Shading, single module strings & microinverters

I was trying to see if I could compare the performance of a string inverter against a micro-inverter for beam and self-shading and I have come up against hurdles that I was wondering if there was any work arounds?

I was using the3D shade calculator.

I believe you are limited to 8 shaded strings. Not really enough strings to model a microinverter system? Is the 8 string limit per project or per sub-array?

In your microinverter example file you state: Because SAM's self-shading model does not account for maximum power point (MPPT) tracking of individual modules, it is not suitable for use with microinverters, so you should set the shading mode for self-shading to "none."

This confuses me a little. I would have thought that you model MPPT tracking for each inverter, whether the string length is 20, 10, 5 or in fact 1? So in your example if you set the modules along row to 1 and the modules along bottom of row to 1, then you have 14 self-shading rows? I know this construction is strange/dumb, but would in not accurately predict what was happening? Once you try to increase the count along the bottom row above 1, this is where it fails as it can’t treat each “string” as individual modules?

Thanks
Mark

Hi Mark,

When you specify beam shading losses by time step (by assigning a loss value to each time step of the simulation, either from your own shade analysis or from SAM's 3D Shade Calculator), SAM allows you to input a column of beam shading loss values for up to 8 strings per subarray: You edit the shading losses for each subarray in a separate window. When you specify different shading losses for different strings, SAM combines the string losses into subarray losses using the method you choose: Average of strings, minimum of strings, maximum of strings, or database lookup (Note that the "Database lookup" method for converting string losses to subarray losses is broken in SAM 2016.3.14 r1, and will be fixed in r2, which we plan to release in the next week.)

That's for shading of the array by nearby objects. The self-shading model is separate and works differently. The beam shading factors reduce the beam irradiance incident on the subarray, while the self-shading model calculates a reduction in the incident diffuse irradiance and in the subarray's DC output. The reason it is not suitable for microinverters, even if you configure the subarray as you describe with one module per row, is that SAM treats multiple inverters like one big inverter with a single input. It does not assign a different input voltage to each inverter. So, if the front module in your configuration is not shaded and operates at a different maximum power point than the shaded modules behind it, it would not model the effect of the front module's inverter operating at a different voltage than the rest of the inverters.

Best regards,
Paul.

Thanks Paul
So beam shading calulation is a simulation, by reducing irradiaton, of what would actually happen? Not the actual effects on voltage and current of the modules and strings and in turn inverters that shading produces?

For micro inverters, could you replicate inter row shading in the beam shading calculation and 3d tool by doubling the row spacing and placing a module shaped objects between the rows? Complicates putting in other near shading objects.

So for microinverters, optimizers, multiple mppt inverters, different inverters in the same project, you have to move away from the single inverter model? Quite a task, no doubt?
Thanks

Hi Mark,

That is correct. The beam shading losses simply reduce the beam irradiance incident on the array according to shading percentages you specify.

For an array with fixed rows (no tracking), you could draw the rows of modules in the 3D Shade Calculator, and use that as a way to calculate row-to-row shading. But that approach would be more time consuming than using SAM's built-in self-shading calculator.

We are in the process of considering how to move away from the single inverter/module paradigm in SAM, and yes, as you say, it is a bit involved.

Best regards,
Paul.