Plasma Copilot: Fueling and Density Advising
Advising on gas-puff and pellet fueling to hold density profiles for the breeder and central-cell/plug density for the burner, inside stability limits.
Density as a shared control axis
Fueling sets density, and density sets a large share of the physics on both machines. On the breeder (Hyperion), density controls the Greenwald-fraction margin, radiated-power balance, and fusion rate at the 85.0 MW design point. On the burner (Aegis / MetroVolt), central-cell and end-plug densities set the ambipolar potential and confinement. The Plasma Copilot advises on the fueling schedule that reaches a target density profile without crossing limits.
Actuators and their dynamics
- Gas-puff servo valves: fast, edge-weighted, high recycling
- Pellet injection: deeper deposition, discrete timing, core fueling
- Breeder: hold n/n_Greenwald below the density-limit margin
- Burner: hold end-plug density for phi confinement, central-cell density for pressure
The copilot proposes a schedule and predicts the resulting profile with the twin's transport model, then checks it against density-limit and radiation constraints. Fueling advice for the breeder is explicitly linked to disruption risk because density-limit proximity is a dominant precursor.
advise_fueling(n_target(rho)):
schedule = mix(gas_puff(t), pellet(t))
n_pred = twin.transport(schedule)
check: breeder -> n/n_G < margin, P_rad/P_heat < f_rad_max
burner -> n_plug >= n_plug_min(phi), n_cc in band
return schedule + margins + UQ (proposal only)
Fueling schedules are proposals to the L1 fueling controllers. The copilot never commands a valve or an injector directly; a change is routed through L4 authorization and, for anything with a plant side-effect, requires operator approval. The tritium and helium-3 product framing (breeder ~4 kg/yr tritium class, ~1.97 kg/yr helium-3) is context the copilot can cite, not a knob it tunes. Fueling advice is validated against the twin's transport module and the frozen design point.