摘要
分析地铁火灾事故发生的危险因素,归纳出了人员、设备、管理、环境21个小类的主要影响因素,并通过神经网络建立起地铁火灾风险评价模型。为克服神经网络易陷入局部最小的缺陷,通过遗传算法对神经网络模型中得到的权值和阈值进行优化。训练结果表明,最终建立了一种地铁火灾风险评价有效可靠的方法。
Risk factors of subway fire accident were analyzed, 21 main factors about people, equipment, management, environ- ment were introduced, and subway fire risk assessment model was built by neural network. In order to overcome the problem of neural net being easy to fall into the local minimum, the weights and thresholds of the neural network model were Optimized by genetic algorithm. The training result showed that the evaluation method of subway fire risk is effective and reliable.
出处
《消防科学与技术》
CAS
北大核心
2016年第6期847-849,共3页
Fire Science and Technology
关键词
风险评价
神经网络
遗传算法
地铁火灾
risk evaluation
neural network
genetic algorithm
subway fire
