Smart Water Meters in Full-scale Drinking Water Distribution Networks (DWDNs)

Abstract

For drinking water companies in Netherlands, the current conventional treatment technique may not be fully capable to cope with potential future problems resulted from climate change, salinization and social developments, such as population growth. Innovative technologies like reverse osmosis and more strict drinking water standards are expected to be applied to provide better water quality to customers. In Oasen drinking water company, reverse osmosis has been applied in Lekkerkerk drinking water treatment station since 2016 to prepare for the future challenges, while potential transition effects like biofilms detachment and resuspension of loose deposits in the drinking water distribution systems might occur under irregular changes caused by supply water switch. A setup named Smart Water Meter was developed and used with integrated functions, such as seizing particles inside distribution networks and monitoring several parameters, to investigate this phenomenon. Although slight and temporary transition effects were captured with the help of this device, the amount of retained materials were not sufficient for transition effects study. According to the company, new treatment stations are being planned where background information is required to be collected and still potential transition effects might occur. The main objective of this research was to optimize the Smart Water Meter for future transition effects study by re-selecting the pore size of the filter bag to make it more sensitive and install the upgraded setups in four different locations under different water quality to have a better understanding on its performance and characteristics. Before being applied into the fields, a stagnation (water retained inside the setup for a long time) test was conducted as well to figure out the feasibility of experiments simplification where water is kept flowing through the setup instead of intermittent mode and possible deterioration of the water quality. 10 microns filter bag is selected in this research after three stages tests. Results reveal that the optimized filter bag with new pore size is able to intercept more particles without arousing sudden pressure drop under normal conditions. However, if this setup could set an early alarm, like a sudden pressure drop, when serious transition effects occur remains to be checked in the future. According to the stagnation test, chemical elements concentration and biological activity increased inside the filter bag, meaning that the water quality was indeed deteriorated when it was stagnated inside the filter bag for 23 hours while the quality could go back to normal after opening both influent and effluent valves on the setup for a while. This also indicates that the simplification of experiments is acceptable. Pore size of the filter bag may require to be reconsidered when the influent pressure in households is not sufficient, otherwise great pressure loss is inevitable. For different water quality in four locations provided by three drinking water companies, the performance of the setup is relatively reliable with a stable 100% removal rate of the particles larger than 29 microns inside the water while for the five elements, little relationship could be found. Besides, the filter bag could remove part of the living cells inside it when water keeps flowing through it. Observation of the filter bag inner surface could provide a general idea of the filtrates composition and by making chemical and biological analysis of them, possible origins of the distribution network harbored materials are conjectured. In this research, most of the distribution network harbored materials might come from detached biofilms as calcium is crucial for biofilm formation while part of the retained materials may also originate from loose deposits due to As and Aeromonas existence in some locations. With the optimized setups and a deeper understanding of its performance under different water quality, the setup could be better applied in the future study of transition effects and early warning systems.