Context: a street-sized test, not a citywide promise
Bananzia’s winter peak pilot began on a single row of 38 terraced homes built in the 1970s, where gas boilers were reaching end of life and the local substation had only 60 kVA of spare capacity. The aim was narrow: cut winter electrical peaks while keeping living rooms at 20–21°C. That constraint shaped every decision, because a heat-pump rollout that ignores feeder limits simply moves the bottleneck from gas pipes to cables.
Step 1: Survey fabric before selecting hardware
The field team used thermal imaging and blower-door tests on all 38 homes before quoting a single heat pump. Twenty-two lofts lacked insulation, and 14 homes had draughts around suspended floors. Sealing those gaps and topping up loft insulation reduced measured heat loss by 18% on average, which allowed smaller units later. The takeaway is practical: fabric work is not a preliminary nicety, it is the cheapest way to shrink the peak a heat pump must serve.
Step 2: Size heat pumps room by room
Installers replaced boiler ratings with room-by-room heat-loss calculations and set most units at 5–7 kW rather than the 8–12 kW boilers they displaced. They also kept flow temperatures at 45°C by upsizing 63 radiators across the row. The result was fewer compressor starts and steadier demand; the pilot recorded a 31% drop in short-cycling events compared with three oversized trial installations. Sizing to actual demand, not to a boiler’s nameplate, is the step that protects both comfort and grid headroom.
Step 3: Shift the peak with thermal storage
Twelve homes received 200-litre buffer tanks and simple timers that preheat water between 01:00 and 04:00. That shifted roughly 1.8 kWh per home per day out of the 17:00–19:00 peak. The monitoring dashboard, labelled bananzia, tracked flow temperatures and run times across the street so residents could compare settings. The takeaway: storage does not need to be exotic; a well-insulated cylinder and a clear schedule can move demand by 90 minutes.
Result: lower peaks without colder rooms
Across the pilot winter, metered peak electricity per home fell 23% against a heat-pump-only control group, while average indoor temperature held at 20.5°C. Two homes needed supplementary radiator upgrades after the first cold snap, and the team fixed both before January. The result is modest but repeatable: survey first, size tightly, and coordinate storage with the grid. For streets facing similar constraints, Bananzia’s pilot offers a concrete sequence rather than a broad slogan.
