Prediction and optimization of performance parameters of solar collectors with flat and porous plates using ANN and RSM: Case study of Shahrekord, Iran

It is necessary to adopt measures to make optimal use of rich sources of solar energy, considering the limit resources of fossil fuels and increasing demand for energy. Using solar collectors is one of the methods to collect solar energy. The aim of this study was an investigation and prediction of the effect of the collector's slope angle and plate type on the collector performance using artificial neural network (ANN) approaches. Two different ANN models, multi-layer perceptron (MLP) and radial basis function (RBF), were applied to predict the performance parameters of solar collectors. The results indicated that the ANN-MLP and ANN-RBF models were acceptable, but the ANN-RBF model with structure of 2-16-5 had higher acceptability for prediction of output parameters at various input variables. The highest R and lowest root mean square error (RMSE) values in training set of ANN-RBF for air speed (AS) were obtained as 0.984 and 0.015, respectively. The results showed that increasing the slope angle from 0 to 60 caused to increase in the amount of received radiation by the collector, and so leading to an increase in the air temperature inside the collector. On the other hand, the maximum value of thermal power (TP) and collector efficiency (CE) for the porous plate were more than the simple plate by 28.2 % and 41.5 %, respectively.