Mathematical and artificial neural network modeling of hot air drying kinetics of instant �C_m� brown rice

Abstract

Modeling moisture content variation under variable hot air dryers is challenging. In this study, mathematical models and artificial neural network (ANN) were investigated for modeling of instant “Cẩm” brown rice drying process. The experiments were done in four levels of hot air temperature (55, 60, 65, and 70 °C). The results demonstrated that among eight mathematical models, the diffusion approach could give the best prediction of moisture ratio during the drying process with the highest R-square and lowest mean square error. Besides, the ANN model with 10 hidden layers also could provide the best-fit model with the same criteria as the mathematical model. Compared with the ANN model, both can give a highly accurate prediction. However, the ANN model could be more beneficial in the up-scale process.