Tag
This paper presents a scalable heterogeneous graph neural network workflow for data-driven optimal power flow surrogate modeling, using distributed training on supercomputers and demonstrating improvements via fine-tuning pretrained models.
Microsoft open-sourced GridSFM, an AI foundation model for power systems. It uses graph neural networks to approximate AC-OPF solving, is topology-agnostic, and can serve as a warm start for exact solvers achieving a 1.45x speedup, while also providing feasibility classification capabilities.