Page 1 of 23

The strength envelope of granular soil stabilised by multi-axial geogrid in large triaxial tests

Lees, AS *
Senior Application Technology Manager, Tensar International, Nicosia, Cyprus.

Clausen, J
Associate Professor, Aalborg University, Denmark.
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

* corresponding author:

PO Box 14751
Nicosia 2456
Cyprus

Tel. +357 99319106

Fax. +357 22519843
email: ALees@[Link]

Abstract

Conventional methods of characterising the mechanical properties of soil and geogrid separately are
Can. Geotech. J.

not suited to multi-axial stabilising geogrid that depends critically on the interaction between soil
particles and geogrid. This has been overcome by testing the soil and geogrid product together as
one composite material in large-specimen triaxial compression tests and fitting a non-linear failure
envelope to the peak failure states. As such, the performance of stabilising, multi-axial geogrid can
be characterised in a measurable way. The failure envelope was adopted in a linear elastic perfectly-
plastic constitutive model and implemented into finite element analysis, incorporating a linear
variation of enhanced strength with distance from the geogrid plane. This was shown to produce
reasonably accurate simulations of triaxial compression tests of both stabilised and non-stabilised
specimens at all the confining stresses tested with one set of input parameters for the failure
envelope and its variation with distance from the geogrid plane.

Keywords: numerical modelling; constitutive model; granular materials; geogrid; large triaxial test

1
 Page 2 of 23

Notation

k, c, a Non-linear failure envelope parameters

c0, a0 Non-linear failure envelope parameters for non-stabilised soil

ct, at Non-linear failure envelope parameters at geogrid plane

Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

m, b Non-linear plastic potential surface parameters

q Deviatoric stress

Δy Vertical influence extent of geogrid

y Elevation of horizontal geogrid plane

E Young’s modulus

Q Axial load

εa Axial strain

ν Poisson’s ratio
Can. Geotech. J.

ρd Dry mass density

Introduction

Stiff, punched and drawn multi-axial (triangular aperture) polypropylene (PP) geogrid was designed
primarily to restrict the movement of soil particles in and around its apertures – a function defined
as stabilisation (IGS 2018) – thereby improving the performance of granular layers in applications
such as highway pavements and railways. Whereas biaxial geogrid (for similar applications) and
uniaxial geogrid (for retaining walls and slopes) tend to be characterised by their tensile stiffness for
reinforcement and tension membrane functions, multi-axial stabilising geogrid has required more
novel means of characterisation. Its stabilising effect on soil needed to be characterised more than
the mechanical properties of the geogrid itself.

Calculation models used in design conventionally characterise the geogrid in terms of its in-air
tensile stiffness or strength properties, together with the mechanical characteristics of the soil
unaltered by the presence of the geogrid except perhaps for interface properties between the two
materials (e.g. Milligan et al. (1989)). This can lead to significant under-prediction of geogrid
performance, particularly for multi-axial geogrids whose primary function is stabilisation. Some

2
 Page 3 of 23

authors have modified input parameters to address the under-prediction compared with physical
testing. For instance, Hatami and Bathurst (2005) added artificial compaction stresses in the
simulation of mechanically stabilized walls and Perkins and Edens (2002) increased geogrid stiffness
in the simulation of pavements.

Furthermore, there is evidence (e.g. Bussert and Cavanaugh (2010)) that the stabilising effect of
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

geogrid extends a significant distance from the geogrid plane, typically 300 mm or more.

Clearly, geogrid-stabilised granular soil needs to be characterised in a measurable way that can be
incorporated into calculation models for reliable a priori predictions of performance. This paper
describes a strength envelope that has been implemented into a finite element analysis (FEA)
program to produce accurate simulations of triaxial compression tests on a geogrid-stabilised
granular soil.

Micro-scale observations

Recent laboratory tests on a ballast box representing a half section of rail track comprising a ballast
Can. Geotech. J.

layer, two sleepers and a rail under cyclic loading (Liu et al. 2016) compared particle movement with
and without multi-axial geogrid. Two “SmartRock” 3D printed ballast particles containing wireless
sensors recording rotational and translational acceleration on three orthogonal axes were placed
250 mm below top of ballast beneath the rail seat and end of sleeper. A harmonic load cycling
between 0 and 130 kN at 1 Hz was applied to the rail in set-ups with and without a stiff, punched
and drawn multi-axial PP geogrid placed at 250 mm depth in the ballast layer. The rotational
accelerations in one axis recorded in the SmartRock located beneath the end of the sleeper are
compared in the cases without and with geogrid in Figure 1. The geogrid reduced the recorded
rotations of the SmartRock significantly due to interlock between the ballast particles and geogrid.

Liu et al. (2017) later showed a significant reduction in particle rotation and translation of a
SmartRock placed within a granular layer but 100 mm above the multi-axial geogrid, providing
further evidence that the restraint on particle movement is transmitted some distance from the
geogrid.

<SUGGESTED FIGURE 1 LOCATION>

Discrete element analysis (DEM) has been used by some authors (e.g. Butlanska et al. (2014); Huang
et al. (2017)) to demonstrate the profound effect of particle rotational restraint on a soil’s macro-
scale shear strength. Consequently, at macro-scale, or when treating soil as a continuum, an

3
 Page 4 of 23

enhanced shear strength of granular soil would be expected resulting from the installation of a
stabilising multi-axial geogrid appropriate for the soil type.

Macro-scale observations
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

Large triaxial compression tests with vacuum-applied confining stress were performed on an air-
dried, crushed diabase rock non-stabilised and stabilised with multi-axial geogrid placed at mid-
height. The crushed rock had the particle size distribution shown in Figure 2 typical of a road base
material and a maximum dry density ρd,max of 2.26 Mg/m3 (at an optimum moisture content of 7.2%)
determined using the 4.5 kg rammer according to Clause 3.6 of BS 1377 (BSI 2002). The stabilising
geogrid was composed of punched and drawn polypropylene with ribs in three directions at 40 mm
pitch and a radial stiffness at 0.5% strain of 360 kN/m.

<SUGGESTED FIGURE 2 LOCATION>

Can. Geotech. J.

The triaxial specimens had nominal dimensions of 1.0 m height and 0.5 m diameter as shown in
Figure 3, giving a ratio of diameter to maximum particle size of 10 to reduce specimen size effects. A
3 mm thick latex rubber sheet was found to be sufficiently robust to act as the specimen membrane
at these relatively low confining stresses for a number of tests without being punctured by the
angular particles. Additional protection was necessary during specimen compaction as described
below. An air-tight connection to the top cap and base disc was formed using large hose clamps.
Specimens were prepared in a steel split compaction mould with the rubber membrane clamped to
the base disc and lightly stretched and clamped temporarily to the top of the mould. Compaction to
at least 95% ρd,max was achieved in 200 mm lifts each of which was compacted by blows of a mallet
against the outside of the mould at the approximate level of the new layer 7 times at 10 equally-
spaced locations around the mould. In spite of being air-dried, a high specimen density was achieved
due to the substantial self-weight of the large specimen. Additional sheets of the rubber membrane
were fitted to the inside of the mould at the blow locations to protect the main membrane from
puncture by the angular particles during compaction. Where geogrid was included in the specimen,
half of the mid-height lift was placed uncompacted, followed by the geogrid cut to a disc of slightly
smaller diameter to avoid the ribs puncturing the membrane, followed by the remaining mid-height
lift and the entire lift with geogrid then compacted in the same way.

4
 Page 5 of 23

<SUGGESTED FIGURE 3 LOCATION>

On completion of specimen preparation, the top cap was fitted and the confining stress vacuum
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

established prior to removing the mould to avoid collapse. The specimen was then lifted briefly just
clear of the base with the hydraulic cylinder to weigh the specimen with the load cell.

Axial displacement was measured at the top cap by means of a potentiometer mounted
independently of the triaxial apparatus in order not to be affected by any compliance strains. The
circumference of the specimen at mid-height was measured by means of a draw-wire transducer
wrapped around the specimen. A data acquisition system recorded time, vacuum pressure, axial
load, axial displacement and circumferential displacement throughout the test. Once the vacuum
pressure had reached equilibrium the axial load was increased to failure or else to an axial strain of
around 10%. An unload/reload loop was included prior to failure to obtain the elastic stiffness. The
hydraulic pressure was controlled to achieve an axial strain rate of 1.5 mm/s, although results were
found not to be sensitive to loading rate as also reported from similar tests on coarse granular
Can. Geotech. J.

materials (e.g. Qian et al. (2018)).

Tests on non-stabilised and stabilised specimens (denoted N and S respectively) were performed at
different confining stresses as shown in Table 1. The deviatoric stress q was calculated from the
recorded axial force taking into account the increased circumference of the specimen during
compression as measured by the draw-wire transducer and the self-weight of the upper half of the
specimen. Membrane corrections were not included since these were not significant in specimens of
this size and strength.

<SUGGESTED TABLE 1 LOCATION>

Plots of averaged deviatoric stress q against averaged axial strain εa at the different confining
stresses for non-stabilised and stabilised specimens are shown in Figures 4 and 5 respectively. An
enhanced peak shear strength in the geogrid-stabilised soil was obtained at all the confining
stresses. The higher confining stress stabilised cases exhibited a drop in strength above about 8%
axial strain due to tearing of the geogrid under high shear stresses across the geogrid plane. The
major and minor principal effective stresses at peak failure were plotted in Figure 6, forming a

5
 Page 6 of 23

markedly non-linear failure envelope at the stress levels tested, particularly in the stabilised case.
While slight non-linearity of the failure envelope was obtained in the non-stabilised soil due to the
effects of dilatancy in the compacted, coarse aggregate, this was significantly more marked in the
stabilised case. This was likely to be caused by the restraint on particle translation and rotation, as
observed in the SmartRock testing, significantly increasing the work done required to shear and
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

dilate the specimen. The enhanced strength may dissipate at confining stresses higher than those
tested here but multi-axial geogrid is typically deployed at shallow depths where confining stresses
are well within this test range.

<SUGGESTED FIGURE 4 LOCATION>

<SUGGESTED FIGURE 5 LOCATION>

<SUGGESTED FIGURE 6 LOCATION>

Can. Geotech. J.

Non-linear failure envelope

A good fit to the peak failure points in principal stress space was achieved using Equation 1
(compressive stresses taken as positive), using the values shown in Figure 6 for the stabilised and
non-stabilised cases. Residuals were apparently randomly distributed and were up to 10 and 23 kPa
in the non-stabilised and stabilised cases respectively. Such stochastic errors were considered most
likely due to variation in density and grading within and between specimens and were minimised as
far as reasonably possible during specimen preparation.

( (
𝜎1 = 𝑘𝜎3 +𝑐 1 ― exp ―𝑎 𝑐
𝜎3
)) Equation 1

The stabilised case failure envelope plotted in Figure 6 represents an averaged strength for the
triaxial test specimen but the restraint on the soil particles would be at a maximum at the geogrid
plane and dissipate with decreasing proximity. Therefore, the failure envelope was considered to
vary (assumed linearly) from a maximum at the geogrid plane to the non-stabilised failure envelope
at a perpendicular distance Δy (assuming a horizontal geogrid plane which in most cases is true),
beyond which the non-stabilised failure envelope prevailed, as illustrated in Figure 7. A linear
variation with proximity to the geogrid could also be adopted with perpendicular distance from a

6
 Page 7 of 23

non-horizontal geogrid plane but with slighter higher computational effort. The non-stabilised failure
envelope can be obtained straightforwardly from shear strength tests on the granular material
without geogrid thereby obtaining the parameters k, c0 and a0, which should be validated by test
simulation to take account of top and bottom boundary conditions. That leaves three input
parameters (ct, at and Δy) that can be determined by further simulation of triaxial compression tests
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

on the same material but stabilised under at least three different confining stresses.

In this way the failure envelope is determined by testing the geogrid and soil as a composite material
which is more appropriate for a stabilising geogrid than trying to relate soil mechanical properties
and characteristics (e.g. particle size distribution, angularity) to geogrid mechanical and geometrical
properties (e.g. aperture size, rib dimensions) when the soil-geogrid interactions are so complex.

<SUGGESTED FIGURE 7 LOCATION>

Can. Geotech. J.

Implementation into FEA

A linear elastic perfectly-plastic (LEPP) constitutive model with the non-linear failure criterion
(Equation 1) was implemented into the Plaxis 2D (Brinkgreve et al. 2018) FEA software, together
with the linear variation with distance from the geogrid plane described in Figure 7. The plastic
potential was defined in a similar way as shown in Equation 2 but did not vary with distance from
the geogrid. m and b are generally taken as equal to k and a giving associated plasticity in non-
stabilised soil but increasing non-associativity in stabilised soil.

( (
𝜎1 = 𝑚𝜎3 + 𝑐0 1 ― exp ―𝑏 𝑐0 ))
𝜎3
Equation 2

The triaxial compression tests described earlier in this paper were simulated in a two-dimensional
axisymmetric FEA model with rough specimen ends. The input parameters (ct, at and Δy) were
determined by trial and error during simulation of the triaxial compression tests until the best fit
with the laboratory data was achieved at all five confining stresses with one set of strength
parameters. A Δy value of 0.35 m was obtained which is consistent with other studies on the
influence extent of geogrid in granular soils (e.g. Bussert and Cavanaugh (2010)). The E values were
also obtained by trial and error to achieve the best fit with pre-failure deformation at each confining

7
 Page 8 of 23

stress level in the non-stabilised specimens. The obtained values were low due to the low confining
stress in the triaxial tests and did not need to be changed for the stabilised specimens.

The pre-failure deformation prediction shown in Figures 8 and 9 was rather approximate due to the
linear elasticity in the constitutive model but the failure load was predicted reasonably accurately at
all confining stresses as also shown in Table 2, using one set of parameters for the non-linear failure
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

envelope in both the stabilised and non-stabilised cases as shown in Table 3. Only the S70 and S40
predictions were slightly low and high respectively which was considered due to variability in each
specimen as also reflected in the failure states of the same tests plotted relative to the failure
envelope using averaged parameters in Figure 6. Only the y value was changed for the non-stabilised
case such that the geogrid plane was positioned well above the mesh and a uniform non-stabilised
strength was adopted in all elements.

<SUGGESTED FIGURE 8 LOCATION>

<SUGGESTED FIGURE 9 LOCATION>

Can. Geotech. J.

<SUGGESTED TABLE 2 LOCATION>

<SUGGESTED TABLE 3 LOCATION>

Conclusion

This paper has used micro-scale observations to demonstrate how multi-axial geogrid restrains
aggregate particles against rotation and translation which manifests itself at the macro-scale as an
enhanced shear strength as demonstrated by large scale triaxial testing.

A non-linear failure envelope was found to provide a good fit to peak failure states plotted in
principal effective stress space for a crushed rock tested in triaxial compression in both the non-
stabilised case and when stabilised with a single layer of multi-axial geogrid placed at mid-height.
When implemented into a LEPP constitutive model in FEA such that the failure criterion varied
linearly with distance from the geogrid plane it was found to produce accurate predictions of triaxial
failure load at all the confining stresses tested with a single set of parameters for the failure
envelope and its variation with distance from the geogrid.

8
 Page 9 of 23

Testing the soil and geogrid together in this way as a composite provided a reliable means of
characterising the mechanical properties of this soil stabilised by a multi-axial geogrid. Further tests
should be performed to verify this approach for other aggregates and multi-axial geogrid types,
perhaps with additional layers of geogrid. The link between the observed micro-scale and macro-
scale behaviour could be verified by including Smartrock devices within the triaxial specimen.
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

References

BSI. 2002. BS 1377-4:1990 Methods of test for soils for civil engineering purposes – Part 4:
Compaction-related tests. British Standards Institution, London, UK.

Brinkgreve, R.B.J., Kumarswamy, S., Swolfs, W.M. and Foria, F. 2018. Plaxis 2018. Plaxis bv, Delft, The
Netherlands.

Bussert, F., and Cavanaugh, J. 2010. Recent research and future implications of the actual behaviour
of geogrids in reinforced soil. ASCE Earth Retention Conference (ER2010), 1-4 August, Bellevue,
Washington, pp. 460-477.
Can. Geotech. J.

Butlanska, J., Arroyo, M., Gens A., and O’Sullivan, C. 2014. Multi-scale analysis of cone penetration
test (CPT) in a virtual calibration chamber. Canadian Geotechnical Journal 51(1): 51-66.

Hatami, K., and Bathurst, R.J. 2005. Development and verification of a numerical model for the
analysis of geosynthetic-reinforced soil segmental walls under working stress conditions. Canadian
Geotechnical Journal 42(4): 1066-1085.

Huang, X., Hanley, K.J., O’Sullivan, C., and Kwok, C.Y. 2017. Implementation of rotational resistance
models: a critical appraisal. Particuology 34: 14-23.

IGS. 2018. Guide to the Specification of Geosynthetics, 5th edition. International Geosynthetics
Society, Jupiter FL.

Liu, S., Huang, H., Qiu, T., and Kwon, J. 2016. Effect of geogrid on railroad ballast particle movement.
Transportation Geotechnics 9: 110-122.

Liu, S., Huang, H., and Qiu, T. 2017. Behavior of geogrid-reinforced railroad ballast particles under
different loading configurations during initial compaction phase. Proceedings of the 2017 Joint Rail
Conference, April 4-7, Philadelphia, PA.

Milligan, G.W.E., Jewell, R.A., Houlsby, G.T., and Burd, H.J. 1989. A new approach to the design of
unpaved roads – Part I. Ground Engineering 22 (April): 25-29.

9
 Page 10 of 23

Perkins, S.W., and Edens, M.Q. 2002. Finite element and distress models for geosynthetic-reinforced
pavements. The International Journal of Pavement Engineering 3(4): 239-250.

Qian, Y., Tutumluer, E., Mishra, D., and Kazmee, H. 2018. Triaxial testing and discrete-element
modelling of geogrid-stabilised rail ballast. Proceedings of the Institution of Civil Engineers – Ground
Improvement 171(4): 223-231.
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.
Can. Geotech. J.

10
 Page 11 of 23

Figure Captions

Figure 1. Recorded “SmartRock” rotational acceleration at 250 mm depth in ballast layer beneath
end of sleeper in laboratory test (Liu et al. 2016).

Figure 2: Particle size distribution of aggregate used in triaxial compression tests.

Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

Figure 3: Large scale triaxial test apparatus.

Figure 4: q v. εa plot from large scale triaxial compression tests on the non-stabilised triaxial test
specimens.

Figure 5: q v. εa plot from large scale triaxial compression tests on the stabilised triaxial test
specimens.

Figure 6: Peak failure states plotted in terms of principal effective stresses.

Figure 7: Linear variation of failure envelope with proximity to geogrid plane.

Figure 8: Comparison of triaxial test simulation outputs with laboratory test data (non-stabilised
cases).
Can. Geotech. J.

Figure 9: Comparison of triaxial test simulation outputs with laboratory test data (stabilised cases).

11
 Page 12 of 23

Table 1: Large scale triaxial test conditions and results summary.

Test no. ρ (Mg/m3) % ρd,max† Confining Deviatoric

stress σ3 (kPa) stress at failure
qf (kPa)
N10 2.170 96.0 11.0 72.0
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

N25 2.245 99.3 24.8 160.3

N40 2.154 95.3 39.0 230.7
N50 2.160 95.5 50.6 307.3
N70 2.193 97.0 69.8 412.0
S10 2.146 95.0 10.4 107.5
S25 2.158 95.4 25.6 220.3
S40 2.154 95.3 38.0 287.4
S50 2.171 96.0 51.2 391.0
S70 2.178 96.3 72.0 520.2
† Assuming dry (aggregate was air-dried)
Can. Geotech. J.
 Page 13 of 23

Table 2: Comparison between measured and predicted failure loads in large scale triaxial tests.

Test no. Q (kN) Q (kN) Error

measured predicted
N10 12.6 13.6 +1.0
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

N25 31.2 31.0 -0.2

N40 45.6 45.8 +0.2
N50 61.7 60.4 -1.3
N70 82.9 80.9 -2.0
S10 19.8 22.0 +2.2
S25 44.5 43.5 -1.0
S40 57.6 60.3 +2.7
S50 78.2 75.6 -2.6
S70 106.0 96.2 -9.8
Can. Geotech. J.
 Page 14 of 23

Table 3: Input parameters to FEA triaxial test simulation.

Parameter Value
k 6.1
c0 56 kPa
a0 2.0
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

m 6.1
b 2.0
Δy 0.35m
y +0.5m
ct 280 kPa
at 10
E 6.0, 8.0, 10.5,
13.0, 18.6 MPa*
ν 0.2
γ 21.5 kN/m3
* E adjusted for the five confining stress levels.
Can. Geotech. J.
 Page 15 of 23

40
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

Particle angular acceleration (rad/s2)

without geogrid
30

20 with geogrid

10

0
0 100 200 300 400 500 600
-10
Time (s)
-20

-30

-40

Figure 1. Recorded “SmartRock” rotational acceleration at 250 mm depth in ballast layer beneath
Can. Geotech. J.

end of sleeper in laboratory test (Liu et al, 2016).

 Can. Geotech. J.
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

Cumulative % passing

100

20
40
60
80

0
0.01
0.1
1
Particle size (mm)
10
100

Figure 2: Particle size distribution of aggregate used in triaxial compression tests.

Page 16 of 23
 Page 17 of 23

Reaction beam
Hydraulic cylinder
Displacement transducer
To vacuum gauge To vacuum pump
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

Load
cell

Rubber membrane Steel top cap

Aggregate (dry) 0.5 m

Geogrid
Draw-wire
transducer

0.5 m

Steel base disc

Rigid base
Can. Geotech. J.

0.5 m dia.

Figure 3: Large scale triaxial test apparatus.

 Page 18 of 23

500

400
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

N10
q (kPa)

300 N25

200 N40
N50
100 N70

0
0.00 0.02 0.04 0.06 0.08 0.10
a

Figure 4: q v. a plot from large scale triaxial compression tests on the non-stabilised triaxial test
specimens.
Can. Geotech. J.
 Page 19 of 23

500

400
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

S10
q (kPa)

300 S25
S40
200
S50
100 S70

0
0.00 0.02 0.04 0.06 0.08 0.10 0.12
a

Figure 5: q v. a plot from large scale triaxial compression tests on the stabilised triaxial test
specimens.
Can. Geotech. J.
 Page 20 of 23

800
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

k = 6.1
600 c = 150
a = 6.0
σ 1 (kPa)

400

k = 6.1
c = 56 non-stabilised
200 a = 2.0

stabilised

0
0 20 40 60 80 100
σ 3 (kPa)
Can. Geotech. J.

Figure 6: Peak failure states plotted in terms of principal effective stresses.

 Page 21 of 23

σ1
Linear interpolation of failure surface At geogrid elevation y
between geogrid plane and vertical
influence extent Δy
Outside influence extent
k
of geogrid
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

Curvature at
ct

Curvature a0
c0

σ3

Figure 7: Linear variation of failure envelope with proximity to geogrid plane.

Can. Geotech. J.
 Page 22 of 23

80
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

N10
60
N25
Q (kN)

40 N40
N50
20 N70
FEA
0
0.00 0.02 0.04 0.06 0.08 0.10
a

Figure 8: Comparison of triaxial test simulation outputs with laboratory test data (non-stabilised
cases).
Can. Geotech. J.
 Page 23 of 23

120

100
Downloaded from [Link] by UNIVERSITY OF OTAGO LIBRARY on 04/22/19. For personal use only.

S10
80
S25
Q (kN)

60 S40
40 S50
S70
20
FEA
0
0.00 0.02 0.04 0.06 0.08
a

Figure 9: Comparison of triaxial test simulation outputs with laboratory test data (stabilised
cases).
Can. Geotech. J.