Determination of the controlling resistance to moisture transfer during drying

In this study the correlation between lag factor G and Page´s mathematical model parameter n was developed. Using the newly developed correlation the presence of internal or external resistance to moisture transfer during thermal drying can be verified. As modeling tools Artificial neural networks (ANN) and Adaptive neuro-fuzzy inference system (ANFIS) were used. The best model for determination and modeling of drying curves was selected using coefficient of correlation, chi-square, mean bias error, root mean square error and relative percentage error. Multilayer perceptron (MLP) ANN model with two hidden layers, nine neurons in each hidden layer, trained in 2500 epochs with hyperbolic activation function was found to be the most suitable for prediction of lag factor G values. The optimal model was used to develop new G-n correlation in polynomial form. When the drying behavior of particular material is described by Page´s model and the parameter n values are in the range 0.6<n