1-D Convolutional Graph Convolutional Networks for Fault Detection in Distributed Energy Systems

TL;DR

This paper introduces a 1-D convolutional graph neural network that effectively detects faults, classifies fault types and phases, and locates faults in microgrids with high accuracy, leveraging spatial-temporal data and transfer learning.

Contribution

It proposes a novel 1-D convolutional graph neural network combining CNN and GCN for comprehensive fault detection and classification in microgrids, with transfer learning to reduce training effort.

Findings

Achieves over 99% accuracy in fault detection

High accuracy in fault classification and location

Outperforms traditional ANN models on microgrid data

Abstract

This paper presents a 1-D convolutional graph neural network for fault detection in microgrids. The combination of 1-D convolutional neural networks (1D-CNN) and graph convolutional networks (GCN) helps extract both spatial-temporal correlations from the voltage measurements in microgrids. The fault detection scheme includes fault event detection, fault type and phase classification, and fault location. There are five neural network model training to handle these tasks. Transfer learning and fine-tuning are applied to reduce training efforts. The combined recurrent graph convolutional neural networks (1D-CGCN) is compared with the traditional ANN structure on the Potsdam 13-bus microgrid dataset. The achievable accuracy of 99.27%, 98.1%, 98.75%, and 95.6% for fault detection, fault type classification, fault phase identification, and fault location respectively.