Adjacent Shells

12 posts / 0 new
Last post

I am modeling a building with multiple wall types therefore I have 3 shells per floor. I am trying to figure out how to have the adjacent walls between shells be modeled as interior instead of exterior. In one of spot they came out as interior, but I am not sure how I did it. I am trying to do as much in the wizard as possible. Does anyone know how to do this??

Thanks for your help.

Sheri Rocha, P.E. , LEEDTM AP

Rocha, Sheri's picture
Offline
Joined: 2011-09-30
Reputation: 0

Sherri:

Get all your other items done that you do on the Wizard level. You need to get to the Detailed Edit, then into the .INP file to manually change those.

I do not think there is a more straightforward (or any other) way.

Maybe someday in a future version.

John R. Aulbach, PE, CEM

John Aulbach's picture
Offline
Joined: 2011-09-30
Reputation: 1

We accidentally stumbled onto a way of doing this.

If the walls on the zoning tab do not line up with the walls on the
shell tab, they will show up as internal walls.

Example 1

Shell:

(0,0) (100,0) (100,100) (0,100)

Zone:

(1,1) (99,1) (99,99) (1,99)

This will create a zone surrounded by 4 internal walls.

Example 2

Shell:

(0,0) (100,0) (100,100) (0,100)

Zone:

(1,0) (99,0) (99,99) (1,99)

Will create a zone with 1 external wall ( on the south side) and 3
internal walls.

Hope this helps.

Jim

Crockett, Jim's picture
Offline
Joined: 2011-09-30
Reputation: 0

Go to the "Building Footprint" wizard screen. The wizard displays a plan
view of your shell. You can right-click on any wall and change it to an
exterior wall (default), an entirely adiabatic wall or a partially adiabatic
wall.

Hope this helps -

Robert

Robert Des Rosiers's picture
Joined: 2011-09-30
Reputation: 0

If you don't want to edit the .INP, in the Detailed Edit you can delete
the exterior wall, and just make a note of its polygon vertex. Then
create a new interior wall at that same vertex.

Christopher J. Sachs

Christopher Sachs's picture
Joined: 2011-09-30
Reputation: 0

Robert - Thanks so much for that heads-up! Some further questions
though...

I've observed that separate shells sharing the exact same vertices will
default the shared walls to be "partially adiabatic" in the same
manner/location you've described...

1. I think this still necessitates going into detailed mode and
making them interior walls associated with the correct adjacent zone, if
you don't want to call them adiabatic. Is that right?

2. This is obviously a real time-saver for a number of situations,
but attached and below is a screenshot of my case-in-point: When
modeling courtyards by using a footprint that cuts through the interior
(as has been described in other posts), it looks like we can right-click
the "cutting" lines that are ~1" apart to define them as adiabatic, but
is there any way to zoom in this view that I'm missing to be sure this
has been done correctly to all of the affected walls? It appears that
you can only select one specific wall when they're that close together
from this zoomed perspective... This screengrab shows purple lines
where I've tried to perform this action, but if you squint you'll see it
appears I have both a purple and black line in some locations, and I
can't seem to right click with enough accuracy to define each of them...

Thanks again Robert and to everyone sharing their insights! Apologies
if my mspaint markup is hard on the eyes!

NICK CATON, E.I.T.

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

Jim:

Will you get the internal heat ransfer between these walls? Or will you care?

John R. Aulbach, PE, CEM

John Aulbach's picture
Offline
Joined: 2011-09-30
Reputation: 1

Not sure. But if the set point on both sides of the wall is the same,
the heat transfer will be negligible. Short answer is 'don't care.'

Crockett, Jim's picture
Offline
Joined: 2011-09-30
Reputation: 0

Jim,

I think the issue-at-hand here, if I'm understanding correctly, is that
there's concern an internal wall created by avoiding the footprint
perimeter will simply use the defined internal wall construction and
transfer to/from the outside... I'm not certain (and I haven't tested
this) that you want to answer "don't care" without verifying this isn't
the modeled behavior.

NICK CATON, E.I.T.

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

I just checked, and it does not create an exterior wall with the
properties of the interior wall. (Not sure that eQUEST will let you
create an internal wall with the outdoors on the opposite side.) It
creates an interior wall with the same space on both sides of the wall
(it's like inserting a wall into the middle of a zone), so unlikely to
have any heat transfer.

Robert had a much better solution, I'd recommend using his.

Crockett, Jim's picture
Offline
Joined: 2011-09-30
Reputation: 0

Sheri,

After reading the subtext of this message chain, what I think you?re trying to accomplish is not possible in the wizard (multiple exterior wall constructions with inter-zone heat transfer). I?m sure this undermines all the work you?ve done thus far, but you?ll save quite a bit of time and have a much better project by adjusting the exterior wall constructions in the detailed interface rather than creating a new shell for each construction. Let me explain and comment on the message history:

? eQUEST will create adiabatic interior walls between two adjacent shells if the shell coordinates line up along a common plane. Therefore, if the end goal of creating multiple shells is to account for inter-shell heat transfer and have multiple exterior wall constructions, some use of the detailed interface is required. You will have to adjust either the interior walls (from adiabatic to standard) or the exterior wall constructions in the detailed interface. It is by far faster to adjust the exterior wall constructions. If you opt to change the interior walls from adiabatic to standard, you?ll be prompted for the adjacent zone for each wall. If you have three shells per floor, it sounds like you?ll need to create a spreadsheet or rename all zones in all shells to something more identifiable than ?EL1 East Perim Zn (G.E2)?.

? In the wizard, when a zone is created that is partially exterior and partially adjacent to another shell, the walls along the plane of intersection will be broken up between interior and exterior walls according to what?s on the other side. Where the two zones in separate shells touch, each zone will be assigned an adiabatic interior wall as mentioned above.

? Per Jim?s example below, for any wall that is drawn in the wizard but does not touch the building footprint or another zone, an adiabatic interior wall will be assigned. The wall is basically there to soothe the user ? eQUEST couldn?t care less if the wall is adiabatic or not there at all.

? By definition, an interior wall (adiabatic or standard) will never communicate thermally with the exterior. For what Nick described below to occur, one would have to use the detailed interface to manually assign an interior wall construction to a wall that is defined as exterior. The conversation has mixed the concepts of wall type and construction ? these are separate properties. Wall type dictates where heat is transferred, construction simply dictates the rate and delay (thermal massing).

? Also by definition, an exterior wall will never communicate thermally between two zones; it can only transfer thermal energy between its parent zone and the exterior

? I used the tons-of-shells approach on my first project and ended up with over 10 shells. eQUEST would crash when I tried to run the simulation. What I didn?t quite realize is that when you create a new shell, eQUEST is going to create new occupancy, lighting, equipment, etc., schedules as well as new constructions for all interior and exterior surfaces. True, each shell?s schedules and constructions could have the same properties, but they are still viewed as separate entities and add a lot of weight to your project file. At four or five shells users will see a noticeable delay in opening, saving, and simulating files. Done to extreme, the project becomes too cumbersome to simulate.

? One can be assured that time, complication, and frustration will be added to any project by trying to do it all in the wizard.

With regard to what you?ve witnessed thus far, I would bet that there?s a conflict between the coordinates/reference location a shell is given on Wizard Screen 1 and where that shell is actually drawn on Screen 2 (assuming you?re tracing a CAD file). Where it did create an internal wall, the common plane between Screen 1 and 2 jived. That?s about as accurately as I can describe it in an email. Try creating two adjacent shells in a test file using only site coordinates and aspect ratios and you?ll see what I mean after inspecting the shells in the detailed interface.

To bring this full-circle, it will be faster and more accurate to create as few shells as possible in the wizard and then use the detailed interface to create/assign additional wall constructions. The most prevalent wall construction should be specified in the wizard to reduce the number of changes in the detailed interface.

Regards,

DAKOTA KELLEY

Dakota Kelley's picture
Offline
Joined: 2011-09-30
Reputation: 1

Echoing Dakota's points, it is always good advice and good simulation
practice to simplify inputs when possible. To stay with the one-shell
approach and still represent the thermal effects of three different
constructions on each wall, you can always take the "area weighted" approach
for the three thermal values (R-value or U-values). This takes some up
front calcs outside of eQuest, but will keep your input values, and model
file much simpler and easier to work with.

For each wall, take the [% area of (Construction 1) X the R/U-value of
(Construction 1) + % area of (Construction 2) X the R/U-value of
(Construction 2) + % area of (Construction 3) X the R/U-value of
(Construction 3)] = The overall R/U-value that you will enter for your wall
construction performance values.

For my personal approach, I would rather do the manipulation work outside of
eQuest, than try to 'muscle' eQuest into doing *all* the details. This is
the approach I would take. Becareful though, if you're doing a complaince
model, you need to be mindful of approach's like this that might
'over-simplify' what you need to achieve.

At least now you have some options of how to get to where you are going...

Pasha

Pasha Korber-Gonzalez's picture
Joined: 2011-09-30
Reputation: 400