International Journal of Pressure Vessels and Piping 101 (2013) 113e142

Contents lists available at SciVerse ScienceDirect

International Journal of Pressure Vessels and Piping

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

Review

Recent progresses in experimental investigation and finite element analysis of

ratcheting in pressurized piping
Xiaohui Chen a, Xu Chen a, *, Dunji Yu a, Bingjun Gao b
a
School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China
b
School of Chemical Engineering and Technology, Hebei University of Technology, Tianjin 300000, China

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

Article history: This article presents an overview of recent progresses in experimental investigation and finite element
Received 11 July 2012 analysis (FEA) of ratcheting behavior of pressurized piping. Ratcheting, namely the cyclic accumulation of
Received in revised form plastic deformation, occurs when the structures are subjected to a primary load with a secondary cyclic
28 October 2012
load if the applied loads are high enough to make the structures yield. Typical piping structures including
Accepted 30 October 2012
straight pipes, elbow pipes and piping joints have been investigated experimentally under mechanical or
thermal cyclic loading. Finite element analyses with several well-developed constitutive models
Keywords:
implemented in the commercial software ANSYS and ABAQUS have been conducted to simulate and
Ratcheting
Cyclic loading
predict the ratcheting behavior of pressurized piping. Based on such experimental and FEA research,
Finite element analysis ratcheting boundaries have been determined with the final aim of aiding the safety design and
Constitutive model assessment of engineering piping structures. Some suggestions for structure ratcheting study are
Pressure piping proposed.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction
In the materials or structures subjected to a cyclic stressing with
non-zero mean stress, a cyclic accumulation of inelastic deforma-
tion will occur if the applied stress is high enough (ensuring that
a yielding occurs), which is called ratcheting or ratcheting effect
(some researchers called it cyclic creep too). Structural ratcheting is
proposed by Hübel [1]. Structural ratcheting can occur owing to
inelastic material behavior under cyclic loading, but more sophis-
ticated than material ratcheting for its inhomogeneous multiaixal
stress state due to various load combinations.
Pressurized piping as the most basic structures in chemical
industries and nuclear power plants are subjected to variable
mechanical and thermal loads which often have a cyclic nature. Let
us consider a simple straight pipe that is under a constant pressure
and suffers cyclic thermal gradient in the radial direction. In this
case, the pipe can be viewed as subjected to a primary load in the
axial and circumferential directions due to the constant pressure
and a secondary cyclic bending moment caused by the cyclic
thermal gradient. If the loads are high enough to make the struc-
ture yield, the plastic deformation may be accumulated cycle-by-
* Corresponding author. Tel.: þ86 22 27408399.
E-mail address: xchen@tju.edu.cn (X. Chen).
cycle until the plastic collapse of the structure occurs. This
phenomenon is known as ratcheting, i.e. structural ratcheting.
Similar to other damage mechanisms such as fatigue and creep,
ratcheting has also been considered in many design criteria for
engineering components and structures, including ASME Code
Section III [2], KTA [3], EN13445 [4], R5 [5] and RCC-MR [6]. These
criteria require the structures to remain below the defined ratch-
eting boundaries where elastic or plastic shakedown occurs [7].
However, current methods to determine the ratcheting boundaries
can be too conservative, or sometimes non-conservative. It is
therefore of great significance to investigate the ratcheting
behavior of such structures and predict it with sound accuracy,
which has already been concerned during the last two decades.
So far, several scholars have reviewed research work on ratch-
eting of materials and structures. Ohno [8] and Kang [9] have
reviewed the progresses in ratcheting research for various mate-
rials with emphasis on phenomenon observation and constitutive
modeling obtained before 1997 and 2008, respectively. In addition,
several components of consistent tangent modulus of several
constitutive models were discussed by Kang [9]. Abdel-Karim [7]
reviewed the literature on shakedown problems of various struc-
tures including 4-bar structure, beam, rotating disc, thin infinite
plate, infinite plate with a central hole, and tube under internal
pressure and variable temperature. According to Abdel-Karim [7],
determination of ratcheting boundaries were limited mostly to the

0308-0161/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijpvp.2012.10.008
114 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142

simple linear kinematic hardening rule (LKH) of Prager and the
nonlinear kinematic hardening of ArmstrongeFrederick (AF). The
advanced kinematic hardening rules introduced in Ref. [7] had been
reported to yield more accurate simulation of ratcheting behavior
[8,9]. Being different from these review papers, recent experi-
mental and numerical progresses of the structural ratcheting for
various piping components are mainly reviewed in this paper. The
process of AF type constitutive models implemented into ANSYS
software is detailedly reviewed. The fourth-rank constitutive
parameters H(i)nþ1 of consistent tangent modulus for AF type
constitutive models are derived. The predictions of structural
ratcheting with advanced kinematic hardening rules are discussed.
The ratcheting boundary determinations for piping components
are reviewed.
In the present article, Section 2 summarizes the experimental
observation of ratcheting behavior of pressurized straight pipes,
elbow pipes, tee pipes and cylinders with lateral nozzles, which are
subjected to the constant internal pressure and quasi-static cyclic
loading or dynamic loading. In Section 3, the methods of finite
element implementation of constitutive models with advanced
kinematic hardening rules are briefly reviewed. Finite element
analyses (FEA) of ratcheting behavior of pressurized piping are
evaluated in Section 4. In Section 5, ratcheting boundaries deter-
mined by such FEA are commented. Finally, some suggestions for
further studies are proposed as a conclusion of the review.
2. Experimental observation
Ratcheting effect, namely the cyclic accumulation of secondary
plastic deformation, occurs when the structures are subjected to
a primary load with a secondary cyclic load if the applied loads are
high enough to make the structures yield. Considering that the
yielding of structures essentially originates from the yielding of
materials in some local parts where stress level is high, the accu-
mulation of plastic deformation of the structures can be
represented by the local ratcheting strain in the following three
forms, which is similar to the case of ratcheting tests for materials,
1
εr ¼ ðε þε Þ (1)
2 max min
or
εr ¼ εmax (2)
or
εr ¼ εmin (3)
where, εmax and εmin are the maximum and minimum plastic
engineering strain in a cycle. To evaluate the evolution of ratcheting
strain, ratcheting strain rate is widely used, which is defined as the
increment of ratcheting strain during each cycle.
So far, ratcheting behaviors of various pressurized piping
structures of different materials have been extensively studied in
the last several decades as shown in Table 1. These structures
include straight pipes and elbow pipes as the typical ones, as well
as other structures such as tee and lateral nozzle. Internal pressure
was the common constant load exerted on these structures. Cyclic
loadings in two control modes, displacement control and load
control were mainly applied in these tests. For displacement
control, cyclic loading was applied in the controlled waveform of
displacement or rotation, while for loading control cyclic loading
was applied in the controlled waveform of force. The detailed
descriptions of the ratcheting behavior for each type of piping
structure under different control modes are presented as follows.
2.1. Ratcheting behavior of straight pipe
The experimental modes to apply cyclic loads to pressurized
straight pipes mainly contain three-point bending [10],

Table 1
Review of published literatures on ratcheting investigation of pressurized piping structures.

Scholar Material Structural type Constant Cyclic loading Control mode

loading
Fujiwaka et al. [33,109] Carbon steel: SA106 GrA Straight pipe IP Static displacement Displacement control
Stainless steel: SA312TP304 Elbow pipe/Tee IP cyclic loading
Seismic load
Gau [14] Carbon steel Straight pipe IP Displacement-controlled Displacement control
304 stainless steel cyclic bending load
Moreton et al. [20,25,26] Mild steel Straight pipe IP Fully reversed cyclic Load control
Stainless steel Elbow pipe bending moments
Corona and Kyriakides [30] Aluminum Straight pipe IP Cyclic bending load Load control
Kulkarni et al. [12,13] SA333 Gr.6 carbon steel Straight pipe IP Cyclic bending load Load control
SS304 stainless steel. Elbow pipe Shake table
Chen et al. [28,29,49,50,52,123] Low carbon steel Straight pipe IP Reversed bending load Load control
Elbow pipe
Rahman et al. [16,17] Alloy steel 4130 Straight pipe IP Cyclic rotation Rotation control
SS304L Elbow pipe Openingeclosing Displacement control
cyclic loading Force control
Yoshida et al. [22] Carbon steel Straight pipe IP Cyclic axial load Load control
Guionnet et al. [23,92] Austenitic stainless Tubular specimen Tensile Cyclic torsional loading Load control
steel (17-12SPH) stress
Rider et al. [24] 304S11 stainless steel En6 Thin-walled cylinders IP Cyclic tensile loading Load control
Low carbon steel
Ichihashi [34] Stainless steel Piping components IP Quasi-static cyclic loading Under sinusoidal
Low carbon steel Dynamic cyclic loading deflection control
Shaking table Under inertial force due to seismic
excitation and sinusoidal excitation
Igari et al. [11] 316FR Straight pipe IP Cyclic moment loading Displacement control
Acker et al. [45] Non indicated Elbow pipe IP In-plane bending Displacement control
Guionnet [23,92] Austenitic stainless steel Tube Tensile Cyclic torsional loading Load control
stress

Note: Internal pressure is abbreviated to IP.

 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
cantilevered bending [11], quasi-three-point bending [12] and four-
point bending [13]. Three commonly-used setups for these modes
are shown in Fig. 1. Fig. 1(a) could be used to observe the ratcheting
behavior of the straight pipe with internal pressure supplied by
a servo-hydraulic testing machine under displacement/load control
[14]. The specimen was fixed on the lower loading beam, which was
stiff enough, through supporting ears, clip boards, pins, bolts and
nuts. The upper loading beam was also stiff enough and attached to
the specimen with clip boards, pins, bolts and nuts. The loading
bars were screwed into the upper loading beam and the loading
block welded on the lower loading beam was clamped in the grip
head of the test machine. Cyclic pullepush of the test machine bent
the specimen reversely. Internal pressure was applied with another
branch of the pressure system. Thin-walled pipes under internal
pressure were in biaxial stress state, and axial bending stresses
from the pullepush action on the specimen were cyclically applied
to achieve non-proportional multiaxial loading.
Pressurized straight pipe subjected to seismic excitation was
tested through the setup in Fig. 1(b) [15]. Seismic excitation tests
were conducted with shake table. Fig. 1(c) showed another kind of
experimental equipment which was used to implement bending
control on the straight pipe [16,17]. The pipe ends were welded to
solid rod at ends, in which part of the solid rod was extended into
the tube. The pure bending moment was applied through four
rollers on each side with a sprocket assembly which was mounted
on two heavy beams. Sprockets were connected to hydraulic
cylinders and load cells through strand chains running over
sprockets. Rotation of the sprocket was obtained through con-
tracting and releasing the hydraulic cylinder. Rotation of the
sprocket in turn rotated ends of the pipe specimen. The system of
Fig. 1. Sketch of the experiment apparatus of straight pipe: (a) Four-point and three-
point bend test of straight pipe under constant internal pressure and cyclic load
[28,29]; (b) Straight pipe shake table test setup [15]; (c) Schematic of the pure bending
device [16,17].
115
roller supported sprockets allowed these sprockets to move in the
horizontal direction. The pipe specimen was connected to a pump
to pressurize and maintain the steady internal pressure during the
test. First, the pipe was subjected to a steady internal pressure.
Then, a symmetric cyclic rotation was applied at both ends.
2.1.1. Ratcheting behavior of straight pipe
Piping systems were observed to experience premature failure
due to fatigue cracks or plastic collapse due to the accumulation of
deformation or strain in piping components (Ranganath et al. [18],
English [19]). Firstly, the deformation of pressurized straight pipe
with cyclic loading is reviewed in the following. Gau [14] studied
four-point cyclic bending of pressurized straight pipe under
displacement control. The results showed that ratcheting strain
increased with progressive cycles and that ratcheting strain of
straight pipe with internal pressure was larger than that without
internal pressure. A series of the tests using mild steel and stainless
steel pipe specimens were conducted by Moreton et al. [20]. These
specimens were pressurized and subjected to fully reversed cyclic
bending moments at a frequency of about 5 Hz. It was observed that
ratcheting strain of the mild steel specimens was larger than that of
the stainless steel specimens. Krämer et al. [21] investigated the
ratcheting behavior of austenitic steel X10 CrNiNb18 9 pipes, and
their test setup with the pipe was given in Fig. 2. One end of the
pipe was fixed to a plate, and a rigid lever arm was welded on the
other end, allowing a bending moment to be induced into pipe
which should be constant in the longitudinal direction, as shown in
Fig. 2. The experimental results showed that the longitudinal and
circumferential strains were mainly recorded in the middle cross-
section of the pipe but only on the outside. In addition, the
results also showed nearly constant load amplitude, which meant
a fairly constant cyclic bending moment.
Yoshida et al. [22] studied mechanical ratcheting in a carbon
steel pipe under combined cyclic axial load and steady internal
pressure. The results showed that the biaxial strain accumulation in
the pipe was influenced by maximum effective stress, stress ratio
and steady stress. Guionnet [23] studied austenitic stainless steel
(17-12SPH) tubular specimens subjected to cyclic torsional loading
under constant tensile stress at 600  C. The experimental results
showed the relationship of the ratcheting strain versus the number
of cycles and stress versus strain hysteresis loops. Rider et al. [24]
experimentally examined pressurized 304 stainless steel and En6
Fig. 2. Schematic test arrangement for the pipe tests [21].
116 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
low carbon pipes under the loading conditions of 1% axial strain
with hoop stresses of approximately 0, 1/4, 1/2 and 3/4 of the initial
uniaxial yield stress. It was found that both stainless steel and low
carbon steel pipes satisfied the proposed ASME ratcheting strain
limit of 5% hoop strain after 10 cycles of 1% axial strain range, for
any value of internal pressure.
Moreton et al. [25] conducted five pairs of tests using cylindrical
specimens with mean diameter/thickness (Dm/t) ratios in the range
9  Dm/t  28. The specimens were subjected to constant internal
pressure and alternating bending moments at typical frequencies of
seismic events. Experimental results demonstrated that the
maximum bending strains would be formed at the top and bottom
outside surfaces of the cylinder when bending moments were
greater than the yield moment. As a result, ratcheting strains would
be therefore largest on these surfaces. The moment level for the
onset of ratcheting was found to be in agreement with theoretical
predictions. Moreton et al. [26] and Zehsaz et al. [27] performed
a series of simulated seismic bending tests on carbon steel pres-
surized cylinders having diameter/thickness ratios in the range
9  Dm/t  28 and 8  Dm/t  28, respectively. It is found that axial
and hoop ratcheting strain occurred on the top and bottom surfaces
of the seismic specimens. However, it should be noted that the
experimental work used a rising amplitude technique which might
effectively reduce the ratcheting strain at any particular dynamic
bending moment. It was possible that those tests conducted at
a low amplitude could harden the material sufficiently to reduce
the ratcheting strains observed at a higher amplitude. In addition,
little fundamental difference was found in the behavior of the
thick-walled and thin-walled specimens. All specimens were found
to develop approximately with the same ratcheting strain rates.
Kulkarni et al. [12,13] observed circumferential ratcheting strain
at the top of the straight pipe where cyclic bending stress was
maximum. The specimen was made of SA333 Gr.6 carbon steel and
under constant internal pressure and cyclic bending load. The
results showed that continuous ratcheting strain changed with the
varying amplitude loading and shakedown behavior did not occur
for the given cycles of loading. Gao et al. [28,29] studied ratcheting
behavior of a pressurized low carbon steel pipe under reversed
bending load with a quasi-three-point bending apparatus. The
results also revealed that ratcheting strain occurred mainly in the
hoop direction, as shown in Fig. 3. It was seen that the ratcheting
Fig. 3. Evolution of ratcheting strain at the central position of straight pipe subjected
to the internal pressure of 20 MPa and cyclic bending load with the amplitude of 23 kN
[28,29].
strain rate decreased with cyclic number but shakedown condition
was not achieved in 50 cycles. Moreover, multi-step bending tests
showed that the ratcheting rate increased with increasing loading
level but reduced or even vanished at a lower bending level
imposed after a higher bending level, as shown in Fig. 4. The
position 1 in Fig. 4 was at the centre of straight pipe. The positions
2, 3, 4 was 135 mm, 175 mm, 215 mm from the centre of straight
pipe, respectively. Rahman et al. [16,17] studied the ratcheting
behavior of pressurized straight pipe of alloy steel 4130 with cyclic
bending loading under rotation control. Axial strain and circum-
ferential strain responses were recorded at the top and side of the
pipe using strain gages. The experimental results indicated that the
increase of internal pressure or the prescribed displacement
increased the ratcheting strain rate.
2.1.2. Ratcheting induced ovalization/local bulging of cross-section
of straight pipe
The ovalization of the pipe cross-section and local bulging was
also observed along with the ratcheting behavior of the straight
pipe. For example, Igari et al. [11] performed ratcheting test of the
straight pipe subjected to deformation-controlled cyclic bending.
Test results showed that the ratcheting strain became significant at
around d/dcr ¼ 0.6 and d/dcr ¼ 0.3, and accelerated the plastic local
buckling mode, in which d was the deformation, and dcr was the
critical deformation. Corona and Kyriakides [30] investigated the
degradation and buckling of circular tubes under cyclic bending
load and external pressure. The results revealed that moment-
controlled cyclic bending with a mean moment led to ratcheting
in curvature as well as in ovalization. External pressure accelerated
the accumulation of ovalization and led to buckling in cycles less
than in the corresponding pure bending cases. Moreton et al. [20]
conducted a series of experiments on mild steel pipe specimens.
The results showed that significant ratcheting strain led to gross
swelling. Kulkarni et al. (2003, 2004) [12,13] reported that ovali-
zation of the pipe cross-section and local bulging was present at
higher loading. Gao [29] observed that the diameter increased in
the horizontal direction and decreased in the vertical direction. The
cross of straight pipe changed from original roundness to ellipse.
Moreover, for the straight pipe subjected to reverse bending
loading, the cross-section of the specimen in each cycle altered
between roundness and ellipse. Rahman et al. [16,17] reported in
Fig. 4. Evolution of ratcheting strain at different positions of straight pipe subjected to
the internal pressure of 15 MPa and cyclic bending load with the amplitude of 23 kN,
28 kN, 33 kN and 28 kN in sequence [28,29].
 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
the straight pipe experiments that increase of internal pressure
increased the ratcheting rate but hardly affected the amplitude of
ovalization. It also showed that mean ovalization was also increased
with the increase in the prescribed displacement.
Jiao and Kyriakides [31] conducted a set of experiments on
super-duplex tubes that was pressurized and then compressed into
the plastic range to a level that initiated wrinkling. It was then
cycled under stress control with a compressive mean stress while
the pressure was kept constant. The combined loads caused
simultaneous ratcheting in the hoop and axial directions as well as
a gradual growth of the wrinkles. At some stage the amplitude of
the wrinkles started to grow exponentially with the number of
cycles leading to localized collapse. The rate of ratcheting and the
number of cycles to collapse depended on the initial compressive
pre-strain, the internal pressure, and the stress cycle parameters all
of which were varied sufficiently to generate an adequate database.
Interestingly, in all cases collapse was found to occur when the
wrinkle amplitude reached a critical value under monotonic
compression.
2.1.3. Effects of ratcheting on fatigue life
Evaluations on piping response under static and dynamic cyclic
loading that produce high stress intensities require careful
consideration of ratcheting and fatigue.
In 1990s, it was found that the ratcheting behavior of the
pressurized straight pipe subjected to seismic excitation could
result in the reduction of the fatigue life of the piping components.
For example, Matzen et al. [32] reviewed numerous test results on
piping components and systems subjected to a high level seismic
excitation, and concluded that ratcheting behavior was a significant
contributing factor in the failure of components and systems.
Fujiwaka et al. [33] obtained the ratcheting response of the straight
pipes made from carbon steel (SA106 GrA) and stainless steel
(SA312TP304) under dynamic cyclic inertia load, respectively. The
results indicated that the failure mode of piping components under
seismic loading was fatigue with ratcheting and the reduction of
fatigue life was attributed to the accumulated ratcheting strain.
Moreton et al. [20] reported that for the mild steel specimens,
significant ratcheting led to gross swelling and eventually resulted
in failure. Although ratcheting behavior was observed in the
stainless steel specimens, it was not so pronounced and failure
appeared to be due to the low cycle fatigue.
Ichihashi [34] performed a series of piping tests, including
piping specimens, simplified piping system tests and large scale
piping system tests. These specimens were made of carbon steel
Fig. 5. Schematic diagrams of the butteweld specimen [36].
117
and stainless steel. For the piping components, the quasi-static
cyclic loading was under sinusoidal deflection control and the
dynamic cyclic loading was under inertial force due to seismic
excitation and sinusoidal excitation. For the simplified piping
systems and the large scale piping systems, seismic excitation tests
were done with shaking table. Results showed that the failure of
piping system was attributed to low cycle fatigue with ratcheting
behavior. Moreover, the fatigue life reduction of the piping
components was attributed to large ratcheting strain. It was also
clarified that the fatigue lives of the piping components at given
strain range were about 1/5 of the fatigue lives of the base material
for carbon steel and about 1/2 of those for stainless steel.
For weld piping joint, Humphreys [35] studied the ratcheting
strain near the weld piping joint, as shown in Fig. 5, and considered
that ratcheting strain was a probable reason for the fatigue failure.
Experimental results also revealed that welding residual stresses
might be responsible for the induction of the ratcheting strain level.
Cheng [36] conducted a set of low cycle fatigue experiments on
butt- and socket-welded piping joints in a cantilever setup. It was
observed that the fatigue cracks in all experiments occurred at the
weld toe where both the strain amplitude and ratcheting strain
were anticipated to be the largest. The ratcheting response of the
welded joints was questionable because of the zero shift charac-
teristics of the strain gages used. Furthermore, comparison of the
welded joint ratcheting responses with those from the cyclic
bending of the straight pipe and material tests indicated that the
residual stresses might induce the ratcheting strain at the welded
joints. Then, Song [37] performed four welded piping specimens
subjected to displacement-controlled cyclic loading with different
number of cycles. The influence of ratcheting behavior on the
fatigue failure of the welded piping joints was investigated. The
stress response of welded pipe specimens showed initial cyclic
hardening until 20 cycles followed by softening, and ratcheting
strain was observed at the weld toe. Microstructure at the weld toe
showed various dislocation types at different numbers of cycles.
In summary, it is the common phenomenon that pressurized
straight pipe was subjected to cyclic bending load, seismic excita-
tion, etc. The observations of the above statement regarding the
significance of ratcheting in seismic loading were of potential
importance in applications to plant. Two control modes were
commonly applied in the experiments, i.e., load control and
displacement control. Typical experimental results were as follows:
(1) ratcheting strain mainly occurred in the hoop direction; (2)
ovalization of the pipe cross-section and local bulging was
observed at higher loading; (3) the ratcheting strain rate increased
118 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
with increasing loading level; and (4) reduced fatigue life was
attributed to accumulated ratcheting strain.
2.2. Ratcheting behavior of elbow pipe
Ratcheting deformation of elbow pipe has been investigated
mainly under dynamic or static loading conditions. Fig. 6 showed
the typical test setup for dynamic loading and static loading tests.
The connecting blocks were screwed into the pipe ends of the
elbow specimen, and the loading bars were connected with the
blocks through pins. Loading bars were clamped in the grip head of
the test machine. Pressurized elbow pipe subjected to seismic
excitation was tested through the setup in Fig. 6(a). Seismic exci-
tation tests were done by shake table. Fig. 6(b) was the experi-
mental apparatus of pressurized elbow pipe subjected to cyclic
static loading. Cyclic pullepush of the test machine opens and
closes the elbow, thus bending the specimen reversely. Internal
pressure was applied with another branch of the pressure system.
In addition, at each position, the 0 and 90 gauge of the three
rosette strain gauges were aligned with the meridional and hoop
directions of the elbow. The strain gauges of elbow pipe were
distributed as the pattern in Fig. 6(c).
2.2.1. Ratcheting behavior of elbow pipe under seismic loads
The test setup in Fig. 6(a) was used to study the seismic char-
acteristics of the pressurized elbow pipe under displacement
control and load control. For example, Boussaa et al. [38] performed
three dynamic tests on pressurized in-plane elbows made from
different steels including two carbon steels and one stainless steel.
The ratcheting strain of the three elbow pipes was measured at
flank, and the strain gauges might be shed from the elbow after one
tenth of the failure lives of the specimens. The results indicated that
Fig. 6. Sketch of the experiment apparatus: (a) Dynamic elbow test setup [42]; (b) Static elbow test setup [16,51]; (c, d) Strain gauge location pattern [16,51].
fatigue-ratcheting was the major failure mode. Yahiaoui et al. [39]
performed sinusoidal seismic excitations in-plane or out-of-plane
directions on a series of sixteen pairs of 90 short and long radius
pressurized welding elbows of carbon steel and stainless steel.
Under in-plane seismic excitations, it was shown that all compo-
nents failed by developing a crack at flanks running in the axial
direction of the long radius elbow pipes. For the short radius elbow
pipes, all components failed by developing a crack at intrados
running in the circumferential direction. Under out-of-plane
seismic excitations, all components failed either by developing
a crack at flanks running in the helical direction about the mid-
circumference and extending from the mid-meridian between
flank and intrados or by a crack at a defect in the weld. In addition,
Yahiaoui et al. [39] reviewed the fatigue-ratcheting approach,
including ASME Code cumulative damage assessment procedure,
Markl type equation, etc.
Yahiaoui et al. [40] tested 16 pairs of long and short radius 90
elbow pipes of carbon and stainless steel under conditions of steady
internal pressure and either in-plane or out-of-plane resonant
bending to simulate seismic conditions. It was shown that the
maximum ratcheting strain occurred at crown and intrados posi-
tions for in-plane bending. For out-of-plane bending, the maximum
ratcheting strain occurred at half-way along the mid-circumference
between the crown and intrados. In addition, the experimentally
obtained axial strains varied in proportion with external loads
under cyclic bending loading superimposed on steady internal
pressure.
DeGrassi et al. [41e43] presented the results of BNL (Broo-
khaven National Laboratory) collaboration efforts on the JNES/
NUPEC (Japan Nuclear Energy Safety Organization/Nuclear Power
Engineering Corporation) Ultimate Strength Piping test program.
Dynamic tests were conducted on elbow piping components in the
 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
program. The specimens were subjected to either sinusoidal or
seismic wave acceleration input motions. The specimens were
cycled until through-wall cracks detected. Experimental results
illustrated that large ratcheting strain occurred at flank. The
eventual failures of all pipes were attributed to fatigue-ratcheting.
Huang et al. [44] described dynamic elbow response characteristics
and analytical study. It was observed that pressure effects were
highly influential on ratcheting direction and that unpressurized
elbows tended to have elbow-closing ratcheting when the sus-
tained moment was applied in the elbow-closing direction.
2.2.2. Ratcheting behavior of elbow pipe under cyclic loads
Fig. 6(b) showed the test setup of the elbow under quasi-static
load. So far, it has been the most common experiment apparatus
that was used to study the ratcheting behavior of the pressurized
elbow. Acker et al. [45] reported the experimental results of ten
pressurized elbows subjected to cyclic in-plane bending load under
displacement control. It showed that the ratcheting strain appeared
as an increase of crown to crown diameter of the elbow pipe and of
circumferential strain at the crown. Moreover, the ratcheting strain
increased as a logarithmic function of the number of cycles, which
was proportional to the square of the secondary stress range and
directly proportional to the constant primary stress value, respec-
tively. Stearns et al. [46] demonstrated the relationship of an
elbow’s bend radius and the magnitude of the ratcheting strain for
the pressurized elbows subjected to cyclic bending loads. Their
results indicated that the effective or von Mises stress must exceed
the yield point before any ratcheting occurred.
Touboul et al. [47] studied nine pairs of elbows subjected to
cyclic in-plane, closing and opening, displacement quasi-static
conditions and constant pressure. During the tests, the flank
diameter and extradoseintrados diameter increased with the
number of cycles, but the extradoseintrados diameter was with
a lower amount of accumulated strain. Moreover, the RCC-MR
efficiency diagram was applicable for ratcheting evaluation, but
the final strain determined by this method might be too conser-
vative. Fujiwaka et al. [33] observed that ratcheting strain increased
with increasing load cycles in static cyclic tests of pressurized
elbow pipe. It was also found that there was no difference in the
shape of the hysteresis loops between the static and dynamic tests.
Moreover, the rate of the fatigue strength reduction was due to the
ratcheting strain for piping component. Then, Wilkins [48]
observed the behavior of butt-welded elbows with a variety of
loading conditions under displacement control. The out-of-plane
and in-plane performance of the elbows was examined. It
showed that hoop and axial ratcheting strain was observed at flank,
intrados and approximately midway between the intrados and the
flank on the mid-section of the elbow.
Chen et al. [29,49,50] experimentally studied multiaxial ratch-
eting for pressurized low carbon steel elbows under reversed
bending. It was shown that the maximum ratcheting strain occurred
mainly in the hoop direction at flanks. In addition, hoop-ratcheting
strain was found at intrados for individual specimen, while no
ratcheting strain was found at extrados in all tests. Moreover, under
multi-step loading conditions, the ratcheting strain rate grew with
the increase of the bending loading level under the same internal
pressure or with the increase of the internal pressure under the same
bending load. Recently, Rahman [16] and Hassan and Rahman [51]
studied that both local (ratcheting strain) and global (e.g., load-
deflection, ovalization) responses of the elbow pipe components,
respectively. The elbow specimens were tested under both
displacement-controlled and force-controlled openingeclosing
cyclic loading, with or without internal pressure. The experimental
results demonstrated that the ratcheting strain occurred at both
flanks, intrados, extrados and midway between extrados and flank
119
versus the number of cycles. The maximum ratcheting strain
occurred at flank. Chen and Chen [52] experimentally investigated
that ratcheting behavior of a 90 -elbow pipe of Z2CND18.12N
austenitic stainless steel under constant internal pressure and cyclic
bending. It was also found that the hoop-ratcheting strain at intrados
and flank position was much larger than that at other positions.
In conclusion, elbow piping was tested under openingeclosing
cyclic loading or seismic loads, with or without internal pressure.
Two control modes, displacement controlled and force controlled,
were employed in the tests. The experimental results can be
summarized as follows: (1) ratcheting strain occurs at flank, intrados
and approximately midway between the intrados and the flank on
the mid-section of the elbow; (2) maximum ratcheting strain occurs
mainly in the hoop direction at flanks; (3) the presence of internal
pressure leads to a significant increase of the ratcheting strain; (4)
failure is attributed to fatigue with ratcheting. Reduced fatigue life
can be estimated by considering the ductility consumption of
material due to accumulated ratcheting strain. (5) The deformation
behavior and the stressestrain characteristics of pipe elbows depend
on geometrical parameters and the type of loading.
2.3. Other pipe structures
In addition to the above-reviewed straight pipe and elbow pipe
researches, the ratcheting behavior of other structures has been
also studied, including lateral nozzle of cylinder, tee pipe, etc. For
example, Fujiwaka et al. [33] investigated the ratcheting behavior
of pressurized tee pipe under static cyclic test and the dynamic test.
Ratcheting strain and strain amplitude were obtained after the
initial several cycles. Ratcheting strain increased with the number
of loading cycles, but strain increment each cycle declined with the
increasing number of loading cycles.
Recently, Yang et al. [53,54] experimentally studied the ratch-
eting strain of pressurized lateral nozzle of cylinder made of 20
carbon steel with a multiaxial fatigue testing system and a self-
designed bending apparatus for lateral nozzle structure (Fig. 7).
Ratcheting strains were acquired by three multi-channel strain
processors with strain gauges. It was found that ratcheting strains
were detected around the acute angle region of the structure, such
as A0, A1, A2, a0, a1, and the first principle strains were mainly
directed to the intersecting weld, as shown in Fig. 8. Maximum
ratcheting strain occurred at the acute angle region of the nozzle.
The ratcheting strain at a0, A0, A1 was larger than that at a1, A2. The
ratcheting strain at a0, A0occurred mainly in the axial of lateral
nozzle and the cylinder, respectively. The ratcheting strain of other
points occurred in a certain angle with axial direction, and pointed
to the weld direction. In addition, the ratcheting strain of pressur-
ized lateral nozzle of cylinder made of 20 carbon steel was studied
under the same internal pressure and different cyclic bending
loading. The ratcheting strain at A0 was the largest, and increased
with the increasing cyclic bending loadings.
So far, straight pipe and elbow piping have been the typical
structures of structural-ratcheting analysis. In addition, tees, lateral
nozzle and weld piping have become research subjects. These
specimens were subjected to the constant internal pressure and
cyclic bending loads or seismic loads under displacement control
and loading control. Experimental results indicated that the
ratcheting strain occurred at the critical locations. Ratcheting strain
mainly occurred in the circumference directions. The eventual
failures of these specimens were attributed to fatigue-ratcheting.
2.4. Thermal ratcheting
The conventional thermal ratcheting is well known as
progressive deformation under combination of constant primary
120 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
Fig. 7. Sketch of the experiment apparatus of lateral nozzle of cylinder [53,54]: (a) Experimental apparatus, (b) Strain gauges location pattern, (c) Location of strain gauges.
and cyclic thermal stresses. Structure components in nuclear
industries are subjected to a temperature front moving cyclically in
the axial direction, circumferential plastic strain may accumulate
with the increase of the number of cycles. This is a thermal-
ratcheting problem induced by a liquid surface moving in
a cylinder. Abdel-Karim [7] and Kang [9] reviewed simply thermal
structural ratcheting. For example, Jahanian [55] investigated the
total accumulation of plastic strains of a thick-walled tube sub-
jected to rapid heating and cooling of the inside of the cylinder over
four cycles. The yield reduction at each cycle was considered. It was
found that incremental deformation or reversed plasticity might
occur, depending on the type of the mechanical loading. Watanabe
et al. [56] conducted a thermal-ratcheting test and a thermal
fatigue test of a 316FR stainless steel cylindrical specimen under
combined axial load and moving axial temperature distribution. In
the experiments, longitudinal- and radial-ratcheting deformations
were measured by the laser speckle strain measurement system.
The results showed that ratcheting strain increased with increasing
number of cycles. The ratcheting strain in base metal part was
larger than that of weld metal parts. In addition, the radial distor-
tion became larger with increasing number of cycles.
Thermal ratcheting of the elbow pipe was performed by Ueda
et al. [57]. Elbow pipe was subjected to sustained primary load and
cyclic thermal transient strains. It was shown that ratcheting strain
occurred at around flanks, while the elbow cross-section was
ovalized cycle-by-cycle causing large circumferential bending
strain at around flanks. Moreover, it was shown by the test and
analyses that the maximum primary membrane stress in an elbow
was an essential parameter for thermal-ratcheting behavior of
elbows. It was verified that the application of thermal-ratcheting
criteria in the ASME Code Case N-47 was considerable conservative.
Couterot et al. [58] studied the pressurized auxiliary spray line
of PWR nuclear power plants submitted to severe temperature
transients during upset conditions. During these tests, tempera-
ture, strain and ovalization of the pipe cross-section were
measured. No significant ratcheting deformation was detected on
all components, except on the 600  200  600 T-piece, where a weak
progressive diameter increase was observed during a few cycles.
For such temperature transients, the RCC-M French design code
rules that prevented the ratcheting deformation hazard had no
consideration for the components with thickness transition.
Except for the examples in Abdel-Karim [7] and Kang [9], Lee
et al. [59] still observed the progressive inelastic deformation under
moving thermal fronts for a Y-shaped welded cylindrical structure.
It was shown that ratcheting strain at the first cycle was highest and
tended to decrease in the subsequent cycles. The ratcheting char-
acteristics of the Y-shaped structure showed that progressive
inelastic deformations of expansion and contraction mode occurred
at the same time, although globally the expansion deformation was
dominant. In addition, thermal ratcheting was affected by weld
residual stress. According to the UK assessment code R6, it was
shown that the effect of residual stress on the global ratcheting
 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
Fig. 8. Ratcheting strains of gauged points [54]: (a) Strain of point A0, (b) Ratcheting strain of point A0, (c) The first principal strain, (d) The first principal strain.
strain was less than 0.1% both axial and circumferential directions.
The effect of the weld residual stress on the ratcheting strain was
wiped out quickly due to the applied loading cycle. Lee et al. [60]
studied the thermal ratcheting of 316L stainless steel cylinder.
The thermal-ratcheting deformation at the reactor baffle cylinder
of the liquid metal reactor can occur due to the moving tempera-
ture distribution along the axial direction as the sodium-free
surface moved up and down under the cyclic heat-up and cool-
down transients. The ratcheting deformation was measured with
the laser displacement sensor and LVDTs after cooling the structural
specimen which was heated up to 550  C with steep temperature
gradients along the axial direction. The temperature distribution of
the test cylinder along the axial direction was measured with 28
channels of thermocouples and was used for the ratcheting anal-
ysis. Moreover, the thermal ratcheting deformation was analyzed
with the constitutive equation of nonlinear combined hardening
model which was implemented as ABAQUS user subroutine and the
analysis results were compared with those of the test. The ratch-
eting deformation shapes obtained by the analysis with the
combined hardening model were in reasonable agreement with
those of the structural tests. Abdel-Karim [7] also reviewed the
effect of kinematic rules on thermal ratcheting analysis. Thus, these
issues will not repeat here.
3. Finite element implementation
Although experimental investigation on the ratcheting behavior
of pressurized piping has been researched extensively as reviewed
above and valuable results have been obtained to understand the
121
general rule of ratcheting deformation and fatigue. In the existing
experimental conditions, it should be noticed that the structure
size and type as well as loading conditions are limited. For the
design of actual complex structures, FEA can be helpful to predict
the ratcheting behavior of pressurized piping based on appropriate
constitutive models. Progresses in constitutive models for simula-
tion and prediction of ratcheting have been mainly reported by
Ohno [61e63], Abdel-Karim [7,61], Kang [9,64e69], and Chaboche
[70e77]. So far models available in commercial FEA software
contain bilinear kinematic hardening (BKH) model [78], multilinear
kinematic hardening (MKIN) model (Besseling [79]; Mroz [80];
Owen et al. [81]) and Chaboche (CH3) model [70,72,75]. However,
the existing finite element packages fail to well simulate ratcheting
of structural components [82]. This shortcoming of FEA in simu-
lating ratcheting behavior is demonstrated to be related to the
deficiency of the constitutive models. Hence it is necessary to
implement advanced constitutive models into FEA software to
achieve better evaluation of structural ratcheting.
Finite element program like ANSYS and ABAQUS are commonly
used for structural response simulation under cyclic loading.
Advanced constitutive model has been incorporated into ANSYS,
ABAQUS, etc. In the recent years, a number of advanced nonlinear
kinematic hardening constitutive models have been proposed by
Chaboche et al. [83], Ohno and Wang [62,63], Chen and Jiao [84],
Chen et al. [85], Jiang and Sehitoglu [86,87], Abdel-KarimeOhno
[61], Dafalias and Popov [88,89], Hassan and Kyriakides [90,91],
Guionnet [92], Delobelle et al. [93], Hassan et al. [94,95], and many
others for improving ratcheting simulation under multiaxial
loading. There are two key points for the advanced constitutive
122 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142

rffiffiffi
models incorporated into finite element program. One is that these 3 vf
models contain the large number of parameters, many of which are dεp ¼ dp½M1  (6)
2 vs
determined through a trial and error curve fitting approach using
several experimental responses at the material level [96]. The qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
where, dp ¼ c ð2=3Þdεp ½M2 dεp , [M2] is a non-zero auxiliary
T

process is tedious and might be erroneous. Then, Rahman [16]

overcame through developing an automated parameter optimiza- diagonal matrix, namely [M2] ¼ diag [1,1,1,1/2,1/2,1/2];
pffiffiffiffiffiffiffiffi
tion system using heuristic search technique (e.g. genetic algo- vf =vs ¼ 3=2ðs  a=sy Þ ¼ n, h is plastic modulus.
rithm). The other is a suitable integration algorithm, which should
satisfy the requirements of robustness, accuracy, efficiency and (3) Strain decomposition
stability. Obviously, it is critical for the numerical calculation.
At present, the proposed stress integration algorithms are ε ¼ εe þ εp (7)
divided into explicit and implicit integration algorithms. Koji c [97]
reviewed stress integration algorithms for inelastic material
models. It indicated that implicit stress integration had more (4) Elastic strain is assumed to obey Hooke’s law
significant advantages in accuracy than the explicit one, and was
widely applied in the numerical implementation of inelastic s ¼ De εe ¼ De ðε  εp Þ (8)
constitutive model. In addition, Kojic observed that the conver-
gence rate in equilibrium iterations was obviously improved, which
was attributed to tangent modulus matrix. In a way, the tangent (5) CH3 model
modulus matrix is evaluated consistent with the implicit stress
integration algorithm. Then, Kobayashi and Ohno [98] clarified X
M
detailedly to incorporate a general form of kinematic hardening a ¼ ai (9)
into finite element program, i.e. AF model [99]. Consistent tangent i¼1
modulus was derived by introducing a set of fourth-rank consti- 2
tutive parameters into discretized kinematic hardening. Backward dai ¼ C ½M dεp  gi ai dp (10)
3 i 2
Euler method was used to discretize kinematic hardening. In
addition, successive substitution and its convergence were also
3.2. Application of radial return method
discussed for implicitly integrating stress. Moreover, Kang [68]
implemented visco-plastic constitutive model with strain range-
Given all constitutive variables at t ¼ tn, convergent solutions
dependent hardening into finite element code, ABAQUS. Backward
contain sn, an, εpn. The elastic predictor is taken to be an elastic
Euler method and implicitly integrating stress were also used in the
tentative stress sTnþ1
subroutine. Consistent tangent modulus was derived for the visco-
plastic constitutive model. Kang [9] reviewed the finite element sTnþ1 ¼ D : εTnþ1 (11)
implementation and engineering application of the developed
models in recent years. In the present article, the review of finite εTnþ1 is the tentative strain
element implementation and application to pressurized piping will
be the focus. εTnþ1 ¼ εnþ1  εpn (12)

Because of elastic isotropy and plastic incompressibility, taking

3.1. Finite element implementation of advanced constitutive model
the deviatoric part of tentative stress sTnþ1
Rate-independent constitutive models have been mainly 1
developed to simulate the ratcheting behavior of structures in the STnþ1 ¼ sTnþ1  tr ðsTnþ1 ÞI (13)
3
last several decades. According to ANSYS help manual [100], rate-
independent plasticity section provides the mathematical rela- According to radial return method, deviatoric stress on the yield
tionship that characterizes the elasto-plastic response of materials. surface corresponds to tentative deviatoric stress
Except for user specified behavior, part of the available material
Snþ1 ¼ STnþ1  2G½M2 Dεnþ1
p
behaviors in ANSYS are explained in greater details, such as bilinear (14)
and multilinear isotropic/kinematic hardening. Therefore, for user
Plastic strain increment Dεpnþ1
specified behavior, CH3 model is taken as an example in the
following to elaborate the implicit stress integration and radial rffiffiffi
3
return method as follows: Dεpnþ1 ¼ Dp½M1 nnþ1 (15)
2
(1) von Mises yield criterion where, Dl is the plastic multiplier, nnþ1 is the normal direction of
yield surface.
3
f ¼ ðs  aÞT ½M1 ðs  aÞ  sy ¼ 0 (4) rffiffiffi
2 Snþ1  anþ1 3 Snþ1  anþ1
nnþ1 ¼ ¼ (16)
where, [M1] is a non-zero auxiliary diagonal matrix, namely
kSnþ1  anþ1 k 2 so
[M1] ¼ diag [1,1,1,2,2,2].
nTnþ1 ½M1 nnþ1 ¼ 1 if fnþ1 ¼ 0:
(2) Flow rule
Substitution of Eq. (16) into Eq. (15) provides

3 hds$ni 3 Snþ1  anþ1

dεp ¼ n (5) Dεpnþ1 ¼ Dp (17)
2 h 2 sy
 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142 123

The calculation of plastic multiplier is detailedly listed in the dDsnþ1

following from CH3 model ¼ De  4G2 L1
nþ1 : Id (30)
dDεnþ1

X
M X
M  
2 dY
anþ1 ¼ an þ DaðiÞ (18) Lnþ1 ¼ 2GI þ
ðiÞ
Hnþ1 þ n
3 dp nþ1 nþ1
nþ1 5nnþ1
i¼1 i¼1 (31)
2 Ynþ1
þ ðI  nnþ1 5nnþ1 Þ
2 3 Dpnþ1
DaðiÞ
nþ1 ¼
p ðiÞ
r g ½M Dεnþ1  gi Dpanþ1 (19)
3 i i 2
H(i)
nþ1 in Eq. (31) is fourth-rank constitutive parameters. In order
to obtain H(i)
nþ1, differentiating Eq. (21) to have
2
aðiÞ aðiÞ Dp g D aðiÞ
nþ1 ¼ n þ ri i ½M2  εnþ1  i p nþ1
g (20)
ðiÞ
3 dqnþ1 ðiÞ
ðiÞ 2 ðiÞ
danþ1 ¼ rgq ½M dDεpnþ1 þ ðiÞ anþ1 (32)
  3 i i nþ1 2 q
2 nþ1
aðiÞ ¼ qðiÞ aðiÞ
n þ ri gi ½M2 Dεnþ1
p
(21)
nþ1 nþ1 3 Differentiation Eq. (22) and use of Eq. (15) allow dq(i)
nþ1 to be
expressed as
1 rffiffiffi
qðiÞ
nþ1 ¼ (22) ðiÞ 2 ðiÞ2
1 þ gi D p dqnþ1 ¼  q g nT : dDεnþ1
p
(33)
3 nþ1 i nþþ1

Eq. (21) subtracted from Eq. (14) gives Substitution of Eq. (33) into Eq. (32) provides

ðiÞ 2 ðiÞ ðiÞ

Snþ1  anþ1 ¼ STnþ1  2G½M2 Dεpnþ1 danþ1 ¼ rgq ð½M2   mnþ1 nTnþ1 Þ : dDεpnþ1 (34)
3 i i nþ1
XM   (23)
2
 qðiÞ a ðiÞ
þ r g ½M  Dε p Here,
i¼1
nþ1 n
3 i i 2 nþ1
rffiffiffi ðiÞ
Substitution of Eq. (17) into Eq. (23) provides ðiÞ 3 anþ1
mnþ1 ¼
2 ri
!,
X
M
Snþ1  anþ1 ¼ STnþ1  qðiÞ aðiÞ
nþ1 n
or
i¼1 rffiffiffi ðiÞ rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
! (24)
X
M ðiÞ 3 anþ1 ðiÞ 3 ðiÞ T ðiÞ
1 þ 3GDp=sy 
ðiÞ
ri gi qnþ1 Dp=sy mnþ1 ¼ ; anþ1 ¼ fa g ½M1 fanþ1 g
2 aðiÞ 2 nþ1
i¼1 nþ1

Snþ1  anþ1 should satisfy yield function, namely Eq. (4), thus For the CH3 model, therefore, fourth-rank constitutive param-
plastic multiplier is expressed as: eters H(i)
nþ1 of consistent tangent modulus has an expression

ðiÞ
" !T !#12 ðiÞ danþ1 2 i ðiÞ
3 T XM XM Hnþ1 ¼ ¼ rgq ð½M2   mnþ1 nTnþ1 Þ (35)
S 
ðiÞ
q a ðiÞ T
½M1  Snþ1 
ðiÞ ðiÞ
qnþ1 an sy dDεnþ1
p
3 i i nþ1
2 nþ1 i ¼ 1 nþ1 n i¼1
Dp ¼ P ðiÞ In order to obtain plastic modulus, substitution of Eq. (5) into Eq.
3G þ M i ¼ 1 ri gi qnþ1 (19) provides
(25)
ðiÞ dsnþ1 $nnþ1 ðiÞ
Iterative calculation is applied in the calculation of plastic danþ1 ¼ ri gi nnþ1  gi dpanþ1 (36)
h
multiplier. Iterative calculation is convergent if satisfying in the
following condition. Then, substitution of Eq. (6) into Eq. (36), and both sides
multiplied by nnþ1, provides
jDpðk þ 1Þ  DpðkÞj  eps (26) rffiffiffi
X
M
3 T XM
Iterative convergence represents tentative success. Then, plastic h ¼ ri gi  n ½M1  gi ai (37)
i¼1
2 i¼1
strain increment, stress and plastic strain will be provided

rffiffiffi For the CH3 model, therefore, plastic modulus h has an

3 expression
Dε ¼ Dpðk þ 1Þ ½M1 nnþ1
p
(27)
2 rffiffiffi
X
M
3 T XM
h ¼ ri gi  n ½M1  gi ai (38)
snþ1 ¼ D : ðεtr 2
nþ1  Dε Þ
p
(28) i¼1 i¼1

In order to improve the accuracy of advanced constitutive

εnþ1 ¼ εn þ Dεnþ1
p p p
(29) models, it is very important that the consistent tangent modulus is
derived. It is necessary to prescribe fourth-rank constitutive
The above process is shown in Fig. 9. parameters H(i)nþ1 of consistent tangent modulus. Therefore, based
Consistent tangent modulus is derived as on AF model, such as BurleteCailletaud model (BC) [101],
124 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142

Fig. 9. Flow diagram of radial return algorithm.

 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142 125

McDowell model (MD) [102,103], Abdel-KarimeOhno (AKO)
model [104], modified AKO model (KMAKO) [105], modified AKO
model (HMAKO) [106], which are described as the unified form of
AF model and shown in Table 2. The fourth-rank constitutive
parameters H(i)nþ1 and plastic modulus of these models are listed in
Table 3.
AF type models are commonly called by coupled modeling of
cyclic plasticity, i.e. the plastic modulus is evaluated using the
consistency condition, kinematic hardening rule, and flow rule.
Thus, it is important to calculate the consistent tangent modulus to
preserve the quadratic convergence rate at a global level of the
finite element calculation, after convergence of the local iteration
for the numerical scheme. The initial consistent tangent modulus
forms the structural stiffness matrix to calculate the trial strain
increment to start the numerical calculation. Then, the subroutines
are implemented with adaptive number of sub increments for
numerical calculation.
ANSYS or ABAQUS has an open architecture to compile the
user’s own constitutive models as a subroutine. According to the
above algorithm processes, user subroutine is compiled by user
own. The subroutine USERMAT or USERPL in ANSYS or UMAT in
ABAQUS is used for linking the plasticity routine.

Table 2
Overview of some kinematic hardening rules.

Constitutive model Description Unified form

AF model [99] AF model is the most famous nonlinear kinematic hardening model. dp(i) ¼ dp
 
vf vf
BC model [101] In 1986, Burlet and Cailletaud modified AF model in order to predict aðiÞ dpðiÞ ¼ aðiÞ : dp
vs vs
the uniaxial ratcheting behavior. Dynamic recovery of AF model was
changed into transient radial recovery.
CH3 model [72,73] In 1991, in order to better predict ratcheting effect, Chaboche added dp(i) ¼ dp, i ¼ 1,2,3
a fourth hardening rule with a concept of ‘threshold’ in CH3 model [70,75].
ai
The ‘threshold’ value of back stress was used to determine whether dpðiÞ ¼ h1  idp i ¼ 4
f ðai Þ
dynamic recovery works.
aðiÞ
OW-I model [62] In 1993, Ohno and Wang proposed a dynamic recovery superposition dpðiÞ ¼ Hðfi Þhdεp : i
ai
model. Ratcheting behavior was simulated with piecewise linear or nearly
piecewise linear kinematic hardening rule. !mi !mi
ai ai aðiÞ
OW-II model [63] Ohno and Wang-II replaced H(fi) in OW I model with . The dynamic dpðiÞ ¼ hdεp : i
ri ri ai
recovery became significant nonlinearly.
!mi
ai aðiÞ
MD model [102,103] McDowell investigated that the parameter mi in OW-II model did not have dpðiÞ ¼ hdεp : i
ri ai
the versatility for the prediction of uniaxial and multiaxial ratcheting effect.
Thus, the parameter mi was
dεp ai Bi
modified in order to simulate the multiaxial ratcheting effect while keeping mi ¼ Ai h : i
dp ai
the capability of predicting uniaxial ratcheting effect.
!mi
ai
JS model [86,87] Jiang and Sehitoglu indicated that CH3 model and OW model poorly predicted dpðiÞ ¼ dp
ri
the multiaxial ratcheting effect, due to lack of multiaxial parameter. The
predicted results of JS model were less than those of OW model II model. !
dεp ai
mi ¼ A0i 2  : A0i ¼ Qi ð1 þ ax ebx RM Þ
dp ai
" #
vf aðiÞ
AKO model [104] In 2000, Abdel-Karin and Ohno proposed another kinematic hardening rule dpðiÞ ¼ mi þ Hðfi Þh :  mi i dp
vs ri
with a blend of OW [62] and AF model [99]. The parameter mi can improve
the prediction of multiaxial ratcheting effect, but shakedown can occur in
uniaxial ratcheting effect.
MCH3 model [83] In 2002, multiaxial parameter d0 was introduced into CH3 model [72] by aðiÞ dpi ¼ ½d0 aðiÞ þ ð1  d0 ÞðaðiÞ : nÞndp i ¼ 1; 2; 3
Bari and Hassan.
ai
aðiÞ dpðiÞ ¼ ½d0 aðiÞ þ ð1  d00 ÞðaðiÞ : nÞnh1  idp i ¼ 4
f ðai Þ
!mi
ai 0 0 aðiÞ
CJ model [84] In 2004, CJ combined BurleteCailletaud model and OW-II model, with aðiÞ dpðiÞ ¼ ½d aðiÞ þ ð1  d ÞðaðiÞ : nÞnhdεp : i
ri ai
multiaxial parameter d0 into the new model.
0 0 0
dd ¼ bðddN  dd Þdp
" #
vf aðiÞ
KMAKO model [105] Non-proportional and temperature-dependent parameter was incorporated dpðiÞ ¼ mi þ Jðp; TÞHðfi Þh :  mi i dp
vs ri
into AKO model by Kang et al., in order to simulate ratcheting behavior under
high temperature and non-proportional conditions. Jðp; TÞ ¼ JN ðTÞ þ ½1  JN ðTÞexp ðbpÞ
vffiffiffiffiffiffiffiffiffiffi
u
m0 u s : s,
mi ¼ m ¼ ; F ¼ 1  t
^ ,
1 þ aF s:s
!mi
vf aðiÞ ci ai aðiÞ
CJK model [85] In 2005, Chen et al. modified the OW (II), multiplying a factor to dynamic dpðiÞ ¼ h : i hdεp : i
v s ai ri ai
recovery term which is dependent on noncoaxiality of the plastic strain rate
and back stress.
" #
vf aðiÞ vf ai c
HMAKO model [106] Halama implemented large transient effect and non-proportion into AKO dpðiÞ ¼ mi þ Hðfi Þh :  mi i dp; mi ¼ hh : i
vs ri vs ai
model.
dh ¼ uðhN  hÞdp; dh ¼ dh1 þ dh2

dh1 ¼ u1 ðhN1  h1 Þdp; dh2 ¼ u2 ðhN2  h2 Þdp

mðpÞ ¼ h ¼ hN1 þ ðh01  hN1 Þeu1 p þ hN2 þ ðh02  hN2 Þeu2 p

 126
Table 3
Fourth-rank constitutive parameters and plastic modulus of various models.

Constitutive model Fourth-rank constitutive parameters of consistent tangent modulus Plastic modulus
rffiffiffi
ðiÞ 2 i ðiÞ 3
AF model [99] Hnþ1 ¼ ri gi qnþ1 ð½M2   mnþ1 nTnþ1 Þ h ¼ rg  ga
3 2
rffiffiffi
ðiÞ 2 ðiÞ 3 T
n ½M1 gða : nÞn
i
BC model [101] Hnþ1 ¼ rgq ð½M2   mnþ1 nTnþ1 Þ h ¼ rg 
3 i i nþ1 2
rffiffiffi
ðiÞ 2 ðiÞ PM 3 T P
i ¼ 1 gi ai ; i ¼ 1; 2; 3
i
CH3 model [72,73] Hnþ1 ¼ ri gi qnþ1 ð½M2   mnþ1 nTnþ1 Þ; i ¼ 1; 2; 3 h ¼ i ¼ 1 ri gi  n ½M1  M
3 " ! 2
ðiÞ 2 i ðiÞ a rffiffiffi
Hnþ1 ¼ ri gi qnþ1 M2   mnþ1 nTnþ1 h1  i i ; i ¼ 4
3 f ðai Þ PM 3 T P ai
i ¼ 1 gi ai h1 
n ½M1  M

X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
h ¼ i ¼ 1 ri gi  i; i ¼ 4
2 f ðai Þ
rffiffiffi
ðiÞ 2 ðiÞ ðiÞ PM 3 T aðiÞ X M
OW-I model [62] Hnþ1 ¼ rgq
i
ð½M2   Hðfi Þmnþ1 5mnþ1 Þ h ¼ i ¼ 1 ri gi  n ½M1 Hðfi ÞhnT : i ga
3 i i nþ1 2 ai i ¼ 1 i i
" !mi ! rffiffiffi !mi
ðiÞ 2 ai ðiÞ ðiÞ PM 3 T a aðiÞ XM
OW-II model [63]
i
Hnþ1 ¼ ri gi qnþ1 M2   mnþ1 5mnþ1 h ¼ i ¼ 1 ri gi  n ½M1  i hnT : i ga
3 ri 2 ri ai i ¼ 1 i i
" !mi ! rffiffiffi !mi
ðiÞ 2 a ðiÞ ðiÞ PM 3 T a aðiÞ XM
MD model [102,103]
i
Hnþ1 ¼ ri gi qnþ1 M2   i
mnþ1 5mnþ1 h ¼ i ¼ 1 ri gi  n ½M1  i hnT : i ga
3 ri 2 ri ai i ¼ 1 i i
dεp ai Bi dεp ai Bi
mi ¼ Ai h : i mi ¼ A i h : i
dp ai dp ai
" !mi ! rffiffiffi !mi
ðiÞ 2 ðiÞ ðiÞ ai PM 3 T a XM
JS model [86,87] Hnþ1 ¼ ri gi qnþ1 M2   mnþ1 nTnþ1 h ¼ i ¼ 1 ri gi  n ½M1  i gi ai
3 ri 2 ri i¼1
! !
dεp ai dεp ai
mi ¼ A0i 2  : mi ¼ A0i 2  :
dp ai dp ai

A0i ¼ Qi ð1 þ ax ebx RM Þ A0i ¼ Qi ð1 þ ax ebx RM Þ

" " #! rffiffiffi " #
ðiÞ 2 ðiÞ ðiÞ aðiÞ PM 3 T PM T :a
ðiÞ
AKO model [104] Hnþ1 ¼ ri gi qnþ1 M2   mnþ1 nTnþ1 mi þ Hðfi ÞhnTnþ1 :  mi i h ¼ i¼1 r i g i  n ½M1  i¼1 g i a i mi þ Hðfi Þhn  mi i
3 ri 2 ri
rffiffiffi
ðiÞ 2 ðiÞ 0 ðiÞ 0 ðiÞ PM 3 T P 0 ðiÞ 0
MCH3 model [83] Hnþ1 ¼ rgq ð½M2   nTnþ1 ½d mnþ1 þ ð1  d Þðmnþ1 : nnþ1 Þnnþ1 Þ; i ¼ 1; 2; 3 h ¼ i ¼ 1 ri gi  i ¼ 1 g i ½d a
n ½M1  M þ ð1  d ÞðaðiÞ : nÞn; i ¼ 1; 2; 3
3 i i nþ1 2
" " #! rffiffiffi
ðiÞ 2 ðiÞ 0 ðiÞ 0 ðiÞ ai PM 3 T P 0 ðiÞ 0 a
Hnþ1 ¼ rgq M2   nTnþ1 d mnþ1 þ ð1  d Þðmnþ1 : nnþ1 Þnnþ1 h1  i ; i ¼ 4 h ¼ i ¼ 1 ri gi  i ¼ 1 g i ½d a
n ½M1  M þ ð1  d ÞðaðiÞ : nÞnh1  i i; i ¼ 4
3 i i nþ1 f ðai Þ 2 f ðai Þ
" # !mi ! rffiffiffi !mi
ðiÞ 2 ðiÞ 0 ðiÞ 0 ðiÞ ai ai PM 3 T P 0 ðiÞ 0 a a 0 0 0
CJ model [84] Hnþ1 ¼ rgq M2  nTnþ1 ½d mnþ1 þ ð1  d Þðmnþ1 : nnþ1 Þnnþ1  hnTnþ1 :
ai
i h ¼ i ¼ 1 ri gi  i ¼ 1 g i ½d a
n ½M1  M þ ð1  d ÞðaðiÞ : nÞn i hnT : i i dd ¼ bðddN  dd Þdp
ai
3 i i nþ1 ri 2 ri
0 0 0 0 0 0
dd ¼ bðddN  dd Þdp dd ¼ bðddN  dd Þdp
" " #! rffiffiffi " #
ðiÞ 2 ðiÞ aðiÞ PM 3 T P T a
ðiÞ
KMAKO model [105] Hnþ1 ¼ rgq Jðp; TÞ M2   mðiÞ T
nþ1 nnþ1
mi þ Hðfi ÞhnTnþ1 :  mi i h ¼ i ¼ 1 ri gi  i ¼ 1 gi ai Jðp; TÞ mi þ Hðfi Þhn :
n ½M1  M  mi i
3 i i nþ1 ri 2 ri
" !mi ! rffiffiffi !mi
ðiÞ 2 ðiÞ ðiÞ ai aðiÞ ci T aðiÞ PM 3 T P ai aðiÞ ci T aðiÞ
CJK model [85] Hnþ1 ¼ ri gi qnþ1 M2   mnþ1 nTnþ1 hnTnþ1 : i hnnþ1 : i h ¼ i ¼ 1 ri gi  i ¼ 1 g i ai
n ½M1  M hnT : i hn : i
3 ri a i a i 2 ri ai ai
" " #! rffiffiffi " #
ðiÞ 2 ðiÞ ðiÞ a ðiÞ PM 3 T P T a
ðiÞ
HMAKO model [106] Hnþ1 ¼ ri gi qnþ1 M2   mnþ1 nTnþ1 mi þ Hðfi ÞhnTnþ1 :  mi i h ¼ i ¼ 1 ri gi  i ¼ 1 gi ai mi þ Hðfi Þhn :
n ½M1  M  mi i
3 ri 2 ri
 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142 127

4. Application of FEA
The advanced constitutive models have been so far widely used
to simulate the ratcheting behavior of the pipe systems such as
straight pipe and elbow. Return mapping algorithm for structural
ratcheting simulation has been commonly conducted by
researchers using the commercial finite element code, such as
ANSYS, ABAQUS, etc. Table 4 lists the various structures which have
been simulated with constitutive models implemented into finite
element software.
4.1. FEA of ratcheting in straight pipe
Because of the symmetry in loading and geometry, only one-
quarter of the entire structure was built in the commercial finite
element software when analyzing the ratcheting behavior of the
straight pipe. Boundary conditions and loading of the model were
detailedly given in references [16,17,28,29]. The straight pipe was
symmetric about XeY plane at the mid-pipe section and also
symmetric to YeZ plane. In addition to the symmetric displace-
ment constraints, displacement in the y-direction was applied to
the central point of the end plate. The inside surface of the pipe and
end plate were subjected to internal pressure. For three-point or
four-point bending apparatus (Fig. 1(a)), reversed bending load in
the y-direction was distributed to nodes at the position corre-
sponding to the central line of loading. For the apparatus in Fig. 1(c),
the straight pipe was subjected to rotation at ends.
4.1.1. FEA with existing plasticity model in finite element software
In order to simulate the ratcheting behavior of various struc-
tures, bilinear kinematic hardening (BKH), multilinear kinematic
hardening (MKIN) and the nonlinear kinematic hardening model
developed by CH3 model are commonly used, which are currently
available in the commercial finite element program. For example,
Gau [14] verified against three sets of experimental results of
straight pipe using the modified BKH model in ANSYS program and
the unmodified ANSYS and ABAQUS code. The modified ANSYS
program showed a significant improvement over the unmodified
ANSYS and ABAQUS code. The unmodified ANSYS and ABAQUS
code with lower elasticeplastic modulus greatly overpredicted
ratcheting strain rate.
Suzuki [107,108] and Fujiwaka et al. [109] simulated the ratch-
eting behavior of the piping component under dynamic analysis
and under static analysis with the general-purpose FEM Code
ABAQUS. The results indicated that the momenterotation rela-
tionship and time history of rotation showed good agreement with
the experimental results.
Kulkarni [12,13] predicted the ratcheting strain of the straight
pipe under three-point and four-point bending modes with CH3
model in ANSYS. The results showed that ratcheting strain only
occurred in the circumferential direction, whereas no strain
accumulation occurred in the longitudinal direction for three-
point and four-point bend analysis. For three-point bending, the
CH3 model overpredicted the ratcheting strain, as shown in
Fig. 10(a). Whereas, for four-point bending, it was seen from
Fig. 10(b) that the CH3 model underpredicted ratcheting strain in
the initial cycles, but CH3 model performed the better in higher
number of cycles.
Zakavi et al. [110] studied the ratcheting behavior of the pres-
surized carbon steel (BS4360-43A) and stainless steel (304L)
straight pipe subjected to the seismic bending moment with the
nonlinear combined hardening model using finite element code,
ABAQUS. The finite element results were compared with those
obtained from the experiments to evaluate the capability of the
proposed AF model with isotropic/kinematic hardening rule to
predict the cyclic loading behavior of the straight pipe, as shown in
Fig. 11. The results showed that the hoop-ratcheting strain pre-
dicted by FE analysis to be close to that found experimentally in all
cases with M=MP0:2  1. Otherwise, FE analysis gave overestimated
values compared with the experimental data.

Table 4
Review of FEA application of some constitutive models to piping structures.

Scholars Constitutive model Structural type Software

Guionnet [23,92] Modified AF model Straight pipe ANSYS
Hassan et al. [82,111] AFeDP model Straight pipe ANSYS
Fujiwaka et al. [33,109] BKH/MKIN model Straight pipe/elbow/tee ABAQUS
Zakavi et al. [110,124] Nonlinear isotropic/kinematic hardening model Straight pipe ABAQUS
Gau [14] BKH/AFeDP model Straight pipe ANSYS/ABAQUS
Kulkarni et al. [13] CH3 model Straight pipe ANSYS
Chen et al. [28,29,49,50,52,123] BKH/CH3/OW-II/MJS/CJK model Straight pipe/Elbow pipe ANSYS
Rahman et al. [16,17] BKH/MKIN/CH3/MCH3/OW-II/AKO/JS model Straight pipe/Elbow pipe ANSYS
Rojí
ceka and Halama [113] KINH/CH3/MAKO model Straight pipe ANSYS
Jiao and Kyriakides [114] DafaliasePopov model Straight pipe ANSYS
Garud et al. [115] MKIN model Elbow pipe ANSYS
Kulkarni et al. [12,13] BKH model Elbow pipe ANSYS
Zhao [117] BKH model Elbow pipe ANSYS
Hassan et al. [15,51,82,112] CH3/OW/Guionnet/modified DafaliasePopov model Straight pipe/Elbow pipe ANSYS/ABAQUS
DeGrassi et al. [41e43] BKH/MKIN/CH3 model Straight pipe/Elbow pipe ANSYS
Stearn et al. [46] BKH model Straight pipe/elbow pipe ANSYS
Shalaby and Younan [118] Modified Guionnet model Straight pipe/Elbow pipe ANSYS
Fujiwaka et al. [33,109] BKH/MKIN model Elbow pipe ABAQUS
Wilkins [48] Multilinear model using kinematic hardening or Elbow pipe ANSYS
isotropic hardening rule
Suzuki et al. [107,108,119e121] BKH/AFeOW model Elbow pipe ABAQUS
Ichihashi [34] BKH/AFeOW model Elbow pipe ABAQUS
Balan and Redekop [122] BKH model Elbow pipe ADINA
Moreton et al. [20,25,26,131] (perfectly plastic model) PP/BKH model Elbow pipe ABAQUS
Zehsaz et al. [27] Nonlinear isotropic/kinematic hardening model Cylinder ABAQUS
Zakavi et al. [110,124] Nonlinear isotropic/kinematic hardening model Cylinder ABAQUS
Weib et al. [125] OW model Welded vessel ANSYS
Yang et al. [53,54] BKH/CH3/OW-II/MJS/CJK model Lateral nozzle of cylinder ANSYS
Dahlberg and Segle [126] BKH/MKIN/CH3 model Cylinder ANSYS
Couterot et al. [58] CH3 model Socket welding Finite element EDF code
128 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
Fig. 10. Circumferential ratcheting strain vs. number of cycles [13]: (a) Four-point bend
test, (b) Three-point bend test.
4.1.2. FEA with advanced constitutive model implemented into finite
element software
To well understand the ratcheting behavior of various structures
under cyclic bending loading and constant internal pressure,
advanced constitutive models have been implemented into finite
element software, such as ANSYS and ABAQUS. Guionnet [23]
developed a new constitutive equation which was based on
a generalized ArmstrongeFrederick equation for the kinematic
variable whose coefficients were functions of both instantaneous
and accumulated plastic strain. The experiments described austen-
itic stainless steel (17-12 SPH) tubular specimens subjected to cyclic
torsional loading under constant tensile stress at 600  C. Compari-
sons between experimental and simulated results showed that the
model reasonably well predicted not only ratcheting but also
hysteresis loops and the cyclic hardening curve. Then, based on an
AF rule in conjunction with the DruckerePalgen plastic modulus
equation, an improved cyclic plasticity model (AFeDP) had been
incorporated into an ANSYS user subroutine. The modified ANSYS
code was first verified by Zhu et al. [111] using a set of data from
a tube subjected to cyclic axial strain and constant internal pressure.
Then, Hassan et al. [112] also showed the AFeDP model in ANSYS
code compared to the experimental data [20], and the results indi-
cated that the AFeDP model was able to simulate the ratcheting
strain of the tube more accurately than the original ANSYS code.
To improve the accuracy of prediction of ratcheting strain of the
piping system, Gao et al. [28,29] applied five constitutive models to
Fig. 11. Large scale for experimental and FE ratchet strains results [110]: (a) Stainless
steel (304L), (b) Carbon steel (BS4360-43A).
predict the ratcheting response of the low carbon steel 20# straight
pipe, including BKH model, CH3 model, OhnoeWang (OW) model,
CheneJiaoeKim (CJK) model and Modified JiangeSheitoglu (MJS)
model, as shown in Fig. 12. These constitutive models were
implemented into ANSYS. By comparison with the experimental
data, it was indicated that BKH model underpredicted the ratch-
eting strain of straight pipe, the CH3 model overpredicted experi-
mental results. The ratcheting strain simulated by the OW model
for straight pipe experiment matched initial well, but increased
with the number of cycles. The MJS model (Bi ¼ 1) gave an
acceptable simulation for different loading parameters. Suitable
value of ci should be improved ratcheting strain rate, the predicted
ratcheting strain by CJK model can match well with experimental
data. Then, Rahman et al. [16,17] investigated the performance of
the constitutive models in simulating the global- and local-ratch-
eting responses of the straight piping subjected to the internal
pressure and cyclic bending loading. BKH, MKIN, CH3, modified
Chaboche (MCH3) model, OW model, CheneJiao (CJ) model and
Abdel-KarimeOhno (AKO) model were used, as given in Fig. 13.
Structural simulation with BKH model failed to produce the steady
 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142 129

Fig. 12. Structural response simulation for straight pipe with [28,29]: (a) BKH model, (b) CH3 model, (c) OW model, (d) MJS model, (e) CJK model.

ratcheting rate at higher cycles. The CH3 model agreed well with
experimental results with a slight overprediction. MCH3 model,
OW, CJ model and AKO model exhibited the same deficiency as the
CH3 model.
Rojí
ceka and Halama [113] employed multilinear kinematic
hardening (KINH) model, CH3 model in ANSYS and modified Abdel-
KarimeOhno (MAKO) model to predict the ratcheting strain of the
straight pipe and compared the numerical simulation with the test
data published in the reference [17]. KINH model, CH3 model are
available in ANSYS. MAKO model was implemented into ANSYS by
writing user subroutine in FORTRAN language. Comparison
between the numerical solution and the experimental observations
of the straight pipe under cyclic loading in elasto-plastic domain
indicated that the CH3 model overpredicted the ratcheting strain in
the case of low ratcheting rate, as shown in Fig. 14(a) and (b). For
higher ratcheting rate it gave better results than other tested
models. For uniaxial material test, it was seen from Fig. 14(c) and (d)
that the MAKO model gave the best results of all evaluated cyclic
plasticity models. However, for multiaxial ratcheting test, in
simulations of experiments the MAKO model did not describe
material behavior correctly. It was valid for trends of curves. All
tested material models showed significant stabilization of strain
response in multiaxial ratcheting case, which is in conflict with
experimental results.
Jiao and Kyriakides [114] used the DafaliasePopov two-surface
nonlinear kinematic hardening model to evaluate the ratcheting
130 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142

Fig. 13. Mean of circumferential strain and amplitude of circumferential strain for straight pipe at top [16,17]: (a) Mean of circumferential strain at top with BKH, MKIN, CH3 models,
(b) Amplitude of circumferential strain at top with BKH, MKIN, CH3 models, (c) Mean of circumferential strain at top with Chaboche models, (d) Amplitude of circumferential strain
at top with Chaboche models, (e) Mean of circumferential strain at top with OhnoeWang models, (f) Amplitude of circumferential strain at top with OhnoeWang models, (g) Mean
of circumferential strain at top with Abdel-KarimeOhno models, (h) Amplitude of circumferential strain at top with Abdel-KarimeOhno models.

behavior of the pressurized straight pipes under axial cyclic
loading. The analysis confirmed the experimental observation that
when cycling started well before the onset of wrinkling, wrinkles
appeared and grew at an accelerated rate when the strain reached
the critical strain under monotonic compression. The numerical
model was also used to simulate thermomechanical cycling of
axially restrained pipelines due to periodic startup and shutdown,
as shown in Fig. 15. The results showed that the ratcheting strain
occurred in hoop direction and that wrinkling also increased
gradually. It was shown that for realistic pressure and temperature
levels the pipes considered were not in danger of either bursting or
developing localized collapse provided the number of cycles
applied was less than a few hundred.
4.2. FEA of ratcheting in elbow pipe
The elbow specimen was commonly modeled with shell
elements. The elbow specimen geometry and loading conditions
 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142 131

Fig. 14. Multiaxial ratcheting test for three material models [113]: (a) Mean in-plane circumferential strain peak (Experiment 4), (b) Amplitude in-plane circumferential strain peak
(Experiment 4), (c) Mean in-plane circumferential strain peak (Experiment 5), (d) Amplitude in-plane circumferential strain peak (Experiment 5).

prescribed were doubly symmetric, thus only one-quarter of the
specimen was modeled in FEA, internal pressure was applied to the
structure inner surface and the reversed bending load in y-direc-
tion was imposed at the central point of the pipe end. Each shell
element section of the model represented the elbow plus an
appropriate length of straight pipe at each end. It was sometimes
assumed that the remaining straight pipes would remain elastic, so
they were represented by the two-dimensional elastic beam
elements (BEAM3). The transition from the straight pipe shell
model to the straight pipe beam model was achieved through
a series of rigid beam elements arranged in a wheel spoke
configuration.
4.2.1. FEA with existing plasticity models in finite element software
The currently available plasticity models in finite element code
including BKH, MKIN and CH3 model mentioned above, have also
been used to simulate the ratcheting behavior of elbow piping. For
example, Garud et al. [115] studied elbow piping response to
dynamic loads using MKIN model in ANSYS. Results demonstrated
that with internal pressure the elbow exhibited much greater local
strain accumulation and larger elbow opening compared to the
unpressurized case. The predicted results were in good agreement
with those of tests. Stearns et al. [46] studied the ratcheting effect of
the elbow pipe using LKH model with ABAQUS. From the results,
the worse mean ratcheting strain of 3 inch and 6 inch elbow’s bend
radii occurred at 90 from the intrados (0 ) on the inside wall,
which was also the location of the largest cyclic strain. However, for
the small bend radii elbow (3 inch), the final failure was at 90 from
the intrados, and for large bend radii the failure was at intrados.
Then, Kobayashi and Fujiwaka [116] applied LKH model in FEA to
simulate pressurized elbows subjected to an in-plane bending
moment by dynamic loads (that simulated seismic excitation). Both
shell and beam elements were used to examine the response of the
piping when the mechanical ratcheting at an elbow subjected to
the internal pressure was induced by dynamic excitation. The
analytical predictions of the hoop ratcheting at the top of the elbow
were in good agreement with the experiment results. Zhao [117]
used shell element model to predict the ratcheting behavior of
the elbow pipe subjected to an internal pressure and multiaxial
loading of prescribed stochastic displacement time series obtained
by the random vibration analyses. The bilinear stressestrain
constitutive model was used in SOLVIA code. It was found that
the largest ratcheting strain happened in the hoop direction at
intrados on the symmetrical cross-section of the elbow pipe.
Shalaby and Younan [118] investigated the large strain and
stress analysis for the elbow pipe subjected to in-plane bending
moments using finite element code, ABAQUS. The elastic-perfectly
plastic (EPP) model was used to eliminate the effect of strain
hardening. The inelastic behavior of the elbow pipe was found quite
different under in-plane closing and opening moments due to the
132 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142

Fig. 15. Comparison of measured and calculated εqpeN responses for the three pressurized experiment [114]: (a) Wrinkle amplitude vs. N, (b) average hoop strain vs. N for three DT
(D/t ¼ 28.4, P ¼ 0.7Po).

large deformations involved and the difference in geometric stiff-
ness of the elbow pipe cross-section. In the closing case much less
loads and deformations were attained in comparison with the
opening case. Fujiwaka et al. [33] also performed the elbow pipe
experiments and FEA using shell element. Multilinear stressestrain
relationship material model and bilinear material model with the
kinematic hardening theory were respectively used for the static
monotonic loading analysis and the static cyclic loading analysis.
The analysis results closely agreed with the experimental results.
Also, it was found that maximum strain developed in the hoop
direction of the inner surface of the elbow pipe.
JNES/NUPEC Ultimate Strength Piping Test Program conducted
a multi-year test and a lot of tests to investigate the deformation
behavior of typical NPP (Nuclear Power Plant) piping systems under
large seismic loads. For example, DeGrassi et al. [41], DeGrassi and
Hofmayer [42] and DeGrassi [43] used classical BKH and MKIN
models as well as CH3 model to predict the elasto-plastic response
for one of NUPEC’s simplified piping system seismic tests, as shown
in Fig. 16. Comparisons of analysis results for each plasticity model
against test results for a static cycling elbow component test and for
a simplified piping system seismic test were presented. Results
demonstrated that BKH model overpredicted the ratcheting strain
at the end of test and achieved shakedown in first several cycles.
MKIN model underpredicted the accumulated strain. CH3 model
produced the best match to test results in terms of predicting both
the ratcheting strain and shakedown.
Recently, Chen and Chen [52] employed the CH3 model in
ANSYS software to simulate the distribution of ratcheting strain in
the elbow. It is shown in Fig. 17 that the simulated results generally
agreed well with the test results. Moreover, the influence of wall
thickness and element type was discussed by considering two sets
of elbow wall thickness and two types of elements, respectively.
The wall thickness sets included the real dimension obtained by
cutting open the elbow after the test. Solid and shell type elements

Fig. 16. Elbow static cycling test SE-4-hoop strain vs. number of cycles at elbow flank [42]: (a) test, (b) BKIN model, (c) MKIN model, (d) CHAB model.
 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142 133

Fig. 17. Comparison of the prediction results and experimental results [52]: (a) At intrados, (b) At flank.

were used for different types of wall thickness entity models.
Compared with the real-dimension thickness entity model, the
equal-wall thickness one yielded overprediction of ratcheting
strain, as shown in Fig. 18. Although the application of shell element
could be less time-consuming, the predicted results seemed to be
more conservative, due to the fact that the model of shell elements
yielded higher ratcheting strain than that of solid elements.
4.2.2. FEA with advanced constitutive model implemented into
finite element software
The above advanced constitutive models have been imple-
mented into finite element software in order to simulate the
ratcheting behavior of the structural components. A testing facility
for quasi-static cyclic and seismic testing of straight and elbow
piping components was developed by Hassan and Matzen [15]. In
order to simulate the ratcheting behavior of piping components,
CH3 model, OW model, Guionnet model and modified Dafaliase
Popov model were implemented into ANSYS or ABAQUS by Has-
san and Matzen [15]. But Hassan and Matzen [15] gave only some
introductions of these constitutive models, and did not provide the
specific simulation results of piping systems.
Wilkins [48] investigated the deformation behavior of the butt-
welded elbows under a variety of loading conditions. Multilinear
model using isotropic and kinematic hardening rules in ANSYS was
applied to simulate the deformation behavior of the structures with
SHELL element 181. In general, the correlation was quite good for
both the global behavior and the local strain measures. The first
quarter cycle performance of both analyses was quite good, but as
deformation continued the isotropic results deviated significantly
while the kinematic hardening model results remained quite good.
With an increase in loading rate, an increased deviation from the
analytical results was observed. Strain rate hardening effect was
taken into account which improved greatly the above correlation.
The LKH rule and advanced constitutive model AFeOW model
were used to simulate the ratcheting behavior of the piping
components. For example, Suzuki et al. [119e121] simulated the
strains at the surface of the elbow and nozzle using ABAQUS shell
element and two types of hardening rule, the LKH and the
nonlinear kinematic hardening rule (AFeOW). Comparison of
simulation results and experimental results indicated that the
strain range on the elbow in the quasi-static loading test by the LKH
rule agreed well with the test result. Furthermore, the simulation
by the AFeOW rule could well predict accumulation of the hoop
strain in the seismic loading test. Then, Ichihashi [34] also reported
that the strain range on the elbow in the quasi-static loading test by
the LKH rule was well predicted against the test result. Moreover,
Balan and Redekop [122] simulated the response of the elbow
specimen under cyclic bending and internal pressure with bilinear

Fig. 18. Comparison of the prediction results using shell elements and solid element [52]: (a) At intrados, (b) At flank.
134 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142

plasticity model in the finite element code ADINA. Simulated
results demonstrated that the shakedown phenomenon of the
circumferential strain at the flank occurred.
Advanced constitutive models were extensively applied by Chen
et al. [28,29,49,50,123] and Hassan [15e17,51,82,112]. For example,
Chen et al. [29,49,50] performed ratcheting simulation by ANSYS
into which OW-II, MJS and CJK rules were implemented using user
subroutine. The ratcheting behavior of elbow piping was simulated
by BKH, CH3, OW-II, MJS and CJK model. Comparison of simulation
and experiments showed that CJK model presented simulation
more reasonably, BKH, CH3, OW-II and MJS rules overpredicted
with experimental results, as shown in Fig. 19. Rahman et al.
[16,112] evaluated the performance of the constitutive models in
simulating the global- and local-ratcheting responses of elbow
piping subjected to the internal pressure and cyclic bending. BKH,
MKIN, CH3, MCH3, OW, CJ and AKO model were used to simulate
the ratcheting behavior of the pipe components, respectively. It was
shown in Fig. 20 that structural simulation with BKH model and
MKIN model failed to produce the steady-ratcheting rate at higher
cycles. Simulation results of CH3 model performed well with
experimental results, but slightly overpredicted, as given in Fig. 21.
Figs. 21e23 showed that modified CH3 model, OW and CJ model

Fig. 19. Mean and amplitude of axial and circumferential strain with [29,49,50]: (a) BKH model, (b) CH3 model, (c) OW-II model,(d)MJS model, (e) CJK model.
 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142 135

Fig. 20. Mean and amplitude of axial and circumferential strain with bilinear and multilinear models [16,112]: (a) Mean of circumferential strain at flank, (b) Amplitude of
circumferential strain at flank, (c) Mean of circumferential strain at flank, (d) Amplitude of circumferential strain at flank.

displayed the same deficiency as CH3 model. These models could
simulate ratcheting rates of ovalization and strain at flank reason-
ably, with a small overprediction. A major deficiency of all the
models was that they all simulated negative ratcheting at intrados
or extrados or at both locations, while the experimental-ratcheting
strain occurred in the positive direction.
4.3. Other structures and loading types
In addition to the representative structures analyzed in the above
two sections, the ratcheting behavior of some other commonly-used
structures in the engineering applications were also studied. To
simulate the ratcheting behavior of the cylinder, Moreton et al. [26]
used ELBOW 31B element type in finite element code, ABAQUS. An
EPP material idealization and linear kinematic hardening were
employed. Both the results obtained from the finite element work
and a previously available analytical solution overestimated the
experimentally obtained ratcheting rates considerably.
The nonlinear isotropic/kinematic (combined) hardening model
was used to evaluate the ratcheting behavior of the cylinder under
aforementioned loading condition by Zehsaz et al. [27] and Zakavi
et al. [124]. Both the finite element results and those obtained from
experiments showed that the initial rate of ratcheting was large and
then it decreased with the increasing cycles. The results also

Fig. 21. Mean and amplitude of axial and circumferential strain with Chaboche and modified Chaboche models [16,112]: (a) Mean of circumferential strain at flank, (b) Amplitude of
circumferential strain at flank, (c) Mean of circumferential strain at flank, (d) Amplitude of circumferential strain at flank.
136 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142

Fig. 22. Mean and amplitude of axial and circumferential strain with OhnoeWang and modified OhnoeWang models [16,112]: (a) Mean of circumferential strain at flank, (b)
Amplitude of circumferential strain at flank, (c) Mean of circumferential strain at flank, (d) Amplitude of circumferential strain at flank.

indicated that FE method and analytical solution gave over-
estimated values compared with the experimental data. The
ratcheting data using FEA also showed the hoop strain ratcheting
decreased with the increasing of frequencies and spacing from the
resonant frequency.
Weib et al. [125] focused on the design by analysis of welded and
non-welded pressure vessel components with respect to combined
ratcheting and low cycle fatigue damage mechanisms. An algo-
rithm for the determination of component fatigue curves was
proposed.
The ratcheting behavior of the lateral nozzle of cylinder made of
20# carbon steel was simulated by Yang et al. [53,54] using ANSYS
code with five typical cyclic plasticity models, which showed that
OW model and its modified models improved the prediction of the
ratcheting strain, and CH3 model overpredicted with the experi-
mental results, as shown in Fig. 24. The predicted results of A0, A1, and
A2 matched well with the experimental results. It was found that the
maximum ratcheting strain occurred at the nozzle side of the acute
angle region in the symmetrical plane for the discussed structure in
multi-directions, and mainly in longitudinal symmetrical plane.

Fig. 23. Mean and amplitude of axial and circumferential strain with Ohno-Wang and Abdel-KarimeOhno models [16,112]: (a) Mean of circumferential strain at flank, (b) Amplitude
of circumferential strain at flank, (c) Mean of circumferential strain at flank, (d) Amplitude of circumferential strain at flank.
 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142 137

In addition, Dahlberg and Segle [126] reviewed the thermal-
ratcheting behavior of Bree cylinder (type A and type B) and the
mechanical ratcheting of the pipe structures [16,17,82,112]. These
literature studies showed the limitations of the most commonly-
used models. The linear and multilinear kinematic models were
unable of representing material ratcheting. The example showed
that the CH3 model in its original form has better capability, but
was also shown to be severely limited. Some of the most
advanced models were more promising, but still there was no
verification for arbitrary and complex load paths. Couterot et al.
[58] made computations of a 2 socket welding with the CH3
model which also showed a weak progressive deformation
behavior.
The current literatures have shown that the ratcheting behavior
of the straight pipe can be simulated by BKH model, MKIN model,
CH3 model, MCH3 model, OW model, CJ model, AKO model, AFe
OW model, AFeDP model, etc. BKH and MKIN model under-
predict the ratcheting strain. CH3 model is available in ANSYS
software, but it overpredicts the ratcheting strain. The advanced
plasticity models with modified CH3 model, OW model, CJ model,
AKO model, AFeOW model and AFeDP model must be imple-
mented in ANSYS or ABAQUS through the option of user

Fig. 24. Ratcheting prediction [54]: (a) A0, (b) A1, (c) A2, (d) a0, (e) a1.
138 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
Fig. 25. Ratcheting boundary [28,29]: (a) Straight pipe, (b) Elbow piping by C-TDF method with CheneJiaoeKim model.
programmable features. These models simulate ratcheting rate
more reasonably, but still exhibit some deficiencies.
5. Ratcheting boundary/ratcheting limit
Structural components subjected to a combination of cyclic and
steady loads can have two types of asymptotic deformation
responses: shakedown, ratcheting (or incremental collapse). Abdel-
Karim [7] reviewed that the initial AF rule and/or Besseling multi-
layer rule were utilized to scrutinize shakedown problem, and that
more advanced hardening rules were almost developed to scruti-
nize shakedown problem. In addition, Abdel-Karim [7] summarized
the alternative techniques which were coupled FEA with shake-
down bounding theorems for calculating the shakedown limit of
complex structures, including ‘the 2Y-solution’, Elastic Compensa-
tion Method (ECM), nonlinear superposition method, etc. These
methods were utilized to determine the elastic shakedown limit.
Unlike for elastic shakedown, only a few computational approaches
have been developed to predict the plastic shakedown and ratch-
eting boundary.
As shown by the above-reviewed researches, ratcheting behavior
has been mainly concerned for structural components. For the
design of components, ratcheting effect is usually not an acceptable
behavior. In order to avoid ratcheting effect, it is necessary that
ratcheting boundary is determined by the ratcheting criteria in
standard code or the proposed simple method. Ratcheting
boundary, namely the set of load combinations at the transition
between shakedown and ratcheting is an essential criterion for
structural components, such as piping system in nuclear industries.
Fig. 26. Ratcheting boundary of lateral nozzle of cylinder [54].
5.1. Determination of ratcheting boundary/ratcheting limit using
standard code
Bree [127,128] first investigated the ratcheting boundary of
a pressurized cylinder with cyclic thermal stresses across the wall
thickness. Then, Bree diagram had been contributed to the forma-
tion of the ratcheting criteria in ASME and KTA. In addition, based
on the test results, the Practical Design Rule or Efficiency diagram
rule in RCC-MR developed at the Centre d’Etudes Nucleaires, France
was an alternative to assume shakedown. Moreover, the Committee
of Three Dimensional Finite Element Stress Evaluation (C-TDF) in
Japan proposed one criterion for shakedown analysis with elasto-
plastic finite element analysis (EPFEA), in which an elastic perfect
plasticity model or a bilinear kinematic hardening rule was used
[129,130]. The criterion is “Variations in equivalent plastic strain at
the end of each cycle should have a decreasing trend and should
become lower than the allowable limit of 104/cycle”. The criterion
is used to determine the ratcheting boundary of the structural
components. The number of cycles required to achieve this value
was not specified, but usually 5 or 10 cycles were needed.
Using experimental data, Moreton et al. [131] proposed
a method to determine ratcheting boundary of the pressurized
elbow under cyclic bending loading. Based on different experi-
mental data in different positions, a regression technique was used
to find a best fit straight line of the experimental data. The inter-
cept of the line with the moment axis (i.e. zero ratcheting strain)
was taken as the moment value required to initiate ratcheting
strain.
Gao et al. [28,29,123], Chen et al. [49,50], and Liang et al. [132]
applied the regression technique to determine the moment value
or internal pressure required to achieve zero ratcheting strain. In
addition, they still applied the C-TDF criterion to determine the
ratcheting boundary of the straight pipe and the elbow pipe in 10
cycles. Ratcheting boundary was determined for straight pipe by
the C-TDF method with the MJS model in Fig. 25(a). The horizontal
axis X in Fig. 25 was the ratio of primary stress and yield stress; the
vertical axis Y was the ratio of the secondary stress and yield stress.
Compared to the experimental data, the determined ratcheting
boundary divided the shakedown region well, while those obtained
by KTA/ASME and RCC-MR showed more conservatism in the range
of the experimental loading conditions. Comparing ratcheting
boundaries obtained by the different methods, it was shown that
the boundary line by C-TDF approaches was less than that of ASME
when X < 0.2. However, the boundary line by RCC-MR allows
a much higher range of secondary stress when X < 0.12 but
a smaller range when X > 0.12 [28]. The ratcheting boundary
of elbow piping was determined by the C-TDF method with
 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
Fig. 27. Ratchet boundary for a straight pipe subjected to cyclic secondary bending coincident with (a) constant primary membrane stress, (b) constant primary bending stress
[136].
CheneJiaoeKim model, which was implemented into ANSYS soft-
ware, as shown in Fig. 25(b).
Fig. 26 shows that the ratcheting boundary of lateral nozzle of
cylinder was determined by the C-TDF method with MJS model,
which was implemented into ANSYS software by Yang et al. [53,54].
Five cycles were chosen as a practical number for the ratcheting
check [133]. Then, orthogonal experimental design method was
used in order to reduce the number of repetitive runs. The principle
of the least squares regression analysis was applied to establish the
relationship between the variables. Finally, the empirical formula of
ratcheting boundary of the lateral nozzle of cylinder was
determined.
The ratcheting boundary of C-TDF method was calculated from
cycle-by-cycles method. However, it required considerable
numerical expense and significant computer effort for complex
structures. In order to avoid these shortcomings, a simple method
should be developed.
5.2. Determination of ratcheting boundary/ratcheting limit using
a simple method
In addition to the ratcheting boundary determination method
discussed in the last section, other simple methods have been also
Fig. 28. Interaction diagram for elbow [139].
139
proposed. In particular, the classical shakedown theorems were
often utilized in the analysis, namely either the theorem of Melan
[134], the lower bound elastic shakedown theorem; or Koiter’s
theorem [135], which provides an upper bound for the elastic
shakedown loads. In many practical cases, the classical shakedown
theorems are not able to predict the relevant boundary between
shakedown and ratcheting (ratcheting boundary). Therefore, an
alternative method to predict the loads at the onset of ratcheting is
desirable. Several computational methods for plastic shakedown
analysis have been proposed, which are reviewed as follows.
Reinhardt [136] implemented a numerical two-step ratcheting
assessment into finite element software. Piping systems were taken
as beam models. The proposed method determined ratcheting
boundary without the need for cyclic analysis. Determined ratch-
eting boundary of the proposed method was found in well agree-
ment with that of theoretical solution, as shown in Fig. 27.
Adibi-Asl and Reinhardt [137e139] proposed a non-cyclic
method for shakedown analysis using a combination of linear
and nonlinear FEA. The proposed method used the elastic modulus
adjustment procedure to determine the ratcheting boundary of
the structural components. The ratcheting boundary of a 90
piping elbow with circular cross-section was investigated. The
results showed that the ratcheting boundary using the proposed
method was a very good agreement with cyclic FEA, as shown
in Fig. 28.
Abdalla et al. [140e145] proposed a simplified technique to
predict the ratcheting boundary of elbow piping. For the 18.0e
25.0 MPa steady internal pressure spectrum, the simplified tech-
nique outcomes showed very good correlation with those of Chen
et al. [50]. However, noticeable disagreement for the lower
magnitudes of steady internal pressure was also found.
Chen et al. [146] studied ratcheting limit of a defective pipeline
subjected to constant internal pressure and a cyclic thermal
gradient. Determination of ratcheting limit was employed by the
linear matching method (LMM) [147e152]. Ratcheting limit used
LMM method was compared with that of the step-by-step analysis,
it gave very accurate ratcheting limits for the defective pipelines
with part-through slots, as shown in Fig. 29. In addition, a defective
pipeline with a small slot and circumferential shallow slot did not
affect the ratcheting boundary and limit load. Moreover, a small slot
and circumferential shallow slot of the size gave essentially a local
stress concentration, which would affect the fatigue life of the
pipeline but would not influence the gross plastic deformation or
the incremental plastic collapse behavior.
As seen from above-reviewed researches, the non-cyclic method
was used to determine the ratcheting boundary of structural
140 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142

the thermal ratcheting of piping system in both experiment and

simulation have been studied.

Acknowledgment

The authors gratefully acknowledge financial support for this

work from the National High Technology Research and Develop-
ment Program of China (863 Program 2009AA04Z403), Ph.D.
Programs Foundation of Ministry of Education of China (No.
20090032110016), and Natural Science Foundation of Hebei Prov-
ince (No. E2011202044).

References

Fig. 29. The ratchet limit boundary for small slot case [146].
components. It was observed that these methods were simple,
efficient and accurate for the given examples.
6. Conclusion
This article reviews the experimental observations and FEA of
the structural ratcheting, which have been reported on the inter-
national publications. So far, the typical components of the struc-
tural ratcheting are straight pipe and elbow piping. Experimental
and simulation results show that cyclic bending of pipes induces
progressive accumulation of ovalization of the tube cross-section
and ratcheting strain. The maximum ratcheting strain occurs at
flanks or at intrados of elbow piping, respectively. Ratcheting
strains mainly occur in the circumferential direction. Such
responses are imperative for understanding the ratchetinge
fatigue failure mechanisms, and developing and validating simu-
lation models for tubular structures. In addition, the results
presented in the literature demonstrate the state of structural
response simulations by several cyclic plasticity models. Despite
that there are considerable advancements in cyclic plasticity
models, they are still not robust enough to simulate the structural
response when model parameters are determined from material
response only. However, development of a robust cyclic plasticity
model will be the most desirable for achieving the ratcheting
behavior simulation tools.
At present, ratcheting boundaries are determined with the final
aim of aiding the safety design and assessment of engineering
piping structures, based on such experimental and FEA research.
C-TDF method is commonly used to determine the ratcheting
boundary of piping components. But it requires the scholar to
repeat cycle-by-cycle in FEA. Therefore, an alternative method to
predict the ratcheting boundary of structural components is
desirable. The proposed non-cyclic methods are used to determine
the ratcheting boundary of structural components. It is shown that
these methods are simple, efficient and accurate for the given
examples. In addition, these methods have made a great contri-
bution for the progress of shakedown theory.
So far, the structural ratcheting of piping system has not been
predicted well by advance constitutive models. Therefore, the more
material properties should be incorporated into these constitutive
models in order to improve the simulated results, such as cyclic
hardening or softening response of materials, residual stress,
material anisotropy, non-proportionality degree, time dependence
and rate dependence, etc. Thermal ratcheting of piping system in
nuclear, electric and chemical industries is induced by cyclic
temperature, thermal expansion and shocks. But few researches on
[1] Hübel H. Basic conditions for material and structural ratcheting. Nucl Eng
Des 1996;162:55e65.
[2] Boiler A, Code PV. Section VIII division I, appendix 2. New York: American
Society of Mechanical Engineers; 2007.
[3] KemtecllIlischer AIlsschu b (KTA). Sicherheitstechnische Regel des KTA, Teil:
Auslegung, Konstruktion und Berchnung, Regeladerungsentwurf; 1995.
[4] EN13445e3. Unfired pressure vessels e part 3: design, annex C: method
based on stress categories; 2002.
[5] R5. Assessment procedure for the high temperature response of structures.
Nuclear Electric plc; 1990. p. 2.
[6] Design rules for class 1 equipment RCC-MR codes, revision. 2002.
[7] Abdel-Karim M. Shakedown of complex structures according to various
hardening rules. Int J Pres Ves Pip 2005;82:427e58.
[8] Ohno N. Recent progress in constitutive modeling for ratcheting. Mater Sci
Res Int 1997;3:1e9.
[9] Kang GZ. Ratchetting: recent progresses in phenomenon observation,
constitutive modeling and application. Int J Fatigue 2008;30:1448e72.
[10] Scavuzzo R, Lam P, Gau J. Experimental studies of ratcheting of pressurized
pipe. J Pres Ves e Trans ASME 1991;113:210e8.
[11] Igari T, Wada H, Ueta M. Plastic buckling and ratcheting of straight pipes
subjected to deformation-controlled montonic and cyclic bending. ASME
PVP 1996;331:47e54.
[12] Kulkarni S, Desai Y, Kant T, Reddy G, Parulekar Y, Vaze K. Uniaxial and biaxial
ratchetting study of SA333 Gr. 6 steel at room temperature. Int J Pres Ves Pip
2003;80:179e85.
[13] Kulkarni S, Desai Y, Kant T, Reddy G, Prasad P, Vaze K, et al. Uniaxial and
biaxial ratchetting in piping materials d experiments and analysis. Int J Pres
Ves Pip 2004;81:609e17.
[14] Gau SJ. Elasticeplastic behavior of pressurized pipe. Ph.D. dissertation.
Department of Mechanical Engineering, University of Akron, Ohio, January
1990.
[15] Hassan T, Matzen V. ASME code and ratcheting in piping components. Final
technical report. North Carolina State Univ.: Raleigh, NC, US. 1999.
[16] Rahman SM. Finite element analysis and related numerical schemes for
ratcheting simulation. Ph.D. dissertation. North Carolina State University;
2006.
[17] Rahman SM, Hassan T, Corona E. Evaluation of cyclic plasticity models in
ratcheting simulation of straight pipes under cyclic bending and steady
internal pressure. Int J Plast 2008;24:1756e91.
[18] Ranganath S, Hwang S, Tagart SW. Piping and fitting dynamic reliability
program. EPRI Nuclear Power Division; 1989.
[19] English W. Piping and fitting dynamic reliability program-fourth semi-
annual progress report; Nov. 1986eApr. 1987. GE Nuclear Energy, NEDC-
3152.
[20] Moreton D, Yahiaoui K, Moffat D, Machin H, Amesbury L. The behaviour of
pressurised plain pipework subjected to simulated seismic loading. Strain
1994;30:63e72.
[21] Krämer D, Krolop S, Scheffold A, Stegmeyer R. Investigations into the
ratchetting behaviour of austenitic pipes. Nucl Eng Des 1997;171:161e72.
[22] Yoshida F, Obataya Y, Shiratori E. Mechanical ratcheting behaviors of a steel
pipe under combined cyclic axial load and internal pressure. In: Society of
Materials Science Proc of the 27th Japan Congr on Mater Res. 1984. p. 19e23.
[23] Guionnet C. Modelization of ratcheting in biaxial experiments. Nucl Eng Des
1989;116:223e30.
[24] Rider RJ, Harvey SJ, Charles ID. Ratchetting in pressurized pipes. Fatigue Fract
Eng Mater Struct 1994;17:497e500.
[25] Moreton D, Yahiaoui K, Moffat D, Zehsaz M. The effect of diameter/thickness
ratio on the ratchetting behaviour of pressurised plain pipework subjected to
simulated seismic loading. Strain J BSSM 1996;32:91e6.
[26] Moreton D, Zehsaz M, Yahiaoui K, Moffat DG. The ratcheting of plain carbon
steel pressurized cylinders subjected to simulated seismic bending: the
effect of the D/t ratio and component with finite element predictions. J Strain
Anal 1998;33:39e53.
[27] Zehsaz M, Zakavi SJ, Mahbadi H, Eslami MR. Cyclic strain accumulation of
plain stainless steel pressurized cylinders subjected to dynamic bending
moment. J Appl Sci 2008;8:3129e38.
 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
[28] Gao B, Chen X, Chen G. Ratchetting and ratchetting boundary study of
pressurized straight low carbon steel pipe under reversed bending. Int J Pres
Ves Pip 2006;83:96e106.
[29] Gao B. Modeling of material multiaxial ratcheting and ratcheting prediction
of pressure piping. Ph.D. dissertation. Tianjin University of Technology; 2005
(in Chinese).
[30] Corona E, Kyriakides S. An experimental investigation of the degradation and
buckling of circular tubes under cyclic bending and external pressure. Thin
Wall Struct 1991;12:229e63.
[31] Jiao R, Kyriakides S. Ratcheting and wrinkling of tubes due to axial cycling
under internal pressure: part I: experiments. Int J Solids Struct 2011;48:
2814e26.
[32] Matzen V, Zhu Y, Hassan T. Review of experiments and related analyses on
piping systems and components. Trans SMiRT 1995;13(3):471e6.
[33] Fujiwaka T, Endou R, Shin-ichi F, Ono S, Oketani K. Study on strength of
piping components under elasticeplastic behavior due to seismic loading.
ASME PVP 1999;387:19e26.
[34] Ichihashi C. Some collapse tests for the large scale models of the nuclear
reactor facilities by shaking table. J Jpn Assoc Earthquake Eng 2004;4:
465e71.
[35] Humphreys AE. Influence of residual stress on the initiation of fatigue cracks
at welded piping joints. Ph.D. dissertation. North Carolina State University;
2004.
[36] Cheng PY. Influence of residual stress and heat affected zone on fatigue
failure of welded piping joints. Ph.D. dissertation. North Carolina State
University; 2009.
[37] Song Y. Ratcheting fatigue failure of welded stainless steel pipe and dislo-
cation microstructure. Master dissertation. North Carolina State University;
2009.
[38] Boussaa D, Labbe P, Tang H. Fatigue-ratcheting analysis of pressurized
elbows. ASME PVP 1993;266:13e21.
[39] Yahiaoui K, Moffat DG, Moreton DN. Damage assessment of piping elbows
loaded by steady internal pressure and dynamic in-plane or out-of-plane
bending. ICPVT 1996;1:361e75.
[40] Yahiaoui K, Moffat D, Moreton D. Pressurized piping elbows under simulated
seismic bending: design code implications. Proc Inst Mech Eng E-J Pro 1996;
210:159e70.
[41] DeGrassi G, Hofmayer C, Murphy A, Suzuki K, Namita Y. BNL nonlinear pre-
test seismic analysis for the NUPEC uitimate strength piping test program.
Trans SMiRT 2003;17. BNL-NUREG-71119-2003-CP.
[42] DeGrassi G, Hofmayer. C. Seismic analysis of simplified piping systems for
the NUPEC uitimate strength piping test program. NUREG/CR-6889 BNL-
NUREG-74978-82005.
[43] DeGrassi G, Nie J, Hofmayer C. Seismic analysis of large-scale piping systems
for the JNES/NUPEC uitimate strength piping test program. NUREG/CR-6983.
Washington, DC: US Nuclear Regulatory Commission; 2008.
[44] Huang CT, Iwan W, Jaquay K, Chokshi N. Cyclic moment response charac-
teristics and seismic margins of elbows. ASME PVP 1998;360:271e82.
[45] Acker D, Toubout F, Autrusson B. Experimental analysis of ratcheting in
elbows. PVP Am Soc Mech Eng 1992;235:87e91.
[46] Stearns JC, Lam PC, Scavuzzo RJ. Incremental plastic strain of a pipe elbow
with varying bend radii. ASME PVP 1993;264:169e74.
[47] Touboul F, Lacire M, Blay N, Blanchard M, Le Breton F. Simplified methods for
the evaluation of the seismic behaviour of piping system for criteria appli-
cation. ASME PVP 1998;364:117e28.
[48] Wilkins JK. Experimental and analytical investigation into the non-linear
behavior of 200 and 400 , 90 , large radius, schedule 10, stainless steel elbows
under monotonic, cyclic and rate dependent loading. Master dissertation.
North Carolina State University; 2002.
[49] Chen X, Gao B, Chen G. Multiaxial ratcheting of pressurized elbows subjected
to reversed in-plane bending. J Pres Eq Syst 2005;3:38e44.
[50] Chen X, Gao B, Chen G. Ratcheting study of pressurized elbows subjected to
reversed in-plane bending. J Pres Ves e Trans ASME 2006;128:525e32.
[51] Hassan T, Rahman SM. Simulation of ratcheting responses of elbow piping
components. ASME PVP 2009;1:103e8.
[52] Chen XH, Chen X. Finite element analysis on ratcheting strain distribution for
pressurized 90 -elbows under cyclic bending. ISSI; 2011. p. 51e55.
[53] Yang L, Dong J, Gao B. Ratchetting study of pressurized lateral nozzle of
cylinder subjected to cyclic in-plane bending. Pres Ves Technol 2010;27:11e
5 (in Chinese).
[54] Yang L. Cyclic plasticity study of pressurized lateral nozzle of cylinder.
Master dissertation. Hebei University of Technology; 2010 (in Chinese).
[55] Jahanian S. On the incremental growth of mechanical structures subjected to
cyclic thermal and mechanical loading. Int J Pres Ves Pip 1997;71:121e7.
[56] Watanabe D, Chuman Y, Otani T, Shibamoto H, Inoue K, Kasahara N. An
experimental validation of the guideline for inelastic design analysis through
structural model tests. Nucl Eng Des 2008;238:389e98.
[57] Ueda M, Kano T, Yoshitoshi A. Thermal ratcheting criteria and behavior of
piping elbows. J Pres Ves e Trans ASME 1990;112:71e5.
[58] Couterot C, Geyer P, Proix JM. Experiment and numerical analysis of the NPP
pressurizer auxilliary spray line submitted to large thermal shocks. Proc PVP
Minneapolis 1994;280:67e72.
[59] Lee HY, Kim JB, Lee JH. Evaluation of progressive inelastic deformation
induced by a moving axial temperature front for a welded structure. Int J
Pres Ves Pip 2004;81:433e41.
141
[60] Lee HY, Kim JB, Lee JH. Test and analysis of thermal ratcheting deformation
for 316L stainless steel cylindrical structure. Korean Soc Mech Engrs 2002;
26:479e86.
[61] Ohno N, Abdel-Karim M. Uniaxial ratchetting of 316FR steel at room
temperature, part II: constitutive modeling and simulation. J Eng Mater e
Trans ASME 2000;122:35e41.
[62] Ohno N, Wang JD. Kinematic hardening rules with critical state of dynamic
recovery, part I: formulation and basic features for ratchetting behavior. Int J
Plast 1993;9:375e90.
[63] Ohno N, Wang JD. Kinematic hardening rules with critical state of dynamic
recovery. Part II: application to experiments of ratcheting behavior. Int J Plast
1993;9:391e403.
[64] Kang G. Ratcheting of cyclically stable materials: II. Implicit stress integration
algorithm and finite element implementation. Eng Mech 2005;22:205e9 (in
Chinese).
[65] Kan Q, Kang G, Qian L. Super-elastic constitutive model considering plas-
ticity and its finite element implementation. Acta Mech Solida Sin 2010;23:
95e105.
[66] Kang GA. visco-plastic constitutive model for ratcheting of cyclically stable
materials and its finite element implementation. Mech Mater 2004;36:
299e312.
[67] Kang G. Finite element implementation of advanced constitutive model
emphasizing on ratcheting. Trans SMiRT 2005;18:224e33.
[68] Kang G. Finite element implementation of visco-plastic constitutive model
with strain-range-dependent cyclic hardening. Commun Numer Meth Eng
2006;22:137e53.
[69] Kang G, Gao Q. Ratcheting of cyclically stable materials: I. experiments and
a viso-plastic constitutive model. Eng Mech 2005;22:206e11 (in Chinese).
[70] Chaboche J. Time-independent constitutive theories for cyclic plasticity. Int J
Plast 1986;2:149e88.
[71] Chaboche J. Cyclic plasticity modeling and ratchetting effects. Constitutive
Laws Eng Mater Theory Appl 1987;1:47e58.
[72] Chaboche J. On some modifications of kinematic hardening to improve the
description of ratchetting effects. Int J Plast 1991;7:661e78.
[73] Chaboche J. Modeling of ratchetting: evaluation of various approaches. Eur J
Mech A-Solid 1994;13:501e18.
[74] Chaboche J. A review of some plasticity and viscoplasticity constitutive
theories. Int J Plast 2008;24:1642e93.
[75] Chaboche J, Dang V. Modelization of the strain memory effect on the cyclic
hardening of 316 stainless steel. Trans SMiRT 1979;5. Paper, L11/3.
[76] Chaboche JL. Constitutive equations for cyclic plasticity and cyclic visco-
plasticity. Int J Plast 1989;5:247e302.
[77] Chaboche JL, Kanouté P, Azzouz F. Cyclic inelastic constitutive equations and
their impact on the fatigue life predictions. Int J Plast 2012;35:44e6.
[78] Prager W. A new method of analyzing stresses and strains in work-
hardening plastic solids. J Appl Mech 1956;23:493e6.
[79] Besseling JF. A theory of elastic, plastic and creep deformations of an initially
isotropic material showing anisotropic strain-hardening, creep recovery, and
secondary creep. J Appl Mech 1958;25:529e36.
[80] Mroz Z. On the description of anisotropic workhardening. J Mech Phys Solids
1967;15:163e75.
[81] Owen D, Prakash A, Zienkiewicz O. Finite element analysis of non-linear
composite materials by use of overlay systems. Comput Struct 1974;4:
1251e67.
[82] Hassan T, Zhu Y, Matzen VC. Improved ratcheting analysis of piping
components. Int J Pres Ves Pip 1998;75:643e52.
[83] Bari S, Hassan T. An advancement in cyclic plasticity modeling for multiaxial
ratcheting simulation. Int J Plast 2002;18:873e94.
[84] Chen X, Jiao R. Modified kinematic hardening rule for multiaxial ratcheting
prediction. Int J Plast 2004;20:871e98.
[85] Chen X, Jiao R, Kim KS. On the OhnoeWang kinematic hardening rules for
multiaxial ratcheting modeling of medium carbon steel. Int J Plast 2005;21:
161e84.
[86] Jiang YY, Sehitoglu H. Modeling of cyclic ratcheting plasticity, part II:
comparison of model simulations with experiments. J Appl Mech e Trans
ASME 1996;63:726e33.
[87] Jiang YY, Sehitoglu H. Modeling of cyclic ratcheting plasticity, part I: devel-
opment of constitutive relations. J Appl Mech e Trans ASME 1996;63:720e5.
[88] Dafalias Y, Popov E. A model of nonlinearly hardening materials for complex
loading. Acta Mech 1975;21:173e92.
[89] Dafalias Y, Popov E. Plastic internal variables formalism of cyclic plasticity.
J Appl Mech e Trans ASME 1976;43:645e51.
[90] Hassan T, Kyriakides S. Ratcheting of cyclically hardening and softening
materials: I. Uniaxial behavior. Int J Plast 1994;10:149e84.
[91] Hassan T, Kyriakides S. Ratcheting of cyclically hardening and softening
materials: II. Multiaxial behavior. Int J Plast 1994;10:185e212.
[92] Guionnet C. Modeling of ratcheting in biaxial experiments. J Eng Mater e
Trans ASME 1992;114:56e62.
[93] Delobelle P, Robinet P, Bocher L. Experimental study and phenomenological
modelization of ratchet under uniaxial and biaxial loading on an austenitic
stainless steel. Int J Plast 1995;11:295e330.
[94] Hassan T, Corona E, Kyriakides S. Ratcheting in cyclic plasticity, part II:
multiaxial behavior. Int J Plast 1992;8:117e46.
[95] Hassan T, Kyriakides S. Ratcheting in cyclic plasticity, part I: uniaxial
behavior. Int J Plast 1992;8:91e116.
142 X. Chen et al. / International Journal of Pressure Vessels and Piping 101 (2013) 113e142
[96] Bari S, Hassan T. Anatomy of coupled constitutive models for ratcheting
simulation. Int J Plast 2000;16:381e409.
[97] Kojic M. Stress integration procedures for inelastic material models within
the finite element method. Appl Mech Rev 2002;55:389e414.
[98] Kobayashi M, Ohno N. Implementation of cyclic plasticity models based on
a general form of kinematic hardening. Int J Numer Meth Eng 2002;53:
2217e38.
[99] Armstrong P, Frederick C. A mathematical representation of the multiaxial
Bauchinger effect. CEGB report no RD/BN 731; 1966.
[100] ANSYS, Inc.. Guide to ANSYS user programmable features (ANSYS release
9.0); 2004.
[101] Burlet H, Cailletaud G. Numerical techniques for cyclic plasticity at variable
temperature. Eng Comput 1986;3:143e53.
[102] McDowell D. Stress state dependence of cyclic ratchetting behavior of two
rail steels. Int J Plast 1995;11:397e421.
[103] McDowell D. An approximate algorithm for elastic-plastic two-dimensional
rolling/sliding contact. Wear 1997;211:237e46.
[104] Abdel-Karim M, Ohno N. Kinematic hardening model suitable for ratchetting
with steady-state. Int J Plast 2000;16:225e40.
[105] Kang G, Gao Q, Yang X. Uniaxial and non-proportionally multiaxial ratch-
eting of SS304 stainless steel at room temperature: experiments and simu-
lations. Int J Nonlin Mech 2004;39:843e57.
[106] Halama R. Ratcheting simulation with modified AbdelKarim-Ohno cyclic
plasticity model. In: Transactions of the VSB  e Technical University of
Ostrava, Mechanical Series LII, vol. 1; 2000. p. 155e162.
[107] Suzuki K. Seismic proving test of uitimate piping strength. ICONE-9; 2001.
Paper no. 155.
[108] Suzuki K. Seismic proving test of uitimate piping strength. ICONE-10; 2002.
Paper no. 22225.
[109] Fujiwaka T, Kobayashi H, Asada Y, Shitara C. Simulation of excessive defor-
mation of piping due to seismic and weight loads. In: ASME Press vess piping
conf, PVP, vol. 439; 2002. p. 345e52.
[110] Zakavi SJ, Zehsaz M, Eslami MR. The ratchetting behavior of pressurized plain
pipework subjected to cyclic bending moment with the combined hardening
model. Nucl Eng Des 2010;240:726e37.
[111] Zhu YM, Hassan T, Matzen VC. Performance of a commercial finite
element program in simulating ratcheting in piping. Trans SMiRT 1995;
13(3):489e94.
[112] Hassan T, Zhu Y, Matzen VC. Simulation of ratcheting in straight pipes
using ANSYS with an improved cyclic plasticity model. ASME PVP 1996;
340:97e104.
[113] Rojíceka J, Halama R. Numerical simulations of pipeline bending tests. Appl
Comp Mech 2008;2:347e56.
[114] Jiao R, Kyriakides S. Ratcheting and wrinkling of tubes due to axial cycling
under internal pressure: part II analysis. Int J Solids Struct 2011;48:2827e36.
[115] Garud YS, Durlofsky H, Tagart S. Analysis and prediction of fatigue-
ratcheting: comparison with tests and code rules. ASME PVP 1993;266:
23e32.
[116] Kobayashi H, Fujiwaka T. Analytical study of the response of piping with
mechanical ratcheting under dynamic loads. In: ASME Press vess piping conf
PVP, vol. 345; 1997. p. 223e8.
[117] Zhao Y. Random vibration and ratcheting analysis based evaluation of
current ASME code nuclear piping design criteria. In: ASME Press vess piping
conf PVP, vol. 345; 1997. p. 19e26.
[118] Shalaby MA, Younan MYA. Nonlinear analysis and plastic deformation of
pipe elbows subjected to in-plane bending. Int J Pres Ves Pip 1998;75:
603e11.
[119] Suzuki K, Namita Y, Abe H. Seismic proving test of uitimate piping strength:
simulation analysis of simplified piping system test. ASME PVP 2003;466:
23e30.
[120] Suzuki K, Namita Y, Abe H, Ichihashi I, Ishiwata M, Fujiwaka T, et al. Seismic
proving test of uitimate piping strength: test results on piping component
and simplified piping system. ASME PVP 2002;1:99e106.
[121] Suzuki K, Abe H. Seismic proving test of uitimate piping strength: uitimate
strength test. ASME PVP 2004;2:201e8.
[122] Balan C, Redekop D. The effect of bi-directional loading on fatigue assess-
ment of pressurized piping elbows with local thinned areas. Int J Pres Ves Pip
2005;82:235e42.
[123] Gao B, Chen X, Chen G. Experiment research on ratcheting of pressurized
pipe subjected reversed bending. Trans SMiRT 2005;18:981e7.
[124] Zakavi SJ, Zchsaz M, Eslami MR. The effect of frequency on the cyclic strain
accumulation of plain stainless steel pressurized cylinders subjected to
dynamic bending moment. Trends Appl Sci Res 2009;4:200e15.
[125] Weiß E, Postberg B, Nicak T, Rudolph J. Simulation of ratcheting and low
cycle fatigue. Int J Pres Ves Pip 2004;81:235e42.
[126] Dahlberg M, Segle P. Evaluation of models for cyclic plastic deformation e
a literature study. Inspecta Technology AB; 2010.
[127] Bree J. Elasticeplastic behaviour of thin tubes subjected to internal pressure
and intermittent high-heat fluxes with application to fast-nuclear-reactor
fuel elements. J Strain Anal Eng 1967;2:226e38.
[128] Bree J. Plastic deformation of a closed tube due to interaction of pressure
stresses and cyclic thermal stresses. Int J Mech Sci 1989;31:865e92.
[129] Asada S, Yamashita N, Okamoto A, Nishiguchi I. Verification of alternative
criteria for shakedown evaluation using flat head vessel. ASME PVP 2002;
439:17e23.
[130] Yamamoto Y, Yamashita N, Tanaka M. Evaluation of thermal stress ratchet in
plastic FEA. ASME PVP 2002;439:3e10.
[131] Moreton DG, Yahiaoui K, Moffat DG. Onset of ratcheting in pressurized
piping elbows subjected to in-plane bending moments. Int J Pres Ves Pip
1996;68:73e9.
[132] Liang T, Chen G, Zhang Q, Chen X. Ratcheting boundary analysis of straight
and elbow piping. Adv Mater Rev 2010;118:131e5.
[133] Vlaicu D. Shakedown analysis of nuclear components using linear and
nonlinear methods. J Pres Ves e Trans ASME 2010;132:021203.
[134] Melan E. Zur plastizität des räumlichen kontinuums. Arch Appl Mech 1938;
9:116e26.
[135] Koiter W. General theorems of elasticeplastic solids. In: Sneddon JN, Hill R,
editors. Prog solid mech, vol. 1; 1960. p. 167e221.
[136] Reinhardt W. Simplified method for the assessment of elasticeplastic
shakedown in piping. Trans SMiRT 2007;19. Paper B05/5.
[137] Adibi-Asl R, Reinhardt W. Ratchet boundary determination using a noncyclic
method. J Pres Ves e Trans ASME 2010;132:021201e9.
[138] Adibi-Asl R, Reinhardt W. Non-cyclic shakedown/ratcheting boundary
determination e part 1: analytical approach. Int J Pres Ves Pip 2011;88:
311e20.
[139] Adibi-Asl R, Reinhardt W. Non-cyclic shakedown/ratcheting boundary
determination e part 2: numerical implementation. Int J Pres Ves Pip 2011;
88:321e9.
[140] Abdalla HF, Megahed MM, Younan Mr YA. Determination of shakedown limit
load for 90 degree pipe bend using a simplified technique. J Pres Ves e Trans
ASME 2006;128:618e24.
[141] Abdalla HF, Megahed MM, Younan Mr YA. Shakedown limits of a 90-degree
pipe bend using small and large displacement formulations. J Pres Ves e
Trans ASME 2007;129:287e95.
[142] Abdalla HF, Megahed MM, Younan MYA. Shakedown limit load deter-
mination for a kinematically hardening 90-degree pipe bend subjected
to constant internal pressure and cyclic bending. ASME PVP; 2007.
p. 405e413.
[143] Abdalla HF, Megahed MM, Younan MYA. Shakedown limit loads for 90-
degree scheduled pipe bends subjected to constant internal pressure and
cyclic bending moments. ASME PVP; 2008. p. 483e492.
[144] Abdalla HF, Megahed MM, Younan MYA. Comparison of pipe bend ratchet-
ing/shakedown test results with the elastic shakedown boundary deter-
mined via a simplified technique. ASME PVP; 2009. p. 77403.
[145] Abdalla HF, Megahed MM, Younan MYA. A simplified technique for shake-
down limit load determination of a large square plate with a small central
hole under cyclic biaxial loading. Nucl Eng Des 2011;241:657e65.
[146] Chen H, Chen W, Li T, Ure J. On shakedown, ratchet and limit analyses of
defective pipeline. J Pres Ves e Trans ASME 2012;134:011202.
[147] Chen H, Chen W, Li T, Ure JM. Effect of circular holes on the ratchet limit and
crack tip plastic strain range in a centre cracked plate. Eng Fract Mech 2011;
78:2310e24.
[148] Chen H, Ponter ARS. A method for the evaluation of a ratchet limit and the
amplitude of plastic strain for bodies subjected to cyclic loading. Eur J Mech
A-Solid 2001;20:555e71.
[149] Chen H, Ponter ARS. Integrity assessment of a 3D tubeplate using the linear
matching method. Part 1. Shakedown, reverse plasticity and ratchetting. Int J
Pres Ves Pip 2005;82:85e94.
[150] Chen H, Ponter ARS. A direct method on the evaluation of ratchet limit. J Pres
Ves e Trans ASME 2010;132:041202.
[151] Chen HF, Ure JM, Li TB, Chen WH. Shakedown and limit analysis of 90 pipe
bends under internal pressure, cyclic in-plane bending and cyclic thermal
loading. Int J Pres Ves Pip 2011;88:213e22.
[152] Ponter ARS, Chen H. A minimum theorem for cyclic load in excess of
shakedown, with application to the evaluation of a ratchet limit. Eur J Mech
A-Solid 2001;20:539e53.