Content uploaded by Myung Kyoon Chung
Author content
All content in this area was uploaded by Myung Kyoon Chung on Jul 03, 2017
Content may be subject to copyright.
1. Introduction
Since galvanized steel strip has very good properties for
corrosion-resistance, paintability, workability and weldabil-
ity, it has been widely used over a variety of industrial fields
such as buildings, bridges, automobiles, and electronic
equipments. In general, the galvanized steel strip is pro-
duced through a continuous hot-dip galvanizing process.
The thickness of the adhered zinc film is controlled by a gas
wiping process. Due to its good productivity and ease of
thickness control, the gas wiping process that was commer-
cialized in the mid-20th century has been used in most of
the continuous hot-dip galvanizing system.1) Usual working
fluid for the wiping process is nitrogen gas. But the gas
wiping process causes two technically serious problems;
namely the edge overcoating (EOC) and check mark.2)
Frequently, the zinc film near the steel strip edge is 1.4–1.8
times thicker than that in the steel strip center region. This
EOC is a chronic problem that causes troubles in coiling
and poor flatness of the steel strip after uncoiling. Ac-
cordingly, a number of researches have been carried out to
reduce the EOC.3–6)
Recently Kim et al.7) investigated the mechanism of the
EOC through a numerical analysis. They found that the
two-dimensional opposing jets that are not blocked by the
steel strip establish alternating vortices in the region outside
the steel strip (z0 in Fig. 1(a)), and that due to such alter-
nating vortices the impinging pressure of the wiping gas
near the steel strip edge becomes significantly lower than
that in the central region of the steel strip. They showed that
this drop in impinging pressure is the main cause of the
EOC. In order to prevent the collision of the two opposing
jets in the region outside the steel strip in z0, an edge baf-
fle plate parallel to the steel strip has often been used like
the one in the experiment of Park et al.2) As the edge baffle
plate and the steel strip are getting closer, the EOC was ob-
served to be reduced significantly. However, technically it is
very difficult to maintain the small gap between the edge
baffle plate and the steel strip. Due to the inevitable vibra-
tion of the up-lifting steel strip, the edge baffle plate and the
steel strip clash each other. This causes unwanted disper-
sion of molten zinc toward the exit of the air knife nozzle
that eventually blocks the nozzle exit.
Gilchrist et al.8) demonstrated that a jet, discharged tan-
gentially to a cylindrical surface, flows along the surface as
shown in Fig. 2. The stream of fluid emerging from a noz-
zle tends to attach the nearby curved surface due to the
Coanda effect. In the present study, this effect is used to
prevent the direct collision of the two opposing jets at
180°C. As shown in Fig. 1(a), the flow direction of the wip-
ing gas is deflected by a cylinder with small diameter in-
stalled tangentially to the exit of the lower lip of the two-di-
mensional jet, which is placed in the region z0 where the
two opposing jets are not blocked by the steel strip.
ISIJ International, Vol. 46 (2006), No. 4, pp. 573–578
573 ©2006 ISIJ
A Noble Gas Wiping System to Prevent the Edge Overcoating in
Continuous Hot-dip Galvanizing
Ki Jang AHN and Myung Kyoon CHUNG
Department of Mechanical Engineering, Korea Advanced Institute of Science and Technology (KAIST), 373-1, Kusong-dong,
Yusong-gu, Daejon 305-701, Korea. E-mail: gijang.ahn@kaist.ac.kr
(Received on December 12, 2005; accepted on January 23, 2006)
A noble method is proposed to prevent the edge overcoating (EOC) that may develop near the edge of
the steel strip in the gas wiping process of continuous hot-dip galvanizing. In our past study (ISIJ
International, Vol. 27 (2003), No. 10, pp. 1495–1501), it was found that the EOC is caused by the alternating
vortices which are generated by the collision of two opposing jets in the region outside the steel strip. In
the present study, the flow field around the gas wiping system has been analyzed numerically and it was
found that when the two opposing jets collide at an angle much less than 180°, the alternating vortices dis-
appear and the impinging pressure on the steel strip surface becomes nearly uniform. In order to deflect
both jets downward by a certain angle, a cylinder with small diameter is installed tangentially to each exit of
the lower lips of the two-dimensional opposing jets. The three dimensional flow field with the proposed de-
vice is analyzed numerically by using the commercial CFD software, STAR-CD. And the coating thickness is
calculated by solving the boundary layer momentum equation with an integral analysis method. In order to
compare the present noble method with the conventional edge baffle plate method to prevent the EOC, the
flow field with edge baffle plates is also calculated. The calculation results show that the tangentially in-
stalled cylinder at the lower lip of the jet exit is significantly more effective than the edge baffle plate.
KEY WORDS: continuous hot-dip galvanizing; edge overcoating; impinging jet; gas wiping.
In the present study, the 3-D unsteady compressible tur-
bulent flow field with the proposed configuration with a
pair of cylinders in the continuous hot-dip galvanizing sys-
tem is simulated by using a commercial CFD software, to
find out how the cylinder affects the flow behavior of the
wiping gas. The simulation results for the impinging pres-
sure and the shear stress on the steel strip surface are used
to calculate the thickness of the adhered zinc film by the
noble integral analysis method of the boundary layer mo-
mentum equation.7) In addition, the flow field generated by
the edge baffle plate is also analyzed to compare its perfor-
mance for reduction of the EOC with that of the cylinder
installed at the lower lip of the air knife nozzle.
2. Numerical Analysis
2.1. Numerical Analysis of 3-D Unsteady Compressible
Flow
In the continuous hot-dip galvanizing process, a heat-
treated steel strip is passed through a molten zinc bath and
the molten zinc is adhered on the steel strip surfaces.
Usually, the thickness of the adhered zinc film is nearly 10
times thicker than desired one, and the excessive zinc is re-
moved by nitrogen gas jets. This wiping gas jet system is
often called the air knife system. Usually, the Mach number
of the nitrogen gas flow is in a range of 0.3–0.6. Therefore,
in addition to the continuity and momentum equations, the
equation of state and the energy equation are needed. Large
Eddy Simulation (LES) is used to solve the turbulent equa-
tions, of which forms can be found in Kim et al.7) The nu-
merical solutions of these equations for 3-D unsteady com-
pressible turbulent flow are obtained by using a commercial
CFD software. Reliability and numerical accuracy of LES
for complex turbulent flows have been well established
through many comparative studies between LES and exper-
imental measurements. Recently, for example, even a sepa-
rated turbulent flow over a forward-backward facing step
has been successfully simulated with LES by Addad et al.9)
They also demonstrated in a different study that the same
LES could be applied to accurately simulate a buoyant
wall-jet.10) Since the present flow geometry is a combina-
tion of separated flow and curved wall-jet, it is expected
that LES is capable of producing reliable computational re-
sults for the flow geometry of our concern.
The calculation domain is shown in Fig. 1(b). The steel
strip is lifted up at the middle of the two air knives. The
width of the steel strip is nearly half of the two-dimensional
nitrogen jet width. At each of the two opposing jets, a cylin-
der of small diameter is installed tangentially to the lower
lip of the jet exit so that the jet flow is deflected downward
in the region outside the steel strip in z0. The boundary
conditions and the flow Mach number in the present study
are given the same values as described in Kim et al.7)
2.2. Numerical Computation of the Coating Thickness
Nitrogen gas jet injected normally onto the surface of the
moving steel strip is divided into two different wall jets as
depicted in Fig. 3. In the upper region (y0), since the
steel strip moves upward, the molten zinc moves in the
same direction at both the wall boundary and the free
boundary. On the other hand, in the lower region (y0), the
molten zinc moves upward near the steel strip surface while
it moves downward near the free boundary. This feature has
been verified by the experiment of Ellen et al.11) Due to the
velocity profile of the nitrogen gas flow near the interface
between the molten zinc and the nitrogen gas, the shear
stress in the nitrogen gas wall jet exerts a wiping force in
the upper region while that in the lower region contributes
to augmenting the film thickness. This difference makes the
film thickness of the upper region thinner than that of the
lower region as shown in Fig. 3. The governing equations
determining the thickness of the molten zinc film are the
ISIJ International, Vol. 46 (2006), No. 4
©2006 ISIJ 574
Fig. 1. Schematic diagram of air knife system and calculation
domain.
Fig. 2. Deflection of air knife jet due to the curved wall surface.
continuity and Navier-Stokes equations. The noble integral
analysis method proposed by Kim et al.7) is also used for
calculation of the film thickness in the present study.
3. Computation Results and Discussion
3.1. Deflection of the Air Knife Jet
In order to find the effect of the cylinder, that is tangen-
tially attached at the lower lip of a jet exit, on the deflection
of the jet flow, the two-dimensional flow field in the im-
pinging region between two opposing jets in the absence of
a steel strip is calculated for different cylinder diameters.
For comparison, the flow field without installing the cylin-
der was calculated first. When there is no cylinder installed
at the lower lip of the air knife nozzle, the jet alternates its
flow direction cyclically upward and downward as shown in
Fig. 4 where T represents the period of such a cyclic
change. However, when a cylinder of small diameter larger
than 3 mm is installed in place, Fig. 5 shows that the alter-
nating vortices do not appear. In this case, a steady flow
field is established, and the jet flow direction does not
change. It can be seen that the deflection angle of the jet de-
pends on the cylinder diameter. When d5mm, the jet flow
deflects more than 90°.
ISIJ International, Vol. 46 (2006), No. 4
575 ©2006 ISIJ
Fig. 5. Effect of cylinder diameter on the deflection of air knife jet (unit in m/s).
Fig. 4. Variation of flow directions of two opposing air knife jets without cylinders at the lower lips (unit in m/s).
Fig. 3. Schematics of gas wiping mechanism.
3.2. Three Dimensional Flow Field in the Impinging
Region
As depicted in Fig. 1(a), two opposing jets are discharged
onto each other and a steel strip moves upward at the mid-
dle of the two jet exits. Since the jet width is twice larger
than the steel strip, the jets impinge normally on the steel
strip in the central region (z0) and they collide with each
other in the outer free space in z0. Figure 6 demonstrates
the velocity contours of the jet flows at various locations in
span-wise direction. Figure 6(a) shows the velocity contour
at the symmetric center plane. Since the jets are discharged
normally onto the steel strip, the velocity contours are near-
ly symmetric with respect to the moving steel strip. At
z1.75 mm the effect of colliding jets in the free space
begins to appear in the velocity contour. Slightly more wip-
ing gas flows downward due to the deflection of the jets in
the free space caused by the presence of the cylinder whose
diameter is d5mm. At z1.75 mm, a large percentage of
the wiping gas flows downward, and at z175mm, almost
all of the wiping gas flows downward. There are no alter-
nating vortices in the flow field.
3.3. Effect of Cylinder Diameter on the Impinging
Pressure and the Coating Thickness
Figure 7(a) compares the mean impinging pressure on
the steel strip surface for different cylinder diameters.
When there is no cylinder, as the steel strip edge is ap-
proached the mean impinging pressure significantly de-
creases from 21 to 16.5 kPa within 100 mm from the steel
strip edge. But, if a cylinder of small diameter is present at
each exit of the air knife system, the impinging pressure re-
mains at almost the same level up to the point 25 mm from
the steel strip edge. When the cylinder diameter is larger
than 3 mm, it was found that such constant pressure level
does not depend on the cylinder diameter. Similar compari-
son can be observed with Fig. 7(b) where the span-wise dis-
tributions of the coating thickness are compared for differ-
ent cylinder diameters. From these comparisons, it may be
ISIJ International, Vol. 46 (2006), No. 4
©2006 ISIJ 576
Fig. 6. Velocity contours at various span-wise locations (unit in m/s).
Fig. 7. Mean pressure and coating thickness distributions for
various cylinder diameters.
concluded that the problem of the EOC can be effectively
removed by installing a cylinder tangentially to each of the
lower lips of the opposing jet exits. It is worth noting that
the coating thickness becomes gradually thinner approach-
ing to the strip edge. Investigating the static pressure field
on the strip surface at the impinging region, depicted in
Fig. 8, it was observed that the cyclic pressure distribution
moves alternatively left and right. The pressure distribution
shown in Fig. 7(a), is the average pressure field of such an
alternatively moving cyclic field. At the end of the strip,
such movement is restricted due to the presence of the
cylinder. This may cause the higher static pressure near the
strip edge.
3.4. Effect of Gap Distance between the Cylinder and
the Steel Strip
At the exit of the two-dimensional jet, the nitrogen gas
spreads up and down two-dimensionally in the central re-
gion (z0), whereas in the free space (z0) the nitrogen
gas stream deflects downward due to the presence of the
cylinder. These two gas streams disperse into each other
span-wisely at the interface of the two streams. Therefore it
is important to investigate the effect of the gap distance be-
tween the cylinder and the steel strip on the coating thick-
ness distribution. Figures 9(a) and 9(b) compare the mean
impinging pressure and coating thickness distributions for
gap distances of g0mm, 2.5 mm and 5 mm. The diameter
of the cylinder is fixed at 5 mm for all cases. As can be seen
in Fig. 9(b), the case with g5mm yields most desirable
thickness distribution among the three cases. When the gap
becomes larger than 5 mm, the EOC begins to appear. This
trend is noted even with the gap distance of g5mm in Fig.
9(b).
3.5. Cylinder and Edge Baffle Plate
Finally in order to compare the EOC performance of the
cylinder installation method with the conventional method
that employs the edge baffle plates, the same analysis has
been carried out for the case of the edge baffle plate.
Figures 10(a) and 10(b) compare the EOC performances of
these two methods. As can be seen from the figures, the
cylinder installation method is far superior to the edge baf-
fle plate. Moreover, it is certain that the installation of
cylinders is much more convenient in maintenance and op-
eration. Experimental data of coating thickness distribution
obtained by Takeish4) are shown in Fig. 10(b). Although the
device to improve EOC is different, the computed coating
thickness distribution with the edge baffle is similar to that
obtained experimentally with the edge mask. Also note the
favorable comparison between the experimental and com-
puted thickness distributions for the case without baffle or
cylinder in Fig. 10(b). This agreement of the computed re-
sult with the experimental measurement may justify the use
of LES to study the EOC in the present work.
4. Conclusions
The EOC is a chronic problem in the continuous hot-dip
galvanizing process that is caused by the alternating vor-
tices which are generated by the collision of the two oppos-
ing jets in the free outside region of the moving steel strip.
In this research, a noble method was devised to prevent the
collision of two impinging jets in the free outside space. In
other words, a small diameter cylinder was installed tangen-
tially to the exit of the lower lip of the two-dimensional gas
wiping jet to deflect the jet flow direction downward in the
region where two opposing jets are not blocked by the steel
strip so that the two jets collide at an angle less than 180°.
In order to investigate the possibility to remove the EOC by
such a flow configuration 3-D unsteady compressible turbu-
lent flow around the gas wiping system has been calculated
using a commercial CFD software, STAR-CD. The flow
structure of the computed three dimensional flow field was
investigated in detail and the calculated impinging pressure
and shear stress data on the steel strip surface were used to
find the galvanized film thickness.
ISIJ International, Vol. 46 (2006), No. 4
577 ©2006 ISIJ
Fig. 8. Instantaneous pressure distribution on the strip surface.
Fig. 9. Mean pressure and coating thickness distributions for
various distances between cylinder and steel strip.
The effect of cylinder installation on the flow deflection
was first investigated by carrying out two-dimensional flow
field analysis. It was found that when the cylinder diameter
is greater than 3 mm, the wiping gas jet indeed is deflected
enough to prevent the occurrence of the alternating vor-
tices. In the three-dimensional calculation, it was concluded
that when the cylinder diameter is greater than 3 mm, the
EOC problem can be greatly improved. Another parameter
in concern was the gap distance between the cylinder and
the steel strip. When the gap distance is about 5 mm, the
EOC performance of the proposed method is most suitable.
Finally, the EOC performance of the cylinder installation
method was compared with the conventional method of in-
stalling edge baffle plates near both edges of the steel strip.
Computed film thickness distributions reveal that the cylin-
der installation to the lower lip of the gas wiping jet exit
shows much better EOC performance than the edge baffle
plate method.
Acknowledgements
This work was supported by the Brain Korea 21 Project
in 2005.
Nomenclature
d:Cylinder diameter
g:Distance between cylinder and steel strip
P
d
:Molten zinc pressure at y
d
T:Period of a cyclic change
v:Absolute molten zinc velocity component
Vs:Steel strip moving velocity
x,y,z:Cartesian coordinate
Greek symbols
d
:Molten zinc thickness
t
q
:Molten zinc shear stress at y
d
REFERENCES
1) J. J. Butler, D. J. Beam and J. C. Hawkins: Iron Steel Eng., 47 (1970),
77.
2) J. R. Park, S. H. Chun and R. B. Park: Korea Surface Eng., 32
(1999), 555.
3) Y. Takeishi, A. Yamauchi and S. Miyauchi: Tetsu-to-Hagané, 81
(1995), 643.
4) Y. Takeishi and H. Morino: ISIJ Int., 40 (2000), 1127.
5) M. Ohbu, K. Asakawa and K. Tano: Tetsu-to-Hagané, 60 (1974),
101.
6) J. M. Buchlin, M. Manna, M. Arnalsteen, M. L. Riethmuller and M.
Dubois: European Coating Symp. on the Mechanism of Thin Film
Coating, World Scientific, Singapore, (1996), 168.
7) S. J. Kim, J. W. Cho, K. J. Ahn and M. K. Chung: ISIJ Int., 43
(2003), 1495.
8) A. R. Gilchrist and D. G. Gregory-Smith: Int. J. Heat Fluid Flow, 9
(1988), 286.
9) Y. Addad, D. Laurence, C. Talotte and M. C. Jacob: Int. J. Heat Fluid
Flow, 24 (2003), 562.
10) Y. Addad, S. Benhamadouche and D. Laurence: Int. J. Heat Fluid
Flow, 25 (2004), 795.
11) C. H. Ellen and C. V. Tu: J. Fluids Eng., 106 (1984), 399.
ISIJ International, Vol. 46 (2006), No. 4
©2006 ISIJ 578
Fig. 10. Comparisons of mean pressure and coating thickness
distributions between cylinder and edge baffle plate.