Synthetic Fault Diagnosis Method of Power Transformer Based on Rough Set Theory and Improved Artificial Immune Network Classification Algorithm

Abstract

According to complementary strategy, this paper presents a new power transformer fault diagnosis method based on rough sets theory (RST)and improved artificial immune network classification algorithm. Through reduction approach of RST information table to simplify expert knowledge and reduce fault symptoms, the minimal diagnostic rules can be obtained. An improved artificial immune network classification algorithm is proposed on the base of them. At first the artificial immune network which both antigens and memory antibodies with class information has been added into are trained to learn the features of fault samples. In this way, the memory antibody cells pool which can represent the fault samples better than those without class information can be obtained. Then the k-nearest neighbor method is used to classify the fault samples. Compared with the IEC three-ratio method and BP neural network (BPNN), the proposed algorithm has better capability to classify single-fault and multiple-fault samples as well as higher diagnosis precision.