Research On Fault Diagnosis For Inverter Of Wind Power Generation System

This subject comes from the National Natural Science Foundation of China "Research on Intelligent fault diagnosis method of the inverter in wind power generation system"(No. 61364010). In this paper, switch faults of power switch device in inverter are analyzed and studied by extracting useful information and selecting proper method.The object of the paper is the common diode-clamped three level inverter. The three-level inverter of the multilevel inverter is analyzed in the wind power generation system by studying the short circuit states of the power switch devices under all circumstances. In order to make it easy to diagnose the faulted devices accurately, the voltage on the diode leg is chosen as the faulted information and being analyzed.Then, the combination of wavelet transform and support vector machine is used for fault diagnosis. Because the fault signature signal is nonlinear transient signal in three-level inverter. The fault signature vector is the energy value of the frequency part of the fault signal which is extracted by the wavelet transform. Afterwards, the support vector machine method is used for fault diagnosis. To address the uncertainty of parameters, the cuckoo optimization algorithm is introduced to seek the optimal combination of the penalty parameter and kernel function parameters. Thus it can diagnose the multi-fault of the three-level inverter accurately. The simulation of the proposed method is carried out in Matlab. Compared with BP neural network, Genetic algorithm optimized support vector machine method and Particle Swarm Optimization support vector machine method, the results show that the proposed method has a shorter optimization time and higher accuracy in diagnosis.Finally, a GUI interface of the three-level inverter fault diagnosis system is built in Matlab, from where the diagnosis results can be observed and got directly and fast.