Vous êtes sur la page 1sur 9

ACME-211; No.

of Pages 9
archives of civil and mechanical engineering xxx (2014) xxxxxx

Available online at www.sciencedirect.com

ScienceDirect
journal homepage: http://www.elsevier.com/locate/acme

Original Research Article

Meshing strategies in FEM simulation of the


machining process
P. Niesony a,*, W. Grzesik a, R. Chudy a, W. Habrat b
a
b

Opole University of Technology, 76 Proszkowska Str., 45-758 Opole, Poland


Rzeszw University of Technology, 12 Powstacw Warszawy Str., 35-959 Rzeszw, Poland

article info

abstract

Article history:

This paper presents some methods of mesh smoothing when using cutting tool inserts with

Received 8 October 2013

complex geometry of the cutting edge and the rake face. Several sets of meshing parameters

Accepted 24 March 2014

are proposed and their inuences on the performance of FEM simulation of the cutting

Available online xxx

process are presented. In addition, both mechanical and thermal characteristics of the

Keywords:

carried out for a Ti6Al4V alloy using TiAlN coated carbide commercial cutting tool insert. It

Turning

was documented that accurate representation of the tool micro-geometry inuences the

Titanium alloy

simulation results.

cutting process are compared for four groups of meshing parameters. The simulations were

FEM simulation

# 2014 Politechnika Wrocawska. Published by Elsevier Urban & Partner Sp. z o.o. All
rights reserved.

Cutting forces
Heat generation

1.
Meshing problems in FEM modeling of
machining process
Modeling of technological processes is an important part of the
implementation of new technology into production process
and in the case of machining processes a dominant role is
played by FEM-based numerical simulation. Its wider and
successful application is now limited by inaccurate constitutive material models and, in particular by the lack of the
relevant input data [6,10]. These data include both geometrical
and physical properties of the machined workpiece and the
cutting tool insert used. The rst group includes geometrical,
predominantly 3D CAD models of these elements with
particular emphasis on the cutting edge radius [1]. The second
group aggregates the choice of constitutive laws and their

constitutive models which characterize the mechanical and


thermophysical properties of the workpiece and tool materials
[3,9,11].
In this paper the focus was made on the denition of
geometrical input data and, in particular, on the numerical
models of cutting tool inserts including active parts performing
the machining process. It is obviously known that in this area
the most important is the meshing procedure of a CAD model
generated. Its goal is to generate a nite solid element dened by
the nodal points represented by the nodes of basic nite
elements whose structure and localization should be adequate
to the simulated cutting process using FEM technique [2,5].
Simulation of the cutting process using cutting tools with
geometrically complex rake faces requires the boundary
minimum and maximum dimensions/scales of basic nite
elements to be dened in order to guarantee the correct

* Corresponding author. Tel.: +48 77 449 8460; fax: +48 77 449 8460.
E-mail address: p.nieslony@po.opole.pl (P. Niesony).
http://dx.doi.org/10.1016/j.acme.2014.03.009
1644-9665/# 2014 Politechnika Wrocawska. Published by Elsevier Urban & Partner Sp. z o.o. All rights reserved.

Please cite this article in press as: P. Niesony et al., Meshing strategies in FEM simulation of the machining process, Archives of Civil and
Mechanical Engineering (2014), http://dx.doi.org/10.1016/j.acme.2014.03.009

ACME-211; No. of Pages 9

archives of civil and mechanical engineering xxx (2014) xxxxxx

iteration/computation process. It is also known that in order to


obtain a CAD model with satisfactory accuracy both the
cutting edge micro geometry (its roundness with the radius rn)
and the grooved chip breaker should be modeled properly.
Taking this fact into consideration the modeling procedure
was essentially focused on the quantication of meshing of
tool wedge CAD model on the results and quality of 3DFEM
simulation of a turning process representing a non-orthogonal
model of the cutting process. The CAD model of a commercial
cutting tool insert was used in this investigation. A Ti6Al4V
titanium alloy, which is one of the most popular aerospace
materials, was selected as the workpiece material. The FEM
simulation undertaken in this study is strongly oriented on the
optimization of technological processes of such difcult-tomachine materials.

2.

Investigation methodology

In the FEM simulations carried out the special focus was on the
accurate cutting tool model as well as adequate constitutive
material model. The tool wedge model was implemented as
the CAD model of a cutting tool insert applied (in the case
study it was CNMG 120412-UP insert produced by Kennametal). The visualization of the complete CAD model implemented is shown in Fig. 1a, whereas the magnied fragment of the
corner with the cutting edge and grooved chip breaker is
depicted in Fig. 1b.
In order to compare the 3D CAD model with the real
conguration of the cutting tool insert used, the insert shape
was measured using an ALICONA IFM optical microscope.
Exemplary dimensioned prole of the cutting wedge of CNMG
120412-UP insert including the cutting edge (detail A) and the
groove (chip breaker) on the rake face is shown in Fig. 2. The
optical image of the insert generated is presented in Fig. 3a. In
addition, Fig. 3b illustrates the localization of cross-sections 1
3 in which the wedge proles were separated and magnied
using MountainsMap and Geomagic Design software.
The comparison of the model and real shape of the insert
gives good shape and dimensional agreement. In particular,
dimensional errors were not higher than 7 mm. It should also
be noted that the thickness of deposited TiAlN coating was
about 3 mm which also inuences the model accuracy.
In these investigations FEM simulations were performed for
standard constitutive model named Power Law (PL). It should
be noted that the ow stress in the cutting zone is predicted
using equations incorporating the material behavior under
high strain rate and temperature [7]. As a result, the
constitutive PL model is mathematically expressed by the
power equation, as follows
ep
s f e p s o QT 1 0
ep

!1=n
(1)

where so is the initial yield stress, ep is the plastic strain, e0p is


the reference plastic strain, 1/n is the strain hardening
exponent and Q(T) is thermal softening index (factor) dened
as a function of temperature according to (2). In Eq. (2) the c0
through c5 are coefcients for the polynomial t, T is the
temperature, Tcut is the linear cut off temperature, and Tmelt is

Fig. 1 CAD model of CNMG 120412-UP cutting tool insert by


(a) Kennametal and (b) tool rake model used in FEM
simulation.

the melting temperature. The Eq. (2a) is dened for T < Tcut,
where Eq. (2b) for T  Tcut.
QT c0 c1 T c2 T2 c3 T3 c4 T4 c5 T5

(2a)



T  Tcut
QT QTcut 1 
Tmelt  Tcut

(2b)

In a typical machining event, extremely high strain rates in


excess of 104 s1 may be attained within the primary and
secondary shear zones, while the remainder of the workpiece
deforms at moderate or low strain rates. In order to account for
a variation in the strain rate sensitivity at low and high strain
rates, AdvantEdge software incorporates a stepwise variation

Please cite this article in press as: P. Niesony et al., Meshing strategies in FEM simulation of the machining process, Archives of Civil and
Mechanical Engineering (2014), http://dx.doi.org/10.1016/j.acme.2014.03.009

ACME-211; No. of Pages 9


archives of civil and mechanical engineering xxx (2014) xxxxxx

Fig. 2 Dimensioned profile of the cutting edge for CNMG 120412-UP insert based on the 3DCAD model.

of the rate sensitivity exponent m while maintaining continuity of stress.


s s f e p 1

e_ p
e_ 0p

!1=m2
1

e_ t
e_ 0p

!1=m1
(3)

Fig. 3 (a) An optical image of CNMG 120412-UP generated


by means of ALICONA IFM and (b) localizations of crosssections illustrated by means of MountainsMap software.

where s is the effective von Mises stress, e_ 0p is a reference


plastic strain rate, m1 and m2 are low and high strain-rate
sensitivity exponents, respectively, and e_ t is the threshold
strain rate which separates the two regimes. The friction
coefcient between the tool and workpiece corresponding to
Coulomb's friction was equal to 0.5.
The parameters for the Power Law model used in these
simulations for the Ti6Al4V titanium alloy are specied in
Table 1. It was not possible of compare these values with those
proposed by Sima [13] and Calamaz [4] because they are for the
JohnsonCook model not available for users in AdvantEdge
package.
One of the important steps in performing FEM simulation is
the denition behind technological parameters, workpiece
and cutting tool material grades, machining operation and
cutting tool conguration as well as meshing criteria for solid
models of the workpiece and the tool insert used.
The accuracy of a FEM simulation is directly related to the
quality of the discretization used. In this sense, structured
meshes are still preferred in a wide range of simulations where
a strict alignment of elements is required by the analysis [8,12].
In this study special attention was made on the cutting tool
model and the workpiece was meshed using a commercial
meshing procedure.
All simulations were performed using commercial FEM
package AdvantEdge FEM [14]. As a result, the meshing
procedure includes a set of boundary, dimensional and
parametric criteria, as well as adaptive criteria.
Several groups of meshing parameters are selected and
described in Table 2.
FEM simulations were performed for the four different
groups of meshing parameters selected in Table 2 for turning
operations of a Ti6Al4V alloy using carbide tools with the
cutting edge radius of rn = 10 mm coated with TiAlN monolayer.
Group A was selected as a reference group in which all
meshing parameters were selected automatically by the
AdvantEdge package. As a result, it was not possible to
decrease the cutting edge radius to be lower than the
minimum meshing size. Three other groups B, C and D
were created by modication of the mesh size and quantitative
parameters. In particular, the minimum mesh size was

Please cite this article in press as: P. Niesony et al., Meshing strategies in FEM simulation of the machining process, Archives of Civil and
Mechanical Engineering (2014), http://dx.doi.org/10.1016/j.acme.2014.03.009

ACME-211; No. of Pages 9

archives of civil and mechanical engineering xxx (2014) xxxxxx

Table 1 Power Law (PL) material model parameters for Ti6Al4V.


Code
PL

s0, MPa
952

Q(T)

e0p

ecut

22.19

0.035

0.12

c0 = 1.822
c1 = 0.00571
c2 = 1.7  105

c3 = 2.164  108
c4 = 6.48  1012
1655 8C
110 GPa
0.3
9.4  106 1/K

Melting temperature
Young modulus
Poisson ratio
Coefcient of thermal
expansion
Density

4430 kg/m3

decreased from 0.01 to 0.005 mm, mesh grading was changed


in such a way that mesh density was decreased from 0.5 to 0.1
and the number of segments per edge unit was increased from
1 to 5 (group C). In the last case, the mesh nose was performed
only within the cutting edge in contrast to group D for which
the mesh renement (concentration) covers the entire volume
of the meshed model. The visualization effects of the cutting
edge region of the tool model resulting from meshing using
different mesh smoothing procedures (meshing parameters) is
presented in Fig. 4.
In Fig. 4, rectangles of 0.277 mm  0.018 mm within which
numbers of nite element edges and nodes were computed for
different meshing groups are marked in red. It can be observed
that the change of the minimum mesh size for group B does
not inuence the number of edges and nodes in relation to the
reference meshing group A. In these cases the meshing area
includes 50 edges and 22 nodes (see Fig. 4a and b). On the other
hand, the representation of the cutting edge for meshing group
B is more accurate, as indicated in Fig. 5.
The change of the number of segments per edge unit for
group c results in increasing the number of nodes (from 22 to
27) and, as a result, the edges of nite elements from 50 to 60.
The most important factor which determines the number of
nodes is mesh grading. For this variant, the meshing area
includes 209 nodes and 81 edges (Fig. 4d).
It is evident from Fig. 4 that the meshing smoothing/
coarsening is reected visibly in the representation of the
crucial/key fragments of cutting tool model, predominantly
the rounded cutting edge and complex rake angle with grooved
chip breaker.

Table 2 Groups of meshing parameters of tool model


tested in FEM simulations.
Group
A
Maximum tool element
size, mm
Minimum tool element
size, mm
Minimum edge length,
mm
Mesh grading [range:
0.11]
Curvature safety [range:
0.25]
Segments per edge [range:
0.25]

Group
B

Group
C

Group
D

0.5

0.5

0.5

0.5

0.01

0.005

0.005

0.005

0.001

0.001

0.001

0.001

0.5

0.5

0.5

0.1

Fig. 4 FEM meshing of the cutting edge area of cutting tool


model using different meshing parameters selected in
Table 2: (a) group A, (b) group B, (c) group C, (d) group D.

Please cite this article in press as: P. Niesony et al., Meshing strategies in FEM simulation of the machining process, Archives of Civil and
Mechanical Engineering (2014), http://dx.doi.org/10.1016/j.acme.2014.03.009

ACME-211; No. of Pages 9


archives of civil and mechanical engineering xxx (2014) xxxxxx

Fig. 6 Average values of measured cutting forces and FEM


predictions for different meshing parameters including
standard deviations.

Fig. 5 Comparison of the cutting edge representations for


meshing using parameters A and B generated in the (a)
vicinity of tool corner and (b) straight part of the cutting
edge.

Fig. 4a illustrates the case when the representation of the


cutting edge according to the meshing variant A causes that its
radius is about 0, i.e. sharp cutting edge is generated. Better
results were obtained for cases B, C and D but they result in

Table 3 Effects of different meshing schemes for


relevant groups of meshing parameters.

SAT le size (CAD


model), kB
STL le size (model
after meshing), kB
Number of edges STL
model
Number of triangles
(STL model coverage)
Average distance of
nodes close to the
edge, mm
Average time of the
3D simulation of
the 30 mm cutting
path

Group
A

Group
B

Group
C

Group
D

9116

20582

20,958

45,925

4196

8874

9067

19,874

22,383

47,328

48,360

105,999

14,922

31,552

32,240

70,666

0.009

0.005

0.005

0.005

91,240 s

66,254 s

75,936 s

518,400 s

Fig. 7 Spectra of Fx component vs. time for (a) meshing


cases A and (b) B.

Please cite this article in press as: P. Niesony et al., Meshing strategies in FEM simulation of the machining process, Archives of Civil and
Mechanical Engineering (2014), http://dx.doi.org/10.1016/j.acme.2014.03.009

ACME-211; No. of Pages 9

archives of civil and mechanical engineering xxx (2014) xxxxxx

substantial increase of STL les, which, in turn, causes that the


computation time also increases. Especially, for group D the
computation time increases about four times in comparison
for a standard FEM procedure. These effects are summarized in
Table 3.
It is obviously known that the cutting edge micro-geometry
substantially inuences the decohesion of the workpiece
material and its accurate model seems to be very important in
process simulation. It is observed based on the effects of the
meshing of the cutting edge area that the worst result was
obtained for meshing parameters selected in group A. On the
other hand, the quality of cutting edge representation
improves when using other meshing cases, i.e. group B, C
and D from Table 2. In particular, the maximum difference of
the model dimensions in the cross-section perpendicular to
cutting edge was equal to 1.027 mm and 1.046 mm in the corner
and straight part of the cutting insert, as shown in Fig. 5.

3.

Results of FEM simulations

First of all, it was assumed that numerical 3D representation of


the tool wedge inuences mechanical and thermal characteristics of the cutting process of a Ti6Al4V alloy. Simulations
were carried out with various meshing parameters (all groups
selected in Table 2) using constant values of machining
parameters cutting speed vc = 90 m/min, feed rate f = 0.1 mm/
rev and depth of cut ap = 2 mm).

It was revealed that simulation time was equal to 18 and


21 h when meshing cases B and C were selected. On the other
hand, for meshing parameters recommended for users in
AdvantEdge package the computation time increases to 25 h
as indicated in Table 3. Moreover, the longer computation time
of about 144 h is required for meshing variant D. It should be
noticed that the computation time depends on the number of
meshing nodes; for the case D meshing density was two times
higher than for other groups. It is clear that such a long time of
simulation will not be recommended for practical applications. In contrast, the resolution of the simulation images is
excellent.
The mechanical characteristics of the turning process were
assessed based on the components of the resultant cutting
force. Their average values corresponding to the four meshing
cases and experimental tests are presented in Fig. 6.
It is shown in Fig. 6 that meshing parameters inuence the
values of the forces simulated. In general, the highest forces
were generated for meshing case A for which the numerical
representation of the cutting edge was not accurate. As a
result, the comparable values of Fy and Fz components were
obtained for other meshing cases and the difference between
them and basic meshing A are in the range of 1117%. This fact
can be easily observed in Fig. 6. Their changes are characterized by means of standard deviations as shown in Fig. 7. It can
be seen in Fig. 7 that standard deviation 3s is distinctly
narrower for simulations with meshing parameters B and the
simulation process is more stable.

Fig. 8 Distributions of heat flux in the cutting zone in 3D arrangements for relevant meshing groups.
Please cite this article in press as: P. Niesony et al., Meshing strategies in FEM simulation of the machining process, Archives of Civil and
Mechanical Engineering (2014), http://dx.doi.org/10.1016/j.acme.2014.03.009

ACME-211; No. of Pages 9


archives of civil and mechanical engineering xxx (2014) xxxxxx

Fig. 9 A scheme of data export using the line equidistant to


the cutting edge.

Concerning the experimental data one observe that they


differ from FEM simulation results. It can be seen in Fig. 6
that the measured values of the Fy and Fz forces are only
slightly higher than predicted values. On the other hand, the
Fx force values were lower than measured ones. In addition,
standard deviations for measurements are very small. It can
be assumed on the base of Fig. 6 that the inappropriate
meshing of cutting tool model causes visible uctuations of

the cutting force signals and this fact seems to be important


in improving the compliance of simulation and experiment
results.
The inuence of the meshing of cutting tool model on the
performance of the cutting process related to the cutting zone
is presented in Fig. 8. Some advanced functions allowing in the
FEM package used were used to make the cross-section across
the cutting zone and simulate the process characteristics for
all meshing groups selected.
It was proven that differences in model meshing inuence
the heat transfer in the cutting zone due to different shapes of
heat uxes. Fundamental difference causes the concentration
of the maximum heat uxes. For meshing scheme A this
concentration is localized in the workpiece, in the vicinity of
the cutting edge. On the other hand, for cases B and C the
generated heat is mainly concentrated in the primary shear
zone and its intensity increases for C meshing model. A similar
heat distribution was obtained for case D but it seems to be the
rst stage of the material removal process.
Similar trends were revealed for temperature distribution, stresses and strain rates in the cutting zone. It is
possible to visualize in 3D space the distribution of the heat
ux in the material decohesion zone adjacent to the active
part of the cutting edge. The rule used in export data line is
shown in Fig. 9. The nal effect of heat ux distribution
obtained for relevant groups of meshing parameters is
presented in Fig. 10. In order to visualize the thermal
behavior of the decohesion zone, the heat ux distribution is
presented in the plane containing the cutting edge itself. It

Fig. 10 Distributions of heat flux in the plane containing cutting edge for different meshing schemes.
Please cite this article in press as: P. Niesony et al., Meshing strategies in FEM simulation of the machining process, Archives of Civil and
Mechanical Engineering (2014), http://dx.doi.org/10.1016/j.acme.2014.03.009

ACME-211; No. of Pages 9

archives of civil and mechanical engineering xxx (2014) xxxxxx

can be observed in Fig. 10 that the maximum value of the


heat ux corresponds to group C. This effect results from the
fact that for this meshing case the segments per edge was
selected to be equal 5, i.e. the maximum possible value (see
Table 3).
As the next step to make quantitative analyses of the cutting
edge segmentation, special numerical procedure including data
export along the line equidistant to the cutting edge within the
material decohesion zone was performed, as shown in Fig. 9.
The nal result regarding both the heat ux and temperature
distributions is shown in Fig. 11.
It was revealed that the most important factor inuencing
the heat ux value is the node density (number of segments
per edge) which was specied as maximum in group C. For this
case of meshing the heat ux predicted was four times higher
than for other groups of meshing parameters. It should also be
noted that the cumulated heat is just before its ow to the tool
body. On the other hand, the average temperature distribution
(computed form the cloud of points shown in Fig. 11b) was

practically the same but with visible oscillations resulting from


extrapolating temperature data from the central part of
decohesion region.
In can be stated, based on FEM simulations performed,
that meshing of the cutting tool model is crucial for
estimation of heat ux, temperature and stresses in the
material decohesion zone. As a result, such analyses should
be carried out as an integral part of FEM based simulations
of machining processes using cutting tools with dened
geometries.

4.

Summary

It can be concluded that CAD model meshing inuences the


results of FEM simulation related to both mechanical and
thermal behavior of the cutting process. In addition, it
inuences computation time which is specially important for
3D process representation. Some conclusions are as follows:
- During meshing of CAD model of the cutting tool insert
special focus should be put on the appropriate representation of the cutting edge and curvilinear rake face with
grooved chip breaker.
- The quality of the representation of all critical areas of the
model was improved by minimizing tool element size.
- Acceptable representation of the tool geometry was
achieved by STL model with increased meshing density
(nodal elements) related to the cutting edge unit.
- It is not necessary to increase the meshing density in the
entire volume of the tool model.
- Modication of the tool model meshing inuences the force
signals and for a certain set of meshing parameters it is
possible to reduce force oscillations.
- Meshing of the cutting tool model seems to a decisive factor
for the simulation of the heat ux, temperature and stresses
in the material decohesion/deformation zone.

Acknowledgment
This study was supported by Polish National Center of
Research and Development (NCBiR)-Project No. PBS1-178595.

references

Fig. 11 Distributions of (a) heat flux and (b) average


temperature in the decohesion zone along the active part of
the cutting edge.

[1] I. Al-Zkeri, J. Rech, T. Altan, H. Hamdi, F. Valiorgue,


Optimization of the cutting edge geometry of coated carbide
tools in dry turning of steels using a nite element analysis,
Machining Science and Technology 13 (1) (2009) 3651.
[2] R. Ambati, H. Yuan, FEM mesh-dependence in cutting process
simulations,
International
Journal
of
Advanced
Manufacturing Technology 54 (14) (2011) 313323.
[3] M. Barge, H. Hamdi, J. Rech, J.-M. Bergheau, Numerical
modelling of orthogonal cutting: inuence of numerical
parameters, Journal of Materials Processing Technology
164165 (2005) 11481153.
[4] M. Calamaz, D. Coupard, F. Girot, A new material model for
2D numerical simulation of serrated chip formation when

Please cite this article in press as: P. Niesony et al., Meshing strategies in FEM simulation of the machining process, Archives of Civil and
Mechanical Engineering (2014), http://dx.doi.org/10.1016/j.acme.2014.03.009

ACME-211; No. of Pages 9


archives of civil and mechanical engineering xxx (2014) xxxxxx

[5]
[6]

[7]
[8]

[9]

machining titanium alloy Ti6Al4V, International Journal of


Machine Tools and Manufacture 48 (34) (2008) 275288.
T. Coupez, H. Digonnet, R. Ducloux, Parallel meshing and
remeshing, Applied Mathematical Modelling 25 (2000) 153175.
P. Eberhard, T. Gaugele, Simulation of cutting processes using
mesh-free Lagrangian particle methods, Computational
Mechanics 51 (3) (2013) 261278.
Z. Gronostajski, The constitutive equations for FEM analysis,
Journal of Materials Processing Technology 106 (2000) 4044.
M.N.A. Nasr, E.-G. Ng, M.A. Elbestawi, A modied timeefcient FE approach for predicting machining-induced
residual stresses, Finite Elements in Analysis and Design 44
(2008) 149161.
FEM-based modelling of the inuence of thermophysical
properties of work and cutting tool materials on the process
performance, 14th CIRP Conference on Modeling of
Machining Operations, Procedia CIRP, vol. 82013, 38.

[10] Sensitivity analysis of the constitutive models in FEM-based


simulation of the cutting process, Journal of Machine
Engineering 13 (1) (2013) 106116.
[11] Modelling of Heat Transfer and Temperature Distribution in
the Cutting Zone for Cutting Tools with Hard Protective
Coatings,
Publishing House of Opole University of
Technology, Opole, 2008 (in Polish).
[12] E. Ruiz-Girones, J. Sarrate, Generation of structured
hexahedral meshes in volumes with holes, Finite Elements
in Analysis and Design 46 (2010) 792804.
[13] M. Sima, T. Ozel, Modied material constitutive models for
serrated chip formation simulations and experimental
validation in machining of titanium alloy Ti6Al4V,
International Journal of Machine Tools and Manufacture 50
(2010) 943960.
[14] Third Wave AdvantEdge User's Manual, 2011, Version 5.8,
Minneapolis, USA.

Please cite this article in press as: P. Niesony et al., Meshing strategies in FEM simulation of the machining process, Archives of Civil and
Mechanical Engineering (2014), http://dx.doi.org/10.1016/j.acme.2014.03.009