Water Distribution Network Reliability Estimation Using the First-Order Reliability Method

Abstract

This paper utilizes the First-Order Reliability Method (FORM) for estimating water distribution network (WDN) reliability during critical loading conditions. The factors that are assumed to be uncertain include pipe roughness coefficients and nodal demands. FORM predictions of WDN nodal reliability compare closely with more accurate Monte Carlo Simulation (MCS) estimates. FORM is shown to automatically determine the critical node in the network in a reliable and efficient manner. The critical node reliability approximation to network reliability is evaluated and it is argued that the approximation is not appropriate in some cases. An improved FORM network reliability measure is proposed and shown to more closely approximate network reliability estimates when two failure modes are considered at the two worst nodes in the network. The findings of this paper should be considered in future WDN reliability-based optimization studies.