Solutions Manual 3

Ans:
Since h has units of length and g has units of (length)(time)-2, let us divide both
sides of the above equation by h g :

t f  h, m, g 

h g h g

The left side of this equation is now dimensionless. Therefore, the right side must
also be dimensionless, which implies that the time of flight cannot depend on the
mass of the object. Thus dimensional analysis implies the following functional
relationship:

h
t 
g

where  is a dimensionless constant. Only one experiment would be required to

estimate , but several trials at various heights might be advisable to obtain a
reliable estimate of this constant. Note that   2 according to Newton's laws of
motion.
CHAPTER 2

2.1 According to Eq. (2.25), the energy required to increase the crack area a unit amount is equal
to twice the fracture work per unit surface area, wf. Why is the factor of 2 in this equation
necessary?

Ans:
The factor of 2 stems from the difference between crack area and surface area.
The former is defined as the projected area of the crack. The surface area is twice
the crack area because the formation of a crack results in the creation of two
surfaces. Consequently, the material resistance to crack extension = 2 wf.

2.2 Derive Eq. (2.30) for both load control and displacement control by substituting Eq. (2.29)
into Eqs. (2.27) and (2.28), respectively.

Ans:
(a) Load control.
P  d  P  d CP   P dC
G      
2 B  da  P 2 B  da  P 2 B da
4 Fracture Mechanics: Fundamentals and Applications

(b) Displacement control.

  dP 
G  
2B  da 

 dP  d 1  
C    dC
  
 da  da C 2 da

  
2

dC P 2 dC
G C 
2B da 2 B da

2.3 Figure 2.10 illustrates that the driving force is linear for a through-thickness crack in an
infinite plate when the stress is fixed. Suppose that a remote displacement (rather than load)
were fixed in this configuration. Would the driving force curves be altered? Explain. (Hint:
see Section 2.5.3).

Ans:
In a cracked plate where 2a << the plate width, crack extension at a fixed remote
displacement would not effect the load, since the crack comprises a negligible
portion of the cross section. Thus a fixed remote displacement implies a fixed load,
and load control and displacement control are equivalent in this case. The driving
force curves would not be altered if remote displacement, rather than stress, were
specified.
Consider the spring in series analog in Fig. 2.12. The load and remote
displacement are related as follows:

T = (C + Cm) P T   C  Cm  P

where C is the “local” compliance and Cm is the system compliance. For the present
problem, assume that Cm represents the compliance of the uncracked plate and C is
the additional compliance that results from the presence of the crack. When the
crack is small compared to the plate dimensions, Cm >> C. If the crack were to
grow at a fixed T, only C would change; thus load would also remain fixed.

2.4 A plate 2W wide contains a centrally located crack 2a long and is subject to a tensile load,
P. Beginning with Eq. (2.24), derive an expression for the elastic compliance, C (= /P) in
terms of the plate dimensions and elastic modulus, E. The stress in Eq. (2.24) is the nominal
value; i.e.,  = P/2BW in this problem. (Note: Eq. (2.24) only applies when a << W; the
expression you derive is only approximate for a finite width plate.)
Solutions Manual 5

Ans:
The through-thickness crack has two tips; an increment of crack growth causes the
crack area to increase by 2B da. The compliance relationship for energy release
rate must be modified accordingly:

P 2 dC P 2 dC
G 
2 B d  2a  4 B da

Equating the above expression with Eq. (2.24) gives

 2 a P 2 a P 2 dC
G  
E 4 B 2W 2 E 4 B da

Solving for compliance leads to

  a
2

C   dC   ada     constant
2
BW E BE  W 

The constant corresponds to the compliance of the uncracked plate. Assuming a

gage length L, the total compliance is given by

  a  L 1  
2 2

Ctot      C  Cm
BE  W  2BWE

where Cm represents the compliance of the uncracked plate and C is the additional
compliance due to the crack. When a << W or a << L, the first term in the above
expression is negligible. Recall the previous problem, where it was argued that
displacement control is equivalent to load control in an infinite plate because C <<
Cm.

2.5 A material exhibits the following crack growth resistance behavior:

R  6.95(a  ao )0.5

where ao is the initial crack size. R has units of kJ/m2 and crack size is in millimeters.
Alternatively,

R  200( a  ao ) 0.5

where R has units of in-lb/in2 and crack size is in inches. The elastic modulus of this material
= 207,000 MPa (30,000 ksi). Consider a wide plate with a through crack (a << W) that is
made from this material.
6 Fracture Mechanics: Fundamentals and Applications

(a) If this plate fractures at 138 MPa (20.0 ksi), compute the following:
(i) The half crack size at failure (ac).
(ii) The amount of stable crack growth (at each crack tip) that precedes failure
(ac - ao).
(b) If this plate has an initial crack length (2ao) of 50.8 mm (2.0 in)
and the plate is loaded to failure, compute the following:
(i) The stress at failure.
(ii) The half crack size at failure.
(iii) The stable crack growth at each crack tip

Ans:
At instability, G = R and dG/da = dR/da. Therefore,

 2 ac
 6.95  ac  ao 
0.5
(1)
E
and
 2
 3.48  ac  ao 
0.5
(2)
E

Thus we have two equations to relate , ac and ao, and we must specify one of these
quantities.

(a)  = 138 MPa

From Eq. (1) above,

 138, 000 kPa 

2

 3.48  ac  ao 
0.5

2.07 10 kPa

8

ac - ao = 145 mm

Substituting into (2) gives

 138, 000 kPa  ac

2

 6.95 145 mm 
0.5

2.07 10 kPa

8

Thus
(i) ac = 290 mm
(ii) ac - ao = 145 mm
(iii) ao = 145 mm
Solutions Manual 7

(b) ao = 25.4 mm

Dividing Eq. (1) by Eq. (2) leads to

ac  2  ac  ao 

Therefore, if ao = 25.4 mm, ac = 50.8 mm and (ac - ao) = 25.4 mm. We can solve
for critical stress by substituting these results into Eq. (1):

 2  0.0508 m 
 6.95  25.4 mm 
0.5

2.07 10 kPa

8

Thus
(i)  = 213,000 kPa = 213 MPa
(ii) ac = 50.8 mm
(iii) ac - ao = 25.4 mm

2.6 Suppose that a double cantilever beam specimen (Fig. 2.9) is fabricated from the same
material considered in Problem 2.5. Calculate the load at failure and the amount of stable
crack growth. The specimen dimensions are as follows:

B = 25.4 mm (1 in) h = 12.7 mm (0.5 in) ao = 152 mm (6 in)

Ans:
At instability, G = R and dG/da = dR/da. Hence,

12 Pc 2 ac 2
Gc  2 3  6.95  ac  ao 
0.5
(1)
BhE

2Gc
 3.48  ac  ao 
0.5
(2)
a

Dividing (1) by (2) gives

ac
 2  ac  ao 
2
Thus
4
ac  ao  203 mm
3
and
8 Fracture Mechanics: Fundamentals and Applications

12 Pc 2  0.203 m 
2

Gc   6.95  203  152 

0.5

 0.025 m   0.0127 m  2.07 10 kPa

2 3 8

Pc = 5.16 kN

2.7 Consider a nominally linear elastic material with a rising R curve (e.g., Problems 2.5 and
2.6). Suppose that one test is performed on wide plate with a through crack (Fig. 2.3) and a
second test on the same material is performed on a DCB specimen (Fig. 2.9). If both tests
are conducted in load control, would the Gc values at instability be the same? If not, which
geometry would result in a higher Gc? Explain.

Ans
The driving force curve for the through crack is linear, while G varies with a2 for
the DCB specimen. Therefore, the two geometries would have different points of
tangency on the R curve, as Fig. S1 illustrates. The Gcvalue for the through crack
would be higher, and this geometry would experience more stable crack growth
prior to failure.

Through
Crack
DCB
G, R Specimen

Gc(1) R

Gc(2)

Crack Size

FIGURE S1 Effect of specimen geometry on instability (Problem 2.7)

2.8 Example 2.3 showed that the energy release rate, G, of the double cantilever beam (DCB)
specimen increases with crack growth when the specimen is held at a constant load.
Describe (qualitatively) how you could alter the design of the DCB specimen such that a
growing crack in load control would experience a constant G.
Solutions Manual 9

Ans:
In a conventional DCB specimen, compliance varies with a3, and energy release is
proportional to a2 when load is fixed. In order for G to remain constant with crack
growth, compliance must vary linearly with crack length. One way to accomplish
this is to taper the specimen width, as Fig. S2 illustrates. Alternatively, the
thickness can be tapered. The latter method is not as effective as the former because
compliance is less sensitive to the thickness dimension; recall that the moment of
inertia of the cross section is proportional to Bh3. Specimens such as illustrated in
Fig. S2, where G is relatively constant over a range of crack lengths, have been used
successfully in laboratory experiments.

FIGURE S2 Tapered DCB specimen

(Problem 2.8).

2.9 Beginning with Eq. (2.20), derive an expression for the potential energy of a plate subject
to a tensile stress  with a penny-shaped flaw of radius a. Assume that a << plate
dimensions.

Ans:
At fracture,
d  dWS
 
d d

For the penny-shaped crack,

WS  2 s a 2
and
dWS
 2 s
d

Combining the above results with Eq. (2.20) gives

10 Fracture Mechanics: Fundamentals and Applications

4 1  2  a f 2 d
2 s  
E d

The above equation must be integrated with respect to crack area to infer the
potential energy. The crack area can be written in terms of the the crack radius, a:

dA  2 ada
and
d 1 d

dA 2 a da

Therefore,

d 8 2 a 2 1  2 

da E
and
8 2 a3 1  2 
  o 
3E
where o is the potential energy of the uncracked solid.

2.10 Beginning with Eq. (2.20), derive expressions for the energy release rate and Mode I stress
intensity factor of a penny-shaped flaw subject to a remote tensile stress. (Your KI
expression should be identical to Eq. (2.44).)

Ans:
At fracture in an ideally brittle material, G  Gc  2 s . Rearranging Eq. (2.20) leads
to

4 1  2  a f 2
2 s   Gc
E
Thus
4 1  2  a 2 4(1 - 2) a 2
G G =
E E

Invoking the relationship between KI and G (Eq. (2.56)) gives

2
KI   a

Solutions Manual 11

which agrees with Eq. (2.44). Note that the plane strain KI–G relationship is
appropriate in this case. The strain parallel to the crack front is zero because the
crack is axisymmetric.

2.11 Calculate KI for a rectangular bar containing an edge crack loaded in three point bending.

P = 35.0 kN (7870 lb); W = 50.8 mm (2.0 in); B = 25 mm (1.0 in); a/W = 0.2; S = 203
mm (8.0 in).

Ans:
The KI solutions in Table 2.4 have the following form:

KI 
P
B W
f a  
W

Inserting a/W = 0.2 and S/W = 4 into the appropriate polynomial in Table 2.4 gives
f(a/W) = 4.70. Thus

KI 
 35 kN  4.70  29, 400 kPa m  29.4 MPa m
0.025 m 0.050 m

2.12 Consider a material where KIC = 35 MPa m (31.8 ksi in ). Each of the five specimens in
Table 2.4 and Fig. 2.23 have been fabricated from this material. In each case, B = 25.4 mm
(1 in), W = 50.8 mm (2 in), and a/W = 0.5. Estimate the failure load for each specimen.
Which specimen has the highest failure load? Which has the lowest?

Ans:
Failure load is inversely proportional to the geometry correction factor, f(a/W):

K Ic B W
Pcrit 
f a W 

Thus it is obvious from Fig. 2.23 that the CCT and DENT geometries have the
highest failure load and the SE(B) geometry the lowest (for fixed B, W, and a). The
calculated failure loads are tabulated below.
12 Fracture Mechanics: Fundamentals and Applications

Geometry f(0.5) Pcrit, kN

SENT 3.543 5.57
SE(B) 10.65 1.85
CCT 1.051 18.76
DENT 1.049 18.80
Compact 9.659 2.04

2.13 A large block of material is loaded to a stress of 345 MPa (50 ksi). If the fracture toughness
(KIc) is 44 MPa m (40 ksi in ), determine the critical radius of a buried penny-shaped
crack.

Ans:
At fracture, KI = KIc. Substituting the above data into Eq. (2.44) gives

2
44 MPa m   345 MPa   ac


ac = 12.8 mm

2.14 A semicircular surface crack in a pressure vessel is 10 mm (0.394 in) deep. The crack is on
the inner wall of the pressure vessel and is oriented such that the hoop stress is perpendicular
to the crack plane. Calculate KI if the local hoop stress = 200 MPa (29.0 ksi) and the internal
pressure = 20 MPa (2900 psi). Assume that the wall thickness >> 10 mm.

Ans:
Applying the principle of superposition (see Example 2.5) results in the following
stress intensity solution for this case:

a
K I  s   p  f  
Q

  0.01 m 
 1.14  220 MPa   28.4 MPa m
2.64

at  = 0°.
Solutions Manual 13

2.15 Calculate KI for a semi-elliptical surface flaw at  = 0°, 30°, 60°, 90°.

 = 150 MPa (21.8 ksi); a = 8.00 mm (0.315 in); 2c = 40 mm (1.57 in).

Ans:
From Fig. 2.19,

  0.008 m 
K I  s 150 MPa  f  
Q

, Degrees s f() KI, MPa m

0.00 1.20 0.632 15.74
30.00 1.12 0.780 18.08
60.00 1.10 0.943 21.36
90.00 1.09 1.000 22.62

2.16 Consider a plate subject to biaxial tension with a through crack of length 2a, oriented at an
angle  from the 2 axis (Fig. 13.1). Derive expressions for KI and KII for this configuration.
What happens to each K expression when 1 = 2?

Ans:
We can apply the principle of superposition separately to KI and KII:

K I  1 cos 2     a   2 cos 2  90     a

  1 cos 2      2 sin 2      a

K II  1 cos    sin     a   2 cos  90    sin  90     a

 1   2  cos    sin     a

when 1   2 , KI  1  a and KII = 0.

14 Fracture Mechanics: Fundamentals and Applications

2.17 A wide flat plate with a through-thickness crack experiences a nonuniform normal stress
which can be represented by the following crack face traction:

p( x )  po e  x / 

where po = 300 MPa and  = 25 mm. The origin (x = 0) is at the left crack tip, as illustrated
in Fig. 2.27. Using the weight function derived in Example 2.6, calculate KI at each crack
tip for 2a = 25, 50, and 100 mm. You will need to integrate the weight function numerically.

Ans:
Figure S3 is a plot of KI versus crack length for the through crack with the
exponential stress distribution given above. Values for three crack lengths are
tabulated below.

KI, MPa m
2a, mm
x = 2a x=0
25 29.21 48.15
50 21.74 57.37
100 11.06 62.65

2.18 Repeat Problem 2.17 with the following crack face pressure profile:

 x 
p  x   po cos  
 50 mm 

At what crack length(s) does KI = 0 at the right tip?

Ans.
Figure S4 is a plot of KI versus crack length for the right tip (x = 2a). The curve
passes through zero at 2a = 33 mm and 88 mm.
Solutions Manual 15

FIGURE S3. Solution to Problem 2.17.

FIGURE S4. Solution to Problem 2.17

16 Fracture Mechanics: Fundamentals and Applications

2.19 For an infinite plate with a through crack 50.8 mm (2.0 in) long, compute and tabulate Keff
v. stress using the three methods indicated below. Assume YS = 250 MPa (36.3 ksi).

Ans:
Applying Eqs. (2.41), (2.70), and (2.81) to the values above results in the following
stress intensity factors for the LEFM, Irwin, and strip yield methods respectively:

Keff, MPa m or ksi in

Stress, MPa (ksi) LEFM Irwin Correction Strip Yield Model

25 (3.63) 7.06 7.08 7.08

50 (7.25) 14.1 14.3 14.2
100 (14.5) 28.2 29.5 29.3
150 (21.8) 42.4 46.8 46.3
200 (29.0) 56.5 68.5 68.9
225 (32.6) 63.6 82.4 86.6
249 (36.1) 70.3 99.1 143
250 (36.3) 70.6 99.9 

2.20 A material has a yield strength of 345 MPa (50 ksi) and a fracture toughness of 110 MPa
m (100 ksi in ). Determine the minimum specimen dimensions (B, a, W) required to
perform a valid KIc test on this material, based on the traditional size requirements in Eq.
(2.88). Comment on the feasibility of testing a specimen of this size.

Ans:
From Eq. (2.88),

2
 110 MPa m 
a, B, W  a   2.5    0.254 m 10.0 in 
 345 MPa 
Therefore,
W  0.508 m (20.0 in)

Testing such a large specimen would impractical because:

• Machining costs would be very high.

• A very large test machine would be required.
• Materials are usually not available in such large section thicknesses. Even if
a section of sufficient size could be produced, its metallurgical properties
would not be representative of a thinner plate of the same material.
Solutions Manual 17

2.21 You have been given a set of fracture mechanics test specimens, all of the same size and
geometry. These specimens have been fatigue precracked to various crack lengths. The
stress intensity of this specimen configuration can be expressed as follows:

P
KI  f (a / W )
B W

where P is load, B is thickness, W is width, a is crack length, and f(a/W) is a dimensionless

geometry correction factor.

Describe a set of experiments you could perform to determine f(a/W) for this specimen
configuration. Hint: you may want to take advantage of the relationship between KI and
energy release rate for linear elastic materials.

Ans:
The stress intensity factor can be inferred from compliance measurements as
follows:

KI 2
E'
G
P 2 dC P2
 2
2 B da B WE '
f2 a
W  
assuming the specimen contains an edge crack, such that dA = Bda. Solving for
f(a/W) gives

 W
f a
BE ' dC
2 d a W 

Thus f(a/W) for the geometry of interest can be inferred by measuring the elastic
compliance as a function of crack length, evaluating dC/d(a/W), and inserting the
result into the above expression. Note that the absolute compliance depends on
specimen size and material properties, but the quantity (B E' C) is dimensionless,
and depends only on a/W.

2.22 Derive the Griffith-Inglis result for the potential energy of a through crack in an infinite
plate subject to a remote tensile stress (Eq. (2.16)). Hint: solve for the work required to
close the crack faces; Eq. (A2.43) gives the crack opening displacement for this
configuration.

The crack opening displacement at a distance x from the center of the crack
(assuming the coordinate system in Fig. 13.2) is given by
18 Fracture Mechanics: Fundamentals and Applications

4
2u y  a2  x2
E

for plane stress loading. The incremental closure work done at a point is as follows:

1 1
d   2 u y  x  Bdx d =
2 2 •2 uy(x) B dx

Thus the decrease in potential energy due to the formation of the crack is given by

4 2 B  2 a 2 B
a

E 0
  o  a 2
 x 2
dx 
E

which agrees with Eq. (2.16).

2.23 Using the Westergaard stress function approach, derive the stress intensity factor
relationship for an infinite array of collinear cracks in a plate subject to biaxial tension (Fig.
2.21).

Ans:
Substituting z* = z - a into Eq. (A2.39) and re-arranging gives

   a  z * 
 sin  
Z  z *   2W 
    a  z *      a  
2 2

sin     sin  
  2W     2W  

Let us now perform a series expansion about z* = 0 on the sin2 term on the left side
of the denominator:

    a  z *      a  
2 2
 a   a  z*
sin     sin     sin   cos   
  2W     2W    2W   2W  W

Substituting this result into the stress function and taking a limit leads to
Solutions Manual 19

 a 
 tan  
 2W  KI
lim  Z  z *   
z * 0  z* 2 z *
W

Solving for stress intensity gives

 a  2W  a 
K I   2W tan    a tan  
 2W  a  2W 

which agrees with Eq. (2.45).

CHAPTER 3

3.1 Repeat the derivation of Eqs. (3.1) to (3.3) for the plane strain case.

Ans:
In plane strain, the displacement of the crack face a distance ry behind the tip is
given by

4 1  2  ry
uy  KI
E 2

Substituting Eq. (2.63) into the above expression leads to

8 1  2  KI 2 4 K I 1  
2 2
4 G
   2u y  KI  
E 12  YS
2 2
3  YS E 3  YS

3.2 A CTOD test is performed on a three point bend specimen. Figure 13.3 shows the deformed
specimen after it has been unloaded. That is, the displacements shown are the plastic
components.

(a) Derive an expression for plastic CTOD (p) in terms of p and specimen dimensions.

(b) Suppose that Vp and p are measured on the same specimen, but that the plastic rotational
factor, rp, is unknown. Derive an expression for rp in terms of p, Vp and specimen
dimensions, assuming the angle of rotation is small.