A digital twin is a dynamic virtual representation of physical assets, processes, or systems.

When searching for a definition of digital twin, you’ll find a lot of definitions that sometimes only differ on details. From those different versions, we derived a definition that captures the essence of digital twins.

A digital twin is a dynamic virtual representation of physical assets, processes, or systems.

It can be used for understanding and interacting with the actual world and shaping the future world.

Or, if you’re looking for a more comprehensive version of the definition:

A digital twin is a dynamic, real-world virtual representation of monitorable, connected (or soon-to-be connected), unique physical assets, processes, or systems. Representing past, present, and future data, digital twins deliver smart solutions by integrating cross-domain services, using smart software (like machine learning, simulation, VR/ AR) to help you understand and interact with the physical world and shape the future with the objective of improving performance and/or reducing risks.’’

The rationale behind our definition of digital twins

A digital twin is a dynamic virtual representation of physical assets, processes, or systems.

    • We prefer the term “… virtual representation of…” over "digital representation of". A good definition should not be circular.

    • We could also use the word software representation – ultimately, it is software that makes the representation real – but software is “only” used for it, so virtual is better.

    • Dynamic refers to the fact that what is represented by the digital twin is changing over time. This can include near-real-time changes like the temperature of an object in the system, but also the state of the system (e.g. from a design point of view) or a maintenance or operational status of an object like a pump.

    • Explicitly relating to something physical is also common across definitions because the relationship with physical assets is crucial to digital twins.

    • A lot of definitions refer to physical assets only and not intangible assets.

    • We do not use the more detailed decomposition of components of a physical asset.

    • Processes are added because the status of an asset is directly related to the processes where they are used.

    • These processes can also be part of a system (as a collection of assets, processes, and more). With a system you can also represent something large, such as a transport system, a whole enterprise, or even an entire city.

    • The hierarchy of assets, processes, and enterprise is too small because a digital twin can cover more than one enterprise. Using systems avoids this issue.

    • System of systems means that in most cases digital twins will add value if data from more domains are related via services on an IT and application level. We call this cross-domain.

    • To prevent the definition from being too specific, we avoid:

    • references to an industry (process industry, water, transport...)

    • references to a specific type of asset (building, pump, rolling stock, ...)

    • references to all kinds of possible functions (simulation, control, insight, ...); you can never include everything

    • references to a maturity level

    • The term ”data” is not included in the definition. Although a digital twin revolves entirely around data, data is already included in the word “representation”.

    • We did not include the words “real-world”, ”real-life” or ”real-time”, or related words around frequency. That’s because our definition already covers time by describing digital twins as dynamic.

    • A digital twin, however, can cover the entire time-axis of past, present (real-life), and future (options).

    • We did not add a goal or outcome in the definition but added this in the subtitle.

Definitions digital twins