ISOSTERIC HEAT AND ENTROPY MODELING OF PISTACHIO CULTIVARS USING NEURAL NETWORK APPROACH

Abstract

Artificial neural network (ANN) method was used to predict the equilibrium moisture content (EMC) of Pistachio cultivars (Ohadi, Kalleghoochi and Fandoghi). ANN model was trained by two training algorithms. After well training of the ANN models, proved that the ANN model was better than the Oswin and Guggenheim–Anderson–de Boer models. The isosteric heat of sorption of pistachio cultivars were predicted by power models applied in this study with about R2 = 0.99. Also a regression model was developed for entropy of sorption. All cultivars had the identical pattern of variation in heat of sorption with identical moisture content levels, which might be because of the identical physical and biological structure and dimensions of these cultivars. At moisture content above 8% dry basis (d.b.), the isosteric heat and entropy of sorption of all cultivars were lower and they were highest at moisture content about 4% (d.b.). PRACTICAL APPLICATIONS The net isosteric heat equations are suggested for use in the computation of heat of sorption of pistachio cultivars whereas both the isosteric heat and entropy equations are essential to compute the humidity during simulation of stored dried pistachio cultivars. If artificial neural network models are developed using experimental equilibrium moisture content data of pistachio, the models developed can be used to predict more accurately the equilibrium moisture contents using computer and also can be used to predict the isosteric heat and entropy more accurately for modeling and simulation of drying and storage of pistachio.