How to Size a Solar PV System for a Manufacturing Site

The single most common mistake in manufacturing solar is sizing the array to fit the roof. It feels intuitive: you have a large industrial roof, so you fill it with panels. But a system sized to roof area, rather than to how and when your plant actually draws power, will spill cheap electricity onto the export market at a fraction of the value you would get by using it yourself. Correct sizing starts with your load profile and works outwards from there.

Why you size to baseload, not roof area

A manufacturing site earns the most from every kilowatt-hour it generates and then consumes on site. That is because self-consumed solar displaces grid electricity at your full import rate, currently around 18 to 32p per kWh for industrial users. Any surplus you cannot use is exported and earns only 4 to 15p per kWh under the Smart Export Guarantee. The gap between those two numbers is the whole game. Every unit you self-consume is worth two to five times more than a unit you export.

That is why the working rule for manufacturing PV is to install 70 to 90 percent of peak daytime demand. Size below that band and you leave savings on the table. Size above it and you push generation past what the plant can absorb during daylight, so the extra panels earn export rates instead of import savings and the payback on that marginal capacity collapses.

The good news is that manufacturing has close to the ideal load shape for this. A typical plant runs a strong, daytime-weighted electrical baseload from compressors, motors, process heat, refrigeration, and machinery. Sites on 24/5 or 24/7 shift patterns align especially well with the solar generation curve, and process loads such as compressors, conveyors, and machining absorb peak midday generation exactly when the array is producing the most. That baseload, not the square metres of steel above it, is what determines the right system size.

Self-consumption is the number that matters

The metric to optimise is self-consumption: the share of what you generate that you use on site rather than export. High self-consumption is what drives the payback figures manufacturers actually see, typically 5 to 7 years, and in some sub-sectors faster still.

The reason self-consumption varies so much between sites is the shape and steadiness of the baseload. A food and beverage plant running refrigeration, chilling, and ovens close to 24/7 has an exceptionally high, flat baseload, which is why those sites see some of the strongest self-consumption and the quickest payback of any manufacturing type. You can see the full sub-sector breakdown on our food and beverage manufacturing page. A site that runs a single day shift with the machines off at weekends will export more and self-consume less, so its optimal system is smaller relative to roof area.

Rooftop solar rarely covers all demand at an energy-intensive site. Most installs offset 30 to 60 percent of annual consumption, with the balance covered by continued grid import, a green power purchase agreement, or on-site battery storage. Chasing 100 percent coverage with panels alone is the wrong target; it forces you well past the self-consumption sweet spot.

Use 12 months of half-hourly data

You cannot size correctly from an annual electricity total. An annual figure hides the shift-by-shift and seasonal variation that determines how much solar the site can actually absorb. Two plants with identical annual consumption can need very different systems if one runs a flat 24/7 load and the other concentrates demand into a single weekday shift.

So the non-negotiable input is at least 12 months of half-hourly meter data, modelled as a load profile shift by shift rather than as an annual average. Your energy supplier can provide this for any half-hourly settled meter. With a full year of half-hourly data you can overlay a modelled solar generation curve on your real demand curve, half-hour by half-hour, and read off exactly how much of each day’s generation you would consume. That is what tells you where the 70 to 90 percent sizing band falls for your specific site.

This is also the honest answer to optimistic payback claims. A model built from your actual half-hourly data, rather than from an annual estimate, is one your finance team can stress-test and feed straight into a capital appraisal. Realistic payback for a UK manufacturing install lands between 4.5 and 7.5 years depending on baseload, tariff, and self-consumption, and only half-hourly modelling tells you where in that range you sit.

Roof area per kWp: the constraint check, not the driver

Once the load profile has set your target size, roof area becomes a feasibility check rather than the starting point. In 2026, using 450W-plus panels in portrait orientation with optimised row spacing, plan for roughly 5 to 6 square metres of roof per kW of installed PV.

System size	Approx. roof area needed	Typical annual generation	Common site type
200 kW	1,000 to 1,200 sqm	185,000 to 200,000 kWh	Engineering, chemical, textile
500 kW	2,500 to 3,000 sqm	460,000 to 500,000 kWh	Mid-size manufacturing plant
800 kW	4,000 to 4,800 sqm	740,000 kWh plus	Larger plant, food and beverage

Those roof-area figures assume clear, unobstructed roof. Rooftop plant, parapets, walkways, and adjacent buildings all reduce usable area, sometimes significantly, which is why a 3D shading study is needed to confirm what a given roof can actually hold. If the load profile calls for 500 kW but the usable roof only takes 380 kW, that gap is a genuine design decision, not a reason to shrink the system to whatever fits. Ground-mounted capacity or a phased second array can close it.

Comparing your load-based target against the roof-area ceiling is the point where you find out whether the site is roof-constrained or load-constrained. Most manufacturing sites, with their 1,200 to 12,000 square metre roofs, are load-constrained, which is exactly why sizing to baseload is the right method.

A worked example: a mid-size factory

Take a representative mid-size manufacturing plant running a two-shift, 24/5 pattern with compressors, conveyors, and machining making up a steady daytime baseload. Suppose the half-hourly data shows a peak daytime demand of around 600 kW across the working day, dropping overnight and at weekends.

Applying the 70 to 90 percent rule to that 600 kW peak points to a system in the region of 420 to 540 kW. Rounding to a buildable 500 kW array:

• Roof area required: roughly 2,500 to 3,000 square metres, which sits comfortably inside a typical 1,500 to 4,500 square metre plant roof.
• Annual generation: around 460,000 to 500,000 kWh in UK conditions.
• Self-consumption: because the daytime baseload comfortably exceeds midday generation, most of that output is used on site, keeping self-consumption high and payback in the 6-year region typical for a manufacturing plant.
• Bill impact: for a 500 kW install, expect roughly £45,000 to £90,000 of annual bill reduction, plus modest export income on the small surplus at weekends.

If this same plant added a night shift, the overnight demand would rise and battery storage would start to earn its place, since stored midday solar could then displace expensive evening and night grid import. As a rule, battery storage begins to pay above roughly 250 kW of PV where night shifts run, where DUoS red-band charges are heavy, or where the site wants to trade flexibility. For a straightforward two-shift daytime load, though, the panels alone do the work, and adding a battery would be premature.

Note what did not drive this decision: the roof. The roof confirmed the 500 kW array would fit, but the demand profile chose the size. Had we started from the roof and filled all 4,000-plus square metres with 800 kW, the extra 300 kW would have generated straight into weekend and shoulder-hour export at a fraction of its self-consumed value.

Bringing it together

Correct sizing for manufacturing solar follows a clear order. Start with 12 months of half-hourly data and model it shift by shift. Read off peak daytime demand and target 70 to 90 percent of it. Check that figure against the roof at 5 to 6 square metres per kW, and treat any shortfall as a design question, not a reason to undersize. Only then consider whether a night shift, red-band exposure, or flexibility markets justify adding storage.

Doing it in that order is what separates a system that pays back in 5 to 7 years from one that spills its best value onto the export meter. The exact numbers for your site depend on your load, your tariff, and your roof, all of which we model before quoting. For a full breakdown of installed costs and finance routes see our cost guide, for the grants and allowances that improve the case see grants and funding, and when you are ready for a sized, priced proposal from your own half-hourly data, request a quote.