本研究從思考維護管理系統應具備的功能和應達成之目標,而勾勒出整合性維護管理系統之全貌,此一整合性維護管理系統(Integrated Maintenance Management System, IMMS)架構,其中包括即時監視(Real Time Monitoring, RTM)子系統、可靠度核心維護(Reliability Centered Maintenance, RCM)子系統、電腦化維護管理子系統(Computerized Maintenance Management Subsystem, CMMS)。獨立的即時監視子系統,缺乏分析的能力,使我們無法進一步提出改善的要求。若僅有可靠度核心維護子系統,則缺乏伴隨設備的原始數據,可靠度核心維護子系統的分析正確性待評核。又若僅考慮電腦化維護管理子系統,而無分析機制,易加重決策者下達決策的負擔,所以三者間不是獨立的,而是需互相依循與配合方臻完善。透過這套架構使整體工廠維護的流程運作達到最佳運作效率,而不僅是膚淺地節省某一操作員或節省某一操作時間而已,以利發揮整合之功能。本研究將複雜的維護系統有條理的說明及模組化,並以本校實驗室之射出成型機為範例說明之。 因為維護週期機制在維護管理系統中佔有重要地位,而維護管理系統需經由維護週期機制求算維護時程,以方便維護人員安排維護工作,所以維護週期機制是不可或缺的部分。本研究發展預防維護數學模式,同時考慮備品置換頻率和備品採購與維護人員人力,並求得在最小成本下之最佳維護時點。由於維護機器數量、備品數量與維護人員數量在實際系統中相當地多,本模式在實際應用上可能會發展成複雜大型化之系統,所以利用遺傳演算法(Genetic Algorithm, GA)來求解問題。本研究分析遺傳演算法之求解品質,利用實驗方式找出最佳參數組合,並設計三組範例分別與數學規劃所求得之最佳解作分析與比較,最後以半導體測試廠測試機台設備為例說明。
This research addresses the overall view of an Integrated Maintenance Management System (IMMS). Its structure consists of three subsystems of the Real Time Monitoring (RTM), Reliability Centered Maintenance (RCM), and the Computerized Maintenance Management System (CMMS). These subsystems are not independent. A stand-alone RTM cannot continuously provide improving suggestions without analytical capability. On the other hand, a RCM may offer ambiguous analyses if it is short of real-time data from equipment. In addition, a tedious CMMS creates the burden for making decisions. Therefore, these three subsystems should be considered together to form an integrated structure. This structure is expected to provide guidance for the maintenance operation flow development of an efficient factory. Modules of the IMMS will be illustrated in details. A mold injection machine is also described in this paper for the argument. As a critical part in Maintenance Management System, The maintenance interval mechanism is responsible to figure out the maintenance schedule for maintenance arrangement. This research not only develops a preventive maintenance mathematical model but also considering the spare parts replacement frequencies and spare parts purchasing and workforce to obtain an optimal maintenance point in the minimized cost. Due to numbers of maintenance machine, spare parts and maintenance workforce in actual system, the model in real practice may become to a much complex system. Therefore, to solve problems by using Genetic Algorithm(GA). This research analyse the solution quality of Genetic Algorithm, try to find the best parameter combination by exercising experimentation. And design 3 examples to analyse and compare to the best solution of mathematical plan respectively. Finally, taking an example of testing facility from semiconductor factory for illustrating.