Fusion Codes

Predictive Transport Modeling

Predictive transport codes forecast plasma profiles for a proposed scenario by combining transport models, sources, and self-consistent equilibrium.

The goal of prediction

Design and scenario studies require predicting what a plasma will do before it exists. A predictive transport code couples a 1D transport solver, a transport model that supplies diffusivities, heating and fueling source models, and an equilibrium solver, then evolves them together toward a self-consistent stationary state or through a full time history.

The coupling loop

Kronos motion — fusion

Profiles determine the pressure and current, which change the equilibrium; the equilibrium changes the geometry and stability, which feed back into the transport model. A predictive run iterates this loop until the pieces are mutually consistent. Loose coupling updates each module in turn; tight coupling solves them together for numerical stability.

Sources and sinks

Auxiliary heating from neutral beams, ion-cyclotron, or electron-cyclotron systems
Alpha-particle self-heating from fusion reactions
Radiative losses from impurities and bremsstrahlung
Fueling from gas puffing and pellet injection

Sensitivity and uncertainty

Because predicted performance depends strongly on the transport model, credible predictive work brackets results across models and scans key assumptions such as the pedestal height and the density profile. A single point prediction without a sensitivity study overstates confidence.

Stiff transport

Turbulent transport is often stiff: above a critical gradient, diffusivity rises steeply, clamping the gradient near a threshold. Stiffness makes core temperatures depend heavily on the edge boundary condition, which is why pedestal prediction is coupled to core prediction in serious scenario modeling.

Predictive transport is a central tool for evaluating whether the Hyperion breeder design point is self-consistent under stated heating and confinement assumptions, always reported with its uncertainty range rather than as a single guaranteed number.