An efficient islanding detection method in microgrids using hybrid SVM based decision tree

In this paper, a hybrid SVMDT classifier is proposed to provide an accurate detection of islanding with less detection time. The input signals are measured from the particular location of DG and the possible features are extracted from it. Some of the optimal features are selected from the extracted features which makes the classification better. Finally with the knowledge of a trained hybrid SVMDT classifier the signal is classified as islanding or non-islanding. The proposed methodology is implemented in MATLAB platform and the performance comparison is done with existing methods used for islanding detection in the distributed generators which has shown a better performance.