Study of the Intelligent Fault Diagnosis System Based on Rough Set

The knowledge rules acquisition is the bottleneck of the fault diagnosis due to uncertainty information during the process of fault diagnosis. Rough set (RS) is a new theory to deal with vagueness and uncertainty information. A model of fault diagnosis knowledge acquisition for the rotating machine based on rough set is presented in this paper. The decision table is formed and the fault attributes are reduced by Showron matrix. The rule degree of confidence and degree of coverage are used as evaluating indictors to judge the reduction rules. The database of the intelligent expert system is updated with these minimum reduced attributes´ sets. This model is applied in the rules acquisition of rotating machine. The attributes number is reduced from 11 to 5. The intelligent fault diagnosis expert system with the new acquisition rules is verified in water-injection sets of Daqing oil field.