Machine learning for aerodynamics simulations

Purpose and goal:

Machine learning (ML) models are revolutionizing aerodynamics research and product development, driving efficiency and innovation to new heights. In the area of aerodynamics, ML offers accurate predictions and optimizations, enabling the design of more aerodynamically efficient structures with improved performance. This technique accelerates the analysis of airflow patterns and the identification of optimal design changes, significantly reducing wind tunnel and computational fluid dynamics (CFD) simulation times.

Expected results and effects:

Obtain models and methods that result in product development projects at an early stage and in an efficient manner benefiting through fast calculation loops and fast feedback. Approach and implementation:

Scania has and performs a large amount of CFD calculations which will be used by RICOS to develop methods and processes to be able to predict air resistance using deep learning models.