Modeling plenum and handling floor thickness

edit

To test you could look at the plenum Zone object's Volume field to see if the slab thickness affects the plenum height. According to the the EnergyPlus Helpdesk it shouldn't.

In EnergyPlus wall thickness only applies to heat transfer/thermal mass.

Geometrically, all walls in EnergyPlus are infinitesimally thin - i.e. they have no thickness. So, once each surface has been placed changing the material thickness will have no impact on zone volume, ceiling height, floor area, shading, or daylighting.

For most modern buildings the choice of inside vs outside vs centerline should have little impact on results so many modelers just pick one a let the volumes be off a little. Using centerlines throughout the model splits the difference. Or some modelers use outer edges for exterior walls and then use centerlines for interior walls.

If you are modeling a very thick wall, such as an old stone building, then you also have thermal mass considerations. If you use the outside edges there will be too much mass, inside will be too little. Again, centerline will split the difference and will be very close to the correct amount of thermal mass (possibly losing some corner mass).

Here are the results of a quick run of @Julien Marrec's scenario using a 10' footprint, 14' floor to plenum height, 3' plenum, and a 1' concrete slab between floors:

The issue of model surface position relative to floor slabs isn't unique to plenums, but I do see your point that with such a short height the percentage change in the volume is much more apparent. While general practice for interior surfaces is to draw the heat transfer surface at the center-line of the construction assembly, maybe in this case you should draw it at the bottom of the slab. This would maintain your plenum volume, but it would mis-represent by 1' how windows interact with the slab, unless you shift your windows by 1' so their position relative to the top of slab is maintained.

For reference I don't think most people follow the exterior surfaces to outside face and interior surfaces to center-line approach. I think this is generally used for walls but not for floors. I generally see the bottom floor drawn at top of slab vs. bottom, and other floor surfaces also at top of slab. So if you measure any floor you get the floor to floor height, and the bldg. height is equal to top of first floor/basement slab to top of roof. This approach in your example would produce a 4' plenum vs. 3'. I'm interested to see what others do. Another plenum related issue is WWR calc.

Could you not model the top of the plenum and the bottom of the next floor with a 1' gap, meaning they are not directly adjacent when you look at the vertexes. This would be similar to modeling a building floor when using a multiplier.

This would respect the volumes of both spaces as well as elevation. It would probably not solve exterior wall area.

Another potential issue with this approach in OpenStudio is the surface matching algorithm. It would not recognize those offset surfaces as 'adjacent' when you intend for them to be adjacent. One possible (untested) workaround is to have all the surfaces truly adjacent, run surface matching algorithm, then offset all surfaces.

In OpenStudio and EnergyPlus, you can model plenums explicitly. Just make a thermal zone for each plenum, don't add any loads to it, and make sure "Part of total floor area = No" so that it doesn't mess up the building summary info. You can also make these plenums supply/return plenums by hooking them up to an air loop. This video shows how to do all of this stuff.

I believe I remember seeing a method a few years ago in a simulation class for eQuest where the instructor modeled the FLR-FLR height as follows:

1. Conditioned Zone (ToF to BoPlenum)
2. Unconditioned Plenum (BoPlenum to BoSlab) Slab being the next level'sfloor/ roof deck
3. UnconditionedPlenum (BoSlab to ToF) This is the floor thickness between two levels and completes the FLR-FLR dimension.

This way you get the correct volume for plenum and zone & the correct-ish surface transfer and height location for the construction accounting for any difference in insulation where the slab meets the exterior wall.

This always seemed like a lot of work when I saw it originally, but would be interesting to test the two cases.

P.S. I know DesignBuilder allows the user to draw the geometry based on Exterior Surfaces (walls, roof, floors, etc.) and then by way of construction assembly thickness can either simulate the "inner" (net) or "outer" (gross) floor areas, surface areas, and volumes indepently. But that is another skinned cat.