Stochastic Modeling of Fluidized Bed Granulation: Influence of Droplet Pre‐Drying

Abstract

A stochastic Monte Carlo method was used to solve the population dynamics of fluidized-bed granulation. Special emphasis was put on the addition of liquid to the process and on the microprocess of droplet pre-drying. Pre-drying can occur during the flight of droplets from the tip of the nozzle to the fluidized bed, especially in top-spray configuration. A single-droplet drying model that considers the change of droplet diameter and of the radial concentration profile of binder in the droplet was developed and validated by experiments in an acoustic levitator. Combined with spray flow field considerations, this model can be used to retrieve an initial droplet diameter from spray measurements. Integrated in the Monte Carlo simulation, it shows that pre-drying can significantly decrease the agglomeration rate due to overspray and premature solidification instead of successful droplet deposition.

A new framework for population balance modeling of spray fluidized bed agglomeration

“…In the present work, however, the MC method is used as a scaled-down virtual spray fluidized bed granulator rather than as a solver of the PBE. A similar approach has recently been presented by Terrazas-Velarde, Tsotsas (2009), Terrazas-Velarde et al (2011a), Terrazas-Velarde, Peglow, and Tsotsas (2011b) and Dernedde, Peglow, and Tsotsas (2011). In our work, micro-processes such as the collision of particles, deposition of droplets, drying of droplets, and the dissipation of kinetic energy according to Stokes criterion are modeled as in Terrazas-Velarde et al (2011b).…”

“…However, in the CNMC method, as explained by Dernedde et al (2011Dernedde et al ( , 2012, only one among all possible events (collisions and droplet depositions) is simulated in a single time step. The number of particles is regulated in each time step.…”

“…Micro-processes such as the collision of particles, addition of binder droplets, and convective drying of droplets are incorporated in the algorithm. The structure of the algorithm is based on the works of Dernedde et al (2011) and Hussain et al (2013), whereas the models for micro-processes are taken from the work of Terrazas-Velarde et al The time step of MC simulation in the present work is based on the works of Dernedde et al (2012) and Hussain et al (2013). It is calculated by taking the reciprocal of the total number of events (i.e., the sum of the total number of collisions and total number of droplet depositions).…”

“…during the simulation. In the recent works of Terrazas-Velarde et al (2009, 2011a, 2011b, Dernedde et al (2011), andTsotsas (2012), two different MC methods, namely the constant volume and the constant number method, have been utilized, respectively. Therefore, it is important to briefly discuss the appropriateness of both methods prior to choosing one of them for simulating the SFBG process.…”

Restoration of particle size distributions from fiber-optical in-line measurements in fluidized bed processes

“…Accordingly the introduction of in-line detection methods of the PSD to fluidized bed granulation as a pronounced transient process has great potential in regard to monitoring tasks, efficiency aspects and quality control. Being governed by interdependent underlying micro-mechanisms, fluidized bed processes combine phenomena such as binder capture, droplet imbibition and solidification, particle collisions and breakage with particle formation (Dernedde et al, 2010;Tan et al, 2006;Thielmann et al, 2008). In most micro-mechanisms the particle size marks a functional quantity, which demands a close tracking either as a measure of product quality or as model input data for process controlling.…”