Electric power quality disturbance classification using self-adapting artificial neural networks

Power quality is recognised as an essential feature of a successful electric power system mainly due to the rapid increase of loads which generate noise and, at the same time, are sensitive to the noise present in the supply system. A technique for classifying electrical power quality disturbance events is presented, based on a self-adapting artificial neural network (SAANN), which has the unique capability of adapting to new disturbance features. In the proposed technique, distinctive feature vectors from disturbance events captured are extracted using the fast fourier and discrete wavelet transforms. The feature vectors are then fed to two SAANN-based classifiers, which classify the captured events into different categories of power quality disturbances. The technique is tested using a number of disturbance events.