Question-and-Answer Resource for the Building Energy Modeling Community
Get s tarted with the Help page
Ask Your Question

radiant floor savings vary a lot Atlanta vs Pittsburgh

asked 2018-01-16 09:36:06 -0600

updated 2018-01-16 10:26:32 -0600

Dear UMH community, I have modeled two nominally identical buildings, one in Atlanta, Ga, the other in Pittsburgh, PA. While the wall & roof constructions differ, using the ASHRAE 90.1 definitions for the respective climate zone, the floor constructions:

  • are identical for the heated floor portion (about 20% of the total floor)
  • are identical for the unheated portion of the floor
  • there is NO underfloor insulation (the heated slab is in contact with the earth below)

My results for Atlanta show substantial energy savings (~35%) compared to an ASHRAE baseline. Results for Pittsburgh, however, show only 6% energy savings. All of the difference is in gas usage for the hot water boiler.

That makes no sense to me. Can you make any suggestions regarding what to look at in order to understand the discrepancy? Thanks so much!

p.s., the IDFs have been extensively compared and I'm confident they're identical except for wall and roof constructions.

edit retag flag offensive close merge delete


I'd start by comparing the weather files first - how do the outdoor conditions compare during the times when the savings happen? I suspect you'll find a trend there. Admittedly I don't know the Pittsburgh climate, but I'd expect Atlanta to be much hotter so require almost no cooling. If you look at absolute values, is Pittsburgh model saving more absolute energy, just a smaller percentage of the total? I'd also pull out conduction values by surface just to confirm that the savings is from conduction through the floor,

bbrannon4 gravatar image bbrannon4  ( 2018-01-16 23:32:13 -0600 )edit

Pittsburgh has ~5400 HDD65, Altanta is ~2900 HDD65. What did you assume for your ground temperatures? What is your ASHRAE baseline system? And what are your boiler efficiencies? How did you set up radiant floor control in your model?

mdahlhausen gravatar image mdahlhausen  ( 2018-01-17 12:59:21 -0600 )edit

Thanks for the feedback; it's making me think differently about the results! answering some of the questions:

  • I'm only comparing heating system energy (natural gas)
  • default ground temps
  • Pittsburgh's in-floor radiant system uses about the same natural gas annually as a forced air system
  • Atlanta's in-floor system uses half the natural gas as the forced air system

My earlier comparison of relative savings (%) was flawed (thanks!); I only need to look at heating energy. Any thoughts re: why Atlanta's heating is half of forced air and Pittsburgh is ~ same?

Jim Dirkes gravatar image Jim Dirkes  ( 2018-01-17 15:15:23 -0600 )edit

Is your baseline a VAV system with reheat? How much reheat do you have in the baseline? I'd imagine the Atlanta system would have more reheat in the summer/shoulder seasons. Also, you should compare heating between the two radiant models, not between the radiant models and their respective baselines.

mdahlhausen gravatar image mdahlhausen  ( 2018-01-17 15:43:47 -0600 )edit

Simple constant volume heating - no reheat

Jim Dirkes gravatar image Jim Dirkes  ( 2018-01-17 15:51:50 -0600 )edit

1 Answer

Sort by » oldest newest most voted

answered 2018-01-18 03:11:59 -0600

bbrannon4 gravatar image

Based on the comments, it sounds like the increased temperature difference between the hot slab (which should be controlled to the same temperature in both climates) is the reasons. Pittsburgh is much colder, so you have a lot of heat loss, meaning that even though the system is more efficient, you are wasting extra heat by injecting it down into the ground. Atlanta, which has much warmer ground temperatures, doesn't have as much wastage and is able to send more heat upward into the space in comparison. You should be able to confirm this by looking at the heat transfer from the slab into the space (by radiation and convection) compared to the heat transfer from the slab into the ground (by conduction).

I think that's you're answer, but just to be complete, the reason you are seeing this is because it's a poor design for a radiant slab to not have insulation. If you add even a little bit of insulation under the radiant system, you prevent that heat from going into the ground (being wasted) which allows all of it to go up into the space. I suspect that a comparison like this would show a similar percentage heating energy savings, but higher absolute savings in locations, like PA, with colder temperatures.

edit flag offensive delete link more


he's using default ground temperatures (18°C/64°F), so the ground loss component should be similar in his model.

mdahlhausen gravatar image mdahlhausen  ( 2018-01-18 11:39:41 -0600 )edit

..and since you are helpfully zeroing in on the likely cause of my discrepancy, I see now that I have differing floor constructions for the different climates - some are standard contact with the ground using 14C ground temps and some are F-factor with different ground temps for each climate. 1. Thanks for helping me see that. 2. Now I need to go back and review whether they are correct, make adjustments as needed and see how that affects results.

Jim Dirkes gravatar image Jim Dirkes  ( 2018-01-18 12:22:47 -0600 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer


Question Tools

1 follower


Asked: 2018-01-16 09:36:06 -0600

Seen: 113 times

Last updated: Jan 18 '18