Experimental investigation on material optimization in railway embankments using biaxial geogrids

The significance of geogrids in the reduction of displacement and the distribution of stresses in soil embankments is beyond dispute. With the increase in population, resources are already under extreme pressure and especially in the infrastructure sector of India. In this study, an attempt has been made to reduce the construction material usage in railway embankments with help of geogrids. Biaxial geogrid has been utilized to optimize layer thickness in railway embankments based on deformation modulus of the embankment. A series of tests in accordance with DIN 18134, have been performed. Results show that the deformation modulus (E_v2) measured from the second cycle of the plate bearing test showed a 31.47% improvement in E_v2 values with comparison to Indian design methodology for a design load of 25 Tonnes (t). To further enhance the results, regression analysis has been used to understand the influence of embankment thickness on E_v2 values. The developed equation is used to predict the performance of the embankment, resulting in optimized embankment configurations. To aid the regression analysis, a machine learning (ML) approach was undertaken to quantify the design variables governing E_v2 and predict E_v2 along with synthesis of alternative embankment configurations. The trained model predicted an embankment height of 650 mm, satisfying all design criteria to achieve target modulus values with minimal thickness and material usage.