Powder mixing plays an important role in many processing stages within the pharmaceutical, catalysis, food, construction, and mineral industries. However due to stringent processing regulations, in the past, the focus of powder mixing research has been directed towards batch mixing. More recently, continuous processing has emerged as a significant process alternative and recent research efforts indicate that a well-controlled continuous mixing process can significantly enhance productivity (Marikh et. al [1]; Muerza et. al [2]). Our research focuses on investigating the design space of continuous mixing operations in order to optimize blending performance. We are targeting the development of a statistical and semi-empirical process models that allow for on-line control and optimization of powder mixing performance.

The continuous mixers considered are manufactured by GEA Buck Systems. The two mixers are cylindrical horizontal vessels with a series of adjustable blades they vary in size, speed, blade design, and motor power. In both units the powder is discharged through a weir in the form of a conical screen. The mixer has the flexibility of changing angles affecting the gravitational forces on the axial transport thereby affecting the powders residence time. Model blends have been formulated using the following materials: Milled Acetaminophen (30µm), Fast-Flo Lactose (75-250µm), and Lactose 125M. Powder samples are retrieved from the mixer outflow and characterized using Near Infrared Spectroscopy (NIR).

Different operating conditions, design parameters and pharmaceutical formulation are investigated. The results show that processing angle as well as rotation rates, affect mixing performance. Interestingly, faster rotation rates increase the intensity of electrostatics effect and reduce blend homogeneity (Portillo et al. [3]). Convective design has also shown to be an important factor in mixing performance as discussed by Laurent and Bridgwater [4]. Blade location, blade alignment, blade design, blade angle are all important factors within the convective design which affect the radial dispersion coefficient and drive powder flow. Another advantage towards utilizing continuous mixing is that segregation has not shown to play a role during the mixing process although electrostatic effects are clearly present.

After characterizing the mixer through a set of experiments, our focus has been to develop computationally feasible mixing process models that will enable the simulation, optimization, and control of mixing processes. Previously we examined a hybrid approach for convective mixing systems (Portillo et al [5]), which is less computationally expensive than an entirely discrete element approach. Nonetheless in order to further decrease computational time we examine alternative on-line modeling approaches. Berthiaux and coworkers [6] have modeled continuous mixing with Markov chain. However the difficulty of these methods is that without pre-existing particle dynamics the transitional matrices are difficult to obtain. In our work we have utilized the experimental data obtained from various process and operating conditions to develop and validate both statistical and empirical-based models using parameter estimation techniques.

We examine two modeling paradigms, one following an empirical modeling approach dependent on experimental data with a substantially low computational cost. The second approach is based on compartment modeling that has been shown to be very efficient for batch mixing (Portillo et al. [7]). Following compartment modeling our results demonstrate that CPU times are affected by the number of particles per compartment, number of compartments, and mixing rates. As an example a simulation with of 250,000 particles and 150 compartments requires only 45 CPU seconds. We can further reduce the CPU time by modifying the number of particles in the system without sacrificing modeling performance. Both modeling approaches are intended to describe blend homogeneity as a function of axial length within the mixer, thereby preventing or reducing undesirable outcomes by monitoring on-line operating conditions.

References:

[1] Marikh K., Berthiaux H., Mizonov V., Barantseva E., 2005, “Experimental study of the stirring conditions taking place in a pilot plant continuous mixer of particulate solids”, Powder Technology, 157, 138-143.

[2] Muerza S., Berthiaux H., Massol-Chaudeur S., Thomas G., 2002, “A dynamic study of static mixing using on-line image analysis”, Powder Technol., 128, 195-204.

[3] Portillo P.M., Ierapetritou M.G., Muzzio F.J., 2006, “Characterization of Continuous Convective Powder Mixing Processes”, submitted for publication to Powder Technology in Feb 2007.

[4] Laurent B.F.C., Bridgwater J., 2002, “Performance of single and six-bladed powder mixers”, Chemical Engineering Science, 57, 1695-1709.

[5] Portillo P.M., Muzzio F.J., Ierapetritou M.G., 2007, “Hybrid DEM-compartment modeling approach for granular mixing“, AICHE, 53, 1, Pages 119-128.

[6] Berthiaux H., Marikh K., Mizonov V., Ponomarev D., Barantzeva E., 2004, “Modeling Continuous Powder Mixing by Means of the Theory of Markov Chains”, Particulate Science and Technology, 22, 4, 379-389.

[7] Portillo P.M., Muzzio F.J., Ierapetritou M.G., 2006, “Characterizing powder mixing processes utilizing compartment models”, International Journal of Pharmaceutics, 320, 14-22.