Fault Diagnosis Combining Power Grid and Communication System Based on Graph Neural Network
Lingfei Qian, Zixin Zhou, Xinzhong Cai, Pengyu Zhu · 2023
In order to accelerate digital development, the power grid needs to enhance its intelligent regulation capabilities to achieve digital transformation. The important breakthrough point is to achieve joint regulation of communication and power grid, and one of the key points is to research and implement fault boundary determination by combining the power grid and communication network. This article refers to the characteristics, historical experience, and data of the power grid and communication network, proposes a digital joint strategy for the power grid and communication network, and constructs the corresponding network topology. A GCN model has been built based on graph neural network theory and PyTorch Geometric (PyG) framework, and a practical case study has been completed to achieve the research goal of “binary classification of fault location corresponding to fault alarm in the power grid or communication network”. Excellent performance has been achieved, which is conducive to the intelligent management and regulation of the power grid and communication network, and can promote the digital regulation and operation of the backbone optical communication system in the large power grid.