Quantitative understanding of the load–deflection mechanisms of a flexible barrier in intercepting debris flows is critical for barrier design, but remains practically challenging due to the difficulties involved in capturing multi-phase, multi-way interactions. This study employs a physics-based coupled computational fluid dynamics and discrete-element method (CFD–DEM) to simulate a flexible ring-net barrier as a permeable, deformable multi-component system by DEM and model a debris flow as a mixture of discrete particles and a continuous slurry by DEM and CFD, respectively. The CFD–DEM coupling framework offers a unified treatment of in-flow solid–fluid interaction, flow–barrier interaction and interactions among barrier components. Numerical predictions of key flow–barrier interactions and cable forces show reasonable consistency with large-scale experiments. Systematic simulations with varying flow–barrier height ratios ε and flow dynamics are performed to examine the evolving mechanisms of load sharing and transmission and quantify the ε-dependent load–deflection modes. The ratio ε is found to bear a strong, positive correlation with the key barrier response in three typical modes. The post-peak barrier deformations experience shrinkages with ε ≤ 0·6 and expansions when ε > 0·6. This study helps to improve understanding of the load–deflection mechanisms for practical design of flexible barriers in mitigating debris flows.
Flood retention lakes (RLs) are widely employed in rural‐urban catchments with low impacts on the natural environment. However, insights are lacking regarding the control of climate conditions on RLs' performances and how they are affected by different geographic configurations. This study applies a 2D hydrodynamic model to perform a catchment‐scale performance assessment of RLs beyond the scope of analytical and hydrological models. We conduct extensive numerical experiments of rainstorm‐induced flooding in a rural‐urban catchment with a constructed RL and blueprinted ones upstream. Results demonstrate an L‐shaped band of satisfactory performance of the current RL in the frequency‐duration diagram, which coincides with short return periods (4 hr), or short durations (<3 hr) and moderate to long return periods (5–50 years); such L‐shaped pattern is also valid for additional RLs and their combinations. With the increase in event size, the first two modes of RL performance (out of four) correspond to effective flood mitigation. When working jointly, RLs with series configurations are more effective in reducing the mainstream flood peaks, while parallel connections provide a greater spatial extent of flood hazard mitigation. For both series and parallel configurations, the upstream‐weighted settings tend to outperform downstream‐weighted ones under more extreme events; the decentralized arrangement in the urban area yields more benefits in spatial flood hazard mitigation compared to the centralized case. The study highlights the critical role of rainstorm severity (with possible spatiotemporal variabilities) in controlling RL performances despite various configurations and hydraulic settings.