What is Right-Sizing?
First– apologies– this is a Can of Worms question. Feel free to ignore– I’ll admit this is a good faith trawl of the hive-mind (as opposed to bad-faith troll)
I was just having a debate with a colleague about what the benefits of right-sizing are, and I got hairy-toe phenomenon (when you stare at your toe too long and you kinda don’t recognize it as a part of your body anymore).
ACCA manual S says 15% for heating and 40% for cooling (that’s from memory– please correct me if I’m wrong)– and that’s relative to Manual J.
The two main arguments come down to efficiency and longevity. Both seem to involve subjective claims and make a lot of presumptions about the nature of the system and the loads. Presumably, if you vary from these Manual S limits too much, you’re sacrificing either or both.
The gravity of this question is that in a jurisdiction that I work, ACCA Manuals are required by code, and the inspectors will compel systems to be ripped out and replaced if the installed system violates Manual S by one iota. I’m not questioning that these numbers are reasonable guidelines for design work, I’m wondering how prudent they are as a legal standard.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Right-sized equipment will keep your home comfortable on the coldest day of the year as well as the hottest day of the year. If the equipment is right sized, it will operate for 60 minutes per hour under peak conditions.
I don't know what you mean by "15% for heating and 40% for cooling." Fifteen percent of what? Forty percent of what?
Thanks Martin--
I assumed and didn't fully articulate the question.
For the 15%/40% factors-- that's specifically 15% and 40% oversizing allowance on the calculated peak loads.
So... following from your definition of right-sizing:
A house had a peak load of 40 kBTUH in the heating season. Let's assume for simplicity that there's no cooling. The installed HVAC system has a nominal rating of 200 kBTUH (this is a real project).
So long as that system runs at 60 minutes per hour on the coldest day of the year, then the system is 'right-sized'. That makes sense to me...
What this seems to be saying is that the load and the nominal ratings are not dispositive to right-sizing.
Did I extrapolate from your statement correctly?
To your point that code enforcement is condemning systems that exceed what is call for in the manual calculations. To date most of the enforcement people are only reading the one number at the end of the calculation. It does not take a genius to change a few of the input numbers and get what ever number you want at the bottom on the form.
You may find this video interesting.
https://www.youtube.com/watch?v=_hAuKtoRxJI&ab_channel=TechnologyConnextras
Walta
Thanks Walta-- yes-- I actually watched that video right when it came out-- that's why it's been on the brain. He has a triple-oversized furnace, and he makes good points about not relying on that sizing for any future changeout. I didn't find it clear that his furnace inappropriately oversized.
And yes-- I understand about adjusting the numbers to nudge towards certain outcomes. I'm not actually asking about compliance per se, but rather the rationale supporting codes which compel this sizing process and outcome.
Pertinent article from Martin circa 2010:
https://www.greenbuildingadvisor.com/article/saving-energy-with-manual-j-and-manual-d
"[...] If you’ve been paying attention to energy-efficiency experts and green-building Web sites, you probably know that it’s important to properly size your HVAC equipment. Most sources repeat the same advice: oversized furnaces and air conditioners cost too much, waste energy, and sometimes provide lower levels of comfort."
Later on-- Martin also puts this in some doubt:
"The conventional wisdom may be wrong, however
Increasing evidence shows that energy experts have exaggerated the negative effects of equipment oversizing, however. Studies have confirmed that oversized furnaces don’t use any more energy than right-sized furnaces. Moreover, newer modulating or two-speed furnaces operate efficiently under part-load conditions, solving any possible problems from furnace oversizing.
Although there are ample reasons to believe that oversized air conditioners are less effective than right-sized equipment at dehumidification, at least one field study was unable to measure any performance improvements or energy savings after replacing an existing oversized air conditioner with a new right-sized unit."
It's now 14 years later-- is there any new data to help resolve these competing claims?
Combustion appliances tend not to be as sensitive to oversizing. In particular traditional boilers held a lot of water, they had a lot of heat capacity and the aquastat was typically set on a 20F range so even if there was a low demand for heat they didn't short-cycle too badly. Also, a bigger boiler doesn't cost that much more, often within a line the only difference between different outputs is the nozzle size on the burner. The penalty for undersizing a boiler is pretty severe, you have to tear it out and replace it, so the bias among installers was toward oversizing, which got us to where we are today where oversizing by a factor of two to three is the norm.
With heat pump equipment all of that is different. Electric motors really don't like to short cycle. Cooling equipment has trouble dehumidifying when it's oversized. And the cost of a heat pump goes up pretty dramatically with capacity.
Cooling equipment has trouble dehumidifying...
Yes indeed. I don't deal with latent load much, but when I do, I like to consider it as its own separate load and separate equipment.
On electric motors and short cycling-- at what frequency is a motor 'short cycling'. And why doesn't it 'like' short cycling? What degrades? (I think I know, but I want you to state the obvious).
I would also think that variable speed heat pumps aren't as sensitive to these particular issues?
Cost for systems is important, but I'm already convinced of this. Why buy more system than you 'need'?
But that doesn't seem an appropriate territory for code.
Code doesn't consider DHW systems in the same way, right? (Perhaps they should?)
Why shouldn't you undersize boilers? I think this is one of those things I forgot to know in the first place.
“ Why shouldn't you undersize boilers?”
Because you’ll be cold if the boiler is smaller than your heat loss. Size it to load.
Several of the questions in their tone suggest a lack of understanding of what sizing means. This one, and the one about a 5-times oversized system running 100% of the time.
Point well taken DC. To be sure, I am naive about something... I'm just trying to suss it out.
On undersizing, you claimed that undersized boilers had to be ripped out and replaced ... Was that just due to comfort? Or was there something more technical with extended runtimes? If it's the former, then I don't totally agree, but I take your point.
On the 5x oversized -- I often see equipment that can modulate that much... I see inspectors fail such projects and until equipment is 'right sized'.
As far as I can tell, the equipment is right sized according to Martin's definition. It will plausibly run 100% of the time on the peak day.
Heck it will even run 100% of the time on most days with significant enough load.
If your house has a heating load of, say, 50,000 BTU/hr, and you install a boiler with an output of 40,000 BTU/hr, you will be cold in the winter because the boiler won't be able to keep the house warm. There's no adjustment or tinkering you can do, you have to replace the boiler with one that has enough output.
It's true that modern compressors modulate and can run at partial output. So it's not run time that's the right metric, it's output. Properly sized equipment should run 100% of the time at 100% output on the design day.
The design day is picked as the 99th percentile temperature, 99% of the time it's warmer than that. While the equipment may have no problem running at 25% output on the design day, what you want is for as much of the heating season to be within the modulation range as possible.
I simply don't agree that there's no adjustment or tinkering one can't do in that case with a 10kbtuh undersizing discrepancy.
Perhaps there's nothing to do with the boiler, but there's all sorts of ways to add supplemental heat.
DC-- you pose a different definition than martins for right sizing... I appreciate that. And that seems to be the most common implied definition out there.
To sharpen my question --whats wrong with wrong sizing? We've covered undersizing, but what about oversizing? I'm interested in efficiency and performance issues. Those are the ones that I see debate and, frankly I see the problems aren't with the sizing per se, but in other system attributes that result in those performance issues.
For sure example-- cycling can be annoying for occupants. I wonder why they even notice that the system is running... It should be virtually unnoticeable. That's a problem in design and install... Not oversizing specifically
I simply don't agree that there's no adjustment or tinkering one can't do in that case with a 10kbtuh undersizing discrepancy.
Perhaps there's nothing to do with the boiler, but there's all sorts of ways to add supplemental heat.
If the customer contracted for a boiler that meets the heating needs of the house, and they paid for a boiler that meets the heating needs of the house, they deserve a boiler that meets the heating needs of the house.
A heat pump running at 20% of capacity on the coldest day is not right sized, even if it can modulate down to 20%. I would have thought this was self evident. This system will be cycling 99% of the time, and short cycling a majority of the time. Note that just because it can modulate to 20% of capacity doesn't imply it's operating at the same efficiency. It's also needlessly expensive.
you say this system will be cycling 99% of the time, and short cycling a majority of the time.
I believe you
how is this different from a single-speed heat pump where we design the system to cycle 99% of the time? indeed, that's called 'right sizing ' according to both Martin and DC.
when does cycling turn into short cycling?
You're right, it's like a single stage system 1/5 the size, except it costs 3-5 times as much. So at best, you've negated the point of having a modulating system. It's intended to be so the system is more efficient at lower loads, not so you can 5x oversize it.
"when does cycling turn into short cycling?"
With a modulating compressor, any cycling is short cycling, it means the load is below the minimum modulation.
With a single-speed compressor, short cycling is when the run time is being extended beyond what the thermostat is calling for by the anti-short-cycling feature of the controller.
THANK YOU DC
Defining short cycling in the way that you did-- THATS something I've never heard articulated before.
It is interesting that there isn't a universal definition...
Indeed it depends on the system type. We're gradually establishing a warrant for the claim...
@ Trevor--
As far as costs are concerned-- I'm not seeing in my market that a 5 ton system costs 5x of a 1 ton system. I'm seeing a cost premium of maybe 10-20% on those different tonnages on the total system installed cost. For the unit cost, I do see 2-3x on those costs, but that's minor in the grand scheme of things. Equipment costs are roughly 1/4 or less of the total system installed cost for my projects.
Paul-- yes... Point taken. Undersizing seems destined to result in comfort issues. But I wonder about how much that would occur in reality... To wit-- the fear of undersizing results in everything being oversized.
There is that discrepancy between calculated and actual loads, so there's a little wiggle room there.
That is, the only way to practically achieve Martin's definition of right sizing in most homes is to undersize a bit. But, that's more of a rounding error in the end.
When I say undersized, I mean undersized in reality. That’s different than undersized to manual J, considering how inaccurate they can be.
A simple solution would be to include resistance backup for a heat pump that’s slightly “undersized” to check the box, then disable the resistance.
For sure example-- cycling can be annoying for occupants. I wonder why they even notice that the system is running... It should be virtually unnoticeable. That's a problem in design and install... Not oversizing specifically
Oversizing is much more than an annoyance. When an electric motor starts, it consumes a lot more power than when it's running. This is called inrush current. Inrush current is highly stressful on all of the electric components, it's basically what kills motors and relays. So short cycling shortens the life of the equipment.
If that motor is attached to a compressor the problem is even worse. When a compressor shuts off at the end of a cycle the gas in it is under pressure. If you restart the compressor right away that pressurized gas presents resistance to the motor, and makes the inrush current quite a bit higher. To prevent this, compressors are designed to slowly release their pressure when they are off, and the control electronics for the compressor are designed to keep it from restarting within a certain time period, typically about five minutes, of when it last stopped. The refrigerant in a compressor contains lubricant, and when the pressure bleeds of the compressor is slightly under-lubricated when it starts. To protect the compressor they're also programmed to have a minimum run time each time they start.
So when the system is oversized and short cycling, the cycles will be longer than what the thermostat calls for. The heating and cooling that is produced has only one place to go, into the house, and so you get swings in temperature that are detrimental to comfort.
Cooling presents a special problem because the system needs to dehumidify. The way dehumidification works is that when warm air passes over the cooling coil, dew forms. The dew drips off of the coil into a tray, where it flows out of a drain and is disposed of. Short cycling disrupts this because the cycle isn't long enough for the dew to form and drip off the coil. When the cycle ends, whatever dew is on the coil just evaporates and goes back into the house air, no dehumidification happens.
A house with an oversized system is uncomfortable. Comfort is the entire reason for being for HVAC.
Excellent.. that's the kind of detail I was looking for.
Do you think this is a problem that merits code language and enforcement?
Code started out as concerning itself only with health and safety, basically fire safety, not mixing potable water with sewage, and making sure the house wouldn't collapse.
Somehow along the way it started taking on usability and functionality. With plumbing minimum drain sizes to avoid clogs, minimum supply line size to ensure flow. With electricity avoiding nuisance trips. Minimum ceiling heights.
About 20 years ago was when things really changed with the introduction of energy codes. This is a pretty big departure from where codes had traditionally been, mandating that houses be built to a certain level of energy efficiency. If that's the code, then mandating that HVAC systems be properly sized is entirely consistent.
Thanks for responding to that question DC--
That's the impetus for sure. And while I don't want to puts words in your mouth, it seems like you'd agree that such codes are prudent?
I'm still thinking about the tradeoffs, especially when it comes to heat pumps. I saw some research recently that presented a paradigm for efficiency that was substantially different than the Capacity=Peak Load paradigm.
If people are interested, I can try to dig it up and link it here. But the gist of it is hinted at below-- size for maximum COP47, not for Peak Load.
That paradigm is definitely controversial. Energy Efficiency is fundamentally a subjective metric.
I've been practicing the ideology around right sizing for pretty much 20 years.
I've download calcs and 100% of them have been rejected in favor of the subs own estimations. This is a nearly universal experience on this forum.
I've been waiting for years for hundreds of see-I told-you-so with respect to comfort complaints and premature equipment failures
But to date, not one has come in
For a problem as rife as oversizing, I was expecting a little more opportunity to be smug and teach the industry a lesson. (Not that I would actually be a jerk about it).
You can see the amount of oversizing I see-- 5x is pretty common.
Something should have come through by now, right?
(Sigh)
We get people here all the time with multisplit installations that won't modulate properly because they're oversized.
The punch line on my 200, 000 BTUH example was that the sizing was all done by the book. I gave the block load of 40kbtuh, but with all the zoning the engineer came out with the 200kbtuh sizing according to the manufacturers sizing software
The punch line on my 200, 000 BTUH example was that the sizing was all done by the book. I gave the block load of 40kbtuh, but with all the zoning the engineer came out with the 200kbtuh sizing according to the manufacturers sizing software
I'm not getting what you're saying. Could you post more detail? I'm not sure how the zoning works into it unless it's a bunch of 1:1 minisplits.
same idea: 20+ zones on a single outdoor 'unit' (really a pair)
ill admit that I didn't double check all the math from their engineering. but total indoor unit capacity exceeded outdoor unit capacity so it made some intuitive sense to me. obviously my intuition here is in doubt...
(reply to #26)
That may not be that bad. What kind of system and what is the rated capacity of the outdoor unit?
What's currently proposed is a City Multi system. Mostly ducted medium static air handlers. Project also has hydronic floors-- so radiant system is heated with an HTP 199kBTUH unit with 119 gallons of storage.
I think if you’re used to oversized furnaces, then you won’t complain when your next furnace is also oversized - even if you do complain, will someone pony up to replace it?
I know that switching from a 90kbtu furnace to a 24kbtu heat pump (still oversized) was great for my own comfort but that’s not swaying many minds.
It drives me crazy when I go to other people's houses and their furnace cycles every five minutes but I guess most people don't notice.
From the contractor’s point of view the only real risk come if he undersizes a unit.
If he undersizes the equipment the customer calls on the coldest day of the year and says you screw up and my house is cold and he can’t win that argument.
If a contractor oversizes the equipment the customer is slightly less comfortable, the customer pays slightly higher fuel bills, the customer buys replacement equipment a few years sooner. He can fast talk his way around those arguments with ease.
I am not advocating for oversizing just pointing out from the contractor’s point of view it is the safe bet. I was able to talk my contractor into a smaller HP by allowing him to install hugely oversized strip heater. It was a win win because I got the right sized HP that covers 99% of my heating and he could be certain I saw never going to call and say my house was to cold.
Walta
There is a proposed code amendment where I am to prohibit exactly the approach that you took for yourself, walta. That is, it would require heat pumps to be sized for the design load without supplemental heat. Precisely for efficiency reasons.
From my limited foray into heat pump systems , oversizing occurs because:
A. A Manual J was not done aggressively (or not done at all). Instead the 20th century thumb rules are used: 1 cfm / ft2 ; 1 ton per 400 cfm.
B. Ductless Head count drives the system size. 5 heads (LR, 3 BdR and a finished basement ) dictates outdoor units with 5 ports (either one OU with 5 ports or a 2-port and a 3 port).
And the smallest capacity heads are 6kBtu, which is 4.5 x the peak demand for the primary bedroom.
C. Minisplit heating capacity is likely greater than labeled capacity , especially with cold climate units.
We’ve refit our ranch and the manual j is 17kBtu for 5F / 75F.
In the first round of quotes the system sizes ranged from 24 kbtu to 60 kBtu.
We found an outfit that would right size. Although the proposal started as an 18kBtu OU with air handler and ducts, they didn’t want to go down to 15 kbtu because there was no air handler at that size.
Ultimately we got two independent 9kBtu mini splits.
One 9kBtu Minisplit can deliver 14.6 kBtu at those conditions. The runs are short, the indoor unit is a wall mount and If we keep the filter clean it delivers. It’s probably enough, at 5F and 65F in the LR and 55F in the bedrooms and 50 in the basement. But … it does get colder, so a second system addresses that as well as redundancy.
In my experience-- it's option B which really drives the sizing in a by-the-book oversizing.
In my work, I'll have relatively small homes (up to 1200 sq.ft.) with 3-6 zones. This is compelled by energy code compliance (which requires an indoor unit in any habitable space >150 sq.ft.)
3-6 zones in 1200sqft will simply not work well. At best you are looking at a COP in the 1.5 to 2 range for heat. I guess if local electricity is cheap, this is not an issue but a very bad idea overall.
I have an oversized multisplit feeding bedrooms and it doesn't dehumidify worth squat, overcools and overheats the space. Luckily it is not primary heat but I have seriously contemplated ripping it all out and replacing it with a properly sized slim ducted unit.
Indeed Akos
And electricity happens to be $0.38-$0.62 per kWh.
Mine is sized for my design load in fact my strips are locked out unless the outdoor temp falls below 7°F maybe one day a year.
Walta
Still likely illegal with this code proposal-- you'd be allowed 2.7 kW per ton of cooling.
BTW-- I don't agree with this proposed code. I think what you have is entirely prudent...
What I'm seeing is all sorts of different subtle and profound differences what people think of as 'right-sizing'. And now code entities are starting to weigh in so that it's not just a guideline or a reasonable standard of practice, but a mandate. This seems a bit of an overreach.
If I were trying to optimize for efficiency (on my terms-- minimize personal cost), I would specifically select the unit that has the highest part-load COP at 47 degrees at that home's rated load @47. That would be for Billings, MT. Then I would probably forego integrated heat strips in favor of space heaters-- just like your video presented... I thought that was a reasonable approach...
Sizing a basic furnace/heat pump (single-speed) system to run 100% of the time at the design load still means that the system is oversized and cycling for 100% of the time many years (since many years don't experience a design temperature). Cycling is just a reality to the nature of the system. Cycling of motors is stressful no matter how often its done-- however there are nonlinear effects/damages that happen with increasing cycling frequency that DC described-- this pushes 'normal' cycling into 'short cycling'. But it still seems to be an arbitrary distinction.
About cycling: variable speed compressors have no particular current spikes at startup. From my observation, the drawback of cycling with theses systems in heating mode is the feeling of a temperature swing inside. This swing can be lowered by using shorter cycles but then there is an energy cost because each time the compressor stops, pressures balance out which cools inside coils and heats up outside coils. At startup, 3mn of run is wasted to get the temperature difference back to a nominal value. Obviously these drawbacks are amplified with single speed systems and there is the added drawback of current spikes and wear due to frequent startup...
To me, the underlying question is what are the objectives of the regulation: these and the wishes of the end user are different, and may also differ from those of the contractor.
- end user wants maximum comfort for the smallest possible cost. The trade-off varies as well as to what extent the initial system cost is taken into account. Some people will tolerate somewhat less comfort to reduce their bill or impact, others want premium comfort and cost is less a problem. Some may want to pay more for the system and pay less energy later etc.
- also the objectives of the regulation may vary: protect the end user from too bad design leading to discomfort or high energy bills, optimize energy usage at global scale including other uses, and here the optimum depends on which energy is locally available when etc.
For example, optimizing between more or less heat-strips and more or less heat pump capacity has to do with a compromise between system and energy cost, comfort, resiliency for the owner and, on the other side, state-wide optimization of the grid.
Another example is the choice between ducted and ductless with many heads. The second may have lower initial cost but is probably less comfortable, less visually appealing and may or may not also be less efficient and durable.
Thank you Casimir!!
This is a thorough and thoughtful consideration...
I share your perspectives on the 'it-depends' kind of answer here...
My clients tend to be much more on the premium comfort/premium cost side of things. But 'comfort' is rather subjective... by comfort, they really mean control, thus 20+ zones when really 3 would do perfectly nicely.
and if 3 zones are needed, overall outdoor unit capacity can be brought much closer to the block load of the structure.
Portable space heaters are a very bad idea! Many people die every year because of them.
Walta
I wish there was an upvote button!
It's an argument that right-sizing HVAC is a health and safety issue.
agreed, I used to work for a nonprofit that helped people recover from such fires.
that said, there are reasonably safe space heaters.
one friend of mine told me about using some GPUs for his house-- distributed around for good measure. she is was mining crypto-- the waste heat was just a bonus. it seems to me that computers don't have the same fire risk as cheap wires and fans...
@ Trevor--
As far as costs are concerned-- I'm not seeing in my market that a 5 ton system costs 5x of a 1 ton system. I'm seeing a cost premium of maybe 10-20% on those different tonnages on the total system installed cost. For the unit cost, I do see 2-3x on those costs, but that's minor in the grand scheme of things. Equipment costs are roughly 1/4 or less of the total system installed cost for my projects.
Hypothetical--
A project has a design load of 24kBTU-- let's just say the design temp is 17F.
The project is out in the foothills, so no gas and propane is a pro-pain-- heat pump it is. Also, electricity is expensive ($0.60/kWh), so she wants to minimize heating bills.
HVAC sub works with Carrier, and there are two options available locally at the warehouse.
A 3ton Carrier unit (Rated at 33kBTUh @ 17 F) and a 2 ton Carrier unit (rated at 24 kBTUH).
AHRI #s are 205272415 and 9892749 respectively.
From a 'right-sizing' point of view, it looks like the latter 2 ton option is optimal, but from a lowest-operating cost perspective, the 3 ton unit seems better. She chooses the latter as a result... for her, it seems that right-sizing is wrong.
Ohhh yes. For efficiency, oversizing is often more efficient!
This hypothetical example is very different from what you were talking about originally, a 1.5X of design load vs a 5X of design load. Choosing a 3 ton unit when a 2 ton unit will do is a reasonable consideration, whereas I'd say choosing a 10 ton unit in the same scenario is unreasonable. I didn't mean to suggest that right sizing meant you cannot exceed the design load at all. Right size is a range, and the minimum is 1X of design load. Maximum is debatable, but optimal is probably in the 1.2X to 1.5X range
Fair enough... But why is 5x different? I suppose the burden of proof is on me to show the efficiency argument...
For the 5x scenario (which is really 10x since they have fully redundant heating systems), the issue is that there's a competing definition of right sizing -- and thats the engineering manual for the system (corresponding design software --
Diamond system builder in this case). To wit-- manufacturer is dictating sizing and that's WAY different from Martin's and DCs definitions here.
Also, while you didn't mean to say that right sizing can't exceed design load, martin and DC did. Their definitions preclude that.
If you had a definition of Right-sizing-- what would it be? Is there an 'oversizing' factor that you would allow as reasonable? Or something else?
Re: reply #64
I can't speak for Martin or DC, but I don't see where either said that. Martin only posted one reply, and it's definitely not in that reply.
I don't agree that it's a "know it when you see it situation". It's a definable quantity based on quantifiable things. I'm not qualified to say what it should be. I have an opinion on the approximate limits; at minimum being able to handle the peak load, with perhaps a safety margin built in, up to some modest amount of oversizing (probably less than 2X). But whatever it is, there's some range that experts in the field should be able to arrive at some consensus on. It's definitely not "know it when I see it", and in fact I don't even think that methodology is appropriate in the porn example. If you can't point to something objective, you will have lots of disagreements over whatever the topic is you're talking about.
Your opinion of Potter Stewart's take there is shared by many!!
It's a tension in the debate-- if you create an objective standard and that's a legal one, then you fall into a 'yah/nein' dialectic where reasonable systems are rendered illegal. Crafting exceptions is hard stuff. If it's subjective, then it's much harder to enforce, though cultural scorn sure helps. Trying to do what we do here-- make right-sizing a kind of norm in the HVAC world is still good work to do.
Martin said:
"If the equipment is right sized, it will operate for 60 minutes per hour under peak conditions."
This means that the HVAC system is sized for the load, and not one iota more.
"This means that the HVAC system is sized for the load, and not one iota more."
It doesn't actually mean that. A 5 ton rated system with a turn down ratio of 5:1 will run 60 minutes per hour all the way from 5 tons down to 1 ton.
In order to mean what you suggest, you'd need the qualifier that it's running at max capacity for the full 60 minutes. That may have been implied, but it can't be inferred.
I agree Trevor!!
I didn't say that Martin meant that Nominal capacity had to match the load. (though he may clarify what he meant if he wants). You were clever to find that wrinkle. But you do seem to be saying that if the turndown ratio is 5:1, then 5x nominal oversizing is still 'right-sizing' according to your definition?
DC did clarify in their (sorry DC-- don't want to presume pronouns) response that they meant was that max capacity equals design load.
You can see I'm trying to elicit more formal language to capture what we actually mean by right-sizing. I know it's a bit academic, but it's also something that we seem to be talking about different things when we say 'right-sizing'
Furthermore, if we don't 'right-size' are we necessarily 'wrong-sizing' ? Or is it more of ... 'left-sizing' LOL!
This conversation leads me to two sizing questions that I don't know the answer to:
1. Typically the same heat pump is used for both heating and cooling. And typically, the nominal heating rating of a heat pump is about the same as the nominal cooling rating. If you live in a very cold place, your heating load can be a lot bigger than your cooling load -- like four times as big. So how do you size? (I live in a place where cooling loads and heating loads pretty much balance.)
2. If you have a modulating heat pump, it has three modes of operation:
a) The load is less than the minimum modulation, so it cycles on and off.
b) The load is within range of the minimum modulation and it runs continuously.
c) The load exceeds the maximum modulation, it runs continuously and supplemental heat is needed.
Modes a and c are inefficient and are to be avoided. Mode b is the most efficient, and my understanding is that the highest COP's are experienced with the lowest modulation. So if what you need is 12000 BTU/hr, you get a higher COP with a 2-ton compressor modulating to 50% than with a 1.5-ton compressor modulating to 66%.
If everything I just said is true, how do you size for optimum efficiency? Presumably you want to oversize at least by a little to avoid mode c. The question becomes at what point do you run into diminishing returns with oversizing, where the improved COP in mode b gets outweighed by spending less time in mode b and more time in mode a?
This doesn't seem like an obvious question to me.
I think cold climate units typically have higher heating capacity than cooling. At least that was the case when I bought my Fujitsu: nominal 9k cooling, 12k heating, max 12k/22k. That worked out about right for my house in climate zone 6. I should note that the majority of my cooling load is latent. Once you start getting up into zone 7 and 8, your cooling load not only goes down, but in most cases it shifts more toward a sensible load. That means the right sizing of the cooling side is less critical. Your cooling season isn't very long or energy intensive. Lots of people get by in those zones without any cooling at all.
"If everything I just said is true, how do you size for optimum efficiency?"
An accurate answer to this is case-specific. You'd have to compare the detailed performance specs of specific models against the needs of the specific building and climate.
"cold climate units typically have higher heating capacity than cooling. "
That's a really good point that I wasn't thinking of, that there are different types of heat pumps for different climates.
It would seem then that part of any efficiency requirement be that a heat pump be appropriate for the climate it's installed in. I could see labeling heat pumps with climate zones they're approved for. I think this would go a long way toward addressing consumer confusion about heat pumps.
YES thank you DC
These are the kinds of questions I'm asking myself. Though I'm in a different and more naive place...
For example -- it's not obvious to me that option A is actually less efficient. Everyone makes that claim, but I can't connect it to anything technical yet. Cycling is annoying, and cycling will degrade the lifetime of equipment, stipulated
But is it actually less efficient? Where does that inefficiency come from?
In my modeling work, there is a cycling efficiency degradation factor added in. And even with that factor, oversized equipment tends to fare better in annual energy consumption. And that's for a single speed heat pump.
I have a notion of where that inefficiency comes from, but that's highly idiosyncratic to each install
I find the advanced sizing tool on the neep site very helpful (https://ashp.neep.org/#!/product_list/). Once you click on a heat pump, there is a icon for "Advanced Data-System Sizing" that lets you input your ZIP and design load/temp details and comes up with a very handy histogram including summary (see attached).
I tried before to input a couple of systems that I know work well and ones that don't. Seems to be something with 70% modulation runs well, bellow 40% you are running into trouble. A couple of percent of backup heat doesn't really budge your yearly energy costs, more than 10% and your equipment is too small.
I think in cold climates cooling season is short enough that lower efficiency by using an oversized heat pump just doesn't matter. Almost all your energy costs are for heating.
Yup! I was consulting this for my equipment selection above!
Thanks Akos. I have found those graphs very instructive. It's where I got the idea of the three modes that I expressed in post #51 -- my "mode A" is their "low load cycling" and my "mode B" is their "load modulating."
However, like almost everything about sizing, the focus is on making sure you have enough capacity and not on efficiency, there's nothing in there about energy usage.
The annual heat usage by temperature -- the blue bars -- is the starting point. What you would also need is COP by temperature -- corrected for the utilization at that temperature. So if, for example, you need 600K BTU/yr at 27F, and at 27F your heat pump modulates to 50%, you'd need to know the COP at 27F and 50% to figure out annual energy usage. And you'd need that COP for every point on the line.
Then, if you had a competing heat pump that only modulates to 40% at 27F, you could do the same kind of analysis and compare the total annual energy usage.
I'm not aware of anywhere where that level of detailed analysis is done.
Well said DC. That kind of modeling is precisely my work. This is also part of the genesis of the question...
In my modeling work (which is , or courses , indicative and not precise)
I'm often getting results that favor non-right-sized systems over right sized ones. Codes require both right sizing and optimal efficiency, and those often aren't the same thing.
"I'm often getting results that favor non-right-sized systems": can you elaborate on this ?
In a mild climate like where I am (official design temp -10°C), inside the city we extremely rarely see temperatures below -5°C, and the daily average is rarely below 0°C. Given the derating of heatpumps, if you size to be at 100% capacity at the design temp, the heatpump will rarely work at more than 50% capacity, and it will be below 30% capacity 50% of the time. 50% capacity near the freezing temperature is ok because of occasional tough conditions (snow, freezing fog). But oversize this a little bit and you get nearly permanent cycling depending on the heat pump modulation capacity. Now oversizing an AWHP with a buffer tank is less a problem that with a multisplit...
There are two problems here: the official design temperature is perhaps too low compared to reality, and contractors tend to overestimate the heat loss, to be safe. In colder climates, I guess that the fraction of time above 50% of the design load increases and it might become interesting to oversize to increase the COP. Is that right ?
Casimir
In my modeling work, I can model specific performance curves for each unit. This modeling includes cycling losses . The modeling is specific to each home and climate, and it also includes cooling, which is a sizing complication we haven't really explored here. I usually get a couple of options from the HVAC sub explore and optimize, but it's working within a very limited/practical framework.
This doesn't answer your question... I don't really have specific insight with it since it's already endogenous to the annual analysis I'm running-- strictly for efficiency.
“Where does the inefficiency come from?”
The first minute of runtime does almost no thermal transfer. All of that energy is expended in creating the pressure differential required to transfer the thermal energy. After the first minute the system is not at it peak efficiency for next 10-15 minutes or so. While the system’s pressure is stabilizing and the computer is finding the optimal opening in the electronic expansion valve for the current conditions.
On my system I can watch the “super heat “ number change as the computer adjusts EEV opening.
Back to your space heater idea the main reason they are dangerous is almost always the set back requirements are ignored. Also small children are almost certainly going to play with one given any opportunity.
The other problem is that space heaters are designed to use 100% of the max current allowed on a 15 amp circuit. This fact tests every connection to its upper limit if everything is not perfectly you get a fire.
Walta
expand on the first paragraph:
first minute is spent on getting cycle going-- adiabatic compression of the refrigeratant. talk about the efficiency of the compressor -- that's where the inefficiency will be since that will be energy that is not turned into work for the compression. if it goes into compression then that's superheated vapor that can be used for heating the conditioned space. and some/most of it will depending on the location of the lineset. otherwise, the energy expended by the compressor that isn't work is lost as heavy to ambient.
perhaps that's too nuanced of a question for here. I should check YouTube
Now THIS is a great article on boiler sizing
https://www.greenbuildingadvisor.com/article/sizing-a-modulating-condensing-boiler
But alas, those are illegal where I work, so I'm waiting for another treatise that covers AWHP
The fact that the compressor is doing work and moving gas does not necessarily mean that the gas is at the correct pressure and temperatures to condense into a liquid in the condenser and boil back to a gas in the evaporator. Only once you have created the correct pressures will the refrigeration cycle start working. The COP will be low for some time until the system stabilizes and the computer fine tunes the opening in the expansion valve for the current conditions.
Yes some of the energy will be recovered when the compressor stops.
Walta
Separate thought...
All the research that I've seen (notably a study from John proctor, but also other NREL research) indicates that cycling from oversized systems has a at most a minor impact on overall efficiency. Is there new research to suggest otherwise?
On the losses at startup, I don't think we'll be able to resolve that here. That's a detailed presentation at an ashrae conference replete with P-V diagrams
One last thought here for the day-- in all this discussion on cycling, we're all looking at discrepancies between load and space heating. I don't see any consideration for the thermal capacitance and transmissivity of the enclosure and its furnishings. I'd be much more worried about cycling in an empty enclosure (except for air) than say... the same enclosure that's a library full of stacks of books (just to cite a real example of thermal mass).
It's one of many reasons that I find the volumetric requirements for heat pump water heater enclosures unsatisfying.
Maybe what is needed is a different rating system for modulating heat pumps.
With a boiler, capacity is simple: it can produce that much and no more. If you need more, you'll be cold. If you need less, it will cycle. Similarly, with a single stage heat pump it only has two modes, off and on. If you need less than it produces when it's on, it has to cycle on and off.
With modulating heat pumps there isn't a single number you can point to. If you look at the NEEP.org ratings there are at least three ratings -- minimum, rated, and maximum. These are given at 95F, 82F, 47F, 17F, 5F and -13F.
Efficiency depends upon both output level and temperature, but what installers and consumers really want is a single number that allows them to answer questions like, "is this unit a good choice for my house?" or "which of unit A or unit B is a better choice for me?"
So imagine a number like SEER, an adjusted efficiency number, that takes into account both temperature and load. Let's call it ACOP, adjusted COP. So you could go to a manufacturer's website and plug in the loads from your Manual J and the design temperatures, and the website would spit out the ACOP. The computation would take into account any efficiency hit from short-cycling, or from having to use heat strips if the heat is undersized. If the unit is so undersized cooling load can't be met then the unit just fails and gets a score of zero. You'd probably want to put in a similar check for oversizing -- perhaps a exceeding maximum percentage of time spent modulating rather than cycling gives you a zero too.
Then code enforcement could be simple: the unit just has to meet a minimum ACOP. What that number is would depend upon the design temperature. How it meets it is up to the manufacturer.
This is effectively what's already happening in my work. Codes and Standards are on the cusp of implementing this where I'm at. The 'minimum' compliance standard is not adopted locally, or by code-- it's federally preempted. It's illegal for the local jurisdiction to adopt more aggressive efficiency standards than what's minimum efficiency in the Energy Policy and Conservation Act. So you have to just perform better than a slightly oversized (1.2x factor) single-speed heat pump that meets EPCA standards.
Manual J isn't going to be good enough here. It needs to be an 8760 hour analysis for it to reasonably capture modulation and cycling. Every house is going to be different. Some recent projects of mine indicate that Manual J is WAYYY off (way oversized for actual load, which should be expected).
There's no problem with having insufficient cooling... though perhaps there should be. Cooling isn't a code requirement in any building I work on, even in the hottest county in the U.S. (I have projects in Death Valley)
Manual J isn't going to be good enough here. It needs to be an 8760 hour analysis for it to reasonably capture modulation and cycling. Every house is going to be different.
This is why I say you'd go to the manufacturer's website, it's not something you just look up in a manual. I've done similar analysis, using the design temperature and the 99th percentile temperature and the annual mean temperature. If temperatures are normally distributed*, the 99th percentile should be 2.34 standard deviations from the mean, and using a standard distribution you can calculate how many hours would be spent at each temperature over the year. Then at each temperature I calculate the heating load from the Manual J, and the COP and output of the heat pump from the published stats. I do a straight line interpolation between the temperatures that are given. What I don't have is how the COP changes with utilization, although that could be estimated from the NEEP data.
You can see one I did here:
https://docs.google.com/spreadsheets/d/13dKnvUmw6gng4cPJtSCISzEPqw5NyqH0hjKFJ40YHO8/edit?usp=sharing
Some recent projects of mine indicate that Manual J is WAYYY off (way oversized for actual load, which should be expected).
A frequent criticism of Manual J is that it's possible to alter the outcome considerably by changing the input assumptions.
*(It turns out it is a hotly debated topic among meteorologists whether temperatures are normally distributed. Since the purpose is to compare different units it doesn't really matter so long as the same methodology is always used.)
Agreed on playing with the inputs--
for my examples, the difference is really when it comes to massive homes (a lot of thermal capacitance) where the Manual J lookup tables are understandably quite conservative. There is work to develop and approve a more dynamic load calculation method (which is already done in the physics engines I work in). You can see a hint of this in the Passive House sizing approach, which is significantly different than Manual J. They would see Manual J and oversizing for their buildings as well given their ratio of load to mass.
Regardless of my ideal approach-- I like the way you're thinking to parlay existing load calculation analysis towards better sizing. Assuming normal distribution seems reasonable enough here...
Your spreadsheet is cool!
Sometime I'll dig up the sizing method that some consulting firm came up with (Ecotope maybe?) for variable capacity heat pumps-- it used a similar temperature bin analysis if I remember correctly. They added the COP interpolation from NEEP to their analysis
I've been thinking more about the issue of how modulation affects COP. I looked at a couple of pages at NEEP. For instance, if you look at this page:
https://ashp.neep.org/#!/product/112104/7/25000/95/7500/0///0
and this page:
https://ashp.neep.org/#!/product/51530/7/25000/95/7500/0///0
You'll see that COP isn't always better at lower outputs, the second one has better COP at maximum output than at minimum. And there doesn't seem to be any fixed relationship between output and COP, if you look at the first one the minimum to maximum ration varies between 65% at 17F to 93% at 5F, both 47F and -13F are between those two values, makes no sense at all.
NEEP gives outputs at four temperatures for at least two output levels. What would probably be a reasonable way to estimate the COP at a particular combination of temperature and output would be to find the two closest temperature points and interpolate between them to estimate minimum and maximum output and COP at that temperature. Then interpolate between minimum and maximum output to get COP at that output.
Some thoughts on modeling the efficiency of cycling:
Assume that the compressor has a minimum run time, a minimum off time, and a lag time at the beginning of each cycle where it is running but not producing anything. For example, assume a compressor has a minimum run time of ten minutes, a minimum off time of five minutes, and a period of one minute where it runs before producing usable flow.
If it's cycling ten minutes on and five minutes off, that nine minutes of producing and one minute of warming up in every fifteen minutes, which means it's producing 60% of minimum capacity. It also means nine minutes of production for every ten minutes of run time, so you'd expect 90% of the COP at continuous minimum output.
If the output is less than that, the off-times get longer but the on-time stays at the minimum of 10 minutes. And for every ten minute of on time you get the same nine minutes of production and one minute of warming up, so you get the same 90% of continuous COP. So for all values of output below 60% you get the same COP, 90% of continuous.
If output is between 60% and 100%, you get longer runs, but the same five minute minimum off time. So imagine you run for 55 minutes and then go off for five. You have the same one minute of warming up, so you get 54 minutes of production. So in 60 minutes you get 54 minutes of production, or 90% of continuous capacity. The compressor is on for 55 minutes and you get 54 minutes of production, so your COP is 54/55= 98.2% of the continuous COP.
So to summarize this example, the COP is constant at 90% of continuous COP from zero to 60% of continuous output, and then it increases until at 100% of continuous output it's 100% of COP. The maximum impact of short-cycling is ten percent, and probably quite a bit less.