How we size a manufacturer's solar system from half-hourly meter data

Ask three installers to size your solar system from your roof area and you will get three answers, none of them right. A manufacturer’s array should be sized to its load, not its roof, and the load is revealed by one dataset: at least twelve months of half-hourly meter readings. Here is what that data tells us and why it changes the answer.

What half-hourly data reveals

Your half-hourly data is a reading of your site’s electricity demand every thirty minutes across the whole year. Plotted out, it shows four things a roof survey never could:

• When your demand actually peaks — for most manufacturers, hard between roughly 07:00 and 18:00, exactly when solar generates.
• How flat your baseload is — a continuous 24/5 site with refrigeration or moulding runs almost flat; a single-shift assembly site is peakier.
• How your load moves with the seasons — some sites draw far more in summer (cooling) or winter (heating), which changes the match to solar output.
• Your true self-consumption potential — the share of a given array’s output you would use on site rather than export.

That last figure is the whole game. Self-consumed solar displaces grid electricity at your full import rate of roughly 22 to 32p per kWh; exported solar earns only the Smart Export Guarantee rate, typically 4 to 15p. So the goal is not the biggest array, it is the array that keeps the most generation on site.

Why we install to 70 to 90 percent of peak daytime demand

The working rule for manufacturing PV is to install 70 to 90 percent of peak daytime demand. Below that, you leave savings on the table; above it, you start generating more at midday than you can use, and the surplus spills onto the grid at a low export rate. The exact point in that band depends on the shape of your load, which is why we model it from your data rather than applying a rule of thumb.

A single-shift manufacturer typically ends up self-consuming 30 to 60 percent of annual demand; a continuous 24/5 or 24/7 site reaches 70 to 90 percent because the baseload never switches off. The flatter and more daytime-weighted your load, the larger the array that still makes sense, and the faster it pays back.

From data to a modelled number

Once we have your data, we model the array, its generation across the year, and the resulting self-consumption, then run a full discounted-cash-flow: simple payback, IRR, NPV at your discount rate, and LCOE. You get the model, not just the headline, so your finance team can stress-test every assumption. Only after that do we survey the roof, because the load tells us the target and the roof tells us how to build it.

If you can pull twelve months of half-hourly data from your supplier or meter operator, that is all we need to start. Request a free feasibility study and we will model your site, or read how the numbers work by sector and in the cost guide.