Simulation vs Experiment
How Kronos uses simulation, what simulation can and cannot certify, and where experiment remains indispensable.
Simulation as a design instrument
Modern fusion design is impossible without high-fidelity simulation: equilibrium and stability solvers, transport models, neutronics, and magnet mechanics. Kronos uses these to close its designs and to produce the public numbers in this library. But a simulation is a hypothesis about hardware, not a substitute for it.
What simulation can legitimately do
- Explore a design space far faster and more safely than hardware.
- Reproduce known experiments to earn calibrated trust (validation).
- Quantify uncertainty so a result carries error bars, not just a point.
- Rehearse integration and control before a machine exists.
What simulation cannot do
Simulation cannot certify a phenomenon the models do not fully capture, and fusion is rich in such phenomena — edge turbulence, disruptions, materials under 14 MeV neutron flux, and tandem-mirror stability among them. A code that has never been checked against the specific regime it is used to predict is extrapolating, and extrapolation is a gate, not a guarantee.
This is why the breeder's Q_sci 3.076 and the burner's plug design are described as design & validation studies. They are well-founded engineering predictions awaiting the only thing that can confirm them: the machines themselves, beginning with FOAK first tritium around 2030.