Contents

NUMERICAL MODELLING OF GRS ............................................................................................. 2

I. Proposed numerical model ........................................................................................................... 2
1.1.1. General ............................................................................................................................... 2
1.1.2. Soil model and input parameters ..................................................................................... 2
1.1.2.1. Back fill material......................................................................................................... 2
1.1.2.2. Facing ......................................................................................................................... 3
1.1.2.3. Geogrid ....................................................................................................................... 4
1.1.3. Geometry, boundary conditions and interfaces ............................................................. 4
1.1.3.1. Geometry ..................................................................................................................... 4
1.1.3.2. Boundary conditions ................................................................................................... 5
1.1.4. Mesh, construction phase and Tolerated error .............................................................. 7
1.1.4.1. Mesh generation.......................................................................................................... 7
1.1.4.2. Construction phase ..................................................................................................... 8
1.1.4.3. Tolerated error............................................................................................................ 8
1.1.5. Validation Results ............................................................................................................. 9
1.1.5.1. Horizontal facing displacement .................................................................................. 9
1.1.5.2. Reinforcement axial strain ........................................................................................ 10
1.1.6. Summary of model validation ........................................................................................ 11

Figure I-1- Stress-strain relation in primary loading for a standard drained triaxial test ....................... 3
Figure I-2- Validation model geometry used in Plaxis ............................................................................ 6
Figure I-3 Position of nodes and stress points in soil elements .............................................................. 7
Figure I-4 Comparison of facing displacement from Plaxis 2D analysis with predicted values reported
by Hatami and Bathurst (2005) ............................................................................................................... 9
Figure I-5 Comparison of axial strain from Plaxis 2D (Left) analysis with measured and predicted
values reported by Hatami and Bathurst (2005) (Right)....................................................................... 10

Table I-1 Facing block interface properties used in Plaxis 2D validation model..................................... 6
Table I-2Coarseness factor re ................................................................................................................. 8
NUMERICAL MODELLING OF GRS

Numerical modelling has been successfully used to investigate a range of parameters that
influence the behaviour of GRS (Hatami and Bathurst, 2005). The finite model program
Plaxis 2D was used to develop a validation numerical model compared to predictive model
using FLAC, where it replicates the construction and instrumentation of three full-scale GRS
retaining walls within a plane strain test facility at the Canadian Royal Military College
(RMC). In the current model, the backfill and the facing blocks were modelled using soil
rectangles and polygons with different criteria. The reinforcement layer was modelled with
structural elements (geogrids). The model was constructed in two stages, the first is the initial
stage and the second is the construction stage, to compare the facing displacement and the %
strain in the geogrids with the results predicted by(Hatami and Bathurst, 2005) in there finite
element model.

I. Proposed numerical model

1.1.1. General

Plaxis 2D is a finite element program that has been used to perform displacement and
stability for various geotechnical structures such as tunnel embankments and retaining walls,
in plane strain or axisymmetric conditions. The program features constitutive models for soil
and geogrid, and allows staged construction, where clusters of finite elements are activated or
deactivated to simulate a particular construction sequence.

1.1.2. Soil model and input parameters

1.1.2.1. Back fill material

The soil used in this study was modelled using a pre-programmed model called the hardening
soil model, that stimulate the behaviour of various types of soil (Schanz, Vermeer and
Bonnier, 1999).

The hardening soil is an elasto-plastic model that feature a relation between axial strain and
deviatoric stress in primary triaxial loading, which is approximated by a hyperbola figure (I-
1). Moreover, a basic feature of the hardening soil model is the stress dependency of soil
stiffness.
The basic characteristics of the model are:

m: stress dependent stiffness according to the power law

: Plastic straining due to primary deviatoric loading

: Plastic straining due to primary compression

: Elastic unloading/reloading

c, φ, : cohesion, friction, dilatancy (respectively)

Figure ‎I-1- Stress-strain relation in primary loading for a standard drained triaxial test

1.1.2.2.Facing

The facing blocks, were modelled as soil rectangles using a linear elastic model that
represents Hooke’s law of isotropic linear elasticity. This model is too limited for the
stimulation of soil behaviour and used for stiff structures in the soil. The model involves two
elastic stiffness parameters:

E’: effective Young’s modulus

’: Effective Poisson’s ratio

 1.1.2.3.Geogrid

Geosynthetic reinforcement should be modelled as a discrete line elements with two

translational degree of freedom as suggested by Plaxis 2D. The geogrids were modelled
using a planar element with perfect elastic-plastic linear stiffness, defined by two parameters:

EA: axial stiffness

Np: the maximum force in plane

Based on the theory proposed by (Bathurst, Allen and Walters, 2002), the axial strain across
the geogrids will be calculated using the axial force at each point, thus :

𝐽𝑐= 𝑇𝑖𝜀𝑖 Equation ‎I- 1

When re-arranged gives:

𝜀𝑖= 𝑇𝑖𝐽𝑐 Equation ‎I- 2

Where:
Jc = Axial stiffness, EA (kN/m)
i = a selected data point
Ti = Axial force (kN/m)
εi = Strain (decimal)

1.1.3. Geometry, boundary conditions and interfaces

1.1.3.1.Geometry

The model was replicated to the model used by (Hatami and Bathurst, 2005), creating a 3.6m
height segmental wall with a face batter of 8 from the vertical, which retained a 5.8 m
backfill soil. The concrete facing blocks were 0.15 m in height and 0.3 m thick. The blocks
were arranged in 24 inclined blocks supported by a steel plate. An interface was created
between the blocks, and the blocks and soil (section 1.1.3.3.). The blocks properties using the
linear–elastic model are: Unit weight =16.8 kN/m3 Young’s modulus = 100 Mpa, Poisson’s
ratio=0.15

The hardening soil model were used to create the backfill soil which is suitable for stress
dependent elastic modulus. The soil was assigned unit weight=22 kN/m3, Young’s modulus=
50 Mpa, peak angle friction=44 , dilation angle=11 , Poisson’s ratio=0.15 and the soil
cohesion was set c=1kpa.The reason of the small value of cohesion was to prevent premature
soil yielding in locally confining pressure zones, and to take account for any possible
apparent cohesion due to moisture in the soil, that’s has been previously reported by
(Cazzuffi et al., 1993) and (Kerry Rowe and Skinner, 2001).

The geogrids used in the wall are biaxial PP, which are manufactured from polypropylene
sheets using extrusions method followed by stretching is both directions until it reaches the
tensile characteristic, the geogrids has rate dependant properties, when tensile loads vary
none linearly with time and strain.

1.1.3.2.Boundary conditions

The boundary conditions used in Plaxis 2D were similar to those used in FLAC by Hatami
and Bathurst (2005). A fixed boundary conditions was modelled directly below the fill
material base using a vertical displacement line with (Ux=Uy=0) to assure full fixity. The
vertical line displacement of the fill the boundary conditions were taken as hinged, were it
allows the displacement in y direction but blocked in the x direction, to allow the fill to
deform, (Ux=0).

However, the boundary conditions under the facing block wall, where a stiff steel plate was
induced in the model with free boundary condition in the x direction, and blocked in y
direction to allow the facing block to slide in the y direction for more accurate displacement
value, in addition an end anchor toe was placed to support the steel plate in x direction. The
axial stiffness of this toe anchor was taken KToe = 4 MN/m based on measures loads and
displacements reported by Hatami and Bathurst (2005) and Huang, Bathurst and Hatami
(2009).(Figure I-2)
 300

150
Facing Blocks

Detail-A-

3600
Back Fill

Batter angle=8° Geogrids

Fixed end anchor

6000

Figure ‎I-2- Validation model geometry used in Plaxis

Interfaces are joint element added to the model to allow interaction of proper modelling of
two different materials such as soil-block and soil-geogrids, or for the same material such as
block-block. The stiffness matrix for interfaces elements was obtained by means of Newton
Cotes integration.

Since the geogrids elements depends on the friction between the geogrid and the surface soil
and due to the shape of the geogrids where there are longitudinal and transversal openings,
the interfaces between soil-geogrids should be modelled with a reduction in interface strength
Rinter=0.95, according to(Gouw, 2018) . The interfaces of soil-block and block-block, were
modelled using Mohr-Coulomb strength specified by Hatami and Bathurst (2005).

Table -I-1- presents the properties used to simulate each interface element within Plaxis 2D.

Table ‎I-1 Facing block interface properties used in Plaxis 2D validation model

Interface Parameter Value

Soil-Block E, (kN/m2) 50x103

c, (kN/m2) 0
φ, (◦) 44
, (◦) 11
kn, (kN/m3) 100x103
ks, (kN/m3) 1x103
Block-Block E, (kN/m2) 100x103

c, (kN/m2) 46
φ, (◦) 57
, (◦) 0
kn, (kN/m3) 1x106
ks, (kN/m3) 40x103

1.1.4. Mesh, construction phase and Tolerated error

1.1.4.1. Mesh generation

Plaxis 2D, provides 2 types of elements 6 nodes and 15-nodes triangular elements figure(I-3)
to model a soil layer. The 15-nodes element was chosen during the validating model in this
study, it provides a 4th order interpolation for displacement and it’s very accurate element that
has produced high quality stress results for various problem (Nagtegaal, Parks and Rice,
1974).

Figure ‎I-3 Position of nodes and stress points in soil elements

The model geometry was automatically divided into finite elements by Plaxis 2D to perform
the analysis. The mesh generator required a global meshing parameter that represented the
target element size, Ie. This parameter was calculated by Plaxis 2D depending on the model
geometry and a coarseness factor, re. The calculation was performed according to the
following equation (Plaxis 2D):

𝐼𝑒 = 𝑟𝑒 × 0.06 × √(𝑋𝑚𝑎𝑥 − 𝑋𝑚𝑖𝑛)2 + (𝑌𝑚𝑎𝑥 − 𝑌𝑚𝑖𝑛)2 Equation I- 3

Where Xmax, Xmin, Ymax and Ymin are the outer dimensions of the model geometry and re is
given by the Table (I-2).

Table ‎I-2 Coarseness factor re

Very Coarse re = 2.00 30-70 elements

Coarse re = 1.33 50-200 elements
Medium re = 1.00 90-350 elements
Fine re = 0.67 250-700 elements
Very Fine re = 0.50 500-1250 elements

Plaxis 2D outlines that the selection of re depends on the shape of the geometry and the
selection of an optional local refinement setting which locally increases mesh coarseness
around points of interest in the geometry. In this study, a very coarse mesh was selected with
local refinement around the geogrid layers at the front face of the wall. This analysis also
produced a refined mesh along the face of the wall which provided a high-quality
displacement field to assess horizontal movement of the wall face whilst retaining model
computational time at a manageable level.

1.1.4.2.Construction phase

Normally the calculation list contains the initial phase that represents the initial situation of
the project and define the starting point of further calculation.

Therefore, another phase was added, this phase was not time dependent and take into
consideration the construction phase of the whole model soil, blocks geogrids and interfaces.

1.1.4.3. Tolerated error

When conducting a non-linear analysis where a finite element of calculation steps are used,
there will be some drift from the exact solution (Plaxis 2D). The purpose of the solution
algorithm used within Plaxis 2D is to ensure that the equilibrium errors remain within the
acceptable bounds (Plaxis 2D). This error limit is linked to a specified value of tolerated
error. During the calculation process, Plaxis 2D continues to undertake iterations until the
calculated errors are smaller than the specified tolerated error value (Plaxis 2D). As a default,
Plaxis 2D suggests 0.01 as a standard error setting and specifies that this may be increased to
0.03 or 0.05 for failure load calculations.
In this study the tolerated error was taken as 0.03 in all cases to ensure that deformations of
the wall face were not underestimated and that the computational time for the model was
retained at a manageable level.

1.1.5. Validation Results

1.1.5.1. Horizontal facing displacement

Figure (I-4) presents a comparison of the predicted facing displacement from Plaxis 2D with
predicted values reported by Hatami and Bathurst (2005)

4.0

3.5

3.0

2.5
Elevatio [m]

2.0

1.5

1.0

0.5

0.0
0 2 4 6 8 10
Relative facing displacment [mm
Plaxis 2D Hatami and Bathrust model

Figure ‎I-4 Comparison of facing displacement from Plaxis 2D analysis with predicted values reported by Hatami
and Bathurst (2005)

The predicted displacement value of the horizontal facing from Plaxis 2D shows a good
match with the predicted value reported by Hatami and Bathurst (2005), figure (I-4). The
displacement start at the bottom of the wall and increased to a peak value that has occurred
at the mid-height of the wall, and start to decrease until the top of the wall. The predicted
displacement of Plaxis 2D showed a fair correlation with the predicted value reported by
Hatami and Bathurst (2005), figure (I-4).

1.1.5.2. Reinforcement axial strain

Figure (I-5) presents a comparison of the predicted axial strain within the reinforcement from
Plaxis 2D with the measured and predicted values reported by Hatami and Bathurst (2005).

0.40
Layer -6-
Strain
[%]

0.20
0.00

0.80
Layer -5-
Strain
[%]

0.40
0.00

0.80
Layer -4-
Strain
[%]

0.40
0.00

1.50
Strain

1.00 Layer -3-

[%]

0.50
0.00

1.00
Layer -2-
Strain
[%]

0.50
0.00

0.80
Strain [%]

0.40
Layer -1-
0.00
0.50 1.00 1.50 2.00
Distance from back of facing [m]

Figure ‎I-5 Comparison of axial strain from Plaxis 2D (Left) analysis with measured and predicted values reported by Hatami
and Bathurst (2005) (Right)

The predicted magnitude of axial reinforcement strain from Plaxis 2D shows a good match
with the predicted value reported by Hatami and Bathurst (2005), figure (I-5). The axial
strain was found to reach it maximium value behind the facing wall and start to decrease to
zero at the end of the reinforcement. The values of the axial reinforecement starin showed fair
correlation with the values reported by Hatami and Bathurst (2005), figure (I-5).

1.1.6. Summary of model validation

Based on the agreements shown between the predicted horizontal facing displacement and
axial reinforcement strain from Plaxis 2D and that reported by Hatami and Bathurst (2005),
Figure (I-4) and Figure (I-5), it was considered that the finite element model was effectively
validated.

Allen, T. M. and Bathurst, R. J. (2014) ‘Design and performance of 6.3-m-high, block-faced geogrid
wall designed using k-stiffness method’, Journal of Geotechnical and Geoenvironmental
Engineering, 140(2), pp. 1–12. doi: 10.1061/(ASCE)GT.1943-5606.0001013.

Bathurst, R. J., Allen, T. M. and Walters, D. L. (2002) ‘Short-term strain and deformation behavior of
geosynthetic walls at working stress conditions’, Geosynthetics International. doi:
10.1680/gein.9.0225.

Cazzuffi, D. et al. (1993) ‘Laboratory Investigations on the Shear Strength of Geogrid Reinforced
Soils’, in Cheng, S.-C. (ed.) Geosynthetic Soil Reinforcement Testing Procedures. West
Conshohocken, PA: ASTM International, pp. 119–137. doi: 10.1520/STP24317S.

Gouw, T. 2018. (2018) ‘International conference’, Journal of the Society of Archivists, 7(5), pp. 341–
342. doi: 10.1080/00379818409514249.

Hatami, K. and Bathurst, R. J. (2005) ‘Development and verification of a numerical model for the
analysis of geosynthethic-reinforced soil segmental walls under working stress conditions’, Canadian
Geotechnical Journal. doi: 10.1139/t05-040.

Huang, B., Bathurst, R. J. and Hatami, K. (2009) ‘Numerical study of reinforced soil segmental walls
using three different constitutive soil models’, Journal of Geotechnical and Geoenvironmental
Engineering, 135(10), pp. 1486–1498. doi: 10.1061/(ASCE)GT.1943-5606.0000092.

Kerry Rowe, R. and Skinner, G. D. (2001) ‘Numerical analysis of geosynthetic reinforced retaining
wall constructed on a layered soil foundation’, Geotextiles and Geomembranes, 19(7), pp. 387–412.
doi: 10.1016/S0266-1144(01)00014-0.
Nagtegaal, J. C., Parks, D. M. and Rice, J. R. (1974) ‘On numerically accurate finite element solutions
in the fully plastic range’, Computer Methods in Applied Mechanics and Engineering, 4(2), pp. 153–
177. doi: 10.1016/0045-7825(74)90032-2.

Schanz, T., Vermeer, P. A. and Bonnier, P. G. (1999) ‘The hardening soil model: Formulation and
verification’, in Beyond 2000 in computational geotechnics. Ten Years of PLAXIS International.
Proceedings of the international symposium, Amsterdam, March 1999. doi: 10.1201/9781315138206-
27.