Platforms take CAE’s digital twin to real life

Jul 1, 2016 | Hot Topics

If you think about it, we are at the convergence of a number of technological trends that will change how we design products, even as those products become complicated combos of mechanical, electronic, software and control systems. But it’s not all technology; how we work will have to evolve, too. Here’s what I mean:

Cheap computing makes possible tasks that used to be out of reach of many. Whether it’s on the desktop or in the cloud (or on your mobile phone or tablet), consider how technology has unchained you from your office and IT infrastructure. You can run a simulation almost anywhere today; check in on it from truly anywhere. Supercomputers and HPC clusters were never affordable for most enterprises and can now be mimicked by cloud computing for dollars per run and not millions per year. That opens endless possibilities, as we see every CAE software vendor racing towards the cloud. New entrants, too, are getting in on this and changing the dynamics of the industry and its business models.

Computing platforms (which I’ve pondered a lot, lately) enable the creation of ecosystems that build far greater value from combinations that any single element could have on its own. One of the main drawbacks of many aspects of the design art, digitally, were the need to move information from one app to another. From 2D to 3D to 2D, from mechanical to CFD simulation, from 3D to fancy visualization. Each move broke the connectivity of the whole and introduced the possibility of error. Platforms exist to smooth those movements, ideally making them seamless and nearly invisible.

The combo of cheap and seamless means that we have the potential to change how we work. Will we, though? That’s to be determined because, as humans, we hate change. To truly use many of these tools, we need to redefine job responsibilities; to stop with the separation between the “NVH department” and “Powertrain department” (to pick on automotive as an example). You can’t separate the powertrain’s performance from the noise, vibration and harshness it creates, as perceived by the car’s occupants. At the ESTECO user conference leading companies discussed how they use cross-functional teams to bring together disciplines, younger and more seasoned workers, and people from multiple programs — but I’m not sure that’s enough. It’s a great start, but I think we need to do more.

Cheap and seamless also means that we can use simulation for more than digital bend-and-break verification at the end of the design cycle. Again, this might require a shift in roles and responsibilities, but wouldn’t it be awesome if designers could use simulation to weed through dozens of design concepts to the few that are worth pursuing? ESTECO’s modeFRONTIER, Red Cedar’s HEEDS and other design optimization are just now starting to be accessible to non-expert users; in a few more years, these tools should be in the hands of every designer/conceptual engineer/whatever the position title will be by then. Design optimization is able to assess a design envelope (read my explanation of modeFRONTIER, here, to learn more) but it can also, because it’s not limited by human perceptions of what’s worked before, suggest novel alternatives to explore further. Early, early in design is when you want that to happen — alternatives that are proposed later are nearly useless since acting on them is seriously expensive.

Platforms also can expose users to complementary technologies. One example: I don’t know of any serious simulation users who don’t also do physical test. Capturing that data and managing the correlations is hard work. Siemens’ new SImcenter ties the LMS test portfolio into an overall architecture for easier CAE/test correlation. Siemens calls Simcenter a predictive engineering analytics (PEA) platform. I’m not sure we need yet another acronym, but Siemens sees PEA as exploiting design exploration, multidisciplinary simulation and test and data analytics — using data from one model as inputs to another model, perhaps one that is built later in the process or that simulates another aspect of usage.

By combining all of Siemens PLM’s CAE capabilities (NX CAE, NX Nastran, LMS Virtual.Lab and LMS Samtech, CD-adapco  STAR-CCM+, Red Cedar HEEDS and probably other brands I’m forgetting) with Teamcenter to manage all of the processes and data, Siemens is creating an environment that takes a product from early-stage simulation all the way to after-sales (perhaps IoT) operational data gathering. As the product moves from one phase of its development or usage life to the next, that digital view is enhanced, not recreated. This is another area where the human comes into play: the tech exists (or soon will), but how do we create job descriptions around managing this process and its data? Is this warranty-related and so falls under service? Or is it design, even if we’re not planning a product upgrade at this point? Siemens can only create the enabling technology; it’s up to companies to use it well.

Platforms like Simcenter also tie into many of the trends that influence end-markets today. Products that used to be primarily mechanical now include electrical and controls systems — cars, airplanes, dishwashers and industrial machinery now probably contain more lines of code than mechanical components. We’re also rapidly moving from products made with traditional materials and production methods to composites and new production techniques. Clearly, how companies design their products must evolve, too, or they risk becoming obsolete. Today’s designers and engineers need easy access to historical data, 1D/2D/3D (and increasing 4D and 5D) data governing their designs and customer requirements, test data and everything else that’s relevant to the design. Critically, all of this information must also remain in-sync with the product and its operating environment. That’s hard to do without a platform (or some serious in-house IT, if not using a commercial platform).

Bottom line: We can talk about digital twins as much as we want, but they’re only useful if they’re an accurate representation of the physical object. Keeping that up-to-date is somewhat easier during the design process, as more is under the control of the engineering/manufacturing enterprise. Good data management and project governance make all the difference as the design moves across the platform from concept to detailed design. The magic comes in keeping that digital twin updated as the product is used, as maintenance does (or doesn’t) take place. Simulating the operating life of a piece of rotating machinery as if it had been properly maintained when it hasn’t been is pointless; the designer needs to know if vibration is introduced by bearings that rub, for example, due to improper lubrication. That can be seriously big data, if sensors are shooting back readings. You can’t do this without a platform that combines as-conceived, as-designed, as-built and as-used data. But imagine if you had access to all of this, presented in a way that’s intelligent and not overwhelming with useless info. What could you do with that? How would your product be better suited to its mission? How could you delight your customer? It’s a cool time to be an engineer!

This weekend is the Fourth of July holiday, celebrating the birthday of the United States. Catch some sun, take in a concert, enjoy family and friends, appreciate all that it is good.