Ventilated air space behind a brick siding

[JF]

Hi,

I downloaded Therm 7.4.3 last week and I have a few questions following a simulation I tried.

The first one concerns the air space behind the brick siding in my wall composition wich is a ventilated area. I used the "Frame cavity slightly ventilated NFRC100". Is that the proper way to modelise this space? If so, do I have to set some boundary values to get a realistic result? For now, I've only set a -31C on the outside boundary and a 23C on the inside boundary.

At the same time, can somebody tell me what means the (23%) in the page header? Does it represent the margin of error? If so, it seems to me to be high... have I done something wrong?

Also, I would like to know why I can't get a value for U-factor for this wall type. The U-factors box shows a N/A and I would like to know what to do to get a value.

Thanks in advance.

JF

[Fabrizio Prato]

Hi JF

According to standard EN ISO 6946 external ventilated cavities (in your case the one behind the brick siding) should not be considered in thermal resistance calculations and the external surface resistance on the last resistant layer (in your case the EPS insulation) should be considered equal to internal surface resistance (Rsi = Rse = 0.13 m2K/W, in Therm that is Internal film coefficient = external film coefficient = 7.692 W/m2K) and the internal non ventilated cavity thermal resistance can be picked, in case of common emissivity of its internal surfaces, from those precalculated by the standard 6946 (in this case for 20mm horizontal flux R=0,175, equivalent conductivity Leq= 0.114 W/mK). Furthermore in case of inhomogeneous layers (in yor case rockwool between steel mullions) the model should always be cut on symmetry axes to be sure the edge is actually adiabatic.

Apart from the above considerations (valid where standard EN ISO 6946 is applicable) using Therm cavity materials could be proper but it would be essential to correctly set "Gravity arrow" ("View/Gravity Arrow") to let Therm do the right calculation: in your file it should be set "Into screen" not "Down".

I attach the modified therm file according to 6946 and an image of the results obtained: I used internal temperature 20°C and external 0°C because in U calculation delta T doesn't matter, it's enough to have 1°C. To know thermal flux in your actual conditions (internal T=24°C external T=-31°C) you can modify my BC definitions.

Despite the presence of about 15cm of rockwool insulation and about 4cm of EPS the actual U factor of this wall is unfortunately rather high. This is due to the fact that the external EPS insulation layer is too thin to mitigate thermal bridges engendered by the steel mullions.

Fabrizio

[JF]

Thanks for the answer,

I worked on the same model and for a reason i don't get, as soon as i modify the T value of a BC, the U-factor becomes N/A. Maybe you can tell me why? (see attached M1_FP_modified.thm)

I would also like to know what does mean the (18%) in the page header (see attached pdf)

I understand that for my M1 wall type, the brick siding should not be considered, but would it be the same situation in the case of a ventilated cavitie behind a fiber-cement panel siding? (see M2 wall type) I'm asking the question because my guess is that the vertical Z-bar on witch the siding is fixed is a part of a thermal bridge that has to be considered in the simulation?

[Fabrizio Prato]

Hi again JF.

Please find answers below.

Salut
Fabrizio

Il giorno mercoledì 1 febbraio 2017 21:26:18 UTC+1, JF ha scritto:

Thanks for the answer,

I worked on the same model and for a reason i don't get, as soon as i modify the T value of a BC, the U-factor becomes N/A. Maybe you can tell me why? (see attached M1_FP_modified.thm)

This is because after changing T values you had to launch again the calculation and Therm automatically reset results display on "Projected Y"; to have your thermal flux and U factor correctly displayed you should set to "Projected X" or, in this case, "Total length". (See attached image)
A little tip: I find it useful to give BCs names reminding their characteristics. (See attached M1_FP_modified.thm)

I would also like to know what does mean the (18%) in the page header (see attached pdf)

I don't know what that percentage value means, certainly not the error value. Perhaps the zoom value of the detail image: only Therm developers can answer this question.

I understand that for my M1 wall type, the brick siding should not be considered, but would it be the same situation in the case of a ventilated cavitie behind a fiber-cement panel siding? (see M2 wall type) I'm asking the question because my guess is that the vertical Z-bar on witch the siding is fixed is a part of a thermal bridge that has to be considered in the simulation?

The Z-bar are thermal bridges only if they pass through the EPS insulation layer (that does not seem to be the case in M2): on the other hand it is certain that their fastening system is punctually crossing the insulation layer, and these are punctual thermal bridges that worsen definitely a bit the thermal transmittance of the wall but that are a little tricky to calculate with Therm. On the contrary if vertical Z-bar actually pass through the insulation layer (in full height) than this one would become an other inhomogeneous layer (like the rockwool one) worsening much more the thermal transmittance of the wall. This case could be modeled as in the attached file M2_FP.thm

In any case, according to EN ISO 6946, any ventilated air cavity and any successive layer toward the outside should not be considered in calculation.

[Sergio F.]

I'm simulating a wall inside with an aluminum blank profile, and comprising an insulating mineral in adjacent areas.
schematizing as "Frame cavity" empty area, the results of U are incorrect, even for the area with insulation, where the transmittance should be about 0:23 W / m² K, while it is approximately 2:56 W / m²K.
I think the error is due to incorrect setting of "Frame cavity".
see attached file.
Thanks for your help.

[Sergio F.]

Sorry but with my smartphone i can't send the file. See the other my recently post thanks

[JF]

Hi,

In fact, the vertical Z-bar that are shown in M2 are fastened to horizontal Z-bar @ 400 c/c inbetween which is placed the EPS insulation. Those horizontal Z-bar are then fixed to the steel studs through the gyplap board.

I didn't know how to show both EPS and horizontal Z-bar so I only drawed EPS.

Is there a way to ask THERM to consider those horizontal Z-bar in the simulation and get a realistic result?

At the same time, do you know if EN ISO 6946 is applicable to north america (Quebec in particular) or if there's any equivalent? My researchs weren't succesful...

Thanks again for your time. it is really appreciated!

JF

[Fabrizio Prato]

Bonjour JF
Answers below.
Salut
Fabrizio

Il giorno martedì 7 febbraio 2017 21:46:11 UTC+1, JF ha scritto:

Hi,

In fact, the vertical Z-bar that are shown in M2 are fastened to horizontal Z-bar @ 400 c/c inbetween which is placed the EPS insulation. Those horizontal Z-bar are then fixed to the steel studs through the gyplap board.

I didn't know how to show both EPS and horizontal Z-bar so I only drawed EPS.

Is there a way to ask THERM to consider those horizontal Z-bar in the simulation and get a realistic result?

Yes, you could also model the vertical section with the horizontal Z bars and do calculations on it. Then you could create 2 materials that are equivalent (have the equivalent thermal conductivity) of the 2 inhomogeneous layers (the one steel/mineral wool and the one steel/EPS) then do again calculations replacing in the horizontal section the steel/EPS layer and in the vertical section replacing the steel/mineral wool layer. Conservatively I would choose the worst result. It would be just a rounding: actually this should be done with a 3D simulation sw.

At the same time, do you know if EN ISO 6946 is applicable to north america (Quebec in particular) or if there's any equivalent? My researchs weren't succesful...

I don't know but being the 6946 an ISO standard, it should be applicable wherever on earth... ;-)

[JF]

Hi again,

I don't know if got it right but I did a simulation with EPS/Zbars only, and another one with Roxul/studs only. I obtain a U factor of 0.9888 for the first one and 0.5642 for the second one (see attached files). I would  now create new material with k values based on those results. Is that right? If so, my only concern would be on the Hc value (film) used in the BC type. I kept the 7.692 W/m²K of the interior film but I guess I would have to change it considering that those component are inside the wall composition???

Can you tell me if I am on the right path of if there's another way to determine the thermal conductivity of two inhomogeneous layers?

Hoping my question is clear...

Thanks again!

JF

[Fabrizio Prato]

Hi JF

You got it right: actually the Hc values in this calculation are irrelevant because to obtain the equivalent thermal conductivity (λeq) of the inhomogeneous layer you should remember that a single layer's thermal resistance is R=s/λ whence λeq =s/R but you have the thermal transmittance of that layer which is U = 1/RT whence RT = 1/ U and also RT = Rsi + Rlayer + Rse, whence again Rlayer = 1/U - Rsi - Rse  (where Rsi = 1/Hci and Rse = 1/Hce are subtracted). With your U values:
- EPS/Zbars layer thermal resistance is Rlayer = (1/0.9888 - 0.13 - 0.13) m2K/W = 0.751 m2K/W and λeq = 0.03874 m / 0.751 m2K/W = 0.052 W/(mk)
- Roxul/studs layer thermal resistance is Rlayer = (1/0.5642 - 0.13 - 0.13) m2K/W = 1.512 m2K/W and λeq = 0.15240 m / 1.512 m2K/W = 0.101 W/(mk)

Now you can create a fictitious homogeneous material to substitute one of the two inhomogeneous layers in the original model and calculate the "real" U value of the wall (remembering that only a 3D simulation would be really correct)

BUT as I wrote in my first answer in case of inhomogeneous layers the model should always be cut on symmetry axes to be sure the edges are actually adiabatic. So to have more reliable results you should remodel the inhomogeneous layers as in the attached modified therm files, where EPS/Zbars layer U value = 0.9285 W/(m2K) whence  λeq = 0.047 W/(mK) and Roxul/studs layer U value = 0.5038 W/(m2K),
whence  λeq = 0.088 W/(mK)

Hoping my answer to be helpful.

Ciao
F.

[JF]

Molte grazie Fabrizio,

Your help is really appreciated. One or two more things, and I leave you alone ;)

First, depending of the model, the “% Error Energy Norm” I get varies from 3.4% to 9.1%” in the “U-factor boxes”. I’m not sure to understand the meaning of that percentage. I observed that it increases with inhomogeneous layers, but how should I interpret those percentage? I guess I have to consider them in my final U-factor result? For instance, if I get a U-factor of 0.1988 with a %EEN of 9.13%, I have to consider that my result is U=0.1988 W/m²K ± 9.13%... is that right? It might look obvious but I just wanna make sure…

 

And finally, in my M1 wall, the brick siding isn’t considered in the calculation, but the mechanical fasteners do have an incidence on my U_T.

I found out in Annex D.3.1 of ISO 6946:2007 that ΔU_f = n_f * χ.

I unfortunately do not have the ISO 10211 mentioned in D.3.1, so I was wondering if you were able to give me a clue on how to find that χ value that I miss? I also did the calculation with the D.3.2 formula ( ΔU_f = α • (λ_f A_f n_f / d₀) • ( R₁ / R_T,h )²) and obtained a result of 0.0136 W/m²K that seems realistic but I’d like to validate with the D.3.1 formula.

 

Merci et bonne journée!

JF

[Fabrizio Prato]

Hi again JF.
I answer as always between your lines.

Ciao
F.

Il giorno mercoledì 22 febbraio 2017 20:30:10 UTC+1, JF ha scritto:

Molte grazie Fabrizio,

Your help is really appreciated. One or two more things, and I leave you alone ;)

First, depending of the model, the “% Error Energy Norm” I get varies from 3.4% to 9.1%” in the “U-factor boxes”. I’m not sure to understand the meaning of that percentage. I observed that it increases with inhomogeneous layers, but how should I interpret those percentage? I guess I have to consider them in my final U-factor result? For instance, if I get a U-factor of 0.1988 with a %EEN of 9.13%, I have to consider that my result is U=0.1988 W/m²K ± 9.13%... is that right? It might look obvious but I just wanna make sure…

No, “% Error Energy Norm” is not the error in percentage which is talked about in ISO 6946, there is some other topics in this forum which explains it, but a value under 10% is acceptable. Personally I always try to reach a value under 5%, but there are some models where little partsare surraounded by materials with quite different thermal conductivity values  (e.g. alu in XPS)

 

And finally, in my M1 wall, the brick siding isn’t considered in the calculation, but the mechanical fasteners do have an incidence on my U_T.

I found out in Annex D.3.1 of ISO 6946:2007 that ΔU_f = n_f * χ.

 

I unfortunately do not have the ISO 10211 mentioned in D.3.1, so I was wondering if you were able to give me a clue on how to find that χ value that I miss? I also did the calculation with the D.3.2 formula ( ΔU_f = α • (λ_f A_f n_f / d₀) • ( R₁ / R_T,h )²) and obtained a result of 0.0136 W/m²K that seems realistic but I’d like to validate with the D.3.1 formula.

 

To obtain the exact punctual thermal transmittance χ (chi) according to ISO 10211 a 3D software simulation is needed, which detects the typical thermal flux 3D deflection of a punctual thermal bridge. Therm can only do 2D simulations, detecting 2D deflection of a linear thermal bridge, so it shouldn't formally be used for this purpose. BUT you can make a model of your mechanical fasteners, obtain a Psi value in W/mK, multiply it for the minimal length of the fastener and obtain a "virtual" χ (chi).

For the ISO 10211 problem please send me a message to archiprato at gmail dot com

Bonne journée à toi!