Single-phase-to-earth fault line selection method for distribution network based on convolutional neural network

Abstract

The invention relates to a method for selecting single-phase ground fault lines of power distribution networks on the basis of convolutional neural networks. The method includes acquiring zero-sequence voltages of busbars and zero-sequence current signals of various feed lines; carrying out continuous wavelet transformation on the various zero-sequence current signals according to set decomposition scales; acquiring time-scale wavelet coefficient gray-level graphs; identifying fault feed lines by the aid of trained convolutional neural network algorithms. The method for selecting the single-phase ground fault lines of the power distribution networks on the basis of the convolutional neural networks has the advantage that the fault feed lines and the non-fault feed lines can be accurately identified by the aid of the method under various fault working conditions when single-phase ground faults occur.

Description

technical field [0001] The invention relates to the field of distribution networks, in particular to a convolutional neural network-based single-phase ground fault line selection method for distribution networks. Background technique [0002] With the improvement of people's living standards and the continuous development of social and economic construction, people's demand for electricity is increasing day by day. As an important part of connecting the transmission system and users, the power distribution system is safe and stable. There is a direct impact on reliability and utility profitability. The probability of a single-phase ground fault in the distribution network is as high as 80%. With the complexity of the distribution network structure, the system lines (including cable lines and cable-line hybrid lines) in the power grid are gradually increasing, resulting in an increase in the distributed capacitance of the system to ground. , the capacitive current also incre...

AI Technical Summary

Problems solved by technology

Different fault lines, different fault locations, different grounding resistances, different fault closing angles and other fault conditions will affect the size and shape of the transient zero-sequence current, thereby affecting the accuracy of fault line selection

It is often necessary to seek multiple feature quantities to characterize the characteristic mode of the single-phase ground fault signal to achieve the purpose of identification, and the classification algorithms applied to the identification of single-phase ground fault feeders in the distribution network mainly use relatively mature machine learning algorithms , these algorithms do not have the ability of self-learning

Images