Performance evaluation of the solar collector is paramount for the design and optimization of the concentrating solar power system. To evaluate the real-time and annual performance of an ultra-high-temperature collector, a comprehensive model combining an optical-thermal model and an artificial neural network was developed. Initially, the optical performance of the collector was derived by an optical model developed by Monte Carlo ray tracing. Subsequently, by integrating the optical model with a computational fluid dynamics model, the optical-thermal performance of the collector was obtained under various operational conditions. Next, a high-precision artificial neural network model was developed using the data obtained from the optical-thermal model, with a predictive accuracy exceeding an R2 value of 0.9999. Finally, the comprehensive model was employed to analyze the real-time and annual collector performance. The results reveal significant fluctuations in the real-time optical-thermal performance throughout the year, while the monthly field efficiency, receiver efficiency, and collector efficiency demonstrate relatively moderate variations throughout the year, remaining within the ranges of 64.3%–74.9%, 76.4%–82.8%, and 52.8%–61.5%, respectively. Furthermore, the annual field efficiency, receiver efficiency, and collector efficiency can reach 69.7%, 81.6%, and 56.8%, respectively. This study offers reliable models and meaningful insights for the performance predictions and improvements of solar collectors.
Read full abstract