Numerical simulation and parameter sensitivity analysis of river sediment movement based on multi-objective optimization

River sediment movement is the core problem in river dynamics, which is of great significance to the design, operation and management of water conservancy projects. In this paper, a numerical simulation method of river sediment movement based on multi-objective optimization is proposed, and the key parameters are identified by parameter sensitivity analysis. A 3D coupled model of water and sediment is established. Based on Navier-Stokes equation and sediment continuity equation, it is solved by finite volume method, which can accurately simulate the velocity field, pressure field and sediment concentration distribution. The NSGA-II algorithm is introduced for multi-objective optimization, and the model parameters are optimized with the aim of minimizing simulation error and maximizing parameter stability. The experimental data comes from the measured data of a river from 2020 to 2023, and is processed by Z-score standardization and outlier elimination. The results show that the optimized model reduces the error of flow rate and concentration by 62.5% respectively, which significantly improves the simulation accuracy. Parameter sensitivity analysis adopts variance decomposition method and Latin hypercube sampling (LHS), and it is recognized that sediment diffusion coefficient and erosion rate are the key parameters that affect the simulation results, while resistance coefficient and deposition rate have little influence. Transactional analysis revealed the nonlinear relationship between parameters, which provided the optimization direction for complex river simulation.