Power transformer fault diagnosis based on genetic support vector machine and gray artificial immune algorithm

A two classifier cascade power transformer fault diagnosis algorithm was proposed to solve the problem of both single and multiple power transformer fault diagnosis. The diagnosis algorithm simulated biological immune system. Support vector machine classified fault or normal state of power transformer as the first classifier. Genetic algorithm optimized kernel function parameter of support vector machine. Gray relation grade calculated affinity between antibody and antigen in artificial immune algorithm. High-frequency variation based on dynamic vaccine mechanism generated a new antibody. Best memory antibody set was trained according to different fault types. Five neighbors integrated decision making method diagnosed power transformer fault representation based on the best memory antibody set. Experiments indicate that power transformer fault diagnosis algorithm combines genetic support vector machine with gray artificial immune and dynamic vaccine mechanism can effectively classifies single and multi-fault of power transformer and raise power transformer fault diagnosis accuracy and diagnosis speed. © 2011 Chinese Society for Electrical Engineering.