Detection and Classification of Power Quality Disturbances in Smart Grids using Detrended Fluctuation Analysis and SVM Classifier

Conventional power system networks are being replaced by smart-grid systems, resulting in reliable, secure, and environmentally friendly systems. Due to the inclusion of renew-able sources, their associated power electronic devices and smart metering systems, monitoring becomes essential for satisfactory operation. This requires fast and accurate detection of causes that produce disturbances. In this article, a typical smart grid has been modelled in MATLAB Simulink and very common few power quality disturbances (PQD) have been created. A novel signal processing technique named detrended fluctuation analysis (DFA) has been performed to prepare a feature matrix which has been used to train a support vector machine (SVM) based machine learning (ML) classifier to classify these power quality issues. A total of 18 PQDs have been considered and using cubic kernel and ten-fold cross-validation technique, an accuracy of 98.9% has been achieved.