Coupled flood and sediment transport modelling with adaptive mesh refinement

Abstract

Coupled flood and sediment transport modelling in large-scale domains has for long been hindered by the high computational cost. Adaptive mesh refinement is one of the viable ways to solving this problem without degrading the accuracy. This goal can be accomplished through mesh adaptation, e.g., mesh coarsening and refining based on the dynamic regime of the flow and sediment transport along with bed evolution. However, previous studies in this regard have been limited to cases either without involving sediment transport or featuring flow-sediment-bed decoupling and the assumption of sediment transport capacity, which are not generally justified. Here, a coupled hydrodynamic and non-capacity sediment transport model is developed on adaptive non-uniform rectangular mesh. The proposed model is validated against experimental tests and numerical results based on the fixed meshes. It is demonstrated that the proposed model can properly capture shock waves, resolve the wetting/drying transition and reproduce morphological evolution. Compared with models based on the fixed meshes, the proposed model features great advantage in computational efficiency and holds promise for wide applications.shallow water flow, sediment transport, adaptive mesh Citation:Huang W, Cao Z X, Pender G, et al. Coupled flood and sediment transport modelling with adaptive mesh refinement.

A Non-Equilibrium Sediment Transport Model for Dam Break Flow over Moveable Bed Based on Non-Uniform Rectangular Mesh

“…In this respect, the computed bed elevations were generally in reasonable agreement with the measurements. Figure 17 shows the calculated and measured water level at four pairs of gauges in a period of 20 s. It can be found that the calculated water levels at G1 and G4 are about 40 mm lower than the measurement, which are also reported by other researchers [27,35]. One possible reason is that strong 3D characteristics of the flow at the corners of the sudden expansion of the flume width cannot be resolved by 2D depth average hydrodynamic model.…”

“…In this section, the benchmark experiment of partial dam-break flow over mobile bed, which was also performed at UCL-Belgium [16] and has been used to test morphological models by many researchers [17,21,27,35,50], is adopted here to further verify the ability of the model to simulate bed evolution of a mobile bed caused by partial dam-break flow. As shown in Figure 13, the flume is 3.6 m wide and about 36 m long.…”

“…In fact, in large-scale simulations, the high-gradient regions where fine grids are required are much less than the whole computational domain [33]. In order to improve computational efficiency and at the same time maintain the required accuracy, non-uniform grids have been recently used to simulate dam break flow over movable bed, either in the form of non-uniform triangular mesh [12,15,34] or rectangular mesh [35,36]. Compared to the triangular mesh, non-uniform rectangular mesh is more convenient to achieve high order accuracy in space, while keeps most advantages of the rectangular mesh [37].…”

A depth‐averaged two‐phase model for debris flows over erodible beds

“…Technically, concern arises from the increase in computational cost, especially when the TPE‐T2‐ED‐CM model is to be applied for modeling large‐scale natural debris flows characterized by a wider particle size‐distribution. In this regard, it is advisable to employ the technique of adaptive mesh refining, which is shown to save computational time efficiently by an order of magnitude (Huang et al ., ).…”

A Two-Dimensional Depth-Averaged Sediment Transport Mobile-Bed Model with Polygonal Meshes

“…The mesh type was, however, limited to the structured, orthogonal curvilinear coordinate system. Huang et al [18] developed an adaptive mesh refinement model for dam-break flows over mobile-bed streams. The model was validated against experimental cases with good agreements.…”