Water distribution network failure analysis under uncertainty

Abstract

Failure analysis of water distribution network gains prime importance as it plays a crucial role in meeting the increasing demand for water in megacities. Water interruption can occur due to instances of component failure such as in pipes and valves. Water distribution network is a complex system comprising a variety of components and their interrelations. To avoid sub-optimization, a comprehensive framework for failure analysis is needed, in order to include all parts of water distribution network. The primary aim of the study was to propose a framework for failure analysis in which fuzzy set theory was combined with fault tree analysis as a useful model to deal with uncertainties resulting from unavailability of data. The proposed framework was applied to a water distribution network in Tehran to demonstrate its validity. Water supply interruption was considered a top event in the fault tree. As a result, the occurrence probability of top event was 0.29. According to sensitivity results, isolation valve failure was the critical minimal cut set with the value of 0.16, which shows a 55% decrease in top event probability. Thus, it is concluded that in decision-making process, a focus should be maintained on the critical minimal cut set, while suggesting corrective actions.