MASSACHUSETTS INSTITUTE OF TECHNOLOGY

DEPARTMENT OF MECHANICAL ENGINEERING

CAMBRIDGE, MA 02139

2.002 MECHANICS AND MATERIALS II

Spring, 2004

Creep and Creep Fracture: Part III

Creep Fracture

c
L. Anand

Mechanisms of Creep Fracture

Figure 1: Creep strain-time curve for constant stress at

constant temperature

Image removed due to copyright considerations.

Figure 2: The upper row refers to low temperatures

( 0.3 TM ) where plastic ow does not depend strongly
on temperature or time; the lower row refers to the
temperature range ( 0.3 TM )in which materials creep.

Figure 3: Schematic of creep fracture mechanisms

Figure 4: Schematic of uniaxial stress versus time to

fracture data

Image removed due to copyright considerations.

Figure 5: Map of isothermal fracture data for

Nimonic-80A

Image removed due to copyright considerations.

Figure 6: Map of isothermal fracture data for 304

stainless steel

Image removed due to copyright considerations.

Figure 7: Map of isothermal fracture data for 316

stainless steel

Image removed due to copyright considerations.

Figure 8: Micrographs of copper plates illustrating the

continuous distribution of creep damage in plates
containing notches and subjected to far-eld uniaxial
tension. Note that it is predominantly the grain
boundaries perpendicular to the applied stress that are
preferentially damaged.

Creep Fracture

1. Creep Rupture Diagram


(1/)logtf
log = log C

C
C
1/
=
tf =
tf

Times to failure are normally presented as creep rupture diagrams. Their application is obvious. If you
know the stress and temperature you can read o the
life; for a given design life at a certain temperature,
you can read o the stress.

2. Monkman-Grant

tf =

= s

1/n
c
 ss

 0

s
/n
= /n  ss
 0

= C

(Monkman and Grant)

tf (css)

0.1 the creep strain to

Typically,
1 and C
fracture 10 %

3. Time-Temperature Equivalence

 Data is given in terms of in psi, tf in hours and T

in degrees Rankine (460 +0 F ).

Example Problem on Creep

A support beam made of 18Cr 8N i stainless steel

is to be used in a chemical reaction chamber operating at 6000C. The beam geometry and loading are
idealized as shown below.

The performance requirements are that

1. The beam is to carry a constant load F = 600 N .

2. No macro-crack formation due to creep fracture in

25 years.
3. Tip-deection not to exceed 4 cm in 25 years.

Determine if the beam meets the performance specications. If either of the failure criteria are not met,
then what is the maximum value of F that the beam
can carry and not fail?

Data for 304 Stainless Steel

 css = B n; B = 1.095 1018, n = 4.5

||
 =  0
s

n

;  0 = 1 109sec1, s = 98M P a, n = 4.5

Image removed due to copyright considerations.

"Master Rupture Curve for 18-8 Stainless Steel."

Solution:

M (x) = F (L x)
0xL
M (x) 1/n
|y|
(x, y) =
sgn(y)
 In
In =

|y|1+1/n dA

(1)

For a rectangular beam

n
h 1/n
2
In =
bh
2+
4n 
2


n
2v
|M |
=  0
sgn(M )
2
x
sIn

(2)

v = 0 at x = 0
BCs: v = 0 at x = 0, and x

|F |
sIn

n

(L x)n+2

Ln+2

1
+ Ln+1x
n+1
n+2
n+2


n
|F |
Ln+2
= |v(x = L)| =  0
(3)
sIn
n+2

v =  0

Check for Macro-crack Formation

From (1) and (2)

M (x)



 2y 1/n

 
sgn(y)

2

(x, y) = n
2+4n bh

Since M (x) = F (L x), maximum moment is at x = 0,

i.e., Mmax = F L, and maximum tensile stress occurs at
y = h/2, we get

max =

FL
n bh2
2+4n

600 N 1 m
4.5
2 3
2+44.5 0.02 0.04 m

max = 83.33 M P a, and since 1 M P a = 145 psi, we get

max = 12, 083 psi. The temperature is 6000C = 11120F =
15720R
From Larson-Miller master curve for 18-8 stainless steel,

log = logp q

p and q are constants

Solve for q:

log(a/b)
q=
; logp = loga + qa
b a
For a = 10, 000 psi , a = 41, 000;
for b = 2, 000 psi , b = 50, 000
5
Therefore, q = log(5)
=
7.77

10
9000
Solve for p: logp = log10, 000 + (7.77 105) 41, 000 =
7.1842

p = 1.53 107
Now, with LM = T (20 + log(tf )),
log = logp qT (20 + logtf )
log

 1/(qT )
p

= log(1020tf )

or
tf = 1020(p/)1/(qT )

where tf is the rupture time in hours, T is the temperature

in degrees Rankine and is the stress in psi.

tf = 1020

1.53 107

1
7.77105 1572

12, 083

8.187
20
3
= 10
1.2662 10
= 2.513 105 hours

Since 1 year = 365 24 = 8760 hours,

tf = 28.69 years
Therefore, a macro-crack will form on the tensile side
of the beam after approximately 28 years. The beam is
safe for 25 years.
Check for Deection

|F |
=  0
sIn

n

Ln+2
n+2

1/n
h

n
bh2
In =
2 + 4n
For b = 0.02 m, h = 0.04 m, n
106 m[3+(1/4.5)], F = 600 N ,  0
98 106 N/m2
=

109

= 4.5, In = 3.0184
= 109 sec1, and s =

600N

4.5

(1m)[2+4.5]

sec (98 106 N2 ) (3.0184 106m[3+(1/4.5)])

m

= 3.7 109 m/s = (3.7 109m/s) t

t = 25 years = 25y 365da/y 24h/da 3600sec/h =
7.884 108 sec

6.5

Therefore,

= (3.7 109)m/s (7.884 108)s = 2.951 m

Too much deection!
A deection of 4 cm would occur after only
0.04m
7 sec = 3000 hours
t=
=
1.08

10
3.7 109m/s
The load has to be decreased substantially. For a total
deection of 4 102 m in 25 years

4 102m
=
= 5.0 1011 m/s
25y 365da/y 24h/da 3600s/h

From

1/n

0)
(/


F = 
Ln+2

n+2

sIn = 231 N = 52 lbf

Alternatively, the ratio of the deections is the ratio of

the deection rates, which in turn are proportional to the
load ratio, raised to the power n = 4.5;


|F |
600 N

|F | = 600 N

4.5

.04
2.951

=


1
4.5

.04m/25y
2.951m/25y

= 600 N (.01355).2222 = 231 N.