Purdue’s community cluster supercomputers propel Hyperloop team’s success
April 10, 2017
Hurtling through a depressurized tube at a speed of 750 miles per hour may sound like science fiction, but – thanks to an idea by billionaire SpaceX founder Elon Musk and implementation by hard-working and creative engineering students at Purdue – it could soon be a realistic mode of transportation.
A team of Purdue students was one of the top finishers in the recent SpaceX Hyperloop Pod Competition, which challenged college students to design a pod that incorporates the revolutionary elements of Musk’s design, including a magnetic levitation system that lets the pod glide on “air bearings” without making contact with the tube.
The Purdue team, one of approximately 30 chosen to advance from a preliminary design phase to the competition on the SpaceX campus earlier this year, used Purdue’s community cluster supercomputers to create and refine their pod design.
The team ran simulations on the Carter and Rice clusters to design an aerodynamically efficient shell with low drag and to build a magnetic levitation system that fit on the pod and maximized lift and minimized drag.
“These simulations are very computation-intensive,” says Aaron Pikus, a senior in aerospace engineering who served as an assistant project manager at the competition and led the design of the pod’s magnetic levitation system. “The nice thing with the clusters was that we could run it there in just 30 minutes and it wouldn’t be on my computer bogging it down. I was able to run a lot of different magnet orientations because of the clusters.”
That ability to test many different configurations paid off. While most other teams at the competition demonstrated a lift-to-drag ratio between seven and ten, Purdue’s lift-to-drag ratio was 13, which translates to improved performance and energy efficiency.
Akanksha Parmar, a graduate student in aerospace engineering, used the clusters to perform finite element analysis to ensure that the pod’s primary structure was sturdy enough to withstand impact at the end of the tube. Parmar hadn’t used Purdue’s clusters before joining the Hyperloop team, but she was able to quickly get up and running, thanks to help from Lev Gorenstein, an ITaP senior scientific applications analyst who served as the team’s research computing liaison.
Hyperloop at Purdue isn’t just an extracurricular activity. Students on the team had the opportunity to earn academic credit through two Hyperloop classes, Hyperloop Design and Hyperloop Build and Test, offered jointly by the schools of Aeronautics and Astronautics, Industrial Engineering, Mechanical Engineering and Electrical and Computer Engineering.
The Hyperloop classes used Purdue’s Scholar cluster, which is dedicated to educational use. Scholar is available at no cost to any professor at Purdue teaching a class that might benefit from high-performance computing.
In addition to time on the clusters and expert guidance, ITaP supplied the Hyperloop team with space on Data Depot to save their data, which they’re taking advantage of as they get ready for future competitions.
At the recent competition, the Purdue team was one of only seven that made it through certain checkpoints and earned the opportunity to test its pod on the mile-long track. Purdue also received an engineering design award.
"It was a great environment and we were very honored by the results,” says Pikus. He’ll be continuing on to graduate school at Purdue in the fall and is now the project manager as the team looks forward to the next Hyperloop competition, which will focus on adding propulsion and optimizing the pod for speed.
To learn more about Purdue’s Community Cluster Program or how ITaP staff can help faculty and students use the clusters for research or education, contact Preston Smith, ITaP’s director of research services and support, firstname.lastname@example.org or 49-49729.
Images of the team's simulations and pod can be found here.