Most industry sectors in dialog with ARC indicate that digital twins will be used to simulate, predict behavior, and optimize products, assets in the field, infrastructure, plants, and production systems across process industries and discrete manufacturing, thereby recognizing the importance of the concept. ARC believes it is also because of the increasing maturity of the technology and implementation options that a significant percentage of companies and organizations are planning to use some form of a digital twin to improve and optimize products, processes, and assets.
At the same time, ARC Advisory Group found that many users, providers, and analysts have developed views of the digital twin that are often in part inconsistent, incomplete, or sometimes overly detailed. As a result, conversations about application or implementation of digital twins often drift off discussing what the digital twin really is instead of focusing on how to apply digital twins to reach business goals. ARC also noted that many users are challenged on how to implement digital twins.
To overcome these challenges, this report discusses the definitions of digital twins, and proposes a simple, operational definition compatible with most other definitions, while delimiting what a digital twin is from its extensions and applications. This report further attempts to lay a foundation for goal-focused work on digital twins by examining key technologies, applications, and maturity measurements. This includes sections on digital twin scope, ecosystems, virtual simulation, data modeling and architectures, use cases and implementation, and maturity models. This “map” of the digital twin landscape helps orient stakeholders, supports digital twin strategy and implementation, and exposes opportunities for synergies from diverse digital twin-related activities.
Across manufacturing, process industries, and infrastructure, digital twins can be used to manage the information related to design engineering, test and validation, commissioning, construction, and operations. Information across the design, build, operate, and maintain lifecycle is connected by a Digital Thread that ultimately enables the implementation of the digital twin. Except in the design and build stages, digital twins, by definition, are in sync with their physical twin across the product/process lifecycle.
Information needs context to be interpretable, understandable, and actionable. Industry agrees that an information model organizing the information in a standardized way greatly helps understanding and communicating about information. Intra-company exchanges could be based on a company-specific data model, and in an ecosystem of equipment and service providers, engineering and construction firms, integrators, and owner-operators, a more widely used, ideally universal, standard is more efficient. The basis of the digital twin is the information model. The more standardized the model, the easier different parties can collaborate in building, operating, and maintaining digital twins.
A variety of applications can build on the information layer, sharing this data and information. The information can be used to build behavior models, adding simulation and prediction capabilities to the digital twin. Examples of applications are design and engineering, 3D modeling, operator training, planning of operations and maintenance scenarios, utility planning, model-predictive control, asset performance management and more.
Successful implementation can range from small ad hoc applications to strategic, company-wide, long-term strategies. Organizations must carefully determine their strategy, based on costs, benefits, timeline, goals, capabilities, and resources. To realize meaningful benefits when implementing a digital twin for manufacturing, process, or infrastructure, organizations must think holistically. Engineering detailed and complex design and operational models, connectivity and synchronization are major achievements but should not distract from focusing on goals and ensuring value capture from applications, and their potential synergies such as optimization of process and assets in a refinery. Doing so will unlock the value of digital twins as tools to help inform business decisions and to develop new business models. Fundamentally changing business models also requires changing analytic models. The results of those analyses will be robust and accurate if the digital twins are designed to deliver them.
Multiple technologies have emerged in recent years that are instrumental in driving the advancement of smart manufacturing in the discrete industries and advanced asset management in the process industries. These technologies include advanced analytics, artificial intelligence (AI) and machine-learning (ML), operational and performance optimization, cyber-physical systems, Industrial IoT and the digital twin. Each of these technologies is changing the face of many industries today, and combinations of them can potentially create even bigger impact on company performance.
Defining Digital Twins
One motivation for this report is to clarify the definition of Digital Twin and reduce confusion caused by the many (and sometimes inconsistent) definitions now being used. Many acknowledge that the first definition was given by Michael Grieves and John Vickers in a presentation in 2001, which has been reported on by the same authors in 2016. VanDerHorn and Mahadevan, in a research article from 2021, identified 46 definitions of the digital twin and analyzed them. They found commonalities among those definitions and formulated “a consolidated and generalized definition for a Digital Twin […]: a virtual representation of a physical system and its associated environment and processes, that is updated through the exchange of information between physical and virtual systems.” This adds the environment to the original definition of Grieves and Vickers, and this makes sense, as the system interacts with the environment.
This definition is simple and yet comprehensive. It has the advantage of being consistent with many definitions, including ARC’s and with the key elements of the definition by the Digital Twin Consortium, which adds “synchronized at specified frequency and fidelity.” The definition includes the closed loop between physical and virtual twin: by actuation the state of the physical twin is modified. Its changed state can be measured (physical) or inferred (physical and virtual), interpreted and analyzed, leading to a decision (virtual), that can be passed as actuation to the physical system, and so on. Accepting this definition, which we recommend, the difference between the digital twin and its applications becomes clear, even if the application is an extension of the digital twin that seems to be an integral part of it. In the section (below) on closing the loop between the virtual and the physical system we provide several examples.
Table of Contents
- Executive Overview
- Defining Digital Twins
- Dimensions of Digital Twins
- Implementing The Digital Twin
- Use Cases
- Conclusions and Recommendations
ARC Advisory Group clients can view the complete report at ARC Client Portal
If you would like to buy this report or obtain information about how to become a client, please Contact Us
 Behavior models are computational models that describe how the variables of interest relate to each other in steady state, or in dynamic conditions, over time.
 The word “physical” is to be associated with the real world in this context. It could be confused with systems that behave according to the laws of physics. A physical system would also encompass state transformations that behave according to laws of chemistry, nuclear physics, etc.