|
Abstract
|
This study presents a mathematical modeling of eggplant slice drying process in a continuous band dryer. The experiments for drying of eggplant slices were conducted at three air temperature levels of 45, 60, and 75 °C, inlet air velocities of 1, 1.5, and 2 m/s, and belt linear speeds of 2.5, 6.5, and 10.5 mm/s. To estimate the drying kinetics of eggplant slices, different mathematical models were utilized to fit the empirical data of thin layer drying. The models were compared based on their coefficients of determination (R2), reduced Chi squares (χ2) and root mean square errors (RMSE) between the experimental and predicted moisture ratios (MR). A feed and cascade forward with back-propagation algorithm was employed to predict the moisture ratio (MR) and drying rate (DR). The effective moisture diffusivity varied from 3.40 × 10−9 to 1.13 × 10−8 m2/s. The activation energy varied from 14.18 to 25.09 kJ/mol. The obtained results show that the feed forward back-propagation network with training algorithm of Levenberg–Marquardt, 4-5-5-2 topology, threshold functions of tansig–tansig–tansig can able to predict the moisture content and drying rate with R2 values of 0.9992 and 0.9726, respectively. Comparison of ANN results with mathematical models revealed that mathematical modeling yields better accuracy to predict the moisture content and drying rate of eggplant.
|