Neural-network based adaptive single-pole autoreclosure technique for EHV transmission systems

Adaptive single-pole autoreclosure offers many advantages over conventional approaches. In this paper an adaptive single-pole autoreclosure technique is developed using artificial neural networks. The data generation, data preprocessing, and feature extraction process required to set up the training/test data for the neural network, and the implementation of the latter are described in detail. A nonfully connected three-layer perceptron is trained by the Extended-Delta-Bar-Delta learning algorithm. The test results demonstrate the ability of this network to distinguish reliably between permanent and transient faults, and in the latter case, the ability to determine the exact arc extinction time. The outcome of the study indicates that the neural network approach can be used as an attractive and effective means of realising an adaptive autoreclosure scheme