In recent years, continuous manufacturing technologies have been widely accepted in the pharmaceutical environment given the many advantages they offer to product quality. Some of the benefits continuous processes offer over batch counterparts include: enhanced process control, reduced solids handling, small equipment footprint, reduced scale up requirements, and more uniform processing . These advantages make continuous manufacturing a desirable technology for many pharmaceutical manufacturers. Nonetheless, there are many challenges to this technology: higher initial cost, extended initial development time, lack of PAT tools available for monitoring continuous processes, and difficulty of implementation for low volume and dosage products . This last challenge is primarily based on the difficulty to accurately feed small amounts of materials to the process and blend them efficiently. Research has focused on understanding the behavior of both feeders and blenders with the aim of improving upon these challenges [2, 3]. During the early phases of continuous manufacturing blending was studied more extensively given the legacy importance in batch processes (i.e., in batch it is one of most important unit operations). However, feeding has the most significant effect in continuous processing as its sets the mass balance for the entire process. Lack of feeder control and understanding, leads to lack of control of any downstream unit operation .
Feeding accuracy is of great importance in continuous processes given the stringent requirements for product composition. To this end, Loss-in-Weight (LIW) feeders under gravimetric control are commonly used. LIW feeders deliver powders by calculating the amount of material lost (i.e., fed out of the unit) and controlling the rate of material delivery accordingly . This strategy provides the feeding accuracy required in pharmaceutical processes. LIW feeders generally consist of a hopper, a flow aid and a powder conveying system, all over a load cell that measures its weight. The hopper contains the material in the feeders and is the receptacle where refills are performed. The flow aid is located at the bottom of the hopper with the purpose of breaking any powder bridges and disrupt the statics to facilitate transport of material into the conveying system. The conveying system, often a screw due to its accuracy over other systems, is used to discharge the material out of the feeder. Previous research focused on studying LIW feeders to understand how these components affect the unit’s performance (i.e., the output flow rate). The conveying system was found to be the most significant component on feeder performance, followed by the refill hopper refill strategy . Material properties were also found to have a significant impact on feeding performance .
Feeders are often not considered in the absence of other processing equipment, therefore development of models for these units is not as extensive compared to other unit operations. Multiple articles focus on the use of Discrete Element Model (DEM) simulations to characterize the effect screw and hopper configurations have on the unit’s performance [7, 8]. The results from these studies are often semi-qualitative in nature and do not allow for development of models that can be used for process modeling (e.g., flowsheet models and control strategy design). In this work we developed a semi-empirical model of feeder behavior using material properties as input parameters. This work is an extension of previously developed empirical models , with the goal of incorporating the behavior and effects of both the flow aid and conveying screws in the model.
The semi-empirical model was developed to simulate feeder behavior, primarily focusing on modeling screw speed and flow rate as a function of time. Frequency variations (i.e., noise) of these variables were characterized using Fourier series, whose coefficients were regressed from experimental data for individual components. Screw speed and flow rate were directly correlated in the model through the use of a feed factor, whose units are mass per screw revolution. Feed factor, defined as the maximum mass fitting in a screw flight, was modeled to account for changes in hopper material properties occurring due to material changes, flow aid systems, feeder emptying and refills. Material properties were characterized before and after the feeder in order to evaluate and model their relationship to screw speed and flow rate. Lowest feeding capacity (LFC) of the feeder was evaluated based on these findings and the developed model. The effect of flow aid systems on the mixing of powders inside the hopper was experimentally characterized using residence time distribution (RTD) experiments. Models to predict mixing behavior in the unit were developed using fundamental chemical engineering models (e.g., continuous stirred tank reactors (CSTR), plug flow reactors or any combination). The study was performed for multiple feeder types with varying the flow aid systems, hopper configurations, and conveying system locations. Operating conditions (e.g., flow rate, refill level) and powder properties were also varied to investigate the effects these parameters had on the mixing. An optimization framework for refill strategy with minimal trial-and-error was developed. The ultimate goal of this work was to develop a model that can be used to predict the behavior of LIW feeders for any material a priori, using the results from powder characterization test. Such model can lead to a significant reduction in development time and minimize product waste.
1. Rogers, A., A. Hashemi, and M. Ierapetritou, Modeling of Particulate Processes for the Continuous Manufacture of Solid-Based Pharmaceutical Dosage Forms. Processes, 2013. 1(2): p. 67-127.
2. Engisch, W.E. and F.J. Muzzio, Method for characterization of loss-in-weight feeder equipment. Powder Technology, 2012. 228(0): p. 395-403.
3. Vanarase, A.U. and F.J. Muzzio, Effect of operating conditions and design parameters in a continuous powder mixer. Powder Technology, 2011. 208(1): p. 26-36.
4. Dai, J. and J.R. Grace, Biomass granular screw feeding: An experimental investigation. Biomass and Bioenergy, 2011. 35(2): p. 942-955.
5. Boukouvala, F., et al., An integrated approach for dynamic flowsheet modeling and sensitivity analysis of a continuous tablet manufacturing process. Computers & Chemical Engineering, 2012. 42(0): p. 30-47.
6. Falk, J., et al., Mass flow and variability in screw feeding of biomass powders — Relations to particle and bulk properties. Powder Technology, 2015. 276(0): p. 80-88.
7. Owen, P.J. and P.W. Cleary, Screw conveyor performance: comparison of discrete element modelling with laboratory experiments. Progress in Computational Fluid Dynamics, an International Journal, 2010. 10(5): p. 327-333.
8. Fernandez, J.W., P.W. Cleary, and W. McBride, Effect of screw design on hopper drawdown of spherical particles in a horizontal screw feeder. Chemical Engineering Science, 2011. 66(22): p. 5585-5601.
See more of this Group/Topical: Pharmaceutical Discovery, Development and Manufacturing Forum