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

Using temperature data to calculate heating load

pwittlief | Posted in General Questions on

I’m wondering if anyone had input on using real world temperature tracking data to determine building heat loss calculations. I’ve got an ecobee thermostat and rooms sensors for each room. I can download and track all of the temperature data for each sensor. I am planning on replacing my furnace in the next year and without knowing the R-values of all the assemblies in my 100+ year old building I would figure actual data would be more accurate than estimating U-factors. If I know what the average drop in temperature is without the furnace running on a night when the outdoor temperature is relatively steady and the volume of air in each room, what would be the formula to determine my BTU loss?

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
    Dana Dorsett | | #1

    For a whole-house heat load it's simpler to estimate the design load from historical fuel-use for sizing replacement equipment:

    Without knowing the thermal mass in each room it isn't going to be possible to model the heat loss based on a room's temperature drop over time using temperature data from the Ecobee. The air volume's thermal mass is miniscule compared to the walls, floors, furniture, cats, fish-tanks, etc. The Ecobee's temperature data would allow you to more accurately calculate the average indoor temperature for making fuel-use based load calculations though.

    It's a shiny kewl object with lots of datalogging features and all, but in the end an Ecobee has no magical properites. I had to work extra hard to convince one person that the Ecobee would not be able to measure the burn cycles on a hydronic boiler oversized for the radiation and short-cycling. The duty cycle of the calls for heat from the Ecobee aren't identical to the burner's duty cycle, since the boiler's controls can (and did) cycle the burner on/off multiple times during a call for heat.

  2. Expert Member
    BILL WICHERS | | #2

    As an engineer who also has two ecobee thermostats, let me add to what Dana said.

    The ecobee has enough data to track rate of rise and figure out how much sooner it needs to fire up the heating system to get to its setpoint at the programmed time. Yes, you can use that data to figure out how much energy it takes to heat your house up, and since you also know how long it took, you can guess at things like insulating values and thermal mass but there is a HUGE but: do you know, precisley, how much of that energy breaks down into the following three loss modes:
    1- radiation through insulation on walls/roofs, even the floor?
    2- loss due to air leakage?
    3- “loss” (not really loss) to overcome the lag caused by the thermal mass of the structure and everything in it?

    It is possible to measure #2 with a blower door test. It’s possible, but very difficult to accurately measure, the heat lost through walls and the like. I’m not aware of anyone ever trying to measure #3. #3 has the effect of a rechargeable battery, slowing both the rate of rise and the rate of fall when changing between two temperature setpoints.

    Your heating Load to maintain a setpoint is just that needed to overcome #1 and #2. But the ecobee only gives you enough data to say “my furnace is way to big/small for my house at x outdoor conditions and y indoor setpoint”. I wouldn’t consider it a tool capable of replacing real modeling software, and it doesn’t measure enough different things to really accurately predict what you’ll see over a year.

    To give some perspective, I’ve been involved with many projects doing real modeling of large datacenter facilities to characterize the cooling system operating parameters for optimization purposes. Many of these facilities run over $100,000 every month in electricity alone, so there is a huge incentive to optimize everything. When we do the studies, we monitor temperatures multiple places within the facility, we monitor the supply and return temperatures for air handlers, the supply and return temperatures for chilled water supplies, the indoor and outdoor humidity levels, the outdoor temperature, and the outdoor solar radiance (which heats the building). We also only care about cooking load since there is no need for heat in datacenters in any climate. We pick shoulder seasons and extremes and optimize from that. The data sets involved are way beyond what the ecobee provides.


    1. kjmass1 | | #4

      Very cool. I enjoy diving in to my ecobee spreadsheets really just to learn more about my heating system, adjusting temperature imbalances among the various rooms of my 1 pipe steam system. Maybe you've done some insulation upgrades and what to quantify it. We'll have a fire once a week and being able to turn off the thermostat in that room which will get to 78 degrees and average out the 3 other sensors to keep the heat going for the rest of the house is a huge help for a 1 zone system. That said thermal mass is real, and just when I thought I had all my rooms dialed in to about +/- 1F, what was working great at 35 degrees is completely different at -3F overnight plus wind. All of a sudden my ecobee was averaging down because one rooms heatloss was so much greater than the other sensor, causing the boiler to keep running/high pressure/rattling vents all night. One room was 71 but the other was 67 messing up the averaging. For people with top/bottom floor disparities, it's a good stopgap solution for both heating and cooling.

      Anyways, my 2 cents.

    2. Jon_R | | #5

      > 2- loss due to air leakage?
      > It is possible to measure #2 with a blower door test.

      A blower door is not a good proxy for the highly variable effects of wind and other causes of pressure. With blower door data, you can only make an inaccurate guess about ACH-natural-max.

      Seasonal or monthly fuel use based calculations suffer from a similar issue - ACH-natural-max and ACH-natural-average are quite different.

      IMO, a plot of daily fuel use vs that day's average temperature for a month can provide reasonaby accurate results and is feasible.

      1. Expert Member
        BILL WICHERS | | #6

        Yes, and that is my point. The data are all interrelated so it’s necessary to measure many parameters over an extended period of time in order to get accurate results.


  3. user-723121 | | #3

    If you know the efficiency of your current furnace you can clock the run times for several hours. Couple this with the Delta T for the period less furnace efficiency loss will give you some idea of the heat loss for your house in Btu's per hour per degree F. Once you have established this you can get some idea of how to size the new furnace using the Design Temperature data for your area.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |