Where is the most efficient place to get the heat for hot water from?
Where is the most efficient place to get the heat for hot water from? This is a question I’ve been kicking around for a while. It came back to top of mind because GBA had a review recently of the Sanco2 water heater ( see https://www.greenbuildingadvisor.com/article/sanco2-heat-pump-water-heater ) which provides domestic hot water from a heat pump located outside the building. The question I’ve been kicking around is, if you’re pulling heat from outside for your hot water via a heat pump, is it inherently more efficient to do it in a single stage with an external compressor, or in two stages with a HPWH that takes heat from inside the house, and a heat pump heating system that replaces that heat from outdoors? We kicked the question around a bit in January in this thread: https://www.greenbuildingadvisor.com/question/heating-domestic-hot-water-with-a-heat-pump-one-stage-or-two , but didn’t really come to a conclusion. And we touched on it a bit in this thread in March: https://www.greenbuildingadvisor.com/question/minimal-radiant-heating-options . So I thought I’d take another whack. I built a spreadsheet model with the following assumptions: 1. Heat pump COP is linearly proportional to the temperature difference between the source of the heat and the destination. 2. All the heat pumps being used — the HPWH, the home heat, and the dedicated external water heater — have the same performance curve. There’s no reason to believe that any one is going to have inherently better performance. 3. In winter, the HPWH pulls heat from the building, which has to be replaced by the heating system. 4. In summer, the HPWH pulls heat from the building. If you’re not using a HPWH, and instead using a dedicated external water heater, the heat that would have been pulled by the HPWH instead needs to be pulled by the AC. I’ve attached two graphs showing winter and summer performance. In the winter, you want to look at the yellow line (external) vs gray line (HPWH). In the summer you want to look at the red line (HPWH) vs gray line (external plus AC). As you can see, the only time the dedicated external water heater gives higher COP is in the winter when temperatures are between about 40F and 70F. I used the performance line from Chiltrix because I had already entered the data for an earlier question about heat pump performance vs temperature. If you look at their performance graph here: https://www.chiltrix.com/Efficiency-curve.jpg you’ll see that the assumption that COP is linearly proportionate to temperature difference is a very good one. I assumed the interior was at 70F and the water heater was at 110F.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
I want to say a word about how I calculated combined COP's. COP is defined as the heat moved divided by the energy required to move it. So for a two-stage system I calculated the energy needed to move one unit of heat for each stage, and then added them up.
For a heat pump water heater, the amount of energy needed to move one unit of heat into the tank is 1/COPWH, where COPWH is the COP of the water heater. The HPWH pulls heat out of the house, that heat comes from outside and the energy needed to move one unit is 1/COPHP, where COPHP is the COP of the heating heat pump. But -- and this is an important but -- when the HPWH creates one unit of hot water, it doesn't pull one unit of heat out of the house. The energy that it consumes -- equal to 1/COPWH -- also goes into the hot water, so the amount of heat it pulls out of the house for one unit of hot water is equal to 1-1/COPWH. The amount of energy it takes to move that in from outdoors is (1-1/COPWH)/COPHP. The total energy used to heat one unit of hot water is then:
(1-1/COPWH)/COPHP+ 1/COPWH.
COP is heat moved over energy used, we're moving one unit of heat so the combined COP is:
1/((1-1/COPWH)/COPHP+ 1/COPWH).
In the summer the calculation is similar. If you have a HPWH, it cools the house. If you have an external water heater that cooling has to be replaced by the AC. How much energy does it take to replace that cooling? It's (1-1/COPWH)/COPHP, same as the heat in the winter. But that needs to be added to the consumption of the external heat pump to get the combined COP.
For the summer calculation, you are essentially assuming that all the water heat makes it into the house? I'd say a large amount exits the house via the drain probably.
I've done some similar calculations on a spreadsheet. At the end of the day, I guess it depends if I end up with the space for a HPWH as to how I ultimately handle it.
No, I'm ignoring the heat of the hot water, it's the same no matter how the water is heated.
What I'm saying is that the heat removed by the HPWH would otherwise have to be removed by the HVAC. So to do an apples-to-apples comparison between the HPWH and a dedicated external heat pump water heater you have to add the energy needed to remove that heat to the dedicated one's calculation (or remove it from the HPWH one, you end up in the same place.)
In the comments to the GBA review of the SANCO2, Tim_O has posted this link to a desuperheater for residential AC units:
https://hvac.place/Heat-Recovery-Units_c_370.html
I'd like to model how that compares to a HPWH but I don't know enough about how they work. I could see it making both the AC and the WH more efficient.
They seem fairly new to the market, I think the parent company for Chiltrix, Hotspot Energy, makes one marketed towards heating pools. They might have a bit more energy info. It's basically a desuperheater from geothermal. Problem is, it's only useful in the summer.
This is fantastic. Thank you!
Thanks. I'm hoping other commenters can "peer review" my analysis and make sure there's not something I'm missing.
We'll let you know in 6 months if we've decided to provisionally accept this for publication.
One of the questions we get from time to time is whether a HPWH should be vented outdoors. Since the air that is vented gets replaced with outdoor air, the answer I give is that it only makes sense when it's heating season, but the outside temperature is warmer than the exhaust from the HPWH -- you're replacing cold air with warmer air from outside.
If you look at the winter performance curve, that's pretty much also the circumstance under which a dedicated external water heater makes sense. So I would argue that the optimum efficiency setup would be to use a HPWH, and have motorized dampers so it can vent outdoors when the outdoor temperature is warmer than the exhaust temperature during heating season.
That's a great idea and a surprisingly simple rule for when to do it.
What is the temperature of the exhaust on these heat pumps?
In my calculations I'm assuming that the COP is proportional to the ambient temperature at each end of the heat pump cycle. If you look at the Chiltrix graph I linked to in the third-to-last paragraph of my original post, the x-axis is outdoor temperature.
So for the heating and cooling heat pump I assumed one end was at outdoor temperature and the other at 70F. For the HPWH I assumed one end was at 70F and the other at 110F. For the dedicated external water heater I assumed one end was at 110F and the other end at outdoor temperature.
I think that's valid, but I'm not sure. Here's a question: The heat pump that's providing house heating isn't operating at room temperature on the inside, it's got to be heating the air to at least 90F and probably more like 105F. Which is close to the same temperature as the water heater. So is it really going to have a higher COP than the water heater? I don't know.
I answered my own question. I took a cheap thermometer and put it on top of the exhaust vent while the hot water HP was running. It was putting out 47 degree air. Just for fun I put it on top of the intake fan. I keep my garage at 60 degrees. The intake air was 57 degrees.
The water temperature is set at 120 degrees.
DC: Are you using the heated water for space heating, DHW, or both? I talked to Corbett Lunsford about using the smallest Chiltrix system for a very small off grid cabin and he said it was still way too big for the buildings load (and the solar PV system).
I know a lot of people have gone away from considering solar thermal due to possible leaks (and it being a complicated system) and then people considered Solar PV to heat water. Now it's air source heat pumps.
Have you done any calculations on that Artic Air/Sunamp Thermino combination I mentioned in another post? The sticking point in all the other system designs seems to be how to efficiently and easily transfer the heat into the water and store it after it's been produced by the heat pump. The Thermino seems to be a pretty good option since it doesn't actually store much (or any) water.
The problem with solar thermal was always that the sun produces the most usable energy when you need it least, and the least when you need it most. It's not coincidence that the coldest time of year is also when the days are shortest. And there's no good way to store the heat produced, even for short periods. With PV systems, if you have net metering the utility is effectively storing energy for you.
>Have you done any calculations on that Artic Air/Sunamp Thermino combination I mentioned in another post?
I'm not trying to model any particular system. Rather I'm trying to figure out the theoretical limits. My underlying assumption is that the same technology would be available for all of the heat pumps in the system so it doesn't particularly matter which technology you pick.
It probably matters what the shape of the performance curve looks like, at least in the winter, because that would move the break-even point.
I read this last night on my phone, but wanted to come back to it once I had my computer open.
I think something might be off in your exterior WH graph. Shouldn't that be the 50*C line from chiltrix directly? 50C is 122F. Their line stops at -16C it looks like, but COP remains above 1, where as your graph drops below 1. At that point, resistance heat takes over, so the line should flat line at 1.
My calculations were similar. I have a spreadsheet based on HDD for heat load, solar output, etc to model my home. In that, I added water heat. Estimating a usage of 6.4kw of water heat daily, which I think is around 35 gallons or so if I remember right. I'm saying the HPWH has a COP of 3.5. I think Rheem might advertise 4 or so actually. R134a has a little different properties than R410a in an outdoor model. I also did not add in the calculation in this portion of the spread sheet that shows the savings on AC. I made the assumption that the blue months are cooling months and the orange ones are heating pumps. But in the shoulder season, the 4200btu output of a HPWH might exceed the cooling load of a small nearly passive house.
At the end of the day, even with my somewhat high electric rates, we are talking about $20/year difference between the two options. If I used the Chiltrix for DHW, I'd have a 3 way valve on the heat pump loop, pump on the DHW tank with a brass plate exchanger between them. I think I could DIY that for under $1000 including the tank. Indirect tanks seem quite expensive, maybe it's because the only ones that exist are longer lasting than a standard $400 electric water heater? With the federal tax credit, the cost to do that is about the same as a HPWH. Assuming you have an A2W heat pump in the first place... If not, according to my spreadsheet, it would take 30 years for the Chiltrix to pay off for just DHW.
I think one reason the lines get so close together is due to how much the exterior HPWH drops off in efficiency as target temperatures rise. If it only has to make 95* water, it's fairly efficient. But hitting 120* water, the COP drops quite a bit. So the system using 95* water for heat along side a HPWH remains pretty close in total system COP.
Thanks for the detailed, thoughtful response.
>At the end of the day, even with my somewhat high electric rates, we are talking about $20/year difference between the two options.
I think that's the big takeaway. The external systems tend to be really, really expensive. They may be less efficient, but even if they turn out to be more efficient it's not going to be by enough to justify the price.
> But in the shoulder season, the 4200btu output of a HPWH might exceed the cooling load of a small nearly passive house.
That's a fair point. I had assumed that 100% of the "coolth" generated by the HPWH would be usable cooling, if that's not the case then the external becomes a little more favorable.
>I think something might be off in your exterior WH graph. Shouldn't that be the 50*C line from chiltrix directly? 50C is 122F. Their line stops at -16C it looks like, but COP remains above 1, where as your graph drops below 1. At that point, resistance heat takes over, so the line should flat line at 1.
Fair enough. I had moved the line over by 40F, comparing heating the inside of a house to 70F with heating water to 110F. But it may well be that's not a valid way of looking at it. Certainly if you have hydronic heat and run it at 50C/122F then the line for heating is exactly the same as the line for DHW and there's no advantage to the two-stage system with a HPWH. For new systems the AWHP manufacturers recommend a low water temperature, because it makes both the efficiency and the capacity better. Chiltrix likes 95F/35C if you can make it work. That puts the DHW line 27F/15C to the right of the heating line. That's less than I had assumed, which moves the crossover point to the left (ie, lower temperature).
To me, it seems like the SANCO2 has a really small business case in a market where A2W heat pumps barely scratch the surface. Even if you use it in the DHW/Radiant combo package, you can't use it for cooling. If I couldn't fit a HPWH in my mechanicals area, I'd more likely look at using solar and a DC heating element. In the winter, it wouldn't work out, but in the summer, you would only need 2-3 panels for most/all of your hot water production. Panels are down at $0.30/watt these days if you can install yourself.
For the A2W side, I intend to use one Chiltrix or similar to feed a ducted unit and non-ducted in the house and a non-ducted in the garage, with hyronics I can combine the systems. When you start to pencil out 3 mini splits, the numbers get closer. In addition, I'd like to heat the slab in the garage and some portion of the house as well. I call it a garage, but it's a shop. If I wasn't planning that, I wouldn't even think about an external heat pump water heater unless it was nearly identical in price to a HPWH.
Thanks for laying this out- I've always been a bit confused about how efficient an indoor heat pump water heater is during heating season. I actually just was in a situation where a client had a hybrid water heater installed in an area that is partly heated by minisplit, but mostly heated by electric resistance space heaters. (I recommended that they switch to resistance heating mode for the heating season). It's interesting to run the numbers and see that if the average COP of heating is even a little above 1 the COP of the water heater operating in that heated air is better than resistance.
If the heat for the water heater is coming from resistance heat you're better off just having a resistance water heater.