APPLICATION OF ARTIFICIAL NEURAL NETWORK (ANN) IN DRYING KINETICS ANALYSIS FOR POTATO CUBES

In this research, drying time was analyzed in the laboratory for the drying of potatoes using a fluidized bed dryer. Drying was performed at the inlet temperatures of 45, 50 and 55 °C and the air velocity of 3.2, 6.8 and 9.1 m.s-1 and bed depth of 1.5, 2.2 and 3 cm. The effects of temperature, air velocity on drying time were investigated. The results showed that with increasing temperature and velocity, drying time and process moisture loss were reduced, and with reducing bed depth, the samples were dried more quickly. The lowest drying time was obtained at the velocity of 9.1 m/s and temperature of 55 ° C, and the maximum drying time was obtained at the velocity of 3.2 m/s and the temperature of 45 °C. With increasing depth, because of high substrate thickness, drying time increased, the lowest drying time was obtained at bed depth 1.5 cm and the maximum drying time was obtained at bed depth 3 cm. Also, the neural network of feedforward type with different neurons in the hidden layer and of the transfer function, Hyperbolic tangent sigmoid and the Levenberg-Marquardt algorithm was used with a number of different Epoch. The best result obtained with Epoch 500 and R2 = 099957 and 20 neurons in the hidden layer.

Fluidized Bed Dryer Artificial Neural Network Potato Drying time