Table partition

[Barbadori]

Hi there,

I always design my layouts in PVCase and I export them into PVsyst. I always use three different types of tables in order to maximise the space available. In landscape, I do tables of 20 panels, 10 panels and 5 panels long. These lengths obviously don't match with the real length of a string (between 25 and 30 modules).

In the partition section in PVsyst, I tend to put 1 in the Nb of rectangles in width (X) as it would be very hard to calculate the exact length of the string based on the configuration of the tables I mentioned above.

My concern is by doing that, am I considering a string with infinitive number of panels? If that is the case, how accurate is the results I am getting when running the simulation?

Could anyone advise what is the best practice when designing tables in PVCase (or similar) to get an accurate partition in PVsyst?

Many thanks in advance.

[dtarin]

I assume you are modeling them four in height? So 4x20, 4x10, 4x5? This is one misalignment between PVsyst and PV plant design; PVsyst doesnt have a way to model fixed tilt with terrain or as-built table sizes and proper partition sizing, leading to an underestimation of electrical shading losses. In your case, I would model as follows:

4x20 Table - 4 h x 1 w partition

4x10 Table - 2 h x 1 w partition

4x5 Table - 1 h x 1 w partition

If half-cell module, 100% electrical effect; if full-cell, maybe less than 100% (when oriented landscape). PVsyst should really work on this shortcoming in the software, as it has been present for years. Allowing and calculating for fractional partitions could be a short term fix.

[Barbadori]

Thanks for the response.

Yeah, I usually do 4 in height but also 6 landscape is a configuration that tend to use quite often.

It is good that the latest versions allows you to do different partition depending on the size of the table, it makes easier to do the partition you mention in your message.

I agree that this is something that PVsyst should be working on. The partition can affect significantly the annual yield.

Regards