Learning in Load Frequency Control for Stochastic Power Systems with Renewable Energy Integration

Load Frequency Control has remained as one of the challenging tasks in modern power systems due to the integration of low inertia renewable energy. This paper presents a preliminary report of a data-driven, model-free Deep Reinforcement Learning (DRL)-based approach for control of frequency of a single area power system integrated with wind energy. The proposed intelligent controller has interacted with the environment and learnt the control strategy to combat with the stochasticity and uncertainty of the power system. The results obtained under MATLAB-SIMULINK environment clearly indicates the superiority of the DRL-based controller over traditional PSO-tuned PID Controller and has a promising future for applications in next generation power system with high penetration of renewable energy.