fe-model for titanium alloy v cutting based
TRANSCRIPT
-
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
1/13
ORIGINAL RESEARCH
FE-model for Titanium alloy (Ti-6Al-4V) cutting based
on the identification of limiting shear stress
at tool-chip interface
Yancheng Zhang & Tarek Mabrouki & Daniel Nelias &
Yadong Gong
Received: 9 December 2009 /Accepted: 5 May 2010# Springer-Verlag France 2010
Abstract Modeling of metal cutting has proved to be
particularly complex, especially for tool-chip interface. The present work is mainly aimed to investigate the limiting
shear stress at this interface in the case of Titanium alloy
(Ti-6Al-4V) dry cutting based on a FE-model. It is first
shown that the surface limiting shear stress was linked to
the contact pressure and the coefficient of friction (CoF). A
relationship between CoF and the limiting shear stress was
given, and the effect of the temperature on the limiting
shear stress was also considered. After that, an orthogonal
cutting model was developed with an improved friction
model through the user subroutine VFRIC in Abaqus/
Explicit software. The numerical results obtained were
compared with experimental data gathered from literature
and a good overall agreement was found. Finally, the
effects of cutting speed, CoF and tool-rake angle on chip
morphologies were analyzed.
Keywords FE cutting model . Limiting shear stress .
Fracture energy. Titanium alloy Ti-6Al-4V
Abbreviations
FE Finite element
J-C Johnson-Cook
SDEG Scalar stiffness degradation
CoF Coefficient of friction
Nomenclature
A initial yield stress (MPa)a half contact width (mm)
ap cutting depth (mm)
B hardening modulus (MPa)
C strain rate dependency coefficient (MPa)
Cp specific heat (J/kg1C
1)
D overall damage variable
D1...D5 coefficients of Johnson-Cook material shear
failure initiation criterion
E1 tool insert Youngs modulus (MPa)
E2 machined material Youngs modulus (MPa)
f feed rate (mm/rev)
fc chip segmentation frequency (kHz)
Fc cutting force (N)
Ff feed force (N)
L characteristic length (mm)
Lc chip segmentation wavelength (mm)
m thermal softening coefficient
m1 contact coefficient
n work-hardening exponent
n1 contact coefficient
p hydrostatic pressure (MPa)
R radius of a cylinder (mm)
Pc contact force per unit length (N/mm)
p0 maximum contact pressure (N/mm2)
qo tangential traction at x = 0 (N/mm2)
tf limiting shear stress (MPa)
tY shear stress calculated by yield stress (MPa)
g shear strain in the primary shear zone
Rn cutting edge radius (m)
T temperature at a given calculation instant (C)
tc cutting time (s)
Tm melting temperature (C)
Tr room temperature (C)
Y. Zhang : T. Mabrouki (*) :D. NeliasUniversit de Lyon, CNRS, INSA-Lyon, LaMCoS, UMR5259,
69621 Lyon, France
e-mail: [email protected]
Y. Gong
School of Mechanical Engineering & Automation,
Northeastern University,
Shenyang 10004, China
Int J Mater Form
DOI 10.1007/s12289-010-0986-7
-
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
2/13
T* homologous temperature T
T Tr = Tm Tr uf equivalent plastic displacement at failure (mm)
u equivalent plastic displacement (mm)
VC cutting speed (m/min)
Vchip local chip sliding speed (m/min)e plastic strain rate (s1)
e0 reference strain rate (s1)e
normalized effective strain ratee
e= e0e0i equivalent strain at the onset of damage
ef equivalent plastic strain at failure
e equivalent plastic strain increment
e equivalent plastic strain
a0 flank angle (deg)
ad expansion coefficient (m/m/C)
g0 rake angle (deg)
l thermal conductivity (W/m/C)
Poissons ratio
KC fracture toughness (MPa ffiffiffiffimp
)
Gf fracture energy (N/m) CoF
density (kg/m3)
I principal stresses (i=1...3) (MPa)
s von Mises plastic equivalent stress (MPa)
Y yield stress (MPa)
* stress triaxiality s p=sw damage initiation criterion
Introduction
For the high strength-to-weight ratio, combined with an
excellent corrosion resistance at high temperature, the
cutting of titanium alloys has recently received considerable
interest due to their wide range of application in aerospace,
automotive, chemical, and medical industry [1].
However, titanium alloys are classified as hard machin-
ing materials because of their high chemical reactivity and
low thermal conductivity [2]. The high chemical reactivity
increases with temperature and produces an early damage
of the cutting tool affecting the surface quality and
increasing the production costs [3]. Their low thermal
conductivity hinders the evacuation of the heat generated
during the cutting process resulting in a temperature rise of the
workpiece [3]. This leads to the characteristic segmented
chip feature even at very low cutting speed, as it was
mentioned by Hou and Komanduri [4]. These authors
predicted the critical cutting speed value corresponding to
the onset of shear localization of Ti-6Al-4V to be approx-
imately equals to 9 m/min whereas this value is equal to
130 m/min when cutting AISI4340 steel. In order to increase
productivity and tool-life in machining of titanium alloys, it
is necessary to study the mechanics of chip segmentation and
the effects of the working parameter variation on material
cutting performance. For that, it is more judicious to build a
reliable FE-model allowing more physical comprehension in
relationship with this type of material cutting.
Nevertheless, the robustness of a given numerical model
is strongly dependent on the work material contact nature at
cutting tool-chip interface, material fracture criterion, and
both tool and workpiece material thermal parameters [58].In the present study, Material fracture energy was put
forward to reduce the mesh dependency and achieve
material degradation during cutting process. Moreover, the
mesh sensitivity was also analyzed in order to obtain the
appropriate mesh size. The limiting shear stress in the friction
model was refined with the maximum shear stress at the tool-
chip contact surface.
Comparison of predicted results (in terms of chip
morphologies and cutting forces) with those obtained from
experimental studies was carried out to validate the
numerical model for positive and negative tool-rake angle.
Finally, effects of cutting speed, coefficient of friction(CoF), and rake angle on the chip morphology were studied
with the numerical cutting model.
Numerical approach
Modeling data and geometrical model
To improve physical comprehension of segmented chip
formation, friction properties during Titanium alloy Ti-6Al-
4V cutting, the commercial software Abaqus 6.8-2 with its
explicit approach was employed. A 2D orthogonal cutting
model was developed as shown in Fig. 1. Linear quadrilat-
eral continuum plane strain element CPE4RT with reduced
integration was utilized for a coupled temperaturedisplace-
ment analysis.
Machining parameters were taken similar to those
adopted by Jiang and Shivpuri [9] and Umbrello [5]. For
the workpiece, the uncut chip thickness was 0.127 mm with
a cutting depth ap=2.54 mm. The WC ISO-P20 cutting tool
considered has normal rake and flank angles of 15 and 6,
respectively. The tool entering and inclination edge angles
were 90 and 0, respectively. The tool-cutting edge radius
was of 0.030 mm. The tool-workpiece interaction was
considered under dry machining conditions.
To optimize the contact management during simulation,
a multi-part model (Fig. 1) was typically developed with
four geometrical parts: (1) Part1the insert active part, (2)
Part2the uncut chip thickness, (3) Part3the tool-tip
passage zone, and (4) Part4the workpiece support.
It should be noticed that the thickness of tool-tip passage
zone (Part3), which is the sacrificial layer, was usually
recommended larger than that of the cutting edge radius
Int J Mater Form
-
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
3/13
[10] to avoid mesh distortion problem. In the proposed model,
the critical layer size above which the element can climb up
rake face without mesh distortion was found equal to0.026 mm. The sacrificial zone was meshed with five
elements to study the contact effect around the cutting edge.
Also, the passage zone (Part3) was divided into three small
parts: Part3-1, Part3-2, and Part3-3. The CoF in Part3-1should
be very small to reduce the influence of the sacrificial zone for
the new formed chip surface, or the chip surface can be torn
out by the sacrificial zone before it enters into contact with the
tool surface. As Part3-2 and Part3-3 are usually the stick zones
[11], the CoF should be set very high. However, there is no
more contact problem for them when they are finally deleted.
For boundary conditions used in the present model, the
cutting tool is assumed to be fixed on its top and right sides,and the workpiece is allowed to move horizontally from the
left to the right while restrained vertically.
Material behaviour and chip formation criterion
Material constitutive model
The material constitutive material model of Ti-6AL-4V
follows the Johnson-Cook (J-C) model [12]. It provides a
satisfactory description of the behaviour of metals and alloys
since it considers large strains, high strain rates and
temperature dependent visco-plasticity. This model isexpressed by the following expression of the equivalent stress.
s A B"n 1 Cln e 1 Tmh ih
1
The J-C material parameters of the workpiece made in
Ti-6Al-4V can be found in Table 1, whereas the physical
parameters of both the workpiece and the tool-insert aregiven in Table 2.
Chip separation criterion
For different software and material constitutive models, the
approach to deal with element damage is different. Indeed,
Umbrello [5] and Jiang and Shivpuri [9] employed the
Cockroft and Lathams criterion to predict the effect of
tensile stress on chip segmentation, but the elements are
deleted at the onset of damage initiation. This induces
instability in numerical simulation. The J-C model is widely
adopted for its satisfactory description of material flowstress, while different approaches are used to achieve the
degradation of material to get segmented chip. Calamaz
et al. [15] introduced the strain softening through
improving the J-C model by adding the strain softening
effects. While the fracture energy as a failure evolution
criterion after damage initiation can well achieve material
degradation, and this approach was adopted for the present
study.
Damage initiation The initiation of damage in the J-C
material model is derived from the following strain
cumulative damage law:
w Xe
e0i2
Vc X
Y
No displacement
in Y-direction
Fixed boundaries
Part2Part1
Rn=30m
25
0
0Part3
Part4
Damagezones
Part2+Part3
26m
Part3-1
Part3-2
Part3-3
:
Fig. 1 Model mesh and bound-
ary conditions
Materials A (MPa) B (MPa) n C m D1 D2 D3 D4 D5
Ti-6Al-4V 1098 1092 0.93 0.014 1.1 0.09 0.25 0.5 0.014 3.87
Table 1 Johnson-Cook Materi-
al Model [13]
Int J Mater Form
-
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
4/13
Where $e is the increment of equivalent plastic strain
during an increment of loading, and e0i the equivalent strain
at the onset of damage and is expressed as following:
"0i D1 D2 exp D3s h i
1 D4 ln e 1 D5Th ih
3The parameters D1, D2, D3, D4, and D5 are experimental
data (Table 1). Damage is assumed to be initiated when the
parameter w=1.
Damage evolution The Hillerborgs fracture energy [16]
was introduced in this model mainly for two reasons:
Firstly to control the material degradation after the damage
initiates (Fig. 2), which makes the failure process more
stable, secondly, to capture high strain localization
during chip segmentation even for relatively large size
element.
When material damage occurs, the stress-strain relation-
ship does no-longer accurately represent the materials
behavior. Continuing to use the stress-strain relation
introduces a strong mesh dependency based on strain
location, such that the energy dissipated as the mesh is
refined. The mesh size in the Calamazs [15] model is
around 2 m near the cutting edge and along the primary
shear zone. The Hillerborgs fracture energy, Gf, was
adopted to reduce the mesh dependency by creating a
stress-displacement response after damage initiation.
Gf Z"pl
f
"pl
0
LsYd"pl
Zuplf
0
sYdupl 4
Where, the characteristic length L is the square root of
the integration point element area based on a plane strain
element CPE4RT.
The scalar stiffness degradation for the linear damage
process used for part 3 is given by:
D Leuf
uuf
5
Where the equivalent displacement is uf 2Gf=sY.
Whereas an exponential damage parameter used for part 2,
evolves according to:
D 1 exp Zu
0
s
Gfdu
6
The formulation of the model ensures that the energydissipated during the damage evolution process is equal to
Gf, and the scalar stiffness degradation approaches to one
asymptotically at an infinite equivalent plastic displace-
ment. In the case of plane strain condition Gf can be
obtained by Gf K2C 1 n2 =E where KC is the fracturetoughness [14].
Mesh sensitivity During machining process, large defor-
mation is a common phenomenon, especially for the
segmented chip. According to Eq. 4, to get this large
deformation with constant element characteristic length L,
the value of plastic strain must be very high, which alsodepends greatly on the local mesh. The characteristic
length L could also be increased to reduce the mesh
dependency. Consequently, the question that can be asked
in the present formulation is: which size is appropriate for
the characteristic length L?
Two constraints can be retained for the evaluation of L:
First, the size should be relative large (to save computing
Physical parameters Workpiece (Ti-6AL-4V) Tool(WC ISO-P20)
Density, (kg/.m3) 4430 15700
Elastic modulus, E (GPa) 110 705
Poissons ratio, 0.33 0.23
Specific heat, Cp (J/kgC) 670 178
Thermal conductivity, l (W/mC) 6.6 24
Expansion coef., d (m/m/C) 9 5
Tmelt (C) 1630
Troom (C) 25 25
Table 2 Workpiece and tool
physical parameters [14]
Damage initiation
( =1, D=0)
Damage
evolution
Material stiffness
is fully degraded
(D=1)E
E
0if
~D
(1-D)E
y
c
d
b
ad
Fig. 2 Typical uniaxial stress-strain response of a metal specimen
[17]
Int J Mater Form
-
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
5/13
time), second the result should be similar or close to
experimental values (in terms of cutting force and chipmorphology). For that, four mesh mean values are
discussed with characteristic length of 6, 8, 12, and
14 m, where the simulated cutting forces are compared
with the experimental ones. The dynamic cutting force
versus the characteristic length is first investigated.
Furthermore, the fracture energy was refined to get the
appropriated material degradation evolution and conse-
quently the real segmented chip morphology.
Figure 3 presents the cutting force sensitivity for
different mesh sizes. For both characteristic length of 12
and 14 m, the computed average cutting forces are 481 N
and 458.5N, respectively, which are far below the value of
550 N measured experimentally under the same working
conditions [5, 9]. Whereas for characteristic lengths of
8 m and 6 m, the cutting force reaches 541 N and 549 N,
respectively, which is close to the cutting force obtained
especially. To contain the computing costs, the mesh size of
8 m will be now chosen.
Limiting shear stress from the aspect of contact
mechanics
The Zorevs temperature independent stick-slip friction
model [11], see Eq. 7, was widely adopted by many
authors [10, 15, 18, 19] to define the friction properties at
the toolchip interface. Zorev advocated the existence of
two distinct toolchip contact regions: In the stick zone
near the tool tip the shear stress tfis assumed to be equal
to the yield shear stress of the material being machined,tY,, whereas, in the sliding region, the frictional stress is
lower than the yield shear stress. Note that a constant
Coulomb friction coefficient is assumed here.
if tf < tY & tf msn ! Sliding regionif tf tY & tf msn ! Stick region
&7
The yield shear stress tY, in Eq. 7 is usually related to the
conventional yield stress Y of the workpiece material
adjacent to the surface [18, 20, 21]. A reasonable upper
X
Z
aaVchip
P Friction Q
Friction Q
Cutting tool
Chip contact zone
VcRn
X
ZY
(a) PcVchip
Y
(b)
X
Z
Yp0
+a
-aap
Pc
For=0
Cylinder 2
Cylinder 1
Cylinder 2-chip
Cylinder 1-cutting tool
(c)
c
Workpiece
Fig. 4 Simplified contact model
at the tool-chip interface
400
450
500
550
600
650
0. 0 46 0. 0 914 0. 1 368 0. 1 822 0. 2 276 0. 2 73 0. 3 184 0. 3 638 0. 4 092 0. 4546 0.5
Cutting time
Cutting
force(N)
mesh 6mesh 8
mesh 12mesh 14
Average experimental force
(ms)
mmmm
Fig. 3 Cutting force sensitivity versus cutting time with different
mesh sizes
Int J Mater Form
-
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
6/13
bound estimation [17] of the yield shear stress can be
calculated by the von Mises criterion:
1
6s1 s2 2 s2 s3 2 s3 s1 2
n o t2Y s2Y
3
8
In which 1, 2 and 3 are the principal stresses, and tYand Y denote the yield stress values of the material in both
simple shear stress and tension, respectively.
However, when the CoF is relatively small, the maxi-
mum shear stress is found beneath the contact surface. If
this maximum value is adopted to define the limiting shear
0.35
0.375
0.4
0.425
0.45
0.475
0.5
0.525
0.55
0.575
0.6
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Maximum shear stress in contact zone
Maximum shear stress at contact surface
/
Y
CoF -
Fig. 6 Ratio of the limiting shear stress and Yversus CoF at a given
temperature
0.028704
0.057407
0.086111
0.11481
0.14
35
2
0.1
7222
0.20
093
0.20093
0.2296
3
0.22963
0.25833
0.28704
0 0.5 1 1.5 2
-2
-1.8
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0.05
0.1
0.15
0.2
0.25
z/a
x/a
p0
(at X = 0.0, Z = -0.72)
max /po = 0.31574
( /po)contact surface
contact zone
(a)
0.06010.091556
0.12301
0.15447
0.1
8592
0.21738
0.21738
0.24883
0.28029
0.28
029
0.31174
0.343
2
0 0.5 1 1.5 2
-2
-1.8
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0.1
0.15
0.2
0.25
0.3
z/a
x/a
p0
(at X = 0.6, Z = -0.45)
max /po = 0.37465
( /po)contact surface
contact zone
(b)
0.11267
0.149
13
0.18559
0.22205
0.22205
0.25851
0.29497
0.33143
0.36
789
0.40435
0.44081
0.44081
0 0.5 1 1.5 2
-2
-1.8
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0.15
0.2
0.25
0.3
0.35
0.4
z/a
x/a
p0
(at X = 0.69, Z = -0.27)
max /po = 0.47727
( /po)contact surface
contact zone
(c)
0.135
51
0.17466
0.2
1382
0.25298
0.25298
0.29213
0.29213
0.33129
0.37044
0.4096
0.44876
0.48791
0 0.5 1 1.5 2
-2
-1.8
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0.15
0.2
0.25
0.3
0.35
0.4
0.45
z/a
x/a
p0
(at X = 0.42, Z = 0.0)
max /po =0.5271
( /po)contact surface
contact zone
(d)
Fig. 5 Contour plot of shear stress in chip contact zone for different CoFs: a =0.0; b =0.2; c =0.4; d =0.48
0
100
200
300
400
500
600
0 100 200 300 400 500 600 700 800 900 1000
CoF= 0
CoF= 0.1
CoF= 0.15
CoF= 0.2
CoF= 0.3
CoF= 0.4
CoF= 0.48
Temperature (C)
f(MPa)
Fig. 7 Limiting shear stress versus temperature for different CoFs
Int J Mater Form
-
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
7/13
stress, the material beneath the contact surface will yield
before the surface material reaches the critical value. So, it
is necessary to redefine the critical limiting shear stress.
The latter is introduced in detail from the point of view of
contact mechanics by considering the relationship between
the CoF and temperature.
Figure 4(a) schematizes the tool and chip contact. The
tool is fixed and the workpiece moves at the cutting speed
Vc from left to right. To analyze the contact condition, a
small region at the interface was considered with a sliding
rate Vchip and a normal force Pc per length unit. More
precisely the contact model was simplified according toFig. 4(b) which consists of a flat surface slider moving from
left to right over a curved profile and with a steady velocity
Vchip. For frictionless elastic contact and when the contact
dimensions are small compared to the size of the contacting
bodies (Hertz assumptions), the solution is well described
by the Hertz theory [22]. The plane strain problem is
equivalent to the contact between two cylinders, see
Fig. 4(c), with a is the half contact width and p0 is the
maximum pressure at the contact center. It is important to
note that the contact pressure is distributed along a semi-
elliptical shape between [-a, a] along the x-axis.
The load at which material yield begins is related to theyield threshold of the softer material (here the workpiece) in
a simple tension or shear test through an appropriate yield
criterion [22].
Frictionless contact between two cylinders
According to Johnson [22], under a normal force Pc (per
length), the stress field distribution in the chip (here
schematized by cylinder2) is given by Eq. 9 below:
sx p p0a m1 1 z2n2
1
m21n2
1
2z
n osz p p0a m1 1
z2n21
m21n2
1
sy
p v sx p sz p
txz p p0a n1
m21z2
m21n2
1
m2
1
p 1
2
a2
x2
z2
4x2z2 0:5 a
2
x2
z2
h in21
p 1
2a2 x2 z2 4x2z2 0:5 a2 x2 z2 h i
8>>>>>>>>>>>>>>>>>>>:9
In the case of plane strain, the principal stresses can be
calculated by Eq. 10.
s1;2 sx p sz p2 ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
sx p sx p2
2 txz p
r
s3 sy
p v sx p sz p
:
8>: 10
Combining Eqs. 810, the shear stress distribution in the
(X, Z) plane shown in Fig. 5 (a) with a maximum shearstress of 0.31574 p0 at a depth z=0.72a. Note that for
frictionless contact, the maximum shear stress is found at a
location far below the surface whereas it reaches the surface
Chip valley
Considering crack
Chip peak
Free surface
New formed
free surface
Segmentation
wavelength
Lc=
h2=
(a) (b)
F ig . 8 C h i p m o r p h ol o g y
obtained at VC=120 m/min,
f=0.127 mm/rev: a computation
considering the scalar stiffness
degradation (SDEG = 0.74), and
b Experimental comparison with
[9]
VC=120 (m/min) Lc (m) Peak (m) Valley (m) Fc [5] (N)
Experiment 140 165 46.5 559
Simulation 133 161.5 48 541.3
Table 3 Comparison between
experimental and numerical
chip geometry
Int J Mater Form
-
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
8/13
when CoF is equal or higher than 0.48. Consequently, for
frictionless contact or when the CoF is lower than 0.48, it is
necessary to consider the shear stress at the contact surfaceas the limiting shear stress, not the maximum one which is
located at the hertzian depth.
Frictional contact between two cylinders
In presence of friction the contribution of the tangential force
Q, acting on each contact surface along the direction opposed
to the motion (see Fig. 4(b)) should be also considered:
sx q q0a n1 2 z2m2
m21n2
1
2x
n osz q q0p0 txz psy
q v sx q sz q
txz q q0p0 sx p:
8>>>>>>>>>:
11
Where q0=p0 is the tangential traction at x =0, and the
suffixes p and q refer to the stress components due to
normal pressure and tangential traction, respectively. Also,it is assumed that the tangential traction has no effect upon
the normal pressure distribution. When superimposed to the
effect of the contact pressure (normal effect) it yields:
sx sx q sx psy sy
q sy
p
sz sz q sz ptxz txz q txz p
8>>>:
12
Figure 5(b) presents the counter plots of the shear stress
in the contact zone between the chip and the tool for a
relatively low CoF, here 0.2. It is noticed that the maximumshear stress is still below the contact interface. This
maximum shear stress moves towards the interface when
Fig. 9 Comparaison between
the present simulations and
experiments [15] of chip mor-
phologies for f=0.1 mm/rev,
(Vc=60 m/min case of a and b)
and (Vc= 180 m/min case
of c and d)
Vc (m/min) Methods Lc (m) Peak h1 (m) Valley h2 (m) Fc (N)
60 Experimentation [15] 100 N/A N/A 557
Simulation 96 136 94 600
180 Experimentation [15] 100 131 62 548
Simulation 96 132 77 580
Table 4 Comparison between
experimental and numerical
chip geometry
Int J Mater Form
-
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
9/13
increasing the CoF as shown in Fig. 5(c) and (d). It reaches
the surface when 0.48. Whereas only one point is
concerned when =0.48 (see Fig. 5(d)).
Figure 6 presents the maximum shear stress conditions at
the surface and below the contact versus the CoF. Results
are normalized by the material yield stress Y. It can be seen
that the maximum shear stresses are totally different when
the CoF is less than 0.48. When the CoF is higher than
0.48, the maximum shear stress is located at the contact
interface and is equal to 1= ffiffiffi3p
sY.
In practical condition, the yield stress Y is updated
according to variations with temperature, strain rate, etc.
So, it is necessary to identify accurately the limiting shear
stress at the contact interface. According to Komanduri
[23], the yield stress of Ti-6Al-4V evolving with tempera-
ture, by taking into account the results given in Fig. 6, it is
possible to draw the limiting shear stresses versus temper-
ature for different CoFs as shown in Fig. 7. It can be
noticed that the higher the CoF, the higher the limiting
shear stress, whereas the latter decreases monotonously
when the temperature increases. Therefore, both CoF and
temperature influence the limiting shear stress. The as-
sumption of constant threshold shear stresses regardless the
frictional properties and temperature is no more valid.
It is important to underline that this figure is drawn only
for CoFs varying between 0 and 0.48. Indeed, for CoF
higher than 0.48, the limiting shear is found at the surface
and is equal to 1=ffiffiffi
3p
(see Fig. 6), and the limiting shear
stress is applied in the cutting model by user subroutine
VFRIC.
Results and discussions
Numerical model validation with experimental works
In order to validate the model developed in the present work,
the numerical results presented are compared with experi-
mental data mainly gathered from literatures [5, 9, 15]. In the
present model, the failure of material is based on Pirondiswork [24], as described in experiment and simulation, crack
starts to take place when the scalar stiffness degradation
(SDEG) is larger than 0.74, so this value can be taken as the
critical value for determining the material failure occurrence.
The simulation carried-out following Jiang and Shivpu-
ris [9] data gives the result shown in Fig. 8(a). It can be
seen that the periodic cracks initiate at the free surface of
the cutting material ahead of the tool and is propagated
towards the tool tip. This result agrees with the works of
Vyas and Shaw [25]. The chip valley is measured from the
boundary of failure zone (SDEG=0.74) to the new formed
free surface for considering crack.A comparison between results obtained by experimenta-
tion and simulation is shown in Table 3. It can be remarked
a good agreement between them. The average cutting force
(Fig. 3 with mesh size of 8 m) is nearly equal to
Umbrellos result [5] with the same cutting parameters.
Another comparison was made with Calamazs [15]
experiment (Fig. 9). In the case of a cutting speed of 60 m/
min, the chip segment morphologies as a result of the
periodic cracks initiation take place both at tool tip and the
free surface ahead of tool tip, where the failure zone at tool
tip is presented in the form of secondary shear zone [26]
(Fig. 9(a)), while the propagation of the failure zone at the
free surface only arrives at half of chip thickness arising
from the lower shear strain rates in the upper region of the
primary shear zone.
When the cutting speed increases to 180 m/min, the
failure zone (with the stiffness degradation larger than 0.74)
goes through the primary shear zone from two sides in and
out, as shown in Fig. 9(c), which is different form
Table 5 Simulation parameters used in the study
Feed rate, f(mm) 0.06, 0.1, 0.127
Cutting speed, Vc (m/min) 60, 120, 180
CoF, 0.3, 0.7, 1
Rake angle, 0=15, 6, 4 Clearance angle, 0=6, 12
Cutting width ap=2.54 mm Cutting edge radius Rn=30 m
Fig. 10 SDEG distribution in
chip according to VC variation
for=0.7 and f=0.127 mm/rev: a
VC=60 m/min; b VC=120 m/min;and c VC=180 m/min
Int J Mater Form
-
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
10/13
Luttervelt [27], who discovered the occurrence of a crack at
the tool tip and its propagation towards the free surface,
which is also different from Vyas and Shaw [25] as
mentioned above for rake angle is positive.
When the cutting speed increases, the stiffness degrada-
tion in the primary shear zone accelerates chip deformationin further step i.e. the chip valley will become smaller
(Table 4), while the chip wavelength and peak nearly dont
change. The cutting forces in these two conditions are
larger than the experimental value for the negative rake
angle, while the errors are less than 10%.
As a summary, it can be said that the new developed model
is in good corroboration with Shivpuris and Calamazs
experiments. It permits to predict the chip morphology and
the cutting force with a good approximation. Moreover, it
seems that the appropriate values of the limiting shear stress
are well expressed in this new proposed model.
It is also found that the chip morphology is different with
positive and negative rake angle cutting tool. Indeed, for
positive rake angle tool, the period cracks initiate at the free
surface of the machined material ahead of the tool and
propagate partway toward the tool tip. For negative rake
angle tool, the period cracks initiate at two free faces in and
out, go through the primary shear zone, and join in together
forming the shear band zone with high cutting speed.
Parametric study of Ti-6Al-4V cutting
With the proposed new model, a numerical parametric
study concerning chip morphology in the in orthogonal
cutting of Ti-6Al-4V was performed. The process simula-tion parameters are listed in Table 5.
Cutting speed effect on chip morphology
The SDEG distribution in the shear zone is shown in
Fig. 10 with increased cutting speed, hereinto the value
above 0.74 directly indicates the crack possibility of the
chip as mentioned in Numberical model validation with
experimental works, which is also a reason for segmented
chip, as described in [8], among other phenomena. When
cutting speed is equal to 60 m/min, the chip presents a
continuous part and a segmented. Nonetheless, the region withthe value of SDEG above 0.74 is just a small zone beneath the
free surface of the cutting material ahead of the tool tip. When
the cutting speed increases to 120 m/min, the segmentation
takes place on the entire chip, and part of the region with the
value of SDEG above 0.74 passes through half of the chip.
When the cutting speed increases to 180 m/min, the regions
with the value of SDEG above 0.74 nearly arrive at the free
surface around tool tip for all segmentations. So, the higher the
cutting speed the higher chip segmentation.
The segmentation frequency fc can be approximated by
the ratio VcLc, where Lc is chip segmentation wavelength as
it is indicated in Fig. 8(b) [28, 29]
fc VcLc
13
The computed chip segmentation wavelengths, for
cutting speeds 60, 120, and 180 m/min (as shown in
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
Cutting speed(m/min)
L
c(mm)
0
5
10
15
20
25
30
35
fc(kHz)
50 75 100 125 150 175 200
For Lc
For fc
Segmentation wavelength
Segmentation frequency
Fig. 11 Effect of VC on Lc and fc for f=0.127 mm, and =0.7
Fig. 12 Temperature distribution with increased for f=0.127 mm/rev, VC=120 m/min: a =0.3; b =0.7; and c =1
Int J Mater Form
-
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
11/13
Fig. 11), are constant, whereas the segmentation frequency
is directly proportional to the cutting speed according to
Eq. 13. This result agrees with the experiment study carried
out by Sun [28].
CoF effect on chip morphology
To study the influence of CoF for the cutting of Ti-6Al-4V,
friction values were taken 0.3, 0.7 and 1 for tool-chip
interfaces. The CoF for sacrificial passage zone (see
(Fig. 1), Part 3) is taken constant. As shown in Fig. 12,
the temperature distribution on chip and tool is greatly
influenced by the CoF variation. It is remarked that the
insert-zone where the temperature is of 581720C
migrates from tool-tip to the rake face-chip contact end,
as CoF varies from 0.3 to 1.
Moreover, it is noted the absence of friction effect oncutting force which has an average value around 540N, as it
is illustrated in Fig. 13. This result can be expected because
the global chip morphology is remained the same with the
CoF variation. Whereas, the feed force does not rise with
the increased CoF, by contrary the values decline when CoF
evolves from 0.3 to 1. This phenomenon mainly attributes
to the thermal softening effect of the high temperature
ahead of tool tip and rake face, and the limiting shear
stresses directly decrease as discussed in Fig. 7.
Rake angle effect on chip morphology
Two feed rates 0.06 and 0.1 mm/rev were checked to study
rake angle effect on chip morphology. According to Fig. 14
it can be underlined that when passing from positive to
0
100
200
300
400
500
600
0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95 1.05
Averageforce(N)
Cutting force
Feed force
CoF -
Fig. 13 Evolution of the average cutting and feed force versus for
f=0.127 mm/rev, and VC=120 m/min
Fig. 14 Chip SDEG distribu-
tion at VC=120 m/min and =
0.7: a g0=15, f=0.06 mm/rev;
b g0=6, f=0.06 mm/rev;
c g0=15, f=0.1 mm/rev; and
d g0=4, f=0.1 mm/rev
Int J Mater Form
-
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
12/13
negative angle the chip segmentation phenomenon becomes
more pronounced.
For the same tool-rake angle, the increase in feed rate
from 0.06 to 0.1 mm/rev implies more chip segmentation.
This can be observed when comparing Fig. 14(a) and (c).
So, the higher the feed rate the higher the material SDEG as
it is also shown by Fig. 14(b) and (d) (see the propagation
of SDEG between 0.74 and 1). It can be also underlinedthat more negative rake angle, higher is the tool-workpiece
contact pressure and so the chip segmentation.
Concluding remarks
The aim of this contribution concerns the a numerical finite
element cutting model for an aeronautic Titanium alloy
referenced as Ti-6Al-4V. The major findings are summa-
rized as follows:
The present work based on an energetic concept andrefined limiting shear stress to build a multi-part cutting
model under Abaqus\Explicit was presented. The
adopted approach can well predict the cutting force
and chip morphology, with the corresponding mis-
match less than 10 % compared with experiments. And
the cutting model can be used to simulate high
coefficient of friction (CoF) and low feed rate with
large cutting edge radius tool without convergence
problem. In order to obtain the appropriate model mesh
size for the present study, a mesh sensitivity analysis
was performed.
The limiting shear stress at the tool-chip contact surfaceis used as a friction model from the aspect of contact
mechanics. It is first shown that the surface shear stress
is linked to the contact pressure and the CoF. A
relationship between the CoF and the limiting shear
stress is given. The effect of the temperature on the
limiting shear stress is also considered.
The presentation of a parametric study is carried out
with the new cutting model. This allowed bring some
physical comprehension accompanying chip formation
according to the variation of cutting speed, CoF, and
rake angle. The simulation result show that the higher
the cutting speed, the more marked chip segmentation phenomenon, the distribution of material scalar stiff-
ness degradation on the chip seems to be located
closely to the tool surface.
A refined measure of the computed chip segmentation
wavelengths show that they are constant regarding to
cutting speed variation, whereas chip segmentation
frequency is directly proportional to it. Moreover, the chip
segmentation is also more pronounced when the insert
geometry passes from positive to negative rake angle.
As CoF varies among 0.3 and 1, the highest temper-
ature located in the insert zone migrates from tool-tip to
the rake face-chip contact end. This temperature
migration has implied a decrease in feed force whereas
the variation of CoF is inactive regarding cutting force
variation.
In the meaning time, the FE model is being tested forother materials in the objective to propose to the scientific
community a robust cutting model.
Acknowledgments The authors would like to acknowledge Prof.
Shivpuri and Prof. Calamaz for providing the experimental data in this
work, and especially for the financial support of China Scholarship
Council (CSC).
References
1. Arrazolaa PJ, Garaya A, Iriarte LM, Armendiaa M, Maryab S,
Matrec FL (2009) Machinability of titanium alloys (Ti6Al4V andTi555.3). J Mater Process Tech 209:22232230
2. Puerta Velasquez JD, Bolle B, Chevrier P, Geandier G, Tidu A
(2007) Metallurgical study on chips obtained by high speed
machining of a Ti-6Al-4V alloy. Mater Sci Eng A 452453:469
474
3. Ezugwu EO, Wang ZM (1997) Titanium alloys and their
machinabilitya review. J Mater Process Tech 68:262274
4. Hou ZB, Komanduri R (1995) On a thermo mechanical model of
shear instability in machining. CIRPAnnManuf Tech 44(1):6973
5. Umbrello D (2008) Finite element simulation of conventional and
high speed machining of Ti6Al4V alloy. J Mater Process Tech
196:7987
6. Childs THC (2006) Friction modelling in metal cutting. Wear 260
(3):310318
7. Bonnet C, Valiorgue F, Rech J, Hamdi H (2008) Identification of afriction model-Application to the context of dry cutting of an AISI
316L austenitic stainless steel with a TiN coated carbide tool. Int J
Mach Tools & Manuf 48:12111223
8. Mabrouki T, Rigal JF (2006) A contribution to a qualitative
understanding of thermo-mechanical effects during chip formation
in hard turning. J Mater Process Tech 176:214221
9. Jiang H, Shivpuri R (2004) Prediction of chip morphology and
segmentation during the machining of titanium alloys. J Mater
Process Tech 150:124133
10. Subbiah S, Melkote SN (2008) Effect of finite edge radius on
ductile fracture ahead of the cutting tool edge in micro-cutting of
Al2024T3. Mater Sci Eng A 474:283300
11. Zorev NN (1963) Inter-relationship between shear processes
occurring along tool face and shear plane in metal cutting. Int
Res Prod Eng ASME, pp 424912. Johnson GR, Cook WH (1983) A constitutive model and data for
metals subject to large strains, high strain rates and high
temperatures. Proc. of the 7th international symposium on
ballistics, the Hague, pp 3148
13. Lesuer DR (2000) Experiment investigations of material models for
Ti-6Al-4V Titanium and 2024-T3 Aluminum. Technical Report
14. http://www.matweb.com/. Accessed 25 August 2008
15. Calamaz M, Coupard D, Girot F (2008) A new material model for
2D numerical simulation of serrated chip formation when
machining titanium alloy Ti-6Al-4V. Int J Mach Tools Manuf
48:275288
Int J Mater Form
http://www.matweb.com/http://www.matweb.com/ -
8/8/2019 FE-Model for Titanium Alloy V Cutting Based
13/13
16. Hillerborg A, Modeer M, Petersson PE (1976) Analysis of crack
formation and crack growth in concrete by means of fracture
mechanics and finite elements. Cem Concr Res 6:773782
17. H.K.S, Abaqus/Explicit theory and user manuals, Version 6.8.2.
Accessed August 8, 2008
18. Haglund AJ, Kishawy HA, Rogers RJ (2008) An exploration of
friction models for the chip-tool interface using an Arbitrary
Lagrangian-Eulerian finite element model. Wear 265:452460
19. Shi GQ, Deng XM, Shet C (2002) A finite element study of the
effect of friction in orthogonal metal cutting. Fin Elem Anal Des38:863883
20. Liu CR, Guo YB (2000) Finite element analysis of the effect of
sequential cuts and tool-chip friction on residual stress in a
machined layer. Int J Mech Sci 42:10691086
21. Kishawy HA, Rogers RJ, Balihodzic N (2000) A numerical
investigation of the chip tool interface in orthogonal machining.
Int J Mech Sci 5:379414
22. Johnson KL (1985) Contact mechanics. Cambridge University
Press, Cambridge
23. Komanduri R, Hou ZB (2001) On thermoplastic shear instability
in the machining of a titanium alloy (Ti-6Al-4V). Metal Mater
trans 33:29953010
24. Pirondi A, Bonora N (2003) Modeling ductile damage under fully
reversed cycling. Comput Mater Sci 26:129141
25. Vyas A, Shaw MC (1999) Mechanics of sawtooth chip formation
in metal cutting. Trans ASME J Manuf Sci Eng 121:163172
26. Rech J, Calvez CL, Dessoly M (2004) A new approach for the
characterization of machinabilityapplication to steels for plastic
injection molds. J Mater Process Tech 152:667027. Luttervelt CAV, Pekelharing AJ (1977) The split shear zone
mechanism of chip segmentation. Ann CIRP 25(1):3337
28. Sun S, Brandt M, Dargusch MS (2009) Characteristics of cutting
forces and chip formation in machining of titanium alloys. Int J
Mach Tools Manuf 49(78):561568
29. Barry J, Byrne G (2001) Study on acoustic emission in machining
hardened steels, Part 1: acoustic emission during saw-tooth chip
formation. Proc Inst Mech Eng Part B: J Eng Manuf 215:1549
1559
Int J Mater Form