Development of subgrade M_r constitutive models based on physical soil properties

Resilient modulus (Mr) is a measure of the stiffness for geomaterials under cyclic loading. The evaluation of Mr for fine-grained subgrade soils can be challenging due its sensitivity to variation in moisture content and stress state. Additionally, pavement subgrades often exist under unsaturated conditions, which further complicates the evaluation of Mr. For unsaturated soils, matric suction needs to be incorporated into Mr models to accurately capture the stress state of the soil. The importance of accurate Mr evaluation in pavement design has long been recognised; however, laboratory measurement of Mr remains inaccessible to most practitioners due to the specialised equipment and personnel needed to evaluate Mr in the laboratory. This study provides a method to evaluate the Mr value from simple soil physical properties that also account for the effects of moisture variation and stress state without needing to perform laboratory repeated load triaxial (RLT) tests. A laboratory investigation, which included performing RLT tests to obtain Mr values at different moisture contents and measuring soil water retention curves (SWRC), was conducted on four different fine-grained soils. A statistical regression analysis was performed to develop prediction models for evaluating regression coefficients (i.e. k1, k2, k3) of Mr constitutive models based on simple and unique soil physical properties (e.g. degree of saturation, activity parameter). The statistical prediction models were developed for two different constitutive Mr models, one of which accounted for the stress state of unsaturated soils by incorporating matric suction. Results of this study show good agreement between the measured and predicted Mr values obtained utilising the statistical regression models.