Advances in Solar Radiation Estimation Techniques: A Comprehensive Rewiew

Abstract

Solar energy is a favored renewable energy source in the energy sector due to its zero emissions, environmental friendliness, low cost, and sustainability. However, meteorological factors such as weather conditions and cloud movements can interrupt solar radiation, potentially leading to undesirable outcomes in the energy sector. Solar irradiance forecasting is crucial for mitigating these adverse effects and supporting the development of renewable energy projects. In this study, the methods employed in the literature for various prediction intervals are classified, and the evaluation results of these predictions are summarized in a table. Also, an example model created with ANFIS for estimating solar radiation is presented. Imagebased and NWP models is perform well for short-term forecast horizons. To predict various time horizons, artificial intelligence-based models such as time series models, deep learning, and machine learning are prefer. Hybrid models that combine multiple methods to achieve higher accuracy are also proposed, although this increases the complexity and cost of the model. There are potential limitations in the field of solar forecasting that arise from model and data characteristics. Therefore, This study aims to guide other researchers in this field by discussing the features, limitations, and results of the models used for solar forecasting. Also, the example of daily solar radiation forecasting provided in this study offers a practical application opportunity for researchers new to this field.