Hybrid method for fault diagnosis based on model and case-based reasoning

Model-based diagnosis which describes the deep knowledge of internal structure and behavior of a system can easily obtain diagnostic knowledge, while its disadvantage is the large amount of calculation. The diagnostic efficiency of case-based reasoning which relies on historical experience is relatively high, but knowledge acquisition is its bottleneck. Therefore, this paper presents a reasoning method combining model-based diagnosis with case-based reasoning to improve the practicability of predicate logic model based diagnosis. The process of case acquisition is achieved by model-based diagnosis and the fault diagnosis result is given through case-based reasoning. This approach overcomes the problem of knowledge acquisition and inherits the high-speed characteristics of case-based reasoning.