Pharmaceutical water systems and the 6D rule: a computational fluid dynamics analysis
Corcoran, Brian G. (2003) Pharmaceutical water systems and the 6D rule: a computational fluid dynamics analysis. PhD thesis, Dublin City University.
Full text available as:
The problem o f piping system dead-legs are frequently encountered in high purity water systems throughout the pharmaceutical and semi-conductor industries. The installation of a pipe tee in sterile process pipework often creates a stagnant dead-leg zone which can result in the formation of bio-film and compromise the entire system. Considerable basic research is required to address the lack of understanding of this problem and to assist during design, manufacture, installation and operation of these critical systems. This study involves the application o f CFD (computational fluid dynamics) techniques to the study of turbulent flow in Pharmaceutical pipe teejunctions.
Numerical models have been developed to initially study divided turbulent flow in a range of standard Pharmaceutical tee-j unctions and then to study dead-leg flow. Numerical predictions were compared with previously presented experimental results based on Laser Doppler Velocimetry. Turbulent models such as the k - e and Reynolds Stress model (RSM) were used to analyse the flow. Dye injection studies highlighted the lack of penetration o f the dead-leg and complex branch flow patterns for both sharp and round entry tees. Hydrogen bubble techniques gave clear evidence of the presence of a slow rotating vortex at entry to each branch and the presence of stagnation zones throughout the dead-legs.
The effect o f mainstream velocity and loop to branch ratios on dead-leg flow patterns was analysed. Stagnation zones were identified within each branch and the presence of a slow rotating cell within the dead-leg resulted in a lack of exchange of mainstream fluid from the distribution to the branch. The 6D-rule was found to be industrially irrelevant. 1 to 2D configurations should be used to avoid stagnation. No configuration resulted in high wall shear stress within the branch and a reduction in branch to loop diameter increased branch stagnation. There was no evidence of exchange of fluid between the loop and branch and all configurations had some quiescent (dormant/inactive) water.
Archive Staff Only: edit this record