Come the Formula 1 United States Grand Prix Race in Austin, all eyes will be on the world's top-ranked drivers — Red Bull Racing Driver Sebastian Vettel, Ferrari's Fernando Alonso, McLaren's Lewis Hamilton and Lotus' Kimi Raikkonen.
But long before the racing season starts, engineers, aerodynamicists and computer scientists begin preparing a strategy for their teams — a strategy that increasingly relies on computational fluid dynamics and the use of high performance computers to model the design and optimize the performance of their cars.
"This is viewed as the pinnacle of motor sports," said Bobby Epstein (B.A. Plan II), chairman of the Circuit of the Americas and lead investor of the Formula One U.S. Grand Prix, which will take place Nov. 16-18. "The budget for these teams is in the hundreds of millions of dollars, and the technology associated with it is phenomenal."
The linchpin of F1's technology is computational fluid dynamics (CFD), a field that utilizes advanced mathematics and computer simulation to model and predict how the laws of physics and racing conditions will affect a race car's performance on race day. The Cockrell School of Engineering at The University of Texas at Austin has an expertise in CFD across disciplines.
Formula 1 race cars — which are similar in design and function to an upside down plane — can take a corner with an acceleration of 4.5 g-force units laterally, and can decelerate with around 6 g-force units.
In this highly visible, high-stakes sport there are heavy demands on CFD and tremendous pressure on engineers to eek out advantages in design and performance, experts said.
"[CFD] is critical as the teams seek to evaluate improvements that can earn them even hundredths of a second in lap time," said Dipankar Choudhury, vice president research for Pennsylvania-based ANSYS Inc.
Choudhury is responsible for heading up the research and design efforts at ANSYS, which provides CFD software and consulting to a number of F1 teams, including Red Bull, Ferrari, Sauber, Force India and others.
"What was once a scientific novelty, [CFD] is now a practical tool," Choudhury said. "They now use simulation as a standard part of the design and optimization process. It has become routine in the last five years, and that's amazing to someone like me."
The types of 'runs' or CFD experiments engineers are utilizing relate to either the design of the race car, or the simulation of race conditions. When done well, this virtual testing allows engineers to simulate — at a relatively high degree of accuracy — how certain components of a car will perform. And it gives the team a better understanding of how a car behaves when faced with changes in conditions such as temperature or humidity.
"CFD can run a snapshot for specific components. Usually, unless it's a pivotal tool, you can't really see what's going on," said Nathan Sykes, who leads the CFD team at Red Bull Technology Ltd's office in the United Kingdom. "CFD lets you see right at a [conceptual] stage before anything has really happened."
Sykes said that over the last decade, the quality of CFD has risen such that F1 teams are relying more and more on this form of modeling. Along with developing Red Bull's CFD strategy, he oversees the architecture of its high performing computers.
"When I first started 10 years ago, there was minimal CFD being done here. We did one or two full models a year," Sykes said. "Now we are doing jobs in the order of magnitude of many hundreds a week."
Frank Makowski, head of F1 Aerodynamics Simulation Technology for Ferrari Gestione Sportiva, says that there are different instantaneous scenarios that CFD captures that wind tunnel testing cannot.
"Not all performance aspects of a modern F1 car can be replicated in a wind tunnel, but because of advances in CFD, a larger portion of the performance envelope of a car can be simulated," Makowski said.
While Ferrari, Red Bull and other teams are conducting more CFD runs, they are also limited by F1 regulations in the amount of CFD that can be used. F1's governing body dictates the number of teraflops [computing power] a team can run, as well as the hours of wind tunnel testing performed each week.
"During a period of stable F1 regulations, designs become mature, so we must search for smaller margins of superiority. Given the limitations on CFD testing, we must seek the truth with great economy," Makowski said. "So one has to refine one's simulation strategy."
Makowski declined to speak about specific runs to not give away any competitive insights.
In the months before a race, engineers will collect volumes of information about different operational aspects. One key area of experiment is turbulent drag, which is created by the downward force on a car and can slow it down.
"The trick is to provide enough vertical load so that the tires are forced to grip the roughness of the track, like two meshing gears," Makowski said.
Engineers will also spend a good deal of time simulating racing conditions for specific tracks. They must optimize a car in a different way depending on the number of turns and straights a track has.
"The thing is, racing conditions are completely different as you go from races in the UK to Singapore," Choudhury said. "You have to make adjustments, so the car behaves optimally in those conditions. And you will have more confidence if those conditions have actually been simulated."
After an event is over, an F1 engineering team will go back and analyze the data collected from the race, along with driver feedback, before starting the process over again.
However, an engineering team's job is never done, Choudhury said, "There is always room for improvement."
Looking ahead, Sykes said Red Bull's five- to 10-year plan is to rely less on wind tunnel testing and more on CFD.
"I think that always has to be the ultimate goal. We want to get to the point where we are running CFD simulations as accurate as track results, and to get to a time frame when we can run many thousands of them a week. That has to be our goal," Sykes said.
Makowski also sees the use of CFD increasing with the costs associated with acquiring CFD information dropping, and turn-around-time for runs becoming shorter.
"You've got the added benefit of simulating things that a wind tunnel can't and that is going to feed this long-term trend," he said. "At the same time, you have to make sure you have a frame of reference to compare to. Wind tunnel testing is not going away."
