Applied Surface Science 359 (2015) 73–81

Contents lists available at ScienceDirect

Applied Surface Science

journal homepage: www.elsevier.com/locate/apsusc

Fractal and probability analysis of creep crack growth behavior

in 2.25Cr–1.6W steel incorporating residual stresses
Mengjia Xu, Jijin Xu ∗ , Hao Lu, Jieshi Chen, Junmei Chen, Xiao Wei
School of Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, PR China

a r t i c l e i n f o a b s t r a c t

Article history: In order to clarify creep crack growth behavior in 2.25Cr–1.6W steel incorporating residual stresses, creep
Received 23 June 2015 crack tests were carried out on the tension creep specimens, in which the residual stresses were generated
Received in revised form by local remelting and cooling. Residual stresses in the specimens were measured using Synchrotron X-
29 September 2015
ray diffraction techniques. The fracture surface of the creep specimen was analyzed using statistical
Accepted 11 October 2015
methods and fractal analysis. The relation between fractal dimension of the fracture surface and fracture
Available online 13 November 2015
mode of the creep specimen was discussed. Due to different fracture mechanisms, the probability density
functions of the height coordinates vary with the intergranular crack percentage. Good fitting was found
Keywords:
Creep damage
between Gaussian distribution and the probability function of height coordinates of the high percentage
Fracture surface intergranular crack surface.
Fractal dimension © 2015 Elsevier B.V. All rights reserved.
Fracture toughness
Gaussian distribution
Residual stresses

1. Introduction complexity in geometry, residual stresses state, microstructure and

2.25Cr–1.6W steel (HCM2S, T23/P23) is usually used to manu-
facture water wall, superheater and reheater tubes, and main steam
pipe in the Ultra Super Critical (USC) Power Plants with the benefit
of enhancing thermal efficiency. Components employed in power
plants are continually exposed to both high temperatures and high
steam pressures. The safety and performance of a thermal power
plant are highly dependent on the integrity of its welded joints
[1]. Welding residual stresses arise as a consequence of the het-
erogeneous application of energy and localized melting [2,3]. The
welding residual stresses would be superimposed on any applied
loads, such as high steam pressures, and generate a complex stress
state on in-service components. They may accelerate creep dam-
age development and cracking initiation at high temperature on
these components. Hence, understanding the combined effects of
welding residual stress and primary loads on the creep damage and
crack initiation plays an important role in assessing the reliability
of high-temperature structural components [4–9].
It is very difficult to obtain an accurate structural integrity
assessment on actual components with residual stresses due to the
∗ Corresponding author.
E-mail address: xujijin 1979@sjtu.edu.cn (J. Xu).
local multiaxial creep ductility [4,5]. Therefore, such assessment
was usually investigated under closely controlled laboratory condi-
tions [4–13]. A new design for a compact tension (CT) specimen can
introduce a tensile residual stress field into the CT notch root area.
The residual stress is generated in compact tension (CT) specimen
by pre-compressing and then unloading [4,6,5]. The crack growth
in this work is driven solely by the residual stress with no exter-
nal stress applied at all. The stresses state of welding can affect the
damage process during creep; however, there is insufficient infor-
mation on the effects of welding residual stresses on fracture mode.
For this purpose, a novel experimental investigation is required to
characterize the fracture surface quantitatively.
The concept that the fractal dimension is related to irregular
objects with a self-similar property provides a basis for the quan-
titative characterization of the tortuosity of fracture surfaces.
Since the fracture surface roughness is predetermined by mor-
phological features which are dependent on the toughness of
the material, it can be deduced that fracture toughness might
also affect fracture surface roughness [14]. Ever since Mandelbrot
et al. first introduced this concept to materials science in 1984
[15], a number of studies have focused on the fractal character-
istics of fracture surfaces and their microstructure [14,16–25]. It
should, however, be noted that the reported correlation between
fractal dimension of fracture surfaces and the fracture toughness

http://dx.doi.org/10.1016/j.apsusc.2015.10.063
0169-4332/© 2015 Elsevier B.V. All rights reserved.
74 M. Xu et al. / Applied Surface Science 359 (2015) 73–81
Table 1
Chemical compositions of 2.25Cr–1.6W steel (mass, %).
C Mn P S Si Cr Mo
0.06 0.37 0.012 0.002 0.28 2.32 0.08
Fig. 1. Geometry of the employed creep specimens.
in steel could be negative [14,16–22] or positive [23,24]. Such
divergent results have been pointed out to be dependent on the
microstructure of fractured materials and measurement meth-
ods [21,25]. For instance, Strnadel et al. [21] reported that causes
could be the differences between the fractal character on differ-
ent scaling levels and the random character of initiation of fracture
mechanisms. In most previous cases, the fractal dimensions of frac-
ture surfaces were calculated based on the profile of the section
perpendicular to the fracture surface. In recent years, the fractal
characterization of fracture surfaces of steels has been presented
by means of image processing (IP), for example, by Tang et al.
[23,24]. The image used for IP was obtained from scanning elec-
tron microscopy (SEM). However, errors are introduced by noise
superimposed onto the material signal when using image analy-
ses from SEM [25]. Strnadel et al. [21] have estimated the fractal
dimension of the drop-weight tear test (DWTT) specimen fracture
surfaces using a three-dimensional (3D) photographing and box-
counting method. However, studies on 3D fractal analysis of the
creep test specimen fracture surface in Cr–Mo heat-resistant steel
were rarely reported.
The aim of the present study is to investigate the creep crack
growth behavior and fracture mode of 2.25Cr–1.6W steel incor-
porating residual stresses. The residual stress was generated in a
tension creep specimen by local remelting followed by cooling.
A quantitative fractographic analysis of fracture surfaces of the
creep specimens was performed in order to investigate the creep
fracture characteristics when taking consideration of the effect of
residual stress distribution. The results of this study will provide
possibilities for an objective and quantitative determination of
cracking growth process and fracture mechanisms in Cr–Mo
heat-resistant steel.
2. Experimental material and test procedures
While creep cracking has been observed in welded engineering
structures exposed to long-term elevated temperature condi-
tions, laboratory studies of this failure mechanism have to rely
W V Nb N Al B Ni Ti Ti/N
1.56 0.22 0.044 0.0035 <0.015 0.0026 0.05 0.009 3.9
on complex welded specimens with welding residual stresses
[4,6,5]. This type of specimen is far from ideal in that the residual
stress field is operator-dependent, difficult to predict and that the
microstructure varies locally. Our aim is to develop a laboratory
creep test specimen into which welding residual stresses could be
introduced, and in which the microstructure could represent the
actual welding structure.
The specimens for the creep test were cut from the 2.25Cr–1.6W
steel. Table 1 presents chemical compositions for 2.25Cr–1.6W
steel. Fig. 1 shows the adopted geometry and dimensions of the
creep specimen. A circular region of 6 mm diameter at the center
of the specimen was remelted completely along the through-
thickness orientation by TIG (tungsten inert-gas arc) welding at
current of 150 A and voltage of 20 V for 5 s and then cooled in
air. When the remelted region solidified, welding residual stresses
arose as a consequence of structure constraint. In this case, remelt-
ing treatments were carried out in an identical manner in two creep
tension specimens (labeled T1 and T2).
The specimen T1 was tested under a constant nominal tensile
stress of 330 MPa and a temperature of 550 ± 1 ◦ C using a lever arm
high temperature creep machine until it broke. The specimen T2
was tested under the same load and temperature condition, but
the test was halted after 180 h. This time was chosen because the
time to failure for specimen T1 was 191 h. The intent was to stop the
test at a point after cracking had initiated and close to failure. The
change of gauge was monitored of two specimens during creep test
using the linear variable differential transducers, as shown in Fig. 2.
Before creep testing, the specimen T1 was used to measure the
residual stresses in the longitudinal direction (the creep tension
direction). The experiment was carried out using synchrotron X-
ray diffraction at BL14B beam line of the Shanghai Synchrotron
Radiation Facility (SSRF), in Shanghai, China. The conventional
sin2 -method was applied. The specimen T1 was irradiated by
a X-ray beam of 23 keV energy, with an approximate dimension
of 0.3 mm × 0.3 mm, positioned in the center line of the remelted
specimen as shown in Fig. 1. The interval between the measured
points was 1 mm. Measurements were made the {2 1 1} reflection.
 M. Xu et al. / Applied Surface Science 359 (2015) 73–81
Fig. 2. Creep curves of the creep specimens: (a) creep strain vs. time for specimen
T1 and specimen T2; creep strain rate vs. strain for (b) specimen T1 and (c) specimen
T2.
The inclination angle was varied between 0◦ and 45◦ , using five
equal steps in sin2 . The measurements of residual stresses were
based on the following equation:
E cot 
ϕ = (1)
(1 + ) sin2
where  represents the diffraction angle and is the tilt angle
between the normal of the lattice plane and that to the speci-
men surface; and E and  are Young’s modulus and Poisson’s ratio,
respectively. The term E/(1 + ) is a constant. The linear relation-
ship in Eq. (1) for residual stress implies that a plot of lattice strain
75
against sin2 will be a straight line, whose gradient is a function
of  ϕ , E and . The value of  ϕ can be calculated from the gradient
of a straight line fit by the least squares method through the data
points for various tilt angles [26].
After creep testing, the fracture surface morphology of the creep
tested specimen T1 was observed with a scanning electron micro-
scope (SEM; FEI Nova Nano-SEM 230, USA). Specimens were then
characterized by a 3D confocal laser scanning microscope (3D
CLSM; Carl Zeiss LSM 700, Germany) with ZEN 2012 software to
obtain quantitative fracture surface maps. To better visualize the
overall three-dimensional (3D) fracture surface morphology and
the spatial relationship, a computer-controlled microstepper stage
of the confocal microscope with a 2.124 ␮m vertical step in depth
was used. The scanning area concentrated in the center of 1.28 mm
in thickness direction, keeping away from the region near surface.
Fracture roughness data were recorded as (x, y, z) files to rebuild a
3D topographic map, as shown in Fig. 3, which reveals the fine struc-
ture and morphological details of the creep fracture surface. They
were used to evaluate the character of fracture surfaces in creep
specimens based on mathematical-statistical method detailed in
the next section. The surface in the middle of specimen T2 was
polished to a 0.25 ␮m finish, and then etched by immersion with
agitation in 4% nital solution for 40 s. An optical microscope ZEISS
imager A1m with AxioCam MRc 5 was used to investigate the posi-
tion of crack initiation and fracture mode.
3. Statistical analysis methods
New methods of evaluating the character of fracture surfaces
in creep specimens were developed based on mathematical-
statistical processing of topographic map of the fracture surface.
The roughness data of fracture surface are composed of a set of 3-D
points with (x, y) denoting 2-D position and the third coordinate (z),
denoting randomly distributed heights which are characterized by
a one-dimensional probability density function f(z). Fracture rough-
ness data can be rebuilt into a 3D topographic map as shown in
Fig. 3. The basal plane is selected at the reference point with the
lowest height coordinate. The core for the estimation of probabil-
ity density function f(z) is to determine the empirical probability
function or the frequency histogram. In this study, the estimation
of the probability function is carried out by categorizing the dis-
ordered set of heights z1 , z2 , . . ., zn into different height intervals.
By defining the number of heights in interval i = 1,2, . . ., K as Ci , the
probability function pi is determined as follows:
Ci
pi = (2)
M
where M is the total number of sorted fracture surface heights.
Fractal analysis method is able to define self-affine or self-
similar structure with various complex and intricate shapes. Fractal
dimension (DS ) is the descriptor of self-similarity. There are many
methods for computing fractal dimension. In this study, three-
dimensional box counting method was utilized for determining
the DS of the topographic map of fracture surface resulting from
the creep test. In the practical calculation of the value DS , the topo-
graphic map has been covered by boxes with edge length ε. The
number of boxes N(ε) was counted as those boxes through which
any part of the surface passes, as shown in Fig. 4. N(ε) is propor-
tional to ε−Ds , where DS is the fractal dimension obtained from box
counting. So the slope of the plot of log N(ε) vs. log(1/ε) is the DS ,
as following:
lim log N(ε)
DS = ε−→0 (3)
log(1/ε)
By choosing different ε, DS can be estimated by using linear regres-
sion [21,27].
76 M. Xu et al. / Applied Surface Science 359 (2015) 73–81

Fig. 3. 3D fracture roughness map of the fracture surface of specimen T1.

4. Results and discussion
Fig. 5 shows the optical micrograph in the region of crack ini-
tiation from specimen T2. It exemplifies the intergranular nature
of creep cracking, which is caused by the linking-up of creep
cavities nucleated along prior austenite grain boundaries [4]. A
number of discontinuous cracks emanated in the location approxi-
mately 2 mm from the center of the remelted region, perpendicular
to the creep loading direction. Here, the dominant creep crack is
approximately 1.2 mm in length. Away from the remelted region,

Fig. 4. Box-counting method.

 M. Xu et al. / Applied Surface Science 359 (2015) 73–81
Fig. 5. Optical micrograph in the region of crack initiation of specimen T2.
cracks were never observed on the surface of specimen T2. Fig. 6
shows typical cross-section optical micrographs acquired from
fractured specimen T1. For the melted region (Fig. 6(a)) and the
coarse grained zone (CGHAZ; Fig. 6(b)), the sample failed inter-
granularly along prior austenite grain boundaries. Secondary cracks
are present behind the fracture surface. In contrast, in the region
of fine grained zone (FGHAZ), intercritical zone (ICHAZ) and base
Fig. 6. Cross-section optical micrographs of fractured specimen T1 in different regions: (a) the remelted region, (b) the CGHZA zone, (c) the FGHAZ zone, (d) the ICHAZ zone,
and (e) the base material zone.
77
material zone (Fig. 6(c)–(e)), the edge of fracture surfaces appears
completely transgranular ductile tearing feature.
Fig. 7 shows the measured residual stress in the longitudinal
direction along the central ligament of specimen T1. Within and
near the remelted zone, the measured residual stresses were ten-
sion, but turned into compression in the base metal. It is because of
that, during heating, thermal gradients lead to greater expansion in
the heat affected zone (HAZ), while contrastively, less in the base
metal farther from the weld. The base metal will still be softened
and thus plastically flow in order to accommodate the expansion
of the HAZ. Subsequently, during the cooling process, the thermal
contraction in the remelted zone and HAZ regions will be partially
resisted by the base metal, resulting in tension in the remelted zone
and HAZ, but compression in the base metal [1–3].
It is noteworthy that the location of the peak tensile stresses
correlates with the crack initiation zone, at approximately 2 mm
distances from the center of remelted region. This location is within
the remelted region and next to the region of CGHAZ.
The typical SEM images in different regions of the creep frac-
ture surface in specimen T1 are shown in Fig. 8, which are related
to different fracture mechanisms. A mixed mode of intergranu-
lar cracking and transgranular ductile tearing was observed in
the crack initiation region (Fig. 8(a) and (b)). The larger ten-
sile residual stress in this region lead to a major of creep crack
78 M. Xu et al. / Applied Surface Science 359 (2015) 73–81
Fig. 7. The measured residual stresses along the central ligament and the DS of the fracture surface of specimen T1.
initiation and propagation along grain boundaries, as seen in
Fig. 8(b). The metallurgical reason of stress-relief cracking for
2.25Cr–1.6W steel has been explained by Nawrocki and Park
et al. [28,29]. The stress-relief cracking in the region of CGHAZ
of 2.25Cr–1.6W steel is controlled by a balance of intergranular
and intragranular carbide precipitation. The precipitation of Fe-
rich M3 C carbides along a prior austenite grain boundary leads to
the formation of denuded zone. As such, the matrix adjacent to
the grain boundary becomes softer than that of the actual grain
boundary. The grain interiors are strengthened by the precipita-
tion of fine MC carbides. Hence, most of the strain that arises from
residual stress relaxation can be concentrated in the soft denuded
zone, causing transgranular cracking [28]. Likewise, for the melted
region, the majority of the coarser precipitates along the grain
boundary were found to be M3 C and M23 C6 with fine and ran-
domly distributed MC carbides in the matrix [29]. The growth of
creep crack leads to the decrease of the effective bearing area of
the specimen, and then the remaining uncracked areas could not
sustain the stress. Therefore, the rest part fractured in the form of
transgranular ductile tearing. In the meanwhile, it was observed
(Fig. 8(c)) that the intergranular crack covered less portion of the
fracture surface in the center of remelted region, as a consequence
of lower residual stresses. A typical dimple ductile tearing can be
observed in the zone of base metal, outside the remelted region of
specimen T1 (Fig. 8(d)). In this region, the residual stresses were
compressive and offset by the applied tensile load. Therefore, there
was not creep crack initiated here. When the crack growth was
significantly accelerated, the unstable fracture occurred. In conclu-
sion, the fracture mode in the creep specimens changed gradually
from high portion of partial intergranular cracking to high portion
of partial transgranular ductile tearing, with the decrease of tensile
residual stresses, and finally to almost pure dimple ductile tearing,
with compressive residual stresses.
The above result is also supported by the fractal analysis. A
3D topographic map of the creep fracture surface is shown in
Fig. 3 for the creep specimen T1. Box-counting method was used to
 M. Xu et al. / Applied Surface Science 359 (2015) 73–81 79

Fig. 8. Fractographies in various regions of the fracture surface of specimen T1: (a) crack initiation region, (b) crack initiation region, (c) crack propagation region, and (d)
unstable fracture region.

determine the fractal dimension DS for the 3D topographic map Table 2

Gaussian fitted values of fracture surface height coordinates distribution in speci-
of the creep fracture surface. The DS was calculated in continu-
men T1.
ous sections of projection area of 1300 ␮m (in the center region of
thickness direction) × 100 ␮m with a constant interval of 100 ␮m Zone y0 A (␮m) 2
through the entire fracture surface. Fig. 7 shows the DS of the A 1.08e−3 1.75 794.72 (102.92)2
fracture surface in relation to the distance from the center of the B 8.77e−4 1.82 782.94 (82.17)2
remelted region. As a graphic interpretation, the values of the DS
in the remelted region have a wavy character as the distance from
quasi-static process, and the ductile fracture surface exhibits very
the center of the remelted region changes. These findings are in
good fractal behavior [22].
conformity with the results obtained by Dlouhý et al. [14], who
According to the DS of the fracture surface in specimen T1,
reported that the wavy variation of the fractal dimension of the frac-
the fracture surface was divided into three zones, as shown in
ture surface suggests that the fracture process occurred in different
Fig. 7. Height coordinates from digitized fracture surfaces of the
ways. Therefore, this observation proves that the creep damage
three zones were statistically processed, and the empirical proba-
grows in a mixed mode of intergranular cracking and transgranular
bility functions pi were determined with the sorting interval width
ductile tearing, as stated above. The creep crack initiates and prop-
z = 2.124 ␮m according to Eq. (1). Fig. 8 shows a graphic rep-
agates along grain boundaries due to local stress–strain conditions.
resentation of the probability function of fracture surface height
Subsequently, the decrease of bearing capacity of specimen leads
coordinates. The empirical probability function of the height coor-
to transgranular ductile tearing. It is noteworthy that the change
dinates from zone A and zone B of the fracture surface has the
trend of the DS is similar to the change of the measured resid-
character of the Gaussian distribution, as following [30]:
ual stresses along the central ligament of specimen T1. The larger
 
tensile residual stress in the crack initiation zone promotes creep A (x − )2
crack initiation and propagation, resulting in higher intergranular f (z) = y0 +  exp −2 (4)

/2 2
crack percentage, in which the DS of the fracture surface exhibits
higher. Similar conclusions were reached by Strnadel et al. [21] where A is a constant; y0 the ordinate offset; the expectation
when testing DWTT broken specimens of X70 steel, where the DS of Gaussian distribution, presenting concentration values of the
of the fracture surface of DWTT specimens decreases with increas- fracture surface height coordinates;  2 the variance of Gaussian
ing ductile fracture percentage. The values of the DS in the region distribution, presenting the discrete of fracture surface height coor-
of base metal remain at approximately 2.13, correlating with typi- dinates. A little value of  2 means a higher concentration degree
cal dimple ductile fracture. In contrast to the wavy character of DS of the fracture surface height coordinates around . The Gaussian
of the fracture surface in the remelted region, the values of DS of fitted values are shown in Table 2.
the dimple ductile fracture surface exhibits stable variation. This Comparing the fittings of zone A (Fig. 9(a)) and zone B (Fig. 9(b)),
result is ascribed to the fact that ductile fracture is closer to the the curve in Fig. 9(b) looks more ‘Gaussian’ than that in Fig. 9(a). In
80 M. Xu et al. / Applied Surface Science 359 (2015) 73–81

function of the fracture surface height coordinates indicates that

the intergranular crack percentage influences the course of func-
tion f(z). The probability function of height coordinates of the high
percentage intergranular crack surface closely fits the Gaussian dis-
tribution.

5. Conclusions

The creep specimens were remelted with a 6 mm diameter at

the center of the specimen, introducing residual stresses as a con-
sequence of structure constraint. Residual stresses distribution and
detailed quantitative fractographic analysis of fracture surfaces of
the creep specimen have demonstrated that residual stresses play
an important role in creep crack initiation and propagation. In addi-
tion, the assessment of statistical characteristics of the 3D creep
fracture surface roughness data provides possibilities for the quan-
titative determination of mechanisms of creep fracture. The main
findings are as follows:

(1) The creep cracks initially originate at the location of the peak
tensile residual stress, with intergranular cracking nature. In
the region of tensile residual stresses, the creep specimen is bro-
ken by the mixed mode of intergranular crack and transgranular
ductile tearing. The creep fracture mode changes gradually from
high portion of partial intergranular crack to high portion of
partial transgranular crack, with the decrease of tensile residual
stresses.
(2) The mixed mode of intergranular cracking and transgranu-
lar ductile tearing may lead to a wavy variation of the fractal
dimension of the fracture surface. On the contrary, the change
of fractal dimension of the fracture surface seemed relatively
stable for unique fracture mode, such as pure dimple ductile
tearing.
(3) Fractal analysis of the fracture surface of creep specimen shows
that the fractal dimension of the fracture surface increases with
increasing intergranular crack percentage. The distribution of
fractal dimension of the creep fracture surface is similar to the
distribution of the measured tensile residual stresses of the
creep specimen, which demonstrates that the tensile residual
stresses promoted the initiation of creep crack.
(4) The intergranular crack percentage influences the probability
function of height coordinates. In the region of higher percent-
age intergranular crack, the statistical distribution of height
coordinates of the fracture surface closely fits the Gaussian dis-
tribution.

Acknowledgement

This work was supported by the National Natural Science Foun-

dation of China (Grant Nos. 51575347, 51405297 and 51204107).
Fig. 9. Probability function and probability density function of Gaussian distribu-
tion.
References

the fracture surface of zone B, where the intergranular crack per-
centage is higher, the probability function of the fracture surface
height coordinates is highly symmetric and centered at 782.94 ␮m
with exactly a single peak. The left and right parts match the gen-
eral shape of the Gaussian distribution. In the fracture surface of
zone A, where the intergranular crack percentage is lower than that
of zone B, the probability function of the fracture surface height
coordinates is centered at 794.72 ␮m but it is not very symmetric
and has several local peaks. Nevertheless, in zone C with pure dim-
ple ductile tearing surface, the probability function of the fracture
surface height coordinates does not fit Gaussian distribution at all.
Therefore, the relation between fracture mode and the probability
[1] A.F. Mark, J.A. Francis, H. Dai, M. Turski, P.R. Hurrell, S.K. Bate, J.R. Kornmeier,
P.J. Withers, On the evolution of local material properties and residual stress
in a three-pass SA508 steel weld, Acta Mater. 60 (2012) 3268–3278.
[2] J.A. Francis, H.K.D.H. Bhadeshia, P.J. Withers, Welding residual stresses in
ferritic power plant steels, Mater. Sci. Technol. 23 (2007) 1009–1020.
[3] M.J. Xu, J.S. Chen, Y.J. Jin, Y.Z. Li, H. Lu, Effects of weld metal strength and
transformation temperature on welding residual stress, Sci. Technol. Weld.
Join. 20 (2015) 208–215.
[4] M. Turski, P.J. Bouchard, A. Steuwer, P.J. Withers, Residual stress driven creep
cracking in AISI Type 316 stainless steel, Acta Mater. 56 (2008) 3598–3612.
[5] L.Y. Chen, G.Z. Wang, J.P. Tan, F.Z. Xuan, S.T. Tu, Effects of residual stress on
creep damage and crack initiation in notched CT specimens of a Cr–Mo–V
steel, Eng. Fract. Mech. 97 (2013) 80–91.
[6] L. Zhao, H.Y. Jing, L.Y. Xu, Y.D. Han, J.J. Xiu, Effect of residual stress on creep
crack growth behavior in ASME P92 steel, Eng. Fract. Mech. 110 (2013)
233–248.
 M. Xu et al. / Applied Surface Science 359 (2015) 73–81
[7] L. Zhao, H.Y. Jing, L.Y. Xu, Y.D. Han, J.J. Xiu, Experimental investigation of
specimen size effect on creep crack growth behavior in P92 steel welded
joint, Mater. Des. 57 (2014) 736–743.
[8] J.W. Zhang, G.Z. Wang, F.Z. Xuan, S.T. Tu, The influence of stress-regime
dependent creep model and ductility in the prediction of creep crack growth
rate in Cr–Mo–V steel, Mater. Des. 65 (2015) 644–651.
[9] G. Chen, G.Z. Wang, F.Z. Xuan, S.T. Tu, Mismatch effect in creep properties on
creep crack growth behavior in welded joints, Mater. Des. 63 (2014) 600–608.
[10] S. Spigarelli, F. Quadrini, Analysis of the creep behaviour of modified P91
(9Cr–1Mo–NbV) welds, Mater. Des. 23 (2002) 547–552.
[11] L. Milovića, T. Vuhererb, I. Blačićc, M. Vrhovacd, M. Stankoviće,
Microstructures and mechanical properties of creep resistant steel for
application at elevated temperatures, Mater. Des. 46 (2013) 660–667.
[12] L. Zhao, H.Y. Jing, L.Y. Xu, J.C. An, G.C. Xiao, Numerical investigation of factors
affecting creep damage accumulation in ASME P92 steel welded joint, Mater.
Des. 34 (2012) 566–575.
[13] L. Zhao, H.Y. Jing, L.Y. Xu, Y.D. Han, J.J. Xiu, Y.X. Qiao, Evaluating of creep
property of distinct zones in P92 steel welded joint by small punch creep test,
Mater. Des. 47 (2013) 677–686.
[14] I. Dlouhý, B. Strnadel, The effect of crack propagation mechanism on the
fractal dimension of fracture surfaces in steels, Eng. Fract. Mech. 75 (2008)
726–738.
[15] B.B. Mandelbrot, D.E. Passoja, A.J. Paullay, Fractal character of fracture
surfaces of metals, Nature 308 (1984) 721–722.
[16] Z.Q. Mu, C.W. Lung, Studies on the fractal dimension and fracture toughness
of steel, J. Phys. D: Appl. Phys. 21 (1988) 848–850.
[17] H. Su, Y. Zhang, Z. Yan, Fractal analysis of microstructures and properties in
ferrite-martensite steels, Scr. Metall. Mater. 25 (1991) 651–654.
[18] Z.G. Wang, D.L. Chen, X.X. Jiang, S.H. Ai, C.H. Shih, Relationship between
fractal dimension and fatigue threshold value in dual-phase steels, Scr. Metall.
22 (1988) 827–832.
81
[19] A. Imre, T. Pajkossy, L. Nyikos, Electrochemical determination of the fractal
dimension of fractured surfaces, Acta. Metall. Mater. 40 (1992)
1819–1826.
[20] H. Nagahama, A fractal criterion for ductile and brittle fracture, J. Appl. Phys.
75 (1994) 3220–3222.
[21] B. Strnadel, P. Ferfecki, P. Židlík, Statistical characteristics of fracture surfaces
in high-strength steel drop weight tear test specimens, Eng. Fract. Mech.
112–113 (2013) 1–13.
[22] S. Ruzicka, P. Haušild, Fractal aspects of ductile and cleavage fracture surfaces,
Eng. Fract. Mech. 77 (2010) 744–752.
[23] T. Wei, Y. Wang, Fractal characterization of impact fracture surface of steel,
Appl. Surf. Sci. 258 (2012) 4777–4781.
[24] Y.Q. Xian, J. Liu, C.H. Zhang, J.C. Chen, Y.T. Yang, L.Q. Zhang, Y. Song, Fractal
characteristics of fracture morphology of steels irradiated with high-energy
ions, J. Nucl. Mater. 461 (2015) 171–177.
[25] E. Charkaluk, M. Bigerelle, A. Lost, Fractals and fracture, Eng. Fract. Mech. 61
(1998) 119–139.
[26] I.C. Noyan, J.B. Cohen, Residual Stress: Measurement by Diffraction and
Interpretation, first ed., Springer-Verlag, New York, 1987.
[27] S. Dutta, A. Das, K. Barat, H. Roy, Automatic characterization of fracture
surfaces of AISI 304LN stainless steel using image texture analysis,
Measurement 45 (2012) 1140–1150.
[28] J.G. Nawrocik, J.N. Dupont, C.V. Robino, J.D. Puskar, A.R. Marder, The
mechanism of stress-relief cracking in a ferritic alloy steel, Weld. J. 82 (2003)
25–35.
[29] K. Park, S. Kim, J. Chang, C. Lee, Post-weld heat treatment cracking
susceptibility of T23 weld metals for fossil fuel applications, Mater. Des. 34
(2012) 699–706.
[30] T.T. Soong, Fundamentals of Probability and Statistics for Engineers, John
Wiley & Sons, England, 2004.