212 Q IWA Publishing 2009 Journal of Water Supply: Research and Technology—AQUA | 58.

3 | 2009

A CFD methodology for the design of rectangular

sedimentation tanks in potable water treatment plants
Ghawi A. Hadi and Jozef Kriš

ABSTRACT

Several modifications with different construction costs were proposed to enhance the Ghawi A. Hadi (corresponding author)
Jozef Kriš
performance in large rectangular sedimentation tanks. A two-dimensional fully mass conservative Department of Sanitary and Environmental
Engineering,
sedimentation model, based on modern computational fluid dynamics theory, was applied to Faculty of Civil Engineering,
evaluate proposed tank modifications of the existing and modified sedimentation tanks. The Slovak University of Technology,
Bratislava Radlinského 11,
usually unknown and difficult to be measured particle density is found by matching the 813 68 Bratislava 15,
Slovakia
theoretical to the easily measured experimental total settling efficiency. The proposed strategy is Tel.: +4212 5927 4615
Fax: +4212 5292 1184
computationally much more efficient than the corresponding strategies used for the simulation of E-mail: hadi.ghawi@stuba.sk
wastewater treatment. It is noteworthy that even small differences in the particle velocity can
cause large changes in the percent of settled particles; in this work, the overall solids removal
efficiency increased when using the modifications from 84.4 and 98.8%, leading to a reduction of
the effluent solids concentration of approximately 85%. The comparison of model predicted
results with the field data collected in tanks before and after modifications shows very good
agreement.
Key words | baffle, computational fluid dynamics, numerical modeling, sedimentation tank, water
treatment

NOMENCLATURE
conditions. CFD is able to yield a “virtual prototype” and
F fractal dimension (m)
a good example of this is in its application to the design of
Seff effective resistance coefficient (-)
compact and more efficient sedimentation tanks for
K effectiveness of particle settling (%)
traditional water treatment plants. CFD can capture the
ni mass fraction (-)
three-dimensional fluid flow inside a tank and thus help to
i particle size group (-)
minimize turbulence and optimize solids separation.
Many researchers have used Computational Fluid
Dynamics (CFD) simulations to describe water flow and
INTRODUCTION
solids removal in settling tanks for sewage water treatment.
The purpose of the sedimentation tank is to allow particles However, works in CFD modelling of rectangular sedimen-
sufficient time to settle out of the water. Over the last two tation tanks for potable water treatment have not been
decades, there has been a large growth in the application of found in the literature. Ghawi & Kris (2007a,b) investigated
Computational Fluid Dynamics (CFD) to the design of the effect of various solids and hydraulic loads on sedimen-
facilities. CFD has many advantages over traditional tation tank performance in water treatment plants. More-
modelling approaches as it is a low-cost, high-speed over, the physical characteristics of the flocs may not be such
technique for evaluating engineering systems that are significant parameters in the flow field of sedimentation
difficult to simulate in a laboratory or under field tanks for potable water, due to the much lower solids
doi: 10.2166/aqua.2009.027
 213 G. A. Hadi and J. Kriš | CFD methodology for design of rectangular sedimentation tanks Journal of Water Supply: Research and Technology—AQUA | 58.3 | 2009

concentrations and greater particle size distributions than manner, which may or may not have any coupling effect
those encountered in wastewater treatment. with fluid momentum. De Clercq & Vanrolleghem (2002)
The aim of this project was originally stated as to mentioned that the Lagrangian model should not be applied
improve the operation and performance of horizontal whenever the particle volume fraction exceeds 10 – 12%.
sedimentation tanks in the Holic water treatment plant The trajectories of individual particles through the
(WTP) which have been identified as operating poorly, by continuum fluid using the Lagrangian approach are
predicting the existing flow and flocculent concentration calculated in FLUENT by the Discrete Phase Model
distribution of the sedimentation tank by means of CFD (DPM). The particle mass loading in a sedimentation tank
techniques. for potable water treatment is typically small, and therefore
The objective of this work was to develop a new CFD it can be safely assumed that the presence of particles does
methodology for the analysis of sediment transport for not affect the flow field (one-way coupling).
multiple particle sizes in sedimentation tanks of potable In addition, the volume fraction of the particles in the
water treatment plants (the place is Slovakia in the Holic tank is of the order of 1024. The turbulent coagulation is
WTP), and to improve the operation and performance of well known to be proportional to this volume fraction, so it
horizontal sedimentation tanks in the Holic WTP which can be ignored under the present conditions. Also, the
have been identified as operating poorly, by analyzing and coagulation due to differential settling can be ignored due to
modeling the two-phase flow regime found in settling tanks the relatively low settling velocities resulting from the low
and comparing the model with experimental data. The two densities of the flocs.
phases present are water, the continuous medium, and floc The settling velocity hindering is insignificant for these
concentration, the dispersed phase. For this study the flow levels of solids volume fraction as can be shown by
field is considered to be isothermal, incompressible and employing the corresponding theories. Moreover, Lyn
without phase change. The CFD package FLUENT 6.3.26 et al. (1992), based on model observations, concluded
was used for the case study of the effect of adding energy that for conditions of relatively small particle concen-
dissipation baffles, a flocculation zone, perforated and non- trations in sedimentation tanks, the coalescence of the
perforated baffles, and modifications to the launders and on flocs does not affect the flow field and the effects on the
the efficiency of solids removal. concentration field and the removal efficiency may be of
secondary importance. Finally, decreases in the size range
relevant to primary separators do not suffer breakage
(Wilkinson et al. 2000).
MATERIALS AND METHODS The final system of particle conservation equations is a
linear one, so the superposition principle can be invoked to
Flow solver
estimate the total settling efficiency. The inlet particle size
The computational fluid dynamics code FLUENT 6.3.26 range is divided into classes with the medium size of each
has been used to carry out the simulations. Either a Eulerian class assumed as its characteristic. Then independent
or a Lagrangian approach can be adopted to model the simulations are conducted for monodisperse particles in
particulate phase. In the literature, Eulerian applications the feed using the individual particle sizes every time. The
are used for almost all diffusion-dominated problems, so overall settling efficiency can be found by adding appro-
strictly speaking they are only suitable for gas or ultrafine priately the efficiency for each particle size.
particle study (De Clercq & Vanrolleghem 2002). Due to Tracks are computed by integrating the drag, gravita-
their versatile capabilities, approaches based on the tional and inertial forces acting on particles in a Lagrangian
Lagrangian method have been applied extensively for frame of reference. The dispersion of particles due to
many two-phase flow problems. turbulence is modeled using a stochastic discrete-particle
In these approaches, the fluid is treated as a continuum approach. The trajectory equations for individual particles
and the discrete phase is treated in a natural Lagrangian are integrated using the instantaneous fluid velocity along
 214 G. A. Hadi and J. Kriš | CFD methodology for design of rectangular sedimentation tanks Journal of Water Supply: Research and Technology—AQUA | 58.3 | 2009

the particle path during the integration. By computing the closer than this distance. Also these measurements were
trajectory in this manner for a sufficient number of limited to an upper solids fraction of about 0.07, around 1.3
representative particles, the random effects of turbulence times the inlet solids fraction.
on particle dispersion may be accounted for.

Particle size distribution—experimental determination

Samples of incoming and effluent suspensions were taken

Sedimentation tank
and analyzed for particle size distribution using the laser
A full-scale rectangular sedimentation tank was investi- diffraction technique, with a Malvern Mastersizer 2000
gated, similar to those used in the potable water treatment analyzer. The location of sampling is very important and
plant of the city of Holic. There are two rectangular tanks depends on the goal of the study. The numerical model is
at the Holic water treatment plant. The plant receives raw compared with data obtained in a settling tank. In this
water from groundwater and its capacity is around 200 l/s. study, mean velocity and solids fraction concentration were
The employed processes include aeration, coagulation – measured at a number of stations along the length and
flocculation, sedimentation, filtration through sand and across the width of the tank.
chlorination. The volume of the sedimentation tank is The time taken to gather all the data for one inlet
500 m3 (Figure 1). The goal when designing this baffle is to condition was around 3 h, so no definitive time was given
achieve some head loss while keeping the velocity for the measurements at each station. Because of this, the
gradients through the ports equal to (or lower than) the numerical simulation was compared with the experimental
velocity gradient in the end of the flocculator, so as to not data at a time when the settled sludge layer in the simulation
break up the flocs. In each tank there is one effluent weir, was at approximately the same height as that found in the
which goes along the longitude direction and is distributed experiment. At this time, the flow field above the settled
over a range of 2 m from the side wall. The multiple sludge layer should be similar in both the experiment and
scrapers driven by a chain move toward the tank upstream the simulation.
hopper to collect the settled sludge into the sludge hopper
located at the beginning of the tank.
The influence of particle structure
Longitudinal velocity was determined by means of
submerged drifters, a method that involves measuring the The settling velocity of an impermeable spherical particle
time taken for a drifter at the measurement depth to traverse can be predicted from Stokes’ law. However, the aggregates
a given distance from which the velocity could be in the water are not only porous but it is well known that
calculated. Accuracy is given to ^ 2 mm/s. they have quite irregular shapes with spatially varying
Solids concentration was measured by an ultrasound porosity. The description of the aggregates as fractal objects
probe. Interference with concentration measurements is the best possible one-parameter description of their
occurred when the probe was placed within 0.2 m of the complex structure, so it has been extensively used in the
walls or sludge scraper; hence no readings were taken literature. The well-known fractal dimension, F, is a

Figure 1 | Schematic representation of the Holic sedimentation tank in 3D.

 215 G. A. Hadi and J. Kriš | CFD methodology for design of rectangular sedimentation tanks Journal of Water Supply: Research and Technology—AQUA | 58.3 | 2009

quantitative measure of how primary particles occupy the is the same for all positions; therefore, a 2D geometry can
floc interior space. But the settling velocity of the aggregate be used to properly simulate the general features of the
depends on its structure both through its effective density hydrodynamic processes in the tank. As a first step, a mesh
and its drag coefficient. These variables must be indepen- was generated across the sedimentation tank. The segre-
dently estimated for the aggregate shapes instead of the gated solution algorithm was selected. The k 2 1 turbu-
settling velocity because settling velocity cannot be directly lence model was used to account for turbulence, since this
entered into the CFD code. As regards the drag coefficient model is meant to describe better low Reynolds numbers
of fractal aggregates, ample information can be found in the flows such as the one inside our sedimentation tank. The
literature; from simulation of the flow inside reconstructed discretization schemes used were the simple for the
flocs using the FLUENT code to purely empirical relations. pressure, the PISO for the pressure – velocity coupling
According to Gmachowski (2005), the ratio of the resistance and the second-order upwind for the momentum, the
experienced by a floc to that of an equivalent solid sphere turbulence energy and the specific dissipation. Rodi (1993)
(Seff) can be expressed as follows: pointed out that, for real settling tanks, the walls can be
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi considered as being smooth due to the prevailing low
 ffi
F 2 velocities and the correspondingly large viscous layer.
Seff ¼ 1:56 2 1:728 2 2 0:288 ð1Þ
2 Consequently, the standard wall functions as proposed by
Launder & Spalding (1974) were used. The water-free
where Seff is the effective resistance coefficient and F is the
surface was modeled as a fixed surface; this plane of
fractal dimension. It is found that aggregates generated
symmetry was characterized by zero normal gradients for
in water treatment processes exhibit a fractal dimension
all variables.
ranging between 2.2– 2.6 (Lee et al. 1996), so the resistance
As a first step, the fluid mechanics problem was solved
coefficient, Seff, varies between 0.84 – 0.94. This estimation
in the absence of particles to find the steady state flow-field.
agrees also with the theoretical results for the drag
The converged solution was defined as the solution for
coefficient of fractal aggregates given by Vanni (2000) who
which the normalized residual for all variables was less
solved the problem in the limit of zero Reynolds number.
than 1026. In addition, the convergence was checked
Contrary to the drag coefficient, the effective density
from the outflow rate calculated at each iteration of the
cannot be estimated at all. Even if the structure (and the
run. The convergence was achieved when the flow rate
porosity) of the aggregate is known, the intrinsic density of
calculated to exit the tank no longer changed.
the primary particles is not known. As the settling velocity is
The settling tank was simulated for a specific set of
not so sensitive to Seff (the sensitivity with respect to
conditions used in the Holic treatment plants for which the
effective density is much larger, since the difference between
particle size distribution at the inlet and outlet and the total
effective and water density determines the settling velocity),
settling efficiency has been experimentally measured. The
the resistance coefficient was fixed at 0.91 and the apparent
23 inlet was specified as a plug flow of water at 0.035 m s21 and
density was then estimated as 1,065 kg m by requiring the
0.065 m s21, respectively, whereas the inlet turbulence
final computed settling effectiveness to coincide with the
intensity was set at 4.6%. The outlet was specified as a
measured settling effectiveness of 91%. This is a typical
constant pressure outlet with a turbulence intensity of 6.0%.
effective density value for the aggregates met in water
treatment applications (Deininger et al. 1998). The water flow rate was 0.06 m3 s21 and 0.08 m3 s21,
respectively.
For simulation purposes, the range of the suspended
solids was divided into 13 distinct classes of particles based
Simulation
on the discretization of the measured size distribution. The
To limit computational power requirements, the rectangu- number of classes was selected in order to combine the
lar settling tank was modelled in 2D. The major assump- solution accuracy with short computing time. Two other
tion in the development of the model is that the flow field numbers, 6 and 15, were tested. While the predictions
 216 G. A. Hadi and J. Kriš | CFD methodology for design of rectangular sedimentation tanks Journal of Water Supply: Research and Technology—AQUA | 58.3 | 2009

obtained using 6 classes of particles were found to be In this way, to predict the overall percent solids removal
different from those resulting from 13 classes, the difference efficiency one needs to know only the particle size
between the predictions made by the 13 and 15 classes distribution in the influent. With this knowledge and the
were insignificant. Therefore, a number of 13 classes were percentages Ki calculated from the 13 simulations, the
selected as suitable ones. Within each class the particle sedimentation efficiency could be calculated for any
diameter is assumed to be constant (Table 1). As can be seen different particle size distribution in the influent.
in Table 1, the range of particle size is narrower for classes
that are expected to have lower settling rates.
The procedure used to determine the overall settling
effectiveness, K, was based on the calculation of the percent RESULTS AND DISCUSSIONS
of solids settled for each particle size class, Ki:
Simulation of existing tank
PK
i¼1 ðni K i Þ
K¼ PK ð2Þ A rectangular sedimentation tank in Slovakia water treat-
i¼1 ni
ment plants was selected to demonstrate the response of
where K is the effectiveness of particle settling and ni is rectangular tanks to different internal geometries. This case
the mass fraction. The settling efficiency for each particle is based on the Holic settling tanks describe by Ghawi &
size class was calculated after the conclusion of the 13 Kris (2007c, 2008). This tank was selected because perfor-
different sedimentation simulations for the standard and mance data are available for model calibration and because
the modified tank, respectively. In each run, only one particle it represents a marginal performance case.
size class was taken into account; all injected particles As shown in Figure 2(a, b), the predicted hydraulic
were considered to have the same diameter corresponding regime typically consists of the upward inlet jet, the influent
to the so-called pivot particle size and assumed to be the density waterfall, a bottom density current and a strong
average of the lower and upper diameters of the class. surface reverse flow in the absence of proper baffling. For a
The effectiveness of particle settling is estimated as the case with a thick sludge blanket, the simulated velocity field
percentage of solids settled over the rate of solids showed that the bottom density current deflects upward
introduced from the inlet. while near the tank bottom a strong reverse sludge flow
appears. According to both the field observations and the
Table 1 | Classes of particles used to account for the total suspended solids in the STs
modeling of the existing process, each of the following
in Holic and Hrinova STs
reasons (or combination of them) may cause the Sedi-
Class Range of particle sizes (mm) Mass fraction mentation Tank (ST) problems, i.e. the flocculant solids
1 10 – 30 0.025 blowing out:
2 30 – 70 0.027 1. The location of the existing weir (distributed in a range of
3 70 – 90 0.039 1 m at the very downstream end of the ST) can cause
4 90 – 150 0.066 very strong upward currents, which could be one of the
5 150 – 190 0.095 major reasons that the flocculant solids were blowing out
6 190 – 210 0.115
around the effluent area.
7 210 – 290 0.126
2. The strong upward flow is not only related to the small
8 290 – 410 0.124
area the effluent flow passes through but also to the
9 410 – 490 0.113
rebound effect between the ST bottom density current
10 490 – 610 0.101
and the downstream wall. The “rebound” phenomenon
11 610 – 690 0.077
has been observed and reported by many operators as
12 690 – 820 0.057
well as field investigators, especially in ST with small
13 820 – 900 0.040
amounts of sludge inventory. A reasonable amount of
 217 G. A. Hadi and J. Kriš | CFD methodology for design of rectangular sedimentation tanks Journal of Water Supply: Research and Technology—AQUA | 58.3 | 2009

(a)
Strong surface reverse flow
y (m)

Upward bottom density current and sludge reverse flow

(b)

30
y (m)

130
230

x (m)

Figure 2 | Flow pattern (a) and sludge blanket (b) (in a vertical section along tank central axis) in existing tank.

sludge inventory can help dissipate the kinetic energy of All effluent launders are aligned with the tank longitudi-
the bottom density current. nal direction. The launders extend from the end wall to
the perforated baffle between B and C.
In the existing operation, the bottom density current
must be fairly strong due to the lack of proper baffling and The predicted flow and solids fields in the tank with
the shortage of sludge inventory in the tank. modifications 1, 2, 3 and 5 are presented in Figure 4(a, b).
The flow pattern shows that influent density waterfall and
surface reverse flow were significantly reduced by the three
Proposed tank modifications baffles. The flocculation baffle eliminates most of the
entrainment flow from the surface clear water layers into
The proposed modifications include: energy dissipation the influent density flow: thus both the surface return flow
baffles, a flocculation zone, perforated and non-perforated along the entire tank surface and the bottom density current
baffles, modifications to the launders and modifications of are substantially reduced (see Figures 2(a) and 4(a)).
the sludge withdrawal facility. In the original project, a total The distribution of the sludge blanket among the three
of five proposed modifications and combinations of them bays has been significantly changed by using two perforated
were tested. The following major modifications adopted in baffles (see Figures 2(b) and 4(b)). In the flocculation zone
final tank construction are presented as: relatively minor solids compression takes place in the local
† Modification 1: inlet flocculation baffle, the distance sludge blanket. The highest sludge blanket occurs in bay A
from tank influent to the baffle ¼ 6 m and the baffle due to the high resistance of the perforated baffle A/B. The
depth ¼ 2.5 m; lowest sludge blanket appears in bay C. The difference of
† Modification 2: a perforated baffle between bays A and B the sludge blanket level between bays B and C is relatively
with slot space of 54% of flow cross-sectional area; small due to the lower resistance of perforated baffle B/C.
† Modification 3: a perforated baffle between bays B and C The predicted flow pattern and solids field in Figure 4(a, b)
with slot space of 66% of flow cross-sectional area; show that solids spill gently over the baffle slots, at a lower
† Modification 4: a conventional baffle between bays A velocity and potential energy head than that in the
and B with baffle depth of 1.75 m below the surface; upstream bays.
† Modification 5: removing existing surrounding effluent Besides the evaluation of the impact of modifications on
weir and adding 4 – 5 new launders in bay C (Figure 3). the sedimentation tank performance, the sedimentation
 218 G. A. Hadi and J. Kriš | CFD methodology for design of rectangular sedimentation tanks Journal of Water Supply: Research and Technology—AQUA | 58.3 | 2009

Figure 3 | Tank with baffle and launder Modification 5.

tank model also helps us to implement cost-effective towards the bottom of the tank, turbulent diffusivity which
sedimentation tank modifications in this study. The model is independent of the particle size and the perfect sink
prediction shows that an optimized baffle design (Modifca- boundary condition). In practice it is expected that the
tions 1, 2 and 3), which needs about half a million dollars in settling efficiency decreases as the particle size decreases,
construction cost for all tanks, can give more than 85% going to a zero (or close to zero) value for Brownian
improvement of effluent quality. However, the construction particles. Although this inconsistency is not exhibited in the
costs were about half a million dollars for all tanks due to case studied here due to the relatively large particle sizes of
the necessity of a large amount steel in the new launder- the feed, it must be considered for the sake of completeness
supporting structure. The baffle modifications show the of the simulation procedure. The easiest way to accommo-
dominant improvement of tank hydraulics behavior, date the realistic behavior of a decreasing settling efficiency
especially in tanks running with a normal sludge blanket as particle size decreases is by incorporating a particle-size-
level. dependent trapping probability in the Lagrangian code. This
Figure 5 presents the predicted percents of solids settled probability should depend on the interparticle (particle
for different tested particle size classes. As can be inferred, deposits) interactions, turbulent diffusivity and gravity. As
the theoretical settling efficiency tends to nonzero (in fact, the particle size decreases the effect of gravity decreases,
relatively large) values as the particle size tends to zero. This leading the probability from unity to the inverse of the
is due to the combined effect of convection (fluid velocity stability ratio well known in the studies of small particle

(a) A Reduced reverse flow and bottom density current due to 3 baffles C

B
y (m)

Reduced density waterfall due to flucculation baffle

x (m)

(b)
Flocculation baffle A, perforated baffle A/B, B/C

A B C
y (m)

30
30 230

x (m)

Figure 4 | Flow pattern (a) and sludge blanket (b) (in a vertical section along tank central axis) in tank with Modifications 1, 2, 3 and 5.
 219 G. A. Hadi and J. Kriš | CFD methodology for design of rectangular sedimentation tanks Journal of Water Supply: Research and Technology—AQUA | 58.3 | 2009

100 Although the increase in the overall effectiveness seems

small it corresponds to an estimated reduction in the solids
90
exiting the tanks of approximately 520 kg d21 or to a
Setling efficiency (%)

80 reduction of about 85% of the solids in classes 1 – 13

70 Modified tank (Table 1) that exit the tank. These values are greater than
Standard tank those reported by other researchers. Huggins et al. (2005),
60
who tested a number of potential raceway design modifi-
50 cations, noticed that by adding baffles the overall percent

40
solids removal efficiency increased from 81.8 to 91.1%,
0 150 300 450 600 resulting in a reduction of the effluent solids of approxi-
Particle size µm
mately 51%. Crosby (1984) used an additional baffle at mid-
Figure 5 | Predicated percents of solids settled for each particle size class. radius extending from the floor upwards to mid-depth and
observed a reduction of 38% in effluent concentration.
deposition from turbulent flows (Kostoglou & Karabelas
1998). Nevertheless, despite the aforementioned improve-
ment, the present CFD model provides a good overall Validation of solids concentration profiles
description of the system behavior.
As far as the CFD model validity is concerned, Figure 6
The percentages presented in Figure 5 result in an
presents a comparison between the experimentally
overall settling efficiency of 84.4 and 98.8% for the existing
measured and the simulated values of the floc size
and modified tank, respectively. As can be seen, the model
distribution in the effluent of the existing tanks. Apparently,
predicts a highly distinct concentration for different classes
there is a good agreement between measured and predicted
of particle: lower removal rates for the smallest and higher
values.
removal rates for the heaviest particles.
Therefore, the improvement in the overall efficiency of
solids removal is only achieved by improving the settling
of particles with lower settling velocities (classes 1– 5 CONCLUSION
(Table 1)). This observation is similar to that obtained by This work deals with the development of a specialized
Huggins et al. (2005), who used a CFD model to evaluate strategy for the simulation of the treatment of potable water
the impact of potential raceway design modifications on the in sedimentation tanks. The strategy is based on the CFD
in-raceway settling of solids. code FLUENT and exploits several specific aspects of the
potable water application (low solids mass and volume
0.35
fraction) to derive a computational tool computationally
Experimental
0.30
much more efficient (due to the independent handling of
Simulated
0.25 flow fields and different particle classes) than the corre-
Mass fraction

sponding tools employed to simulate wastewater settling

0.20
tanks. The present code is modified based on data from a
0.15
real sedimentation tank.
0.10 The 2D fully mass conservative ST model is applied to
predict the tank performance in the existing STs and the STs
0.05
with proposed modifications at the Holic water treatment
0.00
0 50 100 150 200 250 300 plants. The existing Holic STs suffered from relatively poor
Particle size (µm) hydraulic performance which typically occurs in tanks
without proper baffling. The unfavorable hydraulic regime
Figure 6 | Simulated and experimental particle size distribution in the effluent of the
existing tank. includes strong turbulence, a high influent potential energy
 220 G. A. Hadi and J. Kriš | CFD methodology for design of rectangular sedimentation tanks Journal of Water Supply: Research and Technology—AQUA | 58.3 | 2009

and a strong density current due to excessive flow Hydraulic Engineering 2007 with Special Emphasis on the
entrainment. The baffle modifications can considerably Impact of Hydraulic Engineering Construction on the
Environment, Šibenik, Croatia, 4–9 September. Zagreb Faculty
reduce the strength of the density flow and increase the
of Civil Engineering, Zagreb, CD.
solids detention time in the tank; the effluent quality can be Hadi, G. A. & Kris, J. 2007b A numerical model of flow in
improved by more than 85% in some cases. Proper launder sedimentation tanks in Slovakia. In: Proc. 3rd International
modifications can be used to improve local flow pattern PhD Symposium in Engineering, Hungary University of
Pollack Mihály Faculty of Engineering Pécs, Hungary, 25 –26
near the effluent weir and to re-distribute the effluent flow
October. Pollack Periodica, 3(2), 59 –73.
along the tank longitudinal direction. Hadi, G. A. & Kris, J. 2007c Improved, modeling, simulation and
operational parameters of settling tank. In: Proc. 6th
International Conference of PhD Students, University of
Miskolc, Hungary, 12 –18 August. pp. 69 – 75.
ACKNOWLEDGEMENTS
Hadi, G. A. & Kris, J. 2008 Design and optimization of settling
Research Grant VEGA No. 1/3313/06 and project KEGA tanks performances in Slovakia. In: Proc. 20th National and
8th International Scientific and Technical Conference Water
3/5125/07 solved at the Department of Sanitary and
Supply and Water Quality Water, Poland, POZNAN-
Environmental Engineering, Faculty of Civil Engineering, GNIEZNO, 15 –18 June. Part 1, pp 425 –438.
Slovak University of Technology, Bratislava have supported Huggins, D. L., Piedrahita, R. H. & Rumsey, T. 2005 Use of
this paper. computational fluid dynamics (CFD) for aquaculture raceway
design to increase settling effectiveness. Aquacult. Eng. 33,
167 –180.
Kostoglou, M. & Karabelas, A. J. 1998 Comprehensive modeling of
REFERENCES precipitation and fouling in turbulent pipe flow. Indus. Eng.
Crosby, R. M. 1984 Hydraulic Characteristics of Activated Sludge- Chem. Res. 37, 1536 – 1550.
secondary Clarifiers EPA 600/2-84-131, NTIS no. PB-84- Launder, M. & Spalding, D. 1974 The numerical computation of
229665. US EPA, Municipal and Environmental Research turbulent flows. Comput. Meth. Appl. Mech. Eng. 3, 269 –289.
Laboratories, Cincinnatti, OH. Lee, D. J., Chen, G. W., Liao, Y. C. & Hsieh, C. C. 1996 On the
De Clercq, B. & Vanrolleghem, P. A. 2002 Computational fluid free-settling test for estimating activated sludge floc density.
dynamics in wastewater treatment. Med. Fac. Landb. Toegep. Water Res. 30, 541 –550.
Biol. Wet. 67, 15 –19. Lyn, D. A., Stamou, A. & Rodi, W. 1992 Density currents and shear
Deininger, A., Holthausen, E. & Wilderer, P. A. 1998 Velocity and induced flocculation in sedimentation tanks. J. Hydraul. Eng.
solids distribution in circular secondary clarifiers: full scale 118, 849– 867.
measurements and numerical modelling. Water Res. 32(10), Rodi, W. 1993 Turbulence Models and Their Application in
2951 –2958. Hydraulics—A State-of-the-art Review 3rd edition. IAHR
Gmachowski, L. 2005 Aggregate structure and hydrodynamics of Monograph Series, A.A. Balkema, Rotterdam.
aggregated systems. Colloid Surf. A: Physicochem. Eng. Vanni, M. 2000 Creeping flow over spherical permeable aggregates.
Aspects 255, 105– 110. Chem. Eng. Sci. 55, 685 –698.
Hadi, G. A. & Kris, J. 2007a Design and optimization of Wilkinson, D., Waldie, B., Mohamad Nor, M. I. & Lee, H. Y. 2000
sedimentation tank in Slovakia with CFD modeling. In: Proc. Baffle plate configurations to enhance separation in horizontal
10th International Symposium on Water Management and primary separators. J. Chem. Eng. 77(3), 221 –226.

First received 17 March 2008; accepted in revised form 12 September 2008. Available online February 2009