Research On Fault Diagnosis Technology For Power Electronic Circuits

With the proposal of the concept of all-electric-aircraft, the aviation power system becomes increasingly important. Acted as power supply or motor driver, the power electronic circuit is able to provide power for the aviation power systems. In order to improve the reliability of aviation power system and facilitate the development of airborne equipment automatic test technology, the research of power electronic circuit fault diagnosis is of great significance. In this paper, the power electronic circuit fault diagnosis technology is studied, the main research content contains these aspects:(1) Power electronic circuit fault diagnosis algorithms are studied, which contain BPNN and SVM. Single output and multiple output BPNN models are designed in the fault diagnosis methods, respectively. In view of the defect that BPNN is easy to fall into local minimum value, the mind evolutionary algorithm(MEA) is proposed to optimize the initial weights and thresholds; As for SVM, one-vs-all(OVA) and one-vs-one(OVO) algorithms are adopted.(2) Taking DC-DC conversion circuit with feedback control as an example to conduct the study of power electronic circuit fault diagnosis simulations, the components’ equivalent models are established and the typical fault modes of the converter are set. A number of simulation-based experiments under different working conditions are carried out to evaluate the performance of the fault diagnosis methods based on BPNN classifier and SVM classifier. The characteristics and suitable applied occasions of each classifier are summarized based on the simulation results.(3) The DC-DC converter test platform is built. Then, the faults are injected into the converter and the practical measured data are sampled. After that, BPNN and SVM classifiers are used to diagnose the converter. Experimental results further verified the feasibility and effectiveness of the fault diagnosis methods based on these two classifiers.