Buranawatanachoke, B.; (2008) Bioprocess analysis using computational fluid dynamics and ultra scale-down. Doctoral thesis , University of London.

Text U592527.pdf Download (13MB)

Abstract

The primary aim of this study is to develop the means to predict mechanisms and rates of material damage and especially the size of particles making up cell suspensions. Computational fluid dynamics (CFD) is employed to define experimental conditions used in ultra scale-down (USD) techniques and relate these to large-scale operations. The validity of this novel scaling methodology as applied to a large-scale membrane filtration unit and disc-stack centrifuge has been investigated. The current state of the art predicts performance of bioprocess operation but not of material damage itself. For the membrane operation, emphasis has been given mainly to flux rate and fouling, and clarification and dewatering are key parameters during centrifugation. Conversely, the effects of hydrodynamic parameters during the operation on the process material, e.g. shear damage, have not been paid much attention. Though CFD has been currently employed to facilitate predictions of the performance of these operations by a number of research groups, the technique alone cannot provide accurate predictions for a given biomaterial. CFD data thus needs to be integrated with experimental data given by other techniques, e.g. scale-down mimics of process equipment, USD device and etc. This study uses CFD and USD techniques to predict this damage. Following preliminary analysis of fluid stresses, the pump has been identified as the key component of the membrane rig responsible for material damage. CFD was then employed to determine the magnitude of maximum energy dissipation rate (EDR) and provide equivalent rotational speeds of the rotating disc device, or USD device, capable of generating the same engineering conditions. The USD experiment was then carried out following recommended speeds given by CFD and experimental data together with proposed mathematical models was used to predict particle characteristics in terms of size reduction. A verification process was finally performed using full-scale experiments. It was found that CFD could provide essentially similar operating conditions for USD experiments mimicking a large-scale operation, especially for operations of a few hours. An initial investigation has also been carried out on the applicability of the CFD-USD technology to predict material damage in the discharge part of the solid-ejecting disc stack centrifuge. The analysis shows that CFD suffered from difficulties in defining details of the problem domain in order to assess flow fields. Moreover, the exposure to gas phase during the discharge and subsequent droplet formation and an unpredictable shape of slurry made it impossible to investigate this problem further using CFD. As a result, it is concluded that this problem should be addressed by other means. Based on findings of this investigation, prediction guidelines are proposed to facilitate the design process and process development by using a small USD device to assess the engineering environment in a large unit operation. More importantly, the concept may theoretically lay the foundation for scaling any unit operation based on key engineering parameters of the system.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as XMLCSVJSON

Archive Staff Only

View Item