Engineering Applications of Computational Fluid Mechanics Vol. 3, No. 4, pp.

562572 (2009)

CFD ANALYSIS OF INCOMPRESSIBLE TURBULENT SWIRLING

FLOW THROUGH ZANKER PLATE
Yehia A. El Drainy*, Khalid M. Saqr**+, Hossam S. Aly* and Mohammad Nazri Mohd. Jaafar*
* Department of Aeronautical Engineering
** High-Speed Reacting Flow Laboratory
Faculty of Mechanical Engineering, Universiti Teknologi Malaysia,
81200 Skudai, Johor Bahru - Malaysia
+ E-Mail: khaledsaqr@gmail.com (Corresponding Author)

ABSTRACT: Zanker plate is a widely used flow straightener to eliminate turbulence originating from pipe fittings
in experimental fluid flow applications. In this paper, steady state, incompressible, swirling turbulent flow through
Zanker plate has been studied. The solution and the analysis were carried out using finite volume CFD solver
FLUENT 6.2. Five turbulence models were used in the numerical investigation and their results were compared with
the pressure drop correlation of BS EN ISO 5167-2:2003. The turbulence models investigated here are the standard
k-, realizable k-, the Reynolds Stress Model (RSM), the large Eddy Simulation (LES), and the Detached Eddy
Simulation (DES). The results showed that the DES model gave the best agreement with the ISO pressure drop
correlation, therefore, the model was used further to predict streamline patterns, vortex formations and separated
flow regions. The effects of Zanker plate thickness and Reynolds number on the flow characteristics have been
investigated as well.
Keywords:

flow straightener, flow conditioning, Zanker plate, turbulent flow, turbulence modeling, CFD

avoid flowmeter calibration and measurement

errors entail a flow with 2 swirl angle and a ratio
of axial velocity at any point on a given pipe cross
section to the maximum axial velocity at the same
cross section is within 5% of the corresponding
ratio of fully developed flow as measured in the
same pipe after 100 pipe diameter length (ISO,
2003).
In order to achieve these conditions, great lengths
of straight pipes are required between the
flowmeter and the nearest upstream fitting to
decay the vortical motion of the turbulent swirling
flow; these lengths are typically equivalent to 600
times of the pipe diameter (Quazzane and
Benhadj, 2002). Large and costly upstream pipe
assemblies are required to comply with these flow
conditions in order to diminish the crucial
articulated flow metering errors. Alternatively,
turbulent swirling flow can be adapted to meet the
conditions
of
flowmeter
minimum-error
performance by the employment of flow
straighteners/conditioners upstream of the
flowmeter.
Flow straighteners are designed to remove swirl
from the flow, while flow conditioners attempt to
reorder the velocity profile in order to create fully
developed turbulent flow (Baker, 2000). Although
these devices have been used for decades, there
are several questions that need to be answered to

1. INTRODUCTION
Flow measurement is essential in almost all
industrial applications; from food industry to
petrochemical, and oil/gas processing. In some
applications, such as pharmaceutical plants, the
high accuracy of flow metering is unrelentingly
required in order to control the precise chemical
reactions and drug formation processes. On the
other hand, errors in flow measurement can result
in huge cost losses and inefficiency repercussions.
For instance, slight flow metering inaccuracies
can cause enormous monetary losses in the
transfer of oil and natural gas due to the great
volumes involved in these transfers.
The installation conditions of flowmeters have a
more pronounced effect on their performance than
construction. In fact, proper installation
conditions tend to provide fully developed and
swirl/pulsation-free flow upstream of the
flowmeter. This flow is very difficult to measure
without specific precautions in real practice,
because flowmeters are unavoidably installed in
assemblies located in plant pipe works
downstream of pipe fittings. The main
disadvantage of pipe fittings, in this sense, is the
generation of distorted velocity profiles
associated with varying degrees of swirl at the
inlet of flowmeter. Satisfactory stipulations to

Received: 23 Feb. 2009; Revised: 30 May 2009; Accepted: 12 Jun. 2009

562

Engineering Applications of Computational Fluid Mechanics Vol. 3, No. 4 (2009)

plate at different thicknesses. One purpose of this

study is to determine the optimum plate thickness
that would result in a swirl-free flow after the
shortest distance downstream of the straightener
plate. The CAD geometry and numerical model
are validated through a comparison between the
pressure drop resulting from five turbulent models
with the pressure drop calculated by standard
equations. Subsequently, the pressure drop
resulting from the selected turbulence model is
validated with the standard pressure drop at four
different flow rates. The validated turbulence
model is used afterwards to study the effect of
plate thickness on the flow dynamics.

define the geometry that would provide the

optimum performance. Such performance is
characterized by minimum swirl angle, more
stable and repetitive velocity profile as well as
minimum downstream pipe length.
The Zanker flow straightener has been included in
international standards such as ISO 5167 for
several decades. As described in ISO (2003), it is
named after its inventor who described it for the
first time in 1962 (Zanker, 1962). Zanker flow
straightener consists of a thin perforated plate
with various hole diameters designed to eliminate
swirl. The hole diameter is increasing towards the
plate center. The smaller holes are designed near
the plate edge in order to break down the swirl,
which intensifies near the boundary. Its swirlremoving efficiency increases with plate
thickness.
1.1

2. MATHEMATICAL FORMULATION
2.1

Zanker plate has five grades of holes, as described

in Fig. 1 and Table 1. The smaller holes are
concentrated near the plate edge because the
major concentration of eddies and swirl is near
the wall. On the other hand, this radial reduction
of hole diameter assists in stabilizing the velocity
distribution. The geometry of holes is expressed
in terms of plate diameter (i.e. pipe internal
diameter). In the present study, an outer diameter
of 50 mm is adopted for the Zanker plate.

Literature review

There have been several researches on the effect

of flow conditioner installation on the discharge
coefficient of flow metering devices, mostly on
the orifice flowmeter. In the early 1990s, the
effect of flow conditioner location on the orifice
meter discharge coefficient was investigated
(Morrison, DeOtte and Beam, 1992; Brennan
et al., 1991; Gajan et al., 1991). These studies, in
addition to the extensive research reported in
Quazzane (1995) and Laws and Quazzane (1994),
have found that developing flow causes higher
pressure drop across the orifice plate than fully
developed flow, hence, the function of flow
conditioner becomes essential.
Laws and Quazzane studied the effect of Zanker
plate thickness on the temporal mean velocity and
axial turbulent intensity (Laws and Quazzane,
1994). They included the honeycomb explained in
ISO 5167 standard in their experiments. It was
found that a relatively thin plate followed by the
recommended honeycomb section behaves
differently in well behaved and distorted flow
conditions, while a deep perforated plate alone
gives a more repeatable flow condition.
1.2

Geometry of Zanker plate

Study objectives and description

Fig. 1

In this paper, the turbulent swirling flow through

Zanker plate is analyzed through five turbulence
models in order to determine the most suitable
models to simulate such flow problems. Moreover,
the effect of plate thickness on the flow dynamics
is studied numerically. A detailed flow insight is
presented to explain the flow separation and
recovery within the plate holes as well as
visualizing the swirl removal effect of the Zanker

2.2

Description of Zanker flow straightener plate.

Governing equations

Two methods can be employed to render tractably

the Navier-Stokes equations so that the smallscale turbulent fluctuations do not have to be
directly simulated:
563

Engineering Applications of Computational Fluid Mechanics Vol. 3, No. 4 (2009)

Table 1 Details of Zanker plate geometry.
Hole grade

Number of holes

Pitch angle

45

o
o

0.25 D 0.002 5D

18

0.139D 0.001 D

0.56 D 0.005 6D

45o

0.136 5D 0.001D

0.75D 0.0075D

0.11D 0.001D

0.85D 0.008 5D

0.077D 0.001D

0.9D 0.009D

11

29

Standard plate thickness

0.12 D t p 0.15 D

Standard upstream pipe length

17 D L f

Standard downstream pipe length

7 .5 D L S L f 8 .5 D

Reynolds-averaging Navier-Stokes equations

(RANS)

Filtered Navier-Stokes equations (LES)

2.3.1 The standard k- model

The standard k- model is a two equation eddy
viscosity turbulence model (Launder and
Spalding, 1972). In this model, the eddy viscosity
is computed based on the turbulence kinetic
energy k, and the turbulence dissipation rate via:

In this study, incompressible, unsteady, and

isothermal flow is assumed. Based on these
assumptions, the governing equations are as
follows:

(ui ) = 0
x j

t = C

(1)

k2

(3)

Each of these two turbulence scales has its

transport equation. The turbulence kinetic energy
equation is derived from the exact momentum
equation by taking the trace of the Reynolds stress.
This equation can be expressed as:

(u j ) + (ui u j )
t
x i
u

i + u j p + g +
j
x j x i x j
x i i j

u u j ui
k
k

+ ui
=t i +
+

t
xi
x j xi x j xi

(2)
where u is the velocity component in x i direction,
is the density and p is the static pressure. In the
case of RANS modeling, the over-bar denotes
time averaging, while in the case of LES, the
over-bar denotes filtered variable. The stress
tensor i j is an unknown term representing
Reynolds stress tensor ( uiu j ) in the case of

t k

k xi
(4)

The dissipation rate equation, on the other hand,

is obtained using physical reasoning. The
equation is:

+ ui
t
xi
t
u u u
2

C 2
= C 1 t i + j i +
k x j xi x j xi xi
k
(5)

RANS and sub-grid scale stress tensor in the case

of LES. This additional term in the governing
equations is the result of averaging of time
dependent Navier-Stokes equations and needs to
be modeled in order to achieve closure. Details
about
transport
equation
and
closure
approximation are presented in reference (Fluent,
2005).
2.3

0.141D 0.001 D

c
e

x i

Pitch circle diameter

b
d

Hole diameter

The standard k- has five empirical constants C ,

k , , C1 and C 2 with values of 0.09, 1.0, 1.3,
1.44 and 1.92, respectively. These values were
obtained from experiments and computer
optimization. It is worthy noting that these values
are not universal and the k- model requires a
certain amount of fine tuning in order to obtain
correct results.

Turbulence modeling

In this section the five turbulence models used in

the numerical simulation are described briefly.
564

Engineering Applications of Computational Fluid Mechanics Vol. 3, No. 4 (2009)

no longer a constant and it is computed from:

2.3.2 The realizable k- model

Shih and coworkers have proposed a new k-
model in order to improve the performance of the
standard model (Shih et al., 1995). In the new
model, the eddy viscosity formulation satisfies the
mathematical constraint of the positivity of the
normal Reynolds stresses. This is achieved by
making C variable and sensitizing it to the mean
flow (i.e. mean deformation) and the turbulence
(i.e. k and ). Moreover, the realizable k- model
adopts a new transport equation for the
dissipation rate which is based on the dynamic
equation of the mean-square vorticity fluctuation.
The new dissipation rate equation is:

C =

A0 + AsU

(8)

where

A0 =4.0; U = S ij S ij + ij ij ,

8S ij S jk S ki
1

As = 6 cos arccos 6W , W =
S3
3

u
and the vorticity tensor ij = 1 ui j .
2 x j xi

2.3.3 The Reynolds stress model

+ ui
t
xi
= C 1 S +

xi

The Reynolds stress model involves calculation of

the individual Reynolds stresses uiu j using

t
2

C 2
k +
xi

(6)

differential transport equations (Gibson and

Launder, 1978) and (Launder, 1989). The
individual Reynolds stresses are then used to
obtain closure of the Reynolds-averaged
momentum equation, thus, avoiding the use of the
eddy viscosity approximation which was proven
to perform poorly in many types of flows. The
exact form of the Reynolds stress transport
equations can be derived by taking moments of
the exact momentum equation. This is a process
wherein the exact momentum equations are
multiplied by a fluctuating property, the product
then
being
Reynolds-averaged
stress.
Unfortunately, several of the terms in the exact
equation are unknown and modeling assumptions
are required in order to close the equations.
The exact transport equations for the transport of
the Reynolds stresses, is:

The model constants and C 2 have values of

1.2 and 1.9 respectively while C 1 is computed
from:

k
S

C 1 = max 0.43,
k

S + 5

(7)

where S = 2S ij S ij and S ij = 1 j + ui .

2 xi x j
The eddy viscosity is computed as in the standard
model via Eq. (3). However, in order to ensure the
positivity of the normal Reynolds stresses C is

uiuj +
uk uiuj =
uiuj uk + p( kj ui + ik uj +
t424
xk
xk
1
3

Local Time Derivative

142
4 43
4
Cij =Convection

1444442444443
DT ,ij Turbulent Diffusion

u
(uiuj ) uiuk j uj uk i g i uj + g j ui +

3
x
xk
xk 14442444
144
42k444
3 14 4442
44443
Gij Buoyancy Production

xk

DL , ij Molecular Diffusion

Pij Stress Production

u u
u uj
p i + j 2 i
2 k uj um ikm + uium jkm
x

14444244443

x
j
i
k
k
14
4244
3 14243 Fij Production by System Rotation

ij Pressure Strain

ij Dissipatio n

565

(9)

Engineering Applications of Computational Fluid Mechanics Vol. 3, No. 4 (2009)

reduced substantially in order to have a successful

LES computation. Moreover, as the Reynolds
number increases, the computational effort
becomes dependent on Reynolds number in a way
similar to DNS. Therefore, the idea of using a
RANS model near the wall will reduce the
computational cost significantly and make the
computation independent of Reynolds number.
The DES approach combines an unsteady RANS
version with a filtered version of the same model
to create two separate regions inside the flow
domain: one that is LES-based and another that is
close to the wall where the modeling is dominated
by the RANS-based approach (Shur et al., 1999).
The LES region is normally associated with the
high-Re core turbulent region where large
turbulence scales play a dominant role. In this
region, the DES model recovers the pure LES
model. Close to the wall, where viscous effects
prevail, the standard RANS model is recovered.
The switch between the two approaches is
accomplished by an automated criterion and
relieves the user from its specification.

Some terms in these exact equations, namely the

convection, molecular diffusion, stress production,
and production by system rotation do not require
any modeling. However, the turbulent diffusion,
buoyancy production, pressure strain, and
dissipation need to be modeled to close the
equations. Many approximations have been made
to close these equations (Fluent, 2005).
2.3.4 Large Eddy Simulation (LES)
LES can be thought of as an intermediate
technique between DNS and RANS. The former
involves resolving all the scales of turbulence
even the smallest ones. This is accomplished
using numerical schemes designed to minimize
the numerical dispersion and dissipation errors. In
the latter the effect of turbulence fluctuations
appears in the Reynolds stress term. In order to
close the system of equations, this unknown
Reynolds stress term is closed using closure
approximations and experimentally derived
constants. In LES, however, the large scales are
computed, and the small scales are modeled.
Since the small scales tend to be isotropic and
problem independent, there is a bigger potential to
develop universal models for these scales than for
RANS equations.
The governing equations employed for LES are
Eqs. (1) and (2). These equations are obtained by
filtering the time-dependent Navier-Stokes
equations. The filtering process effectively filters
out eddies whose scales are smaller than the filter
width or grid spacing used in the computations.
The resulting equations thus govern the dynamics
of large eddies.

3. SIMULATION DETAILS
3.1

CFD model description

For all the CFD models used in this study, except

for flow insight, a cell size of 1 mm3 is used in the
Zanker plate region, and a size function is applied
to the upstream and downstream flow regions.
The size function has a growth rate of 1.1% with
a maximum cell size of 4 mm3. The Zanker plate
region has 7776 cells, and the upstream and
downstream flow regions have 236042 and
245088 cells respectively. The diameter and
length of the cylindrical computational domain
are 0.05 m and 0.8 m respectively. The constant
cell size in the Zanker plate flow volume is shown
in Fig. 2-a, while the effect of size function is
very obvious on cell size in Fig. 2-b. The
boundary conditions used with the incompressible
CFD model are constant velocity inlet and
ambient pressure outlet. All walls are assumed to
have no-slip boundary conditions.

2.3.5 Detached Eddy Simulation (DES)

LES seems to be a reliable and promising tool for
predicting turbulent flows, especially those that
are dominated by large scale anisotropic vortical
structures. However, computational expenses
prevent LES from being widely adopted by the
industrial community. These computational
expenses lies in the requirement of a very fine
grid and a small time step in the order of the
smallest resolved eddies.
The strategy of coupling LES with statistical
RANS aims to reduce the computational cost of
the numerical simulation, thus making LES
computations more feasible for a wider range of
industrial complex flows. The inadequacy of
implementing LES for wall induced shear flows is
another motivation for LES/RANS coupling.
Closed to the walls, the flow is dominated by the
small eddies. This requires the grid size to be

3.2

CFD model validation

In order to validate the proposed CFD model, and

select the optimum turbulence model to
implement in the presented study, all the
abovementioned models have been run. The inlet
axial and swirl velocity components were
10.85 m/s and 4.1 m/s, respectively. It should be
noted that at the inlet section of the computational
domain, the axial and tangential velocities
566

Engineering Applications of Computational Fluid Mechanics Vol. 3, No. 4 (2009)

(a)

Fig. 2

(b)
3

(a) Meshing of the Zanker flow volume at 1 mm . (b) Meshing of the total flow volume with the size function
applied.

Table 2 Pressure drop across the Zanker plate for five different turbulence models.
Turbulence model

CFD computed P (Pa)

Deviation from standard P

1st order

315.5

49.7 %

quick

290.1

37.9 %

k- Realizable

275.3

30.8 %

RSM

288.6

37.2 %

LES

230.1

9.4 %

DES

220.5

4.8 %

Standard k-

of the DES over the LES is justified by the

inadequacy of implementing LES for wall
induced shear flows which are dominated by the
small eddies. This requires the grid size to be
reduced substantially in order to have a successful
LES computation. On the other hand, DES uses a
RANS based model to calculate the turbulent
stresses near the wall, thus maintaining the grid
requirements to a reasonable level. It is believed
that LES would perform better than DES if the
generated grid allowed for an adequate treatment
of the turbulent eddies near the wall.
Subsequently pressure drop was computed across
the Zanker flow straightener at four different axial
velocities with a constant inlet swirl angle of 20o.
The purpose of this stage of validation is to
confirm the suitability of DES model to solve the
present problem. The computed pressure drop is
plotted against the average inlet axial velocity in
Fig. 3. The maximum deviation between standard
and computed pressure drop is 7.4% at an inlet
velocity of 21.75 m/s.

maintain constant values with respect to the radial

direction up to the wall. The resultant pressure
drop across the Zanker plate was compared for
the five turbulence models with the value
calculated using the standard equation in (ISO,
2003), which calculates the pressure drop as a
function of the dynamic head:
1
P = K v 2
2

(10)

where K is the coefficient of discharge, estimated

at a value of 3 (ISO, 2003). The standard pressure
drop was calculated using Eq. (10), and it was
found to be 210.4 Pa. In contrast, the pressure
drop was computed from the CFD analysis, using
different turbulence models, at distances 1D and
2D upstream and downstream the Zanker plate,
respectively. Table 2 gives the validation results.
The results show that the three RANS models
performed poorly in this type of flow, with the
realizable k- performing slightly better than the
RSM and the standard k- models. The DES
obtained the best result among all the turbulence
models, while the error of LES is approximately
twice that of the DES. The unexpected superiority
567

Engineering Applications of Computational Fluid Mechanics Vol. 3, No. 4 (2009)

4. RESULTS AND DISCUSSION

4.1

Fig. 3

In order to retrieve higher resolution results for

flow insight purpose, a hexahedral cell volume of
0.125 mm3 was implemented in the Zanker plate
flow volume. A user defined mesh adoption
function was used to increase the cell size in both
directions by rate of 1.02% until a maximum size
of 1.0 mm3.
A general view of the flow field is shown in
Fig. 4-a where the velocity vectors are plotted in a
45o inclined longitudinal plane. It is observed that
the flow-aligned recirculation zones, represented
in Fig. 4-b, are formed directly downstream the
Zanker plate in between the main jets issued from
the plate holes. Subsequent to separation, the flow
begins to reattach to the pipe walls at stagnation
points indicated by A and E in Fig. 4-a. On the
other hand, the flow to flow attachment occurs at
the near to core stagnation points indicated by B,
C, and D. Figure 4-c shows the path lines of the
flow through Zanker plate. It can be seen that
swirl motion is eliminated due to the restrictions
provided by the plate holes in Z direction.
Furthermore, the downstream circulation zones
assist in eliminating the swirl motion by limiting
the flow motion in the tangential direction.
In order to visualize the flow conditioning effect
of the Zanker plate, the progress of axial velocity
profile is captured at incremental diameter
distances after the Zanker plate, as in Fig. 5. The
velocity profileplotted on vertical diametrical
lines, is observed to have gradual stabilization
within a distance equal to 5D. At distance
exceeding 5D, the axial velocity profile is
approximately identical. The velocity contours at
three cross-sections at different distances from the
exit section of the Zanker plate are also shown in
Fig. 6. This shows the flow conditioning effect of
the Zanker plate. The fully developed axial
velocity profile has an acceptable qualitative
agreement with comparable profiles discussed in
Xiong, Kalkuhler and Merzkirch (2003),
Spearman, Sattary and Reader-Harris (1996), and
Schluter and Merzkirch (1996). However, due to
the dissimilarity of geometry and boundary
conditions between the case in hand and that in
latter literature, a quantitative agreement cannot
be reached.

Comparison between standard and computed

pressure drop across Zanker plate at four
different axial velocities.

(a)

(b)

(c)

Fig. 4

Flow insight

(a) Flow velocity vectors across the Zanker

plate (b) Recirculation zones (c) Flow
pathlines.

4.2

Significance of plate thickness variation

The effect of plate thickness on eliminating swirl

from the flow field is explained in terms of two
parameters: tangential velocity and swirl angle.
568

Engineering Applications of Computational Fluid Mechanics Vol. 3, No. 4 (2009)

Fig. 5

(1D)

(2D)

(3D)

(4D)

(5D)

(6D)

(7D)

(8D)

Progress of velocity profiles, at incremental diameter distances, downstream of the Zanker plate.

569

Engineering Applications of Computational Fluid Mechanics Vol. 3, No. 4 (2009)

(a)

Fig. 6

(b)

(C)

Planar velocity contours at distances (a) 1D, (b) 3D and (c) 5D from the exit section of the Zanker plate.

(a)

(b)

(c)

Fig. 7

(d)

(a) vp at pipe inlet section,

(b) vp at pipe exit section at
tp = 1 mm, (c) vp at pipe exit
section at tp = 3 mm, (d) vp at
pipe exit section at tp =6 mm
and (e) vp at pipe exit section
at tp = 12 mm.

(e)

maximum plate thickness, the pressure drop was

found to be 250.24 Pa, compared to 210.4 Pa at
the standard thickness. The reason behind is that
pressure drop is highly dependent on the
boundary layer behavior, and in the present case,
the plate thickness, even at maximum value, does
not allow the boundary layer to uniformly
develop in the plate holes. As for the tangential
velocity, the smaller diameter holes near the
boundary flow region modify the tangential
velocity, which has higher values near the walls,
and turn it into axial velocity. That is why the
axial velocity increases at the exit plane.

Figure 7 illustrates the planar velocity vectors at

inlet and outlet sections for Zanker flow
straightener with four different thicknesses. The
swirl angle at the inlet is constant at 20o.
The effect of plate thickness on the swirl angle
and tangential velocity is illustrated in Fig. 8. The
tangential velocity is more affected by the plate
thickness than swirl angle. However, both
decrease significantly when the plate thickness
increases. Compared to the flow pattern, the
increase in plate thickness was found to increase
the pressure drop across the Zanker plate by only
19%, which is relatively insignificant. At the
570

Engineering Applications of Computational Fluid Mechanics Vol. 3, No. 4 (2009)

4.0

10.0

8.0
3.0

6.0
2.0
4.0

Swirl Angle (degrees)

Average tangential velocity (m/s)

Average tangential velocity (m/s)

Swirl angle (degrees)

1.0

Fig. 8

2.0

0.0

0.0
0.0

2.0

4.0

6.0

8.0

10.0

Average tangential velocity

and swirl angle at pipe
outlet for different plate
thicknesses.

12.0

Plate thickness (mm)

5. CONCLUSIONS

NOMENCLATURE

A 3-D numerical investigation of the

incompressible isothermal flow through the
Zanker flow straightener plate has been conducted.
The CFD model was validated through
comparison with the ISO pressure drop
correlation. The comparison showed that DES
turbulence model yielded better agreement than
other examined models for the available
computing resources. A detailed description of
the flow field through the plate has been
presented. A flow conditioning effect has been
revealed through the flow insight. The plate
thickness was found to have a fundamental effect
on the flow quality in terms of swirl removal.
Plate thickness was proved to be inversely
proportional to the swirl angle and tangential
velocity component. The average tangential
velocity at the outlet plane decreased from
1.8 m/s to 0.02 m/s when the plate thickness
increased from 1 mm to 12 mm, respectively, at
the same inlet conditions. As for the swirl angle,
it decreased from 8.9o to 0.08o due to the same
increase in plate thickness. The study suggests
that Zanker plate can be employed as both flow
straightener and conditioner with optimized plate
thickness.

a to e

LS

Lf

tp
v
vP

Hole grades of Zanker plate

The distance between the downstream
surface of Zanker plate and the flow
meter (m)
The distance between the upstream
surface of Zanker plate and nearest
fitting (m)
Pressure drop (Pa)
Density (kg/m3)
Outer diameter of Zanker plate (m)
Zanker plate thickness (m)
Average axial velocity (m/s)
Inplane velocity (m/s)
Swirl angle

REFERENCES
1. Baker RC (2000). Flow Measurement
Handbook: Industrial Designs, Operating
Principles, Performance and Applications,
Cambridge University Press, 2730.
2. Brennan JA, Sindt CF, Lewis MA, Scott JL
(1991). Choosing Flow Conditioners and
Their
Location
for
Orifice
Flow
Measurements. Flow Measurements and
Instrumentations 2(1):1420.
3. Fluent Inc. (2005). FLUENT 6.2 Users
Guide, 2005-01-04.
4. Gajan P, Hebrard P, Millan P, Giovanni A
(1991). Basic Study of Flow Metering Fluids
in Pipes Containing an Orifice Plate, Gas
Research Institute Report, No. 5086-27-1412.

ACKNOWLEDGEMENTS
Partial support was provided by MOSTI (Ministry
of Science, Technology and Innovation,
MALAYSIA) under Science Fund grant 79253.

571

Engineering Applications of Computational Fluid Mechanics Vol. 3, No. 4 (2009)

Measurement and Instrumentation 14(6):249

260.
18. Zanker KJ (1962). The Development of a
Flow Straightener for Use with Orifice-Plate
Flowmeters in Disturbed Flow. Flow
Measurement in Closed Conduits, NEL, 398
415.

5. Gibson MM, Launder BE (1978). Ground

Effects on Pressure Fluctuations in the
Atmospheric Boundary Layer. J. Fluid Mech.
(86):491511.
6. ISO (2003). ISO 5167:2003 Measurement of
Fluid Flow by Means of Pressure Differential
Devices Inserted in Circular Cross-Section
Conduits Running Full.
7. Launder BE (1989). Second-Moment
Closure: Present... and Future? International
Journal of Heat and Fluid Flow 10(4):282
300.
8. Launder BE, Spalding DB (1972). Lectures in
Mathematical Models of Turbulence.
Academic Press, London, England.
9. Laws EM, Quazzane AK (1994). Compact
Installation for Differential Flow Meters.
Flow Measurements and Instrumentations
5(2):7985.
10. Morrison GL, DeOtte RE, Beam EJ (1992).
Installation Effects upon Orifice Flow Meters.
71st Annual Gas Processors Association
Convention, CA, U.S.A.
11. Quazzane AK (1995). Experimental and
Computational Investigation of Flow through
Flow Conditioning Devices. PhD thesis,
Salford University, Salford.
12. Quazzane AK, Benhadj R (2002). Flow
Conditioners Design and Their Effects in
Reducing Flow Metering Errors. Sensor
Review 22(3):223231.
13. Schluter Th, Merzkirch W (1996). PIV
Measurements of the Time-Averaged Flow
Velocity Downstream of Flow Conditioners
in a Pipeline. Flow Measurement and
Instrumentation 7(34):173179.
14. Shih TH, Liou WW, Shabbir A, Yang Z,
Zhu J (1995). A New k- Eddy-Viscosity
Model for High Reynolds Number Turbulent
Flows - Model Development and Validation.
Computers and Fluids 24(3):227238.
15. Shur M, Spalart PR, Strelets M, Travin A
(1999). Detached-Eddy Simulation of an
Airfoil at High Angle of Attack. 4th
International Symposium on Engineering
Turbulence Modeling and Experiments,
Corsica, France.
16. Spearman EP, Sattary JA, Reader-Harris MJ
(1996). Comparison of Velocity and
Turbulence
Profiles
Downstream
of
Perforated
Plate
Flow
Conditioners.
Flow Measurement and Instrumentation
7(34):181199.
17. Xiong W, Kalkuhler K, Merzkirch W (2003).
Velocity and Turbulence Measurements
Downstream of Flow Conditioners. Flow
572