Driving sustainability and Innovation in Battery Thermal Management with AI
A study on how AI can optimise heat exchange in various industries.
At AFRY, we believe innovation is the key to building a more sustainable, efficient world. Our recent collaboration with Chalmers University of Technology exemplifies how AI can transform industries. Together, we explored novel ways to optimise heat exchange systems, driving sustainability and expanding applications beyond the automotive sector. The study report is further down in the article.
The Origin of the Project
A growing challenge within the automotive industry sparked this initiative. With the shift towards electrification, effective battery cooling solutions have become critical. Proper thermal management prolongs battery life, increases vehicle range, enhances vehicle performance, and empowers electric vehicles to reach new heights. However, designing cooling channels that balance thermal efficiency with manufacturability has traditionally been a time-intensive and complex process.
That’s where AI entered the picture.
The technical team sought to explore how machine learning models could streamline this process by eliminating the biases and limitations that come with human design. The goal? To develop a more efficient process for optimising cooling channel geometries while maintaining performance and feasibility.
The Process
This project brought together the expertise of AFRY engineers and the technical skills of five Chalmers University students. Working in two teams alongside AFRY supervisors, the collaboration focused on integrating computational fluid dynamics (CFD) with artificial intelligence.
Explanation of the image above:
Various cooling channel designs were simulated in the project and overlaid on the cooling plate. The simplified battery packs are placed below the cooling plate, and no other interface is assumed between the battery pack and the cooling plate.
Neural Network architecture, showing the inputs on the left, leads to several Fourier layers, followed by 2 separate models to predict temperature and pressure drop.
The genetic algorithm chooses the optimum path for new cooling channels based on given heat loads.
The outcome
The results were impressive: a streamlined, automated process that enhances thermal management with unmatched effectiveness and speed. It ensures optimal performance and reliability in even the most demanding environments while reducing inefficiencies, saving time, and maintaining consistent temperature control—perfect for advanced applications. By eliminating the need for conventional design methods during early concept phase exploration, engineers gain more time to focus on specialised tasks. This innovation marks a step forward in the future of thermal management.
The Power of Collaboration
This project was more than a technical achievement. Pairing AFRY’s seasoned engineers with talented Chalmers students created a dynamic knowledge exchange.
AFRY mentors guided students through complex areas like simulation setup and AI model training, while the students introduced fresh perspectives by automating simulations and experimenting with cutting-edge neural networks.
"The students approached challenges with open minds and curiosity. Seeing them tackle problems in ways we hadn’t considered was inspiring", said Anthony Vivek, Fluid Dynamics Specialist at AFRY.
This two-way mentorship proved invaluable, showcasing the synergy between academia and industry.
Overcoming Challenges
Naturally, every innovation brings its own unique challenges. Running 4,000 CFD simulations demanded significant computing power. AFRY and Chalmers tackled this issue together by leveraging advanced supercomputing resources at the university.
The partnership also involved navigating other complexities, like training AI models with highly detailed physics simulations and ensuring accurate predictions. However, these challenges only strengthened the collaboration, driving both teams to exceed expectations.
Innovation Beyond the Automotive Industry
While the project was initially aimed at EV cooling, its impact has proven far-reaching. Heat exchange systems play a vital role outside the automotive world, from manufacturing and energy production to HVAC systems. Leveraging the insights gained, we're applying this technology to create more sustainable solutions across diverse sectors:
- Manufacturing: Optimising industrial heat exchanger designs to reduce energy and resource use. By leveraging AFRY knowledge in manufacturing and automation, additional constraints can be imposed on the developed AI model, such as minimum tooling radius, etc. Hence, the AI model can not only be used in early design stages but also to improve the redesign of existing products in mass production.
- Energy Efficiency in Buildings: Applying AI to HVAC systems, improving energy use in offices, homes, and public spaces.
According to IEA, 20% of all electricity is used for HVAC and the energy demand for HVAC is growing. Hence, efficiency improvements in HVAC is necessary to achieve net-zero emissions.
These innovations underline the versatility of AI in transforming how heat exchange challenges are approached in different industries, helping businesses cut costs while meeting sustainability goals.
What’s Next for AI in Heat Exchange?
The achievement of this initiative is just the beginning. AFRY is already engaging in discussions with global industry leaders to apply AI insights to large-scale projects across various industries. The possibilities for AI-powered heat exchange optimisation are immense, from proof-of-concept exploration to scaling these solutions globally.
“We’ve only scratched the surface of what’s possible,” AI is helping us improve thermal management, not only making it more efficient but also supporting the shift towards a more sustainable future.” says Anthony Vivek.
If your business relies on heat exchange technology and you’re curious about AI’s potential, we’re here to help.
