Building a Neural Solver for the Grid with OpenGridFM

How can foundation models and neural solvers accelerate steady-state power-system analysis? In this hands-on tutorial, we will introduce OpenGridFM, an open-source framework for developing and deploying neural solvers for power-flow and optimal-power-flow applications.

Participants will work through interactive notebooks to build an end-to-end neural power-system workflow. Using gridfm-datakit, they will generate synthetic but realistic datasets with configurable grid cases, load and generator-dispatch variations, and topology perturbations. They will then use gridfm-graphkit to train and evaluate GENCO, a unified graph neural solver, in a low-code environment and compare its accuracy and computational performance with classical power-system solvers.

The tutorial is intended for power-system practitioners, researchers, and developers interested in applying machine learning to grid analysis, contributing to OpenGridFM, or integrating its tools into their own workflows. No prior experience with OpenGridFM is required; basic familiarity with Python and power-system concepts is recommended.