Power disturbance identification through pattern recognition system

This paper presents an artificial intelligent system to identify and classify the power disturbance waveforms that are obtained from the monitoring system in a power control station. The pattern recognition technique used in this paper is a combination of Bayes' linear classifier and artificial neural network (ANN). Simulated disturbance waveforms are transformed by the fast Fourier transformation and the feature vector is extracted. The weight matrix for ANN is generated by the linear classifier and fed into ANN. The product of the test sample and the weight matrix will be the input of the ANN. The system can identify the power disturbance and it can provide the power surge frequency as well.