基于反赋权与MBCT-SR多维云模型算法岩爆预测研究  

Research on rockburst prediction based on anti-weighting and MBCT-SR multi-dimensional cloud model algorithm

作　　者：宋英华[1,2] 庞昭胜 李墨潇[1,2] 江晨[1,2] 齐石 SONG Yinghua;PANG Zhaosheng;LI Moxiao;JIANG Chen;QI Shi(China Emergency Management Research Center,Wuhan University of Technology,Wuhan Hubei 430070,China;School of Safety Science and Emergency Management,Wuhan University of Technology,Wuhan Hubei 430070,China)

机构地区：[1]武汉理工大学中国应急管理研究中心,湖北武汉430070 [2]武汉理工大学安全科学与应急管理学院,湖北武汉430070

出　　处：《中国安全生产科学技术》2022年第3期39-46,共8页Journal of Safety Science and Technology

基　　金：中央高校基本科研业务费专项资金项目(2020VI003,2021Ⅲ052JC,2021Ⅲ053JC)。

摘　　要：基于岩爆事故的模糊性与随机性特征,为解决在当前岩爆烈度等级预测研究中,通过正向云发生器经验式计算多维云模型特征值进而导致预测结果主观性较强的问题。本文选取应力比T_(s)=σ_(θ)/σ_(t)、岩石脆性指数B=σ_(c)/σ_(t)以及弹性应变储能指数W_(et)作为评价指标,结合机器学习理论,采用多维逆向MBCT-SR云发生器算法对岩爆等级进行预测,通过实例数据计算数字特征,建立多维云模型和动态适应度函数,最后选取优化算法反求最优权重。选取国内外192组岩爆实例数据对建立的模型预测结果进行验证,并与反赋权一维云模型、常规逆向云发生器多维云模型预测结果进行对比。研究结果表明:本文模型对岩爆烈度分级预测的精准率可达89%,岩爆倾向性预测准确率可达到100%,相对于其他模型具有更高的准确性,可为岩爆预测提供更加科学有效、切合实际的评价模型。Based on the fuzziness and randomness characteristics of rockburst accidents,in order to solve the problem of strong subjectivity of prediction results by calculating the eigenvalues of multi-dimensional cloud model through the empirical formula of forward cloud generator in the current research on rock burst intensity prediction.In this paper,the stress ratio T_(s)=σ_(θ)/σ_(t),the rock brittleness index B=σ_(c)/σ_(t),and the elastic strain energy storage index W_(et) were selected as the evaluation indexes,and combined with the machine learning theory,the multi-dimensional reverse cloud generator algorithm was used,and the MBCT-SR algorithm was also used to study the grade prediction of rockburst.The digital characteristics were calculated through the example data,then the multi-dimensional cloud model and the dynamic fitness function were established,and finally the optimization algorithm was selected to reversely solve the optimal weight.192 sets of rockburst case data at home and abroad were selected to verify the prediction results of the established model,and the prediction results were compared with those of the anti-weighting one-dimensional cloud model and the conventional reverse cloud generator multi-dimensional cloud model.The results showed that the accuracy of this model for the classification prediction of rockburst intensity could reach 89%,and the accuracy of rockburst tendency prediction could reach 100%.Compared with other models,it had higher accuracy and provides a more scientific,effective and realistic evaluation model for the rockburst prediction.

关 键 词：岩爆预测 反分析权重 逆向云发生器 多维云模型 

分 类 号：TU457[建筑科学—岩土工程] X936[建筑科学—土工工程]