Dipartimento di Ingegneria informatica, automatica e gestionale

This study presents a resilient model for accurately predicting annual solar radiation in Ghardaia, Algeria, utilizing a locally-sourced database. The model integrates temperature, humidity, wind speed, and pressure as inputs. A combination of machine learning and deep learning techniques, including convolutional neural networks and conventional neural networks, are employed to forecast direct normal irradiance and diffuse solar radiation. This comprehensive approach uses multivariate regression analysis, validated with established databases for high-resolution analysis in data-scarce regions. The findings highlight the model’s effectiveness in providing precise forecasts and outline potential applications for optimizing solar energy use in similar climates.