Author(s): Qijie Li; Dongfang Liang; Reinhard Hinkelmann
Linked Author(s): Dongfang Liang, Reinhard Hinkelmann
Keywords: Drainage systems; Flood risk management; Hydrographic modelling; Physical modelling; Surface-drainage interaction; Urban flooding
Abstract: Accurately quantifying the discharge capacity of urban drainage systems and their dynamic interactions with surface flooding is essential for the design of resilient pipe networks and for managing extreme urban flood events. However, limited high-resolution, site-specific data restricts our understanding of flow conditions within drainage structures and hinders the development of robust modelling frameworks. To address this gap, this study develops an instrumented full-scale physical model that enables interactive hydrographic modelling of coupled surface-drainage systems. The experiments reveal the dynamic response mechanisms between overland floodwater and subsurface drainage networks. Preliminary findings indicate a critical anisotropic behaviour in discharge flow rates as surface water depth increases, suggesting that conventional design criteria may substantially overestimate drainage capacity. A set of general flow characteristics is identified, and a unified interaction regime for drainage systems is proposed based on physically derived quantification. Furthermore, a physics-informed graph-theoretic neural network is developed to enhance the stability and predictive accuracy of urban flood forecasting. Building on these insights, a comprehensive assessment framework for extreme flood risks is established, together with a scenario-based multi-level evaluation method that captures risk transmission pathways across complex urban systems. This interdisciplinary approach provides new evidence and modelling tools to support sustainable urban drainage design and resilience-oriented flood risk management.
Year: 2026