Analysis of the distribution of measured and synthetic DNI databases and its effect on the expected production of a parabolic trough plant

Abstract

The generation of a Typical Meteorological Year (TMY) is a common practice in solar energy projects. Therefore it is desirable that the TMY provides not only a good estimate of the solar resource in the long term, but also an adequate input for the estimation of the performance of the project during its lifetime. The main goal of this work is to analyze and compare the effect of using synthetic and measured time series on the statistical relationships between the annual Direct Normal Insolation (DNI) values and the corresponding electricity generation of a Solar Thermal Electricity (STE) plant. For this purpose we have used two DNI databases: (a) terrestrial database with thirteen years (2000 to 2012) of 5-second measurements for the location of Seville, Spain, and (b) synthetic database generated with the Meteonorm® V 6.1.0.23 software, by applying GHI-DNI conversion models to the measured monthly values of GHI for the same location. We have used the EOS code to estimate the electricity generation of a 50 MW parabolic trough plant with 8 equivalent hours of thermal energy storage (TES) capacity. The results show that the use of synthetic series introduces an additional uncertainty that has to be analyzed for each project, being in this study the differences between the measured and synthetic data around a 2-3% for the annual values and for the main statistical parameters.