文档摘要:为准确判断电热法电石生产工艺中电石炉的爆炸风险等级,提出了一种精准有效的风险评估模型.首先,基于危险与可操作性(HazardandOperability,HAZOP)分析筛选出人、物料、设备、管理四方面的34项爆炸风险因素,考虑到因素间存在非线性关联,采用核主元分析(KernelPrincipalComponentAnalysis,KPCA)进行属性约简,减少冗杂信息的干扰.其次,利用融合了Tent混沌序列、高斯变异与混沌扰动的麻雀搜索算法(ImprovedSparrowSearchAlgorithm,ISSA)寻优核极限学习机(KernelExtremeLearningMachine,KELM)的惩罚系数与核参数,建立KPCA-ISSA-KELM风险评估模型.最后,使用该模型分析83组实例数据,选取其中59组用于模型训练,其余24组用于测试.在测试结果中,该模型正确分类了22组数据的风险等级,判别准确率为91.67%,在各项性能指标上均优于对照模型,表明该模型对电热法工艺电石炉的爆炸风险等级具备高识别精度.
Abstract:Electrothermalcalciumcarbideproductionisacomplexprocessinvolvingmultiplefactors.Asthemainproductionequipment,sealing-typecalciumcarbidefurnacecarriesarelativelyhighexplosionriskthatisdifficulttoquantify.Toaccuratelydeterminetheexplosionrisklevelofcalciumcarbidefurnacesandpreventrelatedaccidents,thispaperproposedanaccurateandeffectiveriskassessmentmodel.First,basedonHazardandOperability(HAZOP)analysisandaccident-causingtheories,34riskfactorsrelatedtocalciumcarbidefurnaceexplosioncoveringpersonnel,materials,equipment,andmanagementwereselected.Duetothecomplexcorrelationbetweenfactors,thispaperintroducedKernelPrincipalComponentAnalysis(KPCA)toperformnonlinearfeatureextractionanddimensionalityreductiononthefactors,achievingreconstructionofthefactorsystemandreducinginterferencefromredundantinformation.Afterthat,incorporatingtentchaoticsequence,gaussianmutation,andchaoticperturbation,theSparrowSearchAlgorithm(SSA)wasimproved.Theresultofthebenchmarkfunctionoptimizationtestshowsthat:theImprovedSparrowSearchAlgorithm(ISSA)issuperiortoSSAandParticleSwarmOptimization(PSO)intermsofoptimizationaccuracyandconvergencespeed.Subsequently,ISSAwasusedtooptimizethepenaltycoefficientCandkernelparametersoftheKernelExtremeLearningMachine(KELM),andtheKPCA-ISSA-KELMriskassessmentmodelwasestablished.83setsoforiginaldatawereobtainedfromseveralenterprises,includingsafetyinspectionresultsandcorrespondingdiscriminationfromexpertsontherisklevelofcalciumcarbidefurnaceexplosion.Amongthem,59setswereusedfortrainingthemodel,andtheremaining24setswereusedfortesting.ProgrammingofmodelswascarriedouthavingtheaidofMATLABsoftware.ThenonlinearfeatureextractionresultshowsthatKPCAobtained11principalcomponentsthatcontained85.86%informationoftheoriginaldata.Moreover,theriskpredictionresultof24testsetsamplesshowsthat:KPCA-ISSA-KELMhasariskidentificationaccuracyof91.67%,thismodelissuperiortoKELM,SSA-KELM,andISSA-KELMintermsofoverallaccuracyandtargeteddiscriminationofrisksateachlevel.Therefore,itisconcludedthattheestablishedmodelcaneffectivelypredicttheexplosionriskoftheelectrothermalprocesscalciumcarbidefurnace,providingareferenceforriskmanagementandaccidentpreventionoftheproductionofcalciumcarbide.
作者:毕颖 马世杰Author:BIYing MAShijie
作者单位:沈阳化工大学环境与安全工程学院,沈阳110142
刊名:安全与环境学报 ISTICPKU
Journal:JournalofSafetyandEnvironment
年,卷(期):2024, 24(6)
分类号:X932
关键词:安全工程 风险评估 电石炉 核主元分析(KPCA) 麻雀搜索算法(SSA) 核极限学习机(KELM)
Keywords:safetyengineering riskassessment calciumcarbidefurnace KernelPrincipalComponentAnalysis(KPCA) SparrowSearchAlgorithm(SSA) KernelExtremeLearningMachine(KELM)
机标分类号:TP309SF276.6
在线出版日期:2024年7月4日
基金项目:核主元分析与优化核极限学习机模型在电石炉爆炸风险评估中的应用[
期刊论文] 安全与环境学报--2024, 24(6)毕颖 马世杰为准确判断电热法电石生产工艺中电石炉的爆炸风险等级,提出了一种精准有效的风险评估模型.首先,基于危险与可操作性(HazardandOperability,HAZOP)分析筛选出人、物料、设备、管理四方面的34项爆炸风险因素,考虑到因素间...参考文献和引证文献
参考文献
引证文献
本文读者也读过
相似文献
相关博文
关键词:安全工程,风险评估,电石炉,核主元分析(KPCA),麻雀搜索算法(SSA),核极限学习机(KELM),
|
|