Classification of Power Quality Disturbances Using Regularized Artificial Neural Network

Power quality disturbances are a major concern during the operation of utility grid. Integration of the renewable energy sources to the grid has further deteriorated the power quality due to use of power electronic converters and drives. Proper detection and classification of the various power quality disturbances is needed to identify the origin and to execute adequate mitigation measures. Here, an Artificial Neural Network (ANN) based classifier is used to identify different power quality disturbances and its performance is evaluated. A data-set consisting of nine different disturbances with eight features extracted using Stockwell Transform from each signal is used for training and testing of the proposed ANN classifier. The performance of proposed ANN Classifier is compared with other classifiers such as Random Forests and Decision Trees from the literature. The performance of ANN classifier was appreciable in low noise level situations and it outperformed other classifiers in high noise level scenario.