A Novel Deep Reinforcement Learning Enabled Multi-Band PSS for Multi-Mode Oscillation Control
Guozhou Zhang, Weihao Hu, Junbo Zhao, Di Cao, Zhe Chen, Frede Blaabjerg
Abstract
To better damp out the multi-mode oscillations in an uncertain environment, a novel multi-band power system stabilizer (MBPSS) is proposed. Compared with other MBPSSs, the proposed controller has a well-balanced structure, and each band is designed to address a target low-frequency oscillation (LFO) mode. A deep reinforcement learning-enabled agent is developed to effectively tune the control parameters that are adaptive to system uncertainties and different operating conditions. Comparative results with other types of PSSs on the IEEE 68-bus system demonstrate that the proposed method has better performance of damping out LFO and robustness against unseen operating conditions and faults.