The Power Electronic Circuit Fault Diagnosis Based On The Neural Network

With the presence of new power electronic products and the growth of quality requirement of system performance,the fault diagnosis of power electronic circuit become more important.The research of the power electronic circuit fault diagnosis not only can improve working efficiency for the test members,but also can short the fault time.The study of fault diagnosis of power electronic circuit have great significance on the application of power electronics technology.In this paper,the neural network method is proposed to study the fault diagnosis of the power electronic circuit.The details as follows:Firstly,taking the three-phase rectifier circuit for instance,the simulation model is established by using MATLAB to obtain output voltage.Then the Fourier analysis method is used to extract the DC component,the amplitude,the second harmonic and harmonic amplitude three times.After these values are normalized,they are set into the BP network to receive six numbers so that fault location and fault point are determined.Due to the BP neural network has local minimum problem,the RBF is used to improve it.The RBF is more fit for the demand which contains the ability of function approximation,classification and learning rate.In order to compare with the BP neural network,the three-phase rectifier circuit is still used as the fault model.After these values are normalized,they are set into the RBF network so that the complexity of calculation is reduced.In practical applications,the use of BP neural network in diagnosing the large scale power electronic circuits will result in the low diagnostic accuracy.In order to improve the diagnostic rate,the method of hierarchical clustering can be applied to diagnose.The main reason is that classification processing isemployed firstly,then the appropriate network model is selected for diagnosis,so more ideal diagnostic results can be acquired.The experiment result shows that this method has merit of higher diagnostic rate,higher reliability and widely application.