Application of Artificial Neural Network in predict the stability number of unsupported conical slope in anisotropic and inhomogeneous clays, Journ

Application of Artificial Neural Network in predict the stability number of unsupported conical slope in anisotropic and inhomogeneous clays, Journal of Science of Lac Hong University, 2022

Nguyen Dang Khoa

Faculty of Civil Engineering, Lac Hong University

Abstract: The paper proposed a correlation equation to determine the stability number of unsupported conical slopes on anisotropic and non-homogenous clays using Machine Learning applications based on the Artificial Neural Network model - ANN. The stability number (N) of unsupported conical slopes on anisotropic and non-homogenous clays was investigated by considering the effects of raising undrained shear strength (m), anisotropic strength ratio (re), the dip angle of slopes (β), and the ratio of slope height and radius at the bottom (H/B). The numerical method based on the finite element technics was applied to investigate N variation due to changes of re, m, H/B and β. The results of numerical models are used as the data for the Artificial Neural Network model to propose the correlation equation between N and input parameters of re, m, H/B and β. The results show that, the proposed correlation equation is well in agreement with results of N from numerical models.

Keywords: Unsupported Conical Slopes, Anisotropic Clays, ANN model, Artifiial Intelligent, Machine Learning.