Impact of Neural Concept’s Aerodynamic AI on Formula 1 Racing
It’s a long way from pedal bikes to Formula 1. But that’s precisely the quantum leap that AI-based startup Neural Concept and its co-founder and CEO, Pierre Baqué, made in just six years.
In 2018, the company’s fledgling software helped develop the world’s most aerodynamic bicycle. Today, four out of 10 Formula 1 teams use an evolution of that same technology.
Along the way, Baqué’s company picked up contracts with aerospace suppliers like Airbus and Safran, earning a $9.1 million Series A raise in 2022. Now at 50 employees, Switzerland-based Neural Concept is working toward a Series B round while its software helps historic F1 teams like Williams Racing find their way back to the top of the world’s premiere form of motorsport.
However, where Formula 1 cars rely on 1,000-horsepower hybrid V6 engines, Baqué’s first practical application of the technology was human-powered.
In 2018, Baqué was a student at the Computer Vision Laboratory of the École Polytechnique Fédérale de Lausanne, where he focused on the application of machine learning to three-dimensional problems.
He was introduced to a man who was leading a team on creating the sixth or seventh version of a specific bicycle, with a goal of breaking bicycle speed world records. This man was Guillaume DeFrance and the team was from IUT Annecy of Université Savoie Mont Blanc. The team had already been through multiple versions of bicycle designs.
Just two days later, Baqué impressed the team by coming back with a bicycle shape that almost resembled the current record-holding design. They asked for more iterations, and the resultant design was pronounced by Baqué as ‘the most aerodynamic bike in the world at present’
The claim is a bold one, but it is supported by numerous world records that were broken in 2019. This is not just about aerofoil-shaped down tubes or dimpled rims to cut drag. This bicycle is fully shrouded with the rider encased in a composite shell, completely shielded from the wind.
Neural Concept Shape (NCS) is a core technology product. This system, which is based on machine learning, offers aerodynamic suggestions and recommendations. NCS is part of the wider computational fluid dynamics (CFD) field, an area where experienced engineers use sophisticated software tools to conduct 3D aerodynamic simulations.
CFD provides a quicker alternative to the traditional method of creating physical models and testing them in wind tunnels. However, it’s heavily reliant on substantial system resources and humans making effective choices.
NCS’s main function is guiding engineers to prevent potential aerodynamic missteps and encouraging them to explore unanticipated avenues. In the “co-pilot mode”, an engineer can input a pre-existing 3D figure, setting a baseline for instance.
Following this, NCS consults its neural network to propose enhancements or changes that represent some potential paths in a 3D choose-your-own-adventure game. The engineer then selects the most promising suggestions, subjects them to more testing and refinement, and proceeds iteratively towards aerodynamic success.
NCS proves to be advantageous not solely in racing but is also pivotal in the automotive and aerospace sectors. Heeding to the slower adaptation curve in the somewhat conservative aerospace industry, Baqué mentioned that the focus has been shifted a bit more towards the automotive industry whose exigency is soaring and is quick to adapt to changes.
Contracts with numerous significant global suppliers such as Bosch and Mahle have been secured by Neural Concept. Ever-increasing importance is placed on aerodynamics in the automotive industry, with manufacturers extensively hunting for highly aerodynamic vehicles that furnish the maximum mileage from the given battery pack size.
However, aerodynamics is not the only concern. NCS also finds its application in the development of components like battery-cooling plates. The enhancement of efficiency in these plates can maintain the battery at its ideal temperature without significantly draining the energy. Baqué added, “There are colossal benefits to be obtained”, implying an improvement in the range as well.
The actual testbed for these technologies has always been the road, but the ultimate research center is Formula 1. Having been a global motorsports phenomenon since 1950, F1 is currently riding on an unparalleled resurgence wave.
The Netflix series “Formula 1: Drive to Survive” has introduced the thrill of F1 to a new crowd. This show mainly focuses on the politics and drama between teams, but the actual success on the track largely relies on aerodynamics. This is where Neural Concepts steps in.
Baqué started his journey with Formula 1 way before Netflix was even considered by Reed Hastings. He quotes, “I have been watching F1 since the era of David Coulthard and Michael Schumacher.”
As of today, parts designed with the help of the software from his company have found their place in the apex of international motorsport. Baqué expressed his feeling of great accomplishment and mentions, “I saw it as a landmark when I started the company. Not just for Formula 1, but to notice that parts designed with the software are on the roads. Yeah, each occurrence gives a feeling of immense pride.”
However, Formula 1 is a highly secretive sport. Among the four teams that collaborate with Neural Concept, only a single team was willing to reveal its identity as a client, and even it was fairly reticent about the whole procedure.
Williams Racing is a highly renowned team in the world of Formula 1. It was established in the year 1977 under the leadership of the racing icon Frank Williams. The team, during the 1990s, was unmatched and clinched five constructors’ world championships. Out of these five, three were consecutive wins spanning from 1992 to 1994.
However, like many sports, a team’s triumph in Formula 1 is not constant but rather cycles. Currently, Williams is going through a significant rebuilding phase. The 2022 season saw the team at the bottom of the rankings, marginally improving to the seventh position last year.
The advancement in technologies like NCS, is assisting the team in regaining their competitive prowess. “The technology is being utilized in several ways, some improvements in our simulation, while we are working on other strategies that could lead to better initial results in CFD,” stated Hari Roberts, the Head of Aerodynamic Technology at Williams.
It’s important to note that CFD simulations are both time-consuming and expensive. This issue is intensified by the Formula 1 protocols that put a cap on a team’s testing abilities. The teams have limited physical time in the wind tunnel and each team is given a predetermined budget for computing time to develop their vehicles.
Advantages can come from any tool that facilitates swift aerodynamic designs for a team, and NCS certainly fits that description. According to Baqué, a comprehensive CFD simulation, which traditionally takes an hour, would only require around 20 seconds with NCS.
Since NCS doesn’t actually conduct physics-based calculations, and instead produces AI-assisted estimations based on a collection of aerodynamic insights, it’s largely unaffected by F1’s strict prohibitions. Roberts explains, “We gain a competitive edge when we can implement methods that let us derive more understanding and hence superior performance from every CFD and wind tunnel experiment.”
However, this doesn’t nullify the fact that the teams have to finance it. Baqué suggests that the expense of NCS fluctuates based on the team’s size and nature of access, but it typically ranges from €100,000 to €1 million annually. Given that F1 teams are also subject to a $135 million annual expenditure constraint, this is a significant obligation.
Roberts from Williams hesitated to associate any specific components or lap time enhancements to the NCS software, but he admitted that it has had an impact on their automobile’s performance. He stated, “This technology forms part of the arsenal we use for car aerodynamic development. Hence, we can’t directly connect lap times to it, but we’re confident it enhances our correlation and the pace at which we can explore new aerodynamic situations.”
The relentless progress of AI is set to go further. There is speculation about artificial entities on the pit wall, taking the lead in race strategy and car setups.
“The rise in the AI/ML sector is astronomical,” said Roberts. “However, it presents a significant dilemma for anyone in technology nowadays. Which new technologies should we commit time to research, cultivate, and implement?”
Such intrigue may not engage the typical “Drive to Survive” audience, but for many F1 enthusiasts, the unseen competition is the ultimate thrill.
Regarding Neural Concept, the business is venturing further into the non-racing aspect of the automotive industry. This involves innovating more efficient electric motors, optimizing cabin heating and cooling systems, and branching out into crash testing.
Baqué stated that the company’s software can assist engineers in optimizing a vehicle’s crashworthiness while eliminating superfluous weight. However, currently, the company can only perform crash simulations on individual parts, not on entire vehicles. “That’s one of the few applications where we’ve been reaching performance limits,” he expressed.
Could this be another potential use for the EU’s rapidly growing AI supercomputing platforms?
Discover the pinnacle of WordPress auto blogging technology with AutomationTools.AI. Harnessing the power of cutting-edge AI algorithms, AutomationTools.AI emerges as the foremost solution for effortlessly curating content from RSS feeds directly to your WordPress platform. Say goodbye to manual content curation and hello to seamless automation, as this innovative tool streamlines the process, saving you time and effort. Stay ahead of the curve in content management and elevate your WordPress website with AutomationTools.AI—the ultimate choice for efficient, dynamic, and hassle-free auto blogging. Learn More
