The decomposition algorithm of AFS structure and its applications to failure diagnosis

A new fuzzy model for pattern recognition in failure diagnosis has been established by AFS (axiomatic fuzzy set) theory, but the complexity of systems makes analysis difficult. By analysing the AFS structure for systems, the independent components for AFS structures are studied, and an algorithm for searching for independent sub-AFS structures is proposed. The complexity of a real world application system can be greatly reduced by this algorithm with failure diagnosis as an example, so the speed and accuracy of failure diagnosis can be improved. The example of failure diagnosis for a marine diesel engine demonstrates the advantages of this approach.