GENCO: A Unified Neural Solver for Power Flow, Optimal Power Flow, and State Estimation

Forum topic · 小凯 · 2026-08-12

Summary

GENCO (GEometric Neural Corrective Optimizer), presented by Alban Puech, Matteo Mazzonelli, and Tamara R. Govindasamy (arXiv:2508.05152), is a unified neural solver for steady-state transmission grid analysis that handles power flow (PF), optimal power flow (OPF), and state estimation (SE) within a single architecture. The authors also release the open-source GridFM Development Framework, which standardizes synthetic data generation and training in a low-code environment, plus large-scale datasets with millions of PF and OPF scenarios across diverse grid topologies for reproducible benchmarking. Evaluated on PFDelta and OPFData benchmarks and validated on real Hydro-Québec SCADA data, GENCO recovers full AC operating states for large-scale PF with DC-PF-level active power balance residuals, running 30x faster than Newton-Raphson and only 2x slower than DC-PF. For OPF, it is 85x faster than IPOPT while outperforming DC-OPF on feasibility, optimality, and runtime. For SE, it is more robust than classical weighted least squares to noisy measurements and network parameter errors.

Paper Overview

Field: Machine Learning
Authors: Alban Puech, Matteo Mazzonelli, Tamara R. Govindasamy
Published: 2026-08-12
arXiv: 2508.05152

Abstract

Foundation models are transforming business workflows and boosting productivity, yet they remain largely absent from engineering domains such as power system analysis, where strict physical consistency must be enforced. GENCO (GEometric Neural Corrective Optimizer) addresses this gap as a unified neural solver for steady-state transmission grid analysis, handling power flow (PF), optimal power flow (OPF), and state estimation (SE) within a single architecture and shared network representation.

To support advances in neural power system solvers, the authors introduce the open-source GridFM Development Framework, which standardizes synthetic data generation and training in a low-code environment. They also release large-scale datasets with millions of PF and OPF scenarios across diverse grid topologies to support reproducible benchmarking.

Key Results

GENCO was evaluated on the PFDelta and OPFData benchmarks against state-of-the-art neural solvers and classical solvers (including Newton-Raphson and IPOPT), and validated on real Hydro-Québec SCADA data.

Power Flow: For large-scale PF, GENCO recovers the full AC operating state—including voltage magnitudes and reactive power that DC-PF cannot provide—while achieving DC-PF-level active power balance residuals. It is 30x faster than Newton-Raphson and only 2x the runtime of DC-PF.
Optimal Power Flow: GENCO is 85x faster than IPOPT while outperforming DC-OPF on feasibility, optimality, and runtime.
State Estimation: GENCO is more robust than classical weighted least squares to noisy measurements and network parameter errors, consistently returning high-quality estimates even when weighted least squares fails to converge.

Significance

The unified architecture and development framework together offer a new approach to large-scale steady-state grid analysis, lower the entry barrier for power system engineers, and mark an important step toward foundation models for the grid.

--- *Auto-collected on 2026-08-12*