摘要
为及时掌握尾矿库风险状态,降低尾矿库溃坝造成的人员伤亡、财产损失和环境污染,建立尾矿库风险评价模型具有十分重要的现实意义。首先,构建尾矿库风险等级评价指标体系,确定指标分级标准;其次,利用博弈论将层次分析法和熵值法结合起来确定指标组合权重;再次,利用rand()函数生成伪随机数作为模型训练数据库,将权重与训练数据对应组合,建立物元可拓模型计算尾矿库所属风险等级;最后构建BWOA-SVM(改进鲸鱼算法(BWOA)优化的支持向量机(SVM)模型)模型对风险等级预测,对该模型进行训练,得到模型预测准确率为98%,与传统SVM相比提升了44.9%。采用山西东沟尾矿库数据验证模型的可行性,将尾矿库数据输入训练好的模型中得到该尾矿库等级为Ⅱ级,与实际情况相同,验证了所提方法的可行性,通过该方法可以确定尾矿库的风险等级,从而为尾矿库风险分级管理提供依据。
In order to grasp the risk status of tailings pond timely and reduce the casualties,property losses and environmental pollution caused by tailings pond dam break,it is of great practical significance to establish a tailings pond risk assessment model.Firstly,the risk grade evaluation index system of tailings pond is constructed,and the index grading standard is determined.Secondly,game theory is used to combine the analytic hierarchy process and entropy method to determine the index combination weight.Once more,rand()function is used to generate pseudorandom number as the model training database.Weight and training data are combined to establish matter-element extension model to calculate the risk level of tailings pond.Finally,a BWOA-SVM(improved whale algorithm(BWOA)optimized support vector machine(SVM)model)model was constructed to predict the risk level.The model was trained,and the prediction accuracy of the model was 98%,which was 44.9%higher than that of the traditional SVM.The feasibility of the model is verified by the data of a tailings pond in Shanxi Province.The data of the tailings pond is input into the trained model,and the grade of the tailings pond isⅡ,which is the same with the actual situation,which verifies the feasibility of the proposed method.The risk grade of the tailings pond can be determined by this method,and provides a basis for the risk classification management of the tailings pond.
作者
荀曦
郑欣
于雁武
许开立
XUN Xi;ZHENG Xin;YU Yanwu;XU Kaili(School of Resources and Civil Engineering,Northeastern University,Shenyang 110819,China;School of Environmental and Safety Engineering,North Central University,Taiyuan 030051,China)
出处
《金属矿山》
CAS
北大核心
2023年第12期211-219,共9页
Metal Mine
基金
“十四五”国家重点研发计划项目(编号:2021YFC3001300)。
关键词
尾矿库
风险评价
物元可拓
伪随机数
BWOA-SVM
tailings reservoir
risk assessment
matter element extension
pseudo-random number
BWOA-SVM
