Open access peer-reviewed chapter

Hydrodynamic Behavior of Flow in a Drinking Water Treatment Clarifier

By Wen-Jie Yang, Syuan-Jhih Wu, Yu-Hsuan Li, Hung-Chi Liao, Chia-Yi Yang, Keng-Lin Shih and Rome-Ming Wu

Published: January 1st 2010

DOI: 10.5772/7113

Downloaded: 3412

1. Introduction

Over 50% drinking water was supplied to the Taiwan’s public by sludge blanket clarifiers (Chen et al., 2003; Lin et al., 2004). The sludge blanket performs dual functions as a filter as well as a particle coagulator. Coagulation, the chemistry-based treatment stage, controls the characteristics of the generated sludge layer, whereas sedimentation, the hydrodynamic treatment stage, controls sludge layer stability. The existence of a sludge blanket in clarifiers is thereby essential to produce quality effluent water. In addition, the flow dynamics is an important parameter for the design of clarifiers.

The use of solids flux theory continues in many studies in the design and operation of sludge treatment processes (Takacs et al., 1991; Daigger, 1995; Wett, 2002). Ekama and Marais (2004) gave a survey on the development of one dimensional (1D) settler modeling. Although application of solids flux theory is good for studying the performance of clarifiers, it does not adequately describe the effect of hydrodynamic behavior in clarifiers (Narayanan et al., 2000). Computational fluid dynamics (CFD) has shown to be a powerful tool for resolving complex practical problems in engineering. (Hsu et al., 2007; Videla et al., 2008; Lin et al., 2008; Tao et al., 2008; Yang et al., 2007) Therefore some studies attempted to simulate full clarifiers via CFD (Deininger et al., 1998; Burger et al., 2005; Fan et al., 2007). Recently Wu et al. (2007) simulated flow pattern in a clarifier with porous medium as sludge blanket by 3D CFD. The first work to utilize a 3D, multiphase flow simulation for a clarifier is by Wu et al. (2008). Weiss et al. (2007) utilized non-Newtonian flow to model a circular secondary clarifier and showed that viscosity of sludge dominates the flow in clarifier. This study attempts to improve clarifier effluent quality by altering its geometric construction. The simulation is based on the sludge blanket clarifier at the Bansin Water Treatment Plant, Taiwan.

Bansin Water Treatment Plant (BWTP) is in Banchiao City, Taipei County, Taiwan. About every 20 minutes sludge blanket overturns somewhere and effluent solid flux increases. The turbidity of the clarified water is generally too high to produce quality clean water after sand filtering. In this work, four types of constructions of clarifiers are simulated by 3D, multiphase flow model to improve clarifier effluent quality. Type Ais the conventional clarifier used in BWTP. Type Bchanges inlet pipe to a large one (1.6 folds). Type Cchanges reaction well angle from 1200 to 900. Type Dchanges reaction well angle from 1200 to 600. As shown in previous research (Yang et al., 2008) that, the velocity of suspension in the reaction well is one of the factors affecting the quality of water discharge, so the improved method of Type Bis expected to slow the velocity of flow by enlarging diameter of inlet pipe. The improvement method of Type Cis the angle of 900 of reaction cover. It is hoped that the backflow could be limited within the reaction well, so that the residence time of suspension in the reaction well can be increased. The improved method for Type Dreverses the whole reaction cover, so the water quality may not deteriorate for reason of easy outflow of the suspension from the reaction well.


2. Geometry and meshes

Figure 1 indicates the geometry of the sludge blanket clarifier. The clarifier is 19×19×5.5 m3, with an impeller with 16 blades and diameter of 3 m. The impeller is located in center top of the clarifier. The inlet pipe is 0.9 m in diameter. The inlet pipe is connected to a draft tube 2.4 m in diameter. The inlet water velocity was typically operated at 0.34 m/s. The draft tube comprises the first reaction zone. The second reaction zone is outside the first reaction zone and inside a reaction well with upper and lower diameters of 8 and 13 m, respectively. The reaction well is 3.8 m high. Figure 2 shows all grids in the clarifier.

3. Governing equations and boundary conditions

3.1. Conservation equations.

Volume fractions represent the space occupied by each phase, and the laws of conservation of mass and momentum are satisfied by each phase individually. The volume of phase q,Vq, is defined by


where αis phase volume fraction occupied by each phase, and


The effective density of phase qis


whereρqis the physical density of phase q. In this work, only water and solid phases are considered.

The continuity equation for phase qis


wherevqis the velocity of phase qandm˙pqcharacterizes the mass transfer from the pto qphase, andm˙qpcharacterizes the mass transfer from phase qto phase p, hencem˙pq=m˙qpand


The momentum balance for water phase wyields


where Pis the pressure shared by all phases, andτw¯¯is the water phase stress-strain tensor


Hereμwandλware the shear and bulk viscosity of water phase,Rsqis an interaction force between water and solid phases.vswis the interphase velocity, defined as follows. Ifm˙sw0,vsw=vs; ifm˙sw0


The momentum balance for solid phase syields


wherepsis the solids pressure. For granular flows in the compressible regime, a solids pressure is calculated independently and used for the pressure gradient term,ps, in the granular-phase momentum equation.

The solids stress tensor contains shear and bulk viscosities arising from particle momentum exchange due to translation and collision. The collisional and kinetic parts are added to give the solids shear viscosity. The solids bulk viscosity accounts for the resistance of the granular particles to compression and expansion.

3.2. Phase interactions.

Rwsdepends on the friction, pressure, cohesion, and other effects, and is subject to the conditions thatRsw=Rwsand

FLUENT uses a simple interaction term of the following form:


whereKsw(=Kws)is the interphase momentum exchange coefficient. The fluid-solid exchange coefficientKswcan be written in the following general form:


whereτs, the particulate relaxation time, is defined as


wheredsis the diameter of particles of phases

For the Syamlal-O'Brien model (1989),


where the drag function and relative Reynolds number have the following forms:


This model is based on measurements of the terminal velocities of particles in fluidized or settling beds.vr,sis the terminal velocity correlation for the solid phase (Garside and Al-Dibouni, 1977):


Lift forces act on a particle mainly due to velocity gradients in the primary-phase flow field. The lift force will be more significant for larger particles. Thus, the inclusion of lift forces is not appropriate for closely packed particles or for very small particles. When a secondary phase paccelerates relative to the primary phase q, the inertia of the primary phase mass encountered by the accelerating particles exerts a virtual mass force on the particles. The virtual mass effect is significant when the secondary pahse density is much smaller than the primary phase density. Therefore virtual mass force is not either considered in this work. In order to make the calculation simple, the acting force between water and particle is dominated by the drag force, which is shown in theRswterm. The other relatively strong acting forces, such as combination between particle and particle by coagulant, can be reflected in the viscosity of solid phase.

3.3. Boundary conditions.

The boundary conditions are as follows:

vs=vw=Vat inlet pipe (15a)αs=0.005at inlet pipe (15b)vs=vw=0, at walls (15c)P= 0, at water/air surface (15d)

Equation (15a) and (15b) state that the inlet suspension is moving at a constant speed and has a solid volume fraction equal to 0.005. Equation (15c) describes no slip boundary conditions for water and solid phases. Equation (15d) describes the gauge pressure at the water surface (top of the clarifier) is zero.

The computational fluid dynamics program FLUENT 6.1 (Fluent Inc., USA) solved the governing equations, together with the associated boundary conditions, using hybrid mesh volumes generated by GAMBIT. This work uses three groups of meshes of different quantities (1 million, 2 million, 2.5 million) for calculation. The relative error of the solid volume fraction at the overflow is smaller than 10%. Therefore the numbers of mesh volumes in the whole clarifier are used about 2,500,000. The calculations were carried out with maximum relative error of 10-4 in fluid velocity evaluation.

4. Results and discussion

4.1. Velocity vector of water flow.

Figure 3 plots the velocity vector of water in the clarifier (inlet velocity = 0.3 m/s, impeller rotation speed = 0.3 rad/s). When inlet water is flowing into the draft tube, it is sucked to the top of the clarifier, owing to the rotating impeller. Then it goes down along the inside of the well and separates into two streams. One stream inside the well makes a strong cycling flow (#1) in the secondary reaction well, the other stream rises along the wall of the clarifier, to water surface, and descends along the outside of the reaction well, making another weaker cycling flow (#2). Although this cycling flow (#2) is weaker than (#1), it is still a strong density current that the flocs will be elutriated, leading to an average 20 minutes overturn as reported in Chen et al.(2003). An idea is to treat the flocs into a strong and dense blanket.

5. Comparison with Bansin Water Treatment Plant (BWTP)

Bansin Water Treatment Plant (BWTP) adopted 16 sludge blanket clarifiers to treat its raw water. This type of clarifier is used by many regions to produce drinking water, but relevant references are very few. Figure 4 is a comparison between observation in BWTP and CFD results. An observer stood beside the clarifier, recorded which areas sludge blanket was floating. Thus in figures 4(a)-(d), blue color means that water surface is clean; while yellow color indicates that sludge suspension was on that surface. Figures 4(a)-(d) were the observation results at 0 s, 600 s, 900 s, and 1200 s. Figures 4(e)-(h) were volume fraction contours calculated by FLUENT at 2400 s, 3000 s, 3300 s, and 3600 s. The patterns are similar between BWTP observation and numerical calculation and that sludge overturns from edge and corner every 20 minutes. Circular clarifier might be a solution to solve this problem.

5.1. Twophase flow.

Figure 5 reveals the contour of volume fraction of solid phase of Eulerian two phase flow as time evolves. The primary phase is water phase, the secondary phase is solid phase with particle size 10 μm and density 1,005 kg/m3. As time goes by, the solid particles were sucked from the draft tube to the top of the impeller, descends along the reaction well inside, be full of the whole reaction well, and overflow to the reaction well outside, making a relatively stable blanket at the bottom of the clarifier, and a dynamic upward particles’ surface. Compared with figures 5a to 5c, it is obviously shown that as time passes, the particles boundary becomes higher (white dash line), i.e., many particles rise and the loading of the following fast filtration becomes heavy.

In order to enunciate the geometric effects on the stability of the sludge blanket, the other three types (Type B, Type C, and Type D) of constructions of clarifiers were studied. Comparing with these figures it is shown that after 3600s’ operation, the lowest particle concentration at the effluent surface happens at the Type Bconstruction, the large inlet pipe.

As shown in Figure 6a, the increase of the caliber size of inlet pipes decelerate the fluid, thus, the intensity of backflow formed is smaller than the original structure (Fig. 3). The flow-up velocity outside of the reaction well is relatively small, and the volume fraction of particles carried upward by fluids is small as well. Finally, the quality of water discharge is favorable.

As shown in Figure 6b, the angle of the reaction well is a right angle, which is equivalent to an impinge flow when the suspension flows downward along the inside of the reaction well and approaches the bottom of the clarifier. The effective backflow cannot be formed, thus, the particles are prone to flow out of the reaction well to raise the concentration. In Figure 6c, it is expected that the suspension will be surrounded by the reversed reaction cover, thus preventing the particles from flowing out of the reaction well. However, a backflow zone is produced at the top of reaction well (as compared to the backflow zone appeared in the middle section of the reaction well in the original structure, Fig. 3). The upper backflow field at this position will hamper the backflow of the suspension under the reaction well so that it does not backflow into the reaction well, but leaving the reaction well. Thus, the concentration of the suspension outside the reaction well is raised.

Effluent upflow velocity (EUV) is proposed as an important parameter affecting clarifier performance (Narayanan et al., 2000). In this simulation work, effluent solid flux can be calculated in advance. The effluent solid flux is displayed in figure 7. For Type Aand Type B, effluent solid flux increase steadily; while for Type Cand Type D, effluent solid flux increase saw-toothed. This might due to the squeezed backflow in top of the reaction well causing unstable flow field outside the reaction well. Before 2500s, effluent solid flux is less than 0.05 kg/m2-s except for Type D. After 2500s, effluent solid flux is increasing dramatically. It is suggested that BWTP could drain sludge away per hour according to this simulation work. At almost all the 3000s’ operations, the effluent solid flux is the lowest in Type Bconstruction.


9. Nomenclature

CDdrag coefficient, -

Kinterphase momentum exchange coefficient, -

dsdiameter of particle, m

fcorrelation factor, -

gacceleration of gravity, m/s2m˙pqmass transfer from pto qphase, kg/m3-s

Ppressure, N/m2

pssolid pressure, N/m2

ResReynols number, -

Rwater/solid interaction, N/m2

Vvolume of each phase, m3

vr,sterminal velocity correlation, -

vwater/solid velocity, m/s

α volume fraction, -

μshear viscosity, kg/m-s

λbulk viscosity, kg/m-s

ρdensity, kg/m3

ρ^effective density, kg/m3

τSparticulate relaxation time, s

τ¯¯stress tensor, N/m2Subscripts

wwater phase

ssolid phase

Figure 1.

Geometry of the clarifier

Figure 2.

Meshes of the clarifier (a) whole clarifier (b) reaction well, inlet pipe, and draft tube (c) blades

Figure 3.

Velocity vector of water flows (inlet velocity = 0.3 m/s, impeller rotation speed = 0.3 rad/s)

Figure 4.

Contour of solid volume fraction on water surface (a)t =0 s (BWTP) (b)t= 600 s (BWTP) (c)t= 900 s (BWTP) (d)t= 1200 s (BWTP) (e)t= 2400 s (CFD) (f)t= 3000 s (CFD) (g)t= 3300 s (CFD) (h)t= 3600 s (CFD)

Figure 5.

Contour of solid volume fraction in clarifier (a)Type A,t= 1200 s (b)Type A,t= 2400 s (c)Type A,t= 3600 s (d)Type B,t= 1200 s (e)Type B,t= 2400 s (f)Type B,t= 3600 s (g)Type C,t= 1200 s (h)Type C,t= 2400 s (i)Type C,t= 3600 s (j)Type D,t= 1200 s (k)Type D,t= 2400 s (l)Type D,t= 3600 s

Figure 6.

Velocity vector of water flows (inlet velocity = 0.3 m/s, impeller rotation speedω= 0.3 rad/s). (a)Type B(b)Type C(c)Type D

Figure 7.

Effluent solid flux of four constructions of clarifiers

6. Conclusions

Blanket floc volumetric concentration is an important parameter in understanding the performance of sludge blanket clarifiers. 3D simulations of a clarifier using four different geometric constructions were studied in this work. From the simulation results of effluent solid flux, it is suggested that under the same daily throughput the large inlet pipe can reduce the flow velocity in the clarifier, hence reduce effluent solid flux and improve the quality of water. According to this simulation results, it is recommended that BWTP has to make large inlet pipe to obtain good quality water under the same daily throughput, or obtain large daily throughput under the same quality water.

The simulation results can only be interpreted qualitatively and not quantitatively since only hydrodynamic behavior is considered. Further work should concern the chemical based treatment, i.e., coagulation of flocs in clarifier. Nevertheless, the simulation results can show tendencies in the flow pattern due to changes in clarifier construction.


The authors would like to thank the National Science Council of the Republic of China, Taiwan, for financially supporting this research under Contract No. NSC 96-2221-032-020.

© 2010 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike-3.0 License, which permits use, distribution and reproduction for non-commercial purposes, provided the original is properly cited and derivative works building on this content are distributed under the same license.

How to cite and reference

Link to this chapter Copy to clipboard

Cite this chapter Copy to clipboard

Wen-Jie Yang, Syuan-Jhih Wu, Yu-Hsuan Li, Hung-Chi Liao, Chia-Yi Yang, Keng-Lin Shih and Rome-Ming Wu (January 1st 2010). Hydrodynamic Behavior of Flow in a Drinking Water Treatment Clarifier, Computational Fluid Dynamics, Hyoung Woo Oh, IntechOpen, DOI: 10.5772/7113. Available from:

chapter statistics

3412total chapter downloads

More statistics for editors and authors

Login to your personal dashboard for more detailed statistics on your publications.

Access personal reporting

Related Content

This Book

Next chapter

Applications of CFD in Natural Gas Processing and Transportation

By Majid Abedinzadegan Abdi, Esam Jassim, Mohammad Haghighi and Yuri Muzychka

Related Book

First chapter

Computational Fluid Dynamics

By Victor Udoewa and Vinod Kumar

We are IntechOpen, the world's leading publisher of Open Access books. Built by scientists, for scientists. Our readership spans scientists, professors, researchers, librarians, and students, as well as business professionals. We share our knowledge and peer-reveiwed research papers with libraries, scientific and engineering societies, and also work with corporate R&D departments and government entities.

More About Us