Determination of fault types and locations in power systems by using intelligent systems

In this study, an approach based on discrete wavelet transform, support vector machines and radial basis neural networks is presented for estimating fault types and location on transmission lines. The current and voltage values measured from only the sending end of transmission line are used for prediction of fault types and locations. The data obtained from prototype power system and from ATP simulations are analyzed by using discrete wavelet transform. Thus, the distinctive features of fault transients are obtained. The wavelet entropy criterion is employed to wavelet detail coefficients for reducing the size of feature vector. Support vector machines which are based on statistical learning theory are used to classify different fault types such as single-phase to ground, two-phase, two-phase to ground and threephase symmetrical faults. Radial basis neural networks which have a simple network architecture and shorter training time than the other neural network types are used for locating faults on transmission line. The various test cases including different fault resistances, pre-fault loads and line in feed are investigated. The obtained results show that the proposed approach is very effective in prediction of fault types and locations. Keywords: Transmission line faults, discrete wavelet transform, support vector machines, radial basis neural networks.