Switching detection/classification using discrete wavelet transform and self-organizing mapping network

The transient caused by the load/capacitor switching is one of the current important power quality (PQ) problems. Especially, the capacitor switching on may lead to the system parallel resonance. The PQ monitoring, on the other hand, is addressed to identify different PQ phenomena. With the help of monitoring result, the PQ engineers may adopt proper control strategies. In this paper, the discrete wavelet transform is used to extract the features of transients caused by the load/capacitor switching. The wavelet coefficients are then served as inputs to the hybrid self-organizing mapping neural network for detecting/identifying switching types and phase angles. The simulation results obtained from a distribution system show the applicability of the proposed method.