A Genetic Programming Based Fuzzy Model for Fault Diagnosis of Power Transformers

In this paper, a fuzzy model based on genetic programming (GPFM) is proposed to diagnose the fault types of insulation of power transformers. The proposed GPFM algorithm constructs the fuzzy relationship between input and output fuzzy variables by genetic programming algorithms. The parameters of memberships of fuzzy subsets and the fuzzy relationship of system are represented by the GP candidates that have the form of tree-like combinations of fuzzy subsets of input variables. Then the best fuzzy function is evolved by genetic operations and evolution. Based on the proposed GPFM algorithms, an insulation fault diagnosis system for power systems is designed to distinguish the insulation fault types of power transformers. Compared with the conditional fuzzy IEC code method, the GPFM algorithm can automatically generate fuzzy relationship between fault symptom with fault types and shows better performances.