Digital twins need trustworthy measurements to earn their name

A detailed computer model can show how a factory machine should operate. A digital twin goes further by connecting a digital representation with information about the physical system, allowing its condition and behaviour to be examined as they change.

The quality of that connection is crucial. Sensors, timestamps and the assumptions inside a model influence whether the digital view reflects reality. A visually impressive display may still be a poor guide if its measurements arrive late or represent only part of the process.

NIST's manufacturing research includes methods for implementing and testing digital twins, along with work on standards. These efforts address a practical challenge: different tools and equipment need a dependable way to describe and exchange information. The intended use also matters, whether it is maintenance planning, process analysis or testing a proposed change.

For a manufacturer, success should be judged against a defined question. Can the model detect a relevant deviation, improve a decision or explain a failure? Checking its predictions against observations helps establish that value. A digital twin becomes useful through evidence of accuracy and fitness for its task, rather than through the realism of its graphics.