摘要：The study deals with the mixing of passive scalars (such as pollutants) in open-channel confluences when the two inflows exhibit different concentrations. The dispersion of such passive scalars is investigated through the analysis of the processes enhancing the mixing in the confluence in order to estimate the length for complete mixing Lm in the downstream branch. The aim of this study is then to establish a correlation between this length for complete mixing Lm and characteristics of the confluence, namely its angle and two hydrodynamic parameters that are the momentum ratio M* of the inflows and the width-to-depth ratio of the downstream branch b/h. In this work, the flow in the confluence is numerically simulated by RANS calculation coupled with the advection-diffusion equation at field scale, i.e. with Reynolds and Froude numbers corresponding to real rivers (Re = 107 and Fr=0.1). The role of other parameters such as bed discordance or bed forms... is not addressed here. The numerical results highlight that the mixing in the confluence and its downstream branch, is enhanced by the presence of secondary currents that are themselves strongly affected by the characteristics of the confluence. The results aim at getting an operative empirical law linking the geometric and hydrodynamic parameters of the confluence with Lm.
其他摘要：The study deals with the mixing of passive scalars (such as pollutants) in open-channel confluences when the two inflows exhibit different concentrations. The dispersion of such passive scalars is investigated through the analysis of the processes enhancing the mixing in the confluence in order to estimate the length for complete mixing Lm in the downstream branch. The aim of this study is then to establish a correlation between this length for complete mixing Lm and characteristics of the confluence, namely its angle and two hydrodynamic parameters that are the momentum ratio M* of the inflows and the width-to-depth ratio of the downstream branch b/h. In this work, the flow in the confluence is numerically simulated by RANS calculation coupled with the advection-diffusion equation at field scale, i.e. with Reynolds and Froude numbers corresponding to real rivers (Re = 107 and Fr=0.1). The role of other parameters such as bed discordance or bed forms... is not addressed here. The numerical results highlight that the mixing in the confluence and its downstream branch, is enhanced by the presence of secondary currents that are themselves strongly affected by the characteristics of the confluence. The results aim at getting an operative empirical law linking the geometric and hydrodynamic parameters of the confluence with Lm.
