Research and implementation of a multi-fault diagnosis method based on case-based reasoning

In this paper, a multiple fault diagnosis method based on case-based reasoning (CBR) is presented. And the characteristics of multiple fault modes are studied, then the case knowledge library is established for multiple faults, finally a two-step case retrieval strategy is proposed. The candidate case generation method based on the membership weights is employed to initially retrieve, which can reduce the number of candidate cases efficiently. Then, through calculating the case similarity using the grey relational analysis, the most probable fault case is obtained. It can avoids the problem of time-consuming computation resulting from the combination explosion of multiple faults. Finally, an example is applied to illustrate the proposed scheme, the results mean that the above method is effective, and it can be adopted in the actual industrial production.