Federated Physics-Grounded Reinforcement Learning for Distributed Stability Control in Smart Grids

2026-07-06

AI Summary

This work addresses transient stability control in smart grids by proposing a decentralized cooperative control method based on federated multi-agent reinforcement learning. The approach identifies physically strongly coupled neighborhoods via Kron-reduced susceptance matrices to construct local observations that embed the grid’s topology. It initializes policies using a classical decentralized controller and employs a federated multi-agent Proximal Policy Optimization (PPO) algorithm within a centralized training with decentralized execution (CTDE) framework to achieve efficient coordination. Evaluated on the IEEE 39-bus system, the method achieves 100% transient stability across eight fault scenarios, reduces average stabilization time by 72.4%, and decreases control power consumption by 7–14 times, all while meeting IEEE/IEC real-time requirements and operating without any central coordinator.

Abstract

Transient stability control in smart grids requires rapid post-fault damping of generator frequency and rotor angle deviations to prevent cascading failures. This paper proposes FedPPO-PG, a Federated Multi-Agent Proximal Policy Optimization framework with Physics-Grounded neighborhoods, which reformulates transient stability control as a cooperative multi-agent reinforcement learning problem optimized directly against closed-loop stability objectives. Each generator hosts an independent local actor augmented with the frequency deviations of its two most strongly coupled electrical neighbors, identified from the post-fault Kron-reduced susceptance matrix. A guided policy initialization phase warm-starts all actors from the classical decentralized controller, while a centralized critic guides advantage estimation under the centralized training--decentralized execution (CTDE) paradigm. Evaluated on a simulation of the IEEE 39-bus benchmark system across five training and three unseen fault contingencies, FedPPO-PG achieves 100% stabilization in all 24 trials, reduces mean stability time by 72.4%, and cuts the control power by 7-14 times compared to the centralized baseline. Each actor executes independently with no central coordinator at deployment, and the per-actor inference latency satisfies the IEEE/IEC 60255-118-1-2018 real-time reporting requirements.

Problem

• Transient stability control
• Smart grids
• Distributed control
• Cascading failures
• Frequency deviation

Innovation

• Federated Reinforcement Learning
• Physics-Grounded Neighborhoods
• Transient Stability Control
• Multi-Agent PPO
• Decentralized Execution