Development of a Multiscale Computational Model for Mass Transfer in Osmotic Dehydration of Nigerian Root Vegetables

Osmotic dehydration is a widely studied pre-treatment for fruit and vegetable processing that reduces water activity and energy requirements for subsequent drying while preserving sensory and nutritional quality. The mass transfer mechanisms governing osmotic dehydration are multiscale in nature, involving simultaneous water and solute transfer at the cellular, tissue, and product scales. Existing single-scale models inadequately capture the structural heterogeneity of tropical root vegetables such as yam, sweet potato, and cocoyam. This study develops an original multiscale computational model for mass transfer during osmotic dehydration of these three root vegetables produced in North Central and South East Nigeria. The model operates at the cell membrane scale using a thermodynamic transport model for cell turgor and osmotic equilibrium, at the tissue scale using an effective medium approximation, and at the product scale using a finite element representation of the macroscopic concentration gradient. The multiscale model is solved numerically using COMSOL Multiphysics and validated against experimental osmotic dehydration data collected at sodium chloride and sucrose concentrations of 10 to 40 percent at temperatures of 30 to 60 degrees Celsius. Experimental validation uses magnetic resonance imaging to visualise internal water distribution profiles during dehydration. The model predicts mass transfer parameters and product quality trajectories with significantly lower root mean square errors than conventional first-order Fick's law models. The study contributes an original multiscale modelling framework and experimental validation methodology to the food engineering literature. Keywords: osmotic dehydration, multiscale modelling, mass transfer, root vegetables, finite element