Identification of power quality disturbances using Artificial Neural Networks

This work presents the investigations carried out on application of Hilbert Huang transform (HHT), back propagation algorithm (BPA), radial basis function(RBF) and locally weighted projection regression (LWPR) for power quality disturbance identification. Features are extracted from the electrical signals by using HHT. HHT method is a combination of empirical mode decomposition (EMD) and Hilbert transform (HT). The output of HT are instantaneous frequency (IF) and instantaneous amplitude (IA). The features obtained from the HHT are unique to each type of electrical fault. These features are normalized and given to the RBF, BPA and LWPR. The data required are collected from textile mills using three phase power quality analyzer at various time durations and places. The performance of the proposed method is compared with the existing feature extraction technique namely Hilbert Transform with Radial Basis Function (HTRBF). The accuracy of results are presented by calculation of percentage error for identification of power quality disturbances, training time duration and testing time duration of algorithms and they are compared with existing algorithm. Simulation results show the effectiveness of the proposed method for power quality disturbance identification.