Research Scope and Decisions

Programme scope

OWNER INTENT: The broader RAMAL-EBX programme may cover direct electrons and converted X-rays over approximately 0.1-10 MeV, multiple source-head and shielding families, particle-resolved three-dimensional fields, and later product-dose applications. This is a programme horizon, not the claim of Paper 1.

Paper 1 scope

APPROVED DECISION — D1: “Paper 1 covers conventional industrial LINAC converted-X-ray operation at 5-7.5 MeV, one bounded converter/source family, and one family of closed ordinary-concrete rooms; the surrogate predicts a predefined three-dimensional radiation-protection field for complete held-out configurations.”

APPROVED DECISION — D2: “The primary field is ambient dose equivalent H*(10), PHITS multiplier -200, normalized as pSv per primary source electron; combined, photon, electron, positron, and neutron pages remain identifiable and operational dose rate is derived separately from declared operating factors.”

APPROVED DECISION — D4: “Use a closed rectangular ordinary-concrete room with width 4-10 m, length 6-16 m, height 3-6 m, primary-floor concrete 100-180 cm, shared secondary barriers 50-150 cm, a centred source, +Y upward, and the beam vertically downward; openings and streaming paths are excluded.” These bounds define the study family, not safe construction dimensions.

APPROVED DECISION — D6: “Use JENDL-5 conditionally as the photonuclear reference, compare it with ENDF/B-VIII.1 and PHITS built-in modelling, and retain neutron reporting in the combined D2 target unless later approved evidence supports another rule.” Approval does not establish that the comparisons have been completed.

Direct-EB prediction, doors, mazes, ducts, penetrations, activation, post-shutdown dose, arbitrary source heads, product dose, and facility-specific compliance lie outside the approved claim.

Amendments

APPROVED DECISION — D3, amendment-open: “Use a broad planar tantalum-water-Type-304-stainless scanned-plate archetype in the Paper 1 coordinate frame (+Y up, beam toward -Y, scan along X, remaining transverse direction along Z). Re-select the dimensions and layer construction from a broad primary-source review and predefined PHITS comparisons; finite-edge behaviour, source-field verification, and photonuclear evidence remain unresolved.”

APPROVED DECISION — D5, amendment-open: “Use a declared normalized one-dimensional scan with a local Gaussian spot. Re-select the spot definition, scan weighting, quadrature resolution, finite-plate interception method, and transport-response measures through the approved D3/D5 procedure.”

APPROVED DECISION — D3/D5 model boundary: Paper 1 will select a dimensioned, finite Ta–water–Type-304-stainless research archetype for PHITS source characterization and shielding calculations. It will not claim to reconstruct a specific commercial accelerator or establish manufacturing readiness, thermal-hydraulic safety, commissioning suitability, or regulatory approval.

APPROVED DECISION — D3 comparison-family boundary: Numerical candidate selection remains inside the Ta–water–Type-304-stainless family. Include the centre-aligned Ta-only source-head limiting case defined below; it is not a candidate. Other researched construction families remain literature context unless evidence from the predefined comparisons requires an explicit amendment.

APPROVED DECISION — D3 candidate-selection rule: Apply predefined exclusion criteria and compare each surviving candidate across separately reported responses. Retain the representative centre unless one eligible candidate dominates it and every other eligible replacement under the directed responses defined below. Incomparable replacements require an amendment. Do not collapse photon output, electron contamination, edge behaviour, neutron contribution, numerical stability, and physical or transport plausibility into an invented weighted score.

APPROVED DECISION — D3 plausibility boundary: Require a physical and transport plausibility screen covering material identity and ordering, source applicability, particle and energy accounting, transmitted-electron behaviour, finite-edge response, and deposited energy per primary electron. Cooling geometry may be declared as model geometry, but candidates are not accepted or rejected as thermally safe. Paper 1 does not establish CFD, flow, pressure, boiling, fatigue, corrosion, fabrication, constructibility, or operational safety qualification.

APPROVED DECISION — D3 limiting case: Use one finite Ta-only limiting case with the same Ta layer, plate extent, source, scan, and scoring definitions as the corresponding full Ta–water–Type-304-stainless candidate. It measures the combined transport influence of water, stainless steel, and their interfaces; it does not separate those influences individually.

APPROVED DECISION — D3 comparison boundary: Select D3 using component-model responses on predefined planes and regions around the finite source head. Do not run every D3 candidate through the complete D4 room. Confirm the selected construction later in the ordered D4 workflow.

APPROVED DECISION — D3 engineering normalization: Compare deposited energy per primary electron and expose explicit linear scaling to separately declared operating conditions. Do not select a universal beam power, current, duty factor, or scan utilization for the source-head archetype.

APPROVED DECISION — D3 energy boundary: Use one fixed Ta–water–Type-304-stainless geometry across the full 5-7.5 MeV Paper 1 interval. Electron energy remains a source condition; source-head dimensions do not vary with energy and do not become surrogate inputs.

APPROVED DECISION — D5 scan weighting: Use a uniform normalized continuous one-dimensional scan density, represented numerically by a converged quadrature. Do not introduce shaped or nonuniform weighting without a later amendment backed by a measured waveform or an explicit new research question.

APPROVED DECISION — D3 finite extent: Derive plate extent from the declared scan interval plus a Gaussian-tail margin. Refine that margin until predefined edge-interception and escape responses stabilize; do not copy one published accelerator's plate size or make plate dimensions surrogate inputs.

APPROVED DECISION — D5 spot boundary: Compare a small evidence-informed range of local Gaussian spot definitions and widths during preliminary component calculations, then approve one fixed definition and width for Paper 1. Spot size does not become a surrogate input, and the selected value is not claimed to be universal.

APPROVED DECISION — D3 construction-evidence boundary: Proceed without first-party production drawings or facility-specific engineering constraints. Define a bounded generic transport-sensitivity candidate set from authoritative material data and applicable published converter studies. Water and Type-304-stainless dimensions are transport parameters, not claims about a buildable cooling channel or production source head.

APPROVED DECISION — D3 screening design: Use a staged two-level 2^3 Ta/water/SS304 transport screen with all eight vertices and one all-factor centre, plus the approved centre-Ta-only limiting case. Evaluate the geometry screen at 5, 6.25, and 7.5 MeV; stage the three evidence-informed spot cases and the three scan-span cases separately; and add axial or interior geometry points only under the approved failed-centre rule. Reuse an immutable centre Result whenever every input and numerical setting is identical across screens rather than executing nominal duplicates.

APPROVED DECISION — D5 scan-span screen: Compare 60, 80, and 100 cm uniform scan spans as a separate sensitivity stage, then approve one fixed Paper 1 span. The endpoints are source-specific industrial examples and 80 cm is a declared midpoint; none is a universal industrial standard. Do not multiply the full Ta/water/SS304 geometry screen by every scan span.

APPROVED DECISION — D3 source-energy model: Use reproducible monoenergetic 5, 6.25, and 7.5 MeV electron sources for D3 selection. This is an idealized generic source, not a claim about a commercial accelerator spectrum. Do not invent an energy-spread distribution; add one only through a later amendment supported by measured or otherwise authoritative source data.

APPROVED DECISION — D3 dimensionless geometry levels: Define layer areal density as tau = rho t and use 7.5 MeV as the fixed reference energy. Screen Ta at 0.25, 0.50, and 0.75 tantalum radiation lengths; water at 0.05, 0.125, and 0.20 of its NIST ESTAR CSDA range; and SS304 at 0.50, 0.875, and 1.25 of its custom-composition ESTAR CSDA range. Use the low and high levels for the 2^3 vertices and the middle levels for the all-factor centre. These are transport-screening coordinates, not stopping guarantees, cooling dimensions, industrial optima, or the final construction.

PRELIMINARY CALCULATION EVIDENCE: The recorded reference scales produce approximate Ta thicknesses 1.024, 2.048, and 3.071 mm; water thicknesses 1.894, 4.734, and 7.575 mm; and SS304 thicknesses 2.978, 5.211, and 7.445 mm. D3_D5_SOURCE_HEAD_REVIEW.md owns the source composition and derivation. Raw ESTAR outputs and full-precision operands remain a provenance gate before Case generation; immutable Cases must store full-precision centimetre values rather than reconstructing them from these rounded labels.

APPROVED DECISION — D5 spot cases: Use circular Gaussian FWHM values 2.36, 5.89, and 11.78 mm for the staged spot-sensitivity calculations, converted from Esplen et al.'s reported 2 sigma = 2, 5, and 10 mm cases. These indirect 8-10 MeV Ta–Al comparator values are sensitivity coordinates, not industrial bounds. Approve one fixed Paper 1 FWHM only after reviewing the component responses.

APPROVED DECISION — D3/D5 numerical-calibration sequence: Before the formal candidate comparison, use preliminary calculations to measure PHITS statistical noise, scan-quadrature refinement, plate-margin refinement, tally-mesh refinement, and response curvature. Approve the comparison tolerances, histories, seed procedure, and stopping rules from that evidence before running or admitting the formal comparison. Do not reuse Study-editor defaults or retrospectively count calibration calculations as accepted candidate comparisons.

APPROVED DECISION — D3 representative construction: Predeclare the all-factor centre as the default final construction. Replace it only if preliminary comparison demonstrates a physical or transport failure, or another candidate strictly dominates it under predefined response directions. The surrounding vertices characterize sensitivity and interactions; they are not opportunities for post-hoc optimization.

APPROVED DECISION — D5 representative spot: Predeclare 5.89 mm FWHM as the representative default. Retain it when the three spot cases show smooth quantitative variation. If spot size produces a qualitative transport change, stop and amend the scope rather than selecting the most convenient result.

APPROVED DECISION — D5 representative scan span: Predeclare 80 cm as the representative default and use 60 and 100 cm as sensitivity bounds. If those bounds expose qualitatively different source fields, stop and reconsider whether scan span must enter the Paper 1 domain. Do not select span to minimize exterior dose or optimize local uniformity.

APPROVED DECISION — D3 strict-dominance directions: A candidate strictly dominates the representative centre only when it is no worse on every directed response and better on at least one: maximize forward photon current at P1; minimize transmitted electron and positron current, unintended side/edge particle escape, and complete-assembly deposited energy under equal incident normalization. Report neutron contribution without minimizing it, because minimizing a source term could conceal shielding-relevant physics. Keep spectra, angular distributions, and spatial profiles separate and undirected.

APPROVED DECISION — D3/D5 qualitative-change rule: Treat a sensitivity change as qualitative when it changes candidate domination or exclusion; introduces or removes a hotspot or edge maximum; reverses an energy or spatial trend; changes the dominant particle contribution; creates a previously absent transmitted-primary or finite-edge escape path; or prevents one common quadrature, mesh, or statistical procedure from covering the compared cases. Later approved tolerances govern ordinary numerical differences.

APPROVED DECISION — D3 transverse extent: Reserve global Z as the conveyor-compatible transverse direction, but include no conveyor or product motion in Paper 1. Derive the converter's finite Z extent from the stationary Gaussian spot plus the same response-converged tail margin used beyond the X scan endpoints. Converter width is not conveyor width and is not an independent surrogate or screening factor.

APPROVED DECISION — D3 plate-margin refinement: With selected Gaussian standard deviation sigma, define X_plate = L_scan + 2 k sigma and Z_plate = 2 k sigma. Evaluate nested margins k = 3, 4, 5, and 6, and stop only under the later approved edge-interception and escape-response tolerance. Gaussian tail probability is a diagnostic, not the convergence criterion.

APPROVED DECISION — D5 scan-execution audit: During calibration, verify at least one composite PHITS multi-source result against the retained separate-position weighted sum. After equivalence passes the approved tolerance, use recorded multi-source relative weights and totfact for formal candidate calculations. Failure to establish equivalence keeps separate-position execution mandatory.

APPROVED DECISION — D3 component quantity boundary: Omit H*(10) from D3 candidate selection. Use particle-resolved surface current, fluence, spectrum, angle, charged-particle contamination, finite-edge escape, and per-material deposition to characterize the source head. Reserve the approved D2 H*(10) target for D4 complete-room calculations.

APPROVED DECISION — D3 photonuclear screening: Use PHITS built-in photonuclear treatment for preliminary D3 screening, report neutron outputs as model-dependent and undirected, and do not disable neutron production. Run the approved JENDL-5, ENDF/B-VIII.1, and PHITS built-in D6 comparison only on the selected construction before D4 calculations.

APPROVED DECISION — D3 reference materials: Use elemental natural tantalum at 16.654 g/cm3, NIST liquid-water composition at 1.000 g/cm3, and the declared nominal Type 304 composition used for the custom ESTAR calculation at 8.000 g/cm3. Retain every elemental mass fraction and its source in the Project record. These are reproducible reference materials, not claims about a manufactured batch.

APPROVED DECISION — D3 component environment: Place the finite source-head assembly in vacuum for component screening, with only explicitly defined tally regions. D4 supplies room air and concrete later; neither an arbitrary air path nor partial shielding belongs to D3.

APPROVED DECISION — D3 tally scope: Omit the optional local stand-off plane P2. Use P0, separate material regions, P1, and finite side/edge surfaces. Add a stand-off plane only through a later amendment tied to a specific verification question and predefined distance.

APPROVED DECISION — D3/D5 record separation: Retain numerical-calibration calculations as preliminary evidence. After their evidence supports and the owner approves tolerances, histories, seed procedure, and stopping rules, create fresh immutable Cases for the formal candidate comparison. Do not reclassify calibration Results as formal comparisons or delete negative and inconclusive calibration evidence.

APPROVED DECISION — D3 dimension authority: Store each dimensionless factor, reference mass scale, density, and derived thickness in centimetres at the full available calculation precision in every immutable Case. Rounded millimetre values are display labels only. Do not require external lookup to reconstruct a Case.

APPROVED DECISION — D3/D5 statistical estimator: Default numerical calibration and comparison to PHITS history-variance mode istdev = 2. Permit batch variance istdev = 1 only when measured memory or runtime makes history variance impractical; use more than ten batches and record the reason, settings, and estimator change.

APPROVED DECISION — D3 formal eligibility: Before comparing responses, require each formal Case to have immutable input and complete provenance; the intended materials, geometry, source, transforms, and particles; successful process and transport completion; every required P0, per-material, P1, and edge tally; no unresolved PHITS diagnostic affecting interpretation; and satisfaction of the approved statistical, mesh, quadrature, margin, energy-accounting, and normalization-equivalence criteria. Retain failures and classify unexplained Results as inconclusive rather than dominated.

APPROVED DECISION — D3 energy accounting: Record incident energy, deposited energy by region, and energy carried across every declared outer boundary. Use calibration calculations to characterize the residual and approve a tolerance before formal comparison. Do not require exact closure and do not omit escaping or cutoff-related contributions from interpretation.

APPROVED DECISION — D3 layer geometry: Model three coextensive finite rectangular layers in beam order Ta -> water -> SS304, surrounded by vacuum. Include no frame, seals, channels, windows, supports, inlet, or outlet. Water is an idealized transport region representing coolant material, not an operating cooling system or a recommended channel.

APPROVED DECISION — D3 limiting-case alignment: In the Ta-only limiting case, keep P0, P1, and every declared outer boundary at the full centre-candidate coordinates and replace the removed water and SS304 volumes with vacuum. Do not move P1 behind Ta or create a second comparison plane.

APPROVED DECISION — D5 spot-screen geometry: Use one common finite plate extent derived from the largest tested Gaussian spot and the converged tail-margin procedure for all three spot cases. After retaining the representative spot, reconverge the final plate extent around that fixed spot.

APPROVED DECISION — D3/D5 preliminary-stage order: Proceed in dependency order: verify source transforms, fixed tallies, material identity, and energy accounting on the centre geometry; calibrate histories, estimator, and tally mesh; refine scan quadrature and verify multi-source normalization; run the common-geometry spot screen; retain or reconsider the representative spot; run the scan-span screen; retain or reconsider the representative span; refine the final plate margin; approve numerical tolerances and stopping rules; then create fresh formal Cases for the geometry screen, centre, and Ta-only limit. Do not run unresolved stages as a multiplied grid.

APPROVED DECISION — D3 failed-centre refinement: If the centre construction fails an approved physical, transport, or numerical criterion, add at most one interior or axial refinement toward the valid side for each implicated factor before selecting a replacement. Do not jump directly to an extreme vertex or expand automatically to a full three-level factorial.

APPROVED DECISION — D3 dominance tie: Replace the centre automatically only when one eligible candidate dominates the centre and every other eligible replacement under the predefined directed responses. If multiple incomparable replacements remain, stop and amend the selection rule before inspecting detailed response maps; do not invent a weighted score or use total stack thickness as a post-hoc tie-breaker.

APPROVED DECISION — D3 energy-accounting boundary: Use one closed six-face accounting boundary for every full candidate and Ta-only limiting case. P0 and P1 are the upstream and downstream faces; four side surfaces enclose the common finite X-Z extent. Record outward and inward particle-energy crossings separately so re-entry is not counted as new source energy.

APPROVED DECISION — D3/D5 application output boundary: Implement a reusable Case-scoped Result Artifact Contract rather than adapting the fixed legacy source-head reducer. Preserve declared T-Cross, diagnostic T-Track, and T-Deposit cards in immutable Case input; require and hash their declared artifacts; expose generic report and table inspection in Analyze. Raw PHITS files outside the GUI are not the supported research workflow.

APPROVED DECISION — D3 vacuum authoring: Add a reusable GUI world-medium choice that supports vacuum while retaining explicit material regions. Do not depend on imported hand-written inputs or substitute synthetic world air for the approved component environment.

APPROVED DECISION — D3/D5 comparison ownership: Do not expand Study/Assessment into a speculative component-experiment framework before the first workflow. Use immutable contracted Results as component evidence and record owner-approved D3/D5 conclusions in this research document. Add a reusable comparison operation only after the exercised workflow demonstrates the repeated operation and required data shape.

APPROVED DECISION — D3 reference materials in the application: Add the exact water and nominal Type-304 definitions selected for this work as reusable, source-cited application material references rather than adding a free-form custom-material provenance field. Keep them scientifically named and generally reusable; Paper 1 selects those references as Project data rather than hardcoding their use in application control flow. Editing a referenced definition continues to make it a custom unreferenced material.

APPROVED DECISION — Result artifact identity: Give each authored tally explicit stable role metadata separate from its human-readable title and output filename. Case publication and Result verification must reject missing or duplicate required roles rather than infer semantics from mutable text.

APPROVED DECISION — preliminary Case generation: Use duplicated Designs and existing GUI Case controls for the first preliminary workflow. Add another reusable control or batch operation only after a specific repeated GUI edit demonstrates the need. Do not pre-build a factorial framework or use repository scripts as the user workflow.

APPROVED DECISION — D3/D5 random seeds: Assign independent recorded seeds to each Case or replica. Use the same predefined seed schedule across candidate groups for procedural consistency, but do not treat cross-candidate differences as paired correlated estimates. Add replicas only when calibration or a close selection boundary requires them.

APPROVED DECISION — D3/D5 numerical-threshold approval: Use preliminary calibration Cases to propose response-specific uncertainty, mesh, quadrature, normalization-equivalence, and energy-residual thresholds before inspecting candidate response maps. Obtain owner approval and freeze those thresholds before creating formal Cases. Do not substitute one universal percentage, PHITS completion alone, current Study defaults, or post-hoc thresholds chosen after seeing the preferred candidate.

APPROVED DECISION — D5 spot retention: Retain the provisional 5.89 mm FWHM only when the 2.36, 5.89, and 11.78 mm sensitivity cases produce no predefined qualitative change. If they do, stop and either retain spot width as an explicit study variable or obtain external source evidence. Never select the width that makes the converter response most favorable or average the three responses.

APPROVED DECISION — D5 span retention: Retain the provisional 80 cm span only when the 60, 80, and 100 cm sensitivity cases produce no predefined qualitative change. If they do, stop and either retain span as an explicit variable or obtain application evidence. Never select the most favorable span or average the three responses.

APPROVED DECISION — D3/D5 formal-run gate: After calibration and sensitivity calculations, present the immutable Results, proposed frozen thresholds, retained source settings, and exclusions for owner approval before creating fresh formal Cases. Do not create them automatically or reclassify preliminary Cases.

APPROVED DECISION — D3/D5 sensitivity-screen energies: Run the three spot-width sensitivity Cases only at 7.5 MeV, the declared energy nearest their indirect comparator evidence. Run the 60, 80, and 100 cm scan-span Cases at 5, 6.25, and 7.5 MeV. The geometry and Ta-only set contributes an expected 30 Cases; reuse its representative 5.89 mm, 80 cm centre Results in both sensitivity screens only when the complete inputs and numerical settings are identical; add two unique spot Cases and six unique span Cases. The expected unique baseline is 38 preliminary Cases before margin refinement, calibration replicas, or a permitted failed-centre refinement. This is a structural estimate, not a target count.

APPROVED DECISION — D3/D5 identical-Case reuse: Reference an existing immutable preliminary Result in every applicable comparison when its complete input, numerical settings, and seed are identical. Do not rerun nominal duplicates to preserve a target Case count. Any deliberate repeat is an explicitly labelled replica with an independent recorded seed.

APPROVED DECISION — D3 plate-margin energies: Evaluate nested k = 3,4,5,6 margins on the representative centre construction at all three source energies. Retain the smallest single common k satisfying every frozen margin criterion; do not give different energies different plate dimensions.

APPROVED DECISION — D3 surface-current convention: Record inward and outward components separately at all six accounting faces. Derive P1 forward photon current and unwanted escape currents from explicit direction filters, with every value normalized per incident primary. Do not use unsigned crossings or infer direction later from filenames.

APPROVED DECISION — D3 Ta-only role: Treat the centre-Ta-only model solely as a limiting-case comparator. It cannot replace the centre or participate in candidate dominance. It remains in the formal calculation set but is not a tenth construction candidate.

APPROVED DECISION — D3 per-material artifact rule: Require deposition artifacts for every non-void material region present in the immutable Case. Full constructions require separate Ta, water, and SS304 artifacts; Ta-only requires Ta alone. Vacuum has no deposition artifact.

APPROVED DECISION — D3 formal Case set: If no amendment changes the ten constructions or three energies, create 30 fresh formal Cases after every gate passes: the eight geometry vertices, centre, and centre-Ta-only limit at each of 5, 6.25, and 7.5 MeV. Freeze spot width, scan span, plate margin, quadrature, histories, tally mesh, and approved thresholds identically wherever applicable. Preliminary baseline Cases cannot enter this formal set; the count follows the approved structure rather than acting as a target.

APPROVED DECISION — first D3 executable calibration: After repairing the approved reusable GUI blockers, create one short-history preliminary centre Case at 7.5 MeV. Verify immutable input, vacuum world, material references, source and tally transforms, every declared artifact role, directional crossings, deposition, PHITS diagnostics, and Analyze inspection before calibrating histories and tally meshes across all three energies or multiplying Cases.

APPROVED DECISION — Paper 1 Project identity: Recreate the Project with ID industrial-eb-x-ray-dose-prediction and display name Industrial EB/X-ray Dose Prediction.

APPROVED DECISION — D3 incident energy: Define incident energy analytically from the normalized monoenergetic source as E_incident = E0 per primary. Because the particles originate on the Ta entrance plane, P0 is not an incident monitor: outward P0 crossings measure upstream backscatter escape, and inward P0 crossings are re-entry diagnostics.

APPROVED DECISION — D3/D5 implementation gate: Before preliminary calculations, implement and verify only the reusable blockers established by the capability audit: referenced materials, GUI vacuum-world selection, preservation of authored tally sections, stable artifact roles, Attempt/Result enforcement, and generic Analyze inspection. Do not add Study expansion or batch generation and do not substitute imports or raw-output workflows.

APPROVED DECISION — completed cleanup: The owner rejected the former D3 Design and its 0.8 mm and 1.2 mm application-verification executions as a basis for the redesigned source and approved their deletion from Paper 1. Their dimensions and observations must not seed candidate selection.

Editable decision table

Edit decisions and reasoning directly on this page. AI updates must preserve explicit decisions and amendments, reconcile implementation and evidence, and report unresolved contradictions. Git history retains previous wording. The owner approved the D3/D5 procedure and bounded screening coordinates through the recorded D3/D5 grilling; D3 and D5 remain amendment-open until preliminary evidence supports approval of the final source definition and construction.

ID	Current status	Approved wording or controlled question
D1	APPROVED DECISION	Bounded converted-X-ray Paper 1 domain and complete held-out configurations.
D2	APPROVED DECISION	H*(10), multiplier -200, identifiable particle pages, intrinsic normalization, and separate operational scaling.
D3	APPROVED DECISION — amendment-open	Broad Ta-water-SS304 scanned plate in the vertical Paper 1 frame; re-select dimensions and construction from primary sources and predefined comparisons; edges, verification, and photonuclear evidence remain open.
D4	APPROVED DECISION	Closed rectangular ordinary-concrete room, declared ranges, centred source, +Y up, beam toward -Y, no openings.
D5	APPROVED DECISION — amendment-open	Normalized one-dimensional scan with local Gaussian spot; re-select spot, weighting, quadrature, interception, and response measures through the D3/D5 procedure.
D6	APPROVED DECISION	For the selected D3 construction, compare conditional JENDL-5, ENDF/B-VIII.1, and built-in modelling before D4 calculations; retain neutron reporting unless later evidence changes the rule.
D7	UNRESOLVED	Grid, RSE, reportable coverage, repeated-run, convergence, completion, and reporting criteria require owner approval.
D8	UNRESOLVED	Verification geometries, reference values, uncertainties, and comparison criteria require owner approval.
D9	UNRESOLVED	Applicable measurements, measurement uncertainties, and physical-validation criteria require owner approval.
D10	UNRESOLVED	Dataset grouping and training, model-selection, and final-test allocation require owner approval.
D11	UNRESOLVED	Global, local, subgroup, uncertainty, underprediction, and acceptable-performance criteria require owner approval.

D3 and D5 retain the archetype and scan concept, not rejected candidate numbers. The broad primary-source review and owner-approved comparison procedure are complete. Amendment completion still requires the predefined preliminary PHITS calculations, frozen evidence-based numerical criteria, and owner approval of one fixed source definition and construction.

Implementation and evidence reconciliation

Current Project records, absent evidence, application-capability claims, implementation-documentation mismatches, and the next executable action are maintained only in Current Project Evidence. The approved decisions and procedures on this page remain authoritative when the volatile evidence state changes.

Approved D3/D5 procedure

APPROVED DECISION — D3 component method: Use one fixed layout for all candidates and the finite Ta-only limiting case: P0 just upstream of Ta; deposition regions for every present non-void material; P1 just downstream of the complete assembly and off coincident boundaries; and finite side/edge surfaces. Use t-cross for surface current/flux, t-track for regional track-length fluence diagnostics, and t-deposit for per-material deposition. Production/yield diagnostics require a later D6-specific authorization. Omit component H*(10) and any arbitrary stand-off plane.

APPROVED DECISION — D3/D5 source method: Use the PHITS-native circular, mathematically unbounded Gaussian source on the Ta entrance plane, with PHITS FWHM as the canonical width convention and an explicit FWHM-to-σ conversion for reporting. Map the local right-handed basis as (u,v,w)=(+global X,+global Z,-global Y); verify source and tally transforms independently. The approved sensitivity coordinates are 2.36, 5.89, and 11.78 mm FWHM, with 5.89 mm retained only if the approved preliminary screen shows no qualitative change.

APPROVED DECISION — D5 scan method: Use an idealized uniform normalized one-dimensional scan along X with fixed normal incidence toward global −Y. Treat spatial density as following delivered charge/current and spatial speed when a physical waveform is available. Do not infer a turnaround dwell law or position-dependent angle law from literature examples; endpoint half weights, when used, are quadrature integration weights only.

APPROVED DECISION — D5 quadrature and audit method: Use nested composite-trapezoidal refinement with N' = 2N − 1, fixed normalized scan charge/current integral, and observable/statistics-aware convergence. Separate-position weighted sums are the audit default; PHITS multi-source is a checked alternative with explicit relative weights and totfact. Use the approved independent recorded-seed policy. Tolerances, histories, and run budget remain unresolved pending preliminary calibration.

UNRESOLVED: The final construction, spot width, scan span, and plate margin remain provisional until the approved preliminary checks confirm their retention conditions. Numerical convergence and audit-equivalence tolerances, histories, stopping rules, run budget, and the meaning of an adequate energy-accounting residual remain open pending preliminary calibration and owner approval. Public Type-304/water/channel engineering bounds are absent but are not prerequisites for the generic transport archetype; no engineering qualification may be inferred. Production/yield diagnostics, activation gates, occupancy criteria, and cooldown assumptions remain excluded from D3 pending a later D6-specific authorization. Applicable measurements or a validated surrogate for edge/spot physical validation remain unresolved under D8/D9; transport verification must not be presented as physical validation. D3 and D5 remain amendment-open until the preliminary evidence is reviewed and the final construction and source definition are approved.