Evaluation of thin-layer drying models and neural network for describing drying kinetics of Lasagnas angustifolia L.

The thin-layer drying behavior of Elaeagnus angustifolia in a laboratory scale hot-air dryer was examined. Drying characteristics of Elaeagnus angustifolias were determined using heated ambient air at temperatures of 50, 60 and 70°C and air velocities of 0.5, 1 and 1.5 m/s. To select a suitable drying curve, 6 thin-layer drying models were fitted to the experimental data. All the models were compared according to three statistical parameter; R², standard error of estimate (SSE) and root mean square error (RMSE). Using some of the experimental data, an ANN, trained by standard Back-Propagation algorithm, was developed to predict MR and DR values based on the three input variables (time, velocity and temperature). Different activation functions and several rules were used to assess percentage error between the desired and the predicted values. According to the results, a two-term drying model has better agreement with experiment. The effect of the drying air temperature and air velocity on the drying model constants and coefficients were also determined. Consequently, the estimating power of the new model was evaluated. The ANN model was able to predict the moisture ratio and drying rate quite well with coefficient of determination (R²) of 0.9993, 0.9992 and 0.9996 for training, validation and testing, respectively. The prediction Mean Square Error was obtained as 0.00355, 0.00930 and 0.0016 for training, validation and testing, respectively.