|Keywords:||Sloan School of Management.; Chemical Engineering.; Leaders for Global Operations Program.|
|Full text PDF:||http://hdl.handle.net/1721.1/90800|
Some biopharmaceutical companies have responded to evolution of the competitive landscape by placing additional emphasis on reducing their costs of manufacturing as a means of maintaining competitiveness. The prototypical current generation biopharmaceutical drug substance manufacturing facility requires a large upfront capital investment. Improving efficiency of use of existing facilities, such as by improving production throughput through the adoption of technology, represents one way in which a company may reduce its costs of manufacturing and/or avoid or delay investments in additional capacity needed to meet future demand. Reducing the variability in the performance of a liquid filling operation taking place during the protein production step is desirable, because it: (1) enables process optimization, including potential throughput expansion, (2) demonstrates control over the process, and (3) improves step yield reproducibility. The technical and economic bases for the implementation of an alternative process control strategy intended to reduce this variability are presented. This strategy involves controlling the fill operation using an inline concentration measurement of the parameter of interest. An engineering-probabilistic approach, consisting of a transient concentration profile model built into a Monte Carlo framework, is applied to predict the variability of the performance of a concentration-based control strategy for filling an agitated, gassed bioreactor. An optimization methodology for selecting an appropriate post-fill target concentration and for quantifying the economic benefit of reducing variability is proposed.