Digital Twin
Definition
A simulation model that maintains a two-way data connection with a physical asset, is continuously updated, and is used for prediction and decision-making. The emphasis is on being online and co-evolving with the asset, unlike one-off offline simulation or static surrogates.
What it is
A digital twin = physical asset + sensor data link + updatable simulation model + decision loop. The point is not "how accurate the model is" but "can the model update as the asset's state changes" and "can predictions actually drive decisions" (maintenance, control, scheduling).
Key points
- Real time is a hard constraint: the twin must produce results inside the decision window, which is exactly why reduction, surrogates and AI acceleration are mandatory.
- Data assimilation is the core technology: continuously feed observations into the model to correct state (Kalman filters, particle filters, variational assimilation).
- Model degradation management: aging equipment and drifting operating conditions slowly invalidate the model, so drift detection and retraining are needed.
- Relation to surrogates: a surrogate is a common twin kernel, but a twin also includes the data link, state update and decision interface.
Common pitfalls / boundaries
- A "3D visualization dashboard" is not a digital twin; without a two-way data loop and predictive capability it does not qualify.
- The value of a twin depends on whether the decision loop is truly closed; monitoring without prediction is only a "digital shadow".