Power distribution network weak feature fault identification method based on transfer learning

Abstract

The invention provides a power distribution network weak feature fault identification method based on transfer learning, and relates to the technical field of power grid detection. The identificationmethod comprises the following steps: firstly, establishing a 20kV neutral point ungrounded AC power distribution network model, setting two data acquisition points at an outgoing line, and constructing a training sample set and a transfer learning sample set; and training a sparse auto-encoder by using the training sample set to realize high accuracy of fault identification, and finally carryingout transfer learning of the network model by using a small number of transfer sample sets so that the accuracy of the algorithm model can reach 98% under a new topological structure. According to theinvention, in power distribution networks of different topology types, high-resistance fault type identification and fault line selection can be realized, and interference signals of capacitor switching and load switching can be distinguished; and the method is simple in principle, high in reliability, small in training sample number and high in generalization ability, and can realize weak feature fault identification in different power distribution network topologies.

Description

technical field [0001] The invention belongs to the technical field of power grid detection, and in particular relates to a method for identifying faults with weak characteristics of a power distribution network based on transfer learning. Background technique [0002] The distribution network has the characteristics of complex topology, flexible and changeable operation mode, and high probability of fault occurrence, among which single-phase ground fault has the highest probability of occurrence. In my country, after a single-phase ground fault occurs in the distribution network, it can continue to operate with the fault for 1-2 hours, but with the expansion of the distribution network scale, the leakage capacitive current of the distribution network to the ground increases significantly. If the system is allowed to operate with a fault for a long time, the excessive fault current is likely to cause the fault to develop further, posing a threat to the safe operation of the ...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a distribution network weak feature fault identification method based on transfer learning, which solves the problems of data hunger and poor environmental interaction in the existing power grid fault identification method

Images