Convective heating of wheat germ. Model development and experimental validation

PENCI, MARÍA CECILIA; MERILES, SILVINA PATRICIA; CURET, SEBASTIEN; RIBOTTA, PABLO DANIEL; BOILLEREAUX, LIONEL

Stuttgart

Congreso; 5th International ISEKI Food Conference. "The Food System Approach: New Challenges for Education, Research and Industry"; 2018

Univeristy of Hohenheim

The relevance of wheat germ lies on its potential as an ingredient in the human diet, which has not yet been fully exploited. For its use in food formulation, it is necessary to treat the germ thermally in order to inactivate enzymes and increase storage time. The aim of the present study is to determine wheat germ properties that best describe the convective heating process. Numerical results were validated with experimental measurements. The nonstationary equation representing heat transfer was solved using a finite element computational code (COMSOL Multiphysics 5.0). The particulate material was considered as a porous media composed of a solid matrix of wheat germ and void spaces filled with air. The thermodynamic properties were calculated using a volume averaging model to account for both solid matrix and fluid properties. A 2D axisymmetric system was considered to describe the heat transfer process using a cylindrical shape (high 5 cm; radius 1 cm). Thermal losses were accounted considering a heat transfer coefficient of 5 W/m2K at the lateral faces. At the upper air-germ interphase the coefficient was estimated by heating a cylindrical aluminum bar in the same equipment and position. To validate the model, samples were placed in a polystyrene mold and experimentally treated in the chamber for 1800 s at 60 °C with a relative humidity of drying air of 30%. Time-temperature profiles were recorded by 6 t-type thermocouples placed at different heights on the main axis of the cylinder. The resulting heat transfer coefficient was 28.4 W/m2K. Specific heat capacity, measured by differential scanning calorimetry varied between 1,91 to 2.45 kJ/kgK. The wheat germ properties that fit the model were bulk density: 413.83 kg/m3, porosity: 0.687, thermal conductivity: 0.245 W/m°C. The root mean square error and root mean square percentage error used to measure the differences between predicted and averaged experimental temperature profile, were 1.41 ± 1.11 and 3.33 ± 1.97%, respectively. Thus, satisfactory model predictions were achieved with the parameters detailed above, with a maximum difference in the temperatures at the surface of the sample.