Open brtietz opened 2 years ago
Best to do https://github.com/NREL/SAM/issues/1146 first, so that we can do stronger output comparisons while solving this issue. Moving to patch 1.
Additional source of large discrepancies: PV smoothing uses the same power number for charging and discharging. In the default case (DC sized battery) these can be different by a factor of 8% due to AC and DC losses.
Battery losses appear to already be included in PVSmoothing output power, AC wiring losses are not included.
Availability losses for the battery are an issue. The battery attempts to charge, availability losses reduce the amount of PV available, battery reduces charging, and so-on in a loop. It probably doesn't make sense to fix this until https://github.com/NREL/SAM/issues/985 splits the PV and battery availability losses.
PV smoothing still has some discrepancies between PV Smoothing Outpower (the output used to calculate violations) and System Power Generated (the final power output of the simulation). A major cause of this is that losses are not provided to the PV Smoothing algorithm. In the defaults case, this causes a constant, 1% difference in output power.
There are two other inputs that need to be set properly for these outputs to align. The user should get warnings to prevent issues: • The “minimum time at charge state” input on the dispatch page being shorter than the ramp interval • If simulating a DC connected battery, the inverter efficiency cutoff could cause trouble.