DC vs Gross DC vs Nominal DC

I think I follow.

I expected that Subarray 1 DC power gross (kW) would be the unclipped output of the panels, but it looks like voltage clipping happens somewhere before that as evidenced by changing the inverter affecting the hourly Subarray 1 DC power gross (kW) values (as seen in the attached)  .  So the general flow would be,

Raw DC output from the subarrays (totaled in Annual DC energy nominal (kWh/yr)) --> Voltage clipping produces Subarray 1-4 DC power gross (kW) (totaled in Annual DC energy gross (kWh/yr)) --> DC losses applied to produce Inverter DC input power (kW) (totaled in Annual DC energy (kWh/yr)? ) --> Power clipping, inverter efficiency, and other inverter parametersapplied to find Gross AC Output.

Is this the correct relationship between the single annual values and the hourly data? If there were no voltage clipping would the nominal DC and gross DC be the same? Or am I misunderstanding what the nominal is?

Thanks, that clarifies things.

So it sounds like if I want to proceed with my Pysam project where I combine more than 4 subarrays (essentially running the JSON files from multiple SAM tabs and combining the results), I should set the inverter arbitrarily high on max voltage so there's no voltage clipping, combine the various Subarray DC power gross (kW) values, apply the DC losses, and then use those hourly values as the Inverter DC input power variable when modeling whatever separate inverter I want to use.

Sorry, one final thing.

I get that nominal DC is POA irradiance total after reflection (IAM) * Module Efficiency. 

And that the Gross DC has voltage clipping applied to it before calculating the power.

But it seems that even without voltage clipping there's still some slight difference. between the two totals. Is there some additional module specific loss being applied here? Something to do with the operating temperatures and real efficiency while nominal DC just goes by a constant value of what's on the label, or something like that?