活塞式氢气压缩机排气温度超限故障诊断方法  

Fault diagnosis method of exhaust temperature exceeding limit of piston hydrogen compressor

作　　者：韩宇恒 胡光忠 夏秋 王平 HAN Yuheng;HU Guangzhong;XIA Qiu;WANG Ping(College of Mechanical Engineering,Sichuan University of Science&Engineering,Yibin 644000,China;Sichuan Provincial Key Lab of Process Equipment and Control,Yibin 644000,China;Key Laboratory of Advanced Manufacturing Technology of Panzhihua City,Panzhihua 617000,China)

机构地区：[1]四川轻化工大学机械工程学院,四川宜宾644000 [2]过程装备与控制工程四川省高校重点实验室,四川宜宾644000 [3]攀枝花市先进制造重点实验室,四川攀枝花617000

出　　处：《机电工程》2023年第12期1973-1982,共10页Journal of Mechanical & Electrical Engineering

基　　金：自贡市重点研发计划项目(2021QYCX03);过程装备与控制工程四川省高校重点实验室开放基金资助项目(GK202205);攀枝花市先进制造技术重点实验室开放基金资助项目(2022XJZD01)。

摘　　要：针对活塞式氢气压缩机各级排气温度超限的故障难以用传统故障的二态性准确描述,以及目前未见能对该故障实施自诊断的故障诊断系统等问题,以DW-(5.6-13)/(0.2-0.5)-20型无油润滑变频活塞式氢气压缩机为研究对象,对相关零部件进行了重要度分析,并设计了故障自诊断系统。首先,对该故障搭建了模糊高木-关野(T-S)故障树模型,并利用基于贝叶斯网络(BN)的T-S故障树分析法对底事件进行了重要度分析,以确定薄弱环节和故障零部件诊断顺序;然后,利用模糊数表征了零部件的故障状态,对排气温度超限故障发生概率进行了预测;最后,设计了基于规则的故障诊断专家系统,并对实例进行了故障原因的推理排查。研究结果表明:基于规则的故障诊断专家系统在对现有故障数据进行分析后,其诊断结果与事实相符,说明该系统能够准确地诊断出故障原因,并进行自诊断;同时,基于BN的T-S故障树分析法,使该系统可对顶事件故障发生概率进行预测,为活塞式氢气压缩机故障诊断提供了一个可行方案。In response to the difficulty in accurately describing the fault diagnosis of exhaust temperature exceeding the limit of a piston type hydrogen compressor using traditional two-state fault diagnosis at each stage,and the lack of application of a fault diagnosis system that could finish self-diagnose for diagnosing the exhaust temperature exceeding the limit,the DW-(5.6-13)/(0.2-0.5)-20 oil-free variable frequency piston type hydrogen compressor was taken as the research object,the importance of relevant components were analyzed,and a fault self-diagnosis system was designed.Firstly,the corresponding fuzzy Takagi-Sugeno(T-S)fault tree analysis model was built for the fault,and the T-S fault tree analysis method based on Bayesian network was used to analyze the importance of the bottom event to determine the weak links and the diagnosis sequence of faulty parts.Then,the fuzzy number was used to represent the fault status of the parts,and it was brought into the T-S fault tree analysis algorithm based on Bayesian-network(BN)to predict the occurrence probability of the primary exhaust temperature overrun fault.Finally,a rule-based fault diagnosis expert system was designed to infer and troubleshoot the fault causes of the instance.The research results indicate that after analyzing existing fault data,the rule-based fault diagnosis expert system s diagnosis results are consistent with the facts,indicating that the system can accurately diagnose the cause of the fault and achieve self-diagnosis.At the same time,combining with the T-S fault tree analysis method based on BN,the system can also realize the prediction of the probability of top event failure,which provides a feasible scheme for the fault diagnosis of the piston hydrogen compressor.

关 键 词：故障诊断 活塞式氢气压缩机排气温度 T-S故障树模型 专家系统 贝叶斯网络 

分 类 号：TH457[机械工程—机械制造及自动化]