Prediction of pomegranate, grape and terebinth drying properties using artificial neural network approach

The objective of this research was the predication of effective moisture diffusivity, energy consumption, drying rate and moisture ratio in fluidized bed dryer of pomegranate, grape and terebinth by using artificial neural network approach. The effective factor to predict the effective moisture diffusivity and specific energy consumption were: product type, air velocity and air temperature. Four factors were used to predict the drying rate and moisture ratio as: Product type, air velocity, air temperature and time. A lot of 27 tests were conducted by a laboratory fluidized bed dryer to created training set for effective moisture diffusivity and specific energy consumption and 2165 tests for drying rate and moisture ratio. Several network and algorithms were used to train and evaluate the patterns. The best neural network feed forward back-propagation topology for the prediction of effective moisture diffusivity and energy consumption were 3-5-5-2 with the training algorithm of Levenberg-Marquardt (LM). This structure is capable to predict effective moisture diffusivity and specific energy consumption With R2= 0.9954 and 0.9805, respectively and mean-square error of 0.00009. Also the highest R2 values to predict the drying rate and moisture ratio were 0.9799 and 0.9996 respectively.