ISEN 658, Homework 2: Deformation Processing

1. For a 2D stress state (σ3 = 0), identify all the points in the σ1 -σ2 space where we see
maximum deviation between the von Mises and Tresca yield criteria. Show that this
maximum deviation is about 15%.
2. Consider a material undergoing plane-strain plastic deformation. Show that the corre-
sponding stress state for this deformation is equivalent to pure shear under superimposed
hydrostatic pressure. Assume von Mises material.
3. Consider Mode I loading of a thin and infinite plate with a pre-existing crack of half-length
a = 25 mm and an applied stress σ0 of 50 MPa. Assuming that the plate is thin enough that
plane-stress condition exists at the surface (i.e., σzz , σxz , σyz = 0), the stress distributions
resulting from the stress concentration take the form:

pa
cos 2θ 1 − sin 2θ sin 3θ


σxx = σ0 2r 2
pa
cos 2θ 1 + sin 2θ sin 3θ


σyy = σ0 2r 2
pa
σxy = σ0 2r
sin 2θ cos 2θ cos 3θ
2

For a soft material like copper with a yield stress of 100 MPa, determine the size and
shape of the plastic zone ahead of the crack (red boundary in the figure). In other words,
plot the contour of the zone ahead of the crack in which material is expected to deform
plastically. One approach to solve the problem is to determine the principal stresses as a
function of r and θ, and then apply von Mises yield criterion to locate the boundary at
which the yielding condition is satisfied. Use Matlab, Mathematica or something similar
for calculations and plotting.
4. Consider forging of a material slab that can be safely approximated as a plane-strain
compression problem (compression along y axis and no flow taking place along the z axis).
Assume von Mises plasticity.

• Determine the effective plastic strain ε̄ as a function of εy .

• Determine the effective flow stress σ̄ as a function of σy .
• Say, the true stress-strain curve of the material is given by the following relationship:
σ = 500ε0.3 (in MPa). Neglecting friction, what is the work done in reducing the slab
height by 50%? Consider the slab dimensions of 25 mm (in y) × 500 mm (in x) ×
2000 mm (in z).
• Plot the expected force-displacement curve (i.e., force along y vs. displacement along
y) during the forging process. Based on this, comment on the minimum forge press
capacity (in N) required to accomplish the job.

5. Consider axisymmetric wire drawing of a non-strain-hardening material (i.e., assume ma-

terial is perfectly plastic) and ignore any frictional effects. Comment on the minimum

1
 number of drawing passes needed to achieve 90% reduction in the wire cross-sectional area.
Does your answer change if the material is strain-hardening? Explain why/why not.
6. Consider axisymmetric wire drawing of a strain-hardening material described by the fol-
lowing (true) stress-strain relationship: σ = 200 + 600ε0.4 (stress units in MPa). Initial
diameter of the wire is 5 mm and ignore any friction-related effects. Consider a multi-step
drawing scheme comprising of 5 drawing passes, where wire cross-sectional area is reduced
by 50% in each pass. Say, you are in charge of selecting a motor that can be used to draw
the wire at a nominal drawing speed of 0.1 m/s. What should be the minimum power the
motor should have to successfully draw the wire?
7. Now, consider plane-strain drawing of a thin plate of initial length l0 , width w and thickness
t0 , where w ≫ t0 . The geometry of plane-strain drawing is similar to the axisymmetric
wire drawing problem that we discussed in the class except that the width of the plate
now remains unchanged (i.e., strain along the width direction is zero). Assume non-strain-
hardening material (i.e., perfectly plastic) and ignore any friction-related effects. For this
case, what is the maximum fractional reduction in thickness that is possible in a single
drawing pass?
8. • For the rolling problem that we discussed in the class, assuming von Mises yield
√ show that the normal force (P ) acting on the rollers is given by P ≈
criterion,
1.15σY R∆hw, where w is the workpiece width, ∆h is the thickness reduction due
to rolling, and R (≫ ∆h) is the radius of the rollers. Assume plane-strain conditions
(i.e., width remains unchanged), ignore friction, and perfectly plastic behavior for the
workpiece material with a yield strength of σY .
• Now, estimate the normal force on the rollers based on the Tresca criterion.
• Show that the normal force on the rollers can be reduced by applying tension on the
strip (i.e., by pulling the strip along the horizontal direction). Derive an expression
for the normal force P if this applied tension is half the yield stress.
9. Consider plane-strain rolling of a metal that is described by the following (true) stress-strain
relationship: σ = 600ε0.5 (stress units in MPa). Assume that the material has been rolled
to 30% reduction in its thickness (and therefore, would have hardened in this process). Say,
we are interested in predicting how this rolled sheet will behave under a uniaxial tension
test. Estimate: (1) stress at which the rolled sheet will plastically yield, and (2) strain at
which the material will form a neck.
10. As a follow-up to Problem 4, comment on by how much percent does the minimum press
capacity requirement (in N) increase in the presence of friction? Consider two cases: full
sliding friction (with µ = 0.1) and full sticking friction. You may assume the material to
be perfectly plastic (meaning that it yields at some constant σY , say 300 MPa, and does
not harden with strain).
11. Although we have assumed that the friction coefficient µ is a known constant in all our
analyses, often it is not known a priori and has to be determined somehow. Let’s say you
were to find the value of µ for a given die-workpiece material combination under plane-
strain compression. Propose how would you go about finding µ if the friction condition
at the interface is that of sliding (Coulomb) friction. Assume you have the capability

2
 to measure the average (normal) pressure and vary the dimensions of the workpiece, and
nothing more. Of course, you have the theoretical results (from the class) in your tool kit.

12. Although we have neglected friction in our analysis of rolling, it plays a very important
role in sheet manufacturing process. In fact, it is the friction in the roll bite that drives the
sheet through the rolls (without friction, the material would just slip, much in a same way
that we cannot walk or drive without some level of friction). A consequence of friction is
that about the midway along the roll-sheet contact (called the neutral point where the roll
surface velocity is equal to the velocity of the material), a friction hill develops, similar
to the case of plane-strain forging. In the presence of these frictional effects, the average
h
roll pressure is given by Proll = 2k µL (exp(µL/h) − 1), where k is the material’s shear yield
stress,√h is the average height of the material in the roll bite, µ is friction coefficient and
L ≈ R∆h is the contact length, with R and ∆h being the roll radius and thickness
reduction, respectively.
(a) Based on the above, show that the friction hill is larger if h is small and R is large. In
fact, the large force pushing back on the rolls under these conditions can result in bending
and deflection of the rolls. If this deflection is sufficient, it can make rolling reductions of
very thin stock almost impossible. For example, that is why thin aluminum foil is produced
by rolling multiple stacks together, and this is also why usually only one side of the foil
(one in contact with the roll) is shiny and not both.
(b) Given all of this, can you suggest at least 4 strategies for reducing the roll forces?

3
 ISEN 658, Homework 2 Solutions: Deformation Processing

1. For a 2D stress state (σ3 = 0), identify all the points in the σ1 -σ2 space where we see
maximum deviation between the von Mises and Tresca yield criteria. Show that this
maximum deviation is about 15%.
Solution

Figure 1: Black curve: von Mises criterion, green: Tresca criterion, red dotted lines represent the
stress states where we observe maximum deviation between the von Mises and Tresca criteria.
The three lines are: σ1 = 2σ2 , σ2 = 2σ1 , and σ1 = −σ2 .

Let us calculate the deviation between Tresca and von Mises criteria for the case when
σ1 = −σ2 (referred to as “pure shear”).
Von Mises criterion:
1
(σ1 − σ2 )2 + (σ2 − σ3 )2 + (σ3 − σ1 )2 = σY2

2
σY
σ1 = √
3
Tresca criterion:
1 σY
τmax = Max {|σ1 − σ2 | , |σ2 − σ3 | , |σ3 − σ1 |} =
2 2
σY
σ1 =
2
σY σ
√ − Y
2
The difference between σ1 predictions by Tresca and von Mises is 3
σY = 0.154 or ∼ 15%.
2

1
 Follow a similar approach for the other two conditions (σ1 = 2σ2 , σ2 = 2σ1 ) as well to
obtain the same result. Alternatively, you can also do the following tensor decomposition
to arrive at the same conclusion. If we consider the stress state for pure shear,
 
σ1 0 0
σij =  0 −σ1 0
0 0 0
   
σ1 0 0 0 0 0
σij = 0 σ1 0 + 0
   −2σ1 0 
0 0 σ1 0 0 −σ1
You can see from the above, the deviatoric component of the stress tensor corresponds to
the case where one principal stresses is twice the other.

2. Consider a material undergoing plane-strain plastic deformation. Show that the corre-
sponding stress state for this deformation is equivalent to pure shear under superimposed
hydrostatic pressure. Assume von Mises material.
Solution
From the class, we know that stress state for a plane-strain plastic deformation problem is
given by:  
σ1 0 0
σij = 0 σ2 0 
0 0 (σ1 + σ2 )/2

This can be written as a sum of hydrostatic (first tensor) and deviatoric (second tensor)
stress components as follows:
   
(σ1 + σ2 )/2 0 0 (σ1 − σ2 )/2 0 0
 0 (σ1 + σ2 )/2 0 + 0 −(σ1 − σ2 )/2 0
0 0 (σ1 + σ2 )/2 0 0 0

From the deviatoric stress component, you can see that the first diagonal term is equal to
the second diagonal term in magnitude but opposite in sign, which is equivalent to the case
of pure shear.

3. Consider Mode I loading of a thin and infinite plate with a pre-existing crack of half-length
a = 25 mm and an applied stress σ0 of 50 MPa. Assuming that the plate is thin enough that
plane-stress condition exists at the surface (i.e., σzz , σxz , σyz = 0), the stress distributions
resulting from the stress concentration take the form:

pa
cos 2θ 1 − sin 2θ sin 3θ


σxx = σ0 2r 2
pa
cos 2θ 1 + sin 2θ sin 3θ


σyy = σ0 2r 2
pa
σxy = σ0 2r sin 2θ cos 2θ cos 3θ
2

For a soft material like copper with a yield stress of 100 MPa, determine the size and shape
of the plastic zone ahead of the crack (red boundary in the figure). In other words, plot the

2
 contour of the zone ahead of the crack in which material is expected to deform plastically.
One approach to solve the problem is to determine the principal stresses as a function of r
and θ, and then apply von Mises yield criterion to locate the boundary at which the yielding
condition is satisfied. Use Matlab, Mathematica or something similar for calculations and
plotting.
Solution
We first use the equations for σxx , σyy and σxy to determine the principal stresses σ1 and
σ2 as a function of r and θ (note that the third principal stress σ3 = 0). Figure below
shows the contour
q plot (units in MPa) for the left-hand-side quantity in von Mises plasticity
criterion: σ̄ = 12 {(σ1 − σ2 )2 + (σ2 − σ3 )2 + (σ3 − σ1 )2 }. σ̄ is highest in the immediate
neighborhood of the crack tip and decays as you move further away from the crack. The
locus where σ̄ = σY (where σY is the yield stress) represents the boundary of the plastic
zone; that is, material located within this boundary is yielded plasticially whereas material
that outside the boundary is only elastically loaded. If we approximate the plastic zone to
be a circle with a radius r, then you can see that r ∼ 1.6 mm.

Figure 2: Contour plot for σ̄. The red line bounds the plastic zone around the crack tip.

4. Consider forging of a material slab that can be safely approximated as a plane-strain

compression problem (compression along y axis and no flow taking place along the z axis).
Assume von Mises plasticity.

• Determine the effective plastic strain ε̄ as a function of εy .

• Determine the effective flow stress σ̄ as a function of σy .
• Say, the true stress-strain curve of the material is given by the following relationship:
σ = 500ε0.3 (in MPa). Neglecting friction, what is the work done in reducing the slab

3
 height by 50%? Consider the slab dimensions of 25 mm (in y) × 500 mm (in x) ×
2000 mm (in z).
• Plot the expected force-displacement curve (i.e., force along y vs. displacement along
y) during the forging process. Based on this, comment on the minimum forge press
capacity (in N) required to accomplish the job.

Solution
Since there is no flow in the z direction, εz = 0. Therefore, εx = −εy .
The effective strain is given by:
q
ε̄ = 2/3(ε21 + ε22 + ε23 )

Plugging in the values of εx , εy and εz , we get ε̄ = √2 εy .

3
From Levy-Mises flow rules, we have

dεz = dλ(σz − (σx + σy )/2)

Since dεz = 0, we get σz = (σx + σy )/2. Since σx = 0, σz = σy /2.

Effective stress is given by:
p
σ̄ = 1/2[(σ1 − σ2 )2 + (σ2 − σ3 )2 + (σ1 − σ3 )2 ]
√
3
Plugging in the values, we get σ̄ = 2 y
σ .
Work done (per unit volume) during plastic deformation is given by:
Z ε̄f
W = σdε
0

For 50% reduction of height, ε̄f = √2 ln(2) = 0.8. Therefore,

3

Z 0.8 Z 0.8
W = σdε = 500ε0.3 dε = 287 MJ/m3
0 0

Since the sample volume is 0.025 × 0.5 × 2 m3 , work done is 7.2 MJ.
Let us try to estimate the force at height h. Consider the initial sample dimensions are
height h0 , breadth b0 (dimension along x) and width w (dimension along z). Since w is
constant for this problem, the sample breadth b at height h is h0 b0 /h. The displacement
(δy) along the y direction is nothing but h0 − h.
We have:
F 2
= σy = √ σ̄
(w)(b) 3
F 2
= √ 500ε̄0.3
(w)(b) 3

4
  0.3
F 2 2
= √ 500 √ εy
(w)(b) 3 3
  0.3
F 2 2 h0
= √ 500 √ ln
(w)(b) 3 3 h
  0.3
Fh 2 2 h0
= √ 500 √ ln
(w)(h0 b0 ) 3 3 h
  0.3
F (h0 − δy) 2 2 h0
= √ 500 √ ln
(w)(h0 b0 ) 3 3 h0 − δy

Now you have an equation for force F as a function of displacement δy along the y direction,
which is plotted in Fig. 3. For 50% reduction in height, δy = 12.5 mm. From the figure,
the minimum force capacity of the press should be around 1,100 MN.

Figure 3: Force vs. displacement curve in plane-strain forging.

5. Consider axisymmetric wire drawing of a non-strain-hardening material (i.e., assume material

is perfectly plastic) and ignore any frictional effects. Comment on the minimum number of
drawing passes needed to achieve 90% reduction in the wire cross-sectional area. Does your
answer change if the material is strain-hardening? Explain why/why not.
Solution
Maximum cross-sectional reduction that is possible in a single pass is 63% (discussed in
class). Therefore, the number of passes (n) needed to reach 10% of initial area (90%
reduction) is:
(0.37)n = 0.1

Solving the above equation gives n ≈ 2.3. Hence at least 3 drawing passes are needed. This
answer is independent of whether the material is perfectly-plastic or strain-hardening. For

5
 the case of perfectly-plastic material, you need to replace σavg (average flow stress) with σY
(yield stress) in our derivation for computing the maximum cross-sectional reduction (rmax )
that is possible to achieve in a single pass. When you do so, you will notice that rmax again
comes out to be 0.63.

6. Consider axisymmetric wire drawing of a strain-hardening material described by the following

(true) stress-strain relationship: σ = 200 + 600ε0.4 (stress units in MPa). Initial diameter
of the wire is 5 mm and ignore any friction-related effects. Consider a multi-step drawing
scheme comprising of 5 drawing passes, where wire cross-sectional area is reduced by 50%
in each pass. Say, you are in charge of selecting a motor that can be used to draw the wire
at a nominal drawing speed of 0.1 m/s. What should be the minimum power the motor
should have to successfully draw the wire?
Solution
For this, you need to figure out in which pass (out of the five passes), the drawing force (in
N) is the highest. Then, the minimum power the motor should have can be calculated from
this force multiplied by the drawing speed (0.1 m/s).
From the class, we know that:
 
Fi Ai−1
= σi = σavg,i ln
Ai Ai

where i represents the pass number and all other quantities have their standard meanings.
Now, let us calculate the average flow stress σavg,i in each pass (units in MPa):
R ln(2)
0
σdε
σavg,1 = = 570.12
ln(2) − 0
R ln(4)
ln(2)
σdε
σavg,2 = = 806.64
ln(4) − ln(2)
R ln(8)
ln(4)
σdε
σavg,3 = = 946.37
ln(8) − ln(4)
R ln(16)
ln(8)
σdε
σavg,4 = = 1054.6
ln(16) − ln(8)
R ln(32)
ln(16)
σdε
σavg,5 = = 1145.5
ln(32) − ln(16)

Note that ln(2) comes from the fact that Ai = (Ai−1 )/2 in this example. Based on this,
you can calculate the drawing force Fi during each pass: F1 = 3, 879 N, F2 = 2, 744 N,
F3 = 1, 610 N, F4 = 896 N and F5 = 487 N. The maximum drawing force (3,879 N) is
therefore realized in the first pass. So, the minimum motor capacity needed is 387.9 W or
0.52 hp.

6
7. Now, consider plane-strain drawing of a thin plate of initial length l0 , width w and thickness
t0 , where w ≫ t0 . The geometry of plane-strain drawing is similar to the axisymmetric
wire drawing problem that we discussed in the class except that the width of the plate
now remains unchanged (i.e., strain along the width direction is zero). Assume non-strain-
hardening material (i.e., perfectly plastic) and ignore any friction-related effects. For this
case, what is the maximum fractional reduction in thickness that is possible in a single
drawing pass?
Solution
Following a similar approach that we discussed in class:
F1 l1
Wext = = σ1
A1 l1
Wint = σavg ε̄f
σ1 = σavg ε̄f
For plane-strain condition, ε̄ = √23 ε1 . The maximum limit on % reduction in cross sectional
area occurs when σ1 = σavg . Let us say ε∗1 represents the strain (along the pulling direction)
for this special condition. Therefore,
2
σ1 = σavg √ ε∗1
3
√
This gives, ε∗1 = 3/2. We know ε1 = ln(l1 /l0 ) = ln(t0 /t1 ); note that l1 /l0 = t0 /t1 for plane-
strain drawing since width is constant. If we define the fractional reduction of thickness
using r, then you can show:  
1
ε1 = ln
1−r
√
Solving for ε∗1 = 3/2 results in a r of 0.57, i.e., maximum percent reduction in the thickness
that can be achieved in a single pass is 57%.

8. • For the rolling problem that we discussed in the class, assuming von Mises yield √
criterion,
show that the normal force (P ) acting on the rollers is given by P ≈ 1.15σY R∆hw,
where w is the workpiece width, ∆h is the thickness reduction due to rolling, and R
(≫ ∆h) is the radius of the rollers. Assume plane-strain conditions (i.e., width remains
unchanged), ignore friction, and perfectly plastic behavior for the workpiece material
with a yield strength of σY .
• Now, estimate the normal force on the rollers based on the Tresca criterion.
• Show that the normal force on the rollers can be reduced by applying tension on the
strip (i.e., by pulling the strip along the horizontal direction). Derive an expression for
the normal force P if this applied tension is half the yield stress.
Solution
Normal stress σ2 estimated from the Tresca criterion will be simply σY (since σ1 = 0 and
σ3 = σ2 /2, which means that the highest principal stress is zero and the smallest principal

7
 stress is σ2 , which√is negative). √
Normal force estimated from the Tresca criteria will be
therefore P = |σ2 | R∆hw = σY R∆hw.
If there is a pulling stress that is equal to half the yield stress, the principal stresses are
now σ1 = σY /2, σ2 , and σ3 = (σY /2 + σ2 )/2). Obviously, σ1 (+ve) is the highest principal
stress and σ2 (-ve) is the smallest principal stress. Based on the Tresca criterion:
1 σY
τmax = |σY /2 − σ2 | = (1)
2 2
Therefore, for this case, σ2 = −σY /2. Hence, if we apply a pull stress of σY /2 in along the
horizontal rolling direction, then the normal load acting on the rollers will reduce by half
(according to Tresca criterion and ignoring any frictional effects).

9. Consider plane-strain rolling of a metal that is described by the following (true) stress-strain
relationship: σ = 600ε0.5 (stress units in MPa). Assume that the material has been rolled
to 30% reduction in its thickness (and therefore, would have hardened in this process). Say,
we are interested in predicting how this rolled sheet will behave under a uniaxial tension
test. Estimate: (1) stress at which the rolled sheet will plastically yield, and (2) strain at
which the material will form a neck.
Solution
During rolling, the material undergoes strain-hardening because of plastic deformation.
Hence, if you conduct a tension test on a rolled sheet, it will yield at a stress that is higher
than the material’s original yield stress in its undeformed state. To estimate the rolled
sheet’s yeild stress, first compute the effective strain (ε̄f ) that has been put into the sheet
sample.
In this example, ε̄f = √2 εy = √2 | ln(0.7)| = 0.41.
3 3

Uniaxial stress-strain relationship for the material is given by: σ = 600ε0.5 . Since for the
uniaxial case, σ̄ = σ and ε̄ = ε (where σ and ε represent true stress and true strain along
the axial or loading direction), you can write this as σ̄ = 600ε̄0.5 . Replacing ε̄ with a value
of 0.41 will result in a value of 358 MPa for σ̄. This basically means that the effective
flow stress σ̄, or alternatively, the uniaxial yield stress of the sheet material that has been
already deformed to an effective plastic strain of 0.41 is 358 MPa.
From Considere’s criterion, the strain at which the original (undeformed) material will exhibit
necking is 0.5 (same as the exponent in the stress-strain power-law relationship). Given
that rolling to 30% thickness reduction is equivalent to deforming the sample under uniaxial
tension to a strain of 0.41, the additional strain that the rolled sheet can accommodate
before necking is 0.09.

10. As a follow-up to Problem 4, comment on by how much percent does the minimum press
capacity requirement (in N) increase in the presence of friction? Consider two cases: full
sliding friction (with µ = 0.1) and full sticking friction. You may assume the material to be
perfectly plastic (meaning that it yields at some constant σY , say 300 MPa, and does not
harden with strain).
Solution

8
 For sliding friction:  
hw µb
P = 2k (e h − 1)
µ
√
k = σY / 3. The maximum value of P will be realized close to the end of the 50% height
reduction, that is when h = h0 /2. At this stage, b = 2b0 .
  
h0 w 4µh0
Pmax = 2k e h0
−1
2µ

From this, you get Pmax = 2.58 × 1011 N, so this is the minimum force capacity the forging
press needs to have to successfully forge the sample.
For the sticking friction case,  
p x
= 1+
2k h
Therefore, force can be estimated from:
Z b/2
P =2 p(wdx)
0

 
b
P = 2kbw 1 +
4h
Similar to the previous case, P will be maximum close to the end of the process, i.e., when
h = h0 /2 and b = 2b0 . Therefore,
 
b0
Pmax = 4kb0 w 1 + = 1.45 × 1010 N
h0

The minimum press capacity needed is therefore 1.45 × 1010 N.

11. Although we have assumed that the friction coefficient µ is a known constant in all our
analyses, often it is not known a priori and has to be determined somehow. Let’s say you
were to find the value of µ for a given die-workpiece material combination under plane-strain
compression. Propose how would you go about finding µ if the friction condition at the
interface is that of sliding (Coulomb) friction. Assume you have the capability to measure
the average (normal) pressure and vary the dimensions of the workpiece, and nothing more.
Of course, you have the theoretical results (from the class) in your tool kit.
Solution
Load (P ) required for plane strain compression under the condition of sliding friction is:
P = 2k(hw/µ) (exp(µb/h) − 1).
We can perform two experiments with different dimensions of the workpiece (say, h1 and b1 ;
and h2 and b2 ). Let us say the corresponding load values for these samples are P1 and P2 ,
respectively. Then,

P1 = 2k(h1 w/µ){exp(µb1 /h1 ) − 1}

9
 P2 = 2k(h2 w/µ){exp(µb2 /h2 ) − 1}
P1 h1 exp(µb1 /h1 ) − 1
=
P2 h2 exp(µb2 /h2 ) − 1
We can estimate the value of µ from the above equation.

12. Although we have neglected friction in our analysis of rolling, it plays a very important role
in sheet manufacturing process. In fact, it is the friction in the roll bite that drives the sheet
through the rolls (without friction, the material would just slip, much in a same way that we
cannot walk or drive without some level of friction). A consequence of friction is that about
the midway along the roll-sheet contact (called the neutral point where the roll surface
velocity is equal to the velocity of the material), a friction hill develops, similar to the case
of plane-strain forging. In the presence of these frictional effects, the average roll pressure is
h
given by Proll = 2k µL (exp(µL/h) − 1), where k is the material’s shear yield stress, h is the
√
average height of the material in the roll bite, µ is friction coefficient and L ≈ R∆h is the
contact length, with R and ∆h being the roll radius and thickness reduction, respectively.
(a) Based on the above, show that the friction hill is larger if h is small and R is large. In
fact, the large force pushing back on the rolls under these conditions can result in bending
and deflection of the rolls. If this deflection is sufficient, it can make rolling reductions of
very thin stock almost impossible. For example, that is why thin aluminum foil is produced
by rolling multiple stacks together, and this is also why usually only one side of the foil
(one in contact with the roll) is shiny and not both.
(b) Given all of this, can you suggest at least 4 strategies for reducing the roll forces?
Solution
(a) Use any plotting software (e.g., MATLAB, Mathematica, etc.) to plot Proll vs. h for
different values of h (for this, keep R and the ratio ∆h/h as constants); and Proll vs. R for
different values of R (keep ∆h and h as constants) to understand the dependence of Proll
on these two variables.
(b) Reduction in the roll forces can be achieved by: (1) reducing R; (2) taking small
reductions in each pass (i.e., small ∆h/h); (3) decreasing k, i.e., by decreasing the yield/flow
stress of the material (this can be done by rolling the material at high temperatures;
this practice is called “hot rolling”); or (4) by reducing the friction coefficient µ through
lubrication.

10