Heat loss calculations for heat pumps: what they are, why they matter, and where most installers go wrong

Heat loss calculations: what they are, why they matter, and where most installers go wrong

Every number you hand a customer traces back to one calculation. Emitter sizes, flow temperature, the capacity of the unit, the efficiency you've promised, the running cost they'll quote back at you in February, all of it is derived from the heat loss. It's the single piece of evidence the whole design stands or falls on, which is why a survey rushed to fit a Thursday afternoon is expensive in ways that don't show up until commissioning.

What the calculation actually works out

A heat loss calculation to BS EN 12831-1:2017 answers one narrow question, room by room: how much heat, in watts, does this space need to hold its design temperature on the coldest design day?

Two sources of loss, added together. Fabric losses run through every external element: walls, windows, roof, floor, and the surfaces touching unheated spaces, as area multiplied by U-value multiplied by the temperature difference between inside and out. Ventilation losses cover the air leaving the room and the cold air replacing it, driven by the air change rate you assign. Thermal bridging and ground contact get their own treatment.

Room-by-room heat loss is the output that matters. The whole-house figure drives heat pump sizing; the per-room watts size the emitters. Both have to be right, and they fail independently. This is a peak load in watts at design conditions, not annual energy in kilowatt hours, two different numbers that get muddled constantly, usually by whoever is comparing a design against an EPC.

Where the number goes wrong

Assumed U-values. Age-band defaults are a starting point, not a survey finding. Take a 20 m² solid wall in a living room at 21°C with a design outside temperature of -2°C. At an assumed 2.1 W/m²K it loses about 970 W. If that wall was internally insulated during a previous refurb and actually performs at 0.6, the real figure is nearer 280 W. One wall, one wrong assumption, and a radiator two sizes too big sitting under the window for the next fifteen years.

The error runs both ways, which is the part worth remembering. Assume insulation that isn't there and the room never gets warm.

Surveys done at speed.

Measuring off an estate agent's floor plan. Not going into the loft. Recording "double glazing" without noting that the units are early nineties and performing nothing like a modern one. Missing that the rear extension is a different construction entirely from the original house. None of these look like errors on the day; all of them land in the calculation as fact.

Ventilation guessed rather than observed. Air change rates are where a lot of quiet inaccuracy lives. An open chimney in a room changes its ventilation loss substantially, and a sealed-up flue behind a plasterboard panel is a very different room from an open one. It takes ten seconds to check and it's rarely in the notes. Same goes for the extract fan someone specified and nobody fitted.

Why the building total isn't the sum of the rooms

Here's the bit worth knowing, because it's where well-meaning caution does damage. Each room is calculated on a worst-case basis, including maximum wind exposure. The wind doesn't hit every elevation at once, and heated rooms transfer heat between themselves, so the building's design heat load comes out lower than the sum of its rooms.

Add a safety margin on top of that and you're padding a figure that already contains its own margin. On a boiler, oversizing was forgiving. On a heat pump it means cycling, a worse seasonal efficiency than you designed for, and a customer whose bills don't match the estimate you gave them.

Why everything downstream inherits the error

The per-room watts feed straight into emitter selection, and radiator output at a heat pump flow temperature is a long way from the catalogue figure. Get the load wrong and every radiator is wrong in the same direction before you've picked a single one.

It's also the reason a full heat loss survey isn't an upsell. The calculation is part of your MCS documentation, it's the thing an auditor reads first, and it's the honest answer when a customer asks why you can't just size from the old boiler.

Book a demo and see a room-by-room heat loss run from postcode to result, with the U-values and ventilation assumptions visible at every step.