Power quality detection with classification enhancible wavelet-probabilistic network in a power system

A model of disturbance detection for harmonics and voltages using wavelet-probabilistic network (WPN) is proposed, which is a two-layer architecture, containing the wavelet layer and probabilistic network. It uses the wavelet transformation (WT) to extract the features from various disturbances and probabilistic neural network (PNN) to analyse the translation patterns from time-domain distorted wave and perform classification tasks. The proposed WPN detects the disturbances of harmonics and voltages, and has been tested for the power quality problems caused by harmonics, voltage sag, voltage swell and voltage interruption. It has also been compared with wavelet networks as well as combined the WT and conventional neural networks. The test results show that this simplified network architecture enhances the classification performance and shortens the processing time for detecting disturbing events.