Prediction of cabbage seed drying at lab and industrial scale using a 'non-equilibrium' sorption isotherm model.

Drying is an important step in seed treatments to ensure high seed quality as well as storability. As drying can also negatively affect seed quality and is a time and energy intensive process, a mechanistic drying model was built to improve trial-and-error research and apply it to fluidised bed trials. First, material characteristics such as the sorption isotherm and diffusivity were measured and modelled. For the isotherms, it was found that a 'non-equilibrium' isotherm described the drying data better than the fully equilibrated measured isotherm. For the 'non-equilibrium' isotherm, the Flory-Huggins model was extended with an addition in glassy state including an elastic parameter by Leibler & Sekimoto. The different sorption behaviour during drying is hypothesized to occur due to accumulated mechanical stresses in the material. Microscopy and X-ray tomography support this hypothesis as air pockets in the seed appear during drying. The single seed model including the 'non-equilibrium' isotherm could predict drying behaviour of thin-layer dried seeds at different drying conditions up to 40 °C very well with a percent error below 10 %. Further, lab- and industrial-scale fluidised bed trials were performed, which were very reproducible and comparable. The uniformity of mixing in fluidised beds was checked to assess the applicability of the single seed model. Mixing was uniform and the single seed model could predict fluidised bed drying well. Thus, the model has high predictive power for moisture decrease for thin-layer as well as fluidised bed drying at lab and industrial scale. Limitations are drying temperatures above 40 °C, due to a not-optimal diffusivity model and no isotherm data. The proposed model can be a useful tool for optimising seed drying processes, if combined with a quality model as a next step.