For Harry Potter and his friends, the Marauder’s Map is a powerful ally in the battle against the forces of the Dark Arts. Not only does the map describe the network of corridors and secret passageways that crisscrosses Hogwarts School of Witchcraft and Wizardry, it also reveals the precise, real-time location of every character within its walls.

Today, companies are using the magic of advanced digital technologies to gain a similarly comprehensive view of their designs, facilities and manufacturing operations – and even of their products once in customer hands. These modern industrial versions of the Marauder’s Map are known as digital twins.

Siemens, whose Product Lifecycle Management (PLM) division is a major provider of the technologies that underpin digital twins, defines them as “a virtual representation of a physical product or process, used to understand and predict the physical counterpart’s performance characteristics.” In practice, a digital twin is not one thing, but many: an aggregation of different data sources and modeling technologies. The approaches and technologies used in a given application vary considerably, depending on the nature of the product and the part of its lifecycle being described.

From blueprints to virtual prototypes

Representing a physical object with a parallel collection of information is nothing new. Designers and engineers have made and stored drawings and bills of materials since before the first industrial revolution. But the development of increasingly powerful computer systems has allowed these representations to become ever more detailed and capable over time.

The transition began in the last decades of the 20th century, as companies replaced drawing boards with computer terminals. The first computer-aided design (CAD) systems were nothing more than digital drawing programs, but these systems quickly evolved. Two-dimensional digital drawings became three-dimensional digital models. Advances in simulation technology allowed engineers to model mechanisms and moving parts, and finite element analysis techniques let them study the distribution of forces through structures or the flow of fluids around them. At the same time, manufacturing engineers began to adopt computer systems for modeling the flow of products through production, helping them optimize the layout of plants and the design of processes.

TWIN TRACK: Advanced digital technologies can give companies a comprehensive view of their designs.

Closing the data loop

The advent of the digital twin required the development of another set of technologies, however: the internet of things. The availability of sophisticated but inexpensive sensors and fast network technologies allowed the previously one-way flow of digital data to become a two-way exchange. For the first time, digital representations of physical objects could become dynamic, growing and evolving alongside the real thing.

As a result, while a product is being manufactured, for example, its digital twin can accumulate information on the batch of raw material used in its construction, the tolerances to which it was machined and the worker who completed each assembly operation. And as that product enters service, the model can go on growing, recording hours of use, operating conditions and maintenance interventions.

Right now, however, the biggest use case for digital twin technologies isn’t products themselves, but the factories that make them. A comprehensive digital model of a manufacturing plant can offer its owner a huge range of potential benefits. Factories are big, complex places: Their design has to accommodate the flow of materials and people, as well as the provision of multiple services including electrical power, water, compressed air and ventilation. They are also highly dynamic. Machines are continually replaced or upgraded, production lines are reconfigured to accommodate new products or boost productivity. And smart machines and advanced automation systems generate a vast and ever-increasing flow of data.

A digital twin helps a factory owner to keep that complexity under control, providing a single home for data that was previously dispersed across different systems and business functions. And it allows companies to manage their facilities in more sophisticated ways: analyzing operational data to identify the root causes of reliability problems and spot productivity improvement opportunities, for example. Research company Gartner predicts that half of all large industrial companies will be using digital twins by 2021, and that those companies will enjoy a 10 percent improvement in efficiency by doing so.

Those benefits aren’t automatic, however. Building and maintaining a digital twin is still a complex and costly process. Many factories incorporate a mixture of new and legacy production assets, and the data describing those assets isn’t always available in an accessible format, if it exists at all. Companies often have to resort to clever tricks in order to build their digital models, such as using laser scanners to capture information on the as-built geometry and location of structures and equipment.

Twins in the wild

Wider application of the digital twin concept creates ethical challenges as well as technical ones. Companies usually own the assets they use in their factories. Once you have sold a physical product to a customer, who owns the rights to its digital twin? Concerns over privacy and the potential misuse of data are already widespread in the worlds of e-commerce and social media. Now consumers are raising the same questions about the growing number of connected products in their lives. Consumer rights advocates are already raising questions about the use of connected toys that collect data on the behavior and preferences of their users, for example.

Questions surrounding the security, privacy and ownership of digital product data remain unresolved in many sectors, but these challenges are unlikely to stop the growth of the approach in new sectors and new applications. It might not be long before a product or production line without a digital twin seems somehow incomplete. Jonathan Ward

Published: January 2019

Images: Siemens