You are on page 1of 5

DESIGN OF DYNAMIC FOOTINGS IN THE PILBARA REGION OF

WESTERN AUSTRALIA1

Simon Hope
Golder Associates Pty Ltd
Level 3, 1 Havelock Street, West Perth, WA, (08) 9213 7600, shope@golder.com.au

ABSTRACT
With the continuation of Western Australia’s resources boom, the construction of specialised mining infrastructure in
remote locations provides interesting and unique challenges to geotechnical engineers. Large ore-processing structures,
such as screening plants and crushing facilities, are designed to vibrate as part of their routine operations. This forces
the footing to withstand dynamic loading that can potentially result in large differential settlements or bearing capacity
failures if not founded on suitable material or designed appropriately.
To assist the structural and mechanical design of these dynamically loaded structures, a dynamic shear modulus value is
required for the soil-structure interaction. The unique geological and geotechnical conditions of the Pilbara region
present specific challenges and yet there is limited published information available regarding methods of assessing the
dynamic shear modulus for these conditions.
This paper presents a summary of relevant published information, investigation methods and a weighted average
method of assessment to estimate the dynamic shear modulus for typical Pilbara ground conditions.

1 INTRODUCTION
The design of dynamic footings requires an assessment of parameters for the design subsurface soil profile. The
methods for assessing the dynamic shear modulus (G) for soils in the Pilbara region will depend on the subsurface
material present at the investigation site. This paper presents a method for calculating a dynamic shear modulus value
for use in the design of dynamic footings, applicable to the subsurface conditions typically encountered in the Pilbara
region.

2 PILBARA SUBSURFACE CONDITIONS


The subsurface conditions in the Pilbara region are greatly diverse due to the size of the region and complex geology
and geomorphology (Hickman, 1983). From recent experience on geotechnical investigations in the Pilbara, the
surficial soil and rock that may be encountered in the zone of influence for a foundation can be generalised as:
 Alluvium – typically silty/clayey sands (fine grained, red-brown, approximately 20% to 30% low to medium
plasticity fines, loose to dense ) or sandy clays (medium to high plasticity, red brown stiff to hard);
 Colluvium – typically silty/clayey sandy gravels (fine to coarse grained, sub angular, approximately 20% to
30% low to medium plasticity fines, often with cobbles, medium dense to very dense);
 Cemented materials – conglomerate or ferricrete (weakly to well iron-cemented Alluvium and Colluvium
deposits);
 Tertiary aged sediments – silt/clay, siltstone/claystone (fine grained, extremely low to medium strength,
residual soil to distinctly weathered);
 Igneous deposits – typically dolerite or granite (fine to medium /coarse grained, very low strength to high
strength, extremely weathered to distinctly weathered, generally increasing in strength and becoming less
weathered with depth);
 Proterozoic metasediments – chert, shale, banded iron (fine grained, layered, high to extremely high strength,
slightly weathered to fresh).
Groundwater is generally not encountered in zones influenced by loading.

3 IN SITU TESTING
The two general in situ methods for assessing the dynamic properties of a soil involve geophysical and geotechnical
tests. The use of geophysical or seismic tests to assess dynamic soil properties is generally accepted as a reliable method

1
This paper was presented at the 2011 Baden Clegg Award

Australian Geomechanics Vol 47 No 2 June 2012 87


DESIGN OF DYNAMIC FOOTINGS IN THE PILBARA REGION OF WESTERN AUSTRALIA SIMON HOPE

of assessing the soil profile. Dynamic parameters of the soil can be derived from measuring the velocity of seismic
waves, including compression or primary (P) waves and shear or secondary (S) waves. These measured velocities are
referred to as P wave velocity (Vp) and S wave velocity (Vs).
Investigation methods can be divided into two distinct areas:- non-intrusive surface wave tests and intrusive down-hole
soil profiling using cased boreholes.
Non-intrusive test methods include Multi-channel Analysis of Surface Waves (MASW) and Refraction Micro-tremor
(ReMi), both of which involve measuring the velocity of surface (shear) waves using uniform spaced geophones. As
these tests record Vs from the ground surface, the confidence and accuracy of the test results generally decrease with
increasing depth. Tests can also be affected or not identify bands of low velocity material in between high velocity
material and vice-versa.
Intrusive methods typically require the drilling of boreholes and installation of PVC casing in uncemented gravels for
down-hole or cross-hole seismic tests to ensure good contact is made between the probe and the soil mass. The cross-
hole seismic test can be performed with two or more boreholes to improve the confidence level of the recorded values.
These tests are generally only limited by the depth of the borehole and can pick up bands of varying velocity material.
As the Pilbara soils are often coarse alluvial gravels or parent rock, most geotechnical investigations include the drilling
of boreholes to perform in situ tests (i.e. standard penetration test and pressumeter), visually logging the material, and
collecting core samples for laboratory testing (i.e. point load index and unconfined compressive strength tests). These
tests are used to assess the elastic modulus at a large strain level, E s, which can then be converted into a shear modulus
value relative to that strain level, Gs, using Equation 3. Using the correlation in Figure 1, G s can be adjusted for smaller
strain levels to provide an assessment of Gd. Alternatively, Kramer (1996) provides a summary of direct correlations
for Gmax from conventional in situ geotechnical tests.
Very small strain ultrasonic pulse velocity tests based on bender elements (Vigianni and Atkinson 1995), can also be
performed during unconfined compressive strength tests of core samples to provide values for V p and Vs. Although,
these are not suitable for soil strength material, i.e. gravels.
As an alternative investigation technique to drilling boreholes, plate load testing of foundations can be performed to
estimate a value of Es.
The test method most applicable to any individual site will vary depending on factors such as foundation material, size
and configuration of the structure, site access and available budget. In addition seismic testing, including a direct
measurement of the shear velocity profile, should ideally be performed in conjunction with conventional geotechnical
testing at the proposed structure location.
Using the shear velocity profile, Gmax can be calculated from the S or P waves using Equations 1 or 2 respectively.
From the stiffness-strain relationship presented in Figure 1, G max can then be factored to estimate Gd, the dynamic shear
modulus for a given strain level, for use in dynamic foundation design.

4 DYNAMIC SHEAR MODULUS


Values for the dynamic shear modulus (G) are generally required for the design and analysis of foundations under
dynamic loading. Values for G can be calculated using both dynamic and static soil parameters, assuming the soil
behaves elastically (Bowles, 1997).
It is customary to use S wave velocity (Vs) when calculating G. A value for G can be obtained using the following
equation:
Where is the density of the soil in kg/m3
(1)
A value for Vs can also be derived from the P wave velocity (V p), using Poisson’s ratio ():
(2)

Values for Vp and Vs are generally derived from in situ geophysical tests as mentioned in the previous Section.
The relationship between the shear modulus of a soil and its stress-strain elastic Modulus (E) is the same for dynamic
conditions as it is in static conditions, which is as follows:
(3)
Values for E can be derived using conventional in situ geotechnical tests.

88 Australian Geomechanics Vol 47 No 2 June 2012


DESIGN OF DYNAMIC FOOTINGS IN THE PILBARA REGION OF WESTERN AUSTRALIA SIMON HOPE

Examples of typical Vs and Gmax profiles from an investigation site in the Pilbara are presented as Figure 1a and Figure
1b respectively
When calculating the dynamic shear modulus, G, using ‘static’ geotechnical tests, the G value will need to be adjusted
due to the non-linear stiffness-strain relationship of a soil. Although this relationship is well known to decrease with
increasing strain, defining this relationship is still the focus of many studies in this field (Atkinson, 2000).
Most studies consider a normalised shear modulus (stiffness) value is derived from dividing G by an assumed
maximum dynamic shear modulus value, G max. This value of Gmax is typically considered to be G at a very small level
of strain (<10-6). For static loads, the level of strain is generally considered to be large and ranges from 0.1% upwards
(e.g. 5 mm displacement over 5 m footing). For vibrating and mechanical structures the level of strain developed is
generally considered to be small and between the range of about 10-3 and 10-2 (Arya et al., 1979).

(1a) (1b)
Figures 1a and 1b: Typical plots of Vs and Gd versus depth for a site in the Pilbara region.
The majority of studies assessing stiffness-strain relationship have to date generally investigated fine grain soils (clays
and silts) and sands, which are generally not applicable for sites in the Pilbara. Although there are areas of recent and
Tertiary aged alluvial deposits in the Pilbara the majority of sites investigated for crushing and screening plants are
typically located in areas of shallow colluvium overlying weathered siltstone/granite or fresh chert/banded iron.
Seed et al. (1986) highlighted the limited amount of published information on the dynamic modulus of gravel materials,
and provided an assessment of the stiffness-strain relationship for a range of gravel materials. The assessment included
a dynamic modulus estimate using cyclic undrained triaxial laboratory tests on crushed rock and natural gravels from
the Californian area. Results published by Seed et al. (1986) identified the stiffness-strain relationship for the full range
of gravel materials to be relatively consistent, if not slightly flatter than the comparisons developed for sands by Seed
and Idriss (1970). Of note, the authors also commented that the variation of material grading and relative density had
little effect on the shear modulus compared with more significant variations from changes in strain.
The range of normalised shear modulus values versus shear strain for gravel soils presented by Seed et al. (1986) has
been reproduced as Figure 2.

Australian Geomechanics Vol 47 No 2 June 2012 89


DESIGN OF DYNAMIC FOOTINGS IN THE PILBARA REGION OF WESTERN AUSTRALIA SIMON HOPE

Figure 2: Variation of shear modulus with shear strain for gravelly soils (Seed et al., 1986)

5 WEIGHTED AVERAGE G
The response of a dynamic foundation under loading is often difficult to model due to the complexities of modelling
soil and structural or dynamic loads simultaneously. Often the soil-structure interaction is modelled using a lumped
mass or lumped parameter method (i.e. springs and dampeners) by mechanical/structural engineering software. Using
this method, values for dynamic shear modulus and Poisson’s ratio are required to calculate spring and dampener
constants to simulate the soil behaviour. As the soil profile is typically non-homogeneous, calculating a weighted
average dynamic shear modulus is one method for representing the behaviour of the soil profile in the zone of influence
beneath a footing.
Assuming the soil behaves elastically, the weighted average dynamic modulus can be calculated by factoring the
dynamic shear modulus values for each stratum based on the Boussenesq stress distribution ratios below a footing as
presented in Figures 3a and 3b. The stress distribution ratio can be described as q/q 0, where q is the stress at any point
below the footing and q0 is the initial or applied stress at the surface.

(3a) (3b)
Figures 3a and 3b: Pressure distributions based on the Boussenesq equation (from Bowles, 1997)

90 Australian Geomechanics Vol 47 No 2 June 2012


DESIGN OF DYNAMIC FOOTINGS IN THE PILBARA REGION OF WESTERN AUSTRALIA SIMON HOPE

Once the soil profile and dynamic soil parameters have been developed for the site, a weighted average dynamic shear
modulus can be calculated. As previously mentioned, the dynamic stress-strain distribution behaves consistently with
the static stress-strain distribution beneath a footing, allowing us to reduce the dynamic shear modulus in proportion to
the stress reduction factors presented in Figure 2.
The proportioning of the modulus values to calculate the weighted average value can be done using finite layer elastic
analysis software or by small squares calculations (effectively assuming numerous thin soil layers).

6 CONCLUSIONS
Based on the limited studies that have been performed on the stiffness-strain relationship for gravel materials, the
relationship can generally be assumed to fit the limits of the sand stiffness-strain curve.
Variations to Poisson’s ratio and the soil density typically have little effect on the dynamic soil parameters, however
variations in strain level can significantly affect the foundation design.
A combination of both geophysical and geotechnical test methods is recommended when assessing soil parameters for a
soil profile.
Due to the comparable stress-strain behaviour between the dynamic and static shear modulus, a weighted average
dynamic shear modulus value can be assessed using the Boussenesq stress distribution below a footing.

7 REFERENCES
Arya, S.C., O'Neill, M.W., and Pincus, G., 1979. Design of Structures and Foundations for Vibrating Machines. Gulf
Publishing Co., Houston
Atkinson, J.H., 2000. Non-linear soil stiffness in routine design. Geotechnique, 50(5), pp. 487-508.
Bowles, J. E., 1997. Foundation Analysis and Design, 5th Edition. McGraw-Hill, New York.
Hickman, A.H., 1983. Geology of the Pilbara Block and its environs. Geological Survey of Western Australia, Bulletin
127
Kramer, S.L., 1996. Geotechnical Earthquake Engineering. Prentice Hall Inc., New Jersey
Seed, H.B. and Idriss, I.M., 1970. Soil moduli and damping factors for dynamic response analysis. Report No. EERC
75-29, Earthquake Engineering Research Center, University of California, Berkeley, California.
Seed, H.B., Wong, R.T., Idriss, E.M., and Tokimatsu, K., 1986. Moduli and Damping Factors for DynamicAnalyses of
Cohesionless Soils. Journal of Geotechnical Engineering, Vol. 112, No. 11, pp. 1016-1032
Viggiani, G. and Atkinson, J.H., 1995a. Interpretation of bender element tests. Geotechnique, 45(1), pp. 149-154.
Viggiani, G. and Atkinson, J.H., 1995b. Stiffness of fine grained soils at very small strains. Geotechnique, 45(2), pp.
249-265.

Australian Geomechanics Vol 47 No 2 June 2012 91

You might also like