Research Article

Case Study of Soil-Geogrid Interaction Simulating Large-Scale Pullout Tests with Numerical Approach

Boon Tiong Chua¹* and Kali Prasad Nepal²

¹Kellogg Brown & Root Pty Ltd, Australia, Level 21/30 Pirie Street, Adelaide, SA 5000, Australia

²School of Engineering, Central Queensland University, 120 Spencer St, Melbourne Victoria 3000, Australia

Submission: October 24, 2024; Published: October 24, 2024

*Corresponding Author: Boon Chua, Kellogg Brown & Root Pty Ltd, Australia, Level 21/30 Pirie Street, Adelaide, SA 5000, Australia Civil Eng

How to cite this article: Boon Tiong Chua* and Kali Prasad Nepal. Case Study of Soil-Geogrid Interaction Simulating Large-Scale Pullout Tests with Numerical Approach. Civil Eng Res J. 2024; 15(2): 555904. DOI 10.19080/CERJ.2024.15.555904

Abstract

Soil-geogrid interaction plays a critical role in the design of reinforced-soil structures and can make a difference between success and failure of geogrid application. Depending on the types and properties of both geogrid and soil, the soil- geogrid interaction mechanism can be very complex and requires in-depth studies. The soil-geogrid interaction is most commonly studied using expensive large-scale experimental pullout tests. In this study, two-dimensional (2D) finite element analysis is conducted using a commercially available program- ABAQUS, to simulate these tests. The model can capture the three-dimensional (3D) aspect of the geogrid considering its equivalent 2D physical geometry. The model is verified using the data from published literature and is found to provide reasonable predictions of load-displacement pullout behaviour. The model is then used to study the influence of transverse members, confined pressure, soil density, fines content, load- transfer mechanism and coefficient and angle of soil friction. The results reveal that these parameters have a significant influence on pullout load-displacement performance. For example, displacement decreases with an increase in confined pressure, anchorage length required is around 400 mm for confined pressure equal or greater than 75 kPa and only short anchorage lengths is required at higher overburden pressures/confined pressures.

Keywords: Geogrid, Pullout test; Interaction coefficient; Finite element model (FEM)

Introduction

The interaction between soil and geogrid is the most fundamental design parameter in the design of reinforced-soil structures to ensure their sound performance in service. This interaction can be very complex as it is affected by a number of factors including, (i) physical and mechanical properties of soil (particle size distribution, shear strength, grain shape, degree of compaction and density), (ii) geogrid reinforcement (type and production method: glued or extruded geogrid, geometric characteristics: rib dimensions, junction type, aperture size and, mechanical properties: tensile strength, tensile modulus, radial secant stiffness) (iv) confining stresses and, (iv) loading conditions and more [1-3]. The geogrid- reinforcement may be provided through a lateral restraint effect through friction and interlock between aggregates, soil and geogrids, tensioned membrane effect and stress dispersion effect [4] as illustrated in Figure 1.

Properties of soil-geogrid interaction can be determined experimentally either from direct shear or pullout tests. However, pullout test is more versatile and appropriate than shear test in that it can be used for geogrid with large aperture due to the size of pullout box. Hence, pullout tests are widely used to study the behaviour of soil-geogrid interface [5-21]. Current understanding of soil-geogrid interaction is based largely on laboratory studies [2]. However, laboratory tests can be influenced by testing procedures, boundary effects and variations in soil properties due to placement and compaction resulting in a wider range of reported results [21-23]. Limit equilibrium is another conventional approach used to design reinforced earth structures.

However, this method cannot be used to predict displacement behaviour of reinforced-earth structures and is usually very conservative [24]. The alternative approach is to use finite element model (FEM) to predict the failure surface and displacement of reinforced-earth structures. FEM overcomes the shortcomings with a more efficient approach to evaluate the soil-geogrid interaction and visualise the distribution of the stresses, strains and displacements. The objective of the paper is to use FEM simulation of pullout tests to study the soil-geogrid interactions and to investigate the influence of several parameters with a wider coverage than available literature.

Literature review

Numerical studies have been conducted successfully by various studies [17,21,24-35] and found that numerical simulations are specifically appropriate for parametric evaluations. A rigorous evaluation of the past studies that have used different constitutive models to simulate the pullout tests is summarised in Table 1. Based on review of literature, there is a lack of performance database to discern the anchorage length and some soil- geogrid interaction studies focused on specific conditions, which may not be suitable in other conditions. Moreover, these studies lack direct industry application. These findings motivated a case study of a large-scale pullout tests with numerical approach to simulate soil-geogrid interaction and to investigate the influence of several parameters with a wider coverage than available literature [36,37].

]

Mechanism of soil-geogrid interaction

The western highlands of Cameroon consist of a zone with steep and very steep slopes and valleys, spreading across three regions of Cameroon namely South-West, West, and North-West regions, and principally on the slopes of the Bambouto Mountains, and the Bamenda Mountains, located in the North-West region. The study area is shown in (Figure 1) which is located between longitude 10°0’0’’ to 10°20’0’’ E and latitude 5°45’0’’ to 6°5’0’’ N and includes the localities of Santa, Bamenda I, and Tubah.

where Pr, Ps and Pb are pullout, frictional and bearing resistances respectively. Equation (1) can be used to determine the pullout resistance for small strain conditions whereby linear elastic material behaviour apply. For large strain before the ultimate load is reached, Equation (1) cannot be used in design because the relative contributions of frictional and bearing resistances are not linearly related to the incremental pullout resistance [26]. The pullout resistance of geogrid can be estimated using Equation (2) as suggested by Moraci et al. [18]:

where f is the coefficient of soil-geogrid interaction mostly range from 0.6 to 0.8 for uniaxial geogrids [39], L is effective reinforcement length, σv is the effective bearing stress between soil and reinforcement and ϕ is the angle of soil friction. σv is the normal stress which is related to overburden pressures/ confined pressures and equal to normal load P over loading area A. Therefore, Equation 2 can be rearranged to show the correlation between overburden pressures/confined pressures and anchorage length.

The full derivation of pullout resistance equation can be found in Jewell et al. [40] and Moraci et al. [18]. Rearranging Equation (2):

In Equation (4), all input parameters can be derived from the pullout tests. It should be noted that Equation (4) only accounts for an average friction coefficient. It does not account for the reduction in the area of geogrid reinforcement including the effective embedded length with pullout displacement and no consideration is given for nonlinear distribution of shear stress developed during the mobilisation of pullout displacement [41].

Numerical modelling of pullout tests

Finite element model

Two-dimensional (2D) planar pullout test model as shown in Figure 2 is developed using a commercially available FEM program- ABAQUS. This model is selected because it is simple and it provides greater efficiency in regard to computation time, execution time, entry and reduction of data, adaptability to pullout test models with more cost-effective solutions.

Pullout box model geometry

A layer of geogrid is inserted at the top of compacted soil in a test box and a second layer of compacted soil is put over it. A surcharge load is applied from the top and bottom of the test box and the geogrid is pulled from the test box. A typical schematic 3D view of the test method is shown in Figure 3. For this study, a 2D planar pullout test model with 153 cm long and 61 cm high in dimension is used. The dimension of the pullout box is chosen to ensure that the soil layer above and below the geogrid is greater than six times the nominal aggregate size [42] thereby mitigating the scale factor and boundary effects.

Boundary conditions

The geogrid is placed in the middle of the pullout test model and is constrained from displacements in Y-directions and no other boundary conditions have been imposed to allow the geogrid to move horizontally. Roller support condition has been assumed on the vertical sides to allow the model to move vertically and restrained horizontally. For the top and bottom surfaces, displacement/rotation boundary conditions have been used. This is achieved by fixing at current position and constraining the degree of freedom in X- and Y-axes. The boundary conditions of the 2D planar pullout test model are shown in Figure 4a.

Load model

During the simulation of pullout test, a constant uniform distribution of normal pressure to simulate confined pressure is applied to the top and bottom of the pullout box. The normal pressure is applied incrementally prior to applying the pullout force as illustrated in Figure 4b as a first load step of the model. The geogrid is then pulled at a constant speed as the second load step of the model. The pullout force is applied at a constant rate until it reaches the maximum applied pullout force of 68 kN/m. In some cases, the simulation terminates before it reaches the maximum pullout force depending on the test conditions. In order to model this process, the pullout force is applied in small increments until the failure occurs. The displacement and load are numerically extracted.

Finite element mesh

There is no clear information available in the literature regarding the optimum mesh size to achieve satisfactory results. Cho et al. [43] found that 25mm quadratic elements to be acceptable. Helwany et al. [44] verified mesh coarseness by assigning all the soil layers the same properties and comparing the FEM results to the Boussinesq’s closed form solution. Mesh sensitivity test is undertaken using different element sizes to ensure that the elements are small enough not to affect the final solutions. A series of numerical analysis has been undertaken to assess the influence of mesh size. FEM meshes with different global seed size ranging from 5mm to 12mm were tried to obtain an appropriate mesh for the simulation of the pullout test that converges to a unique solution. Based on the consideration of accuracy, computation time, and memory storage for the FEM simulations, the optimum element size is found to be 10mm as shown in Figure 5a. The final adopted FEM mesh is illustrated in Figure 5b. An eight-node biquadratic plane strain quadrilateral reduced integration element type is used to discretise the soil and geogrid.

Material constitutive models

Nonlinear constitutive models are needed to simulate the soil under large deformation. Within a typical reinforced soil structure, granular material or coarse sand is normally used as backfill material. The traditional nonlinear elastic method is used to simulate the limiting yield due to tensile stresses generated at the soil-geogrid interface.

Sand and clay soils

Sand exhibits nonlinear, anisotropic, stress-dependent and inelastic properties. The elasto-plastic Mohr-Coulomb model is used for modelling the sand. Mohr-Coulomb model is appropriate for frictional materials such as sand because it is capable of considering the effect of stress history, stress path, dilatancy, and exhibits pressure-dependent yield (the material becomes stronger as the pressure increases) and allows the material to harden and/or soften isotropically. A nominal cohesion value is adopted in this study to enhance the stability of the analysis and avoid convergence issues that may arise due to any singularity. The influence of moisture content on the interface characteristics is eliminated by using dry soil [45]. Mohr-Coulomb model is also used to characterise clay soil. The volumetric changes resulting from shearing of clay soil is characterised by dilatancy angle (ψ). The numerical material properties are taken from published literature and detailed in Table 2.

Geogrid

Geogrid has a 3D structure and cannot be modelled directly using 2D plane-strain FEM. Wilson-Fahmy and Koerner [36] developed a simplified method to convert 2D structure into an equivalent 1D structure while retaining its 2D properties. Another simplification in FEM is to assume geogrid as a continuous equivalent 2D rough planar reinforcement having the total tensile rigidity as the original one while having an equivalent friction angle at the soil interface, which is equal to or smaller than the physical interface angle [27]. But this approach does not account for the discontinuous nature of the geogrid geometry.

In addition, it ignores the unique deformation characteristics of each member during unconfined tensile loading, and the effect of bearing resistance on confined geogrid members [34]. Most recently, Chen et al. [47] developed a new approach to convert the 3D structure into an equivalent 2D structure to replicate uniaxial and biaxial geogrids geometric characteristics while retaining its physical property and assuming the same material property. The procedure to use an equivalent 2D model to approximate 3D structure is documented in Chen et al. [47] research publication. This conversion approach is adopted in this study.

Assuming that the induced strain is very small within the elastic range, it is reasonable to use a linear elastic material to simulate the behaviour of geogrid in a plane-stress condition. For the purpose of this study, a uniaxial high-density polyethylene (HDPE) geogrid ‘Tensar SR2’ reinforcement characteristics are incorporated in the models and the geogrid properties are described in Table 3.

Geogrid-soil interface

Coulomb’s friction model with directional and normal stressdependent friction coefficients is used for shear interactions between the geogrid and surrounding soil [49]. Contact algorithm in ABAQUS incorporates two pairs of interface bonding conditions: one at the upper layer of the geogrid and the other one at the lower layer representing the friction resistance at the sandgeogrid- sand interfaces. A penalty friction, based on Coulomb’s friction law, is used to simulate the frictional behaviour between various soil contact surfaces. The penalty friction is a stiffness formulation of the friction that permits nominal elastic slippage of the surfaces when they should be sticking [33]. This implies that when contacting surfaces are sticking, the magnitude of sliding is limited to the allowable elastic slip. The model contains two material properties: friction coefficient (μ) and the allowable elastic slip parameter (Eslip). The ultimate shearing resistance of geogrids to pullout is related to the average shear stress acting on the specimen (τ) and the effective vertical stress (σv) for a specific friction coefficient. The interface shear strength will not be reached when the relative shear displacement is less than the elastic slip. Based on FEM modelling and verified with pullout test results by past studies, an elastic slip of 0.002 m and the soilgeogrid interface friction coefficient of 0.80 have been adopted in this study.

Model Validation

The suitability of the FEM was validated by comparing it with published data reported by Farrag [50] and Hussein et al. [32] using the same material properties. The calculated loaddisplacement response in this model indicates a non-linear elastic behaviour up to failure at applied load of 68 kN/m taken at the displacement of 15mm (i.e., 14% strain) as shown in Figure 6. It reveals that the measured displacement response is non-linear up to failure (at an applied load of 68 kN/m). The calculated pullout displacement is generally linear in the initial phase of the pullout due to the elasto-plastic nature of the selected material model. It is worth noting that the model successfully predicted the maximum geogrid strength at failure.

Figures 7a and 7b show the comparison of the past experimental and numerical results and simulated results in this study. It shows good agreement for pullout force up to about 50 kN/m. Upon validation, the validated model with an appropriate variation is used to perform actual simulations for studying the influence of geogrid-soil interface behaviour, confined pressure, geogrid transverse members, fines content, friction angle and soil density.

Analysis and Results

The results of the numerical simulation of pullout test conducted to capture the influence of the confined pressure, soil density, geogrid transverse members, fine contents and soil friction angle on soil-geogrid interaction under small displacements and strains are presented in this section.

Influence of confined pressure

Figure 8a shows load-displacement results of the same geogrid specimen under six different confined pressures (20, 49, 75, 100, 150 and 200 kPa). The results suggest that the displacement decreases with an increase in confined pressure for a constant pullout load. This indicates that increasing the confined pressure (overburden) improves the ultimate pullout capacity which is consistent with the findings by Ren et al. [51]. This is attributed to higher confined pressures which cause a compression of the soil close to the geogrid resulting in local densification of the soil in front of the transverse members leading to greater passive earth resistance being mobilised to prevent moving. This leads to an increase in pullout resistance. However, under low confined pressure, pullout softening may not occur because the transverse (bearing) member penetrates into the soil in front gradually during the pullout process. When the soil strength in front of the transverse member is fully mobilised and reaches a critical state value, it will yield leading to no softening behaviour.

Effectively, a state of failure develops in the surrounding soil and the geogrid is pulled out of the soil when the passive (pullout) resistance reaches a local maximum value [13, 52]. Figure 8a suggests that at higher confined pressures exceeding 48 kPa, the simulations terminated at a pullout force of 68 kN/m. Given that peak pullout load for geogrid Tensar SR2 is 79 kN/m as detailed in Table 3, geogrid pullout resistance did not reach its peak value. Unlike in laboratory tests, there is no possibility of modelling geogrid rapture beyond the maximum tensile strength of geogrid. Therefore, the maximum pullout resistance in any pullout test is expected to be less than the actual tensile strength of the reinforcement.

Figure 8b shows that the displacement is non-linear along the length of the geogrid for different values of applied forces. It can be seen that the displacement has two distinct phases. The first phase exhibits moderate displacement when the applied load is 40% (40%Pr) of the maximum pullout force (68 kN/m). There is a sharp increase in displacement beyond this in the second phase. A progressive mobilisation of the displacement along the geogrid length is noted with the incremental pullout force. This is attributed to gradual transfer of resistance force along the longitudinal members to transverse members.

Influence of friction coefficient

The friction coefficient decreases from 1.08 to 0.72 when the confined pressure increases from 20 kPa to 100 kPa as illustrated in Figure 9. The lower confining pressure leads to higher friction coefficient due to greater dilative response of the dense sand under low stresses. This implies that at higher overburden pressures only very short anchorage lengths are required even with a low values of friction coefficient. It is noteworthy to highlight that under 48 kPa confined pressure and friction coefficient of 0.94, the results of FEM simulation are in reasonable agreement with the laboratory pullout test results as discussed earlier in ‘model validation’ section.

Figure 10a shows the effect of friction coefficient on displacement behaviour with incremental pullout force until it reaches the maximum value at constant confining pressure of 48 kPa. There is an increase in pullout capacity with the increase in friction coefficients. However, the model was unstable and terminated at approximately 14mm displacement during simulation. Figure 10b shows the pullout-strain results from six tests carried out on the same geogrid specimen under a range of confined pressures. The results show that the pullout-strain relationship is relatively linear for small strains and the confined pressure has little effect on pullout-strain response. This is attributed to the elasto-plastic nature of the selected material. Therefore, the use of linear relationship for geogrid load-strain tensile response is justified in design where relatively small strains are developed in the linear elastic material.

In this study, there is no obvious peak pullout force and the maximum pullout force of 68 kN/m is taken at a 14% strain. However, for large strains where the geogrid rupture is imminent, solutions incorporating a nonlinear formulation for the geogrid is warranted. For design purposes, the value of secant modulus at relative elongation (strain) of 2% is normally used [53]. In the case of pullout test results, the secant for 2% elongation coincides with the corresponding pullout force of 10 kN/m. Therefore, the numerical result for geogrid Tensar SR2 stiffness at 2% strain is approximately 500kN/m. Based on manufacturer’s literature, the geogrid stiffness for Tensar SR2 at 2% strain is 1096 kN/m. This value is equal to approximately 219% of the modulus value at 2% strain. This suggests that the geogrid used is appropriate for the pullout test conditions and meets the minimum requirement by a factor of 2.19.

Anchor length

Figure 11a illustrates that non-linear displacement developed along the length of the geogrid up to the tensile rapture. The plots in the figure show that the front displacement decreases with an increased confined pressure. The geogrid tensionedmembrane effect is observed mainly in the front end of the pullout and reduced to zero at a distance of around 400 mm to 600 mm from the front end depending on the applied confined pressure. It also shows that for lower confined pressure of 20kPa, the geogrid moves more relative to the soils up to 600 mm. This indicates that the geogrid has no effect on the front displacement beyond a distance of 600 mm from the front end of the pullout. This phenomenon is attributed to the dilation of the soil during mobilisation of the geogrid at low confined pressure. The differences in relative displacement can be attributed to the increase in frictional resistance between the soil and geogrid and the bearing resistance of the geogrid transverse members that results from increase in confining pressure [1, 40].

Figure 11b shows that the axial strain distribution is not constant and therefore the relative displacement along the geogrid length is expected to exhibit similar behaviour. The results indicate that the calculated axial strain in the geogrid becomes insignificant beyond 400 mm from the point of the application of pullout displacement. This implies that geogrid did not have any contribution to the axial strain beyond this point. There is no difference in axial strain at this threshold location, indicating that there is no impact of higher confined pressure exceeding 75 kPa on anchorage lengths. This finding indicates that long anchorage lengths are not required as geogrid axial strain under higher confined pressure (equal or greater than 75 kPa) diminishes to zero.

Influence of transverse members

It is easier to visualise the effects of the soil-geogrid interaction in FEM. Figure 13 shows that the asymmetric wedge is formed ahead and around the transverse members under confined pressure of 48 kPa and pullout displacement of 16.8mm. However, the size of the wedge-shaped shear zone decreases along the length of geogrid away from the front end of the pullout. In addition, the soil in front of each transverse member fails individually and is asymmetrical about the interface The results demonstrate the development of punching shear failure in front of the transverse members and progressively reduce in footprint away from the point of application. At the same time, a low stress region is formed behind the transverse members and results in the softened region. The soil failure can occur in the softened region. It appears that the stress is gradually transferred along the longitudinal members to transverse members as reflected with the increase in stress blocks. Furthermore, the first transverse member underwent the greatest mobilisation of the bearing mechanism.

Influence of fines content

Figure 14a shows the pullout force versus displacement response for the sand with different fines content. As expected, Sand B with low fines content and higher modulus outperformed Sand C with higher fines content and lower modulus. However, Sand A with zero fines and highest modulus has similar performance as that of Sand C. This observation indicates that while fines content has a weak association with the pullout performance, grain size and the modulus value are equally important. For a constant confined pressure of 48 kPa, the plots in Figure 14b shows that the maximum geogrid movement is recorded at the loaded end and the displacement is reduced to zero at a distance around 1000 mm from the front end for Sand A and Sand B.

But the geogrid movement is extended to 1200 mm in the case of Sand C. This is attributed to the presence of higher fines content in Sand C, causing the reduced resistance against displacement of the geogrid. Interestingly, the maximum displacement in Sand A is higher than Sand B, though both soils have the same friction angle. The presence of fines is responsible for the lower pullout resistance in the case of Sand B. Figure 15 shows the effect of geogrid displacement response using Sand C under pullout condition with significant increase in confined pressures (48, 100, and 200 kPa). This is attributed to better mechanical interlocking of soil particles with confinement. It shows that the rupture zone has been forced away from the transverse member towards the front end of the pullout with the exponential increase in confined pressure and is asymmetrical about the interface. It also illustrated that the rapture zone is well clear of the boundary effects.

Figure 16 shows the significant increase in base movement near the geogrid during the pullout simulation with lower confinement pressure. However, the stiffened zones occurred approximately 100 mm above and below the geogrid in all confined pressure cases. The effective stiffened zones within the granular base decrease when extending away from the geogrid which is consistent with the findings by Huang et al. [54].

Influence of soil friction angle

Figure 17 shows numerical estimates of the relationship between the pullout force and displacement for soil friction angle (ϕ) equal to 30◦, 35◦ and 41◦ for geogrid under 48 kPa confined pressure. At higher pullout force exceeding 60 kN/m, there is a marginal increase in displacement with the decrease in friction angle. These results demonstrate that the soil friction angle has an insignificant effect on pullout displacement response of the geogrid. This finding is consistent with the trend reported by Attache and Mellas [46].

Influence of soil density

The density of soil can affect the strength and deformability of reinforced soil, hence the pullout resistance. At higher density, the grain structure of the soil is more compacted in the geogrid openings leading to an increase in the passive earth resistance in front of the transverse members of the geogrid. The contact points between the soil and the geogrid increase with the increase in soil density. For the purpose of this study, sand and clay are used which correlate with high and low density respectively.

Figure 18a shows the effect of soil density on geogrid pullout resistance under constant confining pressure of 48 kPa. The results indicate that for the lower soil density (clay) the peak pullout force is 47 kN/m. It can be seen from the test data that an increase in the relative density of soil results in moderate increase in peak pullout resistance. This observation can be attributed to the fact that dense soils tend to dilate when shear stresses are mobilised along the reinforcement interface. Under confining pressure, the dilatancy is restricted by the surrounding soil resulting in an apparent increase of the pullout resistance. The increases in soil density leads to greater soil-geogrid shear resistance, resulting in the decrease in geogrid displacement and increasing the interface modulus and the pullout resistance.

The displacement along the geogrid is normalised with respect to the front displacement. This explains the differences in strain as shown in Figure 18b. The numerical results show the influence of soil density on the mobilised shear distribution along the reinforcement. The higher soil density leads to a higher shear stress concentration at the vicinity of the point of pullout and the displacement decay to zero at a distance around 600 mm from the front end of the pullout. For the higher soil density, only a half of the geogrid length contributes to the adherence resistance. Geogrid embedded in small grain soil (sand) exhibits higher secant tensile stiffness compared with fine grained soil (clay) thereby influencing the pullout capacity and the displacement. It is worth noting that the interlocking effect of sand-geogrid interfaces is relatively weak compared with coarse grained material and the pullout resistance is dominated by the friction effect which is affected by the tensile strength of the geogrid.

Conclusion

In this study, a 2D planar FEM analysis of soil-geogrid interaction under pullout test conditions is developed using the ABAQUS program. The pullout test models are developed to simulate the micro-mechanism and interaction between soil and geogrid transverse members, peak pullout force, maximum geogrid axial displacement and tensile strain at pullout location including displacement and strain distribution along the geogrid reinforcement. A uniaxial geogrid type “Tensar SR2” reinforcement characteristics are incorporated in the pullout models. The input parameter values into the material constitutive models are obtained from published literature. The numerical results from the simulations of geogrid pullout test models show relatively good agreement with the measured values reported by Farrag [50] and Hussein et al. [32]. It demonstrates that Coulomb’s friction model for interface can reasonably simulate the soil-geogrid interaction when the geogrid slips along the effective embedded length. The rupture zone has been forced away from the transverse members towards the front end of the pullout with the exponential increase in confined pressure. The rapture zone is well clear of side and upper and lower box boundary effects. Therefore, the selected pullout box dimension chosen has eliminated the effect of these boundaries on test results. The shear band caused by the transverse member is asymmetrical about the interface. The observed wedge like shear zone ahead of the transverse member supports the concept of punching failure mode.

A number of results can be drawn from this study. The apparent friction coefficient decreases with the increase in confining pressures and hence the anchorage zone of geogrids may be shorter than usual in practice. The pullout displacement subjected to a constant pullout force decreases with the increase in confined pressure. The displacement and strain profiles along the geogrid are not linearly related to the incremental pullout force. This finding reinforces the need to use active effective embedded length in Equation (3). The calculated result for geogrid “Tensar SR2” pullout test at 2% strain under confined pressure condition is 500kN/m. It demonstrates that the peak pullout strength is mobilised at strain levels much lower than those of the unconfined tensile tests by a factor of 2.19. This suggests that the design criteria of reinforced soil structures should be based on pullout test results rather than the unconfined tensile test results. An increase in the soil relative density results in moderate increase of the peak pullout resistance. This finding illustrates the influential effect of soil types, soil placement and compaction.

The pullout resistance decreases with the increase in the fines content of the soil. This observation indicated that while fines content is the weak link to the pullout performance, the soil grain size and the modulus values are equally important. The soil friction angle has an insignificant effect on pullout displacement response of the geogrids. The coefficient of soil friction has an influence on the pullout capacity and the displacements. This finding reiterates the importance of using the correct soil-geogrid interface properties to model contact interactions in the FEM. Several soil and geogrid parameters have a significant influence on the pullout load-displacement performance without a single parameter showing dominance. The model can capture the 3D response of the geogrid considering its equivalent 2D physical geometry. This demonstrates the effectiveness of using ABAQUS program to capture the physical disturbances at the soil-geogrid interface during the load-transfer mechanism between soil particles and geogrid. The present study preformed only a limited number of test conditions with only one type of geogrid. Further studies with different geogrids and wider coverage of soil types are essential for establishing generalised observation and design recommendations.

References

1. Palmeira EM (2004) Bearing force mobilisation in pullout tests on geogrids. Geotextiles and Geomembranes 22(6): 481-509.
2. Palmeira EM (2009) Soil-geosythetic interaction: Modelling and analysis. Geotextiles and Geomembranes 27(5): 368-390.
3. Minazek K, Mulabdic M (2013) A review of soil and reinforcement interaction testing in reinforced soil by pullout test. Gradevinar 3: 235-250.
4. Guo X, Zhang H, Liu L (2020) Planar geosynthetic-reinforced soil foundations: a review. SN Applied Sciences p. 8.
5. Jewell RA (1980) Some effects of reinforcement on soils. Ph.D. Thesis, University of Cambridge, Cambridge, UK.
6. Dyer MR (1985) Observation of the stress distribution in crushed glass with applications to soil reinforcement. PhD Thesis, University of Oxford, UK.
7. Palmeira EM, Milligan GWE (1989) Scale and other factors affecting the results of pull-out tests of grids buried in sand. Geotechnique 39(3): 519-521.
8. Alfaro MC, Hayashi S, Miura N, Watanabe K (1995) Pullout interaction mechanism of geogrid strip reinforcement. Geosynthetics International 2(4): 679-698.
9. Lopes ML, Ladeira M (1996) Role of specimen geometry, soil height and sleeve length on the pullout behaviour of geogrids. Geosynthetics International 3(6): 701-719.
10. Ochiai H, OtanI J, Hayashi S, Hirai T (1996) The pullout resistance of geogrids in reinforced soil. Geotextiles and Geomembranes 14(1): 19-42.
11. Lopes MJ, Lopes ML (1999) Soil-geosynthetic interaction-influence of soil particle size and geosynthetic structure. Geosynthetics International 6(4): 261-282.
12. Alagiyawanna AMN, Sugimoto M, Sato S, Toyota H (2001) Influence of longitudinal and transverse members on geogrid pullout behaviour during deformation. Geotextile and Geomembranes 19: 483-507.
13. Meyer N, Nernheim A, Emersleben A (2003) Influence of confining pressure, soil density and types of geogrids on soil-geogrid interaction coefficient. e-Conference “Modern Trends in Foundation Engineering Geotechnical Challenges and Solutions”, IITM, India.
14. Abdel-Rahman AH, Abdel-Moniem IM, Ashmawy AK (2007) Utilisation of a large scale testing apparatus in investigating and formulating the soil geogrid interface characteristics in reinforced soils. Australian Journal of Basic and Applied Science 1(4): 415-430.
15. Dixon N (2010) Soil-geosynthetic interaction: interface behaviour. Proceedings of the 9th International Conference on Geosynthetics, Guaruja Brazil 563-582.
16. Zhou J, Chen JF, Xue JF, Wang JQ (2012) Micro-mechanism of the interaction between sand and geogrid transverse ribs. Geosynthetics International 19(6): 426-437.
17. Abdi MR, Zandieh AR (2014) Experimental and numerical analysis of large scale pull out tests conducted on clays reinforced with geogrids encapsulated with coarse material. Geotextiles and Geomembranes 42(5): 494-504.
18. Moraci N, Cardile G, Gioffre D, Mandaglio MC, Calvarano LS,Carbone L (2014) Soil geosynthetic interaction: Design parameters from experimental and theoretical analysis. Transp. Infrastruct. Geotech 1: 165-227.
19. Choudhary AK, Krishna AM (2016) Experimental investigation of interface behaviour of different types of granular soil/geosynthetics. Int. J. of Geosynth. and Ground Eng 2(4): 11.
20. Cardile G, Gioffre D, MoracI N, Calvarano LS (2017) Modelling interfacing between the geogrid bearing members under pullout loading conditions. Geotextiles and Geomembrances 45: 169-177.
21. Hegde A, Roy R (2018) A comparative numerical study on soil-geosynthetic interactions using large scale direct shear test and pullout test. International Journal of geosynthetics and Ground Engineering 4(2): 11.
22. Juran I, Knochenmus G, Acar YB, Arman A (1988) Pullout response of geotextiles and geogrids. In: Proceedings of Symposium on Geotextiles for Soil Improvement. ASCE, Geotech. Special Publication 18: 92-111.
23. Bakeer RM, Abdel-Rahman AH, Napolitano PJ (1998) Geotextile friction mobilisation during field pullout test. Geotextiles and Geomembrances 16(2): 73-85.
24. Yuan Z, Chua KM (1990) Numerical evaluation of the pullout box method for studying soil-reinforcement interaction. Transportation Research Record 1278, National Research Council, Washington, DC 116-124.
25. Yogarajah I, Yeo KC (1994) Finite element modelling of pull-out tests with load and strain measurements. Geotextiles and Geomembranes 13(1): 43-54.
26. Yan S, Feng S, Barr B (1998) Finite element modelling of soil-geogrid interaction dealing with the pullout behaviour of geogrids. The Chinese Society of Theoretical and Applied Mechanics, Beijing, China 14(4): 371-382.
27. Peng F, Kotake N, Tatsuoka F, Hirakawa D, Tanaka T (2000) Plane strain compression behaviour of geogrid- reinforced sand and its numerical analysis. Japanese Geotechnical Society, Soil and Foundations 40(3): 55-74.
28. Duszynska A, Bolt AF (2002) Soil-geogrid interaction in pullout test at 2D deformation conditions. 7th International Conference on Geosynthetics, Nice, France 4 pages.
29. Perkins SW, Edens MQ (2003) Finite element modelling of a geosynthetic pullout test. Geotechnical and Geological Engineering 21: 357-337.
30. Sugimoto M, Alagiyawanna AMN (2003) Pullout behaviour of geogrid by test and numerical analysis. Journal of Geotechnical and Geoenvironmental Engineering, ASCE 129(4): 361-371.
31. Aggarwal P, Gupt KK, Sharma KG (2008) Modelling of pull-out behaviour of geogrid embedded in sand. International Journal of Modelling and Simulation 28(3): 264-270.
32. Hussein MG, Meguid MA, Mowafy Y (2009) On the 3D modelling of soil-geogrid interaction, Geo-Halifax 6 pages.
33. Mosallanezhad M, Taghavi SHS, Hataf N, Alfaro MC (2016) Experimental and numerical studies of the performance of the new reinforcement system under pull-out conditions. Geotextiles and Geomembranes 44: 70-80.
34. Hussein MG, Meguid MA (2016) A three-dimensional finite element approach for modelling biaxial geogrid with application to geogrid-reinforced soils. Geotextiles and Geomembranes 44(3): 295-307.
35. Tang X, Wang Y, Feng Y (2017) Numerical studies on the interface features experiments of geogrids. Asia-Pacific Engineering and Technology Conference 2282-2290.
36. Wilson-Fahmy RF, Koerner RM (1993) Finite element modelling of soil geogrid interaction with application to the behaviour of geogrid in a pullout loading condition. Geotextiles and Geomembranes 12(5): 479-501.
37. Siriwardane H, Gondle R, Kutuk B, Ingram R (2008) Experimental investigation and numerical analysis of reinforced geologic media. 12th International Association for Computer Methods and Advances in Geomechanics, Goa India.
38. Koerner RM, Wayne MH, Carrol RG (1989) Analytic behaviour of geogrid anchorage. Proceedings Geosynthetics ’89, San Diego, CA, IFAI Publishers, St. Paul, MN, USA 525-536.
39. Han J (2015) Principles and practice of ground improvement. John Wiley and Sons, Inc., Hoboken, New Jersey.
40. Jewell RA, Milligan GWE, Sarsby RW, Dubois D (1984) Interaction between soil and geogrids. Symposium on Polymer Grid Reinforcement in Civil Engineering, ICE, London 18-30.
41. Fuller AL (1997) On the comparative behaviour of geogrids in tension and pullout. M.Appl.Sc Dissertation, The University of British Columbia, Canada.
42. Austroads (2009). Guide to pavement technology Part 4G: Geotextiles and Geogrids. AGPT04G-09 Austroads, Sydney, ISBN: 978-1-921551-68-0.
43. Cho Y, Mccullough BF, Weissmann J (1996) Considerations on finite element method application in pavement structural analysis. Transportation Research Record 1539, TRB, National Research Council, Washington DC 1539: 96-101.
44. Helwany S, Dyer J, Leidy J (1998) Finite element analyses of flexible pavements. Journal of Transportation Engineering 124(5): 491-499.
45. Zhang J, Ji M, Jia Y, Miao C, Wang C, et al. (2021) Anisotropic shear strength behaviour of soil- geogrid interfaces. MDPI. Applied Science 11(23): 11387.
46. Attache S, Mellas M (2017) Numerical study of large scale pull-out test of horizontal corrugated strips. International Journal of Geotechnical Engineering.
47. Chen R, Luan M, Hao D (2011) Improved simulation method for soil-geogrid interaction of reinforced earth structure in FEM. Trans. Tianjin Uni 17: 220-228,
48. Bathurst RJ (1990) Instrumentation of geogrid-reinforced soil walls. Transportation Research Board 1277: 102-111.
49. SIMULIA (2013) ABAQUS analysis user’s manual. version 6:13.
50. Farrag K (1990) Interaction properties of geogrids in reinforced soil wall testing and analysis. Ph.D. Dissertation, Louisiana State University, Baton Rouge, LA.
51. Ren F, Huang Q, Liu Q, Wang G (2020) Numerical study on the pull-out behaviour of planar reinforcements with consideration of residual interfacial shear strength. Transportation Geotechnics 35.
52. Abdi MR, Sadrnejad A, Arjomand MA (2009) Strength enhancement of clay by encapsulating geogrids in thin layers of sand. Geotextiles and Geomembranes 27(6): 447-455.
53. DTMR (2022) Transport and Main Roads Specifications MRTS58 Geosynthetics for subgrade and pavement reinforcement. Technical Specification. Queensland: Department of Transport and Main Roads.
54. Huang H, Bhandari A, Yang X (2011) Numerical modelling of geosynthetic-reinforced earth structures and geosynthetic-soil interaction. Geotechnical Engineering Journal of the SEAGS & AGSSEA 42(1): 14.