الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 220 |  |  |
| | | from | 220 | | |
Abstract
The utilization of artificial intelligence (AI) technology for environmental and renewable energy applications has been expanding in recent years. One of the most widely used AI techniques for this purpose is artificial neural networks (ANNs). Due to the shortage of solar measurements for different locations around the world, various solar radiation-forecasting methods have been developed, wherein models that rely on sunshine parameters outperform other meteorological-based models. However, sunshine data is not as widely accessible as other meteorological parameters, such as temperature parameters. Therefore, this study aims to develop temperature-based global solar radiation (GSR) models using ANNs, a commonly utilized approach in machine learning techniques, to predict GSR solely using temperature data. Additionally, the study compares the performance of these models to the commonly used empirical technique.Furthermore, the study aims to develop precise GSR models for twenty-seven sites and the entire region of Egypt that currently lack AI-based models. It also examines the impact of varying validation dataset lengths on the prediction and accuracy of solar radiation models, an aspect that has received little attention. In addition, the study optimizes the design of ANN, one of the most widely used machine learning algorithms, to achieve accurate global solar radiation forecasting while minimizing computational requirements.