证据理论在泵房过程故障诊断中的应用研究  

Application Research on Evidence Theory for the Process Fault Diagnosis of Pumps Control System

作　　者：吴宇恒[1] 税爱社[1] 路申易 王飞[1] 

机构地区：[1]后勤工程学院信息技术与军事物流工程系,重庆401311

出　　处：《自动化与仪器仪表》2017年第3期154-156,共3页Automation & Instrumentation

摘　　要：针对泵房作业过程的典型故障,分析了其对状态参数的影响;应用基于证据理论的故障诊断方法,设计泵房过程故障诊断流程;利用模糊理论在处理不确定问题时的优越特性,构建状态参数的隶属函数,确立证据理论基本概率分配函数模型,实现基于证据理论的泵房作业过程故障诊断。实验结果表明,该诊断方法提高了对冗余信息的利用效率,对泵房作业状态识别准确,具有较好的鲁棒性,降低了漏报率与误报率。Focusing on the typical process faults of pumps control system, their influences on state parameters are analysed. Fault diagnosis methods are applied based on evidence theory. And the fault diagnosis process of pumps control system is designed. By utilizing the superior characteristics of fuzzy theory in dealing with the uncertainty problem, the state parameters of the membership functions are built and the evidence theory basic probability assignment function model is established. All of this is in order to complete the fault diagnosis process of pumps control system The experimental results show that diagnostic method improve the efficiency of redundant information, recognize accuracy of pumps control system work state, have a good robustness and reduce the rate of missing report and false alarm.

关 键 词：证据理论 故障诊断 隶属函数 基本概率分配. 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]