A Thermal Performance Evaluation of a Medium-Temperature Point-focus Solar Collector Using Local Weather Data and Artificial Neural Networks

Abstract

Theoretical investigation of a medium-temperature point-focus solar collector performance using an artificial neural network (ANN) is presented. In practice, the ANN model can be employed to model the performance of solar collectors with complex structures when other models may have difficulties. The objectives of the study were to assess the sensitivity of the ANN estimations to different combinations of input parameters as well as to determine the minimum amount of inputs necessary to accurately model solar collector performance. The data used for training the ANNs were derived analytically employing thermodynamics and energy equations. The study used the principal weather variables and operating parameters affecting the point-focus solar collector performance, which are the monthly average beam solar insolation, monthly average air temperature, monthly average wind velocity, wind incidence angle, and the absorber wall temperature. Among the ANN-models, ANN9 including of all input variables, gave the best estimation with the accuracy of more than 99%. Overall, the smallest value of RMSE = 0.0300 and the highest value of Rtest = 0.9998 were obtained for the ANN9 with two hidden layers and the topology of 5-6-10-1. Satisfactory results for the proposed point-focus solar collector suggest that, with sufficient input data, the ANN method could be extended to estimate the performance of other solar collector designs in different climate regimes and operating conditions.