PVWatts_loss was applied on AC calculation

[yxh289]

Hi,
In the description of pvwatts_losses function, the losses are applied on DC side:
'Calculates DC power losses according the PVwatts model using pvwatts_losses() and self.losses_parameters.`'
However I found that in pvlib.modelchain.ModelChain.pvwatts_losses, PVwatts loss was applied to AC.
'def pvwatts_losses(self):
self.losses = (100 - self.system.pvwatts_losses()) / 100.
self.ac *= self.losses
return self'
When system have high DC/AC ratio, AC power output clips at a lower level as a result of these losses applied on the AC side.

[wholmgren]

thanks @yxh289. pull request welcome. In a rush to finish off 0.6.2 so going ahead with a PR.

[cwhanse]

Concur with the bug report. May have been due to coding following the older PVWatts description, rather than the v5 document. Or a simple typo.

[pasquierjb]

The post-inverter losses such as the AC wiring losses and AC Availability should in theory only apply to the AC power and not be subjected to inverter clipping.

[kandersolar]

@pasquierjb conceptually that might be true, but in this case we're just sticking with the original spec: per Section 11 of the PVWatts Version 5 Manual, "Losses in the system that are not explicitly modeled are provided by the user as a percentage of DC energy."

See also #1350