How to size commercial air conditioning: kW, tonnage and heat load

Why floor area does not work

The most common way commercial air conditioning gets sized badly is a rule of thumb — so many watts or so many pounds per square metre. It fails because floor area is a weak predictor of heat load.

Consider two 400 m² spaces. One is an open-plan office with 40 staff, a full lighting grid, desktop equipment and a south-facing glazed elevation. The other is a warehouse with six staff, high-bay lighting and no windows. Their cooling loads differ by several times. Any single figure that covers both is wrong for both — and the direction of the error is not predictable, which is what makes the shortcut dangerous rather than merely imprecise.

What a real load calculation adds up

A proper calculation works through the heat entering or being generated in each zone.

Occupants. Roughly 100–140 W of sensible heat each at typical office activity, more for physical work. On a 60-person floor that alone is several kilowatts, plus a latent (moisture) component that matters for dehumidification.

Lighting. Essentially all the electrical power drawn becomes heat. This matters less than it once did — an LED relighting scheme can remove a meaningful slice of cooling load as a side effect, which is worth sequencing correctly if both projects are on the table.

Equipment. The item most often got wrong. Use measured draw at the distribution board or PDU, not nameplate ratings — nameplate figures are maximum-possible values and are commonly around double the real draw. Sizing to nameplate is the single most common route to a grossly oversized system.

Solar gain. Depends on glazing area, orientation, glazing specification and shading. On a heavily glazed elevation it can dominate everything else, and it is why zoning should follow the building's aspect rather than its floor plan.

Fabric gain. Conduction through walls, roof and glazing. Large in a poorly insulated building, and the reason a converted mill and a modern unit of the same footprint need different answers.

Fresh air. Ventilation air has to be brought to room condition, and that load is set by the occupancy-driven ventilation requirement independently of the cooling calculation.

Two further refinements matter. Diversity: not every zone peaks simultaneously, so central plant need not equal the sum of zone loads — applied sensibly this saves real money, applied too aggressively it produces a system that fails in the hottest week. And simultaneity: in a multi-aspect building some zones need cooling while others need heating, which is the argument for heat-recovery systems. See commercial air conditioning.

Tons, kilowatts and American sizing charts

A lot of air conditioning information online is American, and American equipment is sized in tons of refrigeration — historically the cooling produced by melting a short ton of ice over 24 hours. The conversion:

1 ton = 12,000 BTU/h = 3.517 kW.

• 3 ton ≈ 10.6 kW
• 5 ton ≈ 17.6 kW
• 10 ton ≈ 35.2 kW
• 20 ton ≈ 70.3 kW

Two warnings. US pricing does not convert — different market, labour rates and regulations — so a dollar-per-ton figure will not give you a useful pound-per-kW one. More importantly, US equipment is frequently sized from rules of thumb per square foot, which is exactly the method this guide is warning against. If you have arrived at a tonnage from an American sizing chart, treat it as a rough sanity check and have the load calculated properly. Our cost guide covers UK pricing in kW.

Why oversizing is not the safe option

There is a persistent instinct that if you are unsure, size up. It is wrong, and understanding why is useful.

An oversized system reaches setpoint quickly, switches off, and switches back on — short-cycling. That wastes energy, because starting is the least efficient part of a cycle. It wears compressors, because start-stop is what damages them. And it dehumidifies badly: removing moisture requires the coil to run long enough for condensation to form and drain, so a system that satisfies temperature in short bursts leaves the space cold and clammy.

Oversized plant also costs more to buy, takes more space, and on A2L refrigerants carries a larger charge — which pushes it toward more frequent F-Gas leak-check bands and can run into BS EN 378 charge limits in small rooms.

The correct treatment of uncertainty is not a bigger machine but better modulation — inverter-driven plant that runs continuously at reduced output — plus designing the plant space so capacity can be added later if the building's use changes.

What to ask a supplier

One question does most of the work: "What heat load did you design to, and how did you calculate it?"

A credible answer describes the inputs — occupancy, lighting, measured equipment draw, glazing and orientation, fresh air — and gives a figure per zone. An answer that reduces to floor area, or to "it is what we normally fit for this size of unit", tells you the design step was skipped. On a multi-zone building the design loads quoted by different bidders diverge far more than their prices do, and that divergence is what you are actually choosing between.

Frequently asked questions

How many kW of air conditioning do I need per square metre?

There is no reliable figure, because load comes from occupancy, lighting, equipment and solar gain rather than from area. A busy glazed office and a sparsely occupied warehouse of identical size can differ several-fold. A per-m² number can sanity-check a budget but cannot size a system, and the direction of the error is not predictable.

What is a 5 ton air conditioning unit in kW?

One ton of refrigeration is 3.517 kW, so 5 tons is about 17.6 kW. Similarly 3 tons is roughly 10.6 kW, 10 tons about 35.2 kW and 20 tons about 70.3 kW. UK suppliers quote in kW, and tonnage figures derived from American per-square-foot sizing charts should be treated as a rough sanity check only.

Is it safer to oversize air conditioning?

No. An oversized system short-cycles — reaching setpoint and switching off repeatedly — which wastes energy, wears compressors and dehumidifies poorly, leaving a space cold and clammy. It also costs more, and on A2L refrigerants a larger charge can push the system into more frequent F-Gas checking bands or run into BS EN 378 room-volume limits.

Why do I need to measure equipment load rather than use nameplate ratings?

Because nameplate ratings are maximum-possible figures and are commonly around double the actual draw. Adding them up produces a heat load far higher than reality, and sizing to it is the most common cause of a grossly oversized system. Measure at the distribution board or PDU instead.

See commercial air conditioning for system selection, installation for how a survey and design run, and costs for indicative UK pricing.