URGENT: Window U-values: Total window assembly v/s Glass only

8 posts / 0 new
Last post
Brad Robinson's picture
Joined: 2011-09-30
Reputation: 0

Indeed there are many ways to model windows! I am kind of lost on how to

I think I'm going to directly response to the Review comment by arguing
that eQUEST captures frame effects.

"It is It is unclear whether the window U-value of 0.26 and SHGC of 0.28
used for the Proposed case accounts for the impact of the window frames
on the whole assembly as required by ASHRAE modeling protocol. Please
provide additional information to confirm that the framed assembly
U-value was used for the Proposed case windows (e.g. showing that the
whole window assembly has been tested by NFRC, or verifying that LBNL
Window5 calculations have been provided for the whole assembly, or
verifying that the frame effects are captured within the energy modeling
software), or revise the model referencing ASHRAE 90.1-2007 Table A8.2
if needed."

I mean, I got 1 or 2 responses in this thread arguing that eQUEST does
indeed capture frame effects. The issue is that no single manufacturer
has heard of NFRC in Lebanon (and there are no documented standards on
frames in Lebanon), and U-values of Table A8.2 are much, much higher
than what's proposed, so using these A8.2 values in my proposed case
will increase my energy consumption considerably!

Best regards,

Omar Katanani

Omar Katanani's picture
Joined: 2011-09-30
Reputation: 0

Hi Omar,

I apologize it took so long for me to look at this too for you. My own
project deadlines have me running...

Here is what I know and interpret from your situation and the thread that
has been created from this discussion.

1) for LEED you will have to submit constructions specifications for both
the selected/installed glazing (glass only) AND the frame systems. In most
cases you can get tables from the frame mfg that will show how the Assembly
U-value is derated with the frame performance on top of the glass
performance values. You should try and get these tables from the fram mfg
if you can (i.e. Kawneer, or other).

2) These window systems in the US are typically NFRC rated, but by my
experience window systems outside the US are not NFRC rated....yours are
most likely not NFRC rated; this forces you to the Appendix A tables to get
your proposed *assembly* U-value that you have to use. Yes this will be
worse performing than your baseline value and thus shall reduce your
proposed model energy savings--in accordance with LEED modeling requirements
(and Appendix G.)

3) From the perspective of eQuest. When you are in the wizard you can
specify your glass/window performance in the wizard mode. When you do this
you can choose "NFRC U-value" or "U-value". When you use the NFRC U-value
method, then eQuest will take this value to account for the frame effects
and thus you do not have to specify frame data in your model inputs. If
you choose the U-value method, then this represents the glass (center of
glass) U-value only, and doesn't account for frame properties or
performance. This info comes from the help files for eQuest/DOE-2.2. The
only difference in eQuest is that using the U-value method assisgns a frame
width to the input properties NFRC U-value method sets frame width input
value at 0.00.

Recommendation: Based what I can pick up on your dilemma, I suggest the
following-- Use the U-value from Appendix A and model this in your
proposed case model. For your eQuest input, select NFRC method (frame
width = 0.0) because the appendix A data is for *assembly* factors including
the frame effects with the glass.

4) Resubmit to LEED that your proposed U-value was incorrectly applied
initially and then reference the Appendix A table data that you used in your
updated model since your projecd doesn't have NFRC rated fenestration

Hope that helps, Cheers,


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

Hi Omar-- one addition;

if you are in detailed edit mode you don't have the option to choose "NFRC
method" so what this means for you is that you have to use the U-value
input method ('simplified' not 'glass library) and then in your navigation
tree go to a window and click on it and go to spreadsheet mode. In
spreadsheet mode, find the "frame width" property and notice if it has a
frame width or if it is 0.00. You will want it all set to 0.00 since your
Appendix A value includes the frame effects already.


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

This is an interesting thread that's related to a topic that has been going around in
ASHRAE TC 4.7 for almost four years, so I'd like to add my two-cents.

1. I've always understood that the fenestration requirements in ASHRAE 90.1 are for the
whole assembly, i.e., fenestration + frame.
2. The NFRC rated values are also for the whole assembly, but those bulk property values
are insufficient for modeling because (a) they do not indicate the angular properties of
the glazing, (b) they do not distinguish between center-of-glass, edge-of-glass, or frame.
3. If the detailed fenestration description is available, one can model it in WINDOW-5/6
and produce a WINDOW-5 file (or what the WINDOW-5 people call the "DOE-2 file"!) and use
that in DOE-2/eQUEST.
4. If the detailed fenestration description is not available, one can model it using the
simple monolithic window model in DOE-2, but the results may differ substantially from 3,
particularly for complex glazing systems; under no circumstances should comparisons be
done mixing the two methods, i.e., comparing a window modeled using WINDOW-5/6 and
another modeled using the simple window model.

So, this gets to the topic floating around in ASHRAE TC 4.7 (Energy Calculations):

5. Although 3 is the most desirable approach, there are many instances where it's
unfeasible, either because
only the bulk properties have been defined, e.g., for ASHRAE 90.1 or Title-24, or the
detailed fenestration descriptions are not available, as in an existing building. For
those instances, would it be helpful to have a computer program that would "reverse
engineer" the layer-by-layer description of the fenestration based on its bulk properties
(assembly U-factor, SHGC) and any other obtainable details, e.g., number of panes, gap
thickness, frame material type, etc. ? A number of us within TC 4.7 have felt "yes", but
I thought I'd canvas people on this thread on what you think.

Please send me e-mail if you wish to see the draft Work Statement.

Joe Huang

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

One thought about the 0-frame and NRFC modeling method is, what about the
additional visible light transmissive areas for what would have been opaque
frame and is now glass? Do others simply apply the same SHGC from COG
properties to the frame, now that the assembly is monolithic?


NYCCTgreen's picture
Joined: 2011-09-30
Reputation: 0

In the simple window model, all the bulk properties, chiefly U-factor and SHGC, are
smeared over a single monolithic layer. The main source
of discrepancy from the detailed layer-by-layer model, however, are not the bulk
properties, but the angular properties of the window.

Joe Huang

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

Arpan, your framing mfg should provide you with adjustment tables for the
U-value, SHGC/SC, and VLT values for the ASSEMBLY value once the frame is
applied. Once you see these frame tables, they will make sense to you.
I've applied my COG values for PPG glazing and Kawneer frames with the
Kawneer frame performance adjustment tables and this method spits out the
whole assembly performane data that is needed for the proposed model.

If you'd like to see an example of these email me and I'll share what I


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