Slice S0380.77: WHC 914 routes primary-loss gate to DHW main; cert 000565 +1175 kWh

SAP 10.2 §4 line 7700 + Table 3 (PDF p.159) defines primary loss
as the loss between "a heat generator (e.g. boiler) connected to a
hot water storage vessel via insulated or uninsulated pipes (the
primary pipework)". The spec keys eligibility off the heat
generator that feeds the cylinder, NOT the space-heating main.

Verbatim spec (Table 3, p.159):

    Primary circuit loss applies when hot water is heated by a heat
    generator (e.g. boiler) connected to a hot water storage vessel
    via insulated or uninsulated pipes (the primary pipework).
    Primary loss is set to zero for the following:
      - Electric immersion heater
      - Combi boiler (including when it is part of a combined heat
        pump and boiler package and provides all the hot water)
      - CPSU (including electric CPSU)
      - Boiler and thermal store within a single casing
      - Separate boiler and thermal store connected by no more than
        1.5 m of insulated pipework
      - Direct-acting electric boiler
      - Heat pump (not combined heat pump and boiler package with a
        non-combi boiler) from PCDB with hot water vessel integral
        to package

    For other cases (indirect cylinders and thermal stores connected
    by uninsulated pipework or more than 1.5 m of insulated
    pipework) the loss in kWh/month is calculated as follows.

    Primary loss = n_m × 14 × [{0.0091 × p + 0.0245 × (1 − p)} × h
                              + 0.0263]

`_water_heating_section` was routing the gate through
`_first_main_heating` (Main 1). Cert 000565 lodges ASHP Main 1 +
gas combi Main 2 + WHC 914 ("from second main system") + 160 L
combined cylinder + cylinder thermostat "N". The cylinder is
heated by Main 2 via uninsulated primary pipework, so the spec
formula applies; the cascade was zeroing (59)m because Main 1's
HP record (or its absence) drove the gate.

Fix: `_primary_loss_override` resolves its `main` via
`_water_heating_main(epc)` (the WHC-914 routing) so the
eligibility check follows the heat generator that physically
incurs the loss. Call site loses the `main` argument; the gate's
internals are unchanged.

Cert 000565 worksheet line (59)m (U985-0001-000565 Block 1):

    128.3772, 115.9536, 128.3772, 124.2360, 128.3772,  41.9160,
     43.3132,  43.3132,  41.9160, 128.3772, 124.2360, 128.3772

sums to 1174.79 kWh/yr. Cascade now matches every month at < 1e-4
kWh under the spec params (uninsulated p=0 from RdSAP §3 age band
A-J default, h_winter=11 / h_summer=3 from "no cylinder
thermostat" row).

Cert 000565 movements at HEAD:
  - sap_score:                   29 ✓ EXACT (unchanged)
  - sap_score_continuous:        29.2905 → 29.1090 (Δ +0.78 → +0.60)
  - space_heating_kwh:           59274.46 → 58725.77 (Δ +266 → −283)
  - main_heating_fuel:           34867.33 → 34544.57 (Δ +157 → −166)
  - hot_water_kwh:                3668.54 →  5154.02 (Δ −86 → +1399)
  - co2_kg:                       6352.61 →  6616.01 (Δ −95 → +168)
  - total_fuel_cost_gbp:          4611.14 →  4627.10 (Δ −69 → −53)
  - ecf:                           5.3073 →   5.3256 (Δ −0.08 → −0.06)

HW pin over-shoots by +1399 as expected — the +1174.79 spec credit
is amplified by the gas combi's 0.88 water efficiency (≈ +1335
kWh) plus cascade coupling (heat-gain feedback into space heating
drops it −549 kWh). The +1399 will be pulled back by the next two
demand-cascade slices:
  - (45)m energy_content over by ~903 kWh (occupancy + shower mix)
  - (56)/(57)m storage-loss over by ~98 kWh + missing solar
    `(57)m = (56)m × (H13 − H12)/H13` adjustment (~298 kWh)

New test pins the full 12-month (59)m tuple for cert 000565's
heating shape at abs=1e-4. Old `_first_main_heating`-based test
(`test_cert_with_hot_water_cylinder_computes_primary_loss_59m_…`)
unchanged — still passes via the WHC-901 fall-through to Main 1.

Test baseline: 547 → 548 pass (new test added) + 9 expected
`test_sap_result_pin[000565-*]` cascade-gap fails. Pyright
net-zero on both touched files (45 baseline = 45 after).

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

domain/sap10_calculator/rdsap/cert_to_inputs.py

@@ -3216,8 +3216,11 @@ def _water_heating_worksheet_and_gains(
     storage_loss_override = _cylinder_storage_loss_override(epc, main)
     # SAP 10.2 §4 line 7700 + Table 3 (PDF p.159) — primary circuit loss
     # (59)m. Only fires for indirect cylinders; HPs with integral
-    # vessels and combi boilers are in the spec's zero list.
-    primary_loss_override = _primary_loss_override(epc, main, primary_age)
+    # vessels and combi boilers are in the spec's zero list. The gate
+    # keys off the *DHW* main (`_water_heating_main`) so WHC 914 ("from
+    # second main system") routes the primary-loss eligibility check
+    # to the heat generator that actually feeds the cylinder.
+    primary_loss_override = _primary_loss_override(epc, primary_age)
     # SAP 10.2 Appendix H — solar HW contribution (63c)m. Only fires
     # when the cert lodges solar HW; orchestrator drives off lodged
     # collector geometry + RdSAP 10 §10.11 Table 29 defaults for
@@ -3260,7 +3263,6 @@ def _water_heating_worksheet_and_gains(
 
 def _primary_loss_override(
     epc: EpcPropertyData,
-    main: Optional[MainHeatingDetail],
     primary_age: Optional[str],
 ) -> Optional[tuple[float, ...]]:
     """Resolve (59)m for `water_heating_from_cert` from the cert + PCDB
@@ -3270,7 +3272,18 @@ def _primary_loss_override(
     comes from RdSAP §3 age-band default (no API field); circulation
     hours h come from Table 3 keyed on cylinder thermostat + separately-
     timed-DHW lodgement.
+
+    The gate keys off the DHW main resolved via `_water_heating_main`
+    (the WHC-914 "from second main system" routing) rather than
+    `_first_main_heating`. SAP 10.2 §4 line 7700 + Table 3 (PDF p.159)
+    define primary loss as the loss between the *heat generator that
+    heats the water* and the storage vessel — so the eligibility check
+    must follow the DHW routing. Cert 000565 (ASHP Main 1 + gas combi
+    Main 2 + WHC 914 + 160 L cylinder) is the cohort case: Main 1's
+    HP record is irrelevant; Main 2's combi feeds the cylinder via
+    primary pipework and incurs the loss.
     """
+    main = _water_heating_main(epc)
     cylinder_present = bool(epc.has_hot_water_cylinder)
     hp_record: Optional[HeatPumpRecord] = None
     if main is not None and main.main_heating_index_number is not None:

domain/sap10_calculator/rdsap/tests/test_cert_to_inputs.py

@@ -1802,6 +1802,83 @@ def test_cert_with_hot_water_cylinder_computes_primary_loss_59m_from_sap_table_3
     )
 
 
+def test_whc_914_dhw_routes_primary_loss_gate_to_second_main_heating_per_sap_table_3() -> None:
+    """SAP 10.2 §4 line 7700 + Table 3 (PDF p.159) primary-loss eligibility
+    is determined by the heat generator that feeds the hot water storage
+    vessel, not by the space-heating main. Cert 000565 lodges Main 1 = ASHP
+    (SAP 224) + Main 2 = gas combi (PCDB 15100) + water_heating_code 914
+    ("from second main system") + 160 L combined cylinder + cylinder
+    thermostat absent ("N"). The cylinder is heated by Main 2 via uninsulated
+    primary pipework, so the Table 3 formula applies with p=0 and the
+    "no cylinder thermostat" hours (h=11 winter, h=3 summer).
+
+    Worksheet line (59)m for U985-0001-000565 (Block 1) reads
+        128.3772, 115.9536, 128.3772, 124.2360, 128.3772, 41.9160,
+         43.3132,  43.3132,  41.9160, 128.3772, 124.2360, 128.3772 kWh
+    summing to 1174.79 kWh/yr. The cascade must route the primary-loss gate
+    to `_water_heating_main` (the WHC-914-resolved DHW main) — gating on
+    `_first_main_heating` mis-keys the gate to Main 1's HP record (or
+    absent record) and zeroes (59)m even though the gas combi → external
+    cylinder pipework physically incurs the loss.
+    """
+    # Arrange — synthesise cert 000565's heating shape: ASHP Main 1 +
+    # gas-combi Main 2 servicing DHW via WHC 914 + cylinder lodged with
+    # no cylinder thermostat. The Elmhurst mapper produces
+    # `main_heating_category=None` on Main 1 when the cert lodges a SAP
+    # code without a PCDB Table 362 reference (see
+    # `_elmhurst_main_heating_category` TODO docstring) — mirror that
+    # shape here so the current `_first_main_heating` routing fails the
+    # gate even though Main 2 (the DHW main) plainly carries the loss.
+    hp_main = MainHeatingDetail(
+        has_fghrs=False,
+        main_fuel_type=29,  # electricity
+        heat_emitter_type=1,
+        emitter_temperature=1,
+        main_heating_control=2206,
+        main_heating_category=None,
+        sap_main_heating_code=224,
+    )
+    combi_main = _gas_boiler_detail(sap_main_heating_code=102)
+    epc = make_minimal_sap10_epc(
+        total_floor_area_m2=_TYPICAL_TFA_M2,
+        habitable_rooms_count=4,
+        country_code="ENG",
+        has_hot_water_cylinder=True,
+        sap_building_parts=[make_building_part(construction_age_band="D")],
+        sap_heating=make_sap_heating(
+            main_heating_details=[hp_main, combi_main],
+            water_heating_code=914,  # DHW from second main system
+            cylinder_size=3,  # Medium = 160 L
+            cylinder_insulation_type=1,
+            cylinder_insulation_thickness_mm=50,
+            cylinder_thermostat="N",  # no cylinder thermostat → h_winter=11
+        ),
+    )
+
+    # Act
+    wh_result, _ = _water_heating_worksheet_and_gains(
+        epc=epc,
+        water_efficiency_pct=0.88,
+        is_instantaneous=False,
+        primary_age="D",
+        pcdb_record=None,
+    )
+
+    # Assert — full 12-tuple matches cert 000565 worksheet (59)m at 1e-4.
+    assert wh_result is not None
+    expected_59m = (
+        128.3772, 115.9536, 128.3772, 124.2360, 128.3772, 41.9160,
+        43.3132, 43.3132, 41.9160, 128.3772, 124.2360, 128.3772,
+    )
+    got_59m = wh_result.primary_loss_monthly_kwh
+    for month_idx, (got, want) in enumerate(zip(got_59m, expected_59m)):
+        assert abs(got - want) < 1e-4, (
+            f"(59)m month {month_idx + 1}: got {got!r}, want {want!r} per "
+            f"SAP 10.2 §4 line 7700 + Table 3 (uninsulated p=0, no "
+            f"cylinder thermostat h=11/3); cert 000565 worksheet line (59)"
+        )
+
+
 def test_air_source_heat_pump_main_heating_zeroes_table_3a_combi_loss_per_sap_4_line_7702() -> None:
     """SAP 10.2 §4 line 7702 worksheet defines (61)m as 'Combi loss for
     each month from Table 3a, 3b or 3c (enter "0" if not a combi