Fault Detection in Complex Distribution Network Based on Hilbert-Huang Transform

Traditional distribution network fault location methods often cannot be effectively applied for the structure of the branch in complex distribution network. A new accurate fault location for the single-phaseground fault in complex distribution network with structure of the branch based on Hilbert-Huang transform was proposed in this paper. First, the distribution network was modeled. The faults on each branch were simulated. The energy characteristics under the branch in a particular frequency band were identified by HHT. Then these energy characteristics were used to train artificial neural networks (ANN).When the energy characteristics of actual fault are inputted, the trained neural network can output the malfunction branch. When the fault branch was determined, using the online fault feature matching method, combined with the genetic algorithm, the precise determination of the distance to fault location in the fault branch can be completed. With combinations of signal processing-Hilbert-Huang transform, artificial neural network and genetic algorithm, the entirely new method was proposed to deal with the problem of fault location in distribution network in this article. The results showed that the method has a good precision and apply to the small current grounding system.