Mechanics of Materials 43 (2011) 705–720

Contents lists available at SciVerse ScienceDirect

Mechanics of Materials
journal homepage: www.elsevier.com/locate/mechmat

Key issues in cyclic plastic deformation: Experimentation

Surajit Kumar Paul a,c,⇑, S. Sivaprasad b, S. Dhar c, S. Tarafder b
a
Research and Development, Tata Steel Limited, Jamshedpur 831 001, India
b
Materials Science & Technology Division, National Metallurgical Laboratory (Council of Scientific & Industrial Research), Jamshedpur 831 007, India
c
Department of Mechanical Engineering, Jadavpur University, Kolkata 700 032, India

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

Article history: Cyclic plastic deformation phenomena include the Bauschinger effect, cyclic hardening/
Received 22 December 2010 softening, strain range effect, loading history memory, ratcheting, mean stress dependent
Received in revised form 19 July 2011 hardening, mean stress relaxation and non-proportional hardening. In this work, different
Available online 2 August 2011
cyclic plastic deformation responses of piping materials (SA333 C–Mn steel and 304LN
stainless steel) are experimentally explored. Cyclic hardening/softening is depends upon
Keywords: loading types (i.e. stress/strain controlled), previous loading history and strain/stress range.
Cyclic plasticity
Pre-straining followed by LCF and mean stress relaxation shows similar kind of material
Ratcheting
LCF
response. Substantial amount of non proportional hardening is observed in SA333 C–Mn
Cyclic hardening steel during 90° out of phase tension-torsion loading. During ratcheting, large amount of
True stress–strain permanent strain is accumulated with progression of cycles. Permanent strain accumula-
Non proportional loading tion in a particular direction causes cross-sectional area reduction and which results
uncontrollable alteration of true stress in engineering stress controlled ratcheting test. In
this work, true stress control ratcheting on piping materials has been carried out in labo-
ratory environment. Effects of stress amplitude and mean stress on the ratcheting behav-
iors are analyzed. A comparison has also been drawn in between the true and engineering
stress controlled tests, and massive difference in ratcheting life and strain accumulation is
found.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction 2005; Mayama et al., 2008; Taleb and Hauet, 2009) and
the degree of damage is influenced by a number of factors
Engineering components are often subjected to cyclic such as type of loading (stress/strain controlled,
load excursions and the cyclic plastic deformation of engi- proportional/non-proportional), presence of stress riser
neering materials thus becomes inevitable. Since the resul- (inclusions, defects, geometrical discontinuities, surface
tant elastic–plastic stress–strain response of the material roughness) and residual stress (large deformation history,
plays a pivotal role in analysis of design and failure of welding effects). In spite of great deal of efforts, especially
the component, it becomes important to understand the in the last two decades, to understand the various aspects
cyclic plastic deformation behavior of engineering materi- of cyclic plasticity and their constitutive modeling, there
als. There are numerous evidences that the process of dam- are still difficulties in translating the knowledge gained
age accumulation due to cyclic loading is greatly to different material systems. One of the primary reasons
influenced by the evolution of dislocation sub-structure for this inability is due to the complex nature of the cyclic
(for e.g., Gaudin and Feaugas, 2004; Zhang and Jiang, plastic phenomena and its strong dependency on material
characteristics, and it thus becomes essential to deal with
all common factors that control the plastic deformation
⇑ Corresponding author at: Research and Development, Tata Steel
in different materials. This paper is aimed at addressing
Limited, Jamshedpur 831 001, India. Tel.: +91 9939632284; fax: +91 657
2271748. some of these key issues that should be considered in deal-
E-mail address: paulsurajit@yahoo.co.in (S.K. Paul). ing with cyclic plasticity. While all these aspects are widely

0167-6636/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mechmat.2011.07.011
706 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720

accepted by the research community in the subject area, 2010), few special phenomena also noticed in anisotropic
we have revisited these issues in light of our recent exper- materials like permanent softening, cross-effect etc. (Verma
imental investigations. et al., 2010) which are not investigated in this study. These
The load excursion on a component can be either in eight basic phenomena thus form essential elements that
terms of stress (as in load based design) or strain (as in dis- describe the details of the hysteresis loops for isotropic
placement based design) and when the magnitude of materials. A complete plasticity model should have the
stress/strain reversal is sufficient, cyclic plastic deforma- capability to predict all of the above cyclic plasticity phe-
tion will occur. Accurate description of inelastic behavior nomena. As cyclic plasticity is a very complex path depen-
of isotropic materials necessitates a thorough understand- dent phenomena, plasticity model can be simplified by
ing of following phenomena. requirement basis and to do this complete understanding
Bauschinger effect: that describes the reduction of yield is necessary. Excluding any major plasticity issue during
stress during reloading the material in reverse direction. modeling may results completely wrong results i.e. stress–
If its influence is not considered in predicting the material strain response. In this work, we have experimentally inves-
cyclic deformation behavior, the hysteresis loop shape will tigated each of these material aspects in order to understand
be deviate from experimental results. the cyclic plastic deformation in two types of primary heat
Hardening/softening behavior: portrays the ability of the transport piping materials. Special emphasis has been given
material to resist the dislocation movement during cyclic to cyclic ratcheting, considering its importance in fatigue
deformation manifested in the form of either an increase life design. The significance of true stress controlled ratchet-
(hardening) or a decrease (softening) in stress amplitude ing experiments over that of engineering stress controlled
during strain controlled loading. ratcheting studies has been examined.
Load history memory: refers to the influence of previous The overall objective of this work is to provide a funda-
deformation histories on the subsequent hardening/soften- mental understanding of the cyclic plastic phenomena,
ing behavior of the material. If not considered, the results considering various materials’ aspects that are experimen-
may yield inaccurate hardening/softening characteristics tally observed. Such understanding is expected to present a
and hence an erroneous stress–strain evolution. basic guideline for development of constitutive relation-
Strain range effect: notify that the linear region of hys- ships for cyclic plasticity. The discussions will be useful
teresis loops (i.e. double of cyclic yield stress) varies with for further evaluation/revision of the existing plasticity
applied strain/stress amplitude. Similar to load history ef- models.
fect, strain range variation is intimately connected to the
stress–strain evolution. If not included, the cyclic deforma-
tion behavior of material over a range of strain amplitude 2. Experimental
levels cannot be predicted accurately.
Mean stress relaxation: represents the stress response of The materials used for the experimental investigation
the material when non-zero mean strain is present during are (i) 304LN austenitic stainless steel and (ii) SA333 C–
cyclic deformation. The extent over which the mean stress Mn steel. Both the materials are used in the primary heat
relaxation would occur depends upon the material charac- transport (PHT) piping of pressurized heavy water reactors
teristic and will influence the resulting stress–strain (PHWRs) in Indian nuclear power plants. Both the materi-
behavior. For accurate prediction of cyclic response, the als were available in the form of pipes with 340 mm outer
mean stress relaxation effect has to be accounted for. diameter  25 mm wall thickness (304LN stainless steel)
Mean stress dependent hardening: refers to the additional and 600 mm outer diameter  50 mm wall thickness
hardening that might occur due to the presence of mean (SA333 C–Mn steel). The chemical composition (wt.%) of
stress. Ignoring this factor will result in an erroneous stress the 304LN stainless steel was: C 0.03; Si 0.65; Ni 8.17;
calculation. Mo 0.26; Cu 0.29; N 0.08; S 0.02; P 0.034 while that of
Non-proportional hardening: represents the additional the SA333 C–Mn steel was: C 0.18; Mn 0.9; Si 0.02; P
hardening behavior of materials under non-proportional 0.02. The nominal mechanical properties of both the steels
loading. It is a material dependent phenomenon and can are listed in Table 1.
change the material’s stress–strain response. The crystal structure of 304LN stainless steel was face
Ratcheting: refers to progressive accumulation of direc- centered cubic (fcc) with austenitic microstructure while
tional strain due to asymmetric stress cycling. It is one of that of the SA333 C–Mn steel was body centered cubic
the greatest concern to the design engineers due to the fact (bcc) with a ferrite-pearlitic structure. Thus the two steels
that apart from cyclic damage induced by load reversals, are expected to exhibit different deformation charac-
the cumulative strain accrual (termed as ratcheting strain) teristics.
in every cycle arising from non-closure of the hysteresis For all the uniaxial experiments solid specimens of
loop can adversely affect the estimated pure fatigue life. 7 mm gauge diameter and 13 mm gauge length were fabri-
As a matter of fact, the magnitude of ratcheting strain cated from the pipe sections such that the loading axis of
accumulation is influenced by the loading conditions such the specimen is parallel to the length of the pipe. For mul-
as mean stress, stress amplitude and prior load history tiaxial experiments, tubular specimens of 25 mm gauge
(Jiang and Sehitoglu, 1994a,b; Paul et al., 2010a,c). length  25 mm outer diameter with 1.6 mm wall thick-
Apart from these eight basic phenomena in isotropic ness were used. The orientation of these specimens with
materials, plastic deformation response dependency on respect to the pipe dimension was similar to those of solid
strain rate and temperature are also observed (Sung et al., specimens.
 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720 707

Table 1
Tensile properties of SA333 C–Mn steel and 304 LN stainless steel.

Material Yield stress (MPa) Ultimate tensile stress (MPa) Uniform elongation (%) Total elongation (%)
SA333 C–Mn steel 304 494 17.18 24.53
AISI 304LN stainless steel 353 671 52.8 69.62

All strain controlled experiments were conducted using

a triangular waveform at a constant strain rate of 103 s1.
a Forward flow curve
Stress
All stress controlled experiments were conducted at a con-
stant stress rate of 50 MPa/s employing a similar wave- Reversed flow curve
form. A 100 kN closed loop servo-electric test system was
used for the uniaxial studies employing a 12.5 mm gauge
length extensometer. A 100 kN/1000 Nm closed loop ser-
vo-hydraulic axial-torsion test frame was used for the mul-
σ0 E 2σ0
tiaxial experiments. A 25 mm gauge length bi-axial
E
extensometer with a travel of ±10% in axial and ±2.5% in 1
1
shear was used for strain measurements. Axial-torsion
Strain
experiments were conducted employing both proportional
and 90o non-proportional loading schemes.
All experiments were conducted in room temperature
and approximately about 200 data points per cycle were
collected for further analyses. Tests were conducted under
software control running on a computer interfaced to the
control system of the testing machine. For true stress con-
trolled tests, the feedback from the extensometer was used 600
to compute true strain and there from the true stress con-
b
tinuously. Applied loads were concomitantly altered to 400
A
maintain true stress amplitude and mean as per the test
200
specification. σ0
Stress, MPa

E 2σ0
C E
0 1
3. Results and discussion B
1

-200 2σ0
During cyclic plastic deformation, different effects are
observed. They are: Bauschinger effect, cyclic hardening/ -400
softening, strain range effect, loading history memory, rat- SA333 C-Mn steel
Strain amplitude 1.6 %
cheting, mean stress dependent hardening, mean stress -600
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0
relaxation and non-proportional hardening. Movement of Strain, %
dislocations is the prime cause of plastic deformation,
and their diverse arrangement with in the material during Fig. 1. Schematic representation of Bauschinger effect: (a) forward and
load cycling creates various effects in cyclic plasticity. All reverse flow test; (b) hysteresis loop.

those effects in cyclic plastic deformation can be deduced

from hysteresis loop shapes which are discussed below. During monotonic tensile loading material is yielded at
point A, but at the time of reloading in compressive direc-
3.1. Bauschinger effect tion it yielded at point B, and in further reloading in tensile
direction it get yielded at point C. Therefore, it can be said
Bauschinger effect is discussed in brief below for the that yield stress (point B) during reloading in reverse direc-
sake of completeness. When materials are loaded uniaxial- tion is lower than the yield stress in loading (point A).
ly in one direction (e.g. in tension) into the plastic regime,
unloaded to zero stress level, then reloaded in reverse 3.2. Cyclic hardening/softening
direction (e.g. in compression), they may yield during the
reloading at a stress level lower than if the reloading were Alteration in the stress–strain response of a material
carried out in the original direction. Fig. 1a is the schematic subjected to cyclic loading is referred as cyclic hardening/
representation of Bauschinger effect (Bauschinger, 1886). A softening behavior. It is often studied by testing the mate-
virgin material starts deformed plastically in forward load- rial under fully reversed stress or strain controlled cycling.
ing after crossing the yield stress (r0). During reversed With controlled strain amplitude, a material is said to have
loading the material will display elastic deformation until exhibited cyclic hardening/softening when an increase/de-
the difference between the current stress and the stress crease in stress amplitude with progressive cycles is ob-
at which the load started to reverse reaches 2r0. First cycle served (Feltner and Laird, 1967a; Eifler and Macherauch,
hysteresis loop of SA333 C–Mn steel is shown in Fig 1b. 1990; Jiang and Zhang, 2008; Paul et al., 2010a,b,c). Under
708 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720

stress controlled cycling, the hardening/softening can be steels show a different combination of initial hardening
related to the increase/decrease in the strain amplitude followed by softening. It may be noted that the two mate-
over a period of time (Eifler and Macherauch, 1990; Paul rials have different crystal structures (SA333-bcc; 304LN-
et al., 2010a). In general a hard material is expected to fcc) and possess different phase constituents (SA333-fer-
cyclically soften and a soft material is expected to cycli- rite–pearlite; 304LN-austenite). Obviously, the dislocation
cally harden. generation and their subsequent movement upon defor-
One of the major factors that control the ability of a mation are liable to be different in the two materials and
material to cyclically harden/soften is its microstructural therefore a different deformation characteristic. The initial
constituents and their participation in the deformation hardening in the SA333 C–Mn steel can be attributed to the
process. The microstructural constituents here refers to rapid multiplication of dislocation upon cyclic deformation
relative volume fractions of different phases present in while that in 304LN stainless steel apart from dislocation
the material, their crystallographic structure and the num- generation to the partial transformation of austenite to
ber of slip systems available, the dislocation density, and martensite (Das and Tarafder, 2009) that impede smooth
the size, shape and distribution of second phase particles. dislocation movement. The subsequent softening can be
As a general rule of thumb, an annealed (read soft) metal/ related to the re-arrangement of the dislocation network.
alloy of high purity would exhibit cyclic hardening due to With progressive deformation, refinement of dislocation
dislocation multiplication (Feltner and Laird, 1967b; Gau- arrangement can lead to finer dislocation cell formation
din and Feaugas, 2004; Zhang and Jiang, 2005; Taleb and (Gaudin and Feaugas, 2004). In effect, the original grain is
Hauet, 2009). Presence of non-shearable precipitates divided into a number of sub-cells, thereby confining the
would additionally contribute to the cyclic hardening (typ- dislocation movement. This indicates the cyclic deforma-
ical example being the h precipitates in Al–Cu alloys) of tion and the hardening/softening behavior depends upon
material due to precipitate-dislocation interaction. On the the microstructural constituents described above and their
other hand, in materials that are hard or have been hard- interaction during the deformation event.
ened, inelastic strain cycling causes the existing dislocation The SA333 C–Mn steel (Fig. 3a) shows initial cyclic
structure to rearrange into a configuration such that there hardening for all strain amplitudes, however, the degree
is less resistance to deformation (Chai and Laird, 1987; of hardening is more prominent at higher strain ampli-
Giordana et al., 2010). The reconfigured structure enables tudes. Similarly, a distinct hardening is observed in
the dislocations to circumnavigate microstructural barriers 304LN stainless steel (Fig. 3b) for strain amplitudes >1%.
thereby promote greater dislocation mobility. The material
is easily deformed or cyclically softens.
The phenomena is very complex in polycrystalline
materials due to the own mechanical properties, ordering a
450
direction and directional properties of each grain. For
Stress amplitude, MPa

example, selected stress–strain hysteresis loops of 304LN

400
stainless steel under fully reversed symmetric strain
amplitude of 2.0% are given in Fig. 2. Shifting of the tip of 350
the hysteresis loop upwards indicates significant cyclic
hardening in this material. However, one of the most con- 300
Total Strain amplitude, %
1.6
venient methods to illustrate the hardening/softening 1.4
1.2
behavior is to examine the manner in which the stress 250 1.0
0.85
amplitude varies with the loading cycles. This is shown 0.7
in Fig. 3a and b for SA333 C–Mn steel and 304LN stainless 200 0.5

steel respectively for various strain amplitudes. Both the 1 10 100 1000
Number of Cycles

600
600
2.0% LCF
b Strain amplitude, %
500 550 0.2
0.5
Stress amplitude, MPa

400 0.77
500 1.0
300
1.2
200 450 1.4
100 1.6
Stress, MPa

0 400 1.8
2.0
-100 350
-200
-300 300
-400
250
-500
304LN stainless steel
-600 1 10 100 1000 10000 100000
-2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 Number of cycles
Strain, %
Fig. 3. Stress amplitude variation with number of cycles for fully reversed
Fig. 2. Hysteresis loops of 304LN stainless steel for fully reversed symmetric uniaxial strain cycling: (a) SA333 C–Mn steel; (b) 304LN
symmetric uniaxial strain cycling (strain amplitude 2.0%). stainless steel.
 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720 709

For lower strain amplitudes (0.2% and 0.5%), both the is responsible for cyclic hardening. Therefore, cyclic hard-
materials exhibited a rapid softening preceded by a rela- ening can be modeled by change in cyclic yield stress i.e.
tively low initial hardening. Although this indicates an isotropic hardening model.
alteration in the dislocation dynamics due to the variation
in strain amplitudes, it would be more appropriate to re- 3.3. Loading history memory
late this change in the hardening/softening behavior to
the loading conditions. Many engineering materials are sensitive to the preced-
Another important issue with regards to modeling cyc- ing load path. Subsequent cyclic plasticity of materials is
lic plasticity is the incorporation of hardening/softening therefore strongly influenced by the prior load history. In
behavior within material models. It has been a general fact, the first evidences of the influence of load history on
practice (Feltner, and Laird, 1967b; Pratt, 1967) to use the cyclic deformation behavior of materials have been re-
the yield stress variation to model the cyclic hardening/ ported four decades ago (Lukáš and Klesnil, 1973; Laird
softening behavior of materials. To verify this, the stress– and Finney, 1975). It is documented that for cross-slip
strain hysteresis loops of different cycles are translated in materials (similar to SA333 C–Mn steel in this work) below
such a manner that the common linear regions of the load- a threshold strain the stress–strain response is history
ing branches of all the cycles coincide and a universal independent. Above this threshold strain, the cyclic plastic
envelope curve can be drawn to describe the loading response is prior history dependent. In case of planar slip
branches. The results are shown in Fig. 4a and b for the materials (e.g. 304LN stainless steel here), the cyclic plas-
SA333 C–Mn steel (at 1.6% strain amplitude) and 304LN ticity is always sensitive to the prior load history. Influ-
stainless steel (for 2% strain amplitude) respectively for se- ences of prior load history on the cyclic hardening/
lected cycles. It appears that increase in the elastic portion softening response of materials are also reported in recent
of the hysteresis loop is responsible for cyclic hardening times (Kunz et al., 2001; Jiang and Zhang, 2008; Jiang and
and thus the hardening/softening can be explained by Sehitoglu, 1994a, b). Generally, prior load history affects
change in the yield stress. It can be said from Fig. 4 that in- initial cyclic hardening/softening behavior of materials;
crease in elastic portion of hysteresis loop (i.e. yield stress) however, the stabilized stress–strain response remains
unaltered (Zhang and Jiang, 2004).
To verify the influence of prior load history on the cyclic
stress–strain response, we performed fully reversed strain
a 800 Strain amplitude 1.6% controlled cyclic deformation studies on (i) specimens ten-
700 sile pre-strained to different pre-determined strains and
Number of cycle
600 1 (ii) on specimens pre-cycling at constant pre-determined
500
2 strain amplitude. The pre-strain cycling was carried out
Stress, MPa

5
400 10 under strain control analogous to the low cycle fatigue pro-
20 cedure given in the experimental section except that the
300 40
70 cyclic deformation was imposed for a pre-determined
200 100
number of cycles. The influence of tensile pre-strain on
100 cyclic deformation behavior of SA333 C–Mn steel is shown
0 in Fig. 5 along with base line fatigue data (without any
-100 SA333 C-Mn steel prior history effect) and that of the pre-strain cycling in
Fig. 6 for the same material. The results in Fig. 5 infer that
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Strain, % after the imposition of tensile pre-strain, the stress re-
sponse during subsequent cyclic deformation under strain
b 1000 304LN stainless steel
800
550
Strain amplitude 1.0%
600 single step
500 5.75% pre-straining
Stress, MPa

400 7.8% pre-straining

Stress amplitude, MPa

14.0% pre-straining

200 450
Number of cycles
100
0 50
30
10 400
5
-200 3
2
1
350
-400
-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
SA333 C-Mn steel
Strain, % 300
1 10 100 1000
Fig. 4. Hysteresis loop at selective number of cycle, first cycles compres- Number of cycles
sive tip translated to origin (0, 0) and other loops are translated in such a
way that their loading branch overlapped with fist cycles hysteresis loop’s Fig. 5. Dependence of cyclic hardening/softening on loading history:
loading branch: (a) strain amplitude of 1.6% for SA333 C–Mn steel; (b) Stress amplitude versus number of cycles plot for SA333 C–Mn steel, at
strain amplitude of 2.0% for 304LN stainless steel. 1.0% strain amplitude and pre-straining of 0.0%, 5.75%, 7.8% and 14.0%.
710 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720

420 translating tensile tips) overlap and form a common enve-

20 cycles 1.6 % LCF followed by 0.7 % LCF
Single step 0.7 % LCF lope curve – termed as Master curve, then the material is
400 said to follow Masing model (Ellyin and Kujawaski,
1984). The results for the two steels in this investigation
are shown in Fig. 7. It can be seen from this figure that both
Stress, MPa

380
304LN stainless steel and SA333 C–Mn steel exhibit non-
Masing behavior. Masing behavior is not a universal phe-
360 nomenon in engineering materials. Plumtree and Abdel-
Raouf (2001) have shown that some materials exhibit Mas-
340 ing behavior while others do not. Fan and Jiang (2004)
noted Masing behavior in a pressure vessel steel at
SA333 C-Mn steel
320
300 °C and 420 °C. Maier et al. (2006) observed Masing
1 10 100 1000 behavior in an ultrafine grained copper.
Number of cycles It is not necessary that all materials would exhibit Mas-
Fig. 6. Dependence of cyclic hardening/softening on loading history:
ing behavior under all experimental conditions. Whether
Stress amplitude versus number of cycles plot for SA333 C–Mn steel, or not a material will exhibit Masing behavior is strongly
single step 0.7% LCF and 20 cycles 1.6% LCF followed by 0.7% LCF. influenced by the distribution of second phase particles
and the manner in which dislocation movement is im-
peded by them. It is shown that Masing behavior will ap-
control is completely altered. The base line data showed a pear in precipitation hardened alloys containing closely
significant hardening upon cyclic deformation, whereas spaced unshearable particles, while in those with particles
after imposition of tensile pres-strain, the material exhibits spaced far apart, a non-Masing behavior is observed
cyclic softening behavior. As the magnitude of the tensile (Plumtree and Abdel-Raouf, 2001). A material system that
pre-strain is increased, the cyclic softening response was exhibits Masing behavior in one load path might show a
also noted to be high. This increase in cyclic softening is non-Masing nature in another load path or vice versa. In
significant for the first few cycles when compared to the an austenitic stainless steel (Ye et al., 2006), Masing behav-
base line cyclic response. This implies that the material ior was observed in incremental load step whereas for a
memorizes the prior loading history and influences the
subsequent hardening/softening response. The material
memory, however, gradually fades away with subsequent
cycles and finally converges with the base line stress re- a 1000 Strain amplitude, %
sponse. Similarly, after pre-cyclic straining (Fig. 6), a 900
0.5
0.7

change in the cyclic stress–strain response of the material 0.85

1.0
800 1.2
may be noted. After the higher amplitude (1.6%) prior load- 1.4
1.6
700
ing history, the material shows significant softening as op-
Stress, MPa

600
posed to the hardening exhibited by the same material in
the absence of prior cycle. The observations noted here 500
on the SA333 C–Mn steel gather support from the investi- 400
gations made by Jiang and Zhang (2008) on 304 stainless 300
steel and, Zhang and Jiang (2005) on OFHC copper. The 200
observations made in this investigation and those at the 100
SA333 C-Mn steel
references cited above suggest that the cyclic hardening/ 0
softening is not only dependent on material but also on 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
the loading magnitude and loading history. Strain, %

Strain amplitude, %
3.4. Strain range effect (non Masing behavior) b 1200 0.2
0.5
0.77
1.0
1000 1.2
Masing (1926) while modeling the Bauschinger effect 1.4
1.6
(1886) using a number of elastic-perfectly plastic elements
Stress, MPa

1.8
800 2.0
each with different yield levels observed that while load
reversal (unloading from peak tensile point as origin) each 600
element was required to deform twice its preceding elastic
range in order to cause complete yielding. Thus, a material 400
that follows the Masing model would have identical load-
200
ing and unloading branches of the hysteresis loops, and
304LN stainless steel
their elastic region remain unaltered with a change in 0
strain amplitude. Experimentally, Masing behavior in a 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
material is verified by bringing the compressive tips of Strain, %
(sometimes the tensile tips) all the stable hysteresis loops Fig. 7. Stable hysteresis loop at different strain amplitude, compressive
belonging to various strain amplitudes to a common origin. tip translated to origin (0, 0): (a) SA333 C–Mn steel; (b) 304LN stainless
If all the loading branches (or unloading branches in case of steel.
 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720 711

companion specimen test (constant strain amplitude test) the magnitude of the non-Masing stress, an additional
a non-Masing behavior was noted. Raman and Padmanab- quantity that is required to shift the corresponding tip to
han (1996) demonstrated that Masing behavior can be ob- the origin. It may be noted that this extension (of linear re-
served in 304LN stainless steel when cold rolled to 30% gion) represents the cyclic expansion of the proportional
prior to testing. Christ and Mughrabi (1996) and Li and limit and therefore the non-Masing behavior or the strain
Laird (1993) showed that in both polycrystal and single range effect can be modeled by change in the cyclic yield
crystal copper, Masing behavior was not present. However, stress. The results show that non-Masing behavior is both
under a special incremental step test (Christ and Mughrabi, material and strain range dependent. It may be noted that
1996) and ramp loading conditions (Wang and Laird, incorporation of non-Masing behavior in cyclic plasticity
1988), Masing behavior could be observed. models is more significant during large range strain
From microscopic point of view Masing behavior relates cycling.
to a stable microstructural condition and dislocation sub- Sometimes, strain energy based approaches are also
structure (Maier et al., 2006) against fatigue cycles. In a successfully used to quantify the non-Masing behavior. It
study of hysteresis loop shapes for different metals and al- should, however, be noted that strain energy approach is
loys (Plumtree and Abdel-Raouf, 2001), it was noted that suitable only for fully reversed loading conditions because
non-Masing behavior was observed for group of metals the plastic strain energy are not sensitive to the mean
and alloys in which cyclic deformation is controlled by ma- stress effects.
trix properties, and dislocation cells are formed at rela-
tively low strain ranges. In addition, Mughrabi and Christ 3.5. Mean stress relaxation
(1997) have pointed out that phase stability is a pre-requi-
site for the Masing phenomena. Investigations on micro- So far we have seen situations wherein the cyclic load
structural changes during cyclic deformation and (or strain) reversals are symmetric. During asymmetric
observations of dislocation cell development (Sivaprasad strain cycling, the presence of a constant mean strain,
et al., 2010a; Gaudin and Feaugas, 2004; Mayama et al., would introduce a mean stress that would relax gradually
2008) showed that both the microstructural phase constit- upon progressive cyclic deformation (Morrow and Sinclair,
uent and the dislocation cell size need not necessarily re- 1958; Fang and Berkovits, 1994; Arcari et al., 2009). Mean
main stable upon cyclic deformation in all materials. The
304LN stainless steel investigated in this work belongs to
this category. Such instability will produce a deviation a 900 Strain amplitude, %
0.5
from Masing behavior. In materials that do not suffer phase 800 0.7
0.85
transformation upon cyclic deformation (such as the 700 1.0
1.2
1.4
SA333 C–Mn steel in this work), the non-Masing behavior 600 1.6

is considered mainly due to heterogeneous dislocation 500

Stress, MPa

arrangement. Observations of dislocation arrangements 400

during cyclic deformation at different strain amplitudes 300
in many materials (Sivaprasad et al., 2010a; Skelton
200
et al., 1997; Gaudin and Feaugas, 2004; Golos and Ellyin,
100
1987) showed a gradual change in dislocation cell struc-
0
ture with the dislocation density (dislocation veins) within
the cell interior gradually depleting. The material within -100 SA333 C-Mn steel

the cell interior is therefore expected to be soft. The cell 0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 3.0
walls representing regions of high dislocation density is Strain, %
expected to possess high yield level. Such a material can
be treated as a composite consisting of hard and soft re- b 800 Strain amplitude, %
0.2
gions strained in parallel as proposed by Masing model 0.5
0.77
(Magnin et al., 1989). Such transient dislocation cell struc- 600 1.0
1.2
ture and/or phase instability can introduce additional yield 1.4
400 1.6
Stress, MPa

levels upon cyclic deformation thereby introducing a devi- 1.8

2.0
ation from Masing behavior. In general, Masing phenom- 200
ena require a stable microstructural event. A
0
microstructural event here refers to both phase stability
and stable dislocation density. Any deviation from this ba- -200
sic requirement is an indication that conformity to Masing
behavior will not be maintained. -400 304LN stainless steel
One of the methods of measure and model the non- -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Masing behavior is by translating the loading branches of Strain, %
stable hysteresis loops of different strain amplitudes in
such a fashion that the common linear regions of all the Fig. 8. Stable hysteresis loop at different strain amplitude, compressive
tip of lowest strain amplitude hysteresis loop translated to origin (0, 0)
loops coincide. This is shown in Fig. 8a and b for the two and other loops are translated in such a way that their loading branch
materials investigated. The extension of the linear region overlapped with lowest strain amplitude hysteresis loop’s loading
for any strain range (taking origin as reference) is therefore branch: (a) SA333 C–Mn steel; (b) 304LN stainless steel.
712 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720

stress relaxation during asymmetric strain cycling with a can be categorized as (i) material ratcheting and (ii) struc-
mean strain of 5.5% and strain amplitude of 1% is shown tural ratcheting and it is necessary to understand the dif-
in Fig. 9. The results infer that the presence of a tensile ference between the two phenomena (Hartwig, 1996).
mean strain influences the tensile peak stress that relaxes Material ratcheting can occur in the absence of structural
with progressive cycles. The compressive peak stress re- effects (i.e., if stress is homogeneously distributed in a
mains unaltered. It will be interesting to also know if mere structure) and thus purely a material related effect and
presence of mean strain influences the peak stress irre- can be examined by material tests and theoretically de-
spective of the manner in which it is imposed or not. To scribed constitutive equations. Hardening behavior, mag-
verify this, we additionally performed an experiment in nitude of mean stress, anisotropy, shape of yield surface,
which a specimen was tensile pre-strained to 5.75% total loading pattern (proportional or non-proportional) and
strain (5.5% plastic strain/residual strain) and was fol- test temperature are some of the variables that would
lowed by imposition of 1% symmetric strain cycling. Com- influence the material ratcheting phenomena. The investi-
parisons of the two results, one is 1% symmetric strain gation presented in this work belongs to this category.
cycling with 5.5% mean strain and other one is tensile Structural ratcheting can occur due to inelastic material
pre-strained to 5.75% total strain (5.5% residual strain) behavior under cyclic loading even if there is no material
followed by 1% symmetric strain cycling, are given in ratcheting (Hartwig, 1996). For example, a purely kine-
Fig. 9d. The results show that the material response is sim- matic hardening material in uniaxial stress state at a con-
ilar in both the cases indicating the fact that mean stress stant temperature can still exhibit ratcheting due to
relaxation depends mainly on the magnitude of mean inhomogeneity of stress state in the structure. Structural
strain irrespective of the manner in which it was induced. ratcheting is generally verified by detailed or simplified
inelastic analysis and this is not covered within the scope
3.6. Ratcheting of this work.
As a matter of fact, a number of investigations have
Ratcheting is a secondary deformation process that pro- been carried out in the last two decades to understand
ceeds in every cycle and liable to severely deteriorate the and model the material ratcheting behavior. Basically, all
performance of a component due to cumulative effect of these investigations can be broadly categorized into (i)
fatigue damage and progressive strain accumulation in a studies on laboratory specimens and (ii) those on full scale
particular direction. It should be noted that ratcheting components (Kulkarni et al., 2003). All these investigations

a 500 SA333 C-Mn steel b 50

400 Mean strain 5.5 %
Strain amplitude 1.0 %
300 40
Mean stress, MPa

200
Stress, MPa

30
100
0 20
-100
-200 10

-300
0
-400 Mean strain 5.5 %
Strain amplitude 1.0 % SA333 C-Mn steel
-500 -10
0 1 2 3 4 5 6 1 10 100 1000
Strain, % Number of cycles

c 500 d 500
Mean stress relaxation
400 (Mean strain 5.5 %, Strain amplitude 1.0 % )
Tensile pre-strain followed by LCF
300 475
Stress amplitude, MPa

(Pre-strain 5.75 %, LCF stran amplitude 1.0 %)

Maximum peak stress
200 Minimum peak stress
Stress, MPa

100 450
SA333 C-Mn steel
0
-100 Mean strain 5.5 % 425
Strain amplitude 1.0 %
-200
-300 400
-400
SA333 C-Mn steel
-500 375
1 10 100 1000 1 10 100 1000
Number of cycles Number of cycles

Fig. 9. Mean stress relaxation at mean strain of 5.5% and strain amplitude of 1.0% on SA333 C–Mn steel: (a) stress–strain plot; (b) mean stress variation with
number of cycles; (c) peak stress variation with number of cycles; (d) change in stress amplitude with cycling and comparison with monotonic pre-straining
5.75% followed by 1.0% symmetric strain cycling.
 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720 713

were carried out to understand experimentally cyclic plas- 600 60

a True stress amplitude, MPa

Engineering ratcheting strain, %

tic deformation behavior and to develop a suitable robust Engineering stress amplitude, MPa
550 True maximum stress, MPa 50
constitutive model which can be plugged into a finite ele- Engineering maximum stress, MPa
ment platform for predicting ratcheting strain and subse- Engineering ratcheting strain, % 40
500

Stress, MPa
quent cyclic plastic deformation. For example, the
30
material ratcheting behavior studied by Hassan and 450
Kyriakides (1992, 1994a,b), Hassan et al (1992), Jiang and 20
Sehitoglu (1994a,b), Delobelle et al (1995), Haupt and 400
10
Schinke (1996), Kang et al. (2002), Zhang and Jiang
350
(2005), Khan et al. (2007), Sivaprasad et al. (2010b) and SA333 C-Mn steel
0
Paul et al. (2010a) under uniaxial and multiaxial loading 300 -10
conditions are a few worth mentioning. In a majority of 0 200 400 600 800 1000 1200
investigations, including the works cited above, the rat- Number of cycles
cheting behavior has been investigated employing a cyclic
waveform under load or engineering stress control that is b 1000 True stress amplitude, MPa

Engineering ratcheting strain, %

Engineering stress amplitude, MPa 100
based on the original dimensions of the specimen. One True maximum stress, MPa
Engineering maximum stress, MPa
possible justification for this consideration could be the 900 80
Engineering ratcheting strain, %
accumulated strains were not sufficient enough to cause

Stress, MPa
800 60
a measurable change in the specimen dimension. However,
the magnitude of strain accumulation during asymmetric 700
40
load/stress cycling depends upon the magnitude of mean
600
stress (or load) and stress (or load) amplitude. For high 20
magnitudes of mean stress and stress amplitude, the accu- 500
mulated strain over a period of time may be reasonable to 304LN stainless steel 0
produce a considerable change in the specimen cross-sec- 400
-100 0 100 200 300 400 500 600 700 800
tional area. If appropriate correction factors are not ac- Number of cycles
counted for these dimensional alterations, the true mean
stress and true stress amplitude are liable to increase Fig. 10. Alteration of true maximum stress and stress amplitude with
uncontrollably, leading to overload failure of the specimen number of cycles in engineering stress control test: (a) SA333 C–Mn steel;
(b) 304LN stainless steel.
rather than due to ratcheting fatigue (Paul et al.,
2010a,b,c). The life prediction will therefore be inappropri-
ate. A systematic investigation on the ratcheting behavior
True strain refers to the ratio of change in gauge length
due to engineering and true stress controlled asymmetric
to instantaneous gauge length which can be described by
cycling is presented in this work. From the engineering
Eq. (1)
point of view, engineering stress controlled ratcheting is
Z L  
important since in actual engineering applications, mea- DL L
surement and regulation of true stress (and strain) will
er ¼ ¼ ln ¼ ln ð1 þ er Þ ð1Þ
L0 L L0
not be possible. At the same time, from materials perspec-
tive, true stress controlled experiments is essential to Assuming metal volume is not changing, true stress can be
understand and model the ratcheting response of material. determined from Eq. (2)
It is often difficult to decide a safe percentage of strain r ¼ Sð1 þ er Þ ð2Þ
accumulation so that dimensional alterations in the speci-
men could be ignored or vice versa. The safest method where, r and er are true stress and ratcheting strain, S and
should be to study the material ratcheting behavior under er are engineering stress and ratcheting strain, L0 and L are
true stress control, since true stress and true strain are the initial and final gauge length of the specimen.
actual manifestation of material response under external Amount of area reduction can be calculated from vol-
loads. In this work, ratcheting behavior of SA333 C–Mn ume incompressibility condition, as shown in Eq. (3)
steel and 304LN stainless steel has been investigated under
DA er
both engineering and true stress control to bring out the ¼ ð3Þ
A 1 þ er
difference in the estimated life.
The evolution of stress and strain during an engineering where, A is the original cross sectional area and DA is the
stress controlled experiment is shown in Fig. 10a for SA333 reduction in cross sectional area at er ratcheting strain.
C–Mn steel with mean stress (80 MPa)–stress amplitude The ratcheting strain accumulation prior to specimen
(310 MPa) combination and in Fig. 10b for 304LN stainless failure was noted to be 50.5% for SA333 C–Mn steel and
steel with mean stress (120 MPa)–stress amplitude 93% for 304LN stainless steel. Accumulation of such huge
(420 MPa) combination to bring out the implications of magnitudes of ratcheting strains produced a 1.505
engineering stress controlled ratcheting experiments. It (SA333) and 1.931 (304LN) times increase in the actual
can be noted that in both the steels a significant magnitude true stress, and a corresponding 33.56% (SA333) and
of ratcheting strain has been accumulated. The ratcheting 48.21% (304LN) reduction the specimen cross-section area.
strain (er) here represents
 the average of the maximum This significant reduction of cross-sectional area produced
and minimum strain emax þe 2
min
of the hysteresis loop. an uncontrollable increase in both the true stress
714 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720

amplitude and true maximum stress, whereas the corre- Therefore, to reach true ultimate tensile stress more engi-
sponding engineering values were maintained at a con- neering ratcheting strain accumulation is required com-
stant level. Stress–strain response of selected loading pared to engineering tensile strain, as engineering
cycles of two different piping materials subjected to true ultimate tensile stress is higher than maximum engineer-
stress controlled uniaxial loading with both positive and ing ratcheting stress. Therefore, it is thought to be the
negative mean stress are shown in Fig. 11. We will discuss probable cause of higher ratcheting strain accumulation
the difference between engineering and true controlled in engineering stress controlled ratcheting than uniform
experiments on ratcheting life subsequently. engineering tensile strain before necking.
As noted in those ratcheting experiments, 93.0% The evolution of ratcheting strain in SA333 C–Mn steel
(304LN) and 50.5% (SA333) elongations are achieved in and 304LN stainless steel under true stress control tests are
engineering stress controlled ratcheting tests where as shown in Fig. 12. It is observed that the ratcheting strain
uniform elongation during tensile test of those materials accumulation occurs in the direction of mean stress that
are around 52.8% (304LN) and 17.18% (SA333). Therefore, is a tensile mean stress leads to strain accumulation in ten-
material can accommodate much higher strain during rat- sile direction and a compressive mean stress produces
cheting without necking. Initiation of micro void is the strain in the compressive direction. A similar observation
probable cause of necking and their coalescence leads to is noted by Jiang and Sehitoglu (1994a,b). Ratcheting rate
failure (Floreen and Hayden, 1970; Han and Margolin, decay is also noted for all loading conditions in Fig. 12. Ini-
1989). Local instability/necking in tensile test starts at tially, the ratcheting strain accumulation rate is higher fol-
the point of UTS (ultimate tensile strength). Therefore, it lowed by a gradual decrease and then stabilizes at a
can be said that necking starts when the true stress reach constant rate throughout the life. It is also documented
a critical value which is equal to true ultimate tensile that both the ratcheting strain and fatigue life increases
stress. Both engineering stress and strain increases with with increasing mean stress.
time during tensile test, but in case of ratcheting maximum The improvement in fatigue life due to increasing mean
engineering stress is maintained a constant value and stress can be explained by plastic strain energy during cyc-
accumulated engineering strain increases with time. Com- lic deformation. The plastic strain energy represented by
bine influence of both engineering stress and strain is re- the area enclosed by stress–strain hysteresis loop signifies
flected in true stress calculation as shown in Eq. (2). the ability of the material to cyclically deform. More the

a 500
1-3 100 500 1000 2000 5000 8200 b 400
4000 2500 1000 500 100 1-3
400 300
True stress, MPa

300 200
True stress, MPa

200 100

100 0
0 -100

-100 -200

-200 -300

-300 3500 7000 -400

Mean stress 80 MPa Mean stress -40 MPa
SA333 C-Mn steel SA333 C-Mn steel Stress amplitude 350 MPa
Stress amplitude 350 MPa -500
-400
0 5 10 15 20 25 -16 -14 -12 -10 -8 -6 -4 -2 0
True strain, % True strain, %

800
c Mean stress 120 MPa
Stress amplitude 420 MPa
d 400 820 600 400 200 100 50
304Ln stainless steel
1-5
600 1-3 50 500 1500 3600
True stress, MPa

200
True stress, MPa

400
0
200

-200
0

-400
-200
Mean stress -60 MPa
304Ln stainless steel Stress amplitude 420 MPa
-400 -600
0 2 4 6 8 10 12 14 16 18 20 22 -15.0 -12.5 -10.0 -7.5 -5.0 -2.5 0.0
True strain, %
True strain, %

Fig. 11. Shifting of hysteresis loop along strain axis during true stress control asymmetric stress cycling: (a) 80 MPa mean stress and 350 MPa stress
amplitude for SA333 C–Mn steel; (b) 40 MPa mean stress and 350 MPa stress amplitude for SA333 C–Mn steel; (c) 120 MPa mean stress and 420 MPa
stress amplitude for 304LN stainless steel; (d) 60 MPa mean stress and 420 MPa stress amplitude for 304LN stainless steel.
 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720 715

a 35
SA333 C-Mn steel
a 14000 Number of cycles
5
0.7
30 Hysteresis loop area

Plastic strain amplitude, %

12000 Plastic strain amplitude
True ratcheting strain, %

25 4 0.6

Hysteresis loop area, MJ/m3

10000

Number of cycles
20
0.5
15 Stress amplitude 350 MPa 3
8000
10 Mean stress 0.4
120 MPa
5 80 MPa 6000 2 0.3
0 40 MPa
0 MPa 4000 0.2
-5 Mean stress -40 MPa
1
-10 2000 0.1
-15 SA333 C-Mn steel
0 0 0.0
-20 -60 -40 -20 0 20 40 60 80 100 120 140
0 2000 4000 6000 8000 10000 Mean stress, MPa
Number of cycles
16
b 7000 Number of cycle
2.5
b 25 304LN stainless steel
6000
Hysteresis loop area, MJ/m2
Plastic strain amplitude, %
14

Hysteresis loop area, MJ/m

20 2.0 12

Plastic strain amplitude, %

5000

Number of cycles
Ratcheting strain, %

3
15 10
4000 1.5
10 8
Stress amplitude 420 MPa
5 Mean stress 3000 6
180 MPa 1.0
0 120 MPa 4
60 MPa 2000
-5 0 MPa 0.5 2
-60 MPa 1000
-10 0
304LN stainless steel
0 0.0
-15 -100 -50 0 50 100 150 200
0 1000 2000 3000 4000 5000 6000 7000 Mean stress, MPa
Number of cycles
Fig. 13. Number of cycles to failure, hysteresis loop area and plastic strain
Fig. 12. True ratcheting strain versus number of cycles: (a) stress amplitude versus mean stress: (a) stress amplitude of 350 MPa and mean
amplitude of 350 MPa and mean stress of 40, 0, 40, 80 and 120 MPa stress of 40, 0, 40, 80 and 120 MPa for SA333 C–Mn steel; (b) stress
for SA333 C–Mn steel; (b) stress amplitude of 420 MPa and mean stress of amplitude of 420 MPa and mean stress of 60, 0, 60, 120 and 180 MPa for
60, 0, 60, 120 and 180 MPa for 304LN stainless steel. 304LN stainless steel.

hysteresis loop area, the material is expected to absorb cheting strain variation with number of cycles by the two
more energy and hence shorter will be the fatigue life methods for SA333 C–Mn steel is shown in Fig. 14a and b
(Xia et al., 1996). The energy calculated from saturated and that of 304LN stainless steel in Fig. 14c. The fatigue life
hysteresis loops is plotted against the applied mean stress of both the materials was more during true stress con-
in Fig. 13a and b for SA333 C–Mn steel and 304LN stainless trolled experiments than under engineering stress control.
steel respectively. The figures also include the correspond- The reasons for the decrease fatigue life during engineering
ing evolution of plastic strain amplitudes. If the zero mean stress control experiments were thought to be due to (i) ra-
stress condition is taken as reference, it may be noted from pid accumulation of ratcheting strain than under true
Fig. 13 that the hysteresis loop area decreases with stress control as can be witnessed in Fig. 14, (ii) a continu-
increasing mean stress irrespective of its direction. At the ous increase in the actual true stress on the specimen as
same time, it is also interesting to see that the plastic strain seen in Fig. 10, (iii) instability and necking produced due
amplitude decreases with increasing mean stress. The re- to large reduction of cross-sectional area. On contrary, dur-
sults thus infer that increasing mean stress leads to de- ing true stress controlled experiments, failure always oc-
crease in both plastic strain amplitude and plastic strain curred by initiation and growth of fatigue cracks. All
energy as a result of which the fatigue life increases. The these observations are true for a positive mean stress.
observation holds good for both the varieties of steel inves- However, when the mean stress is negative (Fig. 14c), it ap-
tigated here. A similar behavior has been reported in a car- pears that the fatigue life increases during engineering
bon steel (Kliman et al., 1980), 304 stainless steel (Turner stress controlled situation when compared to that during
and Martin, 1980; Lorenzo and Laird, 1984; Mughrabi true stress control. The reason for this anomaly is that
et al., 1997), polycrystalline copper (Eckert et al., 1987; Lu- the presence of compressive mean stress drives the rat-
kas and Kunz, 1989) and for polycrystalline nickel (Holste cheting strain in compressive direction (a case contrary
et al., 1994). to the increasing true stress depicted in Fig. 10) as a result
In order to draw the difference between the engineering of which true stress amplitude reduces with progressive
and true stress controlled ratcheting behavior of the mate- cycling, leading to an increase in life.
rial, some of the above tests were repeated under engineer- The comparison brings forth the inadequacy in engi-
ing stress controlled mode while keeping all other neering stress controlled experiments to understand the
experimental conditions identical. A comparison of the rat- ratcheting behavior of materials, specifically involving
716 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720

stress controlled tests would yield acceptable results. Rat-

a 55
Stress amplitude 310 MPa SA333 C-Mn steel
50 Engineering stress controlled tests
cheting strain evolution in the two modes of experiments
45
Mean 80 MPa depends up on the material characteristics and loading
Mean 40 MPa
conditions. For these reasons, it would be a good practice
Ratcheting strain, %

40 True stress controlled tests

Mean 80 MPa
35 Mean 40 MPa if the ratcheting response is investigated under true stress
30 control irrespective of the expected magnitude of strain
25 accumulation.
20
15 3.7. Mean stress dependent hardening
10
5 It is noted during ratcheting that an increasing mean
0 stress led to an improvement in fatigue life. Fatigue life is
1 10 100 1000 10000
intimately connected to the hardening/softening response
Number of cycles
of the material and an elegant way of representing the
55 hardening/softening behavior is to examine the change in
b 50
Mean stress 80 MPa
Engineering stress controlled tests the plastic strain energy. It is seen in Fig. 15 that there is
Amplitude 350 MPa
45 Amplitude 310 MPa a systematic decrease in both the plastic strain energy
Ratcheting strain, %

40 True stress controlled tests and plastic strain amplitude with increasing mean stress.
Amplitude 310 MPa
35 Amplitude 350 MPa This implies that SA333 C–Mn steel and 304LN stainless
30 steel suffer an additional hardening due to the presence
25 of mean stress. For low magnitudes of mean stress the
20 material could accommodate larger plastic strains than
15 for the higher magnitudes of mean stress. The cyclic dam-
10 age imposed on the material at low magnitudes of mean
5
SA333 C-Mn steel
stress is therefore expected to be more than for higher
0 magnitudes of mean stress. This explains the relative low
10 100 1000 10000 ratcheting life of SA333 C–Mn and 304LN stainless steels
Number of cycles
at low magnitudes of mean stress levels.
c 100 Stress amplitude 420 MPa
In order to further understand this mean stress depen-
True mean stress dent hardening, the stress–strain responses of the two
180 MPa
80 120 MPa steels at the saturated cycles are plotted in Fig. 15a and b
Ratcheting strain, %

60 MPa
0 MPa for SA333 C–Mn and 304LN stainless steel respectively. It
60 -60 MPa
Engg. mean stress
may be noted that for the purpose of comparison, all the
180 MPa hysteresis loops in Fig. 15 are translated to a common ori-
40 120 MPa
60 MPa gin. For both the steels, the figure shows that with increas-
-60 MPa
20 ing mean stress (i) the size of the hysteresis loops
gradually decrease and (ii) the non-linear portion of the
0 loading branch becomes steeper. The first observation indi-
304LN stainless steel cates that the plastic strain, represented by the width of
-20 the loop, a manifestation of damage accumulated in the
1 10 100 1000 10000
Number of cycles material, decreases with increasing mean stress. The sec-
ond observation indicates an increase in the hardening re-
Fig. 14. Comparison between true and engineering stress control rat- sponse of the material. It may be noted that the change in
cheting test: (a) stress amplitude of 310 MPa and mean stress of 40 and the hardening response of the material due to mean stress
80 MPa for SA333 C–Mn steel; (b) mean stress of 80 MPa and stress
starts occurring right from the first cycle and a typical
amplitude of 310 and 350 MPa for SA333 C–Mn steel; (c) stress amplitude
of 420 MPa and mean stress of 60, 0, 60, 120 and 180 MPa for 304LN change in the hysteresis loop shape of SA333 C–Mn steel
stainless steel. for various mean stress levels are given in Fig. 15c. The
observation provides evidences that the two steels in this
investigation show an additional hardening that is depen-
large strains and suggests true stress control experiments dent on mean stress.
for test conditions that are expected to produce large strain
accumulation. Referring to Fig. 14, it can be seen that both 3.8. Non proportional hardening
the methods produce almost identical strain accumulation
for the initial few cycles. It is only with the progressive cy- Additional hardening of material under non-propor-
cles, the strain accumulation by the two methods is largely tional loading is called a non-proportional hardening. If
different. The assumption (as in many investigations avail- the stress or strain ratios at different directions vary pro-
able in literature) that for small magnitudes of ratcheting portionally with time then the loading type is called pro-
strain accumulation the investigation employing engineer- portional loading otherwise non-proportional loading. For
ing stress control ignoring the dimensional changes in the example, in case of axial-torsional direction if the phase
specimen seems to be logical. However, it is very difficult difference and the mean values in both the directions are
to specify a cycle/strain limit below which an engineering zero, then it is a proportional. If there is a phase difference
 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720 717

a 800 Stress amplitude 350 MPa

Mean stress
a 1.00
700 0 MPa 0.75
-40 MPa
40 MPa

Shear strain (γ), %

0.50
True stress, MPa

600 80 MPa
120 MPa
500 0.25

400 0.00

300 -0.25

200 -0.50

100 -0.75
SA333 C-Mn steel
0 -1.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2 -0.75 -0.50 -0.25 0.00 0.25 0.50 0.75
True strain, % Axial strain (ε), %

b 1000 304LN stainless steel b 250

200

Shear stress (τ), MPa

800 150
True stress, MPa

100
600 50
0
Stress amplitude
400 420 MPa -50
mean stress
180 MPa -100
200 120 MPa
60 MPa -150
-60 MPa -200
0 MPa SA333 C-Mn steel
0 -250
0.0 0.5 1.0 1.5 2.0 2.5 3.0 -500 -400 -300 -200 -100 0 100 200 300 400 500
True strain, % Axial stress (σ), MPa

c SA333 C-Mn steel

c 600
400 Axial stress-strain (σ-ε)
Shear stress-strain (τ-γ)
400
True stress, MPa

200
Stress, MPa

200
0
0

-200 -200
Stress amplitude 350 MPa
Mean 0 MPa
Mean 40 MPa -400
-400 Mean 80 MPa
Mean 120 MPa
SA333 C-Mn steel
-1 0 1 2 3 4 5 -600
True strain, % -1.00 -0.75 -0.50 -0.25 0.00 0.25 0.50 0.75 1.00
Strain, %
Fig. 15. True stress–strain plot of saturated cycle, compressive tip of
Fig. 16. Symmetric 90° out of phase strain cycling on SA333 C–Mn steel,
hysteresis loops are translated to common origin: (a) stress amplitude of
0.6% axial strain and 0.78% shear strain (a) applied strain (b) Stable stress
350 MPa and mean stress of 40, 0, 40, 80 and 120 MPa for SA333 C–Mn
response (c) Stable stress–strain response.
steel; (b) stress amplitude of 420 MPa and mean stress of 60, 0, 60, 120
and 180 MPa for 304LN stainless steel; (c) First cycle true stress–strain
plot at stress amplitude of 350 MPa and mean stress of 0, 40, 80 and
120 MPa for SA333 C–Mn steel. dent. Doong et al (1990) have reported that no additional
hardening is observed in aluminium alloys but a significant
additional hardening in oxygen free copper, 310 and 304
or mean value between axial and torsional loading then stainless steels. Krempl and Lu (1984) observed that non-
non-proportionality starts. Since the cyclic loading experi- proportional hardening is dependent on loading path. Itoh
enced by many engineering components is expected to be et al (1992) reports that amount of non-proportional hard-
multiaxial, understanding the non-proportional hardening ening is dependent on stalking fault energy of the materi-
becomes important. The extreme level of non-proportion- als. Doquet and Clavel (1996) report that extra hardening
ality can be generated by creating 90o phase difference in FCC solid solutions during 90o out of phase tension-tor-
and therefore in many investigations the non-proportional sion experiments increases with decrease in stalking fault
hardening behavior has been investigated at 90o out of energy. In non-proportional loading, the maximum shear
phase between tension and torsion strain paths. stress plane rotates with time in every cycle thus triggering
Recent investigations have shown that the extra hard- many slip planes and their mutual interaction thus attrib-
ening produced by non-proportionality is material depen- uting to an additional hardening (Shamsaei et al., 2010).
718 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720

a 0.75 a 500
SA333 C-Mn steel

0.50
Axial strain (ε), %

Equivalent stress, MPa

0.25 450

0.00

-0.25 400

-0.50 Axial strain (ε) 0.6%

Shear strain (γ) 0.78% Out of phase
Equivalent strain 0.75% In phase
-0.75 350
-1.00 -0.75 -0.50 -0.25 0.00 0.25 0.50 0.75 1.00 1 10 100 1000
Shear strain (γ), % Number of cycles

b 500 Axial stress-strain (σ-ε) b 480 Axial strain (ε) 0.4%

400 Shear stress-strain (τ-γ) 470 Shear strain (γ) 0.52%
300 460 Equivalent strain 0.5%

Equivalent stress, MPa

200 450
Stress, MPa

100 440
0 430
-100 420
-200 410
400
-300
390
-400
SA333 C-Mn steel 380 Out of phase
-500 In phase SA333 C-Mn steel
-1.0 -0.5 0.0 0.5 1.0 370
Strain, % 1 10 100 1000
Number of cycles
Fig. 17. Symmetric in phase strain cycling on SA333 C–Mn steel, 0.6%
axial strain and 0.78% shear strain: (a) applied strain; (b) stable stress– Fig. 18. Von-Mises equivalent stress alteration with cycles during in
strain response. phase and 90° out of phase symmetric strain cycling for SA333 C–Mn
steel: (a) 0.6% axial strain and 0.78% shear strain (Dc/De = 1.3); (b) 0.4%
axial strain and 0.52% shear strain (Dc/De = 1.3).

In this investigation both in-phase and out of phase

loading (Dc/De = 1.3) with a combination of 0.6% axial
of cycles to failure) is noted. It can therefore be said from
(De) and 0.78% shear (Dc) strain is applied on tubular spec-
Fig. 18 that SA333 C–Mn steel shows non-proportional
imens fabricated from SA333 C–Mn steel pipes. The ap-
hardening in 90° out of phase loading.
plied strain, stable stress response and stable stress–
strain response are plotted for 90o out of phase loading
in Fig. 16a–c. Those for the in-phase loading are given in 4. Conclusions
Fig. 17a–c. Von-Mises equivalent stress amplitude calcu-
lated from Eq. (4) and plotted as a function of number of Eight basic key issues in cyclic plastic deformation phe-
cycles is shown in Fig. 18a for both the in-phase and out nomena on isotropic materials should be included to de-
of phase conditions. Von-Mises equivalent stress ampli- scribe and model the cyclic plasticity i.e. stress–strain
tude calculated at every point for out of phase loading con- hysteresis loop. The phenomena are: Bauschinger effect,
dition shows maximum value in between maximum axial cyclic hardening/softening, strain range effect, loading his-
strain and shear strain. For this combination of tension- tory memory, ratcheting, mean stress dependent hardening,
torsion loading condition, the SA333 C–Mn steel shows mean stress relaxation and non-proportional hardening.
around 14.3% additional hardening (calculated at half of Based on the present investigation to explore key issues
number of cycles to failure) in cyclic plastic deformation on SA333 C–Mn steel and
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 304LN stainless steel, the following conclusions have been
req ¼ r2 þ 3s2 ð4Þ reached:

where, req, r and s are the Von-Mises equivalent stress, ax- (a) Cyclic hardening/softening is not only dependent on
ial stress and shear stress respectively. material but also on the loading condition and load-
Similarly, by maintaining Dc/De = 1.3, in-phase and 90o ing history. With progression of cycling, increase in
out of phase loading with a combination of 0.4% axial 0.52% linear region of hysteresis loops is responsible for
shear strain has been investigated on the same material. cyclic hardening. Pre- tensile straining or pre- cyclic
Variation of Von-Mises equivalent stress amplitude with straining with higher amplitude results cyclic soft-
number of cycles is shown in Fig. 18b and around 13.2% ening whereas as received material shows cyclic
non-proportional hardening (calculated at half of number hardening behavior.
 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720 719

(b) SA333 C–Mn steel and 304LN stainless steel shows Design, ESIS 21. Mechanical Engineering Publication, London, pp. 43–
60.
non Masing behavior (strain range effect). Increase
Eckert, R., Laird, C., Bassani, J., 1987. Mechanism of fracture produced by
in linear region of hysteresis loops (i.e. double of fatigue cycling with a positive mean stress in copper. Mater. Sci. Eng.
cyclic yield stress) with increase in strain amplitude 91, 81.
is the cause of non Masing behavior. Eifler, D., Macherauch, E., 1990. Microstructure and cyclic deformation
behaviour of plain carbon and low-alloyed steels. Int. J. Fatigue 12 (3),
(c) Peak stress in the direction of mean strain decreases 165–174.
with cycles whereas other peak stress not changes Ellyin, F., Kujawaski, D., 1984. Plastic strain energy in fatigue failure. J.
substantially. Mean stress is relaxed with cycles in Engng. Mater. Technol., Trans. ASME 106, 342–347.
Fang, D., Berkovits, A., 1994. Mean stress models for low-cycle fatigue of a
asymmetric strain cycling. nickel-base superalloy. Int. J. Fatigue 16 (6), 429–437.
(d) Pre-straining followed by LCF and mean stress relax- Fan, Z., Jiang, J., 2004. Investigation of low cycle fatigue behavior of
ation shows similar kind of material response. 16MnR steel at elevated temperature. Zhejiang Daxue Xuebao
(Gongxue Ban)/J. Zhejiang Univ. (Eng. Sci.) 38, 1190–1195.
(e) Ratcheting strain accumulation increases with Feltner, C.E., Laird, C., 1967a. Cyclic stress–strain response of FCC metals
increase in mean stress or stress amplitude in both and alloys.–.I. Phenomenological experiments. Acta Metall. 15, 1621–
engineering and true stress controlled experiments. 1632.
Feltner, C.E., Laird, C., 1967b. Cyclic stress–strain response of F.C.C. metals
(f) Decrease in both hysteresis loop area and plastic and alloys—II. Dislocation structures and mechanisms. Acta Metall. 15
strain amplitude with increasing mean stress, (10), 1633–1653.
mean-stress dependent hardening can be noticed. Floreen, S., Hayden, H.W., 1970. Some observations of void growth during
the tensile deformation of a high strength steel. Scr. Metall. 4 (2), 87–
(g) True stress controlled experiments show an increas-
94.
ing ratcheting life and ratcheting strain with Gaudin, C., Feaugas, X., 2004. Cyclic creep process in AISI 316L stainless
increasing mean stress. Mean stress dependent steel in terms of dislocation patterns and internal stresses. Acta
hardening can be thought to responsible for the Mater. 52, 3097–3110.
Golos, K., Ellyin, F., 1987. Generalization of cumulative damage criterion
improvement in ratcheting life. to multilevel cyclic loading. Theor. Appl. Fract. Mec. 7 (3), 169–176.
(h) True and engineering stress controlled tests show Giordana, M.F., Alvarez-Armas, I., Armas, A.F., 2010. Cyclic softening of
significant differences in ratcheting strain and rat- EUROFER 97 at room temperature mechanical and microstructural
behavior. Acta Microsc. 19 (1), 84–88.
cheting life. Large ratcheting life and small ratchet- Han, K.S., Margolin, H., 1989. Void formation, void growth and tensile
ing strain accumulation is observed in true stress fracture of plain carbon steel and a dual-phase steel. Mater. Sci. Eng. A
controlled test over engineering stress controlled 112, 133–141.
Hartwig, H., 1996. Basic conditions for material and structural ratcheting.
test in the presence of tensile mean stress where Nucl. Eng. Des. 162 (1), 55–65.
as reverse is true for compressive mean stress. Hassan, T., Corona, E., Kyriakides, S., 1992. Ratcheting in cyclic plasticity,
(i) SA333 C–Mn steel shows non proportional harden- Part II: multiaxial behavior. Int. J. Plasticity 8, 117.
Hassan, T., Kyriakides, S., 1992. Ratcheting in cyclic plasticity, Part I:
ing in out of phase loading. uniaxial behavior. Int. J. Plasticity 8, 91.
Hassan, T., Kyriakides, S., 1994a. Ratcheting of cyclically hardening and
softening materials, I: uniaxial behavior. Int. J. Plasticity 10, 149.
Hassan, T., Kyriakides, S., 1994b. Ratcheting of cyclically hardening and
Acknowledgements softening materials, II: multiaxial behavior. Int. J. Plasticity 10, 185.
Haupt, A., Schinke, B., 1996. Experiments on the ratcheting behaviour of
AISI 316L (N) austenitic steel at room temperature. ASME J. Eng.
The authors would like to thank The Director, National Mater. Technol. 118, 281.
Metallurgical Laboratory, Jamshedpur, India for providing Holste, C., Kleinert, W., Gurth, R., Mecke, K., 1994. Cyclic stress–strain
all the necessary facilities in carrying out this study. Authors response and strain localization effects under stress-control
conditions. Mater. Sci. Eng. A187, 113.
would also like to acknowledge the Bhabha Atomic Research Itoh, T., Sakane, M., Ohnami, M., Ameyama, K., 1992. Effect of staking fault
Centre, Mumbai, India, for supplying the material. energy on cyclic constitutive relation under non proportional loading.
J. Soc. Mater. Sci. Jpn. 41-468, 1361–1367.
Jiang, Y., Sehitoglu, H., 1994a. Cyclic ratcheting of 1070 steel under
References multiaxial stress states. Int. J. Plasticity 10, 579.
Jiang, Y., Sehitoglu, H., 1994b. Multiaxial cyclic ratcheting under multiple
Arcari, A., Vita, D.R., Dowling, E.N., 2009. Mean stress relaxation during step loading. Int. J. Plasticity 10, 849.
cyclic straining of high strength aluminum alloys. Int. J. Fatigue 31, Jiang, Y., Zhang, J., 2008. Benchmark experiments and characteristic cyclic
1742–1750. plasticity deformation. Int. J. Plasticity 24 (9), 1481.
Bauschinger, J., 1886. Mechanisch-Techanischen Laboratorium, Materials Kang, G., Gao, Q., 2002. Uniaxial and non-proportionally multiaxial
Forum. ratcheting of U71Mn rail steel: experiments and simulations. Mech.
Chai, H.F., Laird, C., 1987. Mechanisms of cyclic softening and cyclic creep Mater. 34 (12), 809.
in low carbon steel. Mater. Sci. Eng. 93, 159–174. Khan, A.S., Chen, X., Abdel-Karim, M., 2007. Cyclic multiaxial and shear
Christ, H.-J., Mughrabi, H., 1996. Cyclic stress–strain response and finite deformation response of OFHC: Part I, experimental results. Int.
microstructure under variable amplitude loading. Fatigue Fract. Eng. J. Plasticity 23 (8), 1285–1306.
Mater. Struct. 19 (2–3), 335–348. Kliman, V., Bily, M., 1980. The influence of mode control, mean value and
Das, A., Tarafder, S., 2009. Experimental investigation on martensitic frequency of loading on the cyclic stress–strain curve. Mater. Sci. Eng.
transformation and fracture morphologies of austenitic stainless 44, 73.
steel. Int. J. Plasticity 25 (11), 2222–2247. Krempl, E., Lu, H., 1984. The hardening and rate-dependent behavior of
Delobelle, P., Robinet, P., Bocher, L., 1995. Experimental study and fully annealed AISI Type 304 stainless steel under biaxial in-phase
phenomenological modelization of ratchet under uniaxial and and out-of-phase strain cycling at room temperature. ASME J. Eng.
biaxial loading on an austenitic stainless steel. Int. J. Plasticity 11, 295. Mater. Technol. 106, 376–382.
Doong, S.H., Socie, D.F., Robertson, I.M., 1990. Dislocation substructures Kulkarni, S.C., Desai, Y.M., Kant, T., Reddy, G.R., Parulekar, Y., Vaze, K.K.,
and nonproportional hardening. ASME J. Eng. Mater. Technol. 112, 2003. Uniaxial and biaxial ratchetting study of SA333 Gr.6 steel at
456–464. room temperature. Int. J. Press. Vessels Pip. 80, 179.
Doquet, V., Clavel, M., 1996. Stacking-fault energy and cyclic hardening of Laird, C., Finney, J.M., Schwartzmann, A., de la Veaux, R., 1975. History
FCC solid solutions under multiaxial nonproportional loadings. In: dependence in the cyclic stress–strain response of wavy slip
Pineau, A., Cailletaud, G., Lindley, T.C. (Eds.), Multiaxial Fatigue and materials. J. Test. Eval. 3, 435–441.
720 S.K. Paul et al. / Mechanics of Materials 43 (2011) 705–720

Li, Y., Laird, C., 1993. Masing behavior observed in monocrystalline copper Raman, S.G.S., Padmanabhan, K.A., 1996. Effect of prior cold work on the
during cyclic deformation. Mater. Sci. Eng. A 161 (1), 23–29. room-temperature low-cycle fatigue behavior of AISI 304LN stainless
Lorenzo, F., Laird, C., 1984. A new approach to predicting fatigue life steel. Int. J. Fatigue 18, 71–79.
behavior under the action of mean stresses. Mater. Sci. Eng. 62, Shamsaei, N., Fatemi, A., Socie, F.D., 2010. Multiaxial cyclic deformation
205. and non-proportional hardening employing discriminating load
Lukáš, P., Klesnil, M., 1973. Cyclic stress–strain response and fatigue life of paths. Int. J. Plasticity 26 (12), 1680–1701.
metals in low amplitude region. Mater. Sci. Eng. 11, 345–356. Sivaprasad, S., Paul, S.K., Das, A., Narasaiah, N., Tarafder, S., 2010a. Cyclic
Lukas, P., Kunz, L., 1989. Effect of mean stress on cyclic stress–strain plastic behaviour of primary heat transport piping materials:
response and high cycle fatigue life. Int. J. Fatigue 11, 55. Influence of loading schemes on hysteresis loop. Mater. Sci. Eng. A
Kunz, L., Lukás, P., Weiss, B., Melisova, D., 2001. Effect of loading history 527 (26), 6858–6869.
on cyclic stress–strain response. Mater. Sci. Eng. A 314 (1–2), 1–6. Sivaprasad, S., Paul, S.K., Gupta, S.K., Bhasin, V., Narasaiah, N., Tarafder, S.,
Magnin, T., Ramade, C., Lepinoux, J., Kubin, L.P., 1989. Low-cycle fatigue 2010b. Influence of uniaxial ratchetting on low cycle fatigue behviour
damage mechanisms of F.c.c. and B.c.c. polycrystals: homologous of SA 333 Gr. 6 C–Mn steel. Int. J. Press. Vessels Pip. 87 (8), 464–469.
behaviour? Mater. Sci. Eng. A 118, 41–51. Skelton, R.P., Maier, H.J., Christ, H.-J., 1997. The Bauschinger effect, Masing
Maier, H.J., Gabor, P., Gupta, N., Karaman, I., Haouaoui, M., 2006. Cyclic model and the Ramberg–Osgood relation for cyclic deformation in
stress–strain response of ultrafine grained copper. Int. J. Fatigue 28, metals. Mater. Sci. Eng. A 238 (2), 377–390.
243–250. Sung, J.H., Kim, J.H., Wagoner, R.H., 2010. A plastic constitutive equation
Masing, G., 1926. ibid., 1926.5, p. 135. incorporating strain, strain-rate, and temperature. Int. J. Plasticity 26
Mayama, T., Sasaki, K., Kuroda, M., 2008. Quantitative evaluations for (12), 1746–1771.
strain amplitude dependent organization of dislocation structures Taleb, L., Hauet, A., 2009. Multiscale experimental investigations about
due to cyclic plasticity in austenitic stainless steel 316L. Acta Mater. the cyclic behavior of the 304L SS. Int. J. Plasticity 25 (7), 1359.
56, 2735–2743. Turner, A.P.L., Martin, T.J., 1980. Cyclic creep of type 304 stainless steel
Morrow, J., Sinclair, G.M., 1958. Cycle-dependent stress relaxation. In: during unbalanced tension–compression loading at elevated
Symposium on Basic Mechanisms of Fatigue. American Society for temperature. Metall. Trans. A 11, 475.
Testing and Materials, Philadelphia, PA, pp. 83–109 (ASTM STP Verma, R.K., Kuwabara, T., Chung, K., Haldar, A., 2010. Experimental
237). evaluation and constitutive modeling of non-proportional
Mughrabi, H., Christ, H.J., 1997. Cyclic deformation and fatigue of selected deformation for asymmetric steels. Int. J. Plasticity (available online
ferritic and austenitic steels: specific aspects. ISIJ Int. 37, 1154. 9 April 2010).
Paul, S.K., Sivaprasad, S., Dhar, S., Tarafder, S., 2010a. Ratcheting and low Wang, Z., Laird, C., 1988. Cyclic stress—strain response of polycrystalline
cycle fatigue behavior of SA333 steel and their life prediction. Int. J. copper under fatigue conditions producing enhanced strain
Nucl. Mater. 401 (1–3), 17–24. localization. Mater. Sci. Eng. 100, 57–68.
Paul, S.K., Sivaprasad, S., Dhar, S., Tarafder, S., 2010b. True stress control Xia, Z., Kujawski, D., Ellyin, F., 1996. Effect of mean stress and ratcheting
asymmetric cyclic plastic behavior in SA333 C–Mn steel. Int. J. Press. strain on fatigue life of steel. Int. J. Fatigue 18 (5), 335.
Vessels Pip. 87 (8), 440–446. Ye, D., Matsuoka, S., Nagashima, N., Suzuki, N., 2006. The low-cycle
Paul, S.K., Sivaprasad, S., Dhar, S., Tarafder, S., 2010c. Cyclic plastic fatigue, deformation and final fracture behaviour of an austenitic
deformation and cyclic hardening/softening behavior in 304LN stainless steel. Mater. Sci. Eng. A 415 (1–2), 104–117.
stainless steel. Theor. Appl. Fract. Mech. 54 (1), 63–70. Zhang, J., Jiang, Y., 2004. A study of inhomogeneous plastic deformation of
Plumtree, A., Abdel-Raouf, H.A., 2001. Cyclic stress–strain response and 1045 steel. ASME J. Eng. Mater. Technol. 126, 164–171.
substructure. Int. J. Fatigue 23 (9), 799–805. Zhang, J., Jiang, Y., 2005. An experimental investigation on cyclic plastic
Pratt, J.E., 1967. Dislocation substructure in strain-cycled copper as deformation and substructures of polycrystalline copper. Int. J.
influenced by temperature. Acta Metall. 15 (2), 319–327. Plasticity 21, 2191.