XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:30am]
(PIC)
[1–17]
[PREPRINTER stage]
Original Article
An investigative study on the performance
of twist roll machine in a continuous cold
strip rolling mill
Proc IMechE Part C:
J Mechanical Engineering Science
0(0) 1–17
! IMechE 2012
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0954406212469149
pic.sagepub.com
Amir Mostashfi1, Mahmoud Kadkhodaei1, Mehrdad Poursina2
and Saeid R Bakhshi3
Abstract
In many modern continuous production lines of steel sheets, a twist roll machine is used to change the strip direction for
shortening the production line. A twist roll machine typically consists of a cylindrical body on which some guide rollers
are mounted in rows to gradually change the traveling direction of the strip when passes over the guide rollers. In this
article, quality of the sheets after exiting an industrial twist roll machine is first investigated. The amount and distribution
of wear on the guide rollers are also assessed by measuring and comparing diameter at different sections of selected
worn and new guide rollers. The specific wear rate as well as friction coefficient for guide rollers made of two different
popular polymers is measured by pin-on-disk wear tests. Details of the strip path on the twist roll machine as well as
contact between the strip and all the guide rollers are specified, and stress distribution in strip and the guide rollers is
studied by finite element analysis. Effects of the guide rollers material and arrangement, the bridle rolls tension, and width
and thickness of the strip on the amount and distribution of wear on the guide rollers as well as the elasto-plastic
response of the strip are studied. The results are utilized to propose techniques for reducing defects on the sheet and
the guide rollers, and finite element simulations show the effectiveness of these techniques.
Keywords
Twist roll, continuous rolling, wear, guide roller, strip defect
Date received: 26 June 2012; accepted: 25 October 2012
Introduction
In recent years, continuation of movement from one
piece of equipment to another one has been investigated in various manners as a factor of rationalization
for manufacture of steel sheets. However, since it
involves various problems regarding the installation
area and other aspects to array a number of pieces
of equipment in a linear manner, in order to enable
arrangement of the equipment in a small area, it is
contemplated to change the traveling direction of
the sheet. This is also necessary for achieving continuation of movement between existing pieces of equipment.1 One of the important equipment in most of the
continuous production lines of steel sheet is twist roll
machine, using which the traveling direction of the
strip is changed to reduce the length of the production
line. Figure 1 shows a schematic of this machine. The
strip generally enters a twist roll machine (TRM) from
above and begins to rotate by passing over some guide
rollers, which are mounted in rows on the cylindrical
body of the machine. The strip exits from below the
TRM while its direction is changed by 90 . In 1918,
Rosen2 presented the first mechanism for rotation of
sheet in rolling mills. This mechanism, based on which
many similar machines have been so far proposed,
consisted of a cylindrical body containing steel balls
to guide the strip motion. However, it was used just in
a few industries because it made much noise and
caused damages to the sheet due to its concentrated
contact with the balls. Later, in 1986, Hashimito
et al.1 founded the current roller-type twist roll,
which has been widely used in steel production
1
Department of Mechanical Engineering, Isfahan University of
Technology, Isfahan, Iran
2
Department of Mechanical Engineering, Faculty of Engineering,
University of Isfahan, Isfahan, Iran
3
Department of Materials Engineering, Malek-Ashtar University of
Technology, Isfahan, Iran
Corresponding author:
Mahmoud Kadkhodaei, Department of Mechanical Engineering, Isfahan
University of Technology, Isfahan 84156-83111, Iran.
Email: [email protected]
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:30am]
(PIC)
[1–17]
[PREPRINTER stage]
2
Proc IMechE Part C: J Mechanical Engineering Science 0(0)
companies. The guide rollers are usually made of
polymer, among which polyamide 66 (PA66) and
polyamide 11 (PA11) are the most common materials.
When a TRM is utilized in a rolling mill, any
improper adjustment in the sheet tension, contact
between the strip and the guide rollers, arrangement
of the guide rollers, or their material may cause imperfections such as formation of crossbow in the sheet or
indented lines on its surface, and excessive or uneven
wear on the guide rollers.
A lot of modifications are thus tried in practice to
reduce these damages, and no comprehensive technique is reported so far due to the complexity of the
machine as well as changes in the sheet properties
such as width, thickness, and material in different
companies and different production schedules of a
rolling mill. Consequently, published researches are
also generally just focused on wear characteristics of
the guide rollers for selection of appropriate materials
to minimize their wear rate. Unal et al.3 investigated
dry sliding wear characteristics of some industrial
polymers against steel counterface with the use of a
pin-on-disc testing machine. They studied the influence of test speed as well as the applied pressure
values on friction and wear behavior of PA66 and
other industrial polymers. They reported that the friction coefficient decreases with increase in the amount
of pressure and that the specific wear rate of PA66 is
less than that of the other polymers. They also found
that the specific wear rate shows a very little sensitivity to the applied pressure and test speed for PA66.
Chen et al.4 investigated the effect of fiber reinforcement on the friction and wear behavior of PA66 running against itself using a twin-disc test rig. They
showed that both the wear and friction properties of
unreinforced PA66 can be improved considerably by
filling it with 20 wt% PTFE. Further, Kukureka
et al.5 found that one of the major benefits of fiber
reinforcement, particularly by using glass, is that it
reduces the coefficient of friction and hence allows
the material to be used for heavier duties without
exceeding the softening point of the matrix. This
increase in duty, however, increases the specific wear
rate leading to a shorter component life. Chen et al.6
studied the mechanical and tribological properties of
PA/PPS blends. They found that the crystalline structure of PA66, PPS, and PA66/PPS blends changes due
to sliding and tribochemical reactions occurred with
the PA66 and the PA66 phases in blends.
In this article, for the TRM of Mobarkeh Steel
Company (MSC) in Iran, defects on the steel sheets
caused by the machine are first studied. Also specific
wear rate and its distribution on the guide rollers are
evaluated through the measurement of the worn guide
rollers diameter at different sections, and the obtained
dimensions are compared with those in the new guide
rollers. As PA66 and PA11 are two popular materials
for the guide rollers in MSC, pin-on-disk tests are
carried out to assess and compare the friction and
wear behavior of these two materials. Pin specimens
are fabricated from sheets crossing over the TRM,
and specific wear rate as well as friction coefficient is
determined for different applied normal loads and
sliding velocities. Geometry of the whole machine
together with the passing strip is created in CATIA
to investigate exact path of the strip on the TRM as
well as its contact with all the guide rollers. This helps
to better understand the origin of uneven wear on the
guide rollers as well as differences in the wear behavior of different guide-roller rows. Finite element simulations of travelling the strip over the TRM guide
rollers for different operational conditions are done
using ABAQUS/Explicit, and stress distribution in
strip and the guide rollers is studied. Finite element
(FE) results as well as the wear test findings are utilized to study the effect of material and arrangement
of the guide rollers, the bridle rolls tension, and width
and thicknesses of the strip on the guide rollers specific wear rate and its distribution as well as the
defects on the strip. Based on these findings, practical
Figure 1. A schematic of a twist roll machine.
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:30am]
(PIC)
[1–17]
[PREPRINTER stage]
Mostashfi et al.
3
techniques are systematically found to reduce damages to the sheet and the guide rollers in the studied
TRM. Effectiveness of these techniques is shown in
the finite element simulations indicating that the
method proposed in this article can be considered as
an organized troubleshooting procedure for a TRM.
Defects of the TRM
Investigating the TRM of Mobarakeh Steel
Company, the following major damages to the sheet
and the guide rollers are identified:
Crossbow formation on the sheet
Although the ingoing strip to the TRM is tried to
have a flat and smooth cross section, its edges are
slightly deformed at exit from the machine leading
to crossbow formation as shown in Figure 2.
Crossbow is a type of surface-to-surface length differential defect, and it is seen in practice that this defect
declines with decrease in the sheet thickness.
strip and the guide rollers to be non-uniform and different from the expected manner. This leads to more
pronounced defects on the sheet. The amount of the
bridle rolls tension, the cross-sectional dimensions of
the sheet, and the guide roller materials mainly affect
the wear rate. In MSC, the guide rollers are made of
PA66 or PA11, and investigations show that these two
polymers exhibit different wear behaviors for the same
process conditions.
Angular wear of guide rollers
In some of the TRM guide rollers, wear pattern is not
flat but rather angular. Angular wear results in the
sheet deviation to one side of its path, and hence
more wear is made on the guide rollers of that side.
Figure 4(a) and (b) shows the flat and angular wear of
the guide rollers, respectively.
Geometrical considerations of the TRM
The TRM guide rollers are gradually worn during
operation due to contact with the travelling strip.
Beside the material waste, decrease in the guide rollers
diameter due to wear causes the contact between the
As shown in Figure 5, the studied TRM consists of
7 guide-roller rows, each of which contains 27 guide
rollers. The traveling direction of steel sheet is changed by its movement in a spiral manner over these
guide-roller rows. The method of numbering the
guide rollers is shown in Figure 5.
The components of TRM are modeled in CATIA
to better understand configuration of the machine and
the strip and to show how the strip rotates over the
machine. Using this 3D model, it is possible to determine the area, length, and angle of contact between
the sheet and the guide rollers. This angle, varying
between 5 and 7.5 , is shown in Figure 6. The contact
angle increases in initial and final guide-roller rows
due to more asymmetrical contact between the strip
and the guide rollers. Due to the guide rollers arrangement on the cylindrical body and also the existence of
the bridle tension on the sheet, outer edges of the
guide rollers have the longest contact with the sheet.
As a result, the amount of wear in this edge is more
Figure 2. Crossbow in the sheet.
Figure 3. Indented lines on the strip surface.
Indented lines on the sheet surface
When the strip passes over the TRM guide rollers,
longitudinal indented lines shown in Figure 3 appear
on its surface. Investigation of this defect indicates
that increase in the bridle rolls tension leads to more
tense indentations on the sheet surface. Measurements
in MSC show that the depth of these indented lines
varies from 0.1 to 1 mm along the strip length. This
defect causes severe decrease in the quality of the produced sheet.
Excessive wear on the guide rollers
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:30am]
(PIC)
[1–17]
[PREPRINTER stage]
4
Proc IMechE Part C: J Mechanical Engineering Science 0(0)
than the other one causing the angular wear of the
guide rollers (Figure 4(b)).
Once the contact angle between the guide rollers of
a row and the raveling strip is determined in the
CATIA model, average of the angles can be calculated
as the theoretical wear angle in that row. The actual
wear angle can be evaluated in practice, too, by measuring the slope on the surface of the worn guide rollers. In Figure 7 the experimental measurements on
PA11 guide rollers are compared with the theoretical
wear angles derived from the CATIA model.
Referring to Figure 5, number of the guide-roller
rows begins from 1, for the first row at entry side,
and ends to 7, for the last row at exit side. As it can
be seen, a very good agreement exists between the
experimental and the simulation results for the
middle roller rows. Differences shown for the initial
and the final rows mainly arise from the following
sources:
1. The arrangement of the guide rollers is so that the
initial and the final rows are subjected to less average contact pressure in comparison to the middle
ones. It will be shown in section ‘Wear behavior of
guide rollers in the TRM’ that decrease in the contact pressure would lead to increase in the amount
of wear especially for PA11.
2. The sheet experiences some fluctuations in practice when passing over the initial and the
final guide-roller rows. This may be considered
as another reason for a less contact pressure
acting on these rollers leading to more wear
on them.
Figure 4. Two wear patterns on the guide rollers: (a) flat
wear and (b) angular wear.
Accordingly, the CATIA model gives a theoretical
wear angle in which only geometrical aspects of the
contacts are taken into account. However, less contact
pressure causes more wear rate than the theoretical
predictions. Consequently, beside the geometrical
details, effects of contact pressure and some other factors are studied in section ‘Wear behavior of guide
Figure 5. A schematic of the TRM guide rollers showing the method of numbering the rows and the rollers of each row.
TRM: twist roll machine.
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:31am]
(PIC)
[1–17]
[PREPRINTER stage]
Mostashfi et al.
5
rollers in the TRM’ to more practically investigate the
wear behaviors of the guide rollers.
It should be noted that deep groove ball bearing
are usually employed as the hinge supports of the
guide rollers in a TRM. As these bearings cannot
compensate any misalignment, using self-aligning
bearings is suggested to allow the guide rollers to
adjust their orientation relative to the sheet surface
leading to less angular wear of the guide rollers.
Moreover, if some guide-roller rows are added
between the rows with high wear, the number of locations where the sheets rotate over the guide rollers is
increased. This leads to smaller contact angle between
the sheets and the guide rollers, and hence less angular
wear is achieved. Figure 7 shows that the guide-roller
rows 1, 2, 6, and 7 experience more angular wear than
the other ones. So, addition of one row between the
rows 1 and 2 and another between the rows 6 and
7 may effectively decrease the amount of angular
wear in the guide rollers.
the tensions of the bridle rolls located before and
after the TRM.
2. Due to the existence of the bridle rolls, entry and
the exit edges of the sheet are so fixed that they
cannot move and/or rotate transversely and vertically. In other words, the sheet is only free to move
and/or rotate longitudinally.
3. The front edge of the sheet moves with the velocity
of 2 m/s at exit from the TRM.
4. Since the weight of the sheet crossing from TRM
is not negligible, the gravity acceleration is defined
for the sheet.
C3D8R (an 8-node linear brick element with reduced
integration and hourglass control) element is used to
Stress analysis of the Sheet
To study the stress distribution in the sheet, finite
element method (FEM) is employed using
ABAQUS/Explicit. The contact of guide rollers with
the traveling sheet is assumed to be surface-to-surface
using penalty contact algorithm where the contact
interaction is according to tangential scheme in
ABAQUS. The boundary conditions are as follows:
1. A distributed tensile stress is applied to the entry
as well as the exit edge of the sheet representing
Figure 7. The angles of wear for different guide-roller rows.
Figure 6. Illustration of the guide rollers angular contact.
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:31am]
(PIC)
[1–17]
[PREPRINTER stage]
6
Proc IMechE Part C: J Mechanical Engineering Science 0(0)
mesh the sheet body. It is important to use a sufficiently refined mesh to ensure that the FEM results
are independent from the element size. A mesh is said
to be converged when further mesh refinement produces a negligible change in the results. Figure 8
shows variations of the maximum effective stress generated in a sheet with 2.8 mm thickness and 900 mm
width subjected to 15 MPa tensile stress for different
numbers of the element. It can be seen that after
around 500,000 elements the maximum effective
stress converges to 238 MPa. All the upcoming FE
results are obtained from a converged mesh.
The steel sheet used in MSC is ST12 with the engineering stress–strain curve shown in Figure 9 for a
simple tension test. This curve is fitted to the following
equation and is converted to true stress–strain curve
to define the material properties in the FE
simulations.
¼ 560"0:174 MPa
ð1Þ
To study the conditions in which the highest
amounts of stresses are applied to the strip, a sheet
with 900 mm width and 2.8 mm thickness is
Figure 8. Illustration of the mesh convergence for the maximum effective stress of a sheet with 2.8 mm thickness and 900 mm width
subjected to 15 MPa tension.
Figure 9. Engineering stress–strain curve for ST12 steel sheet.15
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:31am]
(PIC)
[1–17]
[PREPRINTER stage]
Mostashfi et al.
7
considered to pass over the TRM. According to the
practical production plan, three different tensions per
unit cross section area of the sheet with the amounts
of 12, 15, and 20 MPa acting by the bridle rolls are
investigated since these are the only amounts applied
in Mobarakeh Steel Company. Figure 10 illustrates
the maximum equivalent stress in the sheet for these
three amounts of the forward and backward tensions.
Referring to Figure 9, the yield stress of ST12 is
236 MPa, and Figure 10 shows that the maximum
effective stress in the sheet reaches 236 MPa when a
14 MPa tensile stress is applied by the bridle rolls.
Consequently, permanent deformations will occur in
the traveling strip if the bridle rolls tension exceeds
around 14 MPa. These permanent deformations
appear in the form of indented lines on the sheet
surface in the contact areas of the strip with the
guide rollers. In other words, to avoid creating these
lines on the sheet, the bridle rolls tension must be less
than 14 MPa. A similar investigation is done in Figures
11 and 12, where the maximum effective stress in the
sheet is shown for two thicknesses of 1.5 and 2.8 mm at
a constant width of 900 mm and for the sheet with
different widths of 900 and 600 mm at a constant thickness of 2.8 mm, respectively. These two figures show
that increase in the width and thickness of sheets generally leads to increase in the maximum generated
effective stress in the sheets. Therefore, to prevent
yielding of the sheet, the applied tensile stress by the
bridle rolls must be reduced. In Table 1, the maximum
allowable tensions per unit cross section in order to
limit the maximum generated effective stress to the
Figure 10. Maximum effective stress of the sheet with 900 mm width and 2.8 mm thickness at different tensions.
Figure 11. Maximum effective stress for the sheet with 1.5 and 2.8 mm thicknesses and 900 mm width.
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:32am]
(PIC)
[1–17]
[PREPRINTER stage]
8
Proc IMechE Part C: J Mechanical Engineering Science 0(0)
Figure 12. Maximum effective stress for the sheet with 900 and 600 mm widths and 2.8 mm thickness.
Table 1. Maximum allowable tensile stresses and carrying capacity applied by the bridle rolls to ST12.
600 mm
900 mm
Thickness (mm)
Width (mm)
Maximum allowable
tensile stress (MPa)
Carrying capacity
(kN)
Maximum allowable
tensile stress (MPa)
Carrying
capacity kN)
1.5
2.8
17.2
16
15.48
26.88
15
14
20.25
35.28
Figure 13. Stress distribution on the sheet surface with 900 mm width and 1.5 mm thickness for 15 MPa tension.
yield stress of 236 MPa are shown for steel ST12 at
different widths (600 and 900 mm) and thicknesses
(1.5 and 2.8 mm). Moreover, the corresponding carrying capacities as the allowable tensile loads (multiplication of the stress by the width and the thickness of
the sheet) of the bridle rolls are listed. As it is seen,
although less tensile stresses are allowable for larger
thicknesses and widths, increase in the cross section
gives rise to increase in the amount of resultant carrying capacities. Figure 13 shows the effective stress distribution on a sheet with 900 mm width and 1.5 mm
thickness at 15 MPa tension.
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:33am]
(PIC)
[1–17]
[PREPRINTER stage]
Mostashfi et al.
9
Figure 14. Maximum effective stress for sheets with different widths and thicknesses subjected to the currently applied tensions and
the suggested ones by the bridle rolls.
Figure 15. Variations of the effective stress through the sheet thickness.
Figure 14 compares the allowable tensile stresses of
the bridle rolls recommended in Table 1 with those
currently applied in practice by the bridle rolls in
MSC. By comparing the numbers for both the two
thicknesses, it is found that the appearance of
indented lines on the strips of MSC after being
passed over the TRM is due to the fact that the
applied bridle-roll tensions are greater than the
amounts due to which the sheet begins to yield.
In Figure 15, distribution of the effective stress is
shown through the sheet thickness. It can be seen that
the stress at the bottom and the top layers of the sheet
is higher than that in its middle layer. According to
Beheshti7 and Hira et al.,8 in such a situation, the
sheet will deform to a bent configuration leading to
the appearance of the crossbow defect. Williamson9
recommends sheet leveling in these conditions to prevent the formation of crossbow on the sheets.
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:33am]
(PIC)
[1–17]
[PREPRINTER stage]
10
Proc IMechE Part C: J Mechanical Engineering Science 0(0)
by the load cell shown in Figure 18. The specific wear
rate k reported in this study is calculated according to
the relation
Wear behavior of guide rollers
in the TRM
To study the effects of velocity and applied normal
load on specific wear rate of the guide rollers as well
as to determine the friction coefficient of the guide
rollers with ST12 sheet, pin-on-disc wear tests are performed for two conventional polymeric guide rollers:
PA66 and PA11. A polyamide, commonly referred to
as Nylon, is a polymer containing monomers of
amides joined by peptide bonds. By varying CH2/
CONH ratio, several PAs with varying properties
have been synthesized. This ratio is 10 for PA11 and
is 5 for PA66.10 In Table 2, some properties of PA66
and PA11 are presented.
For pin-on-disc wear testing, according to the
standard provided in Rao et al.,12 discs with 60 mm
diameter and 5 mm thickness made of the two polymers are prepared. The pin specimen is made of the
ST12 steel sheet. In order to provide real contact conditions of the guide rollers with the sheet, the pin is
required to be prepared so that the practical contact
between the sheet surface and the polymeric disc
maintain during the wear tests. Therefore, the pin is
cut from a strip, which is passed over the TRM, so
that its length is equal to the strip thicknesses and its
upper and lower surfaces are those of the strip. To use
such a short pin, a fixture is made to put the pin in a
point-to-point contact with the disc. As shown in
Figure 16, this fixture is made from a 5-mm diameter
rod for connection to the wear testing device and a
mounting plate to install the pin specimen on the fixture. For the pin specimen, the sheet shown in
Figure 17 is used and is mounted on the fixture.
Prior to installation, the pin specimen is bulged to
provide point-to-point contact with the disc.
According to wear test standards, the contact area is
set to be 1 mm2. In Figures 18 and 19, the pin-on-disc
wear testing device and disc specimens cut from the
guide rollers using conventional turning are shown,
respectively. To investigate the effect of normal pressure on specific wear rate and coefficient of friction,
9 tests with different forces from 10 to 110 N (according to the capacity of the testing device)are conducted
for both PA66 and PA11 disks. The weight loss measurement is done after the tests for a sliding distance of
500 m. The sliding speed is set to be 0.083 m/s in order
to be sure that the tests are carried out in a steadystate manner. The friction coefficient is obtained
through the calculation of frictional torque measured
k¼
wv
FN S
ð2Þ
Figure 16. Fixture made for pin grippage.
Figure 17. Pin specimen.
Table 2. Properties of PA66 and PA11 guide rollers.16
Property
Thermal expansion
(K)
Young’s modulus
(MPa)
Shear modulus
(MPa)
Tensile strength
(MPa)
Elongation
(%)
Melting temperature
( C)
PA11
PA66
110–120 106 K
70–100 106 K
1100–1400
1700–2000
450–500
1100–1200
47
80–85
280
120–300
190
260
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:34am]
(PIC)
[1–17]
[PREPRINTER stage]
Mostashfi et al.
Figure 18. Wear test machine.
Figure 19. Disk specimens.
11
where wv, S, and FN are the wear volume (mm3), the
sliding distance (m), and the applied normal load (N),
respectively. The wear volume wv is determined by
measuring the weight loss of the disc by using an analytical balance with the precision of 0.1 mg. Three
repeated sliding tests are carried out for minimizing
data scattering, and the reported friction coefficient
and specific wear rate are the average values obtained
from the three repeated tests.
Variations of the specific wear rate and friction
coefficient with the applied normal load for PA11
and PA66 under dry friction conditions are presented in Figures 20 and 21, respectively. The results
reveal that the specific wear rate and friction coefficient of PA11 considerably decrease with increase in
the applied load, whereas those of PA66 are almost
independent from the applied load. Since different
amounts of contact loads act on different guide rollers of a row and on different points of each guide
roller, the behaviors shown in Figures 20 and 21
indicate that guide rollers made of PA66 will all
show almost flat and equal wear throughout the
TRM while this is not the case for guide rollers
made of PA11. If PA11 guide rollers have to be
employed in a TRM, increase in the contact forces
leads to decrease in the specific wear rate of the
rollers. However, increasing the contact forces
needs the bridle rolls tension to be increased, while
the tension is limited to the amount which leads to
the yielding of the sheet. In other words, increase
in the bridle rolls tension results in decreasing the
guide rollers wear rate, on one hand, and increasing
the risk of the sheet yielding, on the other hand.
Consequently, an optimum amount of the bridle
rolls tension should be applied to satisfy both the
requirements.
Figure 20. Effect of applied load on the specific wear rate of polymer discs under dry sliding conditions (sliding speed: 0.083 m/s).
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:34am]
(PIC)
[1–17]
[PREPRINTER stage]
12
Proc IMechE Part C: J Mechanical Engineering Science 0(0)
Figures 20 and 21 also show that the specific
wear rate and friction coefficient of PA66 are generally less than those of PA11. These all indicate
that the use of PA66 instead of PA11 will efficiently
lead to more even wear as well as less wear rate of
the guide roller, and these benefits give rise to suppression of the surface defects on the traveling strip
over a TRM.
To investigate the effect of speed on wear behavior
of PA11 and PA66, different wear tests at different
typical velocities are carried out with a fixed load of
6.93 kg and the sliding distance of 500 m. Variations
of the specific wear rate and friction coefficient with
speed are respectively shown in Figures 22 and 23. In
general, the specific wear rates for PA11 and PA66 are
obtained in order of 109 mm3/Nm. It is believed that
Figure 21. Effect of applied load on friction coefficient under dry sliding conditions (sliding speed: 0.083 m/s).
Figure 22. Variations of the specific wear rate with velocity for PA66 and PA11 polymers (sliding speed 500 m, applied load 6.93 kg).
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:35am]
(PIC)
[1–17]
[PREPRINTER stage]
Mostashfi et al.
13
these results are associated with the softening temperature of the polymers, which leads to surface
plastic deformation under lower load values.13
Comparison of Figures 22 and 20 shows that the specific wear rate of PA66 and PA11 are not highly influenced by changes in the speed, and this is in
agreement with the findings reported by Liu et al.14
Referring to Figure 23, there will be an average
decrease of about 25% in the friction coefficient of
PA11 against ST12 when the speed is nearly doubled.
This is in agreement with the findings reported by Jia
et al.13 and Wang and Li.11 This figure also shows
that an average increase of around 20% is seen in
the friction coefficient of PA66 against ST12 if the
speed is nearly doubled, and this is in agreement
with the results obtained by Wang and Li11 and
Unal et al.3 However, comparison of Figures 23 and
21 indicates that the effect of changes in the contact force on the variations of friction coefficient is
more pronounced than that of changes in the velocity.
Consequently, the friction coefficient may be assumed
to be almost independent from the sliding velocity.
Practical Interpretation of the
experimental and numerical predictions
The proposed results of the pin-on-disk wear test
mainly show that the contact force is a very influencing parameter on the wear rate of the guide rollers.
Increase in the contact pressure generally results in
decrease in the amount of wear on PA11 and PA66,
but it is seen that PA11 is more pressure-sensitive than
PA66. Wear of the guide rollers evidently affect their
diameter in practice. Accordingly, to more clearly
compare the present experimental and numerical
results, PA11 guide rollers are investigated after
being in service for a continuous period of 6 months
in Mobarakeh Steel Company, and their diameters
are measured to study their specific wear rate after
this working interval. Moreover, the contact forces
acting on each guide roller in the TRM of MSC is
determined with the use of the finite element simulation. Figure 24(a) and (b) shows the amount of the
contact forces on all the guide rollers and diameter of
the worn rollers measured after 6 months, respectively. Comparison of these two figures indicates the
following findings:
1. According to Figure 24(a), the guide rollers of
rows No. 3, No. 4, and then No. 5 are mostly
subjected to the highest contact forces, respectively, and Figure 24(b) shows that the guide rollers
of these rows have larger diameters than the other
ones. In other words, the less wear rates occur on
the guide rollers where the highest contact forces
act and this in agreement with the findings shown
in Figure 20.
2. The obtained contact forces for the sixth row show
the highest variations among its guide rollers as
the maximum load is around 24858.5 N on the
first guide roller and the minimum load is
around 13027.7 N on the ninth roller. According
to Figure 20, it is expected that the most non-uniform wear patters exists among the guide rollers of
this row, and Figure 24(b) approves it as the most
varying diameters are obtained for these rollers. In
contrary, the least variations of the contact forces
are seen among the guide rollers of the fourth row,
and hence the least variations in diameter of the
worn rollers of this row are seen.
Figure 23. Variations of friction coefficient with sliding velocity for PA66 and PA11 polymers (sliding speed 500 m, applied load
6.93 kg).
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:35am]
(PIC)
[1–17]
[PREPRINTER stage]
14
Proc IMechE Part C: J Mechanical Engineering Science 0(0)
Figure 24. Illustration of: (a) applied contact load on guide rollers, (b) diameters of the guide rollers after 6 months working of TRM.
TRM: twist roll machine.
Figure 24(a) and (b) and the above-mentioned
results indicate that if PA11 guide rollers are
employed in the studied TRM, adding a rollers row
after the first one in the current configuration of the
machine as well as adding a row before the last one
will cause the contact loads distribution to be more
uniform. This leads the guide rollers to experience
almost the same amounts of the specific wear rate
causing diameters of the worn rollers to be all closer
to each other. However, inserting new guide-roller
rows between the current rows of the studied TRM
is a difficult task. Any change in the orientation or the
thickness of the guide rollers is no practical too.
Consequently, a more practical and easier technique
is proposed and investigated in the subsequent
section.
A practical modification to the TRM
Considering the limitations of some solutions proposed in the previous section, another technique for
obtaining more even wear on the guide rollers is to
use guide rollers with different initial diameters to
assist in obtaining more uniformly distributed contact
forces. In a guide-rollers row, if the diameter of
the rollers with the highest contact forces is less
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:36am]
(PIC)
[1–17]
[PREPRINTER stage]
Mostashfi et al.
15
than that of the other ones in the same row, a less
contact area is achieved leading to decrease in the
contact stresses acting on these guide rollers. For
instance, referring to Figure 24(a), variations of the
contact force on the guide rollers of the first row
shows general increase from the first roller to the
last one. Accordingly, if a reverse order is maintained
in the initial diameter of these guide rollers, a
more uniform distribution is achieved for their contact forces. Considering this rule, after trying various
distributions for the diameters in the finite element
simulations, the results shown in Figure 25 are
obtained.
As it is seen, from the first guide roller to the last
one in a row, an increasing order for the rows No. 1 to
No. 3 and a decreasing order for the rows No. 5 to
No. 7 is proposed for the diameters. However, as it
was discussed earlier according to Figure 24(a), the
contact forces on the guide rollers of the fourth row
have the least variations, and hence no change in the
diameters of this row is suggested.
The effect of this new arrangement of the diameters is seen in Figure 26(a) and (b) where the stress
distribution on the traveling sheet is shown for the
proposed and the current arrangements, respectively.
It is seen that the sheet would be involved with
more uniform contact stresses, in comparison to the
current situation of TRM, if the proposed arrangement is applied. This gives rise to more uniform
contact forces as well. Figure 27 compares the contact forces acting on the guide rollers in the current
configuration of the machine with those obtained
if the new arrangement is employed. As it is seen,
less variations in the contact forces are seen when
the proposed arrangement is replaced with the
current one.
Conclusion
In this article, for an industrial twist roll machine,
defects on the traveling sheets as well as the amount
and distribution of wear on the guide rollers are investigated. Pin-on-disk wear tests are carried out to
assess and compare specific wear rate as well as friction coefficient of two polymers PA66 and PA11
against ST12 at different conditions. The results
show that the effect of the contact force on the wear
rate and friction coefficient is higher than that of the
sliding velocity. Moreover, PA66 has a smaller wear
rate as well as friction coefficient and its wear behavior is not very sensitive to changes in the sliding velocity and contact force compared to PA11. As these
two polymers are widely used in industrial twist roll
machines, these results indicate the advantages of
PA66 over PA11. Geometry of the machine and the
strip is modeled in CATIA to better understand the
origin of uneven wear on the guide rollers as well as
differences in the wear behavior of different guideroller rows. Finite element simulations of the travelling strip and the guide rollers are done using
ABAQUS/Explicit. It is shown that distribution of
stress through the sheet thickness causes the formation of crossbow and that the sheet yields if high
amounts of tension greater than some allowable numbers are applied by the bridle rolls. Yielding of the
sheet will lead to the appearance of indented lines
on its surface. Maximum applicable tensions before
the sheet begins to yield are determined for different
operational conditions of the studied TRM. Adding
some roller rows between the existing ones is found to
be beneficial in both suppressing the angular wear as
well as obtaining more uniform wear rate of the guide
rollers in each row. Employing self-aligning bearings
Figure 25. The proposed variations of the initial diameters of guide rollers in the studied TRM.
TRM: twist roll machine.
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
16
[30.11.2012–10:36am]
(PIC)
[1–17]
[PREPRINTER stage]
Proc IMechE Part C: J Mechanical Engineering Science 0(0)
Figure 26. Comparison of the effective stress distribution on the sheet for: (a) the proposed arrangement, (b) the current
arrangement.
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016
XML Template (2012)
K:/PIC/PIC 469149.3d
[30.11.2012–10:38am]
(PIC)
[1–17]
[PREPRINTER stage]
Mostashfi et al.
17
Figure 27. Comparison between the rollers contact force in the proposed and the current arrangements.
instead of deep groove ones as the supports of the
guide rollers is also suggested to reduce the uneven
wear of the rollers. If PA11 has to be used in a
TRM, a practical technique to obtain more even
wear on the guide rollers is to use different initial
diameters for the rollers of each row. An arrangement
according to the distribution of the contact forces on
the guide rollers is proposed in this work, and the FE
simulations show that this method effectively results
in more even wear of the rollers. The present study
can be considered as a comprehensive tool in optimization of the performance of a twist roll machine.
Funding
This research received no specific grant from any funding
agency in the public, commercial, or not-for-profit sectors.
6.
7.
8.
9.
10.
11.
Acknowledgment
The technical personnel of Mobarkeh Steel Company are
appreciated for their cooperation in collecting the experimental data and for their instructive discussions.
12.
References
1. Hashimoto K, Nakano T, Arita K, et al. Apparatus for
changing the traveling direction of a web-like material.
Patent No. 4687125, USA, 1987.
2. Roesen O. Web feeding mechanism. Patent No. 1273926,
USA, 1918.
3. Unal H, Sen U and Mimaroglu A. Dry sliding wear characteristics of some industrial polymers against steel counterface. J. Tribol Int 2004; 37: 727–732.
4. ChenYK, Kukureka SN, Hooke CJ, et al. Surface topography and wear mechanisms in polyamide 66 and its
composites. J Mater Sci 1999; 35: 1269–1281.
5. Kukureka SN, Hooke CJ, Rao M, et al. The effect of
fibre reinforcement on the friction and wear of
13.
14.
15.
16.
polyamide 66 under dry rolling-sliding contact.
J Tribol Int 1999; 32: 107–116.
Chen Z, Li T, Yang Y, et al. Mechanical and tribological properties of PA/PPS blends. J Wear 2004;
257: 696–707.
Beheshti A. Investigation and simulation of tension leveling process. MS Thesis, Isfahan University of
Technology, 2007.
Hira T, Abe S and Azuma S. Analysis of sheet metal
bending deformation behavior in processing lines and
its effectiveness. Scientific Report, Kawasaki Steel
Engineering 1988.
Williamson HM. Strain softening during tension leveling of rolled sheet metals. Scientific Report, Institute of
Metals and Materials, Australia, 1986.
Rajesh JJ, Bijwe J and Tewari US. Abrasive wear performance of various polyamides. Wear 2002; 252:
769–776.
Wang YQ and Li J. Sliding wear behaviour and mechanism of ultra-high molecular weight polyethylene.
J Mater Sci Eng 1999; A2(66): 155–160.
Rao M, Hooke CJ, Kukureka SN, et al. The effect of
PTFE on the friction and wear behavior of polymers in
rolling-sliding contact. J Polym Eng Sci 1998; 38(12):
1946.
Jia B-B, Li T-S, Liu X-J, et al. Tribological behaviors of
several polymer–polymer sliding combinations under
dry friction and oil-lubricated conditions. J Wear
2007; 262: 1353–1359.
Liu CZ, Ren LQ, Tong J, et al. Effects of operating
parameters on the lubricated wear behavior of a PA6/UHMWPE blend: a statistical analysis. J Wear 2001;
249: 878–884.
Poursina M. To investigate effective factor in strip
braking in tandem mill with neural network. Scientific
Report, Mobarakeh Steel Company, 2007.
van Krevelen DW and te Nijenhuis K. Properties of
polymers. 4th ed. Sloveina: Elsevier, 2009.
Downloaded from pic.sagepub.com at PENNSYLVANIA STATE UNIV on September 13, 2016