This study presents the integrated use of experimental, numerical (CFD), and machine learning (ML) methods to enhance the thermal efficiency of parabolic trough solar collectors (PTC). In the first phase, experimental tests were conducted on a laboratory-scale PTC designed and manufactured under Elazığ climate conditions; environmental parameters, including solar radiation, ambient temperature, wind speed, and relative humidity, were recorded. Based on the data obtained, the inlet and outlet temperatures of the collector were determined, and the system's thermal efficiency and exergy performance were calculated. In the second phase of the study, the flow and heat transfer behaviour of the PTC was analysed using a three-dimensional CFD model with ANSYS Fluent. In this context, a mesh independence test was performed, and the model reliability was verified with experimental results. Two different methods were investigated to improve performance: waste metal (WM) placed inside the absorber pipe and helical channel modification (MG) in the internal geometry. Numerical and experimental findings show that the WM method increases the outlet temperature by 14–24% and is approximately 1400 times more cost-effective than the MG design. Furthermore, exergy efficiency and CO 2 reduction analyses reveal that the use of WM significantly improves environmental sustainability. In the final stage of the study, PTC outlet temperature prediction models were developed using the M5P and PACE regression algorithms; R 2 values of 0.9932 and 0.9252 were obtained for M5P and PACE, respectively. These results demonstrate that AI-based models are effective in quickly and reliably predicting PTC performance.
Murat Catalakaya (Tue,) studied this question.