Graph Neural Networks for Efficient AC Power Flow Prediction in Power Grids
Seyedamirhossein Talebi, Kaixiong Zhou
TL;DR
The paper tackles the computational intractability of AC-OPF in large power grids by proposing Graph Neural Networks to predict voltage magnitudes and angles directly from grid topology. It models the grid as a graph and compares four GNN architectures (GCN, GATConv, SAGEConv, GraphConv) using IEEE test systems, with Newton-Raphson serving as ground-truth. Results show that GNNs can achieve high accuracy (NRMSE < 0.05, $R^2$ near 1) while offering substantial speedups, and SAGEConv and GraphConv demonstrate strong performance on larger grids. The work demonstrates potential for real-time grid management and outlines plans to scale to massive networks and to analyze the influence of generator buses on predictions.
Abstract
This paper proposes a novel approach using Graph Neural Networks (GNNs) to solve the AC Power Flow problem in power grids. AC OPF is essential for minimizing generation costs while meeting the operational constraints of the grid. Traditional solvers struggle with scalability, especially in large systems with renewable energy sources. Our approach models the power grid as a graph, where buses are nodes and transmission lines are edges. We explore different GNN architectures, including GCN, GAT, SAGEConv, and GraphConv to predict AC power flow solutions efficiently. Our experiments on IEEE test systems show that GNNs can accurately predict power flow solutions and scale to larger systems, outperforming traditional solvers in terms of computation time. This work highlights the potential of GNNs for real-time power grid management, with future plans to apply the model to even larger grid systems.