267239 Predictive Modeling of Hot Melt Extruders

Wednesday, October 31, 2012: 8:51 AM
Allegheny III (Westin )
Andreas Eitzlmayr1, Daniel Treffer1, Gudrun Hoerl2, Sarah Windhab2, Gerold Koscher2 and Johannes G. Khinast1, (1)Institute for Process and Particle Engineering, Graz University of Technology, Graz, Austria, (2)Research Center Pharmaceutical Engineering GmbH, Graz, Austria


During recent years, extrusion processes increasingly attracted attention in the area of pharmaceutical manufacturing. As a continuous manufacturing operation, extrusion provides the potential of increased efficiency and reduced operation costs together with a steady product quality. Depending on the used materials and conditions, different types of pharmaceutical extrusion processes can be distinguished, as hot melt extrusion, wet extrusion and solid lipid extrusion. In contrast to wet extrusion and solid lipid extrusion, hot melt extrusion is characterized by a temperature level above the melting point of the material. Since the production of solid dispersions and solid solutions can be achieved, hot melt extrusion is a potential pathway to increase the bioavailability of poorly soluble drugs.

The most common extrusion devices are co-rotating twin screws. Compared to different types as counter-rotating twin screws and single screws the co-rotating twin screws are often preferred due to their mixing performance and self cleaning screw profiles.

Although melt extrusion processes are known from plastics industry for many years and spatially resolved CFD (Computational Fluid Dynamics) simulations of extruders screws are already achieved today [1], there is still a lack of predictive tools supporting design, scale-up and optimization of extrusion processes and a high amount of empirical methods and experimental effort is required. Finite element and finite volume CFD methods can be used to solve the hydrodynamics of completely filled screws. Partially filled screw sections cannot be addressed by conventional methods, since their facilities to recover free surface flows are very limited. The interactions of the involved physical phenomena are highly complex and first-principle design methods are not available for extrusion processes.


In our work, we are following two different approaches. The first is a one-dimensional (1D) screw model, which describes physical process variables as pressure, material temperature and screw filling ratio along the screw, and follows the idea of a continuous stirred tank reactor cascade with backmixing, as shown by Choulak et al. [2]. This model targets a global, approximate description of the material along the twin-screw. Since a 1D model cannot resolve the three-dimensional (3D) geometry of the screws, it depends on empirical screw parameters which close the gap between 3D geometry and 1D flow description. These parameters can be determined either by experiments or by 3D simulations. The latter is a target of our second approach, which is a spatially resolved screw model based on a relatively new CFD method providing increased potential for problems with complex, moving boundaries as extruder screws.


Results of the 1D model are profiles of filling ratio, pressure, material temperature and power input along the screw. A residence time distribution gives a measure for back-mixing in screw direction. The calculated profiles are tightly related to the geometrical configuration of the screw. Experimental work to validate the results is ongoing.


[1]     K. Kohlgrüber, Der gleichläufige Doppelschneckenextruder, Carl Hanser Verlag, München, 1st edn., 2007, ch. 7, pp. 140 – 142.

[2]     S. Choulak, F. Couenne, Y. Le Gorrec, C. Jallut, P. Cassagnau, A. Michel, Generic Dynamic Model for Simulation and Control of Reactive Extrusion, Ind. Eng. Chem. Res., 2004, 43, 7373-7382.

Extended Abstract: File Not Uploaded