模糊故障诊断方法之研究.PDF

德霖學報 第二十二期 模糊故障診斷方法之研究 民國 97年 6月 1 模糊故障診斷方法之研究 蔡有藤 德霖技術學院 機械系 教授 摘要 隨著科技的進步與自動化技術發展，機台設備漸趨高功能與複雜化，機台設備的保養維護越來 越重要；透過故障診斷系統的建立，縮短停機維護時間，使設備發生故障時能迅速排除故障，為製 造系統管理上的重要課題。本文探討系統故障徵兆(Symptoms)與故障源的建立方法，利用故障樹分 析(Fault Tree Analysis)分析各機具(Tools)可能失效，建立機具失效的判定法則，系統故障徵兆與機 具失效間的對應關係。利用模糊理論建立故障徵兆語意表示，對異警徵兆不明確問題建立故障源與 徵兆之隸數度，發展模糊故障診斷推論策略；此外，結合貝式估計方法，分析故障底事件之失效機 率，以為決定維護優先次序依據。透過二階段故障診斷作業，對徵兆具模糊性的故障問題，提供必 要的故障維修資訊，使維護人員可迅速檢出故障源。此外，導入以診斷維修為導向之設計概念 （Design for Diagnosis and Maintenance），藉此改善設備可靠度、維護度、操作度，提昇設備妥善 率及生產能力。 關鍵字：模糊估計、故障診斷、維護 A study of fuzzy fault diagnosis strategy Yuo-Tern Tsai Department of Mechanical Engineering, De-Lin Institute of Technology Abstract Following the development of production technology, equipment is more and more complex and maintenance is also more important in modern industry. In order to rapidly fix the faults and shorten the time of maintenance, fault diagnosis is more critical in maintenance action. This paper presented the method of fuzzy fault diagnosis to support the development remote fault diagnosis system. The technique of Fault Tree Analysis (FTA) is used to drawn out the possible faults and the symptoms of a system. The fuzzy evaluation is used to process the problems of which the faults and the symptoms are dealing with the uncertainty. The faulted subsystem was first decided by a fuzzy operation. The set theory is used to investigate the failure probabilities of the failure causes in each subsystem. The Bayesian estimation is then taken to evaluate the priority of failure for the causes. The web-based diagnosis framework was also suggested to development far-distance fault diagnosis system. Further, the concepts of design for diagnosis and design for maintenance are also proposed for promoting the availability of system in use. Key words: Fuzzy evaluation, fault diagnosis and maintenance. 德霖學報 第二十二期 2 一、緒論 機台發生故障，維修工程師可依據機台狀況資料，判斷故障源並依據設備製造商提供之故障維 修手冊進行簡易故障排除。然而，相關的診斷與維修知識仍屬工程師的經驗累積，由於工程師或專 家的異動，常使得維修經驗與知識無法傳承下來，導致維修人員專業知識不夠；傳統維修作業面臨 問