Shading electrical loss for diffuse

[S.Faulkner]

It appears that there is no shading electrical loss for diffuse irradiance, and I am wondering why?

I noticed this when modelling a large tracking system with backtracking ie no shading of the beam component, which gives an irradiance loss of about 1% but zero electrical loss (According to module strings, 100% electrical effect). I set albedo to zero so that there is no related loss to confuse matters, ie the irradiance loss is entirely diffuse. I think I have set it up correctly but this lack of electrical loss makes me wonder. I'm expecting the electrical loss in this case to be roughly 1%. If I understand correctly, it should be at least as large as the irradiance loss, and can't be zero since the cells at the lowest point on the module see a smaller portion of the sky than the higher up cells.

If I turn off backtracking there is as expected more irradiance loss (beam as well as diffuse) and an electrical loss that is larger than the irradiance loss, apparently coming only from shading of beam irradiance.

As a check with no backtracking, I set the horizon very high so all the beam irradiance is blocked, and got a similar result - some irradiance shading loss but no associated electrical shading loss.

So, what could be stopping the electrical loss, is this incorrect, or have I misunderstood something? (I have used 100% electrical effect and calculated According to modules strings)

[André Mermoud]

The diffuse shading loss is supposed to be "uniformly" distributed: you don't see any "shades" on the array due to the diffuse componrent.

Now the elctrical losses are due to a mismatch when some cells only are shaded within a string (or an array). This cannot occur with the diffuse.

[S.Faulkner]

I'm sorry but I still don't understand why there isn't electrical loss for diffuse shading. Doesn't the amount of diffuse shading vary with position on the module? Does the electrical effect not apply to small differences in irradiance (such as diffuse) in the way that is does to large (shading of beam irradiance)?

As I understand it, the diffuse shading (after taking account of irradiance behind the POA) varies from zero at the top of the array to a maximum at the bottom, related to the angle above the horizon of the next array that is causing the shading, ie how much of the sky is blocked by the next array. At the top this angle is zero, and increases towards the bottom. If I have read the descriptions correctly, PVsyst calculates the diffuse shading loss as an integral of this varying shading?

As an example, for array "height" 1.976m (one module in portrait), pitch 3m, tilted at 30deg. Diffuse shading factor (excluding loss due to sky behind POA) is 1 at the top, 0.74 at the bottom, and the average across the array is 0.91. If the module performance is determined by the cell with lowest irradiance (electrical effect), then the shading factor for diffuse-irradiance is 0.91 and for diffuse-electrical is 0.81 (0.74/0.91).

As noted earlier this make a difference of roughly 1% to the overall result, which just meets my threshold for being worth assessing.