Analysis of Influencing Parameters of the Improved Model for Rainfall Infiltration in Unsaturated Tailings Soil

Abstract

The rainfall infiltration analysis method is an important method for slope stability forecast and prevention. Slope angle and unsaturated soil layers are not shown in the conventional Green-Ampt (GA) infiltration model. In this paper, based on the GA model of rainfall infiltration in tailings slopes, two aspects of slope angle and the proportion of the transition layer of the wet layer are modified. The rainfall infiltration test of unsaturated tailing soil was conducted using a self-developed large-size tailing slope model test device. The results of GA model, improved GA model, and Richards’ equation calculation model are compared. The results show that the difference between the three models is small in the free infiltration stage, but the infiltration rate is lower than that of the GA model. With the gradual increase of rainfall time into the ponding infiltration stage, the expansion depth and infiltration rate of the wetting front of the improved GA model and Richards’ equation are greater than those of the GA model. The difference between the GA model and the improved GA model and Richards’ equation for the extended depth of wetting front increases with the increase of rainfall duration, while the difference of infiltration rate changes in the opposite trend. The results of the improved GA model and Richards’ equation to calculate the expansion depth and infiltration rate of wetting fronts are consistent, and the difference basically tends to stabilize with the increase of rainfall ephemeris. The results of the improved GA model are closer to the measured data, which can provide a reference for analyzing the rainfall infiltration pattern of open pit tailing dams and slope stability research. On the basis of the improved GA model, the influence of slope angle, rain intensity, initial water content, and saturation infiltration coefficient on rainfall infiltration was analyzed, and the analysis of parameter sensitivity indexes showed that slope inclination and initial water content had a greater influence on the model.