A Q-Learning Based Approach For Voltage Control In Unbalanced Distribution System

In recent years, with the increasing use of renewable energy sources, voltage regulation in power distribution systems has become a greater challenge due to these sources’ intermittent and variable nature. To address this problem, this paper proposes a Q-learning-based method for voltage regulation in power distribution systems that incorporates tap changer voltage regulators when renewable energy sources are present. The voltage deviation problem is formulated as an optimization problem, and Q-learning is used to find near-optimal solutions that can effectively control voltage deviations and improve the overall performance of the power distribution system. The proposed method is applied to power distribution system using OpenDSS incorporated into Python. The simulations are performed on base case IEEE-123 Bus distribution network and with modifications by adding Photovoltaics. From the simulation results we observed that the proposed algorithm is able to solve the overvoltage and undervoltage problems.