Three-Dimensional Stability Analysis of Tailings Pond Based on PSO-Kriging Model

In the process of tailings accumulation, multiple factors lead to obvious spatial differences and uneven distribution of the physical characteristics of tailings. However, a large number of physical experiments are required to study these characteristics, and although these experiments can be repeated, there are systematic errors and high costs, so an approximate model can be constructed for mechanical learning to predict how to solve this problem and become the key to the stability analysis of tailings ponds. In this paper, we start by constructing an improved Kriging model, and determine the actual characteristics of geological parameters through actual tailings pond sampling. Firstly, the physical and mechanical parameters that have a great influence on the stability of tailings ponds, natural bulk density, cohesion, and internal friction angle are selected as interpolation characteristic parameters. By comparing different optimization methods, an improved PSO-Kriging interpolation model was proposed from the perspective of spatial variability. According to the cross-validation between the improved model and the original model, the prediction discrete plots and correlation coefficient values of the PSO-Kriging model and the traditional Kriging model were analyzed. The results show that compared with the traditional model, the correlation coefficient values of the new model on the three tailings parameters are increased by 0.035, 0.151, and 0.117, respectively, indicating that the new model not only improves the accuracy of prediction, but also better captures the prediction trend. On the basis of the construction of the PSO-Kriging interpolation model and the reliability verification of the model, a set of characteristic mechanical parameters of the interpolation points in line with the actual situation of the tailings pond were obtained, the mechanical parameters of the tailings were interpolated, and the three-dimensional numerical model of the tailings pond was established according to the interpolation results, and the seepage characteristics and stability of the dam body under different working conditions were studied.