A Brief Study on Power Restoration of Distributive Networks

Abstract

The method and program complex realisation for restoration scheme search of power consumption in networks based on algorithm processing of network count and artificial neural networks (ANN) are proposed and justified. Both algorithms operate in a competitive mode, smoothing out the shortcomings of each individual method. Work complicated concepts, in particular, concerning application of the generalized error vector for ANN self-training and solution stability are theoretically proved. The software complex developed as part of the 'advisor of the distributive network dispatcher' was tested on a distribution network with 201 buses, 7 sources, and 227 lines and yielded positive results.