Towards Systematic Generalization for Power Grid Optimization Problems

2026-05-03

AI Summary

This work addresses the limited generalizability of traditional approaches that solve AC optimal power flow (ACOPF) and security-constrained unit commitment (SCUC) in isolation, particularly across varying grid topologies. The authors propose the first unified learning framework that jointly models power grid topology and physical interactions through a shared graph neural network backbone, complemented by task-specific decoders for static and time-coupled decision variables. The framework incorporates solver-supervised training, physics-informed unsupervised consistency scheduling, and a transfer mechanism enabling cross-case generalization without retraining. Extensive evaluations on multi-scale power systems demonstrate substantial performance gains over existing learning-based baselines, confirming the method’s effectiveness and robustness in complex power system optimization.

Abstract

AC Optimal Power Flow (ACOPF) and Security-Constrained Unit Commitment (SCUC) are fundamental optimization problems in power system operations. ACOPF serves as the physical backbone of grid simulation and real-time operation, enforcing nonlinear power flow feasibility and network limits, while SCUC represents a core market-level decision process that schedules generation under operational and security constraints. Although these problems share the same underlying transmission network and physical laws, they differ in decision variables and temporal coupling, and prior learning-based approaches address them in isolation, resulting in disjoint models and representations.We propose a learning framework that jointly models ACOPF and SCUC through a shared graph-based backbone that captures grid topology and physical interactions, coupled with task-specific decoders for static and temporal decision-making. Training includes solver supervision with physics-informed objectives to enforce AC feasibility and inter-temporal operational constraints. To evaluate generalization, we assess cross-case transfer on unseen grid topologies for ACOPF and SCUC without retraining, and systematic generalization on the UC-ACOPF problem using unsupervised, physics-based objectives and a power-dispatch consensus mechanism. Experiments across multiple grid scales demonstrate improved performance and transferability relative to existing learning-based baselines, indicating that the model can support learning across heterogeneous power system optimization problems.

Problem

• AC Optimal Power Flow
• Security-Constrained Unit Commitment
• Systematic Generalization
• Power Grid Optimization
• Cross-case Transfer

Innovation

• systematic generalization
• graph-based learning
• physics-informed neural networks
• AC optimal power flow
• security-constrained unit commitment