Module independent VS dependent shading

[Martin]

Hi,

When simulating a particular project and reading the EB, I can see that the module independent shading, which I normally assume to be due to the horizon (far) shading is just below 6%.

Looking to the module dependent shading, I can see that this is at around 16%.

Two aspects are interesting when it comes to this:

1. The total shading loss over the year, presented in the summary, is 23.2%.
2. Removing the horizon from the 3D model reduces the module dependent shading almost entirely.

Knowing how the help file defines independent and dependent shading, I am a little confused by these two contradicting aspects.

Regarding (1), what is the logic inferred here?
Shouldn't the loss presented in module dependent shading be a share of the incoming radiation converted at STC into a certain available energy quantity? How can the relative loss before and after STC conversion be added together, when the loss on module level would be much smaller if corrected for the conversion efficiency?
 

Regarding (2), what is the relation between horizon (module independent) shading and near (module dependent) shading?
It is reasonable to expect that removing the horizon would remove the module independent shading loss. However, as it seems, it actually also removes the module specific shading.
How can this be? I would expect these two losses to be distinguishable and non-interactive.

 

Appreciate any explanation, as I haven't been able to find an elaborate explanation elsewhere.

[developer_mh]

Hi Martin,

could you provide an example project so that we can reproduce the situation?

Thanks a lot in advance, kind regards,

Martin

ps: sorry for answering so late, we are a bit busy with the new release right now..

[Martin]

Hi,

I am afraid the project file is too big for the limit imposed here.

I could send you this by email.

However, the situation should be possible to reproduce with any project using horizon profiles.

I have attached two horizon profiles and corresponding EB's and shading loss as stated in summary.

To me it seems strange that the total shading loss is a sum of two percentage losses, when these are separated by some conversion factor. This is for me the foremost reason for writing this post.

Furthermore, it also seems strange that the module dependent shading entry isn't reserved for the internal shading of the model alone, while the module independent shading is reserved for horizon losses. I would regard the losses of the (near) shading model and geometry of installed collector surfaces in relation to this to be decided on a clear sky day and the horizon losses also, but separately. As it is now, the one influences the other.

I can understand that relative near shading losses are lower when the horizon (far shading) losses are lower, as there is more irradiation available. Similarly, I can understand why higher horizon losses makes the near shading loss higher, as the near shading loss is based only on the actual amount of irradiation reaching the modules. This is how you have defined it. I still don't really agree with the logic, however.

 

DEM_hor.HOR

PVGIS_hor.hor

[developer_mh]

Hi Martin,

yes, you could send me the projects to my mail address which I will send you in a minute via private message. Please understand that for me it is a lot easier to explain what is going on if I have the project file, but I will try to give a general explanation.

What distinguishes the three different measures "Module-independent shading", "Module-specific Partial Shading" and "Yield Reduction due to Shading" is (roughly) the following.

Yield Reduction due to Shading is the result of simulating the whole PV system twice and compare the electrical output of the unshaded PV system with the electrical output of the "real" PV system. We have to simulate twice because the electrical output of a PV system is not proportional to the percentage of shading as this leads to lower module temperatures, lower currents and ohmic losses and so on.

Module-independent shading is the minimum percentage of the diffuse irradiation (not direct!) that is not reaching the modules. Imagine it like this: For each module in the PV system we receive a diffuse shading factor from the 3D visualisation. The diffuse shading factor is a geometrical ratio of the portion of the sky dome that is visible to the module. It is 0 if the whole sky dome is visible (no horizon and no other shading objects) and 1 if the module doesn't see anything of the sky. Now we determine the minimum diffuse shading factor that occurs in the PV system and input this in the calculation of the module-independent shading. When we simulate, the diffuse irradiance on the modules is reduced by this factor for each time step.

Module-specific Partial Shading is the electrical loss of energy in the modules due to shading of diffuse and direct irradiance. That is, for most of the modules there is a diffuse shading fraction left (except for the one with the lowest shading factor) that has to be taken into account, as well as all the shading of the direct irradiance by the horizon or near objects. In contrast to the first section in the Energy Balance, where we list irradiation energies in kWh/m², the second section lists electrical energies in kWh. Due to the electrical configuration of the modules it can happen that a geometrical direct shading factor of 0.05 leads to an electrical loss of 0.33, if a whole module substring is going down. So that's an important aspect here.

Hope that helps... if that doesn't answer your questions fully, please don't hesitate to send me your project files.

Kind regards,

Martin