Waymo and CAE
From finite element analysis (FEA) to computational fluid dynamics (CFD) to plenty of other tools in between, computer-aided engineering (CAE) is a mainstay in the development of everything from individual components to full vehicles. While there are still full prototypes built for testing, the number of these physical objects has been reduced, with the reduction coming from virtual versions.
And talk to anyone who knows something about how the digital realm matches to tangible reality and they’ll tell you that the results are good. Very good.
At the TC Mobility program in San Jose yesterday, Dmitri Dolgov, Waymo chief technology officer, said that the company has “driven ten billion miles in simulation.” This is like having 25,000 vehicles running 24/7 in the Matrix.
Waymo’s Dolgov, CTO (Images: Waymo)
To put that 10-billion into context, realize that according to the most recent statistics from the International Organization of Motor Vehicle Manufacturers, the total number of motor vehicles—everything—on the planet is 1.3-billion. So (yes, this is absurd), if every one of those motor vehicles was equipped for autonomous vehicle development, they would have to drive 7.69 miles to reach 10-billion.
From Peterbilt to Pacifica, Waymo is developing self-driving tech that can be used by these and everything in between
But there is another aspect to this vis-à-vis individual vehicles, and that is that the vehicles in the simulations are all—and this word doesn’t quite give the situation justice—networked such that if there is an “incident” experienced by vehicle X, the information is shared with vehicles Y, Z, etc.: i.e., “machine learning.”
Things, presumably, get better and better with each simulated mile driven.
Of course, just as physical vehicles must still be crash tested, Waymo has to rack up physical miles, which it is doing by running self-driving vehicles on public roads in some 25 cities. It has
Still, its CAE is something that can’t be underestimated for its importance of bringing robust self-driving technology to a vehicle near you.
In two hours or less, you can create fairly sophisticated animations from your CAD system's solid models so that people who know nothing more than how to use Microsoft Word and PowerPoint on their Windows-based computers can better understand a part or assembly design
Translating customer needs and requirements into design and engineering specifications is mostly a matter of listening, seeing, quantifying the qualitative, and then figuring out the possibilities. Yes, there are tools that can help.
General Motors is working with Autodesk on utilizing advanced design software and 3D printing capabilities to develop parts that are not only lighter than those otherwise developed, but which combine what would otherwise be separate parts, thereby reducing manufacturing complexity.