Upper limit on WindowMaterial:SimpleGlazingSystem U-Factor is inconsistent

[mjwitte]

Helpdesk ticket 10120 User notes that WindowMaterial:SimpleGlazingSystem U-Factor maximum is 7.0, but the reported U-Factor in output is never more than 5.828. ASHRAE Standard 90.1-2007, skylight U value should be 6.64 W/m2K, since the building locates in climate zone 4B and the skylight is with curb.

In the Input Output Reference,

Field: U-Factor This field describes the value for window system U-Factor, or overall heat transfer coefficient. Units are in W/m2·K. This is the rated (NFRC) value for U-factor under winter heating conditions. The U-factor is assumed to be for vertically mounted products. Although the maximum allowable input is U-7.0 W/m2·K, the effective upper limit of the glazings generated by the underlying model is around U-5.8 W/m2·K.

So, why is't the max limit for this input 5.828? And is there a way for the model to allow higher u-factor values?

[mjwitte]

Took the WindowTestsSimple example file and made a construction with u-factor input of 7. This resulted in 5.838. So, added another construction with input of 5.838 and this gives 5.815.

[EnergyArchmage]

The upper limit is heat transfer physics -- with the default film coefficients and a flat plane, you'd need a negative conductivity to go any higher. The inconsistiance was intentional so users could input the crazy high values in 90.1, even though they are not physical.

[mjwitte]

@EnergyArchmage Does the 90.1 value include a frame perhaps?

[EnergyArchmage]

yes, frames, 3d fins, projecting garden windows and skylights etc. 90.1 is a minimum performance standard and any window would exceed the highest U-values. This is a case where code minimum is not physical and energy modeling programs are stuck.

[mjwitte]

But we could offer guidance to add a frame to achieve the higher u-value. My preference here is to not allow an input higher than the physical center-of-glass limit or allow it but throw a warning stating that the center-of-glass value will only be 5.8, add a frame to lower it. But then we'd need to add reporting for a total window u-value including frame (which would be useful in general anyway). My concern is that users who are putting in the higher u-value may not notice that it's not actually being modeled as that.

[EnergyArchmage]

Adding a frame isn't really going to help. For a flat plane, the frame also cannot be any higher than U-5.8 in actual performance (unless you use unrealistical and high surface emittance to alter the radiation part of film coefficient.) It is the flatness of surfaces in E+ that restricts how low a level of performance can be attained because the surface film coefficients already have some resistance. 90.1 levels for windows include "Standard" film coefficients. This issue exists for all the window models, not just the simple glazing.

Note also that this issue is more than a reporting issue, because E+ cannot represent the worst window frames out there in existing buildings (so as to model the energy savings from improved windows).

[mjwitte]

But the reported u-factors use the standard film coefficients - no?

[EnergyArchmage]

yes the do.

[jeannieboef]

My two cents: I've just opened my WINDOW 7.2 and created an assembly with U-6.143 as follows...

1000 x 1000 mm Frame: U-6.980 W/m2-K (all round), edge correlation Class 1, from Arasteh,1989 for frame u-values from 1.9 to 0.4 Btu/(h ft2 ºF), i.e. u-values of up to 11 W/m2-K!!! for an aluminium frame without thermal break. Uedge: 5.913 Glass: single pane, Ucenter 5.913, SHGC 0.861, from CLEAR_3.DAT

If I had to use the value of 11 W/m2K for the frame, I could get U-7.0 for the assembly.

To get a 90.1 Climate Zone 1, glazing assb. I can use the 11 W/m2-K frame with SSS20Si.pfg glass out of the library (ID 16122) --> gives me a Uass. 6.887 with SHGC 0.39...getting a combination to give a U-6.81 and a SHGC of 0.25 is a bit impossible...I can get U-6.81 and SHGC 0.30, but I pretty much have to build the pane in the wrong way around. Perhaps there is a combination, but I've not found it.

[xwanguk]

The frame's U-value of 11 W/m2K will give total thermal resistance of 0.091 m2K/W. Back to EnergyArchmage's concerns, all standard total surface resistance (interior plus exterior) would normally be around 0.15, such as ASHRAE 90.1 0.12+0.03=0.15 and CIBESE 0.13+0.04=0.17. This may defer in different countries, however 0.091 would still be very hard to achieve. In other words, it is difficult for window frame to reach a U-value of more than 6W/m2K, unless you use very high (higher than standard) surface coefficients.

[xwanguk]

Oh, yes, metal frame, aluminium for example, with 3D fins can have enhanced thermal coefficient, which leads to a much smaller exterior surface resistance, therefore 11 W/m2K could be possible.

[mjwitte]

This relates to #6530

[shorowit]

With regard to the original question of modeling a skylight with a U-factor of 6.64 W/m2K, it may be that it should be entered as 6.64/1.2 = 5.53 W/m2K, which is below the upper limit of the SimpleGlazingSystem model. See https://github.com/NREL/EnergyPlus/issues/8102.