Modelling Initial Elastic Modulus of Sand-EPS Mixtures Using Evolutionary Polynomial Regression

Due to its low density as well as high strength/density ratio, expanded polystyrene (EPS) has gained significant popularity as a viable approach to soil stabilization. Considering the increasing use of EPS as a lightweight fill material, the development of reliable correlations for predicting its compressibility behavior would be beneficial due to the reduction of the costs connected with conventional laboratory testing. In the current study, undrained-unconsolidated (UU) triaxial compression tests were carried out on sand-EPS specimens prepared with various EPS contents and subjected to different confining pressures. Having observed and analyzed the compressibility responses of the composites, an EPR-based model was proposed and verified for the purpose of predicting the initial elastic modulus. Subsequently, in order to investigate the relative impact of each input variable on the estimated output, sensitivity analysis, as well as parametric study were carried out. The results demonstrate that although both the confining pressure and the EPS volume content have a consequential influence on the predicted initial elastic modulus, the former is a more sensitive component of the model

Keywords:

EPS beads , sand-EPS mixture , compressibility behavior , Evolutionary Polynomial Regression , EPR model