A new soft-computing approach for stability evaluation of inclined circular anchor in natural clays

A novel approach based on the coupling of Finite Element Limit Analysis (FELA) and Artificial Neural Network (ANN) is adopted to investigate the effect of the inclination angle and the natural clay properties on the uplift resistance of inclined circular anchors embedded in anisotropic and non-homogenous clays. Totally 1296 3D FELA models with the Anisotropic Undrained Shear (AUS) model are implemented, and an optimal ANN model is selected for investigating the effects of four design parameters on uplift factor (F_c), including inclined angle (θ), embedment ratio (H/D), anisotropic ratio (r_e), and non-homogenous ratio (m). Specifically, the uplift resistance of inclined circular anchors is strongly affected by H/D, followed by m, θ, and r_e, with importance indexes of 44.1%, 37.9%, 15.0%, and 3.0%, respectively. For practical application, this study proposes a correlation equation between four design parameters and uplift factor (F_c) with high accuracy, which can be helpful in quickly estimating the uplift resistance. Additionally, the design charts that show detailed relationships between F_c and investigated parameters (θ, m, H/D, and r_e) and the failure mechanism of the inclined anchors are also illustrated in this research. These investigations can significantly enhance the understanding of the performance of inclined circular anchors embedded in anisotropic and non-homogenous clays.