基于SCG-BP神经网络的岩爆预测模型  

Rockburst Prediction Model Based on SCG-BPNN

作　　者：张亭[1] 陈佳 高志荣[1] 任育荣 周义升 罗宏伟 ZHANG Ting;CHEN Jia;GAO Zhirong;REN Yurong;ZHOU Yisheng;LUO Hongwei(Northwest Research Institute of Mining and Metallurgy;Sinosteel Maanshan General Institute of Mining Research Co.,Ltd.;Deep Mining Company,Baiyin Nonferrous Group Co.,Ltd.)

机构地区：[1]西北矿冶研究院 [2]中钢集团马鞍山矿山研究总院股份有限公司 [3]白银有色集团股份有限公司深部矿业公司

出　　处：《现代矿业》2023年第6期159-162,共4页Modern Mining

摘　　要：标准的BP神经网络在岩爆预测中表现较差,选用量化共轭梯度法优化BP神经网络进行岩爆预测分类研究。选取应力系数、脆性系数和弹性能量指数为预测指标,以46组工程案例作为数据库,所建立的SCG-BP神经网络预测准确率达80.43%,远高于优化前的54.05%。对模型训练集与测试集的分类误差和分类结果进行可视化,并与标准BP神经网络的预测结果进行对比分析,结果表明优化效果良好。通过在实际工程中应用,表明该岩爆预测模型具有推广使用价值。The standard BP neural network performs poorly in rockburst prediction.The quantitative conjugate gradient method is used to optimize the BP neural network for rockburst prediction classification.The stress coefficient,brittleness coefficient and elastic energy index are selected as the prediction indexes,and 46 groups of engineering cases are used as the database.The prediction accuracy of SCG-BP neural network is 80.43%,which is much higher than 54.05%before optimization.The classification error and classification results of the model training set and the test set are visualized,and compared with the prediction results of the standard BP neural network.The results show that the optimization effect is good.Through the application in practical engineering,it shows that the rockburst prediction model has the value of popularization and application.

关 键 词：岩爆预测 BP神经网络 量化共轭梯度法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TD31[自动化与计算机技术—控制科学与工程]