COP and SCOP: what a heat pump's efficiency figures mean

Every heat pump carries efficiency figures, and they are the numbers most likely to be quoted at you and least likely to be explained. There are two, and they answer different questions.

COP: an instant

The coefficient of performance is the heat delivered divided by the electricity consumed, at one particular moment, under one particular set of conditions. A COP of 4 means four kilowatts of heat for one of electricity.

A COP is meaningless without its conditions, and they are always written alongside it in a form like A7/W35: air at 7 °C, water out at 35 °C, and changing either condition changes the number.

The same machine appears in every row. Nothing about it changed; the job did. This is why a headline COP on a brochure tells you very little, and why the number quoted is invariably A7/W35, which is the kindest condition on the list.

SCOP: a year

The seasonal coefficient of performance averages a whole heating season, weighted by how many hours a typical year spends at each outside temperature. It accounts for the mild autumn days when the machine is barely working and the February fortnight when it is working hard, and for the defrost cycles in between.

SCOP is the honest figure for comparing machines and the right one for estimating a bill. It is also quoted at a flow temperature, and this is where the confusion usually enters.

A SCOP of 4.2 at 35 °C flow and a SCOP of 3.1 at 55 °C flow can describe the same heat pump. Whether your house gets the first or the second depends on your radiators, not on the machine. A quotation offering a SCOP without saying what flow temperature it is at, and without that matching the flow temperature the system is actually designed for, is not telling you what you need to know.

SCOP also depends on climate. European figures come in three climate zones, and the "average" zone, based on Strasbourg, is the one usually published. It is a reasonable proxy for much of Britain.

Where the figures come from

MCS 007, the heat pump product standard, sets out what a unit must demonstrate to be listed, and the testing is done to European standards so that different manufacturers' figures mean the same thing. MCS 026 sets out how SCOP is calculated for the scheme.

The MCS product database publishes output and efficiency at a range of conditions, which is more useful than any brochure. It is a public database, and looking up the specific unit in a quotation takes a minute.

Why installed systems fall short of laboratory figures

Field trials consistently find installed SCOPs below the laboratory figure, and the gap is worth understanding because most of it is avoidable.

Hot water. The seasonal figure is for space heating. Heating a cylinder to 50 or 55 °C is a much larger lift than heating water for underfloor heating at 35, and the weekly pasteurisation cycle above 60 is done partly by immersion heater at a COP of 1. Hot water can be a fifth of a household's heat demand and pulls the annual average down noticeably.

Cycling. An oversized heat pump in a mild autumn cannot turn down far enough to match a small load, so it runs in short bursts. Each start is inefficient, and a machine that cycles heavily can lose a great deal against its rating. This traces straight back to the heat loss calculation.

Flow temperature drift. A system commissioned at 45 °C and later turned up to 55 because one room felt cold has lost a quarter of its efficiency, and often the room was cold for a reason that had nothing to do with the flow temperature.

Auxiliary electricity. Circulation pumps, controls and the fan all draw power. Laboratory figures include some of this; installed systems have more of it, particularly where a buffer tank has added a second pump.

Buffer tanks and low loss headers. Fitted badly, these mix flow and return water, so the heat pump sees a warmer return than the emitters produced and works harder for the same result.

Using SCOP to estimate a bill

The arithmetic is straightforward. Take the heat the house needs in a year, divide by the SCOP, and multiply by the price of electricity.

The heat demand is best taken from what the house currently uses. A house burning 12,000 kWh of gas a year through a boiler running at about 85 per cent seasonal efficiency needs roughly 10,200 kWh of heat. At a SCOP of 3.5 that is about 2,900 kWh of electricity.

Whether that is cheaper than the gas depends on one comparison, and it is the only comparison that matters. Gas at 6p a unit through an 85 per cent boiler costs 7.1p per kWh of delivered heat. Electricity at 25p a unit at a SCOP of 3.5 costs 7.1p. Those two are the same, which is not a coincidence: the break-even SCOP is the electricity price divided by the cost of delivered gas heat, and 25p against 7.1p is about 3.5.

Beat the ratio and the heat pump is cheaper to run. Fall short of it and it is not. Everything else, including whether a house needs new radiators, comes back to whether it moves the SCOP across that line.