Department
Electrical and Computer Engineering
Major
Electrical Engineering
Research Advisor
Kimball, Jonathan W.
Advisor's Department
Electrical and Computer Engineering
Abstract
This research explores the fusion of deep reinforcement learning and virtual synchronous generator control in a bid to enhance microgrid operations. Microgrids typically consist of multiple inverter-based distributed generators (IBDGs) connected in parallel to mismatched line impedances. This results in unequal reactive power sharing which negatively impacts the performance of IBDGs in microgrids. To achieve enhanced control, a solution utilizing deep reinforcement learning (DRL) is proposed. DRL agents are trained to control variables in each IBDG using a well-designed reward function capable of achieving the following objectives: 1.) ensure output voltage of each IBDG remains within the designated operating boundary and 2.) minimize IBDG RPSE. The proposed DRL method is then compared to the classical droop method under various system disturbances. This exploration into the integration of DRL into microgrid applications holds the potential to revolutionize future grid control methods.
Biography
Sophia Strathman is a fourth-year undergraduate electrical engineering student at the Missouri University of Science and Technology with a passion for renewable energy technology and the possibilities of deep reinforcement learning. She actively contributes to research projects focused on microgrids, deep reinforcement learning, renewables, electric vehicle fast charging, and power electronics. She has interned at Spire Energy and Burns & McDonnell and is planning to intern at Oak Ridge National Lab for the summer of 2024. Driven by a desire to make a positive impact, she is eager to continue exploring the frontiers of deep reinforcement learning within microgrid applications.
Research Category
Engineering
Presentation Type
Poster Presentation
Document Type
Poster
Location
Innovation Forum - 1st Floor Innovation Lab
Presentation Date
10 April 2024, 1:00 pm - 4:00 pm
Data Driven Enhanced VSG Control for Microgrids
Innovation Forum - 1st Floor Innovation Lab
This research explores the fusion of deep reinforcement learning and virtual synchronous generator control in a bid to enhance microgrid operations. Microgrids typically consist of multiple inverter-based distributed generators (IBDGs) connected in parallel to mismatched line impedances. This results in unequal reactive power sharing which negatively impacts the performance of IBDGs in microgrids. To achieve enhanced control, a solution utilizing deep reinforcement learning (DRL) is proposed. DRL agents are trained to control variables in each IBDG using a well-designed reward function capable of achieving the following objectives: 1.) ensure output voltage of each IBDG remains within the designated operating boundary and 2.) minimize IBDG RPSE. The proposed DRL method is then compared to the classical droop method under various system disturbances. This exploration into the integration of DRL into microgrid applications holds the potential to revolutionize future grid control methods.
