摘要
基于层次分析法(analytic hierarchy process,AHP)研究地铁接触网系统的致灾机理,建立AHP模型进行风险辨识与权重分析,得到AHP模型指标层各元素的权重系数;基于风险矩阵法(risk matrix method,RMM)提出综合考虑风险因素的后果严重性与发生可能性的风险量化评价体系,得到指标层各元素的风险等级隶属度和评价值。最后,采用BP(back propagation)神经网络方法对评价结果进行训练,验证可靠性评价结果的有效性。结果表明:当采取E2组评价权重,即{0.1,0.2,0.3,0.4}时,BP神经网络训练结果与评价结果的重合度最高,评价效果最佳。
Based on the analytic hierarchy process(AHP),the disaster-causing mechanism of subway catenary system was studied.The AHP model was established for risk identification and weight analysis,and the weight coefficients of each element in the index layer of the AHP model were obtained.Based on the risk matrix method(RMM),a risk quantitative evaluation system was proposed,which comprehensively considers the severity and probability of the consequences of risk factors.The risk grade membership and evaluation value of each element in the index layer were obtained.Finally,the back propagation(BP)neural network was used to train the evaluation results and verify the effectiveness of the risk evaluation matrix.The results show that when the evaluation weights of E2 group are used,i.e.{0.1,0.2,0.3,0.4},the coincidence degree between BP neural network training results and evaluation results is the highest,and the evaluation effect is the best.
作者
许勇
XU Yong(China Railway Siyuan Survey and Design Group Co.,Ltd.,Wuhan 430063,China)
出处
《武汉大学学报（工学版）》
CAS
CSCD
北大核心
2021年第4期375-380,共6页
Engineering Journal of Wuhan University
基金
国家重点研发计划项目(编号:2018YFB2100900)。
关键词
地铁接触网系统
可靠性分析
层次分析法
风险矩阵法
subway catenary system
reliability study
analytic hierarchy process
risk matrix method
