工程装备分布式故障诊断专家系统设计与研究

Design and research on engineering machinery expert diagnostic system

国内工程装备的故障检测与诊断存在诊断对象和知识复杂、历史经验不足等缺点,故障诊断系统的效率不高,实用性较差。设计了一种基于B/S结构的远程分布式故障诊断模糊专家系统。该系统由星形拓扑结构组成,包括现场监测站、现场监测中心和远程诊断中心三个部分。系统的软件开发标准采用分布式组件对象模型标准,实现网络结构的分布式计算和相关通讯及同步。系统开发过程遵循了知识库的知识表示规范、文件规范的制定、知识库管理软件的开发和推理机的开发这一顺序。介绍了系统的推理机组件定义,利用模糊集理论和基于Mamdani算法的模糊推理技术构造了模糊专家系统推理机。应用证明远程分布式故障诊断模糊专家系统解决了分布在广域范围内的故障知识的获取问题,降低了复杂问题的求解难度,提高了系统的整体诊断能力。

Due to the complicated object and knowledge database, as well as lack of historical experience in domestic fault diagnosis of engineering vehicle, the fault diagnosing system is inefficient and sometimes infeasible. Therefore, a kind of remote distributed fuzzy expert diagnostic system based on browser/server structure is designed. It consists of LMU （Local Monitoring Unit）, LMC （Local Monitoring Center） and RDC （Remote Diagnosis Center） with star-type structure. Distributed Component Object Model was adopted as the software development standard to realize the distributed computation, interrelated communication and co-synchronization. The system was accomplished with the order of the preparation of knowledge expression standard and file system standard, development of the knowledge database management system, inference engine. The definition of components of inference engine of fuzzy expert system is introduced. Combined the fuzzy set theory and fuzzy inference technology based on Mamdani algorithm, the fuzzy expert system inference machine is realized. The application shows that the remote distributed fuzzy expert diagnostic system can resolve the problem of knowledge acquisition in wide area, lower the difficulty of problem solving and promote the overall performance of the expert system.