Simulação Computacional para Secagem de Grãos

Abstract

This dissertation presents computational simulations and mathematical modeling from a drying process of soybean of a cultivate Monsoy 7575, whose source of observations comes from a laboratory experiment carried out in thin layers by Guimarães (2005), under specified drying conditions. The calculated results of these simulations were compared with the moisture contend measurements (wet basis) in their respective instants, registered in this experiment. Through predictive statistical analysis of these data, a general mathematical model and equations applied to represent drying with supplementary heating of this experiment were developed in this dissertation; the variables considered were: moisture contend, specific weight and drying time of these seeds. Two methods of determining a coefficient of , component of the general equation of this model, were considered: the first from an average of results and the second calculated at each moment of the drying process; which generated good results in both cases. The calculated moisture contend of the soybean, using a medium coefficient calculated to this equation provenient of the model developed in this work, produced good results, with a correlation coefficient equal to 0,99 in relation to the measurements of the moisture contend registered by Guimarães (2005). Also the drying time curve calculated with an equation developed from this mathematical model, produced good results. It is observed in this work that the general equation, which represents the proposed model, elaborated through the predictive statistical analysis of the experimental data registered by Guimarães (2005), is well indicated for drying simulations of others granular agricultural products. An electronic spreadsheet, elaborated to the production of the soybean equations, from the drying model developed, is appropriate to the production of equations for others species of grains and seeds. This study also demonstrates, as elaborated in the spreadsheet, the statistical procedure used to increase the accuracy in the results, reducing the residue.