RULES ACQUISITION BASED ON ROUGH SET FOR ROTATING WATER INJECTION MACHINERY
FAULT DIAGNOSIS

WANG Hongjun¹  ZHANG Jianmin²  XU Xiaoli¹

(1. Department of Mechanical Engineering, Beijing Institute of Machinery, Beijing 100085；
2. School of Mechanical & Vehicular Engineering, Beijing Institute of Technology, Beijing 100081 )

Abstract: Rough set method is good for dealing with the knowledge under the condition of imprecise and inconsistent information to acquire hidden knowledge and principle. A model of the fault diagnosis and decision rules acquisition for large rotating machinery based on Rough set is put forward. According to the inconsistent and redundant original information, the decision table is formed and the fault attributes are reduced based on Rough set theory. The attributes’ value is reduced by Showron discernibility matrix. Finally the minimum reduced attributes set and diagnosis rules database are established. This model is used for a rotating water injection machinery rules acquisition in practice. The original rough information is reduced to get the fault rules with 25% attributes’ reducing rate for the rotating water injection machinery. Its validity is proved.

Key words: Rotating water injection machinery set Fault diagnosis Rules acquisition Rough set

CLC No: TH113

国家自然科学基金(50375017)、北京市自然科学基金(3042006)和北京市重点实验室基金(030314)资助项目. Received 20051027, received in revised form 20051220