作者简介:姜进成(1977-),男,高级工程师,研究方向为数字化转型。E-mail:jiangjch@shandong-energy.com
DOI: 10.13512/j.hndz.2025.04.24
备注
针对对矿山潜在安全隐患发现不及时导致的应急响应滞后、风险和损失增加等问题,提出基于BES-RBF神经网络的矿山重大灾害风险预警方法,采用核主成分分析(KernelPrincipalComponentAnalysis,KPCA)在复杂的矿山数据中提取出与矿山灾害相关的特征,选取径向基函数(RadialBasisFunction,RBF)神经网络作为矿山重大灾害风险预警模型,并通过秃鹰搜索(BaldEagleSearch,BES)算法优化网络参数值,搭建最优网络模型,将提取到的特征作为训练后网络的输入值,输出矿山灾害相对风险值,划分相对风险值为不同灾害风险等级,实现矿山重大灾害风险预警。实验结果表明,所提方法KS值较高且CPU占用率和内存占用率较低。
To address the issues of delayed emergency response and increased risks and losses due to untimely identification of potential safety hazards in mines,this paper proposes a major mine disaster risk warning method based on a BES-RBF neural network. Kernel Principal Component Analysis(KPCA)is employed to extract features associated with mine disasters from complex mine data. The Radial Basis Function(RBF)neural network is adopted as the warning model for major mine disasters. The Bald Eagle Search(BES)algorithm is utilized to optimize the network parameters,thereby establishing an optimal network model. The extracted features are then used as inputs to the trained network,which outputs the relative risk value of mine disasters. This value is classified into different risk levels to achieve major mine disaster risk warning. Experimental results demonstrate that the proposed method achieves a high KS value while maintaining low CPU and memory occupancy.

采用ξ= 描述主成分贡献,ξ与主成分分量呈正相关。
=0,k=1, 2,⋯, M,则其协方差矩阵D如下所示:
,i=1, 2,⋯, M,由此可得:将(ψ(xj) ψ(xi))表示为Kij,j=1,2,⋯, M,并采用全部Kij构建M×M维矩阵K,由此可将上式表示为下式形式:



表示上次更新后平均分布位置,。










