By Dusty Weis, AEM Strategic Communications Manager

The engine roars, its bone-rattling power enough to make your jaw clench and your eyes water as waves of heat radiate off the cowling. A breakdown now is not an option; there’s too much on the line to let mechanical failure stand between you and your goal.

There are enough similarities between auto racing and heavy equipment that, from an engineering standpoint, there are valuable lessons about machine data and performance that can be shared between fields. The extreme conditions, the brute force, the demand for perfect performance—these are experiences common to these two very different disciplines.

“We’re all students of the data, we’re students of the technology just like you guys are. And there’s a lot of things we can pick up from each other,” says Carlos Gutierrez, an engineer with racing’s Team Penske.

Gutierrez addressed a collection of equipment manufacturing professionals at AEM’s recent Thinking Forward event in Charlotte, North Carolina, which was held at Penske’s headquarters. Members learned about the advanced networks of IoT sensors with which Penske equips its vehicles, swapped data analysis tips and toured the facility where Penske builds and maintains its fleet of stock cars, Indy cars and other high-performance machinery.

As an organization, Team Penske recently celebrated a whopping 500 wins across racing series. “To achieve that is pretty significant,” Gutierrez says. “But from a technology perspective, it’s a really big challenge.”

In Penske’s championship-level approach to winning with data and machine performance, Gutierrez outlines a four-step approach which equipment manufacturers should work to emulate.

Document Machine Data from Every Component

When Will Power flashed across the finish line to claim another Indianapolis 500 trophy for Team Penske this year, you’d think that the engineers who propelled him to victory would have lost themselves in the confetti-covered celebration for a few moments of well-earned elation. But Gutierrez says, even then, there was work to be done.

“As soon as that checkered flag dropped, we were collecting all the information, everything we learned from that event,” Gutierrez says. “We need an accurate record of everything that happened so that, when we come back next year, we can start at the same baseline and make improvements from there.”

Penske’s engineers relentlessly document up to 600 tuning parameters on each car. Lasers provide constant measurements of the car body’s distance from the ground. Hundreds of military-grade sensors measure temperature, inertial forces, vibration, driver behavior and other factors up to a thousand times per second, and then beam that information back to Penske’s engineers in real time.

“Whenever you operate at those frequencies, it allows you to see a macro-level view of what the car is doing,” Gutierrez says. “If you take too few sample points, you might not get the entire picture of what’s actually happening.”

Not only does Team Penske track its machines and its drivers, but it keeps data about its data, cataloging the sensor calibration process using specialized plates upon which the cars must sit. Similar to the heavy equipment industry, their sensors are called upon to function in extreme environments where lower-caliber equipment would be destroyed by weather, vibration or engine temperatures up to 200 degrees Celsius.

But such is the precision required to maximize machine performance in this competitive field.

“We go back to the same race tracks year after year, like you might go back to the same job sites or the same types of machines,” Gutierrez says. “We want to make sure we can reproduce those results when we go back to that track, if it’s a good result. Or if it’s bad result, we can learn from it.”

Consolidate Data Using the Same Standards and Software

If knowledge is power, the data trove from Penske’s 53-year history as a racing team makes it a veritable juggernaut. But, like manufacturers in the heavy equipment industry, Penske faces challenges compiling performance data from different decades and different types of machines in a way that makes potential insights easy to find—and the biggest challenge can be organizational.

“One thing we struggle with is we have a lot of very smart people who work here,” Gutierrez says. “Everybody’s very good at their job, but everybody takes a lot of initiative. That can make it very challenging to communicate between engineering departments.”

The first challenge, Gutierrez says, is making sure that different engineering teams are using the same standards to measure success. In his world, engineers might be designing a car to run in the Indy 500 or the Daytona 500. Similarly, in heavy equipment, engineers might have to account for different types of job sites. Regardless, carefully organizing the data is the only way to provide an apples-to-apples comparison that provides critical insights.

“Whenever people are talking in different languages to each other, it creates arguments on your team that shouldn’t happen,” Gutierrez says. “Sometimes it’s just as basic as agreeing on a common set of performance standards.”

That goes double, then, for the software that engineers use to log and analyze their data. Rather than allowing different teams to design their own software solutions, Gutierrez says Penske delegates the responsibility to a specialized software engineering team, forcing all their teams to use the same software and preventing data from being siloed.

“Now that you know each of these teams is using the same software to calculate, let’s say fuel mileage, it allows us to compete with each other as well between teams,” Gutierrez says. “It’s a competitive advantage.”

Analyze Data in Real Time and After the Fact

With data pouring in at the speed of light from cars traveling at 235 miles an hour, things happen fast for an engineer at Team Penske. Ultimately, Gutierrez says his goal is to use those data to improve safety, minimize errors, maximize performance and inform decisions.

“As the car is going around the track, we can see a lot of these sensors in real time,” Gutierrez says. “That allows us to start acting to prevent, for example, tire pressure failures or overheating issues.”

But there’s only so much a team of human engineers can glean from watching the numbers scroll by in real time. The real work begins when the data—sometimes 10 or even 20 gigabytes worth of data per car per event—are cataloged and analyzed in the weeks following a race.

Measurements from sensors on the car are charted against prior years’ execution on the same track. The performance of various pieces of equipment and gear are mapped against factors like wind direction and track condition. And the car’s position on the track is analyzed to map out an optimum line, which drivers are coached to follow in the next race on that track.

This science has been advanced in recent years by the advent of new machine learning and artificial intelligence techniques; after all, “A.I. in itself is just statistical analysis of data,” Gutierrez notes.

Machine learning platforms can examine massive tables tracking temperature, traffic, weather and track surface, and predict how those factors combined will affect tire pressure, which varies as the race wears on. With this type of analysis, each past race is a dataset that can be used to help pit crews make snap decisions in future races.

“We only have five or six seconds to make an adjustment,” Gutierrez says. “We need to know that, when we make a decision, it’s the right one every time.”

Communicate Machine Data Insights to Operators

Server farms full of data and thousands of man-hours of careful analysis don’t mean anything if the insights aren’t readily available and easy to understand—that’s just as true whether you’re on a job site or at the track.

“When we go to Iowa Speedway, for instance, the lap is 18 seconds long,” Gutierrez says. “So every 18 seconds, we have to make critical decisions, and being able to communicate the insights from our data is very important to us.”

Data insight solutions must provide users with an intuitive interface that allows them to quickly sort data by many different parameters and determine the best course of action for a machine in the field. Benchmarking performance against one’s peers is another helpful tool.

“If I see that Joey Logano has posted the fastest lap of the qualifiers, I can really quickly, without having to ask the team, look at all the documentation and set-up information and apply those changes to my car as well,” Gutierrez says.

And for equipment operators or racecar drivers alike, the process of communicating data insights should start long before they ever step foot in a machine. Penske employs an advanced training and onboarding process, including time spent in simulators and classroom work poring over video and data from other operators’ performance at various tracks.

“There are hundreds of commands they have to know,” Gutierrez says. “If they do these things even one corner too late, it might be the difference between first place and second place… or even first place and the wall.”

In this day and age, Gutierrez says that being able to harness the power of machine data is no longer a perk, but a necessity. In heavy equipment, it can be the difference between a job well done and expensive overages, between optimization and breakdown—between winning and losing.

“If it’s lap 199 of the Indy 500 and Mr. Penske asks me if we can make it to the end of the race, I can’t say, ‘I don’t know,’” Gutierrez says. “You have to have a reliable data solution to help make those decisions in real time.”

AEM members learned about this and other topics at a Thinking Forward event on November 6 in North Carolina.

Subscribe to the AEM Industry Advisor for regular updates about this and other cutting edge topics.