Richard Childress Racing (RCR) employed HPC cloud infrastructure to simulate and fine-tune its race car performance on the track.
- Fast, accurate simulations
- Fewer physical models and wind tunnel tests
- Faster data-driven designs
A NASCAR racing institution since 1969, RCR has fielded teams in NASCAR’s top three racing series, and an RCR subsidiary builds engines for other NASCAR teams. In its 50-year history, RCR has won 15 championships and more than 200 victories, and was the first organization in NASCAR history to win championships across all three national touring series.
Testing can be a real drag
Richard Childress Racing (RCR) is dedicated to using cutting-edge technologies to build winning race cars from the ground up. To stay at the top of the competitive NASCAR racing series, they need to ensure that every part in every car is optimally designed. One important step is to run computational fluid dynamics (CFD) analysis for aerodynamic simulations to ensure peak performance.
However, RCR’s expertise is in high-performance race cars, not writing simulation software or maintaining the high-performance computing (HPC) resources to run it.
Building highly aerodynamic automotive parts requires extensive design testing, both in simulations and in an actual wind tunnel. RCR couldn’t run CFD simulations with the sufficient mesh counts required to obtain the precision results it needed. As a result, it had to test more of its designs in a wind tunnel, a time-consuming and expensive process that costs $2,000 per hour.
RCR had previously outsourced its simulations to a third party, but they often experienced delays and were limited by the finite number of engineers and cluster time available. It could take days of waiting before RCR received its data.
Getting up to speed
RCR worked with Ansys to develop its Adjoint Solver CFD simulation software, providing the extreme accuracy necessary to optimize the aerodynamics and component geometries of its cars.
It also needed a partner with an HPC cloud infrastructure able to handle those computations quickly — without the investment and maintenance required by an in-house solution. By partnering with Nimbix, an Atos company, RCR had the computing resources it needed to optimize its car designs.
RCR used the Nimbix Cloud powered by JARVICE to run Ansys Fluent R2 and R3, leveraging the shape optimization capabilities in Fluent Adjoint Solver to analyze the aerodynamic performance of its designs. By using the Nimbix Cloud and the Adjoint Solver, RCR was able to use a mesh count of over 615 million points to obtain a much greater degree of accuracy than with its previous simulation provider.
For most simulations, the Adjoint Solver uses 400 cores. Ansys engineers also ran multiple cases using 400, 600 and 800 cores to demonstrate the linear scalability of both the Nimbix Cloud infrastructure and the Ansys software.
These results provided RCR engineers with excellent design insights into the shape-sensitive regions of the car and guided them towards intelligent design modifications that improve the aerodynamic performance of RCR’s cars.
“Being able to do those simulations in-house on demand has sped up the process exponentially,” said RCR CFD Engineer Colby Mazzuca. “It literally went from days to hours. The data is often ready faster than I can be ready for it.”
A winning combination
Using the Nimbix Cloud allows RCR to run its CFD simulations with a higher mesh count, to obtain precise results and give its engineers greater confidence in their designs. That translates to spending less time and money to build cars for wind tunnel testing.
RCR engineers now perform CFD simulations in a matter of hours, not days, using the HPC resources of the Nimbix Cloud. Engineers are happy with the accuracy of the Ansys Fluent Adjoint solutions, which are confirmed by data from wind tunnel tests, giving them the confidence to move forward with the results. RCR anticipates faster cars and more trips to the winner’s circle, which will also have a positive effect on its bottom line.
With Nimbix’s user-friendly interface, designers with less technical know-how have no issues using the platform. With a few clicks, they can spin up a cluster with the desired number of cores and memory to push their designs through — without having to learn the technical details of super and accelerated computing infrastructure.
Thanks to RCR’s partnership with Nimbix and Ansys, it can now:
- Run CFD simulations faster to improve the external aerodynamic performance, thereby improving speed on the track.
- Utilize more accurate simulation solutions and reduce the number of physical models (and cost) needed for wind tunnel testing.
- Bring new, improved designs to development faster, based on the numerical data generated by the Adjoint Solver.
“Designers should be designing,” says Mazzuca. “They shouldn’t be maintaining infrastructure and figuring out how to deploy things. Partnering with Nimbix and Ansys allows us to design more aerodynamic race cars, faster and more cost effectively, so we can keep our sights squarely on the winner’s circle.”
Atos’s Nimbix supercomputing suite is a set of flexible and secure as-a-service high-performance computing solutions. It offers industry-first federated supercomputing-as-a-service and dedicated bare metal services, providing customers with added agility for their compute-intensive workloads and expanded consumption models.