Sd value vs. explicit layer

[avp081090]

Hello

Im currently trying to simulate a wooden windowframe. I would like this wooden frame to be paintcoated so that it gets a Z resistance of = 3 [GPa*m^2*s/kg]. I have tried doing this in two ways:
1. Adding a 1 mm explicit layer with a u=600, and no sorption curve and capillarity action.
2. Adding a Sd value of 0.6 m to the boundary surface.

When comparing the two situations i find that the one with the Sd-value has a much higher water content. Has this to due with the Sd-value only resisting transport by diffusion and not capillarity?

I have tried answering the above question my self by disabling the rain. This should, in theory, give identical results in the two situations but unfortunately it didn´t.

What is going on here - which of the two methods are correct? why aren´t the results identical when rain is out of the equation?

Also, the design has some recesses where air layers has been filled in. In the transition between the wooden frame and the air layer, the wooden frame doesn´t have the diffusion resistance because a Sd value can´t be defined in the line between two materials.

so personaly i think the one with the explicit layers added are the way to go.

Sorry for the rather long post but i would appreciate your help. I have attached som pictures of the frame. The pictures are showing water content for both situation 1 and 2. rain has been disabled and i would therefore expect equal results.

[Christian Bludau]

Hello,

the sd-value in the surface conditions only affects the diffusion on the surface, so if you have rain, that can intrude, and can get absorbed/ sucked in. As the drying effect is due to diffusion, so the rain cannot dry out very good.
If you have to consider rain, you have to create a layer for the material...

There is a description about how to add a water repelling layer at the surface. The problem is similar. See "Water repellent treatment of facades": http://www.wufi.de/frame_en_tutorials.html

In the transition between the wooden frame and the air layer, the wooden frame doesn´t have the diffusion resistance because a Sd value can´t be defined in the line between two materials.

No, thats right. The connection of two materials is ideal, there are no resistances between them. Here only the materials values are taken in account.

Both of your added pictures have the exactly the same boundary conditions except the sd-value and no rain? Can you please check again? Please recheck your pictures as well, the scaling is different... you can save and load the scaling colors/settings doing a right click on it in WUFI Motion.

Regards,
Christian

[avp081090]

Sry for the pictures - Here is new ones with same scaling, color coding and intevals. Both have same boundary conditions.
For these simulations i have assummed that surfaces has no capillary abilities and therefore i have disabled rain - is this a legitimate assumption?

The only difference between the two situations is that one of them have an explicit layer (covering the wood material) in the geometry (thick. 1mm) with the following properties:
-Density 460
-porosity 81%
-H.cap of 1880
-Ther.cond. of 0.093
- Diff. Res factor of 600 (corresponding to z-value of 3)
- No defined moisture storeage function.
- No liquid transport coefficients for suction and distribution.
The above can be summed up as the layer having the same properties as Eastern White Pine (database) but without moisture storage function and capillary abilities.

In the other situation a Sd value of 0.6m has been added to the boundary conditions.

Could the difference regarding moisture content in the two situations originate from not having a diff resistance between material and intermidiate airlayer in the situation with Sd-value?

thanks and best regards

[veitner]

Hello,

the heat flux is the same in both cases?

Veit

[avp081090]

the heat flux is the same in both cases?

hmm. do you mean temp. across the frame? i guess it is. The wood material used has a thermal conduction of 0.093 and the layer i have added has a themal conduction of 0.093 so there should be no difference in resistance towards heat as far as i can see - but i haven´t actually checked.

[Christian Bludau]

Hello,

there is a mathematical reason for the difference in the water contents in your two cases.
If you add the diffusion resistence at the surface as a separate layer, it contains a predefined number of grid elements (for thin layers this is the "minimum number of elements", in your case probably 2 for each half-layer). However, at the interface between the membrane and the wood, the material properties of the adjacent elements are averaged. This modifies the properties of the boundary grid element of the membrane (e.g. its effective mu-value). If the membrane consists of a small number of grid elements, this may affect the properties of the membrane as a whole and thus influenc the results.

See the picture. The left cases show the water content with a separate layer, the right cases with "sd-value" in the surface conditions. The layer in the upper left case has 20 elements, the lower left only 4 elements. The case with the higher amount of elements is closer to that one modelling the sd-value as a surface condition.

So the higher your number of elements in the membrane, the closer you can model the effect of a given sd-value.

Regards,
Christian

[avp081090]

Ahh okay, I see.

Is it possible to raise the amount of grid elements in the membrane manually? My membrane thickness is currently 1mm in thickness. maybe I should make it thicker or make more smaller grid elements?

best regards, and thanks for the help.

[Christian Bludau]

You can switch the grid to manual by doing a right click on the grid table.
Regards,
Christian