文章编号：1671-7872(2023)02-0158-08基于FTA−BN的混合动力汽车故障诊断策略万凌初1，牛礼民1,2，胡超1(1.安徽工业大学机械工程学院，安徽马鞍山243032；2.安徽工程大学电气传动与控制安徽省重点实验室，安徽芜湖241000)摘要：为实现对混合动力汽车(HEV)底盘系统多部件进行故障融合诊断，提出一种基于故障树分析和贝叶斯网络(FTA−BN)相结合的故障诊断策略。结合VB语言和SQLServer开源数据库平台设计HEV底盘故障诊断系统，基于构建的HEV底盘系统故障树模型，通过故障树与贝叶斯网络的映射关系将故障事件转换为网络节点，引入专家系统；依据混合模糊推理结合有界深度优先搜索方法，通过对故障征兆的定性和定量分析，基于FTA−BN计算故障事件的可信度，选择可信度高的原因作为诊断结论，且进行实验验证。结果表明：采用提出的FTA−BN故障诊断策略可对HEV底盘系统多部件进行故障融合诊断，故障诊断的准确率为0.850，高于人工经验判断的准确率(0.675)。关键词：混合动力汽车；故障诊断；专家系统；SQLServer；贝叶斯网络中图分类号：U469.79文献标志码：Adoi：10.12415/j.issn.1671−7872.22213FaultDiagnosisStrategyofHybridElectricVehicleBasedonFTA−BNWANLingchu1,NIULimin1,2,HUChao1(1.SchoolofMechanicalEngineering,AnhuiUniversityofTechnology,Maanshan243032,China;2.KeyLaboratoryofElectricDriveandControlofAnhuiProvince,AnhuiPolytechnicUniversity,Wuhu241000,China)Abstract：Torealizefaultfusiondiagnosisformultiplecomponentsofhybridelectricvehicle(HEV)chassissystem,afaultdiagnosisstrategybasedonfaulttreeanalysisandBayesiannetwork(FTA−BN)wasproposed.ThefaultdiagnosissystemofHEVchassiswasdesignedbasedonVBlanguageandSQLServeropensourcedatabaseplatform.BasedontheconstructedfaulttreemodelofHEVchassissystem,thefaulteventswereconvertedintonetworknodesthroughthemappingrelationshipbetweenfaulttreeandBayesiannetwork,andthefaulteventswereintroducedintotheexpertsystem.Basedonthemixedfuzzyreasoningandboundeddepthfirstsearchmethod,throughthequalitativeandquantitativeanalysisoffaultsymptoms,thereliabilityoffaulteventswascalculatedbasedonFTA−BN,andthecauseswithhighreliabilitywereselectedasthediagnosticconclusions,andtheexperimentalverificationwascarriedout.Theresultss...
