Modeling multiple-pipe failures in stormwater networks using graph theory

ABSTRACT One way to enhance the resilience of urban stormwater networks is to identify critical components of the network for targeted management and maintenance strategies. However, tackling multiple-pipe failures poses a significant hurdle due to the computational burden of conventional methods, making it impractical to analyze all potential combinations of pipe failures. To address this research gap, this study proposes a novel method based on complex network analysis to determine critical pipe failure combinations. The method incorporates network topology and estimates flooding impacts based on graph measures instead of solving complex hydraulic equations. The method is tested on two case studies with varying loop degrees and for different failure levels and compared with the state-of-the-art hydrodynamic modeling method (HMM) in terms of accuracy and computational time. Results show that the proposed graph theory method (GTM) can be used to identify the most critical pipe failure combinations. However, the accuracy of the GTM decreases slightly with increasing failure level and with increasing loop degree of the network. A hybrid graph-hydrodynamic model (GHM) is also developed as part of the study which combines advantages from both GTM and HMM.