Audit check: lodged cylinder with unresolvable insulation drops storage loss

Portfolio 814 / scenario 1271 audit (2026-07-03): all 26 properties lodging
has_hot_water_cylinder with insulation type 0 (none) or NULL over-rate their
effective baseline by up to +18 SAP (avg +12.7) because
_cylinder_storage_loss_override returns None for any insulation label it
can't resolve and the worksheet keeps the zero-storage-loss combi default.
Same-lodged neighbours split by 19 effective points (WD5 0DP pids
742215/742216, both lodged 54, effective 50 vs 69) and the inflated members
distorted neighbour-divergence cohort medians (E17 6EB pid 742355 Δ+18).

New HIGH check cylinder-storage-loss-dropped fires on the deterministic
input pattern (cylinder present, insulation type ∉ {factory, loose jacket}
or thickness missing) — 28/338 on the motivating portfolio (the 26 above
plus 2 same-pattern rows with no lodged score). The calculator fix
(uninsulated Table 2 loss, or RdSAP Table 29 age-band default) is separate
follow-up work.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Jun-te Kim 2026-07-02 15:27:15 +00:00
parent cf50c53cad
commit 318a859e88

View file

@ -55,6 +55,26 @@ def _band_of(score: Optional[float]) -> Optional[str]:
return Epc.from_sap_score(round(score)).value
def _int_or_none(value: object) -> Optional[int]:
"""Coerce a jsonb-backed scalar (int / digit-string / None) to int."""
if value is None:
return None
try:
return int(str(value))
except ValueError:
return None
def _float_or_none(value: object) -> Optional[float]:
"""Coerce a jsonb-backed scalar (number / numeric-string / None) to float."""
if value is None:
return None
try:
return float(str(value))
except ValueError:
return None
class Severity(IntEnum):
LOW = 1
MEDIUM = 2
@ -89,6 +109,12 @@ class PropertyAudit:
solar_sap_points: Optional[float] # max SAP a single solar_pv measure earns
solar_bill_savings: Optional[float] # the solar_pv measure's £/yr bill saving
n_measures: int
# Hot-water-cylinder fields from the latest lodged EPC row (None when the
# Property has no epc_property row). Insulation type codes are the RdSAP
# lodgement codes: 0 = none, 1 = factory foam, 2 = loose jacket.
has_hot_water_cylinder: Optional[bool]
cylinder_insulation_type: Optional[int]
cylinder_insulation_thickness_mm: Optional[float]
@dataclass(frozen=True)
@ -342,6 +368,54 @@ def _negative_bill_savings(a: PropertyAudit) -> Optional[str]:
return f"energy_bill_savings £{a.energy_bill_savings:.0f}/yr on £{a.cost_of_works:.0f} of works"
@check("cylinder-storage-loss-dropped", Severity.HIGH)
def _cylinder_storage_loss_dropped(a: PropertyAudit) -> Optional[str]:
"""A lodged hot-water cylinder whose insulation fields cannot resolve a SAP
10.2 Table 2 loss branch insulation type is not factory foam (1) or loose
jacket (2), or the thickness is missing so the calculator's
``_cylinder_storage_loss_override`` (domain/sap10_calculator/rdsap/
cert_to_inputs.py) returns None and the worksheet keeps the zero-storage-loss
default meant for combi / instantaneous systems. The baseline then
OVER-RATES: an UNINSULATED cylinder (type 0, the worst case SAP knows) scores
~+13 SAP better than a properly insulated one instead of much worse.
Provenance: portfolio 814 / scenario 1271 audit (2026-07-03). All 26
properties with ``has_hot_water_cylinder`` and insulation type 0 (23) or NULL
(3) over-rate vs their lodged score Δ+1..+18, avg +12.7 for type 0 while
the type-1/2 population averages Δ2. Same-lodged neighbours split by up to
19 effective points (WD5 0DP: pid 742216 lodged 54 effective 69 vs pid
742215 lodged 54 effective 50, identical building parts), and the inflated
members distorted neighbour-divergence cohort medians (E17 6EB pid 742355
Δ+18; WD5 0DP pushed honest pid 742210 over the flag threshold). Fix is in
the calculator (apply the uninsulated Table 2 loss for type 0, or the RdSAP
Table 29 age-band default when unknown), not in the data.
Fires on the input pattern (deterministic the loss IS dropped for every
such row), not on a tuned delta threshold; the reason string reports the
effective-vs-lodged gap where a lodged score exists so triage can rank by
materiality. A full-SAP cert lodging a manufacturer's declared loss instead
of type/thickness would bypass the bug and false-positive here none exist
on the motivating portfolio; revisit if one appears."""
if a.has_hot_water_cylinder is not True:
return None
if (
a.cylinder_insulation_type in (1, 2)
and a.cylinder_insulation_thickness_mm is not None
):
return None
gap = (
f", effective {a.effective_sap:.0f} vs lodged {a.lodged_sap:.0f} "
f"{a.effective_sap - a.lodged_sap:+.0f})"
if a.effective_sap is not None and a.lodged_sap is not None
else ""
)
return (
f"cylinder lodged with insulation_type={a.cylinder_insulation_type} "
f"thickness={a.cylinder_insulation_thickness_mm} — storage loss dropped "
f"to zero (combi default), baseline over-rates{gap}"
)
# ─────────────────────── runner ───────────────────────
_QUERY = text(
@ -351,12 +425,24 @@ _QUERY = text(
pbp.lodged_sap_score, pbp.lodged_epc_band,
pbp.effective_sap_score, pbp.effective_epc_band, pbp.rebaseline_reason,
pl.post_sap_points, pl.post_epc_rating, pl.cost_of_works,
pl.energy_bill_savings, pl.energy_consumption_savings
pl.energy_bill_savings, pl.energy_consumption_savings,
ep.has_hot_water_cylinder, ep.cylinder_insulation_type,
ep.cylinder_insulation_thickness_mm
FROM property p
LEFT JOIN property_baseline_performance pbp ON pbp.property_id = p.id
LEFT JOIN plan pl ON pl.property_id = p.id AND pl.is_default = TRUE
AND (:scenario_id IS NULL OR pl.scenario_id = :scenario_id)
LEFT JOIN scenario s ON s.id = pl.scenario_id
LEFT JOIN LATERAL (
SELECT e.has_hot_water_cylinder,
e.heating_cylinder_insulation_type AS cylinder_insulation_type,
e.heating_cylinder_insulation_thickness_mm
AS cylinder_insulation_thickness_mm
FROM epc_property e
WHERE e.property_id = p.id
ORDER BY e.id DESC
LIMIT 1
) ep ON TRUE
WHERE (:portfolio_id IS NULL OR p.portfolio_id = :portfolio_id)
AND (:property_id IS NULL OR p.id = :property_id)
ORDER BY p.id
@ -479,6 +565,13 @@ def _load(
solar_sap_points=solar_sap,
solar_bill_savings=solar_bill,
n_measures=n_measures,
has_hot_water_cylinder=m["has_hot_water_cylinder"],
cylinder_insulation_type=_int_or_none(
m["cylinder_insulation_type"]
),
cylinder_insulation_thickness_mm=_float_or_none(
m["cylinder_insulation_thickness_mm"]
),
)
)
return out