Differences between SAM Sandia Performance Model and PVLIB

I am running a simulation of solar generation for every zip code in the U.S. using SAM and its LK scripting interface. I am also performing the same simulation with Sandia's PVLIB ( pvpmc.sandia.gov/applications/pv_lib-toolbox/ ) model. It is my understanding that SAM's Sandia PV Arary Performance Model should yield the same results as PVLIB if the weather inputs are the same. However, I have tested this using TMY3 weather files, and the AC generation output is different, even with the exact same parameter settings. Is there a possible explanation for different results?

Also, are there any available comparisons between SAM's various models (i.e., PVWatts, Sandia, and CEC) that might indicate which of these is the best model to use for countrywide simulations of AC output from residential PV systems?

Hello,

Have you tried comparing the array DC output from the two models rather than the AC output? The system AC output depends on the inverter model and loss inputs that might be different for the two models.

You might also check the loss assumptions in SAM to be sure the irradiance inputs is the same for the two models. For example, SAM has a default soiling loss of 5% that reduces the incident irradiance.

Please see the Case Studies and Validation page for a list of references to studies comparing SAM to different models:

sam.nrel.gov/case-studies

You might also be interested in the SAM Photovoltaic Reference Manual and other technical documents listed on the Performance Documentation page:

sam.nrel.gov/performance

Best regards,
Paul.