Abstract

Purified water forms an integral part of pharmaceutical production. The consistency of water quality produced by purrficatron processes and distributed to points-of-use is of utmost importance Tee-sectlons Installed in distribution loops are commonly used to divert fluid flow at take-off points. However fluid flow restriction at teesection branches can cause piping dead-legs. Dead-legs consist of reglons of stagnant fluid where harmful organisms can proliferate unaffected by the scounng effects of distribution loop flow. This thesis presents a thermo-fluld analysis focusmg upon the fluid dynamics and heat transfer mechanisms occurrmg wthin dead-leg branches A literature review of high purity water system designed etails sanitization methods currently employed in industry with reference to the detrimental effects of dead-legs. Experimentation was performed using a single-loop fluid rig complete with capped 90' tee-section representmg a piplng dead-leg. Analysis of the thermal conditions for various dead-leg configurations was performed mcludmg variations of branch length and diameter. The effect of varying loop velocity was also investigated. The application of non-intrusive analysis techniques was considered. Infrared thermography and surface-mounted thermocouples were used to map surface temperature distribution across a dead-leg branch. Increased temperatures were noted at the base of the dead-leg branch for increasing loop velocities. Comparison of reduced and equal diameter dead-legs for varying branch lengths suggested dead-leg temperature is strongly related to mlet loop velocity. Acceptable thermal responses were noted m 4d dead-legs for loop velocity > 0.94m/s, 2d reduced diameter dead-legs at 1 50m/s and in 2d equal diameter deadlegs throughout the examined velocity range. Although all dead-leg configurations used in analysis adhered to industry recommendations; unsatisfactory thermo-fluid conditions recorded for remalnlng dead-legs suggests revision of accepted regulations. Non-intrusive analyses illustrated greater temperatures at branch md-pomnt compared with base measurements. However the application of techniques was deemed limited due to pipe wall conduction effects.