Get to grips with AI and IoT

The racing engineers at Ducati Corse are demonstrating what can be achieved with smart products: Thanks to fully networked racing motorcycles, the IIoT, and machine learning, these experts are significantly reducing the time required for race-deciding setups.

<<< Start >>>

Marcello Tamietti

<<< End >>>

Marcello Tamietti’s presentation doesn’t feature any roaring engines, the feel of the wind in your hair, or the smell of hot machine oil. That’s because the managing director and global sales lead for Industry X.0 at Accenture isn’t showing any racing motorcycles, even though he is giving a talk on Ducati Corse.

Instead, he’s brought along brightly colored dashboards and explanations about the interplay between sensors, data, and simulations. “Motorcycle races are of course exciting, and networked racing motorcycles are truly spectacular. But for me, the key aspect is ultimately the usefulness of the applied technology,” explains Tamietti, noting that this value is highly significant for industrial companies in particular.

<<< Start >>>

Ducati Racing Analytics App.

Image: Accenture

<<< End >>>

“Our work on behalf of Ducati Corse demonstrates what machine learning can accomplish in the field of engineering—and thereby emphasizes one of the core value propositions of our Industry X.0 approach.” What Tamietti means is that the project he is describing illustrates in concrete terms how digital technology, data analyses, and simulations can minimize the effort required for configuration tests, new configurations, and machine optimization. And this is the precise aim of the project he is presenting.

Better configurations, better testing

Ducati Corse wants to use new procedures to speed up racing-motorcycle tests and reduce the effort involving in the testing process. To this end, the group’s engineers are equipping the machines with up to 100 IIoT sensors that measure performance data while the motorcycle is on the move. This data is then transferred to an analytics engine developed in house at Accenture.

The engine evaluates the machine and operating data in its entirety—including motor settings, the speed, tire and brake temperatures, acceleration, oscillations, vibrations, and grip—and subsequently calculates simulation scenarios for reconfigurations. It then uses easy-to-read visualizations to show the engineers how modifications to the suspension, gearbox, or tires, for instance, will improve a machine’s overall performance.

Recommendations by AI

But that’s not all: The software also uses machine learning to calculate recommendations for the ideal machine configuration for specific circuits. This is made possible by the fact that the algorithms also take into account historical measured values from earlier tests, in addition to the mentioned machine data – and include data on circuits and weather.

Tamietti explains: “The simulations and recommendations are already extremely robust. This is primarily due to the high degree of detail of the data that we use.” According to its own figures, Accenture has already loaded 4,000 circuit segments and around 30 racing scenarios into the engine.

And the effort is already beginning to pay off: The precision and reliability of the recommendations are in fact already saving Ducati Corse’s engineers some work during initial testing; the experts are able to reduce the number and duration of actual test drives—which leaves them more time to improve configurations or coordinate more closely with the drivers.

<<< Start >>>

<<< End >>>

“The MotoGP has 18 circuits, and we need to test as many configurations as possible to ensure that our motorcycles operate at peak performance. Accenture’s solution has already yielded excellent results in the lab. The analytics engine will change our entire testing procedure and help us to get the very best performance out of our machines—no matter the weather or the circuit”, highlights Luigi Dall’Igna, General Manager at Ducati Corse the benefits this offers.

“And this is exactly where this whole approach becomes interesting to industrial companies as well,” Tamietti notes. “Attaining top performance at all times and under any conditions – that’s ultimately something that you don’t just expect of racing motorcycles, but also of production facilities, tools and vehicles, and processes in logistics. Our project shows that analytics and machine learning can deliver these results.”

When it comes to potential applications, the expert mentions branches of industry in which “recipes” and system configurations are constantly changing “such as in the chemical industry, during steel production, and at contract manufacturers.” At the end of the day, the main focus in these industries is also on getting the most out of existing machinery. “Just as it is in racing.”

<<< Start >>>

<<< End >>>

<<< Start >>>

<<< End >>>