Confusing behavior re: shade factors

Comparing outputs from SAM between an old version (2014.11.24) with the latest (2016.3.14 rev 2) I'm seeing some output I don't understand regarding shading. On the shading tab, the new version reports no shading losses, and in the Time Series, shows a constant beam shading factor for the array of 0.95. I notice that there's a variable called "Subarray 1 DC Shading Factor" that appears to properly follow my shading inputs, but it doesn't affect the "POA beam irradiance after shading and soiling" outputs at all.

It does seem like the new version is doing something with the shading, in that if I turn off the shading input, the output increases, but it doesn't actually show up in the outputs that talk about shading.

File attached.

Hello,
There are a couple of different things going on here. I agree that the results are confusing.
The beam shading losses that you specified on the Shading and Snow page by clicking Edit shading for Subarray 1 affect on the POA beam irradiance. For hours that you specified a beam shading loss of 100%, the value of the Subarray 1 POA beam irradiance after shading and soiling (W/m2) output should be zero (note that this is the beam irradiance, not the total irradiance). You can also see that in the Subarray 1 Beam irradiance shading factor (frac): When the beam shading loss input is zero, the beam shading factor should also be zero. SAM applies both that beam shading loss and the soiling loss from the Losses input page. By default, the soiling loss is a constant 5% throughout the year. If you do not change the soiling loss, the resulting beam shading fraction will always 0.95 times whatever the beam shading fraction is for that hour. You should also see a reduction in the POA total irradiance for hours that the beam shading loss is not zero. That all looks fine to me in your file.
The Subarray 1 DC shading factor (frac) output looks suspect to me. I think that value should be zero for all hours in your file because you are providing a single set of beam shading losses for the entire subarray. That output is for a new shading model we introduced in SAM 2016.3.14 that converts beam shading losses for different strings in each subarray to a single set of losses for the entire subarray. It should only apply when you set Strings in Subarray 1 in the Edit Shading window to a value greater than 1. For some reason, even though you set that value to 1, when you choose the default Database lookup for the Method for converting string losses to subarray, SAM appears to be calculating and applying a DC shading fraction from that model when it should be ignoring it. To avoid this issue, in the Edit Shading window where you specify the time series beam shading losses, you should choose an option other than Database lookup, such as Maximum of strings. If you do that, you should see Subarray 1 Beam irradiance shading factor (frac) reflecting the beam shading losses you specified (combined with the soiling losses), and and teh associated reduction in the subarray's POA beam irradiance.
Best regards,

Paul.

Paul,

Thanks for your reply. I'd like to point out that changing nothing but Database Lookup to Maximum of Strings produces a 500 kWh (~7%) discrepancy in the yr 1 output, with "Maximum of Strings" (which I interpret as the worst losses) actually producing more power than the database lookup. Is there any indication as to which would be "correct"?

Thanks!