Elsevier

Water Research

Volume 35, Issue 9, June 2001, Pages 2268-2276
Water Research

Numerical and experimental hydrodynamic studies of a lagoon pilot

https://doi.org/10.1016/S0043-1354(00)00502-9Get rights and content

Abstract

The determination of residence time is of major interest in terms of the characterisation, design and modelling of most environmental engineering processes, where a proper and homogeneous fluid distribution is often essential, and especially in the domain of wastewater or waste material treatment. In this paper, two different numerical methods for obtaining theoretical predictions of residence time distributions using the finite volume method are used. The first one consists of solving a transport equation of the local mean age of the fluid, which is the average time that a fluid particle takes to reach any point of the domain from a supply inlet. The result obtained is a spatial distribution of the local mean age of the fluid, which may be displayed as isocontours in the space domain considered. The second one consists of injecting a virtual particle stream (i.e. a fluid particle having the same density as the surrounding fluid, and treated numerically as a tracer) and measuring the time elapsed between the injection and the termination of the trajectory using a Lagrangian reference frame. The result obtained is expressed as an exit time distribution and may be displayed as a histogram. Finally, a comparison with measurements of mean residence times of tracers in a lagoon is made in order to establish the relevance of this method in concrete form. The prediction enables the instantaneous determination of the geometrical characteristics of the flow that contribute actually to the residence time dispersion. Its appropriate use prior to the design of water treatment and waste material treatment installations should contribute to prevent undesirable flow patterns such as short-circuiting and dead spaces.

Introduction

Lagoons are commonly used in wastewater and waste material treatment to reduce organic pollution as well as bacteriological contamination. This kind of facility seems adapted for the treatment of slurry. Such systems are characterised by low concentrations of purifier microorganisms and require large residence times (Degrémont, 1989). The treatment efficiency or the extent of conversion of waste material depends upon the type of waste, the organic loading, and the environmental conditions such as the temperature and the amount of wind and sunlight radiation. However, it is known that for lagoons, i.e. for channel flows as well as for ponds, the performance of the treatment process may be particularly affected by problems directly related to the residence time and arising from the hydrodynamic flow pattern, such as short circuiting or dead spaces (Ferrara and Harleman, 1981; Polprasert and Bhattarai, 1995). Such problems may be reduced by a better control of the flow.

The determination of residence time is thus of major interest in the design and characterisation of such installations, where a proper and homogeneous fluid distribution is often essential. A good knowledge of the flow, including the determination of residence time distributions (RTD), allows the optimisation of the mass transfer or the degradation of pollution. More generally, for these different processes, the determination of the RTD is required in order to design the optimal geometry of the reactor or to calculate the yield of the reactions with kinetic parameters (Villermaux, 1995; Thereska, 1998; Levenspiel, 1999).

The aim of this study is to characterise the flow in a channel with compartments in order to improve the design of a slurry treatment system, by reducing the dispersion of the mean residence time. Two different modelling approaches leading to information relative to spatial and temporal residence time distributions are presented, and the predictions are compared with the results of a stimulus-response experiment, consisting of local residence time measurements of a tracer along a scale model channel. Solving conservation (mass and momentum) and possibly coupled transport equations linked to the turbulent nature of the flow (whenever it is necessary) provides the spatial distribution of relevant variables which describe the basic flow, using classical computational fluid dynamics (CFD) techniques such as the finite volume method. Once these variables are determined, two decoupled approaches may be used in order to determine both the spatial distribution of the mean residence time and the temporal distribution of exit residence times.

It is shown that in a complex geometry classically encountered in the domains of water treatment and waste material treatment, that the prediction of the spatial mean residence time distribution should enable to prevent undesirable flow patterns due to the geometry of the facility which affect the dispersion of the residence time. Finally, a comparison with measurements of mean residence times spatial distributions and RTD of a tracer (lithium chloride) in a scale model lagoon is made in order to establish the relevance of the simulations in concrete form. The comparison of the mean residence time spatial distribution shows that this variable is satisfactorily predicted even though the temporal RTD at the outlet is less accurately predicted.

Section snippets

Physical model

The first approach consists of solving a steady transport equation of the local mean age of the fluid (Spalding, 1958; Sandberg, 1981; Sandberg and Sjöberg, 1983), which is the average time that a fluid particle takes to reach any point of the domain from a supply inlet. This scalar incorporates the time linked to the movement due to laminar and possibly turbulent diffusion. The result obtained is a spatial distribution of the local mean age of the fluid, which may be displayed as isocontours

Analysis of the hydraulic flow

The grid contains 290×33×27 cells, i.e. 258 390 cells. The boundary conditions applied to the domain are the following :

  • •

    at the flow inlet, the axial velocity equals 1.347×10−5 m s−1 and the mean residence time is set to zero;

  • •

    at the flow outlet, the static pressure is set to zero;

  • •

    on the upper free surface, ∂Φ/∂z=0 where z is the vertical coordinate and Φ stands for the velocity horizontal components and mean residence time, and the vertical velocity component is set to zero;

  • •

    on the vertical free

Conclusion

A transport of the conservation equation of the mean residence time was solved with the finite volume method and using a CFD commercial software, Fluent, that was modified and recompiled. The CFD analysis also provides directly the exit residence time temporal distribution. The spatial distribution of the mean residence time and the exit residence time distribution are predicted in a complex lagoon flow, and are compared with experimental measurements obtained with a pulse injection method

References (17)

There are more references available in the full text version of this article.

Cited by (0)

View full text