Application of Grey Relational Analysis on Fault Diagnosis Assistance System: A Case Study of Chemical Dispensing System

• Main Authors: CHI-TA YEH, 葉啟達
• Other Authors: Ming-Jong Tsai
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/8sec82

碩士 === 國立臺灣科技大學 === 自動化及控制研究所 === 99 === Recently, the typical fault diagnosis of a semiconductor factory’s chemical dispensing system is that an engineer makes fault diagnosis according to alarm status and his own maintenance experience. However, it is difficult for training an experienced engineer, so the diagnosis mistakes or time delays are often happened. This study employs an expert-based system including fault related index factors and relational weighting values through the actual case analysis and expert’s maintenance experience in the factory. And the grey relational analysis inference is also used to establish a fault diagnosis assistance system for a chemical dispensing system which can generate the first three priorities of check items to be diagnosed for a maintenance engineer. This study selects five real cases as the investigation tasks and compares the performance between the proposed diagnosis system and the traditional human work. The implemented results show that the proposed fault diagnosis assistance system can reduce 65.9% check items, and shorten 66% of diagnosis time in average. It is a useful method to improve the fault diagnosis efficiency.