GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Mini-Split Sizing Based on Historical Energy Use

tanzerson | Posted in Mechanicals on

I’m looking to install a pair of Mini-Splits to heat a 1550 sq ft bungalow with finished walkout basement.   It’s 1968 construction.  I am getting a variety of size proposals and  I ran, none of these values seem to align.

Currently we heat with mostly with two gas inserts (about 35kBTU each), one per floor but the primary heating is electric BB (27kW total capacity) which don’t get much use. 

The space on both floors is largely open with a couple of bedrooms that we like cool in the winter and don’t use the BB.  Circulation is by convection through all spaces, i.e. no forced circulation other than the fireplace fans.

My question is how useful are historical energy consumption (combined gas and electricity) at sizing new mini splits.  For example, I used an average of 19kBTU/hr (equivalent) continuous energy in the coldest winter month.   Is this useful in sizing a mini split max capacity and how would I do that?

Any insights will be appreciated.

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. Expert Member
    NICK KEENAN | | #1

    Historical usage should give you an accurate picture.

    By convention, it is assumed that heat load runs linearly with outside temperature. So if inside is at 70F and you need X at 50F (20F difference) , at 30F (40F difference) you'll need 2X and at 10F (60F difference) you'll need 3X.

    So if you have the daily high and low temperatures for that month when you averaged 19K BTU/hr you can calculate what the average temperature was. Heating systems are normally sized for the 99th percentile temperature, so you need to find what that is for where you live. Then extrapolate the observed usage and average temperature to the 99th percentile temperature.

  2. Expert Member
    Akos | | #2
  3. tanzerson | | #3

    Thanks for the replies. I did run two calculations using estimates of NG and E used for heating. I established a base use using the summer months which were very even for 2020 and 2021. We're on lake Ontario and don't use AC so I think that's pretty accurate. Subtracting the assumed base and fireplace insert losses from the NG and E usage for Feb and March in 2020 and 2021 (when I have actual instead of estimated NG bills) and the HDD65 data for the same periods, the calculated design loads are within 12%. NG use is about 85% of heat energy used with the rest from electric BB. I'm pretty comfortable with this result. The highest 99% design (-7f) comes out at about 34kBTU, much lower than what I was being quoted. I'm not decommissioning any of the existing heat sources so extreme lows are not a worry. Thanks again for the support, much appreciated.

  4. Jon_R | | #4

    > I'm not decommissioning any of the existing heat sources

    Then you could omit or reduce the 1.4x multiplier (see link in #2 above).

    1. tanzerson | | #5

      Yes, that's my plan, thanks for the confirmation.

  5. tanzerson | | #6

    Follow up: I used a couple calculators, one from Fujitsu, the other My fuel usage calculation method comes up to about 75% of the calculated load (at -5f about 60kBTU vs 47kBTU). I've double checked the fuel use and HDD data so I'm still confident in those numbers. There of course some assumptions in both methods, but far fewer in the fuel based calc, primarily the efficiency of my gas inserts and hot water down the drain.

    My question: Assuming my data is correct, should I take the fuel use numbers as the most accurate of the two methods?

    Thanks again - this is a great site

    1. Expert Member
      Akos | | #7

      Since fuel use gives you actual real world building load, as long as you have done the math correctly, it is the more accurate estimate.

      P.S. With a 1500sqft bungalow with a 35K heat load at your design temperature, there is a lot of low hanging fruit for energy savings. Some blower door directed air sealing would be a good investment.

  6. Jon_R | | #8

    > most accurate of the two methods?

    Is your fuel use calc with or without the 1.4x multiplier? 25% low is about right for the difference between "design day load" and "Man J load" (with margins for colder than design day, etc). With supplemental heat to cover the difference, you can use either.

  7. tanzerson | | #9

    I had a few corrections to loadcalc and now get much closer results. No multiplier in effect as I have backup heat in place.

    The house is 1500 sqft, but it's 3000 sqft with the walkout included. Both upstairs and down have a 50' wall of mostly glass (like 90%) and the basement floor is uninsulated from 1968 construction. I'm sure I can improve on infiltration around doors, but the windows are newish and pretty tight.

    Of the 1500 sq feet on each level about 1000sq ft is open concept.

    I'm looking at a Mits 4C36k H2i pump with 3 heads (18,18,6)to meet the 99% (-7f) design. 18k in each open area + a 6k head in the master bedroom, on which we rarely close the door.

  8. TrackerBob | | #10

    Thanks for this information!

  9. TrackerBob | | #11

    Custom web software development is a critical aspect of the enterprise IT strategy. For a business to succeed, it must develop a variety of different types of applications, including web and mobile apps. Fortunately, there are many tools and developers available for this process. These tools allow developers to create more customized solutions for businesses. And when your business is ready to move on to more robust, customizable platforms, Custom Software Lab is here to guide you.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |