The Digital Twin is a concept that is being established by the Enterprise CAD software crowd as well as people working in industrial automation. Heard among many a chin-chin with white wine and a smile; the digital twin could refer to just how well this term pairs with a higher “share of stomach” of manufacturer spending. Or it could very well relate to just how well this concept goes down with the Industrie 4.0 crowd along with with a reinheitsgebot beer and dreams of renewed German precision manufacturing hegemony.
The digital twin as a concept refers to the idea that in a digital manufacturing world mass customization coupled with software will mean that each and every single product in a company’s inventory will have a (perhaps) unique file encompassing all of the settings and necessary production information for that one part. A million unique hearing aids will all have accessible trackable digital twins that specify when and how they were created.
One can easily see how as a storage company this would make one salivate. For a manufacturer, as agile Chinese companies nip as one’s Achilles heels a renewed push into integration and complication with a side order of software could be just what staves off the competition and keeps everyone A6’s and A8’s for a while. Integration, software integration and plugging one’s product into the aorta of a firm sound very compelling. Once your process or machine is wrapped around the main artery along with ERP and PLM, they’re never going to rip any of that out. We have cured the patient forever; he need only keep taking our insulin. Forever revenue, annual maintenance fees, and a more high tech product, what’s not to like? What’s more, rather than make a device and sell it once we can write code and sell it lots of times. Perhaps this is a path for our stodgy firm to get a better multiple as well.
For manufacturers afraid of a future that they don’t understand, a vague, fluffy wave of a safety net sounds compelling as well. Track all of the products? Track all parts? Analyze all of the parts. Compliance on everything all of the time. Would you say no? After CRM, PLM, and ERP the digital twin could be the next multi-billion dollar dream of total control through software. For governments, the digital twin represents all of their dreams in one, and it seems like a future surefire innovation subsidy darling. Mittel GmbH, a family firm founded in 1810, the largest manufacturer of specialized hand tools for ski lifts in WestElbe Ostfalen, has implemented the digital twin to produce custom tools more efficiently using digital manufacturing in Germany. Perhaps in this century, European subsidies will produce mountains of code instead of butter.
Perhaps you can feel a slight skepticism on my part towards the digital twin. Just a smidgen maybe. Having worked in software, I’ve never seen it as a refuge for altruists. Historically much of software’s labor-saving potential has seemed to be absorbed by the vendor rather than have been created for the benefit of the client. I do however believe in digital manufacturing, and that 3D printing is quickly becoming a viable mass manufacturing technology for millions of highly detailed end-use parts. For that to happen, something like the digital twin has to exist. I had high hopes for something a bit more elegant, robust and open: a stuff DNA or sDNA where all of a files ingredients, parameters of the design, rights, and attributions are included in all 3D printing files themselves in an open format. I still think that this is a much better idea, but the digital twin with its fluffy enterprise software husk and the meaty, dense interior is a veritable beef wellington of profits compared to the ceviche thin earning potential of adding sDNA to all of the things in a universally accessible and free format.
They will PowerPoint this into our heads, and it will become what we need to want. Assuming then that the digital twin will, therefore, become commonplace I’ve been giving the concept some thought. I believe I’ve come up a much improved (and far more profitable!) addition to the idea. Enter: the Digital Triplet.
During production, each individual part will need one record of its precise making, and this will exist for the life of the product, available for tracking, querying, and analysis. We should keep this virtual sibling for reference, warranty, process optimization, and simulation. But, what could make it even more valuable? If a third sibling were added: a digital version of the product which entails everything that has happened to it throughout its lifetime. By implementing a “separation of concerns” between the recipe of what we thought we made and how the actual thing has been treated and has performed. Through doing this, we can compare what we thought we wanted to make and what we actually obtained. The third sibling would also be editable and expandable by notes from installers, customers, maintenance personnel and can be updated with new information on new replacement parts, telemetry, and IoT as well as other sensor data. Through this way we can after a year compare all of the versions of a product made on a single day with their deployed versions and actually find out how our products are doing in the wild.
By having a digital triplet we can truly have the product, its digital copy and a file of its actual use in the wild as three separate things which can all be analyzed and tracked. Interoperability of environments, systems, software, parts, and people can all be compared and continually updated. A manufacturer can not only see how well its parts have performed but also how well those serviced by Hans did. Manufacturers can track how well products do in certain countries and begin to develop more intelligent hypotheses on product life, maintenance cycles, and real-world performance. Complex systems suffer from concurrent interaction and feedback loops from various real-world forces as well as layers of interacting systems. By logging and tracking all of these in the log triplet, the effects of different firmwares, software updates and versions on part performance and interaction can be gaged.
As more firms adopt iterative versions of product development and more agile engineering methods we will have many more interactions of many different parts upon each other. With 3D printing, especially files, slicing and toolpath generation will have effects on part strength and longevity. With only the digital twin one could tell that a part was made on a particular day with a specific machine. But, you wouldn’t be able to understand what has happened to the part. The digital twin is just the product and its birth but what is its biography? What has happened to it over time? Perhaps if we learn that all of the parts that fail five years later were installed on days which were humid and rainy, then we can begin to understand the effect that life in the real world has had on that particular thing. With the complexities of additive thrown in it is through this process that we can finally begin to understand not just how unique things are made but how they live their lives as products in the real world.