Research on fault line detection for distribution network based on improved fuzzy neural networks algorithm

In the distribution network, neutral ineffectively grounding system is widely used. The applicable scope of traditional single phase fault ground line selection is limited. Various methods of earth fault line selection are combined using fuzzy neural networks(FNN) against the shortcoming of single fault line selection method. Steady-state fundamental component, steady-state fifth harmonic component, transient energy component and transient direction component extracted from zero-sequence current separately by means of FFT and wavelet packet are used as fault features to perform fault line selection. The structure of fuzzy neural networks (FNN) is designed and improved. Back-propagation(BP) algorithm is adopted as training algorithm. At last, a 10 kV distribution networks simulation model is set up by Matlab7.1. Fault ground types, fault location, fault close initial angles, and fault lines are simulated to demonstrate the feasibility of the theory.