Multiple Regression Models for Predicting Stability of Reinforced Soil Slope

Slope failures are prevalent issue in the construction sector. Thus the engineers must use appropriate slope stabilization techniques to reduce the risk of human life and property. This work investigates the efficacy of multiple regression analysis in predicting slope stability, specifically focusing on the slopes in the Kullu district, Himachal Pradesh, India. A total of ۱۶۰ cases with different parameters were analyzed by using the well-known Limit Equilibrium Method (LEM), Morgenstern and Price on PLAXIS LE. Numerical analysis was performed using different nail lengths (۶ m, ۸ m, ۱۰ m, and ۱۲ m) and nail inclinations (۰°, ۵°, ۱۰°, ۱۵°, ۲۰°, ۲۵°, ۳۰°, and ۳۵°), applied to a homogeneous soil slope with ۴۵°, ۵۰°, ۶۰°, and ۷۰° inclinations, respectively. The limit equilibrium analysis may not offer predictive capabilities for future scenarios directly. In contrast, Multiple Regressions (MR) can provide predictive insights based on the historical data, allowing for forecasting of stability under different conditions or design scenarios. The utilization of MR provides the coefficients that quantify the influence of each variable on slope stability, enabling a detailed understanding of how each factor contributes. To develop the prediction models using Multiple Regression Analysis (MRA), the factor of safety values obtained by the numerical method were used. The accuracy of this model was evaluated against the conventional LE methods. The results indicate that multiple regression provides a good predictive performance with an R۲ value equal to ۰.۷۷۴, offering a more nuanced and accurate assessment of slope stability compared to the traditional LE techniques.

کلیدواژه ها:

PLAXIS LE ، limit equilibrium method ، Factor of Safety ، Multiple Regression Analysis