Presented at the WISA 2000 Biennial Conference, Sun City, South Africa, 28 May – 1 June 2000

To baffle or not to baffle – some baffled

solutions
A CFD investigation into the use of baffles in process tanks

J J van der Walt

Magalies Water
Private Bag X82327
Rustenburg, 0300
SOUTH AFRICA

SUMMARY
The reaction that takes place in a process tank is in a large way dependent on the internal hydraulics of the
tank, irrespective of the type of tank. In some tanks mixing is desirable in others plug flow is desirable. In
most tanks it is very hard to achieve either of the two. This is partially due to the fact that the area through
which the fluid enters the tank is small in relation to the tank cross sectional, complex geometries and
changes of fluid direction. Baffles can in some cases offer cost effective and simple solutions. This paper
demonstrates how CFD can be used to improve process tank performance by the introduction of baffles.

1 INTRODUCTION
In flow dynamics a baffle is a device used to alter the flow pattern in a tank. Baffle also means to be puzzled
or perplexed - the state of mind designers are in when the process tank does not perform the way it was
supposed to. Fortunately modifications can be made to the tank after construction to improve its
performance. One of the most cost effective ways to achieve this is by means of a baffle. Unfortunately the
selection of a baffle to suite a specific process tank is not a trivial exercise, especially with only the
conventional design methods at hand. The degrees of freedom in terms of size, vertical and horizontal
position and type of baffle makes it almost impossible to make the right choice the first time. The effect of a
baffles can normally only be seen after the baffle has been installed. Scale models can give some guidance,
but are often plagued by scale effects which means that what happens in the lab does not happen in the plant.
Very few guidelines exist for the optimal use of baffles in process tanks. This shortcoming often results in
baffles performing less effective for the following reasons:

• Conventional wisdom is often proved wrong due to a lack of understanding of the complex flow
patterns in most process tanks.
• Empirical data is not available to give quantitative guidelines.
• Experimental trial-and-error type investigations are time consuming and costly.
What can be done to improve the situation? Before this question is addressed, a short digression on the use of
baffles.
To baffle or not to baffle – WISA 2000, Sun City, South Africa

2. HOW BAFFLES ARE USED

The purpose of introducing baffles varies from tank to tank,
but can include the following:

• In some cases it can be used to force the flow in

such a way that the fluid moves more or less at the
same velocity everywhere in the tank. Examples of
such tanks would include tanks where contact
between water and a disinfectant is required such as
chlorine, ozone and UV.
• Some process tanks require deliberate velocity
gradients to enable collisions between suspended
particles as in the case of a hydraulic flocculator.
Baffles can be used to cause the flow to zig-zag or
to spiral (Figure 1a).
• In other cases baffles are introduced to generate
turbulence and improve the mixing of a chemical
with the bulk fluid such as a static mixer (Figure
1b).
• Most of the above cases use solid baffles to direct
the flow, but baffles can also be used to damp the
inlet jets, by dispersing it over a larger area.
Perforated baffles can be used for this purpose as in
the case of a sedimentation tank (Figure 1c). Figure 1a, 1b & 1c (from top to
bottom) –
Different baffle applications

Two types of baffles can therefore be distinguished:

• Solid baffles, used as obstructions perpendicular to the flow to change the direction of the flow.
• Perforated baffles, used to break up the main jet into a more uniform current over a larger area
without the deliberate intention to change the direction.

In most cases baffles are designed as part of the process tank, but in some cases the reality check requires the
introduction of additional baffles or modified baffles after the construction of the tank. Baffles are then used
to deliberately separate the inlet and outlet to prevent short circuiting or improve the level of plug flow or
mixing. Baffles can play an important part in the improvement of process tanks, and design tools to
determine baffle placement can assist designers. Not only can it potentially save on the initial capital cost of a
process tank, but it can also reduce the operating and maintenance costs. CFD provides a generic framework
by which baffles can be designed.

3. BAFFLE DYNAMICS
3.1 THE IMPORTANCE OF INLETS AND OUTLETS
One of the reasons why process tanks do not always behave according to designers’ intuition, is because of
the mechanisms at play near the inlets and outlets of process tanks. The inlet and outlet of a tank determines
to a large extent the internal flow patterns in the tank. It is not often realized, but mechanisms applicable to
inlets and outlets differ fundamentally.

• Inlets inject momentum (kinetic energy) into a process tank. This changes the mass inside the tank
as well as the velocity of the surrounding fluid. The momentum is a vector and can therefore vary in

2
To baffle or not to baffle – WISA 2000, Sun City, South Africa

terms of intensity as well as direction. The way in which the fluid enters a process tank often
determines to a large extent the way in which the complete tank behaves. In some cases where more
than one phase is present, as in the case of a sedimentation tank, not only kinetic energy is added to
inlet, but also potential energy. This is caused by the density differences between the fluid and the
suspended particles.
• An outlet on the other hand removes mass from the tank, but does not contribute to the removal of
momentum in a similar way as at the inlet. Also the effect of an outlet is only local and does not
normally affect internal hydraulics in the same way as the inlet. An outlet withdraws the fluid
closest to it at the time.
• The fact that the inlet injects momentum and the outlet not, does not make an outlet less important.
It is crucial to carefully select the position of the inlet in relation to the outlet to ensure that the
process tank performs effectively. In many cases the process tank outlet is dictated by the external
piping configuration and not the internal processes. This can have adverse effects on the internal
flow patterns as well as the process efficiency.

This following section demonstrates how CFD can be used to improve process tank efficiency by introducing
baffles. A number of process tanks that are typically used in the water industry are used as case studies to
demonstrate this. The examples include:

• Sedimentation tanks
• Flotation tanks
• Reservoirs and chlorine contact tanks
• Flocculators

The focus is on the insight gained by applying CFD and how it can be used to improve process tank
efficiency. The details of the CFD analysis are not discussed, but can be found in the literature.

3.2 CHLORINE CONTACT TANKS AND CLEAN WATER RESERVOIRS

The purpose of a chlorine contact tank is usually two fold, the one is to act as a operational balancing storage
device and the other is to achieve proper disinfection by forcing contact between the disinfectant and the
water. Clean water reservoirs were traditionally considered to have an emergency and balancing storage
function only, but they also act as large reactors in the distribution system. A number of studies demonstrated
how poor internal hydraulics can influence water quality in a distribution reservoir. It was also demonstrated
how the inlet and outlet configuration influences the internal hydraulics which in turn influences the water
quality in the reservoir. The design objective is therefore to achieve a uniform as possible flow pattern to
ensure that the fluid moves continuously and that sufficient contact between the water and the disinfectant is
achieved. This can only be achieved if the internal hydraulics of the tank is properly understood. Two
problems are usually associated with these types of tanks viz. short-circuiting of the inlet and outlet and the
presence of stagnant areas. Most effective ways these tanks can be improved is by:

• Repositioning the inlet

• Repositioning the outlet
• Introducing internal baffles to force the flow in a certain direction

The CFD analysis of this distribution storage reservoir are done in accordance with a previous study [1]. The
inlet and outlet arrangement was not changed, but three concentric baffles were added to reduce the extent of
the stagnant areas. Figure 2 clearly shows the reduction of total stagnant area (blue areas).

Hanoun et al. [2] also applied CFD to improve the contact efficiency in and existing chlorine contact tank
(clear well) by adding baffles. Marked improvements were achieved with minimal capital outlay. Ratzki et
al. [3] provided cost effective practical guidelines to construct baffles in existing and new reservoirs.

3.3 HYDRAULIC FLOCCULATORS

3
To baffle or not to baffle – WISA 2000, Sun City, South Africa

The primary purpose of a hydraulic flocculator is to generate velocity gradients to stimulate inter particle
collisions. Hydraulic flocculators are principally characterized by their volume (which determines the time of
flocculation) and the water level difference between inlet and outlet (which determines the average velocity
gradient). The underlying assumption is that the velocity gradient is proportional to the head loss. The local
detail inside the flocculator is however disregarded in the process as only the global volume and global
pressure drop is considered. The velocity gradients which are the primary force driving the flocculation
process cannot be optimized with conventional methods as the designer has too many degrees of freedom,
such as the average water depth, the number and spacing of baffles, the length of the gap at the baffle ends,
and the degree to which adjoining baffles overlap.

CFD allows the designer to evaluate the effect of changing the geometric ratios in an attempt to optimize the
baffle placement. The channel length for instance can be stretched. Or the gap at the end of a baffle can be
reduced. This enables the designer to calculate the optimal dimensions of a flocculator for a specific flow rate
as described elsewhere [4]. A typical flow pattern is shown in figure 6.

3.4 SEDIMENTATION TANKS AND SECONDARY CLARIFIERS

A sedimentation tank is a phase separating device that performs the following functions:

• Firstly it needs to leave a clarified layer of water near the surface in the vicinity of the overflow
landers.
• Secondly it needs to settle sludge in the vicinity of the sludge hopper to ensure effective hydraulic
removal. In some tanks mechanical removal is done in which case this requirement is less important.

The following challenges are faced to achieve an effective design:

• The presence of suspended matter at the inlet of a sedimentation tank causes density currents and
complicates the flow patterns of a sedimentation tank.
• Special precautions are required to manage the density currents that occur as a result of the suspended
solids. This is particularly prominent in the case of horizontal settling tanks where only a small
portion of the fluid in the tank will move from the inlet towards the outlet at a uniform velocity.
• Particular attention should be given to the inlet of the sedimentation tank. Baffles of various forms are
used to damp the flow at the inlet to restrict the density current. Fluid is also introduced as low as
possible to limit density currents.
• Particular attention should be given to the positioning and shaping of hoppers.
• Particular attention should be given to the positioning and shaping of overflow launders.

Figure 4a shows a sedimentation tanks where baffles were absent. It is clear that the unrestricted density
current caused the sludge to be washed out of the sludge hopper into the overflow launder.The introduction
of three baffles allowed the flow to enter the sedimentation tank horizontally and closer to the sludge hopper
reduced the effect of the density current. Overflow concentrations reduced significantly as shown in figure
4b.

Numerous studies have been conducted [5] to investigate the optimal placement of inlet baffles (reaction
baffles) in secondary settling tanks.

3.5 OTHER EXAMPLES

Flotation tanks also pose some interesting challenges to the designer. The ideal flotation tank should achieve
the following:

• Mix as much suspended solids as possible with the air bubbles

• Float as much suspended solids to the surface as possible

4
To baffle or not to baffle – WISA 2000, Sun City, South Africa

• Remove the scum layer from surface without breaking the scum layer
• Prevent settling of particles
• Poisition underflow in areas of lowest turbidity

The challenges faced are numerous and are still being discovered as our understanding of the flotation
process improves. Some of the challenges include:

• Positioning and height of inlet baffles

• Positioning of nozzles
• Positioning of clear water outlet pipes
• Optimal depth, width and length
• Optimal recirculation ratio
• Optimal air dosage, saturator pressure and many more

CFD cannot assist in all these aspects, but it can for instance assist in the propagation of the bubble plume
which is affected by the position of the air nozzle, the inlet baffle and the tank geometry. It can also assist in
some of the component details. A recent study achieved improved efficiency by repositioning and sizing of
the inlet baffle [9]. In another study, the inlet trumpets of a COCODAFF process were improved by placing
baffles in the throat of an inlet cone – a small change with dramatic effects [10].

CFD has also been used to investigate the use of baffles to improve process efficiency in static mixers [8],
ozone contactors [7] and UV reactors.

4 CONCLUSIONS
The case considered above demonstrates not only how CFD can be used as a design tool to improve process
tank efficiency, but it also demonstrates how simple baffles can make a difference. With tools such as CFD it
is not a question to baffle or not to baffle, but rather how to baffle.

5 REFERENCES
[1] VAN DER WALT, J.J. & HAARHOFF, J. (1999) Performance indicators for water storage tanks,
(in Water Industry Systems: modelling and optimisation applications ed. Savic, D.A & Walters,
G.A. Volume 1:539-550) Research Studies Press, Herfordshire, England
[2] HANNOUN I. A., BOULOS P.F. & LIST. J.E. (1998) Using hydraulic modeling to optimize contact
time. Journal AWWA, 90(8), 77-87.
[3] RATZKI, P.E. GINN, D. & JOSLYN, B. (1996) Clearwell and reservoir short circuiting: don’t be
baffles, here as some solutions, Proceedings AWWA annual conference, Toronto, Ontario.
[4] VAN DER WALT, J.J. & HAARHOFF, J. (2000) Is CFD an engineering tool or toy?, Proceeding
WISA Biennial Conference, Sun City, South Africa.
[5] KREBS, P. VISCHER, D. & GUJER, W. (1992) Improvement of secondary clarifiers efficiency by
porous walls, Water Science Technology, 26(5-6):1147-1156.
[6] CHATAIGNER, O. GATEL, D. BONNE, P. & CAVARD, J. (1999) Better disinfection through
optimised clearwells, Proceedings annual AWWA conference, Chicago, USA.
[7] HENRY, P.E. & BENNET, M.W. (1996) Sizing of ozone contactors: a case study for the Joseph
Jensen Plant, Proceedings AWWA annual conference, Toronto, Ontario.
[8] JONES, S.C. AMRTHARAJAH, A. & SANKAR, L.N. (1996) Static mixers in coagulation – A
CFD model, Proceedings AWWA annual conference, Toronto, Ontario.
[9] CROSSLEY, I.A. et al. (1999) Optimizing theDAF process utilizing two phase 3D CFD modeling,
Proceeding AWWA annual conference, Chicago, USA.
[10] PATTERSON CANDY, (1999) Personal correspondence with Paul Urban.

5
To baffle or not to baffle – WISA 2000, Sun City, South Africa

ANNEXURE A - FIGURES

Figure 2 – Top) large stagnant zones (blue areas) in reservoir with poor hydraulic
efficiency Bottom) improved hydraulic efficiency by introducing three concentric baffles

6
To baffle or not to baffle – WISA 2000, Sun City, South Africa

Figure 3 – Highly turbulent zones after baffle tips results in high G-values (red areas).
Low G-value areas (blue) are due to straightening of flow after each turn.

7
To baffle or not to baffle – WISA 2000, Sun City, South Africa

Version 2.31 Version 2.31

CASE: OLP_ALT1 CASE: OLP_ALT1
DATE: 20 Jul 1999 DATE: 20 Jul 1999
SCALAR 1 VELOCITY
SCALAR Magnitude
m/s
ITER 2620 ITER 2620
GMAX 2.557e+000 GMAX 3.332e-001
GMIN 1.287e+000 GMIN 2.227e-004
1.600e+000
1.580e+000 8.000e-002
1.560e+000 7.111e-002
1.540e+000 6.222e-002
1.520e+000 5.333e-002
1.500e+000 4.444e-002
1.480e+000 3.556e-002
1.460e+000 2.667e-002
1.440e+000 1.778e-002
1.420e+000 8.889e-003
1.400e+000 0.000e+000

Y Y

Z X Z X

Version 2.31 Version 2.31

CASE: OLP_ALT3 CASE: OLP_ALT3
DATE: 21 Jul 1999 DATE: 21 Jul 1999
SCALAR 1 VELOCITY
SCALAR Magnitude
m/s
ITER 2950 ITER 2950
GMAX 2.519e+000 GMAX 2.557e-001
GMIN 1.407e+000 GMIN 1.697e-004
2.000e+000
1.940e+000 5.000e-002
1.880e+000 4.444e-002
1.820e+000 3.889e-002
1.760e+000 3.333e-002
1.700e+000 2.778e-002
1.640e+000 2.222e-002
1.580e+000 1.667e-002
1.520e+000 1.111e-002
1.460e+000 5.556e-003
1.400e+000 0.000e+000

Y Y

Z X Z X

Figure 4 – The effect of introducing damping baffles. a) Top left and right shows the
solids concentration and velocity vectors where no baffles are present. b) Bottom left and
right shows the solids concentration and velocity vectors where three baffles are
introduced to damp the density current.

8
 To baffle or not to baffle – WISA 2000, Sun City, South Africa

Figure 5a - (left) CoCoDAFF inlet trumpet with baffle in throat. Note the uniform flow patterns in the larger area of the trumpet
where the water enters the CoCoDAFF.
Figure 5b - (right) CoCoDAFF inlet trumpet without baffle in throat. Note the skewness of the flow patterns at the
trumpet outlet [15].