Digital Twin in R&D
The digital replica of a physical object
The use of digital models is a long-established standard in industry. They are used in product development (CAD), digital validation (DMU, simulation), production (CAM) and marketing (photorealistic rendering).
The concept of the digital twin goes even beyond this: it links real products and their behavior with their digital counterparts. Thus, the digital twin is a digital representation of a physical object.
Linking both worlds makes it possible to transfer current information from the real object to the digital model. This way, the digital twin provides benefits in a variety of way. For example, it can make factories more efficient or products more customer-oriented by utilizing usage data for (further) development. The digital twin also enables new business models, such as remote services, condition monitoring or data-based services, e.g. traffic forecasts.
The digital twin or digital model in product development serves as a tool for for visualization and simulation, and accelerates the planning and development process through improved stakeholder interaction and virtual validation based on field data. In addition, field data promotes the understanding of customer behavior for future innovations.
A data layer represents the connection of data objects in existing IT systems that are not directly linked or are only linked to a limited extent. The basic idea is to duplicate data in a separate database, only in exceptional cases, in order to avoid redundancies and to allow for a clear "single source of truth." The goal is to continuously provide all of the required data, independent of the department, and ideally, across companies.
The data layer does not merely link systems; it can also enrich data objects and semantically prepare or translate them for target systems, thus enabling the different data models of various corporate divisions to be networked. As a result, the data layer provides a foundation for the digital twin – the virtual replica of a physical product, process or system. For the increasingly complex product world, this is an indispensable basis for analysis and optimization. As the realization of Industrie 4.0 and IoT becomes more prevalent, and the number of sensors in the real product world that are associated to them rapidly increases, end-to-end data availability enables accelerated product development while being supported by AI systems.
The Digital Twin is not new – it dates back to 1970. This was the year that astronaut John Swigert reported to NASA's Mission Control Center in Houston with the words "Houston, we have a problem." An explosion in the oxygen tank of the service module caused the command module's power, light and water supplies to collapse. NASA engineers determined ways to repair the resulting damage with on-board resources. They did this on a 1:1 copy of the Apollo 13 module, which was located on Earth and called “the twin.“
The term "digital twin" was first coined by Michael Grieves. He used this term in his research in the topic of product lifecycle management (PLM) at the University of Michigan to describe a digital 1:1 image of a real object. Initially, Grieves simply referred to it as a double. Since 2016, digital twins have been successfully implemented in various areas such as the visualization of an engine block or for the port of Rotterdam. But this is only the beginning. Due to the ever-increasing number of sensors – over 20 billion in 2020 – Gartner predicts that half of all the largest industrial companies will be using digital twins productively by 2021. This trend will only continue to accelerate.
There is not just one type of digital twin. Digital twins can be distinguished based on their scope and type of application. There are digital twins for individual components; for example, valves that use a sensor to determine volume flow, or products that consist of several components, or entire systems that allow a complete factory to be mapped as a digital twin.
Different applications can be realized by using digital twins: a pure data twin determines the state of real objects via sensors and displays it in the virtual world. On the other hand, a simulation twin, goes beyond pure visualization - it uses the data that is transmitted through sensors to simulate future behavior. An example of this is the use of machine data (throughput time, idle times, buffer sizes, processing times, etc.) in a simulation to determine possible optimizations (reduction of buffer sizes, increase in output quantity, etc.). A control twin not only simulates possible optimizations, it also intervenes in the real world: based on the determined optimization, the digital twin sends control commands to the real machine. These types of control twins allow smart factories – factories capable of transformation – to be realized in the first place.
UNITY at a Glance
25 years of innovation and digital transformation. Find out more in our annual report.
Rise of the Digital Twin
How the digital twin becomes reality. Lecture by Uwe Uttendorfer, Mercedes-Benz AG and Dr.-Ing. Oliver Hornberg, UNITY.
The Future of Digital Engineering
UNITY connects you to the future of Digital Engineering.
PLM Meets Blockchain
Article about the role of the blockchain in the PLM context (published in automotiveIT 07-2019)
Digitalization of Product Development
Read how to make product development future-proof in our OPPORTUNITY.
Read more about our current topics for experts and decision-makers.