Dates: 2022
Principal Investigator: Dr. Maad Al Owaifeer
Funded by: SDAIA-KFUPM Joint Research Center for Artificial Intelligence

Description: This research addresses the challenges of operating power grids with high renewable energy variability and the growing number of distributed controllable resources (e.g., storage, EVs, smart appliances). Current methods for solving the Optimal Power Flow (OPF) problem are too slow, leading to unreliable or suboptimal grid operation. The project proposes a distributed AI-based optimization framework, where multiple agents compute set points locally without relying on a central controller. The framework uses two neural networks: a Power Flow NN (supervised, modeling grid physics) and an OPF NN (unsupervised, for optimization), with Graph Neural Networks (GNNs) enabling distributed learning across the grid’s network structure. This approach aims to achieve faster, reliable, and cost-effective grid operation under renewable and distributed resource uncertainty.