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

What should I trust more? Design heating load calculation based on gas consumption OR heat loss modeling?

Cafferatak2 | Posted in General Questions on

Should I trust a Design Heating Load calculation based on actual gas usage over a model? The engineer in me says yes, but I’d like some validation from folks with more experience on this.

My objective is to pave as simple a path as possible for my neighbors and I to upgrade or at least offset our Natural Gas use in our townhouse complex. One of many issues we are working on is heat pump sizing. Some of the neighbors have gone to installers for quotes and have gotten some “interesting” sizing information (54,000BTU/hr unit for a 1800SF end unit, Zone 4 climate)

Currently in Canada, you need a heat loss calculation performed to be eligible for some very generous rebates. I’ve had one done, took about 4 hours and includes a blower door test.

Result was about 36,000btu/hr.

Using Dana Dorsett’s 2016 article Replacing a Furnace or Boiler – GreenBuildingAdvisor on calculating heat loss using actual gas use data, I got a design heating load of about 20,000 btu/hr. This does however overlook the fact that the top floor has baseboard heaters and I don’t have any clean data on how much energy they consume. I do know that even on the coldest nights the top floor is only about 3 degrees cooler than the downstairs with the baseboard heaters off. So I find it hard to believe that upstairs baseboards are a 16,000 BTU/hr error. My Gas use data is pretty clean, pulling out the gas fired hot water heater is fairly easy by looking at my gas use during the summer.

Also does anyone know where this calculation is explained outside of Dana’s article? Is this in a textbook somewhere? No slight against Dana, but I’d like to have an independent source for the calculation (Textbook, standard, etc…)  if I am going to be challenging vendors on their size determination. “I got this method from a really smart guy on the internet” has limited credibility these days.
Thank you!

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. kyle_r | | #1

    Did they do a Manual J? Can you post it?

    1. Cafferatak2 | | #10

      All I have are the inputs to whatever their modeling program was. I'll post it here, but I don't know if it will be very useful. Its hard to determine exactly what boundaries were used, but most of it seems reasonable. As far as I can tell.

  2. Expert Member
    Akos | | #2

    Fuel usage is always more accurate for existing structure. About the only unknown is the building balance point, this gets tricky if the place has a lot of south facing windows.

    You won't find information about fuel usage based load calculation as it doesn't exist out there. Man J is intended for new building, as long as the inputs are correct, the result is a reasonable accurate load. The issue when you do a Man J for an existing structure is that lot of these design parameters are not known, it is very easy to guess wrong and end up with a man J 2x or 3x reality. Plus lot of times HVAC techs put their finger on the scale when inputting data to come up with a number they want.

    In your case if the fuel based calc is 20k, I would add on the heat loss for the 3rd floor from the Man J and call it a day. I would guess you'll end up somewhere around 12 to 14 BTU/sqft which is a hell of a lot closer to reality than 30BTU/sqft.

    1. Expert Member
      DCcontrarian | | #4

      I don't think it's correct to add the 3rd floor Man J. The third floor is getting heated by heat that is escaping from the second floor. The OP said the third floor is within three degrees of the rest of the house when the heat is off, so the heat from the second floor must be significant. If the third floor were being heated independently the the heat flow from the second floor would be reduced by an equivalent amount.

      I would argue that the amount that the actual usage should be increased by is minimal -- the difference in heat flow between room temperature and three degrees below.

      1. charlie_sullivan | | #9

        You are right that using the full upstairs heat loss is overly conservative. But on the other hand, counting only the difference in heat attributable to the three degrees isn't right either. The amount of heat the flows from downstairs to upstairs is more when the upstairs is colder. When the upstairs is at the same temperature as downstairs, that upward heat flow is reduced, theoretically reduced to zero.

        So going with the conservative option is the best of our imperfect options.

        1. Expert Member
          DCcontrarian | | #11

          All of the heat that is flowing from the second to the third is already included in the actual gas usage. Enough heat is flowing from second to third to keep it within three degrees.

          If there were a heating emitter on the third floor the heat load on the second floor would be reduced by the amount that the emitter produces.

  3. paul_wiedefeld | | #3

    No official studies, but I have multiple years of daily gas usage readings that show heat loss (really fuel usage) is very linear and summing the data into “months” gets you the same answer. Something I’ve thought about the method but haven’t ever tried to put a number on is that manual heat loss calculations assume every square inch of your home is 70 degrees (or whatever) - implying perfect heat distribution, like a frictionless plane. Where in reality, it matters only what the temperature is by the thermostat, no matter how oversized your furnace is. I don’t think it fully explains why the calculations struggle so much, but it might explain some.

  4. Expert Member
    DCcontrarian | | #5

    The theoretical model underlying both the Manual J and Dana's article is the same: that the heat flowing through an assembly is equal to the temperature difference times the thermal constant of that assembly; the thermal constant is the area of the assembly divided by the r-value. The design heating load is room temperature minus the design temperature times the thermal constant.

    Both techniques are attempts to estimate the thermal constant, and thus calculate the design heating load. Manual J attempts to estimate it by measurement and analysis of the building. Dana's method attempts to estimate it by measuring the actual amount of heat put into the building.

    Both estimates are only as good as the input data. I would say the actual fuel usage method is less subjective and thus less likely to be derailed by the estimator's opinions.

    Note that the underlying assumption -- that heat loss is linear and directly proportional to the temperature difference -- is probably not valid. However, since both methods use the same assumption both methods are equally affected by this inaccuracy. In practice the inaccuracy seems to be small enough that the results given are still useful. As economists like to say, all models are wrong, some are useful.

    1. Cafferatak2 | | #13

      Ahhg! Units, my old nemesis. I was using METRIC HDD (Heating Degree Celcius Days) not Imperial (Heating Degree Farenheight Days).
      Except this takes the Load way down to 11,796 BTU/hr. I know there was that other guy out in Ontario that had a similarly low load, but his place was actually Energy Star rated. This is not a tight, well insulated building. Albeit the all time overnight record cold temperature here is -15C (5F).
      What can I be missing? Can this be true?
      The gas bill said I used 7.5Gj (7,109,005 BTU) from Dec 15 2021 through Jan 19 2022 (36 days)
      Its a 1988 non-condensing Furnace, so 70% efficient. Lots of leaky ducts too.
      Heating Farenheit Degree Days during period is 1011 (@ 65F balance point) Courtesy of
      99% heating design basis is 23.9F (-4.5C)
      (BTU*Efficiency)/(Heating Farenheit Degree Day) = 7109005*70%/1011 = 4,922 BTU/FDay = 205/Fhr
      Heat load = 1.4(ASHRAE sizing factor - unitless)*205 BTU/Fhr*(65F-23.9F)= 11,796 BTU/hr.
      The townhouse is 1800SF over 3 floors. It is a middle unit, but there are plenty of windows in the front and back.
      I went back and used the calculation on other heating periods. Dec 2020 (before I replaced my windows) the max heat loss period was 14,000 BTU/hr.

      1. paul_wiedefeld | | #14

        Rowhouses are inherently well insulated buildings even if they don't check the Energy Star boxes! My rowhouse is 100+ years old, also 1800 sqft, it's not air sealed and the ceiling and walls are uninsulated. Its heat loss was ~15,000btu/hr at 17 degrees Fahrenheit using the fuel loss calculation (performed using the old furnace data). In the attached screenshots, the electrical usage for the entire house during hours below the design temperature is minor - roughly 2.4kwh * 3412 btu/kwh * 2 COP = only 16,400 btu/hr ignoring all the other electrical uses.

        1. Cafferatak2 | | #15

          ok, feeling more confident about this. Might be a tough sell convincing folks they can get away with that small a heat pump. But saving $10k in electrical upgrades is a pretty good carrot.
          Just to confirm, the graphs are from when you are heating with a heat pump?
          What is the heat pump rated to?

          1. paul_wiedefeld | | #16

            Yes, the graphs are with the heat pump. Since the graphs include all electricity, it's conservative. It's a 24kbtu ducted Mitsubishi hyper heat, so rated to -14F or something I'll never see. There's a separate 9kbtu unit in a small room without ductwork which runs some but probably doesn't contribute much heat to the rest of the house.

          2. Expert Member
            Akos | | #19


            This is why I was suggesting in the previous thread about service sizing to get a handle on your actual load. No need to start installing oversized service if not actually needed.

            The folks in the end unit also have to do the math as they have much more exterior surface area and higher heat loss. It will still be well bellow 56k BTU.

            Your load is also small enough that even if you keep in some resistance baseboard heaters to even out temperature, it won't change the overall energy usage much. Much simpler and cheaper than trying to run ducts to the 3rd floor if not already there although this won't get you cooling up there. If there are no ducts and cooling is a must, you can also look at installing a single wall mount in the 3rd floor hallway.

        2. Cafferatak2 | | #17

          Ok so with respect to equipment sizing, you didn't go with a 16 or 20kBTU unit. Why was that? Having lived with the 24kbtu unit for some time would you have chosen the same size again, or gone with something smaller?
          Thanks again for the power use graph. The weather is close to what are record lows in my area and you aren't even cracking 2500kW. Power constraints are significant in my complex so accurate power use information is critical.

          1. paul_wiedefeld | | #18

            I think the only two options were 18kbtu or 24kbtu. The 24kbtu is oversized for sure, but the moisture removal capacity for cooling is 4x the 18kbtu unit and it's an area with humid summers. The minimum heating and cooling capacities are actually pretty similar.

  5. user-5946022 | | #6

    Your water heater likely uses a bit more gas in the winter than in the spring/summer/fall non-heating season, because the water temp comes in colder, so slightly more energy is needed to heat it to the same temp.

    That just means your usage for heating the air is slightly less/ lower btu...

    1. charlie_sullivan | | #7

      And you might use warmer water in the shower in the winter and linger in the warm shower a little longer.

  6. Cafferatak2 | | #8

    Ok, thanks again Folks. What I am hearing is that I am probably closer to the mark using the fuel consumption, given that that far fewer assumptions are required for the calculation. I also like this as it is a pretty simple tool my neighbors can apply to see what their units work out to. Even if they do use a gas fireplace some times, they could run their own experiment and heat their home for a billing cycle without using it. Or even over a week or so if they decide to just read their meter.

    1. Expert Member
      DCcontrarian | | #12

      The heat from the gas fireplace goes into the building too. The only question is whether it's at the same efficiency as the furnace.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |