arXiv · 2611.14716
A Three-Stage Unsupervised Online Physics-Informed Neural Network for Microgrid Optimal Power Flow
Abstract
Optimal allocation of generators is a key challenge in power systems to achieve economical dispatch while satisfying operational constraints. Conventional optimization techniques can be computationally expensive, limiting their applicability in real-time environments, and neural networks often require offline retraining under new operating conditions. This paper proposes a three-stage unsupervised online physics-informed neural network (PINN) framework for optimal power flow (OPF). Simulation results demonstrate that the proposed framework satisfies load demand and system constraints while achieving a $43,665 lower operating cost than fmincon-based optimization under the same operating conditions. The training process analysis further demonstrate the framework's ability to adapt its parameters online to variations in load demand and renewable generation in real time.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Abinash Giri, Yi Hu. 2026-11-10. A Three-Stage Unsupervised Online Physics-Informed Neural Network for Microgrid Optimal Power Flow. https://searcharxiv.com/abs/2611.14716
Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Searcharxiv by Indiciae: BibTeX · RIS.