Abstract:
Distributed implementations of model predictive control (MPC) have received significant attention to address the challenges of large-scale, geographically-dispersed systems, where solving centralized optimizations is not viable or practical for real-time control. The essence of distributed MPC (DMPC) is to tackle the overall control problem by using multiple interacting MPC agents. This distribution has the ability to provide increased scalability, efficiency, and adaptability. The seminar will discuss the application of DMPC in thermal solar parabolic trough plants, which are based on concentrating the sunlight on a receiver to raise the temperature of a heat transfer fluid (HTF). Particularly, we will present DMPC methods to optimize the HTF flow rates that are pumped to different areas of the solar field. In this framework, we will place emphasis on approaches that use varying cooperation scenarios among sets of distributed agents. The works that will be presented are part of project OCONTSOLAR (Optimal Control of Thermal Solar Energy Systems), which is an ERC Advanced Grant that was granted to Eduardo F. Camacho in 2018. This project aims at developing scalable DMPC algorithms to yield a safer and more efficient operation of this type of solar plants, while integrating drones and UGV as mobile sensors to measure variables that are relevant for taking the control decisions.
Bio:
Paula Chanfreut is an Assistant Professor in the Control Systems Technology section of the Department of Mechanical Engineering since 2023. She received her B.Sc. and M.Sc. in Industrial Engineering from the University of Seville (Spain) in 2017 and 2019, respectively. She pursued her Ph.D. in Automation Engineering at the same university from 2018 to 2022. She was a visiting scholar at the University of California at Berkeley in 2021, and at Massachusetts Institute of Technology in 2022. Between 2022 and 2023, she worked for ERC Advanced Grant OCONTSOLAR (Optimal Control of Thermal Solar Energy Systems). Her research is framed within the field of model predictive control (MPC), with emphasis on its non-centralized implementations. She is particularly interested in designing control methods for large-scale interconnected systems, including renewable energy plants, traffic networks, and smart buildings. In this context, she has explored the use of flexible cooperation mechanisms to improve performance and scalability in multi-agent systems based on MPC.
Go to the Events page.