A comparative analysis of air and liquid cooling techniques for battery packs with machine learning insights into immersion cooling

This work presents an integrated approach to optimizing the design of a 21700 cylindrical battery pack with immersion cooling to enhance thermal management and reduce energy consumption under harsh loading conditions. A thorough literature review on air cooling and immersion cooling systems provided foundational insights that informed our approach. Cell-to-pack technology, a widely adopted strategy for electric vehicles, increases the energy and volumetric density of battery packs but requires robust thermal management to maintain temperature uniformity and ensure optimal battery performance. In our study, we developed a high-fidelity finite element model based on experimental data to predict temperature variations and energy consumption across different battery layouts and target temperatures. A Gaussian process-based surrogate model was used alongside a data-driven generative design method employing a variational autoencoder. This combination allowed for mining useful properties from a dataset of existing battery layout designs and performance metrics, facilitating the identification of optimal design configurations. The results demonstrate that our co-design approach not only enhances the effectiveness of immersion cooling systems by ensuring better temperature control but also reduces the system’s energy consumption by 13%. Additionally, candidate designs optimizing the layout decisions significantly lower the cooling costs by 90%, making this method particularly effective for managing the thermal environment of battery packs in electric vehicles. This comprehensive modeling and optimization framework effectively integrates battery design and cooling system performance, paving the way for more efficient electric vehicle technologies.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Mechanical Engineering

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2024

Author dc:creator

• Kabirzadeh, Pouya

Contributors dc:contributor

• Miljkovic, Nenad

• Battery Thermal Management
• Immersion Cooling
• Machine Learning
• Literature Review
• Optimization
• Air Cooling

Statement dc:rights

• Copyright 2024 Pouya Kabirzadeh

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/124704

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F124704