Fault Location Of Distribution Network Based On Neural Network And Reclosing Current Detection

Power distribution automation system has been generalized in Guangzhou power grid for years.Its functions and framework had been improved ever since.Among which the overhead lines automation deployment was started earlier and thus basically meets the demands of deploy quantity and network density.The overhead line power distribution system started in a time when wireless communication is not widely used,some of the pole-mounted switches' communication function are absent and unable to upload data.When it a fault occurs on a distribution line,we can't tell where the fault locates,for these switches have no means to report their status,even through they could have acted correctly and separate the fault.To solve this problem,this article suggests a new way of fault section positioning,which no longer needs to talk to automatic switches.And choose neural network tool as the way to realized it after doing a research on related papers.This paper uses the Simulink tool of MATLAB to simulate the reclosure process of power distribution network,and successfully use the neural network tool from MATLAB to train the neural network and generate a neural network function which successfully judges the fault section.Limitations,applicability and reliability of this method are also discussed in this paper.