Heating and Cooling Loads, and the Emitters That Meet Them

Area times a factor is a screening rule, not a load calculation

A load calculation adds up every path by which heat crosses the envelope: conduction through each wall, roof, floor, window and door at its own transmittance; the air leaking in and having to be conditioned; solar gain through each glazed opening by orientation and shading; and the heat given off by people, lighting and equipment inside.

A rule of thumb replaces all of that with one number per unit of floor area. It works as a screening tool because floor area correlates with envelope area in typical construction — and it fails whenever a building is not typical. Two houses of identical floor area, one with modern insulation and small windows and one with solid walls and a south-facing glass wall, can differ by a factor of three in what they need.

So the honest description of most pages in this family is a first pass. They are the right tool for asking whether a quotation is in the right region, for comparing two options, or for a feasibility budget — and the wrong tool for buying equipment, which is what those pages say.

Oversizing a cooling system makes a room clammy

Cooling equipment removes heat in two forms. SENSIBLE cooling lowers the air temperature; LATENT cooling condenses moisture out of it on the cold coil. Comfort needs both, and the second one needs TIME.

An oversized unit satisfies the thermostat quickly, shuts off, and starts again — short cycling. The coil spends its short runs getting cold and barely any time condensing, so the room reaches temperature with its humidity almost untouched. The result is the familiar complaint of a space that is cold and still feels damp, and the instinctive response of turning the thermostat down makes it worse.

So oversizing is not a safe error in cooling the way it is in most of engineering. It costs comfort, it costs efficiency through cycling losses, and it shortens compressor life. A correctly sized unit runs for long periods at design conditions, which is what it is supposed to do.

The same logic sets the limit on how much a variable-capacity machine helps. An inverter-driven unit can modulate down and run continuously at part load, which is exactly the right behaviour — but only within its turndown range, and a badly oversized inverter unit hits the bottom of that range and starts cycling again.

Heating and cooling loads are different shapes

A heating load is dominated by the terms driven by temperature difference: fabric transmission and infiltration. It peaks at night in the coldest weather, when there is no sun and few internal gains.

A cooling load is dominated by the terms that have nothing to do with outdoor temperature: solar gain through glass, people, lighting and equipment. It peaks in the afternoon, it depends heavily on ORIENTATION, and in a well-insulated modern building the internal and solar gains can exceed the fabric term entirely.

Two consequences follow. The same improvement does not help both: adding insulation cuts a heating load substantially and a cooling load much less, while shading cuts a cooling load and does nothing for heating — and can hurt it by removing winter solar gain.

And a single machine doing both is sized for the larger of two loads it meets at different times of year. Where they differ substantially, the equipment is correctly sized for one season and oversized for the other, which is the practical reason a heat pump in a cooling-dominated climate can be poor at heating and the reverse.

Design conditions are percentiles, not records

Every load calculation needs an outdoor design temperature, and the published values are not the coldest or hottest ever recorded. They are PERCENTILES — a temperature exceeded a stated small fraction of the hours in a year, conventionally around one per cent for cooling and one per cent at the cold end for heating. A UK heat pump design under MCS goes further at the cold end: MIS 3005-D takes the temperature the air stays above for 99.6% of the hours in a year, with no uplift for intermittent heating, or the 99% figure, which it does not exempt from one.

That is a deliberate economic choice. Sizing to the record extreme means sizing for a handful of hours a decade and being oversized for every other hour, with all the cycling penalties in the previous sections. Sizing to a percentile accepts that the design condition will be exceeded occasionally and that the building will drift a degree or two when it is.

It also means a design temperature is a LOCAL figure taken from a published station, and using a national or regional value in a place with a different microclimate is a real error. Elevation, coastal proximity and urban heat island effects all move it.

Safety factors applied on top of a percentile design condition are where oversizing creeps in. Each individual allowance looks modest — a little on the temperature, a little on the area, a round-up to the next equipment size — and they multiply rather than add.

An emitter's rating belongs to a temperature difference

A radiator, a baseboard convector or a fan coil is rated at a stated temperature difference between the water in it and the air around it. Its output is not a property of the unit alone, and it falls steeply as that difference narrows — not linearly, but to a power somewhat above one.

This is the arithmetic behind the most common failure in heat pump retrofits. Emitters sized for a boiler running at 70 or 80 °C (158 or 176 °F) are rated at a large temperature difference; run the same emitters at the 35 to 45 °C (95 to 113 °F) a heat pump wants, and their output falls to about a THIRD of their rating at 45 °C (113 °F) and a seventh of it at 35 °C (95 °F). Worked for a steel panel radiator, exponent 1.3, in a 21 °C (69.8 °F) room: water flowing at 70 °C (158 °F) and returning at 60 °C (140 °F) sits 44 K (79.2 °F) above the room and gives 85% of the radiator's ΔT50 rating. At 45 °C (113 °F) flow and 40 °C (104 °F) return it sits 21.5 K (38.7 °F) above the room and gives 33%, which is 39% of what it gave the boiler. At 35 °C (95 °F) and 30 °C (86 °F) it sits 11.3 K (20.4 °F) above, taken as EN 442's logarithmic mean, and gives 14%, which is 17%. The room that was warm is now not, and nothing about the heat pump is faulty.

The remedy is more emitter, not hotter water, because raising the flow temperature is what destroys the heat pump's efficiency. A low-temperature retrofit therefore means larger radiators, more baseboard length, or a different emitter type — and the length calculation is done at the ACTUAL design flow temperature rather than at the manufacturer's headline rating.

The same relationship explains why baseboard output is quoted per unit of length at a stated water temperature, and why a room with a long external wall and a low flow temperature can run out of wall before it has enough emitter.

A radiant floor is capped by comfort, not by water temperature

Underfloor heating delivers heat from the whole floor, so its output per square metre is modest and its limit is not the boiler or the pipe. It is the FLOOR SURFACE TEMPERATURE, which is capped for comfort and for the floor covering — conventionally around twenty-nine degrees in occupied areas, a little higher in perimeter zones and bathrooms.

That cap sets a maximum output per square metre regardless of anything upstream. A room whose heat loss exceeds it cannot be heated by its floor alone however hot the water is, which is why underfloor heating is paired with well-insulated construction and why a poorly insulated room with a large glazed wall usually needs a supplementary emitter.

Below the cap, pipe SPACING sets the output density, exactly as it does for an electric cable: closer pipes deliver more per square metre and also a more even surface temperature. Wider spacing shows as warm and cool stripes underfoot before it shows as insufficient heat.

The floor covering is the other variable and it is chosen by someone else. A covering's thermal resistance sits between the pipe and the room, so a thick carpet or an underlay can require a materially higher water temperature for the same room output — or, at a fixed water temperature, simply reduce the output. It is worth deciding before the system is designed rather than after.

A fan does not cool a room

A ceiling fan removes no heat from a space. It moves air, and the moving air increases convective and evaporative heat loss from SKIN — so it makes people feel cooler while leaving the room's temperature exactly where it was. Its motor, in fact, adds a small amount of heat.

The useful consequence is that a fan lets the thermostat setpoint rise by two or three degrees for the same comfort, and that saves real cooling energy. The useless consequence follows immediately: a fan running in an empty room is doing nothing at all except consuming electricity and warming the space slightly.

Fan sizing is therefore a rule about air movement over an occupied area rather than a load calculation, and the quantity being matched is blade span against room dimension so that the airflow covers where people are. It cannot be compared with, or substituted into, any of the capacity figures elsewhere on this page.

What replaces the rule of thumb

The standard alternative is a room-by-room load calculation done from the actual construction: each surface with its own transmittance and area, glazing by orientation with its solar heat gain coefficient and shading, a measured or assumed air change rate, and internal gains by occupancy. It is a day's work for a house and it is what equipment should be selected against.

For an existing building there is a better input than any calculation: its own history. Metered fuel consumption over a heating season, or the measured RUN TIME of the existing equipment at a known outdoor temperature, gives the building's actual load including every defect and deviation a survey would miss.

That route also avoids the commonest sizing error in replacement work, which is copying the nameplate of the unit being removed. The old unit was very possibly oversized, and repeating its capacity repeats the mistake — often compounded, because a round-up is applied again.

Most calculators here are screening tools and say so on their pages. They are for asking whether a proposal is plausible before the room-by-room calculation is commissioned, and for comparing options; none of them is a substitute for that calculation, and none of them should be the basis of a purchase. Two are not screening tools. The Room Heat Loss page works one room the way BS EN 12831-1 does, and the Radiator Output (ΔT Correction) page sizes an emitter to a loss you supply. Neither is the whole-house calculation a boiler or a heat pump is selected on.