Simulação da secagem da soja com coeficiente de difusão variável

Abstract

The soybean needs to be dried after harvest to decrease the risk of deterioration and to increase the storage time. The mathematical modeling of this process is important for the simulation of new operating conditions. Two numerical strategies for the solution of the second Fick's law with variable diffusion coefficient are proposed in this paper with the objective to assess the moisture profile during drying. It is noteworthy that, the diffusion coefficient is a nonlinear function, therefore the equation describing the drying process does not have analytical solution. The first strategy is to derive the equation in the radial direction using the product rule and then discretize this direction by Finite Difference. In the second strategy, the original equation is directly discretized in the radial direction and the properties between two points of the mesh are calculated through linear interpolation. In both strategies, after discretization in the radial direction, the systems of ordinary differential equations (ODEs) resulting are integrated in time using Matlab®. The simulations are performed using the experimental conditions reported by Barrozo et al. (2006), which solves the model by Orthogonal Collocation. The results of two strategies are compared with experimental data. These comparisons indicate that both strategies were effective in the soybean drying simulation, but the first strategy provides faster convergence.