Open AccessSingle-phase ground fault identification method for distribution network based on inception model and sample expansion
Author(s) -
Jizhong Zhu,
Kai Chen,
Zhonghan Peng,
Hailong Zhang,
Xingchen Wan,
Dahu Zhu
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1656/1/012008
Subject(s) - sample (material) , identification (biology) , fault (geology) , phase (matter) , network model , computer science , data mining , geology , physics , botany , quantum mechanics , seismology , biology , thermodynamics
A single-phase ground fault identification method based on inception model and sample expansion was proposed in this paper. Based on the basic characteristics of the inception model, a hybrid inception model structure is constructed to enhance the width and depth of the network and reduce network parameters. Combined with the training sample expansion method, the model is fully trained to adapt to the single-phase ground-fault current characteristics of the distribution network. The experimental results show that the accuracy for the hybrid inception model structure to the single-phase ground-fault current of the distribution network is 94.21%. After the sample is expanded, the recognition accuracy rate can be further improved by 97.52%.