Operational Risk-Based Decision Making for Wastewater Pipe Management

Recent advances in the development of wastewater pipe deterioration models allow a better quantification of the uncertainty in the likelihood of failure of pipes. However, the consequence of failure is subjected to many uncertainties that are not easy to be captured and support effective risk-based decision making. One component of the consequence of failure is the effect of the failed components on the network-level performance. An operational risk metric index is defined as a combination of the probability of failure and the effect (consequence) of the failed component at the network level. Three topological graph metrics (betweenness, eigenvector centrality, and principal component centrality) are used to identify the criticality of components in large wastewater systems. Their impact on the network performance is quantified in terms of graph connectivity. Both metrics lead to different rankings of components, making hard the decision-making process. The ordered weighted averaging (OWA) technique is used, as a data fusion technique, to determine a new consequence metric that is an aggregation of the information obtained from individual graph theoretical metrics. This metric is coupled with the likelihood of failure to quantify the operational risk of failure of the components. The uncertainty is propagated through the orness function, which is dependent of the decision-maker's attitude toward risk. The results show that the OWA aggregation reasonably captured the critical areas in the city. Different risk levels are identified and projected onto GIS maps for visualization of the high-risk pipes in the network. The proposed approach is applied on the wastewater system of the city of Calgary.