MechChem Africa January 2017

In the first ‘Materials engineering in practice’ column for 2017, Tony Paterson from the School of Chemical and Metallurgical Engineering at Wits, talks about the work being done to ensure that the pharmaceuticals and food-grade products coming out of our process plants are safe to ingest. Materials engineering in practice: Hygiene and the process plant challenge

W ith increasing urbanisation, the 20 th century introduced process plants to mass-pro- duce pharmaceuticals, food and beverages, focusing on optimising cost, time to market and reliability. The21 st centuryprocessplantisfacedwith more complex needs. The triple bottom line requires a solution involving people, prosper- ity and the planet, with its finite resources, particularly water. Process plants concerned with products forconsumptionarerequiredtomeetincreas- ingly robust legislation demanding reduced (or zero) bacterial/spore counts in the final product. The challenge is how to manage the fabrication of new plants and maintain the old plants to accommodate these changing operating conditions. Hygiene is often seen as both as a reputa- tion and market risk, and as a costly alterna- tivetocurrentpractices.Realistically,hygienic welded fabrication will be more challenging and more costly. Whilst this puts capital budgets under pressure, it should ease operational costs, en- hance productivity and reduce clean-in-place (CIP) andwater needs. The strategic choice is business value and risk-driven. Process plant construction Process plants include a variety of, typically stainless steel, components including factory- made tanks, heat exchangers, columns and pumps. These are interconnected by pipes. During construction, fabricated components are transported to site for assembly. These are linkedonsiteusing thin-walled (wt<3%D), small diameter piping. Connection is usually by welding.

Most connecting pipes are measured on site, ends prepared, aligned and welded. Internal weld imperfections in small-bore pipework are generally inaccessible. Site welding requires skill, is more difficult to control, and more difficult to manage for a variety of reasons. Biofilm formation and control Complete sterility of plant, input materials and water, whilst not a realistic expectation, is the ideal. In the absence of complete steril- ity, other methods need to be considered for plant constructionor refurbishment practices toeliminateormitigate against bacterial load. Private sources of water have declined. Municipality suppliedwater is of inconsistent quality over time and product inputmaterials are often shipped fromsources far away from the process plants. Biofilms form on exposed surfaces of pro- cess plant as thin layers of microorganisms adhering to surfaces. These may be organic or inorganic, together with the polymers that they secrete and biofilms can include harmful bacteria. Biofilm depth increases with increased surface roughness, increased temperature and lower flow speeds and is, therefore, promoted by occluded and dead areas. One source of surface roughness and local occlu- sions are welded joints. Resulting frombiofilmformation, bacteria can grow and be released into the product. Welded joints support biofilm growth where there is: • Inadequate penetration – the inner weld profile leaves crevices (dead areas). • Over penetration or cauliflowering – the inner weld profile is proud (dead areas). • Porosity.

• Cracks. • Misalignment during manufacturing or fabrication – occluded areas. • Surface roughness due towelding process effects across the width of the heat af- fected zone (Laser < CMT < TIG/MIG). CurrentlyCIPmethods areused to reduce the impact of biofilms, but CIP is not completely effective as: • The loss of heat and chemical concentra- tion over distance reduces effectiveness. • The process does not clean hidden and occluded areas. • CIP requires the extensive use of water. Manufactured pipes Manufactured pipes are oval and bow to a greater or lesser extent. Whilst accepted tol- erances exist, the impact on welded joints is significant particularlywith thinwalled pipes. Pre-programmedorbitalTIGweldingisthe preferredmethod of joining pipes. It assumes well-matched faying surfaces because: • The pre-programmed controlling current and travel speed has to allow for the im- pacts on themoltenweldmetal of internal inert gas pressure, gravity and theoverlap- ping wall thickness. • If the faying surface overlap varies, the current will be either too high or too low leading to incomplete fusion or to over- melting leading to poor weld geometry. Research Initial research used E-coli build-up to char- acterise the effects of weld processes and joint geometry. More recent research based on real plant pipe analysis and a pair of math- ematical algorithms using an 80% minimum criterion as suited to orbital welding showed poor results if randomly aligned, but better if aligned through themajor axis. This indicated the likelihood of bacterial build-up at joints. Whilst the results were far better with tightertolerances,movingtoahigherstrength material witha thinnerwall thickness showed the desirability of extremely low tolerances. This places huge challenges on manufactur- ing costs. Current research is concentrating on site- based methods of achieving closely matched faying end surfaces using plastic forming ap- proaches. The intent is to check effectiveness using E-coli build up as the rating criteria. q

Pictures courtesy of JD Cluett: www.arcmachines.com

Left: The ideal orbital weld for minimising Biofilm formation. Right: A manual weld taken from an operating pharmaceutical plant. This weld is unacceptable by any sanitary standard.

34 ¦ MechChem Africa • January 2017