European Journal of Scientific Research

ISSN 1450-216X Vol.80 No.4 (2012), pp.472-478

© EuroJournals Publishing, Inc. 2012
http://www.europeanjournalofscientificresearch.com

Estimation of Interface Resistance between Testing

Chamber and Soil Model using Shear Box Test

Ahmad Safuan A. Rashid

Department of Geotechnics and Transportation
Universiti Teknologi Malaysia, Malaysia
E-mail: ahmadsafuan@utm.my
Tel: +605-5531504; Fax: +607-5566157

Norhazilan Md. Noor

Reliability Engineering & Safety Assessment (RESA)
Universiti Teknologi Malaysia, Malaysia
E-mail: norhazilan@utm.my
Tel: +605-5531626; Fax: +607-5566157

Abstract

This paper performs an estimation of interface resistance between testing chamber

and soil model using shear box test. A different of bearing capacity factor was observed
between the result from the physical and numerical modelling tests for soft soil reinforced
by a group of soil cement columns. It was assumed that the interface resistance produced
during the loading of the footing between the clay bed and the chamber sides (Perspex and
Aluminium walls) contributed to the different. In order to determine the interface resistance
between the test chamber walls (Perspex or aluminium) and the soil surface, a series of
shear box tests was conducted. An interface resistance correction was proposed using the
results from the shear box tests and an upper bound approach. By doing so, the inclusion of
the resistance minimised the differences between the laboratory and numerical modelling
bearing capacity results. As a conclusion, it was found that the shear box test and upper
bound solution provided a better estimation of interface resistance on the physical
modelling test.

Keywords: Shear box test, upper bound.

1. Introduction
Several studies have been conducted to study the performance of soft clay reinforced by a group of soil
cement columns using a small scale physical modelling test (Boussida and Porbaha, 2004a, 2004b). A
different was observed from the numerical and analytical solution in order to validate the bearing
capacity result from the laboratory test, (Boussida and Porbaha, 2004a, 2004b). It was assumed that the
interface resistance produced during the loading of the footing between the clay bed and the chamber
sides contributed to the different.
The estimation of the effect of the wall friction inside the chamber box and the soil model
ground during the loading test was proposed by Boussida and Porbaha (2004a, 2004b). By assuming
the effect of the resistance into three types of contact which are smooth, medium roughness and
Estimation of Interface Resistance between Testing
Chamber and Soil Model using Shear Box Test 473

roughness, they found that the friction contributed about 2.6% to 4% of bearing capacity value.
However, there are still uncertainties in estimating the effect of the resistance such as the value of shear
stress friction between the soil and the chamber wall.
Therefore, in this study, a series of shear box test and upper bound methods has been conducted
to investigate the problem. The wall friction correction was proposed to a series of physical modelling
test results to obtain a better validation result with a numerical model. The physical model tests were
performed to investigate the failure behaviour of the Deep Mixing (DM) method. The study was
consisted a series of fully and partially penetrated DM method under rigid and flexible footings.

2. Physical Modelling Test

A series of small scale (1g) physical modelling test was conducted to study on the performance of the
soft soil reinforced by a group of soil cement columns. The study was consisted a series of fully and
partially penetrated soil cement columns under rigid and flexible footings. The model of a soft ground
(speswhite kaolin clay) was prepared in a rigid box having dimensions 400 mm width, 150 mm length
and 430 mm high as shown in Figure 1. The sides of the box (Aluminium and Perspex at front) were
made from rigid block to prevent lateral movement during the consolidation of the model ground. The
improvement ground was load under undrained condition using a constant rate displacement device.

Figure 1: Equipment and test setup

The results from the laboratory was validated using a new numerical modelling software known
as LimitState:Geo (Smith and Gilbert, 2007; Gilbert et al., 2010) and homogenisation method (Jellali et
al., 2005; Omine et al., 1999). Figure 2 shows the relationship between the Nc values obtained from the
laboratory and the one from LimitState:Geo, represented by the dash line. It was assumed that the
474 Ahmad Safuan A. Rashid and Norhazilan Md. Noor

interface resistance produced during the loading of the footing between the clay bed and the chamber
sides (Perspex and Aluminium walls) contributed to the differences. Therefore, a correction should be
conducted to the laboratory results in order to match the line of equality because the Nc values from the
laboratory work were underestimated.

Figure 2: Comparison between Nc determined by LimitState and Laboratory.

18.0

16.0

14.0
Nc from LimitState

12.0

10.0

8.0

6.0 Unreinforced (Rigid)
Fully Penetrated (Rigid)
4.0 Partially Penetrating (Rigid)
Unreinforced (Flexible)
2.0 Fully Penetrated (Flexible)
Partially Penetrating (Flexible)
0.0
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0
Nc from laboratory

3. Experimental
3.1. Shear Box Tests
Shear box tests were conducted to determine the friction between the test chamber walls (Perspex or
aluminium) and the soil. Two blocks of Perspex and Aluminium were used to represent the same
material from which the chamber walls were constructed. Before conducting the shear box tests, the
surface roughness of the blocks were measured. Since the aluminium wall of the testing chamber was
hardly dismantled, thus only the roughness of the Perspex panel was determined and will be compared
with the roughness of the Perspex block. The roughness is quantified by the vertical differences of a
real surface from its ideal form (surface) over a measurement length and was determined by using a
profilometer. It was found that the roughness of the Perspex and Aluminium blocks was 0.514 µm and
0.942 µm respectively. The result from the representative Perspex block was smaller than the
roughness obtained from the chamber Perspex panel (0.614 µm) which may be due to wear and tear of
the panel as it had been used for several main tests previously.
A block sample of soil was taken from the model ground using a wire cutter. A 100 mm square
and 20 mm deep cutter was used to prepare the sample for the shear box test by pushing it into the soil
block. A pallet knife was used to trim the sample. The Perspex/aluminium block was attached inside
the shear box equipment as bottom part and the 100 mm square soil block was pushed slowly inside the
shear box container as shown in Figure 3. A 50 kPa normal pressure was briefly applied to ensure the
soil was properly in contact with the block. A sustained different normal stress of 4.2 kPa, 6.2 kPa, 9.1
kPa or 11.1 kPa was then applied. The sample was consolidated and Casagrande’s method was used to
determine the completion of consolidation. The shearing test was conducted using a 1 mm per minute
shearing rate in order to conduct the test in undrained conditions (Head, 1982). The moisture contents
of the samples were determined after the tests were completed.
Estimation of Interface Resistance between Testing
Chamber and Soil Model using Shear Box Test 475

Figure 3: Sketch of shear box test setup.

Dial gauge (vertical

movement)

Soil Sample
Normal Stress
Dial gauge (lateral
Water level
movement) Shear Stress

Block (Perspex/aluminium

Figure 4 plots the measured peak interface shear stresses against the normal stress for both of
the blocks. It can be seen from the results that the peak interface shear stresses are similar for the two
blocks. The average peak interface friction angle for kaolin is 28.6o. It was found that the shear stress
friction between the soil and the block (Perspex or aluminium) was approximately half the applied
normal stress. It was found that, at similar moisture content to that in the test chamber (for reinforced
tests), the normal stress applied on the sample was 6.2 kPa and the shear stress was about half the
undrained shear strength of the reinforced soil for the same moisture content. Therefore the friction
between the clay bed and the chamber sides was assumed to be half the undrained shear strength of the
soil. The results from the shear box test will be used to determine the interface of wall friction during
the loading test.

Figure 4: Relationship between shear stress and normal stress from direct shear tests.
476 Ahmad Safuan A. Rashid and Norhazilan Md. Noor

3.2. Derivation of Surface Resistance

As a shallow failure mechanism was observed in the unreinforced and fully penetrated cases, a simple
collapse mechanism consisting of three triangular wedges, as shown in Figure 5 (a), was used to
estimate the effect of interface resistance. For the partially penetrating case, an eight wedge collapse
mechanism (Figure 5 (b)) was used based on block type failure. The discontinuities of velocity induced
on the boundary surfaces were calculated by using a velocity diagram or hodograph. The friction
between the clay bed and the chamber sides was assumed to be half the undrained shear strength of soil
from the shear box test result.

Figure 5: Collapse mechanism using upper bound analysis.

wf 

(a) Three blocks

wf 

(b) Eight blocks

The work done by the internal stresses on the slip planes for the unreinforced and fully
penetrated cases was calculated based on the hodograph diagram as shown in Figure 5 (a). The
geometry of the hodograph shows that
1
Length ; 01 = 03 = 12 = 23 = B , 02 = B (1)
2
Displacement ; 01 = 03 = 12 = 23 = 2δ w f , 02 = 2 δ w f (2)
Where δwf is displacement and B is width of the footing.
As a result, the upper bound for the collapse load, Fu, is:
Fu=6suBL (3)
Where L is length of the slip plane
From the shear box test, the friction between the Perspex and soil, su,wall, was determined half
from the undrained soil strength, su. Calculation for the increment work,δW, with wall friction is given
as below:
⎧1 B B ⎛1 B B ⎞⎫
Extra δ W = 2 ⎨ × 2δ wf × su , wall + 2 × ⎜ × 2δ wf × su , wall ⎟ ⎬ (4)
⎩2 2 2 ⎝2 2 2 ⎠⎭
Estimation of Interface Resistance between Testing
Chamber and Soil Model using Shear Box Test 477

Fu = 6 Bsu L + B 2 su , wall ( )
2 +1 (5)
By taking L = 1.5B and su,wall=0.5su;
Fu=6.80 BsuL (13% increase due to wall friction) (6)
Hodograph diagram as shown in Figure 5 (b) was used to calculate the work done by the
internal stresses on the slip planes. The geometry of the hodograph shows that
B 1
Length; 0, 1&2 = 03 = 05 = B, 45 = , 04 = 34 = B , 1&2,3 = 2 B (7)
2 2
2δw f
Displacement; 0, 1&2 = 05 = 45 =δwf, 04 = 34 = , 03 = 2δwf,
1&2, 3 = 2δw f (8)
As a result, the upper bound for the collapse load, Fu, is:
Fu=11.16suBL (9)
Calculation for the increment work, δW, with wall friction is given as below:
⎪⎧⎛ δ wf ⎞ ⎛ B2 ⎞ ⎛ δ w f B 2 ⎞ ⎪⎫
Extra δ W = 2 ⎨⎜ B 2 × 2 × ⎟ + ( B × δ wf ) + ⎜ × δ wf × 3⎟ + ⎜ 2 × × ⎟⎬
2

⎪⎩⎝ 2 ⎠ ⎝ 4 ⎠ ⎝ 2 4 ⎠ ⎪⎭
Fu=11.16BSuL+6.20B2su,wall (10)
By taken L = 1.5B and su,wall=0.5su;
Fu=13.23BSuL (18% increase due to wall friction) (11)
A resistance correction was proposed for the Nc values from the laboratory tests, an increase of
13% for the untreated and fully penetration cases and 18% for the partially penetrated case using the
upper bound approach. Figure 2 shows the effect of making the correction. Root Mean Square Error
(RMSE) was used to quantify the agreement between Nc data obtained from LimitState:Geo and the
laboratory results with and without the interface resistance correction (Devore and Peck, 1994). The
calculated RMSE was 1.46 from the uncorrected predictions of the software. A better agreement was
obtained with an RMSE value of 0.52 by comparing the corrected predictions with the laboratory
results.

4. Conclusions
It was assumed that the interface resistance between the clay bed and the chamber sides (Perspex and
aluminium walls) contributed to the differences between the laboratory and LimitState:Geo results. A
correction was introduced using a simple collapse mechanism to estimate the effect of interface
resistance on the laboratory results with an assumed friction between the clay bed and the chamber
sides. A better result obtained after the correction was implemented. As a conclusion, it was found that
the shear box test and upper bound solution provided a better estimation of interface resistance on the
physical modelling test.

Notation
B Width of the footing
cuc Undrained shear strength of the column
cus Undrained shear strength of surrounding soil
Fu Upper bound for the collapse load
H Column height
Kc Relative cohesion ratio of the column to the soft soil
L length of the slip plane
Nc Bearing capacity factor
qult Ultimate bearing capacity
478 Ahmad Safuan A. Rashid and Norhazilan Md. Noor

su Undrained soil strength

su,wall Friction between the Perspex and soil
W Work
wf Displacement

References
[1] Boussida, M. and Porbaha, A. (2004a). Ultimate bearing capacity of soft clays reinforced by a
group of columns-application to a deep mixing technique. Soils and Foundations, Vol. 44, No.
3, 91-101.
[2] Boussida, M. and Porbaha, A. (2004b). Bearing capacity of foundations resting on soft ground
improved by soil cement columns. International Conference on Geotechnical Engineering
(ICGE 2004), 173-180.
[3] Devore, J.L. and Peck, R. (1994). Introductory statistics. Minneapolis/St. Paul: West Pub.
[4] Gilbert, M., Smith, C., Haslam, I. and Pritchard, T. (2010). Application of discontinuity layout
optimisation to geotechnical limit, Proceedings of the 7th European Conference on Numerical
Methods in Geotechnical Engineering.
[5] Head, K.H. (1982). Manual of soil laboratory testing volume 2: permeability, shear strength
and compressibility tests. London: Pentech Press.
[6] Jellali, B., Bouassida, M., and de Buhan P. (2005). A homogenization method for estimating
the bearing capacity of soils reinforced by columns. International Journal for Numerical and
Analytical Method in Geomechanics, Vol. 29, 989-1004.
[7] Omine, K., Ochiai, H., and Bolton, M.D. (1999). Homogenization method for numerical
analysis of improved ground with cement-treated soil columns. Proc. of the Int. Conf. on Dry
Dry Mix Methods for Deep Soil Stabilization, 161-168.
[8] Smith, C. and Gilbert, M. (2007). Application of discontinuity layout optimization to plane
plasticity problems. Proceedings of the Royal Society A, Vol. 463, 2461 -2484.