A series of numerical analyses were conducted to determine the effects of geogrids with a granular capping layer on the stabilization of weak subgrade soil. Total vertical settlements and bearing capacity were observed as strength indicators, and strain modulus (EV2) values were calculated as the key performance indicator for monotonic loading on the subgrade models. The modulus results were further analyzed to develop design curves for determination of capping layer thickness, with or without a geogrid, on soft subgrades with low strength below CBR 5.0%
3.1 Strength Properties of Weak Subgrade
Recent research outcomes have highlighted the significance of accurately evaluating the strength properties of the subgrade when a new mechanism introduces for designing the capping layers [37] particularly for low strength subgrades. Investigating the strength characteristics of a weak subgrade, when the CBR is below 5.0%, presents challenges due to poor workability of materials. Hence, validated Plaxis finite element models corresponding to CBR values of 1.5% and 2.5% were utilized to determine the properties of subgrades with different strengths conditions below 5.0% of CBR in the material model during Plaxis analysis. Subsequently, Plaxis models were developed to represent weak subgrade CBR values of 0.5, 1.0, 1.5, 2.5, 3.5, and 5.0%, and were analyzed to determine their strength and stiffness properties. Figure 6 shows the comparison of numerical analysis of low and high subgrade representing CBR 0.5, 1.0, 3.5 and 5.0%. The numerical results illustrate high vertical deformation, approximately 150mm when CBR is below 1.0%, gradually reducing with increasing CBR up to 5.0%. In contrast, a higher concentration of stress from the loading plate can be observed, with values around 405 kPa in the subgrade with a CBR of 5.0%. The stress bulb propagation extends to a large depth, with stresses at around 25 kPa at a depth of approximately 500mm, as observed at the lower limit of the bulb. However, a reduction in stress propagation is observed with decreasing CBR, being lowest in the subgrade with a CBR of 0.5%, with stress only around 40 kPa below the loading plate. The extracted values from the numerical analysis, including ultimate bearing capacity and strain modulus values of all subgrades, were plotted in Figure 7 alongside the results obtained from laboratory tests. The laboratory test results of CBR values of 1.5% and 2.5% show close values, confirming the validity of the numerical models. These models were employed to derive the material properties of weak subgrades for the analysis of strength properties of stabilized subgrades with a capping layer, both with and without a geogrid.
3.2 Strength of Stabilized Subgrade
Validated models (Figure 3 and 4) were employed to estimate the strength properties: bearing capacity and vertical deformation of the stabilized subgrade with a capping layer and, both with and without a geogrid at the interface of the capping layer and subgrade soil. The subgrade properties were varied to compromise the strength as per CBR range of 0.5% to 5.0% and the capping layer thickness was varied as 100, 200, 300, 400 and 500 mm for a single type of material.
Numerical analysis revealed a significant improvement in the stabilized subgrades when a geogrid is placed at the interface of the subgrade and capping layer. Figure 8 represents an example of numerical analysis comparing low and high CBR values of the subgrade (CBR 1.0% and 3.5%, respectively), illustrating the vertical deformations of the stabilized subgrade profiles under the same loading conditions (550 kPa) and with a consistent capping layer thickness of 200mm, both with and without geogrids. Figure 8(a) demonstrates that the reduction in vertical deformation with the inclusion of a geogrid is approximately 26.6% for a 200mm granular capping layer when the subgrade CBR is 1.0%. Similar effects are observed with an increase in subgrade CBR, as shown in Figure 8(b), which exhibits approximately a 23% reduction in surface vertical deformation when the subgrade CBR is 3.5%. However, the effect of the geogrid decreases with an increase in subgrade strength (CBR) as well as with capping layer thickness. The numerical analysis further reveals that the reduction in vertical deformation ranges from approximately 34% to 9.0% when the capping layer thickness increases from 100mm to 500mm for the stabilized subgrade with the lowest CBR of 0.5%. For the stabilized subgrade with a CBR of 5.0%, the reduction ranges from approximately 14% to 2.5% over the same increase in capping layer thickness from 100mm to 500mm.
The bearing capacity was analyzed as the typical strength parameter in numerical analysis, and the load bearing capacity values of the analyzed models were extracted at a vertical deformation of 25mm for standard consistency in comparison, as shown in Figure 9 [41]. The results show that in-situ subgrade with a CBR less than 1.0% can be improved above the respective required minimum strength for constructions, which is 5.0% of CBR, with a 100mm thick capping layer and a geogrid at the interface. However, it requires a 200mm granular capping layer without a geogrid. Considerable enhancement of the bearing capacity (over 1.0 MPa) can be predicted when the capping layer thickness is above 400mm for weak subgrades with a CBR less than 1.0%. However, when the subgrade CBR is 1.0%, a significant improvement in the bearing capacity, exceeding 2.0 MPa, was achieved with a greater capping layer thickness of 500mm and a geogrid at the interface. Further analysis revealed a diminishing effect of the capping layer thickness on the bearing capacity of weak subgrades as the layer thickness increased, as illustrated in Figure 10. The highest percentage increase in bearing capacity, 115%, was observed in subgrade CBR 0.5% with a 100mm capping layer, but the magnitude was lower due to the weakest subgrade, and the effect of the geogrid gradually decreased to 80% and 62% with 200mm and 300mm capping layers, respectively. In contrast, it significantly dropped for the same subgrade strength to approximately 20% with the increase in 400mm and 500mm capping layers, reflecting the reduction of the geogrid reinforcement effect with increases in capping layer thickness on soft subgrade. This trend is commonly observed when the subgrade strength increases from 0.5 to 5.0%. The percentage increase in bearing capacity was reduced from 115% to 25% when the subgrade strength varied from 0.5% to 5.0% for a constant capping layer thickness of 100mm. With the increase in capping layer thickness, the effect of geogrid reinforcement further reduced on the subgrade with high strength of CBR 5.0%, as it was observed to be around 2-4% for 400-500mm capping layer thickness. In summary, Figure 10 illustrates that geogrid reinforcement is highly effective when the in-situ subgrade CBR is less than 2.5% with 100-300mm capping layer thickness, as it shows over a 30% increase in bearing capacity. This agrees with the phenomenon that the influence of geogrid reinforcement decreases with the distance of the loading from the geogrid. Therefore, it is desirable to stabilize weak subgrades with 200-300mm capping layer thickness, although it depends on the requirement of the target strength.
3.3 Stiffness of Stabilized Subgrades
Evaluation of subgrade stabilization using a capping layer typically involves considering ultimate bearing capacity and modulus values. However, recent research emphasizes the importance of analyzing the modulus of stabilized subgrades to ensure consistent and precise assessment [34]. The strain modulus (EV2) was evaluated to analyze the effect of capping layer thickness and geogrid reinforcement on the stiffness properties of the stabilized subgrades. The results of EV2 values concerning different capping layer thicknesses on various subgrade CBR levels (ranging from 0.5 to 5.0%) are depicted in Figure 11. Continuous lines represent EV2 values with the effect of geogrid reinforcement, while dashed lines represent EV2 values with only the capping layer. In the comparison of geogrid reinforcement, the most significant impact was observed for a 100mm capping layer thickness, exhibiting a strain modulus improvement ranging from 1.2 to 1.8 times within the subgrade CBR range of 0.5 to 5.0%. This effect gradually diminishes with increasing capping layer thickness, with improvements ranging from 1.15 to 1.55 for 200mm, 1.05 to 1.23 for 300mm, and minimal effect for 400mm and 500mm thicknesses, resulting in approximately a 1.04 times improvement when the subgrade strength varies from 0.5 to 5.0%. The effect of geogrid reinforcement becomes minimal with increasing capping layer thickness when the in-situ subgrade CBR increases to 5.0%. The effect is significant for 100mm and 200mm capping layer thickness within the subgrade CBR range of 0.5 to 5.0%. However, the maximum improvement of the subgrade with a CBR of 0.5% is less than 35 MPa, even with geogrid reinforcement and a 500mm capping layer. This improvement is not considered significant, as suggested by Adorjányi [4], who recommends enhancing the subgrade modulus to at least 40 MPa with a capping layer to ensure a safer working platform. Furthermore, it was observed that a 100mm capping layer thickness did not achieve stiffness beyond 40 MPa, even with geogrid reinforcement for subgrades with a high CBR of 5.0%. Therefore, the findings recommend improving weak subgrades with a CBR below 1.0% using chemical methods such as lime and/or fly ash mixing to enhance the CBR to at least above 1.0% to achieve the ultimate effect of geogrid reinforcement with a granular capping layer.
3.4 Optimal Capping Layer Thickness for Weak Subgrade
The numerically evaluated EV2 values of stabilized subgrades were further analyzed to develop design-curves to estimate the required thickness of granular capping layer to achieve the desired strength of platform designing on weak subgrade. A correlation was developed for different CBR strength of subgrades between two parameters, subgrade improvement ratio (SIR) and cover ratio which are defined in equations (2) and (3) respectively.
SIR = EV2 (Stabilized subgrade) / EV2 (subgrade) (2)
Cover ratio = Thickness of granular capping layer / Diameter of loading area (3)
(in this study, diameter of loading area = 200mm)
Figure 12 presents six developed curves representing different strengths of subgrades CBR from 0.5 to 5.0%. These curves determine the necessary capping layer thickness to achieve the desired improvement level (EV2) of the subgrade, provided the CBR of the existing subgrade is known. Two curves in each figure represent the subgrade stabilisation with and without geogrid reinforcement. For instance, when it requires to increase the EV2 by 10 times without geogrid reinforcement (SIR=10) for a subgrade with a CBR of 1.0% (corresponding cover ratio=3.1), the calculated capping layer thickness would be 620mm (3.1 x 200mm). When it with a geogrid reinforcement the “cover ratio is 2.5” thereby, the required capping layer thickness is 500mm (2.5 x 200mm).
The developed design graphs illustrate the significant impact of capping layer thickness on subgrade strength, particularly when the CBR is lower. For instance, the EV2 of a subgrade with 0.5% CBR can increase approximately 20 times with a 700mm capping layer and a geogrid in the interface, whereas it only exhibits an increase of about 3 times for subgrades with a strength of 5.0% CBR. This observation suggests that the improvement ratio is less sensitive to variations in layer thickness when the in-situ CBR strength of the subgrade is higher. Furthermore, the graphs demonstrate the high efficiency of geogrid reinforcement in the middle range of subgrade CBR from 1.0 to 3.5%; conversely, they show that the two curves (with and without geogrid reinforcement) are merging for low subgrade CBR below 1.0% and high CBR of 5.0%.
The developed curves are applicable for effectively designing the capping layers on weak subgrades when the CBR is below 5.0%. However, the design is specific to applying high quality base course materials with similar properties to material Type 2.1 as per MRTS05 [33] specifications. In contrast, the mechanism is applicable to develop similar charts for different types of capping layer materials.