In this week's Ask an Engineer segment, we'll be continuing the discussion of the technology of Formula SAE by delving into CFD, or Computational Fluid Dynamics.
CFD is software that simulates fluid dynamics - namely, the motion of fluids like liquids or gases. This process is done through utilizing numerical analysis and algorithms in many different scenarios, some of which can be incredibly complex and involve variables like different fluids, different densities (such as air, water, or oil), different temperatures in fluids that can be heated, and speeds exceeding the speed of sound. Because of the complex numerical analysis required by fluid dynamics, CFD programs like ANSYS are used for a much faster computation when many different factors are involved.
When it comes to designing the aerodynamics of a Formula SAE car, CFD is integral. Alto Ono, a member of Longhorn Racing's aerodynamics subteam, explains that designing an aero kit "is an iterative process by nature. There aren't solvable equations that would tell us exactly what shape the car should be." For the automotive industry, this has traditionally meant creating scale models of different designs and running wind tunnel testing on them to see how each performed. This, however, is a costly endeavor both in terms of time and money.
The introduction of CFD in the automotive industry has meant that professionals are able to create tens to thousands of CAD designs for their products, all of which can be run through CFD software simulating different road, or in the world of motorsport, race conditions in order to optimize designs. A few of the most effective designs are then chosen to be run through physical wind tunnel testing as a way to ensure the product works just as efficiently in real life. While physical testing is the only way to truly ensure a product will produce the results they want, CFD has allowed engineers to focus their tests on the models that are the most likely to work.
Things are different for Longhorn Racing, though. "Because our team does not have regular access to a design tunnel that we can do physical testing in, we rely heavily on creating designs based on data collected from CFD simulations," Ono explains. Because of this, the team creates a few aerodynamics package designs that include changes in different variables that are believed to affect the relationship between the car and the airflow around it. They then can run CFD simulations that show the visual flow paths of air around the car as well as producing hard numbers about downforce and drag. "Our objective is to minimize drag and increase downforce," Ono says. "We experiment over and over with new designs."
A hugely beneficial tool for the team has been ANSYS, which allows calculations to be conducted alongside CFD flow simulations, allowing the aerodynamics subteam to develop a package with both visual and numerical results from their simulations. In addition, The Texas Advanced Computing Center located on the J.J. Pickle Research Campus of UT Austin has very generously donated computation time on Stampede, the seventh fastest supercomputer in the world, to the Longhorn Racing team.
Ono is thankful for the technology and the benefits it provides the team. "In partnering with ANSYS, we were able to get High Performance Cluster editions of ANSYS Fluent running in parallel and were able to greatly cut down our simulation time from three full days to under an hour. This helps our aerodynamics team run a multitude of designs and make small changes during our iterations to fine tune our methods." The process enables them to make multiple changes to their package during a single year's design process. Without CFD technology or ANSYS software, those design changes would have to come on a year-to-year basis, where testing comes from actually building the aerodynamics package and using it on the car.
Computational fluid dynamics is key in both the automotive industry and within Longhorn Racing. By simulating fluid flows - in this case, air - it allows the aerodynamics subteam to create several different iterations of a design in order to find which one provide the least drag and most downforce. Because the team doesn't have constant access to a wind tunnel, CFD and software like ANSYS are the most time- and cost-effective ways to develop an aero package.