Skip to content
Technology How it works Breeder — Hyperion Burner — Aegis Burner — MetroVolt AI-Native Architecture Magnets Fuel cycle Safety Roadmap
Solutions AI & Data Centers Defense & Government Grid & Baseload Neutron Detection Quantum
Learn Technical Library
Proof Publications Whitepapers Technical Library Open Science & Reproducibility The Honest Gates
Company About / Mission Leadership Environment Health & Safety Investors Careers Press Contact
3D Model
AI Architecture › L5 · Applications & Copilots
L5 · Applications & Copilots

Engineering Copilot: Superconducting Magnet Health Reasoning

Reasoning over REBCO strain, temperature margin, and quench precursors for the breeder's 16.84 T magnets and the burner's 26.49 T plug coils.

THE STACK · click to jumpL7Ecosystem & StrategyL6Experience & VisualizationL5Applications & CopilotsL4OrchestrationL3Twin Modeling & AIL2Data FabricL1Control PlaneL0Foundation▲tlmctl▼L5 · APPLICATIONS & COPILOTSAgentic copilots that reason over the machine.1Plasma Copilotscenario design2Engineering Copilotsubsystem analysis3Operations Copilotrunbooks & procedures4Agentic Toolsbounded action-taking5Knowledge BaseRAG over the fabric6Guardrailssafety-boundedMACHINE TIEReads the twin and fabric; proposes actions that route through L4.KRONOS FUSION ENERGYAI-NATIVE S.M.A.R.T. GENERATORAPPLICATIONS & COPILOTSSHEET 07REV. 2026-08L5 · AI-NATIVE STACK
L5 · Applications & Copilots — its place in the stack (left, click any layer) and its internal components (right). Telemetry rises; control descends.

The highest-consequence subsystem

The superconducting magnets are the highest-consequence engineered subsystem on both machines: 16.84 T peak field (8 T on-axis) on the breeder (Hyperion), and 26.49 T plug with 17 T throat on the burner (Aegis / MetroVolt). A quench that is not detected and dumped in time can destroy a coil. The Engineering Copilot provides the slow-timescale reasoning layer above the hardware quench-detection and dump systems.

Division of responsibility

Fast quench detection and the energy dump are hardware functions with microsecond-to-millisecond authority — the copilot never owns them. The copilot reasons on the seconds-to-hours timescale: tracking strain accumulation, temperature-margin erosion, delta-T patterns across pancakes, and joint-resistance trends that predict where a quench becomes more likely, so that operating limits and maintenance can be adjusted before the fast layer must act.

text
magnet_health(coil):
  T_margin = Tcs(B, I, eps) - T_meas      # current-sharing margin
  strain   = eps_meas vs eps_irrev_limit
  dT_map   = distributed_dT(coil)         # hotspot localization
  risk     = quench_precursor_model(T_margin, dT_map, dR_joint)
  if risk rising: propose lower I_op or inspection (via L4)
  return report(T_margin, strain, dT_map, risk, UQ)

When the copilot's reasoning indicates eroding margin it proposes a reduced operating current or an inspection window — a proposal routed through L4, never an autonomous derating. It grounds every conclusion in the strain and quench sensing telemetry and the magnet thermomechanical model, and it defers to hardware for anything on the fast protection path. See also delta-T precursors and hardware quench detection.

Content reviewed August 2026 · design-and-simulation stage