Use of artificial neural network in forecasting optimal distance of enclosures containing PCM-introduced for improving the performance of the evacuated tube solar collectors

Abstract

In this study, the usefulness of using regression-based methods in estimating the optimal fins distance in a PCM finned enclosure of the evacuated tube solar collectors was evaluated. Enclosure containing fins and PCM can absorb more thermal energy from the working fluid of a evacuated tube solar collector than a conventional enclosures. Optimal fins distance is a serious challenge in designing the PCM finned enclosures. Increasing the fins distance reduced the PCM melting rate and consequently decreased absorbed thermal energy content. Also, for the short fins distance, the heat transfer rate within PCM diminished owing to lower convection intensity which led to lower absorbed energy. Based on the artificial neural network and response surface methodology results, the optimal fins distance was estimated with a maximum error of less than 2.36% and 4.31%, respectively. The trend of optimal distance variations was navigated by these methods with the mean square error of 0.0012 and 0.0026. Statistical criteria of R-square for the former method and the latter one were 0.999 and 0.998, respectively. By designing a heatsink in which the number of fins is optimal, the stored energy can be maximized without the heatsink temperature reaching its maximum.