Why are SolarNexus PV estimates different from PVWatts / CSI EPBB calculator?
2016-05 Update: SolarNexus no longer uses Clean Power Estimator's "PV Simulator." All PV simulations are run using PVWatts. There are no longer differences between SolarNexus and PVWatts simulations.
-------------------------------------
SolarNexus uses PVWatts, for the preliminary estimates on the Systems screen. SolarNexus also uses Clean Power Research's (CPR) "PV Simulator" for PV production estimates on the Analysis screen. The production estimates used in all the proposals and analysis are provided by the CPR PV Simulator. PVWatts uses a different simulation models than CPR's "PV Simulator." Although they are similar, they will never produce identical output. CPR's PV Simulator is a paid simulation service that provides arguably more accurate estimates than the free service provided by PVWatts.
Comparing the estimated production results in our Analysis to those from the PVWatts or the CSI EPBB calculator is a complicated question, but luckily one that has been asked before. The short answer is that significant differences are expected both due to using different data sources and differences with the actual simulation models.
Weather Data Sources
The latest PVWATTS simulation model uses a weather data source called TMY2, and optionally TMY3. SolarNexus uses CPR's PV Simulator, which uses TMY3. The main difference between the two is that TMY3 has a much larger number of ground stations that it uses as reference points. As a result, for many locations, a simulation done using TMY3 data will be much more accurate, as it will be able to find a closer ground station to draw typical annual solar irradiance data from.
Comparing PV Simulator to PVWatts
There is one difference to note that will significantly affect results. PVWATTS factors in a number of extra derate inputs to account for factors not included in the component ratings themselves. CPR's analysis of the EPBB calculator suggests that the total of these derate inputs amounts to an approximately 17% effective derate. CPR's PV Simulator has a similar input (PVOutputAdjustment) that is used to factor in all “other losses” associated with a complete PV system (line losses due to wiring, etc.). SolarNexus uses a product of the derate factors that you input as the PVOutputAdjustment for the PV Simulator.
Comparing PV Simulator to the CSI EPBB Calculator
The CSI Go Solar calculator (CSI EPBB calculator) uses the PVWATTS version 2 simulation model. The CSI EPBB calculator is known to be inaccurate, and specifically is known to underestimate real-world production. This fact has been recognized by the authorities in CA. For your reference, here is a link to a report from the CPUC detailing these findings: http://www.cpuc.ca.gov/NR/rdonlyres/CE1D2316-405C-4C94-A805-A68A1988D640/0/2010APA_final.pdf. On average, CPR found that in 2009 the CSI EPBB calculator underestimated production by an average of 9% for systems that they have on record. That calculator continues to use the old TMY1 dataset which is likely to be much less accurate than TMY3. Consequently, you cannot really make an "apples to apples comparisons" between what you see in SolarNexus as estimated output and the Go Solar calculator. We believe our simulation engine using TMY3 data is much more accurate than the CSI EPBB calculator.