Fault Classificaton and Location in Power Distribution Networks using 1D CNN with Residual Learning

In this paper we use the difference of pre-fault and during fault phasor voltages expressed as symmetrical voltages to classify and locate fault in a power distribution network with distributed energy resources. We propose a ID deep learning model comprising residual learning blocks for classifying...

Full description

Saved in:
Bibliographic Details
Published in2024 IEEE Power & Energy Society Innovative Smart Grid Technologies Conference (ISGT) pp. 1 - 5
Main Authors Khattak, Khalid Daud, Choudhry, M.A, Feliachi, A.
Format Conference Proceeding
LanguageEnglish
Published IEEE 19.02.2024
Subjects
Convolutional neural networks
Convolutional Neural Networks (CNNs)
Deep learning
Distribution networks
Fault diagnosis
Fault Location
Load modeling
Neural networks
Power distribution networks
Power systems
Residual networks (ResNet)
Resistance
Smart grids
Training
Online AccessGet full text

Cover

More Information
Summary:In this paper we use the difference of pre-fault and during fault phasor voltages expressed as symmetrical voltages to classify and locate fault in a power distribution network with distributed energy resources. We propose a ID deep learning model comprising residual learning blocks for classifying faults and locating the faulted zone and faulted line. In the proposed model, locating the fault does not rely on first identifying the fault type. We have implemented the model on a modified, IEEE-34 node test feeder to assess the effectiveness of the proposed methodology. We have achieved more than 99% accuracy for fault type classification and faulted zone identification. We have also obtained more than 96% accuracy in identifying the faulted line in the test feeder.
ISSN:2472-8152
DOI:10.1109/ISGT59692.2024.10454243