We are developing software in collaboration with engineers at Škoda Auto that uses artificial intelligence and machine learning and which aims to analyze the simulated airflow around a car, or in other words the Computational Fluid Dynamics (CFD), and discover different relationships between car parts that are not instantly visible. Thanks to this we now know what is the relationship between the airflow around the side-view mirror and the back wheel, or how a change in the shape of the bonnet will change the airflow on the roof.
Such relationships cannot be detected by the human brain – they can only be discovered by analyzing tens of thousands of pieces of information from CFD simulations. The developed software draws attention to different synergies and recommends places where a slight change in the shape could significantly improve the car’s aerodynamic properties. The software is currently being tested on some cars under development. The efforts should result in reducing the drag coefficient, which in turn will reduce fuel consumption.
In addition to aerodynamics, the Data Science Laboratory also collaborates with several other departments at Škoda Auto. If you want to get involved in these research projects, get in touch with the Office for Cooperation with Industry, FIT CTU.
