Design of PSS3B stabilizer using KH Algorithm and Q-Learning for damping Low-frequency Oscillations in SMIB

The main purpose of this paper is to develop a supplementary signal using reinforcement learning (RL) to improve the performance of power system stabilizer (PSS). RL is one of the most important issues in the field of artificial intelligence and is the popular method for solving Markov decision procedure (MDP). In this paper, a control method is developed based on Q-learning and used to improve the performance of a three band PSS (PSS3B) in a single machine infinite bus power system (SMIB). For this purpose, first the parameters of PSS3B are optimized using krill heard (KH) algorithm based on system eigenvalues. Then, using the proposed Q-learning method its performance will improve. The fundamental properties of the proposed Q-learning based control method are its simplicity and independency to system model and operational conditions. In order to evaluate the proposed control method, its dynamic response is compared to conventional PSS (CPSS) and PSS3B. According the simulation results, it is evident that, the developed adaptive controller is superb compared to the other methods in the view of settling time and damping low-frequency oscillations.