Machine Learning for Aerodynamics with Scania

The aerodynamic vehicle design process aims to optimize vehicle shape to minimize aerodynamic drag while relying heavily on computational fluid dynamics (CFD) simulations. This process involves creating a digital model of the vehicle and its surroundings and then simulating how the air flows around and interacts with the car with mathematical equations. Design improvements take time and effort, especially with heavy vehicles such as trucks. However, technological advancements, such as machine learning, deep learning, and artificial intelligence, can significantly enhance the aerodynamic vehicle design process by providing faster, more cost-effective ways to predict aerodynamic performance and propose optimized designs. These technologies can help overcome the limitations of traditional CFD-based approaches and enable design teams to explore a broader design space in a shorter time frame.

Scania is looking for partners to support predicting aerodynamic drag utilizing ML or AI models on large datasets such as existing vehicle simulation data to generate new vehicle designs quicker to improve vehicle efficiency, performance, and sustainability. The partners we are looking for have ML or AI expertise while also possessing previous experience in CFD simulations and vehicle aerodynamics. The envisioned solution should handle large amounts of data with high-performance computational capabilities and would integrate into existing tools and systems for training with new data. In addition, the ability to onboard new users and provide continuous solution support is essential.