Many engineering materials exhibit some cracking behavior under
sustained loading in the presence of an environment (thermal and/or
chemical). The type of cracking
behavior for many chemical environments is referred to as stress-corrosion
cracking behavior. The mechanism for
this attack process has been attributed to the chemical reactions that take
place at the crack tip and to diffusion of reactive species (particularly
hydrogen) into the high stressed region ahead of the crack. The cracking process
has been noted to be a function of time and it is highly dependent on
the environment, the material, and the applied stress (or stress-intensity
factor) level.
For a given material-environment interaction, the
stress-corrosion-cracking rate has been noted to be governed by the
stress-intensity factor. Similar
specimens with the same size of initial crack but loaded at different levels
(different initial K values) show
different times to failure [Brown, 1968; Sullivan, 1972; Chu, 1972], as shown
in Figure 5.1.13.
A specimen initially loaded to KIc
fails immediately. The level below
which cracks are not observed to grow is the threshold level that is denoted as
KIscc.

Figure 5.1.13. Stress Corrosion Cracking Data [Brown, 1968]
If the load is kept constant during the
stress-corrosion-cracking process, the stress-intensity factor will gradually
increase due to the growing crack. As a
result, the crack-growth rate per unit of time (da/dt) increases according to

|
(5.1.7)
|
When the crack has grown to a size so that K becomes equal to KIc,
the specimen fails. This is shown
schematically in Figure 5.1.14. In typical tests, specimens may be loaded to
various initial K’s such as K1, K2, and K3. The time to failure is recorded giving rise
to the typical data point (t1,
K1). During the test, K will increase, as a result of crack extension, from its initial
value to KIc, when final
failure occurs. The times t2 and t3 represent the time to failure for higher K’s such as K2 and K3.

Figure 5.1.14. Stress Corrosion Cracking
The stress-corrosion threshold and the rate of growth depend on
the material and the environmental conditions.
Data on KIscc and da/dt can be found in the Damage
Tolerant Design (Data) Handbook [1994] .
Typical examples of KIscc
and da/dt data presentation formats
are shown in Figures 5.1.15 and 5.1.16.
Figure 5.1.15. KIscc
Data as Presented by the Damage Tolerant Design (Data) Handbook [1994]
Figure 5.1.16. Stress Corrosion Cracking Rate Data for
2024-T351 Aluminum as Presented by the Damage Tolerant Design (Data) Handbook
[1994]
As illustrated in Figure 5.1.17, a
component with a given crack fails at a stress given by
It will exhibit stress-corrosion-crack growth when loaded to
stresses in excess of

Figure
5.1.17. Stress
Required for Stress Corrosion Cracking
In service, stress-corrosion cracks have been found to be
predominantly a result of residual stresses and secondary stresses. Stress-corrosion failure due to primary
loading seldom occur because most stress-corrosion cracks favor the short
transverse direction (S-L), which is usually not the primary load direction. In many materials, the long transverse (T-L)
and longitudinal (L-T) directions are not very susceptible to stress corrosion.
Prevention of stress corrosion cracking is preferred as a
design policy over controlling it as is done for fatigue cracking. This means that stress-corrosion critical
components must be designed to operate at a stress level lower than
in which ai
is the initial flaw size as specified in the Damage Tolerance Requirements of
JSSG-2006. However, if stress corrosion
can occur, it must be accounted for in damage tolerance analyses by using an
integral form of Equation 5.1.7.
Stress-corrosion cracking may occur in fatigue-critical
components. This means that in addition
to growth by fatigue, cracks might show some growth due to stress
corrosion. In dealing with this
problem, the following should be considered:
·
Stress-corrosion cracking is a phenomenon that
basically occurs under a steady stress.
Hence, the in-flight stationary stress level (l g) is the governing
factor. Most fatigue cycles are of
relatively short duration and do not contribute to stress-corrosion
cracking. Moreover, the cyclic crack
growth would be properly treated already on the basis of data for
environment-assisted fatigue-crack growth.
When stress corrosion cracking is expected, the stress corrosion
cracking rate should be superimposed on the fatigue crack growth rate [Wei & Candes, 1969; Gallagher
& Wei, 1972; Dill & Saff, 1978; Saff, 1980].
·
Stress-corrosion cracking is generally confined to
forgings, heavy extrusions, and other heavy sections, made of susceptible
materials. Thus, the problem is
generally limited to cases where plane strain prevails.
·
The maximum crack size to be expected in service is
, where s equals sLT
or sDM,
depending upon the inspectability level (see Section 1.3).
If stress-corrosion cracking is not expected at any crack size,
the l-g stress, s1g, should
be lower than
. With ac given as above, it follows
that complete prevention of stress corrosion extension of a fatigue crack
requires selection of a material for which:

|
(5.1.8)
|