Predictive performance of loss-in-weight feeders for continuous powder-based manufacturing
Description
TitlePredictive performance of loss-in-weight feeders for continuous powder-based manufacturing
Date Created2020
Other Date2020-10 (degree)
Extent1 online resource (xiii, 154 pages) : illustrations
DescriptionContinuous manufacturing for pharmaceutical solid dosage forms has reached a significant milestone over the past decade due to efforts from academia, industry, and regulatory agencies. Advancement in the development of equipment design, process analytical technology, control systems, and modeling tools has facilitated the growing interest in implementation of continuous manufacturing methods in major pharmaceutical companies around the world. The US Food and Drug Administration has also provided regulatory support for the implementation of continuous manufacturing using science- and risk-based approaches. As the pharmaceutical industry modernizes its manufacturing practices and implements more efficient and precise approaches, a more comprehensive evaluation of the process, including unit operations, is needed.
Loss-in-weight feeders, as the first unit operations in a continuous manufacturing design, are responsible for dispensing a given weight of material per unit of time to downstream unit operations accurately and constantly. Disturbances on a given feeder may travel downstream through the process and consequently impact critical quality attributes of the final products, such as potency and content uniformity. Therefore, it is essential to understand the potential aspects that could impact the feeder’s performance and how feeders and downstream unit operations would react in response to the different levels of disturbances and perturbations.
In this work, the effect on feeder performance in a direct compaction (DC) continuous manufacturing line was studied. Studies included the performance of feeders operating in normal gravimetric mode and the effects of feeder refills. Feeder tooling and material properties were both considered to determine their effect on feeder performance. Principal component analysis, a multivariate statistical analysis method, was utilized in order to build up a material library with 30 material flow properties. Partial least squares regression is used to correlate process performance to material flow properties. Multiple near infrared spectroscopy methods were applied to monitor the content uniformity exiting blend uniformity or content uniformity in the final product.
The results obtained from these studies were used to determine the design space for a commercially available loss-in-weight K-Tron KT20 feeder and its dependency on the conditioned bulk density of the powder. Additionally, a methodology was developed to correlate feed rate deviation caused by hopper refill to material flow properties and to create a predictive model.
To understand variations in drug concentration in a continuous direct compaction line, experiments were conducted on how perturbations on the mass flow rate from the feeders transfer down the continuous line, and how much dampening the downstream unit operations can provide to the variability in the mass flow rate. Controlled step changes in concentration and hopper refill operations were performed over short intervals with different blender speeds and total throughputs and the drug concentration in the blend and the tablets was characterized to determine residence time distribution and deviation in concentration.
The results from this work can be applied to the design of a continuous line, the development of a manufacturing plant, and the evaluation of process risks associated with the continuous manufacturing of solid dosage forms. Using the predictive methods enabled by the results discussed here, operators will be able to identify potential failure modes of the feeding operations and facilitate risk assessment for regulatory reporting.
NotePh.D.
NoteIncludes bibliographical references
Genretheses, ETD doctoral
LanguageEnglish
CollectionSchool of Graduate Studies Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.
Statistical Profile