The reason why insulation is more important for heat pumps is because they slowly warm up your space with lower temperature heat. A strong current of high temperature heat is going to feel better comparably in a drafty house.
This is the key difference - our house is well insulated, and I still notice this effect if I try to turn down our heat at night and then back up in the morning. I just stopped doing that and it works just fine.
It would be nice to improve on this aspect of heat pumps having low throughput. It’s common practice (and better for your sleep) to set your temperature back at night but it’s pretty annoying to wait three hours to warm up the place in the morning.
If you're waiting three hours for the place to warm up, your heat pump is undersized. Harking back to Hannahs' comments about the need for a skilled workforce. (And possibly also quality standards in the industry, to avoid just this sort of reputational problem.)
These articles say not to set back with a HP. I don’t know how common this advice is (though I’ve heard it numerous times). I’m just saying they need to innovate if this is still the case with newer models. I guess one option is to have a substantial resistance heater integrated that can provide a heat boost for those that want it.
Heat pumps are measured not by COP in heating but by HSPF, which takes into account defrost efficiency and other items during heating. For operation in cooling mode SEER is the rated metric.
The failure to recognize thus detracts substantially from the article and the author's credibility. Plus ghe obvious failure the view the correlation between unit density and average winter ambient as subject to energy costs and subsidiaries and other factors is a serious omission. In Canada all of these factors come into play, particularly give our wide range of climate zones. The author needs to look further into the matter
I am a mechanical engineer with 40 years of experience. I totally disagree with your premise regarding the use of HSPF vs. COP and I find your questioning of the author's credibility totally inappropriate.
HSPF is a complicated standard adopted by the industry to describe the seasonal average heating performance for one specific climate zone using a single value of merit, https://www.aceee.org/files/proceedings/2004/data/papers/SS04_Panel1_Paper08.pdf, while COP is an easy-to-understand thermodynamic definition used correctly by the author.
I have a heat pump and the manual lists both the HSPF as single number and then has a table of heat output and COP at various temperatures. I find the COP table more informative and useful.
Well, I have to say that this is a rather incoherent response to which a direct reply is difficult. The issue is, in short, whether HEAT PUMPs are suitable for cold climates and HSPF is in fact the thermodynamic performance of a specific machine in a specific climate environment, including defrost efficiencies, something not accounted for in COP. So what's the problem with HSPF? Plus can you comment on the obvious observation that subsidies and fuel rates effect hp sales as much, or more than, hp efficiencies, across national sales figures.
The “S” in HSPF and SEER stands for Seasonal. These are measurements of efficiency over an entire season of heating or cooling. COP is a measurement of efficiency in a particular instant.
I've looked into the issue. Not technically, but in practice since this summer my landlord, with whom I share a house, installed heat pumps (one for his unit and one for mine). I can assure you that here in cold New England, USA, my cost per month thus far during the coldest months of the year have been considerably lower than they were using natural gas in the past.
Occupancy and thermostat are exactly the same. Weather hasn't been normalized, but it also isn't that much different than normal. I also haven't taken into account the fact that gas prices went up this year by about 25% so if we'd still been using gas, it would have been much higher than last year.
Living in the US its surprising how negative opinion in the UK is of heat pumps. Here they are most popular in colder places like Maine where winter heating bills are highest. Those savings are so high because they are replacing expensive oil-fired central heating systems.
The performance of the UK heat pump installations is particularly poor given how mild the UK's climate is compared to Canada with its much colder winters. Perhaps the difference is due to UK homes using pumped hot water central heating vs the forced air systems common in the US and Canada. Heat pumps are less efficient at larger temperature differentials and pumped hot water systems operate at higher temperatures than forced air.
The UK heat pump systems have a COP of 2.5 vs Canada's 3.3 despite the average January daily low of 4C in London and -7C in Toronto. The performance seems pretty poor, especially given how much warmer UK is in winter which should make heat pumps more efficient.
First, I agree with the author that heat pumps work in low temperatures and are excellent technology. I installed a heat pump 3 years ago in Wisconsin and have been very happy with it. In my region, however, electricity is 9 times more expensive than natural gas on an energy basis. The COP of a heat pump cannot come close to over coming this and is significantly more expensive to operate than a 90+% natural gas furnace (~ 3 times more expensive). I've been a little shocked at the cost to operate the heat pump.
I did wonder if in the extremely cold area examples here, they used ground source heat pump technology ( that operate really efficiently at very low temps)
Hannah, thank you for this article. Heat pumps are great in most climates, and are currently the only way to heat and cool a home without hydrocarbons. At least they make it possible to heat and cool homes with renewables. One huge problem in home heating systems is the distribution system. Duct losses can account for losing over 50% of heating capacity (not a typo) which creates a challenge for heat humps. Fixing the duct problem can boost the delivered heat from heat pumps dramatically. Here's more info on that, and a case study I did: https://energyresourcedynamics.substack.com/p/the-duct-problem-americas-building
I looked at heat pumps three years ago in southern Germany, and decided to go for Gas + Solar (with heating support - Heizungsunterstutzung).
This is brilliant at heating lots of water to not a very high temperature. OK, we have had 95C water in summer! But in winter, we still get a lot of heat if we set the radiator inflow temperature to 30C. That keeps the heat store at 28C, and the panels put a lot of heat in if there is sun, even at zero degrees outside. We don't have underfloor heating, but heating the house with 30C water works, till it goes below zero outside. 40C radiators keep the house warm always.
Hence I know a heat pump would work fine, given they are best at delivering 40C water. If anyone wants to know whether a heat pump would work, try lowering the radiator temperatures to 40C for a winter.
But the reason why I didn't go for a heat pump was because electricity was about 5 times the price of gas. In the UK I see gas is 7.5p/KWh and electricity 32p, so you need a CoP of >4 to just break even on running costs.
Lowering costs is only possible with access to variable electricity tariffs. Octopus Go is 9p/KWh from 1am to 5am. Now a heat pump becomes really competitive!
Except, most British homes can't get by being heated just 4 hours at night. Our solar water system in Germany has a 750 litre tank - limited by what fit through the door. It stores about 1KWh per centigrade difference. A heat pump raising it to 50C at night, would store 20KWh before the temperature drops to 30C. Not enough. With a heat pump I would probably want a 1500 litre thermal store, which isn't going to fit in the airing cupboard of a typical British microhouse.
There is Cosy Octopus tariff (04:00 - 07:00 & 13:00 - 16:00): 17.63p / kWh. Then you could use the thermal mass of the house to stay warm. But 17.73p is still expensive.
And in all cases, it means over sizing the heat pump to heat for 4 to 8 hours per day.
And you still have the issue of domestic hot water, which is preferred at 60C. Newer heat pumps can reach that, but many systems in Germany use an immersion heater to raise the temperature from about 50C to 60C.
So .... yes, they're the future. Yes they work. But there is still work to do on making them economically viable.
A question about your COP data: Cold-climate heat pump systems typically have a supplementary heating element (gas or electric-resistance) that kicks in when the outside temp drops below a certain point. Do your COP data represent the performance of "pure" heat pumps, or of heat pumps supplemented by these other kinds of heaters?
I have two heat pumps (one for upstairs, one for downstairs) for our 2700 square foot house in the midwest they have worked well. For our downstairs unit, we do have a dual system with a small gas furnace backup, our thermostat automatically chooses between heat pump and gas furnace based on what will be the lowest cost, per electricity costs and natural gas costs. It ends up choosing the gas heat once the temperature goes below about 20 degrees farenheight (-7c). I also notice that with heat pumps during the cold months, if you turn down your thermostat at night, say to 68 (20c), then bring it back up to 70 (21.5c) in the morning, it will take longer for the heat pump to catch up, so the thermostat will use the gas heat then too. Now I just don't bother turning it down at night.
Quality wise, I do say I slightly prefer the warmer air of the gas heat, but again, if I just keep the thermostat at the desired indoor temperature and don't futz with going down and up, I don't think much about it. To be clear, the heat pump heat is still warm coming out of the vents, just not almost hot like the gas heat is.
It's also worth saying that even when it reached 0 farenheight (-17c) the upstairs was fine and the heat pump did its job, even as the downstairs was using gas. I will have to try as an experiment next cold spell to see how the downstairs heat pump fairs. I have heard anecdotally that the peak heating capacity of gas furnaces is higher, so if it's super cold and especially if you are playing catch up, gas heat would be more convenient.
The upstairs probably is getting help from the warm air from downstairs rising, at least it does in my house with a similar setup, though I use mostly cooling in which case the upstairs unit does most of the work.
Can confirm part of that: I recently had a heat pump installed in a rural town in Canada, and while the pump itself has been great, the only installer I could find was completely incompetent, and does not seem to understand even its basic operation.
I just installed a heat pump in a Minneapolis, MN suburb. Many models, including the Mitsubishi model I put in, work in below 0 deg F ambient temps. My only remaining fossil gas appliance is a fireplace.
Brian, I'm also in the Twin Cities. I would like to reduce my reliance on fossil fuels, but do not think my house would work with a heat pump. Was your house already using forced air?
It strikes me that all the cold climate people who have had success had this combination of factors going for them:
1. Their houses were well-insulated.
2. They were able to use forced air heat.
3. They could potentially tolerate the aesthetics of having a bunch of duct work spider-webbing their homes (if they were switching from radiators to heat pumps).
Nice piece! Thanks for writing. Couple of questions for you. No biggie if you don’t know the answers, but thought you might’ve come across these questions in your research:
- How does capital / installation cost compare between heat pumps and boilers?
- How does capital cost of a heat pump rated for low temps compare to mild temps?
- Is there any other reason why people are complaining about heat pumps in poor insulation buildings? I agree that poor insulation would impact any heating system, but wondering if there’s something in particular to heat pumps that makes poor insulation especially troublesome
Differentiation needs to be made between means of distributing the heat. In a forced air system the heat pump does not put out as high of a discharge temperature as a standard boiler or gas furnace. If the ductwork and distribution isn't modified the for the lower temp it won't feel "comfortable". For a hydronic system with radiators designed for higher water temps the radiators need to be replaced with properly sized low temp units. Of course poor insulation and effective windows will result in poor perceived comfort even with proper dry bulb temperatures.
These factors I believe are what's required for a "highly skilled" workforce that the author is referring to.
I couldn't tell from the text whether the heat pumps also work at very low temperatures without the addition of heating elements. This is crucial for the desired CO2 savings.
As someone who loves radiators - does anyone know if there are good heat pumps that will heat radiators hot enough to function like they do in natural gas settings available in the US? I have seen some available in Sweden, but I happen to be in an old Victorian-era home with radiator heat (no forced air) and very poor insulation that cannot be safely improved without gutting the interior walls and rebuilding them. I was hoping I could find either a geothermal or an air-to-water system in the US...and someone adept at installation.
The reason why insulation is more important for heat pumps is because they slowly warm up your space with lower temperature heat. A strong current of high temperature heat is going to feel better comparably in a drafty house.
This is the key difference - our house is well insulated, and I still notice this effect if I try to turn down our heat at night and then back up in the morning. I just stopped doing that and it works just fine.
It would be nice to improve on this aspect of heat pumps having low throughput. It’s common practice (and better for your sleep) to set your temperature back at night but it’s pretty annoying to wait three hours to warm up the place in the morning.
If you're waiting three hours for the place to warm up, your heat pump is undersized. Harking back to Hannahs' comments about the need for a skilled workforce. (And possibly also quality standards in the industry, to avoid just this sort of reputational problem.)
These articles say not to set back with a HP. I don’t know how common this advice is (though I’ve heard it numerous times). I’m just saying they need to innovate if this is still the case with newer models. I guess one option is to have a substantial resistance heater integrated that can provide a heat boost for those that want it.
https://www.efficiencyvermont.com/blog/how-to/who-knew-8-ways-not-to-use-a-heat-pump
https://www.betterhomesbc.ca/products/heat-pump-operation/
They do have automatic controls
Yep that’s one option. Start the warm up at 4am or whatever.
Heat pumps are measured not by COP in heating but by HSPF, which takes into account defrost efficiency and other items during heating. For operation in cooling mode SEER is the rated metric.
The failure to recognize thus detracts substantially from the article and the author's credibility. Plus ghe obvious failure the view the correlation between unit density and average winter ambient as subject to energy costs and subsidiaries and other factors is a serious omission. In Canada all of these factors come into play, particularly give our wide range of climate zones. The author needs to look further into the matter
I am a mechanical engineer with 40 years of experience. I totally disagree with your premise regarding the use of HSPF vs. COP and I find your questioning of the author's credibility totally inappropriate.
HSPF is a complicated standard adopted by the industry to describe the seasonal average heating performance for one specific climate zone using a single value of merit, https://www.aceee.org/files/proceedings/2004/data/papers/SS04_Panel1_Paper08.pdf, while COP is an easy-to-understand thermodynamic definition used correctly by the author.
I have a heat pump and the manual lists both the HSPF as single number and then has a table of heat output and COP at various temperatures. I find the COP table more informative and useful.
Well, I have to say that this is a rather incoherent response to which a direct reply is difficult. The issue is, in short, whether HEAT PUMPs are suitable for cold climates and HSPF is in fact the thermodynamic performance of a specific machine in a specific climate environment, including defrost efficiencies, something not accounted for in COP. So what's the problem with HSPF? Plus can you comment on the obvious observation that subsidies and fuel rates effect hp sales as much, or more than, hp efficiencies, across national sales figures.
The “S” in HSPF and SEER stands for Seasonal. These are measurements of efficiency over an entire season of heating or cooling. COP is a measurement of efficiency in a particular instant.
COP is widely used to compare and measure heat pumps. The Department of Energy uses it as do many others. - https://www.energy.gov/sites/default/files/2016/04/f30/32212_Shen_040616-1135.pdf
HSPF is another great measure but the core of the article "Heat pumps work in cold climate" stands.
I've looked into the issue. Not technically, but in practice since this summer my landlord, with whom I share a house, installed heat pumps (one for his unit and one for mine). I can assure you that here in cold New England, USA, my cost per month thus far during the coldest months of the year have been considerably lower than they were using natural gas in the past.
Have you normalized the cost differential for weather and usage --thermostat set point and occupancy levels, for example?
Occupancy and thermostat are exactly the same. Weather hasn't been normalized, but it also isn't that much different than normal. I also haven't taken into account the fact that gas prices went up this year by about 25% so if we'd still been using gas, it would have been much higher than last year.
Living in the US its surprising how negative opinion in the UK is of heat pumps. Here they are most popular in colder places like Maine where winter heating bills are highest. Those savings are so high because they are replacing expensive oil-fired central heating systems.
The performance of the UK heat pump installations is particularly poor given how mild the UK's climate is compared to Canada with its much colder winters. Perhaps the difference is due to UK homes using pumped hot water central heating vs the forced air systems common in the US and Canada. Heat pumps are less efficient at larger temperature differentials and pumped hot water systems operate at higher temperatures than forced air.
Can you explain what you mean here by 'poor'?
The UK heat pump systems have a COP of 2.5 vs Canada's 3.3 despite the average January daily low of 4C in London and -7C in Toronto. The performance seems pretty poor, especially given how much warmer UK is in winter which should make heat pumps more efficient.
First, I agree with the author that heat pumps work in low temperatures and are excellent technology. I installed a heat pump 3 years ago in Wisconsin and have been very happy with it. In my region, however, electricity is 9 times more expensive than natural gas on an energy basis. The COP of a heat pump cannot come close to over coming this and is significantly more expensive to operate than a 90+% natural gas furnace (~ 3 times more expensive). I've been a little shocked at the cost to operate the heat pump.
I did wonder if in the extremely cold area examples here, they used ground source heat pump technology ( that operate really efficiently at very low temps)
Hannah, thank you for this article. Heat pumps are great in most climates, and are currently the only way to heat and cool a home without hydrocarbons. At least they make it possible to heat and cool homes with renewables. One huge problem in home heating systems is the distribution system. Duct losses can account for losing over 50% of heating capacity (not a typo) which creates a challenge for heat humps. Fixing the duct problem can boost the delivered heat from heat pumps dramatically. Here's more info on that, and a case study I did: https://energyresourcedynamics.substack.com/p/the-duct-problem-americas-building
Thank you Hannah!
I looked at heat pumps three years ago in southern Germany, and decided to go for Gas + Solar (with heating support - Heizungsunterstutzung).
This is brilliant at heating lots of water to not a very high temperature. OK, we have had 95C water in summer! But in winter, we still get a lot of heat if we set the radiator inflow temperature to 30C. That keeps the heat store at 28C, and the panels put a lot of heat in if there is sun, even at zero degrees outside. We don't have underfloor heating, but heating the house with 30C water works, till it goes below zero outside. 40C radiators keep the house warm always.
Hence I know a heat pump would work fine, given they are best at delivering 40C water. If anyone wants to know whether a heat pump would work, try lowering the radiator temperatures to 40C for a winter.
But the reason why I didn't go for a heat pump was because electricity was about 5 times the price of gas. In the UK I see gas is 7.5p/KWh and electricity 32p, so you need a CoP of >4 to just break even on running costs.
Lowering costs is only possible with access to variable electricity tariffs. Octopus Go is 9p/KWh from 1am to 5am. Now a heat pump becomes really competitive!
Except, most British homes can't get by being heated just 4 hours at night. Our solar water system in Germany has a 750 litre tank - limited by what fit through the door. It stores about 1KWh per centigrade difference. A heat pump raising it to 50C at night, would store 20KWh before the temperature drops to 30C. Not enough. With a heat pump I would probably want a 1500 litre thermal store, which isn't going to fit in the airing cupboard of a typical British microhouse.
There is Cosy Octopus tariff (04:00 - 07:00 & 13:00 - 16:00): 17.63p / kWh. Then you could use the thermal mass of the house to stay warm. But 17.73p is still expensive.
And in all cases, it means over sizing the heat pump to heat for 4 to 8 hours per day.
And you still have the issue of domestic hot water, which is preferred at 60C. Newer heat pumps can reach that, but many systems in Germany use an immersion heater to raise the temperature from about 50C to 60C.
So .... yes, they're the future. Yes they work. But there is still work to do on making them economically viable.
A question about your COP data: Cold-climate heat pump systems typically have a supplementary heating element (gas or electric-resistance) that kicks in when the outside temp drops below a certain point. Do your COP data represent the performance of "pure" heat pumps, or of heat pumps supplemented by these other kinds of heaters?
This is hilarious. Heat pumps are very popular in Sweden.
The main issue might be that houses in England aren't well isolated, if at all.
I have two heat pumps (one for upstairs, one for downstairs) for our 2700 square foot house in the midwest they have worked well. For our downstairs unit, we do have a dual system with a small gas furnace backup, our thermostat automatically chooses between heat pump and gas furnace based on what will be the lowest cost, per electricity costs and natural gas costs. It ends up choosing the gas heat once the temperature goes below about 20 degrees farenheight (-7c). I also notice that with heat pumps during the cold months, if you turn down your thermostat at night, say to 68 (20c), then bring it back up to 70 (21.5c) in the morning, it will take longer for the heat pump to catch up, so the thermostat will use the gas heat then too. Now I just don't bother turning it down at night.
Quality wise, I do say I slightly prefer the warmer air of the gas heat, but again, if I just keep the thermostat at the desired indoor temperature and don't futz with going down and up, I don't think much about it. To be clear, the heat pump heat is still warm coming out of the vents, just not almost hot like the gas heat is.
It's also worth saying that even when it reached 0 farenheight (-17c) the upstairs was fine and the heat pump did its job, even as the downstairs was using gas. I will have to try as an experiment next cold spell to see how the downstairs heat pump fairs. I have heard anecdotally that the peak heating capacity of gas furnaces is higher, so if it's super cold and especially if you are playing catch up, gas heat would be more convenient.
The upstairs probably is getting help from the warm air from downstairs rising, at least it does in my house with a similar setup, though I use mostly cooling in which case the upstairs unit does most of the work.
Can confirm part of that: I recently had a heat pump installed in a rural town in Canada, and while the pump itself has been great, the only installer I could find was completely incompetent, and does not seem to understand even its basic operation.
I just installed a heat pump in a Minneapolis, MN suburb. Many models, including the Mitsubishi model I put in, work in below 0 deg F ambient temps. My only remaining fossil gas appliance is a fireplace.
Brian, I'm also in the Twin Cities. I would like to reduce my reliance on fossil fuels, but do not think my house would work with a heat pump. Was your house already using forced air?
It strikes me that all the cold climate people who have had success had this combination of factors going for them:
1. Their houses were well-insulated.
2. They were able to use forced air heat.
3. They could potentially tolerate the aesthetics of having a bunch of duct work spider-webbing their homes (if they were switching from radiators to heat pumps).
Nice piece! Thanks for writing. Couple of questions for you. No biggie if you don’t know the answers, but thought you might’ve come across these questions in your research:
- How does capital / installation cost compare between heat pumps and boilers?
- How does capital cost of a heat pump rated for low temps compare to mild temps?
- Is there any other reason why people are complaining about heat pumps in poor insulation buildings? I agree that poor insulation would impact any heating system, but wondering if there’s something in particular to heat pumps that makes poor insulation especially troublesome
The issue is not so much insulation, but low radiator area compared to heating requirements.
If you can turn your gas boiler down to 40C boiler outlet temperature, and the house is still warm in mid winter, then a heat pump will work fine.
Octopus are quoting about £9000 for a air source heat pump install (less subsidy). I guess a gas boiler replacement will be £3000 to £4000.
Differentiation needs to be made between means of distributing the heat. In a forced air system the heat pump does not put out as high of a discharge temperature as a standard boiler or gas furnace. If the ductwork and distribution isn't modified the for the lower temp it won't feel "comfortable". For a hydronic system with radiators designed for higher water temps the radiators need to be replaced with properly sized low temp units. Of course poor insulation and effective windows will result in poor perceived comfort even with proper dry bulb temperatures.
These factors I believe are what's required for a "highly skilled" workforce that the author is referring to.
Got it. So just to make sure I understand… you’re saying that:
A) a heat pump system has a lower discharge temperature.
B) bc it has a lower discharge temperature, the distribution system must be appropriately modified or else it won’t feel comfortable
Is that roughly right?
HPs do "work" in colder climates but they also often don't meet their product rated performance. Heat pumps could further improve through good design and installation - https://www.researchgate.net/publication/356749299_Air-to-water_heat_pumps_Review_and_analysis_of_the_performance_gap_between_in-use_and_product_rated_performance
I couldn't tell from the text whether the heat pumps also work at very low temperatures without the addition of heating elements. This is crucial for the desired CO2 savings.
As someone who loves radiators - does anyone know if there are good heat pumps that will heat radiators hot enough to function like they do in natural gas settings available in the US? I have seen some available in Sweden, but I happen to be in an old Victorian-era home with radiator heat (no forced air) and very poor insulation that cannot be safely improved without gutting the interior walls and rebuilding them. I was hoping I could find either a geothermal or an air-to-water system in the US...and someone adept at installation.