Narrow Results

Experiments with uniform sediments of different grain sizes were performed and net transport rates were measured. A comparison between four existing models to predict the net transport rate of uniform sediments was achieved. The prediction of offshore transport was considered since it was significant to predict the erosion of the beach profile. An enhanced PIV technique developed by the present authors, Ahmed and Sato (2001), was utilized to measure the time-dependent moving layer thickness. A relationship between a dimensionless moving layer thickness (δ_s/D), since δ_s stands for the moving layer thickness and D is the mean diameter of sediment, Shields parameter and sediment-flow acceleration parameter (a_m/D), in which a_m is the semi-excursion of water particle, was deduced. A new model was developed based on Dibajnia and Watanabe (1992) conception by improving the definition of moving layer thickness. The new model agreeably predicts the on-offshore transport of uniform sediments.

Experiments with heterogeneous sediments, composed of three different grain sizes, were conducted under asymmetric oscillatory sheetflows. Net selective transport and total net transport of heterogeneous sediments were measured. A sensitivity analysis utilizing a new uniform sediment transport model developed by the present authors (Ahmed and Sato, 2003) was performed. The essence of this analysis was to determine a relevant representative grain diameter of heterogeneous sediments to predict the total net transport. A comparison between transport model prediction of Ahmed and Sato model and four existing models showed that both Ahmed and Sato, and Dibajnia et al. (2001) models were relatively preeminent. Further comparison between two recent selective sheetflow transport models utilizing the present experimental data and available data was performed. The transport mechanism of an individual sediment particle was investigated by means of enhanced PIV technique, Ahmed and Sato (2001), and then the armoring effect was evaluated. A new sheltering and exposure criterion was proposed and applied in Ahmed and Sato (2003) transport model. The selective transport model prediction showed a good agreement with the experimental data.

Laboratory experiments were conducted under sinusoidal sheetflow conditions. By using image analysis, which overcame the demerits introduced by intrusive measurements, the time-varying as well as maximum erosion depths for different flow conditions were estimated. The temporal variation of suspended sand concentration in the sheetflow layer showed high concentration during the deceleration phase and relatively low concentration during the acceleration phase. Rapid sand deposition was observed around flow reversals. The phase lag between the free stream velocity and the sand concentration increased with elevation. The temporal and the spatial distribution of suspended sand concentration revealed an asymmetry in the suspension process, namely relatively long suspension (erosion) and short sedimentation (deposition) reflecting the asymmetry in turbulence. The instantaneous sand particle velocities were estimated using a PIV technique. The mean particle horizontal velocity was found to decrease at the beginning of the acceleration phase, corresponding to the rapid deposition of sand.

A two-phase flow model based on the mass and momentum conservation is presented, which can simulate the fluid and sediment movement on a flat bed under sheetflow conditions. The governing equation in the vertical direction is modified by considering the influence of static normal intergranular stress. Horizontal pressure gradient is modified in terms of sediment concentration. A criterion is introduced to decide the temporal variation of the still bed level during a one-wave cycle. Vertically parabolic eddy viscosity and corresponding sediment diffusion coefficient are assumed. Numerical results include initial validation comparisons with the existing experimental data from Horikawa et al. [1982]. All measured properties, such as the concentration, sediment velocity and sediment flux are reproduced quite well. Further investigations on the experimental data of Delft Hydraulics [Ribberink and Al-Salem, 1995; Dohmen-Janssen, 1999] cover a wide range of pure sinusoidal wave and combined wave/current flow conditions for different sediment sizes. The numerical results are satisfactory with respect to the measured time-varying and time-averaged concentration distributions both in the sheetflow layer and the suspended layer. Taking into account the various experimental measurements, the present two-phase flow model shows the significant superiority over an existing two-phase flow model [Mina and Sato, 2004]. Comparison between the measurement and the present simulation for sediment flux and net transport rate is also performed.