Digital Twin
A digital twin is a live, data-fed model of a physical system that mirrors its state and predicts its behavior.
More than a model
A digital twin is a simulation of a specific physical asset that is continuously updated with data from that asset. It is not a generic design model but a mirror of the actual machine — its as-built geometry, its current state, its accumulated history.
What it enables
- State estimation: infer quantities that are not directly measured.
- What-if analysis: test a proposed action in the model before the machine.
- Predictive maintenance: forecast wear and failures from observed trends.
- Operator training and design feedback on a faithful stand-in.
The data loop
A twin is only as good as the data feeding it and the models inside it. Diagnostics stream in, the twin’s state is corrected to match, and its predictions are checked against what the machine does next. Persistent disagreement points to a model that needs improvement — the twin diagnoses itself.
Fidelity versus speed
A twin must often keep pace with the real machine, which pushes toward fast reduced-order or surrogate models. The engineering art is choosing fidelity high enough to be useful and fast enough to be live, and knowing which questions each level of the twin can honestly answer.
Before and after first hardware
Even before a machine is built, a design-stage twin lets a program exercise operations and control logic against a simulated plant. For the breeder Hyperion, whose construction is planned for Q2 2027, the twin is where operating procedures can be rehearsed long before first tritium.