Development of a novel 2D single coffee bean model and comparison with a 3D model under varying heating profiles

Coffee roasting produces several hundreds of aroma compounds that affect its quality. A good roasting model can not only help to control the required energy to reduce the process cost, but also ensure the product quality by cooperating with process analytical technologies and model predictive control strategies. However, current coffee roasting models are limited to constant heating temperature application due to high computational cost. Current research is developing a 2D roasting model of a single bean under varying heating profiles and is comparing it with the commonly used 3D model. The 2D model is designed to achieve minimum discrepancies with experimental data using estimated parameters of the bean's thermal conductivity and moisture diffusivity. The model performance showed that the 2D model was able to solve much faster than a 3D model, while the absolute difference in the bean temperature prediction between these two models was less than 5 degrees C. The proposed 2D model shows potential in monitoring the roasted coffee bean quality, where it can provide a cost efficient and reliable bean temperature prediction. Practical applications The application of varying heating profiles can optimize the flavor of roasted coffee beans. This study had developed a 2D model that can efficiently simulate under varying heating conditions. The 2D model also provides comparable accuracy to the 3D model and therefore, proving of its potential to be used in the modeling of batch coffee roasting.