Anda di halaman 1dari 12


Advanced technology in creep life prediction and

damage evaluation for creep strength enhanced
ferritic steels
Fujimitsu Masuyama
Graduate School of Engineering, Kyushu Institute of Technology, 1-1, Sensui-cho,
Tobata, Kitakyushu 804-8550, Japan

Creep damage mechanisms of recently developed boiler and steam-turbine steels are not well clarified,
however, creep life prediction technology for such creep strength enhanced ferritic steels, grades 91,
92 122, etc. with tempered martensitic structure including welds has been strongly demanded by
power plant operators and equipment manufacturers. In this paper, the technologies related to the
creep life prediction and damage evaluation for creep strength enhanced ferritic steels being studied
from the various aspects in Japan will be surveyed and presented. The creep life prediction of this
type of steel must be grounded on the findings on microstructural degradation and creep softening in
martensitic structure composed of very fine martensite lath, block, packet and prior austenite grains,
and precipitation and dislocation structures. Physical properties response to the creep degradation and
measurement and detection of localized creep damageystrength would be useful tools to develop
diagnostic techniques for life prediction of creep strength enhanced ferritic steels. Other important
techniques to support the creep life prediction based on creep-strainyrupture data are creep modelling
and data analysis which have been successfully investigated to date.

Keywords: creep life prediction, damage evaluation, creep strength enhanced ferritic steels, martensitic heat
resistant steels, non-destructive damage detection

INTRODUCTION techniques are required for the appropriate life management

of equipment using these steels and for the achievement of
In fossil-fired thermal power plants, elevated steam tempera- life extension [2]. Although non-destructive damage
ture and pressure conditions are essential for increasing detecting techniques have been developed and extensively
efficiency, while frequent shutdowns and startups and load studied with respect to creep damage in these steels, using
swing operations are presently unavoidable in response to not only base metal and cross-weld standard specimens, but
daily and weekly changes in demand for electric power. Such also full-scale component test rigs in Japan, it is essential to
strenuous conditions require materials for high temperature understand the actual creep damage mechanisms and to
components in power plants having high creep strength, high accurately predict the long-term creep rupture strength at
thermal conductivity, and a low thermal expansion coeffi- 100,000 h and service life. In the present paper, advances in
cient. The creep strength enhanced ferritic steels such as technology related to the creep life prediction and damage
grades 91, 92, 122, etc. developed in the 1980s and 1990s evaluation for creep strength enhanced ferritic steels being
having a capability to respond to these requirements have studied from the various aspects such as creep strength
been introduced for practical application during the last extrapolation methods, microstructures, modelling, and
decade [1]. However, a number of creep failure experiences non-destructive and destructive testing in Japan are
with these steels used in superheateryreheater tubes, hot surveyed.
reheat pipes and headers have been reported. It is essential
to evaluate the material conditions and remaining life of the
materials for long-term use in power plant components, but it 2. PREDICTION OF LONG-TERM CREEP
becomes apparent that these steels exhibit different degrada- RUPTURE LIFE
tion characteristics from conventional Cr Mo steels due to
the fine lath structure of the tempered martensite and its In order to develop design stress values, or allowable stresses
complicated creep degradation mechanism in base metal and in the creep regime, it is necessary to predict 100,000 h creep
welds. Accordingly, it is difficult to apply conventional rupture strength. Also to conduct life assessment of compo-
techniques for the diagnosis of remaining life, and new nent materials, the service life must be accurately known at

234 doi: 10.3184/096034011X13123829786079

Creep life prediction and damage evaluation: Fujimitsu Masuyama

Figure 2 Comparison of creep rupture data of Gr.122 to the

Figure 1 Comparison between creep rupture lives of Gr.91 pre- regression curves (solid lines) obtained from the solid data points
dicted by Larson Miller parameter with best fit constant value of below half the 0.2% proof stress [20].
32 to 38 and value of 20 [7].

the given stress and temperature. To predict the 100,000 h allowable stresses is based upon research results on the stress
creep rupture strength or service life at operating conditions dependence of degradation and creep rupture life of creep
the time-temperature parameter (TTP) technique, for strength enhanced ferritic steels. The studies found a break-
instance, the Larson Miller parametric method is commonly down in the stress-time to rupture diagram at a stress of half
used to extrapolate from shorter and higher temperature the 0.2% proof stress, and the data calculated were separated
databases. In the standard equation, T(C log tR), the at this stress using a different Larson Miller parameter
Larson Miller constant C was 20 when this method was constant such as 20 for the lower stress region (regional
introduced for the aging or tempering of steels, but optimized splitting analysis method) to accurately predict the 100,000 h
values from 10 to 40 have subsequently been found to be strength which should be lower than a stress of half the 0.2%
suitable, depending on the material. A value of approxi- proof stress at the temperatures [13 16].
mately 31 was used for Gr.91 steel [3], and 36 for Gr.92 and As mentioned above in the regional splitting method [7],
Gr.122 steels [4], and most researchers and engineers were in the data set is simply separated at the stress of half the 0.2%
consensus for the use thereof. Recently, however, some proof stress. From another point of view it is proposed that
research results indicate that the parameter constant C must the data set be divided into several regions (multi-regional
be reduced to lower values ~20 from the mathematically analysis method), because the apparent activation energy Q
fitted value of around 31 or 36 for the respective steels at low for creep rupture life changes from a high value in the short-
stress conditions [5,6], because the data obtained under lower term to a low value in long-term, while the conventional TTP
stress and higher temperature conditions do not meet the technique ignores the changes in Q, resulting in the over-
predicted values, due to the unexpected degradation for creep estimation of creep rupture strength [17 19]. Figure 2 [20]
strength enhanced ferritic steels with martensitic structure as shows the comparison of actual creep rupture data points for
shown in Figure 1 [7]. The design stresses for a welded joint Gr.122 against the regression curves (solid lines in Figure 2)
are usually developed by multiplying weld strength reduc- by the regional splitting method using the 0.2% proof stress
tion factors, which are determined as ratio values of cross- reported in the reference for this steel [4]. The dash-two
weld creep rupture strength to base metal rupture strength. dots curve in Figure 2 indicates the critical stress of half the
This means that the overestimated base metal strength gives 0.2% proof stress; however this curve is located above the
a higher design stress for a welded joint. In 2004, a Gr.122 dash dot curve which is determined by the multi-regional
longitudinally seam-welded hot reheat pipe experienced analysis method. The horizontal dotted line also shows the
steam leaking at the welds due to a Type IV failure boundary of upper limit causing the static recovery during
mechanism after 33,000 h service, and in the last decade creep with lower stress exponent n and lower activation
Gr.91 seam-welded hot reheat pipe elbows also leaked after energy Q. This proposed method predicts a more conserva-
approximately 80,000 h service. The design stresses for these tive strength for all temperatures compared with single
steels were derived at that time from the extrapolated region analysis or the regional splitting method, and provides
100,000 h strength predicted by the use of a mathematically a fundamental understanding of creep degradation in terms
best fit parameter constant for all creep rupture data, such as of creep activation energy [21,22].
36 and around 31 for Gr.122 and Gr.91 respectively. The 100,000 h creep rupture strength predicted from
The steam leakage accident with Gr.122 provided an temperature-accelerated creep rupture test data gives design
opportunity to start the recalculation of design stress for all stress or allowable stress multiplying the reciprocal of design
kinds of creep strength enhanced ferritic steels worldwide. factors of 1.5 for average strength, however, in the 2001
Re-evaluation of stresses in Japan was started in 2004 at the edition of the ASME Code, an F average concept was
JAPEIC Committee under the METIyNISA (Nuclear and introduced to apply a multiplier Favg instead of the reciprocal
Industrial Safety Agency) sponsorship [8 12]. The creep of design factors of 1.5, defining the Favg such that
strength prediction method used in the re-evaluation of these log Favg 1yn (n is the stress exponent or the slope of MATERIALS AT HIGH TEMPERATURES 235

Creep life prediction and damage evaluation: Fujimitsu Masuyama

Figure 3 Temperature dependence of the ratio sR (105) to sR (106) Figure 4 Creep damage progress and damage detecting methods.
and sR (104) to sR (105) [23].

the log time to rupture versus log stress plot at 100,000 h). microstructural observation. Recently, microstructural
Assuming that the Favg is the ratio of the creep rupture changes have been simulated employing the free energy
strength at 1,000,000 h (sR 106 ) to that at 100,000 h concept as a thermodynamic model for steels [24 26] and
(sR 105 ), Favg and the ratio sR (104) to sR (105) are Ni-based alloys [27]. Figure 5 shows the change in reduced
calculated against temperatures as shown in Figure 3 for system energy, composed of a chemical free energy, an
Gr.91 [23]. From Figure 3 it is evident that the maximum interfacial energy and elastic strain energy [28]. The
temperature of design factor of 1.5 is applicable for the system energies are reduced with creep time in both Gr.91
welds at 540 C for the case of both parameter constants of 20 and Gr.92 towards the saturated values, which are higher
and 36. than those of the equilibrium state, and it is seen that there
are no marked differences between the grip portion and
gauge portion for Gr.91; and there is a larger drop in
3. MICROSTRUCTURAL DEGRADATION DURING energy for Gr.91 than Gr.92 due to the relatively greater
CREEP structural degradation. These results suggest that structural
simulation would be helpful in assessing the material condi-
3.1 Feature of microstructural degradation in creep tions and creep damage.
strength enhanced ferritic steels

Creep strength enhanced ferritic steels use primarily Cr, Mo 3.2 Precipitation behaviour in creep strength enhanced
andyor W, V, Nb and N as the alloying elements, and since ferritic steels
they are tempered martensite, the as-received hardness is
around 220 HV. This hardness declines by only 5% when M23C6 and Laves phase can be observed as precipitates at the
subjected to thermal aging for 30,000 h at 650 C, and tempered martensite grain boundaries using an optical
declines by 20 30% due to creep after 10,000 to 30,000 h microscope, and TEM allows observation of MX as well.
exposure associated with microstructural degradation. MX is extremely fine, and does not show any major change
However, it is very difficult to determine the creep degrada- in particle size resulting from thermal aging or creep. In
tion and material conditions through the optical microstruc- contrast, M23C6 and Laves phase show coarsening due to
ture in the case of creep strength enhanced ferritic steels, thermal aging and creep [29]. There are four kinds of grain
while for conventional Cr Mo steels with ferrite and pearlite boundaries in the structure, i.e., prior austenite, packet, block
structures, or coarse-grained martensiticyfine-grained
bainitic structures for base metal or heat affected zones
respectively, their degradation rating and creep damage
percentage can be determined. Figure 4 illustrates the creep
damage progress for Cr Mo steels and creep strength
enhanced ferritic steels as well as the damage detecting
method applied to the first half of creep life. The micro
cracks and macro cracks in the latter half of stage III and
stage IV of the creep life are detectable by non-destructive
testing and replication techniques. In the case of Cr Mo
steels, microstructural changes (stage I) under the optical
microscope followed by creep cavitation processes (stage II)
can be classified against the creep damage rating, but both
stages in creep strength enhanced ferritic steels are not
distinguishable, and creep damage rating is impossible by Figure 5 Changes in reduced system free energy with time [28].


Creep life prediction and damage evaluation: Fujimitsu Masuyama

strength enhanced ferritic steels, it is said that fine MX

carbo-nitride particles strongly improve creep resistance by
precipitation hardening in the matrix [1]. Therefore the
precipitation behaviour of MX during the heat treatment
process and creep exposure has extensively been investigated
as well as Z phase formation from M2X or Cr(Nb, V)N after
long-term creep exposure [32 34]. Through the studies on
the effect of tempering temperature on the precipitation and
creep strength, M2X particles in the steel tempered at 680 C
disappeared during creep for about 20,000 h, being accom-
plished by the formation of VX which contributed to
maintain higher creep strength without any formation of Z
phase. On the other hand, Z phase formation was observed in
Figure 6 Effect of creep exposure on precipitates diameter dis-
tribution for Gr.91 [31]. the steels tempered at 730 C and 765 C after 12,000
20,000 h creeping conditions, this promoting the creep
degradation [35].

and lath, while prior to the creep exposure, M23C6 is seen at

all grain boundaries and MX is observed at the lath grain 3.3 Stability of martensite lath structures
boundaries and within the lath. After creep exposure, new
Laves phase precipitation is seen at the grain boundaries. It is The creep strength enhanced ferritic steels are composed of
known that the sizes of these precipitates change due to very fine martensite lath structures with approximately 1 mm
heating and creep. According to particle size distribution width at normalized and tempered conditions, and the
measurement results [30] for MX, M23C6 and Laves phase morphology of lath structures correlates with the crystal-
on the grip and gauge potions of 10Cr 1Mo 1W V Nb lography. The morphology and crystallography of martensite
steel, average values on the grip portion are approximately lath structures have been extensively studied by use of
30 nm for MX, 80 nm for M23C6 and 350 nm for Laves metallographic instruments [36]. Particularly EBSD (electron
phase, while on the gauge portion these are approximately backscattering diffraction) technique has recently become a
50 nm, 130 nm and 400 nm, respectively. The distributions strong tool to reveal the crystallographic changes of marten-
are generally normal, with M23C6 demonstrating the greatest site lath structures during creep, as well as the conventional
decline in peak height, and with the peak particle size for techniques of TEM and SEM, because the crystallographic
MX and Laves phase moving in the direction of increase. orientation of the martensitic structures changes with
Looking at the precipitate size distribution measurement recovery and creep deformation. Measuring the mis-orienta-

results [30] in the case of thermal aging at 600 C, Laves tion angles is a way to know the information about grain
phase precipitation is not seen prior to thermal aging, with boundary changes due to creep degradation. Figure 7 [37]
distribution occurring due to the heating. With respect to the shows the changes in lath boundary length per unit area in
Laves phase size distribution, the peak location moves terms of creep strain under several creep conditions and
rapidly towards the maximum in conjunction with heating TEM structures observed at tytR 0:2 and 0.6 of crept
time. Specifically, the average size (that is 70 nm at around conditions. At the beginning of creep, the length of the lath
1,000 h) changes to 400 nm at around 33,000 h. In boundary increases a little corresponding to the generation of
contrast, the average sizes of MX and M23C6
(approximately 50 nm and 80 nm respectively),
do not exhibit any major changes due to heating.
From the above, Laves phase undergoes consider-
able coarsening even without the stress, while it
can be seen that growth without stress for MX and
M23C6 is quite limited, and that even the effect of
stress on MX is quite small. In this context, M23C6
appears particularly sensitive to the action of stress.
Figure 6 [31] presents the precipitate size distribu-
tions before and after creep (interrupted at 0.9 of
tytR ) performed on Gr.91. There is no change in
the average particle size for all precipitates, and
only the precipitates in excess of 100 nm are seen
to experience coarsening. Precipitates in excess of
100 nm correspond to M23C6 and Laves phase, and
it is considered that coarsening of Laves phase is
due to heating during creep, while that of M23C6 is
due to the action of stress during creep exposure. Figure 7 Relationship between lath boundary length and creep strain for Gr.91
Among the precipitates appearing in the creep [37]. MATERIALS AT HIGH TEMPERATURES 237

Creep life prediction and damage evaluation: Fujimitsu Masuyama

Figure 8 Relationship between lath width and creep. Figure 9 Relationship between lath width and strain for Gr.91 [39]
minimum creep rate for Gr.91 [38].

new boundaries within the lath shown in the TEM graph (a),
then it linearly decreases with creep strain due to the away and reach the lath boundary, thus being the direct cause
recovery of lath structure with subgrain growth (TEM of the boundary movement. The lath width would be
graph (b)). inversely proportional to the load stress (lath width being
A TEM photograph [38] in Figure 8 [39] indicates the greater with lower stress), and as a result the lath structure or
structure of Gr.91 with 25% tensile strain at 820 C. subgrains that had grown would have a major effect on
Normalized and tempered as-received material has an extre- hardness. Especially in the case of ruptured material, it is
mely high dislocation density, with an approximately 1 mm predicted that hardness will decline with lower stress. With
width lath structure, but deformation at high temperatures respect to the lath width distribution as well, while a normal
results in a dramatic lowering of the dislocation density, and distribution applies to cases of high stress above approxi-
substantial growth in lath size can be seen. Also, while it is mately 100 MPa, the distribution in the case of low stress is
thought that precipitates originally exist at the lath bound- characterized by two maxima, and mixing of coarse and fine
aries or on other grain boundaries, the photograph clearly grains is reported [41]. Hardness, on the other hand, is
shows that these are dispersed within the grains, and the proportional to the square root of the dislocation density
alignment of precipitates suggests that they had been at the remaining after creep deformation, such that a relationship
original grain boundaries. This type of structure is also can be established between hardness and lath width.
observed after creep [40], and it is considered to be caused Accordingly, because lath growth behaviour differs
by movement of the lath boundaries or subgrain boundaries depending on stress, it is predicted that differences will
during creep. It is already known that the grain size of this also occur in terms of changes in hardness. As a result of
lath (inclusive of subgrains) is related to creep resistance, and investigations [40] into the influence of lath width and
as indicated in Figure 9 [38], creep resistance declines in dislocation density on the hardness of Gr.91, it has been
proportion to the cube of the lath width. The lath width is clarified that increased lath width and reduced dislocation
also strain-dependent, and as shown in Figure 8 [39], this density cause hardness to decline. As lath width and disloca-
increases proportionately with strain until the strain reaches tion density are directly related to creep resistance, measure-
approximately 0.1. It appears that necking occurs in cases ment of hardness can be understood to enable estimation of
where the strain is above ~0.1, and, although the lath width is material creep resistance.
shown as a constant value in order to separate the location of
structural observation from the necking area, it is considered 4. DIAGNOSTIC AND DAMAGE DETECTING
that there is actually a wide range in which the proportional TECHNIQUES FOR CREEP LIFE PREDICTION
relationship holds. It is clear that such lath boundary move-
ment or grain growth is induced by the action of stress and 4.1 Applicability of replication technique
the resulting strain. Based on considerations [39] of this
process, precipitates are strongly bound, with the dislocation It is extremely difficult in the case of creep strength enhanced
networks (serving to increase internal stress) moving as a ferritic steels with a martensitic structure to discriminate
result of the action of stress and reaching the lath boundaries. structural changes corresponding to creep life fraction. As
When these networks are swept by the movement of the lath these steels feature a fine martensite structure, such changes
boundaries, it is thought that they are absorbed, thus cannot be ascertained on an optical microscopic level.
increasing the interface energy. Considering the growth of Furthermore, precipitate distributions obtained from extrac-
the lath structure, given some form of change to the tion replicas do not indicate meaningful changes with respect
precipitate or interaction between the precipitate and the to life fraction, and, because influence is sustained from the
dislocation to which it is bound, the dislocation could shift conditions associated with taking extraction replicas, they are


Creep life prediction and damage evaluation: Fujimitsu Masuyama

Figure 10 Microstructural observation of replica taken from OD surface of internal pressure creep test tube [42].

currently difficult to utilize in life assessment. Accordingly, 4.2 Potential drop method
attention has focused on specific precipitates such as M23C6,
and new approaches such as EBSD analysis [37] using Using an alternating current electrical resistance measure-
samples taken from components as mentioned above are ment apparatus, the difference in potential was determined
needed in order to track changes in these structures. In between the surface of Gr.91 large-scale cross-weld creep
contrast, it has proven possible to observe creep cavities test specimens with the cross-section of 30 mm square [42]; a
using the replication method for components such as super- frequency of 97 Hz was used to determine the difference in
heater tubes, whose surfaces are subjected to multi-axial potential between a damaged location (heat affected zone)
stress conditions, and the resulting cavity numbers clearly and a non-damaged location (base metal). Figure 12 [42]
indicate changes in comparison with life fraction. On the shows the relationship between the potential drop ratio
other hand, almost no cavities are observed on the surface of between the damaged location and non-damaged location,
welded joints, and changes with exposure time are not and life fraction as determined from measurement results
clearly evident; however sub-surface damage takes place comparing with conventional Gr.22 Cr Mo steel: a clear
within the wall of welds. Figure 10 [42] represents observa- relational diagram was obtained for both steels. Thus, by
tions using the replication method of creep cavities on the measuring the potential drop ratio, and using the relational
surface of a Gr.91 tube internally pressurized for a creep test, diagrams obtained for the potential ratio and the life fraction,
with the corresponding plot of cavity density number in life assessment is possible. However the life fraction depen-
Figure 11 [42]. Clear correlation is seen between this number dency of potential drop ratio for Gr.91 is less than that for
density of cavities and life fraction, and this type of figure Gr.22.
can be used for life assessment of superheateryreheater tubes.

Figure 11 Number density of cavity versus creep versus creep life

fraction life fraction curve for Gr.91 [42]. Figure 12 Potential drop ratio curve for Gr.91 [42]. MATERIALS AT HIGH TEMPERATURES 239

Creep life prediction and damage evaluation: Fujimitsu Masuyama

4.4 Measurement of lattice defect response to creep


The positron annihilation method has a capability to measure

the atomic-scale defect in solids, and is used for in-situ
measurements of the thermal equilibrium defect in some
metallic materials [45]. Recently this technique has been
applied to fatigue damage evaluation by measuring positron
lifetime [46], and also to creep damage evaluation devel-
oping equipment for high performance and accurate positron
annihilation life monitoring combined with a high tempera-
ture creep test machine (Shirai, Y. unpublished work). The
equipment has the capability of measuring the changes in
dislocation density and lattice defect due to very fine particle
precipitation during the creep exposure of creep strength
enhanced ferritic steels. The hydrogen desorption profile
analysis method has been developed to evaluate material
conditions and creep damage for creep strength enhanced
ferritic steels [47]. This is basically a technique to use
hydrogen as a tracer of changes in lattice defects including
Figure 13 Anodic polarization curve and relation between peak
precipitation in the material. The steels aged or crept at high
current density versus creep life fraction [44].
temperature are charged with hydrogen by cathodic electro-
lysis and then subjected to thermal desorption spectroscopic
analysis. Experimental results revealed that the amount of
hydrogen desorbed CH , decreased monotonously as ageing
4.3 Electrochemical method proceeded and there was a relatively good correlation
between CH and hardness. It was suggested that the
An example of the conditions for electrochemical polariza- changes in decomposed profiles due to ageing might reflect
tion measurements is introduced as follows [43]. The the dislocation density conditions and precipitation of
measurement surface of specimens is polished to a M23C6, etc. [48].
0.1 mm diamond finish. The surface area exposed to the
test solution is 0.3 to 0.5 cm2 and the remaining surface is 4.5 Hardness method and model for creep life prediction
electrically insulated using paraffin. The electrochemical
polarization measurements are performed in a 1 M-KOH Consideration thus far clearly indicates that hardness drop is
solution at a temperature of 30 + 1 C. The potential of the due to creep, and that the amount of decline is strongly
specimens is scanned from the rest potential ( 400 mV related to life consumption or creep degradation, and that the
versus SCE) at a sweep rate of 0.5 mVys to obtain anodic hardness of tempered martensitic structure is influenced by
polarization curves. In this case, it was seen that the peak the structural conditions of grain boundaries, subgrain size,
current density Ip for Gr.92 increased with thermal aging precipitates, dislocation density and creep strain [48 50].
and creep, and corresponded to the selective dissolution The effect of these microstructural parameters upon hardness
volume of M23C6 and Laves phase. Figure 13 [44] shows is investigated by use of not only macro-ymicro-hardness but
electrochemical measurement results for Gr.122 under also on the nano-indentation technique [51,52]. The creep
similar test conditions indicating two peak current densities strength enhanced ferritic steel in the present context experi-
at around  250 mV and 250 mV, respectively, and ences reduced hardness due to thermal aging, but the amount
increase of these peaks in terms of creep life fraction. of decline is small, and is actually extremely large under the
However there is no effect of stress for the Ip1 which action of stress. That is, reduced hardness in the crept
should be the thermally dominant factor, in contrast the materials is based on structural changes dependent upon
effect of stress is significantly observed for Ip2 if the data stress, and there is no doubt that the assessment of life
are compared between grip portion and gauge portion. from hardness measurement values has a firm basis in
According to other examples of test results for an materials science. The hardness drop measured for Gr.91,
internally pressurized creep test specimen and a cross- both thermally aged and creep-interrupted for base metal and
weld creep specimen [42], and as correlation was recog- welds (minimum hardness in the heat affected zone), taking
nized between peak current density Ip and creep life the Larson Miller parameter as a variable, can be demon-
fraction with respect to base metal, this method was strated as shown in Figure 14 [50]. It is found from Figure 14
considered to be applicable for life assessment. It should that, for the hardness data from the higher stress tests at
be noted, however, that, because there is inhomogeneous 98 MPa and above, there is scatter at relatively lower values
structural change in the weld heat affected zone, there will of the Larson Miller parameter, while for the data from the
be many cases in which there is not a consistent structure lower stress tests, there is scatter at higher parameter values,
within the measurement region. Thus, it will be difficult to thus being divided into two parts (including the weld data).
apply this method to welded joints. The reason why the hardness versus Larson Miller para-


Creep life prediction and damage evaluation: Fujimitsu Masuyama

Figure 14 Relation between hardness drop and hardness ratio and

Larson Miller parameter [51]. Figure 15 Relation between creep life fraction [51].

meter plot can be divided into two parts is considered to be measured are input to calculate the creep life fraction
(i) strain-oriented softening for the higher stress test and (ii) consumed.
stress-oriented softening for the lower stress test, without any
tytR 1y0:150:98  HyH0
dependence of temperature and time. The hardness measured 
on the grip portion corresponds well with that of the aged 1y0:15f0:98  Hgy H expKs6T20 log t
specimens. The hardness (H) measured for the creep 4
specimen linearly reduced with the creep life fraction, and
Using this hardness model, creep lives of creep specimens
there is a linear relationship between hardness and creep life
were predicted, and compared with the observed creep life
fraction of 0.2 to 0.9 for base metal and welds. When the
fractions, and it was found that predicted lives are longer
initial hardness (H0) for base metal and welds is defined as
than the observed ones in an approximate range of factor of 2
the hardness of the grip portion of the base metal and that of
for high stress softening and factor of 3 for low stress
the base metal of the welded joint respectively, the hardness
softening [50].
ratio is shown as being linear against the creep life fraction as
indicated in Figure 15 [50]. The hardness ratio (HyH0) can
be expressed as below. 4.6 Simplified tertiary creep modelling method

In order to assess the creep damage or creep life, a tertiary

HyH0 0:98  0:15tytR 1 creep modelling method, the so-called Omega method [53] is
adopted by API 579, and the logarithm of the strain rate is
where tytR is creep life fraction. well expressed by the following equation over a wide range
The hardness drop (DH) due to creep and thermal aging of tertiary creep, irrespective of stress and temperature.
can be demonstrated as a function of the Larson Miller
ln e ln e 0 Oe or e e 0 exp Oe 5
parameter as shown in Figure 14. Three lines represent the
hardness drop due to creep or ageing for the high stress creep where e is the creep rate, e0 is the initial creep rate, and O is
test, low stress creep test, and thermal aging test respectively. the strain rate acceleration factor (dlneyde).
The logarithmic hardness drop is expressed by following When the duration of primary creep is short enough, the
equation. creep life and creep strain can be expressed by the terms O
and e0 , and the remaining creep life is expressed by
ln DH ln DH0 Ks6LMP 2 following equation.
tR  ti 1yOei 6
where DH is the hardness drop; DH0 is the initial hardness Therefore to obtain the remaining life, the values of O and
drop; Ks is the coefficient for high stress (98 MPa and the instantaneous true creep strain rate ei are needed at the
115 MPa), low stress (71 MPa) and ageing, LMP: given time ti , however these values are sometimes difficult to
T(20 log t); T is the absolute temperature in K; t is the be determined. In order to simplify the remaining life
time in hours. ln DH0 is assumed to be 0. Therefore Eqn (2) prediction by the Omega method it was considered that the
can be expressed in the following form. instantaneous true creep strain rate ei could be estimated
from the average creep rate, eavg , which should be the ratio of
ln DH KsLMP Ks6T20 log t 3 the creep deformation to the operating duration, and thus the
parametric expression of the value of O could be developed.
As the creep life fraction tytR is given by Eqn (4), the initial In this sense, ei is expressed by aeavg (a is the coefficient),
hardness H0 is not necessary, but only the temperature and and O is expressed by the parametric function of stress and
operating hours as well as the hardness value instantaneously temperature as mentioned below [50]. To express the para- MATERIALS AT HIGH TEMPERATURES 241

Creep life prediction and damage evaluation: Fujimitsu Masuyama

3 mm diameter and 0.25 mm thickness for creep tests [57].

Figure 16 [57] shows the results of the small punch creep
rupture test of the base metal plotted in terms of the Larson
Miller parameter compared with data for a uniaxial standard
creep test. Correlating between the load and the stress to
convert the result of the small punch creep test into that of a
standard creep test, the ratio of load to stress was calculated
as 0.43 so that the curves for the small punch creep test and
standard creep test overlapped each other [58].


The creep strength enhanced ferritic steels, grades 91, 92

Figure 16 Results of mall punch creep test and uniaxial standard
creep test [58].
122, etc. with a tempered martensitic structure including
welds has been extensively used in fossil-fired power plant
worldwide for the last decade or so, and development of
creep life prediction technology is strongly demanded
metric expression of the value of O, the Larson Miller
because of several experiences of premature creep failures
parameter is employed and the time to rupture in the
with these steels. Technologies related to creep life predic-
equation is converted to KMG em (the Monkman Grant
tion and damage evaluation for these types of steel are being
relationship), where KMG is the Monkman Grant constant,
studied from several aspects in Japan (such as creep strength
and em is the minimum creep rate. Then since the minimum
extrapolation methods, microstructures, modelling, and non-
creep rate em is expressed by Nortons law i.e. em Asn ,
destructive and destructive testing) and each has been
KMGem becomes KMG Asn . Accordingly the temperature-
surveyed. As a result, it was recommended that the creep
stress parameter, F T; s) is expressed by the following
life prediction of creep strength enhanced ferritic steel must
   be grounded on the findings of microstructural degradation
FT; s T C log KMG yAsn 7 and creep softening in martensitic structures, composed of
very fine martensite laths, block, packet and prior austenite
where C is the parameter constant.
grains, and precipitation and dislocation structures.
The magnitude of the O value increases with decreasing
Mechanical, physical and chemical property response (hard-
both the stress and temperature at temperatures of 500 C and
ness, potential drop, electrochemistry, lattice defect, cavita-
above. The O values can be expressed in terms of the
tion, etc.) to creep degradation, and measurement and
temperature-stress parameter F T; s, and the linear rela-
detection of localized creep damageystrength would be
tionship between both the O value and F T; s estimates
useful tools to develop diagnostic techniques for life predic-
the O value thus predicting the remaining creep life.
tion of creep strength enhanced ferritic steels. Creep life
Accordingly, the deformation record of the equipment for
prediction based on creep strain modelling and creep rupture
given operating hours, and the data bases of the material used
data (including small punch creep test data) analysis would
(such as the Monkman Grant constant and Nortons stress
also be promising.
exponent as well as the operating stress and temperature)
simply provide the remaining creep life.
4.7 Advanced small punch creep test method
[1] Masuyama, F. (2001) History of power plants and progress in
heat resistant steels. ISIJ Int., 41(6), 612 625.
The above mentioned techniques to diagnose and evaluate
[2] Masuyama, F. (2003) Creep rupture behavior and life assess-
damage in materials for creep life prediction are non- ment of advanced 9-12%Cr steel weldments for power appli-
destructive, and the results obtained should commonly be cations. Proc. 2nd Int. Conf. on Integrity of High Temperature
compared with the results from destructive creep tests to Welds, pp. 373 380. London, November 10 12, 2003.
verify the accuracy of prediction. However, the destructive [3] Brinkman, C.R., Alexander, D.J. and Maziasz, P.J. (1990)
tests using standard specimens are costly, due to the cutting Modified 9Cr-1Mo steel for advanced steam generator appli-
of components for sampling. The small punch creep tests can cations. ASME/IEEE Power Generation Conference, 90-JPGC/
contribute to direct life prediction of the subjected material NE-8.
andyor localized interest area, for example the Type IV [4] Masuyama, F. (1995) ASME code approval for NF616 and
failure region of a weld heat affected zone [54,55] using a HCM12A. In: Metcalfe, E. (ed.), New steels for advanced
plant up to 620 C, pp. 98 113. PicA, Drayton.
small specimen sampled from the components by a newly
[5] Kimura, K., Kushima, H. and Sawada, K. (2003) Long-term
developed and unique machine [56]. Generally the small creep prediction of high Cr ferritic creep resistant steels based
punch testing method is conducted using miniature speci- on degradation mechanisms. Proc. 6th Int. Charles Parsons
mens in the form of a thin plate of dimensions 10 mm by Turbine Conf., IOM.
10 mm with 0.5 mm thickness; however, recently an [6] Kimura, K., Sawada, K., Kubo, K. and Kushima, H. (2004)
advanced technique of small punch testing has been devel- Influence of stress on degradation and life prediction of high-
oped to use an even further miniaturized specimen with strength ferritic steels. ASME PVP, 476, 11 18.


Creep life prediction and damage evaluation: Fujimitsu Masuyama

[7] Kimura K., Kushima H. and Abe F. (2002) Degradation and [24] Murata, Y., Koyama, T., Morinaga, M. and Miyazaki, T.
assessment of long-term creep strength of high Cr ferritic (2002) Prediction of the Laves phase morphology in Fe-Cr-
creep resistant steels. Int. Conf. on Advances and Optimization W-C quaternary steels with the aid of system free energy
of Fossil Power Plants, Orland, Florida, March 11 13, 2002. concept. ISIJ Int., 42, 1423 1429.
[8] Kimura, K. (2005) Assessment of long-term creep strength [25] Kunieda, T., Nakai, M., Murata, Y., Koyama, T. and Morinaga,
and review of allowable stress of high Cr creep resistant steels. M. (2005) Estimation of the system energy of martensite phase
Proc. of PVP2005, PVP2005-71039, Denver Colorado, July, in an Fe-Cr-C ternary alloy. ISIJ Int., 46, 1909 1914.
2005. [26] Kunieda, T., Akada, K., Murata, Y., Koyama, T. and Morinaga,
[9] Kimura, K. (2006) Creep strength assessment and review of M. (2007) Estimation of the system energy of the lath
allowable tensile stress of creep strength enhanced ferritic martensite phase in high Cr ferritic steels. Adv. Mater. Res.,
steels in Japan. Proc. of PVP2006-ICPVT-11, PVP2006- 15 17, 690 695.
ICPVT-11-93294, Vancouver, Canada, July, 2006. [27] Tsukada, Y., Murata, Y., Koyama, T. and Morinaga, M. (2008)
[10] Tabuchi, M. and Takahashi, Y. (2006) Evaluation of creep Phase-field simulation on the formation and collapse processes
strength reduction factors for welded joints of modified 9Cr- of rafted structure in Ni-base super alloys. Mater. Technol., 49,
1Mo steel (P91). Proc. of PVP2006-ICPVT-11, PVP2006- 484 488.
ICPVT-11-93294, Vancouver, Canada, July, 2006. [28] Saito, Y., Kita, T., Murata, Y., Morinaga, M., Koyama, T.,
[11] Takahashi, Y. and Tabuchi, M. (2006) Evaluation of creep Hasegawa, Y. and Igarashi, M. (2009) Proc. 47th Symp. on
strength reduction factors for welded joints of HCM12A Strength of Materials at High Temperatures, pp. 81 85.
(P122). Proc. of PVP2006-ICPVT-11, PVP2006-ICPVT-11- [29] Hofer, P., Cerjak, H., Schaffernak, B. and Warbichler, P. (1998)
93294, Vancouver, Canada, July, 2006. Quantification of precipitates in advanced creep resistant 9-
[12] Yoshida, K., Nakai, H. and Fukuda, M. (2006) Regulatory 12%Cr steels. Steel Res., 69, 343 348.
review results on allowable tensile stress values of creep [30] Cerjak, H., Hofer, P. and Schaffernak, B. (1999) The influence
strength enhanced ferritic steels. Proc. of PVP2006-ICPVT- of microstructural aspects on the service behaviour of
11, PVP2006 -ICPVT-11-93294, Vancouver, Canada, July, advanced power plant steels. ISIJ Int., 39, 874 888.
2006. [31] Masuyama, F., Nishimura, N. and Sasada, A. (1998) Precipita-
[13] Kimura, K., Sawada, K., Kubo, K. and Kushima, H. (2004) tion behavior during creep of modified 9Cr-1Mo steel. CAMP-
Influence of stress on degradation and life prediction of high ISIJ, 11, 1245.
strength ferritic steels. Proc. PVP2004, PVP, Vol. 476, pp. [32] Yoshino, M., Mishima, Y., Toda, Y., Kushima, H., Sawada, K.
11 18. and Kimura, K. (2005) Phase equilibrium between austenite
[14] Kimura, K., Kushima, H., Sawada, K. and Toda, Y. (2007) and MX carbonitride in a 9Cr-1Mo-V-Nb steel. ISIJ Int., 45,
Region splitting analysis on creep strength enhanced ferritc 107 115.
steels. Proc. CREEP 8, CREEP2007-26406, pp. 647 654. [33] Suzuki, K., Kumai, S., Toda, Y., Kushima, H. and Kimura, K.
[15] Kimura, K., Sawada, K., Kushima, H. and Toda, Y. (2007) (2003) Two-phase separation of primary MX carbonitride
Stress dependence of degradation and creep rupture life of during tempering in creep resistant 9Cr1MoVNb steel. ISIJ
creep strength enhanced ferritic steels. Proc. 5th Int. Conf. Int., 43, 1089 1094.
Advances in Materials Technology for Fossil Power Plants, [34] Sawada, K., Kushima, H., Kimura, K. and Tabuchi, M. (2007)
EPRI, Marco Island, Florida. TTP diagram of Z phase in 9-12%Cr heat-resistant steels. ISIJ
[16] Kimura, K., Sawada, K., Kushima, H. and Kubo, K. (2008) Int., 47, 723 739.
Effect of stress on the creep deformation of ASME grades [35] Sawada, K., Suzuki, K., Kushima, H., Tabuchi, M. and
P92/T92 steels. Int. J. Mater. Res., 99, 395 401. Kimura, K. (2008) Effect of tempering on Z phase formation
[17] Maruyama, K., Baba, E., Yokokawa, K., Kushima, H. and and creep strength in 9Cr-1Mo-V-Nb-N steel. Mater. Sci. Eng.,
Yagi, K. (1994) Errors of creep rupture life extrapolated by A480, 554 563.
time-temperature parameter methods. Tetsu Hagane, 80, 336 [36] Morito, S., Huang, X., Furuhara, T., Maki, T. and Hansen, N.
341. (2006) The morphology and crystallography of lath martensite
[18] Maruyama, K., Armaki, H.G. and Yoshimi, K. (2007) Multi- in alloy steels. Acta Mater., 54, 5323 5331.
region analysis of creep rupture data of 316 stainless steel, Int. [37] Mitsuhara, M., Morioka, S., Hata, S., Ikeda, K. and Naka-
J. Press. Vess. Pip., 84, 171 176. shima, H. (2009) Analysis of creep degradation behavior in lath
[19] Maruyama, K. (2008) Fracture mechanism map and funda- martensite based on crystallographic orientation changes.
mental aspects of creep fracture in creep-resistant steels, pp. Report of the 123rd Committee on Heat-resisting Metals and
350 364. Woodhead Publishing, Cambridge, UK. Alloys, Japan Society for the Promotion of Science, 50, 37 43.
[20] Maruyama, K. and Yoshimi, K. (2007) Methodology of creep [38] Masuyama, F. and Nishimura, N. (1994) Effect of structures
data analysis for advanced high Cr ferritic steel. Proc. CREEP and residual elements on creep behavior of modified 9Cr-1Mo
8, CREEP 2007-26150, pp. 631 636. steel. In: Oikawa, H. et al. (eds), Strength of materials, pp.
[21] Armaki, H.G., Maruyama, K., Yoshizawa, M. and Igarashi, M. 657 660.
(2008) Prevention of the overestimation of long-term creep [39] Endo, T., Masuyama, F. and Park, K.-S. (2002) Change in
rupture life by multi-region analysis in strength enhanced high hardness and substructure during creep of mod.9Cr-1Mo steel.
Cr ferritic steels. Mater. Sci. Eng., A490, 66 71. Tetsu Hagane, 88, 526 533.
[22] Maruyama, K. and Armaki, H.G. (2009) Cr concentration [40] Sawada, K., Maruyama, K., Hasegawa, Y. and Muraki, T.
dependence of overestimation of long term creep life in (2000) Creep life assessment of high chromium ferritic steels
strength enhanced high Cr ferritic steels. Proc. ECCC Creep by recovery of martensitic lath structure. Key Eng. Mater.,
Conf., pp. 543 552. April 21 23, 2009, Zurich. 171 174, 109 114.
[23] Masuyama, F. (2005) Creep rupture life and design factors for [41] Suzuki, K., Kumai, S., Kushima, H., Kimura, K. and Abe, F.
high strength ferritic steels. Proc. Creep and Fracture in High (2000) Heterogeneous recovery and precipitation of Z phase
Temperature Components Design and Life Assessment during long term creep deformation of modified 9Cr-1Mo
Issues, pp. 983 996. steel. Tetsu Hagane, 86, 550 557. MATERIALS AT HIGH TEMPERATURES 243

Creep life prediction and damage evaluation: Fujimitsu Masuyama

[42] Masuyama, F., Fujita, M., Maruyama, M. and Endo, M. (2002) [50] Masuyama, F., Tokunaga, T., Shimohata, N., Yamamoto, T.
Development of creep life assessment technology for and Hirano, M. (2009) Comprehensive approach to creep life
mod.9Cr-1Mo steel. EPRI/DOE Conf. Advances in life assess- assessment of martensitic heat resistant steels. Proc. ECCC
ment and optimization of fossil power plants, Orlando, Florida. Creep Conf., 31 42.
[43] Komazaki, S., Kishi, S., Shoji, T., Kumazaki, T., Higuchi, K. [51] Sawada, K., Miyahara, K., Kushima, H., Kimura, K. and
and Suzuki, K. (2002) Influence of pre-aging on creep rupture Matsuoka, S. (2005) Contribution of microstructural factors
strength of tungsten alloyed 9%Cr ferritic steel and creep to hardness change during creep exposure in mod.9Cr-1Mo
damage evaluation by electrochemical method. JSME Int. J., steel. ISIJ Int., 45, 1934 1939.
45, 30 38. [52] Jang, J., Shin, S., Komazaki, S. and Honda, T. (2007) A
[44] Abe, Y., Komazaki, S., Mitsueda, T. and Kohno, Y. (2010) nanoindentation study on grain-boundary contributions to
Change in electrochemical property of 12%Cr ferritic steel strengthening and aging degradation mechanisms in advanced
in 1M-KOH solution due to creep. Adv. Mater. Res., 89 91, 12Cr ferritic steel, J. Mater. Res., 22, 175 185.
1 5. [53] Prager, M. (2000) The omega method an engineering
approach to life assessment. J. Press. Vess. Technol. Trans.
[45] Shirai, Y. (1998) Positron annihilation method: thermal and
ASME, 122, 273 280.
non-thermal defects refractory intermetallics. High Temp.
[54] Komazaki, S., Hashida, T., Shoji, T. and Suzuki, K. (2000)
Mater. Process., 12, 57 67.
Development of small punch tests for creep property measure-
[46] Kawaguchi, Y. and Shirai, Y. (2002) Fatigue evaluation of ment of tungsten-alloyed 9%Cr ferritic steels. J. Test. Evaluat.,
Type 316 stainless steel using annihilation line shape analysis ASTM, 28, 249 256.
and b  g coincidence positron lifetime measurement. J. [55] Komazaki, S., Sugimoto, T., Hasegawa, Y. and Kohno, Y.
Nucl. Sci. Technol., 39, 1033 1040. (2007) Damage evaluation of a welded joint in a long-term
[47] Honda, T., Komazaki, S., Sugimoto, T. and Kohno, Y. (2007) service exposed boiler by using a small punch creep test. ISIJ
Microstructure evaluation of high chromium ferritic steels Int., 47, 1228 1233.
based on hydrogen desorption profile analysis. Adv. Mater. [56] Kusumoto, J., Kanaya, A., Kobayashi, T. and Miyamoto, T.
Res., 15 17, 684 689. (2007) Residual life assessment of service-aged fossil power
[48] Endo, T., Masuyama, F. and Park, K.S. (2003) Changes in plant equipment by small punch creep test. Therm. Nucl.
Vickers hardness and structure during creep of mod. 9CR- Power, 58, 1123 1129.
1Mo steel. Mater. Trans., 44, 239 246. [57] Komazaki, S., Kato, T., Kohno, Y. and Tanigawa, H. (2009)
[49] Masuyama, F. (2009) Hardness model for creep life assess- Creep property measurements of welded joint of reduced-
ment of high strength martensitic steels. Mater. Sci. Eng., activation ferritic steel by the small-punch creep test. Mater.
A510 511, 154 157. Sci. Eng., A510 511, 229 233.


Copyright of Materials at High Temperatures is the property of Science Reviews 2000 Ltd. and its content may
not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.