• Aerodynamics
• AI
• Automated
• Machine_Learning
• Optimization

This paper discusses the use of artificial intelligence (AI) and machine learning (ML) in aerodynamics, which is becoming increasingly essential in the automotive industry due to the need to minimize the time required for running simulations and the trend toward reducing CO2 emissions. The paper explains the difference between AI and ML and highlights the challenges faced by designers and aerodynamicists in finding common ground between style and aerodynamic performance during the development phase of road-legal vehicles. The paper proposes a methodology that uses optimization algorithms and a list of parameters to generate a set of elementary deformations and trains an AI algorithm to predict aerodynamic performance within seconds by using a low-fidelity mesh and estimating drag values on the variations. The proposed methodology has the potential to significantly improve the development phase by allowing for multiple cross-combinations and generating the most optimal design.