Formation of discoloured water and turbidity in an unfiltered water distribution system

Despite current management practices discoloured water events in both filtered and unfiltered water distribution systems continue to be a major issue for the water industry worldwide. Limited research has taken place into the formation and prediction of discoloured water events in unfiltered water distribution systems. The most widely adopted measuring techniques currently used by the majority of water companies to identify and measure discoloured water events, namely, discoloured water customer complaints and water quality grab samples, severely underestimate the extent to which discoloured water events occur. Customer complaints and grab samples also give little information as to the cause, time or magnitude of a discoloured water event. It is recommended that water companies not use discoloured water customer complaints or water quality grab samples for the prediction of areas at risk of forming discoloured water in unfiltered water distribution systems with lined or PVC pipes. This study demonstrates that monitoring sites for continuous on-line testing of turbidity and flow rate in series within the water supply network can be used effectively to investigate the operating environment of the system and understand the steady state hydraulics that cause discoloured water events in an unfiltered system. This new understanding can be used in preventative management of discoloured water events. This study investigated discoloured water formation in a 450 mm transfer main in Wantirna water quality zone of South East Water in Melbourne, Australia, as an example of an unfiltered water distribution system with few unlined ferrous pipes. The study found that the majority of discoloured water events could be explained by the erosion of cohesive layers of particles formed on the internal pipe wall when a critical shear stress was exceeded in the pipe. For the system investigated, the critical shear stress was variable and could be estimated from the maximum applied shear stress reached in a particular section of pipe in the preceding 24 hours. It is therefore recommended that operational procedures, particularly isolated flushing procedures, be modified to reduce the occurrence of unplanned incidences of discoloured water when the critical shear stress of a pipe is exceeded. A hydraulic model and the method for estimating the critical shear stress outlined in this thesis could be used to test scenarios of operational procedures to identify pipes with increased likelihood of producing discoloured water. The cohesive layers which contribute to discoloured water events appear to form in less than 1 week. It is therefore also recommended that the mains cleaning techniques using flushing currently employed to remove material causing discoloured water events appear to have no medium or long term benefits for the system investigated and further investigation into the effectiveness of routine flushing programs for the prevention of discoloured water events is required. When the causes of discoloured water are understood, a system can be proactively managed to prevent discoloured water events by managing those causes. This philosophy fulfils contempory water system management approaches such as the Framework for the Management of Drinking Water Quality in the 2004 Australian Drinking Water Guidelines.