Crash Testing Cars on Computers — Monolith AI founder and CEO Dr Richard Ahlfeld

Listen to this article

Reading time: 4 minutes

“When I founded Monolith, I was a machine learning fellow at Imperial College London, working on, among other things, NASA’s Mars rocket,” says Dr. Richard Alfeld, the company’s CEO.

It’s not often you get to talk to someone who helped launch a rocket into space. It’s even rarer that the same person now works with BMW to make their cars safer and with the Jota Le Mans team to make their cars faster.

In fact, thanks to Monolith’s machine learning, BMW cars prepare for crashes that could never be simulated in the real world, and Jota cars are pushed to the limit without the need to turn a wheel.

Practically Reality

“My background is in academia and research, and much of my time at Imperial College London was spent looking at how data science techniques could be applied to engineering,” continues Alfeld.

“It was clear to me that deriving insight from data for decision making is a complex but necessary challenge for engineering organizations. Engineers understand data, are embedded in the business, and have the intuition to know why to process and model data.”

However, although engineers can create models and simulations to test their products, the process can be lengthy.

“As an engineer, you might be tasked with building a model and then using that model to design a virtually perfect product in the hope that a physics-based simulation approach will do the job,” Alfeld says.

“But what I’ve found in industry is that if you want to radically accelerate the speed of new product development, you need a radically different solution to understand the physical challenges that are not yet fully understood.”

A monolithic laptop

However, speeding up the testing process is only one part of Monolith’s work.

“Monolith enables engineers to use AI to solve their toughest physical problems. We do this by enabling engineers to use their existing data from the ongoing product development process,” explains Alfeld.

“They can then build self-learning models that can instantly predict outcomes for challenges that would otherwise have to be solved using extensive, time-consuming physical tests.” These self-learning models learn from existing data and identify patterns to solve challenges that have so far been beyond the scope of what a human engineer or simulation technology can do.

By using massively advanced machine learning, but with a code-free interface, Monolith enables engineers and companies to quickly solve incredibly complex problems.

Engineers at Honeywell Corporation, for example, used Computational Fluid Dynamics to understand gas-fluid dynamics when developing a smart gas meter.

“But the simulations weren’t 100 percent accurate,” Alfeld says, “leaving a critical gap in understanding between simulation and reality. Using Monolith, engineers use statistical machine learning methods to fill the gap, allowing them to immediately and accurately understand the impact of different temperature conditions and gas types in all operating conditions, including extreme and unstable parameters.”

Simulating the impossible

“When faced with intractable physics problems, engineering experts can use Monolith to instantly use their existing data and instantly solve previously intractable ones, and thus reclaim literally weeks or months of their time,” says Alfeld .

BMW uses the software to simulate crash tests with impressive accuracy.

“Crash testing is a highly non-linear, expensive endeavor that relies heavily on traditional R&D methods, using thousands of simulations as well as numerous physical crash tests to capture vehicle performance and ultimately meet global homologation requirements , rightfully so,” says Alfeld.

“But even with sophisticated modeling techniques, because of the complexity of the physics underlying crash dynamics, the results require significant engineering expertise to calibrate to the real world.” Also, physical crash tests can only be conducted in later stages of development when the design is mature enough to produce physical prototypes.

Monolith began working with BMW in 2019, and Ahlfeld says the simulations will continue to improve.

“The engineers built self-learning models using their wealth of existing crash data and were able to accurately predict the force on the human tibia for a number of different types of crashes without physically doing crashes,” Alfeld explains

“Going forward, the accuracy of the self-learning models will continue to improve as more data becomes available and the platform is further integrated into BMW’s engineering workflow.” This means engineers can optimize crash performance earlier in the design process and reduce reliance on time-consuming and expensive testing, while making historical data infinitely more valuable.”

Jota’s racing team also sees a number of benefits from using Monolith’s software.

Jota Sport
The #38 Jota Sport car that finished first in the LMP2 class at the 24 Hours of Le Mans

“Jota uses Monolith to accelerate its design and engineering processes in every area of ​​engineering from vehicle tuning, vehicle dynamics, aerodynamics and tire degradation modeling,” says Alfeld.

“By incorporating Monolith into real-time testing, Jota engineers reduced the number of simulations and tests by 50% and the associated costs by 66%, making faster, better design decisions and streamlining the way validate car data and simulations.”

Accelerating change

Of course, while AI testing is great for simulating events at the edge of possibility, Alfeld believes the real benefits in the coming years will be in the pace of development.

“No matter what new technology is introduced, whether it’s autonomous, connected or electric, engineers will always need to create a fundamentally great car to stay competitive and drive demand – from premium acoustics and greater fuel efficiency to safety and dynamics of the car,” he says.

“AI technology can radically transform vehicle development by enabling engineers to extract the best possible insights and predict outcomes from existing engineering data, and much earlier in the development process. This allows engineers to reach design and engineering decisions faster and more efficiently, giving them time to explore even more design parameters and operating conditions.”

“Ultimately, this means OEMs can bring better vehicles to market faster, which is not just vital to achieving our collective EV ambitions, but allows engineers to do what they love most of all, to create amazing products.”