Reliability Analysis Of Soil Slope Stability Using Ann, Anfis, Pso-Ann Soft Computing Techniques

Abstract

Slope stability is an important concern while constructing any slope or constructing any structure on a soil slope. As soil is heterogeneous in nature due to the process of its formation, slope stability analysis cannot be done without considering the variability in the properties of soil. To consider soil variability, with time the research approach is shifting towards probabilistic approach. In the present paper to study soil slope stability, three soft computing techniques were used: Artificial Neural Network (ANN), Adaptive Network based Fuzzy Inference System (ANFIS), and Particle Swarm Optimization based Artificial Neural Network (PSO-ANN). Because three soil parameters, unit weight (ϒ), cohesion (c) and angle of shear resistance (ϕ), mostly determine the stability of a soil slope, they were employed as input variables for the models, and the factor of safety was used as an output. Models were examined using statistical criteria such as NS, RPD, RMSE, R2, PI, GPI to assess performance. The output of the result showed that although all the models performed well, however, the model PSO-ANN out-performed among the above three models. As a result, PSO-ANN can be utilized to analyze multi-layered soil embankment slope stability as a robust soft computing technique.