Nilson
Nilson
Nilson
This paper covers the general principles of pile load testing, including
objectives of pile testing, importance of planning, various types and methods
of testing, instrumentation, data to be obtained, and interpretation and use
of the test results. Also included are some typical case histories together
with correlative data between standud test methods and the constant rate
of penetration test method.
•THE PURPOSE of pile load testing can be either to prove the adequacy of the pile-soil
system for the proposed pile design load or to develop criteria to be used for the design
and installation of the pile foundation. Tests in the first category are generally routine,
are carried to twice the proposed working load, and are conducted at the start of the
job. Tests to develop design and installation criteria involve more elaborate programs,
and piles are usually tested to failure.
Pile load tests are expensive and can be quite time consuming. For small projects
the cost of pile testing can represent a considerable portion of the overall foundation
cost. In many cases, prior experience combined with adequate subsoil data and sound
judgment can preclude the need for pile testing, especially if the pile design load is
relatively low.
Routine pile load testing is often the decision of the foundation engineer, but may be
required by the general specification or building code having jurisdiction over that type
of construction. The decision to embark on an advance test program to develop design
criteria is usually made by the owner and the foundation engineer and is based on the
scope of the project and the complexities of the foundation conditions. Such test pro-
grams can often result in substantial savings in foundation costs, and these can more
than offset the investment in the test program.
The prime objective of a test program is to produce data to determine the most eco-
nomical and suitable pile foundation, including the pile types to be used, the most effi-
cient or highest working load for each type of pile, the required length for each type of
pile, and the installation methods necessary to achieve the desired results.
Paper sponsored by Committee on Foundations of Bridges and Other Structures and presented at the 49th Annual
Meeting.
74
75
Figure 1. Test load applied directly to pile by using loaded platform and water-filled
interconnected steel tanks.
The load and platform must be kept balanced at all times. Usually timber cribs are
placed under the platform edges to prevent tipping of the load in case the platform be-
comes unstable. Wedges between the timber crib and platform edge are tightened only
while the load is being added or removed. T hese wedges must be kept loose as the pile
settles under the direct load.
Figure 2. Test load applied to pile with hydraulic jack reacting against a test
frame and anchor piles.
77
Most test loads are applied with hydraulic jacks reacting against either a stable
loaded platform or a test frame anchored to reaction piles (Fig. 2); there may also be
some other type of reaction. The use of hydraulic jacks has several advantages. For
example, it is the only practical way to apply load-unload-reload cycles, and hydraulic
jacks are more suitable for uplift tests, lateral tests, and tests on batter piles.
Regardless of the method of load application, the load should be kept constant under
increasing pile deflection. For direct loading this presents no problem. When hydrau-
lic jacks are used this can be accomplished by activating the jack pump with a com-
pressed gas control system. Precautions should be taken to avoid eccentric loading
by carefully centering test beams or jacks and maintaining a balanced load.
Anchor piles or the supports for a reaction load must be placed a sufficient distance
from the test pile to avoid influencing its performance. This minimum distance will
depend on such things as the magnitude of load to be applied and the subsoil conditions.
Such influence could reflect a greater or lesser ultimate bElaring capacity than actual.
It is recommended that, during a lateral load test, an axial compressive load equal
to the minimum design dead load, be applied to the pile. This type of combined test
loading would give a more accurate indication of the actual lateral load capacity of the
pile under service conditions. When a vertical load is applied during a lateral load
test, the pile butt should not be restrained from lateral movement. This can be accom-
plished by using a system of rollers between the vertical load and the pile. The point
of application of the horizontal load should, if possible, simulate in-service conditions.
INSTRUMENTATION
The basic information to be developed from the pile load test is usually the deflection
of the pile butt under the test load. Probably the fundamental method of measuring the
pile butt movement is by reading a target rod (or scale fixed to the pile) with an engi-
neer's transit referenced to a fixed bench mark. In most cases, the degree of accuracy
obtained with this type of instrumentation is sufficient. Quite often, measurements with
the level and rod (or scale) are used as a secondary or backup system to check other
measuring systems.
Direct readings of the pile butt movement (either vertically or horizontally) can be
made by using the mirror, scale, and wire method. A measuring scale is fixed to a
mirror, which in turn is attached directly to the pile or the test plate. A taut wire
passing in front of the scale permits direct readings of pile movement. Consistent
scale readings are obtained by aligning the wire and its image in the mirror. The wire
can be kept taut by a weight and pulley system or by springs.
The most common method for measuring the pile movement is with dial extensometers
mounted on an independent support system, and with gage stems bearing against the top
of the test plate or on angle irons attached to the sides of the pile (Fig. 3). At least 2
dial gages mounted on opposite sides of the pile should be used to compensate for pos-
sible tilting or lateral movement of the pile under load. Sometimes a gage sensitivity
of 0.001 in. is specified, but usually gages reading to 0.01 in. have sufficient accuracy
to meet the normal settlement criteria. With ultra-sensitive dial gages, it is often
impossible to satisfy some of the specification requirements such as "until settlement
stops."
When the instrumentation for a compression test is set up, it is often advisable to
mount dial gages to measure lateral movements of the pile under test. Such movement
could be due to eccentric loading and contribute to the apparent vertical movement of
the pile butt.
The instrumentation system must be supported independently from the loading system
with supports protected from extreme temperature variations, effects of the test load,
and accidental disturbance by test personnel. It is advisable to have a secondary or
backup instrumentation system in case of an accidental disturbance of the primary sys-
tem or the necessity to reset dial gages so that continuity of data is maintained.
Data on load distribution and the elastic behavior of the pile can be obtained with dis-
placement (or so-called "strain") rods or strain gages. This type of instrumentation
can be installed in almost all types of conventional piling but more readily in cast-in-place
78
Hydraulic jack
w/pressure gauge-------!!+--
100 ton c.:ip.:icity
concrete piles. Strain gages or the terminal points of "strain" rods can be located at
various positions along the pile.
In general, strain rods are less complicated, are less subject to malfunction, are
more easily handled by field personnel, and produce direct elastic shortening data over
a long gage lengtti between the terminal point and the pile butt. The proper installation
of strain gages, so as to avoid malfunction and produce reliable data, is an extremely
sensitive operation.
The installation of strain rods or gages results in a physical change in the cross
section of the pile and thus its elastic pr9perties. Although data at frequent intervals
along the pile shaft are desirable, it is sometimes advisable to sacrifice some data in
the interest of practicality. Often a single strain rod to the pile tip is sufficient to pro-
vide the essential information on the elastic behavior of the pile and the basic load dis-
tribution.
TEST PROCEDURES
Most routine tests are carried to one and one-half or twice the proposed design load
for a single pile, or one and one-half times the design load for a pile group. Carrying
the test load any higher merely wastes job time and money. Rarely can such additional
data be used advantageously, such as for redesign, without seriously affecting the job
schedule.
Most test programs that are specifically executed to produce design data should in-
clude testing piles to failure in order to develop the most efficient design. However,
this is not always essential, and definite design decisions can be reached if sufficient
routine testing is done on piles of different types, sizes, shapes, and lengths.
79
The time interval between pile driving and testing depends on the type of pile and
subsoil conditions. For example, sufficient time should be permitted for the proper
curing of cast-in-place concrete piles before they are tested. Where test piles are
driven into cohesive soils, it is advisable to wait several days for the soil to regain its
shear strength, which in all probability was reduced because of the remolding effects
of pile driving.
The test load can be applied in various increments and time intervals. In general,
the load should be applied over an extended period of time, with increments equal to
about 25 percent of the proposed or assumed design load. However, in the interest of
saving time, the increments can be larger during the early stages of the test and, in
the interest of obtaining accuracy, they should be smaller as the total load is increased.
A normal time interval between load increments is from 1 to 2 hours. Frequently,
specifications will require that the load be held until the rate of settlement is less than
some fixed value such as 0.01 in. per hour, but in most cases a maximum time interval
between increments will also be specified. Providing that the pile-soil system has not
failed, the full test load should be held for some period of time, such as 24 or 48 hours.
Specifications will often establish a maximum rate of settlement under full load that
cannot be exceeded over a certain period of time in order for the test to be considered
satisfactory. Where specifications use language such as "until settlement has stopped,"
the impracticality of using highly sensitive dial gages is obvious .
Instrumentation readings should be taken before and after each increment of load and
at sufficient intermediate intervals in order to define the load-time-deflection curves .
When piles are not tested to failure, and after the full test load has been applied, read-
ings are taken at least every 30 minutes for the first 12 hours and every hour there-
after. During removal of the test load, readings should also be taken before and after
each load decrement, and a final rebound reading should be taken about 12 hours after
the full load has been removed.
Among the several special testing techniques available are cycle loading and the con-
stant rate of penetration (CRP) method. When piles are tested to establish the design
load, cycle loading can help determine more accurately the load that satisfies the al-
lowable deflection criteria. Also, cycle loading can provide some indication as to the
distribution of load between friction and end-bearing. Van Weele (1) has suggested a
method by which a plot of the elastic recovery at each unloading cyc le versus load ap-
plied at that cycle is used to separate friction from point-bearing. The curve usually
becomes a straight line soon after the early load increments (Fig. 4). The distance
between the plotted curve and a line drawn through
the origin and parallel to the straight part of the
L O A D IN TO N $
curve represents the portion of the load carried
0 50 !00 150 200 250 by friction. At best, this is only an approxima-
tion:
Cycle loading should not be mandatory for
routine testing because it could add unnecessary
I -,
"'w
I
expense without contributing significantadditional
' u \ data. Such special procedures should be included
~ 021--~-1-...+~+-,..---+-~~-1--~-1
"'
<{
reasonably good correlation with standard test
~
w methods. For the CRP test, a force or load of
sufficient magnitude is applied to the pile to main-
tain a constant penetration rate into the ground.
This means that the applied load might have to
be adjusted as the test proceeds. In general,
the recommended penetration rate is about 0.03
Figure 4. Approximate distribution of test in. per minute for cohesive soils and about 0.06
load between point-bearing and friction (1). in. per minute for granular soils. However, the
80
penetration rate could vary over a rather wide range and still produce satisfac-
tory results.
The CRP method is applicable to friction types of piles, and sufficient testing capac-
ity of the pile-soil system.
Special testing procedures can be used to prorluce specific data. For example, the
distribution of applied load between friction and point-bearing can be approximated by
driving and testing piles of different lengths. Some would be driven just short of the
end-bearing stratum, while others would be driven to full embedment. An uplift test
might also produce approximate data on the amount of load carried by friction.
Another special test procedure would be the casing off of that portion of the test pile
that extends through soils offering temporary support so as to determine the capacity
of the pile - soil system within the permanent bearing strata.
1.·
INTERPRETATION OF RESULTS
f
The basic purpose of the pile test is to determine or verify the safe working load for
the pile-soil system. In most cases, tests are not carried to failure, and some arbi-
trary criterion is applied to determine if the test results are satisfactory. Some of
these criteria are rather vague, such as "where the settlement is disproportionate to
the load" or "where the load-settlement curve breaks." Others are based on a maximum
allowable gross or net settlement that can either be a fixed number, such as 1 in. or
related to the amount of test load applied, such as 0.01 in. per ton.
When definite failure does not occur, such as plunging of the pile into the ground,
some arbitrary definition of "failure" must be used. Such criteria should be realistic-
neither too conservative nor too liberal. The important factors to be considered are
the permissible differential settlement under the design load and safety.
Settlement usually governs and requires consideration of the elastic shortening of
the pile under the design load. Assuming that all the piles are of the same material,
of approximately equal length, and driven into substantially similar soils, the elastic
shortening will be approximately equal for all piles and thus will not contribute to dif-
ferential settlement.
Many methods have been suggested for determining the safe allowable pile load or
for defining the "failure" of the pile-soil system. The application of these various
criteria can produce a wide range of "safe" pile loads from the same test data.
Unless failure actually occurs, it would appear reasonable to define the point of
"failure" by a maximum slope of the load-settlement curve. For example, the failure
load could be defined as the load that results in a slope greater than 0.05 in. per ton
on the gross load-settlement curve or a slope greater than 0.03 in. per ton on the plas-
tic load-settlement curve, whichever is smaller (Fig. 5). This is still an arbitrary
definition of failure: but is a more generalized approach. The total criterion would
'
include a maximum allowable gross
settlement under the design load, with
consideration given to elastic shorten-
ing of the pile and to safety.
Where failure results from some
arbitrary criterion, the factor of safety
could range from one and one-half to
two. Where actual failure of the pile-
soil system is determined by a plung-
ing of the pile into the ground, this
factor of safety could range from two
to two and one-half.
- ----- The complete analysis of the test
results should include consideration
of all factors, such as the elastic be-
havior of the pile (from instrumenta-
Figure 5. Slope criteria for determining "failure" load from tion or cycle loading) and an evaluation
load-settlement curves. of the long- term performance. This
81
could involve an analysis and evaluation of the subsoil data in conjunction with the test
results.
It should be noted that observed settlements made at the top of the pile may not
necessarily indicate downward movement of the pile into the ground. Where high
load tests are performed, the possibility of local failure of the pile above ground sur-
face, or crushing of the grout under the test plate, should be recognized as possible
factors contributing toward observed "settlements."
APPLICATION OF RESULTS
Because it is impractical to test every pile on a project, the results of the testing
must be applicable to other piles to be driven. This is a reasonable and accepted pro-
cedure, providing that the following conditions exist:
1. The other piles are of the same type, material, and size as the test piles;
2. Subsoil conditions are comparable to those at the test pile locations;
3. Installation methods and equipment used are the same as or comparable to those
used for the test piles; and
4. Piles are driven to the same penetration depth or resistance or both as the test
piles to compensate for variations in the vertical position and density of the bearing
strata.
The results of tests on single piles can usually be applied to pile groups, especially
in granular soils. The group effect, if any, depends a great deal ·on the subsoil profile
to some depth below the pile tips. Unless the bearing stratum is relatively thin and
( underlain by deep deposits of soft compressible soils, there should be no detrimental
effects from group loading. However, where the piles receive their principal support
in cohesive soil, group action should be analyzed.
I The application of the results of the advance test program to the foundation design
I ~;-~
and specification can often produce substantial savings in foundation costs. Although,
as a practical measure, the test results would lead to the selection of a single design
loati, the requirements for various types of piles as to size, length, shape, weight per
I· foot (stiffness), installation methods, and driving requirements could vary over a rather
wide range. These differences should be reflected in the specifications and, in turn,
will be reflected in the alternative costs to produce the most economical foundation for
the conditions involved.
where
L = maximum proven design load as indicated by pile load test, and
P = "safe" load capacity determined by the dynamic hammer formula used on the
test-loaded piling before the K-factor is applied.
N11 S11!l1m1n1
obtained by interpolation from the
computed net settlement line value ,__
z
w
of0.25 in. This line is obtained by :i;
w
calculation based on the actual re- --'
,__
,__
corded recovery. w
V>
Constant Rate of Penetration Moahn11m Proven Dul9n Lood• Ult1n11111t1 Beorln 9 Cop11111•/2
Method
In January 1963, Engineering
News- Record published an article Figure 7. Interpretation of results (48-24 hour test method).
by Esrig (5) in which he discussed
anewpile-testing procedure devel-
oped by Whitaker and Cooke (2).
Whitaker termed the new testfiig procedure the constant rate of penetration (CRP)
test. This method requires that the test-loaded pile be forced into the ground at a con-
stant rate with the loads corresponding to specific penetrations being measured.
°o~--,2~0~"""""'•0,---,,.•o:----:•~o~-,o~o~-,~
, o:--~~~ 00
MAXIMUM PROVEN DESIGN LOAD -TON S QUICK TE ST LOAD METHOD
(QUICK TEST LOAD METHOD) • K• FACTOR (PROVEN)
Figure 9. Correlation of proven design load between Figure 10. Correlation of proven K-factor between
standard 48-24 hour and quick test methods. standard 48-24 hour and quick test methods.
0.10-in. settlement respectively. The maximum proven design load is the value used
to establish a K-factor for use with the dynamic hammer formula as stated earlier.
Correlation Studies
From January 1963 until March 1965, llpileloadtestswereperformedbythe Texas
Highway Department by using both the standard 48-24 hour and the quick test methods .
Out of this number of tests, eight were test-loaded to theoretical failure by the 48-24
hour test method. All of the tests were taken to plunging failure with the quick test
method. A s ummary of the data for thes e tests is given in Tables 1, 2, 3, and 4. The
maximum proven design load obtained by the quick test method and the 48-24 hour test
mf!thod are s hown i n Figm ·e 9. The ave rage deviation of maximum p r oven design load
values obtained from the quick test method versus the standard 48-24 hour test method
was about 4 percent.
The K-factors given in Tables 3 and 4 for both load test methods are shown in Figure 10.
Agreement is considered to be very good in all ranges of value.
CONCLUSIONS
Based on the results of these tests,
the Texas Highway Department began
TABLE 1 using the quick test method as the stan-
LOCATION OF PILES TESTED FROM dard testing method in Apr il 1965. A
FEBRUARY 1963 TO NOVEMBER 1964
special provision to Item 405, Test
Te~t
County &tructure
Bent and Loading Piling, was prepared for this
Number Test Pile test method and has been in use by the
Brazoria Chocolate department since that time (4).
Bayou Bent 6, C From the 1963-1965 study as well as
2 Wharton Colorado
River Bent 4 from the department's experience to
3 Arkansas Capano Bay Pile 1 date, the following observations and con-
4 Vicluda MP RR OP I'llc 1
5 Harris Ramp F Bent 9
clusions have been reached relative to
6 Jefferson US-69 and the quick test method:
FM-365 Pile 1
7 Harris Pierce Bent 94 1. A pile load test can be expeditiously
8 Harris 183 Bent 3
9 Harris SP RR UP Bent 2
performed in about 1 hour with resultant
10 Galveston GC and SR savings in money and time ;
RR UP Bent 24 2. Construction delay to the project
11 Harris SP HR UP Bent 2
caused by load testing is greatly reduced;
85
TABLE 2
DESCRIPTION OF PILES, SOIL, AND HAMMER
TABLE 3
AASHO 48-24 HOUR a'EST METHOD
TABLE 4
QUICK TEST METHOD
REFERENCES
1. Van Weele, A. F. A Method of Separating the Bearing Capacity of a Test Pile Into
Skin-Friction and Point-Resistance. Proc. Fourth Internat. Conf. on Soil Me-
chanics and Foundation Eng., London, 1957.
2. Whitaker, T., and Cooke, R. W. A New Approach to Pile Testing. Proc. Fifth
Internat. Conf. on Soil Mechanics and Foundation Eng., Paris, 1961.
3. Standard Specifications for Road and Bridge Construction. Texas Highway Department,
1962.
4. Special Provisions to Item 405 of 1962 Standard Specifications. Texas Highway
Department.
5. Esrig, M. I. Load Test on Pile in as Little as Ten Minutes. Engineering News-
Record, Jan. 31, 1963.