Vulnerability Assessment of Storm Water Drains using Bayesian Belief Network

Urban flash floods have become increasingly frequent and severe, resulting in significant damage to lives, infrastructure, and the economy. A major contributor to urban flooding is the inability of stormwater drainage (SWD) systems to efficiently convey excess runoff, which often leads to localized failures and cascading disruptions across the network. Assessing the vulnerability of such systems becomes crucial for developing effective flood risk mitigation strategies. While conventional hydrodynamic models (e.g., SWMM, MIKE) are essential for predicting flood-related characteristics (e.g., peak flow/depth, duration) and inundation extents, they are limited in their ability to evaluate vulnerabilities within the system under rapidly changing rainfall patterns and account for uncertainties in decision-making processes. These limitations highlight the need for alternative approaches to analyze and address network vulnerabilities under dynamic conditions. The present study explores the potential of the Bayesian Belief Network (BBN) approach to evaluate vulnerabilities within the SWD system. This approach leverages the topological structure of drainage systems to assess interdependencies among components and flood-causing factors to better understand the cascading impacts of localized failures on the system-wide performance of a SWD network. The proposed BBN approach is tested on the Bangalore SWD network to identify critical zones under varying hydraulic loads. By providing probabilistic insights, BBNs enable a more comprehensive understanding of flood risk and improve decision-making under uncertainty. The findings of the study demonstrate the potential of BBN as a powerful tool for urban flood risk assessment and offer a comprehensive framework to strengthen flood resilience and guide infrastructural rehabilitation, planning, and management.