An intelligent model for surface and subsurface water simulation considering the effect of pipeline siltation

Siltation in stormwater pipes is an important contributor to urban flooding, yet its effects on pipe-scale hydraulic resistance and surface–subsurface flow exchange are not adequately represented in existing models. This study integrates the multi-factor coupling probability sedimentation function, Chaos adaptive genetic algorithm for optimization of probabilistic deposition, and attention bidirectional long short-term memory neural network, and proposes an intelligent prediction model for the sedimentation of pipeline siltation particles. On this basis, the instantaneous continuity control equation of fluid dynamic evolution process, instantaneous momentum conservation control equation, sedimentation pipeline flow control equation, and translation control equation are introduced to establish a high‑resolution model of overflow characteristics in sediment‑clogged pipes. Finally, by integrating the water flow exchange control equations of clogged pipes in different scenarios, a flood evolution simulation model considering the influence of pipe clogging is proposed. The comparative analysis results show that when the clogging degree increases from 0 to 0.3, the average accuracy and correlation coefficient evolution intervals of high‑resolution model of overflow characteristics in sediment‑clogged pipes are [0.934, 0.951] and [0.975, 0.986], respectively. Hydraulic analysis further showed that pipe siltation reduced conveyance capacity through the combined effects of effective-flow-area reduction and additional sediment-bed resistance, while the blocked flow increased nonlinearly with siltation degree and became more sensitive as the pipe approached the half-full condition. Compared with traditional flood simulation models, the flood evolution simulation model considering the influence of pipe siltation has high accuracy.