Energy consumption analysis – Harwood Fabrication Ltd

Executive summary

Harwood Fabrication used 30,055 kWh over the 90 days from 1 June to 29 August 2026, which annualises to roughly 121,890 kWh, or about 21.6 tonnes of CO2e a year. At 143 kWh per m² across 850 m², the site's consumption is dominated by a steady overnight draw and a sizeable share of activity outside core weekday hours. The largest single question this data raises is what is running when the factory is closed; agreed capacity is the second, smaller opportunity.

Overnight baseload

The site never drops below roughly 4.4 kW overnight. At the assumed unit rate of 26p/kWh that constant draw represents an estimated exposure of around £10,000 a year. A manufacturing site will always carry some legitimate overnight load: servers, compressors holding pressure, refrigeration, security and emergency lighting. The point is not that 4.4 kW is wrong, it is that nobody has yet attributed it circuit by circuit, so there is no way to say which portion is genuinely required.

Out-of-hours consumption

Around 24% of total consumption falls outside 7:00 to 17:00 on weekdays, an estimated ~£7,500 a year at the assumed rate. Some of that will be shift overrun, cleaning, warm-up and legitimate early starts. The useful exercise is to set your time-clock and building controls schedules against what actually happens on site, then work down from the largest out-of-hours loads rather than trying to tackle everything at once.

Weekend consumption

Weekend days average 125.3 kWh against 414.2 kWh on weekdays, about 30%. That is broadly consistent with a site that shuts down at weekends but keeps essential plant live, and it is largely the same baseload described above. If the half-hourly heatmap shows a repeating weekend block at a particular time, that is worth pulling apart as part of the out-of-hours review.

Agreed capacity vs peak demand

Peak demand over the period was 40.6 kW against an agreed capacity of 120 kVA, leaving roughly 66% headroom. On the face of it, moving toward about 50 kVA while retaining a margin of at least 20% would reduce availability charges by an estimated ~£1,300 a year. Three caveats matter before anyone touches this. Under DCP161, any half-hour that exceeds agreed capacity is billed at penalty rates of roughly three times the standard charge, so the margin has to be real. A capacity change can also shift the site between TNUoS and DUoS charging bands, and that can move costs in either direction. Finally, capacity given up can be expensive or slow to get back if the business adds plant later. This should be quoted and checked with the DNO, not simply requested.

DUoS red-band exposure

Only 10% of consumption falls in the typical weekday 16:00 to 19:00 distribution red band, though exact band times vary by region. That is a favourable shape. It is worth putting in front of suppliers at tender, because a profile weighted away from the expensive delivery hours supports sharper pass-through pricing than a flat-rate quote will reflect.

Data coverage

Ninety days of half-hourly data at 100% interval coverage, with no anomalous days. The findings above rest on a complete dataset.

Recommended next steps

1. Attribute the overnight load: carry out an after-hours walk-round or pull circuit-level data, then put anything non-essential onto 7-day timers or corrected BMS schedules.

2. Compare current time-clock and BMS schedules against actual occupancy, and investigate the largest out-of-hours loads first using the half-hourly heatmap to pinpoint when they occur.

3. Ask the DNO for a capacity-review quote at around 50 kVA, and confirm the TNUoS/DUoS banding effect and the metered kVA maximum demand, including power factor, before committing to any reduction.

4. Cite the low red-band share when the contract next goes to tender, to support a keener pass-through offer.