Estimation of Daily Average Global Solar Radiation with Nonlinear Regression Models Developed Using Some Meteorological and Geographical Parameters

Authors : Selçuk USTA, Cafer GENÇOĞLAN, Serpil GENÇOĞLAN

Pages : 589-597

Doi:10.24012/dumf.1130793

View : 22 | Download : 6

Publication Date : 2022-09-30

Article Type : Research Paper

Abstract :In this study, it is aimed to develop empirical models that can be used in estimation of daily average solar radiation insert ignore into journalissuearticles values(RS); based on some meteorological and geographical parameters. Seven estimation models were developed by nonlinear regression analysis method using various combinations of air temperature insert ignore into journalissuearticles values(T);, relative humidity insert ignore into journalissuearticles values(RH);, extraterrestrial radiation insert ignore into journalissuearticles values(Ra);, saturated insert ignore into journalissuearticles values(es); and actual vapour pressure insert ignore into journalissuearticles values(ea); parameters. The models were created using the long-term average daily meteorological data of Kahramanmaraş province insert ignore into journalissuearticles values(1938 – 2020);. The models were tested both these long-term average data and daily meteorological data measured at Kahramanmaraş Sütçü İmam University insert ignore into journalissuearticles values(KSU); in 2019 and 2020. Long-term average daily actual RS data varied between 4.99 – 32.56 MJ m-2 day-1. The estimated solar radiation values insert ignore into journalissuearticles values(insert ignore into journalissuearticles values(`RS` ); ̂); with the highest correlation insert ignore into journalissuearticles values(r = 0.99); with actual RS data were obtained with the RS_7 model, in which the parameters es, ea, T, RH and Ra were used together. The insert ignore into journalissuearticles values(`RS` ); ̂ values obtained using this model varied between 6.45 to 33.99 MJ m-2 day-1. For the RS_7, which showed the best performance among the seven models, MAPE and RMSE were determined as 4.17% and 0.69 MJ m-2 day-1, respectively. The daily RS values measured in KSU varied between 7.75 – 33.48 MJ m-2 day-1 and 10.51 – 30.23 MJ m-2 day-1 for 2019 and 2020, respectively. The insert ignore into journalissuearticles values(`RS` ); ̂ values closest to the measured RS values were estimated with the RS_7 model. The estimated insert ignore into journalissuearticles values(`RS` ); ̂ values by this model varied between 11.74 – 33.93 MJ m-2 day-1 and 13.93 – 31.57 MJ m-2 day-1 for 2019 and 2020, respectively. MAPE values were determined as 11.33% and 7.54%, respectively. It is concluded that this model can be used to estimates daily average solar radiation and will be an excellent alternative since it is compatible with the Kahramanmaraş conditions.
Keywords : Estimation Model, Nonlinear regression, Solar radiation