Wednesday 12 January 2022

Will heating your house constantly use more energy?

We are advised when we get a heat pump to change the heating schedule to be constant, or nearly so. This is because heat pumps are efficient when supplying gentle heat but not good at heating a home from cold quickly. This is completely the opposite of what we have learned about keeping our bills low when using a gas boiler. So how much are we currently saving, and how is this different with a heat pump? I have investigated this with a model of a semi detached house (using similar models to my work for BEIS [1]). In the boiler case, savings from intermittent heating are substantial - up to 21%. In the heat pump case, the difference is much less - at most 4%. 

Well insulated homes retain heat, leaky homes cool off quickly.

A well insulated home will retain its heat quite well even when the heating is off, and keeping the heating on all the time does not make that much difference. A poorly insulated home rapidly gets colder when the heating is off and keeping the heating on all the time will use a lot more energy. This chart compares temperatures in a (modelled) semi-detached house with filled cavity walls: the base case with moderate roof insulation and double glazing, the other with draught-proofing, more loft insulation, triple glazing and additional wall insulation. In both cases, the heating is from a gas boiler and is on morning and evening during the week and all day at weekends.


Modelled room temperatures in a house with and without insulation measures: 5 days in February.

Keeping the temperature steady uses up to 20% more heat depending on insulation levels. 

As a house gets cooler, it loses less heat, which is how you get savings. 

Rather than keeping the thermostat setting completely constant, you can compromise by having a high setback temperature. So you might have 20°C for the 'on' periods and 17°C the rest of the time, or 19°C. This chart shows the impact on gas bills of running the same house model with increasing setback temperatures with three different levels of insulation. It is very hard to say what the implications would be for your home as there are many factors to consider. For a very inefficient house, with solid walls, the difference would be even greater. 

Annual gas bill increasing with higher setback temperature. The setback 12°C is the same as none, as the house does not get that cold anyway. In the baseline case the average internal temperature rarely goes below 15°C, even in February. Increasing the setback to 17°C uses 7% more gas and to 19°C the total increase is 21%. In the highly insulated case the increase is only 10%

In the base case, running with a setback temperature of 17°C increases the gas bill by 7% and running at 19°C increases it another 13% - this is just the gas bill of course, assuming you are starting with a gas boiler. (The differences are less dramatic if you heat your home all day, instead of morning and evening.)

In the extreme insulation case, there is almost no difference at the setback temperature of 17°C because the temperature does not generally drop that low anyway. The step up to 19°C increases gas use by 10% but in absolute terms this is very small: about 600 kWh/year which at 4p/unit would be £24. 

What does this mean for running with a heat pump?

A heat pump can heat even quite draughty homes, if you have big enough radiators. You can continue to run intermittently, but you will lose efficiency:

  • Your heating system has to work hardest first thing in the morning, in the winter when it is coldest and your system is least efficient. 
  • You may need to run the radiators hot in order to heat up in a reasonable time. This loses more efficiency – quite possibly 10-20%. 
  • You can turn it on earlier to allow more time to heat up – but this exacerbates the first point.

This chart shows how the peak heating demand comes at the coldest part of the day. The heating comes on at 4am to ensure the house is warm by 6am. With a boiler it would take less time to heat up so you could delay the start until 5am. External temperature and heating have been averaged over all days in February (using typical Cambridge weather). This is with a setback temperature of 16°C which is still higher than you would use with a boiler but really you cannot expect a normal heat pump to heat up from less than this in the morning with any sort of usable efficiency.

Taking an average February day (in Cambridge), allowing the internal temperature to drop overnight, the demand for heating (red) is highest in the early morning when the external temperature (blue) is coldest so the heat pump is least efficient

This chart shows how much electricity is used for heating in the heat pump case with different setback temperatures. Needless to say in this model both heating systems, gas and boiler, heat the house adequately. Overall mean temperatures were very similar. The heat pump has been configured with weather compensation so it uses a higher flow temperature the colder the weather outside. This is standard practice, even for high temperature heat pumps. 

Increasing the setback temperature has a small impact on annual energy consumption - at most 4% from 16°C to 19°C 

In several cases, as the setback temperature increases the energy use decreases. This is because the lower heat demand has triggered a lower flow temperature configuration. In the base case, the lowest setback temperature requires a flow temperature of 45°C and the highest setback temperature needs only 40°C. (This is the flow temperature before weather compensation - when the weather is warmer the flow temperature can be as low as 30°C.) Where there is an increase in electricity use it is quite small compared to the boiler case. Between the blue (setback 16°C) and the red (setback 19°C) the increase in demand is at most 4%, and that is actually for the most insulated case. This is because there are no more savings from reducing the flow temperature. Even then the difference in cost is only about £20/year, assuming electricity is 20p/unit.


Intermittent heating with a gas boiler saves energy. In the example house modelled here, which is fairly typical with double glazing, filled cavity walls and a reasonable amount of roof insulation: up to 20% less. The savings would be even higher for a less efficient home.

However, intermittent heating with a heat pump decreases overall efficiency so that the annual energy saving is much less. In the example house, the maximum increase was 4%, amounting to £20/year.

This is because heat pumps get less efficient when they have to produce high temperature heat and when it is very cold outside. With intermittent heating:
  • hotter radiators are required to provide bursts of heat
  • heat demand is highest first thing in the morning when it is coldest outside.
Improving the insulation and air tightness in your home always reduces overall heat demand and as well as making you more comfortable. 

[1] Cost Optimal Domestic Electrification ( Sep 2021


  1. Great analysis and I suggest there is another variable at play here of the thermal mass of a home.
    The mC product part of Q=m C delta T
    A home with a very low thermal mass would heat up to set point in a morning very quickly and cool down quick when the heating goes off so I’d expect more benefit from using a low overnight set point than for a home with a high thermal mass needing higher temps for longer to bring it back up to temperature.
    Can’t see how thermal mass has been modelled but I’d be tempted to take a typical home at say 20 inside zero outside, measure the steady state kw being consumed then knock off the heating and measure how many degrees it falls inside in say 3 hours.
    We know how heat loss behaves and with estimates for intentional and unintentional ventilation we could work out the mC for a typical home then run it again, with sensitivities either way.
    Lots of people using cheaper overnight and off peak electricity of course, adds another dimension and reasons to warm up while its cheaper.

    1. Thermal mass is indeed important. The increase in heating demand with a setback depends on heat loss and thermal mass. Both of these ought to be on the EPC - they are estimated in the calculation process. See my paper here.

  2. It might be interesting to see what effect turning the heating off at peak load time (1630-1930 ish) has? Are there any options that are noticeably optimal?

    With our PH, insolation is as important as external temperature but sadly is not so easily available as a numerical forecast.

  3. Hi. Wow! What a complex analysis. Please allow me to ask you about one more complexity. Did you assume that the boiler efficiency is the same at all heating powers?

    1. Yes the model does assume constant efficiency for boilers. In practice this depends on the boiler. Some will have poor efficiency at low demand. Also weather compensation would make a difference but this model does not include that for boilers, only for heat pumps.

  4. Hi Nicola; do heat pump controllers actually increase flow temperature when recovering from a setback temperature? I'm not sure they're that smart; I thought they only changed between weather comp curves that should (eventually) give the desired internal temperature, i.e. they don't "boost".

    1. I am told that some heat pumps do both weather compensation and load-compensation. For example Vaillant does, Daikin - with the Madoka controller, Viessmann Nibe. However this is by no means universal. I do not 100% guarantee this. However, the simulations behind this blog did not assume such intelligent controls. It adjusted the weather compensation curve to allow for for the need to reheat from setback.


Comments on this blog are moderated. Your comment will not appear until it has been reviewed.