Nuclear Engineering and Technology 53 (2021) 647e656

Contents lists available at ScienceDirect

Nuclear Engineering and Technology

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

Original Article

Determination of true stress-strain curve of type 304 and 316 stainless

steels using a typical tensile test and finite element analysis
Hyeong Do Kweon a, c, Jin Weon Kim b, Ohseop Song c, Dongho Oh c, *
a
Korea Hydro & Nuclear Power Co, Ltd, 70, 1312-beongil, Yuseong-daero, Yuseong-gu, Daejeon, 34101, Republic of Korea
b
Department of Nuclear Engineering, Chosun University, 309 Pilmun-daero, Dong-gu, Gwangju, 61452, Republic of Korea
c
Department of Mechanical Engineering, Chungnam National University, 99 Daehak-ro, Yuseong-gu, Daejeon, 34134, Republic of Korea

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

Article history: Knowing a material’s true stress-strain curve is essential for performing a nonlinear finite element
Received 29 April 2020 analysis to solve an elastoplastic problem. This study presents a simple methodology to determine the
Received in revised form true stress-strain curve of type 304 and 316 austenitic stainless steels in the full range of strain from a
9 July 2020
typical tensile test. Before necking, the true stress and strain values are directly converted from engi-
Accepted 12 July 2020
Available online 30 July 2020
neering stress and strain data, respectively. After necking, a true stress-strain equation is determined by
iteratively conducting finite element analysis using three pieces of information at the necking and the
fracture points. The Hockett-Sherby equation is proposed as an optimal stress-strain model in a non-
Keywords:
Finite element analysis
uniform deformation region. The application to the stainless steel under different temperatures and
Necking loading conditions verifies that the strain hardening behavior of the material is adequately described by
Plastic behavior the determined equation, and the estimated engineering stress-strain curves are in good agreement with
Stainless steel those of experiments. The presented method is intrinsically simple to use and reduces iterations because
Tensile test it does not require much experimental effort and adopts the approach of determining the stress-strain
True stress-strain curve equation instead of correcting the individual stress at each strain point.
© 2020 Korean Nuclear Society, Published by Elsevier Korea LLC. This is an open access article under the
CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction large earthquake loads [2,3]. Containments for radioactive material

Nuclear power plant structures and components are designed so
that the working stress under various operational loading condi-
tions does not exceed the yield or ultimate tensile strength of
materials in general. In the engineering evaluation to ensure the
structural integrity of structures and components, therefore, an
elastic analysis is performed assuming that stresses are linearly
proportional to strains. This is an excessively conservative method
that does not take into account strain hardening even though
plastic deformation occurs significantly in actual materials under
some large loading conditions.
The safety of nuclear power plants against large earthquakes
exceeding design basis has been of paramount concern since the
Fukushima nuclear power plant accident in Japan [1]. Several
studies related to cyclic elastoplastic analysis have been under-
taken in the nuclear industry to realistically evaluate structural
integrities considering the plastic deformation of materials due to
* Corresponding author.
E-mail address: dongho@cnu.ac.kr (D. Oh).
transportation have been conventionally designed based on stress
acceptance criteria. However, the stress-based acceptance criteria
are generally too conservative for energy-limited events, for
example, accidental drop or impact. New design acceptance criteria
based on strain have been proposed, and elastoplastic analysis is
therefore needed to obtain the working strain values under large
deformation conditions [4].
Poisson’s ratio and elastic modulus are the only material prop-
erties required for an elastic analysis; in contrast, it is essential to
have uniaxial true stress-strain information in addition to Poisson’s
ratio and elastic modulus to solve elastoplastic problems where the
relationship between stress and strain is not linear. A stress-strain
curve as the inherent material characteristics is commonly deter-
mined by a tensile test. Generally, the stress and strain means the
true stress and strain, respectively, unless notated specifically as an
engineering stress and strain. The stress-strain relationship in a
plastic zone obtained from tensile testing is classified into two
parts. The neck section of a specimen is uniformly deformed before
necking occurs at maximum loading, but non-uniformly deformed
after necking occurs. According to the literature to date, there have

https://doi.org/10.1016/j.net.2020.07.014
1738-5733/© 2020 Korean Nuclear Society, Published by Elsevier Korea LLC. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/
licenses/by-nc-nd/4.0/).
648 H.D. Kweon et al. / Nuclear Engineering and Technology 53 (2021) 647e656

Nomenclature εf True strain at fracture

s0 Derivative of true stress with respect to true strain
S Engineering stress Ai Engineering stresses estimated from a finite element
e Engineering strain analysis
s True stress Ei Engineering stresses measured from an experiment
ε True strain c2 Chi-square
A Currently measured cross-sectional area Af Cross-sectional area at fracture
P Currently measured load A0 Original cross-sectional area
se Corrected true stress equivalent to the uniaxial value RA Reduction in area at fracture
B Bridgman correction factor scorr Amount of stress to be corrected
su True stress at necking Pfea Load estimated from a finite element analysis
εu True strain at necking Pexpt Load measured from an experiment
sef Corrected true stress at fracture

been a number of attempts to describe the stress-strain relation
and the strain hardening behavior of various metals in the region of
uniform and/or non-uniform deformation. The true stress-strain
curve may be obtained by various means, such as directly con-
verting from engineering stress-strain data, conducting special
tests, performing finite element simulations, or combining these
methods. These methods are dependent on the characteristics of
materials and the range of strain to be considered.
The true stress-strain relationship in the region of uniform
deformation may be identified through typical tensile testing. Such
testing involves a method of simply converting the engineering
stress-strain values, where necking does not yet occur, to the true
stress-strain ones, and then fitting them using an appropriate
equation [5e7]. To identify the stress-strain relation in a non-
uniform deformation region as well as in a uniform deformation
region, there are methods of fitting the true stress-strain data in the
uniform deformation region and then extending the fitted curve to
the non-uniform deformation region [8e12]. In References [5e12],
the stress-strain models for fitting adopt a power law form of
equations such as Hollomon’s [13], Swift’s [14], and Ludwik’s [15],
an exponential form of equations such as Voce’s [16] and Hockett-
Sherby’s [17], or a combination of them.
A direct method for obtaining the true stress-strain data after
necking is to continuously measure the shape of the test specimen
[17e20]. Hockett et al. proposed a more generalized equation by
modifying Voce’s equation, which describes the true stress-strain
relation in a large strain region through compression tests on
alpha uranium, pure iron, and 1100 aluminum [17,18]. Using a
digital image correlation technique, the real-time measurement
and analysis of the deformation shape of a specimen have been
applied to obtain the true stress-strain data in the full range of
strain [19,20]. In order to accurately identify the plastic flow char-
acteristics after necking, there have been methods of acquiring the
true stress-strain data through the consecutive measurement of the
specimen deformation shape and iteratively correcting the true
stress-strain data by performing finite element analysis [21e24].
These are methods to directly acquire the true stress-strain data in
the full range of strain by measuring the specimen deformation
shape. In particular, more reliable results are expected by per-
forming a finite element analysis in conjunction with measure-
ment. There are, however, difficulties in accurately measuring the
deformation shape of a specimen because of the need for specialist
equipment.
Many methods using a finite element analysis without special
measurements after necking have been introduced to obtain the
true stress-strain data over the entire strain range [25e32]. After
assuming the initial true stress-strain curve from existing equa-
tions, studies have iteratively corrected the true stress-strain data
in the non-uniform deformation region using the difference of the
engineering stress values at each strain point between the experi-
ment and the finite element analysis. However, the number of it-
erations may be intrinsically high due to the one-by-one
corrections at each strain point, and the stress-strain equation of
the final true stress-strain curve after completing corrections was
not presented for corresponding material.
Austenitic stainless steel such as type 304 or 316 is widely used
in nuclear power plants because of its strong corrosion resistance,
excellent welding performance, high strength, and other advanta-
geous characteristics. Although some research has been published
regarding the stress-strain relationship of stainless steel, that
research only deals with the uniform deformation [7,33,34], or the
mathematical models presented in relation to the non-uniform
deformation are not confirmed by experimental results [35,36].
The method correcting point-by-point the true stress-strain data in
the range of non-uniform deformation by the use of finite element
analysis has been introduced for stainless steel [24,32]. However,
the operating temperatures and strain rates in nuclear power
plants where stainless steel is used were not taken into account.
This paper presents a simple method to determine the true
stress-strain curve of type 304 and 316 austenitic stainless steels in
the full range of strain without continuous measurement of the
specimen deformation shape in typical tensile testing. The tensile
stress flow and strain hardening behavior of materials are investi-
gated in the post-necking region. A stress-strain model that reflects
the deformation behavior of type 304 and 316 stainless steels is
proposed, and the equations pertaining to the model are finally
determined by comparing the engineering stress-strain curve
predicted from finite element analysis with that of experiment, and
iteratively correcting it.
2. Experiments
The materials to be tested are the type 304 and 316 austenitic
stainless steels in accordance with the material requirements of the
ASME B&PV Code Section II [37], which are broadly used for safety
class components such as piping, reactor vessel internals, and
containments for radioactive material transportation in nuclear
Table 1
Chemical compositions of type 304 and 316 stainless steels.
Material Composition (wt%)
C Si Mn P S Cr Ni Mo
Type 304 SS 0.018 0.31 1.76 0.025 0.0002 18.43 10.24 e
Type 316 SS 0.009 0.38 1.22 0.008 e 16.20 11.20 2.08
 H.D. Kweon et al. / Nuclear Engineering and Technology 53 (2021) 647e656
power plants. Their chemical compositions are listed in Table 1.
The tensile test specimens illustrated in Fig. 1 were machined in
the shape of round bars with a uniform length of 32 mm, a gauge
length of 25 mm, and a diameter of 6 mm for quasi-static testing
and 5 mm for dynamic testing, respectively, in accordance with the
testing requirements of the ASTM E8 standard [38].
Each test was carried out at a room temperature of 20  C and an
operating temperature of reactor coolant system of 316  C,
respectively, in an air environment. The load-line displacement
speed was 1 mm/min assuming the quasi-static loading condition
for the type 304 and 316 stainless steels and 2280 mm/min
assuming the dynamic loading condition for the type 316 stainless
steel.
The engineering stress values were calculated by dividing the
measured load by the original cross-sectional area of the specimen,
and the engineering strains were calculated by dividing the
measured displacement by gauge length. In addition, the yield and
ultimate tensile strengths and the uniform and total elongations
were determined based on the same data collected from testing.
The specimen cross section diameter at fracture was measured to
determine the reduction in area. This is a common and simple
tensile test conducted to obtain the fundamental properties of
materials. The tensile properties resulting from the test are sum-
marized in Table 2, and their engineering stress-stress curves are
shown in Fig. 2.
3. Methodology
3.1. Background
For the prediction of the flow stress of a specimen subjected to
tensile loading, there are two distinct parts of strain hardening
behavior, the before and after necking stages. The first part includes
the data before the onset of the necking, whereas the second part
involves the continuation of the stress-strain curve from the
necking to the fracture. Since a specimen is uniformly deformed as
a whole before necking, it is placed in a uniaxial stress state. After
the necking onset, however, a specimen central region is non-
uniformly deformed and placed in a multiaxial stress state. The
stress and strain are no longer uniformly distributed along the
specimen at high levels of elongation due to the necking formation;
therefore, the engineering stress-strain curve cannot provide
proper information of the physical phenomena of the specimen
involved in the test. In general, it is difficult to accurately describe
both stages with one stress-strain model on account of the inher-
ently different deformation behavior.
Before necking, the true stress-strain values are readily con-
verted from the engineering stress-engineering values obtained
from a typical tensile test using Eqs. (1) and (2), respectively.
Accordingly, the pre-necking stress-strain model does not have to
be found in the form of an equation in order to be used for a finite
element analysis in practical engineering problems.
Fig. 1. Test specimen shape and dimensions.
649
s ¼ Sð1 þ eÞ (1)
ε ¼ lnð1 þ eÞ (2)
where S and e are the engineering stress and strain, respectively,
and s and ε are the true stress and strain, respectively.
In contrast, the true stress-strain values of the necking zone in
the post-necking stage cannot be determined directly based on the
engineering stress and strain due to the stress triaxiality in the
necked area. In this case, the instantaneous true stress and strain
values are calculated in accordance with Eqs. (3) and (4), respec-
tively, using the currently measured load and cross-sectional area
of the necked zone.
P
s¼ (3)
A


A0
ε ¼ ln (4)
A
where P is the currently measured load, and A0 and A denote the
original and currently measured cross-sectional area of specimen,
respectively.
As large amounts of deformation beyond the onset of necking
results in the generation of transverse stresses in the neck region,
Bridgman [39] proposed the correction factor of Eq. (5) in order to
modify the increased axial stress and make it equivalent to the
uniaxial stress. Although the correction factor is the value that is
determined by measuring the radius of curvature and the mini-
mum radius at the specimen neck, they are extremely difficult
variables to be implemented in practice. Alternatively, Bridgman
presented the empirical curve for the correction factor as a function
of true strain at the neck exclusively for a cylindrical specimen.
Rather than using the correction factor curve, Dowling [40] pro-
vided Eq. (6), closely estimating it.
se ¼ Bs (5)
B ¼ 0:0684ðlog10 εÞ3 þ 0:0461ðlog10 εÞ2  0:205ðlog10 εÞ
þ 0:825 ð0:12  ε  3Þ (6)
where se is the corrected true stress equivalent to the uniaxial value
at the neck of specimen, and B denotes the Bridgman correction
factor as a function of true strain only.
As Bridgman’s experiments were carried out at room tempera-
ture under high pressure in quasi-static loading conditions, there are
inherent errors due to the difference between Bridgman’s and the
present testing conditions. Additionally, Eq. (6) was approximated
from empirical data for various kinds of metals; thus, it implicitly
includes some degree of error due to data scattering. For these rea-
sons, when Eqs. (5) and (6) are applied for correction of stress in a
large strain region, additional corrections based on the results of
finite element analysis are iteratively carried out, if necessary.
When the continuous measurement of the curvature and radius
of the necked part by the use of special technique such as digital
image correlation is available, the true stress-strain values may be
determined either directly without finite element analysis [19,20]
or interactively with finite element analysis [21e24]. However,
when the only available data obtained from the typical tensile
testing are the load and displacement values (i.e., engineering
stresses and strains), and the necked sectional diameter at fracture
(i.e., reduction in area), the true stress-strain values in the post-
necking phase should be determined through other means.
650 H.D. Kweon et al. / Nuclear Engineering and Technology 53 (2021) 647e656

Table 2
Tensile properties resulting from testing.

Material Testing Condition Yield Strength (MPa) Tensile Strength (MPa) Uniform Elongation (%) Total Elongation (%) Reduction in Area (%)

Temperature ( C) Speed (mm/min)

Type 304 SS 20 1 226.3 579.5 66.50 85.80 86.39

Type 304 SS 316 1 137.9 410.0 31.80 43.70 77.10
Type 316 SS 20 1 313.6 590.4 54.70 71.40 83.18
Type 316 SS 316 1 148.8 429.4 34.10 46.80 81.29
Type 316 SS 20 2280 363.5 637.0 35.83 54.94 81.89
Type 316 SS 316 2280 183.0 448.6 29.81 45.59 82.26

the non-uniform deformation region for the purpose of this paper.

In the uniform deformation region, the stress-strain values can be
calculated using Eqs. (1) and (2), as described earlier. The true
stress-strain values need to be smoothly connected without a
sudden change at maximum loading, namely the necking point
where the uniform and non-uniform deform regions join. For this
purpose, Considere’s instability criterion of Eq. (12) [41] should be
obeyed at the maximum loading point.

Fig. 2. Engineering stress-strain curves.
3.2. Stress-strain model of type 304 and 316 stainless steels
Several models between stress and strain have been widely used
to describe plastic stress flow and strain hardening behavior. The
most commonly used equations of the models, for instance, Hol-
lomon, Ludwik, and Swift, are the power law forms of Eqs. (7)e(9),
respectively. They intrinsically imply that the stress becomes
infinite when the strain is infinite. On the other hand, some stress-
strain models, such as those of Voce and Hockett-Sherby, show the
exponential forms of Eqs (10) and (11), respectively. These equa-
tions, which have been developed with a fundamentally different
approach, have a distinct feature with asymptotic stress value after
severe deformation.
ds
¼s (12)
dε
In addition to Conside re’s criterion, the true stress (su) and
strain (εu) at the necking point are converted from the engineering
stress and strain, respectively, according to Eqs (1) and (2), and the
true stress (sf) and strain (εf) at the fracture point are calculated
using the value of the reduction in area according to Eqs (3) and (4),
respectively. The equivalent true stress (se) to the uniaxial value is
obtained by applying the Bridgman correction of Eqs. (5) and (6).
Consequently, the true stress-strain values from the necking to the
fracture points may be obtained by determining an appropriate
stress-strain equation using three pieces of information, that is,
dsj ¼ s’ ðεu Þ ¼ su , sjε¼εu ¼ sðεu Þ ¼ su , sjε¼εf ¼ sðεf Þ ¼ sef , where
dε ε¼εu
su and εu are the true stress and strain at the necking point,
respectively, and sef and εf denote the corrected true stress and the
true strain at the fracture point, respectively.
Kweon et al. have proposed Eq. (13) as the optimum stress-
strain model in the non-uniform deform region for SA-508 low-
alloy steel, which is applied to determine the stress-strain values by
the use of three pieces of information obtained from a typical
tensile test [42]. The equation incorporates linear and constant
terms into the commonly used Hollomon model which is shown in
Eq. (7), and retains the nature of the power law form of equation.
s ¼ Hεn þ aε þ b (13)

s ¼ Hε n
(7) where n is the strain hardening exponent of Eq. (7) at the maximum
load, and H, a and b are the coefficients dependent on the nature of
s ¼ Hðε þ ε0 Þn (8) the material.
Fig. 3 shows the results of the application of Eq. (13) to the type
304 stainless steel at room temperature and quasi-static loading
s ¼ s0 þ Hεn (9) condition. It is confirmed that the engineering stress-strain values
of the analysis are significantly different from those of the experi-
s ¼ ss  sr ðemε Þ (10) ment at the high-strain region.
As shown in Fig. 4, the true stress and strain values of the type
  304 stainless steel produced by Eq. (13) are normalized between
s ¼ ss  sr emε
n
(11) the necking and the fracture points to check the sensitivity of the
curve to the strain hardening exponent, n. Even though the expo-
where H, n, ε0, s0, ss, sr, m and n are all parameters to be deter- nent is flexibly adjusted in Fig. 4, the strain hardening of the curve is
mined depending on the nature of the material. insensitive to the change of the exponent, that is, all curves seem to
Although Eqs. 7e11 have been developed to be applied to the be almost overlapping. Moreover, the hardening rates of the curves
full range of plastic stress flow regardless of uniform or non- hardly change with the increase of strain.
uniform deformation, they are intended to be used as the stress- Meanwhile, the Voce and Hockett-Sherby models, which are
strain model to determine the true stress-strain values only in written in Eqs. (10) and (11), respectively, have the inherent
 H.D. Kweon et al. / Nuclear Engineering and Technology 53 (2021) 647e656
Fig. 3. Stress-strain curves resulting from the modified Hollomon equation.
Fig. 4. Normalized true stress-strain curve of Eq. (13) according to exponent, n.
characteristic of being saturated with asymptotic stress, ss, as the
strain increases. Eq. (10) has the exponent fixed at 1 (one), while Eq.
(11) is the generalized form that can flexibly adjust the exponent
according to the nature of material. In Fig. 5, the normalized curves
in the post-necking region obtained from Eq. (11) illustrate that the
strain hardening obviously changes with the variation of exponent
and also the hardening rate decreases gradually with the increase
of strain. As the exponent value increases, the curve becomes more
asymptotical; conversely, as it decreases, the asymptotic property
becomes weaker, resulting similar to a power law form of curve.
This means that the Hockett-Sherby model is more flexible than
other models, and thus can express various types of curves. In
addition, because the stress-strain relations of the 304 and 316
stainless steels are various and dependent on the temperature and
strain rate, as seen in Fig. 2, the Hockett-Sherby model is chosen for
describing the post-necking stress-strain behavior in this study.
3.3. Method for determining true stress-strain curve
The true stress-strain curves of the type 304 and 316 stainless
steels are determined by iteratively performing a finite element
analysis based on typical tensile testing data, as per the following
process.
651
Fig. 5. Normalized true stress-strain curve of Eq. (11) according to exponent, n.
Step 1: A typical tensile test is conducted to acquire fundamental
data such as the engineering stresses and strains, the yield and
ultimate tensile strengths, the uniform and total elongations, and
the reduction in area. Among these data, the engineering stress-
strain values and the reduction in area at fracture are required to
estimate the true stress-strain values.
Step 2: The true stress and strain values up to the necking point
in the uniform deformation region are directly converted from the
engineering stress-strain data obtained in Step 1 using Eqs. (1) and
(2), respectively.
Step 3: The true stress and stain values at fracture are calculated
from Eqs. (3) and (4), respectively, in which the cross-sectional area
at fracture is inversely estimated from Eq. (14). The fracture stress is
corrected by the Bridgman factor of Eqs. (5) and (6).
Af ¼ A0 ð1  RAÞ (14)
where A0 and Af are the cross-sectional areas of specimen at orig-
inal condition and fracture, respectively, and RA is the reduction in
area at fracture.
Step 4: The exponent, n, of Eq. (11) is initially set to about 2.0 and
0.5 for the room and operating temperature conditions, respec-
tively. The initial value of the exponent may be flexibly set
depending on the test material and conditions. Three unknown
parameters of Eq. (11), that is, m, ss, and sr, are determined by
solving three equations, using three pieces of information at both
necking and fracture, namely s’ ðεu Þ ¼ su, sðεu Þ ¼ su and sðεf Þ ¼ sef .
Thus, the true stress-strain values between the necking and the
fracture points are obtained from Eq. (11).
Step 5: A finite element analysis is undertaken to simulate the
tensile testing by inputting the obtained true stress-strain data for
the full range of strain as the plastic property of the material cor-
responding to the specimen. The finite element analysis results in
obtaining the engineering stresses and strains.
Step 6: The estimated and measured engineering stress-strain
curves are compared in shape to confirm the adequacy of the
exponent. The n-value is increased in the case where more strain
hardening is needed, or decreased when less is needed, otherwise,
it is maintained. If the exponent is modified, the true stress-strain
values between the necking and the fracture points are updated
using the changed equation and the finite element analysis is
performed again.
652 H.D. Kweon et al. / Nuclear Engineering and Technology 53 (2021) 647e656

conducted for the quasi-static and dynamic tests, respectively. The

nonlinear geometry option was employed for the analysis proced-
ures to take account of the large inelastic deformation during the
tensile test. Displacement boundary conditions were applied on the
upper end of the specimen to simulate the tensile testing.
Fig. 6 represents the finite element model of the specimen used
for the analysis. Only one-eighth of the specimen is modeled due to
geometrical symmetries to reduce the model size and computing
time. The model is composed of a structured mesh arrangement
with C3D8 (first-order integral and three-dimensional solid) ele-
ments, strategically placing finer meshes along critical regions
where plastic localization is expected to occur. The total number of
elements and nodes of the model are 17720 and 19949,
respectively.

4.2. Results and discussion

The true stress-strain curves of type 304 and 316 stainless steels
under the test conditions of Table 2 were obtained in accordance
with the methodology described in Section 3. In the uniform
deformation region, the true stress and strain values are simply
converted from the engineering stress and strain values, respec-
tively, and in the non-uniform deformation region, they were taken
Fig. 6. Finite element model of specimen.

Step 7: The engineering stresses resulting from the analysis are

compared with those from experiment at each engineering strain. If
the difference between the two engineering stress-strain curves is
not satisfactory, the equivalent true stress at fracture, sef , is cor-
rected depending on the degree of difference. Taking account of the
measuring uncertainty of the specimen cross section diameter at
fracture and the steepness of the engineering stress-strain curve
near fracture, it is recommended that the amount of correction be
not more than the value calculated according to Eq. (15), proposed
by Joun et al. [27]. If the fracture stress is modified, the true stress-
strain values between the necking and the fracture points are
updated using the changed equation and the finite element analysis
is performed again.

Pexp  Pfea
scorr ¼ s  (15)
Pfea

where scorr and s are the correcting amount of the true stress and
the present true stress, respectively, and Pexp and Pfea are the loads
measured from experiment and estimated from analysis,
respectively.
Repeat Step 6 and/or Step 7 until the difference between the
estimated and the measured engineering stress-strain curves is
satisfactory. In the present work, the iterative processes are
finished if the chi-square values are less than five with minimum
three times repetition. However, the criteria for terminating the
process should differ from case-by-case depending on the appli-
cation field and purpose of the work.

4. Application

4.1. Finite element analysis

A nonlinear 3D finite element analysis was performed on the

specimen inputting the determined true stress and strain data into
the plastic material property. The commercially available finite Fig. 7. (a) True stress-strain curves, and (b) corresponding engineering stress-strain
element analysis software, Abaqus/Standard ™ [43], was used for curves with iterations of type 304 stainless steel at room temperature under quasi-
this study. The static and dynamic procedures of analysis were static loading.
 H.D. Kweon et al. / Nuclear Engineering and Technology 53 (2021) 647e656
Fig. 8. (a) True stress-strain curves, and (b) corresponding engineering stress-strain
curves with iterations of type 304 stainless steel at operating temperature (316  C)
under quasi-static loading.
from the determined stress-strain equation. In order to demon-
strate the suitability of the obtained true stress-strain curve, the
estimated engineering stress-strain curve was compared with that
of experiment.
Figs. 7e12 illustrate the true and engineering stress-strain
curves with iterations for a total of six cases of tests, which
include the tested material and condition, the finally determined
re is taken
equation, and so on. As the instability criterion of Conside
into account at the necking point, the true stress-strain curves are
smoothly connected without severe change at the joint of the
uniform and non-uniform deformation regions. This is important
from the view point of the actual use of the true stress-strain curve
for a finite element analysis because the sudden variation of the
curve frequently causes problem in convergence of equilibrium due
to the instability of load at necking. The goodness of the curve
estimated from the analysis is judged by the chi-square value, that
is, the sum of square of the deviations of the estimated values from
the experimental values. The chi-square values have no physical or
absolute meaning in this study. They have only relative meanings
among the estimated curves indicating their quality to the
measured ones. The chi-square values are calculated by Eq. (16).
653
Fig. 9. (a) True stress-strain curves, and (b) corresponding engineering stress-strain
curves with iterations of type 316 stainless steel at room temperature under quasi-
static loading.
X ðA  E Þ2
c2 ¼ i i
(16)
Ei
where Ai and Ei are the engineering stresses from the analysis and
experiment, respectively, at the equally spaced fifty points of en-
gineering strains between the necking and the fracture points.
In all cases, it is certain without any further examination that the
engineering stress-strain values in the uniform deformation region
are almost in perfect agreement with the test results regardless of
iterations. This represents that the true stress-strain up to the
necking point can be accurately determined from the typical tensile
test results.
On the other hand, the difference in two curves in the non-
uniform deformation region is strongly dependent on the path of
flow stress between the necking and the fracture points. The
implied uncertainty of the equivalent stress at fracture, which is
resulted from both the Bridgman correction and measurement of
the cross-sectional diameter of the fractured specimen, may
contribute to the deviation between the analysis and the experi-
ment. As illustrated in Figs. 7e12, the analysis results are generally
approaching the experimental results as the iterations continue.
654 H.D. Kweon et al. / Nuclear Engineering and Technology 53 (2021) 647e656
Fig. 10. (a) True stress-strain curves, and (b) corresponding engineering stress-strain
curves with iterations of type 316 stainless steel at operating temperature (316  C)
under quasi-static loading.
Therefore, iterative analysis is inevitable due to the limited infor-
mation available at the necking and the fracture points. The
exponent, n, of the equations is determined to be 1.7e1.9 at the
room temperature (20  C), but it becomes smaller at the operating
temperature (316  C), resulting in 0.2e1.3. This means that even if
the strain hardening rate in the region near the necking is bigger at
the room temperature than at the operating temperature due to the
high strength and elongation, it decreases more rapidly with the
increase of strain. The load-line displacement speed, that is, strain
rate, does not appear to have a meaningful effect on the determi-
nation of the true stress-strain curve. The parameters of Eq. (11)
corresponding to the determined post-necking stress-strain
curves are summarized in Table 3.
Based on the chi-square values and the whole shapes of curves,
the estimated engineering stress-strain curves are in very good
agreement with those of experiments. It is verified that the post-
necking true stress-strain curve, which is basically needed for a
large plastic deformation analysis, is determined with a relatively
high level of goodness according to the methodology described in
Section 3.
Fig. 11. (a) True stress-strain curves, and (b) corresponding engineering stress-strain
curves with iterations of type 304 stainless steel at room temperature under dy-
namic loading.
5. Conclusion
This study presents a simple methodology to determine the true
stress-strain curve of type 304 and 316 austenitic stainless steels in
the full range of strain from a typical tensile test. The methodology
is summarized as follows:
 Before necking, the true stress and strain values are directly
converted from the engineering stress and strain data, respec-
tively, obtained from testing;
 After necking, the true stress-strain equation is determined by
iteratively conducting finite element analysis using three pieces
of information at the necking and the fracture points obtained
from testing.
The Hockett-Sherby equation is proposed as the optimal stress-
strain model in the non-uniform deformation region. The applica-
tion to the stainless steel under different temperatures and loading
conditions verifies that the strain hardening behavior of the
 H.D. Kweon et al. / Nuclear Engineering and Technology 53 (2021) 647e656 655

simple to use and reduces iterations because it does not require

much experimental effort and adopts the approach of determining
the stress-strain equation instead of correcting the individual stress
at each strain point. This requires less time and cost, which is
practically important to the industry.

Declaration of competing interest

The authors declare that they have no known competing

financial interests or personal relationships that could have
appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by research fund of Chungnam Na-

tional University, Republic of Korea.

Appendix A. Supplementary data

Supplementary data to this article can be found online at

https://doi.org/10.1016/j.net.2020.07.014.

References

[1] G. Saji, Safety goals for seismic and tsunami risks: lessons learned from the
Fukushima Daiichi disaster, Nucl. Eng. Des. 280 (2014) 243e249.
[2] I. Nakamura, N. Kasahara, Excitation tests on elbow pipe specimens to
investigate failure behavior under excessive seismic loads, in: ASME 2015
Pressure Vessel and Piping Conference, July 19-23, 2015. Boston, USA.
[3] P. Sollogoub, The OECD-NEA programme on metallic component margins
under high seismic loads (MECOS): towards new criteria, in: ASME 2017
Pressure Vessel and Piping Conference, July 16-20, 2017. Waikoloa, USA.
[4] S.D. Snow, D.K. Morton, Strain-based acceptance criteria for energy-limited
events, in: ASME 2009 Pressure Vessel and Piping Conference, July 26-30,
2009. Prague, Czech Republic.
[5] C. Mondal, B. Podder, K.R. Kumar, D.R. Yadav, Constitutive description of
tensile flow behavior of cold flow-formed AFNOR 15CDV6 steel at different
deformation levels, J. Mater. Eng. Perform. 23 (2014) 3586e3599.
[6] G. Sainath, B.K. Choudhary, J. Christopher, E.I. Samuel, M.D. Mathew, Appli-
cability of Voce equation for tensile flow and work hardening behaviour of
P92 ferritic steel, Int. J. Pres. Ves. Pip. 132 (2015) 1e9.
[7] L. Li, S. Liu, B. Ye, S. Hu, Z. Zhou, Quantitative analysis of strength and plasticity
of a 304 stainless steel based on the stress-strain curve, Met. Mater. Int. 22
Fig. 12. (a) True stress-strain curves, and (b) corresponding engineering stress-strain (2016) 391e396.
curves with iterations of type 316 stainless steel at operating temperature (316  C) [8] J. Choung, C. Shim, H. Song, Estimation of failure strain of EH36 high strength
under dynamic loading. marine structural steel using average stress triaxiality, Mar. Struct. 29 (2012)
1e21.
[9] R. Schneider, R.J. Grant, N. Sotirov, G. Falkinger, F. Grabner, C. Reichl,
M. Scheerer, B. Heine, Z. Zouaoui, Constitutive flow curve approximation of
material is adequately described by the determined equation, and commercial aluminium alloys at low temperatures, Mater. Des. 88 (2015)
the engineering stress-strain curves estimated from analysis have 659e666.
[10] J. Agirre, L. Galdos, E.S. Argandona, J. Mendiguren, Hardening prediction of
high levels of agreement with those from experiments. diverse materials using the digital image correlation technique, Mech. Mater.
For experimentally determining the true stress and strain 124 (2018) 71e79.
values, it is necessary to measure the evolving curvature radius and [11] Q. Pham, B. Lee, K. Park, Y. Kim, Influence of the post-necking prediction of
hardening law on the theoretical forming limit curve of aluminium sheets, Int.
minimum radius of the necking profile at different stages of a
J. Mech. Sci. 140 (2018) 521e536.
tensile test. This is difficult to implement in practice because of the [12] K.B. Othmen, N. Haddar, A. Jegat, P. Manach, K. Elleuch, Ductile fracture of AISI
need for specialist equipment and the significant amount of 304L stainless steel sheet in stretching, Int. J. Mech. Sci. 172 (2020) 105404.
[13] J.H. Hollomon, Tensile deformation, Transactions of AIME 162 (1945)
experimental work. However, the presented method is intrinsically
268e290.
[14] H.W. Swift, Plastic instability under plane stress, J. Mech. Phys. Solid. 1 (1952)
1e18.
Table 3 [15] P. Ludwik, Elements der Technologischen Mechanik, Verlag Von Julius
Parameters of Eq. (11) corresponding to the determined post-necking stress-strain Springer, Leipzig, 1909, p. 32.
curves. [16] E. Voce, The relationship between stress and strain from homogeneous
deformation, J. Inst. Met. 74 (1948) 537e562.
Material Testing Condition ss sr m n [17] J.E. Hockett, O.D. Sherby, Large strain deformation of polycrystalline metals at
low homologous temperatures, J. Mech. Phys. Solid. 23 (1975) 87e98.
Temperature ( C) Speed (mm/min)
[18] P.E. Armstrong, J.E. Hockett, O.D. Sherby, Large strain multidirectional defor-
Type 304 SS 20 1 2009 1377 0.8712 1.7 mation of 1100 aluminum at 300 K, J. Mech. Phys. Solid. 30 (1982) 27e58.
Type 304 SS 316 1 741.1 885.8 2.826 0.5 [19] S.K. Paul, S. Roy, S. Sivaprasad, H.N. Bar, S. Tarafder, Identification of post-
Type 316 SS 20 1 1617 948.3 1.441 1.9 necking tensile stress-strain behavior of steel sheet: an experimental inves-
Type 316 SS 316 1 954.2 3538 2.857 0.2 tigation using digital image correlation technique, J. Mater. Eng. Perform. 27
Type 316 SS 20 2280 1690 996.8 1.412 1.7 (2018) 5736e5743.
[20] C. Annan, E. Beaumont, Low cycle fatigue of stainless steel plates under large
Type 316 SS 316 2280 1104 653 1.284 1.3
plastic strain demands, Journal of Building Engineering 29 (2020) 101160.
656 H.D. Kweon et al. / Nuclear Engineering and Technology 53 (2021) 647e656
[21] E.E. Cabezas, D.J. Celentano, Experimental and numerical analysis of the ten-
sile test using sheet specimens, Finite Elem. Anal. Des. 40 (2004) 555e575.
[22] M. Kamaya, M. Kawakubo, A procedure for determining the true stress-strain
curve over a large range of strains using digital image correlation and finite
element analysis, Mech. Mater. 43 (2011) 243e253.
[23] L. Wang, W. Tong, Identification of post-necking strain hardening behavior of
thin sheet metals from image-based surface strain data in uniaxial tension
tests, Int. J. Solid Struct. 75 (2015) 12e31.
[24] M. Kamaya, Y. Kitsunai, M. Koshiishi, True stress-strain curve acquisition for
irradiated stainless steel including the range exceeding necking strain, J. Nucl.
Mater. 465 (2015) 316e325.
[25] Y. Ling, Uniaxial true stress-strain after necking, AMP Journal of Technology 5
(1996) 37e48.
[26] G. Mirone, A new model for the elastoplastic characterization and the stress-
strain determination on the necking section of a tensile specimen, Int. J. Solid
Struct. 41 (2004) 3545e3564.
[27] M. Joun, J.G. Eom, M.C. Lee, A new method for acquiring true stress-strain
curves over a large range of strains using a tensile test and finite element
method, Mech. Mater. 40 (2008) 586e593.
[28] K. Zhao, L. Wang, Y. Chang, J. Yan, Identification of post-necking stressestrain
curve for sheet metals by inverse method, Mech. Mater. 92 (2016) 107e118.
[29] F. Sebek, P. Kubik, J. Hulka, J. Petruska, Strain Hardening Exponent Role in
Phenomenological Ductile Fracture Criteria, vol. 57, 2016, pp. 149e164.
[30] M. Saboori, H. Champliaud, J. Gholipour, A. Gakwaya, J. Savoie, P. Wanjara,
Extension of flow stress-strain curves of aerospace alloys after necking, Int. J.
Adv. Manuf. Technol. 83 (2016) 313e323.
[31] M. Paredes, D.F.B. Sarzosa, R. Savioli, T. Wierzbicki, D.Y. Jeong, D.C. Tyrell,
Ductile tearing analysis of TC128 tank car steel under model I loading con-
dition, Theor. Appl. Fract. Mech. 96 (2018) 658e675.
[32] M. Paredes, V. Grolleau, T. Wierzbicki, On ductile fracture of 316L stainless
steels at room and cryogenic temperature level, Materialia 10 (2020) 100624.
[33] K.J.R. Rasmussen, Full-range stress-strain curves for stainless steel alloys,
J. Constr. Steel Res. 59 (2003) 47e61.
[34] I. Arrayago, E. Real, L. Gardner, Description of stress-strain curves for stainless
steel alloys, Mater. Des. 87 (2015) 540e552.
[35] K. Abdella, Inversion of a full-range stress-strain relation for stainless steel
alloys, Int. J. Non Lin. Mech. 41 (2006) 456e463.
[36] W.M. Quach, J.G. Teng, K.F. Chung, Three-stage full-range stress-strain model
for stainless steels, J. Struct. Eng. 134 (2008) 1518e1527.
[37] ASME, Materials, ASME Boiler and Pressure Vessel Code Section II, 2015.
[38] ASTM, Standard Test Methods for Tension Testing of Metallic Materials, ASTM
E8/E8M-09, 2009.
[39] P.W. Bridgman, Studies in Large Plastic Flow and Fracture, Harvard University
Press, USA, 1964.
[40] N.E. Dowling, Mechanical Behavior of Materials, fourth ed., Pearson Education,
England, 2012.
[41] A. Considere, Ann. Ponts Chaussees 9 (1885) 574e605.
[42] H.D. Kweon, E.J. Heo, D.W. Lee, J.W. Kim, A methodology for determining the
true stress-strain curve of SA-508 low alloy steel from a tensile test with finite
element analysis, J. Mech. Sci. Technol. 32 (2018) 3137e3143.
[43] ABAQUS Analysis User’s Guide, Version 2019, Dassault Systems, 2019.