Water Distribution System Reliability Under a Fire Flow Condition: A Probabilistic Approach

In the technical literature concerned with the design of water distribution systems (WDS), the term `fire flow' is often associated with three distinct challenges: to estimate the actual fire flow needed to fight a fire; to estimate the availability of fire flow in a specific WDS; or, to evaluate the potential impacts of fire flow on WDS reliability. The first two challenges are generally addressed from a deterministic perspective, and, although some probabilistic approaches exist for solving the latter one, they seldom address fire flows explicitly. In general, fire flows are considered as part of a unique nodal flow probability distribution for reliability analysis purposes. Yet one significant question remains consistently unanswered: what is the WDS reliability given that a fire situation occurs? The objective of this study is to present a methodology which can be used to evaluate the WDS reliability under a fire condition. To this end, two different probability distributions are assessed for the nodal flows: one representing the base demands and the other representing the fire flow demands. A Monte Carlo method is used to generate the flows, and steady-state hydraulic, simulations are performed to derive both nodal and system reliability indexes for a WDS under a fire condition.