Deep Reinforcement Learning-Based Controller for SOC Management of Multi-Electrical Energy Storage System

The ongoing reduction of the total rotational inertia in modern power systems brings about faster frequency dynamics that must be limited to maintain a secure and economical operation. Electrical energy storage systems (EESSs) have become increasingly attractive to provide fast frequency response services due to their response times. However, proper management of their finite energy reserves is required to ensure timely and secure operation. This paper proposes a deep reinforcement learning (DRL) based controller to manage the state of charge (SOC) of a Multi-EESS (M-EESS), providing frequency response services to the power grid. The proposed DRL agent is trained using an actor-critic method called Deep Deterministic Policy Gradients (DDPG) that allows for continuous action and smoother SOC control of the M-EESS. Deep neural networks (DNNs) are used to represent the actor and critic policies. The proposed strategy comprises granting the agent a constant reward for each time step that the SOC is within a specific band of its target value combined with a substantial penalty if the SOC reaches its minimum or maximum allowable values. The proposed controller is compared to benchmark DRL methods and other control techniques, i.e., Fuzzy Logic and a traditional PID control. Simulation results show the effectiveness of the proposed approach.