Computer Science · Artificial Intelligence
Quantifying Performance Loss Rates of Photovoltaic Modules Using Ground-Based vs Satellite-Based Meteorological Data
Ebrar Özkalay, Alessandro Virtuani, Andrew Fairbrother, A. Skoczek, Gabi Friesen, Claude Ballif
See on the mapAbstract
Accurate assessment of the long-term performance of photovoltaic (PV) systems is critical for manufacturers, investors, plant owners and O&M companies. M any PV systems, especially residential and some commercial/industrial systems, are not equipped with meteorological monitoring systems such as irradiance and temperature sensors. Whereas to calculate the performance ratio (PR), a common performance measure, the plane of array irradiance (GPOA) is an input parameter. The irradiance data is the largest contribution to uncertainty in PR. Therefore, in this study, we first evaluate satellite-based irradiance and compare to ground-based irradiance data. Next, we investigate the accuracy of using the satellite-based meteorological data in long-term performance analysis of three test modules in the absence of ground-based data. We use (1) ground-based and (2) satellite-based data separately as input irradiance. We calculate 60 PR time series and performance loss rates (PLR) for each irradiance data using different filtering methods and performance metrics. As a representative PLR value, we used the mean value (excluding outliers) of the various PLR values obtained as suggested by IEA PVPS Task 13. For the first module, this method gave almost identical PLR values (-0.5% error) when ground- and satellite-based data are used. For the other two modules, the values are not identical, but standard deviations largely overlap. We show that when the mean of PLR values is used, it is possible to accurately evaluate the long-term performance using satellite-based meteorological data.