A Method of Fault Judgment and Fault Phase Selection Based on Convolutional Neural Network

Abstract

The invention relates to a transmission line area internal and external fault judgment and fault phase selection method based on a convolution neural network. Firstly, a simulation model is established according to a double-end power supply system principle diagram, the input of a training sample is obtained, the fault type and area internal and external faults are used as the output of the training sample, likewise, a test sample is generated. Secondly, a network structure is given, and an optimal network structure is obtained by testing the error rate of the sample. Finally, after each time of fault, fault current is obtained and input into a trained network, and whether the fault is an internal fault or not can be judged and the fault phase selection is carried out, and retraining is not needed. According to the method, the convolution neural network output is improved, while a same network is used, the two types of non-independent classification problems of area internal and external fault judgment and fault phase selection are solved at the same time, and the weight sharing of two types of non-independent classification problems is realized. The sampling rate requirement is low, the calculation of various setting values is not needed, the method is not influenced by a system frequency, a fault location, load current, transition resistance and other factors, and a result is accurate and reliable.

Description

technical field [0001] The invention relates to a method for judging internal and external faults of a power transmission line and selecting a fault phase, in particular to a method for judging internal and external faults of a power transmission line and selecting a fault phase based on a convolutional neural network. Background technique [0002] In the power system, fault judgment inside and outside the transmission line area and fault phase selection are essential links for relay protection. The judgment of faults inside and outside the area is of great significance to the selectivity of relay protection devices. Reliable and fast fault phase selection is the key to accurate action of protection. , An important prerequisite for correct fault handling, which is of great significance to the stability of the power system. The judgment of faults inside and outside the transmission line area is mainly realized through the longitudinal protection, which is mainly divided into ...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to overcome the deficiencies of the existing fault judgment and fault phase selection methods inside and outside the zone, and provide a method that uses the same network to solve both internal and external fault judgment and fault phase selection methods by improving the output part of the convolutional neural network. Independent classification method, which requires low sampling rate, does not need to calculate various setting values, is not affected by system frequency, fault location, load current, transition resistance and other factors, the result is accurate and reliable, has no special requirements for equipment, and is convenient for implementation

Images