UCSD

Mechanical and Aerospace Engineering

• Home
• Publications
• Group
• Teaching
• Software
• Lab

---

Learning decentralized frequency controllers for energy storage systems
Z. Sun, Z. Yuan, C. Zhao, J. Cortés
IEEE Control Systems Letters 7 (2023), 3459-3464

Abstract

With the growth of volatile renewable generation, it becomes an increasingly critical and challenging task to maintain the frequency stability of a power system under disturbance. An energy storage system (ESS) is a great resource to support frequency control due to its fast charging and discharging capabilities. In this paper, we design decentralized controllers for ESSs to provide active power control for frequency regulation. We prove that the proposed controller satisfies the state of charge (SOC) restriction of each ESS, and that the closed-loop system converges to a desired equilibrium. Moreover, the controller is parameterized and trained through a reinforcement learning framework to optimize transient performance in terms of the maximum frequency deviation and the cost of control. Simulations of the IEEE 39-bus network verify the proposed controller design with significant transient performance improvement.

pdf

Mechanical and Aerospace Engineering, University of California, San Diego
9500 Gilman Dr, La Jolla, California, 92093-0411

Ph: 1-858-822-7930
Fax: 1-858-822-3107

cortes at ucsd.edu
Skype id: jorgilliyo