0
lower bound equivalent plastic strain to failure
1
upper bound of equivalent plastic strain to failure
\

=
J
J
cos
2
3
2
3 1
2
3 3
3
1
eq. 3
where is the Lodes angle.
Any stress state may be represented, combining the above
quantities, by two stress parameters, called triaxiality (T) and
deviatoric (X), defined as (Khan, 1995 [15]):
3
2
= =
p
T
( )
3
3
2
27
3 cos
J
X = =
eq. 4
Under the basic hypothesis about the uniqueness of the stress
strain law a relation between local values of T and X and the
5 Copyright 2012 by ASME
local strain subsisted. To be noted that T may span in the full
range
\

n
eq. 5
where
0
and
1
represent the fracture strain for the lower and
upper bound respectively. Equation (5) establishes a relation
between the lower and upper bound, with an explicit
dependence from the material, through the hardening coefficient
n. From eq. (5) two important considerations derive. First, the
fracture sensitivity to the Lode parameter, expressed by the
ratio
1
/
0
between the upper and lower bound strain at a given
T, depends directly on the work hardening behaviour: higher
values of n parameter correspond lower values of
1
/
0
ratio as
simply shown in Figure 5.
As second consequence, in principle it may be sufficient to
obtain only one of the bounds to find immediately also the
other.
The proposed CSM model is a more generalized criterion that
can take into account also the materials that not follow the well
known Tresca criterion.
0
2
4
6
8
10
12
14
16
18
20
0 0.2 0.4 0.6 0.8 1
n
1
/
0
Figure 5  Fracture strain sensitivity to hardening coefficient n.
Further improvements can be achieved expressing the loss of
ductility due to the influence of the deviatoric parameter X by
using a generalized criterion proposed Coppola Demofonti
(2008) [16] and implemented in this activity:
1
0
( )
( )
f
p
n
f T
D d
g X
eq. 6
where the two functions f(T) and g(X) are defined as:
2
1
( )
C T
f T C e = eq. 7
( )
1
( )
1
cos cos
6 3
g X
X
=
(
(
eq. 8
Where C
1
, C
2
, n, , , are the models constants to be defined
starting from experimental small scale test activity and
numerical simulations.
When D reaches 1 we have the failure, which is the point on
which material has reached its ductile fracture limit. An
automatic deletion algorithm is also used to eliminate the
elements that reach the critical value. On the other hands, this
dependence is well supported by many experimental evidences
and almost all models take this effect into account.
Nevertheless, according to a void growth and coalescence
description in the final phase of the fracture process, two
different mechanisms may to compete, the internal necking of
ligaments between voids, which is mainly influenced by
mediumhigh values of the stress triaxiality, and the shear
failure which is evident at low stress triaxiality, as described
recently by Bao and Wierzbicki [17] and Barsoum [18].
To use the proposal damage model it is necessary to evaluate
for the material of interest both six different parameters (C
1
,
C
2
, n, , , ) and the extended stressstrain curve. Extended
stressstrain curve is a material curve that is still valid also after
the necking point; this curve can be obtained in several ways as
the inverse calibration of tensile test[16]. So to characterize the
material for the assessment of the ductile failure limit it is
necessary to perform a set of experimental tests such as to
define the fracture locus covering as widely as possible the T
and X parameters range. The tests matrix should be designed
according to some basic requirements. At first, tests should be
able to bring the material to failure, which is easy in tensile but
not so feasible in other ones, like bending tests.
0.0
1.0
2.0
3.0
4.0
0.0 0.5 1.0 1.5 2.0
Triaxiality T
S
t
r
a
i
n
t
o
f
a
i
l
u
r
e
Upper bound X=1
Flattening
Torsion
Tension
RNB10
RNB2
Lower bound X=0
Figure 6  Fracture strain as function of stress parameters T
and X (33MnB5 annealed), [14].
In these cases it may be useful to introduce a stress
concentrator, like a smooth notch, which will induce the ductile
fracture in the desired point but with some modification in the T
and X parameters. The second requirement is to keep the T and
X evolution as constant as possible during deformation. To
place in fact the fracture point in the
f T plane (where
f
6 Copyright 2012 by ASME
is the equivalent plastic strain at fracture, and keeping X as
constant) it is necessary to define a unique value for both T and
X. If the two parameters vary during the strain path, we have to
define an effective or mean value, as defined in Wierzbicki
(2005) [17]:
0
1
( )
f
m
f
T T d
0
1
( )
f
m
f
X X d
eq. 9
where is the equivalent plastic strain and ) ( T , ) ( X are the
actual T and X values during the path. In this case it is necessary
to perform numerical calculations, (i.e. finite elements) to
precisely calculate in each point the T and X evolution by using
the equations (8). So the main effort in the test design is
devoted, in general, to the selection of loading paths that give
the more constant as possible T and X parameters evolution as
function of applied plastic strain. An example of stress
parameters (T and X) that can be obtained from conventional
tests is reported in Table 1, while the Figure 6 shows an
example the obtained strain at the failure function of X and T
for 33MnB5 annealed as reported in the reference.
Table 1. Summary of tests and stress parameters target values,
[14].
Test type
Triaxiality
parameter, Tm
Deviatoric
parameter, Xm
Tensile smooth >0.33 1
RNB >0.6 1
Torsion 0 0
Flattening 0.557 0
Bending T~0.5 0<X<1
5. EXPERIMENTAL ACTIVITY
5.1 Selection of material and Laboratory test
A selection of material grades has been done with the aim to
identify a wide range of materials available for gas pipeline
transportation. The selection was focused, as much as possible,
choosing pipes involved in past full scale burst tests, where the
actual fracture arrest capability has been verified. Moreover, to
complete the set of materials, also new production pipes were
selected. Starting from these needs the following pipes have
been selected:
Longitudinally seam welded UOE pipe no. E411 (X100
OD 48 WT 18.4mm) used in the ENI TAP full
scale burst test. The full scale burst test result for this
pipe was a propagation at high speed
Longitudinally seam welded UOE pipe no. 834 (X100
OD 36 WT 20.0mm) used in the second full scale
burst test within the DEMOPIPE project. The full scale
burst test results for this pipe was a natural arrest [9].
Longitudinally seam welded UOE pipe no. 67284 (X80
56 26.0mm). Pipe used in the full scale burst test of
Long Distances ENI/SNAM project (L.D.). The full
scale burst test results for this pipe was a natural arrest
[19]. It is important to note that this pipe after a new
complete mechanical characterization has been
downgraded to X70 for our activity, as confirmed also
from the data reported in the following. (this pipe comes
from an experimental production)
Longitudinally seam welded UOE pipe no. 31 (X65
OD 56 WT18.7mm) used in the CSM full scale burst
test no. 21. The full scale burst test results for this pipe
was a natural arrest.
Longitudinally seam welded UOE pipe no. 924996 (X80
OD 56 WT 27.7mm). Recent production. Pipe not
involved in any full scale burst test.
Spiral welded pipe, n 930605 (X80 OD 42" WT
18.8 mm). Recent production. Pipe not involved in any
full scale burst test.
5.2 Tensile properties: true stress true strain curve
For each pipe round bar specimens both for longitudinal and
transversal direction have been tested at low strain rate
(according ISO 68921 and ASTM A370). The results of the
tensile tests are reported in Table 2 and Table 3.
Table 2. Tensile properties in transversal direction
Material Rp0.2 %
(MPa)
Rm
(MPa)
Rp0.2 /
Rm
(%)
Ag
(%)
A
(%)
Z
(%)
X100 TAP E411 812 814 99.7 1.0 18.3 76.2
X100 Demopipe 779 812 95.9 4.8 22.0 74.1
X80 L.D. 67284 500 596 83.9 11.2 30.1 74.5
X65 Pipe 61 469 576 81.4 12.0 29.4 67.2
X80 UOE 632 718 88.1 7.9 27.3 76.2
X80 Spiral 638 722 88.4 8.5 27.9 76.8
Table 3. Tensile properties in longitudinal direction
Material Rp0.2 %
(MPa)
Rm
(MPa)
Rp0.2 /
Rm
(%)
Ag
(%)
A
(%)
Z
(%)
X100 TAP E411 702 737 95.2 5.4 23.6 76.3
X100 Demopipe 686 747 91.8 3.6 22.5 76.0
X80 L.D. 67284 448 571 78.0 11.0 32.1 76.4
X65 Pipe 61 467 564 82.5 10.8 29.4 66.0
X80 UOE 553 670 82.5 7.5 29.2 75.6
X80 Spiral 643 737 87.3 6.3 25.4 73.0
To highlight this result a comparison between transversal
directions of nondimensional stressstrain curves is reported in
Figure 7 where it is possible see a relevant difference in work
hardening between the selected materials.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0% 5% 10% 15% 20% 25%
Strain (%)
S
t
r
e
s
s
/
R
p
0
.
2
(
M
P
a
)
X80  Spiral pipe
X80  UOE new production
X65  CSM full scale burst test no.21
X80  Long Distance ENI/SNAM project
X100  Demopipe Project
X100  Eni TAP project
Figure 7  Different strain hardening on transversal specimen.
7 Copyright 2012 by ASME
Moreover in Figure 8 it is possible to see a relevant difference
in work hardening coefficient n among the selected materials.
These coefficients correlate well with the uniform elongation as
reported in Figure 8.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
Work Hardening Coefficient
U
n
i
f
o
r
m
E
l
o
n
g
a
t
i
o
n
(
e
)
X100  ENI TAP Project
X100  Demopipe
Project
X80  UOE new production
X80  Spiral pipe
X80  Long Distance ENI/SNAM project
X65  CSM full scale
burst test no.21
Figure 8  Correlation between work hardening coefficient and
uniform elongation values on transversal specimen.
5.3 Toughness properties by: CharpyV, DWTT shelf
energy and CTOA
Toughness properties were also measured by means both
CharpyV and Pressed Notch Drop Weight Tear (PNDWT)
test shelf fracture energy on specimen machined in transverse
direction and tested at room temperature. Moreover the
measurements of the Crack Tip Opening Angle has been made
using the Two Specimens CTOA Propagation Energy Test,
[20]. The results of these tests are reported in Table 4.
Table 4. Toughness properties of selected steels: CharpyV,
DWTT shelf energy, CTOA.
Material
CharpyV
shelf energy
(J)
DWTT specific
total energy
(J/cm2)
DWTT Specific
prop. energy
(J/cm2)
CTOA
()
X100 TAP E411 289 661 304 2.2
X100 Demopipe 251 792 466 4.3
X80 L.D. 67284 253 909 419 9.2
X65 Pipe 61 126 445 228 5.3
X80 UOE 304 >1476 >715 8.5
X80 Spiral 267 1132 629 6.4
Fracture surfaces of DWTT specimens showed that X100 TAP
and X80 Long Distances ENI/SNAM project materials (L.D.
Material) exhibit several separations in comparison to the other
materials. Concerning the DWTT tests Figure 9 shows the
differences between the selected materials in terms of load
displacement curve normalized with respect to maximum load.
These graphs underline that there are important differences
between the selected materials. This is also confirmed in the
Table 4 where these material are compared in terms of Charpy
V shelf energy DWTT (total and propagation energy) and
CTOA.
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 10 20 30 40 50 60
Displacement from crack onset (mm)
D
i
m
e
n
s
i
o
n
l
e
s
s
f
o
r
c
e
X100 TAP E411
X100 Demopipe 834
X80 Long Distance
X65 CSM pipe 21
X80 UOE New Prod.
X80 Spiral pipe
Figure 9  Loaddisplacement curve normalized respect to
maximum load for DWTT tests.
6. DAMAGE PARAMETERS EVALUATION
6.1 Testing activity to obtain the damage model
parameters.
The above proposed damage model implemented into FE code
(MSC Marc commercial finite element code), that permits the
fracture propagation (by elements release), has to be calibrated
using a series of experimental tests that exhibit different values
of triaxiality level (T) and deviatoric parameter (X) of the stress
state. So, the following specimens have been tested:
Round bar smooth tensile tests
Notched round bar tensile tests (radius = 2 mm)
Notched round bar tensile tests (radius = 10 mm)
Torsion tests
Notched three point bend tests (plain strain)
In parallel to the laboratory tests program, the FE simulations
were carried out for all the tensile specimens (RB, NRB2,
NRB10) and for the bending tests to obtain the trend of the
triaxiality (T) and deviatoric parameter (X), during the
deformation, measured in the center of the specimen. For the
pure torsion tests no numerical simulations have been
performed, in fact both the value of T=0, X=1 and strain at the
failure are directly obtained from experimental data. To
perform the numerical simulations an evaluation of the
extended true stressstrain curve it is needed. To obtain this
information for each steels involved an inverse calibration of
loaddisplacement curve has been carried out [16].
In Figure 10 and Figure 11 are reported the strain limit curves
for the materials that are object of this study are reported, in
particular Figure 11 refers to low/medium grade steels (X65
Pipe 61 and X80 L.D. 67284 actual X70) and Figure 10 to
higher grade steels with lower level of uniform elongation. The
pictures report the upper limit (upper curve) and lower bound
(lower curve) plotted with the same color for the same material.
The first important consideration that can be done looking these
charts is that highest grades exhibit a larger difference between
8 Copyright 2012 by ASME
the lower and upper bound curves. This means that there is
large difference in terms of strain to failure between X=0 curve
and X=1 curves. The influence of the deviatoric parameter for
high grade materials seems to be key point in the failure criteria.
This is also in agreement with the highlighted influence of a
reduced value of strain hardening on the divergence of the
lower and upper lines. This is confirmed both from X80 L.D.
67284 (to be note that this steel nominally X80 resulted actually
more similar to an X70) and from X65, where the lower and
upper bounds are very close. So, the damage model seems to be
able to distinguish the behaviour at rupture for different grades.
This effect can be related to the hardening capability of the
materials.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0.0 0.5 1.0 1.5 2.0
Triaxiality T
S
t
r
a
i
n
t
o
f
a
i
l
u
r
e
X100 2nd Demopipe (834)
X100 Eni TAP project
X80 UOE new prod.
X80 Spiral pipe
Figure 10  Strain limit curves for grade steels X80.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0.0 0.5 1.0 1.5 2.0
Triaxiality T
S
t
r
a
i
n
t
o
f
a
i
l
u
r
e
X65CSM Test 21 (31)
X80 L.D. ENI/SNAM (67284)
,,
Figure 11  Strain limit curves for grade steels < X80.
6.2 Testing for the validation the applicability of
plastic damage as fracture critical parameter
With the aim to setup the damage model for the correct
prediction of fracture propagation, Single Edge Notch Bend
(SENB) specimens have been machined from all steels involved
and then tested at room temperature and at low strain rate (no
dynamic test). The geometry of the samples is similar to DWTT
specimen, but with a longer initial crack; in particular all these
samples have been fatigue precracked up to achieve a total
notch depth of about 15 mm. Finally, a grid of 2 mm side has
been etched on one side of the specimen to measure after the
test the frozen plastic deformation on samples with the aim to
compare these data with the same quantities calculated by
means finite element simulations.
Analysis of the grid deformation has been carried out by using
the Automated Strain Analysis and Measurements Environment
(ASAME) that permits, using a double camera, to evaluate the
deformation in the two principal directions (grid orientation).
This ASAME system is a portable camerabased measurement
tool that measures the surface geometry and strain distribution
on a deformed part. The use of the double camera technique
allows to measure deformation correctly also in the regions that
are not perpendicular to the camera lens, for instance near to the
fracture path where the wall thickness decreases due to the
necking.
An analysis of these data shows some important issues:
the longitudinal deformation in the fracture path is
negligible for all the tested samples. This is a strong
difference respect to the pipe rupture way in the full
scale propagation test where longitudinal component of
the strain is very important (up to 58 %).
in transversal direction the X65 pipe shows the lowest
value and also the extension of the deformation is
smaller than other materials.
X80 L.D. pipe seems to have stored more plastic
deformation for the initiation respect to the propagation
phase.
the other X100 and X80 materials have similar
behaviour in terms of plastic deformation near the
fracture path.
For each sample, during the test, a series of unloading steps has
been performed to measure the amount of energy versus crack
propagation. For each loading step, the area of the load versus
displacement curve has been calculated and the increment in
crack propagation has been measured using the corresponding
pictures.
7. APPLICATION OF DAMAGE MODEL TO
REPRODUCE THE DUCTILE CRACK GROWTH
ON SENB SPECIMENS
The damage model, calibrated for all steels involved in this
EPRG project, has been implemented inside a commercial finite
element code and used as criteria to kill elements to the crack
tip when the value of damage D (defined in equation no.6)
achieves the unit. In particular to carry out these simulations a
3D model has been used and an isotropic description of
material has been used in a first phase of the work.
SENB specimens have been used for the validation of model
that has been calibrated using the small scale test reported in
Table 1.
9 Copyright 2012 by ASME
Generally the results of comparison between experimental test
(SENB samples) and their simulations show that it is possible:
to reproduce the load vs crack growth;
to create a plastic deformation field frozen on the
specimens very similar to the measured strain on broken
SENB specimens;
to generate a shape of crack front very similar to the real
specimen.
As example in the Figure 12 a comparison between
experimental and calculated load vs displacement diagram for
the X100 TAP E411 pipe shown.
Figure 12  Comparison between experimental and calculated
load vs displacement diagram for the X100 TAP E411.
After the test, experimental and calculated plastic deformation
on the specimen surface, are compared and shown in the
subsequent Figure 13 and Figure 14.
Figure 13  Comparison between plastic deformation calculate
by FEA (left side) and plastic deformation measured on sample
surface using grid reading after the test (right side).
Also if the scale of contour is not the same, it is possible to note
that the maximum deformation on fracture path is almost the
same (+0.4 ), also the transversal extension of deformation is
very similar.
Figure 14  Comparison between FEA and experimental crack
growth for two different crack extension values.
Finally, Figure 15 shows the shape of the crack front obtained
during the simulated crack growth. It can be noted that the
calculations reveal a similar typical wedge shape as could be
observed in the experimental tests. This is a consequence of the
different local values in the process failure zone of the value
of T and X, along the thickness of the specimen, with an average
value of T
m
of about 1.2.
Figure 15  Shape of the crack front obtained during the
simulated crack growth.
8. PRELIMINARY STUDY ABOUT THE EFFECT OF
LIGAMENT LENGTH ON DUCTILE CRACK
GROWTH
To quantify the role of the specimen ligament length on the
growth of a ductile crack the simulation of:
A low strain rate failure of a CharpyV specimen. (quasi
static three point bending test)
Single Edge Notch in Tension (SENT) specimen with a
ligament length of about 500 mm,
A section of pipe with a longitudinal crack
have been performed.
Also in these three cases, the results in terms of both expected
shape of crack and plastic deformation are very promising. In
fact, as shown in Figure 16, the crack front of CharpyV
specimen is flat, with an average value of T
m
= 1.71.8. In the
case of SENT specimen and pipe section the shape of crack
front comes back to be more similar to the SENB specimen, but
10 Copyright 2012 by ASME
with an average value of T
m
= 0.6. In Figure 16 the shape in
both cases are reported.
Figure 16  Shape of crack front, for CharpyV specimen
geometry (left) and SENT specimen (right).
These values of average triaxiality (together with the previous
value on the DWTT specimen) can be reported qualitatively in
terms of the value of averages of T
m
at the crack front on the
fracture strain diagram as function of stress parameters T and X,
as shown in Figure 17. It is important to note that X values for
all the samples during the fracture propagation can vary in the
range between 0 and 0.5, and the average values are very close
to zero. A preliminary consideration is that this approach in
terms of parameters T and X could permit to quantify the
distance in terms of toughness prediction among CharpyV
specimen, DWTT specimen and large ligament (pipe). This
issue highlights the influence of sample ligament in material
rupture limit. In particular longer ligaments permit a larger
development of plastic field in front of crack tip with respect to
the smaller ones.
At the same time the increasing of divergence between the X=0
and X=1 curves when the strain hardening decrease, as showed
in Figure 5, can give a justification about the non unique
correlation between DWTT and CharpyV specimens when the
strain hardening of steel decreases.
Figure 17  Location of CharpyV, DWTT, SENT specimen
(ligament 500 mm) and pipe on the fracture strain diagram as
function of stress parameters T and X.
9. CONCLUSIONS
Published results of full scale burst tests on high grade pipeline
steels (within the range from API 5L X80 to X120 operated at
very high hoop stress values 500 MPa) show that the
applicability of CharpyV shelf energy to quantify the resistance
of steel to the ductile fracture propagation event is highly
questionable.
There is a common opinion that one of the main causes of the
misprediction can be related both to reduced CharpyV
specimen geometry (a small ligament can increase the constraint
at the crack tip) and also to the low intrinsic value of ductility
when the grade of the steel increases, with a consequent
reduction of both workhardening and value of the strain at
maximum load.
In order to give a contribution on the understanding and
quantify the cause of this misprediction a new plastic damage
approach to give an answer to these open issues has been
proposed.
The plastic damage model is based on two stress parameters,
called triaxiality (T) and deviatoric (X). The model, after an
initial set up, has been implemented inside a finite element
commercial code and used as fracture parameter criteria to
simulate the fracture process. Its validation has been performed
by using six different large diameter steel pipes with a nominal
grade in the range of X65  X100.
Generally first results are very promising, in fact there is a
substantially agreement between the experimental data and
results obtained by numerical simulation in terms of load vs
displacement curve and local plastic deformation frozen in the
specimen.
Preliminary considerations about the future use of this
methodology show that this approach could permit to quantify
the distance in terms of toughness prediction among Charpy
V specimen, DWTT specimen and large ligament (pipe or
plate). In particular the influence of ligament length of sample
has been showed as influence on deformation of failure of the
material.
An indication of influence of ductility role on the behavior of
material at rupture, in terms of strain hardening of steel, has
been highlighted on stronger influence of deviatoric parameter
on low strain hardening materials that tends to separate the
lower and upper bound curves in failure strain diagram.
The work is in progress and new challenging topics must be
faced, as the role of strain rate in the calibration of proposed
damage model and its applicability to a pipe size.
10. ACKNOWLEDGEMENT
The authors wish to acknowledge the considerable contributions
made by many of their colleagues and their companies, within
11 Copyright 2012 by ASME
the EPRG, towards the development and promotion of the
research project described above.
11. REFERENCES
1. Maxey, W.A. Fracture Initiation, Propagation and Arrest
5th Symposium on Line Pipe Research, Houston, Texas,
November 1974, paper J.
2. AISI technical Report Running Shear Fractures in Line
Pipe Subcommittee of Large Diameter Line Pipe
Producers, September 1974.
3. Fearnehough, G.D., Jones, D.G. Toughness Specification
for Shear Fracture Arrest in Pipelines International
Conference on Analytical and Experimental Fracture
Mechanics, Rome, Italy, June 1980.
4. Bonomo, F. et al. A Survey and tentative Revision of
Ductile Arrest Criteria in Pipelines for Gas Transmission
International Conference on Analytical and Experimental
Fracture Mechanics, Rome, Italy, June 1980.
5. Eiber, R.J., Bubenik, T.A. Fracture Propagation Control
Plan Methodology 9th PRCI/EPRG Joint Technical
Meeting, Houston, Texas, May 1993, paper 20.
6. Eiber, R., Bubenik, T.A., Maxey, W.A. Fracture Control
Technology for Gas Pipelines American Gas Association,
Catalogue No. L51691, 1993.
7. Meleddu, A., Mannucci, G., Di Biagio, M., "Desk study on
existing methods for the evaluation of material fracture
resistance on high grade steel pipelines EPRG project
1362006
8. Papka, S. D., Stevens, J. H., Macia, M. L., Fairchild, D. P.,
Petersen, C. W. FullSize Testing and Analysis of X120
Linepipe, Proceedings of the Thirteen International
Offshore and Polar Engineering Conference 2003,
Honolulu, Hawaii, USA, May 2003.
9. Demofonti, G., Mannucci, G., Di Biagio, M., Hillenbrand,
H.G., Harris, D. Fracture propagation resistance
evaluation of X100 TMCP steel pipes for high pressure
gas transportation pipelines using fullscale burst tests,
4th International Pipeline Technology Conference, Ostend,
May 2004.
10. I Yu Pyshmintsev, T P Lobanova, A B Arabey et al. Crack
arrestability and mechanical properties of 1420mm X80
Grade pipes designed for 11.8MPa operation pressure
Pipeline Technology Conference, Ostend, 1214 October
2009, Paper no: Ostend2009078
11. Demofonti, G. et al. UltraHigh Strength Pipeline
Prototyping For Natural Gas Transmission. DEMOPIPE;
Research Programme of the Research Fund for Coal and
Steel RDT, Contract No 7215PP/048, Final Report, 2004.
12. I Yu Pyshmintsev, A B Arabey, A M Gervasyev, and A N
Boryakova Effects of microstructure andtexture on shear
fracture in X80 linepipes designed for 11.8MPa gas
pressure Pipeline Technology Conference, Ostend, 1214
October 2009, Paper no: Ostend2009078, Paper no:
Ostend2009028
13. Bao, Y., Wierzbicki, T., 2004, On Fracture Locus in the
equivalent Strain and Stress Triaxiality Space, Int. Journal
of Mech. Sciences, 46, 8198.
14. Coppola T., Cortese L., Folgarait P., 2009. The effect of
stress invariants on ductile fracture limit in steels. Eng.
Fract. Mech. 76 (2009) 12881302.
15. A. Khan, S. Huang (1995), Continuum Theory of
Plasticity, John Wiley&Sons, Inc.
16. Coppola T., Demofonti G. Advanced methods for the
strain limit assessment in pipeline applications subjected
to extreme loading 2nd ISOPE Strain Based Design
Symposium, SBD 2008, Paper N. ISOPE2008TPC704
Vancouver, Canada, July 611, 2008.
17. Wierzbicki, T., Bao, Y., Lee, YW., Bai, Y., 2005.
Calibration and Evaluation of Seven Fracture Models, Int.
Journal of Mech. Sciences 47, 719743
18. Barsoum, I, Faleskog, J, 2007. Rupture Mechanisms in
Combined Tension and Shear  Experiment, Int. Jou. Solid
and Structures, Vol 44, pp 17681786
19. G. Mannucci, G. Demofonti, M.R. Galli, C.M. Spinelli
"Structural integrity of API 5L X70X80 steel grade
pipeline for high pressure long distance transmission gas
lines" International Gas Research Conference, 811
November, San Diego, USA, 1998
20. G. Mannucci, G. Demofonti, E. Mecozzi. Mill Test
Techniques for Predicting Crack Arrest Arrestability in
High Toughness Steels, PRCI Project No.PR1829903,
March 2002.