Distributed grid-connected photovoltaic power system emission offset assessment: statistical test of simulated- and measured-based data

This study assesses the pollutant emission offset potential of distributed grid-connected photovoltaic (PV) power systems. Computer-simulated performance results were utilized for 214 PV systems located across the US. The PV systems monthly electrical energy outputs were based on a performance calculator called PVWATTS. Offset emissions of sulfur dioxide (SO2), carbon dioxide (CO2), and nitrogen oxides (NOx) were determined from PV system outputs and average utility emissions data from each state. For validation, the simulated monthly results were statistically compared with measurement-based data (both production and corresponding emissions data) from 29 PV systems installed at different sites across the US. While the data shows high (geographic) variability, the substantial number of measurements allows reliable statistical analysis. The methods are found to give consistent results in spite of the necessity to employ some even quite crude input approximations—such as the use of statewide rather than specific emissions data for the systems. No significant differences between simulated and measured monthly means for any of the pollutants were noted on the basis of individual monthly analyses, though the results for NOx suggest the possible existence of some difference in that case. A more detailed statistical modeling using all monthly data in one combined analysis (allowing improved variability estimation) confirms these conclusions. Even the shorter confidence intervals for expected offsets obtained through the combined analysis show no significant differences between simulated and measured methods for SO2 and CO2. The differences for NOx are statistically significant but consistent—suggesting useful prediction by the simulations via a constant correction factor. As expected, significant differences between months are evident for both simulated and measured offsets. Published by Elsevier Ltd.