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

Nightly temperature drift as a measure of efficiency improvement

GBA Editor | Posted in Energy Efficiency and Durability on

A quick way to measure efficiency improvements for a building energy retrofit is as follows; Pre-retrofit, setback the thermostat after sundown and watch both the interior and outside temperatures for a period of say, 4 hrs. Record the indoor temp. and the average outdoor temp. for this period and calculate the Delta T. Post-retrofit, conduct the same exercise and normalize the Delta T for comparison purposes. The normalized difference in the temperature drift is the improvement in efficiency for the building.

Is anyone using this method or has improvements to it?

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. homedesign | | #1

    I do not understand
    How about an example?

  2. Doug McEvers | | #2


    I may have been a bit vague on the important measurement and that is the fall of the indoor temperature before and after the retrofit.

    Let's simplify things a bit and say both nights we are testing (pre and post retrofit), the outdoor tempearture is 20F. The pre retrofit test starts at sundown and lasts for four hours, the indoor temp. is 70F when we start and we turn the thermostat way down. We record the indoor temperature at the end of 4 hours, let's say it's 64F

    Now we do our energy retrofit, airsealing, insulation, any envelope improvements that are needed and deemed cost effective. Now for the post retrofit test, we turn back the thermostat after sundown, the indoor temperature is again 70F when we start and after 4 hours the indoor temperature is 67F, remember, for the sake of simplicity, the outdoor temp. is 20F so the DeltaT is virtually the same for both nights short of a small fraction increase post retrofit.

    Conclusion, the nightly temperature drift for the same Delta T is 3F post retrofit vs 6F pre retrofit, the heat loss for this building has been cut in half due to the weatherization measures. The normalization comes in when comparing tests on nights with different Delta T's. You do the same thing when comparing energy usage for a building for different years with a different number of heating degree days.

  3. Riversong | | #3

    Sounds like a quick and dirty low-tech comparison tool. But its validity and usefulness would require that "all other things are equal', which includes daytime solar gain and temperature of inside thermal mass, wind speed, cloud cover and degree of radiational cooling, number of occupants and their activity level and type (cooking, bathing, doing laundry, etc.).

  4. GBA Editor
    Martin Holladay | | #4

    I agree with Robert -- there are far too many variables for any simple "normalization" calculation.

  5. Doug McEvers | | #5

    Actually NYSERDA used this very technique to quickly determine the efficacy of some weatherization work in the past. I am not suggesting absolute accuracy but it is a seat of the pants method to get a quick idea of the improved efficiency of a project. I ran the concept by a mentor about 4 years ago and of course he had thought of and used it at least 10 years before.

  6. Doug McEvers | | #6

    Robert and Martin,
    I agree with both of you, this is a rough guage of energy efficiency improvement and I am not suggesting you pay the weatherization crew based on the results. I do think it is a good, quick tool to get an idea of efficiency gains.

    I am one of those very strange individuals that can wake in the middle of a cold winter night, look at the outdoor temperatue and predict within 10 minutes when the setback temperature will be reached and the furnace will start. I know our house that well and it is that predictable. The key number to all of this is the Btu heat loss per hour per degree F for the building, in our case 360.

    Knowing this 360 Btu/hr/F number is all empowering and with it I can figure hourly, daily and yearly heating loads. I can also use 360 x the design temperature to figure my design temperature heat loss. The 360 is building heat loss only and does not include internal or solar gains. The point I was making in my post is building envelope improvements will show up as a reduction in the nightly temperature drift. Because you have tightened and thermally improved the building envelope, the house will more slowly give up it's stored heat and this will show up in the method I described.

  7. Riversong | | #7

    Yes, Doug, but only if "all other things are equal". You can't conduct a scientific multivariate experiment by controlling for only one variable (U-value).

    The tighter and more insulated our homes, the more significant become the internal and solar gains. And convective (wind) and night-time radiational cooling are variables that also are measurable.

    If it were true that knowing the theoretical Total Heat Loss Coefficient was "all empowering", then we energy geeks would all be gods.

  8. Doug McEvers | | #8

    One can always hope!

  9. 9PdBHDcrzT | | #9

    Since we didn't establish a baseline before we insulated, we have a total unscientific way that we used to know if our efforts were worth it. We had a automatic thermostat that cooled the house down, to 62 degrees before 10:00 pm. In order to get the house to that temperature after the insulation, we moved the setback 2.5 hours.

    We also have been tracking energy usage for the last eight years. Interesting enough, over a period of time the temperature averages only change a degree or two. After the insulation, our energy usage dropped both in heating and cooling year over year.

    These are only "feel good" measurements, but they made us want to keep working at it.


  10. davidfay | | #10

    I like Doug's method. I believe it's called a "cool down test" and has been used in building science studies in the past. Sure, you need to control as many variables as you can, like wind speed, occupant behavior, and so on. This can be done either by taking care that they not vary during the study period or by collecting enough data that they average out. In theory, though, a cool down test should enable you to find the R-value (or U if you prefer) for the whole house, based on Newton's Law of Cooling.

    One wrinkle: the temperature decrease as a house cools is not linear (as Newton proved), so in Doug's example you couldn't say that "the heat loss for this building has been cut in half ".


  11. user-723121 | | #11


    If you have metered energy you can calculate the energy improvement by normalizing the heating seasons for which you have records. Your post retrofit energy readings can also be normalized ( heating degree days as a percentage of normal) to quite accurately guage the efficiency improvement.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |