Benchmarking machine learning methods for power flow optimization
ML-OPF-Bench: Benchmarking Machine Learning for Optimal Power Flow
Machine Learning
Summary
Solving optimal power flow problems helps electrical grids run efficiently, but it takes time. The authors created a benchmark to fairly compare machine learning techniques that try to solve these problems faster. They found that simply predicting well doesn’t always mean the solutions are practical or safe. The benchmark also shows how well methods keep working when the grid conditions change and under different computing limits. This helps guide better design of machine learning tools for power grid optimization.
What this means in practice
- •For power system operators: Evaluate and select machine learning methods for faster and reliable optimal power flow computations under varying grid conditions and resource limits.
- •For energy software developers: Integrate and benchmark new learning-based optimal power flow algorithms against standard baselines within a consistent open-source Python framework.
Authors
Xinyi Liu, Xuan He, Danny H. K. Tsang, Yize Chen
Abstract
Machine Learning (ML) methods promise a fast solution process for Optimal Power Flow (OPF). While inconsistent test cases, implementations, and evaluation metrics across existing studies make it challenging to determine which algorithmic advances are most critical for real-world deployment. To this end, we propose ML-OPF-Bench, a unified benchmark for AC- and DC-OPF that evaluates representative ML algorithms under a consistent pipeline, stress-tests them across system sizes, distribution shifts, and resource budgets, and ranks them with a multi-objective framework. We find that prediction accuracy alone is not a reliable indicator of operational feasibility. Under heavily loaded, congested conditions, even the strongest in-distribution performers lose their advantage, while feasibility is maintained largely by post-processing that enforces the target constraints rather than by the underlying pure ML predictor. Data scaling shows that prediction accuracy and constraint violations follow different trajectories, whereas compute scaling shows that returns diminish and that larger models do not consistently perform better. These results expose critical trade-offs among ML methods' speed, accuracy, and feasibility, and offer practical guidance for future ML-OPF design. We open-source the benchmark as an extensible Python package for integrating new learning-based OPF algorithms and evaluating them under the same standard as the existing baselines.