How to evaluate the effect of by-pass diodes in shaded arrays?

[André Mermoud]

Let's consider row arrangement (sheds) systems only. In this case when one row puts its shade on the next one, the bottom cell row is primarily shaded.

Let's call submodule the set of cells protected by one by-pass diode. In most modules (60 or 72 cells), there are 3 by-pass diodes and therefore 3 "sub-modules", usually disposed in length within the module.

Let's suppose that the modules are disposed in landscape, and all modules of the bottom row belong to the same string.

Now it is often believed that when the bottom cell row is shaded, the by-pass diodes will limit the electrical loss to the sub-module row, i.e. the string electrical production will remain 2/3 of the normal production.

This is only true (about) if you have one only string per MPPT (violet curve).

If you have more than 2 strings in parallel, the electrical shading factor is shown on the next figure: this shows that as soon as 1/3 of the sub-modules are affected by the shade, the electrical loss is 100%!

 

Electrical shading factor according to the number of shaded submodules in a string

(this plot may be constructed point by point using "Tools" /"Electrical behaviour of PV arrays")

 

Therefore in rows arrangements, the protection diode don't recover any electrical loss: the full string width (module with) should be considered as electrically inactive as soon as the bottom cell is shaded. With the calculation "according to string" in the 3D model, this means that with row systems the Fraction for electrical loss should always be set to 100%. With "Unlimited sheds", the width of the full string should be considered.

The next figures show how the I/V curves are affected: the first shows a not-shaded string

 

One only string without shadings

 

and when 1/3 of sub-modules are shaded: we loose in voltage, but the current remains the same.

If the string is alone in the array, the Pmpp loss is proportionnal to the number of shaded sub-modules.

 

One string with 1/3 submodules shaded

Now when there are other strings in parallel, the MPP is chosen on the resultant I/V curve, and the voltage Vmpp is the same for all strings: for the shaded string, only the diffuse contributes !

 

1/3 shaded string in a 3-strings array