Drawing No. EH–TH–036 // Thermal Engineering & HVAC
Radiator Sizing Calculator
Reviewed September 2026
Estimate room heat loss, correct radiator output for real water temperatures, and see the catalogue ΔT50 capacity you should select.
01 // Inputs
Radiator operating temperatures
02 // Results
ΔTmean = (Tsupply + Treturn)/2 − Troom.
Radiator correction uses Q/Q50 = (ΔTmean/50)n. The quick heat-loss method is deliberately a screening estimate; use a room-by-room heat-loss calculation for final selection.
Worked example — a 20 m² room
5 × 4 m, 2.6 m ceiling, modern construction, 21 °C against −10 °C, on a 55/45 °C flow and return with a 10% margin.
Step 1 — heat loss. The quick method returns 1046 W for this room. With the 10% margin, the emitter must deliver 1150 W.
Step 2 — mean water temperature above the room. ΔTm = (55 + 45)/2 − 21 = 29.0 K.
Step 3 — correction from the catalogue condition. Radiators are rated at ΔT 50 K under EN 442. The output at any other condition scales as
Φ = Φ50 × (ΔTm / 50)n
with n ≈ 1.3 for a panel radiator. Here (29/50)1.3 = 0.493.
Step 4 — the number you actually order. The calculator gives the required output and the correction factor; divide one by the other to get the catalogue rating:
1150 W / 0.493 = 2333 W at ΔT50
So a room losing about 1 kW needs a radiator catalogued at well over 2 kW.
Flow temperature is the whole story
Hold the room, the heat loss and the margin fixed, and change only the flow and return temperatures:
| Flow / return | ΔTm | Correction | Catalogue rating needed |
|---|---|---|---|
| 75 / 65 °C — traditional boiler | 49.0 K | 0.974 | 1181 W |
| 55 / 45 °C — condensing boiler | 29.0 K | 0.493 | 2333 W |
| 45 / 40 °C — heat pump | 21.5 K | 0.334 | 3443 W |
The same room needs a radiator roughly three times larger on a heat pump than on an old boiler. Nothing about the building changed. This is the single most important fact on the page, and it is why "the radiators are too small" is the most common complaint after a heat-pump retrofit — the radiators were correctly sized for 75 °C water and are being fed 45 °C water.
It cuts the other way too. Keeping the flow temperature low is precisely what makes a condensing boiler condense and a heat pump efficient. A heat pump's COP typically improves by 2–3% for every 1 K reduction in flow temperature. Oversizing the emitters is not waste; it is what buys the low flow temperature, and it pays back every hour the system runs.
Note the flow rate. At 45/40 the water-side ΔT is 5 K rather than 10 K, so the circulation doubles — 198 L/h against 99 L/h for the same duty. Pipework and pump sized for a 20 K drop will not deliver a heat-pump system at 5 K.
Assumptions and where this stops being valid
Arithmetic mean, not log mean. EN 442 defines the reference condition using the arithmetic mean water temperature, and this calculator follows that convention so the result is directly comparable with catalogue data. The true driving difference is logarithmic; at small ΔT the arithmetic mean is slightly optimistic, which is one more reason to keep the margin at low flow temperatures.
The exponent is a characteristic, not a constant. n ≈ 1.3 suits typical steel panel radiators. Column and cast-iron emitters run nearer 1.25–1.3; fan-assisted convectors and low-temperature emitters behave quite differently and the manufacturer's own exponent should be used. It appears in an exponent, so a small error matters more than it looks.
Catalogue ratings assume test conditions. EN 442 output is measured in a standardised test room with free convection. A radiator boxed in behind a sofa, under a deep windowsill or fitted with a decorative cover can lose 10–20% of its output.
No balancing or distribution losses. The figure is the emitter duty for one room. It does not check that the circuit can deliver the flow, that the system is balanced, or that pipe runs are adequate.
The quick heat-loss method is an estimate. For anything beyond a first pass, use a proper room-by-room calculation to EN 12831 with real U-values, thermal bridging and measured infiltration.
Standards and what to do next
Emitter output: EN 442-1 and -2 define radiator testing and the ΔT 50 K reference condition (75/65/20). Any output figure quoted without its ΔT is meaningless — check whether a supplier's number is at ΔT50, ΔT30 or the older ΔT60 basis, because the same radiator carries very different numbers under each.
Heat load: EN 12831-1 for design heat load; CIBSE Guide A for UK design conditions. In North America, ACCA Manual J, with Manual S for equipment and Manual D for distribution.
Low-temperature systems: BS EN 14825 covers seasonal heat-pump performance and shows how strongly efficiency depends on flow temperature.
Next steps: size every room before ordering anything — the largest radiator in the house sets the flow temperature the whole system must run at, so one undersized room forces every other emitter to run hotter. Check the physical space: a 2333 W panel at 600 mm high is roughly 1.4 m long as a double-panel type 22, and it has to fit. Then verify the circuit can carry the flow, and balance the system so each emitter actually sees its design ΔT.
Background & method
Radiator output depends on the temperature difference between the emitter and room. Manufacturer data commonly provides output at a reference ΔT, while low-temperature heat-pump systems need a substantial correction.
FAQ
It is the reference temperature difference commonly used to quote radiator output. Actual output falls as mean water temperature approaches room temperature.
Yes. Enter the lower supply and return temperatures. The correction factor will show why low-temperature systems often need larger radiators.
No. It is a screening estimate. Final emitter sizing should use room-by-room fabric, ventilation and design-temperature heat loss.