PASSIVE ISLANDING DETECTON METHOD FOR PHOTOVOLTAIC INVERTER BASED ON WAVELET PACKET TRANSFORM AND SUPPORT VECTOR MACHINE

Abstract

The disadvantage of the existing passive islanding detection methods is a
long detection time and a large non detection zone (NDZ), an islanding detection
method based on wavelet packet transform and support vector machine (SVM)
was proposed. Firstly, the voltage signal of the point of common coupling is
gathered, and then the voltage signal is decomposed into different frequency
bands by wavelet packet. Secondly, the normalized logarithmic energy entropy of
wavelet packet for each band is calculated, and these features are fed into the SVM
model. Finally, the SVM model is used for determining whether there is an island
phenomenon. The experimental results verify that the proposed method can
detect system islanding operation mode is fast and accurate.