R Systems played a key role in helping the University of Michigan Solar Car Team design its successful solar car, Novum, giving our engineers the capability and flexibility to run crucial aerodynamics simulations.
The aerodynamics simulations, Computational Fluid Dynamics (CFD) tests, require a great deal of computational power, and R Systems provided us with that power. The team needed to assess canopy shape, lower shape, array shape, trailing edge, leading edge, and other characteristics--to develop a strong aerodynamic design, UM Solar Car had to put our best foot forward in regards to all these aerodynamic components, and we had to do it quickly.
During the design cycle, our team works at a rapid pace and on an exacting timeline. If certain benchmarks are not reached, the whole project is in jeopardy. Our engineers work often well into the morning; they know that if designs are not improved upon and finished on time, the car cannot be built on time and cannot be shipped on time and so cannot race. But designs cannot be improved upon in a timely fashion if they cannot be assessed quickly, and on a typical desktop computer, running just one CFD simulation can take an entire day. Over the course of our entire design cycle, the Aerodynamics Division ran over 200 simulations.
With one a day, that would mean 200 days, and the team did not have that kind of time to wait around for results to come in one by one. “When the team needs to iterate over a number of different aerodynamic components of the car, the design process can easily span more than a year,” explains 2017 Head Strategist Alan Li.
But with R Systems’ support, we could complete a CFD simulation in not one day, but two hours. R Systems enabled us to not only run individual simulations more quickly, but also to run multiple simulations simultaneously without slowing down any running parallel to one another.
This allowed us to maintain the fast pace our design and build cycles called for. And even more than that, this speed allowed us to try out a larger number of shape designs and afforded us more time to brainstorm and review results. The flexibility to explore a wider array of possibilities yielded a better car than we could have otherwise designed--it yielded Novum.
As the University of Michigan came close to finalizing Novum’s design, R Systems proved very accommodating and supportive, helping the team when problems with the timeline arose. “We had only half the time to complete the task,” 2017 Aerodynamics Lead Jiahong Min recalls. “So, we contacted R Systems to ask for more resources. They got back to us almost immediately and gave us twice the computing power we normally get, allowing us to complete the design on schedule.” Within the span of just one week, UM Solar Car ran around 100 simulations with the computing power R Systems provided.
One big thing that makes Novum impressive is its aerodynamic prowess and how its aerobody stands out from the rest, departing from solar car design convention. We were able to try something new and take that risk and execute it well because of R Systems’ support.
And during the 2017 Bridgestone World Solar Challenge last fall, R Systems provided our Strategy Division’s Weatherman the capacity to run models while racing in the middle of the Australian Outback. The models take about six hours to run, which was quickly enough for our Strategy Division to analyze the results during the race and make important tactical decisions pertaining to speed and charging locations while on the road.
With the help of R Systems, UM Solar Car stayed on track both during the design cycle and during the race. Thank you, R Systems, for making our historic finish possible!