Probabilistic Slope Stability Analysis on a Heavy-Duty Freight Corridor Using a Soft Computing Technique

This study overviews how to strengthen railway embankments using a geo-grid in the cohesive soil embankment layer and the Bishop method to determine a safety factor using Geo-Studio software. The primary goal of this study is to show the relationship between an embankment’s safety factor with and without a geo-grid. These parameters, such as angle of internal friction, cohesion value, and unit weight for both the subsoil and embankment layer, respectively, pull-out resistance, and tensile capacity for geo-grid, have been used as input in this test. The safety factor has increased continually after altering its features and incorporating a geo-grid into the embankment layer. Based on the Geo-Studio results, the ideal choice was to strengthen the railway embankment by adding a geo-grid to the embankment layer and employing a reliable computational technique to analyse the corridor’s probabilistic slope stability for heavy-duty freight trains. The current method, which has been utilised to undertake a probabilistic study of a high embankment of 12.29 m taken by the Ministry of Indian Railways for a heavy-haul freight corridor, consists of four model analyses: convolutional neural networks (CNN), deep neural networks (DNN), artificial neural networks (ANN), and multiple linear regression (MLR). Performance indicators assessed the models’ performance, such as R 2 , RMSE, RSR, WI, MAE, NS, and PI. According to the analysis of the results, the CNN model outperformed DNN, ANN, and MLR. CNN is, therefore, a trustworthy soft computing technique for determining the safety of a railway embankment slope.