Author(s): Qiang Xu

Linked Author(s):

Keywords: Water distribution systems; Leakage detection; Pipe failure models; GP

Abstract: The leakage from the water distribution system is a serious problem of many cities, which incurs large economic and social cost. However, it is difficult and expensive to exactly locate the leakage, imposing pressure on utility managers to develop annual replacement plans for critical pipes. Thus, a reliable and robust pipe failure model is demanded to assess a pipe's propensity to fail. In particular, the means to assign criticality to an individual pipe segment has been concerned about. In this paper, pipe deterioration is modeled using genetic programming (GP). Symbolic formulae were derived by the data-driven technique. The Beijing City was selected as the case study area and the pipe failure data of 19 years (1987-2005) were analyzed. The relationship between the pipe failure possibility and the pipe properties such as diameter, age, and length was established. The relationship between pipe breaks and hydraulics conditions was also discussed. The model was verified using the re-sampling technique of bootstrapping. Finally, the developed model was applied to assess the criticality of all the pipe segments of the entire water distribution system of Beijing City, based on Arc GIS. The results can be used to support the optimization of the layout of leakage detection devices and the plan-making of the critical pipes' replacement.

DOI:

Year: 2009