(301c) Prediction of Loss-in-Weight Screw Feeder Performance and Quantification of Failure Modes from Attribute Measurements of Pharmaceutical Materials

Many pharmaceutical processes such as continuous blending of active ingredients with excipients rely on highly accurate, robust, and tunable loss-in-weight feeders for control of composition and associated quality attributes. The ability of such feeders to accommodate the wide variety of material properties inherent in formulation components is therefore an important process design constraint. It is desirable to predict throughput accuracy and performance robustness and identify potential feeder failure modes in the absence of long-duration experimental runs, which may consume excessive quantities (10s to 100s of kg) of potentially high value ingredients.

To this end, we have conducted empirical studies to assess feeder throughput accuracy for commercially available loss-in-weight feeder configurations. A variety of components spanning a range of cohesivities, compressabilities and yield stresses have been assessed. Feed failure modes indicative of process robustness have been categorized and statistically quantified to develop predictive relationships with lab measurements of materials properties performed with gram-scale quantities. These include particle size distribution, bulk and tapped density; permeability, morphology and flowability (as assessed by both direct arching tests and by shear cell measurements of flow function and angle of internal friction). Along with equipment geometric descriptors, performance envelopes are defined. Results are evaluated against theoretical considerations and guidance around practical strategies for predictive assessment of feeder performance is outlined.