A Pipe Ranking Method for Water Distribution Network Resilience Assessment Based on Graph-Theory Metrics Aggregated Through Bayesian Belief Networks

The resilience assessment is crucial for many infrastructures, including water supply and distribution networks. In particular, the identification of the 'critical' components (nodes or pipes) whose failure may negatively affect network performances and system resilience is a key issue, with a direct relevance for decision-makers involved in planning, management and improvement activities. Among the multiple methods and tools available, the use of graph-theory metrics is a cutting-edge research topic, as the analysis of topological properties may provide simple yet reliable information on the performance of complex networks. In the present work, we aim to overcome the limit associated to the use of individual graph-theory metrics, identifying a subset of relevant metrics that are directly connected to network resilience properties, using them to perform a 'network degradation analysis' in case of single pipe failure and finally proposing an aggregation of the results using a Bayesian Belief Network. Ultimately, the proposed methodology provides a ranking of the most critical pipes, i.e. those that contribute most to system resilience. A real water distribution network in Italy is used for model development and validation.