Mini-split energy usage
I thought I would offer some energy usage I had this winter with my Mitsubishi mini-split. Some info
Units: Mitsubishi MSZ-FH12NA H2i and MZ-FH15NA H2i heat pumps.
Location: Just North of NYC
99% outdoor temp – 12 degrees F
Heat load – 26,000 BTU
Month Degree Days KWh KWh/Degree day
Apr 12, 2018 824 997 1.20
Mar 14, 2018 817 978 1.19
Feb 12, 2018 1035 1450 1.40
Jan 11, 2018 1248 1773 1.42
Dec 11, 2017 751 870 1.15
I subtracted out of the KWh (estimated) I use for lights, appliances etc. The numbers above are rough estimates, I offer them up for general info, not claiming any accuracy. I will get a smart meter this year, so might be able to get better info next winter.
The numbers appear to make sense as the number of degree days increase, the amount of energy needed increase, as there is less heat to extract from the outside. But I am impressed with the performance of these units at the end of Dec and early Jan. in a 2 week period it never got above freezing and nightly temps approached zero and the units produced adequate heat.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
The formatting got mangled when it was posted, the attachment is more readable.
Were you using base 65F HDD or something else?
Base 60F is a more credible heating/cooling balance point for new code-min construction. High-R houses have to drop even lower.
Any data after mid-March has fairly significant error introduced from the high early spring solar gain in most houses.
This is the website I used for degree days, I have not found what base temp they use. They break it out weekly so I could adjust it according to my billing cycle from the utility company.
https://www.nyserda.ny.gov/About/Publications/EA-Reports-and-Studies/Weather-Data/Heating-Degree-Day-Information
When otherwise unspecified, HDD are almost always base 65F, just as it was back in the 1950s when most homes were not fully insulated, and that was the typical heating/cooling balance point. The weekly HDD doesn't occur on the exact meter reading date boundaries, which introduces a double-digit percentage error month to month.
To be more accurate and to be able to test against mutltiple HDD base temps, find a nearby weather station and download DAILY breakdown HDD spreadsheet here:
http://www.degreedays.net/
To adjust for the time of day the meter was read, sum the HDD from the prior metering reading date through the day prior to the most recent reading, then separately sum beginning the day after the prior reading through the HDD of the most recent reading date. Add them and divide by two to get the average, which is about the best you can do with daily HDD data.
Hi Michael, where about north of NYC? How many levels and sqft are you heating?
Thanks!
-Evan
Located about 20 miles north, White Plains, NY. It is 2 stories but heat pumps just used on the 1st floor only, which is about 950 sq. ft.
WPMICHAEL: Despite proximity, White Plains has quite a bit more HDD than NYC, which has more ocean tempering and a lower elevation. The 99% outside design temp for NYC is +15F, fully 3F warmer than the +12F for White Plains. If you were using NYC's weekly HDD it's skewing things quite a bit too.
Over the past 12 months station KNYWHITE18 in White Plains logged 5131 HDD65 whereas station KNYC in Central Park Manhattan logged only 4649 HDD65 . That's 10% more HDD on an annualized basis, and the month to month error will sometimes be quite a bit more than 10% (in both directions), not even including the error incurred from using weekly HDD rather than daily correlated tightly to the exact meter reading dates.
The end dates of the weekly HDD don't even correlate with the dates on the site you referenced:
https://www.nyserda.ny.gov/About/Publications/EA-Reports-and-Studies/Weather-Data/Heating-Degree-Day-Information
eg, they have data for the week ending 12/09/17 and 12/16/17, but not 12/11/17 indicated in your chart. Was that the meter reading date?
I'd be curious to see what the results are like when some of these errors are reduced by using the weather station data closest to your house, using the exact meter reading dates.
Using KNYWHITE18 data there were:
12/11/17- 1/10/2018 had 1209 HDD
...to your chart's 1248 HDD.
1/11- 2/11 had 989 HDD
...to your chart's 1035.
2/12-3/13 had 708 HDD
...to your chart's 817.
3/14- 4/11 had 723 HDD
...to your chart's 824
Something isn't quite adding up, and I haven't done the averaging to adjust for metering reading time of day, but it's not clear where the numbers in your chart come from.
If the dates in your chart are the meter reading dates (I'm assuming they are) and associated ('other uses' adjusted) kwh billed on that meter reading date, when put up against station KNYWHITE18's HDD data you get:
12/11- 1/10: 1778 kwh / 1209 HDD , or 1.47 kwh/HDD
1/11- 2/11: 1450 kwh / 989 HDD, or 1.47 kwh/HDD
2/12-3/13 : 978 kwh / 708 HDD, or 1.38 kwh/HDD
3/14- 4/11 997 kwh / 723 HDD or 1.38 kwh/HDD
That is a much narrower band with more consistent numbers, only a 7% spread from the highest ratio to lowest ratio, despite the errors from variable 'other uses' that weren't spelled out. In your chart the highest to lowest is a 22% spread.
The March & April billing periods probably 7% lower kwh/HDD than the January & February billing periods is probably due to a combination of increasing solar gain and a few much warmer days such as 20 February which logged only 9.4HDD (indicating an average outdoor temp of +55.6F), and 21 February which logged only 4HDD indicating an average temp of 61F, which might have been very close to your actual heating/cooling balance point. And the average temperatures for the more recent months were also sufficiently warmer to yield a slightly higher average COP on the heat pump.
If you play around with the HDD base temp a bit it might become an even narrower band, but a 7% error on any given month is still not bad given all the unknown variables, so your actual balance point might actually be pretty close to 65F.
Dana I was not aware of the site you posted which would have made the data more reliable. To get the billing cycling to degree days to match, since it was weekly data I divided it by 7 which crudely gives the daily degree days and added or subtracted days to match the billing cycle. Obviously that site you used has more local data and had it broken out on a daily basis. The reason I crunched these number is an exercise to see if the COP increase with temperature which it appears it does.
If you look at this pdf http://meus1.mylinkdrive.com/files/13_5_MSZ-FH_High-Performance_Wall-Mount_Heat_Pump_Systems.pdf on page MSZ-FH-9 it show the COP at 47 degrees to be 4.06 and at 17 degrees 3.19 which is roughly a 25% difference. But the difference for the colder and warmer month is less, based on these numbers. But I don't know if the KWh used and the COP is a linear relationship. But let's face a home is not a controlled environment and there are other factors introduced. I mainly did this out of curiosity. I am impressed by the heat pump technology and how much heat can be produced with a 15/20 Amp circuit.
For reference, estimates for a Mitsubishi Hyper-Heat MXZ-3C30NAHZ2 in Western MT 2100 sq ft, 22000 cu ft (est heat load 28kBTU), with 3 zones ducted SEZ 15K, 12K and 9K. Have a direct vent NG FP (operated at 21000 BTU) in a great room used occasionally but have data when not home for zero NG usage (several full months of unoccupied data).
After various air sealing/insulation projects and south facing 2 car garage door painted gray (was white) have been able to reduce kWh/HDD-65 (when at home, indoor temp set to 68/70 F night/day).
For "monthly" bills in mid-Oct to mid-Mar time span, kWh/HDD-65 (NOWData, NOAA) was above 1.2 (high around 1.4) and now below 1.0. Oct-16-to-Nov-14 2020, 922 HDD (including one day low -7 F, another day low -5 F) and approx. 820 kWh for ~0.9 kWh/HDD-65. Again, estimates from utility bills and subtraction for estimated non-ASHP electrical.
Have a few more tweaks on air sealing/insulation to do, and plans for ~ 4kW thin film solar on standing seam metal roof (2:12 2nd floor south facing shed roof).