Load Frequency Control of Two Area Microgrid Using Reinforcement Learning Controller

Abstract

In this paper, automatic generation control problem is inspected and different modelling parameters are decided. Initially Genetic algorithm (GA) based parameters of proportional-integral-derivative (PID) controller is applied to the model. Then Q learning (Reinforcement learning) applied to control the frequency and tie line power deviation by controlling the power mismatch between generators and loads. Area control error (ACE) are used as objective function for PID and RL controller respectively. Q-learning based reinforcement agent is introduced which takes the action according to averaged ACE values. Parameters like step size, discount rate, and exploration rate decides the effectiveness of the RL scheme.