Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/229401565

A notched cross weld tensile testing method for determining true


stress–strain curves for weldments

Article  in  Engineering Fracture Mechanics · February 2002


DOI: 10.1016/S0013-7944(01)00075-3

CITATIONS READS
75 704

4 authors, including:

Zhiliang Zhang Mons Hauge


Norwegian University of Science and Technology Equinor ASA
325 PUBLICATIONS   6,884 CITATIONS    38 PUBLICATIONS   681 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

mechanics of solar silicon wafer View project

Seminar on environmental assisted fracture in Trondheim, Norway, June 10th. Financed by the Research Council of Norway View project

All content following this page was uploaded by Zhiliang Zhang on 04 December 2020.

The user has requested enhancement of the downloaded file.


Engineering Fracture Mechanics 69 (2002) 353–366
www.elsevier.com/locate/engfracmech

A notched cross weld tensile testing method for determining


true stress–strain curves for weldments
Z.L. Zhang a,*, M. Hauge b, C. Thaulow a, J. Ødeg
ard a
a
SINTEF Materials Technology, Rich. Birkelands vei 1C, N-7465 Trondheim, Norway
b
Statoil Research Centre, N-7005 Trondheim, Norway
Received 31 January 2000; received in revised form 18 June 2001; accepted 21 June 2001

Abstract
Cross weld tensile testing is widely used in the industry to qualify welds. In these conventional testing fracture load is
measured and the location of fracture (weld metal, base metal or heat affected zone) is evaluated. Because the load-
elongation curve depends on the location of fracture and the initial gauge length, it cannot be utilized in the failure
assessment of weldments. Failure assessment of weldments requires input of true stress–strain behaviour for each
material zone. In this paper, a notched cross weld tensile testing method is proposed for determining the true stress–
strain curve for each material zone of a weldment. In the proposed method, cylindrical cross weld tensile specimens,
with a notch located either in the weld metal, base metal or possibly heat affected zone are applied. Due to the notch,
plastic deformation is forced to develop in the notched region. A load versus diameter reduction curve is recorded. It
has been shown that the true strain at maximum load is independent of the notch geometry. Furthermore, the materials
true stress–strain curve can be determined from the recorded load versus diameter reduction curve of a notched cross
weld tensile specimen by dividing a geometry-factor G, which is approximated by a quadratic function of the specimen
diameter to notch radius ratio and a linear function of the true strain at the maximum load. It is found that G is in-
dependent of the material zone length when the homogenous material length is larger or equal to the minimum di-
ameter. Ó 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Load-separation principal; True stress–strain curve; Weldment testing and mismatch effect

1. Introduction

For homogenous thin materials, a method has been recently developed to determine the long range true
stress–strain curve by using tensile specimens with rectangular cross section [1]. This paper presents a new
method for weldments. Weldments are inhomogenous in nature. The mechanical properties of base metal,
weld metal and heat affected zone (HAZ) are different to each other. Assessment of failure behaviour of
weldments requires input of full-range true stress–strain curve for each material zone. Testing of weld
thermal simulated specimens and all weld tensile testing can be carried out. However, cross weld tensile

*
Corresponding author. Tel.: +47-7359-2530; fax: +47-7359-2931.
E-mail address: zhiliang.zhang@matek.sintef.no (Z.L. Zhang).

0013-7944/02/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 1 3 - 7 9 4 4 ( 0 1 ) 0 0 0 7 5 - 3
354 Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366

Nomenclature

r flow stress
rn nominal stress, P =A0
rz true tensile stress, P =A
e true tensile strain
P tensile load
Pmax maximum tensile load
A current cross section area
A0 initial cross section area
D0 initial diameter of the minimum notched cross section
H material zone length
R0 initial notch radius
G geometry factor
r0 yield stress
rPmax true stress at the maximum load
ePmax true strain at the maximum load
rB Bridgman corrected true stress

testing of weldments is cheap and attractive. Cross weld tensile testing has been widely used in the industries
to qualify welds, i.e. whether the failure occurs in the base metal that represents a good scenario, or in the
weld metal or HAZ that is often not wanted. In cross weld tensile testing, load versus elongation in the axial
direction can be recorded (Fig. 1). For obvious reasons, the load versus elongation curve cannot be utilized
in the assessment of failure behaviour of weldments. The curve is strongly dependent on the initial mea-
suring gauge length, the sampling of materials, and the location of necking and final fracture.
In practice the actual strain in different material zones can be measured during cross weld tensile testing,
by attaching strain gauges to the materials or using optical measurements. The application of these methods
was very limited because of the nature of these measurements. The surface strains and deformation cannot
be directly transferred to materials full range stress–strain curve.
The limitations of the current weldment testing methods necessitate the development of alternative
testing method. In this paper, a new method called notched cross weld tensile testing has been proposed to
determine materials true stress–strain curve for each material zone. In this method tensile specimens with a

Fig. 1. (a) Conventional cross weld tensile testing, and (b) typical outputs for initial different gauge length.
Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366 355

Fig. 2. Proposed notched cross weld tensile specimen.

round notch located in the middle of a material zone are tested in tension. Fig. 2 schematically shows the
notched cross weld tensile testing specimens proposed in the paper. Load versus diameter reduction is
recorded for each test. It has been shown by an extensive numerical study that smooth and notched tensile
specimens reach the maximum load at the same true strain. And materials true stress–strain curve can be
obtained based on the so-called load separation principle, by normalizing the load diameter reduction curve
by a geometry factor. The effect of notch geometry, material length as well as hardening behaviour on the
accuracy of the method has been investigated.

2. The load separation principle

In fracture mechanics, there is a so-called load separation principle [2] which states that the load–
deformation behaviour of a fracture mechanics specimen can be represented by a multiplication of two
parts. The first part, g, is a function of geometry, while the second one, h, is a function of plastic dis-
placement which represents the hardening behaviour of a material,
P ¼ gða=wÞhðvp =wÞ; ð1Þ
where P is the load, a is the crack length, w is the specimen width and vp is the plastic displacement, for
example, the load line plastic displacement. The load separation principle is the theoretical basis for ASTM
standard test method for J1c and J–R curve from single specimens [3,4]. It should be noted that Eq. (1) most
often is used before the maximum load or possibly after the maximum load for the cases where the
maximum load is caused by crack growth rather than by diffused necking.
The load separation principle is very interesting in that the material function h once obtained from a
specific geometry can be transferred to any other geometry, because the material is the same.
Recent studies [5] have shown that with modifications the load separation principle works satisfactory
for many examples, and the plastic displacement function can be transferred from one geometry to another.
Moreover, it has been suggested that plastic displacement function can be linked to notched cylindrical
tensile specimens [6,7].
356 Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366

The idea of this paper is based on the load separation principle. The authors are generally interested in
obtaining both the tensile properties and fracture toughness for weldments by using alternative and simple
testing methods. This paper is a first step towards an alternative testing method. It will explore the pos-
sibility of determining the true stress–strain curve for different material zones of a weldment by using a
round notched cross weld tensile specimen. When applying the load separation principle, several critical
questions should be asked, namely, how the notch geometry influences the geometry function g, whether
the geometry function is dependent on materials hardening behaviour, and what is the effect of material
zone length. These questions have been answered in the following sections.

3. Numerical procedure

A series of notched specimens have been analysed by A B A Q U S for materials with various hardening
ability and zone length. Fig. 3 shows the representative finite element meshes used for the specimens. The
minimum diameter for the tensile specimens is 6 mm. Eight-node axi-symmetric elements were used. The
total number of elements varied from 530 to 1115.

Fig. 3. Meshes for the (a) smooth specimen, (b) notched specimens.
Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366 357

The elastic properties for the materials considered in the paper are taken as E=r0 ¼ 500 and m ¼ 0:3. A
power-hardening law material with finite strain theory was applied to the analyses. For the material,
!n
ep
r ¼ r0 1 þ ; ð2Þ
e0

where r is the flow stress, ep is the equivalent plastic strain, r0 is the yield stress, e0 is the yield strain
e0 ¼ r0 =E, and n is the strain hardening exponent. Materials hardening considered in the paper varies from
n ¼ 0:05–0.2.

4. Instability of notched tensile specimens

The strain at maximum load is an instability parameter of the tensile specimen. In this section we will
show theoretically that this parameter is independent of the constraint. Thus, the strain at maximum load
should be the same for different notch configurations. The numerical results are used to verify this hy-
pothesis.
Consider two unit cells, one with uniaxial loading and one with triaxial loading, Fig. 4. The first one
(Fig. 4(a)) represents the smooth tensile specimen where the transversal stress is zero, and the latter (Fig.
4(b)) represents the notched tensile specimen where the transversal stress is proportional to the axial stress.
The instability for Fig. 4(a) occurs when the following condition is satisfied [8]
dP ¼ dð rz AÞ ¼ A drz þ rz dA ¼ 0; ð3Þ
where P is the tensile load and A is the current cross section area of the unit cell. From Eq. (3) we obtain
drz dA
¼ : ð4Þ
rz A
In the case of uniaxial tension, the left side of Eq. (4) becomes d r, where r is the flow stress. The right
r=
side can be approximated by the equivalent strain increment de, such that the instability condition becomes:

Fig. 4. Unit cells illustrating the conditions at the maximum load. c is a constant less than 1.0.
358 Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366

d
r
¼ r: ð5Þ
de
For notched tensile specimens (Fig. 4(b)), Eq. (4) is still valid. By neglecting the elastic component of the
strain, the right side of Eq. (4) is also in this case equal to the equivalent strain increment de. However, rz is
not equal to the flow stress, but rz ¼ r=ð1  cÞ. Nevertheless, by substituting this relation to Eq. (4), we can
again obtain Eq. (5).
For smooth specimen and power law hardening material, diffuse necking will follow the maximum load
and the strain at the maximum load is equal to the hardening exponent [8]
ePmax ¼ n: ð6Þ
Eq. (6) and the above explanations show that the strain at maximum load is an unique material pa-
rameter, independent of the geometrical constraint and equal to the strain hardening exponent if a power
law hardening material is considered.
In the numerical study, specimens with a material with n ¼ 0:1 have been studied first. This material has
also been used as a reference material in the next stage where the effect of material hardening behaviour on
the accuracy of the method is considered.
Fig. 5(a) shows the normalized gross stress versus true strain curves for three notch configurations. The
figure confirms that maximum load occurs at the same strain level for all the geometries and that this strain
is equal to the strain hardening exponent. However, it appears also that the exact value of this strain can be
difficult to determine, especially for the smooth specimen. It must be noted that Eq. (5) is only valid ap-
proximately, because the elastic strain has been neglected. So does Eq. (6).

5. True stress–strain curve from smooth and notched tensile specimens – homogenous material

Fig. 5(b) shows the true stress–strain curves for three cylindrical tensile specimens with different notch
configurations. The true stress is calculated by dividing the current load by the area of the current minimum
notch cross section. Difference between the gross stress and the true stress becomes significant only after the
diffuse necking has occurred. It should be reminded that Fig. 5(a) represents the behaviour of the same
material but different geometry. The conventional way of transferring material data is to determine the flow
stress–strain relation from the true stress–strain curve of smooth specimens by the Bridgman correction [9].
Then, the flow stress can be transferred to the prediction of the notched specimens. Here in this paper, we
invert this concept. Fig. 5(b) shows that the slope of the true stress–strain curves for notched specimens are
very similar to those for the smooth specimens. A normalization procedure has been applied to the true
stress–strain curves shown in Fig. 5(b), i.e. all the curves are normalized by their true stress at maximum
load. The results of the normalization are presented in Fig. 5(c). It is interesting to observe from Fig. 5(c)
that, to a large degree, the three curves are collapsed into one over a quite large range of strain. It should be
noted that the strain at the maximum load for the material used in Fig. 5 is 0.1. The normalized curves are
also plotted in Fig. 5(d), where the true strain has been normalized by the true strain at the maximum load.
In line with the load separation principle discussed in Section 2, the true stress–strain curve for a notched
tensile specimen (Fig. 5) can be expressed by a multiplication of the true stress–strain curve of a smooth
specimen and a geometry factor G:
rNotched
z ðeÞ ¼ GrSmooth
z ðeÞ: ð7Þ
From the above equation, materials true stress–strain curve represented by the smooth specimen can be
obtained from the true stress–strain curve of a notched specimen by

rSmooth
z ðeÞ ¼ rNotched
z ðeÞ=G: ð8Þ
Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366 359

Fig. 5. Results for tensile specimens with n ¼ 0:1: (a) normalized gross stress versus true strain, (b) normalized true stress versus true
strain, (c) true stress normalized by the stress at maximum load versus true strain, (d) normalized true stress versus strain normalized
by the strain at maximum load.

Eq. (8) is schematically shown in Fig. 6(a) and (b). It must be noted that the geometry correction is simply
applied to the stress, rather than the strain.
The general observation of the normalized curves is that the difference between the smooth specimen and
the notched specimens before the maximum load is very small. After the maximum load, however, the true
strain is slightly amplified for the notched tensile specimens, compared with the smooth specimen. This
effect comes from the strain concentration at the notch position. Fig. 5 shows that the true stress–strain
curve estimated from the notched specimens is conservative.
By carrying out a number of analyses of notched specimens for material with hardening n ¼ 0:1, we
obtain the following approximate relation for the geometry factor G (n ¼ 0:1),
   2
D0 D0
G ¼ 1:007 þ 0:18777  0:01313 ; ð9Þ
R0 R0
where D0 and R0 are the initial diameter and notch radius. Eq. (9) is plotted in Fig. 6(c).
360 Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366

Fig. 6. A schematic plot of the proposed method. The true stress curve from a notched specimen (a), to the true stress–strain curve of
the material (smooth specimen) (b,c) the geometry factor G versus the initial diameter/notch radius ratio.

It should be reminded that the true stress–strain curve is not the flow stress–strain relation used in finite
element analyses. A Bridgman correction should be applied to the true stress–strain curve obtained from a
notched tensile specimen [9]:
rB ¼ rSmooth
z =Bðe  ePmax Þ; ð10Þ
where B is the Bridgman correction factor which is a function of the net strain after necking, e  ePmax .
Material’s flow stress-equivalent plastic strain curve can be obtained from the rB  e curve by separating
the plastic strain from the true strain.

6. Effect of material hardening exponent

Eqs. (8) and (9) are based on the material with hardening exponent n ¼ 0:1. It is important to know how
the geometry function G in Eq. (8), is influenced by the material hardening behaviour. Materials with
different hardening exponents have been studied with the smooth and notched specimens. Fig. 7(a) shows
the normalization for a material with n ¼ 0:2. It can be seen that the normalization works very well for this
Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366 361

Fig. 7. True stress–strain curves for n ¼ 0:2. The true stress–strain curve for the three specimens (a), and the normalized curves (b).

material before the maximum load, and even better than the material with n ¼ 0:1. Similar results can be
seen in Fig. 8 for a low hardening material with n ¼ 0:05. The general conclusion is that the normalization
is better for high hardening materials than for low hardening materials.
The effect of hardening on the geometry factor has been studied. The difference, ðGn  Gn¼0:1 Þ=Gn¼0:1
versus hardening exponent is shown in Fig. 9(a), where the material with n ¼ 0:1 has been taken as the
reference material. It seems that the hardening will affect the geometry function. The error for the strong
hardening material with n ¼ 0:2 can be up to about 8% when the geometry factor from the reference
material with n ¼ 0:1 is used. It is obvious that the error of the geometry is dependent on the sharpness of
the notch. For the geometry with D0 =R0 ¼ 3:0, the error is much smaller, less than 4%. It should be
mentioned that the reason to apply a notch in the tensile specimens is to force the plastic deformation
localized in the material zone of interest. It is not necessary to use sharp notch geometry. Results from this

Fig. 8. True stress–strain curves for n ¼ 0:05. The true stress–strain curve for the three specimens (a), and the normalized curves (b).
362 Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366

Fig. 9. (a) Relative difference, ðGn  Gn¼0:1 Þ=Gn¼0:1 as a function of hardening exponent ePmax , (b) Gn =Gn¼0:1 versus ePmax .

study shows that when the notch radius to diameter ratio D0 =R0 is larger than 3.0, quite good accuracy in
engineering sense will be achieved.
Fig. 9(b) plots the Gn =Gn¼0:1 ratio versus different hardening for the notch geometry D0 =R0 ¼ 3. A linear
equation can well fit the Gn =Gn¼0:1 hardening relation:
Gn
¼ 1:053  0:53ePmax : ð11Þ
Gn¼0:1
By combining Eqs. (9) and (11) we obtain the following approximate equation for calculating the geometry
factor G,
"    2 #
D0 D0
G ¼ 1:007 þ 0:18777  0:01313 ð1:053  0:53ePmax Þ: ð12Þ
R0 R0

7. True stress–strain curve from notched cross weld tension – effect of material zone length

So far the focus has been the specimens with homogenous material. This section will focus on weldments.
The conventional cross weld tensile testing uses specimens with rectangular cross section. In the notched
cross weld tensile testing, specimens with round cross section will be used. A bi-material specimen (weld
metal and base metal) is considered. For smooth bi-material specimen, necking will occur either in the weld
metal or in the base metal, depending on the mismatch of yield stress, hardening ability and boundary
conditions. When a notch is applied to the centre of a material zone (Fig. 2) plastic deformation will be
localized in the notched zone, no matter undermatch or overmatch the notched zone material is.
The effect of the material zone length on the true stress–strain curve of a notched specimen is investigated
for a bi-material case where the material in the notched region has lower stress–strain curve than the
surrounding material. In particular, the required length of the notched zone to behave as a homogenous
specimen is evaluated.
Five notched specimens with different material zone length have been analysed. Fig. 10(a) shows the
gross stress versus true strain curves for different material zone length for the case with D0 =R0 ¼ 3:0,
Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366 363

Fig. 10. Effect of material zone length on the stress–strain curve of a bi-material specimen with D0 =R0 ¼ 3:0. Gross stress versus strain
(a), and true stress versus strain (b). The material system is characterized as n ¼ 0:1 and m ¼ 0:5.

hardening exponent n ¼ 0:1, and material mismatch ratio m ¼ 0:5. The m is defined as the ratio of weld
metal yield strength (notch material) to the base metal yield strength. The general observation is that short
material zone will increase the gross stress for a given true strain. The effect of material zone length is
similar to the effect of notch radius, short material zone length corresponds to small notch radius. Material
zone length enforces a material plasticity constraint to the specimen, while the notch representing a ge-
ometry constraint. Both the geometry and material constraint enhance the hydrostatic stress in the notch
region. As discussed in Section 4 above, the hydrostatic stress component does not influence the true strain
at the maximum load. Fig. 10(a) shows that the true strain at the maximum load is independent of the
material zone length H.
The effect of the material zone length on the true stress–strain curve is shown in Fig. 10(b). It can be
observed that when the zone length is close to or larger than the diameter at the notch cross section, the bi-
material specimens behave like a homogenous specimen, and the effect of zone length on the resulting true
stress–strain curve becomes insignificant. Similar observation has been observed for cases with different
ratios of specimen diameter/notch radius, see Fig. 11 for D0 =R0 ¼ 7:5.
This finding is in accordance with the observation of Toyoda [10]. Toyoda found that for a bi-material
specimen, the homogenous behaviour becomes dominant when the zone length is larger than the minimum
diameter.
The effect of material mismatch on the true stress–strain curve has also been investigated. When the ratio
of the initial notch radius to the initial minimum diameter is larger than 1.0, the absolute value of mismatch
has almost no effect on the true stress–strain curve. Example is shown in Fig. 12.

8. Discussion and concluding remarks

It has been found in this study that true stress–strain curves for a given material from both smooth and
notched tensile specimens can be normalized by a respective geometry factor, such that the normalized
curves collapse into one. The geometry factor has been approximated as a quadratic function of notch
geometry parameter D0 =R0 , and a linear function of the true strain at the maximum load. It should be noted
that this normalization covers a wide range of strain after the maximum load has appeared. This finding
gives a basis for an alternative testing method for determining material true stress–strain curve by using the
364 Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366

Fig. 11. Effect of material zone length on the true stress–strain curve of a bi-material specimen with D0 =R0 ¼ 7:5. The material system is
characterized as n ¼ 0:1 and m ¼ 0:5.

Fig. 12. Effect of mismatch on the true stress–strain curve for D0 =R0 ¼ 3:0, and H=D0 ¼ 1:67 and n ¼ 0:1.

notched tensile specimen. Notched tensile specimens have the advantage to force the plastic deformation
localised in the region of interest, such that one can always test the material intends to test.
One interesting finding from this study is that the strain at maximum load is rather independent of the
notch geometry and material zone length. Instability analysis shows that the strain at maximum load is a
very good indicator of material hardening behaviour. In many cases, not the whole stress–strain curve but
the hardening is of interest. Thus the simple notched cross weld tensile testing can give very accurate
hardening parameter.
In this study only bi-material systems have been considered. The method can be in principle applied to
multi-material system, for example to the testing of HAZ material in weldments, as long as the HAZ
mechanical properties are uniform over a certain length. The diameter can be adjusted to satisfy the
condition that the material zone length is equal to or large than the diameter. In the interpretation of HAZ
testing results, care should be taken how representative the sampled HAZ is. HAZ usually has a strong
gradient from the fusion line to the base metal. The typical length of steel HAZ is about 2 mm, while the
Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366 365

Fig. 13. True stress–strain curves for n ¼ 0:1 material at the beginning of loading.

aluminium HAZ length can be as large as 10 mm. Direct testing of tensile specimens with notch positioned
in the HAZ may give representative properties.
Finally, it should be mentioned that results in this study show that the load separation is successful in a
global sense for tensile specimens. Fig. 13 shows the true stress–strain curve for smooth and notched tensile
specimens with material n ¼ 0:1 within the strain range 1%. It can be seen that the yielding pattern for the
smooth one and the notched ones are very different. Smooth specimen has a very sharp transition from
elastic behaviour to plastic yielding because the material has a uniform stress state. Once one material point
starts yielding, the whole specimen is yielding. However, the transition to plastic yielding is rather smooth
for the notched specimens. Only part of the material will yield in the beginning and other part is still elastic.
This makes the transition smooth. Strictly speaking, the three curves within 1% of strain cannot be nor-
malized. If the strain gradients are different in different tensile specimens, the load separation principle is in
a strict sense not applicable. Therefore, the load separation principle is generally valid only in an ap-
proximate sense.

Acknowledgements

The financial support from the Norwegian Research Council through the Strategic Institute Programme
at SINTEF Materials Technology is greatly appreciated.

References

[1] Zhang ZL, Hauge M, Ødeg ard J, Thaulow C. Determining material true stress–strain curve from tensile specimens with
rectangular cross section. Int J Solids Struct 1999;36:3497–516.
[2] Ernst HA, Paris PC, Rossow M, Hutchinson JW. In: Smith CW, editor. Analysis of load-displacement relations to determine J –R
curves and tearing instability material properties. Fracture Mechanics ASTM STP 677. 1981. p. 581–99.
[3] ASTM E813 Standard test method for J1c , a measure of fracture toughness, ASTM annual book of standards v 03.01, 1987.
[4] ASTM E1152 Standard test method for determining J –R curve, ASTM annual book of standards v 03.01, 1987.
[5] Cruz JR, Landes JD. A common format approach for the ductile fracture method using a displacement-based normalization
parameter. Fatigue Fract Engng Mater Struct 1997;20:1305–18.
366 Z.L. Zhang et al. / Engineering Fracture Mechanics 69 (2002) 353–366

[6] Donoso JR, Labbe F. A calibration function for notched cylindrical tension specimens, based on the common format equation:
numerical and experimental data analysis. Engng Fract Mech 1996;54:617–28.
[7] Donoso JR, Landes JD. The common format equation approach for developing calibration functions for two-dimensional
fracture specimens from tensile data. Engng Fract Mech 1996;54:499–512.
[8] McClintok FA, Argon AS. Mechanical behaviour of materials. USA: Addison-Wesley Publishing Company; 1966.
[9] Bridgman PW. Studies in large plastic flow and fracture. New York: McGraw-Hill; 1952.
[10] Toyoda M. Plastic constraint controlling factors and their effect. 2nd Workshop on Constraint Effects on Structural Performance
of Welded Joints, Osaka, Japan, 1994.

View publication stats

You might also like