More Building Science
The two primary ways you can heat your house are by burning a fuel (e.g., fossil gas, fuel oil, wood) or using electricity. Furnaces and boilers distribute the heat from combustion to heat a house. Electric resistance heat makes sense only in limited applications, so heat pumps are the best way to heat with electricity. I love heat pumps and have replaced the furnace in my house with two of them. But let’s get something straight here. To say a heat pump is more efficient than a furnace or boiler is comparing apples to pumpkins.
Missing an important point
Unfortunately, a lot of articles on this topic miss an important point here. Even some reputable sources make this mistake. I just looked at three articles comparing heat pumps to furnaces and boilers and found these claims about heat pump energy use:
- “more than twice as efficient as gas furnaces”
- “a third less energy for the same output”
- “much less energy than furnaces”
How can they make those claims? Well, a good furnace turns about 95% of the energy in the fuel it burns into heat for the house. A heat pump uses its input energy to move about 200 to 300% as much energy into the house in the form of heat. Let’s put some numbers on this to illustrate.
You put 100,000 BTU of fossil gas energy into a furnace, and you get 95,000 BTU of heat delivered to the house. You put 100,000 BTU of electricity into a heat pump, and you get 200,000 to 300,000 BTU of heat in the house. If that’s all you look at, it would seem those statements above are correct. You put the same 100,000 BTU of energy into the two types of heating systems. Yet the heat pump puts 2 to 3 times as much heat into the house. But something’s missing here.
What’s the catch?
The two types of equipment do different things. A furnace uses the heat of combustion from burning a fuel. A heat pump uses electricity to run mechanical equipment that moves heat from one place to another. But we need to go further to find the real reason those quotes above are misleading.
![Coal-fired power plant [by stanze, from flickr.com]](https://www.energyvanguard.com/wp-content/uploads/2021/11/coal-fired-power-plant.jpg)
- Burning gas or coal
- Harvesting the heat from a controlled nuclear reaction
- Taming the kinetic energy of blades spinning in the wind or in a dammed river
- Getting electrons to move when radiant energy from the sun hits a photovoltaic module
In the U.S., a lot of our electricity still comes from the first one in that list. In the case of electricity generated by burning coal, the efficiency at the plant is 35 to 40%. Factor in the losses as the electricity goes from the plant to your house, and you’re down to about 30% efficiency. Gas-fired power plants are better but still only about 50% efficient, with the same transmission and distribution losses.
The bigger picture
When you use that electricity in a heat pump, you get about three times as much heat as electrical energy used. That brings the net efficiency back up to the 90 to 100% range. And that’s about the same as a high-efficiency furnace. What happened to heat pumps being “more than twice as efficient”?
Now, the electricity that gets delivered to your house doesn’t come only from burning gas or coal. The actual mix of primary energy sources depends on where you are. In the Pacific Northwest, much of the electricity comes from clean hydroelectric plants. Quebec has about the cleanest electricity in North America, with 95 percent being hydroelectric. In Wyoming or Georgia, the electricity isn’t so clean.
Apples and pumpkins
Still, making a direct comparison to say a heat pump is more efficient than a furnace isn’t helpful. To do so, you’re comparing primary energy used on site to secondary energy that comes from other energy conversion processes off site.
Don’t get me wrong. I’m definitely not saying you should avoid heat pumps just because the net efficiency might be no better than a gas furnace. I’m just saying you can’t look only at the rated efficiency of the two types of equipment.
Another big difference is that the rates utilities charge for electricity and gas are different. Even when you have a super-efficient heat pump going up against an old, inefficient furnace, the furnace may win on cost. That’s because gas rates in many places are really low.
The good news, though, is electricity keeps getting cleaner as solar and wind grow like mad. So, there are good reasons for going with heat pumps, including that they result in lower carbon emissions than a furnace in almost every state in the U.S. But choosing a heat pump because it’s more efficient than a furnace generally isn’t one of them.
________________________________________________________________________
Allison Bailes of Atlanta, Georgia, is a speaker, writer, building science consultant, and the founder of Energy Vanguard. He has a PhD in physics and writes the Energy Vanguard Blog. He is also writing a book on building science. You can follow him on Twitter at @EnergyVanguard.
Weekly Newsletter
Get building science and energy efficiency advice, plus special offers, in your inbox.
36 Comments
No criticism of the article, but a couple of observations.
Why not measure the Effectivity in Dollars per BTUh or Dollars per KWh? The traditional method is Dollars per Therm. That provides more of a "apples-to-apples" comparison.
During the day a Heat-Pump that pulls it power from Solar PV produces BTUh per dollar is much more than a boiler with an AFUE of 96%. At night when the temperatures drop to -19F a boiler does much better. The third alternative is to store solar heated water in 300 gallon buffer tanks.
The efficiency for boilers comes from the fact that hydronic heating is able to distribute the heat within an enclosure much more effectively than air (due to the specific gravity of the glycol mixture).
For the environmental side - we need to add more scrubbers to energy sources that produce fine particulates. Then work on the overall energy portfolio for the grid.
On the Grid side - we need to factor in the cost of overall grid improvements on both the Transmission and Distribution side. Many communities are not ready to migrate to other sources beyond Wood, Propane or Heat-Oil.
Is there enough copper controlled by the US. to upgrade the infra-structure. (Supply chain issues).
For the residential part of the equation is it possible to down-scale "Carbon Sequestration" so Carbon could be captured out of a boiler instead of being released into the air. We already have condensate neutralizers for boilers - why not have a CO2 capture mechanism for residences. That could also reduce the CO2 emissions from a residence and address part of the perceived problem.
While kWh is a unit of energy, BTUh is one of power.
It's possible I've misunderstood your point about boiler efficiency, but there are plenty of furnaces (air distribution) that have comparable efficiencies to boilers, are there not?
The AFUE from boiler and furnaces can be equivalent, i.e., in the 95-96% range. AFUE is a measure of the generation efficiency.
1. A boiler (hot water system) can heat the home and provide DHW.
2. A furnace can heat and cool (A/C) the home.
3. An Air to Water Heat Pump can heat/cool and provide DHW.
From that perspective, an Air-to-Water heat-pump wins hands-down.
The point, I was making is that an ECM pump can distribute heat/cooling through the house at a lower operational cost than an ECM fan. (Distribution efficiency).
"The good news, though, is electricity keeps getting cleaner as solar and wind grow like mad. So, there are good reasons for going with heat pumps, including that they result in lower carbon emissions than a furnace in almost every state in the U.S. But choosing a heat pump because it’s more efficient than a furnace generally isn’t one of them."
As they say: *A rising tide lifts all boats.*
I don't disagree with your analysis. The marketing is correct if you compare heat pumps to resistance heating but it was carried over and applied to gas incorrectly. I take issue with the overall message though.
Choosing a heat pump because it's more efficient than gas *should* be a factor. Your efficiency will improve over time (and it will accelerate rather than improve linearly). The alternative is locking into that 95-98% efficiency for 20 years. You're right that the marketing should be fixed.
We should also note that many homes will pair heat pumps with rooftop solar.
Great article. Thanks. What is the best resource to see what the mix of electricity generation sources is for my specific utility (which is a rural electric cooperative that is part of Allegheny system)? A while ago they sent around a breakdown, but I would like an update. I burn heating oil so it is a no brainer to replace my boiler with efficient mini splits, especially since my grid tied PV system covers part (but not all) of the additional electricity I use to heat. However, I am also considering an EV, and such a decision may be determined to some degree by the mix of electricity from the grid. Given the huge mark up on all EV cars and trucks at this point, I would rather wait for a while, unless it is clear that my electricity is so clean there is an immediate benefit to the environment to going electric.
The EPA has a "power profiler" tool that lets you enter your zip code and get info for your region.
https://www.epa.gov/egrid/power-profiler#/
It operates on larger regions than your utility, but given the way the grid is managed, that arguably makes sense. The EIA has per-state data https://www.eia.gov/electricity/data/state/emission_annual.xlsx For your utility you might find the data on the state PUC web site, or on the utility's own web site.
When you are replacing a heating system or a vehicle, and expect the new hardware to be used for 10-20 years, it makes more sense to project forward 5-10 years and consider how clean we expect the grid to be moving forward, not just what it is now, so even if the comparison is neck-and-neck now, electric is the better choice. You'll also be helping build capacity in the industry that we need moving forward.
Thanks, I will check this out.
I agree that longer term planning is key and I would expect C02 emissions per kwh to go down. However, in the short term, with EV prices so inflated (around here: $6-$10k mark ups on used Model 3 from a year ago), waiting a year or two won't hurt much
I'd suggest looking here: https://app.electricitymap.org/map
That should give you a pretty good figure for the emission intensity of the grid you're connected to. It looks like you're likely on the PJM interconnect, which would put your intensity around 420 g/kWh.
For an example EV, a base Tesla Model 3 uses 151 Wh/km (ev-database). So at 420 g/kWh * 1.05 (5% grid transmission losses) * 1.05 (5% charging losses), that comes out to 70 g/km (112 g/mi).
Since gasoline combustion releases 2300 g CO2 per litre, that EV would have operational carbon emissions equivalent to a gas car at 0.03 litres/km = 3 l/100km = 78 mpg.
If you know your annual mileage (distance) and average mileage (mpg), you can calculate how much you'd reduce your operational emissions. For easy comparison, you can use this calculator to convert mpg to CO2 g/km.
https://www.unitjuggler.com/convert-fuelconsumption-from-mpg-to-gperkmgasoline.html
To get g/mi, multiply g/km * 1.609.
Thanks, I will check these out.
In some ways, consumer investments in efficient electric HVAC will compete with investments in EV's, as consumers have limited $$. It will be interesting to compare relative benefit of one type of investment over another.
It seems from the materials, the amount of C02 emitted to produce a kwh of electricity is dependent on time of day/peak/off peak use. My electric cooperative pushes electronic thermal storage heating (ETS) systems (Steffes) as a way to shift heating load to all off peak hours. I didn't think it made sense because it was a COP of 1, but maybe it does make a difference only consuming off peak electricity.
How much the ETS makes sense probably depends on your off-peak generation mix (if you're optimizing for CO2) or off-peak rates (optimizing for cost). If you're in an area with a lot of nuclear/hydro and cheap off-peak rates (like Ontario), it could work out. Since a good cold climate ASHP can have a COP over 2 at 5f, you'd want off-peak emissions intensity or cost to be half that of on-peak/mid-peak.
If you're thinking about replacing a heat source for a hydronic system, you may also be able to use an air-to-water heat pump like those from Arctic. You could then add a larger buffer tank to use for thermal storage. (But you might have to change your emitters/radiators to accommodate lower loop temperatures, and that starts getting pricey.)
I was looking at the A2W option a few weeks ago when someone was selling a used Arctic system. The problem is that the emitters (high efficiency radiators or panels and buffer tanks are very pricy.
I was discussing the ETS system with my rural electric utility cooperative. They claim to be part of the Allegheny cooperative and get a lot of power from nuclear and hydro, but I can't verify it. The off peak system runs off a separately metered service that runs only off peak at a rate of 6-cents kwh. Cheap. They use a Steffes furnace-essentially a box of bricks heated up during off peak hours and they emit heat (as a boiler, hot air furnace or point heater) during peak hours. They are often paired with an well insulated H20 heater like a Marathon that runs off the service. I didn't like that fact that it essentially has a COP of 1...However, they did suggest that they would also let me charge an EV off peak and even run heat pumps off peak (using the house as a bit of a heat store) along with the Steffes system. The problem is that the calculations are complicated to get the heating right using the Heat pump off peak only and I think you would need a very tight highly insulated building or you would lose too much energy (and probably have to over heat the house when the heat pumps were running). Work from home also complicates this idea.
Does anyone know of studies on supply temperature and efficiency? For example, is the lower supply temperature (e.g., and perhaps more steady-state operation) of a heat pump system more efficient than the higher supply temperature (e.g., with perhaps less steady-state operation) of a gas furnace?
It seems like system/distribution losses might be greater with higher supply temperatures as heat transfer depends on temperature differential. Just don't know if someone has studied this?
Also, any efficiency differences due to how air is moved past a furnace heat exchanger versus an ASHP air handler heat exchanger?
Edited after a quick search on info from NREL: one study "an exponential drop in efficiency as supply air temperature rises".
Nicol [11] measured duct air flow rates and temperatures to determine steady-state duct efficiency. He calculated the cyclic efficiency assuming an exponential drop in efficiency as supply air temperature rises.
Grot and Harrje [10] give more details on the distribution system studied by Princeton. They insulated a duct system and compared its performance with that of the uninsulated system:
It was found that although the addition of duct insulation did drop the basement temperature by about 5°F, the profiles of the air exit temperatures at the room registers did not change appreciably and that little additional heat was delivered to the living areas during normal transient furnace operations.
Their explanation for this phenomenon is that, during the short on-cycles common in residential furnaces, the major mechanism for cooling the supply air is heat transfer to the metal of the ducts. During the on-time of the furnace and blower, the duct does not lose a significant amount of heat to the surrounding space but the duct sheet metal stores a significant amount of heat. This stored heat is lost by conduction, convection, and radiation after the blower is shut off.
The heating capacitance of ducts has an effect on when and where energy from the furnace is delivered to the living space. Grot and Harrje [10] concluded from their data that "... during intermittent blower operations, the furnace blower is turning off at a time when many of the ducts are just reaching maximum exit air temperatures." They also found that the duct metal itself removed a sizeable fraction of the heat from the supply air during the first minutes of burner operation.
https://www.osti.gov/servlets/purl/6444222
Maybe, for ASHPs, it's time to replace metal ducts with material that is of lesser heat conductivity (and heat capacity)? Metal ducts are probably a throw back to high temperatures from gas and oil burners. Thinking PEX rather than copper in plumbing.
I think when people say they want an efficient heat source what they are really interested in is the one with the lowest cost to operate and they could not care less about how much heat goes out the chimney or how much energy is lost in generation and transmission of the electricity. The answer to the real question is in dollars and that answer is very local and mostly to do with the price and availability of the different fuels at that location.
“I love heat pumps” My guess is if operating costs were no object 90% or more people would rather have a furnace blowing 130° air at them than a heat pump moving 2X the air at 90°.
Great article!
Walta
Usually a furnace blows at a higher velocity, so while it is warmer, it might not feel warmer.
Dr. Bailes quite rightly notes that the % efficiency numbers of heating sources only compare their on-site energy usage. As he points out, this is only one piece of a larger picture. However, it's also the only measurement that is independent of other variables - notably, the cost of electricity/gas; and the off-site emissions of the electric grid/gas network. But those other parts of the picture vary dramatically with location and time. (For example, off-peak electricity in Ontario versus on-peak in Germany.) Because of that variability, % efficiency at point-of-use is the only figure manufacturers can publish with any certainty.
Fortunately, it's not too difficult to calculate the others (costs and emissions) once we know where we're installing the equipment. Cost is pretty straightforward - heating load / efficiency * cost per unit of energy. (This gets more complicated with time-of-use billing, of course.)
GHG emissions is also doable. We can find electric grid emissions intensity data collected in a handy map format here: https://app.electricitymap.org/map It looks like they have data for almost all of the US (by interconnect), about half of Canada (by province), and almost all of Europe (by country). That gives us GHG per unit energy. The math will be GHG intensity * (heating load / efficiency) * 1.05. The 1.05 factor accounts for the average 5% loss in transmission and distribution (per the US Energy Information Administration).
GHG for natural gas is a little harder, since we should account for leaks (aka. fugitive emissions) in the gas transmission and distribution network. A 2020 EPA study (Kirchgessner et al) figured these to be about 1%. But those are methane leaks, and methane has a global warming potential of 28 (that is, 28 times worse than CO2; IPCC AR5). At least the GHG intensity of gas combustion is pretty straighforward: 185 g/kWh. So the math becomes: GHG intensity * (heating load / efficiency) * 1.28.
So as an example, I'm in Alberta, which has a pretty dirty grid at around 450 g/kWh. For a home with a heating load of 7 kW (about 2 tons) comparing between a heat pump with an average CoP of 2.5 (250% efficient) and 95% gas furnace, the math is:
Electricity: 450 g/kWh * (7 kW / 2.5) * 1.05 = 1323 g/hr
Gas: 185 g/kWh * (7 kW / 0.95) * 1.28 = 1745 g/hr
So in this case, the heat pump results in lower emissions. It's also notable that the heat pump still comes out slightly ahead even if you ignore the fugitive methane emissions. (Yes, you could make this analysis even more detailed by accounting for changes in CoP with changes in outside temperature informed by historical hourly weather data for your location.)
We can also re-arrange this math to tell us the breakeven CoP at which a heatpump will have lower emissions:
Breakeven CoP = (electric grid emissions intensity) g/kWh * (furnace efficiency) * 0.0044
In the Alberta example, it comes out to: 450 * 0.95 * 0.0044 = 1.88.
For the US national average of 385 g/kWh (US EIA), the breakeven is 385 * 0.95 * 0.0044 = 1.6
(If you want to ignore methane leaks, you can use 0.0056 instead of 0.0044, but we really should be conscious of methane leaks from natural gas distribution networks, because they're a major source of GHG.)
Basically, we do have the data needed to find a reasonably good answer to the question of whether a heat pump is more efficient than a furnace. That answer depends on where you live, but it's probably 'yes'.
Good morning rndvp6. I am in Airdrie AB, and am considering a dual fuel system in my new home under construction. Considering electricity and nat-gas rates are quite reasonable here in AB, I have downplayed the return on investment of the heat pump vs an air conditioner, but rather focused on being a part of longer term transition to lower carbon emissions. I've only recently begun trying to decipher the heat loss calc's provided by my builder's HVAC contractor, and am really hitting a roadblock on how to quantify a path forward. Any insights as a fellow Albertan?
The numbers used here for transmission and distribution losses are unrealistically high. We need to squelch the myth that they are a major problem.
EIA estimate 5% of electricity put into the grid is lost in T&D. Importantly, that's 5% of the electric energy fed into the system, not 5% of the energy in the fuel burned. https://www.eia.gov/tools/faqs/faq.php?id=105&t=3
The article says T&D losses bring a 35 to 40% efficient plan efficiency down to 30% system efficiency. That would mean 14 to 25% T&D losses. That's off by a factor or 3 to 5!
Also - gas furnaces are horrible at air conditioning ;) why just have a condenser unit cool when it can heat? It would be good to compare the GHGI of both systems as well for a range of energy mixes... and as other comments have noted, costs.
I'm confused as to why only the inefficiencies of the electrical production and distribution are factored in but not those of the gas, oil, or propane used by furnaces? How do they get to stay at 95%, while electricity has to take a hit at each stage?
Excellent point. Exploring/Discovery, Development, Shipping/Transport, Refining, Marketing and all the rest is not energy 'neutral'
Such as the fugitive methane leaks described by rndvp6 above?
On the production question, they are two different efficiencies. The 35% to 40% number is the efficiency of converting a burned fossil fuel like natural gas into electricity. The 95% number is the efficiency of converting burned natural gas into heat. BTW, in converting natural gas into electricity, much of what is not converted to electricity is converted to, you guessed it, heat. If only that heat produced at the electricy plant could be piped to your house....
On the transmission question, the amount of natural gas lost from leaking pipes is debated but probably as a percentage is in the same single digit percentage range as transmission loss from electricity. So maybe that 95% number drops down to 85-90%.
It’s a pretty unfair comparison still - is a 95% efficient furnace or boiler the majority of new installs or replacements? I doubt it. However, almost all new gas generation capacity is combined cycle with efficiencies around 50% and coal generation is declining in the US. Also the methane leakages aren’t the same as electricity since methane is much worse for the climate. I think the article conclusion is correct - it’s probably not worth bothering with the calculation.
It's even better than that - GE is citing 63-64% combined-cycle efficiency for its current generation plants. 50% seems like a reasonable fleet average including older combined-cycle, simple cycle, and converted coal plants.
The switch from coal to gas and renewables in the US has reduced the emissions intensity of electricity (CO2 per unit of energy) by more than 40% since 2000, and it's still dropping: https://emissionsindex.org/
Response to Malcolm, comment #13:
I agree that oil, gas, and propane all have transmission losses due to transport, but the big loss from converting combustable fuel into electricity is a step that direct combustion appliances don't have. And when it comes to making electricity from gas/oil/coal, all that prospecting, mining, refining, shipping, etc. still happens.
But ultimately, a total Life Cycle Assessment should happen to truly determine a systems efficiency. That's a significant undertaking and tough to do comprehensively. And of course cleaner and more efficient grids changes the game significantly.
One other general note: it's common for articles to be written from the perspective of gas being the default heating fuel. Let's not forget the many rural communities that still heat with oil (and/or wood) and commonly have older furnaces/boilers that are well below 95% efficiency.
“[Deleted]”
Malcolm: You're right. I should have mentioned gas losses in the article because that does indeed bring down the overall efficiency of furnaces.
What gas losses? Are you talking about losses when the system isn't operational? That's easily addressed when one turns off the supply valve at the end of heating season.
Distribution losses are shared by all users of NatGas.
Allison is talking about natural gas which is a primary energy source whereas electricity is a secondary energy source. Whatever inefficiencies in Natgas are present for both Natgas residential heating and NatGas electrical generation.
https://www.greenbuildingadvisor.com/article/putting-methane-leaks-into-perspective
https://www.greenbuildingadvisor.com/article/natural-gas-pipelines-are-leaking
https://www.greenbuildingadvisor.com/article/concern-mounts-over-natural-gas-leaks
Ya. it's irrelevant because both power generators and residential users draw from the same distribution system.
Leaks are only relevant when you want to compare different fuel sources for power generation in terms of GHG/Carbon emissions.
For example with regards to NatGas you have the following emissions sources
Extraction
Leaks within the distribution system
Flaring of excess supply (Burning or not)
Burning for use (heating, power generation, etc)
Coal
Extraction
Transportation (typically rail)
Burning
Disposal of coal ash.
Some additional "plugs" for ASHPs
A ng furnace has little to no control over flame temperature, 3400-3600 F. Some furnaces have a modulating gas valve, which may be from 40% to 100% at 1% increments. In advertising: "if you only need a little bit of heat, the 98 will give you only a little bit of heat". Is 40% of capacity really "only a little bit of heat"?
40% of capacity is still likely to give a quite high supply temperature as the 3400-3600 F is pretty much fixed; noting higher blower speed may reduce supply temperature - but still higher supply temperature equates to lower efficiency.
If a ng furnace is sized at some standard such as 1.4x of design load, 40% is 0.56x of 1.4x design load. If a heat pump is sized at the same 1.4x of design load, minimum output can readily be at or below 0.46x of 1.4x design load. The expanded control range (low end), together with a lower supply temperature, give ASHPs (ducted) substantial advantages over ng furnaces (ducted).
And just to throw the wet blanket of cynicism on this party, all of the above discussion assumes reasonably proper installation. Of course we need to assume that to run any meaningful calculations. But in practice, of course systems are routinely not installed properly. I can't remember the last time I saw a ng furnace that was properly sized. 1.4x? More like 3x-4x required sizing. Same thing for minisplits. How many threads are there on this list alone where homeowners have listed the hp systems proposed by their contractors that are oversized by 2x-3x for the whole house and/or 5x for individual heads in individual small bedrooms. And that's sizing at the min/max design temperatures that we only see for a few hours of the year. In reality, I would bet that most installed variable systems run at the minimum capacity, on/off, most of the time. At minimum capacity, many/most HP systems run at higher than their advertised efficiency. That's not the case for natural gas furnaces, though the lower fan speeds do help. Throws another complication in any attempt to really compare the different systems. So once again, we get back to the idea that we need to stop burning dinosaurs, as fast as possible. Investing in new combustion equipment is locking in that decision for 20-30 years.
Even while covered by the wet blanket of cynicism, I'm warm inside with my 1x (exactly sized) ASHP even though it was 7 F last night ;-)
And, wow, "I would bet that most installed variable systems run at the minimum capacity, on/off, most of the time".
I have used G.I.S. insulation and fabrication services and personally my experience was really good and i am quite satisfied with the service.
Fire protection systems
Log in or create an account to post a comment.
Sign up Log in