Modelling of osmotic dehydration of mango (Mangifera indica) by recurrent artificial neural network and experimental design.
Zita N. E. B., Emmanuel A. N., Patrice K., Ismael D., Benjamin Y.
Author Affiliation: Laboratoire des Procédés Industriels, de Synthèse et de l'Environnement (LAPISEN), Institut National Polytechnique Houphouët-Boigny (INP-HB), BP 1313 Yamoussoukro, Côte d'Ivoire.
Research Journal of Agriculture and Biological Sciences 5 : 754-761
Abstract : Osmotic dehydration process was investigated in this paper. Mango slices samples were osmotically dehydrated in different hypertonic solutions of glucose and sucrose at three different temperatures (30, 50 and 60°C) without agitation. A full factorial design, used for experiments, was shown significant effect of sugar concentration on water loss. But no significant effect was observed for sugar type or temperature on this response (water loss), due certainly to small thickness (1 mm) of samples. The modelling of the process with recurrent artificial neural network proves the ability of this method for osmotic dehydration process modelling with good accuracy (R>0.967).