fix(hw-cost): WHC-903 immersion off-peak HW bills at Table 13 high-rate fraction

Electric immersion water heating (WHC 903) on an off-peak tariff billed
100% at the low rate, under-costing the dwelling and over-rating it
(median +0.98 SAP across the off-peak WHC-903 API cohort, n=57).

SAP 10.2 Table 12a "Immersion water heater" row (PDF p.191) routes the
water-heating column to Table 13 (PDF p.197): the high-rate fraction is
a function of cylinder volume V, assumed occupancy N (Appendix J Table
1b) and single-/dual-immersion. The remainder bills at the low rate.
Table 13 Note 2 supplies exact equations equivalent to the rounded grid;
`electric_dhw_high_rate_fraction` evaluates them (validated against the
published 110 L grid cells). Per Note 1 the 10-hour equations cover any
tariff with >=10 hours/day low-rate (so 18-/24-hour use that column).

Immersion code mapping CONFIRMED 1=dual, 2=single via RdSAP 10 §10.5
(PDF p.54 — an immersion is "assumed dual" on a dual/off-peak meter)
cross-checked against the API cohort (code 1 sits 3.6:1 on dual meters;
code 2 on single meters). This INVERTS an earlier handover's unverified
"1=single, 2=dual" note — the dual code carries Table 13's small
fraction, matching the cohort over-rating direction; the single mapping
overshot in a prototype.

API SAP eval: 47.6% -> 48.6% within 0.5; <1.0 62.6% -> 63.8%;
mean|err| 1.586 -> 1.561; 909 computed, 0 raises.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
Khalim Conn-Kowlessar 2026-06-08 15:45:41 +00:00
parent d83c431c7d
commit 43d4c67d12
4 changed files with 288 additions and 6 deletions

View file

@ -105,6 +105,9 @@ from domain.sap10_calculator.tables.table_12a import (
tariff_from_meter_type,
water_heating_high_rate_fraction,
)
from domain.sap10_calculator.tables.table_13 import (
electric_dhw_high_rate_fraction,
)
from domain.sap10_calculator.tables.table_32 import (
additional_standing_charges_gbp,
is_electric_fuel_code,
@ -2265,6 +2268,9 @@ def _hot_water_fuel_cost_gbp_per_kwh(
*,
water_heating_code: Optional[int] = None,
inherit_main_for_community_heating: bool = False,
cylinder_volume_l: Optional[float] = None,
occupancy_n: Optional[float] = None,
immersion_single: Optional[bool] = None,
) -> float:
"""Hot water bills at the *water-heating* fuel's rate. When the
water-heating fuel is electric AND tariff is off-peak, bill at the
@ -2278,10 +2284,18 @@ def _hot_water_fuel_cost_gbp_per_kwh(
{901, 902, 914}) and that main is a PCDB Table 362 heat pump, the
HW bills per SAP 10.2 Table 12a Grid 1 WH column (PDF p.191) the
ASHP/GSHP-from-database row carries a 0.70 high-rate fraction at
7-hour and 10-hour, NOT 100% off-peak low rate. Electric IMMERSION
(WHC 903) is a different Table 12a row (off-peak immersion 0.17 /
Table 13) and stays on the 100%-low-rate fallback until that slice
lands.
7-hour and 10-hour, NOT 100% off-peak low rate.
Electric IMMERSION exception (WHC 903): Table 12a's "Immersion water
heater" row (PDF p.191) routes the WH column to Table 13 (PDF p.197).
The Table 13 high-rate fraction a function of cylinder volume,
assumed occupancy and single-/dual-immersion gives the proportion
billed at the high rate, the remainder at the low rate. Without it
the immersion HW billed 100% at the off-peak low rate, under-costing
the dwelling and over-rating it (median +0.98 SAP across the off-peak
WHC-903 API cohort). Needs `cylinder_volume_l` + `occupancy_n` +
`immersion_single`; absent any of them (no cylinder / volume not
resolvable) it falls back to the 100%-low-rate scalar.
`inherit_main_for_community_heating`: per S0380.173, when WHC
{901, 902, 914} AND main is a heat network, ignore the cert-
@ -2306,6 +2320,21 @@ def _hot_water_fuel_cost_gbp_per_kwh(
)
blended = high_frac * high_rate + (1.0 - high_frac) * low_rate
return blended * _PENCE_TO_GBP
if (
water_heating_code == _WHC_ELECTRIC_IMMERSION
and cylinder_volume_l is not None
and occupancy_n is not None
and immersion_single is not None
):
high_rate, low_rate = _tariff_high_low_rates_p_per_kwh(tariff)
high_frac = electric_dhw_high_rate_fraction(
cylinder_volume_l=cylinder_volume_l,
occupancy_n=occupancy_n,
single_immersion=immersion_single,
tariff=tariff,
)
blended = high_frac * high_rate + (1.0 - high_frac) * low_rate
return blended * _PENCE_TO_GBP
return _off_peak_low_rate_gbp_per_kwh(tariff)
if water_heating_fuel is not None:
return prices.unit_price_p_per_kwh(water_heating_fuel) * _PENCE_TO_GBP
@ -4851,6 +4880,17 @@ _CYLINDER_SIZE_CODE_TO_LITRES: Final[dict[int, float]] = {
2: 110.0, 3: 160.0, 4: 210.0
}
# RdSAP `immersion_heating_type` lodgement codes. Code 1 = DUAL immersion,
# code 2 = SINGLE. Confirmed against RdSAP 10 §10.5 (PDF p.54 — an
# immersion is "assumed dual" on a dual/off-peak meter) cross-checked
# with the API cohort: code 1 sits 3.6:1 on dual meters (40 vs 11 single)
# while code 2 sits on single meters (22 single vs 16 dual). This INVERTS
# the unverified "1=single, 2=dual" annotation in an earlier handover —
# the dual code (1) carries Table 13's small high-rate fraction, matching
# the cohort's over-rating direction; treating code 1 as single overshot.
_IMMERSION_TYPE_DUAL: Final[int] = 1
_IMMERSION_TYPE_SINGLE: Final[int] = 2
# RdSAP 10 §10.5 code 7-11: cylinder insulation type. Empirical mapping
# from the ASHP cohort (all 7 certs lodge code 1, worksheet shows
# "Foam" → factory-applied per SAP 10.2 Table 2 Note 2).
@ -5673,6 +5713,21 @@ def _hot_water_cylinder_volume_l(epc: EpcPropertyData) -> Optional[float]:
return _CYLINDER_SIZE_CODE_TO_LITRES.get(size_code)
def _immersion_is_single(epc: EpcPropertyData) -> Optional[bool]:
"""True for a single immersion, False for a dual immersion, None when
the cert lodges no recognised `immersion_heating_type`. Maps the
RdSAP code (1 = dual, 2 = single see `_IMMERSION_TYPE_DUAL`).
None makes the Table 13 high-rate-fraction caller fall back to the
100%-low-rate scalar rather than guess the immersion configuration.
"""
code = _int_or_none(epc.sap_heating.immersion_heating_type)
if code == _IMMERSION_TYPE_DUAL:
return False
if code == _IMMERSION_TYPE_SINGLE:
return True
return None
def _table_3a_combi_loss_default_applies(main: Optional[MainHeatingDetail]) -> bool:
"""Gate for the Table 3a keep-hot 600 kWh/yr fall-through per SAP 10.2
§4 line 7702. Returns True only when the main heating system is in the
@ -7058,6 +7113,9 @@ def cert_to_inputs(
prices,
water_heating_code=epc.sap_heating.water_heating_code,
inherit_main_for_community_heating=_community_hw_inherit,
cylinder_volume_l=_hot_water_cylinder_volume_l(epc),
occupancy_n=wh_result.occupancy if wh_result is not None else None,
immersion_single=_immersion_is_single(epc),
)
hw_co2_factor = _hot_water_co2_factor_kg_per_kwh(
epc, hw_monthly_kwh_for_factors, _rdsap_tariff(epc),

View file

@ -0,0 +1,70 @@
"""SAP 10.2 Table 13 — high-rate fraction for electric DHW heating.
Sourced verbatim from `domain/sap10_calculator/docs/specs/sap-10-2-full-
specification-2025-03-14.pdf`, page 197 (Table 13). RdSAP10 §10.5 (PDF
p.54) routes electric immersion water heating here via Table 12a's
"Immersion water heater" row, whose Water-heating column reads "Fraction
from Table 13".
The table gives the fraction of DHW electricity consumed at the HIGH
rate for a cylinder with a single or dual immersion heater on an
off-peak tariff; the remainder is at the low rate. Note 2 of the table
supplies exact equations equivalent to the tabulated (rounded) grid
this module evaluates those equations, so no floor-area interpolation
is needed:
7-hour tariff (>= 7 hours/day at the low rate)
Dual: [(6.8 - 0.024 V) N + 14 - 0.07 V] / 100
Single: [(14530 - 762 N) / V - 80 + 10 N] / 100
10-hour tariff (>= 10 hours/day at the low rate)
Dual: [(6.8 - 0.036 V) N + 14 - 0.105 V] / 100
Single: [(14530 - 762 N) / (1.5 V) - 80 + 10 N] / 100
where V is the cylinder volume (litres) and N is the assumed occupancy
(Appendix J Table 1b). Per Note 2 the result is clamped to [0, 1]. Per
Note 1 the 10-hour equations apply to any tariff providing at least 10
hours/day at the low rate (so 18-hour and 24-hour use the 10-hour
column). Heat pumps providing water heating only are treated as dual
immersion (Note 1) out of scope of this helper (callers route those
via Table 12a).
"""
from __future__ import annotations
from domain.sap10_calculator.tables.table_12a import Tariff
def electric_dhw_high_rate_fraction(
*,
cylinder_volume_l: float,
occupancy_n: float,
single_immersion: bool,
tariff: Tariff,
) -> float:
"""SAP 10.2 Table 13 (PDF p.197) high-rate fraction for an electric
immersion DHW cylinder on an off-peak tariff.
`single_immersion` selects the single- vs dual-immersion equation
(RdSAP10 §10.5 p.54: an immersion is assumed dual on a dual meter).
The 7-hour tariff uses the 7-hour equations; every other off-peak
tariff (10/18/24-hour, all >= 10 hours low-rate per Note 1) uses the
10-hour equations. STANDARD has no off-peak split and is rejected
callers must early-return before this fires.
"""
if tariff is Tariff.STANDARD:
raise ValueError("Table 13 high-rate fraction is undefined for STANDARD")
v = cylinder_volume_l
n = occupancy_n
if tariff is Tariff.SEVEN_HOUR:
if single_immersion:
fraction = ((14530 - 762 * n) / v - 80 + 10 * n) / 100
else:
fraction = ((6.8 - 0.024 * v) * n + 14 - 0.07 * v) / 100
else:
# >= 10 hours/day at the low rate (10/18/24-hour) — Note 1.
if single_immersion:
fraction = ((14530 - 762 * n) / (1.5 * v) - 80 + 10 * n) / 100
else:
fraction = ((6.8 - 0.036 * v) * n + 14 - 0.105 * v) / 100
return max(0.0, min(1.0, fraction))

View file

@ -3271,8 +3271,9 @@ def test_hot_water_from_pcdb_heat_pump_bills_at_app_n_wh_high_rate() -> None:
# 10-hour. `_hot_water_fuel_cost_gbp_per_kwh` previously billed any
# electric off-peak HW at 100% low rate (its TODO), over-crediting the
# HP-DHW cat-4 cluster. Electric IMMERSION (WHC 903) is a different
# Table 12a row (off-peak immersion 0.17 / Table 13) and must stay on
# the 100%-low-rate fallback here.
# Table 12a row (Table 13) — without the cylinder volume / occupancy /
# immersion-type inputs (not passed here) it falls back to the
# 100%-low-rate scalar; the Table 13 blend is locked separately below.
from domain.sap10_calculator.tables.table_12a import Tariff
from domain.sap10_calculator.rdsap.cert_to_inputs import (
_hot_water_fuel_cost_gbp_per_kwh, # pyright: ignore[reportPrivateUsage]
@ -3305,6 +3306,56 @@ def test_hot_water_from_pcdb_heat_pump_bills_at_app_n_wh_high_rate() -> None:
assert abs(rate_immersion - 0.0750) <= 1e-6
def test_hot_water_immersion_off_peak_bills_at_table_13_blend() -> None:
# Arrange — SAP 10.2 Table 12a (PDF p.191) "Immersion water heater"
# row routes the WH column to Table 13 (PDF p.197). For an electric
# immersion (WHC 903) on an off-peak tariff with a known cylinder
# volume + occupancy + immersion type, the HW bills at the Table 13
# high-rate fraction blend, NOT 100% at the off-peak low rate. A dual
# immersion (small fraction) bills only a little above the low rate; a
# single immersion (large fraction) bills much closer to the high rate.
# Pre-slice both billed 100% at the 7-hour low rate 5.50 p (£0.0550),
# under-costing the dwelling and over-rating it (median +0.98 SAP
# across the off-peak WHC-903 API cohort).
from domain.sap10_calculator.tables.table_12a import Tariff
from domain.sap10_calculator.rdsap.cert_to_inputs import (
_hot_water_fuel_cost_gbp_per_kwh, # pyright: ignore[reportPrivateUsage]
)
from domain.sap10_calculator.tables.table_13 import (
electric_dhw_high_rate_fraction,
)
high_p, low_p = 15.29, 5.50 # Table 32 codes 32 / 31 (7-hour)
n_occupants = 2.7395 # Appendix J Table 1b N at 100 m²
# Act — 110 L cylinder, occupancy N(100), dual (False) vs single (True).
rate_dual = _hot_water_fuel_cost_gbp_per_kwh(
29, None, Tariff.SEVEN_HOUR, SAP_10_2_SPEC_PRICES,
water_heating_code=903, cylinder_volume_l=110.0,
occupancy_n=n_occupants, immersion_single=False,
)
rate_single = _hot_water_fuel_cost_gbp_per_kwh(
29, None, Tariff.SEVEN_HOUR, SAP_10_2_SPEC_PRICES,
water_heating_code=903, cylinder_volume_l=110.0,
occupancy_n=n_occupants, immersion_single=True,
)
# Assert — each rate equals its Table 13 blend; single > dual; both
# strictly above the 100%-low fallback (5.50 p) it replaces.
frac_dual = electric_dhw_high_rate_fraction(
cylinder_volume_l=110.0, occupancy_n=n_occupants,
single_immersion=False, tariff=Tariff.SEVEN_HOUR,
)
frac_single = electric_dhw_high_rate_fraction(
cylinder_volume_l=110.0, occupancy_n=n_occupants,
single_immersion=True, tariff=Tariff.SEVEN_HOUR,
)
expected_dual = (frac_dual * high_p + (1 - frac_dual) * low_p) / 100
expected_single = (frac_single * high_p + (1 - frac_single) * low_p) / 100
assert abs(rate_dual - expected_dual) <= 1e-9
assert abs(rate_single - expected_single) <= 1e-9
assert rate_single > rate_dual > 0.0550
def test_space_heating_pcdb_heat_pump_without_sap_code_bills_at_app_n_high_rate() -> None:
# Arrange — an API-path heat pump resolves via its PCDB Table 362
# index alone (data_source=1, no Table-4a SAP code lodged), so

View file

@ -0,0 +1,103 @@
"""SAP 10.2 Table 13 — high-rate fraction for electric DHW heating.
Locks `electric_dhw_high_rate_fraction` against the published table grid
and the Note-2 clamp at
`domain/sap10_calculator/docs/specs/sap-10-2-full-specification-2025-03-14.pdf`,
page 197. Table 12a's "Immersion water heater" row (PDF p.191) routes
electric immersion DHW here.
The helper evaluates the Note-2 equations, which the spec offers as an
exact alternative to the rounded grid the pins below check that the
equations reproduce the published 2-dp cells.
"""
from __future__ import annotations
from domain.sap10_calculator.tables.table_12a import Tariff
from domain.sap10_calculator.tables.table_13 import (
electric_dhw_high_rate_fraction,
)
# Appendix J Table 1b occupancy N at a few total floor areas (m²) — the
# anchor for the V/N grid cells below. Computed from the same piecewise
# formula the §4 worksheet uses (water_heating.assumed_occupancy).
_N_AT_TFA_100 = 2.7395 # N(100 m²)
_N_AT_TFA_60 = 1.9816 # N(60 m²)
# Table 13 high-rate-fraction grid cells (PDF p.197), keyed by (floor
# area row, cylinder litres column, tariff, single?) → published value.
_GRID_TOL = 0.005 # the published grid is rounded to 2 dp
def test_table_13_dual_immersion_matches_published_grid() -> None:
# Arrange — SAP 10.2 Table 13 (PDF p.197), 110 L cylinder, dual
# immersion. Floor area 100 m² row: 7-hour = 0.18, 10-hour = 0.10.
# Act
seven = electric_dhw_high_rate_fraction(
cylinder_volume_l=110.0, occupancy_n=_N_AT_TFA_100,
single_immersion=False, tariff=Tariff.SEVEN_HOUR,
)
ten = electric_dhw_high_rate_fraction(
cylinder_volume_l=110.0, occupancy_n=_N_AT_TFA_100,
single_immersion=False, tariff=Tariff.TEN_HOUR,
)
# Assert
assert abs(seven - 0.18) <= _GRID_TOL
assert abs(ten - 0.10) <= _GRID_TOL
def test_table_13_single_immersion_matches_published_grid() -> None:
# Arrange — SAP 10.2 Table 13 (PDF p.197), 110 L cylinder, single
# immersion. Floor area 100 m² row: 7-hour = 0.61, 10-hour = 0.23.
# Single immersion carries a much larger high-rate fraction than dual.
# Act
seven = electric_dhw_high_rate_fraction(
cylinder_volume_l=110.0, occupancy_n=_N_AT_TFA_100,
single_immersion=True, tariff=Tariff.SEVEN_HOUR,
)
ten = electric_dhw_high_rate_fraction(
cylinder_volume_l=110.0, occupancy_n=_N_AT_TFA_100,
single_immersion=True, tariff=Tariff.TEN_HOUR,
)
# Assert
assert abs(seven - 0.61) <= _GRID_TOL
assert abs(ten - 0.23) <= _GRID_TOL
def test_table_13_large_cylinder_single_immersion_clamps_to_zero() -> None:
# Arrange — SAP 10.2 Table 13 Note 2 (PDF p.197): "If these formulae
# give a value less than zero, set the high-rate fraction to zero." A
# 210 L cylinder with single immersion on a 10-hour tariff falls below
# zero (the published 210 L 10-hour column is 0), so the helper clamps.
# Act
fraction = electric_dhw_high_rate_fraction(
cylinder_volume_l=210.0, occupancy_n=_N_AT_TFA_60,
single_immersion=True, tariff=Tariff.TEN_HOUR,
)
# Assert
assert fraction == 0.0
def test_table_13_eighteen_hour_uses_ten_hour_column() -> None:
# Arrange — SAP 10.2 Table 13 Note 1 (PDF p.197): the table applies
# "for tariffs providing at least 10 hours ... at the low rate", so an
# 18-hour tariff resolves to the 10-hour equations, not a separate
# column.
# Act
eighteen = electric_dhw_high_rate_fraction(
cylinder_volume_l=110.0, occupancy_n=_N_AT_TFA_100,
single_immersion=False, tariff=Tariff.EIGHTEEN_HOUR,
)
ten = electric_dhw_high_rate_fraction(
cylinder_volume_l=110.0, occupancy_n=_N_AT_TFA_100,
single_immersion=False, tariff=Tariff.TEN_HOUR,
)
# Assert
assert abs(eighteen - ten) <= 1e-9