Hello all,

Been quite a while since I used any type of building performance software (happily retired) but while standing on a counter in my house to repair a ceiling light fixture (wearing full fall protection, of course :-), I was reminded of a question that has bothered me and I've never had an opportunity to ask some pros.

Our home has 9' ceilings and with the stat @ 70 F, it's no surprise the temperature on high is as much as 10F warmer (ceiling fans blowing upwards help diminish stratification, but they go off at night). Have always wondered how simulation software accounts for the significantly larger delta t at the ceiling/attic interface. Could broadcast my ignorance by speculating but much prefer input from those in the know.

Thanks and best wishes.

Views: 1710

Replies to This Discussion

Commercial grade registers can help a LOT with this, especially if they are located in ceilings or sidewalls. Air can be directed downwards instead of 4 ways like a standard register. Increasing CFM of the furance blower helps since air is circulated better. Correctly sizing furances so the blower is on a larger percentage of the time helps too.

Thanks for your reply Bob, but what I want to learn is how simulation software handles the obviously higher temperature differential that occurs in real life situations. As I recall, all the software I used in days gone by had a single entry for room temperature (ie 70F during occupied heating periods). Would not the heat loss through a ceiling be substantially higher when stratified air at perhaps 80F is adjacent to the ceiling? An inquiring mind wants to know.

And thanks for the tip on the commercial diffusers but I despair that between heat cycles, stratification happens again anyway (unless the aforementioned ceiling fan is operating).

Best wishes.

If you are trying to figure heat load, the furnace should be operating at 100% when you hit design conditions. Stratification will be minimized since air will be circulating. Of course if the space has ceilings over 10' high all bets are off. The bad thing in in a space with super high ceilings is it's difficult to get the bottom warm w/o overheating the top of the space. Air infiltration has a LOT to do with the temperature difference between the lower and upper parts of the space. Any cold air leaking in will find it's way directly to the floor.

Thanks again for your input, Bob.

My question about delta T at the ceiling is not so much about system sizing, where, as you point out, it is reasonable to presume that an operating heating system at design conditions should minimize stratification (ie, 70F inside, 0F outside, for instance). What has always bothered me is that simulation of building performance for the purpose of estimating energy use over the entire heating season may end up understating seasonal fuel consumption because the 70F temperature used as nominal room temperature does not in fact exist at the critical ceiling/attic interface. In other words, it would seem heat loss across a 80F delta T ceiling is surely going to be more than heatbloss at a 70F delta T.

How do simulation programs, such as those used for LEED certification, for instance, account for this situation?

The problem is it's difficult to calculate the temperature difference between the bottom and top of the room. If the HVAC system is designed correctly and infiltration is controlled there should be minimal difference between the upper and lower levels of the room. HVAC systems that use a large volume of warm air such as heat pumps will have less temperature difference than a HVAC system such as gas heat which uses a low volume of very hot air. Commercial registers pointed to the floor in the heating season help a LOT with stratification issues. Cutting the amount of cold air leaking into a building helps a lot.

Point is there are so many factors, it's difficult to calculate how much temp difference there will be in real world conditions. Manual J typically oversizes HVAC units anyways, how many houses have you seen with heat that couldn't keep up?

Hi Steve,

This isn't an answer based upon some inside knowledge about energy software calculations, but more based upon the variables that the software must deal with, everything.  In many cases, the software does its best and then those results are tested against the real world and tweaked as needed to be as close as possible.  Take the base 65° we use instead of the actual 70° house temperature, tweaked I'm sure and it may include a measure for the detail you are discussing.

From average inside temperature, to combustion efficiency and the real R value of our insulation, beyond our tape measure, most of our numbers are a best guess.  One of the details I had to adjust to when first entering the energy profession was to stop looking for accuracy and focus more on consistency.  If our work is consistent, then we can duplicate our results and so can others.  If we strive for accuracy, it is mostly unachievable and in the big picture, our results are the difference between two numbers so accuracy becomes less important.

Most of my work involves manual calculations and I do at times vary little details like this to get my calculated results to match the real world.  But I would more often simply shift the entire house temperature to try and find an average for the set backs.

Run your numbers for a 70° ceiling and then an 80° ceiling and see how much difference it makes.  For a whole house, increasing the temperature to 80° increases the HDD by 45%, which directly translates to energy use.


Bob and Bud,

Thanks again for making time to comment.

Getting back to total energy consumption for a residential heating system, I recall in days of yore, when attempting to match a buildings historical energy usage to the base line simulated home (which would then be modified with the addition of various system and envelope modifications) I routinely had to tweak the simulation in an effort to approach the historical record. Month to month numbers were rarely within 25%, since the software was using an aggregated, long term weather data base, so I concentrated on having the total simulated heating energy agree as nearly as possible with metered seasonal record. As I recall, adjusting air changes was a good number to fiddle with as it also affected cooling energy use. Commercial structures further complicated the process because kW demand, an important element of electrical cost, also had to match up reasonably well .

After arriving at a reasonably accurate simulated/historical energy match, I used the bench-marked home to evaluate the aforementioned modifications, weather (pun intended :-) they be operational changes, insulation, lighting mods, etc. 

As you have observed Bud, consistency is the key word here and I'm looking for an answer to a question  no one other than me (a retired guy with too much time on his hands) is asking. Let's drop the discussion at this point and agree that among myriad variables involved in simulation of annual building energy consumption, Delta T at the ceiling is just one of many, many factors contributing to uncertainty.

"Run your numbers for a 70° ceiling and then an 80° ceiling and see how much difference it makes"

If I still had access to software, I'd do it! Frankly, don't know how such ceiling numbers would be entered, but suspect the "occupied room temp" field has only one entry. Clearly, a room temp of 80F would result in much higher annual heating energy, but again, ceiling temp would probably be 90F.

This exercise takes me back quite a few years and I recall reading in an energy magazine of the time how researchers working on software were having difficulty matching their numbers to reality. They said they finally included the evaporative cooling effect of rain water on the home (what, no consideration of sensible cooling?) and the numbers fell right into place...right :-) Wonder if they also factored in the periodic shadow of a parked UPS truck on the south wall...

Best wishes.

I'll have to remember the UPS truck schedule.
But there is definitely an advantage to having started in this business doing all of the calculations manually.
If you calculated the heat loss through just the ceiling area at 70°, and then the again at 80 degrees, assuming you are in a cold climate, it would be the 45% I mentioned. I'm in a 7,500 HDD zone so I recalculated all of the HDD for inside temps from 30° to 80° so I could adjust my calculations for whatever average inside temp the home owners was using. It became one of those spreadsheets that gets really dog-eared over the years.


Most residential compliance software will not allow you to make adjustments for stratification (in fact, I know of none...).  Much more troublesome software does allow such customization by the user at significantly more effort and probably not more accuracy.  For example, in IES (Integrated Environmental Solutions) or DOE's EnergyPlus, you can horizontally divide the volume of a zone into two or more zones, and then specify temperatures for those zones.  In this case, the lower zone would be set to the thermostat setting and the upper zone would be set somewhat higher.  This is really the only work-around available, and as I said above, probably offers very little, if any improvement in accuracy.  These software tools are also not recognized by residential certification schemes (RESNET and whatnot), and their inputs are incredibly troublesome in comparison to the typical residential compliance softwares (EGauge, REMrate, EPro, etc.).  Plus, the difference is likely to be less than the variability caused by uncertainty in other elements of the model.  For example, if you had 2,000 hours of 70 degree versus 80 degree temperature difference across a 1,000 ft2 ceiling, the difference in total energy use (at 100% efficiency) would only be 200 kWh.  Not insignificant, but certainly not terrible, and most climates would never see that many hours at such high temp differentials...Cheers.   

Hello Brennan,

Thanks for your reply!

I suspected more sophisticated software might exist to take into account stratification, but as you point out, there are so many other imponderables in the process that net effect of efforts at fine-tune the model would be unlikely to raise the accuracy bar. And as Bud has observed, we're really just looking for an acceptably accurate benchmark against which to apply measures (apologies for tortured syntax, but I promised Sister Mary Recalcitrant, my English teacher in 11th grade, I would never end a sentence with a preposition :-).

I own a wireless hi/lo memory, digital thermometer, and as an exercise, will tape it to the ceiling and see what difference there is overnight. Room temps will be observed on the Honeywell.

Film at 11

Best wishes.

Hello all,

Results for my min/max ceiling temperature experiment were inconclusive. Watched the differential temp till I went to bed (it remained within 2F) and awoke to discover that real time temperature was still within 2F. However, ceiling temp overnight had reached 87F at some point, most likely during temperature recovery following set back. During steady state operation, have not noted the dramatic differences I had been talking about earlier. Just climbed atop the counter (at great personal risk, I might add...the things we do for science :-) to retrieve the sending unit from it's 9' perch and noted (qualitatively) the temperature was not detectably warmer than down below. The thermometer agreed, reading 71F above and below.

Perhaps a recording thermometer would provide a more accurate picture, but at this point, I believe I am satisfied that stratification, to the degree I can measure it in my home, is not as dramatic as I had believed, except when the furnace is operating. How the temperature manages to "homogenize" throughout the space during extended "fan off" is a puzzlement

Believe I'll move on to some of the more challenging mysteries of the universe, comfortable in the knowledge that, at least in my case, temperature stratification is not worth further investigation.

Thanks again for everyone who contributed!

Best wishes.

If any of the above software is based on manual J it already has a huge "fudge factor" already built in. Real world conditions dictate that gas furnaces are almost always oversized, cycling even on he coldest of nights. I've never seen a properly operating gas furnace ever struggle to keep up, even the 2 burner models. Doing a run time calculation of the current system is probably the most accurate way to size the new furnace.


Discussion Forum

Ceiling/attic temperature differentials

Started by Steve Waclo. Last reply by Steve Waclo Jan 25, 2013. 12 Replies

Latest Activity

tedkidd replied to Bob Blanchette's discussion Tankless Water Heaters... has their time finally come?
"Yeah. Gas water…"
10 minutes ago
Eric Kjelshus replied to Bob Blanchette's discussion Tankless Water Heaters... has their time finally come?
"With super cheap natural gas prices now, still worth going all electric?  Tankless needs a lot…"
1 hour ago
Stuart Langley liked Chris Clay's photo
6 hours ago
David Butler replied to David Butler's discussion How would you insulate this roof?
"Stuart, I wasn't considering the copula windows... yes, those would indeed need to be…"
7 hours ago
Stuart Langley replied to David Butler's discussion How would you insulate this roof?
"Well, @David Butler, you have seen the property more accurately than I have, but my impression of…"
7 hours ago
David Butler replied to David Butler's discussion How would you insulate this roof?
"@ Stuart,I don't think there's any concern about impact of 4" build-up on elevation…"
8 hours ago
Stuart Langley replied to David Butler's discussion How would you insulate this roof?
"So this building has a few issues with air leaks and air leaks have so much to do with discomfort…"
9 hours ago
Stuart Langley replied to David Butler's discussion How would you insulate this roof?
"John Hatfield has the right idea here for a new roof deck. Polyiso boards always need a good stout…"
9 hours ago


  • Add Photos
  • View All

© 2017   Created by Home Performance Coalition (HPC)   Powered by

Badges  |  Report an Issue  |  Terms of Service