Effective Risk Prediction of Tailings Ponds Using Machine Learning

A tailings pond is a place for storing industrial waste, which is a major hazard source with high potential energy. The stability of the tailings dam is usually evaluated by saturation line height. The measuring sensors for saturation line are very expensive and have poor lightning protection abilities. To solve this problem, this paper takes Jiande tailings pond as the study area. Machine learning models are built after integrating various sensor monitoring data, relying on two sensors to accurately predict the saturation line height to predict the safety of tailings pond. The average R2 of real-time regression of saturation line is 99.16%, and average accuracy of real-time warning is 99.89%.