Transferable Graph Learning for Transmission Congestion Management via Busbar Splitting

2025-10-23

AI Summary

Bus splitting in network topology optimization (NTO) for large-scale power system congestion management poses significant computational challenges due to its mixed-integer nonlinear programming (MINLP) formulation, which is intractable for near-real-time solution; existing machine learning approaches suffer from poor generalization. Method: We propose a heterogeneous edge-aware graph neural network (HE-GNN) that directly learns bus-splitting policies. The model integrates linearized AC power flow constraints and structural features of the MINLP to construct a physics-informed candidate solution space. Contribution/Results: On the GOC 2000-bus system, HE-GNN achieves up to four orders-of-magnitude speedup, generating AC-feasible solutions within one minute with only 2.3% optimality gap. It is the first method to demonstrate strong transferability—delivering high-accuracy predictions across unseen topologies, multi-scale systems, and dynamic operating conditions—thereby significantly enhancing the practicality and robustness of NTO in real-world dispatch applications.

Technology Category

• Search and Optimization: Heuristic Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

Network topology optimization (NTO) via busbar splitting can mitigate transmission grid congestion and reduce redispatch costs. However, solving this mixed-integer non-linear problem for large-scale systems in near-real-time is currently intractable with existing solvers. Machine learning (ML) approaches have emerged as a promising alternative, but they have limited generalization to unseen topologies, varying operating conditions, and different systems, which limits their practical applicability. This paper formulates NTO for congestion management problem considering linearized AC PF, and proposes a graph neural network (GNN)-accelerated approach. We develop a heterogeneous edge-aware message passing NN to predict effective busbar splitting actions as candidate NTO solutions. The proposed GNN captures local flow patterns, achieves generalization to unseen topology changes, and improves transferability across systems. Case studies show up to 4 orders-of-magnitude speed-up, delivering AC-feasible solutions within one minute and a 2.3% optimality gap on the GOC 2000-bus system. These results demonstrate a significant step toward near-real-time NTO for large-scale systems with topology and cross-system generalization.

Problem

• Solving transmission grid congestion via busbar splitting optimization
• Overcoming computational intractability in large-scale network topology optimization
• Improving machine learning generalization across topologies and operating conditions

Innovation

• Graph neural network predicts busbar splitting actions
• Heterogeneous edge-aware message passing captures flow patterns
• Achieves topology generalization and cross-system transferability