Prediction analysis of slope stability due to soft and weak interlayers based on partial least squares method

Abstract In this paper, based on sampling and analysis of a large number soft weak sandwich slope data, several factors that have great influence stability are established, predictive model describing the slopes is established by using now more advanced partial least squares method. Then, for traditional method not suitable non-linear coefficient to prediction, recursive with forgetting factor proposed prediction solve problem lag. Finally, elastic mechanics, fracture mechanics unsaturated soil structure interlayers their strength studied, performance MATLAB-based verified designing orthogonal experiments. The results show predicted values do differ much from finite element calculation, absolute errors all less than 0.15, there 5 non-differences 0.1, accounting 62.5% total groups. relative were 6%. This study shows can deal nonlinear mapping relationship between influencing well make accurate objective slopes.