Simultaneous Demand Review v2 — gas and electricity

npower D.sites, elec week Fri 11–Thu 17 Sep · SEFE gas pack siteRows, gas week Sun 13–Sat 19 Sep · build validation 21 Sep 2026

Electricity move against gas move, by site

No site lands in the top-right “both rising” zone above both 1 MWh floors. The five single-fuel watch items sit in the side quadrants — one fuel up, the other flat or down. Bubble size = weekly electricity load. Click a bubble for the site detail.

Site detail

Both fuels side by side, with the weather each was scored against and the steps to discharge the watch item.

Select a site from the chart, the table or a finding.

Why the weeks don’t line up

Each fuel comes from its own published pack on its own week and weather basis. This is a declared basis difference, not an error, and no correction is applied.

5 of 7 days shared (Sun 13 – Thu 17 Sep).

Elec scored on ΔCDD18 (live forecast feed); gas on ΔHDD15.5 (ERA5) at each site’s own weather point — thermometers ~1.9 CDD18 apart.

Week-phase artefact: ~5% heating season, >8% shoulder season — comparable to the flag floors.

Size of each fuel’s move — sites in the current filter

Largest combined movers first (up to 12). The three elec-up-only sites move on electricity alone; Dumbarton and Hoddinott Hall move on gas alone, with electricity falling.

Findings (RAG) — Workplace review

GREEN

No site shows both fuels rising above the 1 MWh floor.The v1 flag on White City Place is removed: on the grossed-up comparator (prior week incomplete) its gas falls 2.4% (112.5 → 109.8 MWh), so only electricity rose (+12.9 MWh).

AMBER

Three sites watch-listed for electricity up only (≥2 MWh, gas not up).White City Place +12.9 MWh, Birmingham Mailbox +2.5 MWh, Cardiff Central Square +2.4 MWh. None has a matching gas rise; Cardiff gas is flagged Investigate for a 2.6 MWh fall. White City gas is gross of the ~55% on-sale.

AMBER

Two sites watch-listed for gas up only (≥2 MWh, electricity not up).Hoddinott Hall gas 5.2 → 9.1 MWh (+3.9, +75.0%) while Cardiff got milder (ΔHDD −0.7): residual +4.1, flagged Investigate and on the sustained-step watchlist (×2.2 base). Electricity fell 0.7 MWh. The gas pack reads this as plant running when it need not be. Dumbarton gas 0.9 → 4.0 MWh (+3.1, residual +2.7, Investigate) while electricity fell 1.4 MWh, in its last weeks before exit on 30 Sep.

AMBER

Newcastle and Plymouth are the stronger of the below-floor signals.Both are on the gas pack’s sustained-step watchlist (Newcastle ×2.3 base, resid +5.0; Plymouth ×2.1 base, resid +2.2) and electricity also rose. A watchlisted site with electricity rising carries more weight than two small week-on-week moves.

AMBER

Basis difference limits what any flag can prove.Different weeks and thermometers, and a week-phase artefact of the same order as the floors. Week-level coincidence is not simultaneity: half-hourly profiles and BMS trends are needed to discharge any item.

AMBER

Maida Vale and Dumbarton exit on 30 Sep — falls from here are wind-down, not savings.Maida Vale (gas −9.5%, elec −3.2%) and Dumbarton (elec −21.9%) may already show the slow-down. The meters are expected to keep reporting into October: report that separately as post-exit consumption to support closing reads and de-registration, and keep both sites out of baselines and like-for-like totals.

AMBER

Hull pairing unverified.Hull (E: BBC Radio Humberside, MPAN 2300001090040 “BC Radio Humberside”) is present in the 11–17 Sep npower CSV, but the load shape against peers is unresolved. Resolve against FM Schedule 3 and the MPAN register.

Paired screen — detail

Click a column to sort, a row to open the site. Gas prev is the comparator used: grossed-up where the prior week was incomplete (White City, York), otherwise as metered. ΔHDD is each site’s own weather point. “—” for Swindon/Derby gas WoW is the pack’s null (0.0 → 0.0), not 0%. Δ MWh derived from the published one-decimal figures.

Television Centre and Studio 1-3 NOT PAIRED

Television Centre is leased by BBC Studios; BBC Workplace covers its energy costs, so the gas is real demand. The gas pack’s NO_MODEL reason (“Site exited Jan 2025; still invoiced by SEFE — metering, not demand”) is wrong and needs correcting there.

Studio 1-3 is a separate building on the Television Centre site (not W12 Media Village) and may have its own supplies, so its electricity (47.4 → 55.4 MWh, +16.9%) is not paired with Television Centre gas (4.9 as metered / 5.66 grossed-up → 5.3 MWh). The Television Centre gas pairs only with the electricity supply to the same building — possibly the ~33.8 MWh/week line — once the MPAN/MPRN register confirms it.

Pacific Quay EXCLUDED

Main heating and cooling both come from four air-source heat pumps. Gas serves only the restaurant and the ancillary building, so any simultaneous heating and cooling happens inside the electricity and cannot show in a gas-against-electricity screen.

This week: elec 175.1 → 171.6 MWh (−2.0%); gas 3.4 → 3.6 MWh (+5.9%, no model, immaterial). The check for Pacific Quay is ASHP operating mode from the BMS.

Coverage — 11 elec-only and 15 gas-only sites outside the screen

Elec-only (11)

Gas-only (15)

Counts recomputed from the 48 D.sites and 52 siteRows: 35 screened + Pacific Quay + Television Centre + 11 elec-only = 48; 35 + 2 + 15 gas-only = 52. Pairs are made at building level only, where both fuels serve the same building. Studio 1-3 is a separate building at Television Centre and stays elec-only unless it has its own gas supply. Radio Lancashire / Blackburn is paired on name and location — confirm on the register. Most material gas-only sites (Elstree, Perivale, Middlesbrough, Bangor, Chelmsford, Truro) have electricity outside the 48 displayed npower sites: resolve on the MPAN/MPRN register. Salford is campus-level: 1× HV electricity against Bridge House / Dock House / Quay House gas.