Determining typical weather for use in solar energy simulations

Abstract

Predicting the performance of a solar energy system by using simulation methods requires weather data input for the locality involved. The present paper describes a method of analyzing an optional number of years of weather data for a chosen month resulting in a “typical week” which is characterized in terms of solar radiation, ambient dry bulb temperature and wind speed. The “typical week” is allowed to vary in length between 5 and 10 days in the analysis in order to enable selection of a period that best represents a given month according to specified criteria. Verification of the method by comparative computer analysis was performed using two forms of weather data as inputs to the solar energy program “TRYNSYS”. The averaging method when compared to the “typical” weather method resulted in differences of less than 7 per cent. The use of “typical” weather appears to give results at least comparable with more established methods while at the same time providing a broad spectrum of the weather typical of an area. The use of “typical” weather can result in savings in computer time.