Integrated optical-thermal model and deep learning technique to estimate the performance of a conical cavity receiver coupled solar parabolic dish collector

An integrated optical -thermal model was developed by adopting ray tracing using the Monte -Carlo algorithm and Computational Fluid Dynamics (CFD) for evaluating the performance of a 40 m2 parabolic dish collector (PDC) at various inclinations under varying wind characteristics. CFD was used for modeling conjugate heat transfer of the absorbed solar heat to Therminol 66, a heat transfer fluid (HTF) in the conical -shaped receiver tube, while considering the heat losses. The model validation was done against the well -established studies. The sun conditions were modeled using raytracing in SolTrace, adopted for the intricate receiver geometry using ANSYS (R) mesh and MATLAB (R) code. The ray count of 106 was appropriate for raytracing to the meshed receiver surface containing quadrilateral elements. Utilizing a user -defined function, the SolTrace heat flux distribution was used in the ANSYS thermal model as the heat source. The variations in thermal performance with the inclination angle, wind speed of varying directions, and HTF inlet temperature are discussed. For the dish considered, 66-80% of the absorbed solar heat was carried by the HTF under varying inclinations, wind direction, and wind speed. Further, deep learning was introduced for predicting thermal performance. The various activation functions were compared, and the best was implemented. Lastly, the performance of the present model was compared with the existing models.