When Pieter de Buck took on the role of Aerodynamics and Composites Lead of the Carnegie Mellon Racing team last year, he had a clear goal: “improve the team’s knowledge in aerodynamics and advanced manufacturing with composites”. Together with the team, Pieter decided to take a serious look at the aerodynamics of their car and use a 3D CFD for the first time.
Designing the aerodynamic devices of their car, refining them using CFD, and ultimately manufacturing them, was a very challenging experience. However, it was also an experience that paid off. This year, Carnegie Mellon Racing—an FSAE team from Pittsburgh in Pennsylvania—had their most successful season in history, taking the first place at both Formula North and Formula SAE Electric. Here is the short story of their journey to the podium.
Eliminating the Need for HPC with CFD in the Cloud
Carnegie Mellon Racing wanted to be able to run 3D CFD simulations of a full car. Initially, they used 2D CFD to establish trends, but they knew that simulating the full car was extremely important. When it comes to aerodynamics in automotive, every individual part of the car influences and impacts all of the other parts.
They considered various approaches for establishing a workflow for a full car with CFD. Since they lacked the computational power to run simulations locally, the team were drawn towards potential cloud computing solutions. After evaluating the options available, they were most impressed by SimScale. The intuitive user interface and the option of a free trial gave them the opportunity to get the hang of the SimScale platform before committing to it. As a result, the team ended up moving from using ANSYS Fluent to exclusively running all their CFD simulations on SimScale.
Gaining Expertise in Automotive Aerodynamics
Initially, the team invested a significant amount of time in the meshing settings. They found that small changes would have a drastic effect on their mesh quality. Additionally, they put a lot of emphasis on understanding the theory behind CFD. Eventually, they ran around 50 full-car simulations and various simulations on the front/rear wing on their own. A full-car simulation took about 2.5 hours on 32 cores, while meshing took approximately 20 minutes on 32 cores. Once the simulation runs were completed, the post-processing work was focused on investigating velocity plots, pressure plots, streamlines, and turbulent intensity. For example, the team would check if the airflow around wings was attached and if the diffuser was working as expected. The final iterations of the front wing, rear wing, and underbody/diffuser were manufactured and fitted to the car. The decision about which iteration would be the final solution was mostly based on the amount of the downforce that this design variation produced.
Overall, the Carnegie Mellon Racing team noticed a major increase in their understanding of FSAE aerodynamics since they started using CFD from SimScale. The platform allowed them to visualize the airflow around the car and improve its aerodynamic performance.
Defending their First Place
Moving forward, the team wants to iterate on their current design and hopefully improve the car enough to defend their first place on the podium at Formula North and Formula SAE Electric next year. They are planning a rigorous track testing to correlate their simulation results with real-life performance.
“SimScale has allowed my FSAE team to gain a significant understanding of automotive aerodynamics in only a year’s time. The generous sponsorship gave us the freedom to be creative with our design choices and ultimately produce our first serious aerodynamics package. We are excited about continuing our partnership with SimScale in the upcoming year,” said Pieter de Buck.