Professional Documents
Culture Documents
Evaluation of Ultimate Pile Compression Capacity
Evaluation of Ultimate Pile Compression Capacity
net/publication/318461784
Evaluation of Ultimate Pile Compression Capacity from Static Pile Load Test
Results
CITATIONS READS
2 3,889
1 author:
Kedar Birid
Toyo Engineering India Pvt. Ltd.
17 PUBLICATIONS 13 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Kedar Birid on 08 February 2018.
ABSTRACT: The in-situ static pile load testing is often carried out on the test pile by
applying 1.5 times to 2.5 times the design pile capacity. Due to practical and time
constraints, it is not possible to always load and test the pile up to the failure. The
load-settlement behavior of the pile during pile load testing under this loading often
does not reach the ultimate pile capacity. Hence, in order to utilize the maximum
available pile capacity in the design of pile foundations, an extrapolation of load-
settlement data are required to evaluate the ultimate pile capacity. Various methods
were proposed in the past by researchers such as Chin, Decourt, Davisson, De Beer,
Brinch Hansen etc. to evaluate extrapolated ultimate pile capacity. These methods
have been adopted in this paper to estimate ultimate pile capacity using load-
settlement data of 23 nos. static pile load tests performed on driven piles and drilled
shafts. The ultimate capacities evaluated from different methods have been discussed
and compared with each other and with ultimate pile capacities for piles tested up to
the failure. Based on this comparison, it has been observed that when the test load is
high and close to the ultimate load limit, the accuracy in estimated ultimate load can
be achieved by all the methods. However, study on piles tested under very less or
partial load revealed the overestimation of ultimate pile load except for the Davisson
method. Each method estimated different values of ultimate load under different test
loads and no specific method can be recommended based on accuracy to evaluate the
ultimate pile capacity.
INTRODUCTION
For projects involving pile foundations, it is usually necessary to confirm the actual
ultimate compression capacity of the pile with respect to the theoretical ultimate pile
capacity. Often this is confirmed by performing a static load test on the test pile. The
ultimate pile compression capacity can roughly be defined as the load for which rapid
pile movement occurs under sustained or slight increase of the applied load or when
the pile plunges. However, often distinct plunging ultimate load is not obtained during
the test. Therefore, the pile ultimate capacity or failure load must be determined by
some criterion using load-settlement data recorded in the test. Various researchers in
past suggested different methods for evaluation of pile ultimate capacity.
As per Fellenius (2001), the concept of an ultimate load, a failure load or capacity, is a
fallacy, and a design based on the ultimate load is a quasi-concept and of uncertain
relevance for the pile assessment. An old definition of capacity has been the load for
which the pile head movement exceeds a certain value, usually 10 % of the diameter
of the pile, or a given distance, often 1.5 inches. However, these definitions do not
consider the elastic shortening of the pile, which can be substantial for long piles,
while being negligible for short piles. In reality, a movement limit relates only to a
Page 1
movement allowed by the superstructure to be supported by the pile, not to the
capacity of the pile in the static loading test. As such, the 10 % or any other ratio to the
pile diameter is meaningless from both the points-of-view of the pile-soil behavior as
well as the structure. Similarly, 1.5-inch maximum movement criterion can be just
right for the structure, but it has nothing to do with the pile-soil behavior. The
maximum deflection at the design load depends on the sensitivity of the structure to
the movement, the desired rigidity of the foundation, and local experience. However,
it remains necessary to arrive at some acceptable value of ultimate capacity, based on
some method, for the design purpose. As a result, practitioners and academicians have
introduced various pile capacity interpretation methods. Some methods were
established to decouple the effect of pile shaft stiffness from soil resistance, some that
separate side shear from end bearing, and others to try to better understand the pile-
soil behavior. Few of these methods are Davisson offset limit, the DeBeer yield limit,
Hansen ultimate load, the Chin-Kondner extrapolation, Decourt methods and
Mazurkiewicz’s graphical method. The 2003 International Building Code permitted
the use of the original Davisson, Brinch-Hansen 90 percent, Chin-Kondner, and other
methods approved by the building official. Pile Driving Contractors Association
proposed an amendment to the 2006 International Building Code, and Chin-Kondner
was replaced by the Butler-Hoy criterion. Commentary on the code change stated that
extrapolation methods should be avoided (Perko, 2009). Some of these methods have
been considered in this paper to evaluate ultimate pile capacities using load settlement
data of static pile load test conducted on different types of piles of various projects.
Page 2
relations can be derived for computing the capacity or ultimate resistance, Qu,
according to the Hansen 80%-criterion for the Ultimate Load:
Where Qu = capacity or ultimate load, C1 = slope of the straight line (see fig. 2), C2 =
y-intercept of the straight line (see fig. 2)
Chin-Kondner Extrapolation
Chin (1970) proposed an application to piles of the general work by Kondner (1963).
Chin assumes that the relationship between load and settlement is hyperbolic. The
method is similar to the Hansen method. To apply the Chin-Kondner method, divide
each settlement with its corresponding load and plot the resulting value against the
settlement. As shown in Fig. 3, after some initial variation, the plotted values will fall
on straight line. The inverse slope of this line is the Chin-Kondner Extrapolation of the
ultimate load.
Normally, the correct straight line does not start to materialize until the test load has
passed the Davisson Offset Limit. As an approximate rule, the Chin-Kondner
Extrapolation load is about 20% to 40% greater than the Davisson limit. When this is
not a case, it is advisable to take a closer look at all the test data. The Chin method is
applicable on both quick and slow tests, provided constant time increments are used.
Page 3
De Beer Yield Load (Fellenius, 2001)
De Beer (1968) made use of the logarithmic linearity by plotting the load-movement
data in a double-logarithmic diagram as shown in Fig. 6. If the ultimate load is reached
in the test, two line approximations will appear; one before and one after the ultimate
load (provided the number of points allow the linear trend to develop). The slopes are
meaningless, but the intersection of the lines is useful as it indicates where a change
occurs in the response of the piles to the applied load. De Beer called the intersection
the Yield Load. All previously mentioned methods determine a failure load except for
De Beer’s. Therefore, one should distinguish between the failure load and the limit
load to adopt the proper factor of safety.
The pile load-settlement data from different projects in different countries such as
India, Egypt, Nigeria, Singapore and USA are analysed using different extrapolation
methods explained in the previous section. The static load testing has been carried out
on drilled and driven piles with different diameters and lengths. Table 1 summarises
the load and settlement results all the test piles considered in this study.
The pile design capacity mentioned in table 1 has been calculated based on the theory
of plasticity with factor of safety of 2.5 or 3.0 depending on the probable accuracy of
the Geotechnical report as per the author’s experience. The test load has been applied
in the range of 0.86 to 5.5 times the calculated design load. The reason for applying
test load less than 2 times or as low as 0.86 times the design load is that the pile
capacities considered during the actual design of the structure were less based on the
local experience and loadings from the superstructure with respect to the calculated
capacity. The test load with higher ratio up to 5.5 times has been deliberately applied
to test the pile near to its ultimate value.
The estimated values of ultimate pile capacities from various extrapolation methods
are summarised in table 2.
It can be observed from table 2 that Davisson’s method is applicable just to four pile
load tests, i.e. 17.4%, of all analysed cases, while Hansen’s, Chin’s, Mazurkiewicz’s,
Decourt’s and De Beer’s methods were applicable to 17, 23, 23, 23 and 3 tests,
representing 74%, 100%,100%, 100% and 13% respectively of all analysed cases.
Consequently, the only three methods which are applicable to all pile load tests are
Chin’s, Mazurkiewicz’s, and Decourt’s methods.
Out of five piles tested in the Nigeria project, four piles were loaded to failure during
Page 4
the testing. Often the piles are tested up to 1.5 to 2.0 times the design load, during
which there are no signs of ultimate failure load. Hence, the test results of four failed
piles are used to estimate the ultimate load using above various methods and the
values are compared with the actual load at failure. For failed piles, the ultimate load
can simply be found out from plunging curve as shown in Fig. 7a. However, if the test
loading is done only up to 75T, the curve would be obtained as shown in Fig. 7b
which does not show plunging failure. Similar load-settlement curves are drawn only
with partial loads for 4 piles which have failed and the ultimate load are estimated
using various methods. The ultimate load is then compared with the actual failure load
to assess the accuracy of different methods for ultimate load estimation. The test
results are shown in table 3 and in fig 8.
It can be observed from table 3 that the estimated ultimate load value based on partial
test results is 13% - 48% higher than those estimated based on full test data. The
capacities estimated from Chin-Kondner and the Decourt method does not differ by a
large extent. The capacities from Davisson’s method are most conservative and are
lower by 24% to 66% with respect to actual capacities.
Among all the methods, estimation by Mazurkiewicz’s method is more accurate with
variation ranging between -15% to +26%.
Hence, it is suggested to test the pile to maximum possible load and near to ultimate
load, instead of evaluating ultimate load from test with very less test load, as such test
does not indicate any signs of failure and tends to overestimate the pile ultimate load.
Davisson’s assumption that an offset of 3.8 mm + D (inches) /120 from the elastic line
represents the movement necessary to mobilize toe resistance. The soil quake
proposed by Davisson is specifically for driven piles and is not appropriate where soil
resistance beneath the pile toe has not been fully mobilized at the beginning of load
testing. The Davisson study evaluated piles installed by driving where a compressed
soil plug forms during placement. In contrast, cast-in-place piles and other types of
drilled shafts do not compress the soil beneath the pile toe during installation. Thus, a
greater downward movement of the pile toe would be required to mobilize the end
resistance for cast-in-place piles if all other conditions are equal.
Analysis by Zheng et al. (2007) confirmed this based on the results of load tests
performed on the displacement cast-in-place piles (NeSmith and Siegel, 2009).
It is difficult to make a rational choice of the best capacity criterion to use, because the
preferred criterion depends heavily on the extent of test load, one's past experience and
conception of what constitutes the ultimate resistance of a pile.
The Davisson Offset Limit is very sensitive to errors in the measurements of load and
movement and requires well maintained equipment and accurate measurements. This
Page 5
method offers the benefit of allowing the engineer, when proof testing a pile for a
certain allowable load, to determine in advance the maximum allowable movement for
this load with consideration of the length and size of the pile. However, this method
has failed to provide the ultimate capacity in this case study as the pile settlement has
hardly approached near yield or ultimate load during the testing. The Davisson offset
line is thus unable to intercept the load-settlement curve which is required to evaluate
ultimate load.
Mazurkiewicz and Hansen ultimate load methods indicate the most conservative
results, less than the values obtained using Davisson, Chin, and Decourt methods. It is
simple in its construction, more reliable, especially for piles loaded near failure.
However, Hansen method has also failed to provide ultimate load in some cases.
Mazurkiewicz ultimate load values are on average 51% lesser than Chin’s ultimate
load.
Chin’s method is affected by the limit of loading as the pile is loaded near failure, the
greater predicted value of ultimate load. It has also been found that the Chin–Kondner
extrapolation ultimate load is 80% to 500% greater than the Davisson ultimate load.
Chin–Kondner extrapolation ultimate load is also 22% greater than Hansen’s ultimate
Load on average for the 15. pile load tests. A review of the load test data available to
the author indicates that at pile-head deflections of about 5% of the pile diameter,
sufficient mobilization of the pile toe has occurred to allow for a reasonable
extrapolation according to Chin’s method .
The Decourt method has an advantage that a plot prepared, while the static loading test
is in progress, allows the user to ‘eyeball’ the projected capacity directly once a
straight-line plot starts to develop. Extrapolation, by this method, shows ultimate load
4% lesser than Chin’s method.
The De Beer method requiress the pile to be loaded near failure, otherwise the plotted
values of the load settlement fall on approximately one straight line and the limit load
is not defined (Abdelrahman et al., 2003).
It is a sound engineering rule never to interpret the results from a static loading test to
obtain an ultimate load larger than the maximum load applied to the pile in the test.
For this reason, the allowable load cannot and must not be determined by dividing the
limit loads according to Chin-Kondner and Decourt methods with a factor of safety.
CONCLUSIONS
The analysis results into an observation that irrespective of the type of method, more
the pile loaded near the ultimate load, more accurate estimation of the ultimate load
Page 6
can be made. If the test load is very less compared to the ultimate load, variation in
ultimate load can be obtained by different methods. Hence, no conclusion can be
reached about the suitability of method for ultimate load evaluation.
For small and non-complex projects, such level of sophistication or lack thereof, is
acceptable if the uncertainty is covered by a judiciously large factor of safety. For
larger projects, however, this approach is costly. For these, the test pile should be
instrumented and the test data evaluated carefully to work out the various influencing
factors.
Combining an instrumented static loading test with dynamic testing, which can be
performed on many piles at a relatively small cost, can extend the application of the
more detailed results of the instrumented static test.
Design of pile foundation shall not be completely based on the capacity value and
more emphasized on the settlement of the pile under sustained load.
Pile capacity determined from a predefined maximum deflection can depend mainly
on the structural properties and elasticity of the pile, and can have less to do with the
pile-soil behavior. Additionally, load test equipment utilized in pile load testing is
limited in the amount of load that can be applied to the pile and often cannot reach the
full ultimate capacity of the pile.
Only three methods applicable to all tests used in the study are Chin’s,
Mazurkiewicz’s, and Decourt’s methods. Davisson and De Beer methods need the pile
to be loaded to failure to be applicable and thus cannot rely upon for non-failed piles.
As per England (1994) and England & Fleming (1994), all pile testing methods for
determining bearing capacity, from a continuous rate of penetration test to wave
analysis system, appear to introduce complications related to inability of soils to reach
a stable state in terms of effective stress during the load period. Hence, no specific
method of failure load estimation is workable under all the circumstances.
REFERENCES
Page 7
DeBeer E. E. (1968). Proefondervindlijke bijdrage tot de studie van het grensdraag
vermogen van zand onder funderingen op staal. Tijdshift der Openbar Verken van
Belgie, No. 6, 1967 and No. 4, 5, and 6.
Decourt L. (1999). Behavior of foundations under working load conditions.
Proceedings of the 11th Pan-American Conference on Soil Mechanics and
Geotechnical Engineering, Foz DoIguassu, Brazil, August 1999, Vol. 4, pp. 453 -
488.
England M. (1994). New Techniques for Reliable Pile Installation and Pile Behavior
Design and Analysis, Transportation Research Record, Issue Number: 1447,
Publisher: Transportation Research Board, pp. 39-48.
England M. and Fleming W.G.K. (1994). Review of foundation testing methods and
procedures, proceedings of Instn. Civ. Engrs Geotech. Engng, 107, July, pp. 135-
142.
Fellenius B. H. (2001). What capacity value to choose from the results a static loading
test. We have determined the capacity, then what?, Two articles reprinted from
Deep Foundation Institute, Fulcrum, Winter 2001, pp. 19 – 22 and Fall 2001, pp.
23–26.
Hansen J. B. (1963). Discussion on hyperbolic stress-strain response. Cohesive soils.
American Society of Civil Engineers, ASCE, Journal for Soil Mechanics and
Foundation Engineering, Vol. 89, SM4, pp. 241-242.
Kondner R. (1963). Hyperbolic Stress-Straim Response of Cohesive Soils, J of
SMFD, ASCE, Vol.89, SM1, pp. 115-143.
NeSmith W. and Siegel, T. (2009) Shortcomings of the Davisson Offset Limit Applied
to Axial Compressive Load Tests on Cast-in-Place Piles. Proceedings of
Contemporary Topics in Deep Foundations, International Foundation Congress and
Equipment Expo 2009, Orlando, Florida, United States, March 15-19, 2009, pp.
568-574.
Perko H. A. (2009). Helical piles, John Willey & Sons, Inc, pp. 205-214.
Zheng W., Hart T. P. and Roldan R. A. (2007). Load test analysis on augered pressure
grouted displacement piles, Proceedings, 32nd Annual Conference on Deep
Foundations, Colorado Springs, Deep Foundations Institute, pp. 25-36.
TABLES
Table 1. Load and Settlement values of Pile Foundations used in the analysis
Page 8
Dahej,
4 Drilled 500 13.80 32 174.85 10.11 1.81
Gujarat, India
Dahej,
5 Drilled 500 18.00 32 123.19 8.44 1.56
Gujarat, India
Dahej,
6 Drilled 500 18.00 32 79.46 5.73 2.05
Gujarat, India
Dahej,
7 Drilled 500 12.00 34 63.58 2.09 0.49
Gujarat, India
Dahej,
8 Drilled 500 18.00 56 123.19 6.04 2.53
Gujarat, India
Dahej,
9 Drilled 600 21.36 54 107.30 5.94 1.95
Gujarat, India
Dahej,
10 Drilled 500 13.23 35 77.94 7.04 1.99
Gujarat, India
Dahej,
11 Drilled 500 13.53 38 63.58 5.58 2.45
Gujarat, India
Kakinada,
12 Drilled 600 30.00 98 242.19 4.25 3.08
A.P., India
Kakinada,
13 Drilled 600 30.00 98 180.00 3.23 2.16
A.P., India
Alexandria,
14 CFA 600 18.00 170 255.00 1.39 0.40
Egypt
Alexandria,
15 CFA 600 18.00 170 340.00 2.57 0.73
Egypt
Port Harcourt, Driven
16 508 28.00 100 193.00 37.16 20.73
Nigeria Steel Pile
Port Harcourt, Driven
17 508 53.00 95 252.60 30.65 5.62
Nigeria Steel Pile
Port Harcourt,
18 PHC 400 28.00 78 114.10 60.50 50.00
Nigeria
Port Harcourt,
19 PHC 400 28.00 78 112.50 40.70 33.64
Nigeria
Port Harcourt,
20 PHC 400 28.00 78 73.43 50.00 50.00
Nigeria
Pulau Bukom
21 Drilled 300 26.00 70 133.00 14.38 3.50
in Singapore
Port Comfort,
22 ACIP 600 21.33 70 360.00 5.38 1.83
Texas, USA
Port Comfort,
23 ACIP 600 21.33 110 360.00 8.75 4.28
Texas, USA
CFA: Continuous Flight Auger Pile
ACIP: Auger Cast in Place Pile
PHC: Pretensioned Hollow Concrete Pile (Driven)
Page 9
Table 2. Ultimate Pile Capacities from different methods
2 Dahej, Gujarat, India Driven cast in situ NA NA 153.84 146.17 138.15 140 ---
---
3 Dahej, Gujarat, India Driven cast in situ NA NA 144.93 114.28 122.35 100
4 Dahej, Gujarat, India Drilled NA 155 357.14 274.41 386.05 250 ---
5 Dahej, Gujarat, India Drilled NA 120 227.27 197.79 286.99 210 ---
6 Dahej, Gujarat, India Drilled NA NA 133.33 92.45 150.21 98 ---
7 Dahej, Gujarat, India Drilled NA NA 98.03 102.57 113.01 80 ---
8 Dahej, Gujarat, India Drilled NA NA 181.81 172.00 193.02 155 ---
9 Dahej, Gujarat, India Drilled NA NA 169.49 116.69 164.67 120 ---
10 Dahej, Gujarat, India Drilled NA 54 344.82 --- 366.93 94 ---
11 Dahej, Gujarat, India Drilled NA 63 113.63 67.96 130.01 76 ---
12 Kakinada, A.P., India Drilled NA NA 294.11 246.03 282.55 280 ---
13 Kakinada, A.P., India Drilled NA NA 285.71 190.34 301.81 200 ---
14 Alexandria, Egypt CFA NA NA 1000.00 --- 1059.68 900 ---
15 Alexandria, Egypt CFA NA NA 1111.11 --- 1070.09 1000 ---
16 Port Harcourt, Nigeria Driven Steel Pile NA NA 256.41 233.12 261.69 205 ---
17 Port Harcourt, Nigeria Driven Steel Pile NA NA 588.23 --- 612.06 400 ---
18 Port Harcourt, Nigeria PHC 105 NA 121.95 111.52 120.31 110 85
19 Port Harcourt, Nigeria PHC 112 NA 120.48 117.65 113.21 120 120
20 Port Harcourt, Nigeria PHC 73 NA 74.62 81.71 88.50 76 74
21 Pulau Bukom, Singapore Drilled NA NA 416.67 159.39 497.19 360 ---
22 Port Comfort, Texas, USA ACIP NA NA 1111.11 --- 1158.82 650 ---
23 Port Comfort, Texas, USA ACIP NA NA 2500.00 --- 2730.92 500 ---
Page 10
Table 3. Pile ultimate load based on full load test and partial load test data
Davisson 68.0 67.0 0.34 75.0 74.0 0.70 85.0 85.0 0.76 72.0 --- ---
Chin-
256.4 250.0 1.28 121.9 178.6 1.70 120.5 169.5 1.51 74.6 108.0 1.49
Kondner
Hansen 80% 233.1 EG --- 111.5 EG --- 117.7 EG --- 81.7 EG ---
Decourt 261.7 247.0 1.27 120.3 181.1 1.73 113.2 172.4 1.54 88.5 107.8 1.48
Mazurkiewicz 205.0 165.0 0.85 110.0 107.0 1.02 120.0 132.0 1.18 76.0 92.0 1.26
Page 11
FIGURES
Page 12
FIG. 4. Decourt method
Page 13
(a) (b)
FIG. 8. Ultimate load evaluated from partial load test data for failed piles
Page 14