The application of artificial neural networks to predict the performance of solar chimney filled with phase change materials

In this study, the effects of phase change material on a solar chimney were studied both experimentally and numerically. Paraffin wax, as a phase change material, was installed inside the small constructed solar chimney in the campus of the University of Tehran with 1.5 m collector radius, 3 m chimney height and 20 cm chimney diameter. The goal of the paper was to illustrate the application of a properly constructed Artificial Neural Network model to predict the performance of the solar chimney, filled with phase change material. The parameters T-A and T-air were reported as the input and output data, respectively. Then, a multi-layer neural network using MATLAB software was employed to find a relationship between the inputs and outputs. The process had eight inputs and four outputs, and it was shown that the trained network was well-suited for the problem modeling. Comparing the results with the experimental data indicates the accuracy of the analysis. As a result of the research, the correlation between the predicted values by the network and the experiments, for all outputs, was more than 99%. Moreover, the average value of the relative errors for all outputs were less than 3%.