Fundamentals Wagoner&Chenot ProblemSolutions

CHAPTER 1 - PROBLEM SOLUTIONS
The plot below of load vs. extension was obtained using a specimen (shown in the following figure)
of an alloy remarkably similar to the aluminum-killed steel found in automotive fenders, hoods,
etc. The crosshead speed, v, was 3.3x10-4 inch/second. The extension was measured using a 2"
extensometer as shown (G). Eight points on the plastic part of the curve have been digitized for
you. Use these points to help answer the following questions.
900
800
700
(0.10, 630 )
(0.50, 745 )
(0.30, 729)
600
(0.20, 699)
500
400
(0.40, 741.5 )
(0.004, 458)
(0.80, 440 )
(0.0018, 405 )
300
0.03"
D = 3.3 "
0.5"
1.
PROFICIENCY PROBLEMS
Load, pounds
A.
200
G = 2.0"
100
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Extension, inches
a. Determine the following quantities. Do not neglect to include proper units in your answer.
Yield stress
Ultimate tensile strength
Uniform elongation
Engineering strain rate
Young's Modulus
Total elongation
Post-uniform elongation
b. Construct a table with the following headings, left-to-right: Extension, load, engineering
strain, engineering stress, true strain, true stress. Fill in for the eight points on graph.
What is the percentage difference between true and engineering strains for the first point?
(i.e., % = ______ x 100)
What is the percentage difference between true and engineering strains for the last point?
c. Plot the engineering and true stress-strain curves on a single graph using the same units.
Page 2
Fundamentals of Metal Forming - Solution Manual
Chapter 1
d. Calculate the work-hardening rate graphically and provide the ln-ln plot along with the
value of n. How does n compare with the uniform elongation in Part a? Why?
e. A second tensile test was carried out on an identical specimen of this material, this time
using a crosshead speed of 3.3x10-2 inch/second. The load at an extension of 0.30 inch
was 763.4 lb. What is the strain-rate sensitivity index, m, for this material?
SOLUTION:
a.
y =
458 lbs = 30,500 psi E = 27,000 psi = 30 x 106 psi

.030" x 0.5"
0.0018"/2.0"
UTS =
745 lbs
= 49,700 psi
.030" x .5"
e t = 0.80" = 0.40 or 40%

2.0"
e u = 0.5" = 0.25 or 25%

2.0"
e pu = e t e u = 0.40 0.25 = 0.15 or 15%
4 inch/s
e = 3.3 x 10
= 10 4/s
3.3"
b.
c.
True
Strain
True Stress
% Error
eng/true
strain
27000
0.001
27024
0.04%
30533
0.010
30839
0.50%
0.050
42000
0.049
44100
2.48%
0.100
46600
0.095
51260
4.92%
Eng. Strain Eng. Stress
Extension
Load
0.0018
405
0.001
0.02
458
0.010
0.1
630
0.2
699
0.3
729
0.150
48600
0.140
55890
7.33%
0.4
741.5
0.200
49433
0.182
59320
9.70%
0.5
745
0.250
49667
0.223
62083
12.04%
0.8
440
0.400
29333
0.336
41067
18.88%
Chapter 1
Page 3
65000
Stress (psi)
True Stress / Strain

55000
Engineering
Stress / Strain
45000
35000
25000
0.1
0.2
0.3
0.4
0.5
Strain
d.
11.2
Ln. (true stress)
11
10.8
10.6
n=0.13
(all data)
10.4
n = 0.225
(less extreme points)
10.2
10
-8
-7
-6
-5
-4
-3
-2
-1
Ln. (true strain)
The n value (slope of the ln-ln plot) is as follows:

All points: n = 0.13, Wth first and last removed: n = 0.225
The first point must be removed because the elastic strain is a significant part of the total strain and the
last point is meaningless because necking means that the formula to find t, and t cannot be used.
.225 differs from 0.25 because n is true strain so e.225 - 1 = 0.25 = uniform elongation.
Page 4
m=
e.
2.
Chapter 1
lb
ln 763.4
ln p 2/p 1
729 lb
=
= ln 1.047 = .046 = 0.010
ln v2/v1 ln 3.3 x 102/s
ln 100
4.605
3.3 x 104 /s
Starting from the basic idea that tensile necking begins at the maximum load point, find the true
strain and engineering strain where necking begins for the following material laws. Derive a
general expression for the form and find the actual strains.
n
a. = k ( + o )
= 500 ( + 0.05) 0.25
n
b. = o + k ( + o )
= 100 + 500 ( + 0.05) 0.25 (Ludwik)
-B
c. = o (1 - Ae )
= 500 1 - 0.6 exp (-3)
(Voce) 2
d. = o
= 500
(Ideal)
e. = o + k
f.
= 250 + 350
= k sin (B)
= 500 sin (2)
(Swift) 1
(Linear)
(Trig)
SOLUTION:
= k ( + o ) n
a.
d = nk ( + ) n1 = k( + ) n =
o
o
d
n = u + o , u = n o
for
b.
o = 0.05, n = 0.25 u = 0.20
= o + k( + o) n
d
= nk ( + o)n1 = o + k( + o)n =
d
o + k( + o)
n1
+ o n = 0
This is transcendental, so it cannot be solved algebraically.

Let's solve it numerically by Newton's Method for the special case n = 0.25, o = 0.05, o = 100, k =
500.
1
2
H. W. Swift: Plastic Instability Under Plane Stress, J. Mech. Phys. Solids, 1952, Vol. 1, p.1.
E. Voce: The Relationship Between Stress and Strain For Homogeneous Deformation, J. Inst. Met: 1948, Vol. 74, p. 537-562, 760.
E. Voce: The Engineer: 1953, Vol. 195, p.23.
E. Voce: Metallurga: 1953, Vol. 51, p. 219.
Chapter 1
Page 5
F() = o + k( + o) n1 + on = 0
F() = k(n2)( + o )n2 + o n + k( + o) n1
Start from a trial of eu = 0.20 (from Part b)
Step (i)
u(i)
F[u(i)]
F'[u(i)]
u(i+1)
0.20
100
1,414
0.129
0.129
-29
3,078
0.138
0.138
-8.5
2,762
0.142
So, eu 0.142
c.
= o(1AeB)
d
= BAoe B = o(1Ae B) =
d
where X = Ae-B
BX = 1-X
X=
1
1
1
or ln X = ln
ln A B = ln
1+B ,
1+B
1+B ,
u =
e.
u =
1
ln 0.6(4) = 0.29
3
d
= 0 = o =
= o , d
= o + k,
for o = 250, k = 350,

f.
1
ln A
1+B
1
1
1
ln A ln
= ln A(1+B)
B
1+B
B
for A = 0.6 B = 3:
d.
B = ln
(Never stable for constant o .
u = 0
ko
d
= k = o + k = =
d
k
,
u =
350250
= 0.29
350
= k sin B
d
= kB cos B = k sin B
d
,
B = tan B,
1
1
tan B
B
Page 6
for B = 2, k = 500 ,
3.
Chapter 1
1
tan1 2 = 0.22
2
What effect does a multiplicative strength coefficient (for example k in the Hollomon Law, k in
Problem 2.a., or o in Problem 2.c.) have on the uniform elongation?
SOLUTION:
No effect. Because it is only the ratio of strength in one part of the tensile test (i.e. in the neck) to
another (outside the neck), multiplication of has no effect on stability.
4.
For each of the explicit hardening laws presented in Problem 2, calculate the true stress at =
0.05, 0.10, 0.15, 0.20, 0.25 and plot the results on a (ln -ln ) figure. Use the figure to calculate
a best-fit n value for each material and compare this with the uniform strain calculated in Problem
2. Why are they different, in view of Eq. 1.16?
5.
For each of the explicit hardening laws presented in Problem 2, plot the engineering stress-strain
curves and determine the maximum load point graphically. How do the results from this
procedure compare with those obtained in Problems 2 and 4?
SOLUTIONS:
See table and plots. Compare u and n from ln-ln plots
Equation
u (Problem 2)
u (Problem 4)
u (Problem 5)
(from max load)
0.20
0.17
0.20
0.14
0.13
0.14
0.29
0.24
0.29
0.00
0.00
0.00
0.29
0.14
0.29
0.22
0.75
0.22
d ln
(=n)
is not a constant. Only this quantity
The results are different from Problems 2 and 4 because d ln
d
=
at the point at which d
is important, not an average of this quantity over a large range of strains.
The results from Problems 2 and 5 are identical, whether Considere's Criterion is used mathematically
(Problem 2) or whether the hardening equation is plotted in engineering units and the maximum load is
found.
2
520 ( + 0.05) 0.25 520
(at two rates) =
=
0.25
1
500
500 ( + 0.05)
Chapter 1
Page 7
Problem 1-4
Strain
Stress
Stress
Stress
Stress
Stress
Stress
Part a
Part b
Part c
Part d
Part e
Part f
0.05
281
381
242
500
268
155
0.1
311
411
278
500
285
294
0.15
334
434
309
500
303
405
0.2
354
454
335
500
320
476
0.25
370
470
358
500
338
500
ln stress
ln stress
ln stress
ln stress
ln stress
ln stress
ln strain
Part a
Part b
Part c
Part d
Part e
Part f
-2.996
5.639
5.943
5.488
6.215
5.589
5.040
-2.303
5.740
6.019
5.627
6.215
5.652
5.683
-1.897
5.812
6.074
5.732
6.215
5.712
6.003
-1.609
5.868
6.117
5.815
6.215
5.768
6.164
-1.386
5.914
6.153
5.881
6.215
5.822
6.215
slope (n)
0.17
0.13
0.24
0.00
0.14
0.75
(Figure for Problem 1-4 follows.)
6.4
d
6.2
ln (true stress)
5.8
a
5.6
a
b
c
d
e
f
c
5.4
5.2
5
-3
-2.5
-2
ln (true strain)
-1.5
-1
Page 8
Chapter 1
500
engineering stress
400
300
200
a
b
c
d
e
f
100
0
0
0.1
0.2
0.3
0.4
0.5
engineering strain
Figure for Problem 1-4 (upper), for Problem 1-5 (lower).
Problem 1-5
True
Eng.
Eng. Stress Eng. Stress
Eng. Stress Eng. Stress Eng. Stress
Eng. Stress
Strain
Strain
Part a
Part b
Part c
Part d
Part e
Part f
0.01
0.01
245.0
344.0
206.8
495.0
251.0
31.1
0.02
0.02
252.1
350.1
213.2
490.1
251.9
61.4
0.03
0.03
258.1
355.1
219.1
485.2
252.8
90.9
0.04
0.04
263.1
359.2
224.8
480.4
253.6
119.5
0.05
0.05
267.5
362.6
230.0
475.6
254.5
147.0
0.06
0.06
271.2
365.4
234.9
470.9
255.2
173.3
0.07
0.07
274.4
367.6
239.5
466.2
255.9
198.5
0.08
0.08
277.1
369.5
243.7
461.6
256.6
222.4
0.09
0.09
279.5
370.9
247.7
457.0
257.3
244.9
0.1
0.11
281.6
372.0
251.3
452.4
257.9
265.9
0.11
0.12
283.3
372.9
254.7
447.9
258.4
285.5
0.12
0.13
284.8
373.4
257.8
443.5
259.0
303.6
0.13
0.14
286.0
373.8
260.7
439.0
259.5
320.1
0.14
0.15
287.0
373.9
263.3
434.7
259.9
334.9
0.15
0.16
287.8
373.9
265.7
430.4
260.4
348.2
0.16
0.17
288.4
373.6
267.9
426.1
260.8
359.7
0.17
0.19
288.9
373.3
269.8
421.8
261.1
369.7
0.18
0.20
289.2
372.7
271.6
417.6
261.4
377.9
0.19
0.21
289.4
372.1
273.2
413.5
261.7
384.4
0.2
0.22
289.5
371.3
274.6
409.4
262.0
389.3
Chapter 1
6.
Page 9
0.21
0.23
289.4
370.5
275.8
405.3
262.2
392.6
0.22
0.25
289.2
369.5
276.8
401.3
262.4
394.2
0.23
0.26
289.0
368.4
277.7
397.3
262.6
394.1
0.24
0.27
288.6
367.3
278.4
393.3
262.7
392.5
0.25
0.28
288.2
366.1
279.0
389.4
262.8
389.4
0.26
0.30
287.7
364.8
279.5
385.5
262.9
384.8
0.27
0.31
287.1
363.4
279.8
381.7
263.0
378.7
0.28
0.32
286.4
362.0
280.0
377.9
263.0
371.2
0.29
0.34
285.7
360.5
280.1
374.1
263.0
362.4
0.3
0.35
284.9
359.0
280.1
370.4
263.0
352.3
0.31
0.36
284.1
357.4
279.9
366.7
262.9
341.0
0.32
0.38
283.2
355.8
279.7
363.1
262.9
328.5
0.33
0.39
282.2
354.1
279.3
359.5
262.8
315.0
0.34
0.40
281.2
352.4
278.9
355.9
262.6
300.5
0.35
0.42
280.2
350.7
278.4
352.3
262.5
285.1
0.36
0.43
279.1
348.9
277.8
348.8
262.3
268.8
Uniform strain (eng.)
0.22
0.15
0.34
0.00
0.33
0.25
Uniform Strain (true)
0.20
0.14
0.29
0.00
0.29
0.22
Tensile tests at two crosshead speeds (1mm/sec and 10mm/sec) can be fit to the following
hardening laws:
at V1 = 1mm/sec, = 500 ( + 0.05)0.25
at V2 = 10mm/sec, = 520 ( + 0.05)0.25
What is the strain-rate sensitivity index for these two materials? Does it vary with strain? What is
the uniform strain of each, according to the Considere Criterion?
SOLUTION:
ln 520/500
ln 2/ 1
=
= 0.017
m =
ln v2 /v1
ln 10/1
The strain-rate sensitivity is independent of strain because the ratio of stresses at the two strain rates is
independent of strain.
Substituting into the result for Problem 2a gives the uniform true strain in each case:
u (v2) = u (v1 ) = n o = .25 0.05 = 0.20

7.
Repeat Problem 6 with two other stress-strain curves:

at V1 = 1mm/sec, = 550 0.25
at V2 = 10mm/sec, = 500 0.20
Plot the stress-strain curves and find the strain-rate sensitivity index at strains of 0.05, 0.15, and
0.25. In view of these results, does Eq. 1.17 apply to this material?
SOLUTION:
2 550 0.25
=
= 1.1 0.05
1 500 0.20
,
0.05
ln 1.1
m=
ln 10
0.095 + 0.05 ln
= 0.041 + 0.022 ln
2.303
Page 10
Chapter 1
400
350
300
250
200
150
0
0.1
0.2
0.3
True Strain
0.020
0.000
0
-0.020
0.1
0.2
0.3
-0.040
-0.060
True Strain
In this case, the strain-rate sensitivity varies with strain.
Eq. 1.17 applies equally well to Problem 6 or Problem 7 at a given strain rate. The difficulty is that the
equation was derived assuming that tensile stress depends only on tensile strain. However, the effect of
strain-rate sensitivity on the maximum load point is small if m<<n, as is the usual case. However, the
post-uniform elongation depends strongly on even small values of m.
8.
Consider the engineering stress-strain curves for three materials labeled A, B, and C below.
Qualitatively, put the materials in order in terms of largest-to-smallest strain hardening (n-value)
and strain-rate sensitivity (m-value).
e
C
B
A
e
Chapter 1
Page 11
SOLUTION:
e
Strain hardening (based on strain to maximum
load) order: B, A, C.
Strain-rate sensitivity (based on post-uniform
strain) order: A, C, B.
Ductility or formability (based on total strain to
v1 = 10-3 m/s
failure) order: A, B, C.
9.
v2 = 10-2 m/s
e=315 MPa
B
A
e=300 MPa
It is very difficult to match tensile specimens

precisely. For sheet materials, the thickness, width,
and strength may vary to cause a combined
uncertainty of about 1% in stress. Considering this
uncertainty of K's in Problem 6, calculate the range
of m values which one might obtain if one conducted the tests at both rates several times.
SOLUTION:
From Problem 6, we recall that
ln
m=
520
500
ln 10
= 0.017
,
520 1% x 520 = 515 to 525

500 1% x 500 = 495 to 505
but now we consider the range:

and
ln
mlow =
ln 10
ln
So, the combined uncertainty of m is in the range:
mhigh =
515
505
525
495
ln 10
= 0.009
= 0.017 0.009
= 0.026
So, a 1% uncertainty in stress corresponds to a 50% uncertainty in m!
10.
Considering the specimen-to-specimen variation mentioned in Problem 9, it would be very

desirable to test strain-rate sensitivity using a single specimen. Typically, "jump-rate tests" are
conducted by abruptly changing the crosshead velocity during the test. Find the strain-rate
sensitivity for the idealized result shown:
SOLUTION:
ln 315
300
m=
= 0.021
ln 10
11.
In fact, the procedure outlined in Problem 10, while being convenient and attractive, has its own
difficulties. In order to obtain sufficient resolution of stress, it is necessary to expand the range
and to move the zero point. Some equipment does not have this capability. More importantly, the
response shown in Problem 10 is not usual. For the two more realistic jump-rate tests reproduced
Page 12
Chapter 1
below, find m values using the various points marked.

e
D
v2 = 10-2 m/s
C
v1 extrapolated
B
A
A=300 MPa
B=315 MPa
C=330 MPa
D=345 MPa
v1 = 10-3 m/s
e
e
v2 extrapolated
A
v1 = 10-3 m/s
B
C
D
v2 = 10-2 m/s
A=315 MPa
B=310 MPa
C=300 MPa
D=290 MPa
e
SOLUTION:
For the "up jump" in rate:
mB =
ln 315
300
ln 10
= 0.021
,
mC =
ln 330
300
ln 10
ln
= 0.041
,
mD =
345
300
ln 10
= 0.061
For the "down jump" in rate:
mB =
ln 310
315
1
ln 10
mC =
= 0.007
ln 300
315
1
ln 10
290
315
1
ln 10
ln
= 0.021
mD =
= 0.036
It should be apparent that neither the jump or continuous method eliminates the uncertainties.
B.
DEPTH PROBLEMS
12.
If a tensile test specimen were not exactly uniform in cross section, for example if there were
Chapter 1
Page 13
initial tapers as shown below, how would you expect the measured true stress-strain curves to
appear relative to one generated from a uniform specimen? Sketch the stress-strain curves you
would expect.
(a) Uniform gage length

G
(b) General notch
(c) Severe notch
SOLUTION:
Engineering Stress
The presence of a notch tends to concentrate the strain in the reduced gage section such that work
hardening occurs there rapidly. In a more severe notch, the stress state begins to have a lateral
component (tending toward plane strain) which leads to more hardening. Therefore, one might expect
the behavior to appear as shown.
Severe Notch
Mild Notch
Uniform
Engineering Strain
13.
What is the relevance of the 0.2% offset in determining the engineering yield stress?
SOLUTION:
It is simply a convenient number; small enough so that little strain hardening takes place but large
Page 14
Chapter 1
enough to resolve using most tensile testing equipment.

14.
Some low-cost steels exhibit tensile stress-strain curves as shown below. What would you expect
to happen with regard to necking?
e
SOLUTION:
During the first, flat stage one should expect localization to begin. In fact, this happens in a narrow
band called a Luder's band, but as the strain there increases the material in the bank increases and the
flow stress exceeds that of the surrounding material. The bank thus moves outward until the entire
specimen is strained beyond the flat region. After that, straining takes place normally.
15.
It has been proposed that some materials follow a tensile constitutive equation which has additive
effects of strain hardening and strain-rate hardening rather than multiplicative ones:
multiplicative: = F() G()

additive:
= F() + G()
(V2 )
In the first case one investigates G() at constant by examining (V1 ) , as we have done so far.
In the second case, one would watch (V2 ) - (V1 ) . Assume that an additive law of the following
type were followed by a material:
0.030
= 500 0.25 + 25
o
where o is the base strain-rate where the strain hardening law is determined (i.e. a tensile test
conducted at a strain rate of o exhibits = 5000.25).
a. Given this law, determine the usual multiplicative m value at various strains from two
tensile tests, one conducted at o and one at 10 o .
b. Compare tensile results extracted from the additive law provided and the multiplicative
one determined in Part a. [Use the m value obtained from the center of the strain range, at
= 0.125.]
Chapter 1
Page 15
SOLUTION:
= 500 0.25 + 25
0.30
0.25
at = o = 500 + 25
at
= 10 o = 500 0.25 + 25 (10)0.30 = 500 0.25 + 50
ln
m=
500 0.25 + 50
500 0.25 + 25
ln
10
, such that at
= 0.05
= 0.15
= 0.25
= 0.125
m = 0.040
m = 0.031
m = 0.028
m = 0.032
The m value decreases with strain because the stress difference between the two rates is reduced relative
to the overall flow stress.
16.
Use Eq. 1.1-19 (or, equivalently, Eqs. 1.1-20 and 1.1-22) to find the plastic instability for the
strain hardening [f()] and strain-rate hardening [g()] forms specified. In each case m=0.02 and
o =1/sec.
m
g() =
o
a. = f() g() , f() from Problem 2a,
g() =
=
f()
g()
o
b.
, f() from Problem 2c,
g() =
=
f()
g()
o
c.
, f() from Problem 2d,
g() =
=
f()
g()
o
, f() from Problem 2e,
d.
g() =
=
f()
g()
o
e.
, f() from Problem 2f,
f. = f() + g() , f() and g() from Problem 15. (Leave Part f in equation form.)
SOLUTION:
= k ( + o )n
0
ln
ni =
= n ,
ln + o
a.
=
`
n
1-m + o
mi =
ln
ln
=m
u = n - o = 0.205
1-m
Page 16

m
b.
= o (1 - Ae-B)
0
BA exp(-B )
ln
,
ni =
=
ln 1 - A exp(-B )
ln
ln
mi =
=m
A (B+1-m)
= 0.300
u = 1 ln
1-m
B
c.
= o
0
ni =
ln
ln
=0,
mi =
ln
ln
=m
u = 0 (Never stable)
d.
0
ln
ni =
= k ,
ln o + k
= o + k
k
1-m o + k
mi =
ln
ln
=m
u = 1 - o = 0.306
1-m k
e.
= k sin (B)
0
ln
ni =
= B ,
ln tan(B)
=
f.
1-m tan(B)
= k n + B
o
ni =
ln
ln
mi =
ln
ln
=m
u = 1 tan-1 B = 0.225
B
1-m
nk n
k n + B
mi =
ln
ln
Bm
k n + B
Substitution leads to a transcendental equation:

n n-1 - n = B (1-m)
k
o
Chapter 1
Chapter 1
Page 17
which may be solved iteratively if so desired. Note that for an additive law such as
this one, the plastic instability strain depends on strain rate as well as material
constants.

A. PROFICIENCY PROBLEMS
1.
Perform the indicated vector operations using the vector components provided:
a (1, 1, 1) b (1, 2, 3)
ab
ac
bc
a+b
a+c
b+ c
SOLUTION:
Note:
c (-1, 1, -1)
ab
ac
bc
ba
ca
cb
a(bc)
(ab)(ac)
a(b+c)
ab+ ac
a(b+ c)
(ab)+( ac)
a b a i b i = a 1 b 1 + a 2 b 2 + a 3b 3
a b = kij a i b j x k = a 2b 3 a3b 2 x 1 + a 3b 1 a1b 3 x 2 + a 1b 2 a2b 1 x 3
a b = (1, 1, 1) (1, 2, 3) = 1 1 + 1 2 + 1 3 = 6
a c = (1, 1, 1) (1, 1, 1) = 1 + 1 1 = 1
b c = (1, 2, 3) (1, 1, 1) = 1 + 2 3 = 2
a + b = (1, 1, 1) + (1, 2, 3) (2, 3, 4)

a + c = (1, 1, 1) + (1, 1, 1) (0, 2, 0)
b + c = (1, 2, 3) + (1, 1, 1) (0, 3, 2)
x 1 x 2 x3
ab =
1 1 1 =
1 2 3
In a similar way,
a c (1, 2, 1)
b c (2, 0, 2)
b a (5, 2, 3)
c a (2, 0, 2)
c b (5, 2, 3)
11
11
11
x1
x2 +
x = x 1 2x 2 + x 3 (1, 2, 1)
23
13
12 3
Page 2
Chapter 2
a (b c) = 4
(a b) (a c) = 0
a (b + c) = 5
a b+a c= 5
a (b + c) (1, 2, 3)
(a b) + (a c) (1, 2, 3)
2.
Perform the indicated vector operations.

a. Write the
x 1 ,x 2 ,x 3
components of the given vectors
(a,b,c) n terms of the base vectors
provided:
x 1 (0.866, 0.500, 0.000)

x 2 (-0.500, 0.866, 0.000) ,
x 3 (0.000, 0.000, 1.000)
where the components of these base vectors are expressed in the original coordinate
system as follows:
a (1, 1, 1)
b (1, 2, 3)
c ( -1, 1, -1)
or
a = x 1 +x 2 +x 3 b = x 1 +2x 2+3x 3 c = -x 1+x 2 -x 3
b. Perform the following operations using the components of a, b, c expressed in the new
(primed) basis:
a b
a (bc)
a(b+c)
ab
(ab) (ac)
(ab)+(ac)
a+b
c. Construct the rotation matrix [R] to transform components from the original coordinate
system to the primed coordinate system. Is [R] orthogonal? Find the inverse of [R] in
order to transform components expressed in the primed coordinate system back to the
original, unprimed coordinate system.
d. Transform the components of the results found in Part b. to the unprimed coordinate system
and compare the results with the equivalent operations carried out in Part a. (using
components expressed in the original coordinate system).
SOLUTION:
x1 (0.866, 0.500, 0.000)
a.
x2 (0.500, 0.866, 0.000)

x3 (0.000, 0.000, 1.000)
x1 = 0.866 x 1 + 0.5 x2
x2 = 0.5 x 1 + 0.866 x 2
x3' = x3
Chapter 2
Page 3
so that, according to Eq. 2.23:
0.866 0.500 0.000

R = 0.500 0.866 0.000 , and
0.000 0.000 1.000
a1
a2
a3
b1
b2
b3
c1
c2
c3
0.866 0.500 0.000

= 0.500 0.866 0.000
0.000 0.000 1.000
1.366
1
1 = 0.366
1.000
1
0.866 0.500 0.000

= 0.500 0.866 0.000
0.000 0.000 1.000
1.866
1
2 = 1.232
3.000
3
0.866 0.500 0.000

= 0.500 0.866 0.000
0.000 0.000 1.000
0.366
1
1 = 1.366
1.000
1
a (1.366, 0.366, 1.000)

b (1.866, 1.232, 3.000)
c (0.366, 1.366, 1.000)
b.
a b = 6 (Note: the prime notation is used here to remind that the required operations
were carried out using the components expressed in the primed coordinate system.)
a b (0.134, 2.232, 1.000)
a + b (3.232, 1.598, 4.000)
a (b c) = 4.000
(a b) (a c) = 0.000
(ab)+(ac) = 4.000
(a b) + (a c) (1.866, 1.232, 3)
Page 4

0.866 0.500 0.000
R = 0.500 0.866 0.000 ,
0.000 0.000 1.000
c.
Chapter 2
x = R x
1.00 0.00 0.00

0.866 0.500 0.000
0.500 0.866 0.000 = 0.00 1.00 0.00 = I
0.00 0.00 1.00
0.000 0.000 1.000
0.866 0.500 0.000

R R T = 0.500 0.866 0.000
0.000 0.000 1.000
Orthogonal
d.
signed
cofactor
matrix
=
a = R a , or
(i)
(ii)
0.866 0.500 0.000

= 0.500 0.866 0.000 = R
0.000 0.000 1.000
a = a
a b = 6 ; does not change since it is scalar.

0.866 0.500 0.000
a b = 0.500 0.866 0.000
0.000 0.000 1.000
a +b =
(iii)
0.866 0.500 0.000

0.500 0.866 0.000
0.000 0.000 1.000
0.134
1
2.232 = 2 (1, 2, 1) (same result)
1.000
1
3.232
2
1.598 = 3
4.000
4
(2, 3, 4) (same result)
(iv) a (b c) = a (b c ) = 4
(v)
a (b c) = a (b c) = 0
(vi) (a b) (a c) = (a b) (a c) (1, 2, 3)
(vii) a (b + c) = a (b + c) (1, 2, 3)
(viii) (a b) + (a c) = (a b) + (a c) (1, 2, 3)
Should have the same results for every case.
3.
Find the rotation matrices for the following operations:

a. Rotation of axes (i.e. component transformation) 45 about x 3 in a right-handed sense
(counter-clockwise when looking anti-parallel along x 3 ).
b. Rotation of a physical vector 45 about x 3 in a right-handed sense (i.e. the vector moves
Chapter 2
Page 5
counter-clockwise when looking anti-parallel along x 3 ).

c. Rotation of axes (i.e. component transformation) 30 about x 2 in a right-handed sense (i.e.
counter-clockwise when viewed anti-parallel to x 2 ).
d. Rotation of a physical vector 30 about x 2 in a right-handed sense (i.e. the vector moves
counter-clockwise when looking anti-parallel along x 2 ).
SOLUTION:
a.
2
2
2
R = cos 135 cos 45 cos 90 =
2
cos 45 cos 45 cos 90
cos 90 cos 90
2
2
2
2
2
2
R =
2
2
0
a.
0
0
R =
1
c.
b.
4.
cos 0
2
2
2
2
3
0
2
0
0
1
3
0
2
1
2
1
2
3
2
R =
d.
1
0
2
1
2
0
3
2
Perform the matrix manipulations shown.

Find the determinants and inverses of the following matrices:
1 2 3
7 8 9
1 1 1
A = 4 5 6
B = 1 2 3
C = -1 2 3
7 8 9
4 5 6
3 1 -1
-1
-1
-1
b. Multiply A A , B B , and C C
to verify that the inverse has been correctly
obtained.
SOLUTION:
a.
123
A = 456
789
789
B = 123
456
1 1 1
C = 1 2 3
3 1 1
A = 1 (5 9 6 8) 4 (2 9 3 8) + 7 (2 6 3 5) = 0
Page 6
Chapter 2
signed wfactor
matrix
1
,
A =
A
B = A = 0,
since A = 0,
cannot be obtained
does not exist

C
C = 1 (2 3) 1 (1 9) + 1 (1 6) = 4,
b.
A A
C C
, B B
1.25 0.5 0.25

= 2
1
1
1.75 0.5 0.75
; not applicable.
1 1 1
= 1 2 3
3 1 1
100
1.25 0.5 0.25
= 010
2
1
1
001
1.75 0.5 0.75
Yes, inverse has been correctly obtained.

5.
The following sets of basis vectors are presented in a standard Cartesian coordinate system
(x 1 , x 2 , x 3 ).
(1)
x 1 0.707, 0.707, 0.000

Set (1)
(1)
x 2 -0.500, 0.500, 0.707

(1)
x 3 0.500, -0.500, 0.707
(2)
x 1 0.750, 0.433, 0.500

Set (2)
(2)
x 2 -0.500, 0.866, 0.000

(2)
x 3 -0.433, -0.250, 0.866
(3)
x 1 0.866, 0.500, 0.354

Set (3)
(3)
x 2 0.500, 0.866, 0.354

(3)
x 3 0.000, 0.000, 0.866
a. Using vector operations, determine which of the basis sets are orthogonal.
b. Determine the transformation matrices to transform components presented in the original
coordinate system (x 1 , x 2 , x 3 ) to those in each of the other basis systems.
Chapter 2
Page 7
c. Which of the transformation matrices in Part b. are orthogonal? Does this agree with Part
a?
d. Find the transformation matrix to transform components provided in coordinate system (1)
to components expressed in coordinate system (2). Is the transformation matrix
orthogonal?
SOLUTION:
a.
To be orthogonal, the inner product of two vectors should be zero.
Set (1)
(1)
(1)
(1)
(1)
(1)
(1)
x1 x 2 = (0.707, 0.707, 0.000) (-0.5, 0.5, 0.707) = 0.0

x 3 x 2 = (0.500, -0.500, 0.707) (-0.500, 0.500, 0.707) = 0.0
x 1 x 3 = (0.707, 0.707, 0.000) (0.500, -0.500, 0.707) = 0.0
orthogonal
(2)
(2)
(2)
(2)
(2)
(2)
Set (2)
x 1 x 2 = x 2 x 3 = x 1 x 3 = 0.0
Set (3)
x 2 x 3 = 0.991
(3)
(3)
(3)
(3)
(3)
(3)
orthogonal
x1 x 2 = 0.307
x 1 x 3 = 0.307
b.
x 1 (1, 0, 0) ,
x 2 (0, 1, 0) ,
not orthogonal
x 3 (0, 0, 1)
(1)
x 1 (0.707, 0.707, 0.0) = 0.707 (1, 0, 0) + 0.707 (0, 1, 0) + 0.0 (0, 0, 1)

(1)
x 2 (-0.5, 0.5, 0.707) = -0.5 (1, 0, 0) + 0.5 (0, 1, 0) + 0.707 (0, 0, 1)

(1)
x 3 (0.5, -0.5, 0.707) = 0.5 (1, 0, 0) + -0.5 (0, 1, 0) + 0.707 (0, 0, 1)

Set(1)
In a similar way as shown in Exercise 2.5, we obtain
Set (1)
Set (2)
0.707 0.707 0.000

R (1) = 0.500 0.500 0.707
0.500 0.500 0.707
0.750 0.433 0.500
R (2) = 0.500 0.866 0.000
0.433 0.250 0.866
R (3) =
Set (3)
0.866 0.500 0.354

0.500 0.866 0.354
0.000 0.000 0.866
Page 8
c.
(1)
(1) T
Set (1)
0.707 0.5 0.5

= 0.707 0.5 0.5
0.0 0.707 0.707
Chapter 2
100
0.707 0.707 0.0
0.5 0.5 0.707 = 0 1 0
001
0.5 0.5 0.707
orthogonal
R (2) R (2)
Set (2)
(3)
(3) T
orthogonal
=I
1.13 0.99 0.31

= 0.99 1.13 0.31
0.31 0.31 0.75
Set (3)
0.866 0.500 0.000

0.500 0.866 0.000 =
0.354 0.354 0.866
0.866 0.500 0.354

= 0.500 0.866 0.354
0.000 0.000 0.866
1 00
0 10 = I
0 0 1
not orthogonal
i.e. When the transformation is a pure rotation, the transformation is orthogonal.

All of these results agree with Part a.
d.
x (1) = R (1) x ,
Thus,
x (2) = R (2) x
x (2) = R(2) R (1)
x (1)
R (1) is orthogonal R (1)

Since
R (1) (2) = R (2) R (1)
= R (1)
0.750 0.433 0.500

= 0.500 0.866 0.000
0.433 0.250 0.866
=
0.707 0.500 0.500

0.707 0.500 0.500
0.000 0.707 0.707
0.836 0.195 0.512

0.259 0.683 0.683
0.483 0.704 0.521
Check of orthogonality:
T
R R(1) =
0.836 0.195 0.512

0.259 0.683 0.683
0.483 0.704 0.521
0.836 0.259 0.483

0.195 0.683 0.704
0.512 0.683 0.521
100
010 = I
001
orthogonal
For those familiar with matrix manipulation, another proof may be written briefly as follows:
Chapter 2
R (1) R (1)
= R (2) R (1)
R (2) R (1)
= R (2) R (1)
Page 9
R (1) R (2)
= R (2) R (2)
= I
6.
Solve the sets of equations presented below by finding the inverse of the coefficient matrix. (Note
that Part b will require extension of the inversion formula to matrices of size greater than 3x3.):
a.
X1 + 2X 2 + 3X 3 = 10
X1 + 5X 2 - X 3 = 12
X1 + 3X 2 + X 3 = 14
b.
X1 + 2X 2 + 3X 3 + 4X4 = 10
X1 + 5X 2 - X 3 +14X4 = 12
X 1 + 3X 2 + X 3 + X4 = 14
X1 + 4X 2 - 2X 3 - 2X4 = 16
SOLUTION:
a.
K X = F X = K
1 2 3
K = 1 5 1 ,
1 3 1
10
F = 12
14
T
8 2 2
1
1
7 2 1
K =
2
17 4 3
K = 5 + 3 2 + 9 2 15 = 2,
Classical adjoint: transpose of the matrix of cofactors.
X =
b.
1
2
8 7 17
2 2 4
2 1 3
In the same way,
10
12
14
37
= 6
5
4.632
5.789
X =
2.000
0.737
Page 10
Chapter 2
7. Perform the following operations related to eigenvector - eigenvalue problems.

a. Find the eigenvalues and the associated eigenvectors for the following matrices:
1 2 3
1 -1 2
1 2
2 4 5
-1 2 -3
3 1
3 5 6
2 -3 3
b. Find the transformation matrices which change components expressed in the original
coordinate system to ones expressed using the eigenvectors as base vectors. Choose the
direction associated with the maximum eigenvalue to be the new x 1 ' ,the second one x 2 ' ,
and the third one x 3 ' .
c. Treating the columns of the matrices in part a. as vectors, find the equivalent components
expressed in the eigenvector bases from Part b.. [i.e. use the transformation matrices
found in Part c. to find the new components of the tensors in Part a., expressed in the
principal coordinate system.]
SOLUTION:
a.
1 2
= 1 2 6 = 12 + 2 6 = 22 5 = 0
A I = 0 , 3 1
Eigenvalues: 1 = 1 + 6 , 2 = 1 6 ;
(i)
For
= 1 + 6 = 3.449 ,
A 1 I p (1) =
6 2
3 6
(1)
Eigenvectors: p (1) , p (2)
x (1)
1(1)
x2
0
x (1)
1
=
(1)
0
x2
(1)
(1)
1
Let x = 1,
then 6 + 2x
p (1) 15
Normalizing these
(ii)
(1)
2
= 0
1, 26
x2 =
6
2
= (0.632, 0.775)
For = 1 6 = 1.449
6 2
3 6
(2)
x1
x
(2)
2
0
0
,
and if
x (2)
= 1,
1
(2)
Normalizing, we get p (0.632, 0.775)
then x (2)
=
2
6
2
Chapter 2
Check of orthogonality:
original
matrix is not symmetric.)
p(1) p(2) = 0.20 0
Page 11
(Not orthogonal because the
Using the same procedure,

1 2 3
2 4 5
For 3 5 6 ,
(ii)
(1)
b.
c.
(2)
(3)
1 = 6.419, p1 (0.374, 0.577, 0.725)

2 = 0.387, p2 (0.816, 0.577, 0.038)
3 = 0.806, p3 (0.441, 0.577, 0.687)
1 1 2
1 2 3
2 3 3 ,
(iii)
1 = 11.345, p1 (0.328, 0.591, 0.737)

2 = 0.516, p2 (0.737, 0.328, 0.591)
3 = 0.171, p3 (0.591, 0.737, 0.328)
0.632 0.775
0.632 0.775
0.328 0.591 0.737

= 0.591 0.737 0.328
0.737 0.328 0.591
0.374 0.577 0.725
= 0.816 0.577 0.038
0.041 0.577 0.687
Suppose x, y, z are orthogonal eigenvectors of A where eigenvalues are 1, 2, and 3,

respectively, let
x 1 y 1 z1
L = x 2 y 2 z2
x 3 y 3 z3
, where
x1
x x2
x3
y1
y y2
y3
z1
z z2
z3
A x = 1 x
Then from 2.35a,
A y = 2 y
A z = 3 z
A L = D L
D = L
A L
T
T
Here L = T , and D = T A T
We will obtain [D]; the diagonal matrix whose diagonal components are eigenvalues.
For example,
Page 12
0.328 0.591 0.737

D = 0.737 0.328 0.591
0.591 0.737 0.328
8.
123
245
356
Chapter 2
0.328 0.737 0.591

11.345 0
0
0.591 0.328 0.737 =
0 0.516 0
0.737 0.591 0.328
0
0
0.171
Find the new components of the tensors provided below if the coordinate system change
corresponds to a rotation of 30 about x 3 :
0.866 0.500 0.000

R = -0.500 0.866 0.000
0.000 0.000 1.000
1 2 3
T1 2 4 5 = T1 ,
3 5 6
SOLUTION:
T
T = R T R ,
0.866 0.500 0.000
T1 = 0.500 0.866 0.000

0.000 0.000 1.000
T2 = R T2 R
9.
T1 = R T 1 R ,
123
245
356
1 2 3
T2 4 5 6 = T2
7 8 9
T2 = R T R 2
3.482 2.299 5.098

0.866 0.500 0.000
0.500 0.866 0.000 = 2.299 1.518 2.830
5.098 2.830 6.000
0.000 0.000 1.000
4.598 2.232 5.598

= 4.232 1.402 3.696
10.062 3.428 9.000
In calculating contact conditions at an interface, it is often necessary to find the unit vector which
represents the projection of a given vector (usually the displacement of a material point) onto a
plane tangent to the interface. If the normal to the tangent plane is denoted n and the arbitrary
vector is a, find t, the unit tangent vector corresponding to the material displacement. [Express
the result in terms of a, n^ , and simple vector operations.]
SOLUTION:
One possibility is based on the use of vector addition and the dot product:
t=
a an n
a an n
(since t is a unit vector)
Alternatively, one may use the cross product to accomplish the same thing by first defining a
unit vector q , which is orthogonal to n , a , and t :
n an
q= a n t= n q=
an ,
an
B.
DEPTH PROBLEMS
Chapter 2
10.
Page 13
Perform the following operations related to component transformations.

a. Find the transformation for components from basis set (2) to basis set (3), in Problem 5,
above.
b. Find the inverse transformation, that is, one that expresses components in basis set (2) if
they are given in basis set (3).
c. Using the approach shown in Exercise 2.5, verify that transformation matrices found in
Parts a. and b. do, indeed, perform the indicated transformations.
d. Show the matrix form of the tensor transformation for components given in basis set (2) to
those in basis set (3).
SOLUTION:
(2)
a.
xi
(3)
xi
x (2) = R (2) x ,
x = R(2)
x (3) = R (3) x ,
x 3 = R (3) R (2)
R (2)(3) = R (3) R(2)
x2
T
x(2)
0.866 0.500 0.354

0.500 0.866 0.354
0.000 0.000 0.866
0.750 0.500 0.433

0.433 0.866 0.250
0.500 0.000 0.866
1.043 0.000 0.193

0.927 0.500 0.126
0.433 0.000 0.750
R (3) is not used in this transformation, its non-orthogonality is not

Since the inverse of
an issue. However, in part b this requires inverting (rather than transposing) a matrix.
b.
Recall from Part a: R

R (2)(3)
(2)(3)
= R (3) R (2)
= R
(3)
= R (2) R(3)
Where we have used the relationships:

c.
(2) T
R (2)
0.866 0.000 0.223

= 1.732 2.000 0.110
0.500 0.000 1.204
1
= R (2)
and
A B
= B
Let's verify the transformations by considering three vectors, namely those of the original
(o)
basis set x i (let's label them x , y, x here to simplify the notation). We form the matrix
A (corresponding to the tensor A composed of the three vectors) in the usual way, by
putting the components of each basis vector into a column. Since we are considering the
components of the basis set in the basis set, A is the identity matrix:
Page 14
A A
(o)
(o) (o)
x (o)
1 y1 z 1
= x (o)
y(o)
z (o)
=
2
2
2
(o) (o) (o)
x 3 y3 z 3
1
0
0
0
1
0
0
0
1
Chapter 2
= I
(2)
(3)
We then find the coordinates of these three vectors expressed in the x and x basis sets:
0.750 0.433 0.500

A A(2) = R(2) I = 0.500 0.866 0.000
0.433 0.250 0.866
A A(3) = R(3) I =
(in basis set 2)
0.866 0.500 0.354

0.500 0.866 0.354
0.000 0.000 0.866
(in basis set 3)
R (2)(3)
Now, our transformation matrix
must transform the components of any vector
(2)
(3)
expressed in x i to components expressed in x i :
A (3) = R (2)(3) A (2) =
1.043 0.000 0.193 0.750 0.433 0.500

0.927 0.500 0.126 0.500 0.866 0.000
0.433 0.000 0.750 0.433 0.250 0.866
0.866 0.500 0.354
0.500 0.866 0.354
0.000 0.000 0.866
= A(3)
A(3) obtained here with A(3) above shows that the transformation matrix
Comparison of
R (2)(3)
R (3)(2)
R (2)(3)
performs its intended function.
, the inverse of
may be verified
in the same manner.
d.
11.
Shown in a.
Find the rotation matrix for the double rotation of coordinate axes: rotate 90 about x 1 , and then
90 about x 3 .
SOLUTION:
R1 =
1 0 0
0 0 1
0 1 0
0 1 0
R 2 = 1 0 0
0 0 1
Chapter 2
R = R2
0 1 0
R 1 = 1 0 0
0 0 1
1 0 0
0 0 1
0 1 0
X2
0 0 1
1 0 0
0 1 0
0 0 1
R 12 = 1 0 0
0 1 0
12.
Page 15
A cylindrical coordinate system is one which

rotates according to the coordinates of the point
in question. Typically, r, , z represent the
coordinates of a point, with the base vectors
X3 = Z
given as r, , z , for example. The figure at the
right shows such a coordinate system and a

superimposed Cartesian coordinate system
which coincides when =0.
a. Find the transformation matrix to change components expressed in x 1 , x2 , x 3 to ones
expressed in r, , z .
b. Find the cylindrical components of the following vectors expressed in the x 1 , x 2 , x 3
system:
a 10.000,
0.000,
0.000
b 0.866,
0.500,
0.000
c 1.000,
1.000,
0.000
d 1.000,
1.000,
1.000
e 0.000,
1.000,
0.000
f -1.000,
1.000,
1.000
c. Find the magnitudes of the vectors given in Part b., first using the Cartesian coordinate
system components, then using the cylindrical coordinate system. How is the magnitude of
a vector computed in cylindrical coordinates?
d. Is [R] orthogonal for this transformation? Physically, why or why not?
SOLUTION:
r
cos sin
= sin cos
z
0
0
a.
b.
0
0
1
a (10.000, 0.000, 0.000)
x1
x2
x3
from geometry
X1
Page 16
Chapter 2
a = R x
a1
cos sin
a2 = sin cos
0
0
a3
10 cos
10
0 = 10 sin
0
0
0
0
1
In a similar way,
0.866 cos + 0.5 sin

b = 0.866 sin + 0.5 cos
0
cos + sin
d = sin + cos
1
c.
cos + sin
c = sin + cos
0
Cartesian Coordinate
Cylindrical Coordinate
2
2
2
a =
a = 10 + 0 + 0 = 10
Likewise, should have the same results.
R R
cos sin
= sin cos
0
0
cos + sin
f = sin + cos
1
sin
e = cos
0
0
0
1
10 2 cos 2 + 10 2 sin2 = 10
cos sin
sin cos
0
0
0
0
1
= I
d.
The basis sets of each system are mutually orthogonal.
13. Perform the indicated operations related to equation solving.
a. Solve the equations given in Problem 6 by using Gaussian reduction instead of by finding
the inverse. Which to you prefer for large matrices?
b. Given the solutions obtained in Part a., find the inverse of the coefficient matrix.
c. For larger sets of equations, why is it easier to solve by a reduction method?
SOLUTION:
a.
1 2 3
1 5 1
1 3 1
10
1 2 3
12 0 3 4
14
0 1 2
10
1 2 3
2 0 3 4
4
0 23
10
x 3 = 5
2 x = 6
2
10
3
x 1 = 37
Similarly, we should get the same results for the second set. The Gassian reduction method
Chapter 2
Page 17
is much simpler for large matrices because it operates row-by-row and it is not necessary to
keep track of complex expressions.
b.
Solutions given in Problem 6.
c.
For large sets of equations, it is much more complicated to compute the inverse of a matrix,
whereas the reduction method does not involve inverse computation. Less computation is
required in a reduction method.

A.
1.
Calculate the 3-D stress tensor components for the rectangular material shown in the figure, first
in the coordinate system x 1 , x 2 , and x 3 and then in the coordinate system x 1 ', x 2 ', x 3 ' .
1 mm
X2'
1 mm
X3'
200 N
X2
X1
X1'
X3
The angle between x 1 and x 1 is 30o, and the x 3 and x 3 axes are parallel.
SOLUTION:
Before doing the problem formally using the known tensor transformations, let's approach it
from a physical and geometrical standpoint. Because of the equilibrium, we know that the force
transmitted throught the cross-sectional area (1mm2) normal to x 1 is 200N, and the stress vector
N
S (200, 0, 0) mm
2
operating on that same plane is 1
. The other two planes, normal to
x 2 and x 3 , contain the force vector and thus have no associated stress vectors:
S 2 (0, 0, 0)
S 3 (0, 0, 0)
.
Therefore, the stress tensor in the x i coordinate system may be written:
200 0 0
0 0 0 .
0 0 0
The situation in the x i coordinate system may be approached similarly by first considering the
plane normal to x 1 which passes through the rod. The entire 200N of force must be transmitted
through this area, which is now larger because of the incline:
1mm 2
= 1.155mm 2
A 1 =
cos 30
The corresponding stress vector thus has a magnitude of

200 N
S 1 = 200 N2 cos 30 =
= 173.2 N/mm 2 ,
1mm
1.155 mm 2
To find the components of this vector along the x 1 ' and x 2 ' vectors, we simply resolve this
value:
S1 (173.2 cos 30, 173.2 sin 30, 0) = (150, 86.6, 0) N 2
mm
To find the stress vector operating on the plane normal to x 2 , we follow the same procedure:
2
A 2 = 1mm = 2 mm 2,
S 2 = 100 N/mm 2,
sin 30
, so
Page 2
Chapter 3
and the associated stress vector components are

S 2 (100 cos 30, 100 sin 30, 0) = (86.6 N/mm 2, 50 N/mm 2, 0)
And the entire stress tensor in x 1 ' is
S 1' S 2' S 3'
=
150 86.6
86.6 50
0
0
0
0
0
It is much easier and less error-prone to use the known tensor transformation properties to solve
the problem, as follows:
Let be the stress tensor in the material and t the stress vector, then we have in general:
n = t = F
a .
For n = x 1 , we get:
t 1 = x 1 = Fa , so that
t1 =
200 0
0 0
0 0
0
0
0
MPa
Similarly, t 2 = x 2 , t 3 = x 3 , and we can conclude that the stress tensor is

200 0
= 0 0
0 0
0
0
0
MPa
In order to transform these components to those corresponding to the

first find the rotation matrix:
x1
x
= R x
x2
x3
3 1
2
2
1
3
2 2
x1
x2
0 0 1
x3
and then apply the transformation rule for a second-ranked tensor:
= R R
3 1
2
2
1
3
2 2
3
2
1
0 0 1
0 0 1
200 0
1
2
3
2
0
0
x1
coordinate system, we
Chapter 3

100 3
100
3
2
1
0 0 1
2.
1
2
3
2
150 50 3
0
0
= 50 3
0
Page 3
150 86.6 0
50
= 86.6 50
MPa
Given the stress tensor which appears below, find the stress vector acting on planes normal to the
unit vectors n, m, and p , also given.
1
n 3 ( 1, 1, 1)
1 2 3
1
2 2 4
m 6 ( 1, 2, 1)
3 4 3
1 ( , , )
p
110
2
SOLUTION:
t = n
1 2 3
2 2 4
3 4 3
1 2 3
t
2 2 4
3 4 3
1 2 3
t
2 2 4
3 4 3
3.
1
3
1
3
1
3
1
6
2
6
1
6
1
2
1
2
0
6
3
8
3
10
3
8
6
10
6
15
6
3
2
4
2
7
2
3.464
=
4.619
5.774
3.266
=
4.082
5.715
2.121
=
2.828
4.950
Find the principal stresses, the principal directions, and the rotation matrix for transforming
coordinates to the principal coordinate system ( x 1 corresponds to , x 3 corresponds to )

max
for the stress tensors given.
3 -1 0
a. - 1 3 0 , b.
0 0 1
3 0 0
10 -5 5
0 3 - 1 , c. - 5 0 5
0 -1 1
5 5 10
Note: No numerical procedure is required to find the roots of the cubic equations.
min
Page 4
Chapter 3
SOLUTION:
a.
3 1 0
= 1 3 0
0 0 1
, so the eigen equation is:
3 2 1 = 0
3 3 1 1 = 1
So:
3 1 0
1 3 0 = 0
0 0 1
= 1, = 2, = 4 1 = 4, 2 = 2, 3 = 1
n1
0
n 2 = 0 , where (n 1 ,n2 ,n 3) are the components of x1
n3
0
1 1 0
1 1 0
0 0 3
For 1 = 4,
2
2
n3 = 0, n 1 + n 2 = 0, n1 + n 2 = 1 (unit vector)
n1 =
1
,
2
n2 = +
1
,
2
1 1 0
1 1 0
0 0 1
For 2 = 2
n 3 = 0, or
x 1 (
1,
2
1,
2
0)
m1
0
m 2 = 0 , where (m 1,m 2 ,m 3 ) are the components of x 2
m3
0
2
2
m 3 = 0, m1 m2 = 0, m 1 + m 2 = 1 (unit vector)
m1 =
1
,
2
For 3 = 1,
m2 =
2 1 0
1 2 0
0 0 0
1
,
2
m 3 = 0, or
x 2 (
1,
2
1,
2
0)
p1
0
p 2 = 0 , where (p 1 ,p2 ,p 3) are the components of x3
p3
0
2p1 p 2 = 0, p1 + 2p 2 = 0 p 1 = p 2 = 0,
p 1 = 0, p2 = 0, p 3 = 1, or
p23 = 1 (unit vector)
x 3 ( 0, 0, 1)
In order to find the rotation matrix, we first choose a right-hand set from among the various
choices of x 1 , x 2 , x 3
x1 (
x 2 (
1
1
,
, 0)
2
2
1
1
,
, 0)
2
2
x 3 (0,
0, 1)
R =
,
then the required rotation matrix is
1
1
2
2
1
1
2
2
Chapter 3
=
b.
Page 5
3 0 0
3 0 0
0 3 1 , so the eigen equation is
0 3 1 = 0
0 1 1
0 1 1
3 1 1 = 0
So,
1 = 3, 2,3 = 4 2 8 = 3.41, 0.59 1 = 3.414, 2 = 3.000, 3 = 0.586
For 1 = 3.41,
0.41 0
0
0 0.41 1
0
1 2.41
0
n1
n 2 = 0 , where (n 1,n 2 ,n3 ) are the components of x 1
n3
0
n 1 = 0, 0.41 n2 n3 = 0, n 21 + n22 + n 23 = 1
x 1 (0, 0.92, +0.38)
For 2 = 3,
0 0 0
0 0 1
0 1 2
m1
0
m 2 = 0 , where (m 1,m 2,m 3 ) are the components of x 2
m3
0
m 3 = 0, m2 = 0, m 1 = 1
x 2 (1, 0, 0)
2.414 0
0
0 2.414 1
For 3 = 0.586,
0
1 0.414
p1
0
p2 = 0 , where (p1 ,p2 ,p3 ) are the components of x 3
0
p3
p 1 = 0, 2.414 p2 p3 = 0, p 21 + p 22 + p 23 = 1
x 3 (0, 0.38, 0.92)
In order for find the rotation matrix we first choose a right-handed set of eigenvectors:
x 1 (0, 0.92, 0.38)
0 0.92 0.38
x 2 (1, 0, 0)
R =
x 3 (0, 0.38, 0.92)
c.
then the rotation matrix is
10 5 5
= 5 0 5 1 = 15, 2 = 10, 3 = 5
5 5 10
1
0
0
0
0.38 0.92
Page 6
Chapter 3
x 1 0.707, 0, 0.707 , for 1 = 15

.707 0 .707
R = .577 .577 .577
.408 .816 .408
x 2 0.577, 0.577, 0.577 , for 2 = 10

x 3 0.408, 0.816, .408 , for 3 = 5
R R
Check:
4.
1 0 0
= 0 2 0
0 0 3
10.6 0 10.6
R = 5.77 5.77 5.77
2.04 4.08 2.04
R R
15 0 0
= 0 10 0
0 0 5
Find the invariants for the stress tensors shown below:

a.
1.44 0.22 -0.76

1.75 0.35 -0.75
0.22 2.25 -0.38 , b. 0.35 2.50 -0.35 , c.
-0.76 -0.38 2.31
-0.75 -0.35 1.75
1.94 0.38 -0.54

0.38 2.75 0.22
-0.54 -0.22 1.31
SOLUTION:
a.
1.44 0.22 0.76

= 0.22 2.25 0.38
0.76 0.38 2.31
J 1 = 11 + 22 + 33 = 1.44 + 2.25 + 2.31 = 6
J 2 = 11 22 + 22 33 + 33 11 + 23 2 + 31 2 + 122
2
= 1.44 2.25 + 2.25 2.31 + 2.31 1.44 + 0.38 + 0.76 + 0.22 2 = 11
J3 =
b.
1.44 0.22 0.76

0.22 2.25 0.38 = 6
0.76 0.38 2.31
1.75 0.35 0.75

= 0.35 2.50 0.35
0.75 0.35 1.75
J1 = 6
J 2 = 11
J3 = 6
Chapter 3
Page 7
1.94 0.38 0.54

= 0.38 2.75 0.22
0.54 0.22 1.31
c.
J1 = 6
J2 = 11
J3 = 6
5.
Find and solve the characteristic equations for the stress tensors shown in Problem 4. Use the
method followed in Exercise 3.3. (A numerical procedure is required.)
SOLUTION:
J1 = 6
J2 = 11
J3 = 6
a.
0 = 3 6 2 + 11 6 =
i
' i
= 1 exactly. Lucky guess!
To obtain the quadratic equation, perform synthetic long division as shown below.
2 5 + 6
3
2
( 1) 6 + 11 6
3
2
6
5 2 + 11
5 2 + 5
6 6
6 6 , thus the original expression is ( 1) ( 2 5 + 6)
The remaining roots are found by the quadratic formula:

+5 25 24
=
= 2, 3
2
So, the three roots (principal stresses) are:
1 = 3,. 2 = 2, 3 = 1
and the characteristic equation can be written in product form:
3 2 1 = 0
b.
J1 = 6, J2 = 11, J 3 = 6
Page 8
Chapter 3
0 = 3 6 2 + 11 6 =
This characteristic equation is identical to 5.a., thus the two stress tensors are identical except for
a rotation. The principal values must therefore be the same.
c.
J1 = 6, J2 = 11, J 3 = 6
[Same as 5.a. and 5.b.]
6.
Find the principal directions for the stress tensors shown in Problem 4 and find the rotation
matrix which transforms components given in the original coordinate system to ones in a principal
coordinate system. (Assume that the minimum principal stress acts on a plane with normal x 1 '
and the maximum principal stress acts on a plane with normal x 3 ' .)
SOLUTION:
1 = 1
2 = 2
3 = 3
x 1 0.866, 0, 0.5
x 2 0.24, 0.866, 0.433

x 3 0.43, 0.5, 0.75
a.
Taking the three plus signs forms a right-handed system for which the rotation matrix is
0.866 0
0.5
R = 0.24 0.866 0.433 .
0.43 0.5 0.75
1 = 1
2 = 2
3 = 3
x 1 = 0.707, 0, 0.707
x 2 = 0.5, 0.704, 0.50
x 3 = 0.50, 0.704, 0.50
b.
For the "+" signs (one choice of right-handed system):

0.707 0 0.707
R = 0.50 0.704 0.50
0.50 0.704 0.50
1 = 1
2 = 2
3 = 3
c.
x 1 = 0.5, 0, 0.867
x 2 = 0.75, 0.5, 0.43
x 3 = 0.433, 0.867, 0.24
Chapter 3
Page 9
For the "+" signs (one choice of right-handed system):

R =
7.
0.5
0
0.867
0.75 0.5 0.43
0.43 0.867 0.24
Find the spherical and deviatoric components of the stress tensors given in Problem 4. Find the
principal stresses and directions of the deviatoric tensors following the method outlined in Section
3.6. How do these compare with the values for the stress tensor obtained in Problems 5 and 6.
SOLUTION:
stress
spherical
deviatoric
1.44 0.22 0.76

200
0.56 0.22 0.76
0.22 2.25 0.38 = 0 2 0 + 0.22 0.25 0.38
0.76 0.38 2.31
002
0.76 0.38 0.31
a.
d
Spherical Invariants: J 2 1.0
1 d = 2 13
2
3 d
= 2 13
= 2 13
1
2
1
2
1
2
J d3 0 , so
1 = 13 cos 1 0 = 30
2 = 120 + 30 = 150
3 = 30 120 = 90
cos (30) = 1.0

cos (150) = 1.0
1 = 1.0 + 2.0 = 3
cos 90 = 0.0
2 = 1.0 + 2.0 = 1
3 = 0 + 2.0 = 2
Therefore:
The principal directions are found using the principal values with the same result as Problems
5.a. and 6.a.
stress
spherical
deviatoric
1.75 0.35 0.75

200
0.56 0.22 0.76
0.35 2.50 0.35 = 0 2 0 + 0.22 0.25 0.38
0.75 0.35 1.75
002
0.76 0.38 0.31
b.
The results are the same as Problems 5.b. and 6.b.
stress
spherical
deviatoric
1.94 0.38 0.54

200
0.56 0.22 0.76
0.38 2.75 0.22 = 0 2 0 + 0.22 0.25 0.38
0.54 0.22 1.31
002
0.76 0.38 0.31
c.
Page 10
Chapter 3
The results are the same as Problems 5.c. and 6.c.

8.
Find the spherical, deviatoric, principal deviatoric components, and principal directions of stress
for the following cases:
11 = , other ij = 0
Uniaxial tension:
21 = 12 = , other ij = 0
Simple shear:
11 = 22 = , other ij = 0
Balanced biaxial tension:
13 = 31 = A, 21 = 12 = B, other ij = 0
Biaxial shear:
11 = t, 13 = 31 = s, other ij = 0
Tension and shear:
SOLUTION:
stress
spherical deviatoric
00
0 00 = 0
a.
Uniaxial tension
1d =
2
3
11 = , other ij = 0
2d = 3d =
0 00
00
2
3
0 +
0 0
00
0 0
m 1, 0, 0
n 0, 1, 0
(current axes are principal)
p 0, 0, 1
stress
b.
Simple shear
12 = , other ij = 0
00
00 0
00
0 0 = 00 0 + 0 0
0 00
00 0
00 0
1
, 12 ,
2
n 0, 0, 1
12 d = 21 d = ,
ij d = 0
spherical deviatoric
1,
2
1,
2
Chapter 3

stress
spherical
00
2
3
00 = 0
c. Balanced biaxial tension 11 = 22 = 0, other ij = 0
Page 11
000
0 + 0
2
3
0 0
2
3
0 0
deviatoric
2
3
m 1, 0, 0
1d
= 2d
=
3
n 0, 1, 0
2
=
3
3d
(current axes are principal)
p 0, 0, 1
d. Biaxial shear 13 = 31 = A, 12 = 21 = B, other ij = 0

stress
spherical
deviatoric
0 B A
000
0 B A
B 0 0 = 0 0 0 + B 0 0
A 0 0
000
A 0 0
B
1 ,
2
n 0,
B2
+ A
A
A + B
2
2 B 2 + A2
3 =
B 2 + A 2
A
2 B 2 + A2
p 1,
2
1 =
B 2 + A 2
2 = 0
A2 + B 2
A
2 B 2 + A2
e. Tension and shear: 11 = t , 13 = 31 = s , other ij = 0

stress
spherical
deviatoric
t
3
t 0 3
0 0 0
s 0 0
1 =
t +
0 0
t
3
0 0
3 =
0 +
t
3
t 2 + 4 s 2
2
2 = 0
t
2 t
3
t 2 + s2
2
t
3
s
0
t
3
Page 12

m
s
1
, 0, t
,
D
D
0, 1, 0
s

, 0, t 2 ,
D
D
t 1
where D =
where D =
+ s 2
Chapter 3
1
2
t 2 + s2
1
2
(If s t < 0, the minus sign is adopted for the components of p.)
B.
DEPTH PROBLEMS
9.
The reciprocal theorem of Cauchy states that the stress vectors acting on two intersecting planes
have the following property:
s 1 n2 = s 2 n1
where si is the stress vector acting on a plane with normal n i . Show that this principle follows
from the symmetry of the stress tensor, or from the equilibrium condition directly.
SOLUTION:
It is possible to prove the relationship by considering Cauchy's tetrahedron (Exercise 3.1), or by
multiplying all of the required components and comparing the results.
The shortest method is by writing the various terms in indicial notation.
Let
n1 = n, n 2 = m, and s 1 = s and s 2 = t
for simpler notation, then
s = n si = ij nj
t = m t i = ij m j
s m = nm or, s m = si m i = ij n j m i
t n = m n or t n = t i n i = ij m j n i
but, since ij = ji, we can rewrite ij n j m i = ij ni m j, so s m = t n.
10.
Octahedral planes are ones which have normals forming equal angles with the three principal
axes. Find an expression for Sn , the normal components of the stress vector on the octahedral
plane in terms of a) principal stresses and b) arbitrary stress components.
SOLUTION:
Chapter 3
Page 13
x2
n
x1
x3
1
=
0
0
a. In principal axes, as show in the figure above,
Sn = nSn =
1 1 1
3 3 3
Therefore:
where J1 is the first stress invariant.
1 0 0
0 2 0
0 0 3
1
3
1
3
1
3
0 0
2 0 , and n =
0 3
= 1 ( 1
3
+ 2 + 3) p =
1
3
1
3
1
3
1 J (Eq. 2.38)
3 1
b. Since we know that this quantity is invariant to the choice of coordinate system orientation, the
expression in an arbitrary cartesian system is
Sn = p = 13 J 1 = 13 ( 11 + 22 + 33)
11.
Show that the tangential component of the stress vector on the octahedral planes (i.e. the shear
2
J2 '
component) is equal to 3
1
2
, where J2 ' is the second invariant of the deviatoric stress tensor.
SOLUTION:
As shown in Problem 10., the normal component on an octahedral plane is
1
SN = 1 + 2 + 3
3
The tangential component may be found from the relationship
S N 2 + S T 2 = S S = ( n) ( n), or
1
2
2
2
S T 2 = ( n) ( n) S N 2 = 1
3 1 + 2 + 3 9 1 + 2 + 3
9 ST 2 = 2 1 2 + 2 2 2 + 2 3 2 2 1 2 2 1 3 2 2 3 = 6 J2
Page 14
ST = 2
3 J2
So,
12.
Chapter 3
1
2
Problems involving cylindrical symmetry often use cylindrical coordinates r, , z where

1
x
2
= tan -1
r = x21 +x22
y . (Conversely, x1 = r cos , x2 = r sin , x3 = z). For example,
and
consider pure torsion of an elastic, long bar with axis parallel to z , where z is a constant on the
outside of the bar. Find the stress tensor in two alternate Cartesian coordinate systems:
a. One which has x 1 normal to the cylinder axis and is tangent to the cylinder surface, x 2
normal to the cylinder surface, and x 3 the cylinder axis.
b. One which is fixed in space (i.e. independent of ), with x 3 // z , x 1 lying in the =0, z=0
direction, and x 2 lying in the = 2 , z=0 direction.
x2
x1
a.
Reference Axis
Since the axes are identical, the stress components are identical, i.e.
= R R
00 0
00
00
100
= 010
001
100
010 =
001
x2
b.
cos sin
R = sin cos
0
0
x1
0
0
1
(In this case, r and rotate as the point of interest rotates, but hte cartesian system is fixed.)
= R R
cos sin
sin cos
0
0
0
0
1
0 0 0
0 0
0 0
cos sin
sin cos
0
0
0
0
1
Chapter 3

cos sin
= sin cos
sin
cos
Page 15
0
0
sin
=
0
0
cos
sin cos
0
(fixed Cartesian system)

13.
Show that if two roots of the characteristic equation are identical (i.e. degenerate), then any
direction normal to the other principal direction (i.e. the one corresponding to the non-identical
root) is a principal direction. Show that if all three roots are identical, all directions are
principal.
SOLUTION:
a. Assume that the characteristic equation is of the form:
2
o 1 = 0
,
where 1 is a degenerate root. The stress components in the principal axes are
1 0 0
0 1 0
0 0 o
,
where o is the principal stress in the x 3 (3rd principal) direction. A general rotation of
coordinate system about the x 3 axis may be written as follows:
cos sin
= sin cos
0
0
0
0
1
1 0 0
0 1 0
0 0 o
cos sin
sin cos
0
0
0
0
1
1 cos2 + sin 2
1 sin cos + sin cos
1 sin cos sin cos
1 sin 2 + cos2
Page 16
1 0 0
= 0 1 0
0 0 o
for any direction normal to x 3
Chapter 3
b. If all three roots are degenerate, there are many ways to show that any direction is equivalent.
The derivation in Part a. can be done for 1 = o, or one can note that all three roots being
equivalent is the same as the spherical component (i.e. hydrostatic pressure or tension).
14.
It is often convenient to replace one set of forces with another, statically-equivalent set. For
example, consider a triangular element of material (assume unit depth normal to the triangle)
which is assumed to be a small enough piece of a body to feel only a homogeneous stress, ij
(i,j=1,2, assuming that i3=0, where x 3 is normal to the triangle). Use a simple, physicallymotivated procedure to replace ij by three forces, f1, f2, f3 acting at the three corners of the
triangle. Consider the force transmitted by each face.
SOLUTION:
Consider a triangle with normals defined to each side with a magnitude equal to the length of the
side. (For unit depth of the sides in three dimensions, these are area vectors corresponding to the
sides.)
C
3
b
1
where:
A = a
B = b
C = c
A
A, B, C are deduced from a, b, and c by a rotation of -90o, therefore since a + b + c = 0, A + B
fA = A, fB = B, fC = C
+ C = 0. The forces acting on the planes A, B, and C are
and fA + fB + fC = A + B + C = 0 because A + B + C = 0.
To assign these forces to the vertices, let's use the unweighted average (although other choices
might make more physical sense) of the forces on the connected sides:
f1 = 1 fA + fC = 1 A + C
2
2
1
f2 =
f + fB = 1 A + B
2 A
2
1
1
f3 =
f + fC = B + C
2 B
2
15.
Physically, why can the entire material loading at a point be reduced to three orthogonal force
intensities passing through the point? Why do the shear components disappear along these
directions?
SOLUTION:
Chapter 3
Page 17
For simplicity, let's consider a two-dimensional situation first. Similar to Figure 3.1, imagine
making an arbitrary mathematical cut as shown in part (a) of the figure below. We can find the
force acting on one of the cut faces required to maintain equilibrium, part (b). (The opposite
force is required on the other cut face by equilibrium.) Then, glue the first cut back togehther
and using the direction of the force as a guide, make another cut, this one perpendicular to the
force observed on the first cut. Find the new force required for equilibrium and, if necessary,
make another cut perpendicular to the new force. Continue until the current cut and current
force are perpendicular, part (c). If we now relax the forces on the cut plane (and any external
forces required to maintain equilibrium as the cut face is unloaded), can we be assured that the
material is completely unloaded? No, because the direction parallel to the cut face (grey arrow
in part (c)) is unaffected by the cut and therefore we have no information about it. Therefore,
make a cut perpendicular to the final first cut and the force required by equilibrium will by
necessity be perpendicular to the first force, part (d). This simple thought exercise demonstrates
why there are only two independent force intensities passing through a point in a twodimensional body, and why they must be perpendicular.
To extend the exercise to three dimensions, follow precisely the same procedure. Once the first
plane and normal force are found, there remain two perpendicular planes which must have only
normal forces acting on them.
(a)
(c)
(b)
(d)
Although opposite to the usual derivation, it would be possible to derive the symmetry of the
stress tensor by first noting that this result requires the existence the three perpendicular
principal directions and that any rotation of axes from this principal set must produce a
symmetric and real set of tensor components.
16.
The two sets of components presented below correspond to the identical stress tensor, as measured
in two coordinate systems, x 1 , x 2 , x 3 , and x 1 ', x 2 ', x 3 ' . Find the rotation matrix to transform
components from the x i system to the x i ' system, and vice versa. (Hint: First find the rotations to
the common, principal coordinate systems.)
1.000 1.730 1.000
= 1.730 0.750 0.433
1.000 0.433 0.250
SOLUTION:
= R R
0.500 1.414 0.500

= 1.414 1.000 1.414
0.500 1.414 0.500
Page 18
Chapter 3
Define rotation matrices R 1 and R2 such that

principal = R1 R 1 T
principal = R2 R2
Then, we can find [R] in terms of R 1 and R2

T
T
principal = R1 R 1 = R 2 1 R2
R2
R1 R1
Therefore:
R = R2
R2 = R2
R 2 1 R 2
R1 , R = R1
R 2 = 1
R2
We find R 1 and R2 as usual:

=
R1
n (0.707, 0.612, 0.36)

1 = 3
1.0 1.73 1.0
1.73 0.75 0.433 2 = 0 m (0.0, 0.50, 0.866)
1.0 0.433 0.25
3 = 1
p ( 0.707, 0.61, 0.36)
0.707 0.61 0.36

=
0.0 0.50 0.866
0.707 0.61 0.36
So,
0.50 1.414 0.50

= 1.414 1.0 1.414
0.50 1.414 0.50
R2 =
n (0.50, 0.707, 0.50)

1 = 3
2 = 0 m (0.707, 0, 0.707)
1 = 1
p ( 0.50, 0.707, 0.50)
0.50 0.707 0.50

0.707 0.0 0.707
0.50 0.707 0.50
R R2
Therefore:
17.
R1 =
0.71 0.36 0.61

0 0.86 0.51
0.71 0.36 0.61
Imagine that we define a new measure of stress, [S], as a matrix of components relating force
components to area components, but that the force components are defined in two ways: 1. in
terms of a different coordinate system than the area components, or 2. the are transformed
according to a fixed linear operation to represent a new vector in the same coordinate system. a)
Is [S][ symmetric? b) According to these two definitions, does [S] represent the components of
tensor?
SOLUTION:
In either of cases 1 or 2, we note that the new force components (let us call these components gi)
may be obtained from the standard force components fi as follows (note that by "standard" we
mean the components of a force as normally defined in the same coordinate system used to
Chapter 3
Page 19
express vector area components):
g = L f , where [L] is a linear operator (rotation or other transformation matrix)

a. The definition of [S] follows from the expression of [g]:
g
= s a
L f = s a,
f =
s a
= L
s , or
= L
Since [L] is a general, non-symmetric matrix, [s] is in general not symmetric.

b. In order to examine how the new stress measure [S] transforms, let's imagine that we want to
x = R x . In the new coordinate system, our
express [S] in a new coordinate system:
definitions will be expressed as follows:
g = S a , where a = R a
The central question, the one that differentiates Case I from Case II, is: What is the meaning of
g ?
g is found in the new coordinate system by
Case 1 - According to the definition of Case 1,
applying the fixed linear operator, [L], to the components of [f] in the new coordinate system:
g = L f = L R f ,
g is found, we can find how S and S are related:

once this expression for
g = L R f = S a = S R a
f = R
g =
S R a , but note that f
L R
S R
S = L R
S R , or
= L
, so
a
, and therefore
S = L
R L S R
Clearly this last expression is not the proper transformation for tensor components, so [S]
defined as in Case 1 does not represent tensor components.
g is found in the new coordinate system by

Case 2 - According to the definition of Case 2,
simply transforming the compents of g as any other vector in the original coordinate system, i.e.
g = R g ,
Page 20
once we have this expression for
and S :
= S a
g = R g
S R
= R
g , we proceed as before to find the relationship between S
= S R a
S
S
Chapter 3
S R , or
, and therefore
S
= R S R
This last expression is precisely the transformation for tensor components, so [S] defined in Case
2 does represent tensor components. Put more simply, S defined according to Case 2 is a proper
tensor. In fact, representation of stress in this manner is convenient in some cases, where the
force or area vectors are rotated to correspond to deformed or undeformed states in a material.

F
1.
Given:
12
34
a. Find Eij, and ij , the components of the large and small strain tensors, respectively.
b. Using E and directly, find the new length of the vectors OA and AB shown below. Note
that the original vectors are of unit length.
A
0
c. Why are the deformed lengths of O'A' and O'B' different when calculated using the two
different measures of deformation?
SOLUTION:
a. We have from the definitions:
C = F T F = 1 3 1 2 = 10 14
2 4 3 4
14 20
so that the strain tensor E is:
7
E = 1 C - I = 4.5
2
7 9.5
while the small strain tensor is:
= 1 (F T+ F )- I = 0
2
2.5
2.5
3
b. The change of length l of the OA vector is such that:

2
7
1 =9
lA = OA T 2 E OA = 1 0 2 4.5
7 9.5
0
The final length is then: lA =
OA2 + lA = 10
For the OB vector the same approach gives:

2
lB = OB T 2 E OB = 0 1 2 4.5
7
with the final length: lB =
OB2 + lB = 20
7
9.5
0 = 19
1
Page 2
Chapter 4
With the small strain tensor [] the new lengths are:

T
2
l A = 1 + OA 2 OA = 1 + 1 0 2 0 2.5 1 = 1
2.5 0 0
and
T
2
l B = 1 + OB 2 OB = 1 + 0 1 2 0 2.5 0 = 7
2.5 3 1
c. The results are quite different when we use the two strain measures: in fact the use of the []
tensor is not valid here as the strain components are not much less than 1, as required for
accuracy.
2.
Given the figure below for an assumed homogeneous deformation, write down the deformation
gradient, F:
X2
(6.6, 4.6)
A'
O'
(6, 3)
B'
(8.1, 3)
(0, 2) A
O
B
(2, 0)
(0, 0)
X1
SOLUTION:
The unknown deformation gradient is denoted by:

a b 2 = 8.1 - 6 2 a = 2.1
c d 0
2c
3-3
0
and
a b 0 = 6.6 - 6 2 b = 0.6
c d 2
4.6 - 3
2d
1.6
we conclude that:
F = 1.05 0.3
0 0.8
F , with F =
3.
Given:
0.1
0.3
0.7
0.2
0.4
0.8
0.5
0.6
0.9 .
Find: J, E, and .
SOLUTION:
From the equality: [J] = [F] - [I] we get:
-0.9 0.2 0.5
J = 0.3 -0.6 0.6
0.7 0.8 -0.1
F = a b
c d . We must verify:
Chapter 4
Page 3
The strain tensor [E] is writen:

0.1
0.2
0.5
E =1 FT F - I =1
2
2
= 1
200
1+3*3+7*7-100
sym
sym
-0.205
0.35
0.43
0.35
-0.08
0.53
0.3
0.4
0.6
0.7
0.8
0.9
0.1
0.3
0.7
0.2
0.4
0.8
0.5
0.6
0.9
1.0
0
0
0
1.0
0
0
0
1.0
2+3*4+7*8
5+3*6+7*9
2*2+4*4+8*8-100
2*5+4*6+8*9
sym
5*5+6*6+9*9-100
0.43
0.53
0.21
The small strain tensor is:
=1
2
4.
-0.9 0.3 0.7

-0.9 0.2 0.5
0.3 -0.6 0.6 + 0.2 -0.6 0.8
0.7 0.8 -0.1
0.5 0.6 -0.1
-0.9 0.25 0.6

= 0.25 -0.6 0.7
0.6 0.7 -0.1
As shown below, a point in a continuum (O) moves to a new point (O') as shown.
A'
x2
B'
O' (4,2)
1.5 A
B
O
x1
1.5
a. Find the new points A' and B', assuming homogeneous deformation for the following two
cases:
F 1 2 , J 2 1
4 3
3 4
b. For each deformation, find E the large strain tensor.
SOLUTION:
By definition of the deformation gradient (when there is an homogeneous deformation), we
obtain for the first case:
0 = 3
OA OA with: OA = F OA = 1 2
1.5
3 4
6
OB OB with: OB = F OB = 1 2
3 4
and
1.5
0
1.5
4.5
Page 4
so that:
Chapter 4
OA = OO + OA = 4 + 3 = 7
2
6
8
4
1.5
OB = OO + OB =
+
= 5.5
2
4.5
6.5
For the second case we have first:
F = J + I = 3 1
4 4
The transformation verifies:

OA OA with: OA = F OA = 3 1
4 4
0
1.5
1.5
6
OB OB with: OB = F OB = 3 1
4 4
and
1.5
0
4.5
6
OA = OO + OA = 4 + 1.5 = 5.5
2
6
8
so that
OB = OO + OB = 4 + 4.5 = 8.5
2
6.0
8
b. For the first case we have

E =1 F TF - I = 1
2
2
1 3
2 4
1 2 - 1 0
3 4
0 1
4.5
7
7
9.5
and for the second case
E =1
2
5.
J + J
+ J
=1
2
21 + 2 4 + 24
13
13
43
21
43
= 12 9.5
9.5 8.0
Imagine that a line segment OP is embedded in a material which is deformed to a new state. The
line segment becomes O'P' after deformation, as shown below.
P'
x 2
O'
P(2, 1)
O
a. Find the vector components of O'P' :
F 1 2
3 4
b. Find the length of O'P' if:
x1
Chapter 4
Page 5
C= 1 3
3 2
c. Find the components of OP if:

J= 0 2
3 3
SOLUTION:
a. If the deformation is homogeneous we can write:
OP OP such that: OP = F OP = 1 2
3 4
b. According to Eq. 4.19:
2
OP = OP T C OP = 2 1
1 3
3 2
2 = 4
1
10
2 =
2 1
1
5 = 18
8
the final length is then: OP = 18 4.24

c. From [J] we deduce:
OP OP such that: OP = J + I OP = 1 2
3 4
6.
2 = 4
1
10
A homogeneous deformation is imposed in the plane of the sheet. Two lines painted on the surface
move as shown below, with coordinates measured as shown:
x 2
x 2
(3.1, 3.3)
(1.8, 2.1)
(3, 2)
(2, 1.5)
(4.1, 1.9)
(1, 1)
x1
x1
Before
After
a. Find F, the deformation gradient.

b. Starting with F, find C, E, and .
c. Find the principal strains and axes of E.
SOLUTION:
a. The method is similar to that used for Problem 2. We express that the deformation gradient
applied to the two vectors (as rows) gives their transformed coordinates:
a b
1 0.8 = 1.1 0.1
c d
0.5 1.1
-0.1 1.3
Page 6
Chapter 4
The best method is to invert the matrix on the left hand side:
1 0.8 -1 = 1 1.1 -0.8
0.7 -0.5 1
0.5 1.1
so that the unknown deformation gradient is:
a b = 1.1 0.1
1 1.1 -0.8 = 1 1.16 -0.78 1.66 -1.11
0.7 -0.76 1.38
c d
-0.1 1.3 0.7 -0.5 1
-1.09 1.97
b. The usual tensors are computed:
C = 1
0.7
E =1
2
=1
2
1.16 -0.76 1
-0.78 1.38 0.7
3.92-1 -3.99 =
-3.99 5.13-1
1.66-1 -1.11
-1.09 1.97-1
1.16 -0.78 = 1
-0.76 1.38 0.49
1.9232 -1.9536 3.92 -3.99

-1.9536 2.5128
-3.99 5.13
1.46
-2
-2 2.07
+ 1.66-1 -1.09
-1.11 1.97-1
= 0.66 -1.10
-1.10 0.97
(But again in this case the small strain tensor has no precise meaning, as its components are not
small with respect to 1).
c. The principal strains are solutions of the eigenvalue problem:
1.46-
-2
= 0, or (1.46- ) (2.07-) - 4 = 0
-2 2.07-
which is also:
2
- 3.53 - 0.98 = 0 with solutions E1 = -0.26 and E2 = 3.79
The principal axes v 1 and v 2 are such that their components verify:
(1.46 + 0.26) v11 - 2 v21 = 0 v1 = 0.76
0.65
(1.46 - 3.79) v12 - 2 v22 = 0 v2 = -0.65
0.76
7.
At time t, the position of a material particle initially at (X1, X2, X3) is
x1 = X1 + a X2
x2 = X2 + a X1
x3 = X3
Obtain the unit elongation (i. e. change in length per unit initial length) of an element initially in
the direction of x 1 + x 2 .
SOLUTION:
Chapter 4
Page 7
1 a 0
x
F=
, F = a 1 0
X
0 0 1
The deformation gradient is here:
1+a2 2a
C =
and the Cauchy strain tensor is immediately deduced:
2a 1+a2
1
2
1
dS
2
The initial unit vector is:
The length of the final corresponding vector verifies:
2a 1+a2
1
2
1 dS = (1 + a)2 dS2
2
1+a2
ds2 = dS T
C dS = 1 1 0 dS
2 2
2a
so that ds = (1 + a) dS and the unit elongation is: E = a.
8.
X i
Take fixed right handed axes x1, x2, x3. Write down the deformation gradient matrix, xj , for the
deformation of a body from x to X for
a. right handed rotation of 45o about x 1 .
b. Left handed rotation of 45o about x 2 .
c. Stretch by a stretch ratio of 2 in the x 3 direction.
1
d. Stretch by a stretch ratio of 2 in the x 2 direction.
e. Right handed rotation of 90o about x 3 .
Find the total deformation matrix for these motions carried out sequentially. Using this result,
check the final volume ratio.
SOLUTION:
The deformation gradients are computed at each step:
Page 8
Fa =
1 - 1 ,
2
2
1
1
2
2
1 00
Fc = 0 1 0 ,
0 02
Fd =
Fb =
1 0 0
0 1/2 0
0 0 1
1
2
0
0 - 1
2
1
0
1
2
1
2
0 -1 0
Fe = 1 0 0
0 0 1
The combined transformation has a deformation gradient:

0 - 2 2
4 4
2 -1 -1
F = Fe Fd Fc Fb Fa =
2
2
2
2
The Cauchy strain tensor is then evaluated:
1
2.5 1.06 1.06
C = 1.06 1.38 1.12
1.06 1.12 1.38
We see immediately that: C = 1, i. e. the deformation takes place with no volume change.
9.
From the following mapping, find C, U, and R:

x1
X1
2 0 0
x2 = 0 3 4
X2
x3
X3
0 4 -3
Check whether this is a permissible deformation in a continuous body.
SOLUTION:
The Cauchy strain tensor is first calculated by:
2 0 0
2 0 0
4 0 0
C = 0 3 4
0 3 4 = 0 25 0
0 4 -3
0 4 -3
0 0 25
The polar decomposition is written: [F] = [R] [U]; with [U]2 = [C] we deduce:
1 0 0
1
0
0
2
2 0 0
2 0 0
3
4
U = 0 5 0 and R = F U -1 = 0 3 4
0 1 0 =
0
5
5
5
0 0 5
0 4 -3
4
0 0 1
0
-3
5
5
5
But we see that R = - 1 so that [R] is not a rotation, but rather an inversion.
Chapter 4
Chapter 4
10.
Check the compatibility of the following strain components:

x1 2 x1 2 +x2 2
x1 +x2 x1 +2x2
0
x1 +2x2 x1 +x2
x1 2 +x2 2 x2 2
0
0
0
x3
0
x2 x3
Page 9
x2 x3
0
0
SOLUTION:
x 1 +x 2 x 1+2x 2
= x 1+2x 2 x 1+x 2
0
0
0
0
x3
From the strain tensor:

the components of which are first order polynomials, we see that: (see also Problem 15)
2 ij
2 ii 2 jj
+
=2
=0
x2j
x2i
xi xj
so that the field is compatible.
x1 2
x1 2 +x2 2
x2 x3
x2 2
= x1 2 +x2 2
The second strain tensor is:
x2 x3
2 ij
2 2 2 jj
+
=2
=0
2
2
x
x
x
x
j
i
1
2
we observe that:
2 22 2 33
2 23
+
=0=2
x23
x22
x2 x3
2 33 2 11
2 13
+
=0=2
x21
x23
x1 x3
So that this tensor field is also compatible.
B.
DEPTH PROBLEMS
11.
Consider the extension of an arbitrary small line element AB. Start by examining how (A'B')2 is
related to (AB)2 using the small extensional strain along that direction, en. Show that for small
strains and displacements, rotations do not cause extension, i.e. if extensions are zero, strains are
zero.
SOLUTION:
We put: AB = a dl, where a is a unit vector, and write Eq. 4.18 in the form:
Page 10
Chapter 4
AB = (a dl)T C a dl = dl2 aT(I + J)T (I + J) a = AB2 a T (I + J T + J + J T J) a

Using the small strain approximation we obtain:
2
AB AB2 a T (I + J T + J) a = AB2 a T (I + 2 ) a
As the components of the strain tensor are small with respect to unity we can also write:
AB AB (1 + 2 a T a)1/2 AB (1 + a T a)
If we have only an extension en in the x1 direction, the previous equation becomes:
1
AB AB (1 + en ) (where we used a = 0 )
0
If there is zero extension in any direction we have:
AB = AB = AB (1 + a T a)
for any a, that is also:
aT a = 0
If a is a principal direction, the above equality shows that the corresponding principal strain must
be zero, so that we come to the conclusion that the strain tensor must be null.
12.
In sheet forming, one often measures strains from a grid on the sheet surface. Then, one can plot
these strains as a function of the original length along an originally straight line:
Original Line
Strain
Die
Die
PUNCH
Original Position
As shown in the second figure (for a simple forming operation), this originally straight line is
curved and stretched.
If the edges of the sheet do not move (stretch boundary conditions), develop a rule that the
measured strain distribution must follow. Consider that the original sheet length lo becomes l at
some later time.
SOLUTION:
We consider the strain l or e1 (true or engineering definition) corresponding to the elongation in
the initial direction of the line. A small vector dl0 on the initial line will be transformed into a
Chapter 4
Page 11
small vector dl so that its new length verifies: dl = exp(l ) dl0 = (1 + e1 ) dl0 . Therefore the final
length of the line will be:
l0
l=
l0
exp(l ) dl0 =
(1 + e1 ) dl0
13.
Consider a 1" square of material deformed in the following ways:
dx 2
dx2
1
0
dx1
a)
dx1
3
dx 1
30
0
b)
c)
45
0
dx2
dx
1
dx1
dx
2
dx2
d)
Find F for each case.

Find C and E.
Find the principal values and directions of E.
Find the material principal directions after deformation.
Which of these cases (a-d) are mechanically the same under isotropic conditions? Under
general anisotropy?
SOLUTION:
a. We have:
F = 3 0
0 2
C = 9 0
0 4
E =
4
0
0
3/2
The principal strains are obviously: E1 = 3/2, E2 = 4, the principal axes and the material axes
after deformation are the ox1 and ox2 axes.
b. The new vectors are projected on the ox1 and ox2 axes so that we have:
3 3/2 -1 = 9 0 ,
F = 3 3/2 -1
, C = 3 3/2 3/2
3/2
3
3
3/2
3
0 4
-1
The remaining is the same as case a.
c.
F = -3 0 ,
0 -2
C = 9 0
0 4
E =
4
0
0
3/2
E =
4
0
0
3/2
Page 12
Chapter 4

d.
F = 3 2/2 - 2
3 2/2 2
C = 3 2/23 2/2 3 2/2 - 2

- 2
2 3 2/2 2
= 9 0
0 4
E =
4
0
0
3/2

For isotropic materials the four cases a-d are equivalent by definition, as a rotation has no effect
on the material properties. Under general anisotropy case a is equivalent to the undeformed
configuration (if the deformation does not introduce any change in the anisotropic coefficients),
while cases b, c, d correspond to rotations of 30; 180 and 45 respectively, and the material
properties will be rotated by the same rotation.
C = 1 0
0 2 , find F when
14. Given that
a. the principal material axes do not rotate, and
b. the principal material axes rotate by 30o counterclockwise. Express your answers in the
original coordinate system.
SOLUTION:
a. We use the polar decomposition: F = RU, if there is no rotation, then R = I, F = U is symmetric
and C = FF so that here F is easily computed:
F = U = 1 0
0 2
R =
b. The rotation matrix is:
3
2
1
2
-1
2
3
2
The [U] stretch tensor is the same and the deformation gradient is:
3 -1
3 - 2
1 0
2
2
2
2
F = R U=
=
0
2
1
3
1
6
2
2
2
2
15.
Derive a set of compatibility equations corresponding to Eq. 4.52 for the three-dimensional case.
SOLUTION:
A non diagonal (small) strain component is calculated according to:
U U j
ij = 1 ( i +
) for i j
2 xj xi
so that we obtain:
Chapter 4
2 ij
3 Uj
3 Ui
=
+
for i j
xi xj xi x2j x2i xj
which is also, with the definition of the diagonal terms (without summation on i and j):
2 ij
2 ii 2 jj
2
=
+
xi xj x2j
x2i
Page 13

1.
Adapt the demonstration of the Green theorem of Eq. 5.1 to:

- a curve C, and show that:
dg
f
dl = f(B) g(B) - f(A) g(A) C dl
df g dl
C dl
where A and B are the extremities of the curve. Examine the case when the curve is closed
(A=B).
- a surface S defined in the plane (with coordinates x1 and x2), so that:
g
f
f
dS =
f g ni dl g dS
S xi
S
S xi
i = 1 or 2, S is the curve limiting the surface S (with no hole), and n is the unit normal vector
to S.
SOLUTION
We suppose that the curve is represented by parametric equations of the form:
xi = xi (l) for 0 l L, where l is the arc lenght measured from the origine A of the curve. f and g
can be considered as functions of the arc lenght l on the curve and will be denoted as:
F(l) = f(x1 (l), x2 (l), x3 (l)) and G(l) = g(x1 (l), x2 (l), x3 (l))
so that the left hand side of the 1-D equivalent of the Green equation becomes:
L
dg
dG(l)
f
dl =
F(l)
dl
dl
C dl
0
Using a general property of integrals of functions of one variable, we can write:
L dF(l)
L
L d
dG(l)
G(l) F(l) dl = F(L) G(L) - F(0) G(0) =
G(l) dl +
F(l)
dl
dl
0 dl
0
0 dl
If this equation is combined with the previous one, we get:

L
F(l)
0
dG(l)
dF(l)
dl = F(L) G(L) - F(0) G(0) G(l) dl
dl
dl
0
which is the desired equality when we come back to the initial notations.
The proof for the 2-D Green formula is very similar to the proof for the 3-D case. The surface S is
represented in the figure below.
Page 2
Chapter 5
SU
x2
S
SL
xA
xB
x1
The boundary of S is divided into an upper part U, and a lower part L the equations of which
are respectively:
x2 = hU(x1 ) ; x2 = hL(x1 )
xB
g
f
dS =
x2
Using these notations, we can write:
hU(x1 )
f
xA
hL(x1 )
g
dx dx1
x2 2
The above equation is transformed in a similar way as Eq. 5.3 so that:

hU
hL
U
g
f
f
dx2 = fU gU - fL gL g dx2
x2
hL x2
This equation is now integrated with respect to x1:

xB
xB
g
dS =
fU gU - fL gL dx1 x
2
S
xA
f
xA
hU
f
g dx2 dx1
x
2
hL
In a way similar to Eqs. 5.3a and 5.3b, the component nU2 of the normal to U is such that
dx1 = nU2 dl and an analogous equation for the lower part of the boundary. Then we get:
xB
xA
xB
fU gU - fL gL dx1 =
xA
xB
fU gU nU2 dl -
xA
fL gL nL2 dl =
fg dl
When this last result is put into the previous equation we obtain the 2-D form of the Green theorem:
f
g
f
dS =
fg n2 dl g dS
S x2
S
S x2
2.
Apply the 2-D form of the Green theorem (see problem 1) to the functions f = x1 + x2 and g = x1 x2 in the square domain [0,1]2, for i = 1 and for i = 2. Compute directly the integrals and verify
the Green theorem on this specific example.
Chapter 5
Page 3
SOLUTION:
We must show that the following equality holds:
1
g
f
dx2 dx1 =
fg n2 dl 0 x2
S
f
g dx2 dx1
0 x2
The first integral in the above equation is computed according to:

1
(x 1 + x 2) (-1) dx 2 dx 1 = 01- x 1x 2 +
2
x1
x 2=1
dx 1 =
x 2=0
2
x1
x1
1
- (x 1 + ) dx 1 = +
2
2
2
0
= -1
0
On the boundary of the square, the n2 component is non-zero only on the segments corresponding to
x2 = 0 or 1, so that the second integral is:
1
S
fg n2 dl =
(x1 + 1) (x1 - 1) dx1 -
x21 dx1 = -1
Finally the third integral is written:

1
1
1
0
(x 1 - x 2) dx 2 dx 1 =
x 1x 2 0
2
x1
x 2=1
x 2=0
dx1 =
2
x1
x1
1
(x 1 - ) dx 1 =
2
2
2
0
=0
0
Using the numerical values of the three contributions allows verification of the Green theorem for
this particular case.
3.
A 2-D velocity field is given by the expressions:

v1 = a x12 + 2 b x1 x2 + c x1 x2
v2 = ax12 + 2 bx1 x2 + cx1 x2
depending on the 6 parameters a, b, c, a', b', c'. Determine the relations the parameters must
satisfy in order that the velocity field is incompressible. Keeping these relations in mind, show
directly that the material flow through the hatched unit triangle pictured below is equal to zero.
Page 4
Chapter 5
x2
1
x1
SOLUTION:
The partial derivatives are first evaluated by:
v1
v
= 2 a x1 + 2 b x2 ; 2 = 2 b x1 + 2 c x2
x1
x2
They permit us to express the incompressibility equation as:
div(v) = 2(a + b) x1 + 2(b + c) x2 = 0
which can hold in a 2-D domain only if:
a + b = 0 and b + c = 0
The hatched triangular domain is denoted by S, with boundary S; the material flow through S
(with outward normal) is evaluated by;
1
S
v.n dl = 0
v2 dx1 -
(v1 + v2 )
2dx1
2
0
v1 dx1 +
We see immediately that:

1
0
v2 dx 1 =
1
0
a x 1 dx1 =
a
, and
3
1
0
v1 dx 1 =
1
0
c v 2 dx 1 =
c
3
The last integral is evaluated on the side with equation x2 = 1 - x1:

1
0
(v1 + v2 ) dx 1 =
=
1
0
(ax 1 + 2bx 1 (1-x 1) + c(1-x 1 ) 2 + ax1 + 2 bx 1(1-x 1 ) + c(1-x 1) 2) dx 1
1
(a + b + c + a + b + c)
3
The three contributions are added to get (a + b + b + c)/3, which was shown to be equal to zero
when the velocity field is incompressible.
4.
0
A porous material with an initial relative density r < 1 is densified by a uniform rate of volume
change equal to c. Express the law of relative density r as a function of time t, the beginning of
the process corresponding to t = 0.
Chapter 5
Page 5
SOLUTION:
We suppose that the initial volume of porous material is V0 so that the current volume is V = V0 - ct
(with c > 0), and the initial mass is m. By definition the density will be:
1
1
=m= m = m
= 0
V V0 - ct V0 1 - c t/V0
1 - c t/V0
where 0 is the initial density. Dividing each members of the previous equalities by the maximum
density m (which corresponds to a perfectly dense material), we obtain the relative density:
1
r = r0
for t vc0 (1 - r0 )
1 - c t/V0
as a function of time and the initial relative density r0 .
If we consider now that the relative volume rate of change is constant we shall put V/V = - c, so that
with Eq. 5.42 we can write:
/ = r/ r= - div(v) = - V/V = c
The last equation is easily integrated with respect to time:
ln( r ) = ct
r0
and the final expression for relative density change is:
- ln( r0 )
r = r0 exp(ct) for 0 t
c
The plane strain upsetting of a rectangular section with height 2h and width 2a is considered (see
figure below).
-v0
y
2h
5.
2a
The velocity field at any moment is given by :
x
y
vx = + v0 , vy = - v 0
h
h
where vo is the (constant) upsetting velocity.
v0
Page 6
Chapter 5
dK
Express the kinematic energy K in the section, and the time derivative dt . Calculate the local
acceleration of any material point with coordinate (x,y), and the local density of the time
derivative of kinematic energy ; verify that the integral of this density gives the previous value for
dK
dt .
SOLUTION:
h
K=
1 v 2 dS = 4
2
The kinetic energy K is computed by:

after explicit integration one obtains:
1 (v2 x2 + v2 y2 ) dx dy
0 2
0 2
h
h
0 2
3
K = 2 v20 (a + ah)
3
h
dK = 2 v2 (3a2 + h) da + ( -a3 + a) dh
0
dt
dt
h
dt 3
h2
The time derivative of the kinetic energy is:
da
a
= vx (a) = v0 ,
h
Using the expression of the velocity field, we get: dt
which permits us to rewrite the derivative of the kinematic energy:
dh
= vy(h) = - v 0
dt
dK = 8 v3 a3
0 2
dt 3
h
Now the acceleration field is calculated by differentiation of the velocity field:

2
2 v0
dvx v0
x dh
x =
=
(vx )= 2 x ,
dt
h
h dt
h
y =
dv y
v0
x dh
=(v ) =0
dt
h y h dt
and the local rate of change of kinetic energy is:
.v =
2 v30
h3
x2
The integral of the above function in the rectangular domain is:

h
a 2 v3
3
0 2
. v dS = 4
x dx dy = 8 v30 a
3
3
h2
0 h
dK
0
, which is equal to dt .
6.
With the help of the general variational theory outlined in section 5.5, verify that the functional
defined by:
(u) = 1
2
ii )2 + 2
ij
2ij dV -
Ti ui dS
Chapter 5
Page 7
corresponds to the linear elastic problem for a material obeying the Hooke law, with a prescribed
stress vector T on the boundary s.
SOLUTION:
According to Eq. 5.82 we put:

F
F
=0 ;
=
ui
ii
F=1 (
2
ii + 2 ii = ii ;
i
2ij and f = -
ii ) + 2
i,j
Ti ui
i
F
= 2 ij = ij for i j
ij
According to the Hooke linear elastic theory, the last two terms were identified whith the
corresponding stress components, so that Eq. 5.83 is equivalent to the equilibrium equation with
ij
= 0 in
x
j
j
components:
Now f is differented with respect to ui:

which is the desired boundary condition.
ij nj on
j
A cylindrical sample is considered with length L and radius R, subjected to a prescribed tension
force F at its right end, while the left one remains fixed (see figure below).
7.
f
- Ti +
= - Ti
ui
, so that Eq. 5.84 gives:
x3
x1
u=
A linear displacement u is introduced with the form:
x2
u1 = U x1
u2 = V x2
u3 = W x3
Compute the energy functional (Eq. 5.104) for an elastic material with Lam coefficients and
. The minimization of allows to determine explicitly the unknown parameters U, V, W. Show
that the usual simple formulas are obtained when the Young modulus E end the Poisson
coefficient are used. It is recalled that the following equalities hold:
(3 + 2)
E=
=
+
2 ( + )
SOLUTION:
The components of the strain tensor are derived from the given displacement field:
11 = U ; 22 = V ; 33 = W ; ij = 0 for i j
The prescribed stress in the ox1 direction on the right hand side of the sample is:
T1 = F with S = R2
S
Page 8
Chapter 5
Then the functional associated to our problem is written:

F U L dS
(u) = 1
(U + V + W)2 + 2 (U 2 + V2 + W2 dV s S
2
= 1 R2 L (U + V + W)2 + 2 (U 2 + V2 + W2 - F U L
2
The minimization of with respect to the unknown parameters U, V, W is performed by writing:
= 0 = R2 L (U + V + W) + 2 U - F L
U
= 0 = R2 L (U + V + W) + 2 V
V
= 0 = R2 L (U + V + W) + 2 W
W
(3 + 2 ) (U + V + W) =
After summing up the above three equations we have:

U=
so that we obtain:
8.
+
1 F
1
F
=
,
3 + 2 R2 E R2
V= W=
F
R2
F
-1
1
F
=
E R2
2 3 + 2 R2
Find the viscoplastic potential for a material obeying the following constitutive equation:
2
s = 2K( 3) m-1 0 +
m-1
2
where 0 is a (small) positive constant. Show that when << 0 the material constitutive equation
is equivalent to a Newtonian behavior with the general form:
s = 2
and that if >> 0 , then it is equivalent to a Norton-Hoff behavior, Eq. 5.110.
SOLUTION:
The definition of the equivalent strain rate is:
2
from which we deduce immediately:
=2
3
2ij
i,j
d(0 + ) = 4
3
ij dij
i,j
d =
ij
dij
On the other hand, for any viscoplastic potential we have:

which is transformed with the help of the constitutive equation and of Eq. 5.106:
i,j
2 K ( 3)m-1 (0 + )(m-1)/2 ij dij
d =
i,j
Chapter 5

2
Page 9
Using the above result on the differentiation of (0 + ) we obtain:

2
2
2 (m-1)/2 3
2
d =
2 K ( 3)m-1 (0 + )
d(0 + )
4
i,j
This equation is easily integrated, giving:
2
2
= K ( 3)m+1 (0 + )
m+1
(m+1)/2
m+1
- 0
The integration constant was chosen so that (0) = 0. We verify that when 0 = 0 the usual
viscoplastic potential is found (Eq. 5.110).
B. DEPTH PROBLEMS
9.
A potential () is convex if for any 1 and 2 we have:

1 + (1- ) 2 ( 1 ) +(1- ) ( 2 ) for 0 1
Apply the above inequality to 1 = 0 and 2 = and express the derivative of each side at = 0 to
show that:
: ( ) 0
SOLUTION:
The convexity condition is written in the general form:
( 1 + (1 - ) 2 ) (1 ) + (1 - ) (2 ) for 0 1
We put 1 = 0 and 2 = in the previous equation so that:
((1 - ) ) (1 - ) () for 0 1
For = 0 both sides of the above equation are equal to (): therefore it can be rewritten after
substracting () from both sides, dividing by and changing the signs:
((1 - ) ) - ( )
( ) for 0 1
When tends to zero the left side tends to the derivative so that
()
ij () for 0 1
i,j ij
which is one of the desired inequalities, the other one resulting from the positiveness of which is
assumed by hypothesis.
10.
Give the proof for the generalized expression of the variational formulation to mechanical
problems in section 5.5.
SOLUTION:
We start from the functional of Eq. 5.82 and evaluate a small variation:
Page 10

I (u) =
F
u ui i
F dV +
ij ij
Chapter 5
f
u dS
uj j
Taking into account the symmetry of the strain tensor we verify that:
F
F
F ui
F uj
ij +
ji =
( ) +
( )
ij
ji
ij xj
ji xi
Then the Green theorem is used to write:
F (ui )
F
dV =
ui nj dS
x
ij
j
ij
F
( )
ij
ui dV
xj
Recalling that u is prescribed on u and has variation u only on s, the variation of the
functional can be written:
F
( )
ij
F
f F
I (u) =
ui dV +
+
n ui dS
ui i
ui ij j
xj
i
s i
At first we use a displacement field u which is nill on the boundary: the above equation shows that
I = 0 only if Eq. 5.83 holds. Then with displacement fields u which can take any value on the
boundary u, we see that the second integral of the above equation must be nill with the
consequence that Eq. 5.84 is satisfied.
11.
From the variational formulation given by Eq. 5.111, derive the constitutive equation and the
equilibrium equation for a Norton-Hoff viscoplastic material (Eq. 5.112). To obtain the
appropriate form it is suggested to utilize the following mathematical property of an
incompressible vector field v:
if div(v ) = 0, then a vector potential exists so that: v = curl ()
First verify that: curl (div(s )) =0, where s is the deviatoric stress tensor. Then, with the help of
the previous mathematical result, it is possible to show that div(s) is the gradient of a scalar field,
which is identified with the pressure field p.
SOLUTION:
A variation of the functional is calculated according to:
(v) =
2K 3
m-1
Tdi .vi dS
ij ij dV ij
Using the constitutive equation (Eq. 5.111) this is equivalent to:
Chapter 5
(v) =
Tdi .vi dS
sij ij dV
ij
Page 11
The Green theorem is used to transform the above equation in the same way as in Problem 10,
leading to:
sij
sij vi nj dS -
(v) =
s
i,j
ij
xj
vi dV i
Tdi .vi dS
Now v as well as v are incompressible fields. Then it can be shown mathematically that a vector
field does exist so that:
/x2 (3 ) - /x3 (2 )
v1
v = curl () with components
v2
/x3 (1 ) - /x1 (3 )
/x1 (2 ) - /x2 (1 )
v3
At least it is easy to verify that a velocity field v, derived that way from a vector field , is
necessarily incompressible.
The only volume integral occuring in the variational form is transformed using the Green theorem to
give:
- div(s ). curl () dV =
curl (div(s ). dV -
div(s ).( n) dS
If the vector field is chosen with a zero value on the boundary, we obtain:
=
curl (div(s )). dV
which can be always null only if curl (div(s )) = 0. This last condition is mathematically equivalent
to the existence of a function, which is denoted by p for convenience, which satisfies
div(s ) = grad (p).
This relation permits us to substitute div(s) in the initial functional variation:
p
sij vi nj dS -
(v) =
s
i,j
xi
vi dV s
Tdi .vi dS
Using one more time the Green theorem, we can write:

p
xi
vi dV =
p vi ni dS
vi
xi
dV
As v is incompressible the third integral in the above equation is equal to zero, so that:
Page 12
sij vi nj dS -
(v) =
i,j
ni vi dV -
Chapter 5
Tdi .vi dS
The velocity variation v can take any value on the boundary , therefore we obtain the equality:
s
sij nj - p ni =
ij nj = Tdi on
j
Which is the desired boundary condition.

A.
1.
Consider a cubic crystal with c11 = 2, c12 = 1, and c44 = 1.5.

a. Find c12', referred to the new coordinate system, if the x i ' axes are rotated 30o
counterclockwise about x 3 . The x i axes are aligned along the crystal axes.
x2 '
x2
x1 '
30
x1
b. What is the anisotropy parameter for this crystal?

SOLUTION:
R =
3
2
-1
2
1
2
3
2
a. According to Chap. 2, Eq. 2.23, the rotation matrix is:
0
O
1
If the fourth rank notation is used for the anisotropic tensor Eq. 6.9 shows that:
2
c12 = c1122 =
R1m R1n R2o R2p cmnop

m,n,o,p=1
Here the summations on m, n, o and p extend from 1 to 2 only as R13 = R23 = 0 . Now the only
non vanishing terms (with indices 1 or 2) in the c tensor for a cubic material are (see Eq. 6.18):
c1111 = c11 , c2222 = c11 , c1122 = c2211 = c12
c1212 = c2112 = c1221 = c2121 = c44
The requested component of the c' is then:
2
c 1122 = R 11 R 21 c 1111 + R 12 R 22 c 2222 + R11 R 22 c 1122 + R 12 R 21 c 2211

+ R 11 R 12 R 21 R 22 c 1212 + R 12 R 11 R 21 R 22 c 2112
+ R 11 R 12 R 22 R 21 c 1221 + R 12 R 11 R 21 R 21 c 2121
3 1
3 3 1 1
3 1
3
5
3
=2
c 11 + (
+
) c 12 - 4
c 44 = c 11 + c 12 - c 44
4 4
4 4 4 4
4 4
8
8
4
3
5
3
= 2 + 1 - 1.5 = 0.25
8
8
4
Page 2
b. The anisotropy ratio is defined in Eq. 6.28:

2.
A = c 2 c-44c
11
Chapter 6
= 2x1.5 = 3
2-1
12
Repeat Problem 1, b, for c11 = 1, c12 = 2, c44 = 0.5.
SOLUTION:
The same relation as in Problem 1 holds, giving:
c12 = 3 c11 + 5 c12 - 3 c44 = 3 1 + 5 2 - 3 0.5 = 5 = 1.25
8
8
4
8
8
4
4
The anisotropy ratio becomes:
3.
A = c 2 c-44c
11
12
= 2x0.5 = -1
1-2
Use symmetry to reduce the general 6x6 elastic constant matrix to the proper form for a tetragonal
cell as shown. (Note: all angles are 90o and a = b, so there is 90o rotational symmetry about x 2
and 180o rotational symmetry about x3.)
x2
b
a
c
x1
x3
SOLUTION:
The starting point is the orthotropic material where the anisotropic elasticity constants are given
by Eq. 6.17. Additional symmetry is taken into account: the 90 rotation around the x 2 vector.
The transformation is such that:
11 = 1
33 = 3
x1
x3
33 = 3
11 = 1
x2
x2
23 = 4
- 21 = - 6
x3
-x 1
12 = 6
23 = 4
and
We conclude that the following equalities must hold: c11 = c33 , c12 = c23 , c44 = c66
Chapter 6
The anisotropic coefficients are then:
4.
c11
c12
c13
0
0
0
c12
c22
c12
0
0
0
c13
c12
c11
0
0
0
0
0
0
c44
0
0
0
0
0
0
c55
0
Page 3
0
0
0
0
0
c44
a. Is (x, y) = x4 + y4 - 12x2y2 a valid Airy stress function in terms of both

equilibrium and compatibility requirements? Show why or why not.
b. Assuming that the (x, y) in Part a. is correct, find the stress components at the point
(1, 2).
c. Find the strain components for the point (1, 2) assuming plane stress conditions (31 = 0).
Leave your answers in terms of general Young's modulus (E) and Poisson's ration ().
SOLUTION:
4
4
2 2
a. From: = x + y - 6 x y
4
4
4
= 24 ;
= 24 ; 2
= - 48
y4
x2 y2
we compute: x4
which shows that the compatibility condition (Eq. 6.59) is satisfied.
b. The stress components are computed according to:
2
2
2
xx =
= 12 y2 - 12 x2 ; yy =
= 12 x2 - 12 y2 ; xy = = 24 x y
y2
x2
x y
The equilibrium equation is automatically satisfied; that can be verified on this peculiar case:
xx xy
= -24 x + 24 x = 0
+
x
y
xy yy
= 24 y - 24 y = 0
+
x
y
c. The corresponding strain components are as follows
xx = 1 + xx - 1 (xx + yy ) = 12 1 + (y2 - x2 )
E
E
E
yy = 1 + yy - 1 (xx + yy ) = 12 1 + (x2 - y2 )
E
E
E
xy = 1 + xy = 24 1 + xy
E
E
5.
Given: = 3xy5 + 3x5y - 10x3y3

a. Verify that meets the condition for a proper Airy stress Function. Show your work.
Page 4
Chapter 6
b. Find the stresses as functions of x and y that correspond to .

c. Check to see if the stresses in Part b. satisfy equilibrium. Show your work.
d. With a little manipulation using the stresses in Part b and Hooke's Law for plane stress,
obtain the strain functions shown below. Verify that these strain functions satisfy the
compatibility conditions for strains. Show your work.
SOLUTION:
5
5
3 3
a. The Airy function has the form: = 3 x y 5 + 3 x y - 10 x y
We see that the partial derivatives in Eq. 6.59 are:
4
4
4
= 360 x y ;
= 360 x y ;
= -360 x y
x4
y4
x2 y2
so that Eq. 6.59 which expresses the compatibility condition is satisfied.
xx =
2
= 60 (x y3 - x3 y)
y2
yy =
2
= 60(x2 y - 12 x y3 )
x2
xy = b. We deduce the stress components according to:
2
x y
= -15 (x4 + y4 ) + 90 x2 y2
c. The equilibrium equation is automatically verified by stress components deduced from an Airy
function, on our example we have:
xx xy
+
= 60 y3 - 180 x2 y + (- 60 y3 + 180 x2 y) = 0
x
y
xy yy
+
= - 60 y3 + 180 x2 y + 60 y3 - 180 x2 y = 0
x
y
xx = 4 b (xy3 - x3 y)
yy = 4 b (x3 y - xy3 )
xy = b (-x4 - y4 + 6x2 y2 )
(1 + )
where: b 15
E
d. The components of the strain tensor are calculated in the following way:
Chapter 6
xx = 1 + xx
E
yy = 1 + yy
E
xy = 1 + xy
E
When we put
Page 5
- 1 (xx + yy ) = 60 1 + (x y3 - x3 y)
E
E
- 1 (xx + yy ) = 60 1 + (x3 y - x y3 )
E
E
= 1 + (- 15 x4 - 15 y4 + 90 x2 y2 )
E
b = 15 1 +
E , the desired form is obtained.
The compatibility condition is written here:

2 xy
2 xx 2 yy
= 24 bxy + 24 bxy - 2 (24 bxy) = 0
+
-2
y2
x2
x y
6.
Verify that Eqs. 6.38-6.42 follow from the traditional definitions of elastic constants.
SOLUTION:
We consider a uniaxial tensile test in the x direction. Hooke's law in terms of Lam coefficients
(Eqs.6.31a-f) reduces to:
xx = (xx + 2 yy ) + 2 xx
yy = zz = (xx + 2 yy ) + 2 yy = 0
These equations can be compared to Eq. 6.1 - 5 for the same case:
xx = 1 xx ; yy = - xx or yy = - xx
E
E
We obtain the equalities:
xx = E xx = ((1 - 2 ) + 2 ) xx
yy = ((1 - 2 ) - 2 ) xx = 0
From which we deduce by subtracting the second one to the first one:
Then is immediately deduced (provided 1 - 2 0):
E
2 (1 + )
E
2 (1 + ) (1 - 2 )
From the Hooke's equation with the Lam coefficients the equality G = is obvious. The bulk
1 + +
11
22
33
B=3
= 1
11 + 22 + 33
K
modulus B is defined by:
The easiest way to establish the relation is to consider a spherical strain tensor so that
xx = yy = zz = . According to Hooke's law, the stress state is xx = yy = zz = (3 + 2 ) ,
B=+2
3
and the bulk modulus is:
Page 6
xx = 1 xx - (yy + zz) = 1 - 2 xx
E
E
E
The other form of Hooke's equation is here:

B=
so that the bulk modulus becomes:
Chapter 6
xx
E
=
1
2
3
(1
- 2 )
3
xx
E
Finally the relation between , and E, can be written:

2 (1 + ) = E
2 (1 + ) = E
or
(1 + ) (1 - 2 ) = E
(1 - 2 ) =
=
from which we deduce easily:
7.
(3 + 2 )
and E =
2 ( + )
+
Solve Eqs. 6.31 for strains and verify that the result is identical to Eqs. 6.1-5.
SOLUTION:
for i j ij = 1 ij = 1 + ij
E
2
First we see from Eqs. 6.31d-f and Eq. 6.38:

3
Then we add Eqs. 6.31a-c:
kk = (3 + 2 )
k=1
kk or
k=1
kk =
k=1
3
1
3+2
kk
k=1
Eq. 6.31a-c can be rewritten:
2 (3 + 2 )
Using Eqs. 6.38-39 and 6.42, we obtain:
=
2 (3 + 2 ) E , so that the last equation is
ii = 2 ii -
kk
k=1
ii = 1 + ii -
E
E k=1 kk , which is the desired result.
transformed into:
B. DEPTH PROBLEMS
8.
Discuss any material restrictions which apply to the compatibility conditions Eqs. 6.52 and 6.55.
Consider anisotropy, elasticity vs. plasticity or other constitutive equations, possible presence of
body forces, and the possibility of voids and cracks developing during deformation.
SOLUTION:
Eq. 6.52 does not use any additional physical assumption regarding the material. It is only based
on the continuity of the third order derivatives of the displacement field. In the presence of
cracks or a non differentiable boundary, this hypothesis must be analyzed carefully as the
Chapter 6
Page 7
displacement field may be discontinuous, thus violating the idea of a continuum of material upon
which compatibility is based.
On the other hand, Eq. 6.55 is established from Eq. 6.52 and the constitutive equation, so that it
can be valid only for isotropic linear elasticity (with or without body forces). This equation must
not be used for anisotropic elasticity, plasticity, elastoplasticity, etc.... An equivalent form can
be derived for a given relationship between stress and strain for other cases.
9.
a. Show that the elastic work done during small straining is given by
w = 1 cjkl ij kl .
2
b. Write the elastic work in terms of stresses alone.
c. Write results for parts a and b for the isotropic case.
SOLUTION:
a. We suppose for convenience that the strain is applied with a so called radial loading, i. e. we
suppose that loading is performed from 0 to t, and that for any time , with 0 t, the strain
=
= c : = c :
tensor is:
t and the corresponding stress tensor:
t
. During the time
d
d =
increment d the strain increment is:
t , so that the increment of work is deduced with
d
dw = :d = c : :
the help of Eq. 4.54:
t
t
w=
The total work is obtained by time integration:
c: : d = 1 c: : =
1 cijkl ij kl
t
2
t
2
0
ijkl
The assumption of radial loading is not required since, by definition, the elastic response is
conservative, i. e. no work is done in a closed loading cycle returning to the same loading state.
and
b. The compliance tensor s is introduced (see Exercise 6.1) which permits us to write: = s :
1
1
sijkl ij kl
w = s : : =
2
2
ijkl
the work in the form:
c. The isotropic expressions are derived from Eqs. 6.31 and 6.1 - 4:
2
2
w=
+
2ij = -
ii + 1 +
2ij
2 i ii
2
E
2
E
ij
i
ij
10.
a. Show that Hooke's Law for an isotropic material may be written in the following form
ij = 2 ij + ij ij
1 - 2
Page 8
Chapter 6
b. Find the equilibrium equation in indicial form in terms of strains alone.
SOLUTION:
ij = kk ij + 2 ij = 2 (ij + kk ij )
2
a. Hooke's law is written:
with the help of Eq. 6.39 it is easy to obtain the desired result:
b. The equilibrium equation is expressed as:

11.
div(
)i =
ij = 2 (ij +
)
1 - 2 kk ij
ij
kk
= 2 ( ij +
)=0
xj
xj 1 - 2 xj ij
Consider a plane-strain isotropic linear elastic problem defined in a given domain. Assume the
displacement vector field takes the form:
= a x1 2 + 2b x1 x2 + c x2 2
u= 1
2 = a' x1 2 + 2b' x1 x2 + c' x2 2
a. Write the strain tensor as a function of x1 and x2.
b. Compute the stress tensor as a function of x1 and x2 with the isotropic Hooke law.
c. Write the stress equilibrium equation assuming no body forces, and determine the relation
between the coefficients a, b, c, a', b', and c' that we must impose.
d. Calculate the stress vector on each side of a square (see figure below).
x2
1
-1
0
x1
-1
Verify that the external forces on the square are in equilibrium when the condition defined
in Part c is fulfilled.
SOLUTION:
a. The
derivatives
of
the
u1
u1
= 2 a x1 + 2 b x2 ,
= 2 b x1 + 2 c x2
x1
x2
u
u2
= 2 a x1 + 2 b x2 , 2 = 2 b x1 + 2 c x2
x1
x2
displacement
components
are
Chapter 6

=
The strain rate tensor is immediately obtained:
Page 9
2 a x1 + 2 b x2
(b+a) x1 + (c+b)x2
(b+a) x1 + (c+b)x2
2 b x1 + 2 c x2
11 = 2 (a+b) x1 + (b+c) x2 + 4 (a x1 + b x2 )
12 = 2 (b+a) x1 + (c+b)x2
b. The components of stress tensor are now: 22 = 2 (a+b) x1 + (b+c) x2 + 4 ( b x1 + c x2 )
11 12
+
= 2 (a+b) + 4 a + 2 (c+b) = 0
x1
x2
12 22
+
= 2 (b+a) + 2 (b+c) + 4 c = 0
x2
c. The equilibrium equation is written: x1
d. The stress vector on the upper side and its integral on this side are:
12
22
2 (b+a) x1 + (c+b)
2 (a+b) x1 + (b+c) + 4 ( b x1 + c) ,
12
-1 22
dx1 =
4 (c+b)
4 (b+c) + 8 c
On the lower side:
1
4 (c+b)
-12 dx =
- 2 (b+a) x1 - (c+b)
- 12
1
=
-22
- 22
4 (b+c) + 8 c
-2 (a+b) x1 - (b+c) - 4 ( b x1 - c) , -1
On the right hand side:
11
12
2 (a+b) + (b+c) x2 + 4 (a + b x2 )
2 (b+a) + (c+b)x2
11
-1 12
dx2 =
4 (a+b) + 8 a
4 (b+a)
On the left hand side:
1 -
4 (a+b) + 8 a
- 2 -(a+b) + (b+c) x2 - 4 (a + b x2 )
11 dx =
-11
2
=
-12
-1 -12
4 (b+a)
-2 (b+a) - (c+b)x2
,
When all these contribution are added and the result equated to zero we obtain:
4 (c+b) + 4 (a+b) + 8 a = 0
4 (b+c) + 8 c + 4 (b+a) + 0 , which are equivalent to the equations we found in c.
12.
Consider 3 elastic bars of equal length which are pinned at the ends as shown and to which a
vertical force F is applied.
Page 10
Chapter 6
Each bar has a length l, a section area s, a Young's modulus E. We assume the links between the
bars do not permit any torque, the bars remain straight and the contact with the horizontal plane
is frictionless. Compute the variation of length of each bar, with the hypothesis of small
displacements. Consider the case where the variation of length is no longer negligible as
compared to the initial geometry.
SOLUTION:
Qualitatively we can see that a compression force F1 will occur in the two rods AB and AC,
while the horizontal rod BC will be subjected to a tension force F2.
F
A
F1
B
F1
F2
The equilibrium of vertical forces at point A gives: F = 2 F1 sin(60) or F1 = F/ 3 . If the

length
variation
l1
of
AB
or
AC
is
introduced,
we
have
also:
l
F1 = F/ 3 = E a 1 or l1 = F l
l
Ea 3
F2 = F1 cos(60) = F
The horizontal force exerted in B on rod BC by rod AB is:
2 3 and the
l
F2 = F = E a 2 or l2 = F l
l
2 3
2 3Ea
length variation l2 of BC is computed by:
When the displacements are no longer neglected, we suppose that the new lengths of the rods
are:
AB = AC l1 = l + l1 ; BC l2 = l + l2
Chapter 6
F1 =
l + l1
l
=Ea 1
2
2
l
(l + l1 ) - 1 (l + l2 )
4
F =F
2 sin 2
The same approach gives:

F2 = F1 cos = F cos = F
2 sin 4
and for rod BC:
Page 11
l + l2
l
=Ea 2
2
2
l
(l + l1 ) - 1 (l + l2 )
4
We obtain a non linear system of two equations with the unknown l1 and l2 , which can be
solved iteratively with the Newton-Raphson method. Here we assumed Hooke's law and
suppose the sections of the rods did not change: the approach can then be still improved be
considering a nonlinear elasticity law and taking into account the area changes.
13.
Consider a plane-strain compression test of a sample with height h, imposing a displacement u at

the top as shown in the figure, assuming no friction on the plates.
x2
u
x1
Write the displacement field using a linear expression and taking into account the boundary
conditions and the symmetry of the problem: an unknown parameter should be introduced in
the x1 component. Calculate the strain tensor, the stress tensor with the isotropic Hooke law, and
the elastic energy. Show that the solution of the problem can be obtained by minimizing the elastic
energy with respect to the unknown parameter . Verify that the stress on the vertical sides are
equal to zero.
SOLUTION:
u =
The displacement and strain fields are:
u1
=
u2
x1
- u x2
h
; =
-u
h
and the stress tensor is deduced immediately with the help of the Hooke law:
( - u ) + 2
0
h
=
0
( - u ) - 2 u
h
h
The elastic density of energy w is homogeneous in the 2-D domain so that it is easily integrated
W = 1 ( - u )2 + 2 ( 2 - (u )2 ) 2 a h
2
h
h
to give:
Page 12
Chapter 6
where 2 a is the width of the sample. Minimization of the elastic energy yields:
W
= 0 = ( - u ) + 2 2 a h
h
The optimum parameter is:
+ 2 h and the corresponding stress tensor is written:
=
0
- 4
u
+2 h
We see that it is the exact solution of our problem.

14.
Repeat Problem 13 for a cubic crystal where x1 and x2 are oriented along four-fold symmetry
axes.
SOLUTION:
The displacement and strain fields are the same as those in problem 13, the non zero components
of the stress tensor become:
11 = c11 11 + c12 22
22 = c12 11 + c11 22
The elastic energy is derived according to:
W = 1 (c11 211 + 2 c12 22 11 + c11 222 ) 2 a h

2
The minimization of the elastic energy is achieved by putting:
W
= (c11 +c12 (- u )) 2 a h = cc12 u
11 h
h
The final solution in stress is:
11 = 0 , 22 = -
15.
c211 - c212 u
c11 h
Repeat Problem 13 with a 3-D sample and an orthotropic material.
SOLUTION:
The approach is similar to that of problems 13 and 14. We have:
Chapter 6
u1
u2 =
u3

x1
x2
- u x3
h
-u
h
Page 13
The non-zero components of the stress tensor are calculated with Eq. 6.17:
11 = c11 + c12 + c13 (- u )
h
22 = c12 + c22 + c23 (- u )
h
33 = c13 + c23 + c33 (- u )
h
and the elastic energy takes the form:
2
W = 1 4abh c11 2 + c22 + c33 (- u )2 + 2 c12 + 2 c13 (- u ) + 2 c23 (- u )
2
h
h
h
if 2a is the length and 2b the width of the sample. Minimization with respect to the two
unknown parameters gives the linear system of two equations:
W
= 0 = 4abh c11 + c12 + c13 (- u ) 11 = 0
h
W
= 0 = 4abh c12 + c22 + c33 (- u ) 22 = 0
h
= c22 c13 - c12 c23 u

c11 c22 - c212 h
The solution of which is:
16.
= c11 c23 - c12 c13 u

c11 c22 - c212 h
Show that the only incompressible isotropic elastic medium is a liquid by computing the shear
modulus, .
SOLUTION:
An isotropic linear elastic medium which obeys Hooke's law is incompressible if (see
0 = 11 + 22 + 33 = 1 - 2 (11 + 22 + 33 )
Eq. 6.1 - 5):
, which holds when = 0.5.
E
1=
2 2 (+ )
According to Eq. 6.41 we can also write:
, that is = 0 for any finite value of , so
that ij = 0 for any i j. Moreover the diagonal terms of the stress tensor do not depend on the
strain tensor so that by isotropy we have: 11 = 22 = 33 = - p. That corresponds to an inviscid
fluid.
17.
The Airy Stress Function for a screw dislocation in an isotropic crystal is given by
Page 14
Chapter 6
by
ln x2 + y2 ,where b = Burger's vector, a small constant.
4 (1-)
a. Find the stresses and strains for the screw dislocation.

b. Is compatibility satisfied everywhere? Why or why not?
c. Do plane strain or plane stress condition hold?
SOLUTION:
b
4 (1 - ) and r2 = x2 + y2 in order to simplify the expressions. Eq. 6.46 gives
a. We first put
the stress components:
2
2
2
3 x2 + y2
y2 - x 2
y2 - x2
xx = 2 = 2 k y
, yy = 2 = 2 k y
, xy = =2kx
4
4
4
x y
y
r
x
r
r
k=
According to Eq. 6.1 - 5 the strain components are:

xx = 2 k
y 1
y -
1
(3 x 2 + y 2) - (y 2 - x 2) , yy = 2 k
(3 x 2 + y 2) + (y 2 - x 2)
4 E
4
E
E
E
r
r
1+
y2 - x2
xy =
2kx
E
r4
-
y
zz =
4k 2,
E
r
b. We put:
2 2 4 k y
+
=
r2
x2 y2
we have to compute:
2 8 k y (y2 - 3 x2 )
2 - 8 k y (y2 - 3 x2 )
=
and
=
r2
r2
x2
y2
2 2
+
=0
2
2
x
y
so that the compatibility condition:
is fulfilled everywhere except at the origin.
c. The mechanical problem is obviously plane stress but not plane strain.
18.
The elastic fields of a screw dislocation are most simply derived by considering the displacements
in polar coordinates:
y
uz (r, ) = b = b tan-1 x , where b = the Burger's vector, a small constant.
2 2
a. Find the stresses and strains for the screw dislocation.
b. Is compatibility satisfied everywhere? Why or why not?
Chapter 6
Page 15
c. Do plane strain or plane stress conditions hold?

SOLUTION:
a. From the given displacement field we compute the strain and stress tensors in the usual way:
b y
0
0
y
ux = 0
0
0
- b
2 r2
r2
b x
=
=
uy = 0
0
0
0
0
b x
2 r2
r2
y
y
uz = b tan-1
b y b x
- b
b x
0
x
2
0
2
2
r
r
2 r2
2 r2
where we put:
r2 = x2 + y2 and b = b
2
b. Compatibility is necessarily satisfied everywhere (except at the origin) as the strain tensor is
obtained by differentiation of a displacement field.
c.
By examining the strain and stress tensors we conclude that we have both plane strain and
stress tensor fields (with respect to plane Oxy).

A.
1.
What is the meaning of stress states lying outside of yield surface?
SOLUTION:
They are states accessible only by changing the material properties - by strain hardening or heat
treatment perhaps. In either case, the yield surface must change in order to achieve these
stresses.
2.
A researcher has found a way to measure the yield surface of sheet metal rapidly and
automatically. He inserts a sheet into a biaxial testing machine and loads along proportional
paths until, while he measures strains in the two directions, he obtains the 0.2% offset strength.
Then he unloads, chooses a slightly different ratio and does the same thing, until he has generated
many yield points.
Criticize this procedure.
SOLUTION:
The major criticism is that the material properties are different after the first and each subsequent
test because the approximately 0.002 plastic strains accumulate and the material strain hardens.
Therefore, the various yield stresses measured are not for a single material and do not represent a
yield surface at an instant. A secondary criticism is the use of the 0.002 offset strength as the
yield strength. In practice, this is not a very practical problem, but the yield strength measured
in this way could be considerably greater or less than the actual yield.
3.
Critically evaluate the yield functions presented below in terms of isotropy, pressure-dependence,
and existence of a Bauschinger effect. Demonstrate your results. Why would you choose to use
each yield function?
a. Hill's 1 Quadratic Yield Function (1948)
2
2
f = F 2 - 3 + G 1 - 3 + H 1 - 2
b. Hill's 2 Non-Quadratic Yield Function (1979)

M
M
M
f = F 2- 3 + G 1- 3 + H 1- 2
c. Bourne and Hill 3 (3 = 0, not principal axes):
f = 3 x 3 - 6 x 2 y - 6 x y 2 + 4 y 3 + (4 x +21 y ) xy2
R. Hill: The Mathematical Theory of Plasticity, Clarendon Press, 1950, p. 318.

R. Hill: Proc. Roy. Soc., 1949, vol. 198, p. 428.
R. Hill: Math. Proc. Camb. Philos. Soc., 1979, vol. 85, p. 179.
P.B. Mellor and A. Parmar, Mechanics of Sheet Metal Forming, D.P. Koistinen and N.M. Wang, Plenum Press, 1978, p.67.
L. Bourne and R. Hill: Philos. Mag., 1950, vol. 41, p. 671.
Page 2

3
d.
Drucker 4 :
Chapter 7
f = J 2 c J3 , where c = constant
e. Edelman and Drucker 5 :
f =
1
2
c ijkl ij - M ij P kl - M kl P , c ijkl, M = constants, kl = plastic strain
f. Gotoh 6 (3 = 0):
4
f = A o x+ y +A 1 x 4+A 2 x 3 y+A 3 x 2 y 2+A 4 x y 3 +A 5 y 4

+ A 6 x 2 +A 7 x y+A 8 y 2 xy 2+A 9 xy 4, A 1 = constants.
g. Bassani 7 (3 = 0):
f = 1 + 2
+ (1 + 2r) k 1 - 2
,
r, k, M = constants.
h. Jones & Gillis 8 (3 = 0):

f = c 11 x 2 + c 12 x y + c 13 x xy + c 22 y 2 + c 23 y xy + c 33 xy 2, c ij = constants
i. Gupta 9 :
f =
J 2 - p J1 , = constant.
SOLUTION:
a. Not isotropic (unless F = G = H)
f i + p = f i
Pressure-independent [since
]
No Bauschinger effect [since f i = f i ]
Modification of von Mises to handle simple anisotropy - for example, for rolled sheet.
Allows different strains in two lateral directions in tensile test.
b.
Not isotropic (unless F = G = H)

Pressure-independent
No Bauschinger effect
Includes the anisotropy of Part a., but allows further adjustment of shape to account for
higher or lower flow strength in plane-strain or balanced biaxial tension (relative to uniaxial
tension) for a given r value.
c.
Not isotropic
Cannot determine pressure-dependence since only two principal stresses are represented
A Bauschinger effect is inherent (because of the cubic terms)
This is a purely 2-D yield function with fixed constants that was fit to experimental data from
4
5
6
7
8
9
D.C. Drucker: J. Appl. Mech., Trans. ASME, 1949, vol. 16, p. 349.
F. Edelman and D.C. Drucker: J. Franklin Inst., 1951, vol. 251, p. 581.
M. Gotoh: Int. J. Mech. Sci., 1977, vol. 19, p.505.
J.L. Bassani: Int. J. Mech. Sci., 1977, vol. 19, p. 651.
(unpublished)
Y.M. Gupta: Acta Metall., 1977, vol. 25, p. 1509.
Chapter 7
Page 3
tensile tests conducted at various angles in a sheet of metal. It provides a detailed variation
of y vs. but has no consistent 3-D form.
d.
e.
Isotropic (use of invariants proves this)

Pressure-independent (use of deviatoric components demonstrates this)
No Bauschinger effect (the odd invariant, J 3 , has been squared to avoid this.
This is a very interesting generalization of the simplest yield function, J 2 = k , which satisfies
pressure-independence, isotropy, and no Bauschinger effect. Because it is isotropic, it cannot
contribute to fitting strain ratio effects on sheet tensile tests, however. It is similar in concept
to Part b., with F = G = H.
Not isotropic (unless C ijkl = C mnop, M = 0, or ij = 0 )

Pressure-independent (from use of deviatoric components)
Bauschinger effect (unless M = 0)
This is the logical extension to the most general quadratic yield function in terms of
anisotropy, with kinematic hardening depending on total plastic strain.
f.
Not isotropic
Cannot determine pressure-dependence since only two principal stresses are represented
A Bauschinger effect is inherent (because of the cubic power terms)
This is a purely 2-D yield function with fixed constants that was fit to experimental data from
tensile tests conducted at various angles in a sheet of metal. It provides a detailed variation
of y vs. but has not consistent 3-D form.
g.
Isotropic (cannot distinguish 1 and 2 because of absolute values)

Cannot determine pressure dependence since only two principal stresses are represented
A Bauschinger effect is present (unless N and M are even)
Very similar to Part b., but the different power for the two terms means it must be handled
numerically.
h.
See Parts c. and f.
i.
Isotropic (use of invariants)
4.
Pressure-dependent (unless p = 0 )
Bauschinger effect (unless p = 0 )
This yield function is the simplest one which includes a pressure dependence, which is its
purpose. For example, compressible materials such as powders or sponges could be treated
with this function.
Write each of the yield functions in Problem 3 in terms of each of the following definitions of
effective stress:
a. tensile test in the x1 direction, = 1
b. tensile test in the x2 direction, = 2
Page 4
Chapter 7
c. balanced biaxial test, 1 = 2 =

d. shear test, same principal axes, 1 = , 2 =
SOLUTION:
x 1 tension:
a.
G + H 2 = f
x 2 tension:
F + H 2 = f
Bal. Biax:
F + G 2 = f
Shear (assume that = xy occurs 45 from principal axes, i.e. as stress state consisting of xy
is equivalent to 1 = , 2 = :
F + G + 4H 2 = f
b.
x 1 tension: = G + H M = f
x 2 tension:
F + H M = f
Bal. Biax:
F + G M =f
Shear:
c.
F + G + 2 MH M
(where 1 = , 2 = )
x 1 tension: 3 3 = f
x 2 tension: 4 3 = f
Bal. Biax: 5 3 = f
Shear: 3 = f
x 1 tension:
d.
(where 1 = , 2 = )
0 0
= 00 0 ,
00 0
2
3
0
1
3
0
0
0
0
0
13
Chapter 7
J2 = 11 22 + 22 33 + 11 33 = 2 1 2 + 1
3
3
3
13 2 + 23
Page 5
13 2 = 13 2
J3 = 11 22 33 = 2 3
27
1 6 C 4 6 = 1 1 4 C 6 = f
27
729
27
27
x 2 tension:
Same as x1 because of isotropy
Bal. Biax:
00
= 0 0 , =
0 00
1
3
0
0
0
0
0 23
1
3
This is the same as x1 and x2 tension because of a) isotropy and b) no Baushinger effect, so that a
stress state of ij is equivalent to - ij
Shear:
0 0
0 0 , =
0 0 0
0 0
0 0
0 0 0
J2 = 2 + 0 + 0 = 2
J 3 = 11 22 33 = 0 ,
x 1 tension:
2
3
so 6 = f
0
1
3
0
0
0
e.
[same as von Mises]
0
0
13
For simplicity, we assume no initial loading, such that ij = 0, then:

1 C
2
2 1111 3
2 2 + C
2
1122
3
3
1 2 + C 1133 2
3
3
1 2 +
3
C2211 13
2 2 + C
1
2222 3
3
13 2 + C 2233 13
13 2 +
C 3311 13
1
2 2 + C
3322
3
3
13 2 + C3333 13
13 2
= f
Page 6

2 4 C
1111 2 C 1122 + C 1133 + C 2211 + C 3311 + C 2222 + C 2233 + C 3322 + C 3333
18
Chapter 7
=f
x 2 tension:
2 4 C 2 C + C + C + C
2222
1122
2211
3311
1133 + C1111 + C3333 + C1133 + C3311
18
= f
2
Bal. Biax:
18 4 C 3333 2 C1133 + C3311 + C2233 + C3322 + C 1111 + C 2222 + C 1122 + C 2211
Shear:
f.
2 C + C C C
1111
2222
1122
2211 = f,
2
where: 1 = , 2 =
x 1 tension: 4 A 0 + A 1 = f
x 2 tension: 4 A 0 + A 5 = f
Bal. Biax:
4 16 A 0 + A 1 + A 2 + A 3 + A 4 + A 5 = f
Shear: 4 A 1 A 2 + A 3 A 4 + A 5 = f , where: 1 = , 2 =
g.
x 1 tension: N + 1 + 2r K M = f
x 2 tension:
N + 1 + 2r K M = f
Bal. Biax: 2 N N = f
Shear:
h.
1 + 2r k 2 M M = f ,
where: 1 = , 2 =
x 1 tension: c 11 2 = f
x 2 tension: c 22 2 = f
Bal. Biax: c 11 2 + c 12 2 + c 22 2 = f
Shear: c 11 2 c 12 2 + c 22 2 = f, where: 1 = , 2 =
=f
Chapter 7
i.
2
x 1 tension: J2 = 1
3
J1 = ,
therefore:
1
3
p =
1
3
p = f
x 2 tension: (same as x 1 , isotropic)

Bal. Biax:
2
J2 = 1
3 , J 1 = 2 ,
Shear: J 2 = 2
5.
J1 = 0 ,
therefore:
therefore: = f,
1
3
2 p = f
where: 1 = , 2 =
Derive the normality condition for each of the yield functions in Problem 3.
SOLUTION:
a.
d 1 = 2d G + H 1 H 2 G 3
d 2 = 2d F + H 2 H 1 F 3
d 3 = 2d F + G 3 G 1 F 2
b. Assume that 1 > 2 > 3 > 0 (other cases can be handled separately), then:
d 1 = Md G 1 3
M1
+ H 1 2
M1
d 2 = Md F 2 3
M 1
H 1 2
M 1
d 3 = Md F 2 3
c.
M 1
G 1 3
M 1
d x = d 9 x 2 12 x y 6 y 2 + 4 xy 2
d y = d 6 x 2 12 x y + 12 y 2 + 21 xy 2
d xy = d 4 x + 21 y 2 xy
d.
d 1 = d f = d 3 J2
1
J 2
J 3
2
2
2
c
J
3
1
1 = d 3 J2 1 3 c J 3 J2 + 3 2 3
Page 7
Page 8
= d J 2
2 1 2 3 2 c J3 2 1 2 2 2 3 2 2 1 2 2 1 3 + 4 2 3
9
f = d 3 J
d 2 = d
2
2
= d J 2
J 2
J 3
2
2
c
J
=
d
3
J
2 23 c J 3 J2 + 3 1 3
3
2
2
2
2 2 1 3 2 c J3 2 2 2 1 2 3 2 2 1 2 + 4 1 3 2 2 3
9
f = d 3 J
d 3 = d
2
3
= d J 2
Chapter 7
2
J 2
J 3
2
c
J
= d 3 J2 3 23 c J 3 J2 + 3 1 2
3
3
3
2 3 1 2 2 c J3 2 3 2 1 2 2 2 + 4 1 2 2 1 3 2 2 3
9
where, from Eqs. 3.34 and 3.35:

J2 = 1
6
1 2
+ 1 3
+ 2 3
= 13 1 2 + 2 2 + 3 2 13 1 2 + 1 3 + 2 3
1 2 2 2
J 3 = 1 2 3 = 27
1
2
3
2
1
3
3
1
2
= 2 1 3 + 2 3 + 3 3 1 1 2 2 + 1 2 2 + 1 3 2 + 1 2 3 + 2 3 2 + 2 2 3 + 4 1 2 3
27
9
9
P
e. Consider only the start of deformation, when ij = 0, and write in principal axes.
f = Cij i j ,
f = C = C j + 1
ij
i
j
ij
i
j
m
m
m
m
but note that since i = 2 i j k , where i j k , that
Therefore, we can write:

d 1
= 2 2C 11 + C 12 + C 13 + C 21 + C 31 1
d
1 2C22 + C21 + C 23 + C 12 + C 13 2
1 2C33 + C31 + C 32 + C 13 + C 23 3
, or
j
m =
2
1
if j = m
if j m
Chapter 7
Page 9
d 1
= 2 2C11 + C12 + C13 + C21 + C 31 2 1 2 3
3
d
13 2C 22 + C 21 + C 23 + C 12 + C 13 2 2 1 3
1 2C 33 + C 31 + C 32 + C 13 + C 23 2 3 1 2
3
The other components follow by inspection. (For example d 2 would look the same except that
1, 2 , 1
3 , or for d 2 the
the coefficients on the three terms would be, respectively, 3 3
1, 1 , 2
coefficients would become 3 3 3 .
f.
d x
3
= 4A o x + y + 4A 1 x 3 + 3A 2 x 2 y + 2A 3 x y 2 + A 4 y 3 + 2A 6 x xy 2 + A 7 y xy2
d
d y
3
= 4A o x + y + A 2 x 3 + 2A 3 x 2 y + 3A 4 x y 2 + 4A 5 y 3 + A 7 x xy 2 + 2A 8 y xy2
d
d xy
= 2 A 6 x 2 + A 7 x y + A 8 y 2 xy + 4A 9 xy 3
d
g. Assume that 1 > 2 > 0 for simplicity. Other cases may be derived separately.
h.
d 1 = d N 1 + 2
N1
+ 1 + 2r k M 1 2
d 2 = d N 1 + 2
N1
1 + 2r M 1 2
M1
M 1
d x = d 2c 11 x + c 12 y + c 13 xy
d y = d c 12 x + 2c 22 y + c 23 xy
d xy = d c 13 x + 2c 23 y + 2c 33 xy
i.
2
2
J2 = + 1
6 1 2
f = J2
1
2
p J1
+ 12 32 + 22 32
, therefore:
f = 1 J
2 2
1
12
, J1 = 1 + 2 + 3
J2
J1
p
1
1
Page 10
d 1
=
d
2 1 2 3
6 J 2
Chapter 7
2 1 2 3
p =
2
1 2 + 2 3 + 1 3
1
2
d 2, d 3 follow directly
6.
Assume that i3 = 0 and that we are working in principal axes for each yield function in Problem
3. Letting = 2 /1 and = d2 /d1 , find expressions for () and for ().
SOLUTION:
=
a.
F + H 2 H 1
F+ H H
d 2
=
=
d 1
G + H 1 H 2
G+ H H
G+ H +H
= 2 =
1
F +H + H
M1
b. Assuming that 1 > 2 > 0:
M1
M1
d 2 F 2M1 H 1 2
H 1
F
=
=
=
d 1 G M1 + H M1 G M1 + H 1 M1
1
1
2
= ? (Because of the form, is transcendental, so must be found numerically.)
c.
d y 6 12 + 12 2
=
=
d x
9 12 6 2
=
2 1
10 2 + 8 + 12
2 +2
which is found by using the quadratic formula on .

2
2
d 1 = d 1
3 1 + 2 1 2
2 2
2 1 2 243
1
2
d.
, or
= d 2 2 1 1 + 2
2
2
d 2 = d 1
3 1 + 2 1 2
2 12 22 2 1 2
2 2
2 2 1 243
1
2
, or
Chapter 7
= d 2 2 1 1 + 2
d 2
=
d 1
2 22 12 2 1 2
2 1
1 + 2 2 2 1 + 2 2 1 2
81
1 + 2 2 2 1 1 + 2 2 2
81
= hopeless to do in closed form unless the numerator and denomerator can be factored.
=
e.
=
c 12 + c 21 2 1 2 +c 22 2 2 1 + c 23 + c 32 1 2
d 2
=
d 1
2 c 11 2 1 2 + c 12 + c 21 2 2 1 + c 13 + c 31 1 2
c 12 + c 21
2 +c 22 2 1 + c 23 + c 32
2 c 11 2 + c 12 + c 21 2 1 + c 13 + c 31
1
1
2 c 12 + c 21 c 22 c 23 c 32 + 4 c 11 +c 12 +c 21 +c 13 +c 31
c 12 + c 21 + c 23 + c 32 2 c 22 + 2 c 11 +2 c 12 +c 21 c 13 c 31
3
f.
4 A 0 1 + + A 2 + 2 A 3 + 3 A 4 2 + 4 A 5 3
d 2
=
=
3
d 1
4 A 0 1 + + 4 A 1 + 3 A2 + 2 A 3 2 + A 4 3
() is a cubic equation and may be solved numerically or in closed form.
g.
N 1+
d 2
=
=
d 1
N 1 +
N1
1 + 2r M 1
M 1
N1
+ 1 + 2r M 1
M1
() = transcendental
h.
c + 2 c 22
d 2
= 12
2 c 11 + c 12
d 1
c 12 2 c 11
c 12 2 c 22
Page 11
Page 12
d 1 =
2 1 2
2 1 + 2 1 2
2
i.
Chapter 7
2 2 1
p , and d 2 =
2 1 + 2 1 2
2 1 p 2 1 2 + 22 1 2
d 2
=
=
d 1
2 p 2 12 + 2 2 1 2
() appears to be transcendental, and thus must be found numerically.
7.
Construct a full set of useful equations expressed in principal axes for the following yield
functions. Useful equations include the yield function in terms of , the associated flow rule, the
normality equations (forward and inverse), definition of d in terms of and d (where x1 tension
is the standard state), d in terms of d1, d2 and d3, (), and (); where = 2 /1 and
= d2 /d1 (The last sets of equations were derived in Problem 6.)
.
2 2 2
a. von Mises: f = ( 1- 2 ) + ( 1 - 3) + ( 2- 3)
b. Hill quadratic:
f = F( 2- 3) 2 + G( 1 - 3) 2 + H( 1 - 2) 2
c. Hill normal anisotropic: modify function #2 such that the x1 and x2 axes are equivalent
and the strain ratio d2 /d1 in an x1 tensile test is r. (x3 is the sheet-thickness direction).
M M
d. Hill non-quadratic (Case IV), (3 = 0):: f = (1+2r) | 1- 2| + | 1 + 2|
(You may restrict your attention to one octant, where 2>1>0).
f = 1+ 2
e. Bassani 2-D (3 = 0):
+ (1+2r) k 1- 2
SOLUTION:
2
a. von Mises:
x1 - Yield:
Flow Rule:
Normality:
f = 1 2 + 1 3 + 2 3
= 1
2
1 2 + 1 3 + 2 3
d 2
d 3
d 1
2 1 2 3 = 2 2 1 3 = 2 3 1 2
d 1 = 2d 2 1 2 3
d 2 = 2d 2 2 1 3
1
2
d i
i
Chapter 7
Page 13
d 3 = 2d 2 3 1 2
Normality (inverse):
1 2 =
1 d d
1
2
6d
2 3 =
1 d d
2
3
6d
1 3 =
1 d d
1
3
6d
d based on usage shown above

d = 4
Lambda:
d = 32
Effective Strain:
d 1 d 2 + d 2 d 3 + d 1 d 3
= 2 d 1 2 + d 2 2 + d 3 2
3
Ratios
3 = 0
2 + 1
= 2 =
2+
1
x1 - Yield:
1
G +H
1
2
+ H 1 2
G 1 3 + H 1 2
d 2
F 2 3 H 1 2
d3
F 2 3 G 1 3
Normality:
d 1 = 2d G 1 3 + H 1 2
d 2 = 2d F 2 3 H 1 2
d 3 = 2d F 2 3 G 1 3
1 2 =
F 2 3 + G 1 3 + H 1 2
d 1
Flow Rule:
1
2
d 2
= 2 1
2
d 1
f = F 2 3 + G 1 3
b. Hill quadratic:
1
F d 1 G d 2
2 d
1
2 2
Page 14
2 3 =
1
G d 2 H d 3
2 d
1 3 =
1
F d 1 H d 3
2 d
Chapter 7
where = FG + FH + GH
d =
Lambda:
d
based on usage shown above
2 G +H
Effective Strain:
2
+H
d = FG +GFH
+ GH F Gd 2 Hd 3 + G Fd 1 Hd 3 + H Fd 1 Gd 2
G+H + H
= 2 =
1
F+H + H
Ratios 3=0 :
F+H H
d 2
=
d 1
G+H H
f =
2 3 + 1 3 + r 1 2
1
1 +r
c. Hill normal anisotropic:
x1 - Yield:
2 3 + 1 3
+ r 1 2
1
2
Flow Rule:
d 1
d 2
d 3
=
=
2
1 2
3
1 3 + r 1 2
2 3 r 1 2
Normality:
d 1 = d 1 3 + r 1 2
d 2 = d 2 3 r 1 2
d 3 = d 2 3 1 2
1 2 =
1
d 1 d 2
1 + 2r d
1 3 =
1
d 1 r d 3
1 + 2r d
2 3 =
1
d 2 r d 3
1 + 2r d
1
2
Chapter 7
d = (1 +dr)
Lambda:
d =
Effective Strain:
Ratios
Page 15
3 =0
1+r + 1
= 2 =
1
1+r + r
f = 1 + 2r
d. Hill non-quadratic (Case IV):
1
2
1 + r d 2 + d 2 + 2r d d
1
2
1+ r 1 2
1 + 2r
=
1 2
+ 1 + 2
1+r 1
d 2
=
d 1
1+r r
M
where 2 > 2 > 0
1
M
1
21+r
2r + 1 1 2
+ 1 + 2
x1 - Yield:
Flow Rule:
d 1
2r + 1
1 2
M1
+ 1 + 2
d 1 = d
M1
2r + 1
2 1 +r
d 2
2r + 1 1 2
1 2
M 1
M1
1
2 1+r
+ 1 + 2
1 + 2
M1
d 3
2 1 + 2
M1
Normality:
d 2 = d
2r + 1
2 1+r
1 2
d 3 = d 11
+ r 1 + 2
1 2
1
=
d 1 + 2r
1 + 2 =
Lambda:
d =
d
M 1
1
d
1
M1
M 1
1
2 1+r
M 1
1
M 1
d 1 d 2
d 1 + d 2
1
M 1
1
M1
1 + 2
M1
M1
Page 16
d = 1 1 + r
2 1 + 2r
1
M1
Chapter 7
M1
M
d 1 d 2
M
M1
+ 1 1+r
2
1
M 1
d 1 + d 2
M
M1
Effective Strain:
Ratios 3=0 :
1 + 2r 1 +
= 2 =
1
1 + 2r 1 +
1
M1
1
M 1
1
M 1
+ 1
1
M1
M1
M1
+ 1+
1 + 2r 1
d 2
=
=
M1
M1
d 1
1 + 2r 1
+ 1 +
e. Bassani yield function:
f = 1 + 2 N + k 1 2 M
where 1 > 2 > 0
Virtually nothing can be done with this yield function in closed form. In fact, it is not
N 2r + 1
k =
M
(Eq. 7.75),
dimensionally correct, because k must take the dimensionless form
and thus the two terms in the yield function have different units and thus cannot even be added.
8.
Verify that Figs. 7.21 represent the von Mises yield function when balanced biaxial tension and
pure shear are used to define the effective stress.
SOLUTION:
2
2
k = 1 2 + 1 3 + 2 3
a. Balanced biaxial tension: 1 = 2 = , 3 = 0
k = 2 + 2 = 2 2
b. Shear:
same as Eq. 7.19
1 = 2 = ,
3 = 0
k = 2 + 2 + 2 = 6 2
9.
Consider the strain paths A, B, and C shown in the figure.
Chapter 7
2
(2, 2)
C2
(2, 1)
B2
(2,0)
Find the effective strain for each path using the following yield functions:
a. von Mises
b. Hill - Orthotropic, Quadratic
(rRD = 2, rTD = 1.5)
c. Hill - Normal Quadratic
(r = 1.75)
d. Hill - Nonquadratic
(r = 1.75, M = 2.5)
e. Hosford
(r = 1.75, M = 6)
f. Bassani
(r = 1.75, N = 2, M = 4)
SOLUTION:
Path A:
Proportional from (0,0) to (2,1) = 1, 2
Path B:
B1 proportional from (0, 0) to (2, 0), plus

B2 proportional from (2, 0) to (2, 1)
Path C:
C1 proportional from (0, 0) to (2, 2), plus

C2 proportional from (2, 2) to (2, 1)
a. von Mises:
d = 23 d 1 2 + d 2 2 + d 3 2
1
2
or, for a proportional path,
= 23 1 2 + 2 2 + 3 2
A =
Path A:
2 22 + 1 2 + 3
3
1
2
1
2
= 3.06
Page 17
Page 18

B 1: 1 =
Path B:
2 22 + 02 + 2
3
Path C:
1
2
= 2.31
2 02 + 12 + 1
3
B 2: 2 =
C 1: 1 =
2 22 + 22 + 4
3
1
2
= 1.15 B = 3.46
1
2
= 4.00
2 02 + 1 2 + 1
3
C 2: 2 =
Chapter 7
1
2
= 115 C = 5.15
b. Hill orthotropic quadratic (see Eq. 7.56). For proportional paths and r RD = 2, rTD = 1.5 ,
1
2
2
= 0.157 2 1.5 2 3 3 + 1.5 2 1 3 3 + 3 2 1 1.5 2
A = 0.157 2 10.5 + 1.5 13 + 3 2.5
1
2 2
= 3.49
Path A:
2
Path B:
B 1: 1 = 0.157 2 6 + 1.5 10 + 3 4
1
2 2
= 2.58
B 2: 2 = 0.157 2 4.5 + 1.5 3 + 3 1.5
Path C:
C 1: 1 = 0.157 2 15 + 1.5 16 + 3 1
1
2 2
1
2 2
= 4.54
C 2: 2 = 0.157 2 4.5 + 1.5 3 + 3 1.5

c. Hill normal quadratic (see Eq. 7.64) for proportional paths, r = 1.75
= 1.30 1 2 + 2 2 + 1.27 1 2
A = 1.30 2 + 1 + 1.27 2 1
2
Path A:
1
2
1
2
= 1.22 B = 3.80
= 3.57
1
2 2
= 1.22 C = 5.76
Chapter 7
Path B:
1
2
B 1: 1 = 1.30 2 + 0 + 1.27 2 0
2
= 2.60
1
2
B 2: 2 = 1.30 0 + 1 + 1.27 0 1
2
Path C:
1
2
C 1: 1 = 1.30 2 2 + 2 2 + 1.27 2 2
= 1.3 B = 3.90
= 4.70
1
2
C 2: 2 = 1.30 0 + 1 + 1.27 0 1
2
Page 19
= 1.30 C = 6.00
d. Hill non-quadratic (see Eq. 7.69), r = 1.75, M = 2.5:

0.60
= 0.99 0.51 1 2
1.667
+ 1 + 2
1.667
For proportional paths:
Path A:
Path B:
A = 0.99 0.51 1
B 1: 1 = 0.99 0.51 2
1.667
1.667
+ 3 1.667
+ 2 1.667
B 2: 2 = 0.99 0.51 2
Path C:
C 1: 1 = 0.99 0.51 0
1.667
+ 4 1.667
C 2: 2 = 0.99 0.51 1
10.
1.667
1.667
0.60
= 3.11
0.60
= 2.54
+ 1 1.667
0.60
= 1.40 B = 3.81
0.60
= 3.96
+ 1
1.667
0.60
= 1.27 C = 5.23
Assume that the tensile stress-strain curve for a standard material is well-known to be
= 500 0.25 (MPa).

Consider a plane-strain tension test where x1 = RD = principal tensile axis, x2 = TD = zero strain
direction, and 3 = 0.
a. Use the yield functions in Problem 9 and find 1 as a function of 1 for this test.
b. Plot the tensile stress-strain curves and 1-1 curves obtained in Part b on the same graph.
Page 20
c. What is the stress ratio 2/1, for each plane-strain case?

SOLUTION:
, 0,
Parts a and c - Strain ratios:

a. von Mises
3 = 0, d 2 = 0 2 = 21
(see Eqs. 7.20 and 7.3-3)
1 21
= 1
2
Stress ratios: 1, 2, 0
+ 1 2 + 21
2 d 2 + 0 2 + d
1
1
3
= 500 0.25
1 = 2 500
3
2
3
1
2
3 = 500
2 1
0.25
1
2 2
= 23 1
,
= 2 d 1
3
0.25
2
1
3
1 0.25 = 598 1 0.25
b. Hill orthotropic quadratic (see Eqs. 7.54, 7.55b, and 7.56), rRD = 2, r TD = 1.5
3 = 0, d 2 = 0 rRD 2 r RD r TD 1 2 = 0
1 + r TD 2 = r TD 1
2 =
r TD
= 1.5 1 = 0.60 1
2.5
1 + rTD 1
1
2
1
1.5 3
= 0.222
2 2 + 1.5 1 + 3 1 2
2
1
2
2 0.6 + 1.5 + 3 0.4
1
2 2
1
2 2
1 = 0.774 1
= 0.157 2 0 + 3 1 + 1.5 5 1 + 3 2 1
= 0.157 18 + 37.5 + 12
1
2
1 = 1.29 1
1
2
Chapter 7
Chapter 7
So,
0.774 1 = 500 1.29 1
0.25
Page 21
or 1 = 687 1 0.25
c. Hill normal quadratic (Eqs. 7.61b, 7.63, and 7.64), r = 1.75

r = 0.64
d 2 = 0, 2 = 1 +
1
r 1
1
2
= 1 + 0.64 1.28 1 0.64
= 1.68
So,
1
2
1 + 0 + 1.28 0 1
0.76 1 = 500 1.30 1
0.25
1
2
1 = 0.76 1
1 = 1.30 1
1 = 694 1 0.25
d. Hill non-quadradic (see Eqs. 7.67, 7.68, and 7.69), r = 1.75, M = 2.5
d 2 = 0,
2r + 1
1
M1
2r + 1
2 1 +r
1 2
M 1
1
2 r+1
1 + 2
M1
1 2 = 1 + 2
k = 2.72
k 1 1 = k + 1 2
= 0.82 0.53
So,
2.5
, or
2 = k 1 1 = 0.46 1
k+1
+ 0.18 146
0.83 1 = 500 1.32 1
0.25
Part b. - The plots are shown below.
1
2.5 2.5
1 = 0.83 1
1 = 646 1 0.25
= 0.51 + 1
0.60
1 = 1.32 1
Page 22
Chapter 7
Problem 7-10 - Plane Strain Tension

500
X 1-Direction Stress (MPa)
Hill, r=1.75, M=2

Hill, r
RD
=2, r
TD
=1.5, M=2
Hill, r=1.75, M=2.5
300
100
Von Mises
0.1
0.2
0.3
True Strain
11. Repeat Problem 9 for a balanced biaxial test where 1 = 2 and 3 = 0. Also find the strain ratios
for each case.
SOLUTION:
a. von Mises (see Eqs. 7.20 and 7.3-3) isotropic, 1 = 2 , 3 = 2 1

= 1 1 2 + 12
2
1
2
= 1
2
2
2
d = 2
3 1 + 1 + 4 1
1
2
= 2 1
= 500 0.25 1 = 500 2 1
0.25
= 595 1 0.25
b. Hill orthotropic quadratic (see Eqs. 7.54, 7.55, and 7.56), rRD = 2, r TD = 1.5
1
=
1.5 1 + 2
1
2
2 1 2 + 1.5 1 2
1
2
= 0.88 1
r
r
d 2
= RD , or d 2 = RD d 1 = 43 d 1 d 3 = 73 d 1
d 1
r TD
rTD
1
2
= 0.157 2 1.5 43 1 + 3 73 1
= 0.157 162 + 121.5 + 0
1
2
+ 1.5 2 d 1 + 3 73 1
1 = 2.64 1
+ 3 2 1 1.5 43 1
Chapter 7
= 500 0.25 0.88 1 = 500 0.51 1
Page 23
0.25
, or 1 = 724 1 0.25
c. Hill normal quadratic (see Eqs. 7.63 and 7.64) r = 1.75

1 2+ 2
=
1
2.75 1
1
2
= 0.85 1
1 r
d 2
1+r = 1
=
d 1
1 r
1+r
= 2.75 1 + 1 + 3.5 1
2.75
4.5
1
2
1 = 2.35 1
= 500 0.25 0.85 1 = 500 2.35 1
0.25
0.25
, or 1 = 728 1
d. Hill non-quadratic (see Eqs. 7.67, 7.68, and 7.69), r = 1.75, M = 2.5
= 2
1
5.5
1
2.5
1 = 1.01 1
d 1
M1
= 1 M1 = 1
d 2
1
= 0.99 2 1 = 1.98 1
= 500 0.25 1.01 1 = 500 1.98 1
0.25
1 = 587 1 0.25
12.
Plot the 1st quadrant of the yield functions in Problem 9.
SOLUTION:
The plot below is in terms of normalized stresses, where the tensile flow stress in the x1 direction
is taken to be unity.
Page 24
1.4
Chapter 7
Hill Ortho. Quad.
Hill Non-Quad.
Normalized X
2 Stress
1.2
1
Von Mises
0.8
0.6
Hill Norm. Quad.
0.4
0.2
00
13.
0.2
0.4 0.6 0.8

1 1.2
Normalized X 1 Stress
1.4
By plotting, show how r affects Hill's normal quadratic yield function. Take values of r = 1/2, r =
1 (von Mises), r = 2, r = 4 for illustration.
SOLUTION:
2
Normalized X
2 Stress
r=4.0
1.5
r=2.0
r=1.0
1
r=0.5
0.5
0
0
14.
0.5
1
Normalized X
1.5
Stress
1
By plotting, show how M affects Hill's normal non-quadratic yield function. For r = 1, take values
of M = 1.5, M = 2 (von Mises), M = 4, and M = 10.
SOLUTION:
Chapter 7
Page 25
1.4
M=1.5
Normalized X
2 Stress
1.2
1
M=2.0
0.8
M=4.0
0.6
M=10.0
0.4
0.2
0
0
15.
0.2
0.4 0.6 0.8

1
1.2
1.4
Compare the form of Hill's normal non-quadratic yield function and Hosford's yield function by
assuming that r = 2 and finding M in each case such that 1 (balanced biaxial tension) is equal to
1.2 B (uniaxial tension).
SOLUTION:
For r = 2, 1 (balanced biaxial tension) = 1.1

Hill normal non-quadratic (Eq. 7.67)
=
1 2M
6
1
M
1 , therefore 1 = 16
2 1.1 , or M = 2.272 (Hill)
Hosford (Eq. 7.70)

= 13
2 1 , therefore 1 = 16
2 1.1, or M = 1.39 (Hosford)
The corresponding yield functions are thus

Hill:
Hosford:
2.27 = 5 1 2
6
4.25 = 13
4.25
2.27
+ 1 1 + 2
6
+ 2
4.25
2.27
+ 2 1 2
4.25
The following plot of the first quadrant shows the differences of the two representations for the
same r value and balanced-biaxial-to-tensile yield stress ratio.
Page 26
Chapter 7
1.4
Normalized X
2 Stress
1.2
1
0.8
Hill Nonquadratic
Hosford
0.6
0.4
(Avg. Difference = 0.6%)

0.2
0
0.2
0.4 0.6 0.8

1
1.2
1.4
B.
DEPTH PROBLEMS
16.
It has been proposed that the direction of the plastic strain increment, d , is not always normal to
the yield surface. Instead, the proposer suggests that d is intermediate in direction between
d (n), the normal direction, and d , the stress increment direction [i.e.
(n)
d = d + (1-) d, where = 0 < < 1 ]. Criticize this model in terms of stability
arguments and your knowledge of real materials.
SOLUTION:
} Possible
d
(n )
The situation is as shown in the figure, where the choice of , which might be a material
property.
Chapter 7
Page 27
Clearly, the choice of d for all , between 0 and 1 satisfies second-order stability since d lies
(n)
closer in direction to d than does d . In general, the second-order work will be higher, for a
given stress or strain path, which will be exhibited as higher work hardening.
The implication of the new direction of d allows first order stability with some possible
relaxation of convexity, depending on the choice of .
17.
It is convenient to introduce factors which can be used to multiply the effective stress or strain to
obtain the stress or strain in a given state. For example:
1 BB = F (BB) or
1 BB = F (BB)
might be used to find the in-plane strains and stresses in balanced biaxial tension (BB) for a given
yield function at a given hardness of , .
a. Show that F and F are constants with respect to strain for a given yield function when
isotropic hardening is obeyed
b. Find the specific values of F(PS) and F(BB) in terms of r which relate plane-strain tension
to uniaxial tension using Hill's normal quadratic yield criterion.
c. Repeat part b for Hill's normal nonquadratic yield function, finding F and F in terms of r
and M.
SOLUTION:
The form of a yield function for an isotropic hardening material does not change as straining or
hardening proceeds.
a. For a specified stress ratio () for a two dimensional state of stress,

= 1 , 2 = 1 , 1 = f 1 , or
1
= F
where f() and F () are reciprocals fixed by the form of the yield function alone. (Note that
for uniaxial tension, F and f are identically unity because = 1 . A similar argument
follows for F by noting that the plastic work is given by
d = f 1 F d 1 = f F 1 d 1 ,
where f and F are constant functions, i.e. they do not depend on hardening.
Page 28
Chapter 7
b. For Hill's normal quadratic yield in two dimensions 3 = 0 , we use Eqs. 7.63 and 7.64. For
plane-strain (x2 is the direction of zero extension):
0 = d 2 =
d
r , so
2 1 +
r 1
2
r , = , = r
= 1+
2
1
2
r
1+ r 1
1
(PS)
Then we can find F from Eq. 7.63,
2 = 1 2 + 1 +r r
=
1 + 2r ,
1
1 +r
And, similarly for F

1+r
d = 1 + 2r
2
r
2
1 2 12r
+ r 1 + r 1
F (PS) =
1+r
1 + 2r
(PS)
from Eq. 7.64:
d 1 + 0 + 12r
+ r d 1 0
2
1 + r d , F (PS) = 1 + 2r
1
1 +r
1 + 2r
The procedure is identical for balanced biaxial tension, where = 1 and d 1 = d 2 (from Eq.
7.63a and b):
d =
= 1 + 1 12r
+ r 1 1 1
d =
1 + r 2 + 2r
1+r
1 + 2r
1
2
d 1
d =
(BB)
2
=
1 + r 1 , F
1+r
2
2 1 + r d 1, F(BB) =
1
2 1+r
c. The solution fowllows Part b with the use of Eqs. 7.67-7.69.

For plane-strain tension:
2r + 1
d 2 = 0 = 2 =
1
2r + 1
1
M1
1
M1
1
+1
= 1 , where: 2r + 1
+ 1
1
M1
Chapter 7
F (PS) =
1 + 1 + r
2
1
M 1
For balanced biaxial tension

2 1 +r
F (PS) =
1
M
M1
M
2
1 + 1 + r
F (PS) =
,
1 = 2, d 1 = d 2
(PS)
Page 29
M1
M
1
M1
1
=
2 1+ r
1
M
d. Inspection shows that in plane strain tension:
F (PS) =
1
F (PS)
This result follows the definition of effective strain from the principle of equivalent plastic work:
d = 1 d 1 + 2 d 2 + 3 d 3.
For plane-strain tension, 3 = 0, d 2 = 0, so,
d = 1 d 1 = F F d , so F F 1 .
18.
In view of results from Problem 17, what can you say about the complexity of yield function which
would be required to account for the following observations? (Assume only normal anisotropy.)
a. In uniaxial tension,
1 = 500 10.25
In plane-strain tension,
1 = 600 10.25
b. Same as part a, but r = 2 from the tensile test.
c. In uniaxial tension,
In plane-strain tension,
r = 2.
1 = 500 10.25
1 = 600 10.35
SOLUTION:
a. The difference in strength coefficient can be accounted for by a single value of r, so Hill's normal
quadratic theory is sufficient.
b. If r is known independently, then we need a second adjustable parameter, M, to fit the data.
c. None of the standard, isotropic-hardening theories can account for different hardening rates.
See, however, R. H. Wagoner: Metall. Trans. A., 1980, vol. 11A, pp. 165-175 for a modification
of standard theories to account for such behavior.
19.
Various authors have attempted to introduce work hardening parameters, particularly to compare
Page 30
Chapter 7
various strain states or to compare material hardness when true strain is unknown. Here are
three such quantities:
d
d
d
d ln
d ln
=
=
d
d
d
d
d ln
Based on your knowledge of isotropic hardening and results from Problems 17 and 18, which of
these do you think is most suitable for comparing hardening in various strain states?
SOLUTION:
Remember that effective quantities, and , depend on the choice of yield function. For a given
set of 1 1 data (from a plane-strain test, for example), the corresponding - will be
obtained after assumption of a yield function, and the corresponding quantities F and F:
F d 1
d
=
d
F d 1
d
F d 1
d ln
=
=
d
d
d 1
the
d
remains F
d
d 1
d ln 1
d ln
=
= 1
=
d
d
1 d 1
d ln 1
Therefore, the third measure of work hardening is invariant to the choice of yield function,
contrary to the other two.
See R. H. Wagoner: Metall. Trans. A, 1981, vol. 12A, pp. 2142-2145 for a more thorough
discussion of work-hardening measures.

A.
1.
In an effort to refine the upper bound estimate to density, , that a vertical embankment of height h
can sustain, you propose a trial displacement field that is different than shown in Exercise 8.1.
The field is shown below, where the boundary between the slipping and stationary parts of the
embankment is a quarter circle of radius R, from = 0 to /2. Calculate the new upper bound to
the density.
SOLUTION:
Take d, the relative displacement across the boundary, to equal Rd, where d is the
incremental rotation of the quarter circle of material about the center. The internal work per unit
depth of embankment is given by
2R
IW =
kRd .
4
The external work is given by considering a piece of material with differential area, r dr d, on
which a downward gravitational force, fg = gr dr d, acts per unit depth of embankment. The
work increment per unit area of material is then fg r cos d, where the factor cos is required to
project fg onto the same direction as d. The external work is then
R / 2
EW = gd
gR 3d
r cosdrd = 3 .
2
0 0
Equating IW to EW produces the following upper bound,

3k
.
2gR
The minimum upper bound on is given by making R = h, the largest value possible. For this
assumed displacement field,
3k
min =
.
2gh
The "quarter circle" deformation mode produces a better upper bound than the linear shearing off
mechanism presented in Exercise 8.1.
2.
Consider the proposed deformation field shown in Fig. 8.3(a). Discuss why for general 1, 2 the
stress state at the lower tip of triangular block d can not be an equilibrium stress state.
Page 2
Chapter 8
SOLUTION:
Below is a sketch of the lower tip of triangular block d. One can not construct a Mohr's circle
for the stress state at this point. In particular, if you rotate couterclockwise by 180-1-2, the
shear stress must change from a maximum of k on the plane 1 to -k on plane 2. This is not
possible for general 1, 2.
180-1- 2
3.
Construct a lower bound for the indentation load, P, per unit depth, based on a three-sector stress
field. Vary the angle to obtain the best lower bound.
P,
SOLUTION:
The optimal lower bound is given by setting = 90, so that the Mohr's circle construction for
the two sectors, A and B, is two circles of radius k as shown below.
r
B
-4k
A
-2k
rr,
The best lower bound here is P = 4kw.

4.
Using the yield condition, Eq. 8.13, show why hydrostatic loading, 11 = 22 = 33 = , cannot
cause yield on any slip system, regardless of the orientation of the slip system to the loading axes.
Chapter 8
Page 3
SOLUTION:
If the stress state, 11 = 22 = 33 = is substituted into the yield condition, Eq.. 8.13, then
()
() ()
s1() m1() + s()
()
2 m 2 + s3 m3
c
or equivalently,
s() m () ()
c
For slip planes, s and m must be perpendicular to one another, so that sm = 0 for all slip
systems. This statement and the rephrased yield condition above predict that yield can never be
reached under hydrostatic loading.
5.
Experiments which apply hydrostatic loading to engineering materials have shown that materials
such as steel, aluminum, copper, or silicon will yield, although the magnitude of hydrostatic
loading to cause yield is many times that required for simple tension or compression. Discuss why
materials do yield in hydrostatic loading.
SOLUTION:
Typically, engineering grades of materials contain grain boundaries, inclusions, and voids. All
serve to act as stress concentrators, so that under nominal hydrostatic loading, the stress state is
not homogeneous and entirely hydrostatic. Yield can then occur.
6.
A single crystal is indented with a square knife edge as shown below. The candidate slip systems
are oriented at discrete angles 1, 2, and 3.
a. Construct a deformation field out of rigid, sliding triangles, produce the corresponding
hodograph, and determine an upper bound to the indentation load.
b. If the crystal had only two slip planes, could you construct a deformation field? Support
your answer with some sketches. What would the indentation load P be under such a case?
2 =60o
m
3
=0o
m
s
SOLUTION:
a. The deformation field and the corresponding hodograph are:
1=30o
Page 4
Chapter 8
d'
1.5v
d
v/2 3
v/2
c
P, v
c'
3v
2v
2v/ 3
w
C B C' D'
v/ 3
A
w
D
v/ 3
a, b
The external dissipation rate is Pv. The internal dissipation rate is

k[lBC'VBC' + lC'OVC'O + lC'DVC'D + lD'OVD'O +lBCVBC + lCOVCO + lCDVCD + lDOVDO].
Substituting the values for lIJ derived from the deformation diagram and values for vIJ derived
from the hodograph, the internal dissipation rate is approximately 6.93 kwv. Equating the
external and internal dissipation rates gives: P = 6.93 kw.
b. It is not possible to construct a deformation field to this geometry with only two slip directions,
provided that the sample is infinitely thick. Triangular blocks must be constructed for a
successful deformation field, and the three sides of the block will require three slip directions.
7.
In the upper bound analysis of a block of dimensions l1, l2, and l3, which is sheared by an amount
b along a direction s, on a plane with normal m (see Fig. 8.2), the internal dissipation is
IW* = k Ab
and the external dissipation is
EW* = 11* Abm1s1 + 22* Abm2s2
where 11* and 22* are uniform stresses, and A is the area of slipped plane.
a. Sketch the resulting upper bound load to the collapse surfaces in 11*-22* space,
assuming the inclination angle, , shown in Fig. 8.2 equals 30.
b. Although the internal strain is concentrated at the slip plane, define an appropriate
average, or macroscopic strain state for the block, after a slip of b has occurred on the
plane.
c. Do your macroscopic strains in Part b satisfy normality with the collapse surface in Part
a? Explain why. If they do not propose a definition of macroscopic strain that will satisfy
normality.
SOLUTION:
a. We note that k = 11m1s1 + 22*m2s2, where for this problem, m1 = cos 60, s1 = +cos 30, m2 =
cos 30, and s2 = -cos 60. When values of mi and si are substituted into the expression,
Chapter 8
Page 5
4
k
3
where the sign on the RHS is produced by considering positive or negative values of slip on
the plane. The result is shown below.
* ( * )
22 22
(11 - 22*) =
direction of
strain increment
2.3k
*
*
2.3k 11 ( 11)
-2.3k
direction of
strain increment
-2.3k
b. Using Eq. 8.23,
1 Ab
d- ij = 2 l l l (mjsi + misj)
123
so that using the particular values for the mi, si,
3 Ab
d- 11 = 4 l l l
123
3 Ab
d- 22 = m 4 l l l
123
c. The macroscopic strains defined in part (b) do satisfy normality. The strain increments may be
plotted in the 11*-22* diagram in (a), where the 11* axis is constructed parallel to the 11*
axis, and the 22* axis is constructed parallel to the 22* axis. The incremental strains are then at
+135, -45 angles measured counterclockwise from the horizontal axis. They are perpendicular
to the yield surface.
8.
Construct the projection of the f.c.c. yield surface on to the 11-12 stress plane. Compare your
answer to the prediction based on a Tresca yield criterion, 1 3 = 2k, where 1 and 3 are the
maximum and minimum principal stresses, respectively.
SOLUTION:
We note Eq. 8.17 and pick 1 = 11 and 6 = 12 (=21) to be the only non-zero components of
stress. The resulting yield condition becomes
Page 6
Chapter 8
11

11 + 12

12
+
11
12

12

1 11 k
6 11 + 12
12

11

11
11 + 12
12
Here, the assumption is made that the critical resolved shear stress for each plane is k. The
corresponding yield surface is constructed from lines generated from the twelve slip systems
represented in the above equation.
The corresponding Tresca yield condition is generated from |1 - 3| = 2k. This condition states
that the diameter, D, of Mohr's circle, constructed in 11-12 space, can not exceed 2k. The
diameter of Mohr's circle for this stress state is
1
D = 2122 + 2 112 ( = 2k)
Accordingly, the above is an equation of an ellipse for which the 11 intercept is 2k and the 12
intercept is k. It is shown in the figure below.
*
22
6k
k
2k
- 6k
*
6 k 11
- 6k
9.
Imagine that you can load a f.c.c. crystal along any crystallographic direction. Assume that the
potential slip systems are of the type {111}/<110>, as listed in Table 8.1. Find a crystallographic
direction along which the tensile stress to yield is a minimum, and report the minimum value of
tensile stress in terms of c, the critical resolved shear stress to activate slip.
Chapter 8
Page 7
SOLUTION:
Use the yield condition, Eq. 8.13, expressed in the form,
,
() = (s () e (1' ) )(m() e (1' ) ) ()
c
where s() and m() are the slip direction and slip plane normal of slip system (), and e(1') is the
direction along which the tensile stress, , is applied. The minimum is found by making the
factor involving the dot products as large as possible. The largest dot products are obtained by
first choosing e(1') to lie in the plane containing s() and m(). Since s() and m() are
orthogonal to each other, the relative orientations of vectors is depicted below, where s() e(1')
= sin and m() e(1') = cos. Therefore,
c
=
,
sin cos
and the minimum value, = 2c, occurs when = 45, 135, 225, or 315. There are several
possible crystallographic directions of the type e(1')= (s()m())/ 2 . A possible direction for
the {111}/<110> slip systems considered here are [ 2 + 3 2 - 3 2 ]/2 3 .
10.
A biaxial stress state is applied to a single crystal of f.c.c. material. However, the crystal is
oriented so that one applied stress, 1'1', is along the [111] crystallographic direction, and the
other applied stress, 2'2', is along the [1-1 0] crystallographic direction. Assume that all slip
planes of the type {111}/<110> (see Table 8.1) slip when the critical resolved shear stress reaches
c. Construct a projection of the yield surface onto the 1'1'-2'2' stress plane. Comment on the
difference between this projection and the one shown in Fig. 8.4.
SOLUTION:
A relatively simple approach is to revert to the yield condition, Eq. 8.13, and compute directly
the resolved shear stress on each of the twelve slip systems according to
,
() = (s () e (1' ) )(m() e (1' ) )1'1' + (s() e(2' ) )(m () e(2' ) )2' 2' ()
c
where e(1') and e(2') corresponds to the crystallographic directions [111]/ 3 and [1-1 0]/ 2 , and
the vectors s() and m() correspond to the slip directions and slip plane normals, respectively,
of slip system (). The dot products involved project the components of stress, applied along the
1' and 2' directions, on to the s() and m() directions. The twelve slip systems listed in Table
8.1 are used in the yield condition and produce twelve equations of the form,
2' 2'
c (slip systems 7,5)
6
21' 1'
c (slip systems 6,9,11,12) .
3 6
21' 1' 2' 2'
+
c (slip systems 4,8)
3 6
6
The result indicates that two or more slip systems can produce the same yield condition. Slip
systems 1, 2, 3, and 10 have zero resolved shear stress on them, and hence, will never yield.
The yield surface is depicted by the dotted line in the figure below. For comparison, the
Page 8
Chapter 8
projection of the yield surface on to the 11-22 plane is reproduced from Fig. 8.4 and is
indicated by the solid line. The differences in yield surfaces demonstrates that yield behavior
depends on the choice of loading axes relative to the crystal basis.
22 or 2'2'
(3, 5, 8, 12)
6c
(4, 8)
1.5 6c
(2, 4, 7, 11)
6
c
(2, 4, 7, 11)
(7, 5)
(6, 9, 11, 12)

1.5 6c
(1, 6, 9, 10)
(6, 9, 11, 12)
(7, 5)
6c
11 or 1'1'
(1, 6, 9, 10)
(4, 8)
6 c
(3, 5, 8, 12)
A more involved approach is to use Eq. 8.17. The components, i, of stress in that equation refer
to the cube basis of the f.c.c. crystal. However, the applied stress is expressed in terms of
components along the [111] and [1-1 0] directions. Thus, Eq. 8.17 may be used only if the
applied stress state, 1'1' and 2'2', is expressed in terms of the components in the crystal basis.
Let the basis e(1), e(2), and e(3) correspond to the crystallographic directions, [100], [010], and
[001], and the basis e(1') and e(2') corresponds to the crystallographic directions [111]/ 3 and [11 0]/ 2 , as before. Then
ij = A ik A jl kl .
where Aik = e(i) e(k'), ij corresponds to components in the e(i) basis, and 'ij corresponds to
components in the e(i') basis. Accordingly,
1 / 3 1 / 2 *
[Aik' ]= 1 / 3 1/ 2 * ,
1 / 3
0
*
1
1 11
1
2 22
3 33
2' 2' 0
1' 1'
+
=
.
and the computed components of stress are
4
23
3
2 0
0

5
13

1
6 12
Substitution into Eq. 8.17 produces the same three types of yield conditions derived from the
first approach.
11.
A sphere with radius r = 1cm is a single f.c.c. crystal. Slip systems 1, 6, and 12 listed in Table 8.1
Chapter 8
Page 9
are activated, so that a slip of 1m in the corresponding +s directions listed occurs on slip planes
which pass right through the center of the sphere. Calculate the resulting macroscopic strain in
the crystal.
SOLUTION:
The resulting macroscopic strain is given by
()
1 A ()
() ()
() ()
s b
ij =
m i s j + m j si
V
2
where the sum is over slip systems 1, 6, and 12. The slip area As = r2 and the amount of slip
b() = 1m for the three slip systems. Also, V = 4r3/3, where r = 1 cm. Consequently, the
equation for average strain above yields contributions for each of the three active slip systems,
(1)
(6)
(12 )
3
ij =
10 4 m jsi + m is j
+ m jsi + m is j
+ m jsi + m is j
8 6
) (
Substituting values of mj and si for each of the slip systems,
12.
0 1 3
3
4
1 2 0
ij =
10
8 6
3 0 2
[ ]
Consider the portion of the yield surface in Fig. 8.4 that is contributed by slip systems 2, 4, 7, and
11. Using Eq. 8.25, show that operation of any individual or linear combination of these slip
systems produces a strain increment which is perpendicular to that portion of the yield surface
shown in the figure.
SOLUTION:
()
d ()
m is j + m jsi
2
The relevant equation is ij =
()
=
Applying this to slip systems 2 and 11, we find d11
and applying this to slip systems 4 and 7,
()
d11
d ( )
, d (22 ) = 0
6
d ( )
( )
=
, d22
=0
6
Since each individual system produces zero d22, any linear combination of operation of these
slip systems will do the same. When plotted in 11-22 space, the strain increment is
perpendicular to the yield surface lines contributed by slip systems 2, 4, 7, and 11 in Fig. 8.4.
Note that components of strain other than 11 may be non-zero here, but that the projected strain
increment in 11-22 space is still normal to the yield surface in 11-22 space shown.
13.
Equation 8.24 states that external and internal plastic work increments are equal. Show that this
( cd )(), where the components [ ] = [ ,
equation may be converted to the notation, d =
i
2, 3, 4, 5, 6] = [11, 22, 33, 23, 13, 12], and the components [di] are defined by the
identity, Eq. 8.31.
SOLUTION:
Page 10
Chapter 8
()
Equating internal and external plastic work increments produces ij dij = ()
c d
()
or if the contracted notation i, di is used, then the following must hold for each slip system ()
()
.
id ()
= ()
i
c d
Equilibrium requires that the stress state produces a resolved shear stress on each active slip
system such that
()
()
+ 2 (m 2s2 )() + 3(m 3s 3 )()
c = 1 (m1s1 )
()
+ 4 (m 2s3 + m 3s 2 )
()
+ 5 (m1s3 + m 3s1 )
+6 (m1s 2 + m 2s1 )()

When this is used to replace c() in Eq. (*) above, then there equivalence of work only if
d1( ) = (m1s1)( ) d ( )
( )
d (2 ) = (m 2s2 )
d ( )
d (3 ) = (m 3s3)( ) d ( )
d (4 ) = (m 2s3 + m3s2 )( ) d ( )
( )
d (5 ) = (m1s 3 + m 3s1 )
d ( )
()
d (6 ) = (m1s 2 + m 2s1) d ( )
Eq. 8.31 is shown using the correspondence above and invoking that
()
()
() d
.
d ij =
m js i + m is j
2
14.
Use Eq. 8.25 to show that no combination of slip systems can produce dilatation.
SOLUTION:
()
d ()
m js i + m is j
.
2
Eq.. 8.25 states that d ij =
Since d11 + d22 + d33 is the incremental volume change per unit volume, the above relation is
used to find the volume change associated with activated slip systems,
()
d11 + d 22 + d 33 = d ( ) (m1s1 + m 2s2 + m 3s3 )
= d
()
() ()
Since ms = 0 for each slip system , the above relation indicates that d11 + d22 + d33 = 0.
15. Of the twelve slip systems listed in Table 8.1, choose slip systems 1, 2, and 6. Determine whether
Chapter 8
Page 11
these 3 slip systems are independent.

SOLUTION:
Using Eq. 8.23, the respective strain increments contributed by slip systems 1, 2, and 6 in Table
8.1 are
2 0 1
2 1 0
2 0 1
d (1)
d (1)
d (1)
(1)
(2)
(6)
0 2 1 , ij =
1 0 1 , ij =
0 2 1
ij =
2 6
2 6
2 6
1 1 0
0 1 2
1 1 0
.
or equivalently, the short-hand notation, di()=d()ni(), from Eq. 8.22 may be used, where the
are
determined
according
to
Eq.
8.16,
ni()
1
1
1

1
0
1
1 0
1 1
1 0
(2)
(6)
n
n(1)
=
=
=
n
.
i
i
i
6 1
6 1
6 1
1
0
1

0
1
0
The systems are linearly independent if no values of x and y can satisfy xij(1) + yij(2) = ij(6),
for all components ij, or equivalently, if no values of x and y can be found that satisfy xni(1) +
yni(2) = ni(6) for all components i. After some examination, x = -1 and y = 0 appear to satisfy the
above relations for all components except d13 (or equivalently, d5). Accordingly, these three
slip systems are independent, since no combination of x and y can be found to satisfy the above
relations.
[ ]
16.
[ ]
[ ]
Suppose you load in tension along the [3 8 6] direction in a f.c.c. material in which the candidate
slip systems are of the type listed in Table 8.1. The crystal begins to yield at a tensile stress of
10MPa. Assume that all slip systems require the same critical resolved shear stress, c, for
activation. What slip system will be activated first? What is the value of c? To what
crystallographic direction will the tensile axis rotate? What is the axis about which the tensile
axis rotates?
SOLUTION:
The yield condition, Eq. 8.13, can be used to determine that slip system 6 in Table 8.1 will
require the smallest tensile stress to activate it. The yield condition for that slip system is stated
as
[110] [386] [111] [386]
10MPa = c .
2
3
109
109
so that c to yield is approximately 4.5MPa. As indicated in Eq. 8.37, the tensile axis will rotate
toward the slip direction, [110]. The direction, r, about which the rotation occurs is given by
Ts, so that r is parallel to [-6 6 -5].
Page 12
=
17. Derive the results,
Chapter 8
b
sin 2
Ho sin o
sin
cos
by analyzing the 2D slip plane geometry below, where band TT are the increment in slip on the
plane and average strain increment parallel to the current tensile axis, respectively. The other
parameters are labeled in the sketch. Do not use the general formulation associated with the
discussion of Fig. 8.5(a,b), but rather, derive the expression from a trigonometric analysis of the
2-D geometry below.
T
T
= TT
180-1- 2
Ho
s
k
SOLUTION:
The geometry below is used, and application of the law of sines produces
T
To
s
180- o
H
H
= o
sin(180 o ) sin
H
b
= o
sin(o ) sin
Ho
Noting that sin(o-) = sinocos - cossin and differentiating the second relation,
Chapter 8
Page 13
H sin
b= o 2 o
sin
Differentiate the first relation and then divide each side by the originating terms in the first
relation. Substitute H/ H = TT to get
sin
TT .
=
cos
18.
Eq. 8.56 defines the increment, dij*, in components of strain that are expressed in the reference
coordinate system. Derive this relation, beginning with a statement of the Principle of Virtual
Work that ij*dij* = ()d() .
SOLUTION:
We begin with a statement of the Principle of Virtual Work,
( )
=
and note from Eq. 8.48 that since ij* = ji*, then
( ) d ( ) = *ijd *ij
1 * *( ) *( ) *( ) *( )
s
m j + sj mi
2 ij i
).
If the expression for () is substituted in the statement of the Principle of Virtual Work, then
dij* must be defined according to Eq. 8.56.
19.
Suppose you have a similar geometry as in Problem 11, but instead, there are two competing slip
systems, (m(1), s(1)) and (m(2), s(2)). o is the initial angle between s(1) and the initial tensile axis
direction, To, and o is the corresponding initial angle for slip system (2).
Suppose you have documented the angle as a function of macroscopic strain, TT, along the T
(not To!) axis. Explain how you could determine the relative amount of slip, b(1) and b(2), on each
system. Do you have enough information?
SOLUTION:
First, we note the assumption that all m() and s() for both systems are coplanar with T and use
the geometry below.
Page 14
To
s(2)
s (1)
Ho
s (1)
First, note that
Chapter 8
H= b(1) cos + b(2) cos

=
The second condition is holds since the lattice rotates rigidly, relative to the T axis. Also note
that
(1)
(2)
H cos( o ) = Ho + b coso + b cos o
H sin( o ) = b(1) sino + b(2) sin o

Each of these relations may be divided through by Ho, any occurrence of H/Ho in these relations
may be replaced by exp(- TT),. The two equations relate the two unknowns, b(1)/Ho and b(2)/Ho,
to the current tensile strain, - TT, the starting angles o and o, and the current angle . Final
expressions are left to the reader to derive.
20.
Consider a material in which the critical resolved shear stress to shear a plane with normal m()
depends on the stress normal to the plane, according to
(c ) = o f m (i ) ijm (j )
where o is the critical resolved shear with zero normal load, and f is a coefficient of friction.
Such behavior is consistent with a hard sphere atomic picture of materials, in which the activation
barrier to slide particles past one another increases with confining pressure on the material.
Show that when a macroscopic stress with non-zero components 11 and 22 is applied, the yield
condition for a slip system () with slip direction s() and slip plane normal m() is
()
()
() ()
(s1() + f m()
1 )m 1 11 + (s 2 + f m 2 )m 2 22 o
where must be interpreted as either + for both terms (+s1() and +s2()) or - for both terms (s1() and -s2()).
Chapter 8
Page 15
Further, produce a projection of the yield surface onto the 11-22 plane, assuming that 11 and
22 are applied along the cube directions of a f.c.c. single crystal, and that f = 0.1. Compare
your result to the yield surface projection in Fig. 8.4.
SOLUTION:
Under such a case, the yield condition, Eq. 8.13, may be rewritten as
s(i ) ij m(j ) o + f m (i ) ijm (j ) ,
and this may be rearranged into the final form,

()
()
() ()
s1() + f m()
1 m 1 11 + s 2 + f m 2 m 2 22 o
The yield condition above consists of lines in the 11-22 plane, with intercepts 11o and 22o
along the axes given by
o
o
o( )
o( )
11 =
,
=
.
22
s(1 ) + f m1( ) m(1 )
s(2 ) + f m (2 ) m (2)
The yield surface is plotted below with dashed lines, and the thicker dashed lines show the inner
locus. For comparison, the result from Fig. 8.4 corresponding to f = 0 is shown with solid lines.
The case for non-zero f is distorted, so that yield in compression is different than yield in
tension. This occurs since tension lowers the critical resolved shear stress for slip, while
compression increases that resolved shear stress.
22 6 o
18
2,4,7,1
2,4,7
, 11
-1.5
(1,
6,
0)
9, 1
3,5,8,12
0.75
-1
18
11 6 o
1
8,12
3,5,
-1
(1,
6,
9, 1
0)
-1.5
21.
Consider the previous example where the critical resolved shear stress for a slip plane depends on
Page 16
Chapter 8
the stress normal to the plane. Construct a yield function for this case that reduces to Eq. 8.26
when the coefficient, f = 0. Find the normal to the yield surface. Compare your result to dij()
predicted by Eq. 8.23. Is the strain increment normal to the yield surface?
SOLUTION:
An appropriate yield function is given by
2
2
1
( ) = (s i m j + s jm i )( ) ij o f m (i ) ij m(j ) [= 0 at yield].
2
Taking the derivative with respect to ij and applying that () = 0,

( )
( )
() ()
( )
= c (sim j + s jm i ) + f m i m j
.
ij
For comparison, Eq. 8.23 states that

d
d ij =
(m jsi + m isj ) .
2
It appears that the strain increment is different than the normal to the yield surface, in that the
yield surface normal contains the added component, fmi()mj(), where () refers to the
critically stressed slip system on that portion of the yield surface.
22.
Assume that you have a polycrystalline wire with a bamboo type structure, in which each grain
occupies a slice of the wire as shown below. Also, each grain has a single slip system described
by slip plane normal m() and slip direction s(). The critical resolved shear stress to operate any
slip system is 10MPa, and the amount of slip on any plane is assumed to be negligible compared to
the diameter, D, of the wire. A single angle () = Ts() describes the orientation of the slip
system relative to the wire axis T, along which there is an applied stress, . Initially, the
probability, P, of finding a grain with angle () = is the same, regardless of the value of ,
which ranges form 0 to 90.
Sketch the probability of finding a grain with angle ()= , as a function of , after loading the
wire to four different values of tension: = 0, 20, 30, and 40 MPa. To obtain your result, use a
lower bound approach and make the approximation that the stress state in each grain is one of
simple tension, . What aspect of compatibility is not satisfied here?
Chapter 8
Page 17
SOLUTION:
The assumption that the stress state in each crystal is simple tension violates compatibility across
the grain boundaries in the bamboo structure. To begin, we sketch that the probability P of
finding a grain with a certain angle () is uniform, and since may span from 0 to 90, the
initial probability must be 1/90 for all possible angles. This is sketched below and labeled = 0,
20 MPa.
The analysis for other levels of stress is done by imposing that each grain that meets or exceeds
the yield condition will continue to strain, and therefore rotate until the yield condition is
exceeded. When the tensile stress is 20MPa, then we check for any grain orientation angle for
which the resolved shear stress exceeds 10MPa, i.e., for which
= 20MPa cos sin c (= 10MPa )
The yield condition above is just met for any grain at angle 45. As soon as rotation occurs, so
that = 45-, the rotation stops, for the yield condition is no longer met. Therefore, at a
tensile stress of 20MPa, grains with orientation angle = 45 just meet the yield criterion. The
rotation is negligible at this point, assuming that c remains at 10MPa even after slip is activated.
At a tensile stress of 30MPa, the yield condition is stated as: 30 MPa cos sin 10 MPa
Therefore, yield is exceeded for all grains with in between 21 and 69. All such grains
continue to rotate until = 21. Thus, in a bamboo structure with 90 grains (and equally
distributed with angle), 48 of those grains would displace to the 21 position, for a total
probability of 49 grains out of 90 having an angle = 21, zero probability for in between 21
and 69, and P = 1/90 for all other angles.
At a tensile stress of 40MPa, the yield condition reveals that all grains with initial angles
between 15 and 75 must rotate to the 15 position. Therefore, the probability is P = 61/90 at
= 15, zero for all angles between 15 and 75, and P = 1/90 for all other angles. The results are
summarized in the figure below.
Page 18
Chapter 8
P = Probability of finding a grain with angle +/- 0.5o

60/90
= 40MPa
47/90
= 30MPa
= 0, 20MPa
1/90
0
23.
15 21 30
60
69 75
90
orient. angle
Outline the components of a computer program which will calculate the uniaxial stress-strain
response of a f.c.c. single crystal loaded in tension. The tensile axis T is a material one, similar to
the case depicted in Fig. 8.5, and initially, T is parallel to [1 12 11]. The sample is loaded from 0
to 40MPa over a period of 300s, so that the stress rate, dTT/dt is constant. Use a rate-dependent
constitutive relation as described in Eqs. 8.73 and 8.74 with the following parameters:
o( ) = 103 / s
m = 20
o( ) = 10MPa (initially, prior to any slip)

0MPa if =
Case A: h =
5MPa if
5MPa if =
Case B: h =
0MPa if
Note that Case A simulates a larger latent hardening situation, and Case B simulates a larger self
hardening situation.
Produce for each case the tensile stress-tensile strain curves along T, from TT= 0 to
approximately 0.8. Note the direction of T in each case when = 0.8. For each case, discuss
the dominant slip system(s) when = 0 and when = 0.8.
SOLUTION:
The generic outline of the computer program is to
define the time increment, t, to be sufficiently small, e.g., t=0.01s
define the tensile loading rate, dTT/dt = 40MPa/300s.
initialize o() = 10MPa.
define the unit vector T along the tensile axis to be parallel to [1 12 11]. In general, the
components of all quantities will be referred to the crystal basis rather than the reference
(*) basis containing T.
define all twelve sets of mi() and si() according to Table 8.1.
Chapter 8
Page 19
define the components h of the 12 by 12 matrix according to either of the two cases
mentioned. For case A, h = 5MPa if , and h = 0 if = . For case B, h = 0
if , and h = 5MPa if = .
begin an outer loop in which the time is incremented by t until the strain along the
tensile axis exceeds 0.80. In this loop...
-increment TT by t(dTT/dt).
-compute all () according to Eq. 8.13, or equivalently,
( ) = TT s( ) T m( ) T .
)(
-calculate the increment,

20
3 ( )
10
( )
d
=
t .
s o( )
-compute the components,
){
)}
dT = d ( ) m( ) T s( ) s( ) T T .
-in order to define the strain increment, dTT, first compute the work,
dw = |()d()| , and then define
dw
d TT =
.
TT
-update the total strain according to TT = TT + dTT

-update T according to Ti = Ti + dTi
-update the o() according to o() = o() + hd()
-print out TT, TT.
end the outer loop
Results for Case A (larger latent hardening) and Case B (larger self hardening) are shown below.
The tensile flow stress for the larger self hardening case is larger, particularly at larger strains.
For both cases, slip system 6 in Table 8.1 has the larger resolved shear stress initially, and slip
systems 1 and 4 have the second largest. This can be verified by noting in Fig. 8.6 that slip
system 6 has the largest resolved shear stress when the tensile axis is in the vicinity of point P.
For the larger latent hardening case (A), slip system 6 has the largest slip increment, d() over
the entire range considered, from TT = 0 to 0.8. In contrast, the larger self hardening case (B)
displays several transitions, and at TT = 0 to 0.8, slip system 4 has the largest slip increment.
For Case A, the components of T at TT = 0.8 are approximately [0.51 0.80 0.31], and T
continues to rotate toward [1 1 0]. However, the corresponding components for Case B are
approximately [0.49 0.65 0.58]. The larger self hardening in Case B causes the slip system with
the largest current d() to harden substantially, so that no one slip system dominates over the
entire history of loading.
Page 20
Chapter 8
Stress
tensile
axisaxis
(MPa)
stressalong
along
tensile
(MPa)
45
40
35
30
hself>hlatent
hself<hlatent
25
20
15
10
5
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
strainalong
alongtensile
tensile
axis
Strain
axis
24.
Consider a polycrystal with random orientation of crystals. Produce a lower bound to yield in
tension.
SOLUTION:
The simplest lower bound is to assume a trial stress distribution that is a homogeneous, simple
tension state, and that does not violate yield at any point in the material. In that case, In that
case, Eq. 8.42 indicates that the largest tensile stress possible is 2c, corresponding to some
optimally oriented slip system that is with m T = s T = 1/ 2 . Here, c corresponds to the
critical resolved shear stress to activate that slip system.

A.
1.
For the geometry shown in Fig. 9.1b, find the relationship between the friction coefficient and the
ramp angle at which sliding begins.
SOLUTION:
The equilibrium of forces is projected on the x and y axes of the figure below.
fN
fT
=
From the Coulomb friction law we obtain:
2.
x direction:
w sin - fT = 0
y direction: - w cos + fN = 0
Or:
fT = w sin
fN = w cos
fF
w sin
=
= tan (at the start of motion)
w cos
fN
Use the rope formula to obtain the unknown forces in each of the following cases, assuming a
friction coefficient of = 0.2.
100N
100N
motion
(a)
(b)
motion
100N
motion
1/2
1/2
100N
motion
(c)
(d)
F
motion
F
100N
(e)
-1
Page 2
Chapter 9
100N
(f)
1
1/2
F
motion
SOLUTION:
Eq. 9.18 is used repeatedly:
a.
b:
c:
d:
e:
f:
3.
F2 = F1 exp ()
, where is the accumulated "angle of wrap."
= 90 =
, so that: Fa = 100 exp (0.2 ) = 137 N

2
2
= 45 =
, so that: Fb = 100 exp (0.2 ) = 117 N

4
4
1
tan( ) = , = 2 tan -1(0.5) = 0.927, so that: Fc = 100 exp 0.2 (0.927) = 120 N
2
2
1 (left) + 2 (right) = = tan -1 1 =
, so that Fd = 100 exp (0.2 ) = 117 N

4
4
1 =
, 2 = , 3 = , so that: Fe = 100 exp (0.2 ) = 187 N

4
2
1 =
, 2 + 3 = , = , so that: Ff = 100 exp (0.2 ) = 137 N

4
4
2
2
Use the geometries presented in Problem 2 and the following information to find the true strains in
the region where F is applied:
Ao = original cross sectional area of sheet = lmm2
= 200 0.2 (MPa)

Which geometry cannot be deformed as shown?
Chapter 9
Page 3
SOLUTION:
First, calculate the maximum force which can be applied to the strip without breaking it,
assuming a uniaxial tensile state of stress (see Chapter 1):
F = A = 200 MPA =
N
mm
0.2 1mm2 exp(- )
The maximum force occurs at = n = 0.2:
Fmax = 200 MPA =
N
mm 2
0.20.2 1mm2 exp (- 0.2) = 119 N
The rope will not break only for cases b and d. For these cases, the strains may be calculated (by
trial-and-error or Newton iteration, for example) as follows:
Fb or d = 117 = 200 exp(-) 0.2 = 0.13
4. a. A ring compression test is carried out using a standard forging press. The original and final ring
dimensions are shown below. Find the value of m or from these data using Fig. 9.14.
Height (original) = 100mm
Height (final) = 60mm
Inside radius (original) = 200mm Inside radius (final) = 160mm
b. Given the purpose and use of the ring compression test, which limiting form of friction (Coulomb
or sticking) is likely to be the more accurate?
SOLUTION:
a. The height and radius reductions are:
h = 100 - 60 = 40%
100
h
ri = 200 - 160 = 20%
ri
200
These values allow the use of Fig. 9.14 to determine the friction factors:
for a Coulomb friction law: 0.25
for a sticking friction or Tresca law: m 1.
b. Direct compression will involve contact pressures near the flow strength of the material. Then,
according to our simple view of friction regimes, Fig. 9.3, the friction is likely to follow more
closely a sticking friction or Tresca model. To verify which friction law is more accurate,
several tests should be carried out with different reduction ratios in height. With these tests, one
will be able to see if the experimental measurements on the evolution of the radii follow a curve
Page 4
5.
Chapter 9
generated by assuming either of the models presented in Fig. 9.14.

What are the friction factors determined for solid soap, graphite in heavy way oil, and way oil
from Fig. 9.16? What difficulty do you see in applying this test to differentiate lubricant
properties for production applications?
SOLUTION:
In Fig. 9.16 we observe that:
- solid soap corresponds to m = 0.1 at the beginning of the process
and m = 0.05 after;
- graphite in heavy way oil follows rather well the curve with m = 0.1;
- heavy way oil is closer in average to the m = 0.2 curve.
This test is not selective enough for a classification of the industrial lubricants, particularly at the
beginning of the process.
6.a. For the double-backward extrusion test, the test is stopped at two punch strokes and the cups
removed. Given the following experimental results and using Fig. 9.18, determine the friction
coefficients and friction factors for this case.
Punch stroke = 100mm,
h1 = 80mm,
h2 = 30mm
Punch stroke = 200mm,
h1 = 120mm,
h2 = 90mm
b. What can you say about the role of sliding distance in determining friction for this configuration,
lubricant, and material combination?
SOLUTION:
a. For a punch stroke of 100 mm we have (using Fig. 9.18):
h1 80
=
= 2.67 so that m 0.15
h2 300
and for a punch stroke of 200 mm:
h1 120
=
= 1.33 so that m 0.08
h2 90
b. The sliding distance is approximately constant in this test as the radii of the punches decrease so
that the contact with the part takes place on a small area and not on the extruded length.
B.
DEPTH PROBLEMS
7.
Derive an equation or equations that could be solved numerically to find 1 or 2 as a function of

n
H for the test shown in the figure. Assume Coulomb friction, Hollomon hardening ( = k ),
and uniform tensile strain (and tensile stress) in each leg. For simplicity, assume that the pins
have negligible diameter, so that the geometry can be computed simply.
Chapter 9
Page 5
2
F2
F1
F1
SOLUTION:
1 H
Note that the wrap angle is = tan a . Then, start with the rope fromula (Eq. 9.18):
H
F2 = F1 exp = F1 exp tan 1
a
F2 = 2 A = k 2 A 0 exp 2 , F1 = 1 A = k 1 A 0 exp 1
2
where:
Then:
k 2 A 0 exp 2 = k 1 A 0 exp 1 exp , or,
2 exp 2 = 1 exp 1 exp tan
H
a
(Eq. 1)
Because this provides only one equation relating H to 1 and 2, we must use the geometry
before and after deformation to find another relationship. Assuming that the strain is uniform at
all times in each leg, we find the original specimen length (2a+b) in terms of current leg lengths
and strains:
2a + b = a 2 + H2 exp 1 + b exp 2
(Eq. 2)
Eqs. (1) and (2) may be solved simultaneously to find 1(H) and 2(H) for given values of m and
n. Or, they may be combined to find a single equation that can also be solved numerically with
the same result. First we solve for 2:
2 = ln
exp 2
2a + b
=
a2 + H 2
exp 1 , or
b
2a + b
So that finally we obtain a single equation which may be solved for 1(H):
a2 + H 2
exp 1
b
Page 6
Chapter 9
ln
8.
2a + b
a2 + H 2
exp 1
b
2a + b
a 2 + H2
exp 1
b
= 1 n e 1 e
tan 1 H
a
Consider now a real test like the one shown in Problem 7.

a. Why would you expect to see a difference between 1 and 2 even for very narrow strips (to
insure uniaxial tension) and with nearly perfect lubrication (by using rollers, Teflon, and
oil, for example)? How would this effect depend or bend radius and sheet thickness?
b. Given results from the test described in Part a., how would you find the true friction
coefficient from a similar test with fixed pins, normal lubrication, and the same material?
SOLUTION:
a. The bending (and unbending) resistance of the sheet is responsible for an increase of the force
exerted on the right and left hand sides of the sheet as it passes over the pins, and consequently
1 > 2 . For a small radius of bending the local strain in the sheet is higher than for a large radius
and so is the bending energy for a work hardening material. An increase of the sheet thickness
will also increase the bending strain; therefore it will produce qualitatively the same result as a
decrease of the radius of bending.
b. The true friction contribution can be better approximated if the effect of bending can be
eliminated. For that purpose it is possible in the analysis of the test to subtract the bending force
measured by a frictionless test (or very low friction), with rotating pins to further reduce possible
drag.
9.
Use the rope formula and the geometry shown below to determine apparent friction coefficient
from F1 and F2 for the modified OSU Friction Test (see Figure 9.11b and dimensions below).
50mm
100mm
SOLUTION:
We introduce notation similar to that used for Problem 7, as shown in the figure below, and
make the same assumptions about small-radii pins.
Chapter 9
F1 /2
Page 7
F1 /2
F1 /2
F2
F2
F'
F'
F2
F' / sin
a
F'
b
The equilibrium of vertical forces and the rope formula provide two equations relating F1, F2,
and F':
F1
F sin =
F = F2 exp ()
2 ,
1 H
where, as in Problem 7, = tan a . F' is then eliminated from the equations, which are
solved for to obtain the desired expression:
ln
F1
1
1
ln
=
2 sin F2
10.
H2 + a 2 F1
2 H
F2
=
tan -1
H
a
a. Explain qualitatively why and how the friction factor or friction coefficient affects the shape
change in the ring compression test.
b. Could you design a plane-strain compression test similar to the ring compression test by
simply putting a long square rod between flat platens?
SOLUTION:
a. For a very low friction factor, the solution of the ring upsetting problem corresponds to a
homogeneous strain, so that the ratio between the inner and the outer radii will remain constant.
On the other hand, when the friction stress cannot be neglected, its effect will be more important
on the large radius zone because of the fact that it is exerted on a larger surface and the velocity
is higher. The braking effect of friction is therefore more important on the outer radius than on
the inner one. That is why the inner radius increases relatively slower than the outer one, and in
fact decreases for large friction stress.
b. If the rod is sufficiently long, a zero-displacement boundary condition will be enforced by

friction in the long direction. Then, we could watch the growth of the central or contact width as
a function of H, or the ratio of the central width to contact width. For higher friction, the contact
width will increase less relative to the central width (i.e. "barreling" will be promoted).

A.
1.
Calculate the ideal work to strain a unit volume of material under uniaxial tension from
= 0 to = for each of the following hardening laws:
n
a. = K ( + o )
(Ludwik law)
n
b. = o + K ( + o )
(Swift law)
-
c. = o (1- A e )
(Voce law)
d. = K
e. = o + K
(ideal plastic)
(linear hardening)
SOLUTION:
w =
v
a.
b.
c.
w =
v
w =
v
w =
v
d.
w =
v
e.
2.
dw =
o
d =
o d +
K + o d = n K
+ 1 + o
n
K + o d = o +
o d oA
e = o +
K +
o
n +1
n+1
n+1
o A
1
K d = K
o
o d + K
2
K d = + K
2
Calculate the drawing force using the ideal work method for a wire drawing operation from 2 mm
to 1 mm diameter of a Voce material where o = 500 MPa, A = 0.5, = 0.2 and where the
efficiency factor is assumed to be 0.5.
SOLUTION:
w
d = n i , wi =
o 1 A d, = 2 ln
500 0.5
d = 1 500 1.39 +
0.2
0.5
0.2
1.39
2
1
= 1.39
= 783 MPa
Page 2

fd = 783 MPa 1 mm
2
3.
= 615 N (Note that 1 MPa = 1N / mm2 , or 1 MPa mm2 = 1 N
Calculate the extrusion force for the same conditions as Problem 2.
SOLUTION:
Pext = d = 783 MPa , F = 783 MPa 1 mm
4.
Chapter 10
= 2460 N
Calculate the plane-strain drawing stress to reduce sheet thickness from 2 mm to 1 mm for the
material defined in Problem 2. How does this compare to the wire drawing stress computed in
Problem 2?
SOLUTION:
2
ln 2 = 0.80
=
3
d =
5.
1
500 0.5 0.2
500 0.80 +
0.2
0.5
0.80
1 = 430 MPa
Repeat Problem 4 for plane-strain extension (instead of drawing) and compare the result with
Problem 4 on an equal-original area basis.
SOLUTION:
For plane-strain extension (like uniaxial tension), there is no redundant deformation or frictional
work, so d = wi = 215 MPa.
6.
Calculate the total ideal work done for the operation shown below, given the hardening law
shown. Assume the material obeys von Mises yield.
= 500 1 0.6 exp (3) MPa
20mm
10mm
5mm
10mm
SOLUTION:
7.5mm
Volume constancy: 10 x 10 x 20 = 5 x 7.5 x l, l = 53.3

5 = 0.69
1 = ln 53.3 = 0.98
2 = ln 7.5
3 = ln 10
20
10 = 0.29
Chapter 10

2
3
Page 3
1
2
0.98 + 0.29 + 0.69 +
= 1.0
=1
500 1 0.6 exp (3) =
w = V
=0
= 2000mm 3 500 1.0 +
7.
500 0.6
3
e 3 1.0 1
MPa = 810,000 Nmm, or 810 Nm
A steel deforms at high temperature at a constant effective stress of 100 MPa. For a given
forming operation, the strain path may be approximated by two proportional paths, the first from
(1 = 0, 2 = 0) to (1 = 0.5, 2 = 0.25) and the second path from (1 = 0.5, 2 = 0.25) to (1 = 0.6,
2 = 0.5).
a. What is the ideal work per volume of material?
b. What is the tensile strain equivalent to this forming deformation?
c. If you assumed that a single proportional path was followed from the start to finish, how
would the answers to a) and b) change?
SOLUTION:
= 100 MPa
Path I:
1 = 0.5,
I =
Path II:
2
3
1 = 0.1,
II =
2
3
2 = 0.25,
2
3 = 0.75
2
0.5 + 0.25 + 0.75
2 = 0.25,
2
1
2 2
= 0.76
3 = 0.35
2
0.1 + 0.25 + 0.35
= 0.36
wi
v = 100 0.76 + 0.36 = 112 MPa
a.
b.
tensile = 1.12 = effective strain
c. for a proportional path:
2
3
0.6 + 0.5 + 1.1
1
2 2
= 1.10
wi
v = 110 MPa little different because there was no strain reversed, so nearly proportional.
8.
Repeat Problem 7 for a different forming operation for which the initial, intermediate, and final
geometric strains are as follows:
Page 4
Chapter 10
(1 = 0, 2 = 0)
(1 = 0.5, 2 = 0.25)
(1 = 0, 2 = 0)
Initial:
Intermediate:
Final:
SOLUTION:
Path I:
Path II:
I =
2
3
II =
2
3
0.5 + 0.25 + 0.75
0.5 + 0.25 + 0.75
a.
wi
v = 100 0.76 + 0.76 = 152 MPa
b.
tensile = 1.52
c.
1
2 2
For a proportional path:
2
3
= 0.76
1
2 2
= 0.76
0 + 0 + 0
1
2 2
= 0
(No ideal work done because there was no deformation, assuming a proportional path.)
9.
a) Use L'Hospital's Rule to find the plane-strain drawing stress for a frictionless case
starting from Eq. 10.29. Note that as 0, 0. (L'H ospital's Rule states that the
lim f(x)
xa g(x)
limit
of
a
composite
function
in
cases
where
lim f(x) 0, lim g(x) o
lim f(x) , lim g(x)
xa
xa
xo
or xa
may be found by
the ratio of the derivatives:
lim f (x)
f (a)
lim f(x)
=
=
x a g (x)
x a g(x)
g (a)
if these limits exist.

b) Compare the frictionless result obtained in part a) with the drawing stress obtained from
the ideal work method.
SOLUTION:
d =
a.
H 1+B
t
1 to
B
i
As 0,
B 0, so the term (1 + B) becomes 1 as B 0), i.e.
Chapter 10
lim
lim
d = H
o
Bo
f = H 1
so:
to
to
ti
which goes to 0 as B 0
ti
which goes to 0 as B 0
g = B
f (0) = H ln
lim
= H ln
o d
Page 5
to
to
ti
ti
= H ln
ti
to
, and g = 1
Note:
d x
a = a x ln a
dx
to
ti
b. The results are identical.

10.
Following the procedure of Exercise 10.3, derive an expression of the wire-drawing stress for the
Coulomb friction case. Show all of your steps clearly.
SOLUTION:
Everything is the same except the friction law, so

0 = r d 1 + 2 1 + P cot P dr
is replaced by P in Eq. 10.3-3:
The same substitution occurs in Eq. 10.3-6:

r
P = P tan r , or P =
tan 1
We then can remove P from the equilibrium equation:
cot 1
0 = r d 1 + 2 1 + r dr where
tan 1
And we use Eq. 10.3-10 as before to remove r:
0 = r d 1 + 2 1 + 1 dr = r d 1 + 2 1 + 1 dr, or
d 1
dr
=
r
2 1 + 1
Integration completes the solution for a material which does not strain-harden:
Page 6

ro
dr =
r
ri
ln ro =
i
11.
Chapter 10
d 1
2 1 + 1
1
2 1 +
ln
1 + d
or = 1 + 1 r o
i
2(1 + )
Following derivations in the text and in Problems 9 and 10, complete the following table:
Calculation Type
Wire Draw
Round
Extrusion
Pext
Sheet Draw
Sheet Extrusion
Pext
Slab (general , )
Slab (any , small )
Slab ( = o)
Ideal Work
SOLUTION:
Wire Draw, sd
Slab
Slab
(small )
Slab
( = 0)
Ideal Work
ro
1
1+
ri
B+1
B
2(1+)
r
1 o
ri
2B
ro
1
1+
ri
B+1
B
2(1+)
r
1 o
ri
2B
Sheet Draw, sd
H
to
1
1+
ti
H 1+B
B
Sheet Extrusion,Pext
(+1)
t
1 o
ti
H
to
1
1+
ti
H 1+B
B
ro
ri
2 ln
ro
ri
HB ln
to
ti
HB ln
to
ti
2 ln
ro
ri
2 ln
ro
ri
HB ln
to
ti
HB ln
to
ti
cot + 1
tan 1
(+1)
t
1 o
ti
2 ln
Where: B = cot
12.
Round Extrusion,Pext
H = 2
3
For each forming operation in Problem 11, show how much the external stress or pressure
changes by using the small angle approximation.
Chapter 10
Page 7
a. Use "typical values": rin, tin = 30mm, rout, tout,t = 20mm, = 20o, = 0.25,
b. Use "extreme values": rin, tin = 40mm, rout, tout,t = 20mm, = 45o, = 0.5.
SOLUTION:
a. Typical values:
Small :
Large :
= 20, = 0.25, = cot = 0.69
d (wire) =
d (wire) =
0.69 + 1
0.69
1.86
2
1
1 1.86
3
cot + 1 1.69
=
= 1.86
tan 1 0.91
2 0.69
2
3
= 1.050
2 11.86
= 1.086
Wire draw difference = 3.3%
Small :
Large :
d (sheet) =
0.69 + 1
0.69
H 1
2
3
0.69
= 0.598 H
11.86
2
d (sheet) = 11.86
1.86 H 1 3
= 0.637 H
Sheet draw difference = 6.1%
b. Extreme values:
Small :
Large :
= 45,
= 0.5,
d (wire) =
d (wire) =
= cot = 0.5
0.5 + 1
1 21
0.5
3 1 1
1 3
2
2 0.5
2 13
= 1.50
= 1.41
Wire draw difference = 6.3%
Small :
Large :
d (sheet) =
d (sheet) =
0.5 + 1
1
0.69 H 1 2
3 H 1 1
13
2
0.5
= 0.879 H
13
= 1.125 H
cot + 1
1.5
=
= 3
0.5
tan 1
Page 8
Chapter 10
Sheet draw difference = 22%

13. Perform a slab calculation for plane-strain compression of a Hollomon material
( = 600 MPa 0.25 ) subject to Coulomb friction ( = 0.3). Assuming an initial geometry of ho =
100mm and bo = 50mm, plot the friction hill at several values of h/ho (0.75, 0.5, 0.25) and find
Pmax and Paverage as a function of h/ho.
SOLUTION:
b x , = 0.3, h = 100mm, b = 50mm
P = 2 exp 2
o
o
h 2
3
h, b, are found as follows:
b
h
h
ho
0.75
0.5
0.25
75
50
25
ho bo
h
67
100
200
= ln h
ho
-0.29
-0.69
-1.39
= 2 h
3
0.33
0.80
1.61
455
MPa
567
MPa
676
MPa
= 600 0.25
Chapter 10
Page 9
Problem 10-13
8000
Note: These solutions assume Coloumb friction
throughout, which may exceed the interfacial
shear strength of the material at the interface
and thus not be accurate.
Pressure (MPa)
6000
h=25mm
4000
2000
h=50 mm
h=75mm
0
14.
20
40
60
Distance from Center (mm)
80
100
Consider the forging operation shown below. If the die material can withstand a contact pressure
of 1500 MPa and the workpiece exhibits a constant effective stress of 500 MPa, what is the
minimum height to which the billet can be forged. Assume two cases: a) = 0.25, b) sticking
friction.
h0 = 30 mm
b 0 =30mm
SOLUTION:
b
= 0.25: Pmax = 1500 MPa = 2 exp
h
3
a. Coulomb friction,
Constant volume:
1500 MPa =
h o b o = h b, b =
2 500 MPa
3
exp
hobo
900mm 2
=
h
h
0.3 b o h o
h2
Page 10
h min =
900mm 2
1500 MPa
3
2
500 MPa
1
2
Chapter 10
= 16.8mm
Pmax = 1500 MPa = 2 1 + b

2h
3
b. Sticking friction,
h min =
15.
900mm 2
1500 MPa
3
2
500 MPa
1
2
= 16.8mm
Consider the drawing of a sheet which reduces both of its dimensions in the cross sections.
Set up and derive the differential equation which governs the drawing operation. Show all of
your work and leave the resulting equation in the form of the following variables:
P2, P3, dA2, dA3, , , , t, and 1
Considering the symmetry of the operation and assuming a von Mises material, find the
ratio of the stresses 1/2/3.
dA2
top view
30 mm
15 mm
P2 acts on dA2
dA3
side view
30 mm P3 acts on dA3
60 mm
60 mm
end view
START
2t
30 mm
30 mm
END 15 mm
MIDDLE t
SOLUTION:
Force balance in the longitudinal direction:
0 = d 1 2 t 2 + 2 P2 cos + sin d A 2 + 2 P3 cos + sin d A 3
The strain ratios may be found from the geometry:

h = ln t
w = ln 2 t = ln t
l = 2 ln t ,
30
60
30
30
h / w / l = 1 / 1 / 2, (same as uniaxial tension)
so
Chapter 10
Page 11
Therefore, the stress state must be the one for uniaxial tension, or one which differs from that by
a hydrostatic pressure (which does not affect plastic deformation of a von Mises material):
1 / 2 / 3 = P / P / P
B.
DEPTH PROBLEMS
16.
Find the efficiency of a wire drawing operation, as estimated by a slab calculation, assuming the
following parameters:
= 1000 MPa, = 10o , rin = 50mm, rout = 40mm, = 0.25.
SOLUTION:
B +1
1
B
d =
Wire draw slab solution:
ro
ri
2B
where B = cot (small ) = 0.25 cot 10 = 1.42

242 1000 MPa 1 4
d = 1.42
5
Ideal work calculation:
2.84
= 800 MPa
d ideal =
Wi
=
Vol
d = 1000 MPa 2 ln
5
= 446 MPa
4
ideal
Wi
446 MPa
= d
=
= 56%
800 MPa
Wa
slab
d
Efficiency =
17.
rin
For the conditions presented in Problem 16, what is the maximum rout that can be performed?
SOLUTION:
Maximum draw stress = 1000 MPa
r
2.42
1000 MPa 1 o
1000 MPa =
ri
1.42
ro
ri
18.
= 1
min
1.42
2.42
1
2.84
= 0.73, or
2.84
ro
ri
= 1.37
max
a. Repeat Exercise 10.2 for the most general assumption that the redundant work is a fixed
fraction of the friction work:
r
=w
w , 0 1.
f
b. Show that your general expression reduces to the standard expression (Eq. 10.13) when = 0.
c. Show that your expression reduces to the improved expression (Eq. 10.2-10) when = 1.
di *
d. Compute do for the cases in Table 10.2 using your general expression with (i.e. all
non-ideal work is redundant work.) What are the differences compared with standard
expression for these cases?
Page 12
Chapter 10
e. Use your result and your knowledge of scaling of wf and wr with respect to size to predict
di *
whether do would go up or down as the size of the drawing operation increased. (Hint: how
does depend on size?)
SOLUTION:
total
a.
Wa
Wi
1 k
d =
=
=
n+1
n +1
Vol
a = k an , where a is such that Wi + Wr = Wdef =
Wa =
Wi
1
= Wi + Wf + Wr = Wi +
+ 1 Wr
Wr =
a =
n+1
Wi
k
a =
1
n+1
n +1
Wi
k
LDR equation:
ro
ri
Wi, Wdef = 1 +
1 +
= exp
, = where Wi =
n
n+1
for = 0, = 1, LDR = exp

b.
for = 1, =
1+
Wi Wi
k
n+1
n +1 i
k n (n+1) n =
n
n+1
= exp
1 k
n+1,
n +1
i = n + 1 n (n+1)
1
n+ 1 1 +
n+1
2
1+
n
n+1
(same as Eq. 10.13b)
+1
, LDR = exp
n+1
2
2
+1
2
n
n+1
(same as Eq. 10.2-10)
c.
for = , =
d.
, or
= k n+1 in
a = d ,
n +1
2
an+1
n+1
+1
1
, LDR = exp
(same as Eq. 10.13b with = 1 or Eq. 10.2-10)
Chapter 10
Efficiency,
material
Ideal,
brass
Ideal,
steel
Medium,
steel
Low,
steel
Low,
aluminum
V. Low,
zinc
di
1.00
do
*
std.
Page 13
di
do
improved
Difference
(%)
(Eq. 10.13)
0.50
2.12
2.12
0%
1.00
0.25
1.87
1.87
0%
0.70
0.25
1.55
1.87
21%
0.50
0.25
1.37
1.87
36%
0.50
0.15
1.33
1.78
34%
0.25
0.00
1.15
1.65
43%
e. Imagine an operation with ri = 1 and draw length l = 1, and imagine scaling this operation by a
factor C, i.e. ri = C, l = C. The deformation work will be proportional to the volume deformed,
i.e. Wdef = (scale) = C3 Wdef. The friction work will depend on the swept surface, i.e. Wf =
(scale) = C2 Wf. Therefore,
W
for = r as C large operations and 0 as C small operations .
Wf
Therefore, larger drawing operations can achieve somewhat larger draw ratios because friction
plays a smaller role.
19.
a) Derive equations similar to Eqs. 10.57 for the case of sticking friction.
b) Assuming plane-strain conditions, eliminate P from your equations.
c) By making appropriate geometric substitutions, write your equations in terms of t and dt
alone, and and d alone.
SOLUTION:
f =
3 whereas with Coulomb friction f = P .
a. For sticking friction, the friction stress is
With this substitution, Eqs. 10.57a and b may be immediately rewritten:
0 = d 1 t + 2 2P tan d x 1 before N (Neutral Point)
3
0 = d 1 t + 2 2P tan d x 1
3
after N (Neutral Point)
b. For plane-strain conditions (see Eq. 10.32), and small :
Page 14
Chapter 10
2
2
, so 0 = d t t + 2 3 1 1 tan 3 tan d x 1
3
,
where the plus sign is taken before the neutral point and the minus is taken after it.
P = 1 +
c. Using the geometry shown, and defining a new variable t as shown, the following geometric
relationships may be obtained readily:
tt
dt = dt = R sin d
t 2 o = R R cos
2
x1 =
dx 1 =
2 R t t 2 = R sin
t = 2 R R cos + t o
R t dt
2 R t t
dt = 2R sin d
tan =
= R cos d
2 R t t
R t
Then, in order to write the governing equation in terms of and d alone, we make the
appropriate substitutions in the result of Part c:
0 = d t t + 2 1 tan 2 tan dx 1
3
3
1 2 R R cos + t o
R cos d
which, when rearranged, obtains:

0 = 1 R sin d + 2 R R cos + t o d 1 2R cos 1 sin 2 sin d
3
3
R-t'
R
-x1
ti/2
t'
t/2
to/2
In order the write the same expression in terms of t and dt alone, we make a different set of
substitutions:
Chapter 10
0 = d t t + 2 1 tan 2 tan
3
3
tan =
2 R t t
R t
Page 15
dx 1
dx 1 =
R t dt
2 R t t
which, when rearranged, obtains:

0 = d 1 t +2
R t
2Rt t
+ 1 + 2 dt
3
where the minus sign applies before the neutral point and the plus sign afterward, and
and dt' = dt/2.
20.
t' = t 2to
Real rolling operations exhibit several characteristics which are not easily seen in simple slab
analysis. What do you think is the origin of the effects?
a)
Plane sections do
not remain planar
Widening
c)
e)
b)
Internal cracking
Side cracking
d)
before
Crowning
after
SOLUTION:
a. Friction at the rolls draws the sheet through the rolling operation, but since it operates only at the
surface, the central material tends to be restrained by the material behind it.
b. The shear strains established by the gradient of friction force can cause shear banding, and strain
localization through the thickness. In fact, the tensile stress components along the rolling
direction can also be large enough to produce cracking.
c. Widening occurs when the roll contact zone is comparable in the rolling direction relative to the
Page 16
Chapter 10
width of the contract region, so that the operation begins to resemble an upsetting operation.
d. Side cracking can occur because shear strains are set up as the plane strain state in the center
approaches uniaxial compression at the edge. The tensile elongation caused by adjacent material
can also introduce tensile stress cracking.
e. Crowning occurs by roll bending, i.e. the pressure across the roll causes beam-type bending of
the rolls.
21.
Consider a sheet-drawing operation (plane strain) that uses streamlined dies, as shown.
30mm
20mm
a) Use a simple 5-step numerical procedure to find 1 at t = 40mm, 36mm, 32mm, 28 mm,
and 24 mm and 20mm.
tin
t
b) What is the limiting out ratio that can be attained with these dies based on your numerical
procedure? How does this compare with the result for straight-sided dies?
SOLUTION:
a.
d 1 = B 1 H 1 + B
dt
t
1 = B 1 H 1 + B
t
t where H = 100 ksi and B = 0.25 cot [2 (t- 18)]
(Eq. 10.25)
Step #
t(mm)
t (mm)
1 KSi
40
--
--
--
1
2
3
4
5
36
32
28
24
20
0.34
0.47
0.69
1.18
3.58
-4
-4
-4
-4
-4
14.9
17.5
20.1
25.9
34.9
so, d = 114.2 ksi.

b. Interpolating between steps 4 and 5 to find the t where d = 100 KS: yields
t* = 24 100 79.3 4mm = 21.6mm, or

114.2 79.3
ti
to
40
= 1.85
21.6
1 ksi
0
(bdy. cond.)
14.9
32.4
53.4
79.3
114.2
Chapter 10
22.
Page 17
Using a slab analysis similar to the one for the plane-strain compression case, derive an
expression for pressure as a function of radial position for the compression of a cylinder.
Consider cases with a) Coulomb friction and b) sticking friction.
SOLUTION:
h dr
(r+r) h (r+dr) d
d/2
f r dr d
R
r
r h dr d
d/2
h dr
Force balance in the radial direction:

r + d r h r + dr d r h r d 2 h d r d 2 f r d r d = 0
2
r h d r + r h h 2 2 f dr = 0
For axisymmetric flow:
r = , so r =
For isotropic plasticity:
h d r = 2 f d r
For von Mises flow rule:
r = z + = P
f = P
a. Coulomb Case:
2
=
dr
h d P = 2 P d r dP
P
h
dP =
P
P = exp
2
2
d r ln P =
R r , so
h
h
2
Rr
h
b. Sticking Friction Case:
(Coulomb friction)
f =
3
Page 18
h d P = 2 d r dP = 2 d r
3h
3
P = +
23.
Chapter 10
2 hdr
3
dP =
P
2
R r
3h
(Sticking friction)
How would the sheet drawing stresses depend on normal plastic anisotropy (r) based on Hill's
theory?
SOLUTION:
To find the effect of normal anisotropy, we need to substitute Hill's normal quadratic yield
function (Eq. 7.60) in place of von Mises yield, and re-solve for d. [Note, however that we
conventionally choose 3 normal to the sheet plane instead of x2 as was done previously for the
sheet-draw problem.]
For plane-strain (2 = 0) using Eq. 7.61b:

d
d 2 = 0 = 2 3 r 1 2
, so
1+r
2 =
r
1
1 +
1+r
1 +r 3
Substituting into Eq. 7.60 (Hill's yield function):

2r + 1
=
1 3
1+r
(Hill)
Which corresponds to the von Mises result (Eq. 10.22) for r = 1.
3
=
1 3
2
(von Mises)
Therefore, the replacement for Eq. 10.23 becomes
1 +r
3 = 1 HHill , where HHill =
2r + 1
The final result corresponds to Eq. 10.29:
H Hill 1 + B
to B
HHill =
d =
1+
B
ti
,
where
Therefore as r increases, d increases.
1+r
2r + 1

Fundamentals Wagoner&Chenot ProblemSolutions

Uploaded by

Copyright:

Available Formats

Fundamentals Wagoner&Chenot ProblemSolutions

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fundamentals Wagoner&Chenot ProblemSolutions

Uploaded by

Copyright:

Available Formats

CHAPTER 1 - PROBLEM SOLUTIONS

Fundamentals of Metal Forming - Solution Manual

458 lbs = 30,500 psi E = 27,000 psi = 30 x 106 psi

e t = 0.80" = 0.40 or 40%

e u = 0.5" = 0.25 or 25%

e pu = e t e u = 0.40 0.25 = 0.15 or 15%

Eng. Strain Eng. Stress

Fundamentals of Metal Forming - Solution Manual

True Stress / Strain

Ln. (true stress)

Ln. (true strain)

The n value (slope of the ln-ln plot) is as follows:

Fundamentals of Metal Forming - Solution Manual

= 500 ( + 0.05) 0.25

= 100 + 500 ( + 0.05) 0.25 (Ludwik)

= 500 1 - 0.6 exp (-3)

= 500 sin (2)

o = 0.05, n = 0.25 u = 0.20

This is transcendental, so it cannot be solved algebraically.

Fundamentals of Metal Forming - Solution Manual

for o = 250, k = 350,

(Never stable for constant o .

Fundamentals of Metal Forming - Solution Manual

Fundamentals of Metal Forming - Solution Manual

(Figure for Problem 1-4 follows.)

Fundamentals of Metal Forming - Solution Manual

Eng. Stress Eng. Stress

Eng. Stress Eng. Stress Eng. Stress

Fundamentals of Metal Forming - Solution Manual

Uniform strain (eng.)

Uniform Strain (true)

u (v2) = u (v1 ) = n o = .25 0.05 = 0.20

Repeat Problem 6 with two other stress-strain curves:

Fundamentals of Metal Forming - Solution Manual

In this case, the strain-rate sensitivity varies with strain.

Fundamentals of Metal Forming - Solution Manual

It is very difficult to match tensile specimens

520 1% x 520 = 515 to 525

but now we consider the range:

So, the combined uncertainty of m is in the range:

So, a 1% uncertainty in stress corresponds to a 50% uncertainty in m!

Considering the specimen-to-specimen variation mentioned in Problem 9, it would be very

Fundamentals of Metal Forming - Solution Manual

below, find m values using the various points marked.

For the "down jump" in rate:

Fundamentals of Metal Forming - Solution Manual

(a) Uniform gage length

(b) General notch

(c) Severe notch

Fundamentals of Metal Forming - Solution Manual

enough to resolve using most tensile testing equipment.

multiplicative: = F() G()

Fundamentals of Metal Forming - Solution Manual

= 10 o = 500 0.25 + 25 (10)0.30 = 500 0.25 + 50

Fundamentals of Metal Forming - Solution Manual

Substitution leads to a transcendental equation:

Fundamentals of Metal Forming - Solution Manual

CHAPTER 2 - PROBLEM SOLUTIONS

a + b = (1, 1, 1) + (1, 2, 3) (2, 3, 4)