall grid frequency of 60 Hertz (shown at the
16:10:38.6 EDT was the point when the Northeast
bottom of Figure 6.17), but when the swing
entered a period of transient instability and a loss
started, those machines absorbed some of its ener-
of generator synchronism. Between 16:10:38 and
gy as they attempted to adjust and resynchronize
16:10:41 EDT, the power swings caused a sudden
with the rapidly changing frequency. In many
extraordinary increase in system frequency, hit-
ting 60.7 Hz at Lambton and 60.4 Hz at Niagara.
Figure 6.16. Active and Reactive Power and Voltage Because the demand for power in Michigan, Ohio,
from Ontario into Detroit and Ontario was drawing on lines through New
York and Pennsylvania, heavy power flows were
moving northward from New Jersey over the New
York tie lines to meet those power demands, exac-
erbating the power swing. Figure 6.17 shows
actual net line flows summed across the interfaces
between the main regions affected by these
swings—Ontario into Michigan, New York into
Ontario, New York into New England, and PJM
into New York. This shows clearly that the power
swings did not move in unison across every inter-
face at every moment, but varied in magnitude
and direction. This occurred for two reasons. First,
the availability of lines to complete the path across
Figure 6.17. Measured Power Flows and Frequency Across Regional Interfaces, 16:10:30 to 16:11:00 EDT,
with Key Events in the Cascade
86 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
each interface varied over time, as did the amount Majestic 345-kV Substation: one terminal
of load that drew upon each interface, so net flows opened sequentially on all 345-kV lines
across each interface were not facing consistent 16:10:41.8 EDT: Fostoria Central-Galion
demand with consistent capability as the cascade 345-kV line
progressed. Second, the speed and magnitude of
the swing was moderated by the inertia, reactive 16:10:41.911 EDT: Beaver-Davis Besse
power capabilities, loading conditions and loca- 345-kV line
tions of the generators across the entire region. Under-frequency load-shedding in Ohio:
After Cleveland was cut off from Pennsylvania
FirstEnergy shed 1,754 MVA load
and eastern power sources, Figure 6.17 shows the
start of the dynamic power swing at 16:10:38.6. AEP shed 133 MVA load
Because the loads of Cleveland, Toledo and
Detroit (less the load already blacked out) were Seven power plants, for a total of 3,294 MW of
now hanging off Michigan and Ontario, this forced generation, tripped off-line in Ohio:
a gigantic shift in power flows to meet that 16:10:42 EDT: Bay Shore Units 1-4 (551 MW
demand. As noted above, flows from Ontario into near Toledo) tripped on over-excitation
Michigan increased from 1,000 MW to 3,700 MW
shortly after the start of the swing, while flows 16:10:40 EDT: Lakeshore unit 18 (156 MW,
from PJM into New York were close behind. But near Cleveland) tripped on under-frequency
within two seconds from the start of the swing, at 16:10:41.7 EDT: Eastlake 1, 2, and 3 units
16:10:40 EDT flows reversed and coursed back (304 MW total, near Cleveland) tripped on
from Michigan into Ontario at the same time that under-frequency
frequency at the interface dropped, indicating that
significant generation had been lost. Flows that 16:10:41.7 EDT: Avon Lake unit 9 (580 MW,
had been westbound across the Ontario-Michigan near Cleveland) tripped on under-frequency
interface by over 3,700 MW at 16:10:38.8 dropped 16:10:41.7 EDT: Perry 1 nuclear unit (1,223
down to 2,100 MW eastbound by 16:10:40, and MW, near Cleveland) tripped on under-
then returned westbound starting at 16:10:40.5. frequency
A series of circuits tripped along the border
between PJM and the NYISO due to zone 1 imped- 16:10:42 EDT: Ashtabula unit 5 (184 MW,
ance relay operations on overload and depressed near Cleveland) tripped on under-frequency
voltage. The surge also moved into New England 16:10:43 EDT: West Lorain units (296 MW)
and the Maritimes region of Canada. The combi- tripped on under-voltage
nation of the power surge and frequency rise
caused 380 MW of pre-selected Maritimes genera- Four power plants producing 1,759 MW tripped
tion to drop off-line due to the operation of the off-line near Detroit:
New Brunswick Power “Loss of Line 3001” Special 16:10:42 EDT: Greenwood unit 1 tripped (253
Protection System. Although this system was MW) on low voltage, high current
designed to respond to failure of the 345-kV link
between the Maritimes and New England, it oper- 16:10:41 EDT: Belle River unit 1 tripped (637
ated in response to the effects of the power surge. MW) on out-of-step
The link remained intact during the event. 16:10:41 EDT: St. Clair unit 7 tripped (221
6D) Conditions in Northern Ohio and Eastern MW, DTE unit) on high voltage
Michigan Degraded Further, With More
Transmission Lines and Power Plants Fail- 16:10:42 EDT: Trenton Channel units 7A, 8
ing: 16:10:39 to 16:10:46 EDT and 9 tripped (648 MW)
Line trips in Ohio and eastern Michigan: Back in northern Ohio, the trips of the Bay
16:10:39.5 EDT: Bay Shore-Monroe 345-kV Shore-Monroe, Majestic-Lemoyne, Allen Junc-
line tion-Majestic-Monroe 345-kV lines, and the
Ashtabula 345/138-kV transformer cut off Toledo
16:10:39.6 EDT: Allen Junction-Majestic- and Cleveland from the north, turning that area
Monroe 345-kV line into an electrical island (Figure 6.18). Frequency
16:10:40.0 EDT: Majestic-Lemoyne 345-kV in this large island began to fall rapidly. This
line caused a series of power plants in the area to trip
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 87
off-line due to the operation of under-frequency running at maximum mechanical output, they
relays, including the Bay Shore units. When the began to pull out of synchronous operation with
Beaver-Davis Besse 345-kV line between Cleve- the rest of the grid. When voltage in Detroit
land and Toledo tripped, it left the Cleveland area returned to near-normal, the generators could not
completely isolated and area frequency rapidly fully pull back its rate of revolutions, and ended
declined. Cleveland area load was disconnected up producing excessive temporary output levels,
by automatic under-frequency load-shedding still out of step with the system. This is evident in
(approximately 1,300 MW), and another 434 MW Figure 6.19, which shows at least two sets of gen-
of load was interrupted after the generation erator “pole slips” by plants in the Detroit area
remaining within this transmission “island” was between 16:10:40 EDT and 16:10:42 EDT. Several
tripped by under-frequency relays. This sudden large units around Detroit—Belle River, St. Clair,
load drop would contribute to the reverse power Greenwood, Monroe, and Fermi—all tripped in
swing. In its own island, portions of Toledo response. After formation of the Cleveland-Toledo
blacked out from automatic under-frequency island at 16:10:40 EDT, Detroit frequency spiked
load-shedding but most of the Toledo load was to almost 61.7 Hz before dropping, momentarily
restored by automatic reclosing of lines such as equalized between the Detroit and Ontario sys-
the East Lima-Fostoria Central 345-kV line and tems, but Detroit frequency began to decay at 2
several lines at the Majestic 345-kV substation. Hz/sec and the generators then experienced
under-speed conditions.
The Perry nuclear plant is in Ohio on Lake Erie,
not far from the Pennsylvania border. The Perry Re-examination of Figure 6.17 shows the power
plant was inside a decaying electrical island, swing from the northeast through Ontario into
and the plant tripped on under-frequency, as Michigan and northern Ohio that began at
designed. A number of other units near Cleveland 16:10:37, and how it reverses and swings back
tripped off-line by under-frequency protection. around Lake Erie at 16:10:39 EDT. That return was
caused by the combination of natural oscillations,
The tremendous power flow into Michigan, begin- accelerated by major load losses, as the northern
ning at 16:10:38, occurred when Toledo and Ohio system disconnected from Michigan. It
Cleveland were still connected to the grid only caused a power flow change of 5,800 MW, from
through Detroit. After the Bay Shore-Monroe line 3,700 MW westbound to 2,100 eastbound across
tripped at 16:10:39, Toledo-Cleveland were sepa- the Ontario to Michigan border between
rated into their own island, dropping a large 16:10:39.5 and 16:10:40 EDT. Since the system
amount of load off the Detroit system. This left was now fully dynamic, this large oscillation east-
Detroit suddenly with excess generation, much of bound would lead naturally to a rebound, which
which was greatly accelerated in angle as the began at 16:10:40 EDT with an inflection point
depressed voltage in Detroit (caused by the high reflecting generation shifts between Michigan and
demand in Cleveland) caused the Detroit units to Ontario and additional line losses in Ohio.
pull nearly out of step. With the Detroit generators
Figure 6.18. Cleveland and Toledo Islanded, Figure 6.19. Generators Under Stress in Detroit,
16:10:39 to 16:10:46 EDT as Seen from Keith PSDR
6D
88 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Western Pennsylvania Separated from New small area so a power swing is less likely to enter
York: 16:10:39 EDT to 16:10:44 EDT the relay target circle at all, averting a zone 1 trip.
6E) 16:10:39 EDT, Homer City-Watercure Road At 16:10:44 EDT, the northern part of the Eastern
345 kV Interconnection (including eastern Michigan) was
16:10:39 EDT: Homer City-Stolle Road 345 connected to the rest of the Interconnection at
kV only two locations: (1) in the east through the
500-kV and 230-kV ties between New York and
6F) 16:10:44 EDT: South Ripley-Erie East 230 kV,
northeast New Jersey, and (2) in the west through
and South Ripley-Dunkirk 230 kV
the long and electrically fragile 230-kV transmis-
16:10:44 EDT: East Towanda-Hillside 230 kV sion path connecting Ontario to Manitoba and
Responding to the swing of power out of Michigan Minnesota. The separation of New York from
toward Ontario and into New York and PJM, zone Pennsylvania (leaving only the lines from New Jer-
1 relays on the 345-kV lines separated Pennsylva- sey into New York connecting PJM to the north-
nia from New York (Figure 6.20). Homer east) buffered PJM in part from these swings.
City-Watercure (177 miles or 285 km) and Homer Frequency was high in Ontario at that point, indi-
City-Stolle Road (207 miles or 333 km) are very cating that there was more generation than load,
long lines and so have high impedance. Zone 1 so much of this flow reversal never got past
relays do not have timers, and operate instantly Ontario into New York.
when a power swing enters the relay target circle. 6G) Transmission paths disconnected in New
For normal length lines, zone 1 relays have small Jersey and northern Ontario, isolating the
target circles because the relay is measuring a less northeast portion of the Eastern
than the full length of the line—but for a long line Interconnection: 16:10:43 to 16:10:45 EDT
the large line impedance enlarges the relay’s target
circle and makes it more likely to be hit by the 16:10:43 EDT: Keith-Waterman 230-kV line
power swing. The Homer City-Watercure and tripped
Homer City-Stolle Road lines do not have zone 3 16:10:45 EDT: Wawa-Marathon 230-kV lines
relays. tripped
Given the length and impedance of these lines, it 16:10:45 EDT: Branchburg-Ramapo 500-kV line
was highly likely that they would trip and separate tripped
early in the face of such large power swings. Most At 16:10:43 EDT, eastern Michigan was still con-
of the other interfaces between regions are on nected to Ontario, but the Keith-Waterman
short ties—for instance, the ties between New 230-kV line that forms part of that interface dis-
York and Ontario and Ontario to Michigan are connected due to apparent impedance (Figure
only about 2 miles (3.2 km) long, so they are elec- 6.21). This put more power onto the remaining
trically very short and thus have much lower interface between Ontario and Michigan, but
impedance and trip less easily than these long
lines. A zone 1 relay target for a short line covers a
Figure 6.21. Northeast Separates from Eastern
Interconnection, 16:10:45 EDT
Figure 6.20. Western Pennsylvania Separates from
New York, 16:10:39 EDT to 16:10:44 EDT
6F 6G
6E
6F
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 89
triggered sustained oscillations in both power New Jersey, the separation occurred along the
flow and frequency along the remaining 230-kV 230-kV corridors which are the main supply feeds
line. into the northern New Jersey area (the two
At 16:10:45 EDT, northwest Ontario separated Roseland-Athenia circuits and the Lin-
from the rest of Ontario when the Wawa-Marathon den-Bayway circuit). These circuits supply the
230-kV lines (104 miles or 168 km long) discon- large customer load in northern New Jersey and
nected along the northern shore of Lake Superior, are a primary route for power transfers into New
tripped by zone 1 distance relays at both ends. York City, so they are usually more highly loaded
This separation left the loads in the far northwest than other interfaces. These lines tripped west and
portion of Ontario connected to the Manitoba and south of the large customer loads in northeast New
Minnesota systems, and protected them from the Jersey.
blackout. The separation of New York, Ontario, and New
The 69-mile (111 km) long Branchburg-Ramapo England from the rest of the Eastern Interconnec-
500-kV line and Ramapo transformer between tion occurred due to natural breaks in the system
New Jersey and New York was the last major trans- and automatic relay operations, which performed
mission path remaining between the Eastern Inter- exactly as they were designed to. No human inter-
connection and the area ultimately affected by the vention occurred by operators at PJM headquar-
blackout. Figure 6.22 shows how that line discon- ters or elsewhere to effect this split. At this point,
nected at 16:10:45 EDT, along with other underly- the Eastern Interconnection was divided into two
ing 230 and 138-kV lines in northeast New Jersey. major sections. To the north and east of the separa-
Branchburg–Ramapo was carrying over 3,000 tion point lay New York City, northern New Jer-
MVA and 4,500 amps with voltage at 79% before it sey, New York state, New England, the Canadian
tripped, either on a high-speed swing into zone 1 Maritime Provinces, eastern Michigan, the major-
or on a direct transfer trip. The investigation team ity of Ontario, and the Québec system.
is still examining why the higher impedance The rest of the Eastern Interconnection, to the
230-kV overhead lines tripped while the under- south and west of the separation boundary, was
ground Hudson-Farragut 230-kV cables did not; not seriously affected by the blackout. Frequency
the available data suggest that the notably lower in the Eastern Interconnection was 60.3 Hz at the
impedance of underground cables made these less time of separation; this means that approximately
vulnerable to the electrical strain placed on the 3,700 MW of excess generation that was on-line to
system. export into the northeast was now in the main
This left the northeast portion of New Jersey con- Eastern Island, separated from the load it had been
nected to New York, while Pennsylvania and the serving. This left the northeast island with even
rest of New Jersey remained connected to the rest less in-island generation on-line as it attempted to
of the Eastern Interconnection. Within northeast rebalance in the next phase of the cascade.
90 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
7C) The Ontario system just west of Niagara Falls measure leaves the generator in good condition to
and west of St. Lawrence separated from the help rebuild the system once a blackout is over
western New York island: 16:10:50 EDT and restoration begins. When unstable power
swings develop between a group of generators that
7D) Southwest Connecticut separated from New
are losing synchronization (unable to match fre-
York City: 16:11:22 EDT
quency) with the rest of the system, one effective
7E) Remaining transmission lines between way to stop the oscillations is to stop the flows
Ontario and eastern Michigan separated: entirely by disconnecting the unstable generators
16:11:57 EDT from the remainder of the system. The most com-
mon way to protect generators from power oscilla-
By this point most portions of the affected area tions is for the transmission system to detect the
were blacked out. power swings and trip at the locations detecting
If the 6th phase of the cascade was about dynamic the swings—ideally before the swing reaches criti-
system oscillations, the last phase is a story of the cal levels and harms the generator or the system.
search for balance between loads and generation.
On August 14, the cascade became a race between
Here it is necessary to understand three matters
the power surges and the relays. The lines that
related to system protection—why the blackout
tripped first were generally the longer lines with
stopped where it did, how and why under-voltage
relay settings using longer apparent impedance
and under-frequency load-shedding work, and
tripping zones and normal time settings. On
what happened to the generators on August 14
August 14, relays on long lines such as the Homer
and why. These matter because loads and genera-
City-Watercure and the Homer City-Stolle Road
tion must ultimately balance in real-time to
345-kV lines in Pennsylvania, that are not highly
remain stable. When the grid is breaking apart into
integrated into the electrical network, tripped
islands, if generators stay on-line longer, then the
quickly and split the grid between the sections
better the chances to keep the lights on within
that blacked out and those that recovered without
each island and restore service following a black-
further propagating the cascade. This same phe-
out; so automatic load-shedding, transmission
nomenon was seen in the Pacific Northwest black-
relay protections and generator protections must
outs of 1996, when long lines tripped before more
avoid premature tripping. They must all be coordi-
networked, electrically supported lines.
nated to reduce the likelihood of system break-up,
and once break-up occurs, to maximize an island’s Transmission line voltage divided by its current
chances for electrical survival. flow is called “apparent impedance.” Standard
transmission line protective relays continuously
measure apparent impedance. When apparent
Why the Blackout Stopped impedance drops within the line’s protective relay
Where It Did set-points for a given period of time, the relays trip
Extreme system conditions can damage equip-
Figure 6.23. New York and New England Separate,
ment in several ways, from melting aluminum Multiple Islands Form
conductors (excessive currents) to breaking tur-
bine blades on a generator (frequency excursions).
The power system is designed to ensure that if
conditions on the grid (excessive or inadequate
voltage, apparent impedance or frequency)
threaten the safe operation of the transmission 7A
lines, transformers, or power plants, the threat-
ened equipment automatically separates from the
7B
network to protect itself from physical damage. 7C
Relays are the devices that effect this protection.
7E 7D
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 91
the line. The vast majority of trip operations on team has concluded that such schemes would
lines along the blackout boundaries between PJM have been ineffective on August 14.
and New York (for instance) show high-speed
relay targets which indicate that a massive power
surge caused each line to trip. To the relays, this Under-Frequency and
power surge altered the voltages and currents Under-Voltage Load-Shedding
enough that they appeared to be faults. The power
surge was caused by power flowing to those areas Automatic load-shedding measures are designed
that were generation-deficient (Cleveland, Toledo into the electrical system to operate as a last resort,
and Detroit) or rebounding back. These flows under the theory that it is wise to shed some load
occurred purely because of the physics of power in a controlled fashion if it can forestall the loss of
flows, with no regard to whether the power flow a great deal of load to an uncontrollable cause.
had been scheduled, because power flows from Thus there are two kinds of automatic load-shed-
areas with excess generation into areas that were ding installed in North America—under-voltage
generation-deficient. load-shedding, which sheds load to prevent local
area voltage collapse, and under-frequency load-
Protective relay settings on transmission lines shedding, which is designed to rebalance load and
operated as they were designed and set to behave generation within an electrical island once it has
on August 14. In some cases line relays did not trip been created by a system disturbance.
in the path of a power surge because the apparent
impedance on the line was not low enough—not Automatic under-voltage load-shedding (UVLS)
because of the magnitude of the current, but rather responds directly to voltage conditions in a local
because voltage on that line was high enough that area. UVLS drops several hundred MW of load in
the resulting impedance was adequate to avoid pre-selected blocks within urban load centers,
entering the relay’s target zone. Thus relative volt- triggered in stages when local voltage drops to a
age levels across the northeast also affected which designated level—likely 89 to 92% or even
areas blacked out and which areas stayed on-line. higher—with a several second delay. The goal of a
UVLS scheme is to eliminate load in order to
In the U.S. Midwest, as voltage levels declined restore reactive power relative to demand, to pre-
many generators in the affected area were operat- vent voltage collapse and contain a voltage prob-
ing at maximum reactive power output before the lem within a local area rather than allowing it to
blackout. This left the system little slack to deal spread in geography and magnitude. If the first
with the low voltage conditions by ramping up load-shed step does not allow the system to
more generators to higher reactive power output rebalance, and voltage continues to deteriorate,
levels, so there was little room to absorb any sys- then the next block of UVLS is dropped. Use of
tem “bumps” in voltage or frequency. In contrast, UVLS is not mandatory, but is done at the option
in the northeast—particularly PJM, New York, and of the control area and/or reliability council. UVLS
ISO-New England—operators were anticipating schemes and trigger points should be designed to
high power demands on the afternoon of August respect the local area’s sys-
14, and had already set up the system to maintain tem vulnerabilities, based Recommendation
higher voltage levels and therefore had more reac- 21, page 158
on voltage collapse studies.
tive reserves on-line in anticipation of later after- As noted in Chapter 4, there
noon needs. Thus, when the voltage and is no UVLS system in place within Cleveland and
frequency swings began, these systems had reac- Akron; had such a scheme been implemented
tive power readily available to help buffer their before August, 2003, shedding 1,500 MW of load
areas against potential voltage collapse without in that area before the loss of the Sammis-Star line
widespread generation trips. might have prevented the cascade and blackout.
The investigation team has used simulation to In contrast to UVLS, automatic under-frequency
examine whether special protection schemes, load-shedding (UFLS) is designed for use in
designed to detect an impending cascade and sep- extreme conditions to stabilize the balance
arate the grid at specific interfaces, could have between generation and load after an electrical
been or should be set up to stop a power surge and island has been formed, dropping enough load to
prevent it from sweeping through an interconnec- allow frequency to stabilize within the island.
tion and causing the breadth of line and generator All synchronous generators in North America
trips and islanding that occurred that day. The are designed to operate at 60 cycles per second
92 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
(Hertz) and frequency reflects how well load and It must be emphasized that the entire northeast
generation are balanced—if there is more load system was experiencing large scale, dynamic
than generation at any moment, frequency drops oscillations in this period. Even if the UFLS and
below 60 Hz, and it rises above that level if there is generation had been perfectly balanced at any
more generation than load. By dropping load to moment in time, these oscillations would have
match available generation within the island, made stabilization difficult and unlikely.
UFLS is a safety net that helps to prevent the com-
plete blackout of the island, which allows faster Why the Generators Tripped Off
system restoration afterward. UFLS is not effective
if there is electrical instability or voltage collapse At least 265 power plants with more than 508 indi-
within the island. vidual generating units shut down in the August
14 blackout. These U.S. and Canadian plants can
Today, UFLS installation is a NERC requirement, be categorized as follows:
designed to shed at least 25-30% of the load in
steps within each reliability coordinator region. By reliability coordination area:
These systems are designed to drop pre-desig- u Hydro Québec, 5 plants (all isolated onto the
nated customer load automatically if frequency Ontario system)4
gets too low (since low frequency indicates too lit- u Ontario, 92 plants
tle generation relative to load), starting generally
u ISO-New England, 31 plants
when frequency reaches 59.3 Hz. Progressively
more load is set to drop as frequency levels fall far- u MISO, 32 plants
ther. The last step of customer load shedding is set u New York ISO, 70 plants
at the frequency level just above the set point for u PJM, 35 plants
generation under-frequency protection relays
(57.5 Hz), to prevent frequency from falling so low By type:
that generators could be damaged (see Figure 2.4). u Conventional steam units, 66 plants (37 coal)
In NPCC, following the Northeast blackout of u Combustion turbines, 70 plants (37 combined
1965, the region adopted automatic under-fre- cycle)
quency load-shedding criteria and manual load- u Nuclear, 10 plants—7 U.S. and 3 Canadian,
shedding within ten minutes to prevent a recur- totaling 19 units (the nuclear unit outages are
rence of the cascade and better protect system discussed in Chapter 8)
equipment from damage due to a high-speed sys- u Hydro, 101
tem collapse. Under-frequency load-shedding u Other, 18.
triggers vary by regional reliability council—New
York and all of the Northeast Power Coordinating Within the overall cascade sequence, 29 (6%) gen-
Council, plus the Mid-Atlantic Area Council use erators tripped between the start of the cascade at
59.3 Hz as the first step for UFLS, while ECAR 16:05:57 (the Sammis-Star trip) and the split
uses 59.5 Hz as their first step for UFLS. between Ohio and Pennsylvania at 16:10:38.6
EDT (Erie West-Ashtabula-Perry), which triggered
The following automatic UFLS operated on the the first big power swing. These trips were caused
afternoon of August 14: by the generators’ protective relays responding to
u Ohio shed over 1,883 MVA beginning at overloaded transmission lines, so many of these
16:10:39 EDT trips were reported as under-voltage or over-
current. The next interval in the cascade was as
u Michigan shed a total of 2,835 MW the portions of the grid lost synchronism, from
u New York shed a total of 10,648 MW in numer- 16:10:38.6 until 16:10:45.2 EDT, when Michi-
ous steps, beginning at 16:10:48 gan-New York-Ontario-New England separated
from the rest of the Eastern Interconnection. Fifty
u PJM shed a total of 1,324 MVA in 3 steps in more generators (10%) tripped as the islands
northern New Jersey beginning at 16:10:48 EDT formed, particularly due to changes in configura-
u Ontario shed a total of 7,800 MW in 2 steps, tion, loss of synchronism, excitation system
beginning at 16:10:4 failures, with some under-frequency and under-
voltage. In the third phase of generator losses, 431
u New England shed a total of 1,098 MW. generators (84%) tripped after the islands formed,
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 93
many at the same time that under-frequency it back again as some frequencies changed several
load-shedding was occurring. This is illustrated in times by as much as 3 Hz (about 100 times normal
Figure 6.24. It is worth noting, however, that many deviations). Figure 6.25 is a plot of the MW output
generators did not trip instantly after the trigger and frequency for one large unit that nearly sur-
condition that led to the trip—rather, many relay vived the disruption but tripped when in-plant
protective devices operate on time delays of milli- hydraulic control pressure limits were eventually
seconds to seconds in duration, so that a generator violated. After the plant control system called for
that reported tripping at 16:10:43 on under- shutdown, the turbine control valves closed and
voltage or “generator protection” might have expe- the generator electrical output ramped down to a
rienced the trigger for that condition several sec- preset value before the field excitation tripped and
onds earlier. the generator breakers opened to disconnect the
The high number of generators that tripped before unit from the system. This also illustrates the time
formation of the islands helps to explain why so lag between system events and the generator reac-
much of the northeast blacked out on August 14— tion—this generator was first disturbed by system
many generators had pre-designed protection conditions at 16:10:37, but did not trip until
points that shut the unit down early in the cas- 16:11:47, over a minute later.
cade, so there were fewer units on-line to prevent Under-frequency (10% of the generators report-
island formation or to maintain balance between ing) and under-voltage (6%) trips both reflect
load and supply within responses to system conditions. Although com-
Recommendation
each island after it formed. 21, page 158 bustion turbines in particular are designed with
In particular, it appears that under-voltage relay protection, it is not clear why
some generators tripped to protect the units from this is needed. An under-voltage condition by
conditions that did not justify their protection, itself and over a set time period may not necessar-
and many others were set to trip in ways that were ily be a generator hazard (although it could affect
not coordinated with the region’s under-frequency plant auxiliary systems). Some generator under-
load-shedding, rendering that UFLS scheme less voltage relays were set to trip at or above 90% volt-
effective. Both factors compromised successful age. However, a motor stalls out at about 70% volt-
islanding and precipitated the blackouts in age and a motor starter contactor drops out around
Ontario and New York. 75%, so if there is a compelling need to protect the
Most of the unit separations fell in the category of turbine from the system the under-voltage trigger
consequential tripping—they tripped off-line in point should be no higher than 80%.
response to some outside condition on the grid, An excitation failure is closely related to a voltage
not because of any problem internal to the plant. trip. As local voltages decreased, so did frequency.
Some generators became completely removed Over-excitation operates on a calculation of
from all loads; because the fundamental operating volts/hertz, so as frequency declines faster than
principle of the grid is that load and generation voltage over-excitation relays would operate. It is
must balance, if there was no load to be served the not clear that these relays were coordinated with
power plant shut down in response to over-speed each machine’s exciter controls, to be sure that it
and/or over-voltage protection schemes. Others was protecting the machine for the proper range of
were overwhelmed because they were among a its control capabilities. Large units have two relays
few power plants within an electrical island, and to detect volts/Hz—one at the generator and one at
were suddenly called on to serve huge customer the transformer, each with a slightly different
loads, so the imbalance caused them to trip on volts/Hz setting and time delay. It is possible that
under-frequency and/or under-voltage protection. these settings can cause a generator to trip within
A few were tripped by special protection schemes a generation-deficient island as frequency is
that activated on excessive frequency or loss of attempting to rebalance, so these settings should
pre-studied major transmission elements known be carefully evaluated.
to require large blocks of generation rejection.
The Eastlake 5 trip at 13:31 EDT was an excitation
The large power swings and excursions of system
system failure—as voltage fell at the generator
frequency put all the units in their path through a
bus, the generator tried to increase quickly its pro-
sequence of major disturbances that shocked sev-
duction of voltage on the AC winding of the
eral units into tripping. Plant controls had actu-
machine quickly. This caused the generator’s exci-
ated fast governor action on several of these to turn
tation protection scheme to trip the plant off to
back the throttle, then turn it forward, only to turn
94 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Figure 6.24. Generator Trips by Time and Cause
Sammis-Star to Cleveland split from PA Ontario split from West New York to final Ontario separation
Cleveland split to Northeast separation from Eastern Interconnection After all the separations
Northeast separation to first Ontario split from West New York All generator trips
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 95
protect its windings and coils from over-heating. of power off the grid to support the unit’s orderly
Several of the other generators which tripped slow-down and trip due to reverse power flow.
early in the cascade came off under similar cir- This is a standard practice to avoid turbine
cumstances as excitation systems were over- over-speed. Also within New York, 16 gas turbines
stressed to hold voltages up. Seventeen generators totaling about 400 MW reported tripping for loss
reported tripping for over-excitation. Units that of fuel supply, termed “flame out.” These units’
trip for a cause related to frequency should be trips should be better understood.
evaluated to determine how the unit frequency
triggers coordinate with the region’s under-fre- Another reason for power plant trips was actions
quency load-shedding scheme, to assure that the or failures of plant control systems. One common
generator trips are sequenced to follow rather than cause in this category was a loss of sufficient volt-
precede load-shedding. After UFLS operates to age to in-plant loads. Some plants run their inter-
drop a large block of load, frequency continues to nal cooling and processes (house electrical load)
decline for several cycles before rebounding, so it off the generator or off small, in-house auxiliary
is necessary to design an adequate time delay into generators, while others take their power off the
generators’ frequency-related protections to keep main grid. When large power swings or voltage
it on-line long enough to help rebalance against drops reached these plants in the latter category,
the remaining load. they tripped off-line because the grid could not
supply the plant’s in-house power needs reliably.
Fourteen generators reported tripping for under- At least 17 units reported tripping due to loss of
excitation (also known as loss of field), which pro- system configuration, including the loss of a trans-
tects the generator from exciter component fail- mission or distribution line
ures. This protection scheme can operate on stable to serve the in-plant loads. Recommendation
11, page 148
as well as transient power swings, so should be Some generators were trip-
examined to determine whether the protection ped by their operators.
settings are appropriate. Eighteen units—primar-
ily combustion turbines—reported over-current as Unfortunately, 40% of the generators that went
the reason for relay operation. off-line during or after the cascade did not provide
useful information on the cause of tripping in their
Some generators in New York failed in a way that response to the NERC investigation data request.
exacerbated frequency decay. A generator that While the responses available offer significant and
tripped due to a boiler or steam problem may have valid information, the investigation team will
done so to prevent damage due to over-speed and never be able to fully analyze and explain why so
limit impact to the turbine-generator shaft when many generators tripped off-line so early in the
the breakers are opened, and it will attempt to cascade, contributing to the speed and extent of
maintain its synchronous speed until the genera- the blackout. It is clear that every generator should
tor is tripped. To do this, the mechanical part of have some minimum of protection for stator dif-
the system would shut off the steam flow. This ferential, loss of field, and out-of-step protection,
causes the generator to consume a small amount to disconnect the unit from the grid when it is not
performing correctly, and also protection for pro-
Figure 6.25. Events at One Large Generator During tect the generator from extreme conditions on the
the Cascade grid that could cause catastrophic damage to the
generator. These protections should be set tight
enough to protect the unit from the grid, but also
wide enough to assure that the unit remains con-
nected to the grid as long as possible. This coordi-
nation is a risk management issue that must
balance the needs of the grid
Recommendation
and customers relative to 21, page 158
the needs of the individual
assets.
96 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
grew within New York City and neighboring from the eastern New York island. The only
northern New Jersey, the utilities serving the area remaining tie was the PV-20 circuit connecting
deliberately increased the integration between the New England and the western New York island,
systems serving this area to increase the flow which tripped at 16:10:54 EDT. Because New Eng-
capability into New York and the reliability of the land was exporting to New York before the distur-
system as a whole. The combination of the facili- bance across the southwest Connecticut tie, but
ties in place and the pattern of electrical loads and importing on the Northwalk-Northport tie, the
flows on August 14 caused New York to be tightly Pleasant Valley path opened east of Long Moun-
linked electrically to northern New Jersey and tain—in other words, internal to southwest Con-
southwest Connecticut, and moved the weak necticut—rather than along the actual New
spots on the grid out past this combined load and York-New England tie.5 Immediately before the
network area. separation, the power swing out of New England
occurred because the New England generators had
Figure 6.26 gives an overview of the power flows increased output in response to the drag of power
and frequencies in the period 16:10:45 EDT through Ontario and New York into Michigan and
through 16:11:00 EDT, capturing most of the key Ohio.6 The power swings continuing through the
events in Phase 7. region caused this separation, and caused Ver-
7A) New York-New England Transmission mont to lose approximately 70 MW of load.
Lines Disconnected: 16:10:46 to 16:10:54 EDT When the ties between New York and New Eng-
Over the period 16:10:46 EDT to 16:10:54 EDT, the land disconnected, most of the New England area
separation between New England and New York along with Canada’s Maritime Provinces (New
occurred. It occurred along five of the northern tie Brunswick and Nova Scotia) became an island
lines, and seven lines within southwest Connecti- with generation and demand balanced close
cut. At the time of the east-west separation in New enough that it was able to remain operational. The
York at 16:10:49 EDT, New England was isolated New England system had been exporting close to
Figure 6.26. Measured Power Flows and Frequency Across Regional Interfaces, 16:10:45 to 16:11:30 EDT,
with Key Events in the Cascade
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 97
600 MW to New York, so it was relatively genera- high current and overloads in addition to true
tion-rich and experienced continuing fluctuations faults, they will trip out before other lines in the
until it reached equilibrium. Before the Maritimes path of large power swings such as the 3,500 MW
and New England separated from the Eastern power surge that hit New York on August 14. New
Interconnection at approximately 16:11 EDT, volt- York’s Total East and Central East interfaces,
ages became depressed across portions of New where the internal split occurred, are routinely
England and some large customers disconnected among the most heavily loaded paths in the state
themselves automatically.7 However, southwest- and are operated under thermal, voltage and sta-
ern Connecticut separated from New England and bility limits to respect their relative vulnerability
remained tied to the New York system for about and importance.
one minute.
Examination of the loads and generation in the
While frequency within New England wobbled Eastern New York island indicates before 16:10:00
slightly and recovered quickly after 16:10:40 EDT, EDT, the area had been importing electricity and
frequency of the New York-Ontario-Michigan- had less generation on-line than load. At 16:10:50
Ohio island fluctuated severely as additional EDT, seconds after the separation along the Total
lines, loads and generators tripped, reflecting the East interface, the eastern New York area had
severe generation deficiency in Michigan and experienced significant load reductions due to
Ohio. under-frequency load-shedding—Consolidated
Edison, which serves New York City and sur-
Due to its geography and electrical characteristics,
rounding areas, dropped over 40% of its load on
the Québec system in Canada is tied to the remain-
automatic UFLS. But at this time, the system was
der of the Eastern Interconnection via high voltage
still experiencing dynamic conditions—as illus-
DC (HVDC) links instead of AC transmission lines.
trated in Figure 6.26, frequency was falling, flows
Québec was able to survive the power surges with
and voltages were oscillating, and power plants
only small impacts because the DC connections
were tripping off-line.
shielded it from the frequency swings.
Had there been a slow islanding situation and
7B) New York Transmission Split East-West:
more generation on-line, it might have been possi-
16:10:49 EDT
ble for the Eastern New York island to rebalance
The transmission system split internally within given its high level of UFLS. But the available
New York along the Total East interface, with the information indicates that events happened so
eastern portion islanding to contain New York quickly and the power swings were so large that
City, northern New Jersey, and southwestern Con- rebalancing would have been unlikely, with or
necticut. The eastern New York island had been without the northern New Jersey and southwest
importing energy, so it did not have enough sur- Connecticut loads hanging onto eastern New
viving generation on-line to balance load. Fre- York. This was further complicated because the
quency declined quickly to below 58.0 Hz and high rate of change in voltages at load buses
triggered 7,115 MW of automatic UFLS.8 Fre- reduced the actual levels of load shed by UFLS rel-
quency declined further, as did voltage, causing ative to the levels needed and expected.
pre-designed trips at the Indian Point nuclear
The team could not find any way that one electri-
plant and other generators in and around New
cal region might have protected itself against the
York City through 16:11:10 EDT. The western por-
August 14 blackout, either at electrical borders or
tion of New York remained connected to Ontario
internally. The team also looked at whether it was
and eastern Michigan.
possible to design special protection schemes to
The electric system has inherent weak points that separate one region from its neighborings pro-
vary as a function of the characteristics of the actively, to buffer itself from a power swing before
physical lines and plants and the topology of the it hit. This was found to be inadvisable for two rea-
lines, loads and flows across the grid at any point sons: (1) as noted above, the act of separation itself
in time. The weakest points on a system tend to be could cause oscillations and dynamic instability
those points with the highest impedance, which that could be as damaging to the system as the
routinely are long (over 50 miles or 80 km) over- swing it was protecting against; and (2) there was
head lines with high loading. When such lines no event or symptom on August 14 that could be
have high-speed relay protections that may trip on used to trigger such a protection scheme in time.
98 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
7C) The Ontario System Just West of Niagara 7D) Southwest Connecticut Separated from
Falls and West of St. Lawrence Separated from New York City: 16:11:22 EDT
the Western New York Island: 16:10:50 EDT
In southwest Connecticut, when the Long Moun-
At 16:10:50 EDT, Ontario and New York separated tain-Plum Tree line (connected to the Pleasant
west of the Ontario/New York interconnection, Valley substation in New York) disconnected at
due to relay operations which disconnected nine 16:11:22 EDT, it left about 500 MW of southwest
230-kV lines within Ontario. These left most of Connecticut demand supplied only through a
Ontario isolated to the north. Ontario’s large Beck 138-kV underwater tie to Long Island. About two
and Saunders hydro stations, along with some seconds later, the two 345-kV circuits connecting
Ontario load, the New York Power Authority’s southeastern New York to Long Island tripped,
(NYPA) Niagara and St. Lawrence hydro stations, isolating Long Island and southwest Connecticut,
and NYPA’s 765-kV AC interconnection to their which remained tied together by the underwater
HVDC tie with Québec, remained connected to the Norwalk Harbor-to-Northport 138-kV cable. The
western New York system, supporting the demand cable tripped about 20 seconds later, causing
in upstate New York. southwest Connecticut to black out.
From 16:10:49 to 16:10:50 EDT, frequency in Within the western New York island, the 345-kV
Ontario declined below 59.3 Hz, initiating auto- system remained intact from Niagara east to the
matic under-frequency load-shedding (3,000 Utica area, and from the St. Lawrence/Plattsburgh
MW). This load-shedding dropped about 12% of area south to the Utica area through both the
Ontario’s remaining load. Between 16:10:50 EDT 765-kV and 230-kV circuits. Ontario’s Beck and
and 16:10:56 EDT, the isolation of Ontario’s 2,300 Saunders generation remained connected to New
MW Beck and Saunders hydro units onto the York at Niagara and St. Lawrence, respectively,
western New York island, coupled with and this island stabilized with about 50% of the
under-frequency load-shedding in the western pre-event load remaining. The boundary of this
New York island, caused the frequency in this island moved southeastward as a result of the
island to rise to 63.4 Hz due to excess generation reclosure of Fraser-to-Coopers Corners 345-kV
relative to the load within the island (Figure 6.27). line at 16:11:23 EDT.
The high frequency caused trips of five of the U.S.
nuclear units within the island, and the last one As a result of the severe frequency and voltage
tripped on the second frequency rise. changes, many large generating units in New York
and Ontario tripped off-line. The eastern island of
Three of the tripped 230-kV transmission circuits New York, including the heavily populated areas
near Niagara automatically reconnected Ontario of southeastern New York, New York City, and
to New York at 16:10:56 EDT by reclosing. Even Long Island, experienced severe frequency and
with these lines reconnected, the main Ontario voltage declines. At 16:11:29 EDT, the New Scot-
island (still attached to New York and eastern land-to-Leeds 345-kV circuits tripped, separating
Michigan) was then extremely deficient in genera- the island into northern and southern sections.
tion, so its frequency declined towards 58.8 Hz, The small remaining load in the northern portion
the threshold for the second stage of under- of the eastern island (the Albany area) retained
frequency load-shedding. Within the next two sec-
onds another 19% of Ontario demand (4,800 MW) Figure 6.27. Frequency Separation Between Ontario
automatically disconnected by under-frequency and Western New York
load-shedding. At 16:11:10 EDT, these same three
lines tripped a second time west of Niagara, and
New York and most of Ontario separated for a final
time. Following this separation, the frequency in
Ontario declined to 56 Hz by 16:11:57 EDT. With
Ontario still supplying 2,500 MW to the Michi-
gan-Ohio load pocket, the remaining ties with
Michigan tripped at 16:11:57 EDT. Ontario system
frequency declined, leading to a widespread shut-
down at 16:11:58 EDT and the loss of 22,500 MW
of load in Ontario, including the cities of Toronto,
Hamilton, and Ottawa.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 99
electric service, supplied by local generation until off-line to protect themselves from the growing
it could be resynchronized with the western New disturbance. The severe swings in frequency and
York island. voltage in the area caused numerous lines to trip,
7E) Remaining Transmission Lines Between so the isolated area broke further into smaller
Ontario and Eastern Michigan Separated: islands. The load/generation mismatch also
16:11:57 EDT affected voltages and frequency within these
smaller areas, causing further generator trips and
Before the blackout, New England, New York, automatic under-frequency load-shedding, lead-
Ontario, eastern Michigan, and northern Ohio ing to blackout in most of these areas.
were scheduled net importers of power. When the
Figure 6.28 shows frequency data collected by the
western and southern lines serving Cleveland,
distribution-level monitors of Softswitching Tech-
Toledo, and Detroit collapsed, most of the load
nologies, Inc. (a commercial power quality com-
remained on those systems, but some generation
pany serving industrial customers) for the area
had tripped. This exacerbated the generation/load
affected by the blackout. The data reveal at least
imbalance in areas that were already importing
five separate electrical islands in the Northeast as
power. The power to serve this load came through
the cascade progressed. The two paths of red dia-
the only major path available, via Ontario (IMO).
monds on the frequency scale reflect the Albany
After most of IMO was separated from New York
area island (upper path) versus the New York City
and generation to the north and east, much of the
island, which declined and blacked out much
Ontario load and generation was lost; it took only
earlier.
moments for the transmission paths west from
Ontario to Michigan to fail. Cascading Sequence Essentially Complete:
16:13 EDT
When the cascade was over at about 16:12 EDT,
much of the disturbed area was completely Most of the Northeast (the area shown in gray in
blacked out, but there were isolated pockets that Figure 6.29) was now blacked out. Some isolated
still had service because load and generation had areas of generation and load remained on-line for
reached equilibrium. Ontario’s large Beck and several minutes. Some of those areas in which a
Saunders hydro stations, along with some Ontario close generation-demand balance could be main-
load, the New York Power Authority’s (NYPA) tained remained operational.
Niagara and St. Lawrence hydro stations, and One relatively large island remained in operation
NYPA’s 765-kV AC interconnection to the Québec serving about 5,700 MW of demand, mostly in
HVDC tie, remained connected to the western western New York, anchored by the Niagara and
New York system, supporting demand in upstate St. Lawrence hydro plants. This island formed the
New York. basis for restoration in both New York and
Electrical islanding. Once the northeast became Ontario.
isolated, it lost more and more generation relative The entire cascade sequence is depicted graphi-
to load as more and more power plants tripped cally in Figure 6.30.
Figure 6.28. Electric Islands Reflected in
Frequency Plot Figure 6.29. Area Affected by the Blackout
100 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Figure 6.30. Cascade Sequence
1. 6.
16:05:57 16:10:40
2. 7.
16:05:58 16:10:41
3. 8.
16:09:25 16:10:44
4. 9.
16:10:37 16:10:45
5. 10.
16:10:39 16:13:00
Legend: Yellow arrows represent the overall pattern of electricity flows. Black lines represent approximate points of separation
between areas within the Eastern Interconnect. Gray shading represents areas affected by the blackout.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 101
Endnotes 5 New York Independent System Operator, Interim Report on
1 New York Independent System Operator, Interim Report on the August 14, 2003 Blackout, January 8, 2004, p. 20.
the August 14, 2003 Blackout, January 8, 2004, p. 14. 6 Ibid., p. 20.
2 Ibid., p. 14. 7 After New England’s separation from the Eastern Intercon-
3 These zone 2s are set on the 345-kV lines into the Argenta nection occurred, the next several minutes were critical to
substation. The lines are owned by Michigan Electric Trans- stabilizing the ISO-NE system. Voltages in New England
mission Company and maintained by Consumers Power. recovered and over-shot to high due to the combination of
Since the blackout occurred, Consumers Power has load loss, capacitors still in service, lower reactive losses on
proactively changed the relay setting from 88 Ohms to 55 the transmission system, and loss of generation to regulate
Ohms to reduce the reach of the relay. Source: Charles Rogers, system voltage. Over-voltage protective relays operated to trip
Consumers Power. both transmission and distribution capacitors. Operators in
4 The province of Québec, although considered a part of the New England brought all fast-start generation on-line by
16:16 EDT. Much of the customer process load was automati-
Eastern Interconnection, is connected to the rest of the East- cally restored. This caused voltages to drop again, putting
ern Interconnection only by DC ties. In this instance, the DC portions of New England at risk of voltage collapse. Operators
ties acted as buffers between portions of the Eastern Intercon- manually dropped 80 MW of load in southwest Connecticut
nection; transient disturbances propagate through them less by 16:39 EDT, another 325 MW in Connecticut and 100 MW
readily. Therefore, the electricity system in Québec was not in western Massachusetts by 16:40 EDT. These measures
affected by the outage, except for a small portion of the prov- helped to stabilize their island following their separation
ince’s load that is directly connected to Ontario by AC trans- from the rest of the Eastern Interconnection.
mission lines. (Although DC ties can act as a buffer between 8 New York Independent System Operator, Interim Report on
systems, the tradeoff is that they do not allow instantaneous
generation support following the unanticipated loss of a gen- the August 14, 2003 Blackout, January 8, 2004, p. 23.
erating unit.)
102 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
7. The August 14 Blackout Compared With
Previous Major North American Outages
Incidence and Characteristics Figure 7.1. North American Power System Outages,
1984-1997
of Power System Outages
Short, localized outages occur on power systems
fairly frequently. System-wide disturbances that
affect many customers across a broad geographic
area are rare, but they occur more frequently than
a normal distribution of probabilities would pre-
dict. North American power system outages
between 1984 and 1997 are shown in Figure 7.1 by
the number of customers affected and the rate of
occurrence. While some of these were widespread
weather-related events, some were cascading
events that, in retrospect, were preventable. Elec-
tric power systems are fairly robust and are capa-
ble of withstanding one or two contingency
events, but they are fragile with respect to multi-
ple contingency events unless the systems are Note: The circles represent individual outages in North
readjusted between contingencies. With the America between 1984 and 1997, plotted against the fre-
shrinking margin in the current transmission sys- quency of outages of equal or greater size over that period.
tem, it is likely to be more vulnerable to cascading Source: Adapted from John Doyle, California Institute of
Technology, “Complexity and Robustness,” 1999. Data from
outages than it was in the past, unless effective
NERC.
countermeasures are taken.
As evidenced by the absence of major transmis- If nothing else changed, one could expect an
sion projects undertaken in North America over increased frequency of large-scale events as com-
the past 10 to 15 years, utilities have found ways to pared to historical experience. The last and most
increase the utilization of their existing facilities extreme event shown in Figure 7.1 is the August
to meet increasing demands without adding sig- 10, 1996, outage. August 14, 2003, surpassed that
nificant high-voltage equipment. Without inter- event in terms of severity. In addition, two signifi-
vention, this trend is likely to continue. Pushing cant outages in the month of September 2003
the system harder will undoubtedly increase reli- occurred abroad: one in England and one, initiated
ability challenges. Special protection schemes in Switzerland, that cascaded over much of Italy.
may be relied on more to deal with particular chal-
lenges, but the system still will be less able to In the following sections, seven previous outages
withstand unexpected contingencies. are reviewed and compared with the blackout of
August 14, 2003: (1) Northeast blackout on
A smaller transmission margin for reliability November 9, 1965; (2) New York City blackout on
makes the preservation of system reliability a July 13, 1977; (3) West Coast blackout on Decem-
harder job than it used to be. The system is being ber 22, 1982; (4) West Coast blackout on July 2-3,
operated closer to the edge of reliability than it 1996; (5) West Coast blackout on August 10, 1996;
was just a few years ago. Table 7.1 represents some (6) Ontario and U.S. North Central blackout on
of the changed conditions that make the preserva- June 25, 1998; and (7) Northeast outages and non-
tion of reliability more challenging. outage disturbances in the summer of 1999.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 103
Outage Descriptions u Two key 345-kV east-west (Rochester-Syracuse)
lines opened due to instability, and several
and Major Causal Factors lower voltage lines tripped open.
November 9, 1965: Northeast Blackout u Five of 16 generators at the St. Lawrence
(Massena) plant tripped automatically in
This disturbance resulted in the loss of over accordance with predetermined operating
20,000 MW of load and affected 30 million people. procedures.
Virtually all of New York, Connecticut, Massachu-
u Following additional line tripouts, 10 generat-
setts, Rhode Island, small segments of northern
ing units at Beck were automatically shut down
Pennsylvania and northeastern New Jersey, and
by low governor oil pressure, and 5 pumping
substantial areas of Ontario, Canada, were
generators were tripped off by overspeed gover-
affected. Outages lasted for up to 13 hours. This
nor control.
event resulted in the formation of the North Amer-
ican Electric Reliability Council in 1968. u Several other lines then tripped out on
under-frequency relay action.
A backup protective relay operated to open one of
five 230-kV lines taking power north from a gener- July 13, 1977: New York City Blackout
ating plant in Ontario to the Toronto area. When
the flows redistributed instantaneously on the This disturbance resulted in the loss of 6,000 MW
remaining four lines, they tripped out succes- of load and affected 9 million people in New York
sively in a total of 2.5 seconds. The resultant City. Outages lasted for up to 26 hours. A series of
power swings resulted in a cascading outage that events triggering the separation of the Consoli-
blacked out much of the Northeast. dated Edison system from neighboring systems
and its subsequent collapse began when two
The major causal factors were as follows: 345-kV lines on a common tower in Northern
u Operation of a backup protective relay took a
Westchester were struck by lightning and tripped
230-kV line out of service when the loading on out. Over the next hour, despite Consolidated Edi-
the line exceeded the 375-MW relay setting. son dispatcher actions, the system electrically
separated from surrounding systems and col-
u Operating personnel were not aware of the lapsed. With the loss of imports, generation in
operating set point of this relay. New York City was not sufficient to serve the load
in the city.
u Another 230-kV line opened by an overcurrent
relay action, and several 115- and 230-kV lines Major causal factors were:
opened by protective relay action.
104 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
u Two 345-kV lines connecting Buchanan South December 22, 1982: West Coast Blackout
to Millwood West experienced a phase B to
ground fault caused by a lightning strike. This disturbance resulted in the loss of 12,350
MW of load and affected over 5 million people in
u Circuit breaker operations at the Buchanan the West. The outage began when high winds
South ring bus isolated the Indian Point No. 3 caused the failure of a 500-kV transmission tower.
generating unit from any load, and the unit trip- The tower fell into a parallel 500-kV line tower,
ped for a rejection of 883 MW of load. and both lines were lost. The failure of these two
u Loss of the ring bus isolated the 345-kV tie to lines mechanically cascaded and caused three
Ladentown, which had been importing 427 additional towers to fail on each line. When the
MW, making the cumulative resources lost line conductors fell they contacted two 230-kV
1,310 MW. lines crossing under the 500-kV rights-of-way, col-
lapsing the 230-kV lines.
u 18.5 minutes after the first incident, an addi-
tional lightning strike caused the loss of two The loss of the 500-kV lines activated a remedial
345-kV lines, which connect Sprain Brook to action scheme to control the separation of the
Buchanan North and Sprain Brook to Millwood interconnection into two pre-engineered islands
West. These two 345-kV lines share common and trip generation in the Pacific Northwest in
towers between Millwood West and Sprain order to minimize customer outages and speed
Brook. One line (Sprain Brook to Millwood restoration. However, delayed operation of the
West) automatically reclosed and was restored remedial action scheme components occurred for
to service in about 2 seconds. The failure of the several reasons, and the interconnection sepa-
other line to reclose isolated the last Consoli- rated into four islands.
dated Edison interconnection to the Northwest. In addition to the mechanical failure of the trans-
mission lines, analysis of this outage cited prob-
u The resulting surge of power from the North-
lems with coordination of protective schemes,
west caused the loss of the Pleasant Valley to
because the generator tripping and separation
Millwood West line by relay action (a bent con-
schemes operated slowly or did not operate as
tact on one of the relays at Millwood West
planned. A communication channel component
caused the improper action).
performed sporadically, resulting in delayed
u 23 minutes later, the Leeds to Pleasant Valley transmission of the control signal. The backup
345-kV line sagged into a tree due to overload separation scheme also failed to operate, because
and tripped out. the coordination of relay settings did not antici-
pate the power flows experienced in this severe
u Within a minute, the 345 kV to 138 kV trans- disturbance.
former at Pleasant Valley overloaded and trip-
ped off, leaving Consolidated Edison with only In addition, the volume and format in which data
three remaining interconnections. were displayed to operators made it difficult to
assess the extent of the disturbance and what cor-
u Within 3 minutes, the Long Island Lighting Co. rective action should be taken. Time references to
system operator, on concurrence of the pool dis- events in this disturbance were not tied to a com-
patcher, manually opened the Jamaica to Valley mon standard, making real-time evaluation of the
Stream tie. situation more difficult.
u About 7 minutes later, the tap-changing mecha-
nism failed on the Goethals phase-shifter,
July 2-3, 1996: West Coast Blackout
resulting in the loss of the Linden-to-Goethals This disturbance resulted in the loss of 11,850
tie to PJM, which was carrying 1,150 MW to MW of load and affected 2 million people in the
Consolidated Edison. West. Customers were affected in Arizona, Cali-
u The two remaining external 138-kV ties to Con- fornia, Colorado, Idaho, Montana, Nebraska,
solidated Edison tripped on overload, isolating Nevada, New Mexico, Oregon, South Dakota,
the Consolidated Edison system. Texas, Utah, Washington, and Wyoming in the
United States; Alberta and British Columbia in
u Insufficient generation in the isolated system Canada; and Baja California Norte in Mexico. Out-
caused the Consolidated Edison island to ages lasted from a few minutes to several hours.
collapse.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 105
The outage began when a 345-kV transmission the overload and loss of additional transmission
line in Idaho sagged into a tree and tripped out. A lines. General voltage decline in the Northwest
protective relay on a parallel transmission line and the loss of McNary generation due to incor-
also detected the fault and incorrectly tripped a rectly applied relays caused power oscillations on
second line. An almost simultaneous loss of these the California to Oregon AC intertie. The intertie’s
lines greatly reduced the ability of the system to protective relays tripped these facilities out and
transmit power from the nearby Jim Bridger plant. caused the Western Interconnection to separate
Other relays tripped two of the four generating into four islands. Following the loss of the first two
units at that plant. With the loss of those two lightly loaded lines, operators were unaware that
units, frequency in the entire Western Intercon- the system was in an insecure state over the next
nection began to decline, and voltage began to col- hour, because new operating studies had not been
lapse in the Boise, Idaho, area, affecting the performed to identify needed system adjustments.
California-Oregon AC Intertie transfer limit.
June 25, 1998: Upper Midwest Blackout
For 23 seconds the system remained in precarious
balance, until the Mill Creek to Antelope 230-kV This disturbance resulted in the loss of 950 MW of
line between Montana and Idaho tripped by zone load and affected 152,000 people in Minnesota,
3 relay, depressing voltage at Summer Lake Sub- Montana, North Dakota, South Dakota, and Wis-
station and causing the intertie to slip out of syn- consin in the United States; and Ontario, Mani-
chronism. Remedial action relays separated the toba, and Saskatchewan in Canada. Outages lasted
system into five pre-engineered islands designed up to 19 hours.
to minimize customer outages and restoration
A lightning storm in Minnesota initiated a series of
times. Similar conditions and initiating factors
events, causing a system disturbance that affected
were present on July 3; however, as voltage began
the entire Mid-Continent Area Power Pool (MAPP)
to collapse in the Boise area, the operator shed
Region and the northwestern Ontario Hydro sys-
load manually and contained the disturbance.
tem of the Northeast Power Coordinating Council.
August 10, 1996: West Coast Blackout A 345-kV line was struck by lightning and tripped
out. Underlying lower voltage lines began to over-
This disturbance resulted in the loss of over load and trip out, further weakening the system.
28,000 MW of load and affected 7.5 million people Soon afterward, lightning struck a second 345-kV
in the West. Customers were affected in Arizona, line, taking it out of service as well. Following the
California, Colorado, Idaho, Montana, Nebraska, outage of the second 345-kV line, the remaining
Nevada, New Mexico, Oregon, South Dakota, lower voltage transmission lines in the area
Texas, Utah, Washington, and Wyoming in the became significantly overloaded, and relays took
United States; Alberta and British Columbia in them out of service. This cascading removal of
Canada; and Baja California Norte in Mexico. Out- lines from service continued until the entire
ages lasted from a few minutes to as long as nine northern MAPP Region was separated from the
hours. Eastern Interconnection, forming three islands
Triggered by several major transmission line out- and resulting in the eventual blackout of the
ages, the loss of generation from McNary Dam, and northwestern Ontario Hydro system.
resulting system oscillations, the Western Inter-
connection separated into four electrical islands, Summer of 1999: Northeast U.S.
with significant loss of load and generation. Prior Non-outage Disturbances
to the disturbance, the transmission system from
Load in the PJM system on July 6, 1999, was
Canada south through the Northwest into Califor-
51,600 MW (approximately 5,000 MW above fore-
nia was heavily loaded with north-to-south power
cast). PJM used all emergency procedures (includ-
transfers. These flows were due to high Southwest
ing a 5% voltage reduction) except manually
demand caused by hot weather, combined with
tripping load, and imported 5,000 MW from exter-
excellent hydroelectric conditions in Canada and
nal systems to serve the record customer demand.
the Northwest.
Load on July 19, 1999, exceeded 50,500 MW. PJM
Very high temperatures in the Northwest caused loaded all available eastern PJM generation and
two lightly loaded transmission lines to sag into again implemented emergency operating proce-
untrimmed trees and trip out. A third heavily dures from approximately 12 noon into the eve-
loaded line also sagged into a tree. Its outage led to ning on both days.
106 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
During these record peak loads, steep voltage pumping, accompanies hot weather and places
declines were experienced on the bulk transmis- higher burdens on transmission lines. Losing cir-
sion system. In each case, a voltage collapse was cuits contributes to voltage decline. Inductive
barely averted through the use of emergency pro- load is unforgiving when voltage declines, draw-
cedures. Low voltage occurred because reactive ing additional reactive supply from the system
demand exceeded reactive supply. High reactive and further contributing to voltage problems.
demand was due to high electricity demand and
high losses resulting from high transfers across the Recommendations from previous investigations
system. Reactive supply was inadequate because include:
generators were unavailable or unable to meet u Paying special attention to the condition of
rated reactive capability due to ambient condi- rights-of-way following favorable growing sea-
tions, and because some shunt capacitors were out sons. Very wet and warm spring and summer
of service. growing conditions preceded the 1996 outages
in the West.
Common or Similar Factors u Careful review of any reduction in operations
Among Major Outages and maintenance expenses that may contribute
to decreased frequency of line patrols or trim-
The factors that were common to some of the ming. Maintenance in this area should be
major outages above and the August 14 blackout strongly directed toward preventive rather than
include: (1) conductor contact with trees; (2) over- remedial maintenance.
estimation of dynamic reactive output of system
generators; (3) inability of system operators or Dynamic Reactive Output of Generators
coordinators to visualize events on the entire sys-
Reactive supply is an important ingredient in
tem; (4) failure to ensure that system operation
maintaining healthy power system voltages and
was within safe limits; (5) lack of coordination on
facilitating power transfers. Inadequate reactive
system protection; (6) ineffective communication;
supply was a factor in most of the events. Shunt
(7) lack of “safety nets;” and (8) inadequate train-
capacitors and generating resources are the most
ing of operating personnel. The following sections
significant suppliers of reactive power. Operators
describe the nature of these factors and list recom-
perform contingency analysis based on how
mendations from previous investigations that are
power system elements will perform under vari-
relevant to each.
ous power system conditions. They determine and
Conductor Contact With Trees set transfer limits based on these analyses. Shunt
capacitors are easy to model because they are
This factor was an initiating trigger in several of static. Modeling the dynamic reactive output of
the outages and a contributing factor in the sever- generators under stressed system conditions has
ity of several more. Unlike lightning strikes, for proven to be more challenging. If the model is
which system operators have fair storm-tracking incorrect, estimated transfer limits will also be
tools, system operators generally do not have incorrect.
direct knowledge that a line has contacted a tree
In most of the events, the assumed contribution of
and faulted. They will sometimes test the line by
dynamic reactive output of system generators was
trying to restore it to service, if that is deemed to be
greater than the generators actually produced,
a safe operation. Even if it does go back into ser-
resulting in more significant voltage problems.
vice, the line may fault and trip out again as load
Some generators were limited in the amount of
heats it up. This is most likely to happen when
reactive power they produced by over-excitation
vegetation has not been adequately managed, in
limits, or necessarily derated because of high
combination with hot and windless conditions.
ambient temperatures. Other generators were con-
In some of the disturbances, tree contact account- trolled to a fixed power factor and did not contrib-
ed for the loss of more than one circuit, contribut- ute reactive supply in depressed voltage
ing multiple contingencies to the weakening of conditions. Under-voltage load shedding is em-
the system. Lines usually sag into right-of-way ployed as an automatic remedial action in some
obstructions when the need to retain transmission interconnections to prevent cascading, and could
interconnection is high. High inductive load be used more widely.
composition, such as air conditioning or irrigation
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 107
Recommendations from previous investigations presented to operators or coordinators as informa-
concerning voltage support and reactive power tion they can use in making appropriate operating
management include: decisions.
u Communicate changes to generator reactive Recommendations from previous investigations
capability limits in a timely and accurate man- concerning visibility and tools include:
ner for both planning and operational modeling
u Develop communications systems and displays
purposes.
that give operators immediate information on
u Investigate the development of a generator changes in the status of major components in
MVAr/voltage monitoring process to determine their own and neighboring systems.
when generators may not be following reported u Supply communications systems with uninter-
MVAr limits. ruptible power, so that information on system
u Establish a common standard for generator conditions can be transmitted correctly to con-
steady-state and post-contingency (15-minute) trol centers during system disturbances.
MVAr capability definition; determine method- u In the control center, use a dynamic line loading
ology, testing, and operational reporting and outage display board to provide operating
requirements. personnel with rapid and comprehensive infor-
u Determine the generator service level agree- mation about the facilities available and the
ment that defines generator MVAr obligation to operating condition of each facility in service.
help ensure reliable operations. u Give control centers the capability to display to
u Periodically review and field test the reactive system operators computer-generated alterna-
limits of generators to ensure that reported tive actions specific to the immediate situation,
MVAr limits are attainable. together with expected results of each action.
u Establish on-line security analysis capability to
u Provide operators with on-line indications of
available reactive capability from each generat- identify those next and multiple facility outages
ing unit or groups of generators, other VAr that would be critical to system reliability from
sources, and the reactive margin at all critical thermal, stability, and post-contingency voltage
buses. This information should assist in the points of view.
operating practice of maximizing the use of u Establish time-synchronized disturbance moni-
shunt capacitors during heavy transfers and toring to help evaluate the performance of the
thereby increase the availability of system interconnected system under stress, and design
dynamic reactive reserve. appropriate controls to protect it.
u For voltage instability problems, consider fast
System Operation Within Safe Limits
automatic capacitor insertion (both series and
shunt), direct shunt reactor and load tripping, Operators in several of the events were unaware of
and under-voltage load shedding. the vulnerability of the system to the next contin-
gency. The reasons were varied: inaccurate model-
u Develop and periodically review a reactive mar-
ing for simulation, no visibility of the loss of key
gin against which system performance should
transmission elements, no operator monitoring of
be evaluated and used to establish maximum
stability measures (reactive reserve monitor,
transfer levels.
power transfer angle), and no reassessment of sys-
System Visibility Procedures and tem conditions following the loss of an element
and readjustment of safe limits.
Operator Tools
Recommendations from previous investigations
Each control area operates as part of a single syn- include:
chronous interconnection. However, the parties
with various geographic or functional responsibil- u Following a contingency, the system must be
ities for reliable operation of the grid do not have returned to a reliable state within the allowed
visibility of the entire system. Events in neighbor- readjustment period. Operating guides must be
ing systems may not be visible to an operator or reviewed to ensure that procedures exist to
reliability coordinator, or power system data restore system reliability in the allowable time
may be available in a control center but not be periods.
108 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
u Reduce scheduled transfers to a safe and pru- a specified period of time. The time delay
dent level until studies have been conducted to should be long enough to allow the system oper-
determine the maximum simultaneous transfer ator to attempt to reduce line loadings promptly
capability limits. by other means.
u Reevaluate processes for identifying unusual u Review phase-angle restrictions that can pre-
operating conditions and potential disturbance vent reclosing of major interconnections during
scenarios, and make sure they are studied system emergencies. Consideration should be
before they are encountered in real-time operat- given to bypassing synchronism-check relays to
ing conditions. permit direct closing of critical interconnec-
tions when it is necessary to maintain stability
Coordination of System Protection of the grid during an emergency.
(Transmission and Generation Elements)
u Review the need for controlled islanding. Oper-
Protective relays are designed to detect short cir- ating guides should address the potential for
cuits and act locally to isolate faulted power sys- significant generation/load imbalance within
tem equipment from the system—both to protect the islands.
the equipment from damage and to protect the sys-
tem from faulty equipment. Relay systems are Effectiveness of Communications
applied with redundancy in primary and backup Under normal conditions, parties with reliability
modes. If one relay fails, another should detect the responsibility need to communicate important
fault and trip appropriate circuit breakers. Some and prioritized information to each other in a
backup relays have significant “reach,” such that timely way, to help preserve the integrity of the
non-faulted line overloads or stable swings may be grid. This is especially important in emergencies.
seen as faults and cause the tripping of a line when During emergencies, operators should be relieved
it is not advantageous to do so. Proper coordina- of duties unrelated to preserving the grid. A com-
tion of the many relay devices in an intercon- mon factor in several of the events described
nected system is a significant challenge, requiring above was that information about outages occur-
continual review and revision. Some relays can ring in one system was not provided to neighbor-
prevent resynchronizing, making restoration more ing systems.
difficult.
System-wide controls protect the interconnected Need for Safety Nets
operation rather than specific pieces of equip- A safety net is a protective scheme that activates
ment. Examples include controlled islanding to automatically if a pre-specified, significant con-
mitigate the severity of an inevitable disturbance tingency occurs. When activated, such schemes
and under-voltage or under-frequency load shed- involve certain costs and inconvenience, but they
ding. Failure to operate (or misoperation of) one or can prevent some disturbances from getting out of
more relays as an event developed was a common control. These plans involve actions such as shed-
factor in several of the disturbances. ding load, dropping generation, or islanding, and
in all cases the intent is to have a controlled out-
Recommendations developed after previous out-
come that is less severe than the likely uncon-
ages include:
trolled outcome. If a safety net had not been taken
u Perform system trip tests of relay schemes peri- out of service in the West in August 1996, it would
odically. At installation the acceptance test have lessened the severity of the disturbance from
should be performed on the complete relay 28,000 MW of load lost to less than 7,200 MW. (It
scheme in addition to each individual compo- has since been returned to service.) Safety nets
nent so that the adequacy of the scheme is should not be relied upon to establish transfer lim-
verified. its, however.
u Continually update relay protection to fit Previous recommendations concerning safety nets
changing system development and to incorpo- include:
rate improved relay control devices.
u Establish and maintain coordinated programs
u Install sensing devices on critical transmission of automatic load shedding in areas not so
lines to shed load or generation automatically if equipped, in order to prevent total loss of power
the short-term emergency rating is exceeded for in an area that has been separated from the
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 109
main network and is deficient in generation. frames in many cases. Other readjustments
Load shedding should be regarded as an insur- must be used, and the system operator must
ance program, however, and should not be used take responsibility to restore the system
as a substitute for adequate system design. immediately.
u Install load-shedding controls to allow fast sin- u Operators’ authority and responsibility to take
gle-action activation of large-block load shed- immediate action if they sense the system is
ding by an operator. starting to degrade should be emphasized and
protected.
Training of Operating Personnel
u The current processes for assessing the poten-
Operating procedures were necessary but not suf- tial for voltage instability and the need to
ficient to deal with severe power system distur- enhance the existing operator training pro-
bances in several of the events. Enhanced grams, operational tools, and annual technical
procedures and training for operating personnel assessments should be reviewed to improve the
were recommended. Dispatcher training facility ability to predict future voltage stability prob-
scenarios with disturbance simulation were sug- lems prior to their occurrence, and to mitigate
gested as well. Operators tended to reduce sched- the potential for adverse effects on a regional
ules for transactions but were reluctant to call scale.
for increased generation—or especially to shed
load—in the face of a disturbance that threatened
to bring the whole system down.
Comparisons With the
August 14 Blackout
Previous recommendations concerning training
include: The blackout on August 14, 2003, had several
u Thorough programs and schedules for operator causes or contributory factors in common with the
training and retraining should be vigorously earlier outages, including:
administered. u Inadequate vegetation management
u A full-scale simulator should be made available u Failure to ensure operation within secure limits
to provide operating personnel with “hands-on”
u Failure to identify emergency conditions and
experience in dealing with possible emergency
or other system conditions. communicate that status to neighboring
systems
u Procedures and training programs for system
u Inadequate operator training
operators should include anticipation, recogni-
tion, and definition of emergency situations. u Inadequate regional-scale visibility over the
power system
u Written procedures and training materials
should include criteria that system operators u Inadequate coordination of relays and other
can use to recognize signs of system stress and protective devices or systems.
mitigating measures to be taken before condi-
New causal features of the August 14 blackout
tions degrade into emergencies.
include: inadequate interregional visibility over
u Line loading relief procedures should not be the power system; dysfunction of a control area’s
relied upon when the system is in an insecure SCADA/EMS system; and lack of adequate backup
state, as these procedures cannot be imple- capability to that system.
mented effectively within the required time
110 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
8. Performance of Nuclear Power Plants
Affected by the Blackout
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 111
ã At Darlington, load swing on the grid led to plants that shut down or disconnected from the
the automatic reduction in power of the four grid responded automatically to grid conditions.
reactors. The generators were, in turn, auto- All the nuclear plants responded in a manner con-
matically disconnected from the grid. sistent with the plant designs. Safety functions
were effectively accomplished, and the nuclear
ã Three reactors at Bruce B and one at Darling-
plants that tripped were maintained in a safe shut-
ton were returned to 60% power. These reac-
down condition until their restart.
tors were available to deliver power to the
grid on the instructions of the transmission In Phase II, the NWG collected comments and ana-
system operator. lyzed information related to potential recommen-
ã Three units at Darlington were placed in a dations to help prevent future power outages.
zero-power hot state, and four units at Representatives of the NWG, including represen-
Pickering B and one unit at Bruce B were tatives of the NRC and the CNSC, attended public
placed in a Guaranteed Shutdown State. meetings to solicit feedback and recommenda-
tions held in Cleveland, Ohio; New York City,
The licensees’ return to power operation followed New York; and Toronto, Ontario, on December 4,
a deliberate process controlled by plant proce- 5, and 8, 2003, respectively. Representatives of the
dures and regulations. Equipment and process NWG also participated in the NRC’s public meet-
problems, whether existing prior to or caused by ing to solicit feedback and recommendations on
the event, would normally be addressed prior to the Northeast blackout held in Rockville, Mary-
restart. The NWG is satisfied that licensees took an land, on January 6, 2004.
appropriately conservative approach to their
restart activities, placing a priority on safety. Additional details on both the Phase I and Phase II
efforts are available in the following sections. Due
u For U.S. nuclear plants: Ginna, Indian Point 2, to the major design differences between nuclear
Nine Mile 2, and Oyster Creek resumed electri- plants in Canada and the United States, the NWG
cal generation on August 17. FitzPatrick and decided to have separate sections for each coun-
Nine Mile 1 resumed electrical generation on try. This also responds to the request by the
August 18. Fermi 2 resumed electrical genera- nuclear regulatory agencies in both countries to
tion on August 20. Perry resumed electrical gen- have sections of the report that stand alone, so that
eration on August 21. Indian Point 3 resumed they can also be used as regulatory documents.
electrical generation on August 22. Indian Point
3 had equipment issues (failed splices in the
control rod drive mechanism power system) Findings of the U.S. Nuclear
that required repair prior to restart. Ginna Working Group
submitted a special request for enforcement dis-
cretion from the NRC to permit mode changes Summary
and restart with an inoperable auxiliary
feedwater pump. The NRC granted the request The U.S. NWG found no evidence that the shut-
for enforcement discretion. down of the nine U.S. nuclear power plants trig-
gered the outage, or inappropriately contributed to
u For Canadian nuclear plants: The restart of the
its spread (i.e., to an extent beyond the normal
Canadian nuclear plants was carried out in
tripping of the plants at expected conditions). All
accordance with approved Operating Policies
nine plants that experienced a reactor trip were
and Principles. Three units at Bruce B and one
responding to grid conditions. The severity of the
at Darlington were resynchronized with the grid
grid transient caused generators, turbines, or reac-
within 6 hours of the event. The remaining
tor systems at the plants to reach a protective fea-
three units at Darlington were reconnected by
ture limit and actuate a plant shutdown. All nine
August 17 and 18. Units 5, 6, and 8 at Pickering
plants tripped in response to those conditions in a
B and Unit 6 at Bruce B returned to service
manner consistent with the plant designs. The
between August 22 and August 25.
nine plants automatically shut down in a safe
The NWG has found no evidence that the shut- fashion to protect the plants from the grid tran-
down of the nuclear power plants triggered the sient. Safety functions were effectively accom-
outage or inappropriately contributed to its spread plished with few problems, and the plants were
(i.e., to an extent beyond the normal tripping of maintained in a safe shutdown condition until
the plants at expected conditions). All the nuclear their restart.
112 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
The nuclear power plant outages that resulted in detail. Two additional requests for more spe-
from the August 14, 2003, power outage were trig- cific information were made for certain plants.
gered by automatic protection systems for the The collection of information from U.S. nuclear
reactors or turbine-generators, not by any manual power plants was gathered through the NRC
operator actions. The NWG has received no infor- regional offices, which had NRC resident inspec-
mation that points to operators deliberately shut- tors at each plant obtain licensee information to
ting down nuclear units to isolate themselves from answer the questions. General design information
instabilities on the grid. In short, only automatic was gathered from plant-specific Updated Final
separation of nuclear units occurred. Safety Analysis Reports and other documents.
Regarding the 95 other licensed commercial Plant data were compared against plant designs by
nuclear power plants in the United States: 4 were the NRC staff to determine whether the plant
already shut down at the time of the power outage, responses were as expected; whether they
one of which experienced a grid disturbance; 70 appeared to cause the power outage or contributed
operating plants observed some level of grid dis- to the spread of the outage; and whether applica-
turbance but accommodated the disturbances and ble safety requirements were met. In some cases
remained on line, supplying power to the grid; and supplemental questions were developed, and
21 operating plants did not experience any grid answers were obtained from the licensees to clar-
disturbance. ify the observed response of the plant. The NWG
interfaced with the ESWG to validate some data
Introduction and to obtain grid information, which contributed
to the analysis. The NWG identified relevant
The NRC, which regulates U.S. commercial
actions by nuclear generating facilities in connec-
nuclear power plants, has regulatory requirements
tion with the power outage.
for offsite power systems. These requirements
address the number of offsite power sources and
the ability to withstand certain transients. Offsite
Typical Design, Operational, and
power is the normal source of alternating current Protective Features of U.S. Nuclear
(AC) power to the safety systems in the plants Power Plants
when the plant main generator is not in operation.
Nuclear power plants have a number of design,
The requirements also are designed to protect
operational, and protective features to ensure that
safety systems from potentially damaging varia-
the plants operate safely and reliably. This section
tions (in voltage and frequency) in the supplied
describes these features so as to provide a better
power. For loss of offsite power events, the NRC
understanding of how nuclear power plants inter-
requires emergency generation (typically emer-
act with the grid and, specifically, how nuclear
gency diesel generators) to provide AC power to
power plants respond to changing grid conditions.
safety systems. In addition, the NRC provides
While the features described in this section are
oversight of the safety aspects of offsite power
typical, there are differences in the design and
issues through its inspection program, by moni-
operation of individual plants which are not
toring operating experience, and by performing
discussed.
technical studies.
Design Features of U.S. Nuclear Power Plants
Phase I: Fact Finding
Nuclear power plants use heat from nuclear reac-
Phase I of the NWG effort focused on collecting tions to generate steam and use a single steam-
and analyzing data from each plant to determine driven turbine-generator (also known as the main
what happened, and whether any activities at the generator) to produce electricity supplied to the
plants caused or contributed to the power outage grid.
or its spread or involved a significant safety issue.
To ensure accuracy, comprehensive coordination Connection of the plant switchyard to the grid.
was maintained among the working group mem- The plant switchyard normally forms the interface
bers and among the NWG, ESWG, and SWG. between the plant main generator and the electri-
cal grid. The plant switchyard has multiple trans-
The staff developed a set of technical questions to mission lines connected to the grid system to meet
obtain data from the owners or licensees of the offsite power supply requirements for having reli-
nuclear power plants that would enable them to able offsite power for the nuclear station under
review the response of the nuclear plant systems all operating and shutdown conditions. Each
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 113
transmission line connected to the switchyard has power (MVAr) from the generator to the
dedicated circuit breakers, with fault sensors, to switchyard and out to the surrounding grid, help-
isolate faulted conditions in the switchyard or the ing to control the grid voltage decrease. If the
connected transmission lines, such as phase-to- switchyard voltage goes low enough, the
phase or phase-to-ground short circuits. The fault increased generator field current could result in
sensors are fed into a protection scheme for the generator field overheating. Over-excitation pro-
plant switchyard that is engineered to localize tective circuitry is generally employed to prevent
any faulted conditions with minimum system this from occurring. This protective circuitry may
disturbance. trip the generator to prevent equipment damage.
Connection of the main generator to the switch- Under-voltage protection is provided for the
yard. The plant main generator produces electri- nuclear power plant safety buses, and may be pro-
cal power and transmits that power to the offsite vided on nonsafety buses and at individual pieces
transmission system. Most plants also supply of equipment. It is also used in some pressurized
power to the plant auxiliary buses for normal water reactor designs on reactor coolant pumps
operation of the nuclear generating unit through (RCPs) as an anticipatory loss of RCP flow signal.
the unit auxiliary transformer. During normal
plant operation, the main generator typically gen- Protective Features of U.S. Nuclear Power
erates electrical power at about 22 kV. The voltage Plants
is increased to match the switchyard voltage by The main generator and main turbine have protec-
the main transformers, and the power flows to the tive features, similar to fossil generating stations,
high voltage switchyard through two power cir- which protect against equipment damage. In gen-
cuit breakers. eral, the reactor protective features are designed to
Power supplies for the plant auxiliary buses. The protect the reactor fuel from damage and to protect
safety-related and nonsafety auxiliary buses are the reactor coolant system from over-pressure or
normally lined up to receive power from the main over-temperature transients. Some trip features
generator auxiliary transformer, although some also produce a corresponding trip in other compo-
plants leave some of their auxiliary buses powered nents; for example, a turbine trip typically results
from a startup transformer (that is, from the offsite in a reactor trip above a low power setpoint.
power distribution system). When plant power
Generator protective features typically include
generation is interrupted, the power supply auto-
over-current, ground detection, differential relays
matically transfers to the offsite power source (the
(which monitor for electrical fault conditions
startup transformer). If that is not supplying
within a zone of protection defined by the location
acceptable voltage, the circuit breakers to the
of the sensors, typically the main generator and all
safety-related buses open, and the buses are
transformers connected directly to the generator
reenergized by the respective fast-starting emer-
output), electrical faults on the transformers con-
gency diesel generators. The nonsafety auxiliary
nected to the generator, loss of the generator field,
buses will remain deenergized until offsite power
and a turbine trip. Turbine protective features typ-
is restored.
ically include over-speed (usually set at 1980 rpm
Operational Features of U.S. Nuclear Power or 66 Hz), low bearing oil pressure, high bearing
Plants vibration, degraded condenser vacuum, thrust
Response of nuclear power plants to changes in bearing failure, or generator trip. Reactor protec-
switchyard voltage. With the main generator volt- tive features typically include trips for over-
age regulator in the automatic mode, the generator power, abnormal pressure in the reactor coolant
will respond to an increase of switchyard voltage system, low reactor coolant system flow, low level
by reducing the generator field excitation current. in the steam generators or the reactor vessel, or a
This will result in a decrease of reactive power, trip of the turbine.
normally measured as mega-volts-amperes-reac-
tive (MVAr) from the generator to the switchyard Considerations on Returning a U.S.
and out to the surrounding grid, helping to control Nuclear Power Plant to Power
the grid voltage increase. With the main generator Production After Switchyard Voltage
voltage regulator in the automatic mode, the gen- Is Restored
erator will respond to a decrease of switchyard
voltage by increasing the generator field excitation The following are examples of the types of activi-
current. This will result in an increase of reactive ties that must be completed before returning a
114 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
nuclear power plant to power production follow- decreased significantly (more than 24 hours
ing a loss of switchyard voltage. after the trip).
u Switchyard voltage must be normal and stable It may require a day or more before a nuclear
from an offsite supply. Nuclear power plants are power plant can restart following a normal trip.
not designed for black-start capability (the abil- Plant trips are a significant transient on plant
ity to start up without external power). equipment, and some maintenance may be neces-
sary before the plant can restart. When combined
u Plant buses must be energized from the
with the infrequent event of loss of offsite power,
switchyard and the emergency diesel genera-
additional recovery actions will be required.
tors restored to standby mode.
Safety systems, such as emergency diesel genera-
u Normal plant equipment, such as reactor cool- tors and safety-related decay heat removal sys-
ant pumps and circulating water pumps, must tems, must be restored to normal lineups. These
be restarted. additional actions would extend the time neces-
sary to restart a nuclear plant from this type of
u A reactor trip review report must be completed event.
and approved by plant management, and the
cause of the trip must be addressed.
Summary of U.S. Nuclear Power Plant
u All plant technical specifications must be satis- Response to and Safety During the
fied. Technical specifications are issued to each August 14 Outage
nuclear power plant as part of their license by
the NRC. They dictate equipment which must The NWG’s review did not identify any activity or
be operable and process parameters which must equipment issues at U.S. nuclear power plants
be met to allow operation of the reactor. Exam- that caused the transient on August 14, 2003. Nine
ples of actions that were required following the nuclear power plants tripped within about 60 sec-
events of August 14 include refilling the diesel onds as a result of the grid disturbance. Addi-
fuel oil storage tanks, refilling the condensate tionally, many nuclear power plants experienced
storage tanks, establishing reactor coolant sys- a transient due to this grid disturbance.
tem forced flow, and cooling the suppression
pool to normal operating limits. Surveillance Nuclear Power Plants That Tripped
tests must be completed as required by techni- The trips at nine nuclear power plants resulted
cal specifications (for example, operability of from the plant responses to the grid disturbances.
the low-range neutron detectors must be Following the initial grid disturbances, voltages in
demonstrated). the plant switchyard fluctuated and reactive
u Systems must be aligned to support the startup. power flows fluctuated. As the voltage regulators
on the main generators attempted to compensate,
u Pressures and temperatures for reactor startup equipment limits were exceeded and protective
must be established in the reactor coolant sys- trips resulted. This happened at Fermi 2 and Oys-
tem for pressurized water reactors. ter Creek. Fermi 2 tripped on a generator field pro-
tection trip. Oyster Creek tripped due to a
u A reactor criticality calculation must be per-
generator trip on high ratio of voltage relative to
formed to predict the control rod withdrawals
the electrical frequency.
needed to achieve criticality, where the fission
chain reaction becomes self-sustaining due to Also, as the balance between electrical generation
the increased neutron flux. Certain neutron- and electrical load on the grid was disturbed, the
absorbing fission products increase in concen- electrical frequency began to fluctuate. In some
tration following a reactor trip (followed later cases the electrical frequency dropped low
by a decrease or decay). At pressurized water enough to actuate protective features. This hap-
reactors, the boron concentration in the primary pened at Indian Point 2, Indian Point 3, and Perry.
coolant must be adjusted to match the criticality Perry tripped due to a generator under-frequency
calculation. Near the end of the fuel cycle, the trip signal. Indian Point 2 and Indian Point 3 trip-
nuclear power plant may not have enough ped when the grid frequency dropped low enough
boron adjustment or control rod worth available to trip reactor coolant pumps, which actuated a
for restart until the neutron absorbers have reactor protective feature.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 115
In other cases, the electrical frequency fluctuated at least one safety bus at about 01:53 EDT on
and went higher than normal. Turbine control sys- August 15. The following equipment problems
tems responded in an attempt to control the fre- were noted: the Combustion Turbine Generator
quency. Equipment limits were exceeded as a (the alternate AC power source) failed to start from
result of the reaction of the turbine control sys- the control room; however, it was successfully
tems to large frequency changes. This led to trips started locally. In addition, the Spent Fuel Pool
at FitzPatrick, Nine Mile 1, Nine Mile 2, and Cooling System was interrupted for approxi-
Ginna. FitzPatrick and Nine Mile 2 tripped on low mately 26 hours and reached a maximum temper-
pressure in the turbine hydraulic control oil sys- ature of 130 degrees Fahrenheit (55 degrees
tem. Nine Mile 1 tripped on turbine light load pro- Celsius). The main generator was reconnected to
tection. Ginna tripped due to conditions in the the grid at about 01:41 EDT on August 20.
reactor following rapid closure of the turbine con-
trol valves in response to high frequency on the FitzPatrick. FitzPatrick is located about 8 miles
grid. (13 km) northeast of Oswego, NY, in northern New
York on Lake Ontario. It was generating about 850
The Perry, Fermi 2, Oyster Creek, and Nine Mile 1 MWe before the event. The reactor tripped due to
reactors tripped immediately after the generator low pressure in the hydraulic system that controls
tripped, although that is not apparent from the the turbine control valves. Low pressure in this
times below, because the clocks were not synchro- system typically indicates a large load reject, for
nized to the national time standard. The Indian which a reactor trip is expected. In this case the
Point 2 and 3, FitzPatrick, Ginna, and Nine Mile 2 pressure in the system was low because the con-
reactors tripped before the generators. When the trol system was rapidly manipulating the turbine
reactor trips first, there is generally a short time control valves to control turbine speed, which was
delay before the generator output breakers open. being affected by grid frequency fluctuations.
The electrical generation decreases rapidly to zero
after the reactor trip. Table 8.1 provides the times Immediately preceding the trip, both significant
from the data collected for the reactor trip times, over-voltage and under-voltage grid conditions
and the time the generator output breakers opened were experienced. Offsite power was subse-
(generator trip), as reported by the ESWG. Addi- quently lost to the plant auxiliary buses. The
tional details on the plants that tripped are given safety buses were deenergized and automatically
below, and summarized in Table 8.2 on page 120. reenergized from the emergency diesel generators.
Fermi 2. Fermi 2 is located 25 miles (40 km) north- The lowest emergency declaration, an Unusual
east of Toledo, Ohio, in southern Michigan on Event, was declared at about 16:26 EDT due to the
Lake Erie. It was generating about 1,130 mega- loss of offsite power. Decay heat removal systems
watts-electric (MWe) before the event. The reactor maintained the cooling function for the reactor
tripped due to a turbine trip. The turbine trip was fuel. Offsite power was restored to at least one
likely the result of multiple generator field protec- safety bus at about 23:07 EDT on August 14. The
tion trips (overexcitation and loss of field) as the main generator was reconnected to the grid at
Fermi 2 generator responded to a series of rapidly about 06:10 EDT on August 18.
changing transients prior to its loss. This is consis-
tent with data that shows large swings of the Fermi Table 8.1. U.S. Nuclear Plant Trip Times
2 generator MVAr prior to its trip. Nuclear Plant Reactor Tripa Generator Tripb
Perry . . . . . . . . . 16:10:25 EDT 16:10:42 EDT
Offsite power was subsequently lost to the plant Fermi 2 . . . . . . . 16:10:53 EDT 16:10:53 EDT
auxiliary buses. The safety buses were de- Oyster Creek . . . 16:10:58 EDT 16:10:57 EDT
energized and automatically reenergized from the
Nine Mile 1 . . . . 16:11 EDT 16:11:04 EDT
emergency diesel generators. The operators trip-
Indian Point 2 . . 16:11 EDT 16:11:09 EDT
ped one emergency diesel generator that was par-
Indian Point 3 . . 16:11 EDT 16:11:23 EDT
alleled to the grid for testing, after which it
automatically loaded. Decay heat removal systems FitzPatrick . . . . . 16:11:04 EDT 16:11:32 EDT
maintained the cooling function for the reactor Ginna. . . . . . . . . 16:11:36 EDT 16:12:17 EDT
fuel. Nine Mile 2 . . . . 16:11:48 EDT 16:11:52 EDT
aAs determined from licensee data (which may not be syn-
The lowest emergency declaration, an Unusual chronized to the national time standard).
Event, was declared at about 16:22 EDT due to the bAs reported by the Electrical System Working Group (syn-
loss of offsite power. Offsite power was restored to chronized to the national time standard).
116 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Ginna. Ginna is located 20 miles (32 km) north- reactor protection signal tripped the reactor,
east of Rochester, NY, in northern New York on which resulted in turbine and generator trips.
Lake Ontario. It was generating about 487 MWe
before the event. The reactor tripped due to Over- The auxiliary bus experienced the under-
Temperature-Delta-Temperature. This trip signal frequency due to fluctuating grid conditions.
protects the reactor core from exceeding tempera- Offsite power was lost to all the plant auxiliary
ture limits. The turbine control valves closed buses. The safety buses were reenergized from the
down in response to the changing grid conditions. emergency diesel generators. Decay heat removal
This caused a temperature and pressure transient systems maintained the cooling function for the
in the reactor, resulting in an Over-Temperature- reactor fuel.
Delta-Temperature trip. The lowest emergency declaration, an Unusual
Offsite power was not lost to the plant auxiliary Event, was declared at about 16:25 EDT due to the
buses. In the operators’ judgement, offsite power loss of offsite power for more than 15 minutes.
was not stable, so they conservatively energized Offsite power was restored to at least one safety
the safety buses from the emergency diesel genera- bus at about 20:02 EDT on August 14. The follow-
tors. Decay heat removal systems maintained the ing equipment problems were noted: the service
cooling function for the reactor fuel. Offsite power water to one of the emergency diesel generators
was not lost, and stabilized about 50 minutes after developed a leak; a steam generator atmospheric
the reactor trip. dump valve did not control steam generator pres-
sure in automatic and had to be shifted to manual;
The lowest emergency declaration, an Unusual a steam trap associated with the turbine-driven
Event, was declared at about 16:46 EDT due to the AFW pump failed open, resulting in operators
degraded offsite power. Offsite power was securing the turbine after 2.5 hours; loss of instru-
restored to at least one safety bus at about 21:08 ment air required operators to take manual control
EDT on August 14. The following equipment of charging and a letdown isolation occurred; and
problems were noted: the digital feedwater control operators in the field could not use radios; and the
system behaved in an unexpected manner follow- diesel generator for the Unit 2 Technical Support
ing the trip, resulting in high steam generator lev- Center failed to function. Also, several uninter-
els; there was a loss of RCP seal flow indication, ruptible power supplies in the Emergency Opera-
which complicated restarting the pumps; and at tions Facility failed. This reduced the capability
least one of the power-operated relief valves expe- for communications and data collection. Alternate
rienced minor leakage following proper operation equipment was used to maintain vital communi-
and closure during the transient. Also, one of the cations.1 The main generator was reconnected to
motor-driven auxiliary feedwater pumps was the grid at about 12:58 EDT on August 17.
damaged after running with low flow conditions
due to an improper valve alignment. The redun- Indian Point 3. Indian Point 3 is located 24 miles
dant pumps supplied the required water flow. (39 km) north of New York City on the Hudson
River. It was generating about 1,010 MWe before
The NRC issued a Notice of Enforcement Discre- the event. The reactor tripped due to loss of a reac-
tion to allow Ginna to perform mode changes and tor coolant pump that tripped because the auxil-
restart the reactor with one auxiliary feedwater iary bus frequency fluctuations actuated the
(AFW) pump inoperable. Ginna has two AFW under-frequency relay, which protects against
pumps, one turbine-driven AFW pump, and two inadequate coolant flow through the reactor core.
standby AFW pumps, all powered from safety- This reactor protection signal tripped the reactor,
related buses. The main generator was recon- which resulted in turbine and generator trips.
nected to the grid at about 20:38 EDT on August
17. The auxiliary bus experienced the under-
frequency due to fluctuating grid conditions.
Indian Point 2. Indian Point 2 is located 24 miles Offsite power was lost to all the plant auxiliary
(39 km) north of New York City on the Hudson buses. The safety buses were reenergized from the
River. It was generating about 990 MWe before the emergency diesel generators. Decay heat removal
event. The reactor tripped due to loss of a reactor systems maintained the cooling function for the
coolant pump that tripped because the auxiliary reactor fuel.
bus frequency fluctuations actuated the under-
frequency relay, which protects against inade- The lowest emergency declaration, an Unusual
quate coolant flow through the reactor core. This Event, was declared at about 16:23 EDT due to the
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 117
loss of offsite power for more than 15 minutes. several hours due to lack of procedure guidance to
Offsite power was restored to at least one safety address particular plant conditions encountered
bus at about 20:12 EDT on August 14. The follow- during the shutdown. The main generator was
ing equipment problems were noted: a steam gen- reconnected to the grid at about 02:08 EDT on
erator safety valve lifted below its desired setpoint August 18.
and was gagged; loss of instrument air, including
Nine Mile 2. Nine Mile 2 is located 6 miles (10 km)
failure of the diesel backup compressor to start
northeast of Oswego, NY, in northern New York
and failure of the backup nitrogen system,
on Lake Ontario. It was generating about 1,193
resulted in manual control of atmospheric dump
MWe before the event. The reactor scrammed due
valves and AFW pumps needing to be secured to
to the actuation of pressure switches which
prevent overfeeding the steam generators; a blown
detected low pressure in the hydraulic system that
fuse in a battery charger resulted in a longer bat-
controls the turbine control valves. Low pressure
tery discharge; a control rod drive mechanism
in this system typically indicates a large load
cable splice failed, and there were high resistance
reject, for which a reactor trip is expected. In this
readings on 345-kV breaker-1. These equipment
case the pressure in the system was low because
problems required correction prior to startup,
the control system was rapidly manipulating the
which delayed the startup. The diesel generator
turbine control valves to control turbine speed,
for the Unit 3 Technical Support Center failed to
which was being affected by grid frequency
function. Also, several uninterruptible power sup-
fluctuations.
plies in the Emergency Operations Facility failed.
This reduced the capability for communications After the reactor tripped, several reactor level con-
and data collection. Alternate equipment was trol valves did not reposition, and with the main
used to maintain vital communications.2 The feedwater system continuing to operate, a high
main generator was reconnected to the grid at water level in the reactor caused a turbine trip,
about 05:03 EDT on August 22. which caused a generator trip. Offsite power was
degraded but available to the plant auxiliary
Nine Mile 1. Nine Mile 1 is located 6 miles (10 km) buses. The offsite power dropped below the nor-
northeast of Oswego, NY, in northern New York mal voltage levels, which resulted in the safety
on Lake Ontario. It was generating about 600 MWe buses being automatically energized from the
before the event. The reactor tripped in response emergency diesel generators. Decay heat removal
to a turbine trip. The turbine tripped on light load systems maintained the cooling function for the
protection (which protects the turbine against a reactor fuel.
loss of electrical load), when responding to fluctu-
ating grid conditions. The turbine trip caused fast The lowest emergency declaration, an Unusual
closure of the turbine valves, which, through Event, was declared at about 17:00 EDT due to the
acceleration relays on the control valves, create a loss of offsite power to the safety buses for more
signal to trip the reactor. After a time delay of 10 than 15 minutes. Offsite power was restored to at
seconds, the generator tripped on reverse power. least one safety bus at about 01:33 EDT on August
15. The following additional equipment problem
The safety buses were automatically deenergized was noted: a tap changer on one of the offsite
due to low voltage and automatically reenergized power transformers failed, complicating the resto-
from the emergency diesel generators. Decay heat ration of one division of offsite power. The main
removal systems maintained the cooling function generator was reconnected to the grid at about
for the reactor fuel. 19:34 EDT on August 17.
The lowest emergency declaration, an Unusual Oyster Creek. Oyster Creek is located 9 miles (14
Event, was declared at about 16:33 EDT due to the km) south of Toms River, NJ, near the Atlantic
degraded offsite power. Offsite power was Ocean. It was generating about 629 MWe before
restored to at least one safety bus at about 23:39 the event. The reactor tripped due to a turbine trip.
EDT on August 14. The following additional The turbine trip was the result of a generator trip
equipment problems were noted: a feedwater due to actuation of a high Volts/Hz protective trip.
block valve failed “as is” on the loss of voltage, The Volts/Hz trip is a generator/transformer pro-
resulting in a high reactor vessel level; fuses blew tective feature. The plant safety and auxiliary
in fire circuits, causing control room ventilation buses transferred from the main generator supply
isolation and fire panel alarms; and operators were to the offsite power supply following the plant
delayed in placing shutdown cooling in service for trip. Other than the plant transient, no equipment
118 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
or performance problems were determined to be to the grid at about 23:15 EDT on August 21. After
directly related to the grid problems. the plant restarted, a surveillance test indicated a
problem with one emergency diesel generator.3
Post-trip the operators did not get the mode switch
to shutdown before main steam header pressure Nuclear Power Plants With a Significant
reached its isolation setpoint. The resulting MSIV Transient
closure complicated the operator’s response The electrical disturbance on August 14 had a sig-
because the normal steam path to the main con- nificant impact on seven plants that continued to
denser was lost. The operators used the isolation remain connected to the grid. For this review, sig-
condensers for decay heat removal. The plant nificant impact means that these plants had signif-
safety and auxiliary buses remained energized icant load adjustments that resulted in bypassing
from offsite power for the duration of the event, steam from the turbine generator, opening of relief
and the emergency diesel generators were not valves, or requiring the onsite emergency diesel
started. Decay heat removal systems maintained generators to automatically start due to low
the cooling function for the reactor fuel. The main voltage.
generator was reconnected to the grid at about
05:02 EDT on August 17. Nuclear Power Plants With a Non-Significant
Transient
Perry. Perry is located 7 miles (11 km) northeast of
Painesville, OH, in northern Ohio on Lake Erie. It Sixty-four nuclear power plants experienced
was generating about 1,275 MWe before the event. non-significant transients caused by minor distur-
The reactor tripped due to a turbine control valve bances on the electrical grid. These plants were
fast closure trip signal. The turbine control valve able to respond to the disturbances through nor-
fast closure trip signal was due to a generator mal control systems. Examples of these transients
under-frequency trip signal that tripped the gener- included changes in load of a few megawatts or
ator and the turbine and was triggered by grid fre- changes in frequency of a few-tenths Hz.
quency fluctuations. Plant operators noted voltage Nuclear Power Plants With No Transient
fluctuations and spikes on the main transformer,
and the Generator Out-of-Step Supervisory relay Twenty-four nuclear power plants experienced no
actuated approximately 30 minutes before the transient and saw essentially no disturbances on
trip. This supervisory relay senses a ground fault the grid, or were shut down at the time of the
on the grid. The purpose is to prevent a remote transient.
fault on the grid from causing a generator out-of-
step relay to activate, which would result in a gen- General Observations Based on the Facts
erator trip. Approximately 30 seconds prior to the Found During Phase One
trip operators noted a number of spikes on the gen-
The NWG found no evidence that the shutdown of
erator field volt meter, which subsequently went
U.S. nuclear power plants triggered the outage or
offscale high. The MVAr and MW meters likewise
inappropriately contributed to its spread (i.e., to
went offscale high.
an extent beyond the normal tripping of the plants
The safety buses were deenergized and automati- at expected conditions). This review did not iden-
cally reenergized from the emergency diesel gen- tify any activity or equipment issues that appeared
erators. Decay heat removal systems maintained to start the transient on August 14, 2003. All nine
the cooling function for the reactor fuel. The fol- plants that experienced a reactor trip were
lowing equipment problems were noted: a steam responding to grid conditions. The severity of the
bypass valve opened; a reactor water clean-up sys- transient caused generators, turbines, or reactor
tem pump tripped; the off-gas system isolated, and systems to reach a protective feature limit and
a keep-fill pump was found to be air-bound, actuate a plant shutdown.
requiring venting and filling before the residual
All nine plants tripped in response to those condi-
heat removal system loop A and the low pressure
tions in a manner consistent with the plant
core spray system could be restored to service.
designs. All nine plants safely shut down. All
The lowest emergency declaration, an Unusual safety functions were effectively accomplished,
Event, was declared at about 16:20 EDT due to the with few problems, and the plants were main-
loss of offsite power. Offsite power was restored to tained in a safe shutdown condition until their
at least one safety bus at about 18:13 EDT on restart. Fermi 2, Nine Mile 1, Oyster Creek, and
August 14. The main generator was reconnected Perry tripped on turbine and generator protective
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 119
features. FitzPatrick, Ginna, Indian Point 2 and 3, August 22. Indian Point 3 had equipment issues
and Nine Mile 2 tripped on reactor protective (failed splices in the control rod drive mechanism
features. power system) that required repair prior to restart.
Ginna submitted a special request for enforcement
Nine plants used their emergency diesel genera-
discretion from the NRC to permit mode changes
tors to power their safety-related buses during the
and restart with an inoperable auxiliary feedwater
power outage. Offsite power was restored to the
pump. The NRC granted the request for enforce-
safety buses after the grid was energized and the
ment discretion.
plant operators, in consultation with the transmis-
sion system operators, decided the grid was stable.
Although the Oyster Creek plant tripped, offsite Conclusions of the U.S. Nuclear
power was never lost to their safety buses and the Working Group
emergency diesel generators did not start and
As discussed above, the investigation of the U.S.
were not required. Another plant, Davis-Besse,
nuclear power plant responses during the
was already shut down but lost power to the safety
blackout found no significant deficiencies.
buses. The emergency diesel generators started
Accordingly, there are no recommendations here
and provided power to the safety buses as
concerning U.S. nuclear power plants. Some areas
designed.
for consideration on a grid-wide basis were dis-
For the eight remaining tripped plants and cussed and forwarded to the Electric System
Davis-Besse (which was already shut down prior Working Group for their review.
to the events of August 14), offsite power was
On August 14, 2003, nine U.S. nuclear power
restored to at least one safety bus after a period of
plants tripped as a result of the loss of offsite
time ranging from about 2 hours to about 14 hours,
power. Nuclear power plants are designed to cope
with an average time of about 7 hours. Although
with the loss of offsite power (LOOP) through the
Ginna did not lose offsite power, the operators
use of emergency power supplies (primarily
judged offsite power to be unstable and realigned
on-site diesel generators). The safety function of
the safety buses to the emergency diesel
most concern during a LOOP is the removal of
generators.
heat from the reactor core. Although the control
The licensees’ return to power operation follows a rods have been inserted to stop the fission process,
deliberate process controlled by plant procedures the continuing decay of radioactive isotopes in the
and NRC regulations. Ginna, Indian Point 2, Nine reactor core produces a significant amount of heat
Mile 2, and Oyster Creek resumed electrical gener- for many weeks. If this decay heat is not removed,
ation on August 17. FitzPatrick and Nine Mile 1 it will cause fuel damage and the release of highly
resumed electrical generation on August 18. Fermi radioactive isotopes from the reactor core. The
2 resumed electrical generation on August 20. failure of the alternating current emergency power
Perry resumed electrical generation on August 21. supplies in conjunction with a LOOP is known
Indian Point 3 resumed electrical generation on as a station blackout. Failures of the emergency
120 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
power supplies would seriously hinder the ability of the event either automatically shut down as a
of the plant operators to carry out the required result of the grid disturbance or automatically
safety functions. Nuclear plants can cope with a reduced power while waiting for the grid to be
station blackout for a limited time without suffer- reestablished. In addition, the Point Lepreau
ing fuel damage. However, recovery of the grid or Nuclear Generating Station in New Brunswick
the restoration of an emergency power supply is was forced to reduce electricity production for a
needed for long-term decay heat removal. For this short period.
reason, the NRC considers LOOP events to be
potential precursors to more serious situations. The Canadian NWG (CNWG) was mandated to:
The risk of reactor core damage increases as the review the sequence of events for each Canadian
LOOP frequency or duration increases. nuclear plant; determine whether any events
caused or contributed to the power system outage;
Offsite power is considered the preferred power evaluate any potential safety issues arising as a
source for responding to all off-normal events or result of the event; evaluate the effect on safety
accidents. However, if the grid is operated in a and the reliability of the grid of design features,
stressed configuration, the loss of the nuclear operating procedures, and regulatory require-
plant generation may result in grid voltage drop- ments at Canadian nuclear power plants; and
ping below the level needed for the plant safety assess the impact of associated regulator perfor-
loads. In that case, each plant is designed such mance and regulatory decisions.
that voltage relays will automatically disconnect
the plant safety-related electrical buses from the In Ontario, 11 nuclear units were operating and
grid and reenergize them from the emergency die- delivering power to the grid at the time of the grid
sel generators (EDGs). Although the resultant disturbance: 4 at Bruce B, 4 at Darlington, and 3 at
safety system responses have been analyzed and Pickering B. Of the 11 reactors, 7 shut down as a
found acceptable, the loss of offsite power reduces result of the event (1 at Bruce B, 3 at Darlington,
the plant’s safety margin. It also increases the risk and 3 at Pickering B). Four reactors (3 at Bruce B
associated with failures of the EDGs. For these rea- and 1 at Darlington) disconnected safely from the
sons, the NRC periodically assesses the impact of grid but were able to avoid shutting down and
grid reliability on overall nuclear plant safety. were available to supply power to the Ontario grid
as soon as reconnection was enabled by Ontario’s
The NRC monitors grid reliability under its nor- Independent Market Operator (IMO).
mal monitoring programs, such as the operating
experience program, and has previously issued New Brunswick Power’s Point Lepreau Generating
reports related to grid reliability. The NRC is con- Station responded to the loss of grid event by cut-
tinuing with an internal review of the reliability of ting power to 460 MW, returning to fully stable
the electrical grid and the effect on the risk profile conditions at 16:35 EDT, within 25 minutes of the
for nuclear power plants. The NRC will consider event. Hydro Québec’s (HQ) grid was not affected
the implications of the August 14, 2003, Northeast by the power system outage, and HQ’s Gentilly-2
blackout under the NRC’s regulations. The NRC nuclear station continued to operate normally.
is conducting an internal review of its station Having reviewed the operating data for each plant
blackout rule, and the results of the August 14th and the responses of the power stations and their
event will be factored into that review. If there are staff to the event, the CNWG concludes the
additional findings, the NRC will address them following:
through the NRC’s normal process.
u None of the reactor operators had any advanced
warning of impending collapse of the grid.
Findings of the Canadian Nuclear
ã Trend data obtained indicate stable condi-
Working Group tions until a few minutes before the event.
Summary ã There were no prior warnings from Ontario’s
IMO.
On the afternoon of August 14, 2003, southern
Ontario, along with the northeastern United u Canadian nuclear power plants did not trigger
States, experienced a widespread electrical power the power system outage or contribute to its
system outage. Eleven nuclear power plants in spread. Rather they responded, as anticipated,
Ontario operating at high power levels at the time in order to protect equipment and systems from
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 121
the grid disturbances. Plant data confirm the rapid reductions in reactor power. While
following. greater flexibility with respect to use of adjuster
ã At Bruce B and Pickering B, frequency and/or
rods would not have prevented the shutdown,
voltage fluctuations on the grid resulted in some units, particularly those at Darlington,
the automatic disconnection of generators might have been able to return to service less
from the grid. For those units that were suc- than 1 hour after the initiating event.
cessful in maintaining the unit generators u Off-site power was unavailable for varying peri-
operational, reactor power was automatically ods of time, from approximately 3 hours at
reduced. Bruce B to approximately 9 hours at Pickering
A. Despite the high priority assigned by the IMO
ã At Darlington, load swing on the grid led to
to restoring power to the nuclear stations, the
the automatic reduction in power of the four
stations had some difficulty in obtaining timely
reactors. The generators were, in turn, auto-
information about the status of grid recovery
matically disconnected from the grid.
and the restoration of Class IV power. This
ã Three reactors at Bruce B and one at Darling- information is important for Ontario Power
ton were returned to 60% power. These reac- Generation’s and Bruce Power’s response
tors were available to deliver power to the strategy.
grid on the instructions of the IMO.
u Required regulatory approvals from CNSC staff
ã Three units at Darlington were placed in a were obtained quickly and did not delay the
zero-power hot state, and four units at restart of the units; however, CNSC staff was
Pickering B and one unit at Bruce B were unable to immediately activate the CNSC’s
placed in a guaranteed shutdown state. Emergency Operation Centre because of loss of
u There were no risks to health and safety of power to the CNSC’s head office building.
workers or the public as a result of the shut- CNSC staff, therefore, established communica-
down of the reactors. tions with licensees and the U.S. NRC from
other locations.
ã Turbine, generator, and reactor automatic
safety systems worked as designed to Introduction
respond to the loss of grid. The primary focus of the CNWG during Phase I
ã Station operating staff and management fol- was to address nuclear power plant response rele-
lowed approved Operating Policies & Princi- vant to the power outage of August 14, 2003. Data
ples (OP&Ps) in responding to the loss of grid. were collected from each power plant and ana-
At all times, operators and shift supervisors lyzed in order to determine: the cause of the power
made appropriately conservative decisions in outage; whether any activities at these plants
favor of protecting health and safety. caused or contributed to the power outage; and
The CNWG commends the staff of Ontario Power whether there were any significant safety issues.
Generation and Bruce Power for their response to In order to obtain reliable and comparable infor-
the power system outage. At all times, staff acted mation and data from each nuclear power plant, a
in accordance with established OP&Ps, and took questionnaire was developed to help pinpoint
an appropriately conservative approach to how each nuclear power plant responded to the
decisions. August 14 grid transients. Where appropriate,
additional information was obtained from the
During the course of its review, the CNWG also ESWG and SWG.
identified the following secondary issues:
The operating data from each plant were com-
u Equipment problems and design limitations at
pared against the plant design specifications to
Pickering B resulted in a temporary reduction in determine whether the plants responded as
the effectiveness of some of the multiple safety expected. Based on initial plant responses to the
barriers, although the equipment failure was questionnaire, supplemental questions were
within the unavailability targets found in the developed, as required, to further clarify outstand-
OP&Ps approved by the CNSC as part of Ontario ing matters. Supplementary information on the
Power Generation’s licence. design features of Ontario’s nuclear power plants
u Existing OP&Ps place constraints on the use of was also provided by Ontario Power Generation
adjuster rods to respond to events involving and Bruce Power. The CNWG also consulted a
122 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
number of subject area specialists, including maintain the reactor at 60% full power if the gen-
CNSC staff, to validate the responses to the ques- erator becomes disconnected from the grid and to
tionnaire and to ensure consistency in their maintain this “readiness” condition if necessary
interpretation. for days. Once reconnected to the grid, the unit
In addition to the stakeholder consultations dis- can be loaded to 60% full power within several
cussed in the Introduction to this chapter, CNSC minutes and can achieve full power within 24
staff met with officials from Ontario’s Independ- hours.
ent Market Operator on January 7, 2004. As with other nuclear reactors, CANDU reactors
normally operate continuously at full power
Typical Design, Operational, and except when shut down for maintenance and
Protective Features of CANDU Nuclear inspections. As such, while they provide a stable
Power Plants source of baseload power generation, they cannot
provide significant additional power in response
There are 22 CANDU nuclear power reactors in
to sudden increases in demand. CANDU power
Canada—20 located in Ontario at 5 multi-unit sta-
plants are not designed for black-start operation;
tions (Pickering A and Pickering B located in
that is, they are not designed to start up in the
Pickering, Darlington located in the Municipality
absence of power from the grid.
of Clarington, and Bruce A and Bruce B located
near Kincardine). There are also single-unit Electrical Distribution Systems
CANDU stations at Bécancour, Québec (Gentilly-
2), and Point Lepreau, New Brunswick. The electrical distribution systems at nuclear
power plants are designed to satisfy the high
In contrast to the pressurized water reactors used safety and reliability requirements for nuclear sys-
in the United States, which use enriched uranium tems. This is achieved through flexible bus
fuel and a light water coolant-moderator, all arrangements, high capacity standby power gener-
housed in a single, large pressure vessel, a CANDU ation, and ample redundancy in equipment.
reactor uses fuel fabricated from natural uranium,
with heavy water as the coolant and moderator. Where continuous power is required, power is
The fuel and pressurized heavy water coolant are supplied either from batteries (for continuous DC
contained in 380 to 480 pressure tubes housed in a power, Class I) or via inverters (for continuous AC
calandria containing the heavy water moderator power, Class II). AC supply for safety-related
under low pressure. Heat generated by the fuel is equipment, which can withstand short interrup-
removed by heavy water coolant that flows tion (on the order of 5 minutes), is provided by
through the pressure tubes and is then circulated Class III power. Class III power is nominally sup-
to the boilers to produce steam from demineral- plied through Class IV; when Class IV becomes
ized water. unavailable, standby generators are started auto-
matically, and the safety-related loads are picked
While the use of natural uranium fuel offers up within 5 minutes of the loss of Class IV power.
important benefits from the perspectives of safe-
guards and operating economics, one drawback is The Class IV power is an AC supply to reactor
that it restricts the ability of a CANDU reactor to equipment and systems that can withstand longer
recover from a large power reduction. In particu- interruptions in power. Class IV power can be sup-
lar, the lower reactivity of natural uranium fuel plied either from the generator through a trans-
means that CANDU reactors are designed with a former or from the grid by another transformer.
small number of control rods (called “adjuster Class IV power is not required for reactors to shut
rods”) that are only capable of accommodating down safely.
power reductions to 60%. The consequence of a In addition to the four classes of power described
larger power reduction is that the reactor will “poi- above, there is an additional source of power
son out” and cannot be made critical for up to 2 known as the Emergency Power System (EPS).
days following a power reduction. By comparison, EPS is a separate power system consisting of its
the use of enriched fuel enables a typical pressur- own on-site power generation and AC and DC dis-
ized water reactor to operate with a large number tribution systems whose normal supply is from
of control rods that can be withdrawn to accom- the Class III power system. The purpose of the EPS
modate power reductions to zero power. system is to provide power to selected safety-
A unique feature of some CANDU plants— related loads following common mode incidents,
namely, Bruce B and Darlington—is a capability to such as seismic events.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 123
Protective Features of CANDU Nuclear Power of the Class III powered shutdown cooling system;
Plants heat removal to the shutdown cooling heat
exchangers is by means of the Class III powered
CANDU reactors typically have two separate,
components of the Service Water System.
independent and diverse systems to shut down
the reactor in the event of an accident or transients
in the grid. Shutdown System 1 (SDS1) consists of
CANDU Reactor Response to
a large number of cadmium rods that drop into the Loss-of-Grid Event
core to decrease the power level by absorbing neu- Response to Loss of Grid
trons. Shutdown System 2 (SDS2) consists of
high-pressure injection of gadolinium nitrate into In the event of disconnection from the grid, power
the low-pressure moderator to decrease the power to shut down the reactor safely and maintain
level by absorbing neutrons. Although Pickering A essential systems will be supplied from batteries
does not have a fully independent SDS2, it does and standby generators. The specific response of a
have a second shutdown mechanism, namely, the reactor to disconnection from the grid will depend
fast drain of the moderator out of the calandria; on the reactor design and the condition of the unit
removal of the moderator significantly reduces the at the time of the event.
rate of nuclear fission, which reduces reactor
60% Reactor Power: All CANDU reactors are
power. Also, additional trip circuits and shutoff
designed to operate at 60% of full power following
rods have recently been added to Pickering A Unit
the loss of off-site power. They can operate at this
4 (Shutdown System Enhancement, or SDS-E).
level as long as demineralized water is available
Both SDS1 and SDS2 are capable of reducing reac-
for the boilers. At Darlington and Bruce B, steam
tor power from 100% to about 2% within a few
can be diverted to the condensers and recirculated
seconds of trip initiation.
to the boilers. At Pickering A and Pickering B,
Fuel Heat Removal Features of CANDU excess steam is vented to the atmosphere, thereby
Nuclear Power Plants limiting the operating time to the available inven-
tory of demineralized water.
Following the loss of Class IV power and shut-
down of the reactor through action of SDS1 and/or 0% Reactor Power, Hot: The successful transition
SDS2, significant heat will continue to be gener- from 100% to 60% power depends on several sys-
ated in the reactor fuel from the decay of fission tems responding properly, and continued opera-
products. The CANDU design philosophy is to tion is not guaranteed. The reactor may shut down
provide defense in depth in the heat removal automatically through the operation of the process
systems. control systems or through the action of either of
the shutdown systems.
Immediately following the trip and prior to resto-
ration of Class III power, heat will be removed Should a reactor shutdown occur following a load
from the reactor core by natural circulation of rejection, both Class IV power supplies (from the
coolant through the Heat Transport System main generator and the grid) to that unit will become
circuit following rundown of the main Heat Trans- unavailable. The main Heat Transport pumps
port pumps (first by thermosyphoning and later by will trip, leading to a loss of forced circulation of
intermittent buoyancy induced flow). Heat will be coolant through the core. Decay heat will be con-
rejected from the secondary side of the steam gen- tinuously removed through natural circulation
erators through the atmospheric steam discharge (thermosyphoning) to the boilers, and steam pro-
valves. This mode of operation can be sustained duced in the boilers will be exhausted to the
for many days with additional feedwater supplied atmosphere via atmospheric steam discharge
to the steam generators via the Class III powered valves. The Heat Transport System will be main-
auxiliary steam generator feed pump(s). tained at around 250 to 265 degrees Celsius during
thermosyphoning. Standby generators will start
In the event that the auxiliary feedwater system automatically and restore Class III power to key
becomes unavailable, there are two alternate EPS safety-related systems. Forced circulation in the
powered water supplies to steam generators, Heat Transport System will be restored once
namely, the Steam Generator Emergency Coolant either Class III or Class IV power is available.
System and the Emergency Service Water System.
Finally, a separate and independent means of When shut down, the natural decay of fission
cooling the fuel is by forced circulation by means products will lead to the temporary buildup of
124 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
neutron absorbing elements in the fuel. If the reac- Summary of Canadian Nuclear Power
tor is not quickly restarted to reverse this natural Plant Response to and Safety During the
process, it will “poison-out.” Once poisoned-out,
August 14 Outage
the reactor cannot return to operation until the fis-
sion products have further decayed, a process On the afternoon of August 14, 2003, 15 Canadian
which typically takes up to 2 days. nuclear units were operating: 13 in Ontario, 1 in
Québec, and 1 in New Brunswick. Of the 13
Overpoisoned Guaranteed Shutdown State: In Ontario reactors that were critical at the time of
the event that certain problems are identified the event, 11 were operating at or near full power
when reviewing the state of the reactor after a sig- and 2 at low power (Pickering B Unit 7 and
nificant transient, the operating staff will cool Pickering A Unit 4). All 13 of the Ontario reactors
down and depressurize the reactor, then place it in disconnected from the grid as a result of the grid
an overpoisoned guaranteed shutdown state (GSS) disturbance. Seven of the 11 reactors operating at
through the dissolution of gadolinium nitrate into high power shut down, while the remaining 4
the moderator. Maintenance will then be initiated operated in a planned manner that enabled them
to correct the problem. to remain available to reconnect to the grid at the
request of Ontario’s IMO. Of the 2 Ontario reactors
Return to Service Following Loss of Grid operating at low power, Pickering A Unit 4 tripped
The return to service of a unit following any one of automatically, and Pickering B Unit 7 was tripped
the above responses to a loss-of-grid event is dis- manually and shut down. In addition, a transient
cussed below. It is important to note that the was experienced at New Brunswick Power’s Point
descriptions provided relate to operations on a Lepreau Nuclear Generating Station, resulting in a
single unit. At multi-unit stations, the return to reduction in power. Hydro Québec’s Gentilly-2
service of several units cannot always proceed in nuclear station continued to operate normally as
parallel, due to constraints on labor availability the Hydro Québec grid was not affected by the grid
and the need to focus on critical evolutions, such disturbance.
as taking the reactor from a subcritical to a critical Nuclear Power Plants With Significant
state. Transients
60% Reactor Power: In this state, the unit can be Pickering Nuclear Generating Station. The
resynchronized consistent with system demand, Pickering Nuclear Generating Station (PNGS) is
and power can be increased gradually to full located in Pickering, Ontario, on the shores of
power over approximately 24 hours. Lake Ontario, 19 miles (30 km) east of Toronto. It
houses 8 nuclear reactors, each capable of deliver-
0% Reactor Power, Hot: In this state, after approx- ing 515 MW to the grid. Three of the 4 units at
imately 2 days for the poison-out, the turbine can Pickering A (Units 1 through 3) have been shut
be run up and the unit synchronized. Thereafter, down since late 1997. Unit 4 was restarted earlier
power can be increased to high power over the this year following a major refurbishment and was
next day. This restart timeline does not include in the process of being commissioned at the time
the time required for any repairs or maintenance of the event. At Pickering B, 3 units were operating
that might have been necessary during the outage. at or near 100% prior to the event, and Unit 7 was
being started up following a planned maintenance
Overpoisoned Guaranteed Shutdown State: Plac-
outage.
ing the reactor in a GSS after it has been shut down
requires approximately 2 days. Once the condi- Pickering A. As part of the commissioning process,
tion that required entry to the GSS is rectified, the Unit 4 at Pickering A was operating at 12% power
restart requires removal of the guarantee, removal in preparation for synchronization to the grid. The
of the gadolinium nitrate through ion exchange reactor automatically tripped on SDS1 due to Heat
process, heatup of the Heat Transport System, and Transport Low Coolant Flow, when the Heat
finally synchronization to the grid. Approximately Transport main circulating pumps ran down fol-
4 days are required to complete these restart activ- lowing the Class IV power loss. The decision was
ities. In total, 6 days from shutdown are required then made to return Unit 4 to the guaranteed shut-
to return a unit to service from the GSS, and this down state. Unit 4 was synchronized to the grid on
excludes any repairs that might have been August 20, 2003. Units 1, 2 and 3 were in lay-up
required while in the GSS. mode.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 125
Pickering B. The Unit 5 Generator Excitation Sys- delayed because of low suction pressure supply-
tem transferred to manual control due to large ing the Emergency High Pressure Service Water
voltage oscillations on the grid at 16:10 EDT and pumps. Manual operator intervention was
then tripped on Loss of Excitation about 1 second required to restore some pumps back to service.
later (prior to grid frequency collapse). In response
to the generator trip, Class IV buses transferred to Units were synchronized to the grid as follows:
the system transformer and the reactor setback. Unit 8 on August 22, Unit 5 on August 23, Unit 6
The grid frequency collapse caused the System on August 25, and Unit 7 on August 29.
Service Transformer to disconnect from the grid, Darlington Nuclear Generating Station. Four
resulting in a total loss of Class IV power. The reactors are located at the Darlington Nuclear Gen-
reactor consequently tripped on the SDS1 Low eration Station, which is on the shores of Lake
Gross Flow parameter followed by an SDS2 trip Ontario in the Municipality of Clarington, 43
due to Low Core Differential Pressure. miles (70 km) east of Toronto. All four of the reac-
The Unit 6 Generator Excitation System also tors are licensed to operate at 100% of full power,
transferred to manual control at 16:10 EDT due to and each is capable of delivering approximately
large voltage oscillations on the grid and the gen- 880 MW to the grid.
erator remained connected to the grid in manual Unit 1 automatically stepped back to the 60%
voltage control. Approximately 65 seconds into reactor power state upon load rejection at 16:12
the event, the grid under-frequency caused all the EDT. Approval by the shift supervisor to automati-
Class IV buses to transfer to the Generator Service cally withdraw the adjuster rods could not be pro-
Transformer. Ten seconds later, the generator sep- vided due to the brief period of time for the shift
arated from the Grid. Five seconds later, the gener- supervisor to complete the verification of systems
ator tripped on Loss of Excitation, which caused a as per procedure. The decreasing steam pressure
total loss of Class IV power. The reactor conse- and turbine frequency then required the reactor to
quently tripped on the SDS1 Low Gross Flow be manually tripped on SDS1, as per procedure for
parameter, followed by an SDS2 trip due to Low loss of Class IV power. The trip occurred at 16:24
Core Differential Pressure. EDT, followed by a manual turbine trip due to
Unit 7 was coming back from a planned mainte- under-frequency concerns.
nance outage and was at 0.9% power at the time of Like Unit 1, Unit 2 automatically stepped back
the event. The unit was manually tripped after upon load rejection at 16:12 EDT. As with Unit 1,
loss of Class IV power, in accordance with proce- there was insufficient time for the shift supervisor
dures and returned to guaranteed shutdown state. to complete the verification of systems, and faced
Unit 8 reactor automatically set back on load rejec- with decreasing steam pressure and turbine fre-
tion. The setback would normally have been ter- quency, the decision was made to shut down Unit
minated at 20% power but continued to 2% power 2. Due to under-frequency on the main Primary
because of the low boiler levels. The unit subse- Heat Transport pumps, the turbine was tripped
quently tripped on the SDS1 Low Boiler Feedline manually which resulted in an SDS1 trip at 16:28
Pressure parameter due to a power mismatch EDT.
between the reactor and the turbine. Unit 3 experienced a load rejection at 16:12 EDT,
The following equipment problems were noted. At and during the stepback Unit 3 was able to sustain
Pickering, the High Pressure Emergency Coolant operation with steam directed to the condensers.
Injection System (HPECIS) pumps are designed to After system verifications were complete, approv-
operate from a Class IV power supply. As a result al to place the adjuster rods on automatic was
of the shutdown of all the operating units, the obtained in time to recover, at 59% reactor power.
HPECIS at both Pickering A and Pickering B The unit was available to resynchronize to the
became unavailable for 5.5 hours. (The design of grid.
Pickering A and Pickering B HPECIS must be such
Unit 4 experienced a load rejection at 16:12 EDT,
that the fraction of time for which it is not avail-
and required a manual SDS1 trip due to the loss of
able can be demonstrated to be less than 10-3
Class II bus. This was followed by a manual tur-
years—about 8 hours per year. This was the first
bine trip.
unavailability of the HPECIS for 2003.) In addi-
tion, Emergency High Pressure Service Water The following equipment problems were noted:
System restoration for all Pickering B units was Unit 4 Class II inverter trip on BUS A3 and
126 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
subsequent loss of critical loads prevented unit Units were synchronized to the grid as follows:
recovery. The Unit 0 Emergency Power System Unit 8 at 19:14 EDT on August 14, 2003; Unit 5 at
BUS B135 power was lost until the Class III power 21:04 EDT on August 14; and Unit 7 at 21:14 EDT
was restored. (A planned battery bank B135 on August 14, 2003. Unit 6 was resynchronized at
change out was in progress at the time of the 02:03 EDT on August 23, 2003, after maintenance
blackout.) was conducted.
Units were synchronized to the grid as follows: Point Lepreau Nuclear Generating Station. The
Unit 3 at 22:00 EDT on August 14; Unit 2 on Point Lepreau nuclear station overlooks the Bay of
August 17, 2003; Unit 1 on August 18, 2003; and Fundy on the Lepreau Peninsula, 25 miles (40 km)
Unit 4 on August 18, 2003. southwest of Saint John, New Brunswick. Point
Lepreau is a single-unit CANDU 6, designed for a
Bruce Power. Eight reactors are located at Bruce
gross output of 680 MW. It is owned and operated
Power on the eastern shore of Lake Huron between
by New Brunswick Power.
Kincardine and Port Elgin, Ontario. Units 5
through 8 are capable of generating 840 MW each. Point Lepreau was operating at 91.5% of full
Presently these reactors are operating at 90% of power (610 MWe) at the time of the event. When
full power due to license conditions imposed by the event occurred, the unit responded to changes
the CNSC. Units 1 through 4 have been shut down in grid frequency as per design. The net impact
since December 31, 1997. At the time of the event, was a short-term drop in output by 140 MW, with
work was being performed to return Units 3 and 4 reactor power remaining constant and excess ther-
to service. mal energy being discharged via the unit steam
discharge valves. During the 25 seconds of the
Bruce A. Although these reactors were in guaran-
event, the unit stabilizer operated numerous times
teed shutdown state, they were manually tripped,
to help dampen the turbine generator speed oscil-
in accordance with operating procedures. SDS1
lations that were being introduced by the grid fre-
was manually tripped on Units 3 and 4, as per pro-
quency changes. Within 25 minutes of the event
cedures for a loss of Class IV power event. SDS1
initiation, the turbine generator was reloaded to
was re-poised on both units when the station
610 MW. Given the nature of the event that
power supplies were stabilized. The emergency
occurred, there were no unexpected observations
transfer system functioned as per design, with the
on the New Brunswick Power grid or at Point
Class III standby generators picking up station
Lepreau Generating Station throughout the ensu-
electrical loads. The recently installed Qualified
ing transient.
Diesel Generators received a start signal and were
available to pick up emergency loads if necessary. Nuclear Power Plants With No Transient
Bruce B. Units 5, 6, 7, and 8 experienced initial Gentilly-2 Nuclear Station. Hydro Québec owns
generation rejection and accompanying stepback and operates Gentilly-2 nuclear station, located on
on all four reactor units. All generators separated the south shore of the St. Lawrence River opposite
from the grid on under-frequency at 16:12 EDT. the city of Trois-Rivières, Québec. Gentilly-2 is
Units 5, 7, and 8 maintained reactor power at 60% capable of delivering approximately 675 MW to
of full power and were immediately available for Hydro Québec’s grid. The Hydro Québec grid was
reconnection to the grid. not affected by the power system outage and
Although initially surviving the loss of grid event, Gentilly-2 continued to operate normally.
Unit 6 experienced an SDS1 trip on insufficient
General Observations Based on the Facts
Neutron Over Power (NOP) margin. This occurred
while withdrawing Bank 3 of the adjusters in an Found During Phase I
attempt to offset the xenon transient, resulting in a Following the review of the data provided by the
loss of Class IV power. Canadian nuclear power plants, the CNWG con-
The following equipment problems were noted: cludes the following:
An adjuster rod on Unit 6 had been identified on u None of the reactor operators had any advanced
August 13, 2003, as not working correctly. Unit 6 warning of impending collapse of the grid.
experienced a High Pressure Recirculation Water
line leak, and the Closed Loop Demineralized u Canadian nuclear power plants did not trigger
Water loop lost inventory to the Emergency Water the power system outage or contribute to its
Supply System. spread.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 127
u There were no risks to the health and safety of rapid reductions in reactor power. While
workers or the public as a result of the concur- greater flexibility with respect to use of adjuster
rent shutdown of several reactors. Automatic rods would not have prevented the shutdown,
safety systems for the turbine generators and some units, particularly those at Darlington,
reactors worked as designed. (See Table 8.3 for might have been able to return to service less
a summary of shutdown events for Canadian than 1 hour after the initiating event.
nuclear power plants.) u Off-site power was unavailable for varying peri-
The CNWG also identified the following second- ods of time, from approximately 3 hours at
ary issues: Bruce B to approximately 9 hours at Pickering
A. Despite the high priority assigned by the IMO
u Equipment problems and design limitations at to restoring power to the nuclear stations, the
Pickering B resulted in a temporary reduction in stations had some difficulty obtaining timely
the effectiveness of some of the multiple safety information about the status of grid recovery
barriers, although the equipment failure was and the restoration of Class IV power. This
within the unavailability targets found in the information is important for Ontario Power
OP&Ps approved by the CNSC as part of Ontario Generation’s and Bruce Power’s response
Power Generation’s license. strategy.
u Existing OP&Ps place constraints on the use of u Required regulatory approvals from CNSC staff
adjuster rods to respond to events involving were obtained quickly and did not delay the
Table 8.3. Summary of Shutdown Events for Canadian Nuclear Power Plants
Operating Status
at Time of Event Response to Event
Stepback to Reactor Trip
60% Power,
Full Not Available To Turbine
Generating Station Unit Power Startup Operating Supply Grid Trip SDS1 SDS2
Pickering NGS 1 √ (a)
2 √
3 √
4 √ √ (b)
5 √ √ √
6 √ √ √
7 √ √
8 √ √
Darlington NGS 1 √ √ √
2 √ √ √
3 √ √
4 √ √ √
Bruce Nuclear Power 1 √
Development 2 √
3 √ √
4 √ √
5 √ √
6 √ √
7 √ √
8 √ √
aPickering A Unit 1 tripped as a result of electrical bus configuration immediately prior to the event which resulted in a temporary
loss of Class II power.
bPickering A Unit 4 also tripped on SDS-E.
Notes: Unit 7 at Pickering B was operating at low power, warming up prior to reconnecting to the grid after a maintenance outage.
Unit 4 at Pickering A was producing at low power, as part of the reactor’s commissioning after extensive refurbishment since being
shut down in 1997.
128 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
restart of the units; however, CNSC staff was another concerning the use of adjuster rods during
unable to immediately activate the CNSC’s future events involving the loss of off-site power.
Emergency Operation Centre because of loss of These are presented in Chapter 10 along with the
power to the CNSC’s head office building. Task Force’s recommendations on other subjects.
CNSC staff, therefore, established communica-
tions with licensees and the U.S. NRC from Despite some comments to the contrary, the
other locations. CNWG’s investigation found that the time to
restart the reactors was reasonable and in line
Regulatory Activities Subsequent to the with design specifications for the reactors. There-
Blackout fore, the CNWG made no recommendations for
action on this matter. Comments were also made
The actuation of emergency shutdown systems at regarding the adequacy of generation capacity in
Bruce, Darlington and Pickering, and the impair- Ontario and the appropriate mix of technologies
ment of the High Pressure Emergency Coolant for electricity generation. This is a matter beyond
Injection System (HPECIS) at Pickering are events the CNWG’s mandate, and it made no recommen-
for which licensees need to file reports with the dations on this issue.
Canadian Nuclear Safety Commission (CNSC), in
accordance with Regulatory Standard S 99,
“Reporting Requirements for Operating Nuclear Perspective of
Power Plants.” Reports have been submitted by
Ontario Power Generation (OPG) and Bruce
Nuclear Regulatory Agencies
Power, and are being followed up by staff from the on Potential Changes to the Grid
CNSC as part of the CNSC’s normal regulatory
process. This includes CNSC’s review and The NRC and the CNSC, under their respective
approval, where appropriate, of any actions taken regulatory authorities, are entrusted with provid-
or proposed to be taken to correct any problems in ing reasonable assurance of adequate protection of
design, equipment or operating procedures identi- public health and safety. As the design and opera-
fied by OPG and Bruce Power. tion of the electricity grid is taken into account
when evaluating the safety analysis of nuclear
As a result of further information about the event power plants, changes to the electricity grid must
gathered by CNSC staff during followup inspec- be evaluated for the impact on plant safety. As the
tions, the temporary impairment of the HPECIS at Task Force final recommendations result in
Pickering has been rated by CNSC staff as Level 2 actions to affect changes, the NRC and the CNSC
on the International Nuclear Event Scale, indicat- will assist by evaluating potential effects on the
ing that there was a significant failure in safety safety of nuclear power plant operation.
provisions, but with sufficient backup systems, or
“defense-in-depth,” in place to cope with potential The NRC and the CNSC acknowledge that future
malfunctions. Since August 2003, OPG has imple- improvements in grid reliability will involve coor-
mented procedural and operational changes to dination among many groups. The NRC and the
improve the performance of the safety systems at CNSC intend to maintain the good working rela-
Pickering. tionships that have been developed during the
Task Force investigation to ensure that we con-
Conclusions of the Canadian Nuclear tinue to share experience and insights and work
Working Group together to maintain an effective and reliable elec-
tric supply system.
As discussed above, Canadian nuclear power
plants did not trigger the power system outage or
Endnotes
contribute to its spread. The CNWG therefore 1 Further details are available in the NRC Special Inspection
made no recommendations with respect to the Report dated December 22, 2003, ADAMS Accession No.
design or operation of Canadian nuclear plants to ML033570386.
improve the reliability of the Ontario electricity 2 Further details are available in the NRC Special Inspection
grid. Report dated December 22, 2003, ADAMS Accession No.
ML033570386.
The CNWG made two recommendations, one con- 3 Further details are available in the NRC Special Inspection
cerning backup electrical generation equipment Report dated October 10, 2003, ADAMS Acccession No.
to the CNSC’s Emergency Operations Centre and ML032880107.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 129
9. Physical and Cyber Security Aspects of the Blackout
conducted after the release of the Interim Report
Summary and Primary Findings allowed the SWG to validate its preliminary find-
After the Task Force Interim Report was issued in ings and the SWG to make recommendations on
November 2003, the Security Working Group those findings:
(SWG) continued in its efforts to investigate u Interviews and analyses conducted by the SWG
whether a malicious cyber event directly caused indicate that within some of the companies
or significantly contributed to the power outage of interviewed there are potential opportunities
August 14, 2003. These efforts included addi- for cyber system compromise of Energy Man-
tional analyses of interviews conducted prior to agement Systems (EMS) and their supporting
the release of the Interim Report and additional information technology (IT) infrastructure.
consultations with representatives from the elec- Indications of procedural and technical IT man-
tric power sector. The information gathered from agement vulnerabilities were observed in some
these efforts validated the SWG’s Interim Report facilities, such as unnecessary software services
preliminary findings and the SWG found no rea- not denied by default, loosely controlled system
son to amend, alter, or negate any of the informa- access and perimeter control, poor patch and
tion submitted to the Task Force for the Interim configuration management, and poor system
Report. security documentation. This situation caused
Specifically, further analysis by the SWG found the SWG to support the promulgation, imple-
no evidence that malicious actors caused or con- mentation, and enforce-
tributed to the power outage, nor is there evidence ment of cyber and physi- Recommendation
32, page 163
that worms or viruses circulating on the Internet at cal security standards for
the time of the power outage had an effect on the electric power sector.
power generation and delivery systems of the u A failure in a software program not linked to
companies directly involved in the power outage. malicious activity may have significantly con-
The SWG acknowledges reports of al-Qaeda tributed to the power outage. Since the issuance
claims of responsibility for the power outage of of the Interim Report, the SWG consulted with
August 14, 2003. However, these claims are not the software program’s vendor and confirmed
consistent with the SWG’s findings. SWG analysis that since the August 14, 2003, power outage,
also brought to light certain concerns respecting the vendor provided industry with the neces-
the possible failure of alarm software; links to con- sary information and mitigation steps to
trol and data acquisition software; and the lack of address this software failure. In Canada, a sur-
a system or process for some grid operators to ade- vey was posted on the Canadian Electricity
quately view the status of electric systems outside Association (CEA) secure members-only web
of their immediate control. site to determine if the
software was in use. The
After the release of the Interim Report in Novem- Recommendation
responses indicated that 33, page 164
ber 2003, the SWG determined that the existing
it is not used by Canadian
data, and the findings derived from analysis of
companies in the industry.
those data, provided sufficient certainty to
exclude the probability that a malicious cyber u Internal and external links from Supervisory
event directly caused or significantly contributed Control and Data Acquisi-
to the power outage events. As such, further data tion (SCADA) networks to Recommendation
34, page 165
collection efforts to conduct broader analysis were other systems introduced
deemed unnecessary. While no additional data vulnerabilities.
were collected, further analysis and interviews
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 131
u In some cases, Control of its work, the SWG defined a “malicious cyber
Area (CA) and Reliability event” as the manipulation of data, software or
Recommendation
Coordinator (RC) visibil- 35, page 165 hardware for the purpose of deliberately disrupt-
ity into the operations of ing the systems that control and support the gener-
surrounding areas was ation and delivery of electric power.
lacking.
The SWG worked closely with the United States
The SWG’s analysis is reflected in a total of 15 rec- and Canadian law enforcement, intelligence and
ommendations, two of which were combined with homeland security communities to examine the
similar concerns by the ESWG (Recommendations possible role of malicious actors in the power out-
19 and 22); for the remaining 13, see Recommen- age. A primary activity in this endeavor was the
dations 32-44 (pages 163-169). collection and review of available intelligence
related to the power outage of August 14, 2003.
Overall, the SWG’s final report was the result of
The SWG also collaborated with the energy indus-
interviews conducted with representatives of
try to examine the cyber systems that control
Cinergy, FirstEnergy, American Electric Power
power generation and delivery operations, the
(AEP), PJM Interconnect, the Midwest Independ-
physical security of cyber assets, cyber policies
ent System Operator (MISO), the East Central Area
and procedures and the functionality of support-
Reliability Coordinating Agreement (ECAR), and
ing infrastructures—such as communication sys-
GE Power Systems Division. These entities were
tems and backup power generation, which
chosen due to their proximity to the causes of the
facilitate the smooth running operation of cyber
power outage based on the analysis of the Electric
assets—to determine if the operation of these sys-
System Working Group (ESWG). The findings
tems was affected by malicious activity. The SWG
contained in this report relate only to those enti-
coordinated its efforts with those of other Working
ties surveyed. The final report also incorporates
Groups and there was a significant interdepen-
information gathered from third party sources
dence on each groups work products and findings.
as well as federal security and intelligence
The SWG’s focus was on the cyber operations of
communities.
those companies in the United States involved in
In summary, SWG analysis provided no evidence the early stages of the power outage timeline, as
that a malicious cyber attack was a direct or indi- identified by the ESWG.
rect cause of the August 14, 2003, power outage.
Outside of the SWG’s scope was the examination
This conclusion is supported by the SWG’s event
of the non-cyber physical infrastructure aspects of
timeline, detailed later in this chapter, which
the power outage of August 14, 2003. The Interim
explains in detail the series of non-malicious
Report detailed the SWG’s availability to investi-
human and cyber failures that ultimately resulted
gate breaches of physical security unrelated to the
in the power outage. In the course of its analysis
cyber dimensions of the infrastructure on behalf
the SWG, however, did identify a number of areas
of the Task Force but no incidents came to the
of concern respecting cyber security aspects of the
SWG’s attention during its work. Also outside of
electricity sector.
the scope of the SWG’s work was analysis of the
impacts the power outage had on other critical
SWG Mandate and Scope infrastructure sectors. Both Public Safety and
Emergency Preparedness Canada and the U.S.
It is widely recognized that the increased reliance Department of Homeland Security (DHS) exam-
on IT by critical infrastructure sectors, including ined these issues, but not within the context of the
the energy sector, has increased the vulnerability SWG.
of these systems to disruption via cyber means.
The ability to exploit these vulnerabilities has Cyber Security in the
been demonstrated in North America. The SWG Electricity Sector
was comprised of United States and Canadian fed-
eral, state, provincial and local experts in both The generation and delivery of electricity has
physical and cyber security and its objective was been, and continues to be, a target of malicious
to determine the role, if any, that a malicious cyber groups and individuals intent on disrupting this
event played in causing, or contributing to, the system. Even attacks that do not directly target the
power outage of August 14, 2003. For the purposes electricity sector can have disruptive effects on
132 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
electricity system operations. Many malicious systems. The Canadian Electricity Association’s
code attacks, by their very nature, are unbiased Critical Infrastructure Working Group is examin-
and tend to interfere with operations supported by ing similar measures.
vulnerable applications. One such incident
occurred in January 2003, when the “Slammer”
Internet worm took down monitoring computers Information Collection
at FirstEnergy Corporation’s idled Davis-Besse and Analysis
nuclear plant. A subsequent report by the North
American Electric Reliability Council (NERC) After analyzing information already obtained
concluded that although the infection caused no from stakeholder interviews, telephone tran-
outages, it blocked commands that operated other scripts, law enforcement and intelligence informa-
power utilities.1 tion, and other ESWG working documents, the
SWG determined that it was not necessary to ana-
This example, among others, highlights the lyze other sources of data on the cyber operations
increased vulnerability to disruption via cyber of those such as log data from routers, intrusion
means faced by North America’s critical infra- detection systems, firewalls, EMS, change man-
structure sectors, including the energy sector. Of agement logs, and physical security materials.
specific concern to the United States and Cana-
dian governments are the SCADA networks, The SWG was divided into six sub-teams to
which contain computers and applications that address the discrete components of this investiga-
perform a wide variety of functions across many tion: Cyber Analysis, Intelligence Analysis, Physi-
industries. In electric power, SCADA includes cal Analysis, Policies and Procedures, Supporting
telemetry for status and control, as well as EMS, Infrastructure, and Root Cause Liaison. The SWG
protective relaying and automatic generation con- organized itself in this manner to create a holistic
trol. SCADA systems were developed to maximize approach to address each of the main areas of con-
functionality and interoperability, with little cern with regards to power grid vulnerabilities.
attention given to cyber security. These systems, Rather than analyze each area of concern sepa-
many of which were intended to be isolated, now rately, the SWG sub-team structure provided a
find themselves for a variety of business and oper- more comprehensive framework in which to
ational reasons, either directly or indirectly con- investigate whether malicious activity was a cause
nected to the global Internet. For example, in some of the power outage of August 14, 2003. Each
instances, there may be a need for employees to sub-team was staffed with Subject Matter Experts
monitor SCADA systems remotely. However, (SMEs) from government, industry, and academia
connecting SCADA systems to a remotely accessi- to provide the analytical breadth and depth neces-
ble computer network can present security risks. sary to complete each sub-team’s objective. A
These risks include the compromise of sensitive detailed overview of the sub-team structure and
operating information and the threat of un- activities for each sub-team is provided below.
authorized access to SCADA systems’ control
mechanisms. 1. Cyber Analysis
Security has always been a priority for the electric- The Cyber Analysis sub-team was led by the
ity sector in North America; however, it is a CERT® Coordination Center (CERT/CC) at Carne-
greater priority now than ever before. CAs and RCs gie Mellon University and the Royal Canadian
recognize that the threat environment is changing Mounted Police (RCMP). This team was focused
and that the risks are greater than in the past, and on analyzing and reviewing electronic media of
they have taken steps towards improving their computer networks in which online communica-
security postures. NERC’s Critical Infrastructure tions take place. The sub-team examined these
Protection Advisory Group has been examining networks to determine if they were maliciously
ways to improve both the physical and cyber secu- used to cause, or contribute to the August 14,
rity dimensions of the North American power 2003, outage. Specifically, the SWG reviewed
grid. This group is comprised of Canadian and materials created on behalf of DHS’s National
U.S. industry experts in the areas of cyber secu- Communication System (NCS). These materials
rity, physical security and operational security. covered the analysis and conclusions of their
The creation of a national SCADA program is now Internet Protocol (IP) modeling correlation study
also under discussion in the U.S. to improve the of Blaster (a malicious Internet worm first noticed
physical and cyber security of these control on August 11, 2003) and the power outage. This
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 133
NCS analysis supports the SWG’s finding that general nature relating to the sector which was
viruses and worms prevalent across the Internet provided to the North American energy industry
at the time of the outage did not have any signifi- by U.S. and Canadian Government agencies in late
cant impact on power generation and delivery July 2003. This information indicated that
systems. The team also conducted interviews with al-Qaeda might attempt to carry out a physical
vendors to identify known system flaws and attack involving explosions at oil production facil-
vulnerabilities. ities, power plants or nuclear plants on the east
coast of the U.S. during the summer of 2003. The
This sub-team took a number of steps, including
type of physical attack described in the intelli-
reviewing NERC reliability standards to gain a
gence that prompted this threat warning is not
better understanding of the overall security pos-
consistent with the events causing the power out-
ture of the electric power industry. Additionally,
age as there was no indication of a kinetic event
the sub-team participated in meetings in Balti-
before, during, or immediately after the power
more on August 22 and 23, 2003. The meetings
outage of August 14, 2003.
provided an opportunity for the cyber experts and
the power industry experts to understand the Despite all of the above indications that no terror-
details necessary to conduct an investigation. ist activity caused the power outage, al-Qaeda
publicly claimed responsibility for its occurrence:
Members of the sub-team also participated in the
NERC/Department of Energy (DOE) Fact Finding u August 18, 2003: Al-Hayat, an Egyptian media
meeting held in Newark, New Jersey on Septem- outlet, published excerpts from a communiqué
ber 8, 2003. Each company involved in the outage attributed to al-Qaeda. Al Hayat claimed to have
provided answers to a set of questions related to obtained the communiqué from the website of
the outage. The meeting helped to provide a better the International Islamic Media Center. The
understanding of what each company experi- content of the communiqué asserts that the “bri-
enced before, during and after the outage. Addi- gades of Abu Fahes Al Masri had hit two main
tionally, sub-team members participated in power plants supplying the East of the U.S., as
interviews with grid operators from FirstEnergy well as major industrial cities in the U.S. and
on October 8 and 9, 2003, and from Cinergy on Canada, . . . its ally in the war against Islam
October 10, 2003. (New York and Toronto) and their neighbors.”
Furthermore, the operation “was carried out on
2. Intelligence Analysis the orders of Osama bin Laden to hit the pillars
of the U.S. economy,” as “a realization of bin
The Intelligence Analysis sub-team was led by
Laden’s promise to offer the Iraqi people a pres-
DHS and the RCMP, which worked closely with
ent.” The communiqué does not specify the way
Federal, State and local law enforcement, intelli-
the alleged sabotage was carried out, but does
gence and homeland security organizations to
elaborate on the alleged damage the sabotage
assess whether the power outage was the result of
caused to the U.S. economy in the areas of
a malicious attack.
finance, transportation, energy and telecommu-
SWG analysis provided no evidence that mali- nications.
cious actors—be they individuals or organiza-
Additional claims and commentary regarding the
tions—were responsible for, or contributed to, the
power outage appeared in various Middle Eastern
power outage of August 14, 2003. Additionally,
media outlets:
the sub-team found no indication of deliberate
physical damage to power generating stations and u August 26, 2003: A conservative Iranian daily
delivery lines on the day of the outage and there newspaper published a commentary regarding
were no reports indicating the power outage was the potential of computer technology as a tool
caused by a computer network attack. for terrorists against infrastructures dependent
on computer networks, most notably water,
Both U.S. and Canadian government authorities
electric, public transportation, trade organiza-
provide threat intelligence information to their
tions and “supranational” companies in the
respective energy sectors when appropriate. No
United States.
intelligence reports prior to, during or after the
power outage indicated any specific terrorist plans u September 4, 2003: An Islamist participant in a
or operations against the energy infrastructure. Jihadist chat room forum claimed that sleeper
There was, however, threat information of a cells associated with al-Qaeda used the power
134 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
outage as a cover to infiltrate the U.S. from the sub-team sought to determine if the physical
Canada. integrity of these cyber facilities was breached,
whether externally or by an insider, prior to or
However, these claims as known are not consis-
during the outage, and if so, whether such a
tent with the SWG’s findings. They are also not
breach caused or contributed to the power outage.
consistent with congressional testimony of the
Federal Bureau of Investigation (FBI). Larry A. Although the sub-team analyzed information pro-
Mefford, Executive Assistant Director in charge of vided to both the ESWG and Nuclear Working
the FBI’s Counterterrorism and Counterintelli- Groups, the Physical Analysis sub-team also
gence programs, testified in U.S. Congress on Sep- reviewed information resulting from face-to-face
tember 4, 2003, that: meetings with energy sector personnel and
“To date, we have not discovered any evidence site-visits to energy sector facilities to determine
indicating that the outage was a result of activity the physical integrity of the cyber infrastructure.
by international or domestic terrorists or other The sub-team compiled a list of questions cover-
criminal activity.”2 ing location, accessibility, cameras, alarms, locks,
Mr. Mefford also testified that: fire protection and water systems as they apply to
computer server rooms. Based on discussions of
“The FBI has received no specific, credible these questions during its interviews, the
threats to electronic power grids in the United sub-team found no evidence that the physical
States in the recent past and the claim of the Abu integrity of the cyber infrastructure was breached.
Hafs al-Masri Brigade to have caused the black- Additionally, the sub-team examined access and
out appears to be no more than wishful thinking. control measures used to allow entry into com-
We have no information confirming the actual mand and control facilities and the integrity of
existence of this group.”3 remote facilities.
Current assessments suggest that there are terror-
The sub-team also concentrated on mechanisms
ists and other malicious actors who have the capa-
used by the companies to report unusual incidents
bility to conduct a malicious cyber attack with
within server rooms, command and control rooms
potential to disrupt the energy infrastructure.
and remote facilities. The sub-team also addressed
Although such an attack cannot be ruled out
the possibility of an insider attack on the cyber
entirely, an examination of available information
infrastructure.
and intelligence does not support any claims of a
deliberate attack against the energy infrastructure 4. Policies and Procedures
on, or leading up to, August 14, 2003. The few
instances of physical damage that occurred on The Policies and Procedures sub-team was led by
power delivery lines were the result of natural DHS and Public Safety and Emergency Prepared-
events and not of sabotage. No intelligence reports ness Canada. Personnel from these organizations
prior to, during or after the power outage indicated have strong backgrounds in the fields of electric
any specific terrorist plans or operations against delivery operations, automated control systems
the energy infrastructure. No incident reports including SCADA and EMS, and information
detail suspicious activity near the power genera- security.
tion plants or delivery lines in question.
This sub-team was focused on examining the
3. Physical Analysis overall policies and procedures that may or may
not have been in place during the events leading
The Physical Analysis sub-team was led by the up to and during the power outage of August 14,
United States Secret Service and the RCMP. These 2003. Policies that the team examined revolved
organizations have a particular expertise in physi- centrally around the cyber systems of the compa-
cal security assessments in the energy sector. The nies identified in the early stages of the power out-
sub-team focused on issues related to how the age. Of specific interest to the team were policies
cyber-related facilities of the energy sector compa- and procedures regarding the upgrade and mainte-
nies were secured, including the physical integrity nance (to include system patching) of the com-
of data centers and control rooms along with secu- mand and control (C2) systems, including SCADA
rity procedures and policies used to limit access to and EMS. The Policies and Procedures sub-team
sensitive areas. Focusing on the facilities identi- was also interested in the procedures for contin-
fied as having a causal relationship to the outage, gency operations and restoration of systems in the
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 135
event of a computer system failure, or a cyber between the power outage and malicious activity,
event such as an active hack or the discovery of whether physical or cyber initiated.
malicious code.
136 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
At 14:54 EDT, FirstEnergy’s backup server failed. thought the Stuart-Atlanta line was out of service.
AEPR continued to malfunction. The Area Control Grid operators referred to their “Outage Sched-
Error Calculations (ACE) and Strip Charting rou- uler” and informed the EMS Engineer that their
tines malfunctioned and the dispatcher user inter- data showed the Stuart-Atlanta line was “up” and
face slowed significantly. that the EMS engineer should depict the line as in
service in the SE model. At 15:17 EDT, the EMS
At 15:00 EDT, FirstEnergy used its emergency engineer ran the SE with the Stuart-Atlanta line
backup system to control the system and make “live,” but the model again mismatched.
ACE calculations. ACE calculations and control
systems continued to run on the emergency At 15:29 EDT, the MISO EMS Engineer asked
backup system until roughly 15:08 EDT, when the MISO grid operators to call PJM Interconnect, LLC
primary server was restored. to determine the status of the Stuart-Atlanta line.
MISO was informed that the Stuart-Atlanta line
At 15:05 EDT, FirstEnergy’s Harding-Chamberlin tripped at 14:02 EDT. The EMS Engineer adjusted
345-kV line tripped and locked out. FirstEnergy the model, which by this time had been updated
grid operators did not receive notification from the with the 15:05 EDT Harding-Chamberlin 345-kV
AEPR software which continued to malfunction, line trip, and came up with a valid solution.
unbeknownst to the FirstEnergy grid operators.
At 15:32 EDT, FirstEnergy’s Hanna-Juniper
At 15:08 EDT, using data obtained at roughly 345-kV line tripped and locked out. The AEPR
15:04 EDT (it takes roughly five minutes for the SE continued to malfunction.
to provide a result), the MISO EMS engineer con-
At 15:41 EDT, the lights flickered at the
cluded that the SE mismatched due to a line out-
FirstEnergy’s control facility. This occurred
age. His experience allowed him to isolate the
because they had lost grid power and switched
outage to the Stuart-Atlanta 345-kV line (which
over to their emergency power supply.
tripped about an hour earlier at 14:02 EDT). He
took the Stuart-Atlanta line out of service in the SE At 15:42 EDT, a FirstEnergy dispatcher realized
model and got a valid solution. that the AEPR was not working and made techni-
cal support staff aware of the problem.
Also at 15:08 EDT, the FirstEnergy primary server
was restored. ACE calculations and control sys- Endnotes
tems were now running on the primary server. 1 http://www.nrc.gov/reading-rm/doc-collections/news/
AEPR continued to malfunction, unbeknownst to 2003/ 03-108.html.
2 http://www.fbi.gov/congress/congress03/mefford090403.
the FirstEnergy grid operators.
htm.
At 15:09 EDT, the MISO EMS engineer went to 3 http://www.fbi.gov/congress/congress03/mefford090403.
the control room to tell the grid operators that he htm.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 137
10. Recommendations to Prevent or Minimize
the Scope of Future Blackouts
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 139
grids at higher loadings means greater stress on Implementation of the recommendations pre-
equipment and a smaller range of options and a sented below will address all remaining issues,
shorter period of time for dealing with unexpected through the ongoing work of government bodies
problems. The system operator’s job has become and agencies in the U.S. and Canada, the electric-
more challenging, leading to the need for more ity industry, and the non-governmental institu-
sophisticated grid management tools and more tions responsible for the maintenance of electric
demanding operator training programs and certifi- reliability in North America.
cation requirements.
140 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Overview of Task Force Recommendations: Titles Only
Group I. Institutional Issues Related to Reliability
1. Make reliability standards mandatory and enforceable, with penalties for noncompliance.
2. Develop a regulator-approved funding mechanism for NERC and the regional reliability councils,
to ensure their independence from the parties they oversee.
3. Strengthen the institutional framework for reliability management in North America.
4. Clarify that prudent expenditures and investments for bulk system reliability (including invest-
ments in new technologies) will be recoverable through transmission rates.
5. Track implementation of recommended actions to improve reliability.
6. FERC should not approve the operation of new RTOs or ISOs until they have met minimum
functional requirements.
7. Require any entity operating as part of the bulk power system to be a member of a regional reli-
ability council if it operates within the council’s footprint.
8. Shield operators who initiate load shedding pursuant to approved guidelines from liability or
retaliation.
9. Integrate a “reliability impact” consideration into the regulatory decision-making process.
10. Establish an independent source of reliability performance information.
11. Establish requirements for collection and reporting of data needed for post-blackout analyses.
12. Commission an independent study of the relationships among industry restructuring, competi-
tion, and reliability.
13. DOE should expand its research programs on reliability-related tools and technologies.
14. Establish a standing framework for the conduct of future blackout and disturbance
investigations.
Group II. Support and Strengthen NERC’s Actions of February 10, 2004
15. Correct the direct causes of the August 14, 2003 blackout.
16. Establish enforceable standards for maintenance of electrical clearances in right-of-way areas.
17. Strengthen the NERC Compliance Enforcement Program.
18. Support and strengthen NERC’s Reliability Readiness Audit Program.
19. Improve near-term and long-term training and certification requirements for operators, reliability
coordinators, and operator support staff.
20. Establish clear definitions for normal, alert and emergency operational system conditions. Clarify
roles, responsibilities, and authorities of reliability coordinators and control areas under each
condition.
21. Make more effective and wider use of system protection measures.
22. Evaluate and adopt better real-time tools for operators and reliability coordinators.
23. Strengthen reactive power and voltage control practices in all NERC regions.
24. Improve quality of system modeling data and data exchange practices.
25. NERC should reevaluate its existing reliability standards development process and accelerate the
adoption of enforceable standards.
26. Tighten communications protocols, especially for communications during alerts and emergen-
cies. Upgrade communication system hardware where appropriate.
27. Develop enforceable standards for transmission line ratings.
28. Require use of time-synchronized data recorders.
29. Evaluate and disseminate lessons learned during system restoration.
30. Clarify criteria for identification of operationally critical facilities, and improve dissemination of
updated information on unplanned outages.
31. Clarify that the transmission loading relief (TLR) process should not be used in situations involv-
ing an actual violation of an Operating Security Limit. Streamline the TLR process.
(continued on page 142)
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 141
Overview of Task Force Recommendations: Titles Only (Continued)
Group III. Physical and Cyber Security of North American Bulk Power Systems
32. Implement NERC IT standards.
33. Develop and deploy IT management procedures.
34. Develop corporate-level IT security governance and strategies.
35. Implement controls to manage system health, network monitoring, and incident management.
36. Initiate U.S.-Canada risk management study.
37. Improve IT forensic and diagnostic capabilities.
38. Assess IT risk and vulnerability at scheduled intervals.
39. Develop capability to detect wireless and remote wireline intrusion and surveillance.
40. Control access to operationally sensitive equipment.
41. NERC should provide guidance on employee background checks.
42. Confirm NERC ES-ISAC as the central point for sharing security information and analysis.
43. Establish clear authority for physical and cyber security.
44. Develop procedures to prevent or mitigate inappropriate disclosure of information.
Group IV. Canadian Nuclear Power Sector
45. The Task Force recommends that the Canadian Nuclear Safety Commission request Ontario
Power Generation and Bruce Power to review operating procedures and operator training associ-
ated with the use of adjuster rods.
46. The Task Force recommends that the Canadian Nuclear Safety Commission purchase and install
backup generation equipment.
u Reliability standards should allow, where applied more broadly. For example, in the
appropriate, flexibility to accommodate Western U.S. and Canada, many members of the
regional differences, including more stringent Western Electricity Coordinating Council (WECC)
reliability requirements in some areas, but include clauses in contracts for the purchase of
regional deviations should not be allowed to wholesale power that require the parties to com-
lead to lower reliability expectations or ply with reliability standards. In the areas of the
performance. U.S. and Canada covered by the Northeast Power
Coordinating Council (NPCC), parties found not to
u An ERO-proposed standard or modification to a
be in compliance with NERC and NPCC reliability
standard should take effect within the United requirements are subject to escalating degrees of
States upon approval by the Federal Energy scrutiny by their peers and the public. Both of
Regulatory Commission (FERC). these approaches have had positive effects. FERC
u FERC should remand to the ERO for further should examine other approaches as well, and
consideration a proposed reliability standard or work with state regulatory authorities to ensure
a modification to a reliability standard that it
disapproves of in whole or in part, with expla-
nation for its concerns and rationale.
B. Action by FERC NERC and the ERO
In the absence of such reliability legislation, FERC If the proposed U.S. reliability legislation
should review its statutory authorities under passes, the North American Electric Reliability
existing law, and to the maximum extent permit- Council (NERC) may undertake various organi-
ted by those authorities, act to enhance reliability zational changes and seek recognition as the
by making compliance with reliability standards electric reliability organization (ERO) called for
enforceable in the United States. In doing so, in H.R. 6 and S. 2095. For simplicity of presen-
FERC should consult with state regulators, NERC, tation, the many forward-looking references
and the regional reliability councils to determine below to “NERC” are intended to apply to the
whether certain enforcement practices now in use ERO if the legislation is passed, and to NERC if
in some parts of the U.S. and Canada might be the legislation is not passed.
142 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
that any other appropriate actions to make reli- based on a surcharge in transmission rates. The
ability standards enforceable are taken. purpose would be to ensure that NERC and the
councils are appropriately funded to meet their
Action by FERC under its existing authorities changing responsibilities without dependence on
would not lessen the need for enactment of reli- the parties that they oversee. Note: Implementation
ability legislation by the Congress. Many U.S. par- of this recommendation should be coordinated
ties that should be required by law to comply with with the review called for in Recommendation 3
reliability requirements are not subject to the concerning the future role of the regional councils.
Commission’s full authorities under the Federal
Power Act. NERC’s current $13 million/year budget is funded
C. Action by Appropriate Authorities in Canada as part of the dues that transmission owners, gen-
erators, and other market participants pay to the
The interconnected nature of the transmission
ten regional reliability councils, which then fund
grid requires that reliability standards be identical
NERC. This arrangement makes NERC subject to
or compatible on both sides of the Canadian/U.S.
the influence of the reliability councils, which are
border. Several provincial governments in Canada
in turn subject to the influence of their control
have already demonstrated support for mandatory
areas and other members. It also compromises the
and enforceable reliability standards and have
independence of both NERC and the councils in
either passed legislation or have taken steps to put relation to the entities whose actions they oversee,
in place the necessary framework for implement- and makes it difficult for them to act forcefully
ing such standards in Canada. The federal and and objectively to maintain the reliability of the
provincial governments should work together and North American bulk power system. Funding
with appropriate U.S. authorities to complete a NERC and the councils through a transmission
framework to ensure that identical or compatible rate surcharge administered and disbursed under
standards apply in both countries, and that means regulatory supervision would enable the organiza-
are in place to enforce them in all interconnected tions to be more independent of the industry, with
jurisdictions. little impact on electric bills. The dues that com-
D. Joint Actions by U.S. and Canadian panies pay to the regional councils are passed
Governments through to electricity customers today, so the net
International coordination mechanisms should be impacts on customer bills from shifting to a rate
developed between the governments in Canada surcharge would be minimal.
and the United States to provide for government Implementation of the recommendations pre-
oversight of NERC or the ERO, and approval and sented in this report will involve a substantial
enforcement of reliability standards. increase in NERC’s functions and responsibilities,
E. Memoranda of Understanding between U.S. and require an increase in NERC’s annual budget.
or Canadian Government Agencies and The additional costs, however, would be small in
NERC comparison to the cost of a single major blackout.
Government agencies in both countries should
decide (individually) whether to develop a memo- 3. Strengthen the institutional framework
randum of understanding (MOU) with NERC that for reliability management in North
would define the agency’s working relationship America.7
with NERC, government oversight of NERC activi-
ties if appropriate, and the reliability responsibili- FERC, DOE and appropriate authorities in Canada
ties of the signatories. should work with the states, NERC, and the indus-
try, to evaluate and develop appropriate modifica-
tions to the existing institutional framework for
2. Develop a regulator-approved mecha- reliability management. In particular, the affected
nism for funding NERC and the regional government agencies should:
reliability councils, to ensure their inde- A. Commission an independent review by quali-
pendence from the parties they oversee.6 fied experts in organizational design and man-
U.S. and Canadian regulatory authorities should agement to address issues concerning how best
work with NERC, the regional councils, and the to structure an international reliability organi-
industry to develop and implement a new funding zation for the long term.
mechanism for NERC and the regional councils
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 143
B. Based in part on the results of that review, sizes and resources. Some have grown from a reli-
develop metrics for gauging the adequacy of ability council into an ISO or RTO (ERCOT and
NERC’s performance, and specify the functions SPP), some span less than a single state (FRCC and
of the NERC Board of Trustees and the proce- ERCOT) while others cover many states and prov-
dure for selecting the members of the Board.
inces and cross national boundaries (NPCC and
C. Examine and clarify the future role of the WECC). Several cross reliability coordinator
regional reliability councils, with particular boundaries. It is time to evaluate the appropriate
attention to their mandate, scope, structure, size and scope of a regional council, the specific
responsibilities, and resource requirements. tasks that it should perform, and the appropriate
level of resources, expertise, and independence
D. Examine NERC’s proposed Functional Model that a regional reliability council needs to perform
and set minimum requirements under which
those tasks effectively. This evaluation should
NERC would certify applicants’ qualifications
to perform critical functions. also address whether the councils as currently
constituted are appropriate to meet future reliabil-
E. Request NERC and the regional councils to sus- ity needs.
pend designation of any new control areas (or
sub-control areas) until the minimum require- D. NERC’s Functional Model
ments in section D (above) have been estab- The transition to competition in wholesale power
lished, unless an applicant shows that such markets has been accompanied by increasing
designation would significantly enhance reli- diversity in the kinds of entities that need to be in
ability. compliance with reliability standards. Rather than
F. Determine ways to enhance reliability opera- resist or attempt to influence this evolution,
tions in the United States through simplified NERC’s response—through the Functional
organizational boundaries and resolution of Model—has been to seek a means of enabling reli-
seams issues. ability to be maintained under virtually any insti-
tutional framework. The Functional Model
A and B. Reshaping NERC identifies sixteen basic functions associated with
operating the bulk electric systems and maintain-
The far-reaching organizational changes in the ing reliability, and the capabilities that an organi-
North American electricity industry over the past zation must have in order to perform a given
decade have already induced major changes in the function. (See Functional Model text box below.)
nature of NERC as an organization. However, the
process of change at NERC is far from complete. NERC acknowledges that maintaining reliability
Important additional changes are needed such as in some frameworks may be more difficult or more
the shift to enforceable standards, development of expensive than in others, but it stresses that as
an effective monitoring capability, and funding long as some responsible party addresses each
that is not dependent on the industry. These function and the rules are followed, reliability will
changes will strengthen NERC as an organization. be preserved. By implication, the pros and cons of
In turn, to properly serve overarching public pol- alternative institutional frameworks in a given
icy concerns, this strengthening of NERC’s capa- region—which may affect aspects of electric
bilities will have to be balanced with increased industry operations other than reliability—are
government oversight, more specific metrics for matters for government agencies to address, not
gauging NERC’s performance as an organization, NERC.
and greater transparency concerning the functions
of its senior management team (including its One of the major purposes of the Functional
Board of Trustees) and the procedures by which Model is to create a vehicle through which NERC
those individuals are selected. The affected gov- will be able to identify an entity responsible for
ernment agencies should jointly commission an performing each function in every part of the three
independent review of these and related issues to North American interconnections. NERC consid-
aid them in making their respective decisions. ers four of the sixteen functions to be especially
critical for reliability. For these functions, NERC
intends, upon application by an entity, to review
C. The Role of the Regional Reliability Councils the entity’s capabilities, and if appropriate, certify
North America’s regional reliability councils have that the entity has the qualifications to perform
evolved into a disparate group of organizations that function within the specified geographic area.
with varying responsibilities, expertise, roles, For the other twelve functions, NERC proposes to
144 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
“register” entities as responsible for a given func- critical functions for their respective areas. In
tion in a given area, upon application. other areas, capabilities may remain less aggre-
gated, and the institutional structure may remain
All sixteen functions are presently being per- more complex.
formed to varying degrees by one entity or another
today in all areas of North America. Frequently an Working with NERC and the industry, FERC and
entity performs a combination of functions, but authorities in Canada should review the Func-
there is great variety from one region to another in tional Model to ensure that operating hierarchies
how the functions are bundled and carried out. and entities will facilitate, rather than hinder,
Whether all of the parties who are presently per- efficient reliability operations. At a minimum,
forming the four critical functions would meet the review should identify ways to eliminate inap-
NERC’s requirements for certification is not propriate commercial incentives to retain control
known, but the proposed process provides a area status that do not support reliability objec-
means of identifying any weaknesses that need to tives; address operational problems associated
be rectified. with institutional fragmentation; and set mini-
mum requirements with respect to the capabilities
At present, after protracted debate, the Functional requiring NERC certification, concerning subjects
Model appears to have gained widespread but cau- such as:
tious support from the diverse factions across the
industry, while the regulators have not taken a 1. Fully operational backup control rooms.
position. In some parts of North America, such as 2. System-wide (or wider) electronic map boards
the Northeast, large regional organizations will or functional equivalents, with data feeds that
probably be certified to perform all four of the are independent of the area’s main energy man-
agement system (EMS).
Sixteen Functions in NERC’s Functional
Model 3. Real-time tools that are to be available to the
u Operating Reliability operator, with backups. (See Recommendation
22 below for more detail concerning minimum
u Planning Reliability requirements and guidelines for real-time oper-
u Balancing (generation and demand) ating tools.)
u Interchange
4. SCADA and EMS requirements, including
u Transmission service backup capabilities.
u Transmission ownership
5. Training programs for all personnel who have
u Transmission operations access to a control room or supervisory respon-
u Transmission planning sibilities for control room operations. (See Rec-
u Resource planning ommendation 19 for more detail on the Task
u Distribution
Force’s views regarding training and certifica-
tion requirements.)
u Generator ownership
u Generator operations 6. Certification requirements for control room
managers and staff.
u Load serving
u Purchasing and selling E. Designation of New Control Areas
u Standards development Significant changes in the minimum functional
requirements for control areas (or balancing
u Compliance monitoring
authorities, in the context of the Functional
NERC regards the four functions shown above Model) may result from the review called for
in bold as especially critical to reliability. above. Accordingly, the Task Force recommends
Accordingly, it proposes to certify applicants that regulatory authorities should request NERC
that can demonstrate that they have the capabil- and the regional councils not to certify any new
ities required to perform those functions. The control areas (or sub-control areas) until the
Operating Reliability authority would corre- appropriate regulatory bodies have approved the
spond to today’s reliability coordinator, and the minimum functional requirements for such bod-
Balancing authority to today’s control area ies, unless an applicant shows that such designa-
operator. tion would significantly enhance reliability.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 145
F. Boundary and Seam Issues and Minimum FERC and appropriate authorities in Canada
Functional Requirements should clarify that prudent expenditures and
investments by regulated companies to maintain or
Some observers believe that some U.S. regions improve bulk system reliability will be recoverable
have too many control areas performing one or through transmission rates.
more of the four critical reliability functions. In the U.S., FERC and DOE should work with state
In many cases, these entities exist to retain com- regulators to identify and resolve issues related to
mercial advantages associated with some of these the recovery of reliability costs and investments
functions. The resulting institutional fragmenta- through retail rates. Appropriate authorities in
tion and decentralization of control leads to a Canada should determine whether similar efforts
higher number of operating contacts and seams, are warranted.
complex coordination requirements, misalign-
ment of control areas with other electrical bound- Companies will not make the expenditures and
aries and/or operating hierarchies, inconsistent investments required to maintain or improve the
practices and tools, and increased compliance reliability of the bulk power system without credi-
monitoring requirements. These consequences ble assurances that they will be able to recover
hamper the efficiency and reliability of grid their costs.
operations.
5. Track implementation of recommended
As shown above (text box on page 14), MISO, as actions to improve reliability.9
reliability coordinator for its region, is responsible In the requirements issued on February 10, 2004,
for dealing with 37 control areas, whereas PJM NERC announced that it and the regional councils
now spans 9 control areas, ISO-New England has would establish a program for documenting com-
2, and the New York ISO, Ontario’s IMO, Texas’ pletion of recommendations resulting from the
ERCOT, and Québec’s Trans-Energie are them- August 14 blackout and other historical outages, as
selves the control area operators for their respec- well as NERC and regional reports on violations of
tive large areas. Moreover, it is not clear that small reliability standards, results of compliance audits,
control areas are financially able to provide the and lessons learned from system disturbances. The
regions are to report on a quarterly basis to NERC.
facilities and services needed to perform control
area functions at the level needed to maintain reli- In addition, NERC intends to initiate by January 1,
ability. This concern applies also to the four types 2005 a reliability performance monitoring function
of entities that NERC proposes to certify under the that will evaluate and report on trends in bulk
Functional Model (i.e., Reliability Authority, electric system reliability performance.
Planning Authority, Balancing Authority, and The Task Force supports these actions strongly.
Interchange Authority). However, many of the Task Force’s recommenda-
tions pertain to government bodies as well as
For the long term, the regulatory agencies should NERC. Accordingly:
continue to seek ways to ensure that the regional A. Relevant agencies in the U.S. and Canada
operational frameworks that emerge through the should cooperate to establish mechanisms for
implementation of the Functional Model promote tracking and reporting to the public on imple-
reliable operations. Any operational framework mentation actions in their respective areas of
will represent some combination of tradeoffs, but responsibility.
reliability is a critically important public policy B. NERC should draw on the above-mentioned
objective and should be a primary design quarterly reports from its regional councils to
criterion. prepare annual reports to FERC, appropriate
authorities in Canada, and the public on the
status of the industry’s compliance with recom-
4. Clarify that prudent expenditures and mendations and important trends in electric
investments for bulk system reliability system reliability performance.
(including investments in new technolo-
gies) will be recoverable through trans- The August 14 blackout shared a number of con-
mission rates.8 tributing factors with prior large-scale blackouts,
146 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
confirming that the lessons and recommendations suits; and 2) affirm publicly that actions to shed
from earlier blackouts had not been adequately load pursuant to such guidelines are not indicative
implemented, at least in some geographic areas. of operator failure.
Accordingly, parallel and coordinated efforts are Timely and sufficient action to shed load on
needed by the relevant government agencies and August 14 would have prevented the spread of the
NERC to track the implementation of recommen- blackout beyond northern Ohio. NERC has
dations by governments and the electricity indus- directed all the regional councils in all areas of
try. WECC and NPCC have already established North America to review the applicability of plans
programs that could serve as models for tracking for under-voltage load shedding, and to support
implementation of recommendations. the development of such capabilities where they
would be beneficial. However, organizations and
6. FERC should not approve the operation individual operators may hesitate to initiate such
of a new RTO or ISO until the applicant actions in appropriate circumstances without
assurances that they will not be subject to liability
has met the minimum functional
suits or other forms of retaliation, provided their
requirements for reliability action is pursuant to previously approved
coordinators. guidelines.
The events of August 14 confirmed that MISO did
not yet have all of the functional capabilities 9. Integrate a “reliability impact” consid-
required to fulfill its responsibilities as reliability eration into the regulatory decision-
coordinator for the large area within its footprint.
making process.12
FERC should not authorize a new RTO or ISO to
become operational until the RTO or ISO has veri- The Task Force recommends that FERC, appropri-
fied that all critical reliability capabilities will be ate authorities in Canada, and state regulators inte-
functional upon commencement of RTO or ISO grate a formal reliability impact consideration into
operations. their regulatory decision-making to ensure that
their actions or initiatives either improve or at
minimum do no harm to reliability.
7. Require any entity operating as part of Regulatory actions can have unintended conse-
the bulk power system to be a member quences. For example, in reviewing proposed util-
of a regional reliability council if it op- ity company mergers, FERC’s primary focus has
erates within the council’s footprint.10 been on financial and rate issues, as opposed to
The Task Force recommends that FERC and appro- the reliability implications of such mergers. To
priate authorities in Canada be empowered minimize unintended harm to reliability, and aid
through legislation, if necessary, to require all enti- the improvement of reliability where appropriate,
ties that operate as part of the bulk electric system the Task Force recommends that regulators incor-
to certify that they are members of the regional porate a formal reliability impact consideration
reliability council for all NERC regions in which into their decision processes. At the same time,
they operate. regulators should be watchful for use of alleged
This requirement is needed to ensure that all rele- reliability impacts as a smokescreen for anti-
vant parties are subject to NERC standards, poli- competitive or discriminatory behavior.
cies, etc., in all NERC regions in which they
operate. Action by the Congress or legislative bod- 10. Establish an independent source of
ies in Canada may be necessary to provide appro- reliability performance information.13
priate authority.
The U.S. Department of Energy’s Energy Informa-
tion Administration (EIA), in coordination with
8. Shield operators who initiate load shed- other interested agencies and data sources (FERC,
ding pursuant to approved guidelines appropriate Canadian government agencies, NERC,
from liability or retaliation.11 RTOs, ISOs, the regional councils, transmission
operators, and generators) should establish com-
Legislative bodies and regulators should: 1) estab- mon definitions and information collection stan-
lish that operators (whether organizations or indi- dards. If the necessary resources can be identified,
viduals) who initiate load shedding pursuant to EIA should expand its current activities to include
operational guidelines are not subject to liability information on reliability performance.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 147
Energy policy makers and a wide range of eco- transmission system is now being used to transmit
nomic decision makers need objective, factual power over distances and at volumes that were not
information about basic trends in reliability per- envisioned when the system was designed, and
formance. EIA and the other organizations cited that this functional shift has created major risks
above should identify information gaps in federal that have not been adequately addressed. Indeed,
data collections covering reliability performance some commenters believe that restructuring was a
and physical characteristics. Plans to fill those major cause of the August 14 blackout.
gaps should be developed, and the associated
resource requirements determined. Once those The Task Force believes that the Interim Report
resources have been acquired, EIA should publish accurately identified the primary causes of the
information on trends, patterns, costs, etc. related blackout. It also believes that had existing reliabil-
to reliability performance. ity requirements been followed, either the distur-
bance in northern Ohio that evolved on August 14
into a blackout would not have occurred, or it
11. Establish requirements for collection would have been contained within the FE control
and reporting of data needed for area.
post-blackout analyses. Nevertheless, as discussed at the beginning of this
FERC and appropriate authorities in Canada chapter, the relationship between competition in
should require generators, transmission owners, power markets and reliability is both important
and other relevant entities to collect and report and complex, and careful management and sound
data that may be needed for analysis of blackouts rules are required to achieve the public policy
and other grid-related disturbances. goals of reasonable electricity prices and high reli-
The investigation team found that some of the data ability. At the present stage in the evolution of
needed to analyze the August 14 blackout fully these markets, it is worthwhile for DOE and Natu-
was not collected at the time of the events, and ral Resources Canada (in consultation with FERC
thus could not be reported. Some of the data that and the Canadian Council of Energy Ministers) to
was reported was based on incompatible defini- commission an independent expert study to pro-
tions and formats. As a result, there are aspects of vide advice on how to achieve and sustain an
the blackout, particularly concerning the evolu- appropriate balance in this important area.
tion of the cascade, that may never be fully Among other things, this study should take into
explained. FERC, EIA and appropriate authorities account factors such as:
in Canada should consult with NERC, key mem-
bers of the investigation team, and the industry to u Historical and projected load growth
identify information gaps, adopt common defini- u Location of new generation in relation to old
tions, and establish filing requirements. generation and loads
u Zoning and NIMBY15 constraints on siting of
12. Commission an independent study of generation and transmission
the relationships among industry
restructuring, competition, and reli- u Lack of new transmission investment and its
ability.14 causes
DOE and Natural Resources Canada should com- u Regional comparisons of impact of wholesale
mission an independent study of the relationships electric competition on reliability performance
among industry restructuring, competition in and on investments in reliability and
power markets, and grid reliability, and how those transmission
relationships should be managed to best serve the
public interest. u The financial community’s preferences and
their effects on capital investment patterns
Some participants at the public meetings held in
Cleveland, New York and Toronto to review the u Federal vs. state jurisdictional concerns
Task Force’s Interim Report expressed the view u Impacts of state caps on retail electric rates
that the restructuring of electricity markets for
competition in many jurisdictions has, itself, u Impacts of limited transmission infrastructure
increased the likelihood of major supply interrup- on energy costs, transmission congestion, and
tions. Some of these commenters assert that the reliability
148 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
u Trends in generator fuel and wholesale electric- effects on system stability during major
ity prices disturbances.
u Trends in power flows, line losses, voltage lev- u Development of practical human factors guide-
els, etc. lines for power system control centers.
u Study of obstacles to the economic deployment
13. DOE should expand its research pro- of demand response capability and distributed
grams on reliability-related tools and generation.
technologies.16 u Investigation of alternative approaches to moni-
DOE should expand its research agenda, and con- toring right-of-way vegetation management.
sult frequently with Congress, FERC, NERC, state
regulators, Canadian authorities, universities, and u Study of air traffic control, the airline industry,
the industry in planning and executing this agenda. and other relevant industries for practices and
ideas that could reduce the vulnerability of the
More investment in research is needed to improve electricity industry and its reliability managers
grid reliability, with particular attention to to human error.
improving the capabilities and tools for system
monitoring and management. Research on reli- Cooperative and complementary research and
ability issues and reliability-related technologies funding between nations and between govern-
has a large public-interest component, and gov- ment and industry efforts should be encouraged.
ernment support is crucial. DOE already leads
many research projects in this area, through part- 14. Establish a standing framework for the
nerships with industry and research under way at conduct of future blackout and distur-
the national laboratories and universities. DOE’s
bance investigations.17
leadership and frequent consultation with many
parties are essential to ensure the allocation of The U.S., Canadian, and Mexican governments, in
scarce research funds to urgent projects, bring the consultation with NERC, should establish a stand-
best talent to bear on such projects, and enhance ing framework for the investigation of future black-
the dissemination and timely application of outs, disturbances, or other significant grid-related
research results. incidents.
Important areas for reliability research include but Fortunately, major blackouts are not frequent,
are not limited to: which makes it important to study such events
carefully to learn as much as possible from the
u Development of practical real-time applications experience. In the weeks immediately after
for wide-area system monitoring using phasor August 14, important lessons were learned per-
measurements and other synchronized measur- taining not only to preventing and minimizing
ing devices, including post-disturbance future blackouts, but also to the efficient and fruit-
applications. ful investigation of future grid-related events.
u Development and use of enhanced techniques Appropriate U.S., Canadian, and Mexican govern-
for modeling and simulation of contingencies, ment agencies, in consultation with NERC and
blackouts, and other grid-related disturbances. other organizations, should prepare an agreement
u Investigation of protection and control alterna- that, among other considerations:
tives to slow or stop the spread of a cascading u Establishes criteria for determining when an
power outage, including demand response ini- investigation should be initiated.
tiatives to slow or halt voltage collapse.
u Establishes the composition of a task force to
u Re-evaluation of generator and customer equip- provide overall guidance for the inquiry. The
ment protection requirements based on voltage task force should be international if the trigger-
and frequency phenomena experienced during ing event had international consequences.
the August 14, 2003, cascade.
u Provides for coordination with state and provin-
u Investigation of protection and control of gener- cial governments, NERC and other appropriate
ating units, including the possibility of multiple entities.
steps of over-frequency protection and possible
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 149
u Designates agencies responsible for issuing NERC:
directives concerning preservation of records, Specified measures required in that area to
provision of data within specified periods to a help ensure the reliability of the FE system and
data warehouse facility, conduct of onsite inter- avoid undue risks to neighboring systems.
views with control room personnel, etc. However, the blackout investigation did not ex-
amine conditions in FE service areas in other
u Provides guidance on confidentiality of data.
states.
u Identifies types of expertise likely to be needed Task Force:
on the investigation team. Recommends that NERC require FE to review
its entire service territory, in all states, to de-
Group II. Support and Strengthen termine whether similar vulnerabilities exist
NERC’s Actions of February 10, 2004 and require prompt attention. This review
should be completed by June 30, 2004, and the
On February 10, 2004, after taking the findings of
results reported to FERC, NERC, and utility
the Task Force’s investigation into the August 14,
regulatory authorities in the affected states.
2003, blackout into account, the NERC Board of
Trustees approved a series of actions and strategic
and technical initiatives intended to protect the 2. Interim Voltage Criteria
reliability of the North American bulk electric sys- NERC:
tem. (See Appendix D for the full text of the Required that FE, consistent with or as part of a
Board’s statement of February 10.) Overall, the study ordered by FERC on December 24,
Task Force supports NERC’s actions and initia- 2003,19 determine the minimum acceptable lo-
tives strongly. On some subjects, the Task Force cation-specific voltages at all 345 kV and 138
advocates additional measures, as shown in the kV buses and all generating stations within the
next 17 recommendations. FE control area (including merchant plants).
Further, FE is to determine the minimum dy-
namic reactive reserves that must be main-
15. Correct the direct causes of the tained in local areas to ensure that these mini-
August 14, 2003 blackout.18 mum voltages are met following contingencies
NERC played an important role in the Task Force’s studied in accordance with ECAR Document
blackout investigation, and as a result of the find- 1.20 Criteria and minimum voltage require-
ings of the investigation, NERC issued directives on ments must comply with NERC planning crite-
February 10, 2004 to FirstEnergy, MISO, and PJM ria, including Table 1A, Category C3, and Oper-
to complete a series of remedial actions by June 30, ating Policy 2.21
2004 to correct deficiencies identified as factors Task Force:
contributing to the blackout of August 14, 2003.
Recommends that NERC appoint a team,
(For specifics on the actions required by NERC, see
Appendix D.)
joined by representatives from FERC and the
Ohio Public Utility Commission, to review
The Task Force supports and endorses NERC’s and approve all such criteria.
near-term requirements strongly. It recommends
the addition of requirements pertaining to ECAR,
and several other additional elements, as described
3. FE Actions Based on FERC-Ordered Study
below. NERC:
Required that when the FERC-ordered study is
A. Corrective Actions to Be Completed by completed, FE is to adopt the planning and op-
FirstEnergy by June 30, 2004 erating criteria determined as a result of that
The full text of the remedial actions NERC has study and update the operating criteria and
required that FirstEnergy (FE) complete by June 30 procedures for its system operators. If the study
is provided in Appendix D. The Task Force recom- indicates a need for system reinforcement, FE
mends the addition of certain elements to these is to develop a plan for developing such re-
requirements, as described below. sources as soon as practical and develop opera-
tional procedures or other mitigating programs
1. Examination of Other FE Service Areas to maintain safe operating conditions until
The Task Force’s investigation found severe reac- such time that the necessary system reinforce-
tive power and operations criteria deficiencies in ments can be made.
the Cleveland-Akron area.
150 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Task Force: capability may be provided by automatic or
Recommends that a team appointed by NERC manual load shedding, voltage reduction, di-
and joined by representatives from FERC and rect-controlled commercial or residential load
the Ohio Public Utility Commission should re- management, or any other method or combina-
view and approve this plan. tion of methods capable of achieving the 1500
MW of reduction in ten minutes without ad-
4. Reactive Resources versely affecting other interconnected systems.
NERC: The amount of required load reduction capabil-
Required that FE inspect all reactive resources, ity may be modified to an amount shown by the
including generators, and ensure that all are FERC-ordered study to be sufficient for re-
fully operational. FE is also required to verify sponse to severe contingencies and if approved
that all installed capacitors have no blown by ECAR and NERC.
fuses and that at least 98% of installed capaci- Task Force:
tors (69 kV and higher) are available for service Recommends that NERC require MISO’s ap-
during the summer of 2004. proval of any change in the amount of re-
Task Force: quired load reduction capability. It also rec-
Recommends that NERC also require FE to ommends that NERC require FE’s load reduc-
confirm that all non-utility generators in its tion plan to be shared with the Ohio Public
area have entered into contracts for the sale of Utilities Commission and that FE should com-
generation committing them to producing in- municate with all communities in the affected
creased or maximum reactive power when areas about the plan and its potential conse-
called upon by FE or MISO to do so. Such con- quences.
tracts should ensure that the generator would 7. Emergency Response Plan
be compensated for revenue losses associated
NERC:
with a reduction in real power sales in order
Required that FE develop an emergency re-
to increase production of reactive power.
sponse plan, including arrangements for de-
ploying the load reduction capabilities noted
5. Operational Preparedness and Action Plan above. The plan is to include criteria for deter-
NERC: mining the existence of an emergency and
Required that FE prepare and submit to ECAR identify various possible states of emergency.
an Operational Preparedness and Action Plan The plan is to include detailed operating proce-
to ensure system security and full compliance dures and communication protocols with all
with NERC and planning and operating crite- the relevant entities including MISO, FE opera-
ria, including ECAR Document 1. tors, and market participants within the FE
Task Force: area that have an ability to vary generation out-
Recommends that NERC require copies of this put or shed load upon orders from FE opera-
plan to be provided to FERC, DOE, the Ohio tors. The plan should include procedures for
Public Utility Commission, and the public load restoration after the declaration that the
utility commissions in other states in which FE system is no longer in an emergency operat-
FE operates. The Task Force also recommends ing state.
that NERC require FE to invite its system oper- Task Force:
ations partners—control areas adjacent to FE, Recommends that NERC require FE to offer its
plus MISO, ECAR, and PJM—to participate in system operations partners—i.e., control ar-
the development of the plan and agree to its eas adjacent to FE, plus MISO, ECAR, and
implementation in all aspects that could affect PJM—an opportunity to contribute to the de-
their respective systems and operations. velopment of the plan and agree to its key pro-
visions.
6. Emergency Response Resources
NERC: 8. Operator Communications
Required that FE develop a capability to reduce NERC:
load in the Cleveland-Akron area by 1500 MW Required that FE develop communications pro-
within ten minutes of a directive to do so by cedures for FE operating personnel to use
MISO or the FE system operator. Such a within FE, with MISO and neighboring
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 151
systems, and others. The procedure and the op- 11. Emergency Preparedness Training for
erating environment within the FE system con- Operators
trol center should allow control room staff to
NERC:
focus on reliable system operations and avoid
Required that all reliability coordinators, con-
distractions such as calls from customers and
trol areas, and transmission operators provide
others who are not responsible for operation of
at least five days of training and drills using re-
a portion of the transmission system.
alistic simulation of system emergencies for
Task Force: each staff person with responsibility for the
Recommends that NERC require these proce- real-time operation or reliability monitoring of
dures to be shared with and agreed to by con- the bulk electric system. This system emer-
trol areas adjacent to FE, plus MISO, ECAR, gency training is in addition to other training
and PJM, and any other affected system opera- requirements. The term “realistic simulation”
tions partners, and that these procedures be includes a variety of tools and methods that
tested in a joint drill. present operating personnel with situations to
improve and test diagnostic and decision-
9. Reliability Monitoring and System Manage- making skills in an environment that resembles
ment Tools expected conditions during a particular type of
system emergency.
NERC:
Required that FE ensure that its state estimator Task Force:
and real-time contingency analysis functions Recommends that to provide effective training
are used to execute reliably full contingency before June 30, 2004, NERC should require FE
analyses automatically every ten minutes or on to consider seeking the assistance of another
demand, and used to notify operators of poten- control area or reliability coordinator known
tial first contingency violations. to have a quality training program (such as
IMO or ISO-New England) to provide the
Task Force: needed training with appropriate FE-specific
Recommends that NERC also require FE to en- modifications.
sure that its information technology support
function does not change the effectiveness of B. Corrective Actions to be Completed by MISO
reliability monitoring or management tools in by June 30, 2004
any way without the awareness and consent
of its system operations staff. 1. Reliability Tools
NERC:
10. GE XA21 System Updates and Transition to Required that MISO fully implement and test
New Energy Management System its topology processor to provide its operating
personnel a real-time view of the system status
NERC: for all transmission lines operating and all gen-
Required that until FE replaces its GE XA21 En- erating units within its system, and all critical
ergy Management System, FE should imple- transmission lines and generating units in
ment all current known fixes for the GE XA21 neighboring systems. Alarms should be pro-
system necessary to ensure reliable and stable vided for operators for all critical transmission
operation of critical reliability functions, and line outages and voltage violations. MISO is to
particularly to correct the alarm processor fail- establish a means of exchanging outage infor-
ure that occurred on August 14, 2003. mation with its members and adjacent systems
Task Force: such that the MISO state estimator has accurate
Recommends that NERC require FE to design and timely information to perform as designed.
and test the transition to its planned new en- MISO is to fully implement and test its state es-
ergy management system to ensure that the timation and real-time contingency analysis
system functions effectively, that the transi- tools to ensure they can operate reliably no less
tion is made smoothly, that the system’s oper- than every ten minutes. MISO is to provide
ators are adequately trained, and that all op- backup capability for all functions critical to
erating partners are aware of the transition. reliability.
152 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Task Force: ECAR should consult with an expert team ap-
Recommends that NERC require MISO to en- pointed by NERC—joined by representatives
sure that its information technology support from FERC, DOE, interested state commis-
staff does not change the effectiveness of reli- sions, and MISO—to develop better modeling
ability monitoring or management tools in procedures and scenarios, and obtain review
any way without the awareness and consent of future assessments by the expert team.
of its system operations staff.
2. Verification of Data and Assumptions
2. Operating Agreements Task Force:
NERC: Recommends that NERC require ECAR to re-
Required that MISO reevaluate its operating examine and validate all data and model as-
agreements with member entities to verify its sumptions against current physical asset ca-
authority to address operating issues, includ- pabilities and match modeled assets (such as
ing voltage and reactive management, voltage line characteristics and ratings, and generator
scheduling, the deployment and redispatch of reactive power output capabilities) to current
real and reactive reserves for emergency re- operating study assessments.
sponse, and the authority to direct actions dur-
ing system emergencies, including shedding 3. Ensure Consistency of Members’ Data
load. Task Force:
Task Force: Recommends that NERC require ECAR to con-
Recommends that NERC require that any duct a data validation and exchange exercise
problems or concerns related to these operat- to be sure that its members are using accurate,
ing issues be raised promptly with FERC and consistent, and current physical asset charac-
MISO’s members for resolution. teristics and capabilities for both long-term
and seasonal assessments and operating stud-
C. Corrective Actions to be Completed by PJM ies.
by June 30, 2004
NERC: E. Task Force Recommendation for Corrective
Required that PJM reevaluate and improve its Actions to be Completed by Other Parties by
communications protocols and procedures be- June 30, 2004
tween PJM and its neighboring control areas Task Force:
and reliability coordinators. Recommends that NERC require each North
Task Force: American reliability coordinator, reliability
Recommends that NERC require definitions council, control area, and transmission com-
and usages of key terms be standardized, and pany not directly addressed above to review
non-essential communications be minimized the actions required above and determine
during disturbances, alerts, or emergencies. whether it has adequate system facilities, op-
NERC should also require PJM, MISO, and erational procedures, tools, and training to
their member companies to conduct one or ensure reliable operations for the summer of
more joint drills using the new communica- 2004. If any entity finds that improvements
tions procedures. are needed, it should immediately undertake
the needed improvements, and coordinate
them with its neighbors and partners as neces-
D. Task Force Recommendations for Corrective
sary.
Actions to be Completed by ECAR by August
14, 2004 The Task Force also recommends that FERC
and government agencies in Canada require
1. Modeling and Assessments all entities under their jurisdiction who are
Task Force: users of GE/Harris XA21 Energy Management
Recommends that NERC require ECAR to re- Systems to consult the vendor and ensure that
evaluate its modeling procedures, assump- appropriate actions have been taken to avert
tions, scenarios and data for seasonal assess- any recurrence of the malfunction that oc-
ments and extreme conditions evaluations. curred on FE’s system on August 14.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 153
16. Establish enforceable standards for A. Enforceable Standards
maintenance of electrical clearances in NERC should work with FERC, government agen-
right-of-way areas.22 cies in Canada, state regulatory agencies, the Insti-
tute of Electrical and Electronic Engineers (IEEE),
On February 10, the NERC Board directed the
NERC Compliance Program and the regional coun- utility arborists, and other experts from the U.S.
cils to initiate a joint program for reporting all and Canada to develop clear, unambiguous stan-
bulk electric system transmission line trips result- dards pertaining to maintenance of safe clearances
ing from vegetation contact. Based on the results of of transmission lines from obstructions in the
these filings, NERC is to consider the development lines’ right-of-way areas, and procedures to verify
of minimum line clearance standards to ensure compliance with the standards. States, provinces,
reliability. and local governments should remain free to set
The Task Force believes that more aggressive more specific or higher standards as they deem
action is warranted. NERC should work with necessary for their respective areas.
FERC, appropriate authorities in Canada, state reg-
B. Right-of-Way Management Plan
ulatory agencies, the Institute of Electrical and
Electronic Engineers (IEEE), utility arborists, and NERC should require each bulk electric transmis-
other experts from the US and Canada to develop sion operator to publish annually a proposed
clear, unambiguous standards pertaining to main- right-of-way management plan on its public
tenance of safe clearances of transmission lines website, and a report on its right-of-way manage-
from obstructions in the lines’ right-of-way areas, ment activities for the previous year. The manage-
and to develop a mechanism to verify compliance ment plan should include the planned frequency
with the standards and impose penalties for non- of actions such as right-of-way trimming, herbi-
compliance. cide treatment, and inspections, and the report
should give the dates when the rights-of-way in a
Ineffective vegetation management was a major given district were last inspected and corrective
cause of the August 14, 2003, blackout and it was actions taken.
also a causal factor in other large-scale North
American outages such as those that occurred in C. Requirement to Report Outages Due to
the summer of 1996 in the western United States. Ground Faults in Right-of-Way Areas
Maintaining transmission line rights-of-way, Beginning with an effective date of March 31,
including maintaining safe clearances of ener- 2004, NERC should require each transmission
gized lines from vegetation, man-made structures, owner/operator to submit quarterly reports of all
bird nests, etc., requires substantial expenditures ground-fault line trips, including their causes, on
in many areas of North America. However, such lines of 115 kV and higher in its footprint to the
maintenance is a critical investment for ensuring a regional councils. Failure to report such trips
reliable electric system. For a review of current should lead to an appropriate penalty. Each
issues pertaining to utility vegetation manage- regional council should assemble a detailed
ment programs, see Utility Vegetation Manage- annual report on ground fault line trips and their
ment Final Report, March 2004.23 causes in its area to FERC, NERC, DOE, appropri-
ate authorities in Canada, and state regulators no
NERC does not presently have standards for later than March 31 for the preceding year, with
right-of-way maintenance. However, it has stan- the first annual report to be filed in March 2005 for
dards requiring that line ratings be set to maintain calendar year 2004.
safe clearances from all obstructions. Line rating
standards should be reviewed to ensure that they D. Transmission-Related Vegetation Manage-
are sufficiently clear and explicit. In the United ment Expenses, if Prudently Incurred,
States, National Electrical Safety Code (NESC) Should be Recoverable through Electric
rules specify safety clearances required for over- Rates
head conductors from grounded objects and other The level of activity in vegetation management
types of obstructions, but those rules are subject to programs in many utilities and states has fluctu-
broad interpretation. Several states have adopted ated widely from year to year, due in part to incon-
their own electrical safety codes and similar codes sistent funding and varying management support.
apply in Canada and its provinces. A mechanism Utility managers and regulators should recognize
is needed to verify compliance with these require- the importance of effective vegetation manage-
ments and to penalize noncompliance. ment to transmission system reliability, and that
154 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
changes in vegetation management may be needed regulatory authorities in the United States and
in response to weather, insect infestations, and Canada.
other factors. Transmission vegetation manage- Task Force:
ment programs should be consistently funded and Recommends that NERC inform the federal
proactively managed to maintain and improve and state or provincial authorities of both
system reliability. countries of the final results of all enforce-
ment proceedings, and make the results of
17. Strengthen the NERC Compliance such proceedings public.
Enforcement Program.
C. Violations in August 14, 2003 Blackout
On February 10, 2004, the NERC Board of Trustees
approved directives to the regional reliability NERC:
councils that will significantly strengthen NERC’s The Compliance and Standards investigation
existing Compliance Enforcement Program. The team will issue a final report in March or April
Task Force supports these directives strongly, and of 2004 of violations of NERC and regional
recommends certain additional actions, as standards that occurred on August 14. (Seven
described below.24 violations are noted in this report (pages 19-
20), but additional violations may be identified
A. Reporting of Violations by NERC.) Within three months of the issuance
NERC: of the report, NERC is to develop recommenda-
Requires each regional council to report to the tions to improve the compliance process.
NERC Compliance Enforcement Program Task Force:
within one month of occurrence all “significant Recommends that NERC make its recommen-
violations” of NERC operating policies and dations available to appropriate U.S. federal
planning standards and regional standards, and state authorities, to appropriate authori-
whether verified or still under investigation by ties in Canada, and to the public.
the regional council. (A “significant violation”
is one that could directly reduce the integrity of D. Compliance Audits
the interconnected power systems or otherwise
cause unfavorable risk to the interconnected NERC:
power systems.) In addition, each regional Established plans for two types of audits, com-
council is to report quarterly to NERC, in a for- pliance audits and readiness audits. Compli-
mat prescribed by NERC, all violations of ance audits would determine whether the sub-
NERC and regional reliability standards. ject entity is in documented compliance with
NERC standards, policies, etc. Readiness au-
Task Force: dits focus on whether the entity is functionally
Recommends that NERC require the regional capable of meeting the terms of its reliability re-
councils’ quarterly reports and reports on sig- sponsibilities. Under the terms approved by
nificant violations be filed as public docu- NERC’s Board, the readiness audits to be com-
ments with FERC and appropriate authorities pleted by June 30, 2004, will be conducted us-
in Canada, at the same time that they are sent ing existing NERC rules, policies, standards,
to NERC. and NERC compliance templates. Require-
ments for control areas will be based on the ex-
B. Enforcement Action by NERC Board isting NERC Control Area Certification Proce-
NERC: dure, and updated as new criteria are ap-
After being presented with the results of the in- proved.
vestigation of a significant violation, the Board Task Force:
is to require an offending organization to cor- Supports the NERC effort to verify that all
rect the violation within a specified time. If the entities are compliant with reliability stan-
Board determines that the organization is dards. Effective compliance and auditing will
non-responsive and continues to cause a risk to require that the NERC standards be im-
the reliability of the interconnected power sys- proved rapidly to make them clear, unambig-
tems, the Board will seek to remedy the viola- uous, measurable, and consistent with the
tion by requesting assistance from appropriate Functional Model.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 155
E. Audit Standards and Composition of Audit A. Readiness Audits
Teams NERC:
NERC: In its directives of February 10, 2004, NERC in-
Under the terms approved by the Board, the re- dicated that it and the regional councils would
gional councils are to have primary responsi- jointly establish a program to audit the reliabil-
bility for conducting the compliance audits, ity readiness of all reliability coordinators and
under the oversight and direct participation of control areas within three years and continuing
staff from the NERC Compliance Enforcement thereafter on a three-year cycle. Twenty audits
Program. FERC and other relevant regulatory of high-priority areas will be completed by June
agencies will be invited to participate in the au- 30, 2004, with particular attention to deficien-
dits, subject to the same confidentiality condi- cies identified in the investigation of the Au-
tions as the other team members. gust 14 blackout.
Task Force: Task Force:
Recommends that each team should have Recommends that the remainder of the first
some members who are electric reliability ex- round of audits be completed within two
perts from outside the region in which the au- years, as compared to NERC’s plan for three
dit is occurring. Also, some team members years.
should be from outside the electricity indus-
try, i.e., individuals with experience in sys- B. Public Release of Readiness Audit Reports
tems engineering and management, such as Task Force:
persons from the nuclear power industry, the Recommends that NERC require all readiness
U.S. Navy, the aerospace industry, air traffic audit reports to be publicly posted, excluding
control, or other relevant industries or gov- portions pertaining to physical and cyber se-
ernment agencies. To improve the objectivity curity. Reports should also be sent directly to
and consistency of investigation and perfor- DOE, FERC, and relevant authorities in Can-
mance, NERC-organized teams should con- ada and state commissions. Such reports
duct these compliance audits, using NERC cri- should draw clear distinctions between seri-
teria (with regional variations if more strin- ous and minor violations of reliability stan-
gent), as opposed to the regional councils us- dards or related requirements.
ing regionally developed criteria.
F. Public Release of Compliance Audit Reports 19. Improve near-term and long-term
Task Force: training and certification requirements
Recommends that NERC require all compli- for operators, reliability coordinators,
ance audit reports to be publicly posted, ex- and operator support staff.26
cluding portions pertaining to physical and In its requirements of February 10, 2004, NERC
cyber security according to predetermined directed that all reliability coordinators, control
criteria. Such reports should draw clear dis- areas, and transmission operators are to provide at
tinctions between serious and minor viola- least five days per year of training and drills in
tions of reliability standards or related re- system emergencies, using realistic simulations, for
quirements. each staff person with responsibility for the
real-time operation or reliability monitoring of the
bulk electric system. This system emergency train-
18. Support and strengthen NERC’s Reli- ing is in addition to other training requirements.
ability Readiness Audit Program.25 Five days of system emergency training and drills
are to be completed by June 30, 2004.
On February 10, 2004, the NERC Board of Trustees
approved the establishment of a NERC program for The Task Force supports these near-term require-
periodic reviews of the reliability readiness of all ments strongly. For the long term, the Task Force
reliability coordinators and control areas. The recommends that:
Task Force strongly supports this action, and rec- A. NERC should require training for the planning
ommends certain additional measures, as staff at control areas and reliability coordina-
described below. tors concerning power system characteristics
156 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
and load, VAr, and voltage limits, to enable NERC’s near-term requirements for emergency
them to develop rules for operating staff to fol- preparedness training are described above. For the
low. long term, training for system emergencies should
B. NERC should require control areas and reliabil- be fully integrated into the broader training pro-
ity coordinators to train grid operators, IT sup- grams required for all system planners, system
port personnel, and their supervisors to operators, their supervisors, and other control
recognize and respond to abnormal automation room support staff.
system activity.
C. NERC should commission an advisory report by Advisory Report by Independent Panel on
an independent panel to address a wide range Industry Training Programs and Certification
of issues concerning reliability training pro- Requirements
grams and certification requirements. Under the oversight of FERC and appropriate
The Task Force investigation team found that Canadian authorities, the Task Force recommends
some reliability coordinators and control area that NERC commission an independent advisory
operators had not received adequate training in panel of experts to design and propose minimum
recognizing and responding to system emergen- training programs and certification procedures for
cies. Most notable was the lack of realistic simula- the industry’s control room managers and staff.
tions and drills to train and verify the capabilities This panel should be comprised of experts from
of operating personnel. Such simulations are electric industry organizations with outstanding
essential if operators and other staff are to be able training programs, universities, and other indus-
to respond adequately to emergencies. This train- tries that operate large safety or reliability-
ing deficiency contributed to the lack of situa- oriented systems and training programs. (The
tional awareness and failure to declare an Institute of Nuclear Power Operations (INPO), for
emergency on August 14 while operator interven- example, provides training and other safety-
tion was still possible (before events began to related services to operators of U.S. nuclear power
occur at a speed beyond human control). plants and plants in other countries.) The panel’s
report should provide guidance on issues such as:
Control rooms must remain functional under a
wide range of possible conditions. Any person 1. Content of programs for new trainees
with access to a control room should be trained so
2. Content of programs for existing operators and
that he or she understands the basic functions of
other categories of employees
the control room, and his or her role in relation to
those of others in the room under any conditions. 3. Content of continuing education programs and
Information technology (IT) staff, in particular, fraction of employee time to be committed to
should have a detailed understanding of the infor- ongoing training
mation needs of the system operators under alter-
native conditions. 4. Going beyond paper-based, fact-oriented
“knowledge” requirements for operators—i.e.,
The Task Force’s cyber investigation team noted confirming that an individual has the ability to
in its site visits an increasing reliance by control cope with unforeseen situations and
areas and utilities on automated systems to mea- emergencies
sure, report on, and change a wide variety of phys-
ical processes associated with utility operations.27 5. In-house training vs. training by independent
If anything, this trend is likely to intensify in the parties
future. These systems enable the achievement of 6. Periodic accreditation of training programs
major operational efficiencies, but their failure
could cause or contribute to blackouts, as evi- 7. Who should certify trained staff?
denced by the alarm failures at FirstEnergy and
8. Criteria to establish grades or levels of operator
the state estimator deactivation at MISO.
qualifications from entry level to supervisor or
Grid operators should be trained to recognize and manager, based on education, training, and
respond more efficiently to security and automa- experience.
tion problems, reinforced through the use of peri-
odic exercises. Likewise, IT support personnel The panel’s report should be delivered by March
should be better trained to understand and 31, 2005. FERC and Canadian authorities, in con-
respond to the requirements of grid operators dur- sultation with NERC and others, should evaluate
ing security and IT incidents. the report and consider its findings in setting
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 157
minimum training and certification requirements needed, between “normal” and “emergency.” In
for control areas and reliability coordinators. some areas (e.g., NPCC) an “alert” state has already
been established.
20. Establish clear definitions for normal,
alert and emergency operational sys- 21. Make more effective and wider use of
tem conditions. Clarify roles, responsi- system protection measures.29
bilities, and authorities of reliability In its requirements of February 10, 2004, NERC:
coordinators and control areas under
A. Directed all transmission owners to evaluate
each condition.28 the settings of zone 3 relays on all transmission
NERC should develop by June 30, 2004 definitions lines of 230 kV and higher.
for normal, alert, and emergency system condi- B. Directed all regional councils to evaluate the
tions, and clarify reliability coordinator and con- feasibility and benefits of installing
trol area functions, responsibilities, required under-voltage load shedding capability in load
capabilities, and required authorities under each centers.
operational system condition.
C. Called for an evaluation within one year of its
System operators need common definitions for planning standard on system protection and
normal, alert, and emergency conditions to enable control to take into account the lessons from the
them to act appropriately and predictably as sys- August 14 blackout.
tem conditions change. On August 14, the princi-
The Task Force supports these actions strongly,
pal entities involved in the blackout did not have a
and recommends certain additional measures, as
shared understanding of whether the grid was in described below.
an emergency condition, nor did they have a com-
mon understanding of the functions, responsibili- A. Evaluation of Zone 3 Relays
ties, capabilities, and authorities of reliability NERC:
coordinators and control areas under emergency Industry is to review zone 3 relays on lines of
or near-emergency conditions. 230 kV and higher.
NERC: Task Force:
On February 10, 2004, NERC’s Board of Recommends that NERC broaden the review
Trustees directed NERC’s Operating Commit- to include operationally significant 115 kV
tee to “clarify reliability coordinator and con- and 138 kV lines, e.g., lines that are part of
trol area functions, responsibilities, capabili- monitored flowgates or interfaces. Transmis-
ties, and authorities” by June 30, 2004. sion owners should also look for zone 2 relays
set to operate like zone 3s.
Task Force:
Recommends that NERC go further and de-
B. Evaluation of Applicability of Under-Voltage
velop clear definitions of three operating sys-
Load Shedding
tem conditions, along with clear statements of
the roles and responsibilities of all partici- NERC:
pants, to ensure effective and timely actions in Required each regional reliability council to
critical situations. evaluate the feasibility and benefits of un-
der-voltage load shedding (UVLS) capability in
Designating three alternative system conditions load centers that could become unstable as a re-
(normal, alert, and emergency) would help grid sult of insufficient reactive power following
managers to avert and deal with emergencies credible multiple-contingency events. The re-
through preventive action. Many difficult situa- gions should complete the initial studies and
tions are avoidable through strict adherence to report the results to NERC within one year. The
sound procedures during normal operations. regions should promote the installation of un-
However, unanticipated difficulties short of an der-voltage load shedding capabilities within
emergency still arise, and they must be addressed critical areas where beneficial, as determined
swiftly and skillfully to prevent them from becom- by the studies to be effective in preventing or
ing emergencies. Doing so requires a high level of containing an uncontrolled cascade of the
situational awareness that is difficult to sustain power system.
indefinitely, so an intermediate “alert” state is
158 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Task Force: year the real-time operating tools necessary for
Recommends that NERC require the results of reliability operation and reliability coordination,
the regional studies to be provided to federal including backup capabilities. The committee’s
and state or provincial regulators at the same report is to address both minimum acceptable
capabilities for critical reliability functions and a
time that they are reported to NERC. In addi-
guide to best practices.
tion, NERC should require every entity with a
new or existing UVLS program to have a The Task Force supports these requirements
well-documented set of guidelines for opera- strongly. It recommends that NERC require the
tors that specify the conditions and triggers for committee to:
UVLS use. A. Give particular attention in its report to the
development of guidance to control areas and
C. Evaluation of NERC’s Planning Standard III reliability coordinators on the use of automated
NERC: wide-area situation visualization display sys-
tems and the integrity of data used in those sys-
Plans to evaluate Planning Standard III, System
tems.
Protection and Control, and propose, by March
1, 2005, specific revisions to the criteria to ad- B. Prepare its report in consultation with FERC,
dress adequately the issue of slowing or limit- appropriate authorities in Canada, DOE, and
ing the propagation of a cascading failure, in the regional councils. The report should also
light of the experience gained on August 14. inform actions by FERC and Canadian
government agencies to establish minimum
Task Force: functional requirements for control area opera-
Recommends that NERC, as part of the review tors and reliability coordinators.
of Planning Standard III, determine the goals
The Task Force also recommends that FERC, DHS,
and principles needed to establish an inte-
and appropriate authorities in Canada should
grated approach to relay protection for gener- require annual independent testing and certifica-
ators and transmission lines and the use of un- tion of industry EMS and SCADA systems to ensure
der-frequency and under-voltage load shed- that they meet the minimum requirements envi-
ding (UFLS and UVLS) programs. An inte- sioned in Recommendation 3.
grated approach is needed to ensure that at the
local and regional level these interactive com- A principal cause of the August 14 blackout was a
ponents provide an appropriate balance of lack of situational awareness, which was in turn
risks and benefits in terms of protecting spe- the result of inadequate reliability tools and
cific assets and facilitating overall grid sur- backup capabilities. In addition, the failure of FE’s
vival. This review should take into account control computers and alarm system contributed
the evidence from August 14 of some unin- directly to the lack of situational awareness. Like-
tended consequences of installing Zone 3 re- wise, MISO’s incomplete tool set and the failure to
lays and using manufacturer-recommended supply its state estimator with correct system data
settings for relays protecting generators. It on August 14 contributed to the lack of situational
should also include an assessment of the ap- awareness. The need for improved visualization
propriate role and scope of UFLS and UVLS, capabilities over a wide geographic area has been a
and the appropriate use of time delays in re- recurrent theme in blackout investigations. Some
lays. wide-area tools to aid situational awareness (e.g.,
Recommends that in this effort NERC should real-time phasor measurement systems) have been
work with industry and government research tested in some regions but are not yet in general
organizations to assess the applicability of ex- use. Improvements in this area will require signifi-
isting and new technology to make the inter- cant new investments involving existing or emerg-
connections less susceptible to cascading out- ing technologies.
ages.
The investigation of the August 14 blackout
revealed that there has been no consistent means
22. Evaluate and adopt better real-time across the Eastern Interconnection to provide an
tools for operators and reliability coor- understanding of the status of the power grid out-
side of a control area. Improved visibility of the
dinators.30
status of the grid beyond an operator’s own area of
NERC’s requirements of February 10, 2004, direct control would aid the operator in making adjust-
its Operating Committee to evaluate within one ments in its operations to mitigate potential
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 159
problems. The expanded view advocated above adequacy analysis comparable to that outlined in
would also enable facilities to be more proactive in FERC’s December 24, 2003, Order33 to FirstEnergy
operations and contingency planning. concerning the Cleveland-Akron area. The Task
Annual testing and certification by independent, Force recommends that NERC develop a priori-
qualified parties is needed because EMS and tized list for which areas and loads need this type
SCADA systems are the nerve centers of bulk elec- of analysis and a schedule that ensures that the
tric networks. Ensuring that these systems are analysis will be completed for all such load cen-
functioning properly is critical to sound and reli- ters by December 31, 2005.
able operation of the networks.
24. Improve quality of system modeling
23. Strengthen reactive power and voltage data and data exchange practices.34
control practices in all NERC regions.31 NERC’s requirements of February 10, 2004 direct
that within one year the regional councils are to
NERC’s requirements of February 10, 2004 call for
establish and begin implementing criteria and pro-
a reevaluation within one year of existing reactive
cedures for validating data used in power flow
power and voltage control standards and how they
are being implemented in the ten NERC regions. models and dynamic simulations by benchmarking
However, by June 30, 2004, ECAR is required to model data with actual system performance. Vali-
review its reactive power and voltage criteria and dated modeling data shall be exchanged on an
procedures, verify that its criteria and procedures inter-regional basis as needed for reliable system
are being fully implemented in regional and mem- planning and operation.
ber studies and operations, and report the results
The Task Force supports these requirements
to the NERC Board.
strongly. The Task Force also recommends that
The Task Force supports these requirements FERC and appropriate authorities in Canada
strongly. It recommends that NERC require the require all generators, regardless of ownership, to
regional analyses to include recommendations for collect and submit generator data to NERC, using a
appropriate improvements in operations or facili- regulator-approved template.
ties, and to be subject to rigorous peer review by
experts from within and outside the affected areas. The after-the-fact models developed to simulate
August 14 conditions and events found that the
The Task Force also recommends that FERC and dynamic modeling assumptions for generator and
appropriate authorities in Canada require all tar-
load power factors in regional planning and oper-
iffs or contracts for the sale of generation to
include provisions specifying that the generators
ating models were frequently inaccurate. In par-
can be called upon to provide or increase reactive ticular, the assumptions of load power factor were
power output if needed for reliability purposes, overly optimistic—loads were absorbing much
and that the generators will be paid for any lost more reactive power than the pre-August 14 mod-
revenues associated with a reduction of real power els indicated. Another suspected problem con-
sales attributable to a required increase in the pro- cerns modeling of shunt capacitors under
duction of reactive power. depressed voltage conditions.
Reactive power problems were a significant factor NERC should work with the regional reliability
in the August 14 outage, and they were also impor- councils to establish regional power system mod-
tant elements in several of the earlier outages els that enable the sharing of consistent and vali-
detailed in Chapter 7.32 Accordingly, the Task dated data among entities in the region. Power
Force agrees that a comprehensive review is flow and transient stability simulations should be
needed of North American practices with respect periodically benchmarked with actual system
to managing reactive power requirements and
events to validate model data. Viable load (includ-
maintaining an appropriate balance among alter-
ing load power factor) and generator testing pro-
native types of reactive resources.
grams are necessary to improve agreement
Regional Analyses, Peer Reviews, and Follow- between power flows and dynamic simulations
Up Actions and the actual system performance.
The Task Force recommends that each regional During the data collection phase of the blackout
reliability council, working with reliability coor- investigation, when control areas were asked for
dinators and the control areas serving major load information pertaining to merchant generation
centers, should conduct a rigorous reliability and within their area, the requested data was
160 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
frequently not available because the control area Washington, must be designed and operated in
had not recorded the status or output of the gener- accordance with a higher level of reliability than
ator at a given point in time. Some control area would be appropriate for sparsely populated
operators also asserted that some of the data that parts of the country. It is essential that regional
did exist was commercially sensitive or confiden- differences in terms of load and population den-
tial. To correct such problems, the Task Force rec- sity be recognized in the application of planning
ommends that FERC and authorities in Canada and operating criteria. Any move to adopt a
require all generators, regardless of ownership, to national, “one size fits all” formula for all parts
collect and submit generator data, according to a of the United States would be disastrous to
regulator-approved template. reliability . . . .
A strong transmission system designed and oper-
25. NERC should reevaluate its existing ated in accordance with weakened criteria
reliability standards development pro- would be disastrous. Instead, a concerted effort
cess and accelerate the adoption of should be undertaken to determine if existing
enforceable standards.35 reliability criteria should be strengthened. Such
an effort would recognize the geo-electrical mag-
The Task Force recommends that, with support nitude of today’s interconnected networks, and
from FERC and appropriate authorities in Canada, the increased complexities deregulation and
NERC should:
restructuring have introduced in planning and
A. Re-examine its existing body of standards, operating North American power systems. Most
guidelines, etc., to identify those that are most important, reliability should be considered a
important and ensure that all concerns that higher priority than commercial use. Only
merit standards are addressed in the plan for through strong standards and careful engineer-
standards development. ing can unacceptable power failures like the
B. Re-examine the plan to ensure that those that August 14 blackout be avoided in the future.37
are the most important or the most out-of-date
are addressed early in the process.
C. Build on existing provisions and focus on what 26. Tighten communications protocols,
needs improvement, and incorporate compli- especially for communications during
ance and readiness considerations into the alerts and emergencies. Upgrade com-
drafting process. munication system hardware where
D. Re-examine the Standards Authorization appropriate.38
Request process to determine whether, for each NERC should work with reliability coordinators
standard, a review and modification of an exist- and control area operators to improve the effective-
ing standard would be more efficient than ness of internal and external communications dur-
development of wholly new text for the stan- ing alerts, emergencies, or other critical situations,
dard. and ensure that all key parties, including state and
NERC has already begun a long-term, systematic local officials, receive timely and accurate infor-
process to reevaluate its standards. It is of the mation. NERC should task the regional councils to
work together to develop communications proto-
greatest importance, however, that this process
cols by December 31, 2004, and to assess and
not dilute the content of the existing standards, report on the adequacy of emergency communica-
nor conflict with the right of regions or other areas tions systems within their regions against the pro-
to impose more stringent standards. The state of tocols by that date.
New York, for example, operates under mandatory
and more stringent reliability rules and standards On August 14, 2003, reliability coordinator and
than those required by NERC and NPCC.36 control area communications regarding condi-
tions in northeastern Ohio were in some cases
Similarly, several commenters on the Interim
ineffective, unprofessional, and confusing. Inef-
Report wrote jointly that:
fective communications contributed to a lack of
NERC standards are the minimum—national situational awareness and precluded effective
standards should always be minimum rather actions to prevent the cascade. Consistent applica-
than absolute or “one size fits all” criteria. [Sys- tion of effective communications protocols, par-
tems for] densely populated areas, like the ticularly during alerts and emergencies, is
metropolitan areas of New York, Chicago, or essential to reliability. Standing hotline networks,
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 161
or a functional equivalent, should be established investigate future system disturbances, outages,
for use in alerts and emergencies (as opposed to or blackouts.
one-on-one phone calls) to ensure that all key par- B. NERC, reliability coordinators, control areas,
ties are able to give and receive timely and accu- and transmission owners should determine
rate information. where high speed power system disturbance
recorders are needed on the system, and ensure
that they are installed by December 31, 2004.
27. Develop enforceable standards for
transmission line ratings.39 C. NERC should establish data recording proto-
cols.
NERC should develop clear, unambiguous require-
ments for the calculation of transmission line D. FERC and appropriate authorities in Canada
should ensure that the investments called for in
ratings (including dynamic ratings), and require this recommendation will be recoverable
that all lines of 115 kV or higher be rerated accord- through transmission rates.
ing to these requirements by June 30, 2005.
A valuable lesson from the August 14 blackout is
As seen on August 14, inadequate vegetation man- the importance of having time-synchronized sys-
agement can lead to the loss of transmission lines tem data recorders. The Task Force’s investigators
that are not overloaded, at least not according to labored over thousands of data items to determine
their rated limits. The investigation of the black- the sequence of events, much like putting together
out, however, also found that even after allowing small pieces of a very large puzzle. That process
for regional or geographic differences, there is still would have been significantly faster and easier if
significant variation in how the ratings of existing there had been wider use of synchronized data
lines have been calculated. This variation—in recording devices.
terms of assumed ambient temperatures, wind NERC Planning Standard I.F, Disturbance Moni-
speeds, conductor strength, and the purposes and toring, requires the use of recording devices for
duration of normal, seasonal, and emergency rat- disturbance analysis. On August 14, time record-
ings—makes the ratings themselves unclear, ers were frequently used but not synchronized to a
inconsistent, and unreliable across a region or time standard. Today, at a relatively modest cost,
between regions. This situation creates unneces- all digital fault recorders, digital event recorders,
sary and unacceptable uncertainties about the safe and power system disturbance recorders can and
carrying capacity of individual lines on the trans- should be time-stamped at the point of observa-
mission networks. Further, the appropriate use of tion using a Global Positioning System (GPS)
dynamic line ratings needs to be included in this synchronizing signal. (The GPS signals are syn-
review because adjusting a line’s rating according chronized with the atomic clock maintained in
to changes in ambient conditions may enable the Boulder, Colorado by the U.S. National Institute of
line to carry a larger load while still meeting safety Standards and Technology.) Recording and time-
requirements. synchronization equipment should be monitored
and calibrated to assure accuracy and reliability.
28. Require use of time-synchronized data It is also important that data from automation sys-
recorders.40 tems be retained at least for some minimum
In its requirements of February 10, 2004, NERC
period, so that if necessary it can be archived to
directed the regional councils to define within one enable adequate analysis of events of particular
year regional criteria for the application of syn- interest.
chronized recording devices in key power plants
and substations. 29. Evaluate and disseminate lessons
The Task Force supports the intent of this require- learned during system restoration.41
ment strongly, but it recommends a broader
In the requirements it issued on February 10, 2004,
approach:
NERC directed its Planning Committee to work
A. FERC and appropriate authorities in Canada with the Operating Committee, NPCC, ECAR, and
should require the use of data recorders syn- PJM to evaluate the black start and system restora-
chronized by signals from the Global Posi- tion performance following the outage of August
tioning System (GPS) on all categories of 14, and to report within one year the results of that
facilities whose data may be needed to evaluation, with recommendations for
162 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
improvement. Within six months of the Planning estimators, real-time contingency analyzers, and
Committee’s report, all regional councils are to other system monitoring tools.
have reevaluated their plans and procedures to
ensure an effective black start and restoration Further, NERC’s present operating policies do not
capability within their region. specify adequately criteria for identifying those
The Task Force supports these requirements
critical facilities within reliability coordinator and
strongly. In addition, the Task Force recommends control area footprints whose operating status
that NERC should require the Planning Commit- could affect the reliability of neighboring systems.
tee’s review to include consultation with appropri- This leads to uncertainty about which facilities
ate stakeholder organizations in all areas that were should be monitored by both the reliability coordi-
blacked out on August 14. nator for the region in which the facility is located
and by one or more neighboring reliability
The efforts to restore the power system and cus-
coordinators.
tomer service following the outage were generally
effective, considering the massive amount of load
lost and the large number of generators and trans- 31. Clarify that the transmission loading
mission lines that tripped. Fortunately, the resto- relief (TLR) process should not be used
ration was aided by the ability to energize in situations involving an actual viola-
transmission from neighboring systems, thereby tion of an Operating Security Limit.
speeding the recovery. Streamline the TLR process.43
Despite the apparent success of the restoration NERC should clarify that the TLR procedure is
effort, it is important to evaluate the results in often too slow for use in situations in which an
more detail to compare them with previous black- affected system is already in violation of an Oper-
out/restoration studies and determine opportuni- ating Security Limit. NERC should also evaluate
ties for improvement. Black start and restoration experience to date with the TLR procedure and
plans are often developed through study of simu- propose by September 1, 2004, ways to make it less
lated conditions. Robust testing of live systems is cumbersome.
difficult because of the risk of disturbing the sys- The reviews of control area and reliability coordi-
tem or interrupting customers. The August 14 nator transcripts from August 14 confirm that the
blackout provides a valuable opportunity to TLR process is cumbersome, perhaps unnecessar-
review actual events and experiences to learn how ily so, and not fast and predictable enough for use
to better prepare for system black start and restora- situations in which an Operating Security Limit is
tion in the future. That opportunity should not be close to or actually being violated. NERC should
lost. develop an alternative to TLRs that can be used
30. Clarify criteria for identification of quickly to address alert and emergency
operationally critical facilities, and conditions.
improve dissemination of updated
Group III. Physical and Cyber Security
information on unplanned outages.42
of North American Bulk Power Systems
NERC should work with the control areas and reli-
ability coordinators to clarify the criteria for iden- 32. Implement NERC IT standards.
tifying critical facilities whose operational status
The Task Force recommends that NERC standards
can affect the reliability of neighboring areas, and
related to physical and cyber security should be
to improve mechanisms for sharing information
understood as being included within the body of
about unplanned outages of such facilities in near
standards to be made mandatory and enforceable
real-time.
in Recommendation No. 1. Further:
The lack of accurate, near real-time information A. NERC should ensure that the industry has
about unplanned outages degraded the perfor- implemented its Urgent Action Standard 1200;
mance of state estimator and reliability assess- finalize, implement, and ensure membership
ment functions on August 14. NERC and the compliance with its Reliability Standard 1300
industry must improve the mechanisms for shar- for Cyber Security and take actions to better
ing outage information in the operating time hori- communicate and enforce these standards.
zon (e.g., 15 minutes or less), to ensure the B. CAs and RCs should implement existing and
accurate and timely sharing of outage data needed emerging NERC standards, develop and imple-
by real-time operating tools such as state ment best practices and policies for IT and
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 163
security management, and authenticate and that cover self-certification, self-assess-
authorize controls that address EMS automa- ment, and/or third-party audit.
tion system ownership and boundaries.
ã Work with federal, state, and provincial/terri-
Interviews and analyses conducted by the SWG torial jurisdictional departments and agen-
indicate that within some of the companies inter- cies to regularly update private and public
viewed there are potential opportunities for cyber sector standards, policies, and other
system compromise of EMS and their supporting guidance.
IT infrastructure. Indications of procedural and u CAs and RCs:
technical IT management vulnerabilities were
ã Implement existing and emerging NERC
observed in some facilities, such as unnecessary
standards.
software services not denied by default, loosely
controlled system access and perimeter control, ã Develop and implement best practices and
poor patch and configuration management, and policies for IT and security management
poor system security documentation. drawing from existing NERC and government
authorities’ best practices.45 These should
An analysis of the more prevalent policies and include, but not necessarily be limited to:
standards within the electricity sector revealed
1. Policies requiring that automation system
that there is existing and expanding guidance on
products be delivered and installed with
standards within the sector to perform IT and
unnecessary services deactivated in order
information security management.44 NERC issued
to improve “out-of-the-box security.”
a temporary standard (Urgent Action Standard
1200, Cyber Security) on August 13, 2003, and is 2. The creation of centralized system admin-
developing the formal Reliability Standard 1300 istration authority within each CA and RC
for Cyber Security. Both start the industry down to manage access and permissions for auto-
the correct path, but there is a need to communi- mation access (including vendor manage-
cate and enforce these standards by providing the ment backdoors, links to other automation
industry with recommended implementation systems, and administrative connections).
guidance. Implementation guidance regarding ã Authenticate and authorize controls that
these sector-wide standards is especially impor- address EMS automation system ownership
tant given that implementation procedures may and boundaries, and ensure access is granted
differ among CAs and RCs. only to users who have corresponding job
responsibilities.
In order to address the finding described above,
the Task Force recommends:
33. Develop and deploy IT management
u NERC:
procedures.
ã Ensure that the industry has implemented its
Urgent Action Standard 1200 and determine CAs’ and RCs’ IT and EMS support personnel
should develop procedures for the development,
if the guidance contained therein needs to be
testing, configuration, and implementation of tech-
strengthened or amended in the ongoing nology related to EMS automation systems and also
development of its Reliability Standard 1300 define and communicate information security and
for Cyber Security. performance requirements to vendors on a continu-
ã Finalize, implement, and ensure member- ing basis. Vendors should ensure that system
ship compliance of its Reliability Standard upgrades, service packs, and bug fixes are made
1300 for Cyber Security and take actions to available to grid operators in a timely manner.
better communicate and enforce these stan- Interviews and analyses conducted by the SWG
dards. These actions should include, but not indicate that, in some instances, there were
necessarily be limited to: ill-defined and/or undefined procedures for EMS
1. The provision of policy, process, and automation systems software and hardware devel-
implementation guidance to CAs and RCs; opment, testing, deployment, and backup. In addi-
and tion, there were specific instances of failures to
perform system upgrade, version control, mainte-
2. The establishment of mechanisms for com- nance, rollback, and patch management tasks.
pliance, audit, and enforcement. This may
include recommendations, guidance, or At one CA, these procedural vulnerabilities were
agreements between NERC, CAs and RCs compounded by inadequate, out-of-date, or non-
164 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
existing maintenance contracts with EMS vendors ineffective monitoring and control over EMS-
and contractors. This could lead to situations supporting IT infrastructure and overall IT net-
where grid operators could alter EMS components work health. In these cases, both grid operators
without vendor notification or authorization as and IT support personnel did not have situational
well as scenarios in which grid operators are not awareness of the health of the IT systems that pro-
aware of or choose not to implement vendor- vide grid information both globally and locally.
recommended patches and upgrades. This resulted in an inability to detect, assess,
respond to, and recover from IT system-related
cyber failures (failed hardware/software, mali-
34. Develop corporate-level IT security cious code, faulty configurations, etc.).
governance and strategies.
In order to address the finding described above,
CAs and RCs and other grid-related organizations
the Task Force recommends:
should have a planned and documented security
strategy, governance model, and architecture for u IT and EMS support personnel implement tech-
EMS automation systems. nical controls to detect, respond to, and recover
from system and network problems.
Interviews and analysis conducted by the SWG
indicate that in some organizations there is evi- u Grid operators, dispatchers, and IT and EMS
dence of an inadequate security policy, gover- support personnel be provided with the tools
nance model, strategy, or architecture for EMS and training to ensure that:
automation systems. This is especially apparent ã The health of IT systems is monitored and
with legacy EMS automation systems that were maintained.
originally designed to be stand-alone systems but
that are now interconnected with internal (corpo- ã These systems have the capability to be
rate) and external (vendors, Open Access Same repaired and restored quickly, with a mini-
Time Information Systems (OASIS), RCs, Internet, mum loss of time and access to global and
etc.) networks. It should be noted that in some of internal grid information.
the organizations interviewed this was not the ã Contingency and disaster recovery proce-
case and in fact they appeared to excel in the areas dures exist and can serve to temporarily sub-
of security policy, governance, strategy, and stitute for systems and communications
architecture. failures during times when EMS automation
In order to address the finding described above, system health is unknown or unreliable.
the Task Force recommends that CAs, RCs, and
ã Adequate verbal communication protocols
other grid-related organizations have a planned
and procedures exist between operators and
and documented security strategy, governance
IT and EMS support personnel so that opera-
model, and architecture for EMS automation sys-
tors are aware of any IT-related problems that
tems covering items such as network design, sys-
may be affecting their situational awareness
tem design, security devices, access and
of the power grid.
authentication controls, and integrity manage-
ment as well as backup, recovery, and contin-
gency mechanisms.
36. Initiate a U.S.-Canada risk manage-
ment study.
35. Implement controls to manage system
In cooperation with the electricity sector, federal
health, network monitoring, and inci- governments should strengthen and expand the
dent management. scope of the existing risk management initiatives
IT and EMS support personnel should implement by undertaking a bilateral (Canada-U.S.) study of
technical controls to detect, respond to, and the vulnerabilities of shared electricity infrastruc-
recover from system and network problems. Grid ture and cross border interdependencies. Common
operators, dispatchers, and IT and EMS support threat and vulnerability assessment methodologies
personnel should be provided the tools and train- should be also developed, based on the work
ing to ensure that the health of IT systems is moni- undertaken in the pilot phase of the current joint
tored and maintained. Canada-U.S. vulnerability assessment initiative,
and their use promoted by CAs and RCs. To coin-
Interviews and analysis conducted by the SWG cide with these initiatives, the electricity sector, in
indicate that in some organizations there was association with federal governments, should
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 165
develop policies and best practices for effective (Canada-U.S.) study of the vulnerabilities of
risk management and risk mitigation. shared electricity infrastructure and cross bor-
Effective risk management is a key element in der interdependencies. The study should
assuring the reliability of our critical infrastruc- encompass cyber, physical, and personnel
tures. It is widely recognized that the increased security processes and include mitigation and
reliance on IT by critical infrastructure sectors, best practices, identifying areas that would ben-
including the energy sector, has increased the efit from further standardization.
vulnerability of these systems to disruption via u Common threat and vulnerability assessment
cyber means. The breadth of the August 14, 2003, methodologies should be developed, based on
power outage illustrates the vulnerabilities and the work undertaken in the pilot phase of the
interdependencies inherent in our electricity current joint Canada-U.S. vulnerability assess-
infrastructure. ment initiative, and their use promoted by CAs
Canada and the United States, recognizing the and RCs.
importance of assessing the vulnerabilities of u The electricity sector, in association with fed-
shared energy systems, included a provision to eral governments, should develop policies and
address this issue in the Smart Border Declara- best practices for effective risk management and
tion,46 signed on December 12, 2001. Both coun- risk mitigation.
tries committed, pursuant to Action Item 21 of the
Declaration, to “conduct bi-national threat assess-
ments on trans-border infrastructure and identify 37. Improve IT forensic and diagnostic
necessary protection measures, and initiate capabilities.
assessments for transportation networks and other CAs and RCs should seek to improve internal
critical infrastructure.” These joint assessments forensic and diagnostic capabilities, ensure that IT
will serve to identify critical vulnerabilities, support personnel who support EMS automation
strengths and weaknesses while promoting the systems are familiar with the systems’ design and
sharing and transfer of knowledge and technology implementation, and make certain that IT support
to the energy sector for self-assessment purposes. personnel who support EMS automation systems
have are trained in using appropriate tools for
A team of Canadian and American technical diagnostic and forensic analysis and remediation.
experts, using methodology developed by the
Argonne National Laboratory in Chicago, Illinois, Interviews and analyses conducted by the SWG
began conducting the pilot phase of this work in indicate that, in some cases, IT support personnel
January 2004. The work involves a series of joint who are responsible for EMS automation systems
Canada-U.S. assessments of selected shared criti- are unable to perform forensic and diagnostic rou-
cal energy infrastructure along the Canada-U.S. tines on those systems. This appears to stem from
border, including the electrical transmission lines a lack of tools, documentation and technical skills.
and dams at Niagara Falls - Ontario and New York. It should be noted that some of the organizations
The pilot phase will be completed by March 31, interviewed excelled in this area but that overall
2004. performance was lacking.
The findings of the ESWG and SWG suggest that In order to address the finding described above,
among the companies directly involved in the the Task Force recommends:
power outage, vulnerabilities and interdependen- u CAs and RCs seek to improve internal forensic
cies of the electric system were not well under- and diagnostic capabilities as well as strengthen
stood and thus effective risk management was coordination with external EMS vendors and
inadequate. In some cases, risk assessments did contractors who can assist in servicing EMS
not exist or were inadequate to support risk man- automation systems;
agement and risk mitigation plans. u CAs and RCs ensure that IT support personnel
In order to address these findings, the Task Force who support EMS automation systems are
recommends: familiar with the systems’ design and imple-
u In cooperation with the electricity sector, fed- mentation; and
eral governments should strengthen and u CAs and RCs ensure that IT support personnel
expand the scope of the existing initiatives who support EMS automation systems have
described above by undertaking a bilateral access to and are trained in using appropriate
166 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
tools for diagnostic and forensic analysis and 40. Control access to operationally sensi-
remediation. tive equipment.
RCs and CAs should implement stringent policies
38. Assess IT risk and vulnerability at and procedures to control access to sensitive equip-
scheduled intervals. ment and/or work areas.
IT and EMS support personnel should perform reg- Interviews conducted by the SWG indicate that
ular risk and vulnerability assessment activities at some CAs and RCs operationally sensitive
for automation systems (including EMS applica- computer equipment was accessible to non-
tions and underlying operating systems) to identify essential personnel. Although most of these non-
weaknesses, high-risk areas, and mitigating actions essential personnel were escorted through sensi-
such as improvements in policy, procedure, and tive areas, it was determined that this procedure
technology. was not always enforced as a matter of everyday
Interviews and analysis conducted by the SWG operations.
indicate that in some instances risk and vulnera- In order to address the finding described above,
bility management were not being performed on the Task Force recommends:
EMS automation systems and their IT supporting
infrastructure. To some CAs, EMS automation sys- u That RCs and CAs develop policies and proce-
tems were considered “black box”47 technologies; dures to control access to sensitive equipment
and this categorization removed them from the list and/or work areas to ensure that:
of systems identified for risk and vulnerability ã Access is strictly limited to employees or con-
assessment. tractors who utilize said equipment as part of
their job responsibilities.
ã Access for other staff who need access to sen-
39. Develop capability to detect wireless
sitive areas and/or equipment but are not
and remote wireline intrusion and directly involved in their operation (such as
surveillance. cleaning staff and other administrative per-
Both the private and public sector should promote sonnel) is strictly controlled (via escort) and
the development of the capability of all CAs and monitored.
RCs to reasonably detect intrusion and surveil-
lance of wireless and remote wireline access points
41. NERC should provide guidance on
and transmissions. CAs and RCs should also con-
duct periodic reviews to ensure that their user base employee background checks.
is in compliance with existing wireless and remote NERC should provide guidance on the implementa-
wireline access rules and policies. tion of its recommended standards on background
Interviews conducted by the SWG indicate that checks, and CAs and RCs should review their poli-
cies regarding background checks to ensure they
most of the organizations interviewed had some
are adequate.
type of wireless and remote wireline intrusion and
surveillance detection protocol as a standard secu- Interviews conducted with sector participants
rity policy; however, there is a need to improve revealed instances in which certain company con-
and strengthen current capabilities regarding tract personnel did not have to undergo back-
wireless and remote wireline intrusion and sur- ground check(s) as stringent as those performed
veillance detection. The successful detection and on regular employees of a CA or RC. NERC Urgent
monitoring of wireless and remote wireline access Action Standard Section 1207 Paragraph 2.3 spec-
points and transmissions are critical to securing ifies steps to remediate sector weaknesses in this
grid operations from a cyber security perspective. area but there is a need to communicate and
enforce this standard by providing the industry
There is also evidence that although many of the
with recommended implementation guidance,
organizations interviewed had strict policies
which may differ among CAs and RCs.
against allowing wireless network access, periodic
reviews to ensure compliance with these policies In order to address the finding described above,
were not undertaken. the Task Force recommends:
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 167
u NERC provide guidance on the implementation Interviews and analysis conducted by the SWG
of its recommended standards on background further indicate an absence of coherent and effec-
checks, especially as they relate to the screening tive mechanisms for the private sector to share
of contracted and sub-contracted personnel. information related to critical infrastructure with
government. There was also a lack of confidence
u CAs and RCs review their policies regarding
on the part of private sector infrastructure owners
background checks to ensure they are adequate
and grid operators that information shared with
before allowing sub-contractor personnel to
governments could be protected from disclosure
access their facilities.
under Canada’s Access to Information Act (ATIA)
and the U.S. Freedom of Information Act (FOIA).
42. Confirm NERC ES-ISAC as the central On the U.S. side of the border, however, the immi-
point for sharing security information nent implementation of the Critical Infrastructure
and analysis. Information (CII) Act of 2002 should mitigate
almost all industry concerns about FOIA disclo-
The NERC ES-ISAC should be confirmed as the sure. In Canada, Public Safety and Emergency Pre-
central electricity sector point of contact for secu- paredness Canada relies on a range of mechanisms
rity incident reporting and analysis. Policies and to protect the sensitive information related to criti-
protocols for cyber and physical incident reporting
cal infrastructure that it receives from its private
should be further developed including a mecha-
nism for monitoring compliance. There also should sector stakeholders, including the exemptions for
be uniform standards for the reporting and sharing third party information that currently exist in the
of physical and cyber security incident information ATIA and other instruments. At the same time,
across both the private and public sectors. Public Safety and Emergency Preparedness Can-
ada is reviewing options for stronger protection of
There are currently both private and public sector CI information, including potential changes in
information sharing and analysis initiatives in legislation.
place to address the reporting of physical and
cyber security incidents within the electricity sec- In order to address the finding described above,
tor. In the private sector, NERC operates an Elec- the Task Force recommends:
tricity Sector Information Sharing and Analysis u Confirmation of the NERC ES-ISAC as the cen-
Center (ES-ISAC) specifically to address this tral electricity sector point of contact for secu-
issue. On behalf of the U.S. Government, the rity incident reporting and analysis.
Department of Homeland Security (DHS) operates
the Information Analysis and Infrastructure Pro- u Further development of NERC policies and pro-
tection (IAIP) Directorate to collect, process, and tocols for cyber and physical incident reporting
act upon information on possible cyber and physi- including a mechanism for monitoring
cal security threats and vulnerabilities. In Canada, compliance.
Public Safety and Emergency Preparedness Can-
u The establishment of uniform standards for the
ada has a 24/7 operations center for the reporting
reporting of physical and cyber security inci-
of incidents involving or impacting critical infra-
dents to internal corporate security, private sec-
structure. As well, both in Canada and the U.S.,
tor sector-specific information sharing and
incidents of a criminal nature can be reported to
analysis bodies (including ISACs), law enforce-
law enforcement authorities of jurisdiction.
ment, and government agencies.
Despite these private and public physical and u The further development of new mechanisms
cyber security information sharing and analysis and the promulgation of existing48 Canadian
initiatives, an analysis of policies and procedures and U.S. mechanisms to facilitate the sharing of
within the electricity sector reveals that reporting electricity sector threat and vulnerability infor-
of security incidents to internal corporate secu- mation across governments as well as between
rity, law enforcement, or government agencies the private sector and governments.
was uneven across the sector. The fact that these
existing channels for incident reporting—whether u Federal, state, and provincial/territorial govern-
security- or electricity systems-related—are cur- ments work to further develop and promulgate
rently underutilized is an operating deficiency measures and procedures that protect critical,
which could hamper the industry’s ability to but sensitive, critical infrastructure-related
address future problems in the electricity sector. information from disclosure.
168 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
43. Establish clear authority for physical these intelligence collection activities be reported
and cyber security. to government authorities.
The task force recommends that corporations In order to address the findings described above,
establish clear authority and ownership for the Task Force recommends:
physical and cyber security. This authority
u The private and public sectors jointly develop
should have the ability to influence
corporate decision-making and the authority
and implement security procedures and aware-
to make physical and cyber security-related ness training in order to mitigate disclosure of
decisions. information not suitable for the public domain
and/or removal of previously available informa-
Interviews and analysis conducted by the SWG tion in the public domain (web sites, message
indicate that some power entities did not imple- boards, industry publications, etc.).
ment best practices when organizing their security
u The private and public sector jointly develop
staff. It was noted at several entities that the Infor-
and implement security procedures and aware-
mation System (IS) security staff reported to IT
ness training in order to mitigate or prevent dis-
support personnel such as the Chief Information
closure of information by the practices of
Officer (CIO).
elicitation.
Best practices across the IT industry, including u The private and public sector jointly develop
most large automated businesses, indicate that the and implement security procedures and aware-
best way to balance security requirements prop- ness training in order to mitigate, prevent, and
erly with the IT and operational requirements of a detect incidents of surveillance.
company is to place security at a comparable level
within the organizational structure. By allowing u Where no mechanism currently exists, the pri-
the security staff a certain level of autonomy, man- vate and public sector jointly establish a secure
agement can properly balance the associated risks reporting chain and protocol for use of the infor-
and operational requirements of the facility. mation for suspected and known attempts and
incidents of elicitation and surveillance.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 169
these recommendations, which are presented The August 2003 power outage demonstrated that
below. the CNSC’s Emergency Operations Center, IT, and
communications equipment are vulnerable if
45. The Task Force recommends that the there is a loss of electricity to the Ottawa area.
Canadian Nuclear Safety Commission
Endnotes
request Ontario Power Generation and 1 In fairness, it must be noted that reliability organizations in
Bruce Power to review operating pro- some areas have worked diligently to implement recommen-
cedures and operator training associ- dations from earlier blackouts. According to the Initial Report
ated with the use of adjuster rods. by the New York State Department of Public Service on the
August 14, 2003 Blackout, New York entities implemented all
OPG and Bruce Power should review their operat- 100 of the recommendations issued after the New York City
ing procedures to see whether alternative proce- blackout of 1977.
dures could be put in place to carry out or reduce 2 The need for a systematic recommitment to reliability by
the number of system checks required before plac- all affected organizations was supported in various ways by
ing the adjuster rods into automatic mode. This many commenters on the Interim Report, including Anthony
review should include an assessment of any regula- J. Alexander, FirstEnergy; David Barrie, Hydro One Networks,
Inc.; Joseph P. Carson, P.E.; Harrison Clark; F. J. Delea, J.A.
tory constraints placed on the use of the adjuster Casazza, G.C. Loehr, and R. M. Malizewski, Power Engineers
rods, to ensure that risks are being appropriately Seeking Truth; Ajay Garg and Michael Penstone, Hydro One
managed. Networks, Inc.; and Raymond K. Kershaw, International
Transmission Company.
Current operating procedures require independ- 3 See supporting comments expressed by Anthony J. Alex-
ent checks of a reactor’s systems by the reactor ander, FirstEnergy; Deepak Divan, SoftSwitching Technol-
operator and the control room supervisor before ogies; Pierre Guimond, Canadian Nuclear Association; Hans
the reactor can be put in automatic mode to allow Konow, Canadian Electricity Association; Michael Penstone,
the reactors to operate at 60% power levels. Alter- Hydro One Networks, Inc.; and James K. Robinson, PPL.
native procedures to allow reactors to run at 60% 4 See “The Economic Impacts of the August 2003 Blackout,”
of power while waiting for the grid to be Electric Consumers Resource Council (ELCON), February 2,
2004.
re-established may reduce other risks to the health 5 The need for action to make standards enforceable was
and safety of Ontarians that arise from the loss of a supported by many commenters, including David Barrie,
key source of electricity. CNSC oversight and Hydro One Networks, Inc.; Carl Burrell, IMO Ontario; David
approval of any changes to operating procedures Cook, North American Electric Reliability Council; Deepak
would ensure that health and safety, security, or Divan, SoftSwitching Technologies; Charles J. Durkin, North-
the environment are not compromised. The CNSC east Power Coordinating Council; David Goffin, Canadian
Chemical Producers’ Association; Raymond K. Kershaw,
would assess the outcome of the proposed review International Transmission Company; Hans Konow, Cana-
to ensure that health and safety, security, and the dian Electricity Association; Barry Lawson, National Rural
environment would not be compromised as a Electric Cooperative Association; William J. Museler, New
result of any proposed action. York Independent System Operator; Eric B. Stephens, Ohio
Consumers’ Counsel; Gordon Van Welie, ISO New England,
Inc.; and C. Dortch Wright, on behalf of James McGreevey,
46. The Task Force recommends that the Governor of New Jersey.
Canadian Nuclear Safety Commission 6 This recommendation was suggested by some members of
purchase and install backup genera- the Electric System Working Group.
7 The need to evaluate and where appropriate strengthen the
tion equipment.
institutional framework for reliability management was sup-
In order to ensure that the CNSC’s Emergency ported in various respects by many commenters, including
Operations Center (EOC) is available and fully Anthony J. Alexander, FirstEnergy Corporation; David Barrie,
functional during an emergency situation requiring Hydro One Networks, Inc.; Chris Booth, Experienced Consul-
CNSC response, whether the emergency is tants LLC; Carl Burrell, IMO Ontario;Linda Campbell, Florida
nuclear-related or otherwise, and that staff needed Reliability Coordinating Council; Linda Church Ciocci,
National Hydropower Association; David Cook, NERC; F.J.
to respond to the emergency can be accommodated Delea, J.A. Casazza, G.C. Loehr, and R.M. Malizewski, Power
safely, the CNSC should have backup electrical Engineers Seeking Truth; Charles J. Durkin, Northeast Power
generation equipment of sufficient capacity to pro- Coordinating Council; Ajay Garg and Michael Penstone,
vide power to the EOC, telecommunications and Hydro One Networks, Inc.; Michael W. Golay, Massachusetts
Information Technology (IT) systems and accom- Institute of Technology; Leonard S. Hyman, Private Sector
modations for the CNSC staff needed to respond to Advisors, Inc; Marija Ilic, Carnegie Mellon University; Jack
an emergency. Kerr, Dominion Virginia Power; Raymond K. Kershaw,
170 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
International Transmission Company; Paul Kleindorfer, Uni- 18 This subject was addressed in detail in comments by David
versity of Pennsylvania; Michael Kormos, PJM Interconnec- Cook, North American Electric Reliability Council; and in
tion; Bill Mittelstadt, Bonneville Power Administration; part by comments by Anthony J. Alexander, FirstEnergy Cor-
William J. Museler, New York Independent System Operator; poration; Ajay Garg, Hydro One Networks, Inc.; George
James K. Robinson, PPL; Eric B. Stephens, Ohio Consumers’ Katsuras, IMO Ontario; and Vickie Van Zandt, Bonneville
Counsel; John Synesiou, IMS Corporation; Gordon Van Power Administration.
Welie, ISO New England; Vickie Van Zandt, Bonneville 19 U.S. Federal Energy Regulatory Commission, 105 FERC ¶
Power Administration; and C. Dortch Wright, on behalf of 61,372, December 24, 2003.
James McGreevey, Governor of New Jersey. 20 See ECAR website,
8 Several commenters noted the importance of clarifying
http://www.ecar.org/documents/document%201_6-98.pdf.
that prudently incurred reliability expenses and investments 21 See NERC website, http://www.nerc.com/standards/.
will be recoverable through regulator-approved rates. These
commenters include Anthony J. Alexander, FirstEnergy Cor- 22 The need to ensure better maintenance of required electri-
poration; Deepak Divan, SoftSwitching Technologies; Ste- cal clearances in transmission right of way areas was empha-
phen Fairfax, MTechnology, Inc.; Michael W. Golay, sized by several commenters, including Richard E. Abbott,
Massachusetts Institute of Technology; Pierre Guimond, arborist; Anthony J. Alexander, FirstEnergy Corporation;
Canadian Nuclear Association; Raymond K. Kershaw, Inter- David Barrie, Hydro One Networks, Inc.; David Cook, North
national Transmission Company; Paul R. Kleindorfer, Uni- American Electric Reliability Council; Ajay Garg and Michael
versity of Pennsylvania: Hans Konow, Canadian Electricity Penstone, Hydro One Networks, Inc.; Tadashi Mano, Tokyo
Association; Barry Lawson, National Rural Electric Coopera- Electric Power Company; Eric B. Stephens, Ohio Consumers’
tive Association; and Michael Penstone, Hydro One Net- Counsel; Vickie Van Zandt, Bonneville Power Administra-
works, Inc. tion; and Donald Wightman, Utility Workers Union of
9 The concept of an ongoing NERC process to track the America.
implementation of existing and subsequent recommenda- 23 Utility Vegetation Management Final Report, CN Utility
tions was initated by NERC and broadened by members of the Consulting, LLC, March 2004, commissioned by the U.S. Fed-
Electric System Working Group. See comments by David eral Energy Regulatory Commission to support the investiga-
Cook, North American Electric Reliability Council. tion of the August 14, 2003 blackout.
10 This recommendation was suggested by NERC and sup- 24 The need to strengthen and verify compliance with NERC
ported by members of the Electric System Working Group. standards was noted by several commenters. See comments
11 See comments by Jack Kerr, Dominion Virginia Power, and by David Barrie, Hydro One Networks, Inc.; Carl Burrell, IMO
Margie Phillips, Pennsylvania Services Integration Ontario; David Cook, North American Electric Reliability
Consortium. Council; and Eric B. Stephens, Ohio Consumers’ Counsel.
12 The concept of a “reliability impact consideration” was 25 The need to verify application of NERC standards via
suggested by NERC and supported by the Electric System readiness audits—before adverse incidents occur—was noted
Working Group. by several commenters. See comments by David Barrie,
13 The suggestion that EIA should become a source of reliabil- Hydro One Networks, Inc.; David Cook, North American Elec-
tric Reliability Council; Barry Lawson, National Rural Electric
ity data and information came from a member of the Electric Cooperative Association; Bill Mittelstadt, Bonneville Power
System Working Group. Administration; and Eric B. Stephens, Ohio Consumers’
14 Several commenters raised the question of whether there Counsel.
was a linkage between the emergence of competition (or 26 The need to improve the training and certification require-
increased wholesale electricity trade) in electricity markets ments for control room management and staff drew many
and the August 14 blackout. See comments by Anthony J. comments. See comments by David Cook, North American
Alexander, FirstEnergy Corporation; F.J. Delea, J.A. Casazza, Electric Reliability Council; F.J. Delea, J.A. Casazza, G.C.
G.C. Loehr, and R.M. Malizewski, Power Engineers Seeking Loehr, and R.M. Malizewski, Power Engineers Seeking Truth;
Truth; Ajay Garg and Michael Penstone, Hydro One Net- Victoria Doumtchenko, MPR Associates; Pat Duran, IMO
works, Inc.; Brian O’Keefe, Canadian Union of Public Ontario; Ajay Garg and Michael Penstone, Hydro One Net-
Employees; Les Pereira; and John Wilson. works, Inc.; George Katsuras, IMO Ontario; Jack Kerr, Domin-
15 NIMBY: “Not In My Back Yard.” ion Virginia Power; Tim Kucey, National Energy Board,
16 Several commenters either suggested that government Canada; Stephen Lee, Electric Power Research Institute; Steve
agencies should expand their research in reliability-related Leovy, personal comment; Ed Schwerdt, Northeast Power
topics, or emphasized the need for such R&D more generally. Coordinating Council; Tapani O. Seppa, The Valley Group,
See comments by Deepak Divan, SoftSwitching Technol- Inc.; Eric B. Stephens, Ohio Consumers’ Counsel; Vickie Van
ogies; Marija Ilic, Carnegie Mellon University; Hans Konow, Zandt, Bonneville Power Company; Don Watkins, Bonneville
Canadian Electricity Association; Stephen Lee, Electric Power Administration; and Donald Wightman, Utility
Power Research Institute; James K. Robinson, PPL; John Workers Union of America.
Synesiou, IMS Corporation; and C. Dortch Wright on behalf of 27 This reliance, and the risk of an undue dependence, is
Governor James McGreevey of New Jersey. often unrecognized in the industry.
17 The concept of a standing framework for grid-related 28 Many parties called for clearer statement of the roles,
investigations was initiated by members of the Electric Sys- responsibilities, and authorities of control areas and reliabil-
tem Working Group, after noting that the U.S. National Aero- ity coordinators, particularly in emergency situations. See
nautics and Space Administration (NASA) had created a comments by Anthony J. Alexander, FirstEnergy Corporation;
similar arrangement after the Challenger explosion in 1986. Chris Booth, Experienced Consultants LLC; Michael
This framework was put to use immediately after the loss of Calimano, New York ISO; Linda Campbell, Florida Reliability
the shuttle Columbia in 2003. Coordinating Council; David Cook, North American Electric
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 171
Reliability Council; F.J. Delea, J.A. Casazza, G.C. Loehr, and Delea, J.A. Casazza, G.C. Loehr, and R.M. Malizewski, Power
R.M. Malizewski, Power Engineers Seeking Truth; Mark Engineers Seeking Truth; Stephen Fairfax, MTechnology,
Fidrych, Western Area Power Authority; Ajay Garg and Inc.; Ajay Garg and Michael Penstone, Hydro One Networks,
Michael Penstone, Hydro One Networks, Inc.; Carl Hauser, Inc.; Shinichi Imai and Toshihiko Furuya, Tokyo Electric
Washington State University; Stephen Kellat; Jack Kerr, Power Company; Marija Ilic, Carnegie Mellon University;
Dominion Virginia Power; Raymond K. Kershaw, Interna- Frank Macedo, Hydro One Networks, Inc.; and Tom
tional Transmission Company; Michael Kormos, PJM Inter- Wiedman, Consolidated Edison. Several commenters
connection; William J. Museler, New York Independent addressed issues related to the production of reactive power
System Operator; Tapani O. Seppa, The Valley Group, Inc.; by producers of power for sale in wholesale markets. See com-
John Synesiou, IMS Corporation; Gordon Van Welie, ISO ments by Anthony J. Alexander, FirstEnergy Corporation;
New England, Inc.; Vickie Van Zandt, Bonneville Power K.K. Das, PowerGrid Corporation of India, Limited; F.J. Delea,
Administration; Kim Warren, IMO Ontario; and Tom J.A. Casazza, G.C. Loehr, and R.M. Malizewski, Power Engi-
Wiedman, Consolidated Edison. Members of the Electric Sys- neers Seeking Truth; Stephen Fairfax, MTechnology, Inc.;
tem Working Group initiated the concept of defining an and Carson Taylor, Bonneville Power Administration.
“alert” status, between “normal” and “emergency,” and asso- 32 See pages 107-108.
ciated roles, responsibilities, and authorities. 33 U.S. Federal Energy Regulatory Commission, 105 FERC ¶
29 The need to make better use of system protection measures
61,372, December 24, 2003.
received substantial comment, including comments by James 34 The need to improve the quality of system modeling data
L. Blasiak, International Transmission Company; David Cook,
and data exchange practices received extensive comment. See
North American Electric Reliability Council; Charles J.
comments from Michael Calimano, New York ISO; David
Durkin, Northeast Power Coordinating Council; F.J. Delea,
Cook, North American Electric Reliability Council; Robert
J.A. Casazza, G.C. Loehr, and R.M. Malizewski, Power Engi-
Cummings, North American Electric Reliability Council; F.J.
neers Seeking Truth; Ajay Garg and Michael Penstone, Hydro
Delea, J.A. Casazza, G.C. Loehr, and R.M. Malizewski, Power
One Networks, Inc.; Gurgen and Spartak Hakobyan, personal
Engineers Seeking Truth; Mark Fidrych, Western Area Power
study; Marija Ilic, Carnegie Mellon University; Shinichi Imai,
Administration; Jack Kerr, Dominion Virginia Power; Ray-
Tokyo Electric Power Company; Jack Kerr, Dominion Virginia
mond K. Kershaw, International Transmission Company;
Power; Stephen Lee, Electric Power Research Institute; Ed
Frank Macedo, Hydro One Networks, Inc.; Vickie Van Zandt,
Schwerdt, Northeast Power Coordinating Council; Robert
Bonneville Power Administration; Don Watkins, Bonneville
Stewart, PG&E; Philip Tatro, National Grid Company; Carson
Power Administration; and David Zwergel, Midwest ISO.
Taylor, Bonneville Power Administration; Vickie Van Zandt,
35 Several commenters addressed the subject of NERC’s stan-
Bonneville Power Company; Don Watkins, Bonneville Power
Administration; and Tom Wiedman, Consolidated Edison. dards in various respects, including Anthony J. Alexander,
30 The subject of developing and adopting better real-time FirstEnergy Corporation; Carl Burrell, IMO Ontario; David
Cook, North American Electric Reliability Council; F.J. Delea,
tools for control room operators and reliability coordinators
J.A. Casazza, G.C. Loehr, and R.M. Malizewski, Power Engi-
drew many comments, including those by Anthony J. Alexan-
neers Seeking Truth; Charles J. Durkin, Northeast Power
der, FirstEnergy Corporation; Eric Allen, New York ISO; Chris
Coordinating Council; Ajay Garg and Michael Penstone,
Booth, Experienced Consultants, LLC; Mike Calimano, New
Hydro One Networks, Inc.; Jack Kerr, Dominion Virginia
York ISO; Claudio Canizares, University of Waterloo
Power; James K. Robinson, PPL; Mayer Sasson, New York
(Ontario); David Cook, North American Electric Reliability
State Reliability Council; and Kim Warren, IMO Ontario.
Council; Deepak Divan, SoftSwitching Technologies Victoria
36 See Initial Report by the New York State Department of Pub-
Doumtchenko, MPR Associates; Pat Duran, IMO Ontario; Bill
Eggertson, Canadian Association for Renewable Energies; lic Service on the August 14, 2003 Blackout (2004), and com-
Ajay Garg and Michael Penstone, Hydro One Networks, Inc.; ments by Mayer Sasson, New York State Reliability Council.
Jack Kerr, Dominion Virginia Power; Raymond K. Kershaw, 37 F.J. Delea, J.A. Casazza, G.C. Loehr, and R.M. Malizewski,
International Transmission Company; Michael Kormos, PJM “The Need for Strong Planning and Operating Criteria to
Interconnection; Tim Kucey, National Energy Board, Canada; Assure a Reliable Bulk Power Supply System,” January 29,
Steve Lapp, Lapp Renewables; Stephen Lee, Electric Power 2004.
Research Institute; Steve Leovy; Tom Levy; Peter Love, Cana- 38 The need to tighten communications protocols and
dian Energy Efficiency Alliance; Frank Macedo, Hydro One improve communications systems was cited by several
Networks, Inc.; Bill Mittelstadt, Bonneville Power Adminis- commenters. See comments by Anthony J. Alexander,
tration; Fiona Oliver, Canadian Energy Efficiency Alliance; FirstEnergy Corporation; David Barrie, Hydro One Networks,
Peter Ormund, Mohawk College; Don Ross, Prince Edward Inc.; Carl Burrell, IMO Ontario; Michael Calimano, New York
Island Wind Co-op Limited; James K. Robinson, PPL; Robert ISO; David Cook, North American Electric Reliability Coun-
Stewart, PG&E; John Synesiou, IMS Corporation; Gordon Van cil; Mark Fidrych, Western Area Power Administration; Ajay
Welie, ISO New England, Inc.; Vickie Van Zandt, Bonneville Garg and Michael Penstone, Hydro One Networks, Inc.; Jack
Power Administration; Don Watkins, Bonneville Power Kerr, Dominion Virginia Power; William Museler, New York
Administration; Chris Winter, Conservation Council of ISO; John Synesiou, IMS Corporation; Vickie Van Zandt,
Ontario; David Zwergel, Midwest ISO. The concept of requir- Bonneville Power Administration; Don Watkins, Bonneville
ing annual testing and certification of operators’ EMS and Power Administration; Tom Wiedman, Consolidated Edison.
SCADA systems was initiated by a member of the Electric 39 See comments by Tapani O. Seppa, The Valley Group, Inc.
System Working Group. Also, see comments by John
Synesiou, IMS Corporation. 40 Several commenters noted the need for more systematic
31 The need to strengthen reactive power and voltage control use of time-synchronized data recorders. In particular, see
practices was the subject of several comments. See comments David Cook, North American Electric Reliability Council;
by Claudio Canizares, University of Waterloo (Ontario); Ajay Garg and Michael Penstone, Hydro One Networks, Inc.;
David Cook, North American Electric Reliability Council; F.J. and Robert Stewart, PG&E.
172 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
41 The importance of learning from the system restoration ea.doe.gov/pdfs/21stepsbooklet.pdf (“21 Steps to Improve
experience associated with the August 14 blackout was Cyber Security of SCADA Networks”).
stressed by Linda Church Ciocci, National Hydropower Asso- 45 For example: “21 Steps to Improve Cyber Security of
ciation; David Cook, North American Electric Reliability SCADA Networks,” http://www.ea.doe.gov/pdfs/
Council; Frank Delea; Bill Eggertson, Canadian Association 21stepsbooklet.pdf.
for Renewable Energies; Stephen Lee, Electric Power 46 Canadian
Research Institute; and Kim Warren, IMO Ontario. reference: http://www.dfait-maeci.gc.ca/
42 The need to clarify the criteria for identifying critical facili- anti-terrorism/actionplan-en.asp; U.S. reference: http://www.
whitehouse.gov/news/releases/2001/12/20011212-6.html.
ties and improving dissemination of updated information 47 A “black box” technology is any device, sometimes highly
about unplanned outages was cited by Anthony J. Alexander,
FirstEnergy Corporation; and Raymond K. Kershaw, Interna- important, whose workings are not understood by or accessi-
tional Transmission Company. ble to its user.
43 The need to streamline the TLR process and limit the use of 48 DOE Form 417 is an example of an existing, but
it to non-urgent situations was discussed by several underutilized, private/public sector information sharing
commenters, including Anthony J. Alexander, FirstEnergy mechanism.
Corporation; Carl Burrell, IMO Ontario; Jack Kerr, Dominion
Virginia Power; Raymond K. Kershaw, International Trans-
mission Company; and Ed Schwerdt, Northeast Power Coor-
dinating Council.
44 NERC Standards at www.nerc.com (Urgent Action Stan-
dard 1200, Cyber Security, Reliability Standard 1300, Cyber
Security) and Joint DOE/PCIB standards guidance at www.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 173
Appendix A
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 175
Security Working Group Protection, Homeland Security Council (U.S.
Government)
Co-Chairs
William J.S. Elliott, Assistant Secretary to the James McMahon, Senior Advisor (New York)
Cabinet, Security and Intelligence, Privy Council Colonel Michael C. McDaniel, Assistant Adju-
Office (Government of Canada) tant General for Homeland Security (Michigan)
Robert Liscouski, Assistant Secretary for Infra- John Overly, Executive Director, Division of
structure, Department of Homeland Security Homeland Security (Ohio)
(U.S. Government)
Andy Purdy, Deputy Director, National Cyber
Canadian Members Security Division, Information Analysis and
Curt Allen, Director Corporate Security, Manage- Infrastructure Protection Directorate, DHS
ment Board Secretariat, Office of the Corporate
Chief Information Officer, Government of Kerry L. Sleeper, Commissioner, Public Safety
Ontario (Vermont)
Gary Anderson, Chief, Counter-Intelligence- Arthur Stephens, Deputy Secretary for Informa-
Global, Canadian Security Intelligence Service tion Technology, Office of Administration
(Government of Canada) (Pennsylvania)
Michael Devancy, Deputy Chief, Information Steve Schmidt, Section Chief, Special Technol-
Technology Security, Communications Security ogies and Applications, FBI
Establishment (Government of Canada) Richard Swensen, Under Secretary, Office of
James Harlick, Assistant Deputy Minister, Public Public Safety and Homeland Security
Safety and Emergency Preparedness Canada (Massachusetts)
(Government of Canada) Simon Szykman, Senior Policy Analyst, Office of
Peter MacAulay, Officer in Charge of Technolog- Science and Technology Policy (U.S.
ical Crime Branch, Royal Canadian Mounted Government)
Police (Government of Canada)
Ralph Mahar, Chief, Technical Operations, Sci- Nuclear Working Group
entific and Technical Services, Canadian Secu- Co-Chairs
rity Intelligence Service (Government of Canada) Nils Diaz, Chairman, Nuclear Regulatory Com-
Dr. James Young, Commissioner of Public Secu- mission (U.S. Government)
rity, Ontario Ministry of Public Safety and Secu- Linda J. Keen, President and Chief Executive
rity (Ontario) Officer, Canadian Nuclear Safety Commission
U.S. Members (Government of Canada)
Sid Casperson, Director, Office of Counter Ter- Canadian Members
rorism (New Jersey) James Blyth, Director General, Directorate of
Vincent DeRosa, Deputy Commissioner, Director Power Regulation, Canadian Nuclear Safety
of Homeland Security, Department of Public Commission (Government of Canada)
Safety (Connecticut) Duncan Hawthorne, Chief Executive Officer,
Harold M. Hendershot, Acting Section Chief, Bruce Power (Ontario)
Computer Intrusion Section, Federal Bureau of Robert Morrison, Senior Advisor to the Deputy
Investigation (U.S. Government) Minister, Natural Resources Canada (Govern-
Kevin Kolevar, Chief of Staff to the Deputy Sec- ment of Canada)
retary of Energy, Department of Energy (U.S. Ken Pereira, Vice President, Operations Branch,
Government) Canadian Nuclear Safety Commission (Govern-
Paul Kurtz, Special Assistant to the President ment of Canada)
and Senior Director for Critical Infrastructure
176 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
U.S. Members Dr. G. Ivan Moldonado, Associate Professor,
David J. Allard, CHP, Director, Bureau of Radia- Mechanical, Industrial and Nuclear Engineering;
tion Protection Department of Environmental University of Cincinnati (Ohio)
Protection (Pennsylvania) David O’Brien, Commissioner, Department of
Frederick F. Butler, Commissioner, New Jersey Public Service (Vermont)
Board of Public Utilities (New Jersey) David O’Connor, Commissioner, Division of
Sam J. Collins, Deputy Executive Director for Energy Resources, Office of Consumer Affairs
Reactor Programs, Nuclear Regulatory and Business Regulation (Massachusetts)
Commission Gene Whitney, Policy Analyst, National Science
Paul Eddy, Power Systems Operations Specialist, and Technology Policy, Executive Office of the
Public Service Commission (New York) President (U.S. Government)
J. Peter Lark, Chairman, Public Service Commis- Edward Wilds, Bureau of Air Management,
sion (Michigan) Department of Environmental Protection
(Connecticut)
William D. Magwood IV, Director, Office of
Nuclear Energy, Science and Technology,
Department of Energy (U.S. Government)
This report reflects tireless efforts by hundreds of individuals not identified by name above. They include
electrical engineers, information technology experts, and other specialists from across the North American
electricity industry, the academic world, regulatory agencies in the U.S. and Canada, the U.S. Department of
Energy and its national laboratories, the U.S. Department of Homeland Security, the U.S. Federal Bureau of
Investigation, Natural Resources Canada, the Royal Canadian Mounted Police, the Bonneville Power
Administration, the Western Area Power Administration, the Tennessee Valley Authority, the North Ameri-
can Electric Reliability Council, PJM Interconnection, Inc., Ontario’s Independent Market Operator, and
many other organizations. The members of the U.S.-Canada Power System Outage Task Force thank these
individuals, and congratulate them for their dedication and professionalism.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 177
Appendix B
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 179
cyber and security experts were established to electric companies affected by the blackout, and
investigate the causes of the outage. The electric data from the nuclear and security associated enti-
system investigative team was comprised of indi- ties. The investigative teams also interviewed
viduals from several U.S. federal agencies, the appropriate individuals to learn what they saw
U.S. Department of Energy’s national laboratories, and knew at key points in the evolution of the out-
Canadian electric industry, Canada’s National age, what actions they took, and with what pur-
Energy Board, staff from the North American Elec- pose. In recognition of the sensitivity of this
tric Reliability Council (NERC), and the U.S. elec- information, Working Group members and mem-
tricity industry. The overall investigative team bers of the teams signed agreements affirming that
was divided into several analytic groups with spe- they would maintain the confidentiality of data
cific responsibilities, including data management, and information provided to them, and refrain
determining the sequence of outage events, sys- from independent or premature statements to the
tem modeling, evaluation of operating tools and media or the public about the activities, findings,
communications, transmission system perfor- or conclusions of the individual Working Groups
mance, generator performance, NERC and regula- or the Task Force as a whole.
tory standards/procedures and compliance,
system planning and design studies, vegetation After publication of the Interim Report, the Task
and right-of-way management, transmission and Force investigative teams continued to evaluate
reliability investments, and root cause analysis. the data collected during Phase I. Continuing with
Phase I criteria, confidentiality was maintained in
Additional teams of experts were established to Phase II, and all investigators and working group
address issues related to the performance of members were asked to refrain from independent
nuclear power plants affected by the outage, and or premature statements to the media or the public
physical and cyber security issues related to the about the activities, findings, or conclusions of the
bulk power infrastructure. The security and individual Working Groups or the Task Force as a
nuclear investigative teams also had liaisons who whole.
worked closely with the various electric system
investigative teams mentioned above. Relevant U.S. and Canadian Legal
Framework
Function of the Working Groups
United States
The U.S. and Canadian co-chairs of each of the
The Secretary of Energy directed the Department
three Working Groups (i.e., an Electric System
of Energy (DOE) to gather information and con-
Working Group, a Nuclear Working Group, and a
duct an investigation to examine the cause or
Security Working Group) designed investigative
causes of the August 14, 2003 blackout. In initiat-
assignments to be completed by the investigative
ing this effort, the Secretary exercised his author-
teams. These findings were synthesized into a sin-
ity under section 11 of the Energy Supply and
gle Interim Report reflecting the conclusions of
Environmental Coordination Act of 1974, and sec-
the three investigative teams and the Working
tion 13 of the Federal Energy Administration Act
Groups. For Phase II, the Interim Report was
of 1974, to gather energy-related information and
enhanced with new information gathered from the
conduct investigations. This authority gives him
technical conferences, additional modeling and
and the DOE the ability to collect such energy
analysis and public comments. Determination of
information as he deems necessary to assist in the
when the Interim and Final Reports were com-
formulation of energy policy, to conduct investiga-
plete and appropriate for release to the public was
tions at reasonable times and in a reasonable man-
the responsibility of the U.S.-Canada Task Force
ner, and to conduct physical inspections at energy
and the investigation co-chairs.
facilities and business premises. In addition, DOE
Confidentiality of Data and Information can inventory and sample any stock of fuels or
energy sources therein, inspect and copy records,
Given the seriousness of the blackout and the reports, and documents from which energy infor-
importance of averting or minimizing future mation has been or is being compiled and to ques-
blackouts, it was essential that the Task Force’s tion such persons as it deems necessary.
teams have access to pertinent records and data DOE worked closely with Natural Resources Can-
from the regional transmission operators (RTOs) ada and NERC on the investigation.
and independent system operators (ISOs) and
180 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Canada the interviews, the interviewers sought informa-
Minister Dhaliwal, as the Minister responsible for tion and data about control room operations and
Natural Resources Canada, was appointed by practices, the organization’s system status and
Prime Minister Chrétien as the Canadian Co-Chair conditions on August 14, the organization’s oper-
of the Task Force. Minister Dhaliwal worked ating procedures and guidelines, load limits on its
closely with his American Co-Chair, Secretary of system, emergency planning and procedures, sys-
Energy Abraham, as well as NERC and his provin- tem security analysis tools and procedures, and
cial counterparts in carrying out his responsibili- practices for voltage and frequency monitoring.
ties. When NRCan Minister R. John Efford Similar interviews were held later with staff at
assumed his role as the new Canadian Co-Chair, Ontario’s Independent Electricity Market Opera-
he continued to work closely with Secretary Abra- tor (IMO) and Hydro One in Canada.
ham and the three Working Groups. On August 22 and 26, NERC directed the reliabil-
Under Canadian law, the Task Force was charac- ity coordinators at the ISOs to obtain a wide range
terized as a non-statutory, advisory body that does of data and information from the control area coor-
not have independent legal personality. The Task dinators under their oversight. The data requested
Force did not have any power to compel evidence included System Control and Data Acquisition
or witnesses, nor was it able to conduct searches (SCADA) logs, Energy Management System (EMS)
or seizures. In Canada, the Task Force relied on logs, alarm logs, data from local digital fault
voluntary disclosure for obtaining information recorders, data on transmission line and generator
pertinent to its work. “trips” (i.e., automatic disconnection to prevent
physical damage to equipment), state estimator
Oversight and Coordination data, operator logs and transcripts, and informa-
The Task Force’s U.S. and Canadian coordinators tion related to the operation of capacitors, phase
held frequent conference calls to ensure that all shifting transformers, load shedding, static var
components of the investigation were making compensators, special protection schemes or sta-
timely progress. They briefed both Secretary Abra- bility controls, and high-voltage direct current
ham and Minister R. John Efford (Minister (HVDC) facilities. NERC issued another data
Dhaliwal, Phase I) regularly and provided weekly request to FirstEnergy on September 15 for copies
summaries from all components on the progress of of studies since 1990 addressing voltage support,
the investigation. During part of Phase I, the lead- reactive power supply, static capacitor applica-
ership of the electric system investigation team tions, voltage requirements, import or transfer
held daily conference calls to address analytical capabilities (in relation to reactive capability or
and process issues important to the investigation. voltage levels), and system impacts associated
The three Working Groups held weekly confer- with unavailability of the Davis-Besse plant. All
ence calls to enable the investigation teams to parties were instructed that data and information
update the Working Group members on the state provided to either DOE or NERC did not have to be
of the overall analysis. Conference calls also submitted a second time to the other entity—all
focused on the analysis updates and the need to material provided would go into a common data
ensure public availability of all inputs to the base.
development of recommendations. Working For the Interim Report the investigative team held
Group members attended panels and face-to-face three technical conferences (August 22, Septem-
meetings to review drafts of the report. ber 8-9, and October 1-3) with the RTOs and ISOs
and key utilities aimed at clarifying the data
Electric System Investigation Phase I received, filling remaining gaps in the data, and
Investigative Process developing a shared understanding of the data’s
implications.
Collection of Data and Information from ISOs,
Utilities, States, and the Province of Ontario Data “Warehouse”
On Tuesday, August 19, 2003, investigators affili- The data collected by the investigative team was
ated with the U.S. Department of Energy (DOE) organized in an electronic repository containing
began interviewing control room operators and thousands of transcripts, graphs, generator and
other key officials at the ISOs and the companies transmission data and reports at the NERC head-
most directly involved with the initial stages of the quarters in Princeton, New Jersey. The warehouse
outage. In addition to the information gained in contains more than 20 gigabytes of information, in
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 181
more than 10,000 files. This established a set of objective was to identify industry procedures that
validated databases that the analytic teams could are in place today for collecting information fol-
access as needed. lowing large-scale transmission related power out-
ages and to assess those procedures in terms of the
Individual investigative teams conducted their
August 14, 2003 power outage investigation.
activities through a number of in-person meetings
as well as conference calls and e-mail communica- They sought to:
tions over the months of the investigation.
u Determine what happened in terms of immedi-
Detailed investigative team findings will be
ate causes, sequence of events, and resulting
included in upcoming technical reports issued by
consequences;
NERC.
u Understand the failure mechanism via record-
The following were the information sources for ings of system variables such as frequency, volt-
the Electric System Investigation: ages, and flows;
u Interviews conducted by members of the u Enable disturbance re-creation using computer
U.S.-Canada Electric Power System Outage models for the purposes of understanding the
Investigation Team with personnel at all of the mechanism of failure, identifying ways to avoid
utilities, control areas and reliability coordina- or mitigate future failures, and assessing and
tors in the weeks following the blackout. improving the integrity of computer models;
u Three fact-gathering meetings conducted by the u Identify deeper, underlying factors contributing
Investigation Team with personnel from the to the failure (e.g., general policies, standard
above organizations on August 22, September 8 practices, communication paths, organizational
and 9, and October 1 to 3, 2003. cultures).
u Three public hearings held in Cleveland, Ohio; Sequence of Events
New York City, New York; and Toronto,
Ontario. More than 800 events occurred during the black-
out of August 14. The events included the opening
u Two technical conferences held in Philadel- and closing of transmission lines and associated
phia, Pennsylvania, and Toronto, Canada. breakers and switches, the opening of transform-
ers and associated breakers, and the tripping and
u Materials provided by the above organizations
starting of generators and associated breakers.
in response to one or more data requests from
Most of these events occurred in the few minutes
the Investigation Team.
of the blackout cascade between 16:06 and 16:12
u All taped phone transcripts between involved EDT. To properly analyze a blackout of this mag-
operations centers. nitude, an accurate knowledge of the sequence of
events must be obtained before any analysis of the
u Additional interviews and field visits with oper- blackout can be performed.
ating personnel on specific issues in October
2003 and January 2004. Establishing a precise and accurate sequence of
outage-related events was a critical building block
u Field visits to examine transmission lines and for the other parts of the investigation. One of the
vegetation at short-circuit locations. key problems in developing this sequence was
u Materials provided by utilities and state regula- that although much of the data pertinent to an
tors in response to data requests on vegetation event was time-stamped, there was variation from
management issues. source to source in how the time-stamping was
done, and not all of the time-stamps were synchro-
u Detailed examination of thousands of individ- nized to the National Institute of Standards and
ual relay trips for transmission and generation Technology (NIST) standard clock in Boulder, CO.
events. Validating the timing of specific events became a
large, important, and sometimes difficult task.
Data Exploration and Requirements
This work was also critical to the issuance by the
This group requested data from the following con- Task Force on September 12 of a “timeline” for the
trol areas and their immediate neighbors: MISO, outage. The timeline briefly described the princi-
MECS, FE, PJM, NYISO, ISO-NE, and IMO. The pal events, in sequence, leading up to the initia-
data and exploration and requirements group’s tion of the outage’s cascade phase, and then in the
182 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
cascade itself. The timeline was not intended, reclosing of outages facilities during the
however, to address the causal relationships sequence of events, load shedding, generation
among the events described, or to assign fault or redispatch, and combinations of load shedding
responsibility for the blackout. All times in the and redispatch.
chronology are in Eastern Daylight Time. u Compared transaction tags for August 14, to
System Modeling and Simulation Analysis show how they matched up with those of other
days in 2003 and 2002.
The system modeling and simulation team
u Analyzed the transactions and generation dis-
(SMST) replicated system conditions on August patch changes used to bring replacement power
14 and the events leading up to the blackout. The for the loss of Eastlake 5 generation into
modeling reflects the state of the electric system. FirstEnergy, to determine where the replace-
Once benchmarked to actual conditions at ment power came from.
selected critical times on August 14, it allowed
analysts to conduct a series of sensitivity studies u Analyzed the performance of the Interchange
to determine if the system was stable and within Distribution Calculator (IDC) and its potential
limits at each point in time leading up to the cas- capability to help mitigate the overloads.
cade. The analysis also confirmed when the sys- The SMST began its efforts using the base case
tem became unstable and allowed analysts to test data and model provided by FirstEnergy as its
whether measures such as load-shedding would foundation.
have prevented the cascade.
The modeling and system studies work was per-
This team consisted of a number of NERC staff and formed under the guidance of a specially formed
persons with expertise in areas necessary to read MAAC-ECAR-NPCC (MEN) Coordinating Group,
and interpret all of the data logs, digital fault consisting of the Regional Managers from those
recorder information, sequence of events record- three regions impacted by the blackout, and their
ers information, etc. The team consisted of about respective regional chairmen or designees.
40 people involved at various different times with
additional experts from the affected areas to Assessment of Operations Tools, SCADA/EMS,
understand the data. Communications, and Operations Planning
Overall, this team: The Operations Tools, SCADA/EMS, Communica-
u Created steady-state power flow cases for
tions, and Operations Planning Team assessed the
observed August 14 system conditions starting observability of the electric system to operators
at 15:00 EDT through about 16:05 EDT (when and reliability coordinators, and the availability
powerflow simulations were no longer ade- and effectiveness of operational (real-time and
quate), about the time of the Sammis-Star day-ahead) reliability assessment tools, including
345-kV outage. redundancy of views and the ability to observe the
“big picture” regarding bulk electric system condi-
u Compiled relevant data for dynamic modeling tions. The team investigated operating practices
of affected systems (e.g. generator dynamic and effectiveness of operating entities and reliabil-
models, load characteristics, special protection ity coordinators in the affected area. This team
schemes, etc.). investigated all aspects of the blackout related to
u Performed rigorous contingency analysis (over operator and reliability coordinator knowledge of
800 contingencies in Eastern Interconnection system conditions, action or inactions, and
run) to determine if the system was within oper- communications.
ating within thermal and voltage limits, and
within limits for possible further contingencies The Operations and Tools team conducted exten-
(N-1 contingencies) prior to and during the ini- sive interviews with operating personnel at
tial events of the blackout sequence. the affected facilities. They participated in the
technical investigation meetings with affected
u Performed sensitivity analysis to determine the operators in August, September and October and
significance of pre-existing conditions such as reviewed the August 14 control room transcripts
transmission outages in Cinergy and Dayton, in detail. This group investigated the performance
and the earlier loss of Eastlake unit 5 of the MISO and FirstEnergy EMS hardware and
generation. software and its impact on the blackout, and
u Performed “what-if” analysis to determine looked at operator training (including the use
potential impacts of remedial actions such as of formal versus “on-the-job” training) and the
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 183
communications and interactions between the reported on patterns and conclusions regarding
operations and information technology support what caused generation facilities to trip. The team
staff at both organizations. identified any unexpected performance anomalies
or unexplained events. The team assessed genera-
Frequency/ACE Analysis tor maintenance practices in the affected area as
The Frequency/ACE Team analyzed potential fre- compared to good utility practice. The team ana-
quency anomalies that may have occurred on lyzed the coordination of generator under-
August 14, as compared to typical interconnection frequency settings with transmission settings,
operations. The team also determined whether such as under-frequency load shedding. The team
there were any unusual issues with control perfor- gathered and analyzed data on affected nuclear
mance and frequency and any effects they may units and worked with the Nuclear Regulatory
have had related to the cascading failure, and Commission to address U.S. nuclear unit issues.
whether frequency-related anomalies were con-
The Generator Performance team sent out an
tributing factors or symptoms of other problems
extensive data request to generator owners during
leading to the cascade.
Phase I of the investigation, but did not receive the
Assessment of Transmission System bulk of the responses until Phase II. The analysis
Performance, Protection, Control, in this report uses the time of generator trip as it
Maintenance, and Damage was reported by the plant owner, or the time when
the generator ceased feeding power into the grid as
This team investigated the causes of all transmis- determined by a system monitoring device, and
sion facility automatic operations (trips and synchronized those times to other known grid
reclosings) leading up to and through to the end of events as best as possible. However, many genera-
the cascade on all facilities greater than 100 kV. tion owners offered little information on the cause
Included in the review were relay protection and of unit trips or key information on conditions at
remedial action schemes, including under- their units, so it may never be possible to fully
frequency load-shedding and identification of the determine what happened to all the generators
cause of each operation and any misoperations affected by the blackout, and why they performed
that may have occurred. The team also assessed as they did. In particular, it is not clear what point
transmission facility maintenance practices in the in time each reported generator trip time reflects—
affected area as compared to good utility practice i.e., when in the cycle between when the generator
and identified any transmission equipment that first detected the condition which caused it to trip,
was damaged as a result of the cascading outage. or several seconds later when it actually stopped
The team reported patterns and conclusions feeding power into the grid. This lack of clear data
regarding what caused transmission facilities to hampered effective investigation of generator
trip; why did the cascade extend as far as it did issues.
and not further into other systems; any
misoperations and the effect those misoperations Vegetation Management
had on the outage; and any transmission equip-
For Phase I the Vegetation/Right of Way Team con-
ment damage. Also the team reported on the trans-
ducted a field investigation into the contacts that
mission facility maintenance practices of entities
occurred between trees and conductors on August
in the affected area compared to good utility
14 within the FirstEnergy, Dayton Power & Light
practice.
and Cinergy service areas. The team also exam-
Assessment of Generator Performance, ined detailed information gained from data
Protection, Controls, Maintenance, and requests to these and other utilities, including his-
Damage torical outages from tree contacts on these lines.
These findings were included in the Interim
This team investigated the cause of generator trips Report and detailed in an interim report on utility
for all generators with a 10 MW or greater name- vegetation management, posted at http://www.
plate rating leading to and through the end of the ferc.gov/cust-protect/moi/uvm-initial-report.pdf.
cascade. The review included the cause for the
generator trips, relay targets, unit power runbacks, The team also requested information from the
and voltage/reactive power excursions. The team public utility commissions in the blackout area on
reported any generator equipment that was dam- any state requirements for transmission vegeta-
aged as a result of the cascading outage. The team tion management and right-of-way maintenance.
184 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Beginning in Phase I and continuing into Phase II, event being considered could have occurred with-
the Vegetation/ROW team looked in detail at the out the proposed cause (or combination of causes),
vegetation management and ROW maintenance the proposed cause or combination of causes is
practices for the three utilities above, and com- dropped from consideration and other possibili-
pared them to accepted utility practices across ties are considered.
North America. Issues examined included ROW
legal clearance agreements with landowners, bud- Root cause analysis typically identifies several or
gets, tree-trimming cycles, organization structure, even many causes of complex events; each of the
and use of herbicides. Through CN Utility Con- various branches of the analysis is pursued until
sulting, the firm hired by FERC to support the either a “root cause” is found or a non-correctable
blackout investigation, the Vegetation/ROW team condition is identified. (A condition might be con-
also identified “best practices” for transmission sidered as non-correctable due to existing law,
ROW management. They used those practices to fundamental policy, laws of physics, etc.). Some-
evaluate the performance of the three utilities times a key event in a causal chain leading to the
involved in August 14 line outages and also to major event could have been prevented by timely
evaluate the effectiveness of utility vegetation action by one or another party; if such action was
management practices generally. feasible, and if the party had a responsibility to
take such action, the failure to do so becomes a
On March 2, 2004, FERC released CN Utility Con- root cause of the major event.
sulting’s “Utility Vegetation Management Final
Report” (see http://www.ferc.gov/cust-protect/ Phase II
moi/uvm-final-report.pdf).
On December 12, 2003, Paul Martin was elected as
the new Prime Minister of Canada and assumed
Root Cause Analysis
responsibility for the Canadian section of the
The investigation team used a technique called Power System Outage Task Force. Prime Minister
root cause analysis to help guide the overall inves- Martin appointed R. John Efford as the new Minis-
tigation process in an effort to identify root causes ter of Natural Resources Canada and co-chair of
and contributing factors leading to the start of the the Task Force.
blackout in Ohio. The root cause analysis team
Press releases, a U.S. Federal Register notice, and
worked closely with the technical investigation
ads in the Canadian press notified the public and
teams providing feedback and queries on addi-
stakeholders of Task Force developments. All
tional information. Also, drawing on other data
public statements were released to the media and
sources as needed, the root cause analysis verified
are available on the OETD and the NRCan web
facts regarding conditions and actions (or inac-
sites.
tions) that contributed to the blackout.
Several of the investigative teams began their
Root cause analysis is a systematic approach to
work during Phase I and completed it during
identifying and validating causal linkages among
Phase II. Other teams could not begin their investi-
conditions, events, and actions (or inactions) lead-
gation into the events related to the cascade and
ing up to a major event of interest—in this case the
blackout, beginning at 16:05:57 EDT on August
August 14 blackout. It has been successfully
14, 2003, until analysis of the Ohio events before
applied in investigations of events such as nuclear
that point was completed in Phase I.
power plant incidents, airplane crashes, and the
recent Columbia space shuttle disaster. System Planning, Design and Studies Team
Root cause analysis is driven by facts and logic. The SPDST studied reactive power management,
Events and conditions that may have helped to transactions scheduling, system studies and sys-
cause the major event in question are described in tem operating limits for the Ohio and ECAR areas.
factual terms, and causal linkages are established In addition to the data in the investigation data
between the major event and earlier conditions or warehouse, the team submitted six comprehen-
events. Such earlier conditions or events are sive data requests to six control areas and reliabil-
examined in turn to determine their causes, and at ity coordinators, including FirstEnergy, to build
each stage the investigators ask whether the par- the foundation for its analyses. The team exam-
ticular condition or event could have developed or ined reactive power and voltage management poli-
occurred if a proposed cause (or combination of cies, practices and criteria and compared them to
causes) had not been present. If the particular actual and modeled system conditions in the
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 185
affected area and neighboring systems. They analysis. They looked independently into many
assessed the process of assessing and approving issues, conducting additional interviews as
transaction schedules and tags and the coordina- needed. The team distinguished between those
tion of those schedules and transactions in violations which could be clearly proven and
August, 2003, and looked at the impact of tagged those which were problematic but not fully prov-
transactions on key facilities on August 14. Simi- able. The SP&C team offered a number of conclu-
larly, the team examined system operating limits sions and recommendations to improve
in effect for the affected area on August 14, how operational reliability, NERC standards, the stan-
they had been determined, and whether they were dards development process and the compliance
appropriate to the grid as it existed in August program.
2003. They reviewed system studies conducted by
FirstEnergy and ECAR for 2003 and prior years, Dynamic Modeling of the Cascade
including the methodologies and assumptions This work was conducted as an outgrowth of the
used in those studies and how those were coordi- work done by the System Modeling and Simula-
nated across adjoining control areas and councils. tion team in Phase I, by a team composed of the
The SPDST also compared how the studied condi- NPCC System Studies-38 Working Group on
tions compared to actual conditions on August 14. Inter-Area Dynamic Analysis, augmented by rep-
For all these matters, the team compared the poli- resentatives from ECAR, MISO, PJM and SERC.
cies, studies and practices to good utility Starting with the steady-state power flows devel-
practices. oped in Phase I, they moved the analysis forward
across the Eastern Interconnection from 16:05:50
The SPDST worked closely with the Modeling and
EDT on in a series of first steady-state, then
System Simulation Team. They used data pro-
dynamic simulations to understand how condi-
vided by the control areas, RTOs and ISOs on
tions changed across the grid.
actual system conditions across August 2003, and
NERC Tag Dump and TagNet data. To do the volt- This team is using the model to conduct a series of
age analyses, the team started with the MSST’s “what if” analyses, to better understand what con-
base case data and model of the entire Eastern ditions contributed to the cascade and what might
Interconnection, then used a more detailed model have happened if events had played out differ-
of the FE area provided by FirstEnergy. With these ently. This work is described further within Chap-
models they conducted extensive PV and VQ anal- ter 6.
yses for different load levels and contingency
combinations in the Cleveland-Akron area, run- Additional Cascade Analysis
ning over 10,000 different power flow simula- The core team for the cascade investigation drew
tions. Team members have extensive experience upon the work of all the teams to understand the
and expertise in long-term and operational plan- cascade after 16:05:57. The investigation’s official
ning and system modeling. Sequence of Events was modified and corrected as
appropriate as additional information came in
NERC Standards, Procedures and Compliance from asset owners, and as modeling and other
Team investigation revealed inaccuracies in the initial
The SP&C team was charged with reviewing the data reports. The team issued additional data
NERC Operating Policies and Planning Standards requests and looked closely at the data collected
for any violations that occurred in the events lead- across the period of the cascade. The team orga-
ing up to and during the blackout, and assessing nized the analysis by attempting to link the indi-
the sufficiency or deficiency of NERC and regional vidual area and facility events to the power flows,
reliability standards, policies and procedures. voltages and frequency data recorded by Hydro
They were also directed to develop and conduct One’s PSDRs (as seen in Figures 6.16 and 6.25)
audits to assess compliance with the NERC and and similar data sets collected elsewhere. This
regional reliability standards as relevant to the effort improved the team’s understanding of the
cause of the outage. interrelationships between the interaction, timing
and impacts of lines, loads and generation trips,
The team members, all experienced participants which are now being confirmed by dynamic mod-
in the NERC compliance and auditing program, eling. Graphing, mapping and other visualization
examined the findings of the Phase I investigation tools also created insights into the cascade, as
in detail, building particularly upon the root cause with the revelation of the role of zone 3 relays in
186 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
accelerating the early spread of the cascade within additional comment on public websites main-
Ohio and Michigan. tained by the U.S. Department of Energy and Nat-
ural Resources Canada (www.electricity.doe.gov
The team was aided in its work by the ability to
and www.nrcan.gc.ca, respectively). The tran-
learn from the studies and reports on the blackout
scripts from the meetings and conferences were
completed by various groups outside the investi-
also posted on these websites.
gation, including those by the Public Utility Com-
mission of Ohio, the Michigan Public Service u Members of all three Working Groups partici-
Commission, the New York ISO, ECAR and the pated in public forums in Cleveland, Ohio
Public Service Commission of New York. (December 4, 2003), New York City (December
5, 2003), and Toronto, Ontario (December 8,
Beyond the work of the Electric System investiga- 2003).
tion, the Security and Nuclear investigation teams
conducted additional analyses and updated their u The ESWG held two technical conferences, in
interim reports with the additional findings. Philadelphia, Pennsylvania (December 16,
2003), and Toronto, Ontario (January 9, 2004).
Preparation of Task Force
u The NWG also held a public meeting on
Recommendations
nuclear-related issues pertaining to the black-
Public and stakeholder input was an important out at the U.S. Nuclear Regulatory Commission
component in the development of the Task Force’s headquarters in Rockville, Maryland (January 6,
recommendations. The input received covered a 2004).
wide range of subjects, including enforcement of The electric system investigation team also devel-
reliability standards, improving communications, oped an extensive set of technical findings based
planning for responses to emergency conditions, on team analyses and cross-team discussions as
and the need to evaluate market structures. See the Phase I and Phase II work progressed. Many of
Appendix C for a list of contributors. these technical findings were reflected in NERC’s
Three public forums and two technical confer- actions and initiatives of February 10, 2004. In
ences were held to receive public comments on turn, NERC’s actions and initiatives received sig-
the Interim Report and suggested recommenda- nificant attention in the development of the Task
tions for consideration by the Task Force. These Force’s recommendations.
events were advertised by various means, includ- The SWG convened in January 2004 in Ottawa to
ing announcements in the Federal Register and the review the Interim Report. The SWG also held vir-
Canada Gazette, advertisements in local news- tual meetings with the investigative team leads
papers in the U.S., invitations to industry through and working group members.
NERC, invitations to the affected state and
provincial regulatory bodies, and government Similarly, the ESWG conducted weekly telephone
press releases. All written inputs received at conferences and it held face-to-face meetings on
these meetings and conferences were posted for January 30, March 3, and March 18, 2004.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 187
Appendix C
List of Commenters
The individuals listed below either commented on the Interim Report, provided suggestions for recom-
mendations to improve reliability, or both. Their input was greatly appreciated. Their comments can be
viewed in full or in summary at http://www.nrcan.gc.ca or at http://www.electricity.doe.gov.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 189
Fernandez, Rick Personal comment
Fidrych, Mark Western Area Power Authority (WAPA) and Chairman of the NERC Operating Committee
Furuya, Toshihiko Tokyo Electric Power Co., Inc.
Galatic, Alex Personal comment
Garg, Ajay Hydro One Networks Inc.
Goffin, Dave Canadian Chemical Producers Assocation
Gruber, William M. Ondrey Attorney
Guimond, Pierre Canadian Nuclear Association
Gurdziel, Tom Personal comment
Hakobyan, Spartak and
Personal comment
Gurgen
Han, Masur Personal comment
Hauser, Carl School of Electrical Engineering and Computer Science, Washington State University
Hebert, Larry Thunder Bay Hydro
Hilt, Dave NERC
Hughes, John P. ELCON
Imai, Shinichi Tokyo Electric Power
Jeyapalan, Jey K. Jeyapalan & Associates, LLC
Johnston, Sidney A. Personal comment
Kane, Michael Personal comment
Katsuras, George Independent Electric Market Operator of Ontario
Kellat, Stephen Personal comment
Kerr, Jack Dominion Virginia Power
Kerr, Jack Best Real-time Reliability Analysis Practices Task Force
Kershaw, Raymond K. International Transmission Company
Kolodziej, Eddie Personal comment
Konow, Hans Canadian Electricity Association
Kormos, Mike PJM
Kucey, Tim National Energy Board (Canada)
Laugier, Alexandre Personal comment
Lawson, Barry National Rural Electric Cooperative Association
Lazarewicz, Matthew L. Beacon Power Corp.
Lee, Stephen Electric Power Research Institute
Leovy, Steve Personal comment
Linda Campbell Florida Reliability Coordinating Council
Loehr, G.C. Power Engineers Seeking Truth
Love, Peter Canadian Energy Efficiency Alliance
Macedo, Frank Hydro One
Maliszewski, R.M. Power Engineers Seeking Truth
McMonagle, Rob Canadian Solar Industries Assocation
Meissner, Joseph Personal comment
190 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Middlestadt, Bill Bonneville Power Administration
Cuyahoga County Board of Commissioners, and member, Community Advisory Panel; panel cre-
Milter, Carolyn
ated for Cleveland Electric Illuminating Co. (later First Energy)
Mitchell, Terry Excel Energy
Moore, Scott AEP
Murphy, Paul IMO Ontario
Museler, William J. New York Independent System Operator
O’Keefe, Brian Canadian Union of Public Employees
Oliver, Fiona Canadian Energy Efficiency Alliance
Ormund, Peter Mohawk College
Pennstone, Mike Hydro One
Pereira, Les Personal comment
Phillips, Margie Pennsylvania Services Integration Consortium
Rocha, Paul X. CenterPoint Energy
Ross, Don Prince Edward Island Wind Co-Op
Rupp, Douglas B Ada Core Technologies, Inc.
Sasson, Mayer New York State Reliability Council
Schwerdt, Ed Northeast Power Coordinating Council
Seppa, Tapani O. The Valley Group, Inc.,
Silverstein, Alison Federal Energy Regulatory Commission
Spears, J. Personal comment
Spencer, Sidney Personal comment
spider Personal comment
Staniford, Stuart Personal comment
Stephens, Eric B. Ohio Consumers’ Counsel (OCC)
Stewart, Bob PG&E
Synesiou, John IMS Corporation
Tarler, Howard A. On behalf of Chairman William M. Flynn, New York State Department of Public Service
Tatro, Phil National Grid Company
Taylor, Carson Bonneville Power Administration
van Welie, Gordon ISO New England Inc.
Van Zandt, Vickie Bonneville Power Administration
Warren, Kim IMO Ontario
Watkins, Don Bonneville Power Administration
Wells, Chuck OSISoft
Wiedman, Tom ConEd
Wightman, Donald Utility Workers Union of America
Wilson, John Personal comment
Winter, Chris Conservation Council of Ontario
Wright, C. Dortch On behalf of New Jersey Governor James E. McGreevey
Zwergel, Dave Midwest ISO
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 191
Appendix D
The Board of Trustees recognizes the paramount importance of a reliable bulk electric system in
North America. In consideration of the findings of the investigation into the August 14, 2003
blackout, NERC must take firm and immediate actions to increase public confidence that the
reliability of the North American bulk electric system is being protected.
A key finding of the blackout investigators is that violations of existing NERC reliability standards
contributed directly to the blackout. Pending enactment of federal reliability legislation creating a
framework for enforcement of mandatory reliability standards, and with the encouragement of the
Stakeholders Committee, the board is determined to obtain full compliance with all existing and
future reliability standards and intends to use all legitimate means available to achieve that end. The
board therefore resolves to:
• Receive specific information on all violations of NERC standards, including the identities of
the parties involved;
• Take firm actions to improve compliance with NERC reliability standards;
• Provide greater transparency to violations of standards, while respecting the confidential
nature of some information and the need for a fair and deliberate due process; and
• Inform and work closely with the Federal Energy Regulatory Commission and other
applicable federal, state, and provincial regulatory authorities in the United States, Canada,
and Mexico as needed to ensure public interests are met with respect to compliance with
reliability standards.
The board expresses its appreciation to the blackout investigators and the Steering Group for their
objective and thorough work in preparing a report of recommended NERC actions. With a few
clarifications, the board approves the report and directs implementation of the recommended actions.
The board holds the assigned committees and organizations accountable to report to the board the
progress in completing the recommended actions, and intends itself to publicly report those results.
The board recognizes the possibility that this action plan may have to be adapted as additional
analysis is completed, but stresses the need to move forward immediately with the actions as stated.
Furthermore, the board directs management to immediately advise the board of any significant
violations of NERC reliability standards, including details regarding the nature and potential
reliability impacts of the alleged violations and the identity of parties involved. Management shall
supply to the board in advance of board meetings a detailed report of all violations of reliability
standards.
Finally, the board resolves to form a task force to develop guidelines for the board to consider with
regard to the confidentiality of compliance information and disclosure of such information to
regulatory authorities and the public.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 193
Overview of Investigation Conclusions
The North American Electric Reliability Council (NERC) has conducted a comprehensive
investigation of the August 14, 2003 blackout. The results of NERC’s investigation contributed
significantly to the U.S./Canada Power System Outage Task Force’s November 19, 2003 Interim
Report identifying the root causes of the outage and the sequence of events leading to and during the
cascading failure. NERC fully concurs with the conclusions of the Interim Report and continues to
provide its support to the Task Force through ongoing technical analysis of the outage. Although an
understanding of what happened and why has been resolved for most aspects of the outage, detailed
analysis continues in several areas, notably dynamic simulations of the transient phases of the
cascade and a final verification of the full scope of all violations of NERC and regional reliability
standards that occurred leading to the outage.
194 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Overview of Recommendations
The Board of Trustees approves the NERC Steering Group recommendations to address these
shortcomings. The recommendations fall into three categories.
Actions to Remedy Specific Deficiencies: Specific actions directed to First Energy (FE), the
Midwest Independent System Operator (MISO), and the PJM Interconnection, LLC (PJM) to correct
the deficiencies that led to the blackout.
1. Correct the Direct Causes of the August 14, 2003 Blackout.
Strategic Initiatives: Strategic initiatives by NERC and the regional reliability councils to strengthen
compliance with existing standards and to formally track completion of recommended actions from
August 14, and other significant power system events.
2. Strengthen the NERC Compliance Enforcement Program.
3. Initiate Control Area and Reliability Coordinator Reliability Readiness Audits.
4. Evaluate Vegetation Management Procedures and Results.
5. Establish a Program to Track Implementation of Recommendations.
Technical Initiatives: Technical initiatives to prevent or mitigate the impacts of future cascading
blackouts.
6. Improve Operator and Reliability Coordinator Training
7. Evaluate Reactive Power and Voltage Control Practices.
8. Improve System Protection to Slow or Limit the Spread of Future Cascading Outages.
9. Clarify Reliability Coordinator and Control Area Functions, Responsibilities, Capabilities
and Authorities.
10. Establish Guidelines for Real-Time Operating Tools.
11. Evaluate Lessons Learned During System Restoration.
12. Install Additional Time-Synchronized Recording Devices as Needed.
13. Reevaluate System Design, Planning and Operating Criteria.
14. Improve System Modeling Data and Data Exchange Practices.
Market Impacts
Many of the recommendations in this report have implications for electricity markets and market
participants, particularly those requiring reevaluation or clarification of NERC and regional
standards, policies and criteria. Implicit in these recommendations is that the NERC board charges
the Market Committee with assisting in the implementation of the recommendations and interfacing
with the North American Energy Standards Board with respect to any necessary business practices.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 195
Recommendation to Remedy Specific Deficiencies
Recommendation 1. Correct the Direct Causes of the August 14, 2003 Blackout.
NERC’s technical analysis of the August 14 blackout leads it to fully concur with the Task Force
Interim Report regarding the direct causes of the blackout. The report stated that the principal causes
of the blackout were that FE did not maintain situational awareness of conditions on its power system
and did not adequately manage tree growth in its transmission rights-of-way. Contributing factors
included ineffective diagnostic support provided by MISO as the reliability coordinator for FE and
ineffective communications between MISO and PJM.
NERC will take immediate and firm actions to ensure that the same deficiencies that were directly
causal to the August 14 blackout are corrected. These steps are necessary to assure electricity
customers, regulators and others with an interest in the reliable delivery of electricity that the power
system is being operated in a manner that is safe and reliable, and that the specific causes of the
August 14 blackout have been identified and fixed.
Recommendation 1a: FE, MISO, and PJM shall each complete the remedial actions designated
in Attachment A for their respective organizations and certify to the NERC board no later than
June 30, 2004, that these specified actions have been completed. Furthermore, each
organization shall present its detailed plan for completing these actions to the NERC
committees for technical review on March 23-24, 2004, and to the NERC board for approval no
later than April 2, 2004.
Recommendation 1b: The NERC Technical Steering Committee shall immediately assign a
team of experts to assist FE, MISO, and PJM in developing plans that adequately address the
issues listed in Attachment A, and other remedial actions for which each entity may seek
technical assistance.
196 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Strategic Initiatives to
Assure Compliance with Reliability Standards and to Track Recommendations
Recommendation 2a: Each regional reliability council shall report to the NERC Compliance
Enforcement Program within one month of occurrence all significant1 violations of NERC
operating policies and planning standards and regional standards, whether verified or still
under investigation. Such reports shall confidentially note details regarding the nature and
potential reliability impacts of the alleged violations and the identity of parties involved.
Additionally, each regional reliability council shall report quarterly to NERC, in a format
prescribed by NERC, all violations of NERC and regional reliability council standards.
Recommendation 2b: Being presented with the results of the investigation of any significant
violation, and with due consideration of the surrounding facts and circumstances, the NERC
board shall require an offending organization to correct the violation within a specified time. If
the board determines that an offending organization is non-responsive and continues to cause a
risk to the reliability of the interconnected power systems, the board will seek to remedy the
violation by requesting assistance of the appropriate regulatory authorities in the United States,
Canada, and Mexico.
1
Although all violations are important, a significant violation is one that could directly reduce the integrity of the
interconnected power systems or otherwise cause unfavorable risk to the interconnected power systems. By contrast, a
violation of a reporting or administrative requirement would not by itself generally be considered a significant violation.
Approved by the Board of Trustees 5
February 10, 2004
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 197
Recommendation 2c: The Planning and Operating Committees, working in conjunction with
the Compliance Enforcement Program, shall review and update existing approved and draft
compliance templates applicable to current NERC operating policies and planning standards;
and submit any revisions or new templates to the board for approval no later than March 31,
2004. To expedite this task, the NERC President shall immediately form a Compliance
Template Task Force comprised of representatives of each committee. The Compliance
Enforcement Program shall issue the board-approved compliance templates to the regional
reliability councils for adoption into their compliance monitoring programs.
This effort will make maximum use of existing approved and draft compliance templates in order to
meet the aggressive schedule. The templates are intended to include all existing NERC operating
policies and planning standards but can be adapted going forward to incorporate new reliability
standards as they are adopted by the NERC board for implementation in the future.
When the investigation team’s final report on the August 14 violations of NERC and regional
standards is available in March, it will be important to assess and understand the lapses that allowed
violations to go unreported until a large-scale blackout occurred.
Recommendation 2d: The NERC Compliance Enforcement Program and ECAR shall, within
three months of the issuance of the final report from the Compliance and Standards
investigation team, evaluate the identified violations of NERC and regional standards, as
compared to previous compliance reviews and audits for the applicable entities, and develop
recommendations to improve the compliance process.
In conducting its investigation, NERC found that deficiencies in control area and reliability
coordinator capabilities to perform assigned reliability functions contributed to the August 14
blackout. In addition to specific violations of NERC and regional standards, some reliability
coordinators and control areas were deficient in the performance of their reliability functions and did
not achieve a level of performance that would be considered acceptable practice in areas such as
operating tools, communications, and training. In a number of cases there was a lack of clarity in the
NERC policies with regard to what is expected of a reliability coordinator or control area. Although
the deficiencies in the NERC policies must be addressed (see Recommendation 9), it is equally
important to recognize that standards cannot prescribe all aspects of reliable operation and that
minimum standards present a threshold, not a target for performance. Reliability coordinators and
control areas must perform well, particularly under emergency conditions, and at all times strive for
excellence in their assigned reliability functions and responsibilities.
198 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Recommendation 3a: The NERC Compliance Enforcement Program and the regional
reliability councils shall jointly establish a program to audit the reliability readiness of all
reliability coordinators and control areas, with immediate attention given to addressing the
deficiencies identified in the August 14 blackout investigation. Audits of all control areas and
reliability coordinators shall be completed within three years and continue in a three-year
cycle. The 20 highest priority audits, as determined by the Compliance Enforcement Program,
will be completed by June 30, 2004.
Recommendation 3b: NERC will establish a set of baseline audit criteria to which regional
criteria may be added. The control area requirements will be based on the existing NERC
Control Area Certification Procedure. Reliability coordinator audits will include evaluation of
reliability plans, procedures, processes, tools, personnel qualifications, and training. In
addition to reviewing written documents, the audits will carefully examine the actual practices
and preparedness of control areas and reliability coordinators.
Recommendation 3c: The reliability regions, with the oversight and direct participation of
NERC, will audit each control area’s and reliability coordinator’s readiness to meet these audit
criteria. FERC and other relevant regulatory agencies will be invited to participate in the
audits, subject to the same confidentiality conditions as the other members of the audit teams.
Ineffective vegetation management was a major cause of the August 14 blackout and also contributed
to other historical large-scale blackouts, such on July 2-3, 1996 in the west. Maintaining
transmission line rights-of-way (ROW), including maintaining safe clearances of energized lines
from vegetation, under-build, and other obstructions2 incurs a substantial ongoing cost in many areas
of North America. However, it is an important investment for assuring a reliable electric system.
NERC does not presently have standards for ROW maintenance. Standards on vegetation
management are particularly challenging given the great diversity of vegetation and growth patterns
across North America. However, NERC’s standards do require that line ratings are calculated so as
to maintain safe clearances from all obstructions. Furthermore, in the United States, the National
Electrical Safety Code (NESC) Rules 232, 233, and 234 detail the minimum vertical and horizontal
safety clearances of overhead conductors from grounded objects and various types of obstructions.
NESC Rule 218 addresses tree clearances by simply stating, “Trees that may interfere with
ungrounded supply conductors should be trimmed or removed.” Several states have adopted their
own electrical safety codes and similar codes apply in Canada.
Recognizing that ROW maintenance requirements vary substantially depending on local conditions,
NERC will focus attention initially on measuring performance as indicated by the number of high
voltage line trips caused by vegetation rather than immediately move toward developing standards for
2
Vegetation, such as the trees that caused the initial line trips in FE that led to the August 14, 2003 outage is not the only
type of obstruction that can breach the safe clearance distances from energized lines. Other examples include under-build
of telephone and cable TV lines, train crossings, and even nests of certain large bird species.
Approved by the Board of Trustees 7
February 10, 2004
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 199
ROW maintenance. This approach has worked well in the Western Electricity Coordinating Council
(WECC) since being instituted after the 1996 outages.
Recommendation 4a: NERC and the regional reliability councils shall jointly initiate a program
to report all bulk electric system3 transmission line trips resulting from vegetation contact4.
The program will use the successful WECC vegetation monitoring program as a model.
Recommendation 4b: Beginning with an effective date of January 1, 2004, each transmission
operator will submit an annual report of all vegetation-related high voltage line trips to its
respective reliability region. Each region shall assemble a detailed annual report of vegetation-
related line trips in the region to NERC no later than March 31 for the preceding year, with the
first reporting to be completed by March 2005 for calendar year 2004.
Vegetation management practices, including inspection and trimming requirements, can vary
significantly with geography. Additionally, some entities use advanced techniques such as planting
beneficial species or applying growth retardants. Nonetheless, the events of August 14 and prior
outages point to the need for independent verification that viable programs exist for ROW
maintenance and that the programs are being followed.
Recommendation 4c: Each bulk electric transmission owner shall make its vegetation
management procedure, and documentation of work completed, available for review and
verification upon request by the applicable regional reliability council, NERC, or applicable
federal, state or provincial regulatory agency.
Should this approach of monitoring vegetation-related line outages and procedures prove ineffective
in reducing the number of vegetation-related line outages, NERC will consider the development of
minimum line clearance standards to assure reliability.
The August 14 blackout shared a number of contributing factors with prior large-scale blackouts,
including:
• Conductors contacting trees
• Ineffective visualization of power system conditions and lack of situational awareness
• Ineffective communications
• Lack of training in recognizing and responding to emergencies
• Insufficient static and dynamic reactive power supply
• Need to improve relay protection schemes and coordination
3
All transmission lines operating at 230 kV and higher voltage, and any other lower voltage lines designated by the
regional reliability council to be critical to the reliability of the bulk electric system, shall be included in the program.
4
A line trip includes a momentary opening and reclosing of the line, a lock out, or a combination. For reporting
purposes, all vegetation-related openings of a line occurring within one 24-hour period should be considered one event.
Trips known to be caused by severe weather or other natural disaster such as earthquake are excluded. Contact with
vegetation includes both physical contact and arcing due to insufficient clearance.
Approved by the Board of Trustees 8
February 10, 2004
200 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
It is important that recommendations resulting from system outages be adopted consistently by all
regions and operating entities, not just those directly affected by a particular outage. Several lessons
learned prior to August 14, if heeded, could have prevented the outage. WECC and NPCC, for
example, have programs that could be used as models for tracking completion of recommendations.
NERC and some regions have not adequately tracked completion of recommendations from prior
events to ensure they were consistently implemented.
Recommendation 5a: NERC and each regional reliability council shall establish a program for
documenting completion of recommendations resulting from the August 14 blackout and other
historical outages, as well as NERC and regional reports on violations of reliability standards, results
of compliance audits, and lessons learned from system disturbances. Regions shall report quarterly to
NERC on the status of follow-up actions to address recommendations, lessons learned, and areas
noted for improvement. NERC staff shall report both NERC activities and a summary of regional
activities to the board.
Assuring compliance with reliability standards, evaluating the reliability readiness of reliability
coordinators and control areas, and assuring recommended actions are achieved will be effective
steps in reducing the chances of future large-scale outages. However, it is important for NERC to
also adopt a process for continuous learning and improvement by seeking continuous feedback on
reliability performance trends, not rely mainly on learning from and reacting to catastrophic failures.
Such a function would assess large-scale outages and near misses to determine root causes and
lessons learned, similar to the August 14 blackout investigation. This function would incorporate the
current Disturbance Analysis Working Group and expand that work to provide more proactive
feedback to the NERC board regarding reliability performance. This program would also gather and
analyze reliability performance statistics to inform the board of reliability trends. This function could
develop procedures and capabilities to initiate investigations in the event of future large-scale outages
or disturbances. Such procedures and capabilities would be shared between NERC and the regional
reliability councils for use as needed, with NERC and regional investigation roles clearly defined in
advance.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 201
Technical Initiatives to Minimize the Likelihood
and Impacts of Possible Future Cascading Outages
NERC found during its investigation that some reliability coordinators and control area operators had
not received adequate training in recognizing and responding to system emergencies. Most notable
was the lack of realistic simulations and drills for training and verifying the capabilities of operating
personnel. This training deficiency contributed to the lack of situational awareness and failure to
declare an emergency when operator intervention was still possible prior to the high speed portion of
the sequence of events.
Recommendation 6: All reliability coordinators, control areas, and transmission operators shall
provide at least five days per year of training and drills in system emergencies, using realistic
simulations5, for each staff person with responsibility for the real-time operation or reliability
monitoring of the bulk electric system. This system emergency training is in addition to other
training requirements. Five days of system emergency training and drills are to be completed
prior to June 30, 2004, with credit given for documented training already completed since July
1, 2003. Training documents, including curriculum, training methods, and individual training
records, are to be available for verification during reliability readiness audits.
NERC has published Continuing Education Criteria specifying appropriate qualifications for
continuing education providers and training activities.
In the longer term, the NERC Personnel Certification Governance Committee (PCGC), which is
independent of the NERC board, should explore expanding the certification requirements of system
operating personnel to include additional measures of competency in recognizing and responding to
system emergencies. The current NERC certification examination is a written test of the NERC
Operating Manual and other references relating to operator job duties, and is not by itself intended to
be a complete demonstration of competency to handle system emergencies.
The August 14 blackout investigation identified inconsistent practices in northeastern Ohio with
regard to the setting and coordination of voltage limits and insufficient reactive power supply.
Although the deficiency of reactive power supply in northeastern Ohio did not directly cause the
blackout, it was a contributing factor and was a significant violation of existing reliability standards.
In particular, there appear to have been violations of NERC Planning Standard I.D.S1 requiring static
and dynamic reactive power resources to meet the performance criteria specified in Table I of
5
The term “realistic simulations” includes a variety of tools and methods that present operating personnel with situations
to improve and test diagnostic and decision-making skills in an environment that resembles expected conditions during a
particular type of system emergency. Although a full replica training simulator is one approach, lower cost alternatives
such as PC-based simulators, tabletop drills, and simulated communications can be effective training aids if used
properly.
Approved by the Board of Trustees 10
February 10, 2004
202 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Planning Standard I.A on Transmission Systems. Planning Standard II.B.S1 requires each regional
reliability council to establish procedures for generating equipment data verification and testing,
including reactive power capability. Planning Standard III.C.S1 requires that all synchronous
generators connected to the interconnected transmission systems shall be operated with their
excitation system in the automatic voltage control mode unless approved otherwise by the
transmission system operator. S2 of this standard also requires that generators shall maintain a
network voltage or reactive power output as required by the transmission system operator within the
reactive capability of the units.
On one hand, the unsafe conditions on August 14 with respect to voltage in northeastern Ohio can be
said to have resulted from violations of NERC planning criteria for reactive power and voltage
control, and those violations should have been identified through the NERC and ECAR compliance
monitoring programs (addressed by Recommendation 2). On the other hand, investigators believe
these deficiencies are also symptomatic of a systematic breakdown of the reliability studies and
practices in FE and the ECAR region that allowed unsafe voltage criteria to be set and used in study
models and operations. There were also issues identified with reactive characteristics of loads, as
addressed in Recommendation 14.
Recommendation 7a: The Planning Committee shall reevaluate within one year the
effectiveness of the existing reactive power and voltage control standards and how they are
being implemented in practice in the ten NERC regions. Based on this evaluation, the Planning
Committee shall recommend revisions to standards or process improvements to ensure voltage
control and stability issues are adequately addressed.
Recommendation 7b: ECAR shall no later than June 30, 2004 review its reactive power and
voltage criteria and procedures, verify that its criteria and procedures are being fully
implemented in regional and member studies and operations, and report the results to the
NERC board.
The importance of automatic control and protection systems in preventing, slowing, or mitigating the
impact of a large-scale outage cannot be stressed enough. To underscore this point, following the trip
of the Sammis-Star line at 4:06, the cascading failure into parts of eight states and two provinces,
including the trip of over 531 generating units and over 400 transmission lines, was completed in the
next eight minutes. Most of the event sequence, in fact, occurred in the final 12 seconds of the
cascade. Likewise, the July 2, 1996 failure took less than 30 seconds and the August 10, 1996 failure
took only 5 minutes. It is not practical to expect operators will always be able to analyze a massive,
complex system failure and to take the appropriate corrective actions in a matter of a few minutes.
The NERC investigators believe that two measures would have been crucial in slowing or stopping
the uncontrolled cascade on August 14:
• Better application of zone 3 impedance relays on high voltage transmission lines
• Selective use of under-voltage load shedding.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 203
First, beginning with the Sammis-Star line trip, most of the remaining line trips during the cascade
phase were the result of the operation of a zone 3 relay for a perceived overload (a combination of
high amperes and low voltage) on the protected line. If used, zone 3 relays typically act as an
overreaching backup to the zone 1 and 2 relays, and are not intentionally set to operate on a line
overload. However, under extreme conditions of low voltages and large power swings as seen on
August 14, zone 3 relays can operate for overload conditions and propagate the outage to a wider area
by essentially causing the system to “break up”. Many of the zone 3 relays that operated during the
August 14 cascading outage were not set with adequate margins above their emergency thermal
ratings. For the short times involved, thermal heating is not a problem and the lines should not be
tripped for overloads. Instead, power system protection devices should be set to address the specific
condition of concern, such as a fault, out-of-step condition, etc., and should not compromise a power
system’s inherent physical capability to slow down or stop a cascading event.
Recommendation 8a: All transmission owners shall, no later than September 30, 2004, evaluate
the zone 3 relay settings on all transmission lines operating at 230 kV and above for the
purpose of verifying that each zone 3 relay is not set to trip on load under extreme emergency
conditions6. In each case that a zone 3 relay is set so as to trip on load under extreme
conditions, the transmission operator shall reset, upgrade, replace, or otherwise mitigate the
overreach of those relays as soon as possible and on a priority basis, but no later than
December 31, 2005. Upon completing analysis of its application of zone 3 relays, each
transmission owner may no later than December 31, 2004 submit justification to NERC for
applying zone 3 relays outside of these recommended parameters. The Planning Committee
shall review such exceptions to ensure they do not increase the risk of widening a cascading
failure of the power system.
A second key finding with regard to system protection was that if an automatic under-voltage load
shedding scheme had been in place in the Cleveland-Akron area on August 14, there is a high
probability the outage could have been limited to that area.
Recommendation 8b: Each regional reliability council shall complete an evaluation of the
feasibility and benefits of installing under-voltage load shedding capability in load centers
within the region that could become unstable as a result of being deficient in reactive power
following credible multiple-contingency events. The regions are to complete the initial studies
and report the results to NERC within one year. The regions are requested to promote the
installation of under-voltage load shedding capabilities within critical areas, as determined by
the studies to be effective in preventing an uncontrolled cascade of the power system.
The NERC investigation of the August 14 blackout has identified additional transmission and
generation control and protection issues requiring further analysis. One concern is that generating
unit control and protection schemes need to consider the full range of possible extreme system
conditions, such as the low voltages and low and high frequencies experienced on August 14. The
team also noted that improvements may be needed in under-frequency load shedding and its
coordination with generator under-and over-frequency protection and controls.
6
The NERC investigation team recommends that the zone 3 relay, if used, should not operate at or below 150% of the
emergency ampere rating of a line, assuming a .85 per unit voltage and a line phase angle of 30 degrees.
Approved by the Board of Trustees 12
February 10, 2004
204 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Recommendation 8c: The Planning Committee shall evaluate Planning Standard III – System
Protection and Control and propose within one year specific revisions to the criteria to
adequately address the issue of slowing or limiting the propagation of a cascading failure. The
board directs the Planning Committee to evaluate the lessons from August 14 regarding relay
protection design and application and offer additional recommendations for improvement.
Ambiguities in the NERC operating policies may have allowed entities involved in the August 14
blackout to make different interpretations regarding the functions, responsibilities, capabilities, and
authorities of reliability coordinators and control areas. Characteristics and capabilities necessary to
enable prompt recognition and effective response to system emergencies must be specified.
The lack of timely and accurate outage information resulted in degraded performance of state
estimator and reliability assessment functions on August 14. There is a need to review options for
sharing of outage information in the operating time horizon (e.g. 15 minutes or less), so as to ensure
the accurate and timely sharing of outage data necessary to support real-time operating tools such as
state estimators, real-time contingency analysis, and other system monitoring tools.
On August 14, reliability coordinator and control area communications regarding conditions in
northeastern Ohio were ineffective, and in some cases confusing. Ineffective communications
contributed to a lack of situational awareness and precluded effective actions to prevent the cascade.
Consistent application of effective communications protocols, particularly during emergencies, is
essential to reliability. Alternatives should be considered to one-on-one phone calls during an
emergency to ensure all parties are getting timely and accurate information with a minimum number
of calls.
NERC operating policies do not adequately specify critical facilities, leaving ambiguity regarding
which facilities must be monitored by reliability coordinators. Nor do the policies adequately define
criteria for declaring transmission system emergencies. Operating policies should also clearly specify
that curtailing interchange transactions through the NERC Transmission Loading Relief (TLR)
Procedure is not intended as a method for restoring the system from an actual Operating Security
Limit violation to a secure operating state.
Recommendation 9: The Operating Committee shall complete the following by June 30,
2004:
• Evaluate and revise the operating policies and procedures, or provide interpretations,
to ensure reliability coordinator and control area functions, responsibilities, and
authorities are completely and unambiguously defined.
• Evaluate and improve the tools and procedures for operator and reliability
coordinator communications during emergencies.
• Evaluate and improve the tools and procedures for the timely exchange of outage
information among control areas and reliability coordinators.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 205
Recommendation 10. Establish Guidelines for Real-Time Operating Tools.
The August 14 blackout was caused by a lack of situational awareness that was in turn the result of
inadequate reliability tools and backup capabilities. Additionally, the failure of FE’s control
computers and alarm system contributed directly to the lack of situational awareness. Likewise,
MISO’s incomplete tool set and the failure of its state estimator to work effectively on August 14
contributed to the lack of situational awareness.
Recommendation 10: The Operating Committee shall within one year evaluate the real-time
operating tools necessary for reliable operation and reliability coordination, including backup
capabilities. The Operating Committee is directed to report both minimum acceptable
capabilities for critical reliability functions and a guide of best practices.
The efforts to restore the power system and customer service following the outage were effective,
considering the massive amount of load lost and the large number of generators and transmission
lines that tripped. Fortunately, the restoration was aided by the ability to energize transmission from
neighboring systems, thereby speeding the recovery. Despite the apparent success of the restoration
effort, it is important to evaluate the results in more detail to determine opportunities for
improvement. Blackstart and restoration plans are often developed through study of simulated
conditions. Robust testing of live systems is difficult because of the risk of disturbing the system or
interrupting customers. The August 14 blackout provides a valuable opportunity to apply actual
events and experiences to learn to better prepare for system blackstart and restoration in the future.
That opportunity should not be lost, despite the relative success of the restoration phase of the outage.
Recommendation 11a: The Planning Committee, working in conjunction with the Operating
Committee, NPCC, ECAR, and PJM, shall evaluate the black start and system restoration
performance following the outage of August 14, and within one year report to the NERC board
the results of that evaluation with recommendations for improvement.
206 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Recommendation 11b: All regional reliability councils shall, within six months of the Planning
Committee report to the NERC board, reevaluate their procedures and plans to assure an
effective blackstart and restoration capability within their region.
A valuable lesson from the August 14 blackout is the importance of having time-synchronized system
data recorders. NERC investigators labored over thousands of data items to synchronize the
sequence of events, much like putting together small pieces of a very large puzzle. That process
would have been significantly improved and sped up if there had been a sufficient number of
synchronized data recording devices.
NERC Planning Standard I.F – Disturbance Monitoring does require location of recording devices for
disturbance analysis. Often time, recorders are available, but they are not synchronized to a time
standard. All digital fault recorders, digital event recorders, and power system disturbance recorders
should be time stamped at the point of observation with a precise Global Positioning Satellite (GPS)
synchronizing signal. Recording and time-synchronization equipment should be monitored and
calibrated to assure accuracy and reliability.
Time-synchronized devices, such as phasor measurement units, can also be beneficial for monitoring
a wide-area view of power system conditions in real-time, such as demonstrated in WECC with their
Wide-Area Monitoring System (WAMS).
Recommendation 12a: The reliability regions, coordinated through the NERC Planning
Committee, shall within one year define regional criteria for the application of synchronized
recording devices in power plants and substations. Regions are requested to facilitate the
installation of an appropriate number, type and location of devices within the region as soon as
practical to allow accurate recording of future system disturbances and to facilitate
benchmarking of simulation studies by comparison to actual disturbances.
Recommendation 12b: Facilities owners shall, in accordance with regional criteria, upgrade
existing dynamic recorders to include GPS time synchronization and, as necessary, install
additional dynamic recorders.
The investigation report noted that FE entered the day on August 14 with insufficient resources to
stay within operating limits following a credible set of contingencies, such as the loss of the East
Lake 5 unit and the Chamberlin-Harding line. NERC will conduct an evaluation of operations
planning practices and criteria to ensure expected practices are sufficient and well understood. The
review will reexamine fundamental operating criteria, such as n-1 and the 30-minute limit in
preparing the system for a next contingency, and Table I Category C.3 of the NERC planning
standards. Operations planning and operating criteria will be identified that are sufficient to ensure
the system is in a known and reliable condition at all times, and that positive controls, whether
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 207
manual or automatic, are available and appropriately located at all times to return the Interconnection
to a secure condition. Daily operations planning, and subsequent real time operations planning will
identify available system reserves to meet operating criteria.
Recommendation 13a: The Operating Committee shall evaluate operations planning and
operating criteria and recommend revisions in a report to the board within one year.
Prior studies in the ECAR region did not adequately define the system conditions that were observed
on August 14. Severe contingency criteria were not adequate to address the events of August 14 that
led to the uncontrolled cascade. Also, northeastern Ohio was found to have insufficient reactive
support to serve its loads and meet import criteria. Instances were also noted in the FE system and
ECAR area of different ratings being used for the same facility by planners and operators and among
entities, making the models used for system planning and operation suspect. NERC and the regional
reliability councils must take steps to assure facility ratings are being determined using consistent
criteria and being effectively shared and reviewed among entities and among planners and operators.
Recommendation 13b: ECAR shall no later than June 30, 2004 reevaluate its planning and
study procedures and practices to ensure they are in compliance with NERC standards, ECAR
Document No. 1, and other relevant criteria; and that ECAR and its members’ studies are
being implemented as required.
Recommendation 13c: The Planning Committee, working in conjunction with the regional
reliability councils, shall within two years reevaluate the criteria, methods and practices used
for system design, planning and analysis; and shall report the results and recommendations to
the NERC board. This review shall include an evaluation of transmission facility ratings
methods and practices, and the sharing of consistent ratings information.
Regional reliability councils may consider assembling a regional database that includes the ratings of
all bulk electric system (100 kV and higher voltage) transmission lines, transformers, phase angle
regulators, and phase shifters. This database should be shared with neighboring regions as needed for
system planning and analysis.
NERC and the regional reliability councils should review the scope, frequency, and coordination of
interregional studies, to include the possible need for simultaneous transfer studies. Study criteria
will be reviewed, particularly the maximum credible contingency criteria used for system analysis.
Each control area will be required to identify, for both the planning and operating time horizons, the
planned emergency import capabilities for each major load area.
Recommendation 14. Improve System Modeling Data and Data Exchange Practices.
The after-the-fact models developed to simulate August 14 conditions and events indicate that
dynamic modeling assumptions, including generator and load power factors, used in planning and
operating models were inaccurate. Of particular note, the assumptions of load power factor were
overly optimistic (loads were absorbing much more reactive power than pre-August 14 models
indicated). Another suspected problem is modeling of shunt capacitors under depressed voltage
208 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
conditions. Regional reliability councils should establish regional power system models that enable
the sharing of consistent, validated data among entities in the region. Power flow and transient
stability simulations should be periodically compared (benchmarked) with actual system events to
validate model data. Viable load (including load power factor) and generator testing programs are
necessary to improve agreement between power flows and dynamic simulations and the actual system
performance.
Recommendation 14: The regional reliability councils shall within one year establish and begin
implementing criteria and procedures for validating data used in power flow models and
dynamic simulations by benchmarking model data with actual system performance. Validated
modeling data shall be exchanged on an inter-regional basis as needed for reliable system
planning and operation.
During the data collection phase of the blackout investigation, when control areas were asked for
information pertaining to merchant generation within their area, data was frequently not supplied.
The reason often given was that the control area did not know the status or output of the generator at
a given point in time. Another reason was the commercial sensitivity or confidentiality of such data.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 209
Appendix E
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 211
Appendix F
Electricity Glossary
AC: Alternating current; current that changes peri- often used to compensate for reactive load and
odically (sinusoidally) with time. help support system voltage. A bank is a collection
of one or more capacitors at a single location.
ACE: Area Control Error in MW. A negative value
indicates a condition of under-generation relative Capacity: The rated continuous load-carrying
to system load and imports, and a positive value ability, expressed in megawatts (MW) or
denotes over-generation. megavolt-amperes (MVA) of generation, transmis-
Active Power: See “Real Power.” sion, or other electrical equipment.
Adequacy: The ability of the electric system to Cascading: The uncontrolled successive loss of
supply the aggregate electrical demand and energy system elements triggered by an incident. Cas-
requirements of customers at all times, taking into cading results in widespread service interruption,
account scheduled and reasonably expected which cannot be restrained from sequentially
unscheduled outages of system elements. spreading beyond an area predetermined by
appropriate studies.
AGC: Automatic Generation Control is a computa-
tion based on measured frequency and computed Circuit: A conductor or a system of conductors
economic dispatch. Generation equipment under through which electric current flows.
AGC automatically responds to signals from an Circuit Breaker: A switching device connected to
EMS computer in real time to adjust power output the end of a transmission line capable of opening
in response to a change in system frequency, or closing the circuit in response to a command,
tie-line loading, or to a prescribed relation usually from a relay.
between these quantities. Generator output is
adjusted so as to maintain a target system fre- Control Area: An electric power system or combi-
quency (usually 60 Hz) and any scheduled MW nation of electric power systems to which a com-
interchange with other areas. mon automatic control scheme is applied in order
to: (1) match, at all times, the power output of the
Apparent Power: The product of voltage and cur- generators within the electric power system(s) and
rent phasors. It comprises both active and reactive capacity and energy purchased from entities out-
power, usually expressed in kilovoltamperes side the electric power system(s), with the load in
(kVA) or megavoltamperes (MVA). the electric power system(s); (2) maintain, within
Blackstart Capability: The ability of a generating the limits of Good Utility Practice, scheduled
unit or station to go from a shutdown condition to interchange with other Control Areas; (3) main-
an operating condition and start delivering power tain the frequency of the electric power system(s)
without assistance from the bulk electric system. within reasonable limits in accordance with
Good Utility Practice; and (4) provide sufficient
Bulk Electric System: A term commonly applied
generating capacity to maintain operating reserves
to the portion of an electric utility system that
in accordance with Good Utility Practice.
encompasses the electrical generation resources
and bulk transmission system. Contingency: The unexpected failure or outage of
a system component, such as a generator, trans-
Bulk Transmission: A functional or voltage classi-
mission line, circuit breaker, switch, or other elec-
fication relating to the higher voltage portion of
trical element. A contingency also may include
the transmission system, specifically, lines at or
multiple components, which are related by situa-
above a voltage level of 115 kV.
tions leading to simultaneous component outages.
Bus: Shortened from the word busbar, meaning a
Control Area Operator: An individual or organi-
node in an electrical network where one or more
zation responsible for controlling generation to
elements are connected together.
maintain interchange schedule with other control
Capacitor Bank: A capacitor is an electrical device areas and contributing to the frequency regulation
that provides reactive power to the system and is of the interconnection. The control area is an
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 213
electric system that is bounded by interconnec- owned (by a federal agency, crown corporation,
tion metering and telemetry. State, provincial government, municipal govern-
ment, and public power district).
Current (Electric): The rate of flow of electrons in
an electrical conductor measured in Amperes. Element: Any electric device with terminals that
may be connected to other electric devices, such
Curtailability: The right of a transmission pro-
as a generator, transformer, circuit, circuit
vider to interrupt all or part of a transmission ser-
breaker, or bus section.
vice due to constraints that reduce the capability
of the transmission network to provide that trans- Energy Emergency: A condition when a system or
mission service. Transmission service is to be cur- power pool does not have adequate energy
tailed only in cases where system reliability is resources (including water for hydro units) to sup-
threatened or emergency conditions exist. ply its customers’ expected energy requirements.
Demand: The rate at which electric energy is Emergency: Any abnormal system condition that
delivered to consumers or by a system or part of a requires automatic or immediate manual action to
system, generally expressed in kilowatts or mega- prevent or limit loss of transmission facilities or
watts, at a given instant or averaged over any des- generation supply that could adversely affect the
ignated interval of time. Also see “Load.” reliability of the electric system.
DC: Direct current; current that is steady and does Emergency Voltage Limits: The operating voltage
not change sinusoidally with time (see “AC”). range on the interconnected systems that is
acceptable for the time, sufficient for system
Dispatch Operator: Control of an integrated elec-
adjustments to be made following a facility outage
tric system involving operations such as assign-
or system disturbance.
ment of levels of output to specific generating
stations and other sources of supply; control of EMS: An energy management system is a com-
transmission lines, substations, and equipment; puter control system used by electric utility dis-
operation of principal interties and switching; and patchers to monitor the real time performance of
scheduling of energy transactions. various elements of an electric system and to con-
trol generation and transmission facilities.
Distribution: For electricity, the function of dis-
tributing electric power using low voltage lines to Fault: A fault usually means a short circuit, but
retail customers. more generally it refers to some abnormal system
condition. Faults are often random events.
Distribution Network: The portion of an electric
system that is dedicated to delivering electric Federal Energy Regulatory Commission (FERC):
energy to an end user, at or below 69 kV. The dis- Independent Federal agency that, among other
tribution network consists primarily of low- responsibilities, regulates the transmission and
voltage lines and transformers that “transport” wholesale sales of electricity in interstate
electricity from the bulk power system to retail commerce.
customers. Flashover: A plasma arc initiated by some event
Disturbance: An unplanned event that produces such as lightning. Its effect is a short circuit on the
an abnormal system condition. network.
Electrical Energy: The generation or use of elec- Flowgate: A single or group of transmission ele-
tric power by a device over a period of time, ments intended to model MW flow impact relating
expressed in kilowatthours (kWh), megawatt- to transmission limitations and transmission ser-
hours (MWh), or gigawatthours (GWh). vice usage.
Forced Outage: The removal from service avail-
Electric Utility: Person, agency, authority, or
ability of a generating unit, transmission line, or
other legal entity or instrumentality that owns or
other facility for emergency reasons or a condition
operates facilities for the generation, transmis-
in which the equipment is unavailable due to
sion, distribution, or sale of electric energy pri-
unanticipated failure.
marily for use by the public, and is defined as a
utility under the statutes and rules by which it is Frequency: The number of complete alternations
regulated. An electric utility can be investor- or cycles per second of an alternating current,
owned, cooperatively owned, or government- measured in Hertz. The standard frequency in the
214 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
United States is 60 Hz. In some other countries the usually not-for-profit and can advise utilities
standard is 50 Hz. within its territory on transmission expansion and
maintenance but does not have the responsibility
Frequency Deviation or Error: A departure from
to carry out the functions.
scheduled frequency; the difference between
actual system frequency and the scheduled sys- Interchange: Electric power or energy that flows
tem frequency. across tie-lines from one entity to another,
whether scheduled or inadvertent.
Frequency Regulation: The ability of a Control
Area to assist the interconnected system in main- Interconnected System: A system consisting of
taining scheduled frequency. This assistance can two or more individual electric systems that nor-
include both turbine governor response and auto- mally operate in synchronism and have connect-
matic generation control. ing tie lines.
Frequency Swings: Constant changes in fre- Interconnection: When capitalized, any one of the
quency from its nominal or steady-state value. five major electric system networks in North
America: Eastern, Western, ERCOT (Texas), Qué-
Generation (Electricity): The process of produc- bec, and Alaska. When not capitalized, the facili-
ing electrical energy from other forms of energy; ties that connect two systems or Control Areas.
also, the amount of electric energy produced, usu- Additionally, an interconnection refers to the
ally expressed in kilowatt hours (kWh) or mega- facilities that connect a nonutility generator to a
watt hours (MWh). Control Area or system.
Generator: Generally, an electromechanical Interface: The specific set of transmission ele-
device used to convert mechanical power to elec- ments between two areas or between two areas
trical power. comprising one or more electrical systems.
Grid: An electrical transmission and/or distribu- ISAC: Information Sharing and Analysis Centers
tion network. (ISACs) are designed by the private sector and
Grid Protection Scheme: Protection equipment serve as a mechanism for gathering, analyzing,
for an electric power system, consisting of circuit appropriately sanitizing and disseminating pri-
breakers, certain equipment for measuring electri- vate sector information. These centers could also
cal quantities (e.g., current and voltage sensors) gather, analyze, and disseminate information from
and devices called relays. Each relay is designed to Government for further distribution to the private
protect the piece of equipment it has been sector. ISACs also are expected to share important
assigned from damage. The basic philosophy in information about vulnerabilities, threats, intru-
protection system design is that any equipment sions, and anomalies, but do not interfere with
that is threatened with damage by a sustained direct information exchanges between companies
fault is to be automatically taken out of service. and the Government.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 215
metering and billing electricity customers. The capable of handling additional applications
retail price for a kWh varies from approximately 4 between participants.
cents to 15 cents. At a 100% conversion efficiency,
one kWh is equivalent to about 4 fluid ounces of Normal (Precontingency) Operating Procedures:
gasoline, 3/16 pound of liquid petroleum, 3 cubic Operating procedures that are normally invoked
feet of natural gas, or 1/4 pound of coal. by the system operator to alleviate potential facil-
ity overloads or other potential system problems
Line Trip: Refers to the automatic opening of the in anticipation of a contingency.
conducting path provided by a transmission line
by the circuit breakers. These openings or “trips” Normal Voltage Limits: The operating voltage
are to protect the transmission line during faulted range on the interconnected systems that is
conditions. acceptable on a sustained basis.
Load (Electric): The amount of electric power North American Electric Reliability Council
delivered or required at any specific point or (NERC): A not-for-profit company formed by the
points on a system. The requirement originates at electric utility industry in 1968 to promote the
the energy-consuming equipment of the consum- reliability of the electricity supply in North Amer-
ers. See “Demand.” ica. NERC consists of nine Regional Reliability
Councils and one Affiliate, whose members
Load Shedding: The process of deliberately account for virtually all the electricity supplied in
removing (either manually or automatically) pre- the United States, Canada, and a portion of Baja
selected customer demand from a power system in California Norte, Mexico. The members of these
response to an abnormal condition, to maintain Councils are from all segments of the electricity
the integrity of the system and minimize overall supply industry: investor-owned, federal, rural
customer outages. electric cooperative, state/municipal, and provin-
cial utilities, independent power producers, and
Lockout: A state of a transmission line following
power marketers. The NERC Regions are: East
breaker operations where the condition detected
Central Area Reliability Coordination Agreement
by the protective relaying was not eliminated by
(ECAR); Electric Reliability Council of Texas
temporarily opening and reclosing the line, possi-
(ERCOT); Mid-Atlantic Area Council (MAAC);
bly several times. In this state, the circuit breakers
Mid-America Interconnected Network (MAIN);
cannot generally be reclosed without resetting a
Mid-Continent Area Power Pool (MAPP); North-
lockout device.
east Power Coordinating Council (NPCC); South-
Market Participant: An entity participating in the eastern Electric Reliability Council (SERC);
energy marketplace by buying/selling transmis- Southwest Power Pool (SPP); Western Systems
sion rights, energy, or ancillary services into, out Coordinating Council (WSCC); and Alaskan Sys-
of, or through an ISO-controlled grid. tems Coordination Council (ASCC, Affiliate).
Megawatthour (MWh): One million watthours. OASIS: Open Access Same Time Information Ser-
vice (OASIS), developed by the Electric Power
Metered Value: A measured electrical quantity Research Institute, is designed to facilitate open
that may be observed through telemetering, super- access by providing users with access to informa-
visory control and data acquisition (SCADA), or tion on transmission services and availability,
other means. plus facilities for transactions.
Metering: The methods of applying devices that Operating Criteria: The fundamental principles
measure and register the amount and direction of of reliable interconnected systems operation,
electrical quantities with respect to time. adopted by NERC.
NERC Interregional Security Network (ISN): A Operating Guides: Operating practices that a Con-
communications network used to exchange elec- trol Area or systems functioning as part of a Con-
tric system operating parameters in near real time trol Area may wish to consider. The application of
among those responsible for reliable operations of Guides is optional and may vary among Control
the electric system. The ISN provides timely and Areas to accommodate local conditions and indi-
accurate data and information exchange among vidual system requirements.
reliability coordinators and other system opera-
tors. The ISN, which operates over the frame relay Operating Policies: The doctrine developed for
NERCnet system, is a private Intranet that is interconnected systems operation. This doctrine
216 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
consists of Criteria, Standards, Requirements, subgroups, and the Regional Councils to achieve
Guides, and instructions, which apply to all Con- bulk electric system reliability.
trol Areas.
Post-contingency Operating Procedures: Oper-
Operating Procedures: A set of policies, practices, ating procedures that may be invoked by the sys-
or system adjustments that may be automatically tem operator to mitigate or alleviate system
or manually implemented by the system operator problems after a contingency has occurred.
within a specified time frame to maintain the
Protective Relay: A device designed to detect
operational integrity of the interconnected electric
abnormal system conditions, such as electrical
systems.
shorts on the electric system or within generating
Operating Requirements: Obligations of a Control plants, and initiate the operation of circuit break-
Area and systems functioning as part of a Control ers or other control equipment.
Area.
Power/Phase Angle: The angular relationship
Operating Security Limit: The value of a system between an AC (sinusoidal) voltage across a cir-
operating parameter (e.g. total power transfer cuit element and the AC (sinusoidal) current
across an interface) that satisfies the most limiting through it. The real power that can flow is related
of prescribed pre- and post-contingency operating to this angle.
criteria as determined by equipment loading capa- Power: See “Real Power.”
bility and acceptable stability and voltage condi-
tions. It is the operating limit to be observed so Power Flow: See “Current.”
that the transmission system will remain reliable
Rate: The authorized charges per unite or level of
even if the worst contingency occurs.
consumption for a specified time period for any of
Operating Standards: The obligations of a Control the classes of utility services provided to a
Area and systems functioning as part of a Control customer.
Area that are measurable. An Operating Standard
Rating: The operational limits of an electric sys-
may specify monitoring and surveys for
tem, facility, or element under a set of specified
compliance.
conditions.
Outage: The period during which a generating Reactive Power: The portion of electricity that
unit, transmission line, or other facility is out of establishes and sustains the electric and magnetic
service. fields of alternating-current equipment. Reactive
Planning Guides: Good planning practices and power must be supplied to most types of magnetic
considerations that Regions, subregions, power equipment, such as motors and transformers. It
pools, or individual systems should follow. The also must supply the reactive losses on transmis-
application of Planning Guides may vary to match sion facilities. Reactive power is provided by gen-
local conditions and individual system erators, synchronous condensers, or electrostatic
requirements. equipment such as capacitors and directly influ-
ences electric system voltage. It is usually
Planning Policies: The framework for the reliabil- expressed in kilovars (kVAr) or megavars (MVAr),
ity of interconnected bulk electric supply in terms and is the mathematical product of voltage and
of responsibilities for the development of and con- current consumed by reactive loads. Examples of
formance to NERC Planning Principles and reactive loads include capacitors and inductors.
Guides and Regional planning criteria or guides, These types of loads, when connected to an ac
and NERC and Regional issues resolution pro- voltage source, will draw current, but because the
cesses. NERC Planning Procedures, Principles, current is 90 degrees out of phase with the applied
and Guides emanate from the Planning Policies. voltage, they actually consume no real power.
Planning Principles: The fundamental character- Readiness: The extent to which an organizational
istics of reliable interconnected bulk electric sys- entity is prepared to meet the functional require-
tems and the tenets for planning them. ments set by NERC or its regional council for enti-
ties of that type or class.
Planning Procedures: An explanation of how
the Planning Policies are addressed and imple- Real Power: Also known as “active power.” The
mented by the NERC Engineering Committee, its rate at which work is performed or that energy is
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 217
transferred, usually expressed in kilowatts (kW) or establishes a security policy for these Control
megawatts (MW). The terms “active power” or Areas and their interconnections, and coordinates
“real power” are often used in place of the term emergency operating procedures that rely on com-
power alone to differentiate it from reactive mon operating terminology, criteria, and
power. standards.
Real-Time Operations: The instantaneous opera- Resistance: The characteristic of materials to
tions of a power system as opposed to those opera- restrict the flow of current in an electric circuit.
tions that are simulated. Resistance is inherent in any electric wire, includ-
Regional Reliability Council: One of ten Electric ing those used for the transmission of electric
Reliability Councils that form the North American power. Resistance in the wire is responsible for
Electric Reliability Council (NERC). heating the wire as current flows through it and
the subsequent power loss due to that heating.
Regional Transmission Operator (RTO): An orga-
nization that is independent from all generation Restoration: The process of returning generators
and power marketing interests and has exclusive and transmission system elements and restoring
responsibility for electric transmission grid opera- load following an outage on the electric system.
tions, short-term electric reliability, and transmis- Right-of-Way (ROW) Maintenance: Activities by
sion services within a multi-State region. To utilities to maintain electrical clearances along
achieve those objectives, the RTO manages trans- transmission or distribution lines.
mission facilities owned by different companies
and encompassing one, large, contiguous geo- Safe Limits: System limits on quantities such as
graphic area. voltage or power flows such that if the system is
operated within these limits it is secure and
Regulations: Rules issued by regulatory authori- reliable.
ties to implement laws passed by legislative
bodies. SCADA: Supervisory Control and Data Acquisi-
tion system; a system of remote control and telem-
Relay: A device that controls the opening and sub- etry used to monitor and control the electric
sequent reclosing of circuit breakers. Relays take system.
measurements from local current and voltage
transformers, and from communication channels Schedule: An agreed-upon transaction size (mega-
connected to the remote end of the lines. A relay watts), start and end time, beginning and ending
output trip signal is sent to circuit breakers when ramp times and rate, and type required for deliv-
needed. ery and receipt of power and energy between the
contracting parties and the Control Area(s)
Relay Setting: The parameters that determine involved in the transaction.
when a protective relay will initiate operation of
circuit breakers or other control equipment. Scheduling Coordinator: An entity certified by an
ISO or RTO for the purpose of undertaking sched-
Reliability: The degree of performance of the ele-
uling functions.
ments of the bulk electric system that results in
electricity being delivered to customers within Seams: The boundaries between adjacent electric-
accepted standards and in the amount desired. ity-related organizations. Differences in regulatory
Reliability may be measured by the frequency, requirements or operating practices may create
duration, and magnitude of adverse effects on the “seams problems.”
electric supply. Electric system reliability can be
Security: The ability of the electric system to with-
addressed by considering two basic and func-
stand sudden disturbances such as electric short
tional aspects of the electric system, Adequacy
circuits or unanticipated loss of system elements.
and Security.
Security Coordinator: An individual or organiza-
Reliability Coordinator: An individual or organi-
tion that provides the security assessment and
zation responsible for the safe and reliable opera-
emergency operations coordination for a group of
tion of the interconnected transmission system for
Control Areas.
their defined area, in accordance with NERC reli-
ability standards, regional criteria, and subregion- Short Circuit: A low resistance connection unin-
al criteria and practices. This entity facilitates the tentionally made between points of an electrical
sharing of data and information about the status circuit, which may result in current flow far above
of the Control Areas for which it is responsible, normal levels.
218 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Shunt Capacitor Bank: Shunt capacitors are Subtransmission: A functional or voltage classifi-
capacitors connected from the power system to an cation relating to lines at voltage levels between
electrical ground. They are used to supply kilovars 69kV and 115kV.
(reactive power) to the system at the point where Supervisory Control and Data Acquisition
they are connected. A shunt capacitor bank is a (SCADA): See SCADA.
group of shunt capacitors.
Surge: A transient variation of current, voltage, or
Single Contingency: The sudden, unexpected fail- power flow in an electric circuit or across an elec-
ure or outage of a system facility(s) or element(s) tric system.
(generating unit, transmission line, transformer, Surge Impedance Loading: The maximum
etc.). Elements removed from service as part of the amount of real power that can flow down a
operation of a remedial action scheme are consid- lossless transmission line such that the line does
ered part of a single contingency. not require any VArs to support the flow.
Special Protection System: An automatic protec- Switching Station: Facility equipment used to tie
tion system designed to detect abnormal or prede- together two or more electric circuits through
termined system conditions, and take corrective switches. The switches are selectively arranged to
actions other than and/or in addition to the isola- permit a circuit to be disconnected, or to change
tion of faulted components. the electric connection between the circuits.
Synchronize: The process of connecting two pre-
Stability: The ability of an electric system to main- viously separated alternating current apparatuses
tain a state of equilibrium during normal and after matching frequency, voltage, phase angles,
abnormal system conditions or disturbances. etc. (e.g., paralleling a generator to the electric
Stability Limit: The maximum power flow possi- system).
ble through a particular point in the system while System: An interconnected combination of gener-
maintaining stability in the entire system or the ation, transmission, and distribution components
part of the system to which the stability limit comprising an electric utility and independent
refers. power producer(s) (IPP), or group of utilities and
IPP(s).
State Estimator: Computer software that takes
System Operator: An individual at an electric sys-
redundant measurements of quantities related to
tem control center whose responsibility it is to
system state as input and provides an estimate of
monitor and control that electric system in real
the system state (bus voltage phasors). It is used to
time.
confirm that the monitored electric power system
is operating in a secure state by simulating the sys- System Reliability: A measure of an electric sys-
tem both at the present time and one step ahead, tem’s ability to deliver uninterrupted service at
for a particular network topology and loading con- the proper voltage and frequency.
dition. With the use of a state estimator and its Thermal Limit: A power flow limit based on the
associated contingency analysis software, system possibility of damage by heat. Heating is caused by
operators can review each critical contingency to the electrical losses which are proportional to the
determine whether each possible future state is square of the real power flow. More precisely, a
within reliability limits. thermal limit restricts the sum of the squares of
real and reactive power.
Station: A node in an electrical network where
one or more elements are connected. Examples Tie-line: The physical connection (e.g. transmis-
include generating stations and substations. sion lines, transformers, switch gear, etc.) between
two electric systems that permits the transfer of
Storage: Energy transferred form one entity to electric energy in one or both directions.
another entity that has the ability to conserve the Time Error: An accumulated time difference
energy (i.e., stored as water in a reservoir, coal in a between Control Area system time and the time
pile, etc.) with the intent that the energy will be standard. Time error is caused by a deviation in
returned at a time when such energy is more use- Interconnection frequency from 60.0 Hertz.
able to the original supplying entity.
Time Error Correction: An offset to the Intercon-
Substation: Facility equipment that switches, nection’s scheduled frequency to correct for the
changes, or regulates electric voltage. time error accumulated on electric clocks.
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 219
Transactions: Sales of bulk power via the trans- facilities used for the transmission of electric
mission grid. energy.
Transfer Limit: The maximum amount of power Trip: The opening of a circuit breaker or breakers
that can be transferred in a reliable manner from on an electric system, normally to electrically iso-
one area to another over all transmission lines (or late a particular element of the system to prevent it
paths) between those areas under specified system from being damaged by fault current or other
conditions. potentially damaging conditions. See “Line Trip”
Transformer: A device that operates on magnetic for example.
principles to increase (step up) or decrease (step Voltage: The electrical force, or “pressure,” that
down) voltage. causes current to flow in a circuit, measured in
Transient Stability: The ability of an electric sys- Volts.
tem to maintain synchronism between its parts
when subjected to a disturbance and to regain a Voltage Collapse (decay): An event that occurs
state of equilibrium following that disturbance. when an electric system does not have adequate
reactive support to maintain voltage stability.
Transmission: An interconnected group of lines Voltage Collapse may result in outage of system
and associated equipment for the movement or elements and may include interruption in service
transfer of electric energy between points of sup- to customers.
ply and points at which it is transformed for deliv-
ery to customers or is delivered to other electric Voltage Control: The control of transmission volt-
systems. age through adjustments in generator reactive out-
put and transformer taps, and by switching
Transmission Loading Relief (TLR): A procedure
capacitors and inductors on the transmission and
used to manage congestion on the electric trans-
distribution systems.
mission system.
Transmission Margin: The difference between Voltage Limits: A hard limit above or below which
the maximum power flow a transmission line can is an undesirable operating condition. Normal
handle and the amount that is currently flowing limits are between 95 and 105 percent of the nomi-
on the line. nal voltage at the bus under discussion.
Transmission Operator: NERC-certified party Voltage Reduction: A procedure designed to
responsible for monitoring and assessing local deliberately lower the voltage at a bus. It is often
reliability conditions, who operates the transmis- used as a means to reduce demand by lowering the
sion facilities, and who executes switching orders customer’s voltage.
in support of the Reliability Authority. Voltage Stability: The condition of an electric sys-
Transmission Overload: A state where a transmis- tem in which the sustained voltage level is con-
sion line has exceeded either a normal or emer- trollable and within predetermined limits.
gency rating of the electric conductor.
Watthour (Wh): A unit of measure of electrical
Transmission Owner (TO) or Transmission Pro- energy equal to 1 watt of power supplied to, or
vider: Any utility that owns, operates, or controls taken from, an electric circuit steadily for 1 hour.
220 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Appendix G
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 221
222 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 223
224 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 225
226 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G
Mr. James W. Glotfelty
Director, Office of Electric Transmission
and Distribution
U.S. Department of Energy
1000 Independence Avenue SW
Washington, DC 20585
Enclosed is the Final Report of the Security Working Group (SWG) supporting the United States
- Canada Power System Outage Task Force.
The SWG Final Report presents the results of the Working Group`s analysis of the security
aspects of the power outage that occurred on August 14, 2003 and provides recommendations for
Task Force consideration on security-related issues in the electricity sector. This report
comprises input from public sector, private sector, and academic members of the SWG, with
important assistance from many members of the Task Force’s investigative team. As co-chairs
of the Security Working Group, we represent all members of the SWG in this submittal and have
signed below.
Sincerely,
G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G 227
Attachment 1:
U.S.-Canada Power System Outage Task Force SWG Steering Committee members:
Bob Liscouski, Assistant Secretary for Sid Caspersen, Director, Office of Counter-Terrorism
Infrastructure Protection, Department of Homeland (New Jersey)
Security (U.S. Government) (Co-Chair)
James McMahon, Senior Advisor (New York)
William J.S. Elliott, Assistant Secretary to the
Cabinet, Security and Intelligence, Privy Council John Overly, Executive Director, Division of Homeland
Office (Government of Canada) (Co-Chair) Security (Ohio)
Steve Schmidt, Section Chief, Special Technologies James Harlick, Assistant Deputy Minister, Office of
and Applications, FBI Critical Infrastructure Protection and Emergency
Preparedness
Kevin Kolevar, Senior Policy Advisor to the Secretary,
DoE Michael Devaney, Deputy Chief, Information
Technology Security Communications Security
Simon Szykman, Senior Policy Analyst, U.S. Office of Establishment
Science &Technology Policy, White House
Peter MacAulay, Officer, Technological Crime Branch
Vincent DeRosa, Deputy Commissioner, Director of of the Royal Canadian Mounted Police
Homeland Security (Connecticut)
Gary Anderson, Chief, Counter-Intelligence – Global,
Richard Swensen, Under-Secretary, Office of Public Canadian Security Intelligence Service
Safety and Homeland Security (Massachusetts)
Dr. James Young, Commissioner of Public Security,
Colonel Michael C. McDaniel (Michigan) Ontario Ministry of Public Safety and Security
228 G U.S.-Canada Power System Outage Task Force G August 14th Blackout: Causes and Recommendations G