Summary: UCLA researchers have developed a computational method to rapidly generate vehicle designs with maximum aerodynamic efficiency.
Background: The global shift towards electric vehicles (EVs) has opened new opportunities in automobile design with stylistic and functional benefits. Unlike combustion engine vehicles, EVs lack a large engine block, which provides more flexibility in shaping the vehicle’s geometry. This design flexibility can be leveraged to significantly reduce wind resistance on EVs. This is a critical benefit as it enhances vehicle efficiency and range, which remain a practical limitation to the widespread adoption of EVs. However, calculating geometries that optimize aerodynamics is computationally expensive and often relies on physical experiments to identify the best geometry. There is an unmet market need to make new computational modeling tools that can predict EV aerodynamic behavior rapidly without sacrificing accuracy.
Innovation: Researchers in the UCLA Department of Mechanical and Aerospace Engineering have developed a new data-driven approach to aerodynamic modeling and optimization. By training a machine learning algorithm on a pre-existing dataset of detailed vehicle geometries and their performance data, the researchers can accurately predict drag coefficients for new vehicle models. In addition, the software generates many design iterations of new vehicles to quickly identify the optimal configuration, reducing drag by up to 11% in SUV design simulations. This new tool has great potential to revolutionize industrial design optimization for a wide array of industries, namely the production of efficient and range-enhanced EVs.
Potential Applications: • EV design and rapid prototyping • Simulation-assisted manufacturing • Battery range optimization • Future aircraft and ship design
Advantages: • Reduced computation costs • Rapid design iteration • Independence from physical models
Development-To-Date: Using the software, researchers have simulated designs that reduce drag by 11%.
Related Papers: Aerodynamics-guided machine learning for design optimization of electric vehicles
Reference: UCLA Case No. 2024-221
Lead Inventor: Professor Kunihiko “Sam” Taira, UCLA Department of Mechanical and Aerospace Engineering