Heat pump sizing
Hello,
I am having difficulties sizing my mini split ductless heat pumps for our 1926 house, the walls are uninsulated double brick. Some data: Climate zone 4a ( Washington DC, 3800 Hdd), basement 690 sqf, first floor 890 sqf, second floor 860 sqf, attic will be insulated to R38 and air sealed with closed cell spray foam. There is practically no insulation in the basement, air leakage is typical for house that age. I understand that I should improve the building envelope first, but that is really tough with a brick wall. So for now, I am thinking of the following, a15 K single zone heat pump for the first floor, which is largely open plan, a 24 K multizone tri head for basement and second floor with a 7K head in the basement and two 9K heads in the second floor bedrooms ( second floor has 4 smallish rooms, so I am placing two heads on either end of the house to balance heat distribution)
I have done simple manual j calculations which suggest my total heat load is on the order of 45 to 55 K, that sounds too large to me! Am I undersizing the system way too much as planned? Also is the 15 K adequate for the first floor with its almost 900 sqf? Note I would be happy with some undersizing, to ensure I am not oversized once I get myself to insulate the walls…
Finally, a question regarding multiple single zones vs one big multizone. The single zones on paper have much higher seer and hspf that suggests many small single zone units, but Dana suggests to limit number of compressors to reduce stand by losses? What is the right approach?
I appreciate any comments and suggestions on sizing as I do not seem to get much hard information from my HVAC company. Thank you !
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Save the closed cell foam for the basement. Air-seal the attic with open cell foam or caulk & gun-foam at half the cost or less. Air seal & insulate the basement with 2" closed cell down to at least 2 feet below grade, or install R15 rigid foam on the walls in the basement.
A heat load of 45-55K @ +17F (the approximate 99% outside design temp for D.C.) for an ~1800 house is quite a bit to the high side but not impossible. An 8" thick 2-wythe brick wall -no-cavity with about an inch of plaster & lath will come in at about R3.5, or a U-factor of about U0.28. If it has empty 2x4 framing or some furring between the brick and finish wall it could be even lower. If you have a heating history on the place, with some wintertime fuel bills with the exact meter reading dates and fuel quantities and the exact ZIP code we could put an upper bound on the whole-house load using the existing equipment as the measuring instrument.
What is the actual stackup and thickness of the walls, from the exterior air film to the interior air film?
Figure out what the heat load is PER ROOM, and then by floor. A 15K Mitsubishi FH15NA is good for at least 18,000 BTU/hr @ +17F, as is the Fujistu 15RLS3 . That's probably enough but it's really a matter of calculating the load for that floor.
A 7K head may or may not be enough for the basement. If the basement is insulated it most definitely would be. Odds are that a 9K head is too big for any one bedroom, but maybe.
Dana, thank you very much. Between October 14 and January 19, I have used 260 therms for heating the house to 62 degrees on average (59 at night ,65 during the day). My zip code is 20011 and I have a 40 year old gas boiler, I reckon maybe 70 percent efficiency? How do I use that information to compute my heating load? Regarding the open cell for the attic, I would have to insulate between the rafters on the underside of the roof sheathing in an invented roof, that is why I was thinking of using closed cell foam, aka to avoid condensation. I appreciate your help with the heat load calculation. And the brick+plaster is about 9 inches thick, I am not sure what the stackup is .
The nameplate combustion efficiency of the boier (DOE BTU divided by the input BTU) is probably 80%. If it's ridiculously oversized for the load and has corroded heat exchanger plates it may be running at 70% or less, but let's assume 80%. That means out of the 260 therms that went into the burner only 0.8 x 260= 208 therms went into the heating system.
Given your thermostat setpoints I'll run the calc against a base-60F instead of the more usual base 65fF.
The nearest weather station with a complete data for that time period on degreedays.net is KDCWASHI60. Downloading spreadsheet of the daily base 60F HDD and adding up the HDD from 15 October to 19 January it comes to 1090 HDD.
So your heating system delivers at most 208/1090=0.1908 therms/HDD, or (x 100,000 BTU/therm =) 19,080 BTU/HDD.
With 24 hours in a day that becomes 795 BTU per degree-hour.
With a 99% outside design temp of +17F, and a presumptive heating/cooling balance point of 60F that's 60F-17F= 43F heating degrees. But to meet code requirements for begin able to maintain 68F you'd have to add about 5F heating degrees to that, call it 48F heating degrees, or even 50F heating degrees if your design temp is more like 15F.
50F heating degrees x 795 BTU/degree hour= 39,750 BTU/hr of implied heat load at 15F outside, 70F inside, not more.
And that's a strict upper bound for the whole load, including the distribution losses. For the load to be even 50,000 BTU/hr would require a boiler efficiency greater than 100%(!)..
The boiler didn't magically get more efficient, and you probably use at least a few 10s of therms heating domestic hot water, and the boiler has some miles on it, probably doesn't run better than 75% even steady-state, and less when accounting for idling losses. If the boiler is really running 70% as-used effeciency instead of 80% the upper bound would be 39,750 BTU/hr x 7/8= 34,781 BTU/hr.
Based on those number you probably won't need more than 35,000 BTU/hr of heat pump output, which COULD be single 2.5 ton multi-split, 3 tons at the absolute outside. The question then becomes how it breaks down by zone, to size the heads/cassettes appropriately.
With a 35K heat load at +15F that's about 635 BTU/hr of load for every degree below 70F, so the load at 47F, a temperature difference of 23F, so you'd be running 23F x 635F= 14,605 BTU/hr if you kept it that warm on 47F days. If you're talking 65F indoors it would be only 18F degree difference and load of 18F x 635= 11,430 BTU/hr. The minimum modulation of most multi-splits at 47F outside, 70F inside is about 6000-7000 BTU/hr, so with a single multi-split the compressor should be modulating rather than cycling even at temps above 50F if you size the heads/cassettes reasonably for their respective zones.
Thank you very much, Dana. That is very useful information!