Partial Discharge Measurements Under DC Test Voltage: Application Note
Partial Discharge Measurements Under DC Test Voltage: Application Note
Author
Dr.-Ing. Kay Rethmeier | [email protected]
Date
Feb, 2009
Related OMICRON Product
MPD 600
Application Area
DC Test Voltage / Polarity Reversal
Version
v2.0
Document ID
ANP_10037_ENU
Content
1
Introduction ..................................................................................................................................... 3
Please use this note only in combination with the related product manual which contains several important safety
instructions. The user is responsible for every application that makes use of an OMICRON product.
OMICRON electronics GmbH including all international branch offices is henceforth referred to as OMICRON.
OMICRON 2010. All rights reserved. This application note is a publication of OMICRON.
All rights including translation reserved. Reproduction of any kind, for example, photocopying, microfilming, optical
character recognition and/or storage in electronic data processing systems, requires the explicit consent of OMICRON.
Reprinting, wholly or in part, is not permitted.
The product information, specifications, and technical data embodied in this application note represent the technical
status at the time of writing and are subject to change without prior notice.
We have done our best to ensure that the information given in this application note is useful, accurate and entirely
reliable. However, OMICRON does not assume responsibility for any inaccuracies which may be present.
OMICRON translates this application note from the source language English into a number of other languages. Any
translation of this document is done for local requirements, and in the event of a dispute between the English and a nonEnglish version, the English version of this note shall govern.
Short Cut
1.
2.
3.
4.
5.
6.
7.
8.
9.
Set up HV PD circuit as shown in IEC 60270, use additional ohmic divider for test voltage
measurement
Choose the "1"-Tab (NOT the DC-Tab) in Large Scope View, calibrate charge
Deactivate auto-gain function during calibration and for the PD measurement
Move PD threshold level clearly above ambient noise level
Activate "Go"-Button in DC-Tab to calibrate voltage measurement with V-Tab, use ohmic divider
Select "Internal Trigger" as voltage trigger source
Set up all diagrams in DC-Tab
Record stream (from the Q-Tab), if required
Export PD pulses to a TXT-File (in the DCTab) for further Excel evaluation, if required
Introduction
Partial discharge analysis is an approved tool in the spectrum of non-destructive diagnostic measurement
methods for all kinds of high voltage apparatus. As the PD pulses are of a highly complex nature, the
gathered PD data needs interpretation. Over the last decades the knowledge about interpretation of
measuring results has improved more and more. A common tool is the evaluation of PD using phaseresolved patterns, showing PD magnitude and quantity over the phase angle of the test voltage (q-n-). The
resulting PD patterns can be analyzed by experts following some rules of interpretation [1].
For DC PD measurements no phase information exists. Therefore the distinction between PD pulses and
external interference is more demanding. As the total number of PD events is mostly lower compared to AC
measurements, it becomes even more important to assign the pulses to their sources. Facing new records
in test voltage levels, as recently demonstrated by testing UHV (Ultra-High-Voltage) DC transformers of a
+/-800kV DC transmission link [2], diagnostic PD measurements become even more important with
higher dielectric field strengths.
Using a highly sophisticated fully digital PD measuring system new evaluation techniques can be realized to
solve the problem of discriminating pulse sources. With the possibility to record even long-term PD
measurements of up to several hours to ensure the generation and accumulation of space charges, an
in-depth analysis of the gathered PD data can be performed after the measurement with different
approaches, allowing a detailed concentration on the short-time period of the polarity reversal from minus to
plus, for instance, or concentrating on the rising edge of a surge voltage shape.
Evaluation of DC PD measurements
With no usable phase information the possibilities of evaluating PD activity under DC test voltage is very
limited. Therefore the main measures are the time dependency of the charge level (charge q over time t)
and the number of PD pulses within a defined time interval (Charge Count Diagram). This kinds of
diagrams can provide an overview over the whole measuring time (as 60 minutes, e.g.) or can spot on
shorter time frames allowing a closer view on short-term PD behavior. A typical diagram is shown in figure 1,
taken from the MPD 600 Expert Mode software.
Besides a stable PD activity at about 30 pC, some single pulses with higher charge levels up to 60 pC can
be seen in the upper diagram Q(t). The two lower diagrams in figure 1 show the number of PD pulses
above a specific threshold level, within a certain sliding time interval. Two of these counter diagrams can be
set with different parameters, according to the requirements of international standards, such as IEC 62199
for HV bushings [3], for instance. This standard limits the number of allowed pulses to a maximum of
10 greater than 2000 pC over the last 30 minutes of the withstand voltage test, as well as a maximum of
10 pulses greater than 2000 pC over any gliding 10 minutes window during the polarity reversal test.
In figure 1, the red curve shows the number of PD pulses with more than 1 pC charge level, counted over
the last minute. On the other hand, the second, darker curve in the lower diagram of figure 1, shows the
number of PD pulses higher than 50 pC over the last 5 minutes.
Another statistical method is the evaluation of the density function of the discharge magnitudes [4]. The
resulting diagram (see fig. 2) shows the number of PD events (or if normalized, the charge density, related to
the maximum value or to the sum of all values) over a group of discharge magnitude classes.
In this type of diagram, internal PD will show a density maximum at the smallest discharge magnitudes,
while corona discharges will have a maximum at a higher
charge level, decaying to both sides. Besides these statistical approaches, also single pulses can be
analyzed, using an oscilloscope function. Figure 3 shows a single corona pulse recorded by the MPD 600
Time Domain Scope.
Typical parameters to be analyzed are the polarity, the magnitude, the rise time, the decay time and the
pulse width, respectively. Depending on the position of the measuring impedance, the pulse polarity can be
used to distinguish between external pulses and internal PD activity.
Also in frequency domain, the pulse spectrum can be visualized for further investigation. The FFT behavior
allows discrimination of multiple PD sources and ambient noise, as shown in figure 4.
Besides looking on fast moving single curves (the red curve in fig. 4, for instance), multiple curves can be
overlaid by using a statistical FFT visualization tool. A color-coded diagram is formed, where the color
represents the spectral density. As an example, figure 4 shows a minimum of 3 different pulse types (no
ambient noise recorded, as trigger level was set above noise level).
By using 3 PD filters with different center frequencies for PD detection simultaneously, the 3CFRD method
allows a discrimination of different pulse sources [5]. Similar to the 3PARD method (where the 3 axes of the
diagram represent the 3 phases, 120 phase-shifted) the 3 Center Frequency Relation Diagram plots the PD
results of each filter to the belonging axis. The forming clusters represent single PD mechanisms, which can
be analyzed independently, also during the PD measurement in real-time.
While the red and the blue circle in figure 5 represent PD activity, the green ellipsis represents external
interference. These interference pulses can be ignored for further PD evaluation of the test sample.
4.1
HV set-up
The PD measurement circuit itself is similar for DC and AC. It consists of a test sample, a coupling capacitor
and a measuring impedance (quadripole). General information is given in the IEC 60270 [7]. As one major
difference, the coupling capacitor cannot be used to divide the test voltage. Therefore the V-output of the
quadripole must not be connected to the MPD's V-input port. A second high-voltage path is needed,
consisting of a high-voltage resistor and the MPD's input resistor of the V-input port. Figure 6 provides a
schematic overview of the test set-up.
If the measuring impedance is placed in the earth path of the test sample (instead of being placed in the
earth path of the coupling capacitor), also the polarity of the detected pulses can be taken into account to
distinguish between external noise and internal PD.
4.2
Charge calibration
To calibrate the charge acquisition a charge injector or calibrator has to be set to the desired value, as
100 pC, for instance. The two calibrator wires have to be connected to the leads of the test sample. In the
PD-Tab of the software the "Auto gain" function should be selected in order to choose an optimal
preamplifier setting.
The calibration pulses shall be visible in the Small Scope View (oscilloscope) and in the Large Scope View
(PRPD pattern) then. The PD filter settings have to be set in accordance with the valid standards, ensuring a
maximal signal-to-noise ratio (SNR). When a filter setting is chosen, the calibrator's charge value has to
be entered into the related text field "Q IEC (target)" of the software's Q-Tab (see fig. 7).
After pressing the "Compute" button (see also fig. 7, red arrow), the displayed measured charge level shall
be identical to the one injected. Figure 8 shows the Large Scope View after a successful calibration process.
The numeric QIEC value shall show the desired charge value, the PD events (vertical light blue lines) shall be
drawn up to the related grid line. In this case, the calibrator CAL 542 with a pulse repetition rate
of 300 Hz leads to 6 equidistant pulses (marked with 6 red arrows in figure 8) for a line frequency of 50 Hz.
For DC PD measurements typically a low PD repetition rate (down to only a few pulses per hour) is
expected. Therefore the "Auto gain" function from the PD-Tab of the software has to be deactivated
ALLREADY DURING THE CALIBRATION PROCESS, as to be shown in figure 9.
The slider of the pre-amplifier then remains at a fix position for the whole measuring time. This ensures that
PD pulses of the expected charge level are measured correctly at any time. With the "Auto gain" function
activated (as usually when measuring PD at AC test voltage) the pre-amplifier would try to improve the SNR
by switching to a higher slider position if no PD pulses were detected over the last few seconds. As a
consequence it is not possible to measure the first following PD pulse (after a period without PD activity)
with the correct charge level as the input of the preamplifier is over driven. With only a few pulses
expected for the whole PD testing time it is of highest importance to measure every single PD pulse with the
correct charge value. After the calibration process the calibrator has to be turned off and removed.
4.3
Voltage calibration
The MPD 600 can acquire the test voltage level for AC and for DC, respectively. As the high voltage has to
be attenuated to match the maximum voltage level of 60 V at the V-input port of the MPD, a voltage divider
has to be used. While for AC PD measurements the coupling capacitor can be used as the high voltage
component of the divider, for DC a high-ohmic resistor is required. Together with the MPD's built-in low
voltage resister of 10 M_ it will act as an ohmic voltage divider, capable to divide DC voltages.
Note: if only a limited assortment of high-voltage resistors is available, additional low-voltage resistors can
be connected in parallel to the MPD 600 V-input port to reduce the maximum voltage at this point.
Figure 10 shows a simple set-up. Of cause, the power rating of this resistor has to be taken into
account.
As the MPD normally shows the peak and RMS value of the applied test voltage, the software has to be
switched to the DC mode by pressing the "Go!" button in the DC-Tab (see figure 11).
As a result the voltage display of the Measured Quantities Display switches to DC, as to be seen in
figure 12.
The VDC display will show the uncalibrated value. With the known ratio of the ohmic divider, this numeric
value can be entered in the belonging "Divider Factor" text field of the V-Tab (see fig. 13). The "Enter" button
of the keyboard has to be pressed to update the divider factor finally.
Now DC test voltage should be applied to the test circuit in a way, that no breakdown or partial discharges
are expected. The reading of a certified calibrated instrument has to be compared with the MPD's reading
(fig. 12). Normally, the calibration has to be adjusted in steps as long as the displayed DC voltage value
differs from the applied voltage value. Therefore the numeric value of the text field "Divider Factor" in the
V-Tab (fig. 12) has to be changed manually, as the automatic "Compute" function of the V-Tab does not
work in DC mode.
The voltage calibration is finished when the required precision is reached. Finally, the voltage trigger source
should be changed to "internal trigger" with 50 Hz internal trigger frequency, for instance (see figure 14).
This will prevent the MPD software from determining the not existing test voltage phase.
4.4
4.5
Diagram set-up
Besides the trend-diagrams of the Small Scope View also available for AC PD measurements (f PD(t), PD
frequency over time, for example), the PD software allows the usage of two additional special
types of DC diagrams during the DC PD measurement, available through the DC-Tab (see fig. 16).
These diagrams will be shown in the Large Scope View when selecting the DC-Tab there (see figure 1).
The charge-trend diagram Q(t), showing the development of the PD level over time, can be set in a
similar way as already known from the AC mode. The additional text field "Threshold" was introduced as the
only new parameter defining the threshold of pulses considered in the resulting diagram. This threshold
value is independent from the main threshold value in the Q-Tab and should be set higher. When replaying
MPD data streams this value can be altered afterwards, as well as all values in the DC-Tab.
Two charge count diagrams can be set up independently, selectable from a pull-down menu in the DC-Tab.
These diagrams count PD pulses higher than a specified reference value over a freely selectable
gliding time window.
As a typical application, these diagrams can be used to monitor the PD activity over the last 30 minutes of a
DC withstand voltage test.
For all three diagrams the time base can be chosen independently.
4.6
As the export function can be activated on-line during the measurement, it is recommended to activate the
text-file export when replaying a recorded data stream.
Example diagrams are shown in figures 23 to 25 of the following chapter.
The Large Scope View shows the test voltage curve (red) going from plus to minus. During this negative
slope single PD pulses from the test object can also be seen in this diagram, represented by single blue
dots. During the longer lasting DC plateaus (60 minutes of positive pre-stress, 60 minutes of negative test
voltage after polarity reversal phase; not shown in this diagram) no PD activity could be detected.
As the whole PD measurement can be recorded as an MPD data stream, all relevant diagrams can be
generated afterwards within the MPD software when replaying the stream file. As a second measuring
example, figure 19 shows a total overview of all PD data recorded during a DC PD measurement of more
than 150 minutes, including a polarity reversal after about 90 minutes.
The blue curve shows major PD activity of more than 70 pC after about 90 minutes (see red mark) which
exactly matches the polarity reversal start of the test voltage. As the main focus lies on the PD activity within
this short time interval, the replay-software allows zooming into the relevant time region, as shown in
figure 20.
It can be seen that the PD activity (smoothed blue curve) occurs in two local maximums, representing the
two steps of the test voltage's polarity change process from minus to zero (zone A) and from zero to plus
(zone B). Limiting the PD data to this time frame (black arrow between time cursors in figure 20) allows
generating DC evaluation diagrams out of the relevant PD data only. So the number of pulses exciting a
certain discharge level could be counted easily. Figure 21 shows the charge trend diagram and the two
pulse count diagrams.
The optional text file export ensures an easy access to the PD pulse data, allowing a further post-processing
in MS Excel, for instance (see fig. 22).
As an example, some curves generated with the exported PD data by Excel are shown in the next
figures.
List of Literature
[1] CIGRE Working Group 21.03: "Recognition of discharges", Electra. v11. 61-98, 1969
[2] B. Ehmann :"Stromrichtertransformatoren fr die leistungsstrksten und lngsten
Energiebertragungsstrecken der Welt", Netzpraxis, Jg. 48,2009, p. 14ff
[3] IEC 62199, "Bushings for d.c. application", Version 2004, 1st Edition
[4] U. Fromm: "Interpretation of Partial Discharges at dc Voltages", IEEE Transactions on Dielectrics
and Electrical Insulation, Vol. 2 No. 5, October 1995
[5] K. Rethmeier: "Separation of Superposed PD Faults and Noise by Synchronous Multi-Channel
PD Data Acquisition", 17th Biennial IEEE International Symposium on Electrical Insulation,
Vancouver, Canada, 2008
[6] Omicron electronics GmbH: "MPD600 Product Manual", Austria, 2008
[7] IEC 60270, High-voltage test techniques Partial discharge measurement, Version 2000,
3rd Edition
Americas
OMICRON electronics Corp. USA
12 Greenway Plaza, Suite 1510
Houston, TX 77046, USA
Phone: +1 713 830-4660
+1 800-OMICRON
Fax: +1 713 830-4661
[email protected]
Asia-Pacific
OMICRON electronics Asia Limited
Suite 2006, 20/F, Tower 2
The Gateway, Harbour City
Kowloon, Hong Kong S.A.R.
Phone: +852 2634 0377
Fax: +852 2634 0390
[email protected]
www.omicron.at www.omicronusa.com