Rough set and fuzzy wavelet neural network integrated with least square weighted fusion algorithm based fault diagnosis research for power transformers

In this paper, a rough set (RS) and fuzzy wavelet neural network (FWNN), integrated with least square weighted fusion algorithm based fault diagnosis for power transformers, using dissolved gas analysis (DGA) is proposed. This approach takes advantage of the knowledge reduction ability of rough set and good classified diagnosis ability of FWNN, integrated with least square (LS) weighted fusion algorithm. The rough set is used as a front of FWNN, integrated with LS weighted fusion algorithm to simplify the input of FWNN and mine the rules whose “confidence” and “support” satisfy a preset criteria. The mined rules are used as a diagnosis knowledge base to offer fault diagnosis service for power transformers. FWNN, integrated with LS weighted fusion algorithm, is used to diagnose the case that cannot be diagnosed by mined rules by rough set. The FWNN input is simplified by rough set reduction, and its learning rate is improved greatly. FWNN, integrated with LS weighted fusion algorithm, on one side, can much better improve the diagnosis accuracy, when the output vector of single FWNN has the similar element. On the other hand, its diagnosis accuracy cannot be limited by the neural network hidden layer number and correlated training parameter. The mechanism has good classified diagnosis ability. The advantages and effectiveness of this method are verified by testing.