The Digital Twin Consortium, a member-supported industry group comprising private companies, government bodies and academic institutions, defines a digital twin as “a virtual representation of real-world entities and processes, synchronized at a specified frequency and fidelity.” What is key is that the digital twin is a live data connection, “synchronized at a specific frequency and fidelity.”.
“Often, people confuse a simulation model with a digital twin. What makes the model a digital twin is the two-way information flow from that physical asset to the twin,” says Prith Banerjee, chief technology officer, Ansys.
Synchronized with live data, the twin faithfully reflects the history and characteristics of the physical product, from its manufacturing defects to its accumulated wear and tear. A twin that fulfills this requirement could be the primary data source for the enterprise.
In the popular lingo of product lifecycle management (PLM), it could become “a single source of the truth.” Different stakeholders should be able to rely on it for the maintenance schedule, replacement part number, supplier info, structural simulation and other functions.
“While there are many elements that might make up the comprehensive digital twin—3D geometry, analysis models, simulations, requirements, manufacturing data, supplier information and so on—the comprehensive digital twin includes all of the information required to define your product, how the product will be produced and how it will function in use,” says Dale Tutt, vice president of Industry Strategy, Siemens Digital Industries Software.
Can this vision ever be realized in practice? Some experts point out that while possible, the data-management burden should not be underestimated.
While NVIDIA’s Omniverse has seen adoption momentum, it’s by no means the only metaverse to house digital twins. Other tech giants like Facebook, Microsoft, Apple and Amazon are also potential providers, competing to be the landlords of your industrial digital twins. So most likely, we will have not one metaverse but various metaverses, housing industrial digital twins belonging to different manufacturers.
Banerjee expects almost all major cloud providers to compete for market share in digital twin custodianship. That means digital twin operators will likely face a dilemma in picking a platform.
It’s not an easy decision. It amounts to making a long-term investment in digital infrastructure for mission-critical items.
“Each company will have to make that decision,” says Banerjee.
The decision will most likely be influenced by existing relationships, the reputation and longevity of the platform provider, and the specific features that cater to the digital twin operator’s industry.
Ansys offers Ansys Twin Builder, a technology stack for creating and maintaining analytics-driven, simulation-based digital twins.
“You can build a digital twin based purely on data analytics, but our value proposition is, we can offer you simulation. You can collect temperature variations with time stamps, but if you need to understand why your product is heating up during certain parts of the day, we can give you the technology to simulate and analyze how and why it’s heating up,” explains Banerjee.
“Siemens supports USD in the Siemens Xcelerator platform and is collaborating with NVIDIA with the next version of the format,” says Tutt. “There are many methods of integrating the data within the digital twin across organizations and ecosystems.
For example, in Germany, Siemens is exploring how the Asset Administration Shell can be used to exchange data for machinery manufacturing.”
“We at Ansys are following the development of OpenUSD very carefully, to see if and when we should join,” Banerjee says.
Adds Gusev, “Given its impressive ability to encapsulate various types of data, as well as their relationships and properties, OpenUSD has strong potential as a future standard language for digital twins.”
But Gusev doesn’t think it can replace other formats like STL or OBJ, already entrenched in workflows like 3D printing and VR applications, respectively.
“Different applications often have varied requirements and data usage patterns, which are reflected in their internal data formats. In other words, although OpenUSD could become an interoperable data format, it may not necessarily replace proprietary formats optimized for specific use cases,” says Gusev.