A New Hybrid Fault Diagnostic Method for Combining Dependency Matrix Diagnosis and Fuzzy Diagnosis Based on an Enhanced Inference Operator

摘要

This paper is devoted to the fault diagnosis of electronic systems by combining logical signals, such as built-in test output, and analog signals, such as voltage, current and temperature. First, the basic inference principles of dependency matrix (D matrix) diagnosis and fuzzy diagnosis are introduced, and the characteristics of their inference operators are summarized. Then, the similarities and differences between the two diagnostic methods are analyzed. Based on the judgement of close degree, a new enhanced inference operator is defined to suit both the D matrix and the fuzzy relation matrix (R matrix). A DR matrix is defined to describe the mixed relationships between faults and the two types of signals. Based on the enhanced inference operator and the DR matrix, a new hybrid diagnostic method is established. Finally, a signal modulating circuit is used to verify the effectiveness of the enhanced inference operator on the D matrix, the R matrix and the DR matrix, which demonstrates the high efficiency of the enhanced inference operator and the feasibility of the new hybrid diagnostic method.