Breathable roofing felt

[johannbjorn]

Dear support,

I am a bit confused by how a membrane is simulated vapour open to the outside but water tight.

I have a d=1.6mm PVC roofing felt (PROTAN SE) on the exterior surface of an unventilated roof-structure.
The Sd-value of the PVC membrane is 22m (ISO 12572:2001)
Watertight at 10kPa according to 1928:2000 (A)

Option 1: (including felt-element)
I build an 1.6mm element to the exterior roof boundary in WUFI.
I find the vapour diffusion resistance factor : µ = sd/d = 22m/0.0016m =13750.
I apply the µ-value to the element.
I apply liquid transport coefficient (A-value) (suction and redistribution) to zero.
Next would be to configure the exterior boundary condition of the roof.
There I should change the Sd-value to zero or 22m?

Option 2: (excluding felt-element)
Only edit the Sd-value of the exterior boundary and change it to 22m.
Would option 2 give same results as option 1?
Is there something else to consider?
What about the water-tightness is it not neccessary to include the water tightness of the invisible membrane-element (A-value)?

I. DIFFUSION:
i) How can vapour pass through the roofing membrane if the vapour diffusion resistance is so high?
ii) i.e. is vapour only passing from inside of the structure to the outside or is it both ways? Is this directly related to the pressure of the ambient air? but the WUFI guideline state:
,, µ is therefore practically independent of temperature and pressure, i.e. it is a constant which only depends on the material in question"
ii) If the relative humidity is high outside all year, let's say 90%, would this not mean that the relative humidity inside the structure would always be the same as the exterior air? i.e. if we have wood rafters inside the structure would this not make it deteriorate relatively quick?

[Thomas]

I have a d=1.6mm PVC roofing felt (PROTAN SE) on the exterior surface of an unventilated roof-structure.
The Sd-value of the PVC membrane is 22m (ISO 12572:2001)
Watertight at 10kPa according to 1928:2000 (A)

Option 1:
I build an 1.6mm element to the exterior roof boundary in WUFI.
I find the vapour diffusion resistance factor : µ = sd/d = 22m/0.0016m =13750.
I apply the µ-value to the element.
I apply liquid transport coefficient (A-value) (suction and redistribution) to zero.
Next would be to configure the exterior boundary condition of the roof.
There I should change the Sd-value to zero or 22m?

Hi johannbjorn,

yes, that is one possible option.

Do NOT set the Sd-value to 22 m in this case, because the diffusion resistance of the roofing felt is already accounted for by including it as a layer in the assembly. Therefore set Sd = 0 m, otherwise the diffusion resistance of the roofing felt would be accounted for twice.

Option 2:
Only edit the Sd-value of the exterior boundary and change it to 22m.
Would option 2 give same results as option 1?
Is there something else to consider?
What about the water-tightness is it not neccessary to include the water tightness of the invisible membrane-element (A-value)?

This is the other possible option. Both options should give the same results.

If you wish to include the material in the assembly (option 1), you need the full material data set for this material (such as bulk density, thermal capacity, thermal conductivity, ...), even though most of these material parameters will not be relevant for the simulation result.

However, if the layer only has a small number of properties which are really relevant for the hygrothermal simulation, it may be possible to omit the layer from the assembly and to take account of its few relevant properties by adjusting some other settings. A typical case is a thin membrane acting as a vapor barrier or a vapor retarder. The only two relevant properties of these thin membranes are their vapor diffusion resistance and their water tightness.

By setting the Sd-value to 22 m, you have accounted for the diffusion resistance of the roofing felt. The water-tightness can be taken into account by "switching the rain off": Simply set the "Adhering fraction of rain" in the surface transfer coefficient dialog to zero.

The short wave radiation absorptivity should be set to the absorptivity of the roofing felt (but that's true in both options), and if you use the calculation mode "explicit radiation balance" the long-wave radiation emissivity is also needed and should also be set to the value of the roofing felt (and that is also true in both options).

I. DIFFUSION:
i) How can vapour pass through the roofing membrane if the vapour diffusion resistance is so high?

An Sd-value of 22 m is not really high. The vapor flux through such a membrane may be small, but it adds up over the weeks, months and years...

For example, if on one side of a 1 m thick air layer the vapor pressure is 1871 Pa (corresponding to 20°C/80%), and on the other side 936 Pa (corresponding to 20°C/40%), then the resulting steady-state vapor diffusion flow through the air layer is

g = 1.93e-10 * (1871 - 936)/1 = 1.80e-7 kg/(m2 s)

So 0.18 milligram of water vapor are passing through each square meter each second. That is not much, but after one day it's 16 grams, and after one year it's 5.6 kilograms. May not be negligible.

Your roofing felt has an sd-value of 22 m and thus the same diffusion resistance which a 22 m thick air layer would have. The vapor flow through the roofing felt is thus by a factor 22 smaller than in the example above (assuming the same conditions).

ii) i.e. is vapour only passing from inside of the structure to the outside or is it both ways?

The water vapor diffusion flow is driven by vapor pressure differences. If the vapor pressure is higher outside than inside, vapour will diffuse from the outside towards the inside. If the vapour pressure is higher inside than outside, vapour will diffuse from the inside towards the outside. These conditions may change from hour to hour, depending on the variations of the outside climate and the inside climate. It is WUFI's task to compute the variable vapor flow under these variable conditions...

Is this directly related to the pressure of the ambient air? but the WUFI guideline state:
,, µ is therefore practically independent of temperature and pressure, i.e. it is a constant which only depends on the material in question"

Water vapor diffusion flow is driven by differences of the water vapor partial pressures on both sides of the layer under consideration, not by differences of the total ambient pressure.

(Differences of the total pressure of the ambient air may drive convective air flows through the component if there are leakages. This is not diffusion, however).

There is a very slight influence of the total ambient pressure on the strength of the diffusion flow because it slightly affects the density of the air through which the diffusing molecules must travel. However, this influence is very small, and in any case the ambient pressure is slightly affecting the strength of the diffusion flow, it is not driving the diffusion flow.

ii) If the relative humidity is high outside all year, let's say 90%, would this not mean that the relative humidity inside the structure would always be the same as the exterior air? i.e. if we have wood rafters inside the structure would this not make it deteriorate relatively quick?

In general, no. Let's consider a case with constant conditions for simplicity: The moisture exchange between inside and outside happens across the roofing felt. The roofing felt is permeable to water vapor but not to liquid water. Moisture exchange therefore happens exclusively by vapor diffusion (not by capillary liquid transport) and is therefore driven exclusively by vapor pressure differences.

If the vapor pressures inside and outside are different, this difference causes vapor diffusion transport which tries to equalize the vapor pressures. After a while the vapor pressures inside and outside will be the same (in our simplified constant case), say 2105 Pa. The relative humidity is the vapor pressure divided by the saturation vapor pressure corresponding to the current temperature. If the outside temperature is 20°C, the relative humidity of the outside air is 2105/2339 = 90 %. If the inside temperature of the roof is 25°C, the relative humidity in the roof is 2105/3170 = 66 %.

The wood rot risk is determined by the relative humidity, not the vapor pressure, therefore the wood in this roof will be safe.

If the roof had the same temperature as the ambient air, then the relative humidities would be the same. In an unoccupied house (i.e. no heating) deep in the forest (i.e. no solar warming) the inside temperature will be close to the outside temperature - this house may deteriorate quickly.

In reality, the conditions are usually not constant enough to achieve complete vapor pressure equalisation. WUFI's task is to compute the hourly state of the roof under variable conditions, balancing hourly moisture gain and moisture loss in order to evaluate the moisture state resulting under these conditions. And if capillary moisture transport occurs at the same time (driven by differences in relative humidity, not by differences in vapor pressure), things become even more complex. But WUFI knows how to do this...

Regards,
Thomas

[johannbjorn]

Dear Thomas,

I am so deeply grateful for your detailed and explanative answer and explanation on this topic.

My understanding on diffusion flow has been expanded much higher and I am now more confident in applying this to constructions and understanding WUFI's task to compute the variable vapor flow under these variable conditions.

Thank you very much,

All the best, Jóhann