| Tailings ponds are a major danger source with high potential energy,and there are a large number of them in my country.The risk assessment of tailings ponds is of great significance to mine safety management.This paper took the tailings pond as the research object,selects the evaluation indexes of the tailings pond based on the literature survey method,the accident statistics method and the fishbone diagram method,and constructs the evaluation index system of the tailings pond.Then,two risk assessment models were constructed for the risk assessment of tailings ponds,namely the Genetic Algorithm-Extreme Gradient Boosting(GA-XGB)model and the Analytic Hierarchy Process-Fuzzy Comprehensive Evaluation(AHP-FCE)model.The main research work and results are as follows:(1)Establishment of tailings pond safety evaluation index system.Through the statistical analysis of tailings pond accidents,the fishbone diagram analysis of the causes of the accidents,and the consideration of tailings pond document policies,the key factors affecting the safety of tailings ponds were selected based on the index selection principle,and five first-level indicators were extracted.And 20 secondary indicators,and finally built a tailings pond risk assessment system.(2)Establishment of risk assessment model based on GA-XGB tailings pond.The coupled model constructed by optimizing XGB by GA had the advantages of fast convergence,good generalization and strong robustness.This paper attempted to introduce the GA-XGB algorithm model into the field of tailings pond risk assessment.Based on the methods of ROC model evaluation and confusion matrix model evaluation,the GAXGB model had better performance compared with Bayes,SVM,Tree,and XGB algorithms.Well,the AUC was the highest value of 0.915,and the prediction accuracy was improved by 2.51%.(3)Establishment of risk assessment model based on AHP-FCE tailings pond.The model can not only reflect the ambiguity of the evaluation process,but also quantified the subjective experience,and had strong logic and practicability.The coupling model adopt the tailings pond safety evaluation index system,and the Axi gold mine tailings pond was used as a sample to pass the feasibility verification,and then it was applied to the Xiaomaliu tailings pond.Based on the evaluation results of Xiaomaliu tailings pond and with reference to the actual operation,some safety management suggestions were put forward.The results showed that GA-XGB algorithm model and AHP-FCE model were feasible,which can accurately predict the safety level of tailings pond and provide reference and basis for the safety management of tailings pond. |