The Research On Some Methods And Application Of Fault Location Of Distribution Feeders In Networks With Ungrounded Neutral

Fault location of distribution feeders in networks with ungrounded neutral is a well-known difficult problem in the world. Many experts and scholars are engaged in the research of the fault location methods and devices. The device already existed can't satisfy the practical need because of the characteristics of the fault signals and the limit of technique. This paper aims to purpose the method of fault location of distribution feeders in networks with ungrounded neutral and its realization.This paper firstly analyses the fault location methods already existed, and analyses the wavelet neural network specially. Then the paper makes simulation of the fault process of distribution feeders in networks with ungrounded neutral using the MATLAB, gets the characters and change rules of the system before and after the fault occur which is the theory basis of research. Because of the benefit of the SVM, we introduce the SVM, clustering into the fault location. Then proposes the fault location method based on clustering FSVM, the method is good at precision fault location in theory.Due to the characteristics of the networks of the system and the technique nowadays, this paper using the fuzzy theory and information fusion to propose the fault location method based on fuzzy information fusion. The method is good in fault location. We realize the method based on fuzzy information fusion into a device. We think of the complexity of locale, anti-jamming of the device and real time demands of the fault location device. This paper does a good job in hardware and software design. The device based on the proposed method is good at precision fault location, high credibility, high adaptability, easy to operation, full function and simple maintenance. Go well with tree network with several branches.