INTERNATIONAL SOCIETY FOR

SOIL MECHANICS AND

GEOTECHNICAL ENGINEERING

This paper was downloaded from the Online Library of

the International Society for Soil Mechanics and
Geotechnical Engineering (ISSMGE). The library is
available here:

https://www.issmge.org/publications/online-library

This is an open-access database that archives thousands

of papers published under the Auspices of the ISSMGE and
maintained by the Innovation and Development
Committee of ISSMGE.
 Geotechnical and Geophysical Site Characterisation 5 – Lehane, Acosta-Martínez & Kelly (Eds)
© 2016 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-2-7

Observing Friction Fatigue on Calcareous Material

D. Denes
Foundation Specialists Group, Victoria, Australia.

ABSTRACT: This case study describes the results of dynamic testing and CAPWAP® analysis on open end-
ed steel tubular piles driven into calcareous material. PDA monitoring of the complete driving process has en-
abled observation of a progressive breakdown in shaft friction, otherwise known as friction fatigue. A series
of CAPWAP® analyses, undertaken at varying penetrations, demonstrate the progressive loss of pile shaft re-
sistance during driving to a residual strength value. Previous research has demonstrated that friction fatigue is
associated with pile length, and more recent investigations show that it is more likely associated with cyclic
motion. The mechanism involved is discussed that provides an explanation for this loss of shaft resistance
which is thought to be attributed to the progressive degradation of localised materials at the pile/soil interface
due to the continuous shearing that is generated as the pile is driven. Using CAPWAP® to evaluate the pro-
gressive changes in the distribution of shaft resistance provides important insight into the phenomenon of this
break down in shaft resistance at a macro level. Nevertheless, it appears that shaft resistance degrades to a re-
sidual value.

1 INTRODUCTION The pile tested is a 508mm outside diameter open

ended steel tube with a 12.7mm wall thickness with
The behaviour where a soil loses resistance around no toe thickening or plate. A 1.3t hydraulic drop
the shaft during driving or cyclic loading is termed hammer with a maximum drop height of 1.3m was
friction fatigue. This phenomenon was first observed used for installation for the entire pile penetration
with piles installed in calcareous sands as described (LP) length of 39 meters.
in literature since 1973 (Angemeer et al) and has
since then been the subject of academic study (e.g.
Murff 1985, Poulos 1989, White and Bolton 2002). 2 BACKGROUND
By using dynamic testing methods to measure the
complete installation of a pile, the variation of shaft 2.1 PDA-CAPWAP®
resistance distribution as a function of pile penetra- Dynamic pile testing, also known as Pile Driving
tion can be inferred. Analyzer (PDA) testing, is a technology which was
This case study focuses on a project off the north- developed in 1960 (Smith) and commercially availa-
west coast of Australia. The first dynamic test pile ble in the 1970’s by Pile Dynamics, Inc. (PDI).
occurred at a penetration of 23.5m and several days Strain and accelerometers gauges are attached to a
later at a penetration of 43.5m. At both penetrations, pile and capture strain/acceleration time records for
the driving set was 4mm per blow using the same each hammer/pile impact. The information recorded
hammer and drop height. It is this difference in is both the stress-wave from the hammer and the re-
penetration and the similar set per blow that raised flected stress-waves generated along the pile length
issues with the shaft frictional resistance. Therefore as well as from the pile toe. These stress-waves can
a particular pile (F1) was chosen for detailed as- be interpreted using wave analysis programs such as
sessment to be monitored over an installation length CAPWAP®, in order to determine the total capacity
of 39 meters. The aim of the study is to observe fric- and the distribution of the capacity along the shaft
tion fatigue using PDA and undertake CAPWAP® length and at the pile toe. The program CAPWAP®,
modelling to determine if there is a reduction/fatigue developed by Goble and Rausche (1979), is general-
of shaft resistance around the pile. ly accepted to be the definitive method on interpret-

1199
ing dynamic pile records. It is based on modelling nominated depth, the larger the decrease in shaft re-
the pile as an elastic body, with both mass and stiff- sistance;
ness, and the surrounding soil as a series of elasto- a repetition of shearing the soil at the pile/soil inter-
plastic static resistance springs and linear dashpots. face produced by movement of the pile (this is re-
Actual soils do not follow simple elasto-plastic be- ferred to as a cycle – i.e. a pile hammer blow);
haviour and are more typically non-linear, but
CAPWAP® in most cases has shown proven relia-
bility providing that a sensible and realistic model
has been applied. To make a rational and valid
CAPWAP® model, the engineer should have all in-
formation at hand, including hammer details, driving
records, set measurements, geotechnical conditions
and pile details. The theory of PDA and CAPWAP®
will not be further discussed within this paper.

2.2 Pile Design

Prior to the realisation that calcareous sands behave
differently to terrigenous sands, determination of
shaft resistance was based on standard or ‘conven- Figure 1. h/D Effect
tional’ methods which are still currently used today.
These methods were initially applied to all types of
sands, both silica and calcareous. It was not until the
1970’s (Angemeer 1973) and the 1980’s (i.e. North
Rankine A – Senders et al. 2013) that pile test results
observed shaft resistance values well below the
‘conventional’ design.
The standard conventional methods involve the
use of overburden stress and a frictional co-efficient
between the pile type, soil and the earth pressure co-
efficient, K. (Johannessen and Bjerrum 1965, Chan-
dler 1968 adopted slightly different approach). Other
methods involve correlations between collected in- Figure 2. Hollow Structure – Coralline Rock
situ testing (such as SPT). A number of foundation
textbook even advise to limit the maximum skin
friction, which is proven incorrect (Kulhawy 1984). 3 CALCAREOUS SANDS
More modern methods for shaft resistance are based 3.1 What is it?
on CPT (Kelly and Wong 2005) and on the expan-
sion cavity expansion theory where there is an in- Calcareous material is a generic description given to
crease in radial stress related to dilation effects. Us- granular soil composed of calcium carbonate. It is
ing the cavity expansion approach removes the ‘K’ formed from shells and skeletal remains from ben-
value and the overburden stress, both being replaced thos organisms (bottom dwelling) such as coral,
with the effective radial stress acting on the shaft at molluscs and calcareous algae. Closer inspections of
failure and the critical state sand interface angle of these biogenous sediments reveal a porous hollow
friction, which is developed when the soil at the in- structure, with a rough surface (Figure 2).
terface has ceased dilating or contracting (i.e. zero 3.2 Engineering Properties
volume change).
Calcareous sands are special because they do not be-
have like normal sands. They have a higher void ra-
2.3 Friction Fatigue tio due to the hollow structure and therefore tend to
crush more readily compared to lithogenous sedi-
Friction fatigue is the destruction of the surrounding ments (quartz sand). These sands may be found in
material due to excessive shear stresses, thought to cemented or uncemented states. At first there is
be related to the following: some strength to the grain due to the material and
the ‘h/D’ effect (Randolph et al 1994). (Figure 1). potential cementation effects. But once a certain
This is associated with the decrease of horizontal ef- pressure is exceeded, it crushes quickly. Trying to
fective stress (radial/normal stress) as the pile pene- determine the initial nature of calcareous materials is
trates deeper. Hence the higher the h/D ratio below a quite difficult. For example, undertaking SPT testing
involves damaging the possible cementation bonds.

1200
The friction angle is generally quite high, typically 4 FULL MONITORING OF PILE F1
above 35 degrees and often exceeding 50 degrees
(Murff 1987). It is noted that even partial cementa- The final penetration for Pile F1 was 39m below sea
tion can cause ‘arching’ around the pile, hence the bed level. PDA monitoring commenced from 6m.
sand can effectively self-support itself, not applying Data collected was of good quality and four strain
a great deal of lateral pressure to a pile. Poulos gauges were used for the last 33m of driving due to
(1989) showed that under a series on monotonic slight hammer alignment issues. There was a total of
loading on calcarenite core samples, that the normal 7226 hammer hits (or cycles) from the start of driv-
stress reduces as shear displacement increases. ing. The set vs penetration is shown in Figure 3. As
These tests are quite important as they demonstrate the hammer drop height is consistent, the set per
the volume change characteristics for calcareous ma- blow variation gives an indication of the strength of
terial where initial contraction leads to a continued the material at the toe of the pile. Hence a smaller set
reduction in the post peak strength at the interface. may indicate stronger material, or higher cementa-
tion.
3.3 Friction Fatigue and Calcareous Material
White and Bolton (2002) discuss that friction fatigue
is related to pile movement. Experiments observed
the decrease in horizontal stress at a nominated
depth below ground level as the pile toe passes this
nominated point and continues further beyond. They
observed that a zone of sand beside the pile under-
went volume reduction reducing the normal stress.
White and Lehane (2004) undertook laboratory ex-
periments using instrumented piles to determine if
friction fatigue is related to the h/D ratio. The piles
were installed in one of three ways: 1. Monotonic; 2.
Jacked – cycles of fixed downward displacement
and 3. Pseudo-dynamic – comprising of a downward Figure 3. Set vs Penetration
then upward displacement. Results from the Mono-
tonic loading showed only a minor decrease in shaft 4.1 Site Geotechnical Model
resistance with increasing depth. However, the
jacked and pseudo-dynamic showed significant deg- Due to the remote location of the work site, a ge-
radation of shaft friction with the results showing otechnical investigation was not undertaken at the
that it is related to the number of cycles that had oc- work site location. Therefore the geological envi-
curred beyond the soil horizon, not the h/D ratio. ronment is based on a desktop study, local observa-
This observation is quite unique because it elimi- tions and geotechnical reports in similar environ-
nates the h/D effect alone based on the monotonical- ments. The geological age is Late Quaternary to
ly installed piles. However, for piles that are in- Recent. It is composed of bedded marine calcarenite,
stalled under cyclic condition, it is shown that it is thick reefs, calcareous sand and talus (limestone)
the number of cycles that influences the reduction in deposits. The generalised profile for the test site is as
shaft reduction. follows: Upper 15m comprising of calcareous sandy
sediments, carbonate sands, coral, phosphate sedi-
3.4 Shear Resistance Values for Calcareous ments, and talus. Areas may be cemented. Below
Material 15m – porous limestone layers, varying degrees of
Observations (Angemeer et al 1973) from a series of cemented sand layers, sand layers composed of cor-
load tests on driven piles showed the capacities were al, coralline limestone, shells, carbonate sands and
approximately 20% of those predicted by conven- silts.
tional methods. Murff (1985) listed a number of test 4.2 PDA Observations of Pile F1
results from 1973 to 1985 and observed mean peak
values ranging between 13.4 to 20.3 kPa. Aggarwal PDA testing captures the downward generated com-
et al (1997) further defined skin friction resistance pressive stress wave (from the hammer impact) and
relating to percentage of calcium carbonate, where the subsequent reflections. The data is generally pre-
above 45% shaft resistance is approximately 28kPa. sented by viewing force/velocity time records, or
Poulos (1989) provided peak skin friction values more commonly used view of Wave-Up (WU) and
ranging between 10 to 20kPa for uncemented mate- Wave-Down (WD). The WU trace gives a visual
rial. Residual friction values ranged between 5- guide of the dynamic soil resistance around the shaft
20kPa for various cementation levels. Ghazali et al of the pile. As the wave travels down (WD) the pile,
(1990) discussed limiting skin friction to 20kPa. the pile moves downwards and displaces the soil
around which ‘slips’ past the pile around the perime-

1201
ter. This slipping is referred to as mobilisation, al approach of using WU to indicate the shaft re-
which is displacement vs shear resistance. This soil sistance model is not applicable from when the un-
resistance is partially reflected as compression loading section begins. Therefore, if there is a weak-
waves where these reflected waves are cumulative er layer, CAPWAP® will model this.
added together to generate the WU response. To fur-
ther demonstrate this, Figure 4 shows the WU re-
sponse and an indication of the shaft resistance on
the pile. As the WU line increases, the shaft re-
sistance on the pile increases. Therefore if the pile is
PDA monitored for the full penetration, then by
comparing the WU response, we are able to visually
assess how the shaft resistance around the pile be-
haves.
The Wave-Up response for various penetrations of
pile F1 is shown in Figure 5. The response has been
modified to only show length of penetration (i.e. the
pile toe response has been removed). The changes in
the response are quite dramatic at certain locations.
For example the response at LP 13.7 and 17m are
much higher compared to the other response. This Figure 5. Wave-Up Response for Various Penetrations
correlates with the sets being lower per blow indicat-
ing harder ground conditions. Once the pile pene-
trated beyond 17m, the response drops of signifi-
cantly. This example demonstrates that the shaft
resistance is reducing as the penetration is increasing
due to driving. At the final penetration, it is noted
that the WU response seems to come to a final ‘re-
sidual’ line following the same trend for 33m.

Figure 6. Shaft Resistance Values Vs Pile Penetration

4.3 CAPWAP® Analysis and Observations for Pile

F1
A CAPWAP® Analysis was undertaken for each of
the penetrations as shown in the PDA Wave Up re-
Figure 4. Interpretation of Wave-Up Response sponse. The total unit shaft resistance at each meter
of shaft vs penetration is shown in Figure 6. The re-
The WU response eventually becomes a downward sults are representative of the PDA data, which show
slope once the penetration reaches 25-30m and be- similar trends to the WU responses. Using the depth
yond. Generally, the WU increases with respect to at 13m as an example, the modelled total unit shaft
increasing skin friction resistance which represents resistance based on the data from LP13.7m is
that the shaft resistance is increasing with depth. 30.7kPa. At the same depth location but looking at
When the WU starts to ‘flatten’ or even reduce this the results from the CAPWAP® at LP38m, the total
may indicate that the shaft resistance is decreasing in shaft resistance is 7.8kPa. This is a decrease of ap-
strength or it is associated with pile unloading (an proximately 75%.
upward movement of the gauges as the pile bottom In all CAPWAP® analysis, it is the bottom 1-
is still moving downwards), or a combination of 2meters of the pile that has the highest unit shaft re-
both. When there is unloading in PDA data, a sistance. This may be expected because the pile has
CAPWAP® analysis is required to model the effects just been driven into virgin material, where the sur-
of unloading on the WU response to determine if the rounding soil is trying to be ‘pushed’ to the side of
decrease in the response is caused by lower shaft re- the pile. As the calcareous sand particles are trying
sistance values. Therefore, the straight forward visu- to dilate, there is an increase in horizontal stress

1202
which increases the shear stress which initiates col-
lapse of the sand. Associated with this collapse is
volume reduction. This is also evident, as the unit
shaft resistances 4m above the toe (or an approxi-
mate h/D ratio of 8) are up to 40% lower compared
to the initial value.
From the CAPWAP analysis, the values below are
given as total kPa values acting on the external sec-
tion of the pile. Therefore, depending on what per-
centage is acting internally, the values may be less:
Peak: 20.7-35kPa, Residual (LP3-7): 1.3kPa, (LP15-
20) 9kPa, (LP22-30) 5.5kPa, Average 5kPa, which is Figure 8. Percentage Shaft Loss vs h/D Ratio
similar to what has been observed in literature.
To determine if friction fatigue is related to pene- Summarising the above information, the best vis-
tration (the h/D ratio) or to the number of cycles, fur- ual trend is seen in Figure 7 (Percentage Peak shaft
ther investigation is required using the CAPWAP re- loss Vs Cycles) where all 3 penetrations seem to
sults. The following depth locations below sea bed align to approximately 60% (a 40% decrease). After
level will be used for this analysis, which are: 5, 15 that, the 5 and 25m locations follow each other,
and 25m depths. For this, the total external shaft re- while the 15m location takes a slightly different
sistance values have been converted to percentage path. Figure 8 (Peak Shaft loss compared to h/D ra-
decrease from peak shaft values using the initial first tio) still shows a visual trend for all 3 locations, but
peak shaft value determined at each depth. is more scattered. It appears that friction fatigue is
Figure 7 compares the % shaft resistance decrease better associated with the number of cycles applied
vs the number of blows encountered below each compared to h/D ratio based on visual assessment of
nominated depth location. This shows that at each of trending data. It is also interesting to note that both
the 3 penetrations, there is a break down in shaft re- methods observe a ‘residual’ value, which is a loss
sistance ranging from 70 to 90% from its initial shaft in shaft resistance from 70 to 90% of its original
resistance value over a period of 3000 to 4000 cy- peak value.
cles. All three depths show an approximate linear The above review is quite unique to observe a pile
decrease starting at 100% which flattens out to a po- with over 7200 cycles. These observations seem to
tential ‘residual’ value. Figure 8 shows the % shaft correlate with the data observed from White and Le-
resistance decrease vs the h/D ratio. This now shows hane (2004). Unfortunately, laboratory models can-
that the decrease in unit shaft resistance ranges from not reproduce this type of analysis due to scaling ef-
70-90% for an h/D ratio ranging from 21 to 33, an fects and model size limitations, with respect to pile
average ratio of 27. It is noted for both that a trend displacements per blow.
begins to appear where the unit shaft resistances
show a similar degradation/fatigue pattern for both.
5 CONCLUSION

This case study focused on the observation of fric-

tion fatigue in calcareous material using Pile Dy-
namic Testing (PDA) to observe the changing nature
of the shaft resistance by recording the response
wave for every cycle. The response wave is generat-
ed by in situ shaft resistance and the cycles are gen-
erated from a driving hammer. Results from Pile F1
showed that PDA is an interesting and suitable vehi-
cle for observations of friction fatigue and provides
an insight to what is occurring in real time. CAP-
WAP® modelling of the pile at various penetrations
allowed the determination of peak and residual shaft
resistance which correlated quite well to existing lit-
Figure 7. Percentage Shaft Loss vs No. of Cycles erature from the past 40 years. The shaft resistance
values allowed further analysis by comparing the fa-
tigue in the shaft resistance values to be compared to
the number of cycles as well as the penetration, or
h/D ratio, to observe if there is a relationship with ei-
ther method. Overall, the number of cycles com-
pared to shaft degradation seemed to best represent

1203
friction fatigue which was based on the observation Smith, E.A.L. (1960). Pile driving analysis by the wave equa-
of 3 different penetrations and comparing the unit tion. Journal Soil Mechanics and Foundations, ASCE, 86
(SM4). pp. 35-41.
shaft resistance which showed better trends com- White, D.J., Bolton, M.D., (2002). Observing Friction fatigue
pared to h/D. This case study has provided a direct on a jacked pile. Centrifuge and Constitutive Modelling:
and unique insight into the phenomenon of friction Two extremes. Springman S.M. (ed) pub Balkema, Rotter-
fatigue. dam.
White, D.J., Lehane, B.M. (2004), 'Friction fatigue on dis-
placement piles in sand', Geotechnique, 54, 10, pp. 645-658
6 ACKNOWLEDGEMENTS / INFORMATION

I would like to thank Dr. Julian Seidel for his time,

support and encouragement. A great mentor. This
paper is based on my Masters Specialists subject,
referenced below.

7 REFERENCES

Angemeer, J., Carlson, E. and Klick, J.H. (1973). Techniques

and results of offshore pile load testing in calcareous soils.
Proceedings, Fifth Annual Offshore Technology Confer-
ence, Houston, Texas, 677-692.
Denes, D. (2015). Observing Friction Fatigue on Calcareous
Material: A Progressive Breakdown of Shaft Resistance for
Piles driven into Calcareous Material as observed during
Dynamic Pile Testing. Masters Subject. UNSW.
Ghazali, F.M., Baghdadi, Z.A. and Mansur, O.A. (1990). A
comparative study of pile foundations in coral formations
and calcareous sediments. JKAU Eng. Vol 2. pp. 3-17.
Goble, G.G. and Rausche, F. (1979). Pile driveability predic-
tions by CAPWAP®. Proc. Int. Conf. on numerical Meth-
ods in Offshore Piling, ICE, London, pp. 29-36.
Johannessen, I. J. and Bjerrum, L (1965). Measurement of the
compression of a steel pile to rock due to settlement of the
surround clay. Proceedings of the 6th International confer-
ence on Soil Mechanics and Foundation Engineering, Mon-
treal, 2, pp. 261 to 264.
Kelly, R. B. and Wong, P.K. (2005). Case Study: Assessment
of the shaft capacity of a driven pile in sand using CPT
based methods. Australian Geomechanics. Vol 40, No. 4.
December. pp.73-80.
Kulhawy, F. (1984). Limiting tip and side resistance: Fact of
fallacy, analysis and design of pile foundations, ASCE,
1984, pp. 80-98.
Murff, J.D. (1987). Pile capacity in calcareous sands: State of
the art. Journal of Geotechnical Engineering, Vol 113, No5,
May. pp. 490-507.
Nauroy, J.F. and LeTirant, P. (1983). Model tests of piles in
calcareous sands. Offshore Engineering Practice, ASCE,
pp356-369.
Nauroy, J.F. and LeTirant, P. (1985). Driven piles and drilled
and grouted piles in calcareous sands. Proceedings Seven-
teenth Annual Offshore Technology Conference, Houston,
Texas. pp. 83-91.
Pile Dynamics, Inc. (2015). Manual for the PDA 8G and PDA-
S Software.
Poulos, H.G. (1989). The mechanics of calcareous sediments,
John Jaeger Memorial Address. Australian Geomechanics.
pp. 8-40.
Randolph, M.F., Dolwin, J. and Beck, R. (1994). Design of
driven piles in Sand. Geotechnique, 44, No3, pp. 427-448.
Senders, M., Banimahd, T., Zhang, T., Lane, A. (2013). Piled
Foundations on the North West Shelf. Australian Geome-
chanics, Vol 48, No4. December. pp149-160.

1204