Water Safety Plan and Fault Tree Analysis for the management and risk assessment of a drinking water supply system

This work of thesis studies the approach of the Water Safety Plan, that is applied to a part of the drinking water supply network managed by AIMAG S.p.A. The Water Safety Plans have been introduced by the World Health Organization and have been identified as the most effective way to systematically guarantee the safety of the drinking water system, the quality of the water supplied, and the protection of the health of consumers. Through a preventive and qualitative analysis that includes risk assessment and management, the Water Safety Plans make it possible to guarantee over time the absence of potential physical, biological, and chemical hazards in the water available for consumption. The result obtained through the application of the Water Safety Plan involves the construction of a risk matrix, that allows the development of an improvement plan for the control of each hazard and associated risk (in priority scale). After that, the previously performed analysis is supported through the application of statistical and probabilistic techniques used in the reliability analysis to optimize the control activities based on a risk assessment. The first one is the Fault Tree Analysis, which is a deductive technique supported by Boolean algebra that focuses on a particular undesirable event and provides a method for identifying possible causes. In parallel with the application of the Fault Tree Analysis, the Monte Carlo method is applied, which is a computational method based on random sampling of certain variables to obtain numerical results. Finally, the result obtained through the application of both methodologies is the definition of the probability of failure of some nodes and internodes that are part of a portion of the drinking water supply network managed by AIMAG S.p.A. At the end, the work of thesis lead to results which prove that the qualitative approach of the Water Safety Plans can be effective, especially if supported by different types of analysis