Özet:

The world’s most important source of energy, solar energy, is researched under different subjects in various fields. In particular, since the decrease of fossil fuels, the value and importance of solar energy has increased further. The first research on solar energy began in the first quarter of the 20th century, and the first study of such studies was that the solar radiation was predicted against the time of sunrise. The objective of this article is to present and implement a new method of polynomial analysis (analysis) in the nature of the forecast of solar radiation. However, since the polinom resolution will remain insufficient in predicting solar radiation, the Angström-Prescott approach with a linear (linear) model feature with Polinom was suggested. The basic principle of the PoLin (POlinom-LINeer) model is to separate the release (periodicity) from the data and then apply the Angström-Prescott model to cleared data. The results of the approach presented within Diyarbakır from the south-eastern Anadolu region of Turkey have been given the necessary advice by comparing the results of the approach with the ANFIS, HarLin and Angström-Prescott models. The output of the PoLin model was found more successful than the famous (classic) Angström Prescott, HarLin and ANFIS models.

Anahtar Kelimeler:

Özet:

The world’s most important energy source, solar energy, is being investigated in a variety of areas under different fields. Especially since the decline of fossil fuel resources, the importance of the solar energy has increased even more. Initial researches on solar energy started in the first quarter of the 20th century and solar irradiation was estimated versus sunshine duration. This study suggests similar procedure to harmonic analysis application to solar irradiation and sunshine duration data. Basis of the methodology is combined application of the POlynomial and classical LINear regression analysis. Therefore, it is referred to PoLin model as a hybrid model. It isolates first the periodicity from the daily values, and then linear regression analysis is applied first to order stationary data. PoLin results are compared with the classical Angström-Prescott, HarLin, and ANFIS models. In the application, solar irradiation site is considered from solar energy potential location in Turkey, namely, at Diyarbakır. Estimations by PoLin model appears more successful than ANFIS, HarLin and Angström-Prescott approaches.