Ground Temperature

rahmiandarini · Posts: 43

Hi,

I am now doing simulation for a building in a tropical country. The weather type is very hot and humid. The building consists of 2 storeys, and for the ground floor, the floor material is ceramic/porcelain & cement sand render. I found the U-value is around 4 W/m².K, and the ground temperature is 14°C. The simulation result shows that there is a big heat loss from the floor which cause there is no cooling load in the first floor. I want to proof that my model is correct. How should I do that? Is it possible to model several layers of ground and see the heat transfer effect between these layers and between the uppermost layer to the first floor of the building?

Thanks,
Rahmi

AJ Lester · Posts: 71 Location: Auckland

Hi Rahmi,

I would suggest that you probably want to use the EnergyPlus Slab program to calculate your ground temperatures if you need reliable ground temperatures. I have no personal experience using the program, but have had a quick look at the documentation for it and it seems like it's what you need.

If anyone would like to confirm this/suggest an alternative, please do.

All the best Rahmi,

Andrew Lester

lklawrie · Posts: 76

Yes, that would be best. Another suggestion I have heard is to use a ground temperature approximately 2C less than your interior temperature.

The ground beneath a building is disturbed ground and does not have the same temperature as would be measured or estimated from the annual weather data.

collinar · Posts: 5

I guess the ground temperature beneath a building is disturbed by the active heating and cooling taking place inside the building. However, if the building uses no active heating or cooling, is the ground temperature still disturbed? It depends. If the building site's natural ground cover which cools by evaporation, has been replaced by a dry roof, then there is a large net heat gain for the top floor. Of course, if you have a green roof, (and walls?), as well as no active heating or cooling, and a well ventillated living space, then I would guess these heat gains can be avoided, and the ground temperature may remain un-disturbed.

I too am designing for a tropical environment. If the annual average air temperature is 28 degrees C., then the un-distrubed deep ground temperature is also 28 C. If active direct expansion A/C cooling is used to maintain a temperature of 25 C. inside the building, then the disturbed ground temperature under the slab will NOT be 2C less than your interior temperature. It will probbly be 1C greater than your interior air temperature.

For multi-story structures, only the top floor experiences direct roof heat load, and only the ground floor experiences direct ground temperature effects. (of course sun exposed walls also need to be considered)

rahmiandarini · Posts: 43

Dear Andrew, Linda and Collinar,

Thanks a lot for your suggestion.
I have tried using CalcSoilSurfTemp Program, and simulate the weather data which has
- min temp : 21.4°C
- max temp : 35.2°C
- average temp : 27.53°C
and I choose 1 (heavy saturated) and 6 (covered and moist) --> I'm still confused with this option, is there any explanation of this types?
The result shows that the annual average soil surface temperature is 28.68°C and the amplitude of Soil Surface Temperature is 2.399°C.
For me this is logically acceptable (as Linda suggests that the difference is about 2°C with indoor air), however, I still have question:
1.In designbuilder, is this value (annual average soil surface temperature) is the value has to be entered in Ground Monthly Temperature?
2.For a simulation location which has a big amplitude of soil surface temperature, how we know these varied values (what temperature in what month?)

Thanks,
Rahmi

lklawrie · Posts: 76

The CalcSurfSoilTemp program is only to be used for EarthTubes. The Slab or Basement programs are the ones you want to use.

Also, as I told someone on the EnergyPlus Support Group forum:

By the way, the Slab program is really for conditioned spaces.

What you may wish to do is look at the "other side coefficients" model. This is described in the IOref as well as the example file GroundTempOSCCompactSched.idf Another feature is to have a core and perimeter set of temperatures -- that is not shown in the example files but can be done using a "door" in the floor with one set of temperatures where the slab floor (perimeter) sees a different set. Some of this is also described in the Auxiliary Programs document on Ground Heat Transfer.

Here you would set your slab as a surface that sees temperatures on the other side that may be more realistic for you. You would still need to estimate the temperatures used but you may be able to "play" with those temperatures more easily and get your measured(?) temperatures inside more realistically.

Ground heat transfer is a real bugaboo for simulation -- one can make the results be almost anything you want (or don't want) by setting ground temperatures too low or too high.

rahmiandarini · Posts: 43

Dear Linda,

Thanks for your explanation, however now I'm confused with my own simulation

. In my understanding, I need to do simulation using EnergyPlus by exporting my idf file. But, then which ground temperature I used?
Or, I need to create a new file (i.e Slab case file)?

I'm very sorry for the inconvenience.

Rahmi

lklawrie · Posts: 76

Not sure what DB's plans are for this kind of endeavor. I think you will be "out" in EnergyPlus land running. The slab program is a separate program and creates the Ground Temperatures (or Other Side Coefficients) for you to use within the "standard" EnergyPlus idf file.

rahmiandarini · Posts: 43

Dear Linda,

Do you mind to give a short explanation about the example file 'GroundTempOSCCompactSched.idf'? I have open this file and simulated it using my weather data, but I don't understand how to interpret the result.

Thanks,
Rahmi

Carlos · Posts: 25

Hi Linda and Rahm.

I'm the one who asked the similar question in E+ group.

I will ask in Rahm's messages since we are both interested in the same issue.

1) I if didn't missunderstood, if we are working with DB and want to use OtherSideCoeff, have to export the DB file to an idf file, run the file in E+ and... then we cannot get the modified file back to DB to go on working (DB has not an option for OtherSideCoeff, so this feature would be lost)?

2) Moving between the programs sounds time consuming and kind of confussing. The result may be similar if instead of using OtherSideCoeff we just 'play' directly with the 'Ground Temperatures' of DB?

3) If in a conditioned space, the ground temperatures are about 2ºC less the inside temperatures, would it be reasonably to think that in an unconditioned space, the ground temperatures are 2ºC less the average inside temperature? If it is so, we could iterate simulations changing ground temperatures until finally get a result where the condition aproximatelly meets. If not, it doesn't matter if we use ground temperatures or other side coeffients, just guessing the temperatures would imply an enormous risk (as you said "one can make the results be almost anything").

4) Are the 'undisturbed' 0.5m depth temperatures ruled out for the unconditioned space? The building is disturbing the ground, but not as much as a conditioned space, so, sorry for insisting, but, what would be the physical meaning of using the 0.5m temperatures with a ground construction of 0.5m of earth?

5) In another part of my work, I will condition the space. Then, I will use Slab program temperatures with no ground construction. But the results of running the Slab programs shows that SlabExample used Inside Temperatures as 18ºC for Jan-March, 20ºC for Apr-June, 22ºC for Jul-Oct and 20ºC for Nov-Dec. I tryed to open SlabExample.idf to modify these temperatures but appeared 'too many errors' and closed. How do I modify the inside temperatures?

6) For the inside temperatures I should use in the slab program: I set to heat at 19ºC and cool at 28ºC. For example, in certain month, it could be hourly temperatures ranging from 20 to 34ºC (if unconditioned), so the conditioned space would have temperatures from 20 to 28ºC, so, what would be the inside temperature used for that month? Also, it could happen that 20 hours a day are under 28ºC and only 4 hours need to be cooled, or maybe 20 hours need to be cooled, how does that affect the choice of the 'inside temperature'?

Regards,

Daniel

lklawrie · Posts: 76

Rahmi,

I have asked the developer of the file to explain but not sure when he will get back to me.

AOG · Posts: 79

Dear all,

I want to share my own little experience in this topic.

We could describe the center of the problem as follow: the building affects ground temperatures beneath it, and ground temperatures affect temperatures inside the building. The less insulated the slab is, the greater reciprocal affectation we get. In terms of simulations (if we are using DesignBuilder-EnergyPlus and Slab), it means a paradox: in order to calculate ground temperatures (Slab) you need to know building internal temperatures, in order to calculate building internal temperatures (DesignBuilder-EnergyPlus)… you have to know ground temperatures!

If you have a permanently conditioned building the problem is solved, as you know reasonably building internal temperatures. In my understanding the problem begins when you have a building that is conditioned just for certain periods (as Carlos says in point 6), and become really significant when the building is not conditioned.

Well, I have used an approach that implies a series of iterations between packages:

1. Run a first Slab simulation using comfort conditioning temperatures as internal building temperatures.
2. Run a first DesignBuilder simulation using obtained ground temperatures.
3. Run a second Slab simulation using monthly internal temperatures obtained within DB.
4. Run a second DB simulation using previously obtained ground temperatures.
5. Run a third Slab simulation using previously internal temperatures obtained within DB.

After point 5 differences will be very slight, but you can follow if want. I know it could be a little thorny, but it could be necessary just once for a site. Anyway, after some similar studies one could think more good intuitive values, which in some cases are accurate enough.

The image below shows results obtained with this procedure for three building alternatives in Chihuahua, Mexico: conditioned building, unconditioned building with good insulation level and unconditioned building with bad insulation level. The model is obviously the same in the three cases: a one story 36m2 building with 10cm concrete ground slab (the analysis is intended for very small buildings).

Well I hope it helps. Anyway if someone has a better approach I'd like to know it. Also, I'd like to know what does EnergyPlus team think (this is for you Linda Smile

) and correct me if the procedure is wrong.

Regards,

Arturo

lklawrie · Posts: 76

3) If in a conditioned space, the ground temperatures are about 2ºC less the inside temperatures, would it be reasonably to think that in an unconditioned space, the ground temperatures are 2ºC less the average inside temperature? If it is so, we could iterate simulations changing ground temperatures until finally get a result where the condition aproximatelly meets. If not, it doesn't matter if we use ground temperatures or other side coeffients, just guessing the temperatures would imply an enormous risk (as you said "one can make the results be almost anything").

4) Are the 'undisturbed' 0.5m depth temperatures ruled out for the unconditioned space? The building is disturbing the ground, but not as much as a conditioned space, so, sorry for insisting, but, what would be the physical meaning of using the 0.5m temperatures with a ground construction of 0.5m of earth?

Both of the above are plausible. The ground temperatures on the weather file are determined heuristically using the ground / soil property shown in the header. You certainly could try #4 with those temperatures or one of the other depths easily.

Same with #3 except it could be difficult to determine realistic internal, unconditioned temperatures though I would try with adiabatic floors (no heat transfer) for starters.

The other, iterative approach looks fine to me -- though it may never "converge". Sad

But should yield good results.

Andy Tindale · Posts: 1148 Location: Stroud

I've thought for some time that it should be possible to provide 'good enough' estimates of monthly ground temperature for input to E+ using a simple rule of thumb. We could set up an interpolation between 2 cases:

1. Undisturbed ground temperatures from epw file
2. Inside temperature - 2

For infinitely large buildings the temperature will be 2 and for infinitely small buildings it is 1. Other buildings lie somewhere between 1 and 2, the interpolation being done using the ratio of ground adjacent perimeter to floor area as a measure of building impact on ground temperature.

It could be an interesting MSC research project to derive a correlation from running detailed simulations of many different sized and shaped building footprints.

Andy

Carlos · Posts: 25

Hi Ramh, Arturo, Linda and Andy.

After all, there are now several specific options Very Happy

Confirmed, 5 heads think better than 1. I will try them for my project and then try justify a little which one seems the best to my project. I'll talk about the results when I have them.

Arturo, what's your e-mail? Rahmi and me have been talking about the topic directly by mail. I have some questions about your method (which seems interesting). I'm already working on my master thesis, but as Andy says, it would be an interesting research the correlation he mentions. I'm not thinking in a MSC research, but in a paper. If you and Ramhi are interested, I think we could do some simulations about the different options for these ground temperatures for different building shapes and sizes and weather files and do a paper. What do you think?

Regards and thank you very much for all of you,

Daniel

P.D.

Arturo, yo también soy mexicano compadre, de Zacatecas, y los lugares que estoy simulando son calientes secos, siendo uno, precisamente Chihuahua.

Login