Comparison of different Methods to derive appropriate Meterological Data Sets for Yield Assessment for Concentrating Solar Power Plants

Abstract

In this study a selection of methods to generate typical meteorological year datasets that characterize the climate of a specific site is tested with regard to their suitability in assessing the long term solar resources in solar power plant financing. Three methods are explained and investigated how good they match real conditions during a 10 year measurement period on two sites in South Spain. This test is done by comparing hourly direct normal irradiance intensity frequency distributions and parabolic trough power plant simulations against the 10-year-average. The study reveals that the long term average is well represented by all investigated methods but the uncertainties in single years are unacceptably high to base a power plant financing solely on the datasets created using these methods. This paper is a contribution to currently ongoing standardization efforts within the CSP community trying to emerge a common method of solar resource assessment.