THE SELF-CLEANING VELOCITY IN PRACTICE

The Dutch drinking water companies are constructing velocity based self-cleaning residential drinking water distribution systems (DWDS). Field studies with particle counters have shown that these DWDS indeed do not foul. Laboratory studies have shown the settlement and re-suspension of particles in water mains under constant flow conditions. However, the relation between mains fouling and hydraulic conditions under realistic (variable) flows has not been determined. In the presented study, the effect of variable flow velocities on particles in a real residential DWDS was studied through a detailed analysis of turbidity measurements during flushing in combination with a detailed EPANET network model. Firstly, each pipe stretch was flushed with 1.5 m/s for three turnovers and most of the pipes appeared to be clean after the first turnover. This means that it was possible to link the measured turbidity to the location in the pipe from where it was re-suspended. Secondly, an all-pipes EPANET network model was filled with realistic demand patterns from the end-use model SIMDEUM; a small hydraulic time step (0.01 h) was used. This allowed for determining the maximum daily velocity occurring in each pipe stretch. The combination of the first and second step led to a graph of resuspended turbidity against maximum daily velocity. The study showed that in this residential DWDS there is a threshold value for the maximum velocity of 0.2 to 0.25 m/s above which the pipes remain clean. Thus, the existence of the self-cleaning velocity was demonstrated. This study helps in understanding particle behavior in DWDS.