Dynamic modelling for optimisation of pellet softening

Mathematical models for the crystallisation process in pellet reactors have been developed during recent decades. In this paper, these models are extended with new insights in the pellet softening process. The diffusion of the supersaturated water to the pellet surface is included in the model and the fluidisation model of the bed is modified. The new model is calibrated in a pilot plant setup. Calibration results are validated in two different full-scale plants. The validation process shows that, even under basic process operation, the model can be used to identify a malfunctioning apparatus and identify undesired process conditions. For the normal situation the model gives satisfactory results in predicting fluidised bed porosity and water quality parameters such as calcium, pH, conductivity and m-alkalinity. The validated model is used to show that pellet size control in the reactor has significant influence on performance with respect to the water quality parameters. Maintaining the pellet size in the bottom of the reactor at 0.8 mm compared with 1.4 mm reduces the supersaturation of calcium carbonate in the water after the reactor by 50%, but increases the consumption of seeding material by 550%.