Control of Multiple Voltage Control Devices in a Distribution System using Double DQN

Many photovoltaic (PV) systems are installed in the distribution system, which complicates the voltage profile; voltage fluctuations due to PV cause frequent LRT and SVR operations and shorten equipment life. Therefore, we believe that voltage control can be achieved by understanding the voltage distribution in detail by estimating the state based on the information from the currently installed measurement devices. Using machine learning from the collected information, high-speed control can be achieved. Therefore, we propose a method to perform voltage control while considering the wear and tear of the devices using double DQN.