Exploring the impact of nanoparticle aggregation in parabolic trough solar collectors with a neural network-based predictions for enhanced thermal performance

Parabolic Trough Solar Collectors (PTSCs) efficiently harness solar thermal energy, making them highly suitable for industrial heating, electricity generation, and water desalination processes. However, their thermal performance is often limited by thermal losses. To address this, the present study investigates how the presence and absence of nanoparticle aggregation affect the heat transfer characteristics of a Casson nanofluid flowing through the cylindrical absorber tube of a PTSC. The dimensional form of the proposed model is converted into a standardized non-dimensional form with the help of adequate transformations. The non-dimensional form of ordinary differential equations (ODEs) is solved utilizing the Runge-Kutta method with the integration of shooting technique. Furthermore, a machine learning-based regression analysis using an artificial neural network is employed to develop a predictive model for thermal transmission with high accuracy. The Levenberg-Marquardt algorithm is applied in conjunction with well-structured datasets for training, testing, and validation phases. The outcomes reveal that the effects of magnetization, slip, and Casson parameters are more pronounced in the case of without aggregation than with aggregation. Moreover, an essential impact of the study is observed in solar power plants, food processing, solar thermal desalination, and solar cooling.