See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/265403529

What constitutes a good PDA test?

Article

CITATIONS READS

5 790

2 authors, including:

Garland Likins
Pile Dynamics, Inc.
38 PUBLICATIONS   241 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Garland Likins on 19 May 2015.

The user has requested enhancement of the downloaded file.

What constitutes a good PDA test?
Likins, G.
President, Pile Dynamics Inc., Cleveland Ohio USA
Rausche, F.
President, GRL Engineers Inc., Cleveland Ohio USA

ABSTRACT: High strain dynamic pile testing (HSDPT) is routine practice, and has many benefits that make its
use highly attractive. However, it requires careful preparation, attention to detail and testing engineer skill to
properly perform these tests.
Depending on his or her experience level, the test engineer may or may not realize when results are valid, or
when capacity results may be only a lower bound. It is widely known that the capacity of driven piles may change
with time after installation and for that reason restrike tests are generally recommended. It is also generally stated
that the set per blow should be at least 2 mm, or otherwise the dynamic test may only yield a lower bound solution.
To investigate the validity of these recommendations, GRL’s database of dynamic test results is compared with
static load tests, and various comparisons are made for both driven and drilled piles. Conclusions from this study
and recommendations for practice are drawn.

1 INTRODUCTION 2 DATA REQUIRED

In the several decades since the beginning of research Dynamic pile testing should follow ASTM D4945
on modern dynamic testing of piles in 1964, much guidelines to obtain the dynamic force and velocity
progress has been made in the tools used such as the on the pile, often two diameters below the pile top.
Pile Driving AnalyzerÒ (PDA) and CAPWAPÒ Usually measurements are made with reusable strain
(Likins, 2008). High Strain Dynamic Pile Testing and acceleration sensors. Obtaining the average force
(HSDPT) has become an indispensable tool initially requires using strain transducers attached to the pile in
for the installation and testing of driven piles, and in diagonally opposite pairs to assess and compensate for
recent decades to also assess the capacity of drilled bending and obtain the axial response. While at least
shafts. Many codes recognize the procedures of two accelerometers are also suggested, for good data
ASTM D4945 and highlight its value by including the velocity signals are practically identical even when
mention of this test method. During installation of bending is severe (e.g. bottom graphs in Fig. 1 each
driven piles, the benefits include providing actual show two velocities). Thus, one well performing
information based on actual measurement of the accelerometer is the minimum required. The
event to assure a safe and efficient installation at a sensors are typically bolted to the pile. Loose
rather modest cost. As a predictor of ultimate static attachments must be avoided. While Fig. 1b shows
capacity of the pile or drilled shaft, results from data with good signals, Fig. 1a depicts a strain record
HSDPT have conclusively demonstrated good where a loose bolt or concrete anchor corrupted one of
correlation with results from static load testing the strain signals. The loose strain transducer, F2
(Likins, 2004), and provide substantial cost savings dashed curve in the middle graph of Fig. 1a, gives a
when compared with Static Load Test (SLT) costs. As signal similar to an accelerometer since the
a result of the good correlation experience, HSDPT is unrestrained end of the strain transducer acts as a
often used to supplement static tests on larger projects, mass responding with inertia to the pile
and on smaller projects often to provide some quality acceleration. A loose bolt on an accelerometer (not
assurance that the installation is sufficient by serving shown) would result in two different velocity signals
as a replacement for static tests. But like any other tool, with one being delayed and with different frequency
HSDPT is only beneficial if the test is performed content. In Fig. 1b, the steel pipe pile has bending so
properly. Several considerations go into obtaining a that the two strains on opposite sides of the pipe are
good test. quite different in magnitude (middle graph of Fig. 1b),

Science, Technology and Practice, Jaime Alberto dos Santos (ed) 403
 and the casting conditions and methods, the general
recommendation is for 4 strain measurements. The
four individual signals, and opposite pairs, can be
compared to assess data quality. If one sensor
produces poor results, then its signals and those of
its opposite mate can be ignored and the signals of the
remaining two sensors used. Since the number of
blows is usually relatively low when testing drilled
or augured piles and because each impact is critical,
having backup sensors assures useable data for every
applied impact.
The pile properties, such as pile cross sectional
area, density, modulus, wave speed, and length, must
be known. Modulus is calculated from the pile
material density and observed wave speed. If the
pile is composite (e.g. concrete filled steel pipe
pile), the composite properties are used. If the pile
is non-uniform along the length (e.g. area or material
changes) then this must be noted.
Figure 1a. Data with a loose strain transducer.
3 SIGNAL MATCHING ANALYSIS (SMA)

A valid dynamic load test must be evaluated by signal

matching (e.g. CAPWAP) analysis. For non-uniform
piles (drilled shafts and augercast piles are often
non-uniform) such analysis is essential since
simplified methods (e.g. the Case Method used by
the Pile Driving Analyzer) calculate capacity
assuming a uniform pile. The signal matching also
develops a soil model for the dynamic event (quakes
and damping), and better distinguishes between
static and dynamic resistances, without depending
on unreliable soil investigation descriptions.
Geotechnical engineers sometimes erroneously
assume that the SMA should incorporate results
from soil borings in its solution. Such a process
would totally defeat the purpose of the HSDPT
and lead to useless results (a) because the HSDPT
should provide additional information, (b) because
the results from soil exploration and dynamic
testing do not match each other (soil properties at
the time of testing and at the strain levels applied are
different from those occurring during the soil
sampling and/or testing), and, most importantly, (c)
often the closest soil boring is not always
Figure 1b. Good data with tight bolts.
representative of the soil condition at the tested
pile (and frequently does not even extend to the pile
toe).
and many times different in shape and magnitude, yet Another advantage of SMA is the simulated static
the average force (solid line in top graph 1b) shows load test graph relating applied load (capacity) to
good proportionality with the velocity. For a series of displacement which can then be correlated to static
blows from a pile driving hammer, good data will be load test results; viewing a series of CAPWAP results
reasonably consistent from blow to blow (if the for the early restrike blows also provides interesting
hammer performs consistently), and the early insights when evaluating the capacity (Rausche,
portion of the force and velocity records (e.g. at 2008). Other benefits include evaluation of pile
first peak input) should be proportional by the pile integrity and modeling of splices, and, when
impedance EA/c. questions arise, an accurate investigation of tension
In the case of HSDPT for drilled or augured piles, and compression stresses along the shaft and at the pile
because of the non-uniformity of the concrete or grout toe.

404 Ó 2008 IOS Press, ISBN 978-1-58603-909-7

4 ADEQUATE IMPACT DEVICE

For most situations, the best test is obtained when the

pile experiences a net final set (permanent
displacement) as a result of the impact. If the pile
to be tested is a driven pile, usually the pile driving
hammer is used as the impact device. The hammer is
often selected using a ‘‘wave equation analysis’’ to
determine its adequacy (Hussein, 1996) for pile
installation to avoid overstressing or excessive blow
counts, i.e. insufficient penetrations per blow.
In cases where the set per blow is anticipated to be Figure 3. Ratio of CAPWAP to max applied load in SLT vs. set
very small, e.g. when the soil develops a high soil setup per blow.
following installation, a larger hammer may be needed
during restrike to mobilize and prove the required pile
capacity. Fig. 2 shows the relative ratio of predicted the maximum applied load, indicating CAPWAP and
CAPWAP solution to SLT as a function of set per blow even the Davisson criteria to which it is often
for the FHWA database reported at the 1996 compared both are generally conservative compared
Stresswave conference (Likins, 1996). The SLT with this reserve strength.
result was evaluated with the Davisson offset Testing of drilled shafts and augered (CFA) piles
procedure. As mentioned earlier, it has long been requires a minimum drop weight of 1 to 2% of the
recommended that for driven piles a net permanent ultimate test load required (Hussein, 1996). Of course,
set per blow of at least 2 mm be achieved in order to larger drop weights with lower drop heights are likely
mobilize the full soil strength. Although this advice to be acceptable too, and are generally preferred
does not seem to be confirmed by this database, it is whenever practical.
still felt to be good advice and recommended The best data is generated when the pile top and
procedure. On the other hand, net sets greater than impact device both have flat surfaces which impact
8 mm per blow are not desirable either, since they uniformly so that local stress concentrations are
make the analysis more difficult, and more sensitive to minimized. For concrete piles or drilled shafts, a
errors in dynamic resistance. In Fig. 2, there are few cushion of soft material (e.g. usually plywood) is
data points above the recommended upper limit of inserted between the impact device and the pile top
8 mm set, but generally the correlation is still good to smoothen any surface roughness and reduce the
with one exception of overprediction by a ratio factor peak stress. It is also strongly recommended that the
of 1.32. Interestingly, in this one case, a Rapid Load pile top is extended by typically 1 m or at least one
Test (e.g. Statnamic) was conducted between the diameter length at the time of initial installation and
dynamic restrike and the static load test, pushing with the same concrete as used for the pile shaft. This
the pile an additional net settlement of 0.22 m pile top extension should include some external
which according to the closest soil profile easily reinforcement in the form of a thin walled tube.
could have pushed the toe from a sand layer into a Forces can also be measured by instrumenting the
clay layer, thus corrupting the correlation. Also known drop weight with an accelerometer to measure
interesting in Fig. 2 are the relatively conservative the deceleration and using Newton’s Law (F ¼ ma)
predictions for pile sets between about 2 and 8 mm per (Robinson, 2002).
blow. In refusal situations, i.e. when the set per blow is
As demonstrated in Fig. 3, very few dynamic test less than 2 mm, a number of steps can be taken to
results from CAPWAP have a prediction in excess of assure meaningful results. First, as described in
Rausche et al. (2007), the maximum pile toe
displacement should be calculated by SMA. If the
pile toe displacement plus the permanent set of prior
impacts exceeds D/60 (D being pile width or
diameter), it can be concluded that the resistance
has been activated. Secondly, if the soil is
sensitive and looses capacity as it activates more
resistance near the toe during the restrike test,
‘‘superposition’’ may lead to better capacity
estimates (Hussein, 2002). Superposition may use
soil resistance components from either early and
late restrike tests or from end of pile installation
and early restriking. Thirdly, in the case of drilled
Figure 2. Ratio of CAPWAP to SLT prediction versus permanent piles whose end bearing is in a relaxed state, the pile
set per blow. may have to be ‘‘driven to capacity’’ with a series of a

Science, Technology and Practice, Jaime Alberto dos Santos (ed) 405
few blows usually with increasing energy per blow. It
then has to be checked, however, whether or not the
necessary penetrations to achieve a significant end
bearing would be less than allowable movements.

5 TIME OF TESTING

‘‘When the pile is tested’’ is closely related to the

‘‘purpose of the pile test’’. If the issue is determining
dynamic stresses during pile driving to confirm
adequate procedures to prevent overstressing and
thus damage, to determine alternate installation Figure 4. Time of dynamic test.
procedures (e.g. changed cushions, reduced stroke,
etc), or monitor the installation efficiency for a pile presents the dynamic CAPWAP (CW) result to SLT
driving hammer to confirm the driving criteria, then result plotted as a function of ‘‘time ratio’’ (time ratio
tests during pile driving are appropriate. For a concrete is the ratio of days for the restrike (BOR) after the
pile where tension stresses in easy driving is a concern, installation, to the number of days for the SLT after
the pile is often tested for the full installation length. installation). A time ratio of 1.0 would then be
For steel piles, testing only the last portion of driving is represent having the restrike and SLT on the same
generally sufficient. day. Fig. 5 shows the same data designated by soil type
When determination of the long term ultimate (key labels ‘‘toe, shaft’’: where SA is sand, SI is silt,
capacity is the reason for the test, since soil and CL is clay). Admittedly the classification is
properties often change with time after installation, subjective, particularly for shafts in a layered soil;
both driven piles and drilled piles are usually evaluated fine grained sands were classified as silt in this study
during a dynamic test (restrike) following some wait since their drainage appears to behave similarly).
period after installation. While some researchers have It is interesting to review Figs. 4 and 5. Looking at
suggested using end of drive data with a Measured the data where BOR was less than 5 days, they all have
Energy Approach to estimate long term service time ratios of less than 0.4 meaning the SLT was run
capacity of the pile, the premises of this procedure considerably later in time than the restrike. Most data
are not correct and the correlation is worse than that of was in sand, or silt, with the silt cases generally
a proper HSDPT test during restrike (Rausche, 2004). yielding low capacity ratios, likely because the full
For piles with much of their capacity coming from setup had not yet been developed by the time of the
cohesive soils where setup (e.g. a capacity increase) relatively early PDA restrike test. The highest point is
can be very significant, the wait before restrike might the aforementioned controversial data point with the
be many days or even weeks to simulate the long term intervening Rapid Load Test.
service conditions and take maximum economic For BOR data of 5 to 9 days, although the wait time
advantage. Piles driven into weathered shale or was longer, the time for the SLT was delayed even
dense saturated silts may experience a capacity more so that the time ratios were generally 0.25 to 0.55
decrease (e.g. ‘‘relaxation’’); testing after a few and since most of the cases are in silts where continued
days is usually sufficient to identify this problem. setup is expected, it is not surprising that the CAPWAP
The selection of the duration of wait period to SLT ratio are usually less than unity, particularly for
between end of installation and restrike depends on the lowest time ratio cases.
many factors. Certainly the soil type has a major For the data where the restrike was 10 or more days
influence. It is well known that clays can exhibit after installation, most cases were also in finer grained
pore pressure effects causing effective stress silt or clay soils as might be expected from the longer
changes and should require longer wait periods due wait times. For the data with time ratio of about 0.5,
to the reduced porosity. While clean coarse sands may
drain quickly, often well graded sands, particularly
fine sands and silty sands may have sufficient fine
content to retain excess pore pressures for longer
times. Sands may also experience arching effects as
the pile experiences lateral motions during driving that
create an oversized hole; with time, the hole
diminishes and normal earth pressures reestablish
themselves. Calcareous soils require longer times
for the soil structure to recrystallize.
Most Static Load Tests are performed at least 7 days
after installation. In fact, in the 1996 database, only 6
piles had 3 or fewer days wait before the SLT. Fig. 4 Figure 5. CAPWAP vs SLT by soil type and time ratio.

406 Ó 2008 IOS Press, ISBN 978-1-58603-909-7

most CAPWAP data result in underprediction of SLT * If the dynamic loading device (e.g. pile driving
Davisson result due to incomplete setup in silt or clay hammer) is of insufficient energy capacity
soils. The notable exception with capacity ratio 1.335 estimates may only be a lower bound if the
is for an open profile H pile in clay where the end resulting set per blow is very small.
bearing was likely overestimated; caution is therefore Superposition methods may then be used to
warranted when CAPWAP would suggest high end project the full bearing capacity.
bearing for H piles where the soils are known to be * Since pile or shaft capacity usually changes with
clay. Most cases with a time ratio of greater than unity time after installation, and dynamic tests reflect
have a CAPWAP prediction within 10% of the SLT the capacity at the time of the test, sufficient wait
Davisson value. time after installation is required for a good
Results from this investigation suggest that for finer dynamic pile test. The wait after installation for
grained soils, or soils with porosity that behave as finer coarse grained soils may be short, while the wait
grained soils, waiting longer for the dynamic test is for piles installed in fine grained soils should be at
beneficial to the dynamic capacity prediction by least a week or two to gain the maximum economic
CAPWAP, while piles in clean sands might be advantage.
tested relatively early with good results. These
findings confirm long held beliefs and suggestions
about the time of testing. REFERENCES
Testing at end of driving and during restrike, or
multiple restrikes for special test piles at different wait ASTM D4945 (2000). ‘‘Standard Test Method for High-Strain
times (and evaluated by extrapolation for longer Dynamic Testing of Piles’’.
Hussein, M.H., Likins, G. E. and Rausche, F. (1996). ‘‘ Selection
waiting times on a log time scale for cohesive of a Hammer for High-Strain Dynamic Testing of
soils), can be useful in determining a proper driving Cast-in-Place Shafts’’. Fifth International Conference on
criteria and assure sufficient capacity (using short term the Application of Stress-wave Theory to Piles
restrikes of production piles, provided the piles (STRESSWAVE ’96): Orlando, FL.
display the same capacity gain trend line). Drilled Hussein, M.H., Sharp, M. and Knight, W.F. (2002). ‘‘The Use of
Superposition for Evaluating Pile Capacity. Deep
shafts and augercast (CFA) piles naturally require at Foundations 2002 an International Perspective on Theory,
least a week (and usually longer) wait time after Design, Construction, and Performance’’, Geotechnical
casting concrete prior to the dynamic testing to Special Publication No. 116, O’Neill M. W., and
allow the concrete to harden sufficiently. This time Townsend, F. C. Eds., American Society of Civil
period is then generally adequate for the soils Engineers: Orlando.
disturbed by the drilling or augering process to Likins, G. E., Piscsalko, G., Rausche, F. and Roppel, S. (2008).
‘‘PDA Testing: 2008 State of the Art’’ Eighth Int’l Conf. on
regain their pre-drilling strength. the Application of Stresswave Theory to Piles, Lisbon.
Likins, G. E. and Rausche, F. (2004). ‘‘Correlation of CAPWAP
with Static Load Tests’’. Proceedings of the Seventh
International Conference on the Application of Stresswave
6 CONCLUSIONS Theory to Piles: Petaling Jaya, Selangor, Malaysia.
Likins, G. E., Rausche, F., Thendean, G. and Svinkin, M. (1996).
A good dynamic pile test requires certain minimum ‘‘CAPWAP Correlation Studies’’. Fifth International
requirements which can be summarized as follows. Conference on the Application of Stress-wave Theory to
Piles (STRESSWAVE ’96): Orlando, FL; 447–464.
* Measurements must be of good quality or the Rausche, F., Likins, G. E., and Hussein, M. H. (2008). ‘‘Analysis
dynamic test results are meaningless. Sensors of Post-Installation Dynamic Load Test Data for Capacity
Evaluation of Deep Foundations’’ ASCE Geotechnical
must be firmly attached to the pile, and their Special Publication ‘‘From Research to Practice in
calibrations known. Geotechnical Engineering’’–Honoring John Schmertmann.
* Interpretation of the data for uniform driven piles Eds: Jim Laier, David Crapps, and Mohamad Hussein.
may be straightforward in the field for stresses Rausche, F., Robinson, B. and Likins, G. E. (2004). ‘‘On the
induced in the pile, evaluation of damage, or Prediction of Long Term Pile Capacity From End-of-Driving
assessing hammer energy transfer to the pile. Information. Current Practices and Future Trends in Deep
Foundations’’, Geotechnical Special Publication No. 125,
* Interpretation of dynamic test data for pile DiMaggio, J. A., and Hussein, M. H., Eds, American
capacity, or analysis of non-uniform piles and Society of Civil Engineers: Reston, VA.
drilled shafts, must be made using a signal Robinson, B., Rausche, F., Likins, G. E. and Ealy, C. (2002).
matching analysis such as CAPWAP. ‘‘Dynamic Load Testing of Drilled Shafts at National
* To assess ultimate pile capacity, the impact device Geotechnical Experimentation Sites. Deep Foundations
2002’’, An International Perspective on Theory, Design,
used must be of sufficient weight to push the pile, Construction, and Performance, Geotechnical Special
at least temporarily, a sufficient distance to activate Publication No. 116, O’Neill M. W., and Townsend, F. C.
the passive soil resistance. Eds., American Society of Civil Engineers: Orlando, FL.

Science, Technology and Practice, Jaime Alberto dos Santos (ed) 407
View publication stats