New mass-based population balance model including shear rate effectsapplication to struvite recovery

Laburpena

Struvite (MgNH4PO4·6H2O) precipitation is a promising solution for phosphorus recovery in wastewater treatment plants. Controlled struvite precipitation can help to reduce eutrophication in the receiving waterways, fight global phosphorus scarcity and reduce operational problems generated by the uncontrolled precipitation of the mineral in the pipes. Due to the generated interest, the description of the precipitation process has been already included in existing wastewater treatment modelling libraries. However, following the classic wastewater treatment modelling approach, the process has been generally included as a one-step kinetic model. This one-step model type is limited for technological design and optimization purposes, as it does not include information about the mechanisms by which the precipitation occurs, nor the particle size distribution, a key variable for the performance of struvite as an effective fertilizer. Therefore, the aim of this thesis has been to upgrade existing one-step kinetic models by developing a mathematical model that could describe in detail the mechanisms occurring in struvite precipitation in order to be able to predict the resulting particle size distribution. This model is a population balance model in which hydrodynamic effects have been considered. The population balance model has been constructed according to Ceit’s plant wide model methodology, guaranteeing mass and charge balance. Therefore, it can be combined with the simulation of other unit processes used to describe wastewater treatment plants in a systematic and straightforward way. A sensitivity and collinearity analysis performed in the thesis, demonstrated that the model is coherent in its structure and valid to represent struvite precipitation processes. In order to incorporate the hydrodynamic effects to the model, results obtained in an experimental campaign where struvite precipitation was analysed under different mixing and saturation conditions in two different experimental set-ups, were used. Obtained results showed that a higher mixing intensity could be linked with a faster pH decay, an increasing particle density and lower particle size. These effects were included in the population balance model using a calibration procedure based on Bayesian Monte Carlo techniques. From the calibration procedure, new kinetic laws were proposed for struvite nucleation and growth, where the effect of the hydrodynamics had been decoupled by explicitly including the shear rate as a process variable.