Voltage Sags Source Detection and Location in Distribution Power System with DG

The purpose of this paper is to describe a new detection and source location method of voltage sags in distribution power system with DG. The proposed method allows us to detect voltage sags and locate sag source from inside or outside when measurement transients are present in the recorded data from DG installations. S-transform is applied to transients signal processing first. Then, features for voltage sag source classification are exacted in reference to criteria of the some existed methods. Support vector machine is applied to establish the optimal hyper-plan for classifying sag source from upstream or downstream inside by machine learning. Last, the established binary classifier is used to classify upstream or downstream, and then determine the sag source from DG system or outside. Simulation in wind farms is presented according to the proposed method.