A depth-averaged two-phase model for fluvial sediment-laden flows over erodible beds

Fluvial sediment-laden flow represents a class of fluid-solid two-phase flows, which typically involve multi grain sizes, interphase and particle-particle interactions, and mass exchange with the bed. However, existing depth-averaged models ignore one or more of these physical aspects. Here a physically enhanced, coupled depth-averaged two-phase model is proposed for fluvial sediment-laden flow, which explicitly incorporates all these aspects and also turbulent Reynolds stresses. A well-balanced numerical algorithm is applied to solve the governing equations of the model. The present model is benchmarked against a series of typical cases, concerning refilling of a dredged trench, bed aggradation due to sediment overloading, and flood flows due to landslide dam failure. It features encouraging performance as compared to measured data and a quasi single-phase mixture model. The present model reveals that the larger the grain size, the slower the sediment fraction transports, which concurs with prior findings from experimental observations and field data. Also, the fluid phase Reynolds stresses are considerable where the flow rapidly varies, whilst the solid phase Reynolds stresses are negligible if sediment concentration is sufficiently low.