Effects of signal processing on the relative standard deviation in powder feeding characterization for continuous manufacturing

Successful continuous powder feeding is essential for continuous manufacturing and its application in pharmaceutical production. This work focuses on sampling and processing of catch scale (mass) data for the purpose of characterization of continuous powder feeding processes. Specifically, it addresses the sampling frequency and the selection of appropriate filtering (cutoff frequencies) and their effect on statistical measures of catch scale data based on simple test signals representative of continuous feeding. The paper provides suggestions for defining comparability between differently acquired and processed datasets. Specifically, relevant statistical measures which should be reported are mentioned and a reasoning for the appropriate choice of filtering algorithms is provided.