Due to a variety of contaminants in floodwater, exposure to urban pluvial flooding may pose a health risk to humans. In-sewer defects may cause increased pluvial flooding, possibly increasing health risks. This paper addresses the impact of in-sewer defects on urban pluvial flooding and, subsequently, on infection probabilities for humans. As such, it provides a necessary input for risk-informed sewer maintenance strategies in order to preserve the hydraulic performance of a sewer system. Critical locations in sewer networks can be safeguarded through detecting changes in hydraulic properties of the sewer system, by using monitoring equipment or alternative inspection methods. Two combined sewer systems in The Netherlands with different characteristics are studied. The catchment-wide average infection probability was calculated using Quantitative Microbial Risk Assessment (QMRA) and flooding frequencies from Monte Carlo simulations with a hydrodynamic model. For the studied catchments, it is concluded that the occurrence of flooding is significantly affected by sediment deposits and, consequently, the infection probability as well. The impact of sediment deposits on infection probabilities depends on sewer systems characteristics and varies within the catchment. The results in this paper also demonstrate that further research on the relationship between flood duration and infection probabilities is required.

Underground water infrastructure is essential for life in cities. The aging of these infrastructures requires maintenance strategies to maintain a minimum service level. Not all elements are equally important for the functioning of the infrastructure as a whole. Identifying the most critical elements in a network is crucial for formulating asset management strategies. The graph theory is presented as a means to identify the most critical elements in a network with respect to malfunctioning of the system as a whole. As opposed to conventional methods, the proposed method does not rely on iterative hydraulic calculations; instead, the structure of the network is taken as a starting point. In contrast to methods applied in practise, the results are independent on the chosen test-load. Because of the limited calculation effort, the method allows the analysis of large networks that are now, for practical reasons, beyond the scope of methods applied so-far.