Issues of operational integrity in membrane drinking water plants

In the last five years over 100 microfiltration (MF) and ultrafiltration (UF) drinking water plants have been installed in Europe with a combined output of some 1,500 ML/day. The impetus behind this growth is concern over drinking water quality, regulatory pressure, the rapidly declining costs of membrane systems, and consumer pressures for non-chemical treatment. Although the principal justification for membrane plants is their capacity to remove human pathogens more effectively than conventional treatment, the ability of large plants to retain an integral barrier is not well understood. Currently available technology can monitor a full-scale membrane plant for integrity against passage of bacteria and parasites but not viruses. In essence the plant operator has to rely on the membrane manufacturer – on an assumption that if the membrane is not compromised at a level of 1 to 2 μm (the current practical level of measurement) it is also not compromised for virus removal. This paper quantifies the loss of integrity that can occur from membrane fiber failure. It explains the mathematical models used to describe bypass flow through compromised fibers and correlates the results with laboratory tests. Both are compared with artificially compromised fibers in a large full-scale operating plant. Under the worst case scenario where a fiber breaks close to the pot (collection end of the filtration module or element) the relative loss of integrity between alternative process designs from a single broken fiber can differ by as much as 2 log reduction values. The analysis demonstrates the need for monitoring methodology that that can track incremental changes in integrity to allow scheduled rather than emergency maintenance. It highlights the need for regulatory authorities to approve membrane systems based on actual operating performance in preference to laboratory data.