A Wavelet-ANN Technique for Locating Switched Capacitors in Distribution Systems

With an increase in usage of sensitive electronic loads and adjustable speed drives, utilities and customers have paid more attention to power-quality (PQ) problems. Shunt capacitor switching is one of the PQ problems that causes a wide band of high-frequency transients in voltage and current signals which harmfully affect these types of sensitive loads. This paper presents an efficient technique for locating switched capacitors in the industrial distribution systems using discrete wavelet transform (DWT) integrated with a feedforward artificial neural network (FFANN). The technique relays on combining the three-phase currents at the customer side to construct one modal current signal, and then the DWT is used to capture the high-frequency current transients contained in this modal signal. The energy contained in the high-frequency current transients and the transient duration, converted into the number of samples, for two DWT decomposition levels, is used to feed an FFANN to classify the switched capacitor. The simulations are performed using PSCAD/EMTDC and the results are then interfaced to MATLAB, where the technique is implemented. It is found that the technique gives reliable results for locating switched capacitors.