The most far-sighted automotive companies are taking advantage of increasingly cheap computing power to leverage simulation work more heavily in their vehicle development.
The technology behind SIMULIA PowerFLOW simulations is referred to as ‘digital physics’ and it has always been computer-intensive. Computational Fluid Dynamics (CFD) simulation involves analyzing effects at a micro level that cumulatively become relevant to the whole, which takes a lot of computing power to get right.
But the dramatic fall in the cost of computer power over the last ten years means that manufacturers can invest in world-class computing and still yield big savings in product development, by using simulation to reduce the amount of physical testing.
We’ve seen computing CPU/hour costs drop every year over the last 15-20 years. Meanwhile the available computer capability at the same cost has increased, so you're looking at a gain of computing capability for the same price.
Some in the automotive industry are taking advantage of this opportunity and are acquiring significant computational resources to boost the amount of their simulation work. This helps them to explore more design variations, identify problems earlier, and at the same time find better solutions. The decreasing cost of simulation enables modern optimization techniques that can explore large design spaces and identify optimal solutions that could never be identified with other techniques.
On the other hand, one doesn’t need to purchase computational resources to benefit from high performance computing. SIMULIA's cloud-based simulation service, SIMULIA ExaCLOUD, provides access to the hardware without the associated costs. So a small truck manufacturing firm, for example, can plug into a system that's equal to or better than that of any large OEM. This is the same system that PowerFLOW engineers use in their daily work when they provide aerodynamic, thermal management, aeroacoustic, and climate control services to their customers.
This would be unthinkable 15 years ago – and for that we have to thank the ever-improving cost vs power relationship of modern computing.