Author(s): Alessandro Pagano; Raffaele Giordano; Ivan Portoghese; Michele Vurro; Umberto Fratino
Linked Author(s): Alessandro Pagano, Umberto Fratino
Keywords: Vulnerability assessment; Drinking water infrastructure; Bayesian networks; Decision Support System; Emergency management
Abstract: Drinking water quality and availability is a life safety issue, being essential for economical, social and sanitary reasons. The damage or the malfunction of any element of a water system, as well as the corruption of the resource quality, could therefore determine significant effects on the served population and on all the other dependent resources and activities. The reliability and the robustness of a water distribution system against physical failures and contamination events have become central tenets of its design, operation and management. Particularly, the assessment of the susceptibility of an infrastructure to specific threats, mainly constituted by extreme natural events as well as by human malevolent or accidental acts, namely its “vulnerability”, is nowadays fundamental for the management of emergencies. The present work summarizes the first results of a research activity that is being carried out by the Water Research Institute of the National Research Council (IRSA-CNR), supported by the Italian Department of Civil Protection (DPC), aiming at developing a vulnerability assessment methodology for drinking water infrastructures, from source to tap, in case of external threats. The objective is to provide the DPC with a simple and effective Decision Support System, helpful for properly managing emergencies, defining optimal strategies and interventions. The proposed methodology exploits the enormous potentialities of Bayesian Belief Networks as Decision Support Systems, particularly when a probabilistic analysis is required and when data of different nature have to be integrated (literature results, field data, expert judgment, …). The structural complexity of water distribution networks, constituted by interconnected components which regulate, treat and convey water, is resolved through the introduction of specific vulnerability sub-models. Once defined a conceptual functioning scheme for each single component of the infrastructure, its vulnerability is estimated with a specific Bayesian Network, depending on the main structural and environmental features influencing system response to physical damages and contaminations. The selected features are related through weighted causal relationship, in order to quantify the influence of specific variable states on the final vulnerability judgment. The proposed methodological approach reveals really powerful, as it allows a quick and effective estimate of the vulnerability of the whole infrastructure in case of emergencies. Nevertheless it could reveals also a useful tool for planning and management purposes. The integration of this mathematical tool with a Geographical Information System, which is currently under development, will provide the stakeholders with a precious framework for investigating the behavior of drinking water infrastructures and for decision making. In the present paper the main aspects of the methodology, as well as its most relevant strengths and the weaknesses, are discussed, with particular attention to the future developments and applications of the research activity.
Year: 2013