Fault Diagnosis of Power Transformer Based on DGA and Information Fusion

Aiming to improve fault diagnosis performance of power transformer, a novel approach based on dissolved gas analysis and information fusion technology is presented in this paper. The diagnosis results obtained by grey relation analysis, fuzzy logic, and back propagation neuron network are employed as basic probability assignments of D-S theory. Then, compatibility ratio is used to solve the possible conflicts among different sub-evidences. Finally, fault diagnosis result is obtained thorough decision-making criterial. The experimental result suggest that the proposed approach has better performance than that of conventional methods, and fault diagnosis accuracy up to 93.18%.