Abstract
An experimental drinking water distribution system (DWDS) was used to evaluate the evolution of particle size distribution (PSD) and basic quality parameters of ultrafiltered water with or without pre-ozonation. An ultrafiltration (UF) module was set up, associated with a pre-ozonation system (3.7 g O(3) /m(3) ). The permeate was circulated in the DWDS (300 m; 0.9 m/s) with 0.4 mg/L of chlorine, and the analysis of the PSD was performed using a β-variable mathematical model. A better control of membrane fouling was obtained with pre-ozonation, and PSD was necessary to observe water quality differences between permeates and in the DWDS. A decrease in particle concentration of 1.8 logarithms was obtained with the application of UF membranes, while a decrease of only 1.2 logarithms was obtained with pre-ozonation. The system without pre-ozonation showed a higher efficiency at removing smaller particles (around 2 μm), with the absence of particles larger than 23 μm during both stages. The PSD revealed a worsening of water quality in the DWDS with an increase of particles smaller than 5 μm during the application of UF membranes, while with pre-ozonation, all particle sizes analyzed increased their concentration. PRACTITIONER POINTS: Pre-ozonation led to a better control of membrane fouling, but a worsening of permeate quality according to particle size distribution. Pre-ozonation does not improve the turbidity, dissolved organic carbon or UV254 removal capacity of ultrafiltration during drinking water treatment. Particles size distribution reveals the deterioration of water quality in a drinking water distribution system better than turbidity or DOC. Ozone prior to ultrafiltration membranes led to a worsening of permeate quality, more significant in the drinking water distribution system.