Pursuing sustainable and efficient energy solutions, a team of researchers from Teesside University, Mudhafar Al-Saadi and Professor Michael Short, has proposed an innovative multiagent AI-based control system for plug-and-play batteries in DC microgrids. The methods and experimental studies, published in the journal Batteries and several IEEE papers and articles, outlines a promising approach to enhance the management of power-storage flow in microgrids, particularly addressing the challenges posed by dynamic vehicle-to-grid (V2G) charging applications.
Energy is fundamental to the creation and sustenance of life. The primary motivation behind this research is the urgent need for sustainable, low-emission energy systems for applications in industry, society and transportation. These systems aim to tackle climate change and fossil-fuel scarcity through decarbonization, digitalization, and decentralization of electrical power systems. Traditional power-flow management techniques often fall short in handling the dynamic and decentralized nature of modern power distribution networks. The innovative approach proposed by Al-Saadi and Professor Short offers a solution by leveraging multiagent reinforcement learning (MARL), an emerging AI-based technique for improving automated decision-making by learning complex input-output relationships, to improve the efficiency and reliability of power storage in DC microgrids.
Mudhafar Al-Saadi explained, “The influence of DC infrastructure on the control of power-storage flow in microgrids has gained significant attention. Our research aims to address the potential loss of charge-discharge synchronization and the subsequent impact on control stabilization.”
The researchers highlight that efficient power-flow management is crucial for integrating renewable energy sources and ensuring the sustainability of decentralized power networks. One of the key challenges in this domain is the accurate synchronization of the charge and discharge cycles of batteries, which is often compromised under real environmental conditions, most prominently due to sustained high load variations coupled with battery heterogeneity and degradation, uncertain power system topology due to dynamic plug-in/plug-out insertions and removals of battery elements, infrastructure influences, and environmental (temperature) influences. The proposed multiagent-based control system compensates for these variations in real-time, ensuring a balanced and stable power flow.
In their experiments, the researchers demonstrated significant improvements in various performance metrics. The proposed system achieved reduced convergence times, enhanced output-voltage balance, reduced power consumption, and improved power-flow balance. These results underscore the effectiveness of the proposed control system in real-world scenarios, where dynamic load conditions and varying infrastructure influences are common.
“The real-time compensation for DC infrastructure influences with plug-and-play insertions/removals of dissimilar batteries in a larger, aggregated storage system is a key factor in the success of our control approach,” stated Professor Short. “Our system adapts to unknown and time-varying DC infrastructure influences and battery types without the need for initial estimates of key parameters, ensuring the reliability and sustainability of the microgrid.”
“EV batteries and chargers present key storage assets to ensure grid balance and peak shaving in Vehicle-to-Grid (V2G) and Demand Response applications and can aid decarbonisation efforts. Our approach can improve V2G performance and prevent unnecessary degradation of EV battery capacity and lifetime.”
The research team employed a combination of simulation and hardware-in-the-loop studies to validate their proposed system. They used realistic conditions, including day-long continuous variations in load demand and dynamic switching of heterogeneous battery connections, to test the robustness and effectiveness of their control system.
In conclusion, the multiagent-based control system proposed by Al-Saadi and Professor Short represents a significant advancement in the management of power-storage flow in dynamic DC microgrids. By addressing the challenges posed by dynamic load conditions and infrastructure variations, this innovative approach offers a practical and efficient solution for modern power distribution networks. The findings from this study pave the way for further research and development in the field of decentralized energy systems, contributing to the global efforts towards sustainable and low-emission energy solutions.
Journal Reference
Al-Saadi, M., & Short, M. (2023). Multiagent-Based Control for Plug-and-Play Batteries in DC Microgrids with Infrastructure Compensation. Batteries, 9(12), 597. https://doi.org/10.3390/batteries9120597
Short, M. and Al-Saadi, M. (2024) “Transient Recovery of Energy Storage Balance in DC Microgrids with MARL-Based Power Control”. In: Proceedings of the 10th IEEE/IFAC International Conference on Control, Decision, and Information Technology (CoDIT 2024), Valetta, Malta, July 2024.Al-Saadi, M. and Short, M. (2023) “Plug-and-Play MARL for SoC and Power Balance Regulation in Heterogeneous BESSs”. In: Proceedings of the 3rd IEEE International Conference on Signal, Control and Communication (SCC), Hammamet, Tunisia, pp. 1-6, December 2023
About The Authors
Michael Short is Professor of Control Engineering and Systems Informatics and Associate Dean for Research and Innovation within the School of Computing, Engineering and Digital Technologies at Teesside University in the UK. He is also a visiting Professor at VIT Chennai in India. He holds a BEng degree in electronic and electrical engineering (1999, Sunderland) and a PhD degree in AI and robotics (2003, Sunderland). Michael’s research interests encompass aspects of control engineering and systems informatics applied to smart energy systems and manufacturing/process industries. He is PI or co-I on fourteen completed or ongoing funded research and innovation projects and has authored over 190 reviewed publications in international conferences and journals. He has supervised ten PhD completions, has won eight best paper awards and a full member of the IET, the IEEE and fellow of the HEA.
Mr Al-Saadi is final-year PhD student in ‘Optimization of energy control and management of micro grids’ under the supervision of Prof Michael Short within the School of Computing, Engineering and Digital Technologies at Teesside University in the UK. Mr Al-Saadi holds a BSC in the General Electrical Engineering from the University of Baghdad in Iraq and a BSC with hons in Electrical Electronic Engineering from Leeds Metropolitan University in the UK. He also holds an MSC with merit in Control and Electronics engineering from Teesside University in the UK. He has authored over 10 reviewed publications in international conferences and journals and is currently awaiting his PhD viva.
