Creep Notes MAAE 4102 -- 2013

Creep refers to the continuous plastic deformation of a material with time It is

accelerated by the increase in the stress level or temperature

Primary Creep:Strain hardening more rapid than strain softening decreasing rate of strain
Secondary Creep:Equilibrium between strain hardening and strain softening constant rate longest
period of time
Tertiary Creep:Stress level increased due to reduction in area - fracture

Creep Curve for various stresses

Creep Curve for various Temperatures

Empirical Expressions for Creep Data

Various expressions have been used to represent constant stress creep data
All expressions have the form:- c = F(s, T, t)
Most generally used expression :- c = B m tn

(1)

Known as the Norton-Bailey relationship

B, m, and n are empirical constants
Norton- Bailey law incorporates both primary and secondary creep
m > 1

(typically 4 20 for most metals)

N < 1

( 0.2 0.5

typically ~ 1/3)

T < Tm

If primary creep is ignored ( when T > Tm - secondary creep dominates)

c=tBm
Or

Creep Relaxation in a Bolt

Often associated with creep and is termed Stress Relaxation i.e. time dependent decrease
in stress at constant strain

Bolt initially tightened to a stress of o producing a strain of o

i.e. t = 0 ,

o , = o

Assume that the plates are rigid

After time t the effect of creep induces plastic deformation
Or creep strain c, which allows a relaxation of elastic strain
i.e.

0
(3)

Or
Substituting in (2)

(4)

In practice the stress relaxation is more rapid that that predicted by this equation because
the effects of primary creep have been ignored.
4

Example
Two flanged cover plates for a pipe connection with a flow area of 0.15m2 are held
together by 16 equally spaced steel bolts of 20 mm diameter. What should be the initial
tightening stress and strain in the bolts in order that the connection remains tight after one
year? Assume that the flanges are rigid and that a factor of safety of 1.5 is required.
Assume also that the operating conditions are t = 450oC and the pressure P = 0.6 MPa .
After what period should the bolts be retightened to prevent leakage? Take E = 170 GPa
and the creep law for the bolt steel is
Two flanged cover plates for a pipe connection with a flow area of
0.15m are held together by 16 equally spaced steel bolts of 20 mm diameter. What
should be the initial tightening stress and strain in the bolts in order that the connection
remains tight after one year? Assume that the flanges are rigid and that a factor of safety
of 1.5 is required. Assume also that the operating conditions are t = 450oC and the
pressure P = 0.6 MPa . After what period should the bolts be retightened to prevent
leakage? Take E = 170 GPa and the creep law for the bolt steel is
4.2 10

( units MPa and hours)

( units MPa and hours)

Solution
Part(a)
Total force taken up by bolts = (0.6 x 106)(0.15) N = 9.0 x 104 N
Thus the minimum stress in each bolt for the connection to remain tight after one year
must be
9
16

10
4

1.5

20

17.90

1.5

26.86

(The safety factor is 1.5)

1

Using
365

24

4.2

10

170 10

1
1
4.1 1 26.86

Simplifying
.

56355

34.08
5

1
.

Part (b)
The joint will start to leak when the relaxed stress reaches the minimum stress in the bolt
due to the internal pressure (i.e. at 17.90 MPa)
1
4.2

10

170

10

1
1
4.1 1 17.90

t = 47612 hours = 5.44 years

1
34.08

EXAMPLE:

A delayed action contact switch

When the pin is removed, the compressed spring causes a tensile stress in the previously
unstressed lead rod. The gap between the contact points decreases steadily due to creep.
Calculate the delay time if the free length of the spring is 40 mm and the stiffness is 10
.
N/mm. For the lead
5 10
/hr.

When the spring recovers by a distance x

Spring force

10 40

25

10 15
3.5

Tensile stress in the lead

Now
But

10

10

.
.

20.3877 0.6
Or

4.905
.
.

i.e

10
.

0.6

t = 0.118 hours = 7.1 minutes

7

0.6

9.621

Creep in Bending of a Rectangular Beam

Assumptions:
(a) Plane sections remain plane
(b) Longitudinal fibres experience only single axial stress
(c) Creep behaviour is the same in tension and compression
Creep strain Equation
But

for a beam in bending

Therefore
Or

(i)

Equilibrium of external and internal moments is given by

2

2
1

(ii)
where C is a property of the cross section; it may be viewed as the 1

moment of area

combining (i) and (ii)

(iii)
( Compare this with the elastic analysis

Integrating (ii)
(iv)
and substituting for

, using (i), results in:

(v)

Since
(vi)
At a given cross section the

occurs at d/2
(vii)

No creep when m =1 (elastic condition) thus (vi) becomes

To obtain the deflections:
Curvature

Therefore
(viii)
Equation (viii) only gives creep deflection
Total deflection = Initial elastic deflection + creep deflection

The effect of creep is to relax the outer fibres and increase the core bending stress

10

Example: Creep deflection of a cantilever beam with an end load

(rectangular cross-section)

Bending Moment
For a beam with Norton-Bailey creep behaviour

2
2

1
3

Boundary Conditions
At

x = L, v = 0,

Thus the deflection after time t is

1

At x = 0

Initial elastic deflection

Therefore the tip deflection after time t is

11

Creep in Turbine Blades

Tensile Stress at a radius r due to the centrifugal pull of the material beyond r

(i)
Strain-Displacement relationship

0 (since blade is thin and narrow)

Therefore Elastic Displacement

(ii)

Assuming that the blade now undergoes creep displacement (stationary creep stage)

Substituting (i)
12

Therefore Total Creep Displacement:

where the numerical value of the integral depends on the value of m.

Total displacement at the blade tip = elastic displacement + Creep displacement
Say the total clearance at the blade tip is

Then

= elastic radial displacement at the disc periphery

+ elastic displacement of the blade
+ creep displacement of the blade

(The creep displacement of the disc can often be ignored because the disc is usually at relatively
lower temperatures)

13

Creep in Multi-axial Stress Systems

Many components such as pipes, pressure vessels etc. are subject to creep conditions under multiaxial stress states
Need to be able to relate creep data from simple tests to these conditions
Some of the laws governing plasticity are also applicable to creep;
(i)
(ii)
(iii)

Principal stresses and strains are coincident in direction

Plastic deformation occurs at constant volume
Max shear stresses and max shear strains are proportional
0

Constant volume
From (ii)

(1)

Rearranging

(2)
2
3
Expressing in terms of stationary creep rate

(3)
Where is function relating the three principal stresses to the uniaxial creep condition.

Defining the equivalent stress eq using the Von Mises yield criterion

For the uniaxial case

2 = 3 = 0 : eq = 1

14

(4)

Using Norton Bailey

Thus

(5)

Example: A Thin-walled tube under internal pressure

;

;
0

3
2
Subst into (5)

0
0

3
2

NOTE: No creep in the axial direction - this has been verified experimentally

15

EXAMPLE:
A spherical pressure vessel, 350 mm mean radius and 6 mm wall thickness, is made of a stainless
steel. The service pressure is 6 MPa and the operating temperature is 450 oC. At the operating
temperature, the creep parameters in the Bailey-Norton relationship are B = 1.47 x 10-22 and m =
6.8 (units are MPa and hours). Estimate the service life of the vessel if the radial expansion is not
to exceed 1.75 mm.
E = 170 GPa, = 0.3 at 450oC

For a thin-walled sphere

;

175

Elastic stresses
Elastic hoop strain

175

0.3 175

7.026

10

Total Permissible hoop strain (Elastic + creep)

7.026

10

4.2794

10

Now consider creep eqn

Ignoring primary creep

2

2
2 2 4.2794

1.47

10

t = 32542 hours = 3.71 years

16

10

Larson- Miller Parameter

Approach so far - Empirical equation of state
Creep strain
= f (, T, t)
e.g Bailey-Norton Law

B, m and n are empirical constants

Creep strain rates can be shown to vary significantly (from 100% to 150%) for even very small
changes in m.
This has serious implications for the accuracy required in determining the empirical constants
It is also impractical to run long-term tests to obtain reliable creep data
Most creep testing involves short term tests, ranging from a few hours to a few hundred hours.
The results are usually presented in the form of creep-rupture plots.

Extrapolation of short-time stress-rupture data to long times in design: --- use Larson-Miller
parameter PLM
Larson and Miller investigated over 28 different materials for creep and rupture; they postulated
that for each combination of material and stress level, there was a unique value of a parameter,
PLM, that is related to temperature and time by
(6)
C empirical constant; taken to be 20 for most materials

17

T* - absolute temperature T* =T +273 oCelcius

This essentially reduces the family of log vs tR curves for different temperatures to a single line
between log and PLM

Example of use of Larson-Miller Parameter

At 100 MPa and 800oC creep failure in a material occurs in 100 hours. At the same stress
what is the creep life at 1000 oC ?
Solution:
assume C=20.
T = 800 oC = 1073 K
PLM = T(C + log t)
PLM = 1073(20 + log 100)
PLM = 23,606
At T = 1000oC = 1273 K
23,606 = 1273(20 + log t)

or

t = 0.035 hours

Note: In this example the temperature was increased from 800oC to 1000oC, but the decrease in
life was a factor of 30,000 times. This is typical of thermally activated processes for which the
rate of the process increases exponentially with temperature.

18

Example:
Creep rupture data for a stainless steel are obtained from a series of tests
Temperature (o C)
Stress (MPa)
Rupture time tR (hours)

425
480
344
277
10120 3154

565
168
1054

650
127
98

730
85
31

815
58
10

Using the Larson-Miller parameter determine the maximum stress that may be applied to a
component made of the same steel operating at 375 oC for 15,000 hours before rupture.
Calculation of PLM = T*(20 + log10 TR) for the above data
PLM x 10-3
Log10

16.755 17.694 19.293 20.298 21.556 22.848

2.537
2.442
2.225
2.104
1.929 1.763

Plotting Log10 vs (PLM x 10-3)

5

4
3
2
1
0
1

11

16

21

Assuming straight line fit between Log10 and (PLM x 10-3) gives
Log10 = 4.569 -1.2657 x 10-4 PLM

----- (1)

At the service conditions T = 375oC for 15,000 hours

PLM = (375 + 273)( 20 + log10 15000) = 15.667 x 103
Subst in (1)

Log10 = 4.569 -1.2657 x 10-4 (15.667 x 103)

Log10 = 2.676
= 472.3 MPa

19

26