GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter 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

Sizing HVAC Equipment Method

prairieburner | Posted in General Questions on

Reading the 99% load calculation thread reminded me of a thread I started on another forum many years ago about determining the right size equipment using my natural gas bill (since our home uses natural gas for only the furnace and the stove).  The question was answered/solve, I think is pretty accurately, by AC_Hacker.

I’m wondering what folks here think of this method.

The thread is here:  most importantly post 51

Here is the relevant parts:

 me:  gas bill for the period of 12/20/2013 – 1/22/2014; this period includes the polar vortex and the rest of the extra cold month. In that period we used 72.42 therms or 2.19 therms per day. That comes to 9125 btu/hour average

AC_Hacker:  So here’s what we know:

  • gas bill for the period of 12/20/2013 – 1/22/2014
  • 72.42 therms used during that period
  • Location: Woodstock, IL

Here’s what he wants to know

  • Size of GSHP

First off we can convert the Therms of Nat Gas to BTUs, so assuming 90% efficient furnace, that would be 6,516,244 BTU for that billing period.

So I went to my FAVORITE Degree Day site and got daily HDD data for Midtown, Woodstock, IL, USA, for a time period that included his gas bill. I used 68F as the base temperature. Bear in mind that this is daily data measured at his location during the period of his gas bill. This is what I think physics looks like.

Description:,"Fahrenheit-based heating degree days for a base temperature of 68F"  Source:," (using temperature data from"  Accuracy:,"No problems detected"  Station:,"Midtown, Woodstock, IL, US (88.45W,42.32N)"  Station ID:,"KILWOODS3"    Date,HDD  2013-12-20,38  2013-12-21,40  2013-12-22,42  2013-12-23,60  2013-12-24,66  2013-12-25,47  2013-12-26,46  2013-12-27,38  2013-12-28,31  2013-12-29,44  2013-12-30,68  2013-12-31,64  2014-01-01,56  2014-01-02,60  2014-01-03,62  2014-01-04,43  2014-01-05,59  2014-01-06,82  2014-01-07,75  2014-01-08,68  2014-01-09,66  2014-01-10,36  2014-01-11,35  2014-01-12,34  2014-01-13,33  2014-01-14,42  2014-01-15,54  2014-01-16,43  2014-01-17,55  2014-01-18,56  2014-01-19,46  2014-01-20,41  2014-01-21,67  2014-01-22,63

The first thing I did with the data was to graph it out to see what the numbers looked like…

Well, no major revelations here, but those Heating Degree Days sure do jump around a lot in Woodstock, IL.

Next, I added up all the degree days and found what percent of the total each days data represented. Then from that calculated data, I found out how many BTU/day worth of natural gas would have been consumed each day (assuming 90% efficient gas furnace). Lastly I looked at the average BTUs generated each day (more HDD, more nat gas).

So, it is the rightmost column that I find to be particularly interesting, because it is what his furnace was actually putting out on average per hour during the Arctic Vortex that we all know and love so well.

So, it looks to me that if one were to choose a GSHP system that is just slightly smaller than the maximum required, a 12,000 BTU/hr would be an intelligent choice, with some kind of auxiliary system to fill in the gap.

There will be daily differences that this method does not capture, but I think it has considerable merit in sizing a GSHP, especially considering the cost of earth work, and how a seat of the pants guess could be the difference between a $20,000 system and a $50,000 system.

As a final note, I think that this method should be applied to a wider time frame, as I did see one HDD from an earlier date that should, itself be chosen as the maximum data point. That particular outlier HDD make all of these high HDD values shown above relevant, and none are outliers.

What do you think about the accuracy of this method? 


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

    This is more accurate than a Manual J.

    Imagine you needed to estimate the weight of a building. Which would you trust more: A. Measure the building, estimate its volume from the measurements, estimate its density from the materials, and estimate the weight from the estimated volume times the estimated density; or B. Jack the building up and measure the force on each jack, add them together to get the total weight.

    Method A is like a Manual J, Method B is like using utility bills.

    The biggest issue with using actual usage is the quality and granularity of the data. If you can isolate just the fuel usage of the furnace you'll get better results. Most furnaces run as either off or on, so measuring run-time is a good proxy for fuel usage.

    It wouldn't take a very complicated device at all to measure and log the number of minutes the system was on in each hour, and the outdoor temperature for that hour. With a few hundred hours of data in the coldest and hottest part of the year you'd have a very good picture of the energy profile of the house.

    I'd love to see that sort of logging as a mandated feature in new equipment. One of the issues with Manual J is that it's very rare to go back after installation and see if the real-world performance matches the model. Usually it only happens if the system doesn't work at all.

    1. john_m1 | | #8

      That the J is not revisited is an excellent point. Unless codes require that, and states/municipalities enforce it going back is the only way to know for sure. Its a thorny problem as once you ask the question, then you have to ask what comes next...

  2. Jon_R | | #2

    I log hourly furnace burner on-time, indoor temp, outdoor temp, sky conditions and wind. Current best fit (with an attempt to exclude solar gain) is:

    btu used = deltaT * 675.00 - 100 (internal) + deltaT * 6.00 * wind

    Note that wind has a large effect on infiltration and that infiltration can be a significant factor in heat loss. Even more so for solar gain. Meaning that results from longer term averages won't be accurate at 5am on a windy night, even if you account for 5am being colder.

    Also note the very low internal gain, suggesting that the most accurate base is just below my 70F interior.

  3. Expert Member
    Dana Dorsett | | #3

    >"I used 68F as the base temperature."

    That only makes sense if the indoor temps are maintained at 73F or so. Internal heat sources (warm mammalian bodies, lights, computers, TVs etc.) and solar gains add up. Most reasonably tight 2x4 framed houses would have a heating/cooling balance point (the base temp) about 5F cooler than the average indoor temps. That is how base 65F HDD became the default average back in the 1950s. Tighter 2x6 framed house with low-E glass can have balance points 8-10F cooler than the

    Have you also seen this bit o' bloggery?

    You're really looking for the BTU per degree-HOUR, not degree-DAY for making these calculations, and need to use something more appropriate than the indoor temp as a base temp (unless it's an unoccupied building with no windows.)

    >"So, it is the rightmost column that I find to be particularly interesting, because it is what his furnace was actually putting out on average per hour during the Arctic Vortex that we all know and love so well.

    So, it looks to me that if one were to choose a GSHP system that is just slightly smaller than the maximum required, a 12,000 BTU/hr would be an intelligent choice, with some kind of auxiliary system to fill in the gap."

    That "...maximum required..." in your table is the daily average load, which is well below the peak HOURLY load and it's not a subtle difference.

    Sizing a GSHP for the AVERAGE would require a thermally buffering mass (like a large tank of water), since even during a Polar Vortex disturbance cold snap the peak load can be substantially more the average daily load when the varying outdoor temps, solar gains and wind washing effects are all factored in. In my area the daily cold temps might hit -15F overnight during a severe cold snap while during the day the outdoor temp can rise to 0F or even +5F, usually accompanied by clear skies and higher than average solar gains. The difference in daily peak load and average load under those conditions is WELL into double-digit percentages.

    1. charlie_sullivan | | #4

      I looked up OP's location, McHenry County IL, and the heating design temp is -2 F. So that is actually milder than the daily average temperature on Jan 6, 2014, included in their data, which was 82 degree days on a 65 base temperature.

    2. Expert Member
      NICK KEENAN | | #5

      Right, I glossed over a bit in my answer.

      The Manual J model assumes that there is a linear relationship between outdoor temperature and heat demand. The idea is you log those two variables for enough time to get a decent sample set, then calculate the slope and intersect of the line that best fits the observed data. Then plot what the heat demand would be at your outdoor design temperature.

      You can also empirically determine what your break-even temperature is, it's where the line crosses the temperature axis (heating is zero).

      This works a lot better for heating than it does for cooling. In cooling you have to worry about latent load and solar contribution.

  4. john_m1 | | #6

    I am honored so much time has been spent on this.
    I live about 35 miles SE of McHenry.
    Wind makes a huge difference for me.
    I definitely think there is more than one way to do this.
    The J I guess is a starting point. Its kind of like buying a pickup truck so you can haul one sheet of plywood once a year. Rent the truck only when you need until it no longer makes sense

  5. prairieburner | | #7

    Very interesting reading all the replies so far.

    The jist I'm getting: better than manual j but more variables than hdd are needed to increase accuracy.

  6. Mark_Nagel | | #9

    Occupant "performance" is a huge dynamic/factor. As has been noted, things could look good on paper but... I can attest that living in the country vs living in the city presents a considerably different dynamic. Country living has my wife and I in and out of the home quite a bit (out on the property). Big conundrum for me/us: I've contemplated whether a vestibule could mitigate the impacts of such actions.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |