Reliability assessment of water distribution systems by the pressure-dependent network analysis (PDNA) : case study (Nonsan network in Korea)

Reliability is an integral of all decisions regarding water distribution system layout, design,operation and maintenance. Providing reliability for water distribution systems is complicateddue to the many factors that affect reliability, the inherent nonlinear behaviour of the system andits consumers, and due to the different conflicting objectives facing a water distribution systemutility. Although the reliability of water distribution systems has received considerable attentionover the last two decades, there is still no common, acceptable, reliability measure or reliabilityassessment methodology.Of the several reliability methodologies, Pressure-Dependent Network Analysis (PDNA) used inthis study is based on the Demand-Driven Available-Demand-Fraction (DD-ADF) proposed byOzger. For assessing reliability of model network, the PDNA computer programme has firstbeen developed using the EPANET Toolkit functions and Visual Studio 6.0. C++ programminglanguage. PDNA can use as a descriptive tool for predicting hydraulic performance to findsupply shortfalls in an iterative manner. Next, the reliability of three model water distributionnetworks using the PDNA was simulated and also analysed under various what-if scenarios toassess their impact on the reliability.Especially, this study assessed the reliability and availability of the Nonsan water distributionnetwork in Korea, which was a first local water supply system to be taken over by Korea WaterResources Corporation (KOWACO) that is responsible for the multi-regional water supplysystems. For an effective model analysis, pipe failure rate was examined based upon thehistorical data and various field data through the literature review. Moreover, the reliability in thefuture was analysed by using the life distribution models: Homogenous Poisson process (HPP)and Non- Homogenous Poisson process (NHPP).Finally, the water quality such like chlorine residual and water age was analysed under bothregular and irregular conditions with the model network using the EPANET hydraulic solver. Theanalysis of an abnormal network condition was based on the critical pipe failure found by PDNAprogramme. Those results were illustrated that water quality was large influenced by networkcomponent failure