Deep Learning-based Multi-fidelity Surrogate Modeling for Prediting Automotive Aerodynamic Performance
Research Project, 2026
This project supports Sweden's climate goals by accelerating the aerodynamic optimization of electric vehicles through AI-driven methods. It focuses on developing a multi-fidelity deep learning surrogate model that combines low- and high-fidelity data to predict aerodynamic performance with near high-fidelity accuracy, but at significantly lower computational cost. The project addresses key challenges in scaling to complex 3D geometries and training with limited high-quality data, aiming to enable faster and more efficient design processes in the automotive industry.
Participants
Simone Sebben (contact)
Chalmers, Mechanics and Maritime Sciences (M2), Vehicle Engineering and Autonomous Systems
Selpi Selpi
Chalmers, Computer Science and Engineering (Chalmers), Data Science and AI
Alexey Vdovin
Chalmers, Mechanics and Maritime Sciences (M2), Vehicle Engineering and Autonomous Systems
Funding
Chalmers Area of Advance Transport
Project ID: IKB 95418010
Funding Chalmers participation during 2026
Related Areas of Advance and Infrastructure
Sustainable development
Driving Forces
Transport
Areas of Advance
C3SE (-2020, Chalmers Centre for Computational Science and Engineering)
Infrastructure