Closed-Loop Digital Twins Are Key to Harnessing Data Complexity, Says Siemens PLM Software

In addition to being highly accurate and multifaceted, digital representations of real-world products and processes must provide feedback to the value chain, the company believes.

“The digital future of industry is no longer some far-flung promise,” said Tony Hemmelgarn, president and CEO of Siemens Digital Industry Software. In his keynote address at the Realize LIVE conference — held earlier this month in Detroit, Michigan — Hemmelgarn presented the arrival of the complexities of digital transformation as a source of both opportunities and stresses.

To survive the digital industrial revolution currently under way, organizations will have to redefine themselves at a faster pace than ever before, and find ways to make sense of “an explosion of data.” By 2025, Hemmelgarn explained, we’ll have to cope with 100 billion connected devices, “each with a dozen or so sensors collecting data, so you’ve got smarter products, you’ve got intelligent manufacturing, all producing data — and ironically, creating a much more complex world that we have to deal with.”

While it’s tempting to believe that more data always leads to clearer insights and better decisions, that’s not the case. Access to masses of data can instead lead the unwary to draw wrong conclusions, or to see connections between factors where none exist. “Studies have shown that many of the correlations are just flat-out false, they don’t make sense,” Hemmelgarn warned.

But that doesn’t mean that big data should be avoided: “We believe the true disruptors won’t be those trying to limit complexity,” he said. “We believe that businesses want to move faster, they want to lower the cost of development, lower cost of production, respond to customer demands, create new business models, and out-innovate the competition and be the leader. These businesses … will use complexity as a competitive advantage.”

A ‘True’ Digital Twin

So if the correct way to handle the rising flood of data is not to reduce the flow, how should enterprises respond? According to Hemmelgarn, “leveraging complexity requires elimination of the barriers of data that moves through the lifecycle of product development.” That means taking advantage of new technologies and embracing deep integration of solution sets: “For example, how can I use computational fluid dynamics to create a generative design of a component?”

And continued investment in a highly accurate, closed-loop digital twin is key. “A true digital twin is not an option, it’s a necessity,” stressed Hemmelgarn. The digital model which represents a real-world product, process, or production system must be tightly integrated with its counterpart, constantly changing to keep pace with any changes to the real thing, and “closing the loop” by sending feedback that can be used to improve design and production processes. “A key part of leveraging data from the utilization of the product is to understand the assets while in operation,” Hemmelgarn explained, so asset information can be captured with Maximo, an asset management system from Siemens partner IBM.

For accuracy’s sake, these digital twins must also be able to represent the mechanical, electrical, and software aspects of whatever they mirror. “If you can’t represent that, how can you have a digital twin?” Hemmelgarn asked. “If all you do is represent CAE, how can you have a digital twin that allows you to make decisions in confidence? You can’t, because the product’s much more complex than that.”

Domain Knowledge and Ecosystem Assistance

In addition to digital twins, Hemmelgarn presented other essential ways to integrate data into a product development lifecycle: leveraging domain knowledge and relying on an ecosystem of partners. In addressing the former, Hemmelgarn advocated the Mendix platform, which reportedly enables users to develop software without having to be software developers. “The challenge we found was, you’ve got all the data, we’ve got to start building applications — and the very applications we need to build, need the domain knowledge of the people that are understanding the data.” With the Mendix platform, “I can access the information, display it, and I never had to write a piece of code — this is where we’re headed with our solutions going forward.”

Part of effectively leveraging domain knowledge is personalizing environments, and to do that, Siemens is turning to artificial intelligence (AI) to learn and adapt to the way users work. “It almost creates … a personalized assistant to help you. It provides you with the best choices based on what you’re doing; it tailors the interface to the way you work.” NX software already features an adaptive user interface; “you’ll see this more and more coming into some of our applications,” said Hemmelgarn.



And to enable a robust ecosystem, Siemens believes an open, flexible, cloud-based solution is essential — but Hemmelgarn was careful to point out that it’s not the only option for customers. “All our products are in the cloud, but we will not force our customers — we will not force them — to go to subscription or go to cloud. We’ll go at the pace you want to go in doing this.”

The open environment supports the flourishing of ecosystems such as an additive manufacturing network, and rapid application development through Mendix. “It’s almost impossible to do this alone,” Hemmelgarn emphasized.



In closing, Hemmelgarn urged attendees, “Don’t try to manage complexity — it’s not going to go away. You want to find a way to leverage that complexity, you want to take advantage of it.”