Entropy aware EnKF framework for reliable power quality prediction in smart grids

In this paper a novel predictive technique is proposed for power quality (PQ) event classification in smart grid. Here Ensemble Kalman filter (EnKF) based approach and Entropy based trust modelling are utilized to design hybrid learning model that can classify PQ events during non-stationary environment. Here both PQ real and synthetic data had been used to mitigate the scarcity of real time data. To improve the learning ability of the proposed technique an entropy-driven weighting method is suggested. This technique dynamically assesses and adjusts the impact of synthetic data during the filter update and ensure the presence of only high-quality augmented samples, used to estimate the state. To get time-frequency features, the adaptive Q-factor Wavelet Transform (AQWT) is used in this method. The proposed method has shown better performance compared with the existing method.