The digital twin is so much more than just a piece of virtualization software. It’s essentially a data switchboard for digital ecosystems, consisting of manufacturers, suppliers and customers, and is the key to data-based business models with completely new value propositions.

<<< Start >>>

The digital twin is a decisive building block for Industry 4.0. Image: Accenture/Mackevision

<<< End >>>

As the tanker comes up to the oil platform, reality shifts: The ship, the rig—the entire world—rotate, vanish, and give way to an underwater view on the pipes that run underneath the setting. Name tags and data callouts appear out of nowhere to indicate the current status of the pipes in view and highlighting predicted issues, and columns of sensor readouts begin to scroll across it all.  

"Now we can see the bottom half of the platform’s Digital Twin”, explains Sophie Ackers, a technology consultant within Accenture’s Resources practice. “Here is where we show all the key in-field info in real-time." The entire scenery isn’t real, of course—it's a virtual, high-res 3D replica of a real oil rig (and incoming ships) in the North Sea, enriched with real-time data inputs from the physical equipment. It’s also the very future of how industrial sites and equipment will be managed. 

The Future of how industrial sites and equipment will be managed

Digital Twins like the one Sophie Ackers is showing are already making their way into a wide range of industrial settings, enabling engineers to simulate, optimize, test and operate equipment in entirely new ways to predict errors before they happen, reduce cost and risks, and increase safety and sustainability. 

As the name implies, a digital twin is a virtual replica of an actual process, service or product. It’s possible to “virtualize” almost anything, from machines to factories, cars and their driving characteristics and even entire cities. Most man-made objects can be captured and recreated using virtual methods on a computer. Developers can then change functions, test new settings, and calculate complex simulations at a fraction of the costs of traditional methods. 

Digital twins are “data bundles” that generate in real time and constantly update, and several factors are required to create them: 

  1. Inventory information from the company’s IT systems, such as design and master data relating to the object that is being digitized.
  2. IoT sensors on the machine, system or product, which continually record operating data.
  3. A Cloud platform containing a digital twin platform, where the data is brought together.
  4. Analytics algorithms that continually examine the platform data, “understand” it, and make their findings available to other applications.
  5. Where possible and expedient, 3D models and visualizations of a twin, so that it's represented within virtual reality applications.

Free flow of data via the digital thread

Digital twins are not built for singular use during specific periods but instead “accompany” their real-life counterparts throughout entire life cycles. This enables the digital twin to consistently supply valuable information on its day-to-day use, which in turn enables the continual optimization of a machine while it’s in operation. Manufacturers have many uses for digital twin feedback, such as to improve future products. 

The unimpeded flow of data from the “real” machine to the IT systems of users and manufacturers and back is known as the “digital thread”—it remains invisible as it continuously runs alongside the physical product. Present-day digital twin applications demonstrate the incredible potential of this technology and how it generates new data-driven business models. 

Accenture subsidiary Mackevision, for instance, is an expert in computer-generated imagery and has developed a digital twin of a motor block for automaker Daimler. The replica was solely based on existing design data and permits developers at Daimler to simulate the behavior of motor components at various speeds. It also facilitates cooperation between engineers and designers across departments and streamlines numerous work processes. 

Similarly, GE recently leveraged the power of digital twin technology to an impressive degree. The company improved the yields of GE wind farms by developing a digital twin for each wind turbine based on twenty different items of configuration data. The data collected from a turbine during operations is evaluated in real time, compared with the output data of other turbines on the farm, and optimized on that basis. This approach has increased wind turbine efficiency by up to 20 percent, according to GE. 

Dassault Systèmes implemented a digital twin on a significantly larger scale, however, with a virtual replica of the entire city state of Singapore. This digital twin may look like a humble 3D map at first glance, but in truth it serves as a comprehensive “playing field” for the city’s planners. For example, it can estimate the impact of construction work on traffic, precisely predict the levels of noise pollution generated by high-speed rail lines, or perform simpler tasks such as testing traffic light phases. 

A boost for new business models

The examples mentioned above highlight how digital twins optimize processes during product development, enhance the efficiency of systems while in operation, and make it easier to coordinate conflicts between targets in complex systems. This is already revolutionary in and of itself, yet the true significance of digital twins lies in the new value propositions that are accessible to customers. 

Those who use virtual methods to recreate the behavior of a product or service during its operation can also continually improve it with the help of simulated effects. This forms the basis for new data-driven business models that create tangible added value for users e.g. a car engine that consumes less fuel and is subject to less wear and tear as its performance constantly adapts to driving styles and traffic conditions. 

The almost perfect simulations of the real world produced by digital twins help to minimize business risks during the development and marketing of new products. Significantly less effort is required to build and test physical prototypes, saving time and money, and mistakes are less likely to occur during the planning stage of complex projects: a factory build is adapted to the individual conditions and production processes of a manufacturer, and the interplay between all entities is tested in advance.

Simplified cooperation

One of the digital twin’s key strengths is how it can simulate extremely complex systems in real time. In doing so, it breaks down the boundaries between departments and companies and compiles data from a wide range of areas on a single platform. 

It could, for example, enable an oil and gas company to keep a constant overview of a platform’s current production data, know its partner’s and vendors current inventory for specific components (like, say, scaffolding parts), and stay up to date with the utilization of other rigs on the other side of the world. A completely new degree of transparency is therefore established with respect to all processes connected to the production of the company—throughout the entire value creation chain. 

The digital twin also provides new opportunities for the entire innovation process with partners. Teams of developers, researchers, designers and scientists can work together on projects regardless of their location and develop and improve a product’s virtual model together. This type of co-innovation shortens development cycles significantly and simplifies coordination between stakeholders. 

Similar benefits exist for customer service teams: in the future, service employees who previously had to visit a customer site to fix a defective elevator will be able to identify the cause of a fault by examining the elevator’s digital twin and fix the fault remotely. To prevent breakdowns, the virtual copy of an elevator could use operating data to independently detect the early signs of a fault. 

So how do we tap into this technology?

It’s comparatively easy to create virtual replicas of products and machines using design and operating data, but a significantly greater number of data points are required for more complex simulations. Increasing the use of smart products embedded with IoT sensors will increase the amount of available operating data, leading to more precise insights from digital twins. 

Companies that collect this operating data via a digital platform and decide to offer their customers new value propositions on this basis should ask themselves: What is the best way to proceed? The answer, of course, is to “Think small”. 

It’s advisable to initially avoid using digital twins to enable ecosystems of new services for customers by way of large-scale strategies and ambitious plans. Instead, approach this technology in a manner akin to agile software development i.e. focus on minimum viable products. To make the most of the digital twin’s functionality during the early stages of implementation, develop products that offer users a concrete and tangible value proposition on a day-to-day basis despite minimal configuration. 

<<< Start >>>

Digital Twin

Frank Riemensperger about the critical building block to Industry X.0 at Hannover Messe 2018.

<<< End >>>

Subscribe to Accenture's Industry X.0 magazine Blog Subscribe to Accenture's Industry X.0 magazine Blog