How much efficiency does a short-cycling heat pump lose?
DCcontrarian
| Posted in Energy Efficiency and Durability on
I know that the general advice is heat pumps shouldn’t be oversized, because it leads to short-cycling. At the same time, many cold-climate heat pumps have higher COP’s at their minimum output than at their maximum output, particularly in moderate weather. So if you just look at rated COP it seems that oversizing actually makes the heat pump more efficient. Is there a methodology for factoring the impact of short-cycling.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Search and download construction details
Replies
One proposal: model it. I know that energy plus and a few other engines (don't know about ekotrope) have built in algorithms for cycling efficiency degradation, but it would be important to also make sure you have the specific performance map input for that unit.
What is the cycling efficiency degradation based on? I haven't tracked down the literature --I don't know.
I think there are elements of this that aren't tracked in this analysis too, e.g. lineset length.
Curious for others' insight and input on this question...
It's coming up as a code question in my work. E.g.-- is it actually more efficient to oversize than to 'right-size'? That's going against the catechism of 'thou shalt not oversize' because it'll kill your efficiency. Fwiw I think they're both right.
Short cycling really doesn’t affect efficiency so much. It actually helps in above freezing weather as it allows the outdoor to defrost naturally when not running. Especially if sunny outside.
Do you have parameters and/or limits for heat pump sizing that you use?
Just thinking out loud --If short cycling has doesn't hurt efficiency, then it would seem that this a tradeoff of part load efficiency and equipment life of short cycling.
Of course, theres the latent load side to deal with, but perhaps that can and should be solved with a dedicated dehu. E.g. one box for sensible, another box to remove latent, and perhaps a third to add it back in?
I've never seen a proposal for separating latent and sensible that wasn't a bad idea.
The problem is they can't be separated. Rather, while you can do sensible without latent, you can't do latent without sensible. And if you try to do latent without sensible, while also doing sensible separately, you end up doing a great deal of work that simply wouldn't exist if you just did them together.
Elaborate, if you would dc.
Anecdote: I installed a supplemental dehu in my house in Atlanta many years ago and it worked great. Maybe I was doing it wrong, but I just didn't see any issues in my experience that I can connect with your experience.
Mechanical dehumidification means cooling the air below its dew point so that humidity condenses out. The difference between a dehumidifier and an air conditioner is what they do with the heat they extract, the air conditioner dumps it outside and the dehumidifier dumps it back inside the building.
If you are running an air conditioner, and when it provides the necessary sensible cooling it is not providing enough latent cooling, then it's OK to run a dehumidifier too, to get more latent cooling. That situation is pretty common up and down the east coast.
What's not OK is to run a cooling device that doesn't provide dehumidification, and then make up for that with a dehumidifier. You're paying to cool the same air twice and heating it up in between.
My Rheem communicating thermostat allows me to watch all the unit’s sensors and operating parameters in real time.
For the first minute or so of the run cycle the HP makes almost zero heat mostly it is building up pressure differential required to make the magic happen. I believe the superheat parameter is telling me about how efficiently the unit is operating at any given moment. A lower the super heat is more efficient. After a cold start I see a 2 dight super heat over the next 15 minutes or so the computer slowly opens the electronic expansion valve and get it down to 2or 3°.
What I am saying is if you only look at the power graph from the breaker it looks like the HP goes from off to peak efficiency in 10 milliseconds and that is simply not the case.
I have the opposite point of view for HP sizing the way I see it you get a more comfortable system if the HP is undersized with back up to cover the very rare coldest hours. If the unit is smaller, it cycles less under low load conditions making it even more efficient when the COP are likely to 4-7. The smaller unit will always be running closer to its max speed and making warmer air so you feel more comfortable most of the time. When the back up heat does kick in the outdoor temps will likely be in the single digits and the HP’s COP will be at its lowest 1.5-2.5 or so. This gives you more run time with the highest COPs.
The biggest problem with oversizing happens when to start cooling and the run times are so short to dehumidify the home. Forcing the residences to lower the set point and giving you the classic complaint of a cold and clammy house.
Walta
Thanks walta. What you're saying makes sense to me. Would you speculate on the impact to your systems overall efficiency for that startup vs. runtime?
Supplemental commentary
As background for my context: the title 24 code (which is the applicable code for my work) will be adopting specific provisions to disallow your sizing approach due to concerns about supplemental heating. Comfort wasn't a formal criterion in this change. I wouldn't be surprised if this concern gets raised in other code processes
Do you have to size for more than 100% of the 99th percentile heating load? Because even then you're going to have 88 hours a year of needing supplemental heat.
I do get the systemic concern of everyone turning on supplemental heat at once during a cold snap.
Dc, not yet. But my understanding is that that's coming. Specifically that heat pumps will be required to be sized to achieve the design heating load without supplementary heat. And that corresponding design temperature must at most the winter median of extremes. Lower is fine.
Absolute zero? Acceptable.
"And that corresponding design temperature must at most the winter median of extremes. "
So I understand what that means, here in DC the 99th percentile heating temperature is 22F. According to the NWS in a typical year we get one hour at 6F and zero hours at 5F. So the system would have to be sized for 6F, rather than the 22F that Manual J uses?
A house that has a heating load of 30,000 BTU/hr at 22F would be 42,632 at 6F. So basically a 50% bigger heat pump when you factor in the derating from 22F to 6F.
So this gets to the question I'm asking in this thread. My estimate is that in that house a four-ton Mitsubishi HyperHeat would start short-cycling at 38F, and a three-ton would start at 43F. So what is the impact of those extra hours of short-cycling? Does it more than make up for almost never running supplemental heat?
I don't know.
My wild guesses at numbers don’t seem likely to add much to the conversation. I don’t have a way to convert my super heat numbers to COP. The first minute is totally wasted and it is 15 minutes before it gets back to peak performance.
As for title 24 all I can say is I am glad there is no enforcement where I live. It sounds like they are trying to make a one sizes fit the world rule. I am pretty sure that my electric providers peak summer usage is about 4X the winter peak because so many people heat with gas but every one has AC.
Walta
I'm looking at the 2025 revision of Part 6 of Title 24 and it says:
Heating design temperatures shall be no lower than the 99.0 percent Heating Dry
Bulb or the Heating Winter Median of Extremes values.
The part in bold is new for 2025. It seems you're welcome to use a warmer design temperature if you wish.
I don't know what the "or" in there means, it seems like poor drafting.
Indeed. Well done,. But youre looking at 15 day language, not the final version. I know the intent was that it should be "no greater than". That may not make it into the 2025 code in a month, but we'll see. If not, then 2028.
That 'or' structure was definitely commented on. Yet it remains...
As it reads it's an anti-oversizing measure, not anti-undersizing. Since the 99th percentile temperature is always going to be warmer than the median extreme, the added language -- whatever it means -- seems to change nothing, you can't oversize by more than the median extreme.
What would make sense to me would be having both undersizing and oversizing limits, that the design temp has to be between the median extreme and the 99th percentile.
I concur actually. But then I couldn't figure out what all the hand wringing was about the rest of the heat pump language. And through personal correspondence that I'm not able to share here, it was pointed out that this was a big error. The intent of this code was absolutely to prevent undersizing, not prevent oversizing, and they just missed this particular part.
Note later on that it references Manual S for System selection, and then effectively undermines the oversizing limits with the statement 'There is no limit on the maximum heating capacity [of heat pumps]'
Later on, there is a limit on heat strips-- no more than 2.7 kW per ton which intended to address comfort issues during defrost. In the 2028 code, there are likely going to be some significant clarifications on all of this.
Here's the paper which backhandedly describes the problem statement:
https://www.aceee.org/sites/default/files/proceedings/ssb24/pdfs/Applicability%20of%20Cold%20Climate%20Heat%20Pumps%20in%20California.pdf
This is my gut feeling as well, just observing heat pumps that it takes a few minutes for them to reach full output after cycling on.
Most tend to have some sort of control for minimum cycle length, so under low load conditions they will overshoot the thermostat rather than short-cycle. This is good for efficiency (and equipment life) and probably doesn't effect comfort.
The units I have seen are happy to turn the compressor off the moment the thermostat is satisfied and maybe run the fan for a bit.
The short cycle the equipment manufacture is concerned about is that the unit stay off long enough for the systems pressure to equalize before the unit restarts.
Walta
OK, so then the thermostat overshoots in the other direction.
Yes it could under shoot when off.
The short off cycle the manufacture is concerned about is an off cycle of less than 3 minutes so the compressor is not damaged.
Just how far can the house get from a satisfied at it set point in less than 3 minutes?
Walta
You also have watch your local code. Ours treats a heat pump that can fully carry the house the same as a fuel burner for envelope requirements.
As soon as you add resistance heat, the building is now treated as electric heat and you need to bump up R values. Not that there is anything wrong with more insulation, but does add cost.
Not a direct answer to your question DC, but I remembered this presentation and thought you may find it interesting. It doesn't directly pertain to Size, but rather to other performance factors for heat pumps.
https://drive.google.com/file/d/1X8Q78jKffLXLK8Ndeif7_h1yZFb4aAjP/view
quoting from the presentation:
"Low-Load efficiency is important to seasonal performance in all climates.
Minimum Capacity COP@47°F appears to be the best indicator"
That is-- chose the best load load efficiency OVER other factors, specifically low temperature capacity (even with supplemental heat impact) if you're interested in overall equipment efficiency.
A few thoughts
The problem with undersizing is that heat pumps lose capacity at low temperatures. When dealing with a 1 percent temperature night, you are simultaneously losing heating capacity and you have undersized the heating system. This seems a poor choice with those lines racing towards zero, compared to a fossil fuel system, which may be undersized, but retains that capacity at any temperature.
The solution to me would seem to be more than one heat pump. While small houses with very high insulation values can operate on one unit, most houses need something to cover bedrooms or upstairs. This gives you the opportunity to stagger them in shoulder seasons. Run the downstairs one or the bedrooms one and the frig running and the coffee maker put out enough heat to keep the main area warm.
I don't think you could convince me to buy a 'large' heat pump to run a whole house. i would sooner have two small ones unless compressor siting was an overwhelming issue.
If one had a very low heat load such that only the smallest heat pump made real sense, putting some of those 200 dollar electric wall heaters in the bedrooms for the 1 percent nights might make economic sense. Odds are it is already over insulated if the heat load is that low.
We have four Fujistu cold climate HPs between 3/4 - 1 1/2 ton in zone 4a. During the shoulder seasons we pretty much just run the 3/4 ton ductless in the main living area. It's just about runs non-stop at a very low speed. It's very comfortable. The 1 1/2 ton ducted is sized for an addition on piers that has a large amount of glass. It short cycles alot unless it's below 20f without sun.