Numerical Analysis of the Effluent Dispersion in Rivers with Different Longitudinal Diffusion Coefficients

Abstract

The knowledge of pollutants dispersion in water bodies is a matter of concern in water quality control, especially when a new industrial development is installed e.g. near riverbanks. To predict pollutants dispersion in rivers, analytical, experimental and in-situ measurement can be performed. However, analytical estimation usually results in low accuracy, while experimental or in situ measurement are quite expensive in time and equipment. Hence, Computational Fluid Dynamics (CFD) approach is other alternative that can be used to obtain simple and accurate results for mass transport in rivers. In other words, it is a good alternative to analyse pollutants dispersion. As it is known, longitudinal diffusion coefficient (E) has strong influence on pollutants spreading into the water body. Therefore, the purpose of this paper is to analyse the effects of E on the mass transport of a conservative pollutant in rivers and channels via CFD. Contaminant dispersion is carried out by a scalar advection-diffusion transport equation that represents the conservation of mass. The velocity and pressure fields are calculated, considering an incompressible fluid, through the Navier-Stokes and the continuity equations. Numerical and analytical results, for one-dimensional (1D) flow, are compared in order to obtain the concentration field, over time and space, using different parametric equations. The concentration field showed significant differences of concentration peak and arrival time of the plume depending on the equation used to predict E. Numerical results, for two-dimensional (2D) flow, are compared with the experimental data from Modenesi et al. (2004). Such analyses are necessary to establish an appropriate correlation between simulated and real channel. The use of different parametric equations for the E in a 2D channel reveals significant differences of concentration peak and arrival time of the plume. As expected, the numerical results of the transport of pollutants show the dependence on the parameterization of the longitudinal dispersion coefficient. The one that best represents the distribution of pollutants is that proposed by Kashfipour & Falconer.

Keywords

• computational fluid dynamics (cfd); longitudinal diffusion coefficient; parametric equations.