Application of Generalized Regression Neural Network in Predicting the Performance of Natural Convection Solar Dryer

Abstract

The performance evaluation of a natural convection solar dryer is a complex one because of the transient and non-linear nature of atmospheric conditions. In this comparative study, a smart neural network -based tool was developed for estimating the performance of such a transient nature solar dryer. For this purpose, a series of experimental studies are conducted through four successive days and compared with the generalized regression neural network (GRNN) modeling. GRNN architecture proposed in this study consists of three inputs (time duration, irradiance, and ambient temperature) and four outputs (drying chamber temperature, the mass of moisture removed, drying rate, and dryer efficiency). Such generalized regression neural network architecture was trained, tested, and validated with real-time experimental variable data sets. The results of the GRNN model are in good agreement with experimental results. The overall accuracy of the proposed GRNN model in predicting the performance is 96.29%.