Research on the Application of Graph Neural Network in Topology Optimization of Distribution Network

Graph Neural Networks (GNN) is a type of neural network specifically designed for processing graph structured data, which is very suitable for topology optimization in power systems. GNN can effectively capture the relationships between nodes and edges in network topology through graph convolution operations, providing accurate representation and analysis capabilities of network structure. This article provides a detailed introduction to the construction process of distribution network topology models and graph neural network models, including the construction of datasets, model training and optimization methods, and the clarification of feature extraction methods. The effectiveness of GNN model in distribution network topology optimization was verified through case studies. By conducting in-depth discussions on the experimental results, analyzing the differences between the results and expectations, proposing directions for model improvement, and exploring future research directions and possibilities.

I. Introduction

The optimization of distribution network topology is of great significance for the stability, reliability, and economy of power systems ^[1]–[2], Through reasonable network configuration, the distribution network can quickly switch power supply paths in the event of faults, ensuring the continuity and stability of power supply, thereby improving power supply reliability ^[3]–[5]. Reasonable network topology optimization can balance load distribution, reduce voltage fluctuations and current imbalances, and improve power quality ^[6]–[8]. Through topology optimization, flexible adjustment of network structure can be achieved to cope with dynamic changes in load and access to distributed power sources, ensuring efficient operation of the distribution network. It can also achieve intelligent scheduling, automated management, and real-time monitoring of the distribution network, improve the intelligence level of the power grid, and enhance the emergency response capability of the distribution network.