Field test investigations into distributed fault modeling in water distribution systems using transient testing

Abstract

Current condition assessment techniques are focused on leakage, however this forms only part of the story with regard to the efficiency and condition of pipelines. The neglected phenomena are distributed faults including pipe wall deterioration and blockages, where blockage refers to any build up ranging from increased pipe roughness to complete obstruction that may be caused by debris, sedimentation, tuberculation, biofilms or valves. The objective of this research was to develop a non‐invasive condition assessment technique for water distribution systems (WDS) that has the ability to detect distributed faults including blockages. Inverse Transient Analysis (ITA) has been identified as a potential method but requires developments in modeling techniques for distributed faults. Blockages can be caused by the build up of many different materials each with their own properties. Extended blockages also form complex flow routes, which cannot accurately be captured by the current approach unless extremely fine discretisation is incorporated, increasing computational effort very significantly. The current approach models extended blockages as reduced diameter sections of pipeline. While a blockage does reduce the diameter, the material properties of the blockage significantly differ from that of the pipeline. A viscoelastic element may be used to account for the change in pipe material and hence response due to the blockage material. Field tests have been conducted on the Adelaide metropolitan water distribution system in streets that have been identified as having significant potential to form blockages. The transient tests indicated that the complex response of the pipe to transient excitation suggests that they are indeed suffering from substantial problems with blockages. The location and identification of zones of differing condition to the remainder of the pipeline, and hence having the potential for distributed faults, have been determined using inverse transient analysis.