Physics · validation

Validation, verification & UQ.

Nothing here asks for trust. The engines re-derive the frozen design points on load, the anchor regression pins every number, and the deposits let anyone reproduce the record from scratch.

Frozen-anchor regression — 19/19

kronos_toolkit.verify.run_regression()

The engine must reproduce every frozen anchor within its stated tolerance: 5 breeder anchors (config 22021), 5 burner anchors (the M-45 design point), the tritium surplus law, the analytic neutronics surrogate, and 5 high-fidelity surrogate checks. This table is the live output, re-run at build time — source in engine/kronos_toolkit/verify/, raw results in data/regression.json.

CheckExpectedGotatolVerdict
breeder.Q3.0763263.0763260.0001PASS
breeder.P_fus_MW85.04057785.0405770.5000PASS
breeder.Ip_MA9.6564839.6564830.0100PASS
breeder.T_kg_yr3.8362293.8362290.0100PASS
breeder.f_n0.7971200.7971200.0001PASS
burner.Q_E1.3181201.3181210.0001PASS
burner.P_fus_MW4298.5000004298.4993040.5000PASS
burner.P_n_MW233.950000233.9509440.2000PASS
burner.f_n_pct5.4426005.4426190.0100PASS
burner.phi_i_keV248.750000248.7481540.0500PASS
surplus.1.341.6900001.6912890.0200PASS
surplus.1.803.9800003.9795030.0200PASS
neutronics.local_tbr1.1517001.1517000.0010PASS
neutronics.net_tbr0.7424000.7424000.0010PASS
hifi.gyrokinetic0.6896550.6896550.0000PASS
hifi.neoclassical0.6324560.6324560.0000PASS
hifi.struct_fea0.7850000.7850000.0000PASS
hifi.kinetic_mirror0.4283880.4283880.0000PASS
hifi.nucdata18.35300018.3530000.0010PASS

The in-browser engine on these pages runs the same check on load (self_check()) before any control unlocks.

Validation · Breeder

Validation Suite

Every check in the reproducibility suite (V01–V23, 23 checks). PASS = reproduces/holds; REQUIREMENT = an open item stated honestly; INFORMATIVE = a scan finding. Click any check for its full claim, criterion and computed value.

23 shown · 14 PASS · 4 open
IDCheckCategoryVerdict
V01Hyperion frozen design point reproducesdesign-pointPASS
V02Gain is SOLVED not scaledgainPASS
V03Fuel follows purpose (min Ip)fuel-tradePASS
V04Closing-config countsfuel-tradePASS
V05Toroidal field is not a second leverfield-leverPASS
V06Neutron multiplication is D-T onlyblanketPASS
V07TBR is a design leverblanketPASS
V08Startup inventory inside civil precedentfuel-cyclePASS
V09Heating power density is physicalpower-densityPASS
V10betaN respects the no-wall limitstabilityPASS
V11Peak field within demonstratedmagnetPASS
V12Uckan / enclosed-current identitycurrentPASS
V13Ash basis is a declared assumptionhonestyINFORMATIVE
V14Triangularity declaration (H27)honestyREQUIREMENT
V15External current drive is requiredcurrent-driveREQUIREMENT
V16He-3 spectrum-destruction rulefuel-cycleINFORMATIVE
V17The current requirement is structuralfuel-tradePASS
V18Design-space minimum-current findingdesign-spaceINFORMATIVE
V19Divertor exhaust is OPEN not adverse (H24)exhaustREQUIREMENT
V20Negative triangularity eliminates ELMs (H25)stabilityPASS
V21NT is not a detectable performance lever (H26)confinementINFORMATIVE
V22Steady-state and damage-life-limited (H29)operational-honestyREQUIREMENT
V23Per-MW tritium rates must name fuel and basis (H35)fuel-cycleINFORMATIVE
VALIDATION_MANIFEST.csv · reproduced from DOI 10.5281/zenodo.21746157. Requirement-class items are shown as plainly as the passes — that honesty is the point.
🔒 Validation Suite is Kronos-only. The overview above is public — the full data, tables, live calculations and engineering detail on this page require a Kronos team sign-in.
Verification · Breeder · Hyperion

Verification & V&V

The systems-engineering layer above the physics: every top-level requirement traced to its verification method, the multiphysics analysis that addresses it, and the specific deposit checks that are its evidence. Status is honest about maturity — nothing is hardware-tested because nothing is built yet; “verified” here means verified by analysis and reproducible validation against the frozen model.

Requirements
10
top-level, traced
Analysis-verified
4
frozen + reproducible check
Model-only
5
needs full FEA / CFD / MCNP
Open
1
needs a test article / demo
Requirements traceability matrix
10 shown · 4 verified · 5 model-only · 1 open
IDRequirementSubsystemMethodAnalysisValidation evidenceStatus
REQ-B01Fusion gain Q ≥ 3.0 at the design pointPlasma physicsAnalysisPhysics Engine ↗V01V02ANALYSIS-VERIFIED
REQ-B02Tritium breeding — on-model, thin margin (net TBR ≥ 1.0; lever 1.03–1.20)Blanket & fuel cycleAnalysisNeutronics ↗V06V07V08V16V23MODEL-ONLY
REQ-B03Peak on-conductor field ≤ 20.1 T (REBCO limit)Magnet systemsAnalysisStructural ↗V05V11ANALYSIS-VERIFIED
REQ-B04MHD stability — β_N below the no-wall limitPlasma physicsAnalysisStructural ↗V10V20ANALYSIS-VERIFIED
REQ-B05Plasma current sustained (I_p, current drive)Current driveAnalysisElectromagnetic ↗V12V15ANALYSIS-VERIFIED
REQ-B06First-wall lifetime ≥ 2 fpy at 1.97 MW/m²First wall / PFCAnalysisNeutronics ↗V09MODEL-ONLY
REQ-B07Divertor exhaust within target material limitsDivertor / PFCAnalysisThermal ↗V19MODEL-ONLY
REQ-B08Structure survives disruption/EM + seismic loadsStructure / vesselTestStructural ↗— none yetOPEN
REQ-B09Magnet quench protection (hot-spot ≤ 200 K)Magnet systemsAnalysisElectromagnetic ↗— none yetMODEL-ONLY
REQ-B10End-of-life waste ≤ Class C low-level wasteMaterials / safetyAnalysisNeutronics ↗V13V14V22MODEL-ONLY
Interface control register (ICD)

Where two subsystems meet, the controlling parameter and its frozen value are the interface. These are the interfaces an as-built machine must hold to the design.

IDInterfaceControlling parameterFrozen valueOwner
ICD-B1TF magnet ↔ vessel structurePeak field / reacted hoop load16.84 T · hoop reacted by coil caseStructural ↗
ICD-B2Plasma ↔ first wallCombined heat + neutron flux1.97 MW/m²Neutronics ↗
ICD-B3Blanket ↔ tritium plantTritium throughput~4.0 kg/yr · TBR ≥ 1.05Neutronics ↗
ICD-B4Divertor ↔ coolant loopPeak surface heat flux≤ 10 MW/m² at 90% radiatedThermal ↗
ICD-B5Cryostat ↔ magnet cold massInsulating vacuum / temperature1.2×10⁻⁸ Pa · 4.5 KThermal ↗
Digital thread
Requirement
10 traced
Multiphysics analysis
4 domains
Validation evidence
V01–V23
3-D component
As-Simulated overlay
Open deposit
DOI
Requirements traceability over the frozen model · validation evidence reproduces from the deposit DOI 10.5281/zenodo.21746157. Nothing is hardware-verified — nothing is built. “Analysis-verified” = the requirement is met by analysis and a reproducible deposit check; “model-only” needs a full solver; “open” needs a test article. SYSTEMS ENGINEERING · NOT A COMPLIANCE RECORD
🔒 Verification & V&V is Kronos-only. The overview above is public — the full data, tables, live calculations and engineering detail on this page require a Kronos team sign-in.
Uncertainty Quantification · breeder

UQ / Monte-Carlo Studio

2,000 Monte-Carlo samples run through the real evaluator, varying the genuinely uncertain inputs (H₉₈, T_i0, He-4 ash, Greenwald fraction, TBR) around the frozen point. Regenerated at the freeze from dt_evaluator.py — not the pre-freeze scan.

mean 4.0759 ± 2.332 · p05–p95 1.6386–8.7738 · frozen 3.0763

Distribution 2000 samples

p05p95meanfrozen1.2417.88Fusion gain Q

Sensitivity of Q tornado · |Δ|

H₉₈6.25
T_i01.43
He-40.93
Greenwald0.19
blanket0.00
Swing of Q when each input moves p05→p95 (others frozen). H₉₈ dominates.
P(Q ≥ 3)
57%
of 2000 samples
P(Q ≥ 3.076 frozen)
50%
at or above frozen
Biggest lever
H₉₈ confinement
ΔQ 6.3
Monte-Carlo over the real evaluator at the frozen design point · reproduced from DOI 10.5281/zenodo.21746157. H₉₈ dominates the spread — confinement quality is the design's biggest uncertainty.
🔒 UQ / Monte-Carlo Studio is Kronos-only. The overview above is public — the full data, tables, live calculations and engineering detail on this page require a Kronos team sign-in.
Live version The interactive version of this panel runs in the team Lab — the static record is shown here.
Metrology · Breeder · Hyperion

Metrology · As-Built vs As-Designed

The layer that closes the digital thread: measured geometry checked against the frozen design, with each deviation propagated back into the multiphysics analysis and the requirement it affects.

⚠ Representative — no as-built article exists (nothing is built). The as-built values below are synthetic deviations at typical fabrication tolerances, shown to exercise the workflow. This tab is the receiving structure for real CMM / laser-tracker / photogrammetry / NDE data at manufacture.

Measurement systems
CMM
Coordinate measuring
component · ±5 µm
machined-part conformance
Laser tracker
Portable interferometric
assembly · ±25 µm/m
coil & module placement
Photogrammetry
Optical, large-volume
structure · ±0.1 mm
vessel & envelope as-built
NDE
UT / RT / PT
volumetric
welds · joints · bond integrity
Controlled features
10
on the ICD
In-tolerance
6
conform to frozen design
Out-of-tolerance
3
need MRB disposition + re-analysis
Pending
1
await a fabricated article
Dimensional conformance — controlled features
10 shown · 6 in-tol · 3 out-of-tol · 1 pending
FeatureSystemNominalTolAs-builtΔAnalysisStatus
Major radius R₀Laser tracker1.200 m±2 mm1.201 m REP+1.30 mmPhysics Engine ↗IN-TOL
Minor radius aLaser tracker480.00 mm±1.5 mm480.60 mm REP+0.60 mmPhysics Engine ↗IN-TOL
TF coil radial positionLaser tracker500.00 mm±0.5 mm500.70 mm REP+0.70 mmElectromagnetic ↗OUT-OF-TOL · MRB
TF-to-TF spacing (×16)Laser tracker405.00 mm±0.5 mm405.30 mm REP+0.30 mmElectromagnetic ↗IN-TOL
First-wall gapCMM20.00 mm±1 mm20.40 mm REP+0.40 mmThermal ↗IN-TOL
Blanket module fit-upCMM5.00 mm±1 mm6.20 mm REP+1.20 mmNeutronics ↗OUT-OF-TOL · MRB
Divertor target alignmentLaser tracker100.00 mm±0.3 mm100.42 mm REP+0.42 mmThermal ↗OUT-OF-TOL · MRB
Vessel roundness (datum)Photogrammetry0.00 mm±3 mm2.10 mm REP+2.10 mmStructural ↗IN-TOL
CS concentricity (datum)Laser tracker0.00 mm±0.4 mm0.31 mm REP+0.31 mmElectromagnetic ↗IN-TOL
Vessel weld penetrationNDEStructural ↗PENDING
Deviation → analysis impact (loop closure)

Every out-of-tolerance feature is propagated to the analysis field it perturbs and the requirement it puts at risk — the as-built geometry re-enters the model rather than sitting in a report.

TF coil radial positionΔ +0.70 mm (tol ±0.5 mm)Electromagnetic ↗REQ-B03 ↗
+0.2% field ripple → local hoop stress +~3%; still inside the 20.1 T / 316LN margin, but dispositioned.
Blanket module fit-upΔ +1.20 mm (tol ±1 mm)Neutronics ↗REQ-B02 ↗
Wider inter-module gap → neutron streaming; TBR debit — must re-run MCNP with the as-built gap.
Divertor target alignmentΔ +0.42 mm (tol ±0.3 mm)Thermal ↗REQ-B07 ↗
Strike-point shift → peak target heat flux +~8%; erodes the already-tight 90%-radiated margin.
As-built values are representative (synthetic, at typical fab tolerances) — the workflow, not a measurement record. Out-of-tolerance items route to Material Review Board (MRB) and re-enter the multiphysics analysis. RECEIVING STRUCTURE · NO ARTICLE MEASURED
🔒 Metrology · As-Built vs As-Designed is Kronos-only. The overview above is public — the full data, tables, live calculations and engineering detail on this page require a Kronos team sign-in.
Context

Benchmarks vs. known devices

Where the Hyperion breeder sits against reference tokamaks. This is honest positioning, not a superiority claim: Hyperion is a compact spherical tokamak optimised to breed tritium, neutrons and helium-3 — not to maximise Q. The burner is a D–³He tandem mirror, a different confinement class, compared separately by net electric output.

Major radius R₀ (m)
ITER
6.2
Hyperion
2.4
SPARC
1.85
NSTX-U
0.94
Headline field (T)
on-axis for ITER/SPARC/NSTX-U · peak-at-coil for Hyperion
Hyperion
16.84
SPARC
12.2
ITER
5.3
NSTX-U
1
Plasma current Iₚ (MA)
ITER
15
Hyperion
9.66
SPARC
8.7
NSTX-U
2
Fusion gain Q
SPARC
11
ITER
10
Hyperion
3.42
NSTX-U
0
Fusion power (MW)
ITER
500
SPARC
140
Hyperion
85.0
NSTX-U
0
The burner is a different class
The D–³He tandem-mirror burner (Aegis · MetroVolt) is not a tokamak and is not measured by the metrics above. Its figure of merit is net electric output: engineering gain Q_E ≈ 1.31 at the closing design point, delivering net +104 / +850 / +2832 MWe across central-cell lengths of 55 / 440 / 1400 m — subject to the plug-density requirement documented on the Physics and Methodology pages.
Reference values are the devices' published headline figures. Field row: on-axis for ITER / SPARC / NSTX-U; peak-at-conductor for Hyperion's HTS coils — the two bases are not directly comparable. Hyperion figures are the frozen computed design point (Q 3.076, P_fus 85.0 MW, I_p 9.66 MA).
Check us — the five reproducibility tests

Reproducibility

Every headline claim rests on a pre-registered test that regenerates from the open deposit. These are the five that matter — shown with their verified verdicts. You don't have to trust us; you can run them.

Verdicts below reproduce from the deposit (Tier-1 byte-identical where marked). The heavy re-runs execute on the record machine via the published runbook and the kronos-py / kronos-ml / kronos-quantum environments — not in this browser. Integrity (Test 1) was last re-verified here: 182 / 182 OK.

Tests
5
pre-registered, reproduce from deposit
Integrity (live check)
182/182
md5 re-verified here
Byte-reproducible (Tier-1)
3 / 5
integrity · calibration · quantum
Honesty
no crossover
quantum stated as not-yet
The five tests
#TestWhat it checksVerified resultTierRegenerate from
1Artifact integrity182 recorded artifacts vs their md5182 / 182 OKTier-1 · byterepro_driver.py --verify (kronos-py)
2RL safety clampEnvelope escapes over 150,000 control steps0 escapes / 150kTier-2 · pre-registeredclamp_activation_stats.csv (kronos-ml)
3Calibration spineModel-to-data improvement factors1691× / 36.8× / 7475×Tier-1 · bytecalibration_loop.csv (kronos-ml + MC engine)
4Quantum advantageFault-tolerant resource crossover this decadeno crossover (classical wins)Tier-1 · bytecrossover_verdict.csv (kronos-quantum)
5CBF safety certificateControl-barrier faces + calibration-through-proof3 / 4 faces · 10 / 10Tier-2 · pre-registeredtrack_R3_1 formal-safety (kronos-ml)
Reproduce it yourself
1. Download the deposit (DOI on publication) and unzip.   2. Build the three environments — kronos-py, kronos-ml, kronos-quantum (conda-forge; versions pinned in the runbook).   3. python track12_repro/repro_driver.py --verify → expect 182 OK.   4. Regenerate track-by-track (repro_driver.py prints the plan) and re-verify. Tier-1 must return byte-identical; Tier-2 (trained-model metrics) reproduces within the pre-registered 5-seed tolerance — a few Tier-2 files reporting CHANGED after retraining is expected, not a failure.
The point is auditability, not authority: "certified" is replaced by "reproducible" throughout. Heavy runs need the record hardware (56-core CPU); the browser shows verdicts + provenance only.
Evidence · Breeder

Evidence, Validation & IP

Every headline reproduces from the open deposit. Validation is public; the science is falsifiable. IP posture is open results with protected implementations.

Reproducibility & readiness
Reproducibility
0.0e+00
bitwise rel diff
Validation
23
checks · V01–V23
Open deposit
CC BY 4.0
DOI …21746157
Patent
US 12,009,112
granted
Validation suite · click a check for detail
IDCheckResultVerdict
V01Frozen design point reproducesworst rel 0.0e+00 (bitwise, NumPy 2.x)PASS
V02Gain is SOLVED not scaledQ 3.0763 from power balancePASS
V03Fuel follows purpose (min I_p)D–T 4.93 MA vs D–D 15.52 MAPASS
V04Closing-config countsD–T no-wall 213/10800PASS
V05–V23 (19 more)field/blanket/current-drive checksPASS
Intellectual property · open science, protected implementations
Open science
All physics, code and data are deposited under CC BY 4.0 — anyone can reproduce the design point.
2 Zenodo DOIs
Granted patent
US 12,009,112 — issued. Foundational method protection.
GRANTED
Protected implementations
Filed — Omnibus, U.S. Application No. 64/128,097 (Patent Pending): magnets, DEC, fuel cycle, control. Open results, protected build.
open but protected
Breeder deposit · DOI 10.5281/zenodo.21746157 · reproduced from the open Zenodo/GitHub deposit (CC BY 4.0). Canonical freeze: FROZEN_2026-08_PUBLICATION.csv. · IP: open science, protected implementations (internal prior-art/FTO analysis not shown).
KRONOS FUSION ENERGY · Kronos Model Conceptual design and simulation study; no machine has been built