Mathematical modeling of wheat drying with input from moisture movement studies using magnetic resonance imaging

Canada annually produces around 50 Mt (million tonnes) of grains and oilseeds worth about 6 billion dollars. Drying is the most important method for preservation of grains. Wheat was selected because it is a major crop in Western Canada and is staple food for most of the world population. An MR imaging probe (Helmholtz configuration, 7 mm internal diameter) was designed and constructed to specifically fit a single wheat kernel for drying studies at temperatures ranging from ambient to about 60°C. Prior to MRI, wheat kernels were preconditioned to a known moisture level. Individual kernels were placed into the imaging probe to be placed into the bore of the vertical MRI magnet and drying was started and continued for 4 h at 30, 40, and 50°C at a constant nitrogen gas flow (∼0.23 m s-1). Two-dimensional MRI data were acquired continuously using a Hahn spin-echo pulse sequence and saved at equal time intervals (about 10 min based on the set-up parameters) without interrupting the drying process. Samples of intact kernels (with all three components: pericarp, germ, and endosperm), mechanically scarified kernels with incisions in the pericarp, and kernels with the germ removed were dried at the above-stated temperatures. A calibration curve of MRI image intensity versus the actual moisture content of the grain was also obtained, using NMR spectra of grains at different, known, moisture contents. Mathematical models based on the real-time conditions from the present research were developed to simulate simultaneous heat and mass transfer during wheat drying.; The MR images clearly showed a non-uniform moisture distribution inside the intact-wheat kernel before and during drying. Further analysis of the MR images revealed that moisture loss from the seed parts differed significantly during drying and was dependent upon the grain components. The rates of moisture change in the pericarp, endosperm, and germ during drying were clearly different. The rate of moisture loss was slower from the endosperm region, whereas the pericarp region dried faster at the initial stages of drying. Furthermore, water had the tendency to move from the endosperm towards the germ to finally move out from the grain. These observations are an important consideration to develop accurate grain drying models with well-defined initial and boundary conditions. In case of the mechanically-scarified kernels, water was released relatively faster from the scarified regions of the pericarp, as compared to the intact kernels. This is expected because the intact pericarp behaves as a moisture barrier after the initial drying and hinders the movement of water from the kernel. In the case of germ-removed kernels, water moved out in a uniform manner from the kernel. A proper synthesis of the results from these analyses led to the development and validation of a mathematical model to follow the moisture distribution inside a grain kernel as a function of drying time and initial moisture content. The model was developed based on non-uniform moisture distribution in the wheat components at the beginning of drying. The new grain drying model is expected to significantly advance the control of mass transfer during drying to maintain both the quality of the product and the economy of the process.