Estimation of the Effect of Shape and Temperature on Drying Kinetics Using MLP

Abstract

The kinetics of drying is described in the article using artificial neural networks (multilayer perceptron MLP). Drying curves for vegetables are possible to obtain theoretically on the basis of the equations of mass transfer in a porous material. A key role in these equations is played by the effective coefficient of water diffusion in the form of liquid, vapor or jointly as liquid and vapor. The diffusion coefficient which depends both on moisture content in the material and temperature should be determined experimentally. The drying kinetic curve in this article is treated as a time series dependent on the state of material prior to drying and on the constant K characterizing process parameters such as drying temperature and describing the material, e.g., its shape and moisture content. Constant K characterizes the analyzed material from the drying point of view because it contains a diffusion coefficient and depends on the shape of material. The following materials were analyzed: beetroot and potato dried in the form of cubes, slices, chips, and strips. Experiments were carried out in a laboratory oven dryer at process temperature 50, 60, 70, 90, and 106°C.

Keywords:

• Drying kinetics
• Modeling
• Multilayer perceptron
• Product shape
• Vegetables