5

th
Australasian Congress on Applied Mechanics, ACAM 2007
10-12 December 2007, Brisbane, Australia
A method of approximate tool wear analysis in cold roll forming
Alexander S. Galakhar
1
, Paul A. Meehan
1
, William J.T. Daniel
1
, and Shi Chao Ding
1
1
Division of Mechanical Engineering, the University of Queensland, Brisbane, QLD 4072 Australia
Abstract: A method of approximate tool wear evaluation is proposed for cold roll forming (CRF). The
method uses simple assumptions for approximate tool wear analysis and allows estimation of roll
profile change caused by wear. Boundary conditions are obtained from the model based on a new
relaxation method. The strip material obeys the rigid-perfect plastic model. A roll-strip sliding velocity
distribution is calculated from roll geometry. A roll-strip vertical contact pressure component is
assumed to be constant within the contact area. The method is a first step in the development of a
reliable approach to tool wear prediction in CRF.
Keywords: approximate wear analysis, cold roll forming, tool wear.
1 Introduction
Rolls replacement in CRF mill followed by equipment tuning is a significant cost to production in CRF [2].
This cost will be reduced if the rolls are designed under an equal wear criterion [5] in application to CRF.
The equal wear criterion is important as tool wear affects not only production time losses in CRF, but
also the energy effectiveness of the entire CRF process. Unfortunately, tool wear remains unstudied in
CRF. The only previous work related to wear in CRF was devoted to surface wear of a formed
stainless strip [4]. This did not take into account tool wear due to the complexity of experimental and
predictive numerical modelling of tool wear in CRF.
2 Tool wear analysis in cold roll forming
Tool wear during the CRF process is determined by i nteraction between forming rolls and a formed
strip that has a complex nature. The three main factors that affect the overall tool wear and forming roll
profile are roll-strip contact pressure distribution, sliding and forming roll surface hardness.
Roll-strip contact pressure has highly non-uniform distribution in the roll-strip contact. The forming pressure
has one component normal and two orthogonal components tangential to the formed strip. The roll-strip
slip velocities in contact have two components in the strip-related curvilinear coordinates with a non-
linear distribution. The forming roll surface hardness depends on the treatment of forming roll surface.
The accurate analysis of roll-strip contact pressure distribution as well as roll-strip contact sliding
velocities distribution is based on a solution to a complex geometrical and structural problem.
However, if only the overall tool wear is estimated, a simplified tool wear model can be used. The
predictive model must have input parameters obtained using standard material test procedures and
results must be validated with measurements of forming roll profile change due to wear. Details of this
simplified approximate tool wear model are described in the following.
2.1 Approximate model of contact pressure distribution
The robust and reliable Archards equation was chosen as a wear model for metal surfaces (1):
* c s
P V w
K
t Hr

, (1)
where w
*
is wear of roll material in a considered point,
3
2
m
m
1
1
]
,
t is time, [s],
K is an experimentally determined dimensionless surface contact wear coefficient [3],
P
c
is a contact pressure, [MPa] ,


V
s
is a roll -strip contact area average sliding velocity,
m
s
1
]
,
Hr

is a forming roll surface hardness, [MPa],

The formula (2) was developed by commonly used empirical table data to convert Hr from HRC units
into MPa (Hr

) with a deviation not greater than 6% for hardness values in the range 1 62 HRC:
Hr

= 7.537 10
-5
Hr
4
3.018 10
-3
Hr
3
+ 0.205 Hr
2
+ 10.487 Hr + 513.935. (2)
The overall tool wear is calculated by the contact pressure averaged by the area of the roll -strip
contact surface obtained with the relaxation method for roll forming problems [1].
2.2 Roll geometry analysis
Roll geometries are reconstructed by matrices of nodes and connectivity of the mesh of roll-strip
contact reference surface. These matrices are created by data of the concave or convex roll-strip
contact obtained with the relaxation method [1]. Each forming roll is described in cylindrical
coordinates, where the axis coordinate coincides with 0y axis (fig. 1).
h
2
h
2
h
2
h
2

h
2
1
cos
.
h
2
1
cos
.
r
c
v
r
c
x
i
i+1
( ) y z i i ,
( ) y z
i i +1 +1
,
h
z
y 0
r
c
x
m
a
x
r
c
v
m
in

Fig. 1. Building forming rolls geometries by the strip
reference surface obtained with relaxation method
The forming roll radii are calculated in each point y
i
of the axis by the formulas (3):
( )
( )
max min
2 cos
cx i cx i
i
h
r r z z

;
( )
( )
min min
2 cos
cv i cv i
i
h
r r z z

,
(3)
where r
cxi
and r
cvi
are the radii of convex and concave forming rolls in i
th
point respectively.
The area of a quadrilateral mesh element of a roll-strip contact is calculated by the expression (4)
obtained from a well known formula of vector product of diagonals for the quadrilateral mesh element:
( ) ( ) ( )
2 2 2
1 2 2 1 2 1 1 2 1 2 2 1
1
2
s y z y z x z x z x y x y + + . (4)
where x
i
, y
i
and z
i
are components of vector representation of i
th
diagonal of the mesh element (i = 1, 2) .
The projection area of a quadrilateral mesh element onto the plane 0xy is expressed by the formula (5):
1 2 2 1
1
2
xy
s x y x y . (5)
2
1
1
1
cos ;
1 tg
tg
i
i
i i
i
i i
z z
y y

+
+

A volume of a forming roll is found as a sum of n elementary volumes bounded between planes
perpendicular to roll axis (see fig. 2). The elementary volume is determined by the formula (6):
( ) ( )
2 2
1 1 1
3
i i i i i i i
V r r r r y y

+ + +
+ + . (6)
The corresponding worn out forming roll radii (see fig. 2) are calculated by the formula (7):
1
w
i worn i
i
V
r r
V
, (7)
where V
w
is a volume of worn out roll material in i
th
point,
3
m 1
]
.
... ...
1 2 3 i n
yn
y
1
r
1
r
2 r
3
r
i r
i+
1
r
i+
2
r
i+
3
r
n
V1V2 Vi
1 2 3 i n
y1 yn
... ... V1V2 Vi
r
i r
i+
1
r
i+
2
r
i+
3
r
1
r
2
r
3
r
n

Fig. 2. Calculation of a roll volume
2.3 Tool wear prediction
The roll-strip contact pressure is calculated by the forming force applied to the forming roll. The forming
force necessary for plastic bending the strip is calculated from principle of virtual work. The selected stress-
strain relation for the strip materi al is assumed to be pure rigid plastic shown in fig. 3 a). The simplified
analytical model of plastic strip bending with radius r is shown in fig. 3 b). The Kirchhoff Loves
hypotheses are assumed for the selected simplified analytical bending model.

y
-
y

0

a)
O
r
h

b)
Fig. 3. a) Stress-strain relation and b) Analytical model of formed strip radius
for the simplified wear modelling
The bending moment for pure rigid plastic deformation of the strip is calculated by the formula (8):
y
h
y
h
y
h
y
h
h
h
y y p
h
z z dz z dz z dz z M

4 2
1
2
1
2
0
2
2
2
0
2
0
2
2
2
2
0
, (8)
The strip forming angle increment (see fig. 3 b) is calculated as a difference between strip bend
angles before and after the roll-strip contact area.


The forming force is expressed with the formula (9):
( )
( )
max min
2max 2min 1max 1min
z z
f p
f
c M
F x x
z z


+
, (9)
where z
1min
, z
1max
are the minimal and maximal z coordinates of the roll-strip contact before for ming;
z
2min
, z
2max
are the minimal and maximal z coordinates of the roll-strip contact after forming;
x
min
, x
max
are the minimal and maximal x coordinates of the nodes in the strip bending area
corresponding to the considered forming roll stand;
c
f
is an empirical coefficient of dynamic resistance to forming.
The dependence of yield stress on the velocity of deformation has an exponential nature for steel [6].
This fact allows us to express the empirical coefficient of dynamic resistance to forming using the
formula (10):
1 2
1 exp( )
f f
c k k v + + , (10)
where k
1
= 1.5 and k
2
= 4.5 are the empirical coefficients;
v
f
is a forming velocity.
These values suit for the forming conditions chosen by Suzuki et al. [7] with velocities up to 0.5 m/s.
The average contact pressure is calculated with the formula (11):
c f xy
P F s

, (11)
where
xy
s

is the contact area projection onto plane 0xy.

The sliding velocity is calculated in each point of a roll-strip contact by the formulas (12):
v
scx i
= v
f

max cx i cx
r r v
cx i
;
v
scv i
= v
f
+ v
cv i

( )
min
2
cvi cv
r r ,
(12)
where v
scx i
, v
scv i
are the roll-strip sliding velocities on the forming surface of convex and concave
rolls respectively in the segment i of the surface;
v
cx i
, v
cv i
are the roll-strip linear velocities on the forming surface of convex and concave rolls
respectively in the segment i of the surface.
The sliding velocity obtained for each node of the mesh for roll-strip contact area is averaged by the area of each
mesh element in the following way. The average node sliding velocity is found for each element. Each
element of the roll-strip contact is subdivided into 2 triangles. The average sliding velocity is multiplied by
the area of each triangle. These products are summarized and divided by entire element area. The
resultant average roll-strip sliding velocity is used for tool wear calculation.
The time period for wear calculation is obtained by the formula (13):
formed f
t L v . (13)
The volume of worn out material is calculated for each element of the roll-strip contact by the formula (14):
*
w
w
V t s
t

. (14)
2.4 Example of tool wear prediction
As most of wear is accumulated on a concave forming roll, the concave forming roll was chosen to
estimate the accuracy of the proposed model. The obtained result was compared with the published
experimental data on the sliding contact region geometry [7] (see fig. 4 with the result comparison for
the concave roll of the first roll forming stand with a bended angle of formed profile
1
equal to 75 and
the inter-stand distance same to the experimental data as it has an effect on a roll-strip contact area).


Centre
Edge Edge
120
-30 30 -60 60 0
10
-10 F
o
r
m
i
n
g

d
i
r
e
c
t
i
o
n
2 b: Width of sheet, mm
Experimental data
Relaxation method
Bending part
On surface of concave roll
15
RH(0, 0, 0)
t = 3.0 mm (2
1 roll
) =150
Inter-stand distance = 400 mm Forming speed 30 m/min

Fig. 4. Comparison of the result with the experimental data [7]
Lack of coincidence of the calculated data with the experiment in the centre of the formed strip is not
essential for the proposed approximate tool wear model as the roll-strip sliding velocity is assumed to
be zero on the web of the formed channel section and does not affect the tool wear.

Fig. 5. Standard coil of strip dimensions
Wear of a concave forming roll
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
-59 -44 -30 -15 0 15 30 44 59
Roll axis related coordinate, mm
R
a
d
i
u
s

c
h
a
n
g
e
,

m

Fig. 6. Estimated wear along the concave forming roll
profile after forming standard coil of strip
If the values of b = 60 mm, t = 3.0 mm, standard coil outer diameter d
o
= 1850 mm and inner diameter
d
i
= 580 mm are substituted into the formula (15) the obtained length of coiled strip l is 808 m.
( )
2 2
2 2
2 2
0
1
ln
2 2 2 4 2 2 2 2 2 4
o i
d d
t
i i i i
d d d d t t t t t t
l
t

1 _
_ _ _
1 + + + + + + + +

1 , , ,
,]
. (15)


If one coil of mild steel strip shown in fig. 5 is formed according to experimental conditions [7] the total
wear estimated for the concave forming roll will be 1009.9 mm
3
for the roll geometry used in the
experiment [7]. Assumed that the forming rolls have surface hardness of 60 62 HRC and the surface
wear coefficient is 5 10
-4
in Archards wear equation.
The calculated forming force is 7.5 kN in the approximate model versus 10.17 kN measured for the
same forming conditions [7] that gives a 26% error in forming force calculation. Change of forming roll
profile due to wear is shown in fig. 6.
4 Discussion
The relaxation method allows finding the contact area very approximately. The actual length of the strip bend
is also longer than estimated with this method. This difference contributes to the 26% error in forming force
calculation along with the friction force that is neglected in this approximate model.
Also, t he proposed tool wear model does not take i nto account the non-linearity of roll-strip contact pressure
distribution. The dependence of forming force upon forming velocity is empirical. It depends on physical
properties of the formed material and should be tuned via field measurements to avoid an extrapolation error
of the proposed empirical formula based on results of experimental studies [6]. However some important
observations can be made concerning the total wear results.
The change of radius of the forming roll shown in diagram (fig. 6) depends on the roll radius value and the
width of the contact zone in a particular point of the roll. The larger is the roll radius in the contact point the
larger is the linear velocity of the roll surface that affects roll-strip sliding. The wider the roll-strip contact zone is
the longer the sliding through the roll-strip contact. These two effects significantly affect the final forming roll
radius. The wear is greater on the peripherals of the roll width due to higher relative sliding velocity and wider
contact area.
3 Conclusions and recommendations
The method of approximate tool wear analysis described in this paper is robust as it is based on
energy conservation principle. Clear geometrical assumptions and use of conventional Archards
model of physical wear make it reliable approach to quick tool wear estimation based on the results of
the CRF process analysis with a new relaxation method [1]. The surface contact wear coefficient
required for Archards model should be experimentally determined for each pair of contacting materials.
References
1. Ding, S.C., W.J.T. Daniel, and P.A. Meehan. A New Relaxation Method for Roll Forming Problems.
in III European Conference on Computational Mechanics (ECCM - 2006). 2006. Lisbon, Portugal.
2. Halmos, G.T., Roll Forming Handbook, ed. G.T. Halmos. 2006: CRC Press. 584.
3. Jendel, T., Prediction of wheel profile wear - comparisons with measurements. Wear, 2002. 253: p. 89-99.
4. Mtt, A., P. Vuoristo, and T. Mntyl, Friction and adhesion of stainless steel strip against tool
steels in unlubricated sliding with high contact load. Tribology International, 2001. 34: p. 779-786.
5. Pczelt, I. and Z. Mrz. Contact optimization problems associated with the wear process. in XXI
International Congress of Theoretical and Applied Mechanics (ICTAM 2004). 2004. Warsaw,
Poland: Springer Verlag.
6. Pearce, R., Sheet Metal Forming. The Adam Hilger series on new manufacturing processes and
materials, ed. J. Wood. 1991, Bristol, Philadelphia, New York: Adam Hilger.
7. Suzuki, H., et al., Experimental Investigation on Cold-Roll-Forming Process II: distribution of contact
pressure on interface between sheet metal and forming rolls in roll-forming process through
tandem mills. 1976, The Institute of Industrial Science, the University of Tokyo: Tokyo. p. 56.