(304a) Experimentally Validated Model Based Optimization of Filter Drying Process | AIChE

(304a) Experimentally Validated Model Based Optimization of Filter Drying Process


Sahni, E. - Presenter, University of Connecticut
Strong, J. C. - Presenter, Abbott Laboratories
Buzon, B. - Presenter, Pfizer Inc.
Chaudhuri, B. - Presenter, University of Connecticut

Drying is traditionally defined as the unit operation which converts liquid, solid or semi-solid feed material into a solid product of significantly lower moisture content. It is one of the most commonly used unit operations in pharmaceutical, catalyst, polymer, paper, food and mineral industries in the preparation of dry granules by thermally removing volatile solvent from the wet solid. The knowledge of granular flow, mixing and heat transfer patterns in a filter dryer is critical to optimize the design and operation of the equipment which is mainly employed in drying of API(s) in pharmaceutical manufacturing. Our study focuses on quantitative investigation of mixing and heat transfer in a filter dryer in the quest to determine the optimum drying conditions. We use experiments and discrete element method (DEM) based numerical model, to improve the understanding of agitated vacuum drying process, and hence develop quantitative approach for optimization. We did not study the filtration operation done by this filter dryer and have only studied the drying operation. Most of the DEM-based heat transfer work has been either two-dimensional or in static granular beds. To the best of our knowledge no previous work has used three-dimensional DEM to study heat transfer in granular materials in agitated filter dryer. Once the dryer vessel is charged with the slurry (1L), nitrogen is purged through the drying chamber. Low pressures are generally used to keep the granular cake from becoming so compressed that the crystals fuse together. The impeller of the dryer suspended at the center is rotated by a set of computer-controlled motor drives. The impeller can rotate in both directions to better dig into the cake to help break it up, remove it from the filter media and finally to dry. In our experimental study, glass beads (2.4 -2.9 mm) or lactose is used with appropriate amount of ethanol as the solvent. Before the main experimental run, the impeller is rotated so as to ensure uniform wetting of the granular system. The drying chamber is purged with pure nitrogen flow at 0.1 L/min. The granular bed is then dried by applying heat to the inner wall of the vessel. Silicone oil/propylene glycol is used to heat the jacket sorrounding the external wall of the vessel. During each batch run, at repeated intervals, the impeller is stopped and a thermocouple is placed through one of the ports in the dryer to measure the temperature of the bed as a function of time. At the same time, the samples of wet glass beads are withdrawn using a cup sampler and stored in tared vials. The vials are then dried under house vacuum (26½ in.) at 323 K overnight and analyzed for estimating the loss in weight in the drying experiments. The change is weight data leads to the measurement of loss of drying and the drying rate as a function of material properties and operational conditions of the filter dryer. Series of experiments are performed to check the effect of various system variables: impeller speed, fill ratio of the dryer and the wall temperature of the dryer on the drying performance of the two granular systems (glassbeads or lactose with ethanol). Discrete Element Method (DEM) is employed to simulate granular flow, mixing and heat transport in the filter dryer. DEM explicitly considers inter-particle and particle-boundary interactions, providing an effective tool to solve the transient heat transfer equations. Typical system with glass beads and lactose powder are numerically simulated using appropriate material properties and validated by the experimental findings. A parametric study for simulations is also being performed to check the effect of wall temperature (318 K, 338 K, 353 K), fill ratio (25%, 50%, 75%), impeller speed (5rpm, 12rpm, 25rpm) and particle size on drying performance. The solvent concentration and temperature profile from simulation are in consensus with our experimental results. In both experiments and simulations, higher wall temperature showed an increase in the drying rate and much higher rise in average bed temperature, for both glass beads and lactose systems, thereby decreasing the total time for drying operation. Also at given wall temperature and air velocity, the increase in fill volume (bed depth) resulted in a decline in the drying rate. Furthermore, as the rotational speed of the impeller decreases, the heat transfer in the granular bed increases. The heat transfer coefficient are also calculated for different parameters under study. Hence the optimal drying condition is determined be operating at an impeller speed of 5 rpm, wall temperature of 353 K and with 50% fill load.