Application of optimization techniques in development of new protection algorithms for MV distribution feeders

Application of new architectures of Artificial Neural Networks (ANN) in detection of arcing high impedance faults in MV distribution networks is presented in the paper. These new computational schemes can be used to reliable HIF detection in ungrounded, grounded using a Petersen coil and grounded via resistance distribution networks. The algorithms are tested using data obtained from computer simulations of the network modelled in EMTP-ATP. Small number of necessary neurons in a developed ANN, short measuring data window and easy to interpret output signals are the main advantages of the proposed approach. Satisfactory results of ANN performance were observed for all examined HIF cases.