Carbon-foamed closed-cell glass insulation blocks and slabs · 8. Foreground Data Collection, Calculation, and Quality Rules

8. Foreground Data Collection, Calculation, and Quality Rules

Data Collection Protocols

For each existing collection protocol that attributes manufacture of included production, packaging or observed-service assets across uses in this route (excluding amortization of the manufactured reference product itself), additionally retain a durable asset identity, its measured full physical mass and acquisition/installation/retirement stock history, original manufacture burden vector, all historical and current beneficiaries, matched-unit service readings, the evidenced whole installed-service denominator D, supported remaining-service basis and uncertainty, prior assigned shares and remaining original manufacture balance. Carry the ledger across reporting windows and custody/provider changes. Collect actual repair, replacement, recovery and treatment separately; these records do not normalize physical movements or create an additional asset purchase. These are added fields and conservation obligations for the existing applicable protocols, not a new assumed capital inventory.

protocol_id	process_id	flow_role	record_type	raw_fields	collection_method	unit	frequency	temporal_coverage	site_scope	aggregation_rule	quality_evidence
cp_mfg	manufacture	raw materials utilities and auxiliaries	foreground production records	supplier and whole delivery state; batch recipe; assay and moisture; each raw material mass; original meters and units; actual gas composition state density and LHV; electric supply voltage geography; campaign start stop idle failure rework; raw and non-WIP charge stocks; inherited exchange vectors; inbound transport per material and mode	Calibrated scales and meters with receipt/recipe reconciliation and direct equipment logs over the entire real route. Record whole purchased preparations once; validate actual operator flow sheet and individual real fuel/material names. Transport has actual mass distance and mode with separate real fuel/emissions if operated on site, or matched transport-service dataset outside this unit inventory. Retain undivided raw totals. convert_raw and attribute_raw precede the single normalize_exchange result per 1 kg reference flow; collection performs no division.	kg; MJ; m3 at declared gas state; original meter units	each batch and meter interval	complete representative period with all zero-output campaigns	declared site and equipment boundary through first release	per 1 kg reference flow	calibration; invoices; actual recipe; supplier specification; metering boundaries; stock vectors
cp_balance	manufacture	physical mass phase species and emission records	measurements and laboratory records	each inlet and outlet species; original volume or rate; time; temperature pressure humidity and compressibility; actual exhaust moisture and O2; retained cell gas and residual carbon; glass oxide and sulfate sulfide speciation; wet and dry waste; dissolved and suspended content; real wastewater receiving medium; actual returned water temperature enthalpy; opening closing melt powder block WIP and fluid charge; vector history	Independent calibrated mass/volume sampling and actual gas speciation across one consistent control volume. Separate stack from fugitive releases and each receiving medium. Distinguish measured species from analytical indicators; determine gas reactions, bound/free water and phase movement with independent records rather than residual fitting. Validate physical mass and elements before applying burden allocation. Retain undivided original physical/species observations. Use physical_gates separately; only attributed exchange totals enter normalize_exchange once per 1 kg reference flow.	kg by species; original m3 or rate; measured state and enthalpy	each campaign stock boundary and representative species sample	same manufacture period and stock history	manufacturing physical control volume and actual treatment outlets	per 1 kg reference flow	sampling reports; detection limits; gas calibration; independent elemental assays; moisture and flow-state records
cp_release	manufacture	first_release	quality release and net weighing	unique release identifier; product configuration; composition and carbon route; actual dimensions and density; closed-cell evidence; thermal conductivity and compressive test method results; applicable standard; anneal cutting and test record; net equilibrated mass; tare and surface condition; first-release time; destructive test loss; closing FG membership	Weigh actual uncoated accepted blocks using calibrated scales at recorded equilibrium condition; remove removable packaging and contamination, retain actual cell gases/residual carbon. Link each unique release to tests, lot, geometry and failures. Measure mass and external volume for the same configuration independently rather than importing nominal catalogue density. Collect raw Q in kg only; release_denominator defines Q and first_release represents 1 kg reference flow. No normalization of raw release records during collection.	kg; measured dimensions and test units	each block or traceable weighed lot at first release	same selected first-release cohort	first release after all required manufacturing quality tests	per 1 kg reference flow	scale calibration; unique release register; test results; tare; traceable configuration and equilibrium state
cp_service	service	packaging FG returns and services	cohort stock movement and meter records	original release cohort and Q; original delivery and reversal; opening closing FG; gross dispatch; returns received and sent; breakage; actual repair additions removals; each packaging empty attached WIP attached FG reusable stocks; purchases supplier-included receipts external returned receipts gross outgoing separate returns sent waste; wood paper moisture; actual repair board nail mass and origin; full article versus constituent; inherited exchange vector; raw energy water and fluid charges	Weigh each real article/configuration and moisture condition and reconcile mutually exclusive physical stocks and movements. Keep complete FG and packaging gross ledgers even for old cohorts; link return/reversal to original delivery, real repairs and inherited burden. Meter actual storage/handling/packing and cleaning services. Add real repair materials and gas/water changes as individual rows before claiming complete coverage, without inventing them. Collect undivided raw quantities. Reconcile packaging_ledger and physical_gates before attribute_raw; use normalize_exchange once per 1 kg reference flow of the original served cohort and its original Q. Older-cohort service remains separate from unrelated current manufacturing results.	kg; MJ; original article counts and moisture	each movement and matched service meter interval	selected release cohort and its original stock history	post-release physical and service gate	per 1 kg reference flow	stock and dispatch records; return links; calibrated article weights; actual repair evidence; inherited vectors; service meter coverage

Calculation Rules

rule_id	Applies to	Formula or rule	Inputs	Output	source_ids
convert_raw	all inventory rows	Convert each actual raw record to its row unit and declared state before stock/burden attribution. Electricity MJ = kWh * 3.6. Integrate only actual mass/volume rates over measured time; cumulative meter differences and already integrated volumes are amounts and receive no second time multiplier. Use independent substance-specific density at measured state for mass/volume, and actual article mass per configuration for counts. Gas native secondary factors never replace measured density or actual LHV.	raw readings; unit factors; measured density and state; article configuration	converted raw totals	bau-foam-glass-pcr-2025
attribute_raw	all inventory rows	Distinguish physical movement quantities from burden/consumption numerators before attribution. For each manufacture exchange j reconcile opening unreleased raw/WIP and unassigned installed/service vector + actual converted additions = corresponding closing unreleased/unassigned vector + first-release transfer B_j^Q + evidenced assignments to other real cohorts/functions. Manufacturing burden A_j is B_j^Q after documented subdivision and any genuine residual allocation; it includes all newly first-released FG even when held at closing, never subtracting that FG vector. Ordinary waste, offcuts, failed lots and destructive tests retain their caused manufacturing and handling/treatment burdens: physical export or sale alone cannot deduct their inherited burden as transferred elsewhere or create a credit. Any assignment elsewhere needs traced distinct cohort/function and the applicable allocation or burden-handover decision. FG history separately reconciles opening FG vector + B_j^Q + actual returns at original dispatch vector + actual service increments = closing FG vector + gross dispatch vector + other traced removals/destinations. Original manufacture is restored on return for tracing, not added again. Service burden A_j includes each actually performed increment for the original served cohort using original Q, even if serviced goods remain in closing FG; correct only disjoint unconsumed/unassigned service or charge inventories. For explicit purchase/receipt, gross dispatch, return and waste movement cards A_j is the converted actual physical movement; maintain burden/consumption vectors separately, without netting unused stock against a gross receipt. Each B retains that exchange numerator/unit, never stock kg substituted for every burden; never correct already stock-adjusted consumption twice. Internal transfers cancel once in the consolidated account.	converted raw totals; exchange-specific stock vectors; cohort and return links; allocation records	A_j attributable raw exchange totals	ilcd-attributional-2010
release_denominator	first_release	Q = sum of measured net mass in kg of uniquely first-released blocks/slabs of the selected product cohort after required quality tests. Use recorded equilibrium state and remove packaging and surface contamination; keep true retained cell gas and residual carbon. Pre-release failures/test destruction are excluded; their burdens remain. First-released closing FG is already in Q. Dispatch, return, repair and retesting add no Q.	cp_release; unique release registry; net mass; quality tests	Q in kg	foamglas-t3plus-epd-2026; foamglas-t3plus-pds-2025; bau-foam-glass-pcr-2025
normalize_exchange	all inventory rows	q_j = A_j / Q for Q > 0. Execute this single division after convert_raw and attribute_raw; q_j is the exchange per 1 kg reference flow for the same release cohort. For Q = 0 retain A_j and raw burdens without a unit result. Flow cards and protocol aggregation only reference this same computed result and never divide again.	A_j; Q; convert_raw; attribute_raw; cp_release	q_j per 1 kg reference flow	ilcd-attributional-2010; bau-foam-glass-pcr-2025
physical_gates	all inventory rows	Manufacturing: opening raw + opening WIP + opening non-WIP charges + external supplied materials + measured gross inlet gas species = closing raw + closing WIP + closing charges + Q + each physical exported waste + each measured outlet species + each actual product return. FG service: opening FG + first-release internal receipt + external returned FG + actual repair additions + measured physical uptake = closing FG + gross dispatch + standalone external FG return + scrap + repair removals + measured gas/water loss. Apply consistent wet/phase/thermal bases and independently measured relevant elements and chemical species; one residual is not two measurements. Consolidation cancels manufacture-to-service receipt. Manufacture closing excludes already first-released FG; the combined gate cancels internal Q and first-release vectors, not completed manufacturing or performed service. Gross carrier-gas observations serve physical closure; report actual environmental withdrawal or new process release with independent origin, reaction, retention, phase and native-flow convention evidence. Equal chemical names do not make resource-input and emission UUIDs interchangeable or establish a characterization offset. Do not add closing FG to production Q or report retained cell gas as external emission.	cp_mfg; cp_balance; cp_release; cp_service; independent species and phase measurements	separate and consolidated physical and elemental gates	foamglas-t3plus-epd-2026; bau-foam-glass-pcr-2025
packaging_ledger	all inventory rows	For EACH actual packaging article: opening empty + opening attached WIP + opening attached FG + opening reusable + new purchases + supplier-included receipts + external returns received = closing empty + closing attached WIP + closing attached FG + closing reusable + with gross physical shipments + standalone external returns sent + waste. All categories are mutually exclusive and each receipt counted once. Add real article water uptake/loss and actual repair-part additions/removals as separate measured terms when needed; keep complete pallet and repair components distinct without counting constituents twice. Internal reuse cancels; burden allocation and inherited exchange vectors remain separate. Packaging never enters Q or automatic co-product credit.	cp_service; article stocks; physical receipts and dispatch; moisture; real repairs; inherited vectors	each article physical ledger and burden reconciliation	foamglas-t3plus-epd-2026; bau-foam-glass-pcr-2025; ilcd-attributional-2010
asset_manufacture_share	existing capital/reusable manufacture attribution	For each included original asset manufacture vector B, f_current = u_current / D only after the persistent ledger verifies the supported whole installed-service denominator, all beneficiaries and prior assigned shares. Attribute f_current * B once to the actual beneficiary before the existing single reference normalization. Check prior_shares + all current/new beneficiary shares <= 1; carry the remaining original balance. Capacity revisions reconcile the historical ledger. Full physical asset/packaging mass and actual gross stocks/returns never receive f_current. Unknown D or history leaves the final attributed result unavailable, without suppressing actual repair/replacement/service burdens.	asset_id; original_B; u_current; D; historical_beneficiaries; prior_shares; capacity_evidence	conserved attributed manufacture vector

Data Quality Requirements

requirement_id	Applies to	Requirement	Evidence
dq_identity	all inventory rows	Every row retains full native restrictions and actual matched provider/state; UUID is not a default amount, conformity certificate or proof of recipe. Unresolved auxiliaries remain single named exchanges.	Native direct records; provider specification; actual product tests
dq_period	all inventory rows	Retain complete raw meter coverage with zero-output failures start/stop and idle. Separate representative production data from legacy supplier examples and manufacturer marketing claims.	Raw readings; calibrated logs; cohort register
dq_species	all physical gates	Measure phase and molecular species independently of total mass/element balance. Cell-retained CO2 and residual carbon, sulfate/sulfide, actual free/reaction/structural water and exhaust O2 require genuine evidence. Keep indicators and molecular species distinct.	Independent analytical speciation; phase records; uncertainty and detection limits
dq_stocks	all stocks and service	Each raw/WIP/FG/charge/packaging stock has physical quantity and inherited vector per exchange. Preserve original release and delivery reversals; actual rework additions and moisture/species changes are not silently ignored.	Opening closing measurements; cohort links; vector history; real repair documents
dq_unknown	dataset completion	Unknown, absent, measured zero and below detection are different states. Missing actual core route, terminal or unresolved causal relationship stays a true hold. Provider matching, unbound auxiliaries and future site activity quantities remain explicit ordinary production gaps; no fabricated default or scientific approval.	Gap register; applicability decisions; raw observations

9. Validation Rules

rule_id	Applies to	Rule	source_ids
validate_terminal	reference product and route	Verify actual unfaced closed-cell carbon-foamed glass slab/block from the entire manufacturing route and actual release tests, dimensions, density and surface condition. Generic Glass (formed & finished) does not certify closed cells, building suitability, thermal conductivity, compressive strength or any legal compliance. Do not apply the entire CPC leaf to one product or replace carbon-black foaming with a different foaming agent.	foamglas-route; foamglas-t3plus-epd-2026; foamglas-t3plus-pds-2025; bau-foam-glass-pcr-2025
validate_identity	each inventory row	Check whole direct native basename, all synonyms/comments, CAS, grade, route, geography, form/particle structure, medium, type, every property and exact reference chain before UUID adoption. Retain each restriction in the description and actual provider evidence. Isolated taxonomy defects are disclosed separately from physical consistency; hard CAS/route/media contradictions are not waived.	bau-foam-glass-pcr-2025
validate_physical	manufacturing service and consolidated gates	Close gross manufacturing and FG-service physical gates independently and then cancel matching internal transfers. Include all actual species, water phases, cell-retained gas/carbon, dust and moisture using independent measured composition and phase evidence. Keep physical mass/element conservation independent of burden allocation. Unknowns and imbalance retain explicit uncertainty/hold; never fit two unknown reactions by one residual.	foamglas-t3plus-epd-2026; bau-foam-glass-pcr-2025
validate_normalization	all raw quantities	Convert units/state first, reconcile stock and historical vectors and allocate second, execute normalize_exchange once as A_j/Q last. Collection retains undivided raw data; cards and aggregation reference that same result and never divide it again. Q=0 retains all real raw burdens without a unit result. Manufacturing production Q, gross shipments and net deliveries are distinct. Require the attribute_raw distinction between first-release manufacture, already performed service, unassigned inventories and physical movement; neither closing FG nor ordinary waste export automatically removes caused burdens.	ilcd-attributional-2010; bau-foam-glass-pcr-2025
validate_packaging	each packaging article	Require complete mutually exclusive opening/closing empty, WIP-attached, FG-attached and reusable stocks; new purchases; supplier-included physical receipts; external returns received; packaging with gross physical shipment; standalone external returns sent and waste. Count only one physical receipt per article. Internal reuse cancels and does not change Q. Identify real wood/paper moisture, nails, replacement boards and other actual repair parts separately from complete pallet purchases; do not invent repairs.	foamglas-t3plus-epd-2026; bau-foam-glass-pcr-2025; ilcd-attributional-2010
validate_coverage	dataset completion	Actual unknown core manufacturing/terminal or causal burden relationship prevents a completed dataset claim. Preserve failed retrievals, source edition differences, mechanical applicability mismatches and unknown/absent/zero/below-detection separately. Ordinary future site measurements, provider matching and auxiliary UUID gaps remain explicit data-production requirements; technical projection/measurement checks do not establish scientific, translation, mapping or publication approval.	foamglas-t3plus-epd-2026; bau-foam-glass-pcr-2025; ilcd-attributional-2010
validate_asset_manufacture_conservation	existing capital/reusable manufacture attribution	For every included capital/reusable manufacture attribution verify same-asset continuity, whole installed service in matched units, all historical/current beneficiaries, prior assigned manufacture shares and cumulative share <= 1. Reject a reset by period, cohort, provider, owner or repair. Check distinct real replacement and incremental repair/treatment vectors separately and preserve gross physical mass. A missing denominator, unexplained previous allocation or incomplete beneficiary ledger requires explicit review and an unavailable final attributed amount; structural inspection cannot claim empirical completeness.

10. Published Dataset Profile

Field	Value
dataset_role	secondary_dataset; background_dataset
downstream_use	Qualified upstream actual uncoated cellular-glass product at declared first-release state; separately reported original-cohort packaging/handling services as applicable
allowed_use	Compatible composition route state density performance and provider boundary; transparent measured kg-to-volume conversion where needed
excluded_use	Automatic EN15804 or product certification; arbitrary material-grade thermal equivalence; replacing full glass manufacture with repacking; coated product or gravel; using gross shipment or net delivery as Q; automatic avoided-energy or moduleD credit
required_metadata	uncoated closed-cell block or slab and declared function; carbon-black particle grade structure and supplier route; glass constituents and whole supplied forms; cullet origin treatment and burden handover; actual melting milling mould foaming anneal cutting route; density and external dimensions measured for same configuration; thermal and mechanical test method results and applicable product standard; unique first quality release cohort and original Q; equilibrium surface moisture and retained-cell gas state; raw WIP FG and non-WIP charge quantities with exchange history; failure idle start-stop and zero-output coverage; actual gas composition pressure temperature density LHV and meter state; electricity geography voltage mix and provider boundary; independent mass element molecular-species and water-phase evidence; packaging article composition moisture stocks gross movements and real repairs; original delivery reversal and returned FG incremental rework links; actual external water-return state enthalpy and supplier convention; allocation subdivision causality residual gate and period evidence
required_quality_disclosure	Original raw quantities and conversions; actual coverage and failures; every physical gate and exchange-specific stock vector; molecular and phase evidence; original cohort service boundary; all identity/provider limitations; unresolved auxiliary UUIDs; unknown and detection-limit distinctions
update_trigger	Changed core chemistry forming or product state; coating/facing; actual supplier route/form/geography; gas reference state; electricity mix/voltage; quality standard; stock cohort or allocation method

11. Data Sources

Source id	Type	Reference	Used for
foamglas-route	handbook	Owens Corning FOAMGLAS. What is cellular glass insulation? https://www.foamglas.com/en/advice-center/general-advice/what-is-cellular-glass-production-process	Primary complete manufacturing description from recycled glass and minerals through carbon milling mould foaming controlled cooling and cut finished block. Current50% recycled-content statement is not a recipe default.
foamglas-t3plus-epd-2026	dataset	Pittsburgh Corning Europe NV; IBU. FOAMGLAS T3+ EPD-PCE-20260415-IBA1-EN issued2026-08-18 valid to2031-08-17. https://www.foamglas.com/-/media/project/foamglas/public/shared/files/certifications/b---common-certificates/epd-environmental-product-decl/epd-foamglas-t3plus-en.pdf	Full publisher PDF actual2026 edition: printedpp1–4 product raw ingredients13-stage diagram uncoated release packaging and2023 two-site declared1m3 context. Source density91kg/m3 differs from nominal95; constituent percentages temperatures life-cycle results mass allocation and module-D scenarios are case-specific and not generic defaults.
foamglas-process-brochure	handbook	Pittsburgh Corning. FOAMGLAS insulation general brochure; publisher retained legacy process illustration p5. https://www.foamglas.com/-/media/project/foamglas/public/corporate/foamglascom/files/brochures/industry/brochures-2019/brochure---foamglas-insulation---general-brochure.pdf	Complete primary industrial brochure with manufacturing diagram atp5 and adjacent industrial-use context. Historical>66% recycled claim and furnace temperatures are not transferred; this is not an independent second manufacturer or proof of one contemporary plant/order.
foamglas-t3plus-pds-2025	handbook	Owens Corning. FOAMGLAS T3+ Product Data Sheet version08.05.2025. https://www.foamglas.com/-/media/project/foamglas/public/shared/files/foamglas-b/foamglas-t3-plus/pds/pds-foamglas-t3plus-int-en.pdf	Full2-page primary building slab terminal dimension-table performance-metadata categories and application limitations. No catalogue density dimensions conductivity strength binder absence or service-life claim becomes an unmeasured site value.
bau-foam-glass-pcr-2025	official_guidance	Bau EPD GmbH. PCR2.22.2.3 Foam glass Part B version1.0 dated2025-09-09 §§2.5–2.8 3.1 5.4.5–5.4.6. https://www.bau-epd.at/fileadmin/user_upload/PCR_product_specific_A2_English/BAU-EPD-PKR-B-2.22.2.3-Foam_glass-Version1.0-20250909-EN15804_A2-English-website.pdf	Independent programme guidance on ingredient/auxiliary disclosure process and packaging identities declared-unit conversion physical mass/water balance and energy double counting. Adjacent aggregate standard and99.8% recycled waterglass/glycerine example are different products/routes.1m3 programme default permits transparent other units; not evidence of carbon-black recipe or EN15804 compliance for this PCR.
ilcd-attributional-2010	official_guidance	European Commission JRC. ILCD Handbook General guide for LCA Detailed guidance First edition March2010 §§7.9.2–7.9.3. https://eplca.jrc.ec.europa.eu/uploads/ILCD-Handbook-General-guide-for-LCA-DETAILED-GUIDANCE-12March2010-ISBN-fin-v1.0-EN.pdf	Attributional multifunctionality only: dedicated/virtual subdivision then exchange-specific determining physical causality and only genuinely residual actual-state contemporaneous common-process-gate market-value allocation. No industrial process recipe factor default percentage or automatic substitution credit.