Note: Descriptions are shown in the official language in which they were submitted.
CA 02417739 2003-O1-30
-1-
Title: CONTROL OF STRESS CORROSION CRACKING GROWTH BY
OPERATIONAL PRESSURE CONTROL
Field of the invention
This invention relates to stress corrosion cracking and, in
particular, stress corrosion cracking in steel.
Background of the invention
Stress corrosion cracking (SCC) deterioration of steel pipelines
is a major problem for the oil and gas transmission industry. There are two
types of SCC occurring in pipeline steels: high pH SCC and near neutral pH
SCC.
Throughout the years, many efforts have been made to identify
the mechanisms responsible for SCC in pipeline steels, in developing viable
techniques for inspecting pipelines for these detects, and for managing the
failure risks. Little work, however, has been done to develop means of
proactively slowing or arresting the growfih of existing cracks in pipelines
or to
reduce crack growth due to daily operation.
SCC in pipeline steels results from a synergistic interaction of a
corrosive medium with susceptible steels under operating stress. For buried
steel pipes this synergistic interaction appears to continue despite the fact
that
the pipelines are nominally protected against environmental deterioration by
catholic protection (CP)
Only about 5% of near neutral pH SCC cracks have the potential
to propagate to failure. The rest become dormant with a depth usually less
than 10% of the wall thickness. A problem is to determine what causes this
small fraction of cracks to grow and how this growth can be minimized.
Room temperature creep is often of significant importance in
structural materials. Its occurrence, for example, may be an important factor
CA 02417739 2003-O1-30
-2-
contributing to the crack growth during stress-corrosion cracking. For
pipeline
steels used in gas transmission or for structural materials for aerospace
application, room temperature creep deformation near the crack tip may result
in a time dependent crack growth.
Room temperature creep is a consequence of time-dependent
dislocation glide. It normally exhibits features of work hardening. The creep
strain-rate has its highest value at the start of the creep, and decreases
with
time until an eventual exhaustion.
The creep occurring under cyclic loading behaves differently
compared to that under static loading. Cyclic loading can cause either an
enhancement or a decrease of creep deformation referred to as cyclic creep
acceleration and retardation, respectively. The former is more frequently
reported than the latter in the literature. It has been shown that a virtual
exhaustion of creep under static loading can be dramatically revived by cyclic
loading. As a whole, however, the creep behaviour under a combination of
static and cyclic loading may be influenced by many factors. Because of its
complicated nature, the current understanding on room temperature creep
behaviours is still insufficient, particularly under alternating static and
cyclic
loading conditions.
Summary of the invention
In accordance with the instant invention, it has been determined
that dynamic strain is a key factor in influencing the growth of cracks either
of
high pH type or near neutral pH type. A crack that lacks dynamic strain may
cease to propagate or may propagate slowly. Therefore, control of the growth
of cracks may be obtained by controlling the dynamic strain of the crack tip.
A strain in the material is generated due to the movement of
"dislocations". The strain at the crack tip can be produced either by pre-
existing mobile dislocations in the steel, which are generated during the
CA 02417739 2003-O1-30
-3-
fabrication of a pipe, or mobile dislocations which are generated by high
stresses at the crack tip. If these dislocations are released in a way that
constantly breaks the oxidation film, or maintains an enhanced dislocation
rate at the crack tip, a crack will propagate at a high rate. On the other
hand, if
these mobile dislocations are exhausted or essentially exhausted, a crack
may cease to grow, or may grow slowly. Therefore, crack growth may be
controlled by essentially exhausting and, preferably by exhausting, the mobile
dislocations at the crack tip, preferably in a short time.
In accordance with the instant invention, the mobile dislocations
may be exhausted through a combined cyclic and static loading process,
which may be used to not only exhaust the mobile dislocations at the crack tip
in a pipe, such as a steel pipe, but to also blunt a sharp crack due to the
generation of significant strains at the crack tip in a short time. A blunt
crack
will reduce the stress concentration at the crack tip and the potential of the
crack growing.
In accordance with another aspect of the instant invention, the
mobile dislocations may be exhausted through the application of a static load.
The total treatment time may be from about to 1 to about 48
preferably from about 2 to about 24 and most preferably from about 4 to about
20 hours. The actual time required to essentially exhaust the dislocations
will
vary depending upon a number of factors including the strain rate, the number
of cycles the metal is exposed to during the cyclic loading, the severity of
the
strain, and the time for which the static load is held. While the actual
amount
of time may be longer, a shorter time period is preferred since it accelerates
the rate at which the material may be treated.
Accordingly, in accordance with one aspect of the instant
invention, there is provided a process for treating a metal member comprising:
a) applying a series of cyclic loads to the member; and,
b) subsequently apply a static load to the member.
CA 02417739 2003-O1-30
-4-
In one embodiment, from 1 to 1000 cyclic loads are applied at a
stress level R of from 0.1 - 1 wherein
R= minimum stress/maximum stress for each cycle.
In another embodiment, the cyclic loads are applied at a strain
rate of from 10-3 per second to 10-6 per second.
In another embodiment, the maximum stress per cyclic may be
from 5% higher than the daily operating stress to 130% of the design hoop
stress of the metal.
In another embodiment the static load is applied for 1 to 24
hours.
In another embodiment, from 40 to 1000 cyclic loads are applied
at a stress level R of from 0.1 - 1 and the static load is applied at from 5%
higher than the daily operating hoop stress to 130% of the design hoop stress
of the metal. Preferably, the cyclic loads are applied at a stress level R of
from
0.5 - 1 and more preferably at 0.7 - 1. Alternately, or in addition, the
static
load are preferably applied at from 5% higher than the daily operating hoop
stress to 110% of the design hoop stress of the metal.
In accordance with another aspect of the instant invention, there
is also provided a process for treating a metal member comprising applying a
sufficient series of cyclic loads to the member at a severity and frequency
which, together with a subsequent a static load which is also applied to the
member reduce stress corrosion cracking in metal.
In accordance with another aspect of the instant invention, there
is also provided a metal member which has been treated by applying a
sufficient series of cyclic loads to the member at a severity and frequency
which, together with a subsequent a static load which is also applied to the
member, reduces stress corrosion cracking in the metal member.
In one embodiment, the metal member has been treated by
applying from 0 to 1000 cyclic loads at a stress level R of from 0.1 - 1
wherein
CA 02417739 2003-O1-30
-5-
R= minimum stress/maximum stress for each
cycle.
In another embodiment, the metal member has been treated by
applying the cyclic loads at a strain rate of from 10-3 per second to 10'6 per
second.
In another embodiment, the metal member has been treated by
applying cyclic loads which have a maximum stress per cyclic from 5% higher
than the daily operating stress to 130% of the design hoop stress of the metal
member.
In another embodiment, the metal member has been treated by
applying a static load for 1 to 24 hours.
In another embodiment, the metal member has been treated by
applying the static load at from 5% higher than the daily operating stress to
130% of the design hoop stress of the metal member.
In another embodiment, the metal member is a pipe.
In accordance with another aspect of the instant invention, there
is also provided a process for treating a metal member comprising applying a
static load which is in the range of from 5% higher than the daily operating
stress to 110% of the design hoop stress of the metal to the member for 1 to
24 hours.
In one embodiment, the static load is applied at from 5% higher
than the daily operating hoop stress to 110% of the design hoop stress of the
metal.
The loading pattern to control strain in accordance with the
instant invention may be implemented using any technique known in the art.
Generally, the loading may be applied by pressurizing the interior of the
pipe,
such as by sealing one end of a section of pipe and applying a gas
compressor at the other, or, alternately, attaching gas compressors at the
ends of a pipe. The process may be applied not only to pipes which have
CA 02417739 2003-O1-30
-6-
been used (e.g. buried pipes) with existing SCC cracks, but also to newly
installed pipes for crack initiation mitigation. Further, the process may be
used
in conjunction with hydrostatic testing for crack detection so as to operate
the
process in such a way that a pipe line will not experience an increased crack
growth during its subsequent use after passing the hydrostatic testing.
In accordance with the broad aspect of this invention, the
process is applicable to any metal to prevent stress corrosion cracking. In an
alternate embodiment, it is applicable to metal, which has been fabricated
into
cylindrical sections for the production of, e.g., a pipe line. In accordance
with
another aspect of the invention, the process is applicable to a pipeline
constructed from steel.
Brief descrilption of the drawings
These and other advantages of the instant invention will be
more fully and completely understood in connection with the following
description of the preferred embodiment in which:
Figure 1(a) is a stress-strain curve of X-52 pipeline steel;
Figure 1(b) is a schematic illustration of the loading procedure
for creep testing;
Figure 2 is a comparison of creep curves obtained under static
and cyclic loading with various R-ratios;
Figure 3(a) is a comparison of creep curves obtained under
various loading conditions (a) (curve 1, starting with cyclic loading with tn
= 2 s
and N~ = 2 cycles; curve 2, starting with static hold with t,, = 2 s and N~ =
2
cycles; curve 3, starting with cyclic loading with tn = 1200 s and N~ = 100
cycles.) and curve 4;
Figure 3(b) is an enlarged scale version of curve 3 of Figure
3(a);
CA 02417739 2003-O1-30
-7-
Figure 4(a) is a graph showing creep deformation under
subsequent static loading following 1-60 cycles of cyclic loading;
Figure 4(b) is a graph showing creep deformation under
subsequent static loading following 80-7200 cycles of cyclic loading;
Figure 5 is a graph showing the dependence of cumulative
creep strain (the sum of cyclic and static creep strain) on the number of
cycles
of the prior cyclic loading;
Figure 6 is a creep curve under static loading following initial
7200 cycles of cyclic load, showing the incubation time for the burst of creep
strain;
Figure 7(a) is a graph showing the influence of the number of
cycles on the incubation time;
Figure 7(b) is a graph showing the creep strain accumulated
within 300 min under subsequent static loading;
Figure 8(a) is a graph of creep curves produced during repeated
cyclic and static loading;
Figure 8(b) is a graph showing partial curves of Figure 8(a) on
an enlarged scale;
Figure 9(a) is a graph of creep curves under alternating static
and cyclic loading;
Figure 9(b) and (c) are graphs, each of which shows a portion of
the curve of Figure 9(a) on an enlarged scale;
Figure 10 is a graph showing cyclic stress-strain hysteresis
loops for the initial few cycles of cyclic loading immediately following
initial
loading;
Figure 11 is a graph showing variation of cyclic plastic strain
range and cyclic creep strain per cycle with the number of cycles following
initial loading;
CA 02417739 2003-O1-30
Figure 12 is a graph showing two duplicate crack growth tests
showing the reduced crack growth rate as a result of creep deformation
produced by a 10 % over load and hold for 24 hour; and,
Figure 13 is a scanning electron microscope image showing a
stress corrosion crack that became blunt and branched due to the loading
process producing immediate creep deformation at the crack tip.
Detailed descriation of the invention
In accordance with the instant invention, a metal member, e.g. a
pipe, is treated by applying a static load for an extended period of time and,
optionally, applying a cyclic load prior to the static load.
It has been determined that cyclic loading has a significant effect
on subsequent static creep (i.e. when the static load is applied). The nature
of
the cyclic loading, as well as the nature of the subsequent static loading may
vary depending on a number of factors including the type of metal which is to
be treated, whether the material forms a protective passivating layer during
corrosion exposure and the geometric dimension of the cracks. If the pipe has
a protective passivating layer, it is preferred to perform the plastic
deformation
quickly so as to minimize the time that the layer is ruptured also to allow
the
protective layer to reform.
The cyclic loading which may be used in accordance with the
instant invention is defined by the number of cycles as well as the severity
of
the cycles. With respect to the number of cycles, the process may be
conducted with from about 0 to about 1000, and preferably from about 40 to
about 1000 cycles. The cycling may be achieved in less than about two days
and may use a strain rate of from about 10-3 per second up to about 10-6 per
second. Preferably, the cycling is conducted in from about 104 per second to
about 10-5 per second. It will be appreciated that the treatment time will be
significantly longer at a strain rate of 10-6 compared to a strain rate of
10'3.
CA 02417739 2003-O1-30
_g_
The severity or the stress level of each cycle may be measured
by R wherein
R = minimum stress/maximum stress for the cycle
While a large variation in R may be utilized, (e.g. it may vary
from about 0.1 to about 1 ), preferably R is from 0.5 to about 1 and, more
preferably, greater than about 0.7.
As the process is to control the strain at the crack tips, the
loading process may be conducted with very mild changes to the operating
pressure. An advantage of this process is that the cyclic loading would be
unlikely to cause any mechanical damage to the part of the pipe free of
cracks. Further, the cyclic loading will not cause the crack to propagate by a
fatigue mechanism.
The maximum stress per cycle is preferably higher than the daily
operating hoop stress for the system. The daily operating hoop stress is the
hoop stress that the metal will be exposed to during normal use. For example,
in the case of a pipe, the daily operating hoop stress is the pressure that
the
pipe would be exposed to during daily as it is used to transport a fluid and
is
based on the pressure of the fluid traveling in the pipe. Preferably, the
maximum stress per cycle could be from about 5% higher than the daily
operating hoop stress up to about 130% of the design hoop stress, more
preferably, from about 5% higher than the daily operating hoop stress up to
about 110% design hoop stress and most preferably from about 5% higher
than the daily operating hoop stress up to about 105% design hoop stress.
The design hoop stress may be determined based on the specified minimum
yield stress of the material and, typically, is set at about 80% of the
minimum
yield stress of the material and must be consistent with the codes and
standards such as ASTM 1003-95, or CSA 2662-99 (Canadian Standard), or
CFR 49 part 192 (Code/Standard of the United States)
Subsequent to the cyclic loading, the pipe is subjected to static
loading for an extended period of time. The static loading may be held for
CA 02417739 2003-O1-30
-10-
from about 1 to about 24 hours, preferably from about 3 to about 20 hours,
more preferably from about 3 to about 10 hours and most preferably from
about 3 to about 6 hours. The static load should not exceed the maximum
cyclic stress, and is preferably equal to the maximum cyclic stress. It will
be
appreciated that about 80% of the creep deformation occurs in about the first
1 -2 hours. The static loading is preferably conducted for a sufficient period
of
time to essentially blunt the crack tip.
The static loading may be conducted from about 5% higher than
the daily operating hoop stress up to about 130% of the design hoop stress,
preferably from about 5% higher than the daily operating hoop stress up to
about 110% of the design hoop stress and, more preferably, from about 5%
higher than the daily operating hoop stress up to about the design hoop
stress.
If the process utilizes only static loading, then the static toad is
conducted from about 5% higher than the daily operating hoop stress up to
about 110% the design hoop stress and, preferably, from about 5% higher
than the daily operating hoop stress up to about the design hoop stress.
Accordingly, in accordance with this aspect of the invention, the static
loading
is applied at a pressure less than that used for hydrostatic testing and may
be
applied for from about 1 to about 24 hours, preferably from about 3 to about
20 hours, more preferably from about 3 to about 10 hours and most preferably
from about 3 to about 6 hours.
If the crack tip has a protective passivating layer, then it is
preferred to conduct the plastic deformation relatively quickly so that the
protective layer may reform as soon as possible. In this way, the protective
layer will be disrupted for the shortest possible time thereby permitting the
protective layer to protect the pipe from other forms of corrosion.
Examales
CA 02417739 2003-O1-30
-11-
The material used in these examples was X-52 pipeline steel.
The chemical composition (in wt.%) is 0.07 C, 0.8 Mn, 0.016 P, 0.27 Si, 0.28
Cu, 0.09 Ni, 0.05 Cr, 0.012 Cb, 0.019 Ti, 0.031 AI, 0.0015 Ca, balance Fe.
The steel was annealed at 600°C for 1.5 h followed by furnace
cooling. The
steel exhibits a microstructure with prevailing equiaxed ferrite grains (~ 12
m).
Standard round tension specimens, with a diameter of 6 mm and a length of
36 mm in the reduced section, were used for mechanical testing. The long
dimension of the samples coincided with pipe longitudinal direction.
Specimens were polished to 600 grit sand paper before mechanical testing,
except that several samples were mirror-polished for slip observation on the
surface.
The stress-strain curve of the steel is shown in Figure 1 (a),
which exhibits significant yield point. The upper yield strength and the lower
yield strength are ~ 435 and 388 MPa, respectively. The yield point elongation
is ~ 2%. All the tests were carried out in a temperature-controlled laboratory
(22°C) on a servo-hydraulic testing machine (INSTRON 8516), which was
computer-controlled with the Instron WaveMaker-Runtime software. The
strain was measured using a strain gauge extensometer with a gauge length
of 25 mm, which was attached to the specimen with rubber bands. All the
specimens were initially loaded following the path, as described in Figure 1
(b).
For all tests, the loading strain rate was controlled to be ~1 x 10-2s-1. The
loading was first under displacement-control up to the end of discontinuous
yielding to avoid a sudden increase in strain, and then under load-control up
to the end of discontinuous yielding to avoid a sudden increase in strain, and
then under load-control to reach 440 MPa. The creep stress for static creep
tests and the peak stress for cyclic creep tests were equal to the flow stress
reached in initial loading. This loading process (referred to as 'initial
loading')
produced an initial strain ('pre-strain') of ~ 2.8%, which indicates that the
creep tests were conducted under a stress in the work-hardening region
(beyond the discontinuous yielding region).
CA 02417739 2003-O1-30
-12~-
After loading to 440 MPa, further tests were conducted under
one of the following three conditions: (1) under constant stress (pure static
creep); (2) under cyclic stress (pure cyclic creep); and (3) under a
combination of static and cyclic stress. One example of the creep process is
shown in Figure 1 (b) by the dashed line, in which the creep load is cyclic at
first and static next. The constant stress for static creep and the maximum
stress in cyclic creep were kept equal to 440 MPa. For cyclic creep, while the
maximum stress remained constant, the stress ratio (R-ratio = minimum
stress/maximum stress) varied from 0.1 to 0.90. The cyclic creep was
performed using a triangle waveform and at a frequency which varied with R-
ratio, keeping the loading rate constant (105 MPa s-1). The frequency was
0.13 for R = 0.1, 0.24 for R = 0.5, 1.2 for R = 0.9, respectively. If not
specified
otherwise, the R-ratio used was 0.1. All the data were recorded automatically
using a computer.
When a crack is present in the pipe, the crack tip in the pipe is
usually subjected to a stress higher than the yield strength of the material
because the crack tip tends to magnify the stress. This is true even if the
pipe
material is exposed to a hoop stress well below its yield strength.
Accordingly,
in order to simulate conditions at a crack tip, Examples 1 - 4 were conducted
using a smooth metal specimen so as to simulate the conditions at a crack tip.
Accordingly, the test stress was above the yield strength of the material (the
stress at which a material starts to deform plastically).
Example 1: Pure static and pure cyclic creep
The pure cyclic and pure static creep curves obtained are
compared in Figure 2, where the cumulative creep deformation as measured
by creep strain is plotted against creep time (the frequency was 0.13 for R =
0.1, 0.24 for R = 0.5, 1.2 for R = 0.9, respectively). For the cyclic loading,
the
strain rate was 5 x 10-4 and the time for cyclic loading was 500 minutes. The
CA 02417739 2003-O1-30
-13-
number of cycles was about 3900 cycles when R=0.1, 7200 cycles when
R=0.5, and 36000 cycles when R=0.9. For the static loading, the loading
stress was equal to the peak cyclic stress.
With the maximum stress equal to the static creep stress, the
cyclic creep deformation is much smaller compared to that of static creep for
the steel. With identical maximum stress, the cumulative cyclic creep
deformation decreases with decreasing R-ratio (i.e. increasing stress range).
As the initial loading processes (independent of cyclic or static creep tests)
were identical, the monotonic pre-strain for all creep tests is almost
identical,
excluding the influence on pre-strain of the results.
It should be noted that, for cyclic creep deformation, only the
maximum strain of each cycle is taken into account (i.e. is plotted in Figure
2).
(As a load cycle begins and ends at maximum stress, the maximum strain in a
cycle is the strain at the end of the cycle, i.e. at the second maximum stress
of
the cycle.) The use of the maximum strain in each cycle makes it possible to
directly compare the level of cumulative creep strain generated under pure
cyclic and pure static loading, since the creep stress in static loading is
equal
to the maximum stress in cyclic loading. Moreover, it also allows to present
the cumulative creep strain data without interruption for the cyclic loading
with
peak stress hold.
Example 2: Cyclic creep with peak-stress hold
With an inclusion of periodical hold at the maximum stress into
the cyclic load scheme, the amount of creep deformation depends on the
length of holding time (tn), the number of cycles in a unit (N~) and the
sequence of static holding and cyclic load (Figure 3a). The creep load for
curves 1 and 2 in Figure 3(a) is nearly identical, both with tn = 10 s and N~
= 2
cycles, except that the first block in creep load is cyclic for curve 1 and
static
for curve 2.
Without being limited by theory, based on the difference in creep
deformation for these two curves, it is believed that the nature of first
loading
CA 02417739 2003-O1-30
-14-
block (cyclic or static) of creep load is very important for cumulative creep
deformation as measured by creep strain. In these two cases, the creep
deformation generated is larger than that under pure cyclic loading, but
smaller than that under pure static loading (Figure 3a).
An extraordinary loading case (curve) 3 is shown in Figure 3(a),
see also Figure 3(b) For curve 3, about 100 loading cycles were performed at
a strain rate of 5 x 10-4. Subsequently, a static load was applied at stress
equal to the peak cyclic stress for about 2 hours. A remarkable jump in creep
deformation is observed in the first static hold (20 min) following the first
block
of cyclic loading with 100 cycles. . In contrast to curves 1 and 2, the creep
deformation for curve 3 is much larger than that produced in the pure static
creep (Figure 3a).
Example 3: Cyclic-load-induced burst of creep deformation under
subsequent static loading
For specimens that experience first cyclic loading and then static
loading (see the dashed line in Figure 1 (b) for loading procedure),
independent of the number of cycles of prior cyclic loading, there is a sharp
increase in creep strain in static loading following the initial cyclic
loading
(Figures 4a and 4b).
The burst of creep deformation in subsequent static loading
occurs normally after an incubation time (Figure 6). The creep strain in
subsequent static loading and the incubation time are dependant on the
number of cycles of prior cyclic loading. As shown in Figure 7(a), the
incubation time increases with the number of cycles performed prior to the
static creep. The creep strain in subsequent static loading increases at first
rapidly with the number of prior cyclic loading and reaches the maximum at =
1000 cycles (Figure 7b). It then decreases with the number of prior cyclic
loading to a level that is no longer sensitive to the number of prior cyclic
loading.
CA 02417739 2003-O1-30
-15-
The application of up to 40 cycles, prior to the application of
pure static creep, reduces the cumulative creep deformation (sum of the cyclic
and static creep strain in this case), compared to that of the pure static
creep.
This retardation effect becomes more significant with an increasing number of
cycles of the prior cyclic load from 1 to 20 cycles, and less significant with
increasing cycles up to 40 cycles (Figure 5). When the number of cycles is
more than 40 cycles, the cumulative creep strain is larger than the pure
static
creep strain (Figure 5).
Example 4: Creep behavior after static creep exhaustion
After the specimen has crept to exhaustion under static creep,
further cyclic loading with the maximum stress equal to the static creep
stress
does not produce a noticeable increase in creep deformation (Figure 8(a) and
Figure 9(a)). For example, as shown in Figures 8(b) and 9(b), only an
insignificant increase in creep deformation occurs in the initial few cycles,
which can be seen on an enlarged scale in Figure 8(b) and Figure 9(b). In
contrast to the cyclic loading directly following the initial loading, the
cyclic
loading after static creep exhaustion does not cause any increase in creep
deformation during subsequent static load (Figure 8(b) and Figure 9(c)).
The steel in this study experiences cyclic creep retardation (i.e.
the creep strain was limited). This cyclic creep retardation is, most of all,
due
to the cyclic hardening effect. In the present 'pull-pull' condition, the
cyclic
loading caused cyclic stress-strain hysteresis loops (Figure 10). The change
of the cyclic plastic strain range (AB in Figure 10 for the first cycle) or
the
cyclic strain amplitude (half of AB), can serve as a measure of whether a
material experiences cyclic hardening or softening. It can be seen from Figure
11 that the cyclic plastic strain range decreases with increasing number of
cycles for the initial few cycles, indicating a cyclic hardening effect. The
hardening effect diminishes at ~ 40 cycles, beyond which the change of the
cyclic plasticity range is negligible.
The cyclic creep strain per cycle (CD for the first cycle and DE
for the second cycle as shown in Figure 10) decreases with increasing
CA 02417739 2003-O1-30
-16-
number of cycles, particularly in the early stage of cyclic loading up to 40
cycles (Figures 10 and 11 ). It is reduced to a relatively low level as the
cyclic
creep approaches its exhaustion.
The change in cyclic plastic strain range and in cyclic creep
deformation can be traced back to the change in substructure, which is, in
turn, attributed to cyclic plastic and creep deformation. The monotonic pre-
strain has provided relatively large number of mobile dislocations, resulting
in
a significant static creep. The dislocations due to pre-strain may exist in
the
form of dislocation bundles, braids, incipient cell structures or their
mixture.
From the slip observation on the specimen surface, the pre-strain has induced
parallel slip lines, indicating the occurrence of planar slip. Parallel slip
lines
are also formed on the surface during the static creep following the initial
loading. Without being limited by theory, it is believed that the dislocation
slip
during the static creep following the initial loading exhibit similar
characteristics as those during initial loading.
The dislocation cells are effective barriers to dislocation
movement. Without being limited by theory, it appears that with increasing
number of cycles, the initial mobile dislocations add into the dislocation
cells,
resulting in a significant reduction in the number of mobile dislocations,
reducing the creep rate. On the other hand, the unloading portion of the
cyclic
loading, especially in the first few cycles, strongly reduces the dislocation
velocity (which is relatively high due to initial loading). This will also
cause a
smaller cumulative cyclic creep deformation if compared to the pure static
creep. The two aforementioned effects should be more significant for smaller
R-ratio (larger extent of unloading), but becomes less pronounced by
introducing a peak stress hold to the cyclic loading scheme.
Based on the results obtained in the present study, it is believed
that the initial cyclic loading may affect the subsequent static creep through
two competitive mechanisms. Firstly, the cyclic loading may cause hardening
of the steel, reducing the dislocation mobility as mentioned above. Secondly,
dislocation rearrangement will occur during subsequent static loading,
CA 02417739 2003-O1-30
-17-
releasing dislocation cells formed during prior cyclic creep deformation. The
dislocation cells may be in a quasi-stable state under cyclic load, but become
destabilized/collapsed under the subsequent static loading, which releases a
large number of mobile dislocations available for plastic deformation.
If the number of cycles is small (up to 40 cycles), the former
factor may be predominant, i.e. the dislocation mobility is reduced due to
cyclic hardening. The contribution from the latter factor is minor, as the
number of cycles is still insufficient for the development of cell structure
that
will cause a significant increase in creep strain during the subsequent static
loading. As a whole, the entire creep strain will be reduced.
With increasing number of cycles (more than 40 cycles), cell
structures may be better defined. Despite the reduction of dislocation
mobility
due to cyclic loading, dislocation rearrangement and the cell collapse under
subsequent static load may result in a relatively large number of mobile
dislocations, leading to a burst of creep deformation. Associated with the
process, multiple or cross-slip will be the nature of dislocation movement.
Under the circumstances, the net increase of creep strain in subsequent static
loading is comparable to or much larger than the pure static creep strain. The
dislocation configuration incurred by cyclic loading is dependent on the
number of cycles, so is creep deformation under subsequent static load.
This cyclic-load induced burst of strain under subsequent static
loading needs certain incubation time. Like strain accumulated during the
burst, the incubation time is dependent on the number of cycles prior to
static
creep loading. With increasing number of cycles of the prior cyclic loading,
the
dislocation cells are getting more stabilized, resulting in increased
incubation
time. Compared to the incubation time, the dependence of accumulated burst
strain on the number of prior load cycles is more complicated. This is because
the latter depends not only on the stability of the dislocation cells, but
also on
the characteristics of the cells, such as the size, the shape of the cells and
the
dislocation density of cell wall, which may change continuously during cyclic
loading.
CA 02417739 2003-O1-30
-18-
No matter the static creep is performed as the first loading event
(Figure 9) or as the second on following cyclic creep deformation (Figure 8),
upon the exhaustion of static creep, further deformation cannot be produced
by alternating the loading conditions, as long as the peak stress is not
increased. It seems that the substructure arising from the exhausted static
creep may be stable enough to prohibit further change in the subsequent
loading.
Example 5: Reduced crack growth rate due to a loading process
producing creep deformation
Figure 12 is a graph of two duplicate crack growth tests showing
the reduced crack growth rate as a result of creep deformation produced by
an over load and hold for 24 hour.
In the first portion of the graph, i.e. for the first 170 hours, the
pipe was subjected to accelerated loading (at R=0.6 and cyclic frequency
f=0.005 Hz) to determine the crack growth rate. Thereafter, the pipe was
treated in accordance with one method of this invention. Thereafter, the pipe
was again subjected to accelerated loading to determine the crack growth rate
after the treatment according to the instant invention.
In accordance with the method of the instant invention, the pipe
specimen was subjected to an overload (an increase in load) about 10%
higher than the peak cyclic stress prior to the overloading and the subsequent
hold (static load) at the overload stress for 24 hours. After overload and
hold,
the test was resumed with the cyclic loading conditions identical to those
prior
to overload and hold.
As shown in Figure 12, process of the instant invention itself did
not cause noticeable crack growth, while the crack growth rate after the
overload and hold is more than one order of magnitude smaller than that
before the overload and hold. It was observed that, when an overload of about
10% was applied without subsequent hold, the growth rate during the
resumed cyclic loading after overloading was not noticeably reduced.
CA 02417739 2003-O1-30
-19-
Figure 13 shows the crack morphology for a specimen with a
30% overload and static hold at the overload stress for 24 hours and
subsequent accelerated crack growth testing at cyclic loading at R =0.6 and
cyclic frequency f=0.005 Hz for 40 days. The growth rate was reduced by two
orders of magnitude after the overload and the static hold. The crack
appeared blunt at the position of overload and hold, and became branched
during propagation under the accelerated cyclic loading after the overload and
hold. This is produced by claimed loading process producing creep
deformation at the crack tip.
