Research On Fault Diagnosis Of Power Electronic Circuits Fused With Wavelet Packet Transform And Artificial Immune Algorithm

Power electronics technology in power transmission,power system,a variety of ac-dc power supply and other fields has a wide range of applications and prospects,so the safety of power electronics circuit has higher requirements.In order to find practical and efficient technology for fault diagnosis of power electronic circuit and minimize all kinds of adverse effects that may occur when power electronic circuit faults,it has been widely concerned in the power electronic industry.The artificial immune algorithm developed on the basis of the biological immune system retains the strong information processing ability and its unique mechanism of the biological immune system,and provides an efficient and feasible way for the field of fault diagnosis.Therefore,it has been highly valued.Aiming at the nonlinear characteristics of power electronic circuits.This paper firstly analyzes the significance of the research on fault diagnosis of power electronic circuit and the current research results.Based on the uniqueness of power electronic circuit and the current research status,a fault diagnosis model of power electronic circuit based on artificial immune algorithm is proposed.Taking three-phase bridge full-control rectification circuit as an example,the working process of the three-phase bridge full-control rectification circuit is described in detail,and 22 fault types of the circuit are simulated by Matlab/Simulink respectively.The output voltage waveform containing circuit fault information is obtained as sample data.Then,wavelet packet transform was used to extract the features of the sample data.Then,dimension reduction was carried out based on the separable analysis based on Fisher's criterion to make the fault features of the sample data more obvious.Then,the sample data were normalized and divided into training samples and test samples.For the established artificial immune algorithm diagnosis model,with the traditional immune cloning algorithm and the immune cloning algorithm improved by gaussian variation and roulette algorithm as the fault classifier,using the training sample training the fault classifier,then testing trained fault classifier with test sample,the results show that the improved immune clone algorithm effectively improves the circuit fault diagnosis accuracy and reduce the time of diagnosis.