Study of Two-Dimensional Shear Effects on Flow and Pollutant Transport in Meandering Channels

Abstract

In this study, the depth-averaged two-dimensional model was used to simulate the shear effects on meandering river flows and pollutant transport with reasonable accuracy and efficiency. To validate the numerical model, large-scale experiments were conducted in the River Experiment Channel (REC) large-scale meandering channels, in which the data of flow by acoustic Doppler current profiler (ADCP) measurement and pollutant transport by tracer tests in meandering channels were collected. Using velocity measurements, the effect of the secondary current on the primary flow distribution was analyzed. First, the relations of the secondary flow strength to the depth to radius-of-curvature of the channel and channel roughness were found. Then, the vertical profile equation for the secondary flow was developed reflecting the nonlinear term effects on the secondary flow which were omitted in the previous studies. The proposed equation generated a vertical profile that showed a decrease in the maximum secondary flow strength at the top and bottom of the profile, which is different from the existing equations. The proposed velocity profile equation was inserted into the momentum equations with the dispersion stress method for the two-dimensional flow solver HDM-2D, in order to induce the shear effect of secondary flow, which is normally neglected in the depth averaging process. The simulation results using the proposed equation to the dispersion stress method showed that the simulation of the Rozovskii channel showed improvement over the dispersion stress model using deVriend and Kikkawas velocity equation, which was based on the linear behavior between the secondary flow and primary flow, as well as over the model with no-dispersion stress in the distributions of primary flow velocity. The validation results with REC meandering channels revealed that the 2D hydrodynamic model with dispersion stress term gave a better fit of primary flow distribution to the experimental data than the simulation without the dispersion stress term. To find the characteristics of pollutant transport in meandering channels, the two-dimensional dispersion coefficients were calculated using the vertical velocity profiles which represents shear dispersion from the ADCP measurements, and the results were compared with the values calculated by applying 2D stream-tube routing procedure to the concentration curves obtained from 2D transient tracer experiments. The velocity driven two-dimensional dispersion coefficient in rivers which was calculated using the shear flow dispersion coefficient equation developed by Fischer et al. (1979) is easy to obtain since it did not require the tracer test experiments. The results showed that non-dimensional longitudinal dispersion coefficient by velocity profile ranges from 4 to 6 which is close to Elders result, while non-dimensional transverse dispersion coefficient ranges from 0.05 to 0.4. However, the dispersion coefficients calculated using the 2D stream-tube routing procedure were quite large: 4-5 times larger than velocity driven values for longitudinal dispersion coefficients, 1-3 times larger for the transverse dispersion coefficients. These differences could be explained by the fact that the concentration-driven dispersion coefficient included the mixing effects due to the irregularities of the channel, storage zones and numerical dispersion while velocity-driven coefficients only accounted for shear flow effects. Then, CTM-2D advection-dispersion model was applied to simulate mixing in meandering channels and the dispersion stress term improved accuracy by 1%. The calibrated dispersion coefficients by the CTM-2D model were between the velocity-driven results and concentration driven results. The simulation results proved the applicability of the CTM-2D model in reproducing concentration curves in meandering channels.