WT and DT-based Protection Scheme for Hybrid Microgrid System

Abstract

The rising power demands accompanied by drastic alterations in the environmental conditions have led to the installation of more and more renewable based DG sources into the present distribution system. Owing to the rising penetration of distributed generation (DG) sources in the distribution network, the protection has become more challenging. This paper describes the wavelet transform (WT) and decision tree (DT) based protection scheme for the distribution lines of hybrid microgrid under islanding mode of operation. The pre-processing and feature extraction of the instantaneous values of the signal obtained at the relaying bus is carried out using WT to obtain the significant features for training the DT based classifiers to perform fault detection, fault classification and faulty line identification. The proposed scheme has been tested for a wide range of fault and no-fault scenarios. The performance of the proposed classifier is compared with linear support vector machine (SVM) and linear discriminant (LD) classifiers in terms of the classification accuracy.