In the last decade, machine learning (ML) technology has been transforming daily lives, industries, and various scientific/engineering disciplines. In particular, ML technology has resulted in significant progress in neural network models; these enable the automatic computation of problem-relevant features and rapid capture of highly complex data distributions. We believe that ML approaches can address several significant new and/or old challenges in urban drainage systems (UDSs). This review paper provides a state-of-the-art review of ML-based UDS modeling/application based on three categories: (1) operation (real-time operation control), (2) management (flood-inundation prediction) and (3) maintenance (pipe defect detection). The review reveals that ML is utilized extensively in UDSs to advance model performance and efficiency, extract complex data distribution patterns, and obtain scientific/engineering insights. Additionally, some potential issues and future directions are recommended for three research topics defined in this study to extend UDS modeling/applications based on ML technology. Furthermore, it is suggested that ML technology can promote developments in UDSs. The new paradigm of ML-based UDS modeling/applications summarized here is in its early stages and should be considered in future studies.

Urban drainage systems (UDSs) continue to face challenges, despite numerous efforts to improve their sustainability through design, planning, and management. The goal of such initiatives is to avoid and minimize flooding as well as maintain the UDS’s sustainable functionality, which can be analyzed using a stormwater management model (SWMM). In this study, a multiobjective automatic parameter-calibration (MAPC) framework was developed based on the SWMM. It consisted of three steps: sensitivity analysis (Step I), objective selection (Step II), and SWMM parameter calibration (Step III). The proposed MAPC framework was verified using the Yongdap drainage network located in Seoul, South Korea. The resultant MAPC framework demonstrated that the system characteristics (such as percent of impervious area and hillslope) and problems in UDS design, planning, and management can be well reflected by the corresponding model. The MAPC framework proposed in this study can contribute to UDS modeling sustainability.

A traditional urban drainage system (UDS) planning approach generally considers the most probable future rainfall scenario. However, this single scenario (i.e., scenario-optimal) planning approach is prone to failure under recent climatic conditions, which involve increasing levels of uncertainty. To overcome this limitation, an alternative is to consider multiple scenarios simultaneously. A two-phase multi-scenario-based UDS planning approach was developed. Scenario-optimal solutions were determined for a set of scenarios in Phase I, as the traditional planning approach, while common elements across the scenarios were identified and used to consider components-wise regret cost concept for Phase II optimization, from which a compromise solution was sought. The storm water management model was dynamically linked with the harmony search algorithm for each phase optimization model. The proposed approach was demonstrated in the planning of the grid-type drainage networks of S-city. The compromise solution was compared with the scenario-optimal solutions (Phase I) with respect to cost effectiveness and system performance under scenarios that were not considered in the planning phase.

With climate change, the frequency and severity of localized heavy rainfalls are increasing. Thus, for urban drainage networks (UDNs), particularly those in aging cities such as Seoul, Republic of Korea, flood risk management challenges are mounting. Conventional design standards typically apply uniform roughness coefficients based on new pipe conditions, neglecting the ongoing performance degradation from physical influences. This study introduces a methodology that systematically incorporates pipe age and size into roughness coefficient scenarios for higher-accuracy 1D–2D rainfall–runoff hydrologic–hydraulic simulations. Eleven roughness scenarios (a baseline and ten aging-based scenarios) are applied across seven UDNs using historical rainfall data. The most representative scenario (S3) is identified using a Euclidean distance metric combining the peak water-level error and root mean square error. For two rainfall events, S3 yields substantial increases in the simulated mean flood volumes (75.02% and 76.45%) compared with the baseline, while spatial analysis reveals significantly expanded inundation areas and increased flood depths. These findings underscore the critical impact of pipe deterioration on hydraulic capacity and demonstrate the importance of incorporating aging infrastructure into flood modeling and UDN design. This approach offers empirical support for updating UDN design standards for more resilient flood management.