I am aware that heat pumps are able to move more heat into a building per unit of energy input (the "coefficient of performance" or COP) when the source (outside) temperature is higher. My question is whether I should set my programmable thermostat to bring the house to a warmer temperature during the warmer afternoon, or whether this increase in efficiency will be outweighed by the additional heat loss during the extra hours of increased temperature differential with the outside.

I have a two-zone heat pump system. For the upstairs I currently have the thermostat set to increase the temperature by about 4 degrees a few hours before bedtime; this is about 3 hours after sunset. I also have the downstairs thermostat set to increase the temperature by a few degrees at sunset.

I noted recently how quickly the outside temperature falls after sunset on these sunny but colder days. Since the heat pump COP is higher when it is warmer outside, would I use less energy by running the heat pump to create the temperature increases before sunset rather than later when it is colder?

For concrete numbers, let's assume that the temperature is 10 degrees Fahrenheit colder three hours after sunset than 1 hour before sunset. (On a warm winter day, it would be 50 in the afternoon and 40 after dark; on a colder day it might be 10 or 20 degrees colder in both instances.) Let's also use 64 degrees inside for computing the heat loss, and assume I am raising the temperature by 4 degrees 4 hours earlier.

If there is a significant savings possible, I could also raise the temperature by even more during the day to delay the time when the house cools to the nighttime indoor setting and starts running the heat pump at the less efficient nighttime temperatures.

  • 2
    Is this an air-source heat pump, or some other source (ground, water)? I think from your question it must be air-source, but wanted to check first.
    – 410 gone
    Dec 2, 2013 at 6:23
  • 1
    Yes, it is air-source, otherwise as you and Flyto point out there would not be a large source temperature difference from day to night. Dec 2, 2013 at 13:58

2 Answers 2


Strictly speaking, your heat pump's CoP is not affected by the outside temperature, but by the difference between inside and outside temperatures. Given that there is a fairly narrow range of interior temperatures that we might want, that is nearly equivalent, but it does mean that there is only a limited extent to which this "pre-heating" approach can work - once you warm the house a certain amount above your normally preferred temperature, the CoP of the pump will be reduced in the same way as it is when it is cold outside.

Ultimately, as you have identified, whether this strategy is worthwhile will depend very much upon the house and the heat pump: if the house is very well insulated, and thus heat loss from it is small, then the suggested scheme ought to work. If not, then the additional heat loss from the house being at a higher temperature may be greater than the benefit gained from the higher CoP. It may be possible for others to give a rule of thumb or an answer from experience, but the only way that I think you can sure is to test it, and even that is tricky:

For a period of time (perhaps a week) try this "pre-heating" strategy. Record the exterior temperature that day and evening, and the amount of energy used by your heating system, each day. For another similar period of time use a more conventional strategy (have the heating on when you want the warmth), making sure that the thermostat is set to the same level at the same time except for the pre-heat period. Record the same items.

Identify similar days between the two data sets, and compare the energy use.

Difficulties that may be faced:

  1. Depending on your metering arrangements, isolating the energy used by the heating system from other load in the house.
  2. If the house is well-insulated, having the same exterior temperature on two days does not necessarily mean that the house starts at the same interior temperature.

EDIT: I've assumed an air-source heat pump. As Energynumbers has noted, if it's ground source then this strategy is almost certainly unnecessary, as the ground temperature should change little on an hour-to-hour timescale.

  • I was being intentionally sloppy with my wording when I only referred to the outside temperature and not the difference, but +1 for pointing out the limits this puts on raising the inside temperature. Dec 2, 2013 at 14:01
  • Luckily, I do have an energy monitor with channels for the HVAC system so I can attempt to measure a difference. The difficulty, as you say, will be determining equivalent days. It also occurs to me that even if this scheme only breaks even on energy use there is a benefit to having more comfortable temperatures for more of the day. I am still hoping someone will be able to answer with actual heat pump COP curves. Dec 2, 2013 at 14:05
  • You'd need both the COP curves and a (perhaps back-of-envelope) calculation of heat loss for your house. But yes, if somebody can come up with figures I'd be interested!
    – Flyto
    Dec 2, 2013 at 14:48
  • Well with the energy monitor I should be able to determine the heat replacement rate (actually just how long the heat runs, I'll still need COP to convert to actual heat loss). Insolation is difficult to control for but I could crunch the numbers for the overnight hours if I find an accurate temperature profile. Dec 3, 2013 at 2:47
  • Oh, good point :-)
    – Flyto
    Dec 3, 2013 at 7:57

Yes it is more efficient, but there may be better ways. As Flyto comments, there is a limit to how warm you run the house. You don't want to live in a Heat Battery.

If you assume a 10 degree C fluctuation in temperature outside, and an average of 30 C colder outside than inside, then yes, a chunk of heat pulled in during the warm part of the day would cost about 1/3 less.

It may be effective as a co-strategy to run the house cooler. If you are willing to wear a sweater during the morning chill, you will have more 'room' to warm the house during the warm part of the day. It may also be effective to reduce heat to some parts of the house. If you don't have zoned heat, you can put old phone books on top of the heat registers.

Increasing the thermal mass in your house can also help. This gives you more heat per degree of temperature increase. It is difficult to retrofit large amounts of mass into an existing structure, but if your house is already slab on grade, it may be as simple as changing from carpet to tile or painted concrete. Dark colours where the floor is sunny can help.

If you are willing to make a major committment, consider radiant heat in the floor. This requires exposing the concrete floor, laying down tubes, and casting another 1.5" or so of concrete on top. Now the entire slab is thermal mass, and it has enough mass that you can heat it through the warm part of the day.

Another possibility is to figure out a way to pre-heat the air reaching the heat pump. Some clever work with a thousand feet of black plastic pipe, and a PV water pump may do wonders. If the heat pump can be converted to using water as it's source, then coils of pipe on the roof to heat a tank of water with the heat pump pulling your house heat out of the tank could be very effective.

However some local experience: I live in Alberta. We have a 2500 square foot house with 2 bathrooms. The house is conventional 2x6 construction, not especially tight. The bathrooms are heated with ciculating hot water. The rest of the house is heated with wood. Yesterday the outside temperature ranged from -10 to 0 C. It was a sunny day. We have most of our windows on the sunny side. No fires yesterday.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.