What is "Digital Manufacturing?"
It’s manufacturing that takes advantage of data, simulations, and computing on a much greater scale than ever before. In Digital Manufacturing, anything you can do virtually, you should do virtually. It’s faster, cheaper, easier, and more reliable.
You can think of it as a new evolution of production. The last such evolution occurred in the early 1900s, when most industrial societies replaced Craft Manufacturing (building things one at a time, usually by hand) with Mass Production (building many of the same thing simultaneously).
Check the About page for more info on Digital Manufacturing.
How do all the terms and acronyms (prediction, MS&A, etc.) fit into Digital Manufacturing?
Digital Manufacturing is the big picture, everything else makes it possible – either tools or concepts that make up the whole. On its own, Digital Manufacturing is an idea, an umbrella that covers a huge array of stuff.
Of all the acronyms and terms you’ve heard, the most common is probably Modeling, Simulation, and Analysis, or MS&A. For some reason people often say MS&A when they mean Digital Manufacturing, probably because MS&A represents a huge spectrum of tools itself. They’re not technically the same – Digital Manufacturing is broader – but you might hear them used interchangeably.
What is "Modeling, Simulation, and Analysis?"
MS&A is a catch-all term for tools that simulate physical effects in a digital environment rather than in the real world. Basically it comes down to the following:
- It’s often a combination of software and hardware
- It runs simulations on computer models such as CAD diagrams
- The more information a model contains – dimensions, shape, weight, composition, etc. – the more precise the simulation will be
- Once the simulation is complete (usually many times), it takes all the outcomes and examines them for statistical events of interest
- It may then recommend some change to the model, to encourage or prevent a possible outcome.
MS&A can simulate anything it has parameters for: the movement of gasses, liquids, or heat; how much strain something can take before it breaks; what happens when an object strikes another at a certain angle and velocity; trends in large systems such as the weather or stock market.
How accurate are computer simulations compared to real-world testing?
If the right data goes into a simulation, then its findings are just as reliable and accurate as real-world experiments. It’s also a lot cheaper no matter what. Modeling and simulation works more efficiently and gives back equally trustworthy results.
Traditional methods are only considered wonderful because they’ve been around for a long time, and because we’re conditioned to prefer things we can touch. It’s human nature: if you can’t touch it, it’s not “real.” But today’s MS&A can mimic real world conditions with near-total accuracy. And working in the virtual world is much more affordable than working in the physical one.
How does this technology differ from Computer-Aided Design (CAD)?
Digital Manufacturing software does look like CAD software, and shares some of the same terms. But they are NOT the same.
CAD is a design tool. It creates digital models that define something’s parameters: the shape, dimensions, components, and materials that make up a part, for example. What CAD can’t do is optimize those models, or simulate physical actions on them, or assess whether the model is ideal for its purpose. That’s the realm of Computer Aided Engineering, or CAE… which for our purposes is basically the same thing as Digital Manufacturing processes.
Activities like MS&A require models, of course. They need something to optimize against. So in a way, you can’t get to CAE without CAD; you can’t optimize without designing first.
I've heard the term "High Performance Computing" (HPC), which makes me nervous. Do I need that? How much does a setup like that cost?
You do not need it, which makes life much easier.
As to how much it costs, it’s like asking how much a car costs – a car’s price depends on what you want. A modest HPC system will set you back maybe $5,000-$10,000, and it’ll probably be about four or five times more powerful than a really good desktop PC. Some of the really epic supercomputer clusters found at national labs and universities cost hundreds of millions of dollars, and are ridiculously fast.
If you take one thing away from this, take away the promise that nobody needs to buy their own HPC system to do Digital Manufacturing. You can if you want to, but it is not a prerequisite.
Those epic supercomputers? They sit around doing nothing from time to time. When you can perform hundreds of billions of operations a second you tend to get your work done quickly. Their owners are happy to rent out idle time for manufacturers like you – just send them the files, their HPC chomps through it, and you pay only for the computing power you use. And there are still other alternatives that make it possible for companies of any means to access the kind of HPC power that was once limited to the uber-rich.
Would Digital Manufacturing be useful for companies that only build to print?
Yes! If a firm does build to print, then modeling and simulation can provide a big competitive advantage: quality control through virtualized testing, improvement of workflows and manufacturing cell layout, trend analysis to reduce costs in raw materials.
A build to print operation wouldn’t use it the same way a product design outfit would, but it could definitely be beneficial.
What is the Grid trying to sell me?
No, seriously, what is the Grid trying to sell me?
Once it all gets rolled out, the Grid will have a marketplace where you can invest in software, processing power, training and expertise. But that doesn’t mean the Grid is just a sales tool, or at least, it’s not intended as some profit-minded e-Commerce monster.
The Grid Initiative is a collaboration. Hopefully, it’s owned by everybody, because the only way it will work is if everybody, or almost everybody, gets aboard. It’s managed by the National Center for Manufacturing Sciences, a nonprofit founded to provide solutions that benefit North American industry. But the decisions are steered by the Digital Manufacturing Strategic Interest Group, and that’s open to everybody – you, me, software vendors, computer manufacturers, whoever. Competitors sit side by side on it, that’s part of the point. No one interest dominates.
So yes, money can change hands on the Grid (eventually, it’s still young), but we see it not as selling this stuff so much as making it as easy as humanly possible for manufacturers to get access to it.
I've seen a ton of references to "small" and "medium" manufacturers (SMMs) on this site. What does that mean?
Technically, an SMM is any manufacturer that employs fewer than 500 people. Presumably the “small” is because they’re smaller than big manufacturers. Which kind of like saying the Milky Way galaxy is small compared to the local supercluster.
We use the term because it’s what everyone knows, but it’s neither flattering nor fair. SMMs are anything but small. They’re colossal. People tend to picture the huge OEMs when they think of manufacturing. But SMMs do 80% of the R&D and employ twice as many people. The whole global economy depends on “small” manufacturing, but – unlike the OEMs – these companies don’t operate with much margin for error. They have to do the most with the least, they have to run thinner and leaner every year, and so it’s tougher for them to adopt new tools on a large scale. Which is the real reason why Digital Manufacturing, which has been around for a long time now, hasn’t already been widely adopted.
Our objective with the Grid is to make the barriers as low as possible, make it as friendly and affordable and on-your-terms as we possibly can, because those “small” manufacturers are impossibly huge. When they win everyone wins.