摘要
提出一种基于遗传算法优化的BP神经网络(GA-BP)的园区应急能力可靠性分析模型。通过故障树分析对化工园区应急救援能力的可靠性进行量化并作为GA-BP神经网络的输出值,以故障树中的基本事件为依据,进行分类总结,建立化工园区应急救援能力层次分析指标体系;从日常生产状态下的应急系统维护与事故时的应急处置能力这两个准则层考虑,指标层元素分别对应故障树的基本事件,并计算其相对于目标层的复合权重;从样本组中,选取训练样本和测试样本验证基于遗传算法优化的神经网络可靠性分析模型的可行性,并与传统BP神经网络的分析数据进行对比;结果表明,经过遗传算法优化后的BP神经网络平均误差和均方误差均较小。得出GA-BP神经网络的分析结果与故障树更加接近,而且相对于故障树分析减少了复杂的建树过程,具有更高的易用性。
A reliability analysis model of campus emergency capability based on BP neural network（ GA-BP） optimized by genetic algorithm is proposed.Through the fault tree analysis,the reliability of the emergency rescue capability of the chemical park is quantified and the output value of the GA-BP neural network is used. Based on the basic events in the fault tree,the classification of the emergency rescue capability of the chemical park is established. From the daily production status of the emergency system maintenance and emergency response capacity of the two criteria layer to consider,the indicator layer elements corresponding to the fault tree of the basic events,and calculate its relative to the target layer of the composite weight; from the sample group,The training samples and the test samples are validated by neural network reliability analysis model based on genetic algorithm optimization,and compared with the analysis data of traditional BP neural network. The results show that the BP neural network average error after genetic algorithm optimi-zation Mean square error is small. Therefore,the GA-BP neural network analysis results are closer to the fault tree,and the analysis of the fault tree reduces the complexity of the process,with higher ease of use.
出处
《安全、健康和环境》
2017年第10期41-43,56,共4页
Safety Health & Environment
关键词
化学工业园区
应急救援能力
可靠性分析
遗传算法
BP神经网络
chemical industry pa rk
emergency res-cue ability
reliability analysis
genetic algorithm
BP neural network
