CORRUGATED FLUTE PUBLISHED-REFERENCE RECONCILIATION
Version: 2.1.0
Verification date: September 15, 2026
Publication: Green Bay Dock Door Repair Research
Reference page: https://greenbaydockdoorrepair.com/research/corrugated-flute-chart/

WHAT THIS RELEASE IS
A source-labeled compilation and publication audit. There are 44 reference rows
from eight publishers or author groups. Three rows are examples from Table 1 of
Rogalka, Grabski and Garbowski (Sensors, 2023), with image-derived dimensional
estimates. This project made no new physical board measurements.

The separate unit-pairing audit covers five pages from four publishers. Arvco
has two pages. This is a selected audit, not a prevalence study or copying-history
investigation. The manufacturer file distinguishes four corporate, plant or
registered-declaration scopes; omissions do not prove product unavailability.

INPUTS (keep beside the Python script)
flute-source-records.csv: 44 attributed reference rows; original labels, units,
source titles/URLs, source dates where visible, date basis, verification date,
evidence tier and field values. Empty means absent/untranscribed, never zero.
unit-mismatch-inputs.csv: Five primary page observations used in the unit audit.
manufacturer-source-audit.csv: Four source scopes used in the manufacturer tables.

REPRODUCTION
Requires Python 3.9 or later; standard-library modules only.
  python flute_geometry_v2.py
Optional separate output directory:
  python flute_geometry_v2.py --out recomputed

The script writes five derived CSVs and the consolidated JSON. It checks unique
source identifiers, required provenance, corpus counts and numerical convergence.
It never fetches websites, updates source dates or performs source re-verification.
Original input CSVs are not overwritten. The script does not regenerate this ZIP.

OUTPUTS
- derived-cross-convention.csv: Eight Summa pitch profiles converted to counts.
  Literal overlap with an approximate FBA count is not a tolerance test.
- derived-image-study.csv: Three image-derived examples, count conversions,
  thickness differences, liner sums, residuals and hypothetical sine-model factors.
- unit-mismatch-audit.csv: Five source observations with exact conversions and
  differences; percentage denominator is the stated 3.2 mm.
- derived-fractions.csv: Five PakFactory fractions converted exactly to mm, rounded
  to one decimal, with the arithmetic difference recorded separately.
- derived-fefco-counts.csv: Five original FEFCO groups converted to counts per foot.
  F/G/N is preserved as one grouped reference, not three independent observations.
- corrugated-flute-reference.json: Inputs, outputs, constants, metrics and limits.

SELECTED FIELD DEFINITIONS
record_id: Stable identifier within this release.
publisher: The named source publisher or study author group, not a laboratory count.
source_date: Date visibly stated by the source; blank if not established.
date_basis: What that date means; unrelated page dates are not substituted.
date_checked: Source verification date, not a measurement or production date.
tier: PRIMARY_PUBLICATION, PRIMARY_STUDY, or AUDIT_ONLY.
evidence_role: publisher_reference, published_image_estimate, or publication_audit_only.
verification_symbol: Star means directly checked against the named publication.
verification_status: Bounded verification; does not imply independent physical testing.
height_mm / pitch_mm: Published reference or image-estimate fields as labeled.
board_caliper_mm: Publisher caliper references; for PRIMARY_STUDY rows it stores the
  source's image-estimated board thickness, NOT an ISO pressure-gauge measurement.
reported_thickness_mm: Published thickness where a more specific basis is not established.
thickness_inches: A published fraction, not a universal nominal order specification.
flutes_per_foot / flutes_per_metre: Original count units; qualifiers remain in notes.
upper_liner_mm / lower_liner_mm / medium_material_thickness_mm: Image-study inputs.
take_up_factor: Published indicative medium length/area ratio, not a total-paper cost.
glue_g_m2_per_glue_layer: Grams per square metre PER GLUE LAYER, as FEFCO specifies.

CALCULATIONS AND ROUNDING
1 inch = exactly 25.4 mm; 1 foot = exactly 304.8 mm.
Flutes per foot = 304.8 / pitch in mm.
Flutes per foot = flutes per metre * 0.3048.
A count per 300 mm must be multiplied by 1.016 to become a count per foot.
11/64 inch = 4.365625 mm. (4.365625 - 3.2) / 3.2 * 100 = 36.42578125%.
Reported display value: 36.4%, relative to 3.2 mm.
Image residual = board thickness - flute height - upper liner - lower liner.
A residual is an arithmetic result, not a diagnosis of glue or nesting geometry.
Sine model: y = (h/2) sin(2 pi x/p); normalized arc length over one period.
Simpson integration at 20,000 and 10,000 intervals checks numerical convergence.
This is a model using image-derived inputs, NOT measured paper consumption and
NOT a validation against matched FEFCO production specimens.
Decimal/fraction arithmetic precedes display rounding. Longer output decimals do
not increase the precision of the published source values.

SOURCE AND USE BOUNDARIES
The publication is primary evidence of what its publisher printed. That alone does
not make every printed value a universal technical specification. Wikipedia and
WebstaurantStore rows are audit-only; neither supplies authoritative dimensions
for the main chart. Wikipedia fraction cells that could not be directly transcribed
remain blank; PakFactory's independently inspected fractions support the main chart.

Full paid ISO, ASTM, DIN and carrier documents were not reviewed or redistributed.
The page states only the verified public catalogue/guidance scope. No claim that no
standard contains flute dimensions is made. No market-share percentage is supplied.

Rights in third-party publications remain with their rights holders. This release
does not assign them a new license. The cited Sensors article is CC BY 4.0; its
Table 1 inputs are attributed, and the compilation's calculations are identified
as adaptations rather than as the authors' own derived results. No source PDFs,
paywalled documents or copied manufacturer diagrams are distributed.

RELEASE CHANGES FROM THE SUPPLIED 2.0.0 DRAFT
- Five audited pages correctly distinguished from four publishers; no causal copying claim.
- Visible page dates distinguished from unavailable dates and search-index timestamps.
- Missing PCG B/E caliper entries added; five directly checked PakFactory fractions added.
- Raw/derived values separated; FEFCO's F/G/N group preserved and glue denominator restored.
- Image-estimated dimensions distinguished from physical gauge measurements.
- Sine-model comparisons recomputed and described as unmatched illustrative comparisons.
- Corporate/plant/declaration scopes kept separate; capacity distinguished from output.
- Unsupported market shares, standards exclusions and physical-cause claims removed.

REFRESH RULE
Recheck the source itself before changing date_checked or the visible verification
date. A data edit requires a versioned release and rerun of dependent calculations.
A script rerun, calendar change or link check alone is not full source verification.
