See discussions, stats, and author profiles for this publication at: [Link]

net/publication/318672602

Driven Pile Design Methods in Weak Rock

Article · September 2017

CITATION READS

1 2,644

4 authors:

Victor Terente Jamie Irvine

Cathie Associates Cathie Associates
12 PUBLICATIONS   24 CITATIONS    6 PUBLICATIONS   9 CITATIONS   

SEE PROFILE SEE PROFILE

Ismael Torres Christophe Jaeck

Cathie Associates Cathie Associates
6 PUBLICATIONS   3 CITATIONS    11 PUBLICATIONS   75 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Design of Piles in Weak Rock - follow up paper - SUT conference, London 2017 View project

ammendment of author names (mine was incorrect in the original paper) View project

All content following this page was uploaded by Victor Terente on 03 October 2018.

The user has requested enhancement of the downloaded file.

DRIVEN PILE DESIGN METHOD FOR WEAK ROCK

V Terente, I Torres, J Irvine and C Jaeck

Cathie Associates Ltd., Newcastle, UK

Abstract

There is currently limited guidance regarding the design of driven piles in weak rock. Due to the absence of
design codes and the conservative design methods currently employed, the capacity of the pile sections embed-
ded into rock are often under predicted; the resistance to driving is also not well understood and usually over
predicted to ensure appropriate installation equipment is mobilised.

Significant driven pile penetrations are now possible in weak rock. Available pile test results often demonstrate
that the pile capacity is significantly higher than the current methods would anticipate. However, a general
explanation of the measured values has not been attempted to date.

The aim of this paper is to discuss the results of available pile testing data and to propose a new generalised
approach for estimating shaft capacity. A better understanding of the capacity of driven piles in weak rock has
the potential to allow significant optimization of pile lengths and to reduce installation risks and costs.

1. Introduction weathered mudstones) and has also proven to

There is no generally accepted pile design method for provide reasonable predictions in very weak
driven piles in weak rock, and no commonly accepted to weak calcareous sedimentary rock
design codes for the assessment of driven pile shaft (Thomas et al, 2011):
capacity in those materials. Shaft capacity is key to
piled foundations in weak rock where the critical load Qs    su  As (2)
case is in tension, as it is the case in many offshore
jacket structures. Irvine et al (2015) summarised the Where su = undrained shear strength (UCS/2);
two main current approaches for estimating shaft ca-  = adhesion factor.
pacity in weak rock as follows:
Using the main text method of API RP2GEO
 Effective stress method: (2014), the adhesion factor can be calculated
as:
Qs  K   ' v  tan   As  f l  As (1) 𝑈𝐶𝑆 −0.5
0.5 ( )
2∗𝑝′0
𝛼 = 𝑚𝑖𝑛𝑖𝑚𝑢𝑚 { −0.25
(3)
𝑈𝐶𝑆
Where: Qs = shaft capacity, K = coefficient of 0.5 ( )
2∗𝑝′0
lateral earth pressure; 𝜎𝑣′ = effective overbur-
den stress; δ = friction angle between the rock Where: UCS = unconfined compressive
and pile wall; As = surface area of the pile strength of the rock mass; p’0 = effective over-
shaft; fl = limiting skin friction. burden pressure.

 Total stress method where a pile-rock adhe- Rock mineralogy, anisotropy, weathering state, rock
sion factor is calculated. This is generally ac- mass properties (particularly strength and stiffness)
cepted for mudstones (particularly completely and in situ stress are all factors affecting pile capacity.
In addition, both disturbance of the rock mass and the Thomas et al (2011) undertook an assessment of the
change of in situ stresses induced during driving will uplift capacity of tubular piles driven into variably ce-
also affect the short term and long term capacities. mented, very weak to weak sedimentary rock with
But the quantification of these effects is not well doc- varying calcium carbonate content. The rock strength
umented to date. generally ranged from 0.3MPa to 4MPa, with an av-
erage strength of ~0.8MPa; the uplift tests indicated a
While it is important to compare the current design total uplift capacity of 4,610kN with an average shaft
practices with known pile test data, it is very relevant capacity of 87kPa.
to propose design method(s) which allow(s) a more
accurate assessment of shaft capacity for driven piles. The unit shaft capacity derived from those tests and
The proposed methods should take into account not others has been plotted as a logarithmic scale against
only the rock material strength or frictional proper- UCS values and compared with the capacity curves
ties, but also rock mass properties, in situ stresses, and of grouted piles published in CIRIA report 181 - Piled
mechanisms associated with the driving process. Foundations in weak rock (2004).

2. Driven pile capacity in weak rock – published Figure 1. Adhesion factor (α) against rock material UCS
literature data and back analysis
Published pile loading test results are sparse and only
a few relevant pull out tests are reported in the last
few decades. Some relevant papers are summarised
below.

Beake and Sutcliffe (1980) observed ultimate shaft

resistances of 170kPa and 300kPa from tension tests
on steel tubular piles driven into weak (3.2MPa to
4.7MPa UCS) carbonate siltstones and sandstones in
the Arabian Gulf.

Rodway and Rowe (1980) investigated the tension ca-

pacity of two steel H-piles driven into medium strong
sandstone in Sydney Harbour, Australia. The rock
strength was indicated to reach in excess of 20MPa.
The pile penetration depths were noted to be rela-
tively limited (only 1m to 1.5m). However, the piles Figure 1 illustrates that the relationship between the
developed significant tension capacity which ex- rock material UCS from test data and the adhesion
ceeded the measurement capacity of the extraction factor  is not linear and potentially follows a similar
equipment. The most conservative (low) estimation pattern (reverse exponential) to the one displayed by
of the skin friction was ~230kPa but the average skin grouted piles; hence at lower material UCS, a larger
friction was probably much higher. ratio of  can be expected. A regression analysis of
lower bound test results shows the following expo-
ARGEMA (1994) summarises several papers regard- nential relationship:
ing driven piles in carbonate rocks and details some
internal research, which is not publicly available:    0.11  UCS 0.5 (3)
 Settgast (1980) agrees with the findings of Where: adhesion factor; UCS = unconfined com-
Beake and Sutcliffe (1980). pressive strength of the rock mass.
 Burt et al (1980) and Puech et al (1988) indi-
cate shaft friction in coarse cemented gravel Based upon these results, the exponential relationship
of 10% to 20% of rock UCS. for the derivation of an adhesion factor outlined
 Internal ARGEMA experiments (Montarges, above is proposed for estimating the shaft capacity of
1988) with small diameter piles in limestone driven piles in weak rock:
(UCS 5MPa to 25MPa) which indicate shaft
friction values of 7% to 9% of UCS. Qs    UCS  As  f l  As (4)
Where: Qs = shaft capacity; As = surface area of the masses. Gonzalez de Vallejo & Hijazo (2008) pro-
pile shaft; fl = limiting skin friction for piles driven in posed K values (derived mainly from Western Euro-
weak rock. pean data) generally ranging between <1 to 5 depend-
ing on the intensiveness of the tectonic history
The adhesion factor significantly decreases at higher suffered by the rock mass (measured by the Tectonic
UCS values and is similar, although more conserva- Stress Index or ‘TSI’):
tive, than the one used for pile design in hard clays
based on API RP2GEO (2014) and DNV 30.4 (1991). Table 1. Gonzalez de Vallejo and Hijazo (2008). Indicative K
value ranges for tectonised rock masses.
Tectonic Stress level K value from database
Other factors, apart from UCS, could influence the
mobilised shaft capacity as there is no obvious rela- Low <1
Medium 1 - 1.5
tionship with UCS only. In situ radial stresses and set-
High 1.5 - 2
up effects after driving could be higher in weaker Very High 2 - >5
rocks due to the significant relaxation of the rock
mass and therefore potentially higher lateral stresses. In many areas, rocks affected by the Hercinian of Al-
It is therefore important to assess the in situ stresses pine orogeny are at or near the surface. Values of K
of the rock mass in order to explain the potential in- in excess of 1.5 are therefore common in weak rocks
fluence of high lateral stresses. in the North Sea. There are other factors such as
weathering, stiffness of individual strata, faulting and
3. In situ stress distribution in rock masses anisotropies associated with bedding (Amadei et al,
The determination of the in situ horizontal rock mass 1988; Amadei, 1996) which will influence the final K
stress is key for the assessment of shaft capacity using values to be adopted for design.
a frictional approach. Horizontal in situ stress in a The exact value of in situ stresses can be measured
rock mass is defined as: with pressuremeter testing to ensure that those effects
are accounted for.
𝜎′ℎ = 𝐾 ∙ 𝜎 ′ 𝑣 = 𝐾 ∙ 𝛾′ ∙ 𝑧 (5)
4. Recent experience with pressuremeter tests in
Where: 𝜎′ℎ = in situ effective horizontal stress before weak rock
piling, K = coefficient of earth pressure, 𝜎𝑣′ = in situ Experience from several offshore wind farms across
effective overburden stress, γ' = effective rock mass the North Sea illustrates that K values below rock
density, z = relevant depth. head can range between 1 and 8 for both clastic and
carbonaceous rocks. The figure below illustrates
The K value for rock masses should be obtained from measured values of K (conservatively obtained as lift-
pressuremeter tests during the ground investigation off pressure versus estimated overburden pressure for
stage. However, in the absence of test data empirical High Pressure Dilatometer (HPD) tests) obtained in
formulas have been proposed to determine this pa- the North Sea and Irish Sea for mudstone, siltstones
rameter, taking into account the in situ stresses of the and chalk rock masses.
rock and the rock mass young modulus. Sheorey
(1994) proposes a formulation based on the average Figure 2. K values vs. Depth below seafloor.
deformation modulus of the rock mass, as follows:
1
𝐾 = 0.25 + 7 𝐸ℎ (0.001 + ) (6)
𝑧

Where: K = coefficient of earth pressure, Eh = Rock

mass Young’s Modulus, z = relevant depth.

As this equation is defined for the deeper Earth crust,

it should be used with caution for shallow depths (less
than 50m). Eh values for rock masses can be estimated
from the Hoek-Brown failure criterion (Hoek et al,
2002) in order to define preliminary K values of the
rock mass. However, in the opinion of the authors this
formulation may significantly overestimate K values
at the shallow depths relevant to piling for weak rock
The individual K profile is likely to be specific to a were installed in the Irish Sea through Holocene de-
windfarm area and bounded by major regional struc- posits, glacial till and into Triassic Mercia Mudstones
tural geology and formation boundaries. consisting of mudstones, siltstones and sandstone
units. Key pile characteristics are summarised below.
5. Steel-rock friction
The friction generated at the rock and steel interface Table 2. Pile and Hammer Details.
is a key mechanism affecting the shaft resistance of Parameter Value
driven piles. Ziogos et al (2015a) studied and dis- Pile Diameter 50" (1270mm)
cussed some parameters affecting the shear strength Wall Thickness 45mm
of this interface, which include: UCS, state of weath- Pile Steel Grade S335
ering of the rock mass and steel roughness. Ziogos Final Pile Penetration Variable
tested the shear interface between mortar blocks and
steel. Mortar strengths ranging between 15MPa to Soil and rock layering depths are illustrated in the fig-
>50MPa and normal stress levels up to 200kPa were ure below. The rock cores were typically described as
tested. It was found that the UCS normalised shear extremely weak to very weak dark reddish mudstone,
stress of the interface generally increases with de- with variable weathering profiles.
creasing compressive strength values and increasing
steel roughness; this behaviour was attributed to the Figure 3. Soil Profile.
localised shear failure of the interface at higher com-
pressive strength values.
Ziogos et al (2015b) reported variations of steel
roughness and their influence on the residual interface
friction angle. They concluded that values of the peak
interface friction angle (p determined by the USBR
6258-09 test (USBR, 2009) are generally ranging be-
tween 0.6 and 0.8 of the ‘basic’ rock friction angle
(. Values of residual interface friction angle
(rwere reported between 0.41 to 0.73 of p for Old
Red sandstone and Caithness sandstones and steel av-
erage roughness (Ra) ranging between 0.4-7.2μm.
The roughness of offshore piles is typically between
15-25µm. 

6. Stress modifications during driving

The process of pile driving into weak rock and asso-
ciated stress changes along the shaft immediately af-
ter driving and long term is not well understood. The
equivalent change of local radial stiffness based on
cavity expansion theory was discussed by Jardine et
al (2005) for sands. More complex mechanisms asso-
ciated with radial stresses could be associated with
the level of fracturing of the rock mass, rock block
displacements, the crushability of the rock material,
the use of pile driving shoes and cyclic degradation
effects. However, there is currently no published in-
formation to assess those effects in detail.
The pile driving process was monitored with Pile
8. Case study Driving Monitoring (PDM) techniques and back ana-
A case study of an offshore jacket structure, driven lysed with CAPWAP software to derive estimated
into Mercia Mudstone can be used to illustrate the ef- shaft capacities. Back analysed capacities were well
fective stress approach for weak rock pile design ver- in excess of the initial pile design, based on a cohesive
sus ‘traditional’ methods based on an alternative ‘to- approach. The following lower bound geotechnical
tal stress method’ contained in the API RP2GEO parameters were used for the pile design and the back
(2011) and described in section 1. A series of piles analysis of the capacities based on an alternative ef-
fective stress approach.
 9. Conclusions
Table 3. Geotechnical Parameters. Existing data regarding capacity tests of driven piles
Rock Design Parameters in weak rock has been presented and discussed in this
Interface paper. The existing data suggest that weaker rock
Unit
Soil Type UCS* GSI K Rock-Steel masses are likely to generate a higher proportion of
Weight
[MPa] [-] [-] Friction shaft capacity than allowed for in the traditional de-
[kN/m3]
Angle [°]
MM** 21-24 1-2 25-45 2.5 25-35
sign methods, and that the lower bound fit of the ex-
isting data follows a reverse exponential.
*Lower bound UCS obtained from UCS tests and point load data.
**MM Mercia Mudstone.
A method to calculate shaft capacity on the basis of
an effective stress approach has been presented and
Where: GSI = Geological Strength Index. illustrated with a case study. However, it is acknowl-
edged that this design approach is heavily reliant on
As it can be seen in the estimated shaft capacity plot in situ testing, rock mass parameters and good ground
presented below, an effective stress approach pre- investigation data. This information is likely to come
sented a better explanation of the significant shaft ca- in the project at a later point in the design process,
pacity estimates reported by PDM. In the absence of during Final Investment Decision (FID) or detailed
pull out tests these results should be treated with cau- design; therefore, a staged approach to design, based
tion, however, it is clear that increases in shaft capac- on an increasing level of rock mass testing is recom-
ity with depth cannot be explained with UCS test re- mended:
sults alone.
1. An initial estimate of shaft capacity based on
Figure 4. Comparison of estimated and in situ pile shaft fric-
tion (in situ as back analysed from PDM data). typical UCS ranges and an ‘adhesion factor
approach’ or the application of the total
stress method as detailed in equation 2.
2. An intermediate approach based on UCS and
rock mass parameters obtained from individ-
ual boreholes to estimate the in situ lateral
stresses and the application of the effective
stress method as detailed in Section 3.
3. Use of in situ testing and HPD data applica-
ble to each formation and application of the
effective stress method.
This approach should also be complemented with on-
shore pile testing in the same formations if possible
or monitoring of piles installed at an early stage of
development (for instance met mast). An illustration
of this approach for wind farm projects is summarised
in the figure below.

Figure 5. Proposed staged design approach.

The under estimation of shaft friction may be due to

the available UCS and GSI description under-repre-
senting the rock mass characteristics.
10. References Settgast, R.H. (1980). Marine Pile Load Testing in
Amadei B., Swolfs H. and Savage W.Z. (1988). Carbonate Rocks. Proc. 12th Annual OTC, Hou-
Gravity-induced Stresses in Stratified Rock ston, Paper OTC 3868.
Masses. Rock Mechanics and Rock Engineering Sheorey P.R. (1994). A Theory for In Situ Stresses in
21; 1-20. Isotropic Rock. Int. J. Rock Mech. Min. Sci. & Ge-
Amadei B. (1996). Importance of Anisotropy When omech. Abstr. Vol. 31; 23-34.
Estimating and Measuring In Situ Stresses in Thomas J., van den Berg M., Chow F. and Maas N.
Rock. Int. J. Rock Mech. Min. Sci. & Geomech. (2011). Behaviour of driven tubular steel piles in
Abstr. Vol. 33, No. 3; 293-325. calcarenite for a marine jetty in Fujairah, United
API, (2011), API Recommended Practice 2Geo Ge- Arab Emirates. Frontiers in Offshore Geotechnics
otechnical and Foundation Design Considerations, II, London. Taylor and Francis Group; 549-554.
API RP2GEO, USBR (2009). Procedure for Determining the Angle
ARGEMA (1994). Foundations in Carbonate Soils; of Basic Friction (Static) Using a Tilting Table
English Edition. Technip. Paris. Test; Designation USBR 6258-09.
Beake R.H. and Sutcliffe G. (1980). Pipe pile driva- Ziogos A., Brown M.J., Ivanovic A. and Morgan N.
bility in the carbonate rocks of the Southern Ara- (2015). Rock-steel interface testing and considera-
bian Gulf. Proc. International Conference on tions for gravity foundations for tidal energy gen-
Structural Foundation on Rock, Vol 1; 133-149. erators. Proc. Int. Symp. Frontiers in Offshore Ge-
Burt N.J. and Harris R.P. (1980). “Design installation otechnics III, ISFOG 2015. London: Taylor and
of belled pile foundation”, Proc. 12th OTC, Houston, Francis, 1245-1250.
OTC 3872. Ziogos A., Brown M.J., Ivanovic A. and Morgan N.
CIRIA (1999). Report 181 - Piled foundations in (2015). Interface shear characteristics of Scottish
weak rock. London. rock samples from sites with tidal energy potential.
DNV, (1992), Classification Notes – Foundations, Proc. XVI European Conference on Soil Mechan-
DNV No. 30.4. ics and Geotechnical Engineering, ECSMGE
DNV (2014). DNV-OS-J101 Design of Offshore Wind 2015. ICE Publishing; 1357-1362.
Turbine Structures. Det Norske Veritas.
Gonzalez de Vallejo L.I. and Hijazo T. (2008). A new
method of estimating the ratio between in situ rock
stresses and tectonics based on empirical and
probabilistic analyses. Engineering Geology, Vol-
ume 101, Issues 3–4; 185–194.
Hoek E., Carranza-Torres C. and Corkum B. (2002).
Hoek-Brown criterion – 2002 edition. Proc.
NARMS-TAC Conference, Toronto, 2002, 1; 267-
273.
Irvine J., Terente V., Lee L.T. and Comrie R. (2015).
Driven pile design in weak rock. Proc. Int. Symp.
Frontiers in Offshore Geotechnics III, ISFOG
2015. London: Taylor and Francis, 569-574.
Jardine R., Chow F., Overy R. and Standing J. (2005).
ICP design methods for driven piles in sands and
clays. Thomas Telford.
Montarges R. (1988). Frottement des pieux battus
dans les formations carbonatées. IFP 36018 (In-
ternal report)
Puech A., Becue J.P. and Colliat J.L. (1988). Ad-
vances in the design of piles in non-cemented to
weakly cemented carbonate formations. Proc. In-
ternational Conference on Calcareous Sediments,
Vol 1; 305-312.
Rodway and Rowe (1980) Rodway R.L. and Rowe
R.K. (1980). The uplift Capacity of Steel Piles
Driven into Hawksebury Sandstone. Proc. 3rd Aus-
tralian and New Zealand Conference on Geome-
chanics; 109-114.

View publication stats