摘要
为了在矿井瓦斯爆炸灾变发生后,快速确定瓦斯爆炸冲击波的压力、温度、有毒有害气体等致灾因子在井巷网络中的传播情况。利用CFD数值模拟或爆炸实验获得瓦斯爆炸冲击波的压力、温度、有毒有害气体等致灾因子传播大数据,将影响瓦斯爆炸传播的因素以及观测点等参数作为人工神经网络的输入节点,压力、温度等致灾因子作为输出节点,建立瓦斯爆炸致灾因子传播快速预测机器学习模型,解决CFD数值模拟的建模、计算及数据分析处理等过程耗时大、不适应灾变应急的快速响应等问题。研究结果表明:在给定爆炸位置和爆炸当量的均直巷道,获得任一点的爆炸冲击波压力、温度以及有毒有害气体所需时间是瞬时的,人工神经网络平均训练误差为6.92%,有训练样本的验证误差为5.24%,无训练样本的验证误差为6.88%。
In order to quickly determine the propagation of hazard factors such as the pressure,temperature,and toxic and harmful gases of gas explosion shocking wave in the mine roadway network after the occurrence of mine gas explosion,the big data about the propagation of hazard factors such as the pressure,temperature,and toxic and harmful gases of gas explosion shocking wave were obtained by using CFD numerical simulation or explosion experiments.The factors affecting the propagation of gas explosion and the parameters such as observation points were taken as the input nodes of artificial neural network(ANN),and the hazard factors such as pressure and temperature were taken as the output nodes,then a machine learning model for the rapid prediction on the propagation of gas explosion hazard factors was established,so as to solve the problems of large time-consuming and unsuitable for rapid response to catastrophe emergency in the modeling,calculation,and data analysis and processing processes of CFD numerical simulation.The results showed that at the uniform roadway with given explosion location and explosion equivalent,the time required to obtain the pressure,temperature,and toxic and harmful gas of explosion shock wave at any point was instantaneous.The average training error of ANN was 6.92%,the validation error of sample with training was 5.24%,and the validation error of sample without training was 6.88%.
作者
刘剑
曲敏
黄德
高科
邓立军
刘学
LIU Jian;QU Min;HUANG De;GAO Ke;DENG Lijun;LIU Xue(College of Safety Science and Engineering,Liaoning Technical University,Huludao Liaoning 125105,China;Key Laboratory of Mine Thermo-motive Disaster and Prevention,Ministry of Education,Liaoning Technical University,Huludao Liaoning 125105,China)
出处
《中国安全生产科学技术》
CAS
CSCD
北大核心
2020年第8期11-17,共7页
Journal of Safety Science and Technology
基金
国家重点研发计划项目(2017YFC0804401)
国家自然科学基金项目(51774169,51574142)。
关键词
瓦斯爆炸
致灾因子
人工神经网络
分类器
应急决策
gas explosion
hazard factor
artificial neural network(ANN)
classifier
emergency decision