基于PSO-SVM的矿井输送带火灾危险程度预测模型研究  

Prediction Model of Mine Tape Fire Hazard Based on PSO-SVM

作　　者：王伟峰 杨博[1,2] 甘梅 任立峰[1,2] 康付如 刘韩飞 WANG Weifeng;YANG Bo;GAN Mei;REN Lifeng;KANG Furu;LIUHanfei(College of Safety Science and Engineering,Xi′an University of Science and Technology,Xi′an 710054,China;Key Laboratory of Coal Fire Disaster Prevention and Control in Shaanxi Province,Xi′an University of Science and Technology,Xi′an 710054,China;China Coal Technology and Engineering Group Chongqing Research Institute Limited,Chongqing 400000,China)

机构地区：[1]西安科技大学安全科学与工程学院,西安710054 [2]西安科技大学陕西省煤火灾害防治重点实验室,西安710054 [3]中煤科工集团重庆研究院有限公司,重庆400000

出　　处：《煤炭技术》2024年第5期212-216,共5页Coal Technology

基　　金：陕西省重点研发计划(2021SF-472,2022QCY-LL-70);陕西省秦创原“科学家+工程师”队伍建设项目(2023KXJ-052)。

摘　　要：矿井带式输送机输送带火灾严重威胁煤矿安全生产。矿井输送带火灾风险监测预警方面存在指标内在关联性差、火灾不同发展阶段的标志性气体和特征温度不明确、火灾风险预警指标缺失等难题。因此,采用热重红外联用仪和锥型量热仪装置,研究了矿井输送带火灾热解燃烧的特性,热解初始温度段,首先生成较多的CO_(2)、H2O以及少量CO、HCl,温度升至252℃时,HCl生成量开始迅速增加,298℃时HCl的生成量达到峰值1.26%;CO在416℃时生成量开始迅速增加,485℃时CO的生成量达到峰值0.29%;CO_(2)在整个热解过程中产生量最大,总结出CO、CO_(2)、HCl作为输送带火灾监测预警指标。建立了基于粒子群算法优化支持向量机的模型,并对模型进行了最优参数与效果的研究分析。为了验证模型的准确性和可靠性,采用最小二乘误差指标将PSOSVM与SVM预测结果进行对比。The coal mine belt conveyor belt fire seriously threatens the safety of coal mine production.There are problems in the monitoring and warning of coal mine belt fire risks,such as poor internal correlation of indicators,unclear landmark gases and characteristic temperatures at different stages of fire development,and missing fire risk warning indicators.Therefore,the characteristics of coal mine belt fire pyrolysis and combustion were studied using a thermogravimetric infrared spectrometer and a cone calorimeter device.In the initial temperature range of pyrolysis,a large amount of CO_(2),H2O,and a small amount of CO and HCl were first generated.When the temperature rose to 252℃,the amount of HCl generated began to rapidly increase,and at 298℃,the peak amount of HCl generated reached 1.26%;The production of CO began to rapidly increase at 416℃,and reached a peak of 0.29%at 485℃;CO_(2) is generated the most during the entire pyrolysis process,and CO,CO_(2),and HCl are summarized as indicators for monitoring and warning of tape fires.A model based on particle swarm optimization for support vector machine optimization was established,and the optimal parameters and effectiveness of the model were studied and analyzed.In order to verify the accuracy and reliability of the model,the least squares error index was used to compare the prediction results of PSO-SVM and SVM.

关 键 词：输送带火灾 支持向量机 监测预警 粒子群算法 

分 类 号：TD75[矿业工程—矿井通风与安全]