Complicated Facades

[rip_xellos]

Hello guys, its my first post on the forum. I'm having quite some difficulty in replicating an existing system on PVsyst for research purposes.

 

 

You can see the String layout here and the actual building below

https://www.google.no/search?q=Oseana&biw=1680&bih=949&source=lnms&tbm=isch&sa=X&ved=0ahUKEwiWz_K8jfrLAhVCuxQKHdcmBsMQ_AUIBigB#imgrc=fOlDxxj0NPASKM%3A

As you can see the number of modules in Strings is not the same and the Strings are in anywhere from 2 to 6 rows depending on which string you consider. My issue arises when I can't place modules belonging to a string in two different orientations. It makes no sense to do that because it would cause mismatch losses but that is exactly what I want to study.

I don't see a way to do this in PVsyst.

[André Mermoud]

First, please be aware that this image is not directly issued from PVsyst.

If it is based on a PVsyst configuration, it has been redrawn.

Now for reproducing this configuration, you have to know the details of the project (more exactly the calculation version) which was used as the basis of this configuration.

There were probably several sub-arrays, with different MPPT inputs and string lengths.

Then, you have to reproduce the 3D scene that has been used, i.e. the exact shape of the façade and the holes.

[rip_xellos]

Hello Andre, I have tried to contact the engineers to get detailed calculations but have not received any response from them. I have completed the 3D scene with several variations.

However I was having difficulty distributing the strings among 4 inverters with 8 orientations so I had to reduce the number to 4 orientations.

4 orientations

With this current 4 orientation model I am able to achieve a system close to the original with 359 modules ( original has 363 modules ), The string layout I think is hard to replicate and since the engineers wont get back to me, I have to proceed with what I have.

 

The aim of my project was to find out mismatch losses because of the shape of the building ( google - Oseana, Norway), but now that I'm simulating with 4 different orientations with no strings crossing over into different orientations(impossible in PVsyst) I don't know how to proceed. Do you have any suggestion for studying mismatch?