mirror of
https://github.com/Hestia-Homes/Model.git
synced 2026-06-08 11:17:27 +00:00
a532f75db0
Per SAP 10.2 spec p.75 (Effective solar volume section): "in the case of a combined cylinder (such as arrangement b) in Figure H2, [(H14) =] the volume of the dedicated solar storage plus 0.3 times the volume of the remainder of the cylinder." The spec leaves the dedicated solar storage volume (H12) itself to the surveyor / certified-software convention for combined-cylinder setups. Empirical pattern across 4 Elmhurst worksheets establishes the combined-cylinder default H12 = 1/3 × cylinder volume (H13): cert 000565 worksheet: H12 = 53 L, H13 = 160 L (ratio 0.331) cert A worksheet: H12 = 37 L, H13 = 110 L (ratio 0.336) cert B worksheet: H12 = 37 L, H13 = 110 L (ratio 0.336) cert C worksheet: H12 = 37 L, H13 = 110 L (ratio 0.336) This matches the broader f-chart literature convention for "solar pre-heat zone" sizing in stratified tanks: roughly the lower third of the cylinder is reserved for solar pre-heat, the upper two-thirds for boiler-heated water. The Elmhurst-certified rounding is to the nearest integer litre (53.33 → 53; 36.67 → 37). `_solar_hw_monthly_override` now derives H12/H13 from `epc.has_hot_water_cylinder + sap_heating.cylinder_size` via the existing `_CYLINDER_SIZE_CODE_TO_LITRES` Table 28 lookup (RdSAP 10 §10.5). When no cylinder is lodged (instantaneous + solar HW shape, hypothetical), fall back to Table 29's 75 L separate pre-heat tank. Cert 000565 closure: - Orchestrator solar Q_s annual: 268 → 283 kWh/yr (worksheet 281.35, Δ +1.73, 0.6% error). Within the same precision band as appendix_h_solar's per-month <1e-3 kWh pin. - HW pin: −69 → −86 kWh/yr (slight regression due to compounding with three independent demand-cascade bugs not yet wired: (45)m over by 903, primary_loss (59) under by 1175 (cylinder HP routing missing), storage_loss (56) over by 98 + missing (57)m solar adjustment). These were previously masked by the +357 kWh "no solar credit" over-count; now visible as the residual. Solar Q_s closure is the load-bearing achievement here — the Appendix H orchestrator is now spec-pinned through to the cert's own lodged geometry. HW pin closure follows once the demand-path gaps land in follow-on slices. Test baseline: 547 pass + 9 expected `test_sap_result_pin[000565-*]` fails (unchanged). Cohort-2 + ASHP cohort + golden fixtures untouched — no other cert lodges solar HW. Pyright net-zero (34 errors baseline → 34 errors post-change). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> |
||
|---|---|---|
| .. | ||
| addresses | ||
| sap10_calculator | ||
| sap10_ml | ||
| tasks | ||
| postcode.py | ||