Multi-Timescale Voltage and Reactive Power Control Method Combining Fast PV PCS and Slow Transformer Tap Response Using Deep Reinforcement Learning

Authors:
Publisher website
Google Scholar
Abstract
In maintaining the power system voltages, one of the major voltage regulators, the tap-changing transformers in a substation, play an important role during relatively slow load changes. In recent years, as a countermeasure against global warming, the introduction of renewable energy such as photovoltaic power generation system (PV) into the power system has progressed rapidly, and along with this, the amount of net load fluctuations has also changed rapidly as most of PV are connected on the customer side. As a result, it is becoming difficult to maintain the system voltages by controlling the low-speed tap change of transformers only. To suppress fast voltage fluctuations caused by PV output fluctuations, a lot of research work on reactive power control using power conditioning system (PCS) of PV is underway. However, it is difficult to coordinate high-speed PCS reactive power control with low-speed tap change of transformers. In this paper, the authors have proposed a multi-timescale control method that can coordinate low-speed tap change of transformers and high-speed PCS power factor (PF) based reactive power control using deep reinforcement learning (DRL). The effectiveness of the proposed method is verified by numerical simulation using a power system model that considers a large-scale PV interconnection.

This publication has 7 references indexed in Scilit: