Introduction
The performance, longevity, and safety of lithium-ion batteries, the cornerstone of electric vehicles, are critically dependent on maintaining a narrow optimal operating temperature range.
Conventional Battery Thermal Management Systems (BTMS), typically based on complex liquid cooling, are reactive and consume significant parasitic energy, reducing overall system efficiency. This research project shifts thermal management from reactive to predictive to provide a more efficient and reliable solution.
This project will investigate a novel BTMS architecture that pairs solid-state thermoelectric Peltier elements with an advanced artificial intelligence (AI) control system. Peltier elements offer significant advantages, including high reliability, silent operation, and precise, bidirectional (heating and cooling) temperature control. Critically, they also offer rapid response rates compared to traditional fluid-based cooling systems.
The core of this project is the development of a hybrid AI controller designed to minimise the parasitic energy consumption of the Peltier-based BTMS, while also giving more accurate, real-time, thermal control of the battery cells.
This controller will consist of two synergistic components:
Predictive, to forecast the battery's future thermal load based on real-time operational (e.g. throttle position and recent usage) and environmental data.
Decision, to proactively control the Peltier element. By using the predictive algorithms’ anticipated thermal needs, the decision algorithm will operate the Peltier element in its most efficient profile to pre-emptively manage temperature fluctuations of the battery cells.
Methodology and Framework
The methodology involves developing a high-fidelity, multi-physics simulation environment in MATLAB/Simulink to model the coupled electrochemical-thermal dynamics of the battery cells and the thermoelectric properties of the Peltier element. This virtual testbed will be used to train and validate the AI controller across a range of dynamic vehicle drive cycles and environmental conditions. The performance of the final AI-driven system will be experimentally verified using test packs, battery analysers and environmental chambers within the Energy Technology Research group.
The anticipated contribution of this research is a validated framework for an intelligent, energy-efficient, and highly reliable solid-state BTMS. By demonstrating the viability of AI-controlled thermoelectric devices, this work aims to advance battery management technology, leading to safer, longer-lasting, and more efficient electrical energy storage systems.