Islanding detection research for multiple inverter grid-connected photovoltaic system

In view of the insufficiency of commonly used islanding detection methods of multi grid-connected inverters, the islanding detection scheme based on wavelet and neural network is proposed. In this method, wavelet transform is adopted to extract feature vectors from the voltage of PCC point and the output current of inverter, and then pattern recognition is exerted by BP neural network to determine whether there is an island phenomenon. The results of simulation and experiment show that the detecting time is short and the non-detection zone is very small. This method doesn't deteriorate the output power quality and the accuracy and reliability is very high. ©, 2015, Science Press. All right reserved.