A PRACTICAL REAL-TIME POWER QUALITY EVENT MONITORING APPLICATIONS USING DISCRETE WAVELET TRANSFORM AND ARTIFICIAL NEURAL NETWORK

Abstract

Determining the events that affect Power Quality (PQ) disturbances is remarkable for consumers. The most important aspects in the assessment of PQ disturbances are real-time monitoring of PQ disturbances and their fast interpretation. In this study, Artificial Neural Networks (ANNs) was used as a classifier benefiting from estimated parameters in PQ disturbances based on Discrete Wavelet Transform (DWT) on the real-time environment for determining the disturbances in power systems. Voltage signals (sag, swell, interruption, transient, harmonic and normal) used in this study were recorded from real grids. DWT was used for featuring the extraction and calculation of the wavelet coefficients, and subsequently, calculated energy levels were used as an input to ANN. The results revealed analyzing the real data processed with DWT and ANN with 100% accuracy proved the superiority of this study. Based on the results of this study, identification of real-time PQ disturbances provided an important advantage for the firms and industry. Particularly, the reasons for the failures in the system related to PQ disturbances were simultaneously diagnosed, as well.