Arcing fault identification using combined Gabor Transform-neural network for transmission lines

This paper presents an intelligent identification scheme for transient faults in transmission systems using Gabor Transform (GT) and Artificial Neural Network (ANN). The successful discrimination between arcing and permanent faults can be then utilized for realize a reliable operation of autoreclosure systems. The proposed algorithm employs the GT as a signal processing technique and the ANN for pattern recognition and classification processes. The use of GT is motivated by the fact that the Gabor elementary functions have distinctive an optimal localization property in the joint time and frequency domains, which leads to an optimal feature extraction. The extracted GT coefficients are used as the inputs to a three layer feedforward ANN. The generalization capabilities of neural networks together with the GT are expected to discriminate between arcing and permanent fault cases successfully. The fault behavior is simulated by ATP/EMTP where the arc model is realized using universal arc representation. Finally, the possibility of hardware implementation of the proposed scheme is visualized in order to verify its practicality and suitability for real field operation. The results show that combining of GT along with ANNs achieves an excellent performance to discriminate between arcing and permanent faults with eliminating the impacts of fault resistance, fault location and fault inception angle as compared with conventional discriminators.