Contingency-constrained economic dispatch with safe reinforcement learning
Michael Eichelbeck, Hannah Markgraf, Matthias Althoff
TL;DR
The paper tackles contingency-aware economic dispatch in renewable-rich microgrids by integrating a provably safe reinforcement learning controller with a formal safety layer. It encodes a time-dependent islanding contingency through set-based backwards reachability and enforces safety via action projection into a safe set represented by constrained zonotopes, enabling real-time operation. The approach is demonstrated on a residential microgrid using real-world measurements, showing that the safe agent satisfies islanding constraints and yields higher performance than a baseline without full safety guarantees. This work provides certifiable safety for RL-based economic dispatch in zero-carbon power systems and offers a tractable framework for future extensions to more complex network models.
Abstract
Future power systems will rely heavily on micro grids with a high share of decentralised renewable energy sources and energy storage systems. The high complexity and uncertainty in this context might make conventional power dispatch strategies infeasible. Reinforcement-learning based (RL) controllers can address this challenge, however, cannot themselves provide safety guarantees, preventing their deployment in practice. To overcome this limitation, we propose a formally validated RL controller for economic dispatch. We extend conventional constraints by a time-dependent constraint encoding the islanding contingency. The contingency constraint is computed using set-based backwards reachability analysis and actions of the RL agent are verified through a safety layer. Unsafe actions are projected into the safe action space while leveraging constrained zonotope set representations for computational efficiency. The developed approach is demonstrated on a residential use case using real-world measurements.