HPC enabled crash testing for automobile safety

HPC Engineering

In my previous blogs I've highlighted how high performance computing (HPC) has become a powerful tool aiding automobile design. HPC has been particularly important in the realm of simulated crash test simulation and this blog focuses on the rapid improvements being made in this field.

And now, new HPC-powered simulation technologies are poised to further increase the safety of vehicles, as advances in modeling software and a host of related technologies make their way out of R&D and onto the factory floor.

The first way that crash tests have made cars become safer is by learning to account for the complexity of the physics that occur during a collision. Until recently, crash simulations were mainly designed to support the development of the “body-in-white,” the external portion of the car composed of welded sheet metal. Today, newer crash test simulations from companies like ESI, who ran the very first simulated crash test in partnership with Volkswagen in 1985, process tens of millions of elements, simulating how impact from a variety of angles will affect each and every piece and component of a car. Models from other leaders in the field, like those developed by Cray and Altair, have made similar advancements, contributing to impressive improvements in car safety, demonstrated in this safety video from the Euro New Car Assessment Program. The vastly improved ability to simulate car crashes has also helped in the development of new car crash technology, making for example the deployment of air bags less injurious to the head, neck, and chest, amongst many other improvements.

The increasing level of fidelity that digital car crash simulation provides can be used to gather information about not just how cars react to a crash, but also they have greatly improved our understanding of how the human body reacts in a crash as well. An alliance of seven automakers formed the Global Human Body Models Consortium (GHBM) to unify body modeling research and further advance crash safety. The group has assembled a “family” of 13 body models, ranging in size from children to large adults, with the most complex containing over 2.2 million elements and 400 anatomical components. So far, the detail available from these models have helped the consortium evaluate over 40 crash injuries. The data from human body modeling has already begun to help cars get safer, playing a key role in Mercedes-Benz adopting front seat bolsters that inflate “within a fraction of a second” to move the passenger away from the door when car sensors detect an imminent side impact, as well as other recent development, like knee airbags. The next generation of these models, which will better account for muscles, joint, and individual organs, will allow researchers to further optimise the construction of the car chassis, passenger compartment, and components for safety.

Improved digital simulations don’t just make cars safer, they also help manufacturers keep costs down. Running a physical crash test is expensive, at the very minimum costing $50,000, and often costing much more when the cost of a modern crash test dummies, which can cost upward of half a million dollars, are factored in. Running a numerical crash tests costs about 0.02% of that. The ability to economically test prototype cars is especially important for modern automobile manufacturers who are now designing cars for global audiences of varying size and body type. For example, the average Australian man is 178.4 cm tall, while the average Chinese man is 169 cm tall. This difference in height and build has a documented effect on the injury a passengers sustains in a car crash, and numerical car crash simulation has made accounting for these difference much less costly than running separate physical crash tests for each body type.

Continued advancements in HPC-powered car crash simulation doesn’t mean that the crash test dummy is out of the picture, yet. Just as the designs of F1 car manufacturers must still be validated in the real world using wind tunnels, the insights gleaned from virtual crash testing must be verified by real world crash testing, especially as most automotive regulators haven’t started accepting numerical data as a proof of a vehicles safety. The necessity for real world crash testing means that automotive companies both big and small have continued to invest heavily in traditional crash testing facilities. At the same time, some automobile manufacturers are starting to look forward to the day when the crash test dummy is out of work, and the realm of vehicle safety is dominated completely by HPC - a reality that looks more plausible with each new breakthrough.

Written by Nick Dale

See Nick Dale's blog

Nick is Senior Director at Verne Global and leads our work across HPC and specifically its implementation within discrete and process manufacturing. He is based in our London headquarters.

Related blogs

Iceland provides the power behind Germany's most pioneering AI start-ups

This week has seen the announcement of Analytic Engineering, a pioneering German AI engineering firm, choosing Verne Global’s data center in Iceland as the location for their intensive computing. This represents another impressive AI and Machine Learning client win for us, following DeepL joining us just before Christmas.

Read more

Deep learning for autonomous cars

Since Sebastian Thrum and his team used machine learning to win the DARPA Grand Challenge in 2005, machine learning and deep learning have been an integral part of developing autonomous vehicle technology. Great progress is being made, but complex questions remain. My latest blog looks at these issues.

Read more

Trends Advancing Industrial HPC

As we build-up to SC18, Verne Global is delighted to welcome Brendan McGinty, Director of Industry for the National Center for Supercomputing Applications (NCSA), University of Illinois at Urbana-Champaign, as a Guest Blogger. In his first blog Brendan looks at the industry trends around supercomputing and industrial HPC, and how these are advancing innovation in key markets.

Read more

We use cookies to ensure we give you the best experience on our website, to analyse our website traffic, and to understand where our visitors are coming from. By browsing our website, you consent to our use of cookies and other tracking technologies. Read our Privacy Policy for more information.