Accepted Manuscript Not Copyedited: Assessment of Fracture Toughness Using Small Punch Tests of Pre Notched Specimens
Accepted Manuscript Not Copyedited: Assessment of Fracture Toughness Using Small Punch Tests of Pre Notched Specimens
Accepted Manuscript Not Copyedited: Assessment of Fracture Toughness Using Small Punch Tests of Pre Notched Specimens
Nomenclature:
The stresits
ε The total strain
0 The yield stress
n The hardening parameter for the nonlinear term
The yield offset
d
Pmax The maximum load
ite
ed
Abbreviation List:
py
SPT Small Punch Test
GTN Gurson-Tvergaard-Needleman
Co
3D 3 dimension
VVF Void Volum fraction
ot
tN
Assessment of Fracture toughness using small
punch tests of pre‐notched specimens
rip
Yangyan Zheng 1,2, Xiao Chen 1,2, Zheng Yang 1, Xiang Ling 1*
sc
Abstract: In this paper line and ring‐notched Small Punch Test (SPT) specimens were studied,
a 3D model of a ring‐notched SPT specimen was established using the contour integral
method, and the validity of the model was verified using ring‐notched specimens. The stress
ed
and strain fields were analyzed using numerical simulations of a ring‐notched SPT specimen,
and the change in the stress gradient during deformation was considered. To verify the finite
pt
element model, the results of the numerical simulations were compared with those of
three‐point bending tests and a Gurson‐Tvergaard‐Needleman (GTN) model. Compared with
ce
the line‐notched specimen, the ring‐notched specimen was more suitable for notch
propagation analysis and fracture toughness evaluation. The results of the numerical
Ac
simulations were in good agreement with those of the experiments, which showed that the
numerical model used in this study was correct. For a notch that initiated when the load
reached its maximum value, the value of the J integral was 335×10‐6 kJ/mm2, and at time
0.85Pmax, the value of the J integral was 201×10‐6 kJ/mm2, and the difference from the result
of the three‐point bending test was 14.4%. For a notch that initiated during the stretching
deformation stage, the relevant fracture toughness was 225×10‐6 kJ/mm2, and the difference
from the result of the three‐point bending test was 3%.
Key Words: Small Punch Test, Fracture Toughness, Pre‐Notched, Integral Contour Method
1 Introduction
Many petrochemical, aviation, aerospace and nuclear power plant components serve in
high‐temperature, high‐pressure, and corrosive environments [1]. Due to their long‐term
operation under harsh conditions, some of the containers, pipes and other components are
subject to material deterioration or even failure [2,3], which has disastrous consequences and
causes significant economic losses. Therefore, to avoid various failure phenomena, it is
necessary to quantify the performance of these materials. The traditional testing methods are
non‐destructive testing and conventional sampling, but they have their own limitations. The
former is rapid and non‐destructive, but the information obtained is neither comprehensive
d
nor accurate. Data from the latter are accurate and reliable, but due to the cumbersome
ite
sampling technique and the sample size, are large enough to significantly damage the
components in service [4]. This makes small test technology (such as small tensile test
specimens and small fatigue test pieces) relevant. The development of SPT technology is
ed
particularly noticeable [5‐8] because SPT samples are tiny (the diameters of circular
specimens and the side lengths of square specimens are between 3 mm and 10 mm, and such
py
specimens are between 0.25 mm and 0.50 mm thick), as shown in Fig. 1. Therefore, the
advantages of SPT technology make it possible to determine the material properties of
Co
in‐service components by sampling.
Ф3‐10mm
ot
tN
ip
0.25‐0.5mm
r
sc
Fig.1 The specimen of SPT
At present, several scholars have attempted to use SPTs to evaluate fracture toughness;
nu
their studies have essentially been based on the idea of empirical fitting. Mao and Takahashi
[9,10] first attempted to determine the fracture toughness of a material using an SPT, which
provided an empirical equation of for the equivalent fracture strain, JIc, or the fracture stress,
Ma
KIc. In recent studies, a few authors have introduced notched SPT specimens with different
shapes and dimensions. Ju [11] and Lacalle [12] used different modifications of notches
through the thicknesses of square plate specimens, and Cuesta [13] proposed V‐notched
ed
samples. Turba [14] introduced a circular notch in a disk, which resulted in an axially
symmetric stress state that led to the development of a near‐plane strain condition. The basis
pt
of the empirical fit is the relationship among the fracture strain, the energy absorbance and
the fracture toughness of the SPT specimen, which leads to different best‐fit equations for
ce
different materials; the best‐fit equations for the same materials developed at different
research institutions are also different [15,16]. The use of empirical fitting is largely influenced
Ac
by the type of specimen, the clamp size and the type of test machine [17].
In this study, SPTs of specimens with two different types of notch were conducted. And
different notch shape was investigated to make sure which notch shape is more suitable for
evaluating the fracture toughness of SPT specimen. A finite element model of a pre‐notched
SPT specimen was established, and the fracture toughness of the material was evaluated
using the contour integral method. Void volume fraction is computed by means of the GNT
model. In order to validate the methodology proposed in this paper, the fracture toughness
obtained from pre‐notched SPT specimen were compared with conventional three point
bending tests and GTN numerical results. The capabilities of the notched SPT to characterize
fracture toughness was proposed.
2 SPT Test
The basic principle underlying the SPT, as shown in Fig. 2, is that a thin sheet specimen is
punched by a ball with a constant speed. The central displacement of specimen are read
throughout the process, from elastic plastic deformation to fracture failure. A series of
material data points, such as the strength, plasticity, and fracture toughness, are obtained [18].
The test material used was Q345R, which had good comprehensive mechanical
properties and performed well during the process. The mechanical properties test was
d
performed according to stand of GB/T 228-2010. The mechanical properties of the Q345R
ite
used in this paper are listed in Table 1. True stress and strain curve of the Q345R is shown in
Fig. 3, and the chemical compositions of the Q345R are shown in Table 2. Pre‐notched
specimens containing the two types of notch shown in Fig. 3 were obtained by electrical
ed
discharge machining. Figure 4 shows a notched section with 40°angled notch tips and
0.5mm notched depth. Because the damage of the sample is initiated from the lower surface,
py
so the surface with notch was put downward. Three annular notch and one straight notch
tests were conducted in this paper.
Co
Table 1 Mechanical properties of Q345R at room temperature
Material E(MPa) s(MPa) b (MPa) Elongation (%)
Q345R 211862 0.3 329.5 529 28.75
ot
Table 2 Chemical composition of Q345R (wt%)
tN
Element C S P Mn Si
Content <0.2 0.0014 0.01 1.2456 0.202
ip
Φ 2.5
Punch
r
sc
Up Die
nu
Φ 2.4
Ma
Disc Specimen
ed
Fig. 2 Schematic diagram of SPT device.
ce
Ac
700
600
500
Stress(MPa)
400
300
d
100
ite
0
ed
0.00 0.05 0.10 0.15 0.20 0.25 0.30
Strain(%)
py
Fig. 3 True stress and strain curve of Q345R
Co
ot
tN
Φ10
r ip
sc
nu
Φ2.4
Ma
0.5
40°
ed
Fig.4 Two differently notched SPT specimens
pt
3 SPT results and discussion
Normal small punch, annular notch and straight notch tests were conducted, and the
ce
morphologies of the fracture surfaces are shown in Fig. 5, Fig. 6 and Fig. 7. The initial notches
in the specimens containing normal and annular notch were the opposite of a cup‐shaped
Ac
protrusion, as shown in the red circles in Fig. 5 and Fig. 6. The surface of the normal specimen
developed micro‐cracks during the loading process, and the pre‐notched specimens failed
along the pre‐notched tips and formed cup‐shaped protrusions, as shown in Fig. 6.
Figure 6 and Figure 7 shows that the notch along the line–notch appeared to develop an
opening mode fracture due to the stress concentration. The load‐displacement curves
obtained for the three different samples are shown in Fig. 8. They show that the results for the
ring‐notched specimens and normal specimens were similar and exhibited four typical stages.
At the same time, the strength decreased, and the maximum load and the maximum
displacement decreased significantly due to the pre‐notching. After the line‐notched
specimen entered the plastic stage, its carrying capacity decreased significantly due to the
stress concentration on the line notch. The line‐notched specimen was not axially symmetric,
which made it difficult for a neck to form. The dimple of the fracture was smaller, which
shows that a large number of voids did not coalesce. The line‐notched specimen was not
suitable for notch propagation analysis and fracture toughness evaluation. Therefore, the
ring‐notched specimen was chosen for the following research.
d
ite
ed
py
Co
Fig.5 The morphology of the normal SPT specimen after fracture
ot
tN
r ip
sc
nu
Ma
Fig. 6 The morphology of the ring‐notched SPT specimen after fracture
ed
pt
ce
Ac
Fig. 7 The morphology of the line‐notched SPT specimen after fracture
1500 Normal
Ring-notched
1200 Line-notched
Load (N)
900
600
d
300
ite
0
0.0 0.5 1.0 1.5 2.0 2.5
ed
Displacement (mm)
py
Fig. 8 Comparison of load‐deformation curves for three types of sample
4 Finite element model
Co
In this study, a finite element model of the Ramberg‐Osgood model [19] was established
based on the idea of the contour integral method.
The Ramberg‐Osgood equation is as follows: ot n 1
E ( )
tN
0 , (1)
where is the stress, is the strain, 0 is the yield stress, n is the hardening
ip
n is 5.4 which fitted from true stress‐strain curve.
sc
The ring notch was 2.4 mm in diameter and 0.2 mm deep. To make the finite element
model more similar to the experimental conditions, this study used simplified rigid bodies as
nu
the upper and lower clamps that constrained the sample. And large strain deformation was
used in the analysis. Additionally, the model and the applied boundary conditions were
axisymmetric. Therefore, a 1/2 axisymmetric model was established to simplify the mesh and
Ma
reduce the computational cost. And the 3D model was built in ABAQUS as shown in Fig.9.
Throughout the calculation, the mesh in the notch tip region was a key factor. Because
the singularity at the notch tip in the model made the stress and strain fields consistent with
ed
their actual values, the value of the J integral obtained was more accurate. In a 3D model, a
hexahedron (brick) of 20 nodes or 27 nodes. To create a singular fields, 20‐node bricks and
pt
27‐node bricks can be used with a collapsed face, as shown in Fig. 10 and Fig. 11. Move the
midside nodes to the 1/4 point nearest the crack tip. The size of the crack tip elements
ce
influences the accuracy of the solutions. The smaller the radial dimension of the elements
from the crack tip, the better the stress, strain, etc. results will be. So the mesh must be refined
Ac
in the vicinity of the crack tip to get accurate stress and strains.
A small circular tube around the notch tip was divided in the model to generate a spider
mesh. Swept hexahedral meshes were used at the notch tip and near the center of the
specimen, and a medial axis algorithm was used to regulate the element shape. Sweep
meshing with hexagonal elements and an evolutionary algorithm were used for the rest of the
specimen. The overall grid type of the model is C3D20R with a total of 11490 units and 53623
nodes. The friction between ball and specimen was 0.2.
d
ite
ed
Fig.9 3D model of a notched SPT specimen
py
Co
ot
tN
r ip
sc
Fig. 10 Hexahedron of 20 nodes collapsed to a surface
nu
Ma
ed
pt
ce
Ac
Fig. 11 Hexahedron of 27 nodes collapsed to a surface
5 Finite element analysis results
5.1 Model verification
Figure 12 shows the results of the experiments and the numerical simulations. The elastic,
plastic and final stages are almost entirely coincident, which proves the effectiveness of the
finite element model. The good agreement during the necking and fracture stages also
provides a good basis for accurately predicting the fracture toughness of the material.
900
750
600
Load(N)
450
300 FEM
EXP1
d
150 EXP2
ite
EXP3
0
ed
0.0 0.4 0.8 1.2 1.6
Displacement(mm)
py
Fig. 12 Comparison of the experimental and numerical load‐displacement curves
5.2 Analysis of the stress and strain field
Co
The stress and strain fields were analyzed based on a numerical simulation of a
ring‐notched SPT specimen, and the change in the stress gradient during deformation was
ot
considered. Figure 13 shows the stress distribution in the whole model and the partial
enlarged view present the stress distribution of notch tip zone.
tN
As shown in Fig. 13, the maximum equivalent stress appeared in the notch tip zone,
where stress concentration, plane strain and small‐scale yielding occurred in the 3D small
punch specimen with ring notches. The maximum von Mises stress was approximately 1008
ip
MPa, which was much greater than the tensile strength of the material (529 MPa). The critical
condition for notch propagation was satisfied. In addition, Fig. 13 shows a very obvious stress
r
sc
gradient near the notch tip which meet the characteristics of the pressure distribution in
quasi‐static stationary notch tip.
nu
Ma
ed
pt
ce
Ac
Fig. 13 Distribution of the stress field of a ring‐notched SPT specimen
The maximum equivalent plastic strain also occurred in the notch tip zone due to the
stress concentration (Fig. 14). The value of the maximum equivalent plastic strain was 3.441.
The results of the numerical simulation show that the notch tip zone was in a plane strain
state that gradually extended to a plane stress state. Additionally, the figure shows that the
notch tip had a small opening because the notch propagation condition had been satisfied.
Figure 14 shows a very obvious strain gradient near the notch tip which also meet the
characteristics of the strain distribution in quasi‐static stationary notch tip.
d
ite
ed
py
Fig. 14 Distribution of the strain field of a notched SPT specimen
Co
5.3 J integral
Seven contours were described in the finite element model. Figure 15 shows the
ot
variations in the J integrals along the seven contours with displacement. The J integral of each
contour increases with the load, and the seven curves coincide very well. This provides
tN
additional evidence that the contour integral method is suitable for calculating the J integral
because the J integral is independent of the contour.
ip
400
Load-Displacement
900 contour1
r
contour2
sc
10-6J (kJ/mm2)
750 contour3 300
contour4
nu
600 contour5
Load (N)
contour6 200
Ma
450
contour7
300
100
ed
150
pt
0 0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
ce
Displacement (mm)
Ac
Fig. 15 J‐ integral
5.4 Determination of fracture toughness
The value of the J integral when notch propagation initiates is the material’s fracture
toughness, as shown in Fig. 15. However, it is difficult to accurately identify the initiation of
notch propagation in the actual situation. Assuming that the notch propagated when the load
reached its maximum value, the corresponding value of the J integral was 335×10‐6 kJ/mm2,
which was greater than the value, 235×10‐6 kJ/mm2, obtained from the three‐point bending
test(According to GB/T21143-2007). This is because the notch began to propagate before the
maximum load was applied [20].
In order to find out the fracture toughness, the notch initiation point also need. However,
it is hard to define the notch initiation point in the SPT. And according to the literature [20], a
notch tip gradually opens, and the notch propagates when the load reaches 0.85 times its
maximum value and finally, extends through the thickness of the specimen. In Fig. 15, at
0.85Pmax=779N, the corresponding displacement is 1.01 mm, which corresponds to an
average J integral of 201×10‐6 kJ/mm2. This result is close to the value obtained from the
three‐point bending test, which shows that it is appropriate to consider notch initiation to
occur at 0.85Pmax.
d
The load and displacement curves of the SPT show that the deformation process
ite
generally consists of four main stages: the elastic deformation stage, the plastic deformation
stage, the stretching deformation stage and the plastic destabilization stage. Many scholars
ed
have shown that after the stretching deformation stage, the voids substantially grow and
coalesce, micro‐cracks appear, and cracks gradually propagate as the load increases.
py
The Gurson‐Tvergaard‐Needleman model (GTN) is a constitutive equations considering
micro‐voids damage. The GTN constitutive equation was applied in the simulation of a
normal SPT specimen to help determine the fracture toughness each of the notched SPT
Co
specimens. The GTN constants listed in Table 3 was get from reference [21]. Figure 16 shows
the void volume fraction (VVF) compared to the load‐displacement curve obtained from the
numerical simulations. It can be observed that as the load‐displacement curve changed, the
ot
voids in the elastic and plastic deformation stages changed very little, whereas those in the
stretching deformation stage grew quickly, and many voids coalesced during this stage.
tN
Table 3 GTN constant for Q345R
q1 q2 q3 f0 fc fF εN SN fN
ip
of the specimen had reached the fourth stage, plastic destabilization. When the third stage
was selected for notch initiation, the resulting fracture toughness was too large.
nu
1600
0.04
1400
Ma
Load
1200
VVF 0.03
1000
Load (N)
ed
VVF
800 0.02
pt
600
400
ce
0.01
200
Ac
0 0.00
0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1
Displacement (mm)
Fig. 16 The value of VVF compared to the load‐displacement curve
Under shear stress, there was a great deal of dislocation movement and crystal glide
inside the specimen. There were many impurities and voids in the grain boundaries, which
distorted the crystal lattice. As the external force increased, the stress concentration in the
material changed gradually, and the deformation was transferred from one region to the next,
which means that the notch initiated before the voids grew substantially. Therefore, it is
reasonable to suppose that notch initiation occurred during the stretching deformation stage.
Unlike the elastic and plastic stages, the stretching deformation stage had obvious linear
characteristics and was easy to identify.
In this study, the curve departed from the plastic stage and increased linearly at
approximately 1.05 mm, and the corresponding average value of the J integral was 225×10‐6
kJ/mm2, which is in good agreement with the results of the three‐point bending test.
6 Conclusion
d
Line and ring‐notched specimens were studied in this paper. Compared to the line
ite
notched specimen, the ring‐notched specimen was more suitable for notch propagation
analysis and fracture toughness evaluation. A 3D model of a ring‐notched SPT specimen was
established to determine the J integral and the fracture toughness using the contour integral
ed
method. To verify the finite element model, the results of the numerical simulations were
compared to those of the three‐point bending test and a GTN model. The conclusions are as
py
follows:
(1) Compared to the line notched specimen, the ring‐notched specimen was more
Co
suitable for notch propagation analysis and fracture toughness evaluation. Therefore, the ring
notched specimen was studied in this paper.
(2) The results of the numerical simulations were in good agreement with those of the
experiments, which shows that the numerical model described in this paper is valid. The
ot
change in the J integral of the notch tip with the load‐displacement curve was displayed. For a
tN
notch that initiated when the load reached its maximum value, the corresponding value of the
J integral was 335×10‐6 kJ/mm2, and at time 0.85Pmax, the corresponding value of the J
integral was 201×10‐6 kJ/mm2, which differed from the result of the three‐point bending test
ip
by 14.4%.
(3) A GTN model was used to evaluate the voids in the SPT specimen during
r
deformation. The voids changed very little during the elastic and plastic deformation stages
sc
but grew quickly during the stretching deformation stage, and many of them coalesced
during this stage. Therefore, it was reasonable to select the stretching deformation stage as
nu
the notch initiation point, and the corresponding fracture toughness was 225×10‐6 kJ/mm2,
which differed from the result of the three‐point bending test by 3%.
Ma
References
[1] Kassner ME. Fundamentals of creep in metals and alloys (Second Edition) [M]. Elsevier Science,
pt
2008.
[2] Evans R W, Wilshire B. Creep of metals and alloys [M]. London: The Institute of Metals, 1985.
ce
[3] Sakthivel T, Panneer S, Laha K. An assessment of creep deformation and rupture behavior of
9Cr‐18W‐0.5Mo‐VNb (ASME grade 92) steel [J]. Materials Science and Engineering A, 2015, 640: 61‐71.
Ac
[4] Dobes F, Milicka K. Application of creep small punch testing in assessment of creep lifetime [J].
Materials Science and Engineering A, 2009, 510‐511:440‐443.
[5] Garcia TE, Rodriguez C, Belzunce FJ, etc. Estimation of the mechanical properties of metallic
materials by means of the small punch test[J]. Journal of Alloys and Compounds, 2014, 582:708‐717.
[6] Xu YF, Guan KS. Evaluation of fracture toughness by notched small punch tests with Weibull stress
method [J]. Materials and Design, 2013, 51:605‐611.
[7] Turba K, Hurst R, Hahner P. Evaluation of the ductile‐brittle transition temperature in the NESC‐I
material using small punch testing [J]. International Journal of Pressure Vessels and Piping, 2013,
111‐112:155‐161.
[8] Finarelli D, Carsughi F, Jung P. The small ball punch test at FZJ [J]. Journal of Nuclear Materials, 2008,
377: 65‐71.
[9] Mao X, Takahashi H, Kodaira T. Supersmall punch test to estimate fracture toughness JIc and its
application to radiation embrittlement of 2.25Cr‐1Mo steel [J]. Material Science Engineering. 1992,
AI50:231‐236.
[10] Mao X, Takahashi H. Development of a further‐miniaturized specimen of 3mm diameter for tem
disk small punch tests [J]. Journal of Nuclear Material, 1987, 150:42‐52.
[11] Ju JB, Jang J, Kwon D. Evaluation of fracture toughness by small‐punch testing techniques using
sharp notched specimens[J]. International Journal of Press Vessels and Piping. 2003; 80:221‐228.
[12] Lacalle R, Alvarez JA, Cicero S. From archeology to precious metals: four applications of small
d
punch test [J]. Metall J 2010; LXIII: 59‐68.
ite
[13] Cuesta II, Rodriquez C, Belzunce FJ, etc. Analysis of different techniques for obtaining
pre‐cracked/notched small punch test specimens [J]. Engineering Failure Analysis, 2011, 18:2282‐2287.
[14] Turba K, Gulcimen B, Li YZ etc. Introduction of a new notched specimen geometry to determine
ed
fracture properties by small punch testing [J]. Engineering Fracture Mechanics, 2011, 78: 2826‐2833.
[15] Fould JR, Voytowite PJ, Pamell TK, et al. Fracture toughness by small punch testing [J]. Journal of
py
Testing and Evaluation, 1995, 23:3‐10.
[16] Saucedo‐Munoz ML, Liu SC, Hashida T, etc. Correlationship between JIc and equivalent fracture
Co
strain determined by small‐punch tests in JN1, JJ1and JK2 austenitic stainless steels [J]. Cryogenics, 2001,
41:713‐719.
[17] Bulloch JH. A study concerning material fracture toughness and some small punch test data for low
alloy steels [J]. Engineering Failure Analysis, 2004, 11: 635‐653.
ot
[18] Manahan MP, Argon AS, Harling OK. The development of a miniaturized disk bend test for the
determination of postirradiation mechanical properties [J]. Journal of Nuclear Material, 1981, 104:
tN
1545‐1550.
[19] Xie Q, Liu YB, Xu ZX. The determination of the hardening coefficient and exponent for
ip
Ramberg‐Osgood’s power hardening material [J]. Journal of Wuhan Transportation University, 1996,
20(3): 319‐322.(In Chinese)
r
[20] Cuesta II, Alegre JM. Determination of the fracture toughness by applying a structural integrity
sc
approach to pre‐cracked Small Punch Test specimens [J]. Engineering Fracture Mechanics, 2011, 78:
289‐300.
nu
[21]Lu YY, Chen L,Guan KS. Determination and Validation of Gurson‐Tvergaard Model Parameters for
Finite Element Simulation of Small Punch Test [J]. Applied Mechanics and Materials, 2015, 750: 59‐68.
Ma
ed
pt
ce
Ac