應用資料探勘技術於設備失效分析

Applying Data Mining Techniques for Equipment failure analysis

劉伃樺(Yu-Hua Liu)

指導教授 ： 陳凱瀛

國立臺北科技大學/管理學院/工業工程與管理研究所/碩士(2008年)

摘要

設備是公司重要的資產，高附加價值的設備更需要妥善維護及管理以滿足公司之生產營運需求。設備失效會導致產能減少影響公司利潤。但設備保養及維護成本昂貴。電子化維護的概念被提出，希望能降低維護成本。電子化維護透過即時監控系統取得資料進行設備維護保養，達到降低設備維護保養成本的目的。即時監控傳回大量資料，要從大量資料中獲取有用的資訊必須依賴資料探勘技術。SQL Server 2005提供一個從資料庫、資料倉儲到資料探勘的完整流程。
大部分的資料探勘技術被用在商業智慧上。本研究將以台電公司火力發電廠的資料為例，使用SQL Server 2005對資料加以整合、分析建立一電子化維護模組。使用類神經、決策樹、貝式分類等資料探勘技術對資料加以分析，從中獲取設備失效關鍵因素建立失效樹。當設備發生失效時，根據失效樹對失效狀況進行診斷，並對失效狀況即時作出回應。

關鍵字

資料探勘 ； 設備失效分析 ； 商業智慧

並列摘要

Equipment, especially those with substantial added value, is an important asset of a company and need to be carefully maintained. The failure of equipment leads to the capacity reduction and affects the profit. However, the equipment maintenance can be costly. E-maintenance, for instance, has recently emerged as a way to reduce the cost. Specifically, e-maintenance is executed via the real-time monitor system. The data mining technology is required to better use the data from this system. SQL Server 2005 provides a complete solution from data source to data mining.
However, the data mining technology is mostly used in business intelligence. This thesis provides a novel way of applying this technology in building e-maintenance model with SQL Server 2005 that provides an “all-in-one” technology solution. The thesis takes the Hsiehho thermal power plant of Taiwan Power Company, Ltd as its example. We use SQL Server 2005 to integrate data from the real-time monitor system. Through the data mining technology such as decision tree, Naïve Bayesian classifier and neural network analyze the data. According to the result of data mining, we can build a fault tree that help diagnose the reasons of failure and provide real-time responses to handle the failure.

並列關鍵字

data mining ； equipment failure analysis ； business intelligence

參考文獻

[34]李家隆，資料粒化技術用於設備失效分析，碩士論文，台北科技大學工業工程與管理系，台北，2006。

[1]Lee, J., “E-manufacturing- fundamental, tools, and transformation,” Robotics and Computer Integrated Manufacturing, vol. 19, no, 6, 2003, pp. 501-507.

[5]Wang, L., Chu, J., and Fu, Y.F., “Advanced maintenance strategy for power plants- introducing intelligent maintenance system,” Proceeding of the 6th world congress on intelligent control and automation, China, 2005, pp. 7444-7448.

[6]Rosovska, I., Chebel-Morello, B., Zerhouni, N., “Process of s-maintenance: decision support system for maintenance intervention,” Emerging Technologies and Factory Automation, vol. 2, 2005, pp. 679-686.

[10]Han, J., and Kamber, M., Data mining: concept and techniques, U.S.A., Academic Press, 2001.