What draws people to engineering and technology is generally the effort to create the best possible product within a series of technical or physical constraints. That same effort to solve a problem underlies the technology behind a Formula 1 race team, although like a race car, engineering decisions have to happen quickly and there is enough tolerance for risk to make a wild experiment every now and again.
After talking to the Lotus Formula One team last Thursday before the team took second at the Formula 1 US Grand Prix, it honestly seems like the best job in the world. Every two weeks during the season (or sometimes every week), the race team is out testing their hunches on the track after hours of testing in computer simulations. Then on weekends those hunches are tested on a racetrack. The whole effort brings together technologies as diverse as computational fluid dynamics and 3D printing. Even big data gets a day in the sun, although the one lament heard repeatedly was about poor bandwidth.
It seems that while the cars can achieve speeds of 200 miles per hour, the Lotus F1 team has to endure a data rate of just 10 Mbps feeding the servers and exporting that data at the track. That’s right, your broadband is probably faster than what these guys buy at the track. When I visited the track last week I was a guest of Avanade, a technical partner of the team. Avanade, a Microsoft consultancy, provides about 20 full-time staff members who work directly with the Lotus team. They are about half of the whole Lotus IT shop, which eats up about 10 percent of the team budget.
Here’s a few facts from the Lotus team:
- Each car has more than 300 sensors
- Every lap those sensors generate and send 25 MB of data
- Races last between 50 and 70 laps depending on the track.
- The Lotus team uses a proprietary wireless protocol to send data from the car to the race engineers
- Those engineers are running simulations that take into account the car’s data, the position of the other cars on the track, weather and a variety of other factors to build out its strategy.
- Those simulations take about 90 seconds — or the length of a typical lap, and the team wants to get it to 60 seconds so it can make better calls on when to bring a driver in.
- About 60 percent of the data generated by the car is used in that race with 40 percent being used later.
As you can see this is a cycle of information generation and reaction that is trying to get as close to real time as possible. When Graeme Hackland, the chief information officer of Lotus F1 team, said his engineers are tracking weather and analytics on how other drivers perform in certain types of weather — using that to make split second calls on whether a driver should come in for a tire change or service — he is really talking in seconds.
Yet, this ignores the other data story happening around F1. In addition to tactical race-day decisions, the Lotus team is also looking to data and high performance computing to build the actual car. And here the interplay between 3D printing and better performing hardware means that the Lotus team can generate a lot more data.
For example, the F1 organization limits each team to 40 teraflops of high performance computing time each week. That used to mean that the team had to choose between time in the wind tunnel testing the physical parts and time simulating the parts. But in the last year it has managed to do the work it used to need 40 teraflops to accomplish in 25, thanks to new chips from Intel (some years it’s AMD) and better software.
So for Lotus, the net is now it can run its wind tunnel all the time. That means more data to crunch, and thanks to 3D printing, it can now test more parts to supply even more data that feed back to improving the car. Before the team adopted wide scale 3D printing a few years ago, it made about 600 parts per week and tested them. Now that it uses 3D printed materials and doesn’t have to wait for parts to be cast, it tests thousands.
This just-in-time manufacturing isn’t just used for test parts, some pieces of the race day vehicle are printed using 3D printers such as the head protection bar on the car, which can take 14 tons of weight on it, and is printed using a titanium powder.
“We don’t tell the driver,” joked Patrick Louis, the CEO of the team.