Power quality disturbance detection and classification using wavelets and artificial neural networks

This article develops a method to detect and classify power quality problems using a novel combination of digital filtering, wavelets, and artificial neural networks. The method is developed for voltage waveforms of arbitrary sampling rate and number of cycles, using a large variety of power quality events:simulated with MATLAB(R) software, in addition to sampled waveforms from utility monitoring and EMTP(R) simulations. Power system monitoring, augmented by the ability to automatically characterize disturbed signals, is a powerful tool for the power system engineer to use in addressing power quality issues. This is a step toward the goal of automating the real-time monitoring, detection and classification of power signals.