Accounting for plenums/technical ceilings

7 posts / 0 new
Last post

Hi all,

How do people generally account for plenums or technical ceilings? I'm not
referring here to plenums used for air circulation, but closed ceiling
space that is not considered as part of a volume to heat or cool for load

- Do you model them as separate spaces?
- Do you ignore them (but how then do you preserve the building's geometry)?
- Some other option?
- And for those using EnergyPlus, do you use the "ceiling height" field of
the Zone object? The documentation says that this generates a warning if
the ceiling height and calculated ceiling height differ too much, whatever
too much is.

The building I'm working on has a 1.1 m (3.6 ft) difference between the
space ceiling height and the height of the floor above, so it's not
something I can ignore.

Thanks for your feedback,

Patrick Bivona's picture
Joined: 2012-11-12
Reputation: 0

Hello Patrick,
The plenum areas of the building certainly need to be taken into account in the model and there are cooling and heating loads associatiated with plenums even if they are not conditioned. While they may not be used as a return plenum, there will still be a heat transfer between the plenum and all of the adjacent surfaces, especially the exterior wall and the ceiling. These heat transfers can also affect the occupied space load as there is an exchange through the ceiling. To answer your first question, plenums are typically modeled as separate spaces, yes. The height of the plenum should equal the floor to floor height minus the ceiling height. All of the math should add up so that the exterior wall heat transfers are correctly accounted for.

I recommend consulting the ASHRAE 90.1 users manual for more information on accounting for different definitions of spaces and partition types.


Shanta Tucker

Shanta Tucker's picture
Joined: 2011-10-02
Reputation: 0


In DOE-2-based simulations modeling a plenum can have +- issues. For example if you model a space as a plenum then the loads in/out of the space do not show in LOADS only in SYSTEMS as a heat gain/loss to/from the plenum only when the air is flowing in the plenum. Also, there is no thermostat setting for the plenum and the load is not included in the sum of space loads.

On the other hand if you model it as an unconditioned zone then you can see the floating temperature, but it is not conditioned.

Finally, you can also model it as a space with a wider dead band so the system never comes on/off.

So, either way has its pros and cons.

This is easy to test with the SAMP1e.inp file that is distributed with DOE-2.

Jeff S. Haberl, Ph.D., P.E., FASHRAE, FIBPSA

Jeff Haberl2's picture
Joined: 2011-10-02
Reputation: 0


Thanks for your answers. I found a specific section, 2.2.9, in the COMNET
guidelines Shanta directed me to:

(...) Because of the leakage through the ceiling (typically suspended), the

I guess the part I'm interested in is acceptable simplifications. And
perhaps modelling tricks, that might be software specific.

Now a bit of context. I'm working on a 40 storey high building. That's
quite a few plenums to model, which add a lot of surfaces simply by being
there and by breaking up spaces that could have been grouped vertically
into continuous thermal zones. I was wondering if there was a better way of
approaching this to keep the model simple and the surface count down. Or do
I just have to suck it up?

An example of simplification I thought of, with EnergyPlus, is to model
space and plenum as a single space but setting the Ceiling Height field of
the Zone object to the height below plenum. It's acceptable in terms of air
changes in the space according to the documentation. It removes a lot of
surfaces and allows to group spaces into continuous vertical thermal zones.
But a part of the heat transfers through external walls that would normally
affect the plenum will be accounted for in the space itself. Acceptable
simplification? Unnecessary optimisation?


Patrick Bivona's picture
Joined: 2012-11-12
Reputation: 0

I've been reading the postings on this thread, but have resisted chiming
in until now :-)

Although modeling a plenum is appealing from the point of physical
verisimilitude, there are actually some drawbacks depending on which
simulation program is being used. With DOE-2, adding a plenum on the
top floor converts the ceiling to an internal surface, for which DOE-2
models only conductive heat flows, thus losing the direct radiant
exchange between the roof bottom and ceiling top. I'm still
investigating the impact this has on the simulations, but the
preliminary indications are that it underestimates the impact of
measures affecting the roof reflectivity, such as a "Cool Roof".

The other thought I have is that while the COMNET description quoted
below is physically correct, much of it (stratification, leakage through
the ceiling, etc.) are ignored in the building simulations. For all
intensive purposes, a plenum in DOE-2 or EnergyPlus is simply an
unconditioned or a partially conditioned zone if it contains ducts (and
the ducts are modeled).

As far as modeling tricks, adding a plenum to DOE-2 or eQUEST does not
add much of a modeling burden if a FLOOR-MULTIPLIER is used to describe
a typical floor + plenum. Similarly, with EnergyPlus, you can create a
ZoneList containing all the spaces on one floor plus a plenum above, and
then give the Zone List a multiplier.

If you want to omit the plenum, I would still make the exterior wall
area include the plenum height, but maybe reduce the air volume of the
space as you've described. I know that's not a problem in DOE-2, but I
don't know whether EnergyPlus would allow you to do so, i.e., the input
zone volume is less than that enclosed by the surfaces. I don't think
you'll go terribly wrong, but I would still think that having a plenum
would be preferable.


Joe Huang's picture
Joined: 2011-09-30
Reputation: 406

1. Came to learn about plenums.

2. Learned the phrase "physical verisimilitude."

3. Will be spending the rest of the day digging for an excuse to slip this phrase into everyday conversation....

Thanks, Joe!

The implications for rooftop radiant exchange are intriguing by the way and could immediately impact a few current projects of mine... Would be interested to learn the results of your investigation!

More to the original topic, I agree there's more than one right answer with respect to liberties of simplification a modeler can reasonably leverage. So long as you do account for the corresponding perimeter envelope loads and are mindful of the conditioned volume, I think you're on solid ground in many cases to omit the plenums explicitly. On the flip side (speaking specifically to the eQuest interface), especially if you are using floor multipliers, you'd likely spend much less time actually modeling the plenums.

For further reading, the [equest-users] archives have many threads discussing plenums, interstitial floors, and heat transfer implications. You might pick up some new ideas and food for thought there ;).



Nick-Caton's picture
Joined: 2011-09-30
Reputation: 605

For ceiling plenums under conditioned spaces or well insulated roofs, I generally agree that they are not worth the effort. Whenever I try to get smarter about modeling plenums I come back to the lack of information on plenums in 90.1 ECB and Appendix G. These are clearly design decision but how should they be accounted for. It is a strategy to have a plenum, to have ducted returns, to have recessed lighting, etc?

Paul Riemer, PE, LEED AP BD+C

Paul Riemer's picture
Joined: 2011-09-30
Reputation: 0