TY - EJOU AU - Tan, Shuxin AU - Yan, Wei AU - Zhao, Lei AU - Zhang, Xianglin AU - Man, Ziqiang AU - Lu, Yu AU - Liu, Teng AU - Zhong, Gaoyue AU - Liu, Weiqun AU - Shi, Linjun TI - Enhancement of Frequency Regulation in AC-Excited Adjustable-Speed Pumped Storage Units during Pumping Operations T2 - Energy Engineering PY - 2025 VL - 122 IS - 12 SN - 1546-0118 AB - The integration of large-scale renewable energy introduces frequency instability challenges due to inherent intermittency. While doubly-fed pumped storage units (DFPSUs) offer frequency regulation potential in pumping mode, conventional strategies fail to address hydraulic-mechanical coupling dynamics and operational constraints, limiting their effectiveness. This paper presents an innovative primary frequency control strategy for double-fed pumped storage units (DFPSUs) operating in pumping mode, integrating an adaptive parameter calculation method. This method is constrained by operational speed and power limits, addressing key performance factors. A dynamic model that incorporates the reversible pump-turbine characteristics is developed to translate frequency deviations into coordinated adjustments in speed and power during pumping operations. The research thoroughly analyzes the influence of control parameters on the frequency response dynamics. Additionally, the paper introduces a deep reinforcement learning (DRL)-based optimization framework, which enables real-time tuning of control parameters in response to changing rotor speed and frequency states. This method strategically manages the utilization of kinetic energy while ensuring compliance with operational safety constraints. The effectiveness of the proposed strategy is validated through simulation studies conducted on a four-machine, two-area DFPSU system. These studies demonstrate the strategy’s potential for improving frequency regulation performance under a variety of operating conditions, highlighting its effectiveness in optimizing energy storage and frequency control in power grids. KW - Doubly-fed pumped storage unit; primary frequency control; adaptive control; variable coefficient; deep reinforcement learning; smart grid DO - 10.32604/ee.2025.068692