Dear WUFI Users,

I am wondering how the case of a photovoltaic installation on a sloping roof (10 degrees of pitch) with a 4 cm air void under the installation should be properly simulated?

I have seen articles about flat roof and rack PV installation (model using correction coefficients of 0.3 and 0.5), one seems to me that my case is different. The installation completely covers the roof surface, there is a 4 cm air void between the wind proofing layer and the PV installation.

I made a detail model including the void and the photovoltaic panel. I made the panel as a 1 mm surface (vapour retarder) with changed parameters so that the thermal conductivity properties and other corresponded to the real PV panel.

However, the question arose what type of void should be used: "Luftschicht 40 mm" or "Luftschicht 40 mm; ohne zasätzl. Feuchtespeicherung"? I made calculations for both voids, but the results contradict the description of the phenomenon of moisture accumulation in the material.

According to the database, the void "Luftschicht 40 mm" generates a non-ralistically high water content in the material, so it is recommended to use "Luftschicht 40 mm; ohne zasätzl. Feuchtespeicherung."

The another question, what air exchange should I use in the air void? From my calculations I came up with 7/h, is this a realistic value?

The outdoor air climate was selected from the WUFI database and assigned to the exterior surface of the PV panel. The inner climate as normal moisture load according to the external climate and the sides of the detail as adiabatic.

With this setting, the relative humidity in the void "Luftschicht 40 mm; ohne zasätzl. Feuchtespeicherung" is almost all the time 100%, while with the void "Luftschicht 40 mm" it varies from 18% (summer) to 89% (winter). The variant with a void "Luftschicht 40 mm" seems more real, but I find the moisture behaviour contradictory to the description seen in the WUFI database.
I will be grateful with explanation of this phenomenon and advice on how best to simulate such a case.

kind regards

Thomas

## The air void under the PV installation

### Re: The air void under the PV installation

Hi Thomas,

I guess you like to simulate the roof construction underneath instead the explicit PV installation.

At the moment we are running some field tests to evaluate the behavior of sloped roof constructions with PV-installation. But up to now the trails still running, so we can give you only little information.

Due to the fact, that there are many unknown parameters, the explicit simulation of such constructions under consideration of all layers are always a little bit difficult – and a validation by measurements is necessary. Instead, I would recommend simplified models with effective surface transfer coefficients.

So here is a few information how I would simulate this construction – some insights from the current field tests mixed up with existing parameters.

At first, please see our guideline for the hygrothermal simulation of ventilated pitched roof constructions:

https://wufi.de/en/service/downloads#ve ... ched_roofs

This is basically the procedure for the simulation of pitched roofs – if you do not have a second ventilated air layer in below your watertight felt or metal, this might be able to catch the situation already quite well.

As you already mentioned there are some existing reduction factors to adapt the radiation parameters for the simulation of flat roofs with PV-rack installations. These are the coefficients out of the “WTA-Merkblatt 6-8”. For the shortwave absorption the factor is 0.3 and for the longwave emission the factor is 0.5.

Combining the information of the guideline with these factors will generate a first approach of effective transfer parameters to simulate your roof construction with the PV-installation. However, as far we could see up to now, the results may far on the safe side. This is caused by the longwave emission which is still too high for the plane-parallel installation of the PV-modules on top of the roof surface. An additional reduction of the emission e.g., by a factor of 0.2 brings the hygrothermal behavior in a range we have observed during our field tests.

As soon as we have more reliable information, we will publish it.

If you decide to proceed with the explicit layers, there are a few more things to consider.

At first the air exchange between the tiles and the sub-roof, here the ACH is in a range between 50 to 150 1/h, depending on the construction. These values where measured during filed tests on different sloped roof constructions. (But I still recommend using the effective transfer parameters instead of an ACH!)

The next thing is the air exchange between the PV-installation and the tiles. I assume that the value of 7 1/h is far to low. Due to the free air gaps on eaves, ridge and both sides I guess the values are higher then underneath the tiles.

A further point is the energy which is discharged by the PV-Modules. These are around 20% which are not available for heating up the construction. This factor isn´t easy to consider because it´s a radiation dependent parameter, so a constant heat sink won´t give you correct values.

Coming to the air layers:

The problem is that probably condensation occur in the gap between the tiles and the PV-modules. The air layer without an additional moisture storage function isn´t able to store this water. This leads to saturation of the air layer and therefore this might lead to convergence problems. Here I would recommend proceeding as described here:

viewtopic.php?f=32&t=1777&p=5226#p5226

Hope this helps you solving your simulation.

Kind Regards

Philipp

I guess you like to simulate the roof construction underneath instead the explicit PV installation.

At the moment we are running some field tests to evaluate the behavior of sloped roof constructions with PV-installation. But up to now the trails still running, so we can give you only little information.

Due to the fact, that there are many unknown parameters, the explicit simulation of such constructions under consideration of all layers are always a little bit difficult – and a validation by measurements is necessary. Instead, I would recommend simplified models with effective surface transfer coefficients.

So here is a few information how I would simulate this construction – some insights from the current field tests mixed up with existing parameters.

At first, please see our guideline for the hygrothermal simulation of ventilated pitched roof constructions:

https://wufi.de/en/service/downloads#ve ... ched_roofs

This is basically the procedure for the simulation of pitched roofs – if you do not have a second ventilated air layer in below your watertight felt or metal, this might be able to catch the situation already quite well.

As you already mentioned there are some existing reduction factors to adapt the radiation parameters for the simulation of flat roofs with PV-rack installations. These are the coefficients out of the “WTA-Merkblatt 6-8”. For the shortwave absorption the factor is 0.3 and for the longwave emission the factor is 0.5.

Combining the information of the guideline with these factors will generate a first approach of effective transfer parameters to simulate your roof construction with the PV-installation. However, as far we could see up to now, the results may far on the safe side. This is caused by the longwave emission which is still too high for the plane-parallel installation of the PV-modules on top of the roof surface. An additional reduction of the emission e.g., by a factor of 0.2 brings the hygrothermal behavior in a range we have observed during our field tests.

As soon as we have more reliable information, we will publish it.

If you decide to proceed with the explicit layers, there are a few more things to consider.

At first the air exchange between the tiles and the sub-roof, here the ACH is in a range between 50 to 150 1/h, depending on the construction. These values where measured during filed tests on different sloped roof constructions. (But I still recommend using the effective transfer parameters instead of an ACH!)

The next thing is the air exchange between the PV-installation and the tiles. I assume that the value of 7 1/h is far to low. Due to the free air gaps on eaves, ridge and both sides I guess the values are higher then underneath the tiles.

A further point is the energy which is discharged by the PV-Modules. These are around 20% which are not available for heating up the construction. This factor isn´t easy to consider because it´s a radiation dependent parameter, so a constant heat sink won´t give you correct values.

Coming to the air layers:

The problem is that probably condensation occur in the gap between the tiles and the PV-modules. The air layer without an additional moisture storage function isn´t able to store this water. This leads to saturation of the air layer and therefore this might lead to convergence problems. Here I would recommend proceeding as described here:

viewtopic.php?f=32&t=1777&p=5226#p5226

Hope this helps you solving your simulation.

Kind Regards

Philipp