Reducción de la incertidumbre en datos de radiación solar para la mejora de la simulación de sistemas fotovoltaicos
PV system simulations are used to estimate the energy yield of new installations and assess the performance of PV materials in different regions. This thesis focuses on reducing the uncertainty of these simulations by quantifying and decreasing the uncertainty in solar radiation data, which currentl...
Guardado en:
Autor principal: | |
---|---|
Otros Autores: | |
Formato: | text (thesis) |
Lenguaje: | eng |
Publicado: |
Universidad de La Rioja (España)
2018
|
Acceso en línea: | https://dialnet.unirioja.es/servlet/oaites?codigo=184952 |
Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
Sumario: | PV system simulations are used to estimate the energy yield of new installations and assess the performance of PV materials in different regions. This thesis focuses on reducing the uncertainty of these simulations by quantifying and decreasing the uncertainty in solar radiation data, which currently accounts for around 50% of the total uncertainty.
Simulations seldom use solar radiation measurements due to the scarcity of ground sensors. However, the uncertainty in measurements is the basis of most solar radiation studies. We found that low-cost photodiodes present substantially larger uncertainties than thermopile pyranometers if they are inadequately calibrated. The uncertainty further increased due to operational failures, which were very common in regional and agricultural networks, leading to uncertainties in measurements higher than those of the best radiation databases. Moreover, these defects were not detected by the most widely used QC methods, such as the BSRN tests. Hence, we developed a new QC procedure, the BQC, that identified most operational defects and some equipment errors by analyzing the stability of the deviations between several radiation databases and measurements.
Solar radiation estimations are customarily used to assess PV systems due to their extensive spatiotemporal coverage and high resolution. We verified that databases from geostationary satellites, such as SARAH or NSRDB, should be preferred to assess the solar resource because they present the smallest bias and uncertainty. We have also evaluated the potential of reanalyses to complement satellite-based data in high latitudes. We confirmed that former ERA-Interim and MERRA reanalyses should be avoided, but we found that ERA5 and COSMO-REA6 are valid alternatives to satellite-based databases. These results validated the incorporation of both reanalyses in the online simulator PVGIS. However, users should take into account their limitations; primarily the strong dependence of their deviations on the atmospheric transmissivity due to the incorrect modeling of clouds. The analysis of the uncertainty propagation through PV simulations confirmed that SARAH should be preferred to assess PV systems in Central and South Europe, whereas it revealed that ERA5 is the best alternative in Northern Europe. We also found that cloud-related errors in reanalyses amplified the bias through the simulations. These amplifications should be accounted for selecting databases because their magnitude is sometimes larger than the bias of solar radiation estimations |
---|