EVIDENCE & METHODOLOGY / 8 SEPTEMBER 2026

A useful starting point.
With visible limits.

The Virtual Test Bed helps you identify a modeling workflow, inspect evidence and compare eligible tools. All reactor geometry is conceptual. No simulation, benchmark or VTB coupling has been executed for this catalog.

What the catalog counts

The snapshot has 167 canonical records: 38 fusion-specific, 58 fusion-enabling, 35 supporting, 28 fission-adjacent, 4 historical and 4 unresolved. These include solvers, frameworks, interfaces, libraries, geometry tools and reference collections. They are not 167 independent fusion solvers.

All 141 named current source rows and 39 archive rows are accounted for. Aliases are merged while distinct software distributions retain separate records. Internal workbook assignees are not upstream maintainers.

How much evidence was reviewed?

75 primary profiles
Fuller capability, access and evidence descriptions.
33 primary descriptions
Shorter source-backed descriptions; deeper review remains incomplete.
50 metadata reviews
Repository identity, description and activity do not establish benchmark validity, build support or model fit.
5 pending records and 4 unresolved identities
Excluded from recommendations. Missing evidence is a review task, not a negative scientific score.

Readiness is provisional editorial judgment

Method morton-discovery-readiness-0.1 rates 18 records in six comparable workflow cohorts. There is no community voting, maintainer endorsement or scientific accuracy score. Access friction does not judge scientific importance.

Score = (25D + 25R + 20A + 15I + 15S) / 4, where each dimension is 0–4. Null means unreviewed; it never becomes zero. Scores are recalculated after eligibility filters. Ties are labeled, with a stable name/ID ordering for display.

Documentation · 25% weight · ordinal anchors
  1. 0: Absent usable guidance established
  2. 1: Minimal overview
  3. 2: Partial guide / limited examples
  4. 3: Substantial guidance and examples
  5. 4: Broad model, input, installation and tutorial coverage
Reference evidence · 25% weight · ordinal anchors
  1. 0: None established for reviewed scope
  2. 1: Descriptive demonstration
  3. 2: Published use / limited comparisons
  4. 3: Documented benchmark or comparison cases
  5. 4: Structured verification and validation resources with reproducibility context
Access · 20% weight · ordinal anchors
  1. 0: Unavailable to target user established
  2. 1: Approval, license or eligibility friction
  3. 2: Public route with unresolved conditions
  4. 3: Public distribution with reuse/setup conditions
  5. 4: Public source/distribution with a clear access route
Integration · 15% weight · ordinal anchors
  1. 0: Unusable interface established
  2. 1: Bespoke or opaque setup
  3. 2: Specialist/manual adaptation
  4. 3: Documented programmable or coupled interfaces
  5. 4: Packaged, documented reusable integration for the role
Stewardship · 15% weight · ordinal anchors
  1. 0: Retired with no usable steward established
  2. 1: Historical owner only
  3. 2: Identifiable project, limited current evidence
  4. 3: Current activity/release and support route
  5. 4: Reviewed release/contribution/support process and continuing activity

Eligibility comes before a score

  1. Select a reactor family, component and research question.
  2. Apply the workflow’s reviewed model and geometry constraints. Broad physics tags alone do not establish compatibility.
  3. Exclude unresolved, pending, historical, archived and inapplicable candidates.
  4. Apply access and hardware requirements. Unknown eligibility stays in a separate review list.
  5. Rank only candidate roles with a complete rating in the selected comparison cohort.
  6. Show up to three. Fewer than three, or none, is a valid result.

Companion libraries and framework stages never fill a missing solver slot. A commercial model-building environment does not pass an open-source-only filter. A local model does not become a global model by changing a label. The ready-made application filter removes general toolkits and environments; the remaining tools still require scientific setup.

New implementation checks refreshed selected upstream access and hardware references on 8 September 2026. The original CSV is preserved. A documented CPU-build supplement is shown on the relevant tool’s detail page; it is not a claim of a tested build on this computer.

Time and coupling are explicit

The fourth dimension is a selected operating phase or load stage. A phase changes relevant questions, not numerical results. There is no common time axis spanning a target pulse, cooldown and years of exposure.

A solid coupling arrow means documented upstream evidence; a dashed arrow means a proposed interface. Each connection exposes its exact evidence type: mechanism, example, framework composition or program integration. Every Morton VTB package remains not built or run.

A future package needs a versioned geometry/source manifest, units and coordinates, material and nuclear-data versions, mesh association, conservative spatial transfer, physical source normalization, temporal interpolation, uncertainty and domain-reviewed acceptance criteria. Display geometry cannot be used as solver-ready input without independent conversion and validation.

Find every research workflow

  1. W01 · Neutron loading and shielding — Where is neutron and photon energy deposited, and what reaches protected components?
  2. W02 · Coolant-channel heat removal — Can this channel remove an imposed heat load within thermal and hydraulic limits?
  3. W03 · Isotope retention and permeation — How much hydrogen isotope remains in a material or crosses a barrier?
  4. W04 · Divertor heat-load chain — How does an edge-plasma scenario load a plasma-facing component?
  5. W05 · Local turbulent transport — What local turbulent fluxes follow from this equilibrium and gradient set?
  6. W06 · Stellarator equilibrium and coils — Does the magnetic configuration meet equilibrium and coil-design objectives?
  7. W07 · Heating and current-drive deposition — Where does launched wave power couple to the plasma?
  8. W08 · Tokamak profile evolution — How do plasma profiles change under heating and current-drive controls?
  9. W09 · Disruption and runaway evolution — How does a disruption scenario alter runaway-electron and plasma evolution?
  10. W10 · ICF radiation-hydrodynamics — Which modeling family represents laser energy deposition and evolving target plasma?
  11. W11 · ICF chamber pulse response — How does a declared chamber load translate into a thermal response and recovery requirement?
  12. W12 · Coil thermal and mechanical response — How do declared electromagnetic and radiation loads affect a coil assembly?
  13. W13 · Shutdown inventory and maintenance — How do irradiation and cooling history change inventories and shutdown sources?
  14. W14 · Irradiated material state — How do defect populations and material response evolve under an exposure history?
  15. W15 · Vacuum pumping and neutral gas — What pressure and gas propagation follow from a molecular-flow pumping arrangement?
  16. W16 · Whole-plant design trade study — How do declared engineering and plasma assumptions alter a conceptual plant trade-off?
  17. W17 · Heat-to-electric plant dynamics — How might thermal storage, heat exchange and controls respond to a changing load?
  18. W18 · Geometry and mesh preparation — Can one approved component description support the required solver representations?
  19. W19 · Uncertainty and reproducibility — Which assumptions dominate the answer, and can another team reproduce it?

Source dates, attribution and corrections

Repository activity, release tags, documentation dates and research check dates are separate facts. A blocked request or an old release tag does not establish abandonment. Exact license terms must be checked for the selected version and intended use.

Public code and institutional names are factual, source-backed attribution attached to canonical records. They do not imply a partnership, endorsement, funding award or availability through Morton.

To correct an entry, use the existing contact page and include its stable code ID, the disputed field and a public primary source. Corrections need editorial review and a dated change record.