Discrimination of Power Quality disturbances using combined mathematical transforms and Artificial Neural Network

With the increased use of non-linear load and sensitive electronic equipment, power quality (PQ) monitoring has become an important issue in power system. This paper presents an artificial neural network (ANN) based approach for PQ monitoring. The signals with power quality disturbance are transformed to time-frequency domain using mathematical transforms such as Fourier transform and wavelet transform. The input features of the network are extracted from the transformed signal. The extracted features after normalization are given to a feedforward neural network trained by the backpropagation algorithm. The training and testing data required to develop the ANN are generated through simulation. The combined mathematical transformation and artificial neural network-based approach is able to classify the power quality disturbances accurately.