The Digital Twin
A digital twin mirrors the design and, in time, the built machine, linking simulation to measured behavior for design and validation.
A model of the machine
A digital twin is a detailed computational model of the machine that runs alongside the physical program. In the design phase it embodies the frozen physics, Q_sci 3.076, 85.0 MW, 9.66 MA, 16.84 T, negative triangularity -0.30, so the design can be explored and interrogated. Once hardware exists, the twin is updated against measured behavior.
What the twin is for
- Represent the frozen design point consistently across the program
- Explore operating scenarios and sensitivities before hardware exists
- Compare predicted behavior to measured results at each gate
- Carry lessons from FOAK forward into NOAK design
Fidelity and honesty
A digital twin is only as good as its physics and its correspondence to reality. During the design phase the twin is a simulation, not evidence of performance. Its representations, including plasma shape, must match the canonical design; for example the negative triangularity must be shown correctly as an outward-opening plasma, not omitted or reversed.
Role in validation
As gates are passed, measured data refines the twin, and the refined twin improves predictions for later units. This closes the loop between simulation and hardware described in the validation program, and it is where the twin earns its keep, by improving as the machine teaches it.
The digital twin models a machine not yet built. It is a design and analysis tool, and its outputs are design values until validated against hardware at the gates.