基于SOM的易燃液体运输安全风险预警

摘要

In order to reduce the risk of railway dangerous goods transportation and the probability of accident for improving the quality of transportation management,a new method is proposed for the risk early warning of rail way dangerous goods transportation safety based on the self-organizing competitive neural network (SOM) technology.Through Hazard and Operability Analysis (HAZOP) of dangerous goods transport process,a risk early warning system of railway flammable liquid transportation safety is established,and the statistical analysis algorithm is used to assess the correlation between the factors affecting the safety risk of each transport link.Combined with the actual situation of freight transportation in Taiyuan Bureau,a safety risk early warning model of railway flammable liquid transport is established based on SOM and simulated by sample data and test data.The results show that the model can play a good role in early warning during the transportation of dangerous goods,which can provide reference for the safety early warning technology of railway dangerous goods transport.%为了降低铁路易燃液体运输风险和事故概率,提高运输管理质量,提出一种基于自组织竞争神经网络(SOM)技术的铁路易燃液体运输安全风险预警方法.首先,通过对易燃液体运输过程的危险与可操作性识别分析,建立铁路易燃液体运输过程安全风险预警体系;然后,采用统计分析算法评定各运输环节安全风险影响因素之间的相关程度;最后,结合太原局货运实际,构建了基于SOM的铁路易燃液体运输安全风险预警模型,并采取样本数据和测试数据进行了模拟验证.研究结果表明,该模型在易燃液体运输中能够较好起到预警作用,可为铁路危险货物运输安全预警技术提供参考.