Control-Room 3D Digital-Twin Overlay

A live 3D rendering of the machine with L3 twin state painted onto real geometry — flux surfaces, thermal fields, and neutron flux where an operator expects to find them.

Geometry as the primary coordinate system

The control-room centerpiece is a real-scale 3D model of the machine — the breeder's spherical-torus vessel or the burner's tandem-mirror axis — onto which the KRONOS-CTRL twin projects its state vector. Rather than forcing operators to read strip charts and reconstruct geometry in their heads, the overlay puts every field where it physically lives: last-closed flux surface on the plasma boundary, temperature on the first wall, neutron fluence on the blanket, strain on the REBCO magnet windings.

The overlay is a render of twin output, not an independent physics model. It subscribes to the twin's ~1-10 ms state stream and re-skins geometry at display rate (30-60 fps). Because it is downstream of the twin, it inherits the twin's confidence scores and can dim, hatch, or flag any region the twin is uncertain about — see uncertainty visualization.

Layer compositing

Each physical quantity is a toggleable layer with its own color grammar (thermal in orange, structural/magnet in gray-blue, plasma in the flux palette). Operators compose the scene they need: during shape control the breeder view foregrounds the flux-surface layer and the negative-triangularity boundary; during a plug excursion the burner view foregrounds end-plug density and ambipolar potential contours along the axis.

{
  "overlay_layers": {
    "flux_surfaces":   {"source":"twin.mhd.psi",   "rate_hz":50, "palette":"flux"},
    "wall_temperature":{"source":"twin.thermo.T_wall","rate_hz":10,"palette":"thermal"},
    "neutron_flux":    {"source":"twin.neutronics.phi_n","rate_hz":2,"palette":"neutron"},
    "magnet_strain":   {"source":"twin.thermo.eps_rebco","rate_hz":10,"palette":"structural"}
  },
  "confidence_channel": "twin.confidence.per_region"
}

Selecting any surface opens a detail card with the twin's local estimate, its confidence, and the diagnostics that constrained it (a link back to model provenance). This keeps the 3D scene glanceable while making one click of drill-down always available. The same overlay runs in training-simulator mode against replayed or synthetic shots so operators build the same spatial intuition before first plasma.