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

net/publication/374260350

Application of axial load tests in the Netherlands to offshore pile design

Conference Paper · September 2023

CITATIONS READS

0 111

4 authors:

Kevin Duffy Kenneth Gavin

Delft University of Technology Delft University of Technology
6 PUBLICATIONS 2 CITATIONS 210 PUBLICATIONS 4,743 CITATIONS

SEE PROFILE SEE PROFILE

Mandy Korff Dirk de Lange

Deltares Deltares
45 PUBLICATIONS 417 CITATIONS 19 PUBLICATIONS 53 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Kevin Duffy on 28 September 2023.

The user has requested enhancement of the downloaded file.

 Innovative Geotechnologies for Energy Transition | The Society for Underwater Technology

Application of axial load tests in the Netherlands to offshore pile design

K.J. Duffy & K.G. Gavin
Department of Geoscience & Engineering, TU Delft, the Netherlands
D.A. de Lange & M. Korff
Deltares, Delft, the Netherlands

ABSTRACT: This paper describes axial load tests on three full-scale driven precast piles in the Netherlands.
The piles were founded in dense to very dense river-deposited sands, a soil that is widespread across the
Dutch North Sea sector. The deposit is characterised by cone penetration test (CPT) tip resistances of up to 90
MPa and offers a detailed insight into pile response in realistic offshore conditions. Each test pile was
incrementally loaded to failure under compression, while fibre optic sensors measured the changing
deformation of the pile. The analysis and interpretation of the load test data focussed on how the three slender
piles behaved at large shaft and base resistances. Notably, the piles mobilised base and shaft resistances
greater than currently prescribed limiting resistances in design standards, thereby highlighting some
overconservatism present when designing piles in dense sand.

1 Introduction on these piles, bringing more certainty to designers

The new ISO/API design method (Lehane et al., and contractors when designing piles in dense on-
2020) has changed how driven piles under axial shore and offshore sand deposits.
loading are designed offshore. To calibrate the de-
sign method, high quality static load tests were com-
piled (Lehane et al., 2017), building on previous da- 2 Limiting resistances in design
tabases such as the ZJU-ICL database (Yang et al.,
2015) and the UWA database (Schneider et al., Divergence in national and international design
2008). However, these databases still lack data from standards has led to different calculation approaches
very dense sands, primarily because of the difficulty for dense to very dense sand. In general, many de-
in safely and economically mobilising the required sign standards limit the calculated base resistance qb
failure loads in these conditions. and shaft resistance qs, (e.g. Table 1). Some stand-
This data gap has created uncertainty regarding ards, such as in the Netherlands, impose a strict limi-
pile behaviour in very dense sand. In response, many tation. Other design methods, like in China or the
design methods cautiously limit the amount of re- older ISO/API method, adjust the limiting resistance
sistance that a pile can mobilise. While this can be a based on the relative density or the CPT cone tip re-
pragmatic response to the unknown, it can also in- sistance qc. Furthermore, Table 1 does not convey
troduce overconservatism into the design. This in some of the implicit limitations in design methods.
turn, increases the financial and environmental cost For example, a weighted average of cone resistances
of pile fabrication whilst increasing the likelihood of around the pile base is used to determine the design
pile damage and pile refusal during installation. cone resistance qc,avg and correspondingly, the pile
To investigate pile response in very dense sand, a base resistance. Many different algorithms have
test site was established at the port of Rotterdam in been proposed to determine this qc,avg (e.g. van
the Netherlands. Three driven precast piles were in- Mierlo and Koppejan, 1953; Boulanger and DeJong,
stalled at least eight pile diameters into a dense to 2018). However, each algorithm treats the high cone
very dense sand layer. Each pile was instrumented resistances differently and often limits the base re-
along their full length with fibre optic sensors, giv- sistance in a way that is not explicitly described in
ing a detailed insight into the pile shaft and base re- design standards.
sponse. This paper presents the load tests performed

2234
 Session 22 - Project & Design Case Studies II

Figure 1: Lifting the load test frame into place at the Amaliahaven test site

The origins of limiting resistances are partially approach these dense to very dense soil conditions
rooted in the critical depth theory (Poulos, 2001). In with caution.
a set of full-scale (Vesic, 1970) and model pile ex-
periments (Kerisel, 1961; Robinsky and Morrison, Table 1: Limiting resistances for driven precast piles in CPT-
based design codes. Where the limitation is dependent on the
1964; Vesic, 1965), it appeared that the average relative density or qc, the value in very dense sand has been
shaft resistance reached a limiting value for depths presented
more than twenty pile diameters from the top of the Location Standard Base [MPa] Shaft [kPa]
pile. However, the theory has since been heavily re- Belgium NBN-EN None 150
futed (Kulhawy, 1984; Kraft, 1991; Fellenius and 1997-2
China JGJ 94- Non-linear 125
Altaee, 1995), including by the original author them- 2008 reduction
selves (Kulhawy, 1996). This was because installa- France NF P 94- None 150
tion-induced residual stresses and scale effects of the 262
model tests were not fully assessed at the time. Oth- Netherlands NEN 15 150
9997-1
er uncertainties, such as apparent diameter- Offshore API RP 2A 12 115
dependent scale effects (Chow, 1997; White and Offshore ISO/API None None
Bolton, 2005), contributed to the apprehension over
pile response at high resistances.
In the Dutch design code, alternative reasons 3 Amaliahaven pile test site
were cited for implementing limiting resistances.
One concern was that pile driving could reduce the Dense sands have been a persistent challenge in the
high horizontal stresses in overconsolidated soils port of Rotterdam (de Gijt and Broeken, 2013). In
and therefore reducing the amount of base resistance 2013, the port was extended into the North Sea by
available (te Kamp, 1977). At the time, precast piles the creation of the Maasvlakte II peninsula. The ex-
were also not prestressed during manufacturing. This tension opened a large amount of land for develop-
meant that piles had to meet minimum size require- ment and many kilometres of deep-sea quay walls
ments to avoid damage or buckling during transpor- are now being designed and constructed. These large
tation, staging and installation. Because of these re- earth-retaining structures require thousands of foun-
quirements, reaching test loads beyond the dation piles and so even minor adjustments to their
prescribed limiting resistances was practically in- design can be hugely beneficial, both financially and
conceivable at the time (te Kamp, personal commu- environmentally.
nication). To investigate the pile behaviour in these dense
Indeed, the geotechnical and structural design of sands, a test site was established at the harbour of
piles has since progressed significantly and modern Amaliahaven (Figure 1). Eleven piles were installed:
design approaches are gradually shifting away from three driven precast piles, four driven cast-in-situ
strict limiting resistances (Fleming et al., 2008). piles and four screw displacement piles. These tests
However, the dearth of high quality full-scale load meet the requirements for high quality test databases
tests has meant that designers and contractors must

2235
 Innovative Geotechnologies for Energy Transition | The Society for Underwater Technology

such as the ZJU-ICL (Yang et al., 2015) or the

ISO/API databases (Lehane et al., 2017).
This paper focusses on the load tests of the driven
precast piles. More information and discussion on all
test piles is to be disseminated into journal papers in
due course.

3.1 Local geology

The river Maas is the focal point of Rotterdam. It is
a part of the Rhine-Maas-Scheldt delta system, a del-
ta which has shaped the Netherlands not just cultur-
ally and economically, but also geologically. The
dynamic nature of the delta and the Dutch coastline
has largely dictated the geological depositional pro-
cesses, resulting in a range of marine, lagoonal and
fluvial soils across the region (Hijma et al., 2012).
Around Rotterdam, newer Holocene soils are pre-
sent to a depth of around 20m. In the eastern part of
the port, these soils consist of thick layers of soft
clay. Moving towards the western coast, on the other
hand, interlayered clays and sands tend to dominate.
These deposits include formations commonly found Figure 2: CPTs performed at each driven precast pile prior to
in the North Sea, such as the Naaldwijk, Nieuwkoop installation
and Southern Bight Formations (Rijsdijk et al.,
2005). 3.2 Site investigation
Under the Holocene soils is a Pleistocene epoch
sand known as the Kreftenheye Formation. The for- Before installation, at least three CPTs were per-
mation was deposited by the Rhine-Maas river sys- formed two metres away from each test pile loca-
tem and can be described as a medium dense to very tion, including one CPT on the location itself (Figure
dense poorly sorted calcareous coarse silica sand, 2). Three CPTs were performed 1.5 metres away
frequently gravelly. The upper boundary of the for- from pile DP3 after installation, although no signifi-
mation is often capped by a bed of stiff clay known cant variation was observed compared to the pre-
as the Wijchen Member (Autin, 2008). installation CPTs. Boreholes were also performed
The Kreftenheye Formation is found not just in across the entire harbour and a large amount of la-
Rotterdam but also across much of the western boratory tests were carried out on the retrieved sam-
Netherlands, such as under the Hague, Utrecht and ples.
parts of Amsterdam. In these cities, the formation is Dredged sands are present down to 14m depth.
the primary load-bearing layer for many pile founda- Underlying these sands is a naturally deposited sand
tions and so it is of substantial engineering im- layer, followed by interlaminated clays and sands
portance. The outwash of the Rhine-Maas river sys- belonging to the Naaldwijk formation. The
tem has also meant that the formation is present Kreftenheye Formation begins at a depth of 28m.
throughout the Dutch North Sea sector (Rijsdijk et The formation is first capped by 1m of stiff clay
al., 2005; Hijma et al., 2012), generally found 5 to with qc equal to around 1.5MPa. The resistance of
10m below seabed level with thicknesses of around the sand layer then builds up to an average of around
10m. 50MPa that is relatively constant with depth. Some
areas of the site feature thin weak laminations within
this sand layer, although none of these were evident
in the vicinity of the driven precast piles.

3.3 Geometry of the test piles

The driven precast piles (DP1, DP2 and DP3) meas-
ured 400mm square (equivalent diameter Deq =
450mm) and with a length of 32m, giving a slender-
ness ratio L/Deq of 70. To minimise the number of

2236
 Session 22 - Project & Design Case Studies II

hammer blows on the pile, water jetting was per-

formed in the upper layers during pile driving. This
jetting stopped at least two metres above the lower
sand layer to avoid any adverse effects on the prima-
ry load-bearing layer. The three piles were installed
at least 7Deq into this lower sand layer (Table 1).
In terms of the soil surrounding the pile base, all
three piles have design cone resistances qc,avg of be-
tween 33 and 38MPa when using the 4D/8D Dutch
averaging method (van Mierlo and Koppejan, 1953).
The range of this qc,avg is much narrower when using
an update to the Boulanger and DeJong (2018) filter
method (de Boorder et al., 2022) with qc,avg around
44–46MPa. Figure 3: Response of the pile head during load testing

Table 2: Test pile properties

DP1 DP2 DP3 4.1 Mobilised base resistances
Side length [m] 0.4 0.4 0.4
Embedded length [m] 31.74 31.29 31.80 As with the overall load-displacement response, the
Age at pile test [days] 28 30 78 piles behaved very similarly at the pile base (Figure
qc,avg (Dutch method) 38.1 33.5 34.0 4). Locked-in residual stresses were already present
qc,avg (Filter method) 45.5 44.0 45.3
at the start of load testing, measuring 10MPa at the
pile base. This is already a substantial contribution
3.4 Instrumentation and load test procedure to the total base capacity, that is, 20–25% of the de-
During each load test, deformation of the pile was sign cone resistance qc,avg when determined by the
measured to help distinguish between the pile base filter method (Table 2). From this initial stress, a fur-
and shaft resistances. This was done using two types ther 20MPa of base resistance was mobilised during
of fibre optic sensing techniques: Brillouin Optical testing, reaching a total base capacity of 30MPa,
Frequency Domain Analysis (BOFDA) and Fibre 66% of qc,avg.
Bragg Grating (FBG). The BOFDA system provided Figure 5 compares the measured resistances to
distributed strain readings, whereas the FBG system those in the ISO/API database, without residual
provided discrete strain readings. Raman sensing stresses included. The Amaliahaven pile tests reach
was also used to apply temperature compensation base stresses 10MPa higher than the existing
between the reference measurement and the residual maximum base stress in the database. A clear linear
load measurement, as described in Duffy et al. relationship is seen across all cone resistances
(2022). measured, well beyond any of the limitations
The piles were tested under axial compression us- provided in Table 1.
ing a test frame tied in by grout reaction anchors. In-
cremental loading was applied, with the duration of
each increment determined based on the creep rate
of the pile. Failure was defined as when the pile base
displacement reached at least 10% of the pile’s
equivalent diameter.

4 Results

All three piles reached peak loads of around 8MN

during testing. Piles DP2 and DP3 were loaded di-
rectly to this peak load, whereas pile DP1 was tested
with load cycles after each of the first eight load Figure 4: Mobilised base resistance, including residual stresses
steps. Nevertheless, Figure 3 shows that the three
piles responded identically in terms of their initial
stiffness. All piles experienced plunging failure
within a pile base displacement of 10% Deq.

2237
 Innovative Geotechnologies for Energy Transition | The Society for Underwater Technology

around 1% of the equivalent diameter. A relatively

brittle failure is then seen at loads between 200–
250kPa. Interestingly, strong dilatancy is exhibited
in pile DP3, tested two months after DP1 and DP2.
This dilatancy reduced the shaft resistance from a
peak of 260kPa to just under 220kPa.
Considering the shaft responses in the context of
the design, the founding sand layer mobilised
resistances twice as high as some of the prescribed
limiting resistances (Table 1). This comes despite
the negative shear stresses that had to be overcome
at the start of the load test. In a similar fashion to the
measured base resistances, the results suggest that
imposing limiting resistances would lead to
unnecessary overconservatism when designing for
the pile shaft resistance.

Figure 5: Comparison of the Amaliahaven results with the

ISO/API database. The Dutch 4D/8D averaging method was 5 Conclusions
used to determine qc,avg
This paper presents the outcomes of high-quality
4.2 Mobilised shaft resistances axial load tests on driven precast piles at a nearshore
site in the Netherlands. The piles were founded in
Taking the average shaft resistance qs,avg for each dense to very dense sand and extensively
layer, some variation can be seen between the soil instrumented with fibre optics, providing a detailed
layers during each load test (Figure 6). Negative insight into the mobilised base and shaft resistances.
shear stresses were mobilised before load testing, In the load-bearing sand layer, high base and
acting in equilibrium with the 10MPa of residual shaft resistances were mobilised, surpassing the
base stress. These negative shear stresses are most limitations present in several design codes such as
significant in the lower sand layer, ranging from −40 the API RP 2A design method. In the context of
to −60kPa. quay walls at the port of Rotterdam or offshore
The load tests mobilised very low resistances of structures in the North Sea, for example, the limiting
around 20kPa in the upper sand layer, despite resistances would introduce a great deal of
average qc values of around 15MPa. In the overconservatism into the foundation design. Indeed,
interlaminated layer, the mobilised resistance was at while it is preferable to include a factor of safety,
least twice as high, measuring peak resistances of 50 appropriately delineating where these factors of
to 120kPa. These results come with the caveat that safety are to be applied is crucial for efficient
the shaft resistance in the two layers was likely design. By introducing overly conservative design
affected by both water fluidisation and friction
components, piles that are larger than necessary are
fatigue. However, discerning these phenomena
developed. This is turn, creates more risk during
within the constraints of field testing means that it is
installation and increases the financial and
difficult to quantify the individual impact of each on
environmental impacts associated with pile
the piles.
fabrication, transport and installation.
Across the lower sand layer, much higher shaft
resistances were reached. Mobilisation of the peak
resistance rapidly occurs, within a displacement of

2238
 Session 22 - Project & Design Case Studies II

Figure 6: Shaft resistances mobilised in each soil layer. The stiff clay has been incorporated within the interlaminated clay and sand
layer

de Gijt J. and Broeken ML (2013). Quay Walls – Second Edi-

tion. Boca Raton, USA: CRC Press. (ISBN: 978-1-138-
5.1 Acknowledgements 00023-0).
Hijma MP, Cohen KM, Roebroeks W, Westerhoff WE and
The authors acknowledge the support of the Port of Busschers FS (2012). Pleistocene Rhine–Thames land-
Rotterdam Authority for financing the Amaliahaven scapes: geological background for hominin occupation of
pile tests, in addition to the remaining InPAD project the southern North Sea region. Journal of Quaternary
Science 27(1): 17–39.
partners Deltares, Fugro, Gemeente Rotterdam, te Kamp WGB (1977). Sonderen en funderingen op palen in
NVAF, Rijkswaterstaat and TU Delft. The authors zand [CPTs. Proc. Sondeer Symposium. Utrecht, the
are also grateful for the support of Buildwise, De Netherlands, 119–133 [in Dutch].
Klerk and Mariteam for assisting with the test pro- Kerisel J. (1961). Fondations profondes en milieux sableux.
Proc. of the 5th Int. Conf. Soil Mechanics and Foundation
gramme. Engineering. Paris, France, 73–83 [in French].
Kraft M. (1991. Computing axial pile capacity in sands for off-
shore conditions. Marine Geotechnology 9(1): 21–92.
6 References Kulhawy FH (1984). Limiting tip and side resistance: fact or
fallacy? Conf. on the Analysis and Design of Pile Founda-
Autin WJ (2008). Stratigraphic analysis and paleoenvironmen- tions, San Francisco, USA: 80–98.
tal implications of the Wijchen Member in the lower Rhine- Kulhawy FH (1996). Discussion: Critical depth: how it came
Meuse Valley of the Netherlands. Netherlands Journal of into being and why it does not exist. Proc. Inst. of Civil En-
Geosciences 87(4): 291–307. gineers − Geotechnical Engineering, 119: 244–245.
de Boorder M., de Lange D. and Gavin, K. (2022). An alterna- Lehane BM, Lim JK, Corenuto P, Nadim F, Lacasse S, Jardine
tive CPT averaging procedure to estimate pile base capaci- RJ and van Dijk BFJ (2017). Characteristics of unified da-
ty. Proc. 11th Int. Conf. Stress Wave Theory and Design tabases for driven piles. Proc. 8th Int. Conf. Offshore Site
and Testing Methods for Deep Foundations, Rotterdam, the Investigation Geotechnics, London, United Kingdom: 162–
Netherlands. 194.
Boulanger RW and DeJong JT (2018). Inverse filtering proce- Lehane B et al. (2020). A new “unified” CPT-based axial pile
dure to correct cone penetration data for thin-layer and tran- capacity design method for driven piles in sand. Proc of the
sition effects. Proc. 4th Int. Symp. on Cone Penetration 4th Int. Symp. on Frontiers in Offshore Geotechnics, Aus-
Testing, Delft the Netherlands, 25–44. tin, USA: 463–477
Chow FC (1997). Investigations into the behaviour of dis- van Mierlo W. and Koppejan A. (1953). Lengte en
placement piles for offshore foundations. Ph.D thesis, Im- draagvermogen van heipalen. Bouw (3) [in Dutch].
perial College London, United Kingdom. Poulos HG (2001). Piled raft foundations: design and applica-
Duffy KJ, Gavin KG, de Lange DA and Korff M (2022). Re- tions. Géotechnique 51(2): 95–113.
sidual stress measurement of driven precast piles using dis- Rijsdijk KF, Passchier S, Weerts HJT, Laban C, van Leeuwen
tributed fibre optic sensors. Proc. 11th Int. Conf. Stress RJW and Ebbing JHJ (2005). Revised Upper Cenozoic stra-
Wave Theory and Design and Testing Methods for Deep tigraphy of the Dutch sector of the North Sea Basin: to-
Foundations, Rotterdam, the Netherlands. wards an integrated lithostratigraphic, seismostratigraphic
Fellenius BH and Altaee AA (1995). Critical depth: how it and allostratigraphic approach. Netherlands Journal of Geo-
came into being and why it does not exist. Proc. Inst. of sciences 84(2): 129–146.
Civil Engineers − Geotechnical Engineering, 113(2): 107– Robinsky EI and Morrison CF (1964). Sand displacement and
111. compaction around model friction piles. Canadian Ge-
Fleming WGK, Weltman A, Randolph M and Elson K (2008). otechnical Journal 1(2): 81–93.
Piling Engineering – Third Edition. London, United King- Schneider JA, Xu X. and Lehane BM (2008). Database as-
dom: Taylor & Francis (ISBN: 978-0-415-26646-8). sessment of CPT-based design methods for axial capacity
of driven piles in siliceous sands. Journal of Geotechnical
and Geoenvironmental Engineering 134(9): 1227–1244.

2239
 Innovative Geotechnologies for Energy Transition | The Society for Underwater Technology

Vesic AS (1965). Ultimate loads and settlements of deep foun-

dations in sand. Proc. Symp. on Bearing Capacity and Set-
tlement of Foundations, Durham, USA: 53–68.
Vesic AS (1970). Tests on instrumented piles, Ogeechee River
Site. Jounral of the Soil Mechanics and Foundations Divi-
sion 96(2): 561–584
White DJ and Bolton MD (2005). Comparing CPT and pile
base resistance in sand. Proc. of the Institution of Civil En-
gineers - Geotechnical Engineering 158(1): 3–14.
Yang ZX, Jardine RJ, Guo WB and Chow FC (2015). A new
and openly accessible database of tests on piles driven in
sands. Géotechnique Letters 5(1): 12–20.

2240
View publication stats