Lijing Xie

Forschungsberichte aus dem
wbk Institut für Produktionstechnik
Universität Karlsruhe (TH)
Estimation Of Two-dimension Tool Wear
Based On Finite Element Method
Forschungsberichte aus dem
wbk Institut für Produktionstechnik
Universität Karlsruhe (TH)
Hrsg.: Prof. Dr.-Ing. Jürgen Fleischer
Prof. Dr.-Ing. Hartmut Weule
Lijing Xie
Estimation Of Two-dimension Tool Wear
Based On Finite Element Method
ISSN 0724-4967
Band 120
© wbk Institut für Produktionstechnik
Universität Karlsruhe (TH)

alle Rechte vorbehalten

Druck: Schnelldruck Ernst Grässer, Karlsruhe
Tel: 0721/61 50 50

ISSN 0724-4967



Vorwort des Herausgebers
Der rasche Fortschritt der Produktionstechnik und der weltweite Wettbewerb um
technisch-wirtschaftliche Spitzenpositionen machen einen intensiven Austausch von
Wissen und Erfahrung zwischen Universitäten und der Industrie erforderlich. In die-
sem Sinne soll im Rahmen dieser Schriftenreihe in zwangloser Folge über aktuelle
Forschungsergebnisse des Instituts für Werkzeugmaschinen und Betriebstechnik der
Universität Karlsruhe berichtet werden.
Die Forschungsaktivitäten des Instituts umfassen neben der Untersuchung und Opti-
mierung von Bearbeitungsverfahren, Maschinenkomponenten und Fer-
tigungseinrichtungen insbesondere Aufgabenstellungen, die durch Nutzung
informationsverarbeitender Systeme eine Verbesserung der Leistungsfähigkeit
fertigungstechnischer Einrichtungen und deren informationstechnisch-
organisatorische Einbindung in automatisierte Produktionssysteme ermöglichen.

Prof. Dr.-Ing. Jürgen Fleischer Prof. Dr.-Ing. Hartmut Weule

Estimation Of Two-dimension Tool Wear Based on Finite
Element Method




Zur Erlangung des akademischen Grades eines
Doktors der Ingenieurwissenschaften
von der Fakultät für Maschinenbau
der Universität Karlsruhe (TH)
genehmigte

Dissertation

von
M. Sc. Lijing Xie
aus China

Tag der mündlichen Prüfung: 05. 02. 2004
Hauptreferent: Prof. Dr.-Ing. Jürgen Schmidt
Korreferent: o. Prof. Dr.-Ing. Dieter Spath
Acknowledgement



The present research work was carried out at Institut für Produktionstechnik (WBK) in
University of Karlsruhe (TH) since Nov. 2000. The last three years has been a
precious experience for me, with excellent learning, intense research work and
interesting activities. I feel very fortunate to have an opportunity to concentrate on the
interesting research field of manufacturing industry and get to know so intelligent,
friendly, and active persons. In this period, I get uncountable unselfish help from
them.

I would like to express my thanks to Prof. Dr.-Ing. Hartmut Weule and o. Prof. Dr.-Ing.
Jürgen Fleischer for their kindly concern in my living and work.

Especially, I want to express my deepest appreciation and thanks to my supervisors,
o. Prof. Dr.-Ing. Jürgen Schmidt and o. Prof. Dr.-Ing. Dieter Spath, for their support,
their careful reviews of my papers and dissertation, and their highly appreciated
instruction.

Thanks to Prof. Siqin Pang and Prof. Xibin Wang for their constantly encouragement
and help.

This thesis is finished under the cooperation with scientists in Institut für Werkstoffe I,
special thanks are given to Dipl.-Ing. Frank Biesinger for kindly offering the
developed material subroutine.

Thanks to all the members in group FT, I was touched by their friendship. Especially,
Mr. Dr.-Ing. Jörg Söhner, Mr. Dipl.-Ing. Carsten Schmidt and Mr. M. Sc. Anurag Jain
for the helpful suggestion and discussion in the research and help in personal living.

Thanks to all the members in the institute for the unforgettable happy time in the past
three years, especially Dr.-Ing. Ivan Tzitzelkov for solving many problems in my
simulation work, Mr. Michael Heinz for the warm-hearted assist and patient
instruction in my experiment work, Mr. Klaus Simon for offering instruction about
measuring basic knowledge and helping me to look for the best measuring method,
Mr. Thomas Hildenbrand for preparing experimental condition and troubleshooting in
the turning experiment. Thanks to Mrs. Margarethe Schüßler for teaching me
Deutsch language voluntarily.

At last, I want to thanks my husband, Dan, and my family and Dan’s for their love and
support. They give me the strength over all the problems in my research.







Karlsruhe, in December 2003 Lijing Xie
Table of Contents I
Table of Contents

Chapter 1 Introduction.............................................................................. 1
1.1 State Of Art: Finite Element Simulation Of Cutting Process.................................. 3
1.1.1 Numerical Aspects................................................................................................. 5
1.1.1.1 Approach ......................................................................................................... 5
1.1.1.2 Mesh Adaptivity .............................................................................................. 6
1.1.2 Mechanical Aspects .............................................................................................. 8
1.1.2.1 Contact And Friction ...................................................................................... 8
1.1.2.2 Material Constitutive Model ........................................................................ 11
1.1.2.3 Chip Separation............................................................................................ 13
1.2 Technical Background About Tool Wear................................................................. 16
1.2.1 Wear Types In Metal Cutting ............................................................................. 17
1.2.2 Wear Mechanism................................................................................................. 18
1.2.3 Tool Wear Model ................................................................................................. 19
1.3 Research Of Tool Wear With Finite Element Methods......................................... 22
1.3.1 Comparison Between FEM Method And Empirical Method ......................... 22
1.3.2 State Of Art: Numerical Implementation Of Tool Wear Estimation.............. 24
1.3.2.1 Tool Wear Estimation With The Combination Of Analytical Method And
FDM............................................................................................................................. 24
1.3.2.2 Tool Wear Estimation With FEM................................................................ 27
1.3.2.3 Summary Of Literature................................................................................ 29

Chapter 2 Objective And Approach........................................................ 31
2.1 Objectives..................................................................................................................... 31
2.2 Approach...................................................................................................................... 32

Chapter 3 Chip Formation Simulation Technology................................. 34
3.1 Introduction .................................................................................................................. 34
3.1.1 Explicit Algorithm In Chip Formation Simulation............................................. 34
3.1.1.1 Dynamic Analysis Procedure ..................................................................... 34
3.1.1.2 Thermal Analysis Procedure ...................................................................... 35
3.1.2 Stability Limit ........................................................................................................ 36
3.2 Continuous Chip Formation Simulation................................................................... 37
Table of Contents II
3.2.1 Limitation Of The Existing Chip Formation Models........................................ 37
3.2.2 Advantages Of The New-developed Chip Formation Model ........................ 39
3.2.3 Adaptive Meshing Technique In ABAQUS/Explicit ........................................ 40
3.2.3.1 Boundary Region Types.............................................................................. 40
3.2.3.2 Geometry Features...................................................................................... 41
3.2.3.3 Curvature Refinement ................................................................................. 41
3.2.4 Analysis Steps...................................................................................................... 42
3.2.4.1 Initial Chip Formation................................................................................... 43
3.2.4.2 Chip Growth .................................................................................................. 45
3.2.4.3 Continuous Steady-state Chip Formation ................................................ 45
3.2.5 Results & Discussion .......................................................................................... 48
3.2.5.1 Stress Analysis ............................................................................................. 48
3.2.5.2 Plastic Strain Analysis ................................................................................. 49
3.2.5.3 Strain Rate..................................................................................................... 50
3.2.5.4 Temperature Analysis.................................................................................. 51
3.2.5.5 Verification With Experimental Data.......................................................... 53
3.3 Chip Formation Simulation For Milling Operation.................................................. 54
3.3.1 Chip Separation ................................................................................................... 55
3.3.1.1 Shear Failure Criterion ................................................................................ 55
3.3.1.2 A Numerical Method To Determine Strain At Failure............................. 56
3.3.2 Chip Formation Modeling ................................................................................... 58
3.3.3 Result & Discussion ............................................................................................ 59
3.3.3.1 Stress Analysis ............................................................................................. 59
3.3.3.2 Cutting Temperature.................................................................................... 61
3.3.3.3 Cutting Force Analysis................................................................................. 62
3.4 Summaries & Conclusion .......................................................................................... 63

Chapter 4 Heat Transfer Analysis In Metal Cutting ................................ 64
4.1 Introduction .................................................................................................................. 64
4.2 General Considerations ............................................................................................. 64
4.2.1 Geometry And Mesh ........................................................................................... 64
4.2.2 Heat Flux............................................................................................................... 65
4.3 In Turning Operation................................................................................................... 66
4.3.1 Modelling............................................................................................................... 66
Table of Contents III
4.3.2 Results & Discussion .......................................................................................... 68
4.4 In Milling Operation..................................................................................................... 70
4.4.1 On Workpiece....................................................................................................... 70
4.4.1.1 Modelling ....................................................................................................... 70
4.4.1.2 Results & Discussion................................................................................... 72
4.4.2 On Tool.................................................................................................................. 73
4.4.2.1 Modelling ....................................................................................................... 73
4.4.2.2 Results & Discussion................................................................................... 75
4.4.2.3 Application Of Preheated Cutting Tool ..................................................... 77
4.5 Summaries & Conclusion .......................................................................................... 81

Chapter 5 Estimation Of Tool Wear In Turning Operation...................... 82
5.1 Introduction .................................................................................................................. 82
5.2 Tool Wear Calculation Program Design.................................................................. 82
5.3 Modelling Procedure................................................................................................... 83
5.3.1 Chip Formation And Heat Transfer Analysis................................................... 84
5.3.1.1 Normal Pressure........................................................................................... 84
5.3.1.2 Sliding Velocity ............................................................................................. 84
5.3.1.3 Tool Temperature......................................................................................... 86
5.3.2 Wear Rate Calculation........................................................................................ 86
5.3.3 Nodal Move Direction.......................................................................................... 86
5.3.3.1 Dividing Node................................................................................................ 87
5.3.3.2 On Rake Face............................................................................................... 87
5.3.3.3 On Flank Face .............................................................................................. 88
5.3.4 Cutting Time Increment Calculation ................................................................. 89
5.3.4.1 Flank Wear Calculation Subroutine........................................................... 90
5.3.4.2 Cutting Time Increment Searching Procedure ........................................ 90
5.3.5 Nodal Displacement ............................................................................................ 91
5.3.6 Tool Geometry Updating .................................................................................... 92
5.3.6.1 Step 1: Initial Tool Wear Profile ................................................................. 92
5.3.6.2 Step 2: Adjustment....................................................................................... 93
5.4 Results & Discussion.................................................................................................. 94
5.4.1 Tool Wear ............................................................................................................. 94
5.5 Summaries & Conclusion .......................................................................................... 97
Table of Contents IV

Chapter 6 Estimation Of Tool Wear In Milling Operation........................ 98
6.1 Introduction .................................................................................................................. 98
6.2 Tool Wear Calculation Program Design.................................................................. 99
6.3 Modelling Procedure................................................................................................. 101
6.3.1 Chip Formation Analysis................................................................................... 101
6.3.2 Heat Transfer Analysis ..................................................................................... 103
6.3.3 Nodal Average Wear Rate Calculation .......................................................... 103
6.3.3.1 Discussion About The Calculation Method Of Nodal Average Wear
Rate ........................................................................................................................... 103
6.3.3.2 Classification Of Workpiece Node........................................................... 105
6.3.4 Nodal Move Direction........................................................................................ 107
6.3.4.1 Dividing Node.............................................................................................. 107
6.3.4.2 On Rake Face............................................................................................. 108
6.3.4.3 On Flank Face ............................................................................................ 108
6.3.5 Cutting Time Increment Calculation ............................................................... 108
6.3.5.1 Flank Wear Calculation Subroutine......................................................... 108
6.3.5.2 Cutting Time Increment Searching Procedure ...................................... 109
6.3.6 Tool Geometry Updating .................................................................................. 111
6.4 Results & Discussion................................................................................................ 111
6.5 Summaries & Conclusion ........................................................................................ 113

Chapter 7 Summary And Prospect....................................................... 114
7.1 Summaries................................................................................................................. 114
7.2 Prospect ..................................................................................................................... 116

References........................................................................................... 118

Nomenclature I
Abbreviation


AI Artificial Intelligence
ALE Arbitrary Lagrangian Eulerian
CBN Cubic Boron Nitride
FDM Finite Difference Method
FE, FEM Finite Element Method
HSC High Speed Cutting
KT Depth of crater wear
VB Width of flank wear (mean)
VC Maximum wear of nose radius
VN Notch wear
Introduction 1

Chapter 1 Introduction



Machining operations comprise a substantial portion of the world’s manufacturing
infrastructure. They create about 15% of the value of all mechanical components
manufactured worldwide [Merc-98]. Because of its great economic and technical
importance, a large quantity of research has been carried out in order to optimize
cutting process in terms of improving quality, increasing productivity and lowering
cost.
Tool wear influences cutting power, machining quality, tool life and machining cost.
When tool wear reaches a certain value, increasing cutting force, vibration and
cutting temperature cause surface integrity deteriorated and dimension error greater
than tolerance. The life of the cutting tool comes to an end. Then the cutting tool
must be replaced or ground and the cutting process is interrupted. The cost and time
for tool replacement and adjusting machine tool increase cost and decrease
productivity. Hence tool wear relates to the economic of machining and prediction of
tool wear is of great significance for the optimization of cutting process.
At present, the prediction of tool wear is performed by calculating tool life according
to experiment and empirical tool life equations such as Taylor’s equation or its
extension versions. Although Taylor’s equation gives the simple relationship between
tool life and a certain cutting parameters, e.g. cutting speed, and is very easy to use,
it gives only the information about tool life. For the researcher and tool manufacturer
tool wear progress and tool wear profile are also concerned. Tool life equation gives
no information about the wear mechanism. But capability of predicting the
contributions of various wear mechanism is very helpful for the design of cutting tool
material and geometry. In addition, such tool life equations are valid under very
limited cutting conditions. For example, when tool geometry is changed, new
equation must be established by making experiment.
Some researchers concentrate on the study of wear mechanism and investigate the
mathematical relationship between wear due to various wear mechanisms and some
cutting process variables such as relative sliding velocity of workpiece material along
tool face, cutting temperature of tool face and normal pressure on tool face. Some
tool wear equation related to one or several wear mechanisms are developed, such
as Usui’s tool wear equation.
Introduction 2
In the recent decades, with the emergency of more and more powerful computer and
the development of numerical technique, numerical methods such as finite element
method (FEM), finite difference method (FDM) and artificial Intelligence (AI) are
widely used in machining industry. Among them, FEM has become a powerful tool in
the simulation of cutting process. Various variables in the cutting process such as
cutting force, cutting temperature, strain, strain rate, stress, etc can be predicted by
performing chip formation and heat transfer analysis in metal cutting, including those
very difficult to detect by experimental method. Therefore a new tool wear prediction
method may be developed by integrating FEM simulation of cutting process with tool
wear model.
Introduction 3

1.1 State Of Art: Finite Element Simulation Of Cutting Process

Chip formation is the essential phenomenon in the cutting process. It is the basic of
the research on physical phenomena-cutting force, cutting temperature, tool wear,
chatter, burr, built-up-edge, chip curling and chip breakage.
According to a comprehensive survey conducted by the CIRP Working Group on
Modelling of Machining Operations during 1996-1997 [Lutt-98], among the 55 major
research groups active in modelling, 43% were active in empirical modelling, 32% in
analytical modelling and 18% in numerical modelling in which finite element
modelling techniques are used as the dominant tool. In recent years, application of
finite element in metal cutting develops rapidly because of its advantages and the
development of powerful computer [Atha-98][Sand-98].
Compared to empirical and analytical methods, finite element methods used in the
analysis of chip formation has advantages in several aspects [Zhan-94]:
• Material properties can be handled as functions of strain, strain rate and
temperature;
• The interaction between chip and tool can be modelled as sticking and sliding;
• Non-linear geometric boundaries such as the free surface of the chip can be
represented and used;
• In addition to the global variables such as cutting force, feed force and chip
geometry, the local stress, temperature distributions, etc can also be obtained.
Finite element method has been used to simulate machining by Klamecki [Klam-73],
Okushima [Okus-71], and Tay et al [Tay-74] since the early 1970s. With the
development of faster processor with larger memory, model limitations and
computational difficulty have been overcome to some extent. In addition, more
commercial FE codes are used in chip formation simulation, including: NIKE2
TM

[Stre-85], ABAQUS/Standard
TM
[Shi-02], MARC
TM
[Behr-98a], DEFORM 2D
TM
[Özel-
00b] [Cere-99], FORGE 2D
TM
[Ng-99] [Mona-99], ALGOR
TM
, FLUENT
TM
,
ABAQUS/Explicit
TM
[Baca-00] and LS DYNA
TM
[McCl-02].
Great progress has been made in this research field: Lagrangian approach is used to
simulate the cutting process including incipient chip formation state [Shet-00];
segmental chip formation is modelled to simulate high speed cutting [Bäke-00] [Bäke-
02] [West-01], hard-turning [Guo-02] [Usui-84] or large negative rake angle [Ohbu-
Introduction 4
03], 3D simulation is performed to analyse oblique cutting [Leop-98] [Klam-73] [Lin-
00] [Cere-00] [Guo-02], etc.
















Fig. 1.1 Modelling research trends [Ng-02a]

A diversity of cutting tool and workpiece materials is used in the simulation of cutting
process. For example, the modelled cutting tool materials include uncoated carbide
[Lin-01b], coated carbide [Mona-99], CBN [Özel-02], cermet, ceramic cutting tool and
diamond [Ohbu-03]. The modelled workpiece materials include carbon steel [Behr-
98b] [Gu-02], composite [Arol-02], high alloy steel [Ng-02a], cast iron, ductile iron
[Chuz-03a] [Chuz-03b], etc.
The effect of tool geometry on the chip formation process is studied, mainly including
varing rake angle [Shih-96] and tool geometry. The studied tool geometries include
sharp, chamfered [Shat-01b] [Mova-02] and round edge [Ozel-02] [Kim-99], chip
breaker [Dill-00], and worn cutting tool [Li-02] [Shih-93].
The mainly simulated cutting types include tuning [Behr-99], milling [Özel-00a],
drilling, microscopic cutting of single abrasive grain in grinding [Ohbu-03]. Orthogonal
cutting is the most frequently simulated cutting type [Stre-93].
Introduction 5
In addition, the influences of sequential cutting [Liu-00] and microstructure of
workpiece material [Chuz-03a] [Chuz.03b] on chip formation are studied.
Except the normally discussed variables cutting force, cutting temperature and
stress, residual stress [Yang-02] [Shih-93], tool wear [Söhn-01b] [Yen-02], tool
performance [Ahma-89], burr formation [Guo-00], chip breakage [Maru-02], chip flow
angle [Stre-02], etc are investigated as well.

1.1.1 Numerical Aspects

The implementation of cutting process simulation is based on numerical theory and
technique. Their development is helpful to improve the capability of the simulation.

1.1.1.1 Approach

Several approaches are supplied for numerical modelling: Lagrangian, Eulerian and
Arbitrary Lagrangian Eulerian (ALE).

Eulerian Approach

In Eulerian approach, the mesh is fixed spatially and the material flows through the
mesh. Eulerian approach is suitable to analyse the steady state of cutting process,
not including the transition from initial to steady state cutting process, varying cutting
thickness in milling operation or serrated chip in high-speed-cutting because it is
unable to simulate free surface conditions. Cutting process analysis with Eulerian
approach requires less calculation time because the workpiece model consists of
fewer elements. That is the reason why before 1995 the applications of Eulerian
approach in chip formation analysis overrun those of Lagrangian approach. But
experimental work is often necessary in order to determine the chip geometry and
shear angle, which is an unavoidable part of geometry modelling.

Lagrangian Approach

In Lagrangian approach, the mesh follows the material. Because the deformation of
the free surface of the chip can be automatically treated by elastic-plastic material
deformation, Lagrangian approach can be used to simulate from initial to steady state
of cutting process. But in order to extend the cutting time until steady state, a long
Introduction 6
workpiece is needed in geometry modelling, which increases the calculation time. In
order to perform chip separation, chip separation criteria and realization method are
necessary.

Arbitrary Lagrangian Eulerian Approach (ALE)

ALE approach combines the features of pure Lagrangian and Eulerian approach, in
which the mesh is allowed to move independently of the material. It is an effective
tool for improving mesh quality in the analysis of large deformation problem. Many
commercial FE codes introduce ALE approach by adjusting mesh based on different
mesh adaptivity.
The adaptive meshing technique in ABAQUS/Explicit belongs to ALE approach. It
can be used to analyse not only Lagrangian problem but also Eulerian problem. By
giving suitable mesh control parameters, the whole process from initial to steady
state can be simulated without the need of chip separation criterion or any chip
geometry data from experiment. Furthermore, it is not necessary to extend the size of
workpiece model. Hence the calculation time is not increased.

1.1.1.2 Mesh Adaptivity

Three types of mesh adaptivity are designed to create a new spatial discretisation
and improve mesh quality: h-adaptivity, p-adaptivity and r-adaptivity [Kalh-01].
• H-adaptivity changes the size of the mesh. The new mesh has different
number of elements and the connectivity of the nodes is changed.
• In p-adaptivity the degree of the interpolating polynomial is changed.
• R-adaptivity is based on relocation of the nodes, without altering the topology
(elements and connectivity) of the mesh.
For example, adaptive meshing technique in ABAQUS/Explicit is accomplished by
using R-adaptivity. During meshing nodes are moved to more favourable positions to
improve mesh distortion. In addition, solution-dependent meshing is supplied to
concentrate mesh towards the developing boundary concave, e.g., chip separation
area in the vicinity of the cutting edge, and produce local mesh refinement in this
area.
But it is found that only the application of r-adaptivity is not sufficient to maintain the
mesh quality. Therefore some FE codes, e.g. Deform-2D and AdvantEdge employ
Introduction 7
the combination of r- and h-adaptivity. Mesh is refined where great difference in the
gradients of a certain solution is detected between elements. For example, Marusich
et al propose to refine mesh according to plastic work rate in each element [Maru-95];
Owen et al use an error estimator based on the rate of fracture indicator to produce a
fine mesh in high plastic deformation area and the regions where material failure is
going to take place [Owen-99].
Chip separation is produced during meshing and mesh refining. In addition, the
contact at tool-chip interface can be improved as well.

Introduction 8
1.1.2 Mechanical Aspects

The development of metal cutting theory helps people get more and more correct
understanding in mechanical aspects of cutting process including contact and friction,
material property, chip separation, etc. The modelling of these aspects influences the
accuracy of cutting process simulation.

1.1.2.1 Contact And Friction

Friction behaviour on the tool face determines the cutting power, machining quality
and tool wear. It plays an important role in metal cutting.

Development Of Friction Model In Metal Cutting

The nature of friction between two dry sliding surfaces was described by Amontoms
in 1699 [Amon-99]. He put forward that the coefficient of friction µ is independent of
apparent area of contact A and applied normal load . In 1785, Coulomb [Coul-85]
approved and developed these laws by proposing that the coefficient of friction is
substantially independent of the sliding velocity. Accordingly a constant coefficient of
friction is expected on the tool face in metal cutting process.
n
F

const
F
F
n
f
= = µ (1.1)

where is the friction force.
f
F
However in metal cutting process, it is generally observed that the mean coefficient of
friction on the tool face varies considerably with the change in cutting speed, rake
angle and so on. This results from the extreme conditions of metal cutting area where
the normal pressure at tool-chip interface is very high.
According to Eq. 1.2 proposed by Finne and Shaw [Finn-56], the ratio of the real area
of contact A
r
to the apparent area of contact A approaches or reaches 1 under cutting
conditions, which is different from the application conditions of Coulomb’s
assumption.

Introduction 9
BN r
e
A
A

− =1 (1.2)

where N is normal force.
Based on the assumption of shearing action within the workpiece material, Zorev
proposed the distribution of shear and normal stress on the rake face as shown in
Fig.1.2 [Zore-63]. The chip-tool interface is divided into sticking and sliding regions. In
sticking region, adjacent to the cutting edge,
A
A
r
approaches unity under very high
normal stress, and shear stress is believed equal to shear strength of the workpiece
material. In sliding region,
A
A
r
is less than unity, and the coefficient of friction is
believed constant.













Fig. 1.2 Stress distribution on tool-chip interface

Plenty of evidence from worn tools, from quick-stop sections and from chips showed
the coexistence of sticking and sliding at tool/chip interface under many cutting
conditions [Tren-77].
Some advanced testing technologies, e.g. photoelastic measurements [Rice-60] or
split tool dynamometers [Kato-72] [Chil-98], are used in experiments to discover the
form of stress distribution on the rake face. But these techniques are limited when the
stresses very close to the cutting edge are determined.
Introduction 10

Applied Friction Models In Cutting Process Simulation

In the finite element analyses of metal cutting, various approaches are used in the
modelling of friction. Constant coefficient of friction based on Coulomb’s friction law is
used in most cases. Normally the coefficient of friction µ is calculated by using
Eq.1.3 according to the cutting force , thrust force , and rake angle
c
F
t
F α .

α
α
µ
tan
tan
t C
C t
F F
F F

+
= (1.3)

Ng and his co-operators performed orthogonal cutting tests under different cutting
conditions to establish a linear relation between the coefficient of frictionµ , cutting
speed , rake angle
c
v α , and feed , given by Eq. 1.4, by using Regression analysis
[Ng-02b].
f

c
v f 0002 . 0 888 . 3 00446 . 0 034 . 1 − − − = α µ (1.4)

Liu et al [Liu-00] determined the coefficient of friction by performing simulation using
different values and carrying out the sensitivity study on the coefficient of friction.
When Zorev’s sliding-sticking friction model is employed in the simulation, the
division of the two regions is determined by two methods: one is to prescribe the
length of each region [Shih-95] [Wu-96] [Shat-00], the other is to determine the
sliding and sticking region automatically by program according to a criterion [Zhan-
94] [Guo-00], given by Eq. 1.5.

) , min(
s
τ µσ τ = (1.5)

where
s
τ is the shear flow stress of the chip material;
τ is friction stress;
σ is normal stress.
Introduction 11
Iwata et al [Iwat-84] proposed the relationship given by Eq. 1.6 after put forward a
method to test friction between newly created surfaces and tool material.

|
|
.
|

\
|
|
.
|

\
|
=
V
V
H
p H µ
τ
07 . 0
tanh
07 . 0
Mpa (1.6)

where
V
H is the Vickers hardness of the workpiece material;
p is contact pressure in MPa.
A frictional shear factor is introduced into the relationship in order to make the
calculated results agree with those of experiment.
Yang and Liu [Yang-02] proposed a stress-based polynomial model of friction, given
by Eq. 1.7.


=
=
=
4
0
n
n
n
n
a σ τ (1.7)

0
a , , , and are determined by fitting experimental stress curve on rake
face.
1
a
2
a
3
a
4
a

1.1.2.2 Material Constitutive Model

The accuracy of the finite element analysis is severely dependent on the accuracy of
the material mechanical properties.

Influence Factors Of Material Property

Experiments shows that material properties, e.g. stress-strain relationship, are
affected by the strain rate and temperature during material forming process with
plastic deformation. For the same value of strain, the stress is higher at higher strain
rate due to the viscous effect during plastic deformation and lower at higher
temperature due to material softening, as shown in Fig. 1.3. This overstress effect by
strain rate is more pronounced as the temperature increases [Shih-91]. In metal
cutting process, temperature, strain and strain rate are very high. Thermal-
Introduction 12
viscoplastic material constitutive model is necessary for the finite element analysis of
metal cutting.







(a) overstress effect (b) material softening





S
t
r
e
s
s

[
N
/
m
m
2
]





S
t
r
e
s
s

[
N
/
m
m
2
]

Strain [%] Strain [%]
Temperature [K]
Fig. 1.3 Material property curve

Many researchers are making efforts to establish such material constitutive models
for different workpiece materials through experimental [Kopp-01], analytical or
simulation methods [Shat-01a] [Özel-00b] [Batz-02]. Based on their supports, a
material model database has been developed by Söhner et al [Söhn-01a].

Material Constitutive Model For Mild Carbon Steel

The main workpiece materials used in the following research are mild carbon steel
CK45 and AISI1045.

• For CK45

The material constitutive model developed by O. Vöringer is used, which is described
by Eq. 1.8 and Eq. 1.9.

( )
m
n
v
T
T
T
|
|
.
|

\
|
|
|
.
|

\
|
− =
∗ ∗
0
0
1 , σ ε σ & (1.8)

with
) (
0
0
0
pl
kIn
G
ε
ε
&
&

= T (1.9)
Introduction 13

where the constants for CK45 are: m=1.78, n=0.53, , ,
and . k is Boltzmann constant and T is temperature in Kelvin [Schu-
00].
ev G 58 . 0
0
= ∆
1 5
0
10 29 . 7

× = s ε&
MPa 1352
0
=

σ
In the simulation of cutting process, a user material subroutine based on this material
constitutive model is employed.

• For AISI1045

To describe the material property of AISI1045, the Johnson-cook constitutive
equation is used.

( )
( )
|
|
.
|

\
|
+
|
|
.
|

\
|


|
|
.
|

\
|
|
.
|

\
|
+ =
− −
2
700 00005 . 0
1000
ln 1
T
room melt
melt n
ae
T T
T T
C B
ε
ε σ
&
(1.10)

where B=996.1, C=0.097, n=0.168, a=0.275, T
melt
=1480°C [Kopp-01], σ is the
effective stress in MPa, and T is temperature in °C.

1.1.2.3 Chip Separation

In the cutting process, with the cutting tool advancing into the workpiece, the
workpiece material is separated into two parts. The unwanted part forms the chip. By
chip separation, a new workpiece surface is formed on the created part.
The realization of chip separation is one of the main problems in the simulation of
chip formation process. Normally it includes two aspects of consideration: chip
separation criterion and model realization.

Chip Separation Criterion

The chip separation criteria used by researchers can be categorized as two types:
geometrical and physical.
Geometrical criteria define geometric parameters, e.g. a distance value. When the
distance between the nearest workpiece node on the moving path of the cutting edge
Introduction 14
and the cutting edge is equal to or smaller than this given distance value, chip
separation takes place [Shih-95].
Physical criteria is related to some physical meaning of chip separation. They are
based on physical parameters such as stress [Iwat-84], strain energy density [Lin-99]
or effective plastic strain [Shir-93]. When such physical parameter reaches a critical
value, material failure takes place. The most reliable critical value is obtained by
performing experiments, although sometimes it is defined at random. A critical value
considering multi-influencing factors, for example, temperature- and strain rate-
dependent strain at failure will provide a better simulation result.
According to the investigation on both types of criteria made by Huang and Black
[Huan-96], neither had a substantial effect on chip geometry, distribution of shear
stress, effective stress or effective plastic strain in the chip and in the machined
surface. However, the magnitude designated for these criteria did have a major effect
on mesh distortion together with the value of maximum shear stress, and the
effective stress in the machined surface [Ng-02a].

Model Realization

There are several methods to model chip separation in finite element mesh. They are
related with the applied software.
• Element removal [Cere-96]
When chip separation criterion, normally physical criterion, is reached, material
failure happens and the element carries no stress any more as if they do not exist.
Such element can be removed and does not display.






Fig. 1.4 Element removal [Behr-98b]




Introduction 15
• Node debond [Shi-02] [Shet-00] [Shet-03]
The chip and the workpiece are two separated parts. They are perfectly bonded
together through some pair of nodes along the prospective parting line. The chip
separation can be geometrical, physical or their combination. When chip separation
criterion is reached, debond of the node pair takes place and the two nodes move in
different direction.






Fig. 1.5 Node debond

• Node splitting [Shih-95]
Chip separation is realized by element separation in front of cutting edge. The two
neighbouring elements have common node before separation. When the separation
criterion is met, for example, a node is very close to the cutting edge. Element
separation takes place and a new node is created at the same position; two nodes
overlap together and connect to two different elements. Through the further
movement of the cutting tool, the two elements move in different direction and lose
contact.






Fig. 1.6 Node splitting [Behr-98b]

• Mesh adaptivity [Arra-02]
Chip separation is performed by mesh refinement in the separation zone by
increasing the number of elements or relocation of the nodes.
Introduction 16
1.2 Technical Background About Tool Wear

Prediction of tool wear is complex because of the complexity of machining system.
Tool wear in cutting process is produced by the contact and relative sliding between
the cutting tool and the workpiece and between the cutting tool and the chip under
the extreme conditions of cutting area; temperature at the cutting edge can exceed
1800°F and pressure is greater than 2,000psi [John-01]. Any element changing
contact conditions in cutting area affects tool wear. These elements come from the
whole machining system comprising workpiece, tool, interface and machine tool:

Tool wear
economy, workpiece quality, process security
material
texture
structure
material
properties
interface
friction
cooling
lubricant
cutting param.
contact
tool
cutting material
coating
geometry
machine
design
dynamics
Tool wear
economy, workpiece quality, process security
material
texture
structure
material
properties
interface
friction
cooling
lubricant
cutting param.
contact
tool
cutting material
coating
geometry
machine
design
dynamics

Fig. 1.7 Influencing elements of tool wear [Söhn-01b]

• Workpiece: It includes the workpiece material and its physical properties
(mechanical and thermal properties, microstructure, hardness, etc), which
determine cutting force and energy for the applied cutting conditions.
• Tool: Tool material, tool coatings and tool geometric design (edge preparation,
rake angle, etc) need to be appropriately chosen for different operations
(roughing, semi-roughing, or finishing). The optimal performance of a cutting
tool requires a right combination of the above tool parameters and cutting
conditions (cutting speed, feed rate, depth of cut, etc)
• Interface: It involves the interface conditions. In 80% of the industrial cutting
applications, coolants are used to decrease cutting temperatures and likely
Introduction 17
reduce tool wear. Increasingly new technologies, such as the minimum liquid
lubrication, have been developed to reduce the cost of coolant that makes up
to 16% of the total machining costs [Walt-98].
• Dynamic: The dynamic characteristic of the machine tool, affected by the
machine tool structure and all the components taking part in the cutting
process, plays an important role for a successful cutting. Instable cutting
processes with large vibrations (chatters) result in a fluctuating overload on the
cutting tool and often lead to the premature failure of the cutting edge by tool
chipping and excessive tool wear.

1.2.1 Wear Types In Metal Cutting

Under high temperature, high pressure, high sliding velocity and mechanical or
thermal shock in cutting area, cutting tool has normally complex wear appearance,
which consists of some basic wear types such as crater wear, flank wear, thermal
crack, brittle crack, fatigue crack, insert breakage, plastic deformation and build-up
edge. The dominating basic wear types vary with the change of cutting conditions.
Crater wear and flank wear shown in Fig. 1.8 are the most common wear types.


Fig. 1.8 Wear types [Lim-01]

• Crater wear: In continuous cutting, e.g. turning operation, crater wear normally
forms on rake face. It conforms to the shape of the chip underside and
reaches the maximum depth at a distance away from the cutting edge where
highest temperature occurs. At high cutting speed, crater wear is often the
factor that determines the life of the cutting tool: the tool edge is weakened by
the severe cratering and eventually fractures. Crater wear is improved by
Introduction 18
selecting suitable cutting parameters and using coated tool or ultra-hard
material tool.
• Flank wear: Flank wear is caused by the friction between the newly machined
workpiece surface and the tool flank face. It is responsible for a poor surface
finish, a decrease in the dimension accuracy of the tool and an increase in
cutting force, temperature and vibration. Hence the width of the flank wear
land VB is usually taken as a measure of the amount of wear and a threshold
value of the width is defined as tool reshape criterion.

1.2.2 Wear Mechanism

In order to find out suitable way to slow down the wear process, many research
works are carried out to analyze the wear mechanism in metal cutting. It is found that
tool wear is not formed by a unique tool wear mechanism but a combination of
several tool wear mechanisms.
Tool wear mechanisms in metal cutting include abrasive wear, adhesive wear,
delamination wear, solution wear, diffusion wear, oxidation wear, electrochemical
wear, etc. Among them, abrasive wear, adhesive wear, diffusion wear and oxidation
wear are very important.
• Abrasive wear: Tool material is removed away by the mechanical action of
hard particles in the contact interface passing over the tool face. These hard
particles may be hard constituents in the work material, fragments of the hard
tool material removed in some way or highly strain-hardened fragments of an
unstable built-up edge [Boot-89].
• Adhesive wear: Adhesive wear is caused by the formation and fracture of
welded asperity junctions between the cutting tool and the workpiece.
• Diffusion wear: Diffusion wear takes place when atoms move from the tool
material to the workpiece material because of the concentration difference.
The rate of diffusion increases exponentially with the increase of temperature.
• Oxidation wear: A slight oxidation of tool face is helpful to reduce the tool
wear. It reduces adhesion, diffusion and current by isolating the tool and the
workpiece. But at high temperature soft oxide layers, e.g. Co
3
O
4
, CoO, WO
3
,
TiO
2
, etc are formed rapidly, then taken away by the chip and the workpiece.
This results in a rapid tool material loss, i.e., oxidation wear.
Introduction 19
Under different cutting conditions dominating wear mechanisms are different. For a
certain combination of cutting tool and workpiece, the dominating wear mechanisms
vary with cutting temperature, as shown in Fig.1.9. According to the temperature
distribution on the tool face, it is assumed that crater wear is mainly caused by
abrasive wear, diffusion wear and oxidation wear, but flank wear mainly dominated
by abrasive wear due to hard second phase in the workpiece material.










Diffusion
Abrasion
W
e
a
r

Cutting temperature
(cutting speed, feed, etc)
Oxidizing
Adhesion
Abrasive wear
Adhesive wear
Diffusion wear
Chip
Oxidizing wear
Tool
vv c c
Workpiece
Fig. 1.9 Wear mechanism [Köni-84]

1.2.3 Tool Wear Model

Many mathematical models are developed to describe tool wear in quantity. They can
be categorized into two types: tool life models and tool wear rate models.
• Tool life models: This type of wear models gives the relationship between tool
life and cutting parameters or variables. For example, Taylor’s tool life
equation [Tayl-07], reveals the exponential relationship between tool life and
cutting speed, and Hastings tool life equation describes the great effect of
cutting temperature on tool life [Hast-79], see Table 1.1. The constants n, C
T
,
A and B are defined by doing a lot of experiments with cutting speed changing
and fitting the experimental data with the equation. It is very convenient to
predict tool life by using this equation. In various sizes of cutting database,
Taylor’s tool life equation and its extension versions under different cutting
conditions appear most frequently.
Tool life equations are suitable to very limited range of cutting conditions. As
the new machining technologies, e.g. high-speed-cutting or dry cutting, are
getting spread in manufacturing industry, the existing tool life equations need
Introduction 20
to be updated with new constants and a lot of experimental work has to be
done. In addition, except that tool life can be predicted by these equations, it is
difficult to get further information about the tool wear progress, tool wear
profile or tool wear mechanisms that are sometimes important for tool
designers.
• Tool wear rate models: These models are derived from one or several wear
mechanisms. They provide the information about wear growth rate due to
some wear mechanisms. In these modes, the wear growth rate, i.e. the rate of
volume loss at the tool face (rake or flank) per unit contact area per unit time
(mm/min), are related to several cutting process variables that have to be
decided by experiment or using some methods [Kwon-00].
Table 1.1 Tool wear models










Usui’s model, which was derived from
equation of adhesive wear
[Usui et al., 1978]:
dW/dt = A
.
σ
t
V
S
exp(-B/T)
- (A, B = constants)
- dW/dt = rate of volume loss per unit
contact area per unit time (mm/min)
- σ
t
, T = normal stress and temperature
- A, B = wear characteristic constants
Hastings tool life equation (Hastings et al, Hastings tool life equation (Hastings et al,
1979): 1979):
T T
B B . .
L = A L = A
( (A, B A, B = = constants) constants)
Takeyama & Murata’s model, considering
abrasive wear and diffusive wear (1963):
dW/dt = G(V
c
, f) + D
.
exp(-E/RT)
(G, D = constants)
Taylor’s tool life equation:
V
c
.
L
n
= C
T
(n, C
T
= constants)
“Differential” Tool Wear Rate Models Empirical Tool Life Models
Usui’s model, which was derived from
equation of adhesive wear
[Usui et al., 1978]:
dW/dt = A
.
σ
t
V
S
exp(-B/T)
- (A, B = constants)
- dW/dt = rate of volume loss per unit
contact area per unit time (mm/min)
- σ
t
, T = normal stress and temperature
- A, B = wear characteristic constants
Hastings tool life equation (Hastings et al, Hastings tool life equation (Hastings et al,
1979): 1979):
T T
B B . .
L = A L = A
( (A, B A, B = = constants) constants)
Takeyama & Murata’s model, considering
abrasive wear and diffusive wear (1963):
dW/dt = G(V
c
, f) + D
.
exp(-E/RT)
(G, D = constants)
Taylor’s tool life equation:
V
c
.
L
n
= C
T
(n, C
T
= constants)
“Differential” Tool Wear Rate Models Empirical Tool Life Models
V
c
= Cutting speed
L = Tool life
E = Process activation
energy
R = Universal gas constant
T = Cutting temperature
f = Feed
V
S
= Sliding velocity
V
c
= Cutting speed
L = Tool life
E = Process activation
energy
R = Universal gas constant
T = Cutting temperature
f = Feed
V
S
= Sliding velocity
C,
λ
v
c
T
θ
C, λ
T
n
c
C T v = ⋅
A T
B
= ⋅ θ
( ) ( ) θ R E D f v G dt dW
c
− ⋅ + = exp ,
( ) θ λ σ − = exp
s t
v C dt dW
θ
v
s
Usui’s model, which was derived from
equation of adhesive wear
[Usui et al., 1978]:
dW/dt = A
.
σ
t
V
S
exp(-B/T)
- (A, B = constants)
- dW/dt = rate of volume loss per unit
contact area per unit time (mm/min)
- σ
t
, T = normal stress and temperature
- A, B = wear characteristic constants
Hastings tool life equation (Hastings et al, Hastings tool life equation (Hastings et al,
1979): 1979):
T T
B B . .
L = A L = A
( (A, B A, B = = constants) constants)
Takeyama & Murata’s model, considering
abrasive wear and diffusive wear (1963):
dW/dt = G(V
c
, f) + D
.
exp(-E/RT)
(G, D = constants)
Taylor’s tool life equation:
V
c
.
L
n
= C
T
(n, C
T
= constants)
“Differential” Tool Wear Rate Models Empirical Tool Life Models
Usui’s model, which was derived from
equation of adhesive wear
[Usui et al., 1978]:
dW/dt = A
.
σ
t
V
S
exp(-B/T)
- (A, B = constants)
- dW/dt = rate of volume loss per unit
contact area per unit time (mm/min)
- σ
t
, T = normal stress and temperature
- A, B = wear characteristic constants
Hastings tool life equation (Hastings et al, Hastings tool life equation (Hastings et al,
1979): 1979):
T T
B B . .
L = A L = A
( (A, B A, B = = constants) constants)
Takeyama & Murata’s model, considering
abrasive wear and diffusive wear (1963):
dW/dt = G(V
c
, f) + D
.
exp(-E/RT)
(G, D = constants)
Taylor’s tool life equation:
V
c
.
L
n
= C
T
(n, C
T
= constants)
“Differential” Tool Wear Rate Models Empirical Tool Life Models
V
c
= Cutting speed
L = Tool life
E = Process activation
energy
R = Universal gas constant
T = Cutting temperature
f = Feed
V
S
= Sliding velocity
V
c
= Cutting speed
L = Tool life
E = Process activation
energy
R = Universal gas constant
T = Cutting temperature
f = Feed
V
S
= Sliding velocity
C,
λ
v
c
T
θ
C, λ
T
n
c
C T v = ⋅
A T
B
= ⋅ θ
( ) ( ) θ R E D f v G dt dW
c
− ⋅ + = exp ,
( ) θ λ σ − = exp
s t
v C dt dW
θ
v
s


In Table 1.1, the right column shows two tool wear rate models, which are
obtained from literatures.
Takeyama & Murata’s model is developed by considering the combination
action of abrasive wear and diffusive wear. Therefore the equation sums two
parts up. One part shows that abrasive wear is influenced by the cutting speed
and feed. Another part including universal gas constant and tool temperature
describes diffusive wear.
Usui’s model is derived from Shaw’s equation of adhesive wear [Usui-78c].
Except the constants A and B, Usui’s equation includes three variables: sliding
velocity between the chip and the cutting tool, tool temperature and normal
Introduction 21
pressure on tool face. These variables can be predicted by FEM simulation of
cutting process or combining analytical method and FDM. Therefore Usui’s
equation is very practical for the implementation of tool wear estimation by
using FEM or by using the combination of FDM and analytical method.
When tungsten carbide tools are used to machine carbon steels, crater wear
on rake face was assumed mainly caused by adhesive wear. According to
cutting experiment, Usui determined the constants for such cutting conditions
and validated this model by the prediction of crater wear.
The latter study showed that this equation is able to describe flank wear as
well, which mainly results from abrasive wear [Kita-88]. All the points for flank
wear and crater wear defined by experiment distribute along two characteristic
lines with different gradients, which intersect at the critical temperature of
around 1,150K. The experimental points for crater wear usually lie on the line
in the higher temperature range, whereas those for flank wear are usually
distributed around the line in the lower temperature range.
The constants in tool wear rate models are depending on the combination of
workpiece and cutting tool material. Table 1.2 shows the charateristic
constants in Usui’s equation for the combination of carbon steel and carbide
tool that obtained from literature [Kita-89]. They are introduced in the later tool
wear estimation models.

Table 1.2 : Characteristic constants for carbon steels [Kita-89]
C [m
2
/MN] λ [K]
K
f
1150 ≥ θ K
f
1150 < θ K
f
1150 ≥ θ K
f
1150 < θ
2
10 198 . 1

×
9
10 8 . 7

×
4
10 195 . 2 ×
3
10 302 . 5 ×



Introduction 22
1.3 Research Of Tool Wear With Finite Element Methods

1.3.1 Comparison Between FEM Method And Empirical Method

Based on tool wear rate models, the estimation of tool wear profile progress with the
cutting process can be implemented by predicting cutting process variables using
finite element method. Its advantages and disadvantages are shown by the
comparison with the empirical method in Table 1.3.

Table 1.3 Comparison of FEM method and empirical method
Compared aspects Empirical method FEM method
Environment
requirement
Special machine, tool,
workpiece, personnel for
cutting tests
Powerful computer, tool
wear rate model and FEM
code
The procedure of
calculating tool wear
Cutting tests and regressive
analysis
Obtaining tool wear rate
model by experiment or
from literature, running the
program with tool wear rate
models under new cutting
conditions
Application under
new cutting
conditions
New experiments have to
be carried out to update the
constants of tool life models
If only tool wear rate model
is updated according to new
cutting conditions, the
program can be used again
Time The development of new
tool life models is time
consuming;
Whereas the prediction of
tool wear with the tool life
model is very efficient

The time for developing the
entire program is relative
long.
The time for calculating the
tool wear with the program
depends on the
performance of computer
Wear mechanism Wear mechanism is not Yes, even the contributions
Introduction 23
considered of the main mechanisms
can be calculated
Workpiece material Uneven material distribution
result from impurity, heat
treatment, work hardening
Homogeneous material
model, thermal visco-plastic
material
Tool material Uneven material properties
result from impurity, heat
treatment, etc
Homogeneous material
model, ideal elastic material
Medium Sensitive to the cooling
method, coolant type,
cooling effect, etc
The types of heat emission
through tool face and
workpiece surface under
various cooling conditions
and their FEM
implementation have to be
considered
Vibration of machine-
tool-workpiece
system
The constants are sensitive
to the vibration of the
system
Not considered at present
Predicted wear
parameters
Very limited information can
be obtained, for example
only tool life is predicted
with Taylor’s tool life
equation
Comprehensive information
about tool wear including
crater wear profile, flank
wear profile, VB, KT, VC
(for 3D), VN (for 3D), etc
can be predicted
Cutting type Tool life models under
various cutting type can be
developed
At present, only tool wear
prediction in turning and
milling operations are
studied.
For different cutting types,
the tool wear program may
need adjusting according to
the characteristic of relative
Introduction 24
motion of cutting tool and
workpiece
Requirement on the
user
No special requirement At present, except basic
knowledge about metal
cutting theory, user needs
the basic knowledge about
FEM chip formation, heat
transfer analysis
Application at
present
Used in the real production For research and education
Quality of the
prediction
Quantitative Qualitative


1.3.2 State Of Art: Numerical Implementation Of Tool Wear Estimation

Tool wear estimation with Finite Element Method is developed from tool wear
estimation with the combination of analytical method and Finite Difference Method
(FDM).

1.3.2.1 Tool Wear Estimation With The Combination Of Analytical Method And
FDM

Usui’s Research-Prediction Of Crater Wear

The earliest reported research work on tool wear estimation with the combination of
analytical energy method and FDM was performed by E. Usui et al in 1978. He first
derived a characteristic equation of crater wear theoretically by combining
M.C.Shaw’s adhesive wear, temperature-dependent material asperity hardness and
temperature-dependent Holm’s probability, given by

) exp( θ λ σ − =
s t
v C
dt
dw
(1.11)

Then he verified the equation experimentally.
Introduction 25
a.) Implementation Procedure

The chip formation, sliding velocity of the chip and cutting force are predicted through
energy method proposed in previous papers [Usui-78a] [Usui-78b] [Usui-78c].
By using the predicted cutting force and tool-chip contact length together with an
assumption of an exponential normal stress distribution and a triangle or trapezoidal
frictional stress distribution on the tool face, the frictional stress is calculated.
The temperature distribution within the chip and the tool at steady state is obtained
with FDM by considering the heat source on the shear plane and on rake face.
The characteristic constants of the equation for the combination of carbon steel and
P20 are determined with the aid of the predicted temperature, stress on tool face and
the measured wear by curve fitting.
Then computer calculation of crater wear is carried out by using the characteristic
equation, and the predicted distribution of the stress and the temperature.

b.) Result

The predicted crater wear was reported in good agreement with the measured in
experiment in depth and contour except some discrepancy in the location of the
deepest portion.

c.) Limitations

• When using the energy method to predict the chip formation and cutting force,
orthogonal cutting data about shear stress on shear plane, friction angle and
shear angle are needed, the prediction of crater wear cannot be carried out
without making experiment in advance.
• The energy method is developed based on single shear plane for the cutting
tool with sharp cutting edge. The effect of cutting edge preparation, such as
round cutting edge, or rounded cutting edge due to wear on the tool wear
cannot be considered.

Introduction 26
Kitagawa’s Research-Prediction Of Flank Wear

By analysing the flank wear characteristics of tungsten carbide tools in turning plain
carbon steels at steady-state cutting without a built-up edge experimentally, Kitagawa
finds that flank wear can be described by the same characteristic equation, Eq. 1.11,
for crater wear. Tool wear consists of two characteristic lines with different gradient,
which intersect at the critical temperature of around 1,150K. The experimental points
for crater wear usually lie on the line in the higher temperature range, whereas those
for flank wear are usually distributed around the line in the lower temperature range.

a.) Implementation Procedure

In the prediction, the sliding velocity of workpiece material on the flank wear land is
assumed equal to the cutting speed.
The values of cutting force, thrust force and chip contact length obtained from
orthogonal experiment must be given beforehand. By prescribing a triangle
distribution of frictional stress along the tool-chip contact length with maximum value
at the cutting edge and neglecting the contribution of stress on flank face to the
cutting force and thrust force, the frictional stress is calculated.
On the flank wear, the frictional stress at the cutting edge is set equal to the
maximum value on rake face, and frictional stress on the other sites is arbitrary set.
Normal stress on flank wear is set equal to frictional stress.
Then the temperature on flank wear land is predicted by considering the heat
generated on the flank wear, rake face and in the shear plane using FDM.
The wear rate on the flank wear is calculated according to the predicted
temperature, arbitrary set normal stress and sliding velocity. Normal stress on flank
wear is adjusted continuously until a uniformly distributed wear rate is achieved
everywhere on the flank wear land.

b.) Result

It was reported that the predicted tool life, temperature and mean stresses on the
flank wear land are in reasonable agreement with experiment even with changing
cutting speed, feed and workpiece material.

Introduction 27
c.) Limitations

• The prediction method is developed under the assumption of no crater wear
formed on the rake face. This limits its application to low cutting speed range.
• The assumption of uniform wear rate on flank wear excludes the formation of
rounded edge due to wear that is often observed in experiments.
• The prediction method is not applicable to the cutting tool with any edge
preparation because of the assumption of stress distribution on rake face
• The prediction is based on cutting force, thrust force and chip contact length
obtained from orthogonal experiment. These values vary with the development
of flank wear. Whether the predicted crater wear is sensitive to the frequency
of measuring these values during the development of flank wear is very
important for its application perspective.

1.3.2.2 Tool Wear Estimation With FEM

Yen And Söhner’s Research (FEM)

Although in a paper in 1999, J. Monaghan and T. MacGinley claimed that they
performed tool wear analysis based on a wear function related to normal stress and
sliding velocity by predicting stress distribution within coated and uncoated carbide
tool with and without chip breaker using commercial FEM code-FORGE2 [Mona-99],
no implementation procedure, clear predicted tool wear profile and wear value are
described or provided. It is suspected that only the tendency and possibility of tool
wear distribution were analysed qualitatively.
Hence the earliest reported research of tool wear estimation in quantity with FEM
was done by Y. C. Yen and J. Söhner et al since 2001.
According to the paper in 2002 [Yen-02] and the dissertation of Söhner [Söhn-03],
the numerical implementation of the integration of tool wear rate models with FEM
calculations to predict the evolution of the tool wear was performed by using
commercial FE code DEFORM-2D.

a.) Implementation Procedure

Usui’s wear model is used to calculate the wear rate of the uncoated carbide tool in
cutting carbon steel.
Introduction 28
The complete procedure includes four phases. In the first phase, a coupled thermal-
viscoplastic Lagrangian cutting simulation combined with an introduced special
simulation module, ‘Konti-cut’, which can prolong the cutting simulation to a sufficient
long cutting time, is used to perform chip formation analysis until mechanical steady
state is reached. In the second phase, pure heat transfer analysis for the tool is
performed to attain thermal steady state in the tool. Both the chip formation and heat
transfer analyses are performed with commercial FE code DEFORM-2D. With the
values of nodal temperature, normal stress and sliding velocity under steady-state
cutting condition provided by the first two phases, the nodal wear rate is calculated in
the third phase. Then new tool geometry accounting for tool wear is calculated based
on the user input for a cutting time increment. In the last phase, the tool geometry
model is updated by moving nodes.

b.) Result

Simulation study was made with worn tool initially including a pre-defined wear land
of 0.06mm on the flank face. The wear rates of flank wear and crater wear are of the
same order, the location of the maximum wear rate and the low wear rate close to
tool radius are consistent with the experimental result.
When a sharp tool is used, the predicted wear rate on the flank face is one order of
magnitude smaller than that on the rake face, while crater wear and flank wear occur
simultaneously at a similar wear rate in the experiment. This problem was improved
by using a new tool wear model especially developed for the simulated cutting
condition [Fran-02].

c.) Limitations

• The tool geometry was updated manually, instead of being performed
automatically according to a certain algorithm.
• The selection of a suitable cutting time increment is very difficult to perform for
a user without doing experiment in advance.
Introduction 29

1.3.2.3 Summary Of Literature

According to the above literature analysis, some conclusions can be obtained:
• The advantages of tool wear estimation with FEM over tool wear estimation
with the combination of analytical method and FDM are considered in several
aspects, as shown in Table 1.4.
• Because of the short history of the research on tool wear estimation with FEM,
only 2D tool wear of uncoated carbide tool cutting carbon steel workpiece
AISI1045 was studied. The cutting type is limited turning operation and
orthogonal cutting.
• Only the commercial FE code DEFORM-2D is used in tool wear estimation.
However, the simulation of cutting process is assumed more suitable to be
performed with explicit method because of the large deformation, impact and
complex contact problem. The study should be carried out with some FE code
using explicit method and providing good development platform as well, for
example, ABAQUS.
• At present, numerical implementation of tool wear estimation is only developed
for the cutting process with steady state. The end of tool life in intermittent
cutting, for example, milling operation, is mainly caused by progressive tool
wear. Tool wear estimation in intermittent cutting, is different from turning
operation because of the lack of steady state. Therefore the estimation of tool
wear should be studied by developing new simulation procedure.
Introduction 30
Table 1.4 Comparisons between tool wear estimation with FEM and tool wear
estimation with the combination of analytical method and FDM
Compared
aspects
With the combination of
analytical method and FDM
With FEM
Realization Analytical method, e.g. energy
method;
Assumption and simplification of
the cutting condition;
Tool wear rate model
FEM chip formation analysis;
FEM heat transfer analysis;
Tool wear rate model
Predicted
wear value
Only crater wear or only flank
wear
Crater wear and flank wear
simultaneously
Tool For crater wear estimation, tool
without flank wear,
For flank wear estimation, tool
without crater wear.
Edge preparations are not
considered
Crater wear, flank wear and
edge preparation can be
included in tool geometry
model
Experimental
data
Yes, cutting force, tool-chip
contact length, etc
No
Applicable
conditions
Conventional cutting speed Conventional cutting and
HSC
Prospective Limited A necessary supplement to
the empirical method




Objective And Approach 31
Chapter 2 Objective And Approach

2.1 Objectives

The objective of this research is to develop methodology to predict tool wear and tool
life in cutting process using finite element simulations. The study is not limited to
turning operation, the prediction of tool wear in milling operation is considered as
well.
Because of the complexity of tool wear mechanisms and forms in real cutting
process, the study at present will be concentrated on two-dimension tool wear
estimation of uncoated carbide tool in dry cutting mild carbon steel.
This tool wear estimation method is performed by predicting tool temperature, sliding
velocity of chip and normal stress on tool face with FEM simulation of cutting
process. Therefore to achieve the objective, FEM simulation of turning and milling
process are studied at first, including chip formation analysis and pure heat transfer
analysis. Several modeling tools are used in order to accomplish the entire research
project.

Objective: 2D, uncoated carbide tool, turning and milling operation Objective: 2D, uncoated carbide tool, turning and milling operation
Continuous chip
formation analysis
model
General chip
formation analysis
model
Thermal steady
state analysis
Cyclical thermal
balance state
Analysis model
Tool wear
estimation program
Tool wear
estimation program
TURNING
OPERATION
MILLING
OPERATION
ABAQUS/Explicit
Fortran
ABAQUS/Standard
Fortran
Objective-oriented
programming
language: Python
MODELING TOOL












Fig. 2.1 Objective and modeling tool

Turning operation is a steady-state process when continuous chip is formed. The
implementation of tool wear estimation is relatively easier and studied first. Based on
Objective And Approach 32
the obtained experience in turning operation, the methodology of tool wear estimation
in milling operation is discussed by analyzing the feature of milling operation. Two
different tool wear estimation models are developed, one is for turning operation,
another for milling operation.

2.2 Approach

Although the tool wear estimation models for turning and milling operations are
different, the calculation procedure are similar and mainly composed of chip
formation analysis, heat transfer analysis, wear calculation and tool geometry
updating, as shown in Fig. 2.2.

The j
th
calculation cycle The j
th
calculation cycle
















Tool temperature θ
Normal pressure σ
t
,
Sliding velocity v
c
Wear calculation
( ) θ λ σ − = exp
s t
v C w&
Updated tool node file
W
i
d
t
h
o
f

f
l
a
n
k

w
e
a
r

V
B

[
m
m
]
∆t
2
∆t
3
∆t
4
t
0
t
1
t
2
t
3
t
4
Cutting time t [min]
Tool reshape criterion (eg. 0.2mm)
∆t
1
W
i
d
t
h
o
f

f
l
a
n
k

w
e
a
r

V
B

[
m
m
]
∆t
2
∆t
3
∆t
4
t
0
t
1
t
2
t
3
t
4
Cutting time t [min]
Tool reshape criterion (eg. 0.2mm)
∆t
1
Chip formation analysis
Heat flux
Temperature
Heat tranfer analysis
Nodal
displacement
Tool geometry updating
t
j-1
Time increment ∆t
(specified by user or
searched by program)
∆t
Nodal displacement
) , ( ) , ( ) , ( j i j j i j i
D t w w
r
&
r
⋅ ∆ ⋅ =
Tool temperature θ
Normal pressure σ
t
,
Sliding velocity v
c
Wear calculation
( ) θ λ σ − = exp
s t
v C w&
Tool temperature θ
Normal pressure σ
t
,
Sliding velocity v
c
Tool temperature θ
Normal pressure σ
t
,
Sliding velocity v
c
Tool temperature θ Tool temperature θ
Normal pressure σ
t
,
Sliding velocity v
c
Normal pressure σ
t
,
Sliding velocity v
c
Wear calculation
( ) θ λ σ − = exp
s t
v C w&
Wear calculation
( ) θ λ σ − = exp
s t
v C w&
Updated tool node file Updated tool node file
W
i
d
t
h
o
f

f
l
a
n
k

w
e
a
r

V
B

[
m
m
]
∆t
2
∆t
3
∆t
4
t
0
t
1
t
2
t
3
t
4
Cutting time t [min]
Tool reshape criterion (eg. 0.2mm)
∆t
1
W
i
d
t
h
o
f

f
l
a
n
k

w
e
a
r

V
B

[
m
m
]
∆t
2
∆t
3
∆t
4
t
0
t
1
t
2
t
3
t
4
Cutting time t [min]
Tool reshape criterion (eg. 0.2mm)
∆t
1
Chip formation analysis
Heat flux
Temperature
Heat tranfer analysis
Heat flux
Temperature
Heat tranfer analysis
Nodal
displacement
Tool geometry updating
Nodal
displacement
Tool geometry updating Tool geometry updating
t
j-1
Time increment ∆t
(specified by user or
searched by program)
∆t
Nodal displacement
) , ( ) , ( ) , ( j i j j i j i
D t w w
r
&
r
⋅ ∆ ⋅ =
Time increment ∆t
(specified by user or
searched by program)
∆t ∆t
Nodal displacement
) , ( ) , ( ) , ( j i j j i j i
D t w w
r
&
r
⋅ ∆ ⋅ =
Fig. 2.2 Approach and procedure of tool wear estimation

The study process for turning operation includes:

Stage 1: chip formation analysis
A new chip formation modeling method for continuous steady state chip formation is
developed. It can simulate the entire chip formation process from initial chip
Objective And Approach 33
formation, chip growth to steady state by making use of Arbitrary Lagrangian Eulerian
technique in ABAQUS/Explicit.
Stage 2: heat transfer analysis
In order to save the calculation time, the temperature distribution in the cutting tool at
thermal steady state is studied by performing pure heat transfer analysis. The
concerned modeling problem is discussed.
Stage 3: tool wear estimation modeling
Through the previous stages, normal stress, sliding velocity and tool temperature at
steady state can be obtained. Then the tool wear estimation modeling is studied. It
includes the calculation of wear rate at steady state, the searching method of a
suitable cutting time increment, the calculation of nodal displacement due to wear
and tool geometry updating.

The study process for milling operation includes:

Stage 1: chip formation analysis
The chip formation modeling method in milling operation is studied. It simulates the
chip formation process in the first milling cycle.
Stage 2: heat transfer analysis
In order to analyze the variation of tool temperature in the further milling cycles, the
cooling of the workpiece is studied, then pure heat transfer analysis of the tool is
performed for several milling cycles.
Stage 3: tool wear estimation modeling
Through the previous two stages, normal stress, sliding velocity, and tool
temperature can be obtained. Then the tool wear estimation modeling is studied. It
includes the calculation of average wear rate in one selected milling cycle, the
searching method of a suitable cutting time increment, the calculation of nodal
displacement due to wear and tool geometry updating.

Chip Formation Simulation Technology 34
Chapter 3 Chip Formation Simulation Technology



3.1 Introduction

Optimisation of the cutting process requires comprehensive knowledge about the
relation between cutting process and the combination of cutting parameters, cutting
tool and workpiece. In this chapter, chip formation process is simulated using
commercial FEM code, ABAQUS/Explicit.

3.1.1 Explicit Algorithm In Chip Formation Simulation

The chip formation simulation is performed using explicit method. In the simulation
the entire cutting process is discretized into many small time increments. In every
small time increment, dynamic and thermal analysis procedures are based on the
implementation of an explicit integration rule.

3.1.1.1 Dynamic Analysis Procedure

Dynamic analysis procedure is performed with the following algorithm.
• Nodal calculation
Accelerations are calculated by satisfying the dynamic equilibrium at the beginning of
the increment:

) (
) ( ) (
1
) ( i i i
I P M u − =

& & (3.1)

where is the acceleration at the beginning of the increment i,
) (i
u& &
M is the diagonal or lump mass matrix,
) (i
P is externally applied load,
and is internal load.
) (i
I
Then the accelerations are integrated through time using the central differential rule.

) (
) ( ) 1 (
)
2
1
( )
2
1
(
2
) (
i
i i
i i
u
t t
u u & & & &
∆ + ∆
+ =
+
− +
(3.2)
Chip Formation Simulation Technology 35

The velocities are integrated through time.

)
2
1
(
) 1 ( ) ( ) 1 (
+
+ +
∆ + =
i
i i i
u t u u & (3.3)

• Element calculations
Element strain increment, ε d , is computed from the strain rate, ε& , which is decided
according to the velocity of nodes.
Then stress, σ , is computed from the material constitutive equation.

) , , ( θ ε ε σ & f = (3.4)

3.1.1.2 Thermal Analysis Procedure

In the chip formation analysis, the stress analysis is dependent on the temperature
distribution and the temperature distribution depends on the stress solution. Fully
coupled thermal-stress analysis is employed.
In the analysis, heat transfer equations are integrated using the explicit forward
difference time integration rule.

N
i i
N
i
N
i
t
) ( ) 1 ( ) ( ) 1 (
θ θ θ
&
+ +
∆ + = (3.5)

where is the temperature at node N.
N
θ
The values of are computed at the beginning of the increment by
N
i) (
θ
&

) ( ) (
) ( ) (
1
) (
J
i
J
i
NJ N
i
F P C − =

θ
&
, (3.6)

where
NJ
C is the lumped capacitance matrix;
J
P is the applied nodal source vector;
J
F is the internal flux vector.
The explicit integration rules are realized in both dynamic and thermal analysis
procedures by using lumped mass matrix and capacitance matrix. The heat transfer
Chip Formation Simulation Technology 36
and mechanical solutions are obtained simultaneously by an explicit coupling.
Therefore no iterations or tangent stiffness matrices are required.

3.1.2 Stability Limit

The central difference and forward difference integrate constant accelerations,
velocities and temperature increments per unit time. In order to produce accurate
result, the time increment must be quite small so that the integrated variables are
nearly constant during an increment. The time increment must be smaller than a
stability limit otherwise the solution becomes numerically unstable. For coupled
thermal-stress analysis, the stability limit can be calculated by

)
2
,
2
min(
max max
λ w
t ≤ ∆ (3.7)

where is the highest frequency in the system of equations of the mechanical
solution response and
max
w
max
λ is the largest eigenvalue in the system of equations of the
thermal solution response.
The ABAQUS/Explicit solver supplies the default time incrementation scheme, which
is fully automatic and requires no user intervention.

Chip Formation Simulation Technology 37
3.2 Continuous Chip Formation Simulation

Continuous chip is very common when most ductile materials, such as wrought iron,
mild steel, copper, and aluminium, are machined. Cutting under these conditions is
steady-state process with steady chip shape, cutting force and temperature. Many
chip formation models were developed for these cutting conditions with different FE
codes. They are based on different approaches: Lagrangian or Eulerian.

3.2.1 Limitation Of The Existing Chip Formation Models

This continuous chip formation process cannot obtain very satisfactory simulation
result because of the limitations of the existing models using ABAQUS FE code:
• Most models take chip formation as a Lagrangian problem. Under the
consideration of reducing calculation time, the length of the workpiece is often
very small, only enough to produce a steady chip shape. If analysis of the
further cutting process is required, a longer workpiece has to be used in the
simulation, which increases calculation time direct proportionally.
• In most chip formation models, chip separation is realized by element removal
or node debond. A small crack is always created before tool edge, as shown in
Fig. 3.1. But evidences from cutting experiments show that for the quasi-
continuous chip formation that takes place in machining ductile materials
under favourable cutting conditions, crack occurs along the shear direction
[Didj-97]. Only for discontinuous chip formation and chip formation with build-
up edge, crack is found ahead of the cutting edge.









(a) Node debond [McCl-02] (b) Element removal [Ng-02a]
Fig. 3.1 Cracks formed before tool edge in the simulations
Chip Formation Simulation Technology 38

• The chip separation path is often predetermined instead of formed
automatically by the deformation of workpiece material under the cutting
action. When round edge cutting tools are used, the position of the
predetermined chip separation path has influence on the volume of material to
be cut away. Whether the cutting force components, residual stress, etc
change with the position of the separation path or not needs analysing as well.
• A chip separation criterion is necessary. In ABAQUS, chip separation criterion
is given by defining material failure criterion. The failure parameters are
material dependent and different parameters are required for Johnson-cook
and other material models. Although many material constitutive models are
provided for the commonly used material by literatures, the failure parameters
are seldom given as well. Failure parameters relate to the successful
implementation of chip formation simulation. They should be determined
experimentally instead of being given at random. This limits the usage of
many material constitutive models.
• When the steady-state chip formation process is modelled as a Eulerian
problem with ABAQUS/Explicit, as reported by Arrazola et al [Arra-02], the
conflict between the cutting time to reach steady state and the length of the
workpiece model limited by calculation time is solved. Steady-state analysis is
performed by prolonging the cutting time without increasing the length of the
workpiece. In addition no shear failure criterion is required. But an initial chip
geometry must be given according to experiment or experience in machining.
When the given initial chip geometry is not suitable, chip geometry cannot
adapt itself when the chip tends to swell up or to shrink too much from the
given initial geometry, as shown in Fig. 3.2.


(a) When the given initial chip is thinner than in experiment, the chip swells up
from the initial geometry
Chip Formation Simulation Technology 39


(b) When the given initial chip is thicker than in experiment, the chip shrinks from
the initial geometry
Fig. 3.2 Problems in the chip formation analysis with an unsuitable initial chip
geometry [Arra-02]

3.2.2 Advantages Of The New-developed Chip Formation Model

Due to the limitations mentioned above, a new continuous chip formation model is
developed with ABAQUS/Eplicit. This model has the following advantages:
• Workpiece geometry in the model only stands for a control area. Workpiece
material is unlimited and flowing through this control area continuously. User
can prolong the cutting time without changing the size of the control area.
• Chip separation is realized with adaptive meshing technique supplied by
ABAQUS/Explicit; no shear failure criterion or material failure parameters are
required. Most of the material constitutive models can be used in this model.
• Chip separation is performed by the deformation of the workpiece material,
instead of forming crack along a predetermined path.
• Good contact is maintained in the cutting tool edge area. No obvious crack is
formed.
• This model is especially suitable to simulate the cutting process with round
edge tool or chamfered tool.
• No initial chip geometry is required. Chip formation simulation includes the
entire process from initial chip formation, chip growth, to steady state.

3.2.3 Adaptive Meshing Technique In ABAQUS/Explicit

This model is developed based on adaptive meshing technique in ABAQUS/Explicit.
It combines the features of pure Lagrangian and Eurerian analysis and can be used
Chip Formation Simulation Technology 40
to both Langrangian, e.g. initial chip formation, and Eulerian problems, e.g. steady-
state chip formation.

3.2.3.1 Boundary Region Types

Adaptive meshing is performed in adaptive meshing domains, which can be either
Lagrangian or Eulerian.






E: Eulerian boundary region
S: Sliding boundary region
L: Lagrangian boundary region
Material flow direction

S
L
L
E
E
S
Boundary region edge
Workpiece
Tool
Chip
Fig. 3.3 Boundary regions in chip formation model

The boundary regions of the adaptive meshing domain can be either Lagrangian,
sliding or Eulerian. In Lagrangian boundary region, the mesh is constrained to move
with the material in the direction normal to the surface of the boundary region and in
the directions perpendicular to the boundary region edge. In sliding boundary region,
the material is constrained to move with the material in the directions normal to the
boundary region, but it is completely unconstrained in the directions tangential to the
boundary region. Eulerian boundary regions can be defined only on the exterior of a
geometry model and the material flows across the boundary, as in a fluid flow
problem. Mesh on the Eulerian boundary regions are fixed in space using spatial
mesh constrains, and material flow velocity across the boundary is defined by
boundary conditions. When the adaptive meshing domain is Eulerian type, only
sliding and Eulerian boundary regions can be defined, for example the workpiece in
Fig. 3.3.

Chip Formation Simulation Technology 41
3.2.3.2 Geometry Features

On boundary regions where the angle θ between the normals on adjacent element
faces is greater than an initial geometric feature angle θ
I
given by the user, geometry
features are detected initially. Adaptive meshing cannot be performed well across
such geometry features because the nodes cannot move across the geometry
features unless they flatten. Therefore it is sometimes necessary to deactivate the
geometry features by defining a greater initial geometry angle.







θ
n n
n
n
θ ≤ θ
I

θ > θ
I

(a) (b)
Fig. 3.4 Geometry Features (a) Geometry feature is detected, no mesh flow is
permitted past the corner (b) No geometry feature is detected, mesh flow is permitted


3.2.3.3 Curvature Refinement

During adaptive meshing, mesh-smoothing algorithms based on minimizing element
distortion tend to reduce the mesh refinement in area of concave curvature,
especially as the curvature evolves. Having sufficient mesh refinement near highly
curved boundaries is very important to model the detail of the chip shape near the
chip separation area. To prevent the natural reduction in mesh refinement of areas
near evolving concave curvature, solution-dependent meshing is used to focus mesh
gradation toward these areas automatically by defining the curvature refinement
weight α
c
a high value, for example, 1 =
c
α .

Chip Formation Simulation Technology 42

(a) α
c
=0 (b) α
c
=1
Fig. 3.5 Effect of curvature refinement

3.2.4 Analysis Steps

The entire continuous chip formation process is performed with a complete modeling
procedure from initial chip formation to the realization of steady state, which consists
of three analysis steps, including initial chip formation, chip growth, and steady-state
chip formation as described in detail in the following parts. The first two analysis
steps supply steady chip geometry for the steady-state chip formation analysis step.
During all the chip formation steps, coupled thermo-stress analyses are performed.
The simulated cutting condition is given in Table 3.1.

Table 3.1: Cutting condition
Cutting type Orthogonal cutting, turning operation, dry cutting
Work material Mild carbon steel AISI1045
Tool material Uncoated carbide WC-Co
Tool geometry
° − = 7
o
γ , ° = 7
o
α , mm r 0245 . 0 =
ε
Cutting parameters
min / 300m v
c
= , , mm a
p
2 = r mm f / 145 . 0 =

In the finite element model, the workpiece has a size of 0.6×3.2mm, which is meshed
with 4-node bilinear coupled temperature-displacement plain strain elements
CPE4RT. In order to save calculation time, only the part of the cutting tool near the
cutting edge is included in the chip formation modelling. Moreover in the first two
steps, the cutting tool is defined as a rigid body, whereas in the last analysis step the
cutting tool is modelled as a deformable body in order to obtain better analysis result
and more comprehensive analysis data.
Chip Formation Simulation Technology 43

3.2.4.1 Initial Chip Formation

This analysis step aims to form initial chip geometry. This chip formation process is
modelled as a Lagrangian problem.








(a) Initial geometry and mesh (b) Initial chip geometry formed at t=0.18ms
c
v
E
D
C
B
A
Fig. 3.6 Initial chip formation analysis

The initial workpiece geometry is designed to have a concave at the top-right corner
under the consideration of seeding more nodes along the concave surface (see Fig.
3.6(a)). The boundary of the workpiece consists of only Lagrangian boundary
regions. During the initial chip formation process, the Lagrangian boundary region on
the top surface of the workpiece traces the chip material continuously and forms the
shape of the chip.
At the cutting edge, chip material separates with workpiece material. Only very fine
mesh can show exactly the shape of this area. In order to maintain the mesh
refinement in this area while the chip formation process continues, the initial
geometry feature angle should be given a suitable value. According to this value, the
four corner points A, B, C and D in Fig. 3.6(a) can be detected as geometry features
but corner point E and the workpiece nodes on the surface ED should not be taken
as geometry features. The value is defined by calculating the angle between the
normals on adjacent element faces in chip separation area, as shown in Fig. 3.7,
finding out the maximum angle value, then taking an angle value between this
maximum angle value and 90 deg (because the angles at point A, B, C and D are
about 90 deg). Adaptive meshing can be performed on any boundary regions except
point A, B, C and D.
i
θ

Chip Formation Simulation Technology 44

Workpie
Tool
θ
i








Fig. 3.7 Determination of initial geometry feature angle

In addition solution-dependent meshing is used to focus mesh toward the chip
separation areas automatically by setting the curvature refinement weight α
c
to unity.
At the beginning the cutting tool is at the right side of the workpiece. The workpiece is
fixed and the cutting tool is moving in the negative x-direction
1
. With the cutting tool
advancing into the workpiece, elements along the concave surface extend and
compose the outside surface of the chip. After 0.18ms an initial chip is formed. Fig.
3.6(b) shows the mesh after the initial chip is formed.

3.2.4.2 Chip Growth

After the initial chip is formed, chip growth analysis step is performed. This analysis
step aims at forming steady chip geometry. In this step the chip formation process is
treated as a Eulerian problem.
A user program is developed with Python language. It reads the variables about
nodal coordinate, nodal temperature, etc of the workpiece and the cutting tool from
the selected time point of the initial chip formation analysis step, when a desirable
initial chip shape is produced. Then it writes them into the model files of the chip
growth analysis step, including node input file, initial temperature input file, etc.
Therefore the initial state of the workpiece and the cutting tool in this step remains
the state at the selected time point of the former analysis step. For example, in this
simulation the initial state information is read from the former analysis step at 0.18ms.

1
In all figures of this paper, x-direction is pointed to the right side and y-direction to the top of the page.

Chip Formation Simulation Technology 45
In this step, the relative movement between the cutting tool and the workpiece is
performed by the movement of workpiece material. The cutting tool is fixed in space.
The workpiece mesh in Fig. 3.8 represents only a control area. The left and right
boundary of the control area are defined as Eulerian boundary regions, whose mesh
is fixed in x-direction by using mesh constrain definition, but material flows into the
control area continuously from the left boundary at the cutting speed and flows out of
from the right boundary, as indicated with the small arrows in Fig. 3.8(a). The top and
bottom boundary are sliding boundary regions. The movement of the mesh on the
bottom boundary is constrained in y-direction, indicated with small triangle in Fig.
3.8(a). But the movement of the mesh on the top boundary is not constrained; the
mesh will adjust itself to fit in with the developing chip geometry.
Initial geometry feature angle is defined in the same way as explained above. The
curvature refinement weight α
c
is set to unity.
Fig. 3.8(b) shows that the chip is growing with the material flowing into the control
area.


(a) Initial geometry and mesh (b) Chip growth at t= 0.09ms
Fig. 3.8 Chip growth analysis

3.2.4.3 Continuous Steady-state Chip Formation

In the second analysis step, with the chip growing to a certain extent, the mesh of the
chip extends too much in the direction of chip growth so that adaptive meshing
cannot solve mesh distortion problem any more. The steady-state chip formation
analysis step is designed for simulating the further cutting process. In this step, the
cutting process is treated as a Eulerian problem as well. According to the definition of
Chip Formation Simulation Technology 46
the workpiece mesh, two methods can be used in the continuous steady-state chip
formation step: mesh modification and model regenerating.

Method 1: Mesh Modification

During the growth of the chip in the second analysis step, the chip geometry near the
chip root becomes steady since a certain time point. The state of the workpiece and
the cutting tool at this time point is written into the model files of steady-state cutting
analysis step. For example, the mesh in Fig. 3.9(a) is read from chip growth analysis
step at t=0.09ms, see Fig. 3.8(b).
In order to allow the chip to flow out of the mesh and grow unlimitedly instead of grow
visually with the mesh, the mesh at the top boundary of the chip is defined as a
Eulerian boundary region. The coordinates of the nodes on this boundary are
adjusted to locate these nodes along a vertical line in order to facilitate the definition
of mesh movement constrain in x-direction, as shown in Fig. 3.9(a). The mesh of the
boundary is not constrained in y-direction; its position and size can adjust with the
chip automatically in y-direction.
The definition of boundary regions and conditions for the other part of the workpiece
control area are similar to those in the second analysis step. Initial geometry feature
angle and curvature refinement weight α
c
are defined in the same way as explained
above.
Because the cutting tool is a deformable body, its movement is fixed by defining
constraint in x-direction at the right boundary and in y-direction at the top boundary.
Fig. 3.9(b) shows the formed mesh at 1ms.









(a) Initial geometry and mesh (b) Mesh at t=1ms
Fig. 3.9 Mesh modification
Chip Formation Simulation Technology 47

Method 2: Model Regenerating

This method is especially important when the cutting tool has some special geometry,
e.g. crater wear, and good contact between the workpiece and the cutting tool in
these areas is desired.
ABAQUS/Explicit supplies only r-adaptivity. When the mesh concentrates in the
cutting edge area according to solution-dependent meshing rule, the mesh in other
area becomes coarse. But sometimes fine mesh along the whole tool-workpiece and
tool-chip interface is required. Model regenerating supplies an approach to improve
the contact problem.
The information necessary for model regenerating includes chip thickness and tool-
chip contact length. They can be obtained from the former two analysis steps, initial
chip formation and chip growth analysis step, in which a steady chip shape is formed.
The regenerated workpiece model in Fig. 3.10(a) has a chip connected to the
workpiece. The chip is 0.3mm thick, and 0.5mm long. The length of the chip should
be determined carefully. It is larger than the tool-chip contact length. But when the
chip is too long, it will complicate the definition of mesh constraint.
The mesh at the top boundary of the chip is defined as a Eulerian boundary region.
The mesh movement is constrained in y-direction. The mesh position and size in x-
direction will be adjusted with the chip automatically.









(a) Initial geometry and mesh (b) Mesh at t=1ms
Fig. 3.10 Model regenerating

Chip Formation Simulation Technology 48
Very fine mesh is given along the entire tool-chip interface. This ensures good
contact between the cutting tool, the workpiece and the chip throughout the entire
steady-state analysis process, as shown in Fig. 3.11.


Contact
problem
Contact
problem

(a) Contact problem is created by using mesh modification


(b) Contact is improved by using model regenerating
Fig.3.11 Contact status along tool-chip interface

3.2.5 Results & Discussion

3.2.5.1 Stress Analysis

Fig. 3.12 shows the stress distribution in the three analysis steps. Maximum stress is
located in the primary shear zone; the workpiece material undergoes serious shear
plastic deformation in primary shear zone and becomes chip.
Chip Formation Simulation Technology 49
Further the underside of the chip undergoes high stress because of the contact and
friction with the tool face when sliding away.
The newly formed machined surface has contact and friction with the round edge and
sometimes a small part of the flank face. This results in a high stress in the workpiece
material beneath the cutting tool edge.


(a) Initial chip formation analysis, t=0.18m (b) Chip growth analysis, t=0.3ms


(c) Steady state analysis, t=1ms
Fig. 3.12 Stress distribution (MPa)

In steady state analysis step, the cutting tool is modelled as a deformable body; very
high stress is observed in the small part of the cutting tool directly under the tool-chip
contact area.

3.2.5.2 Plastic Strain Analysis

Fig. 3.13 shows that the workpiece material undergoes serious plastic deformation in
primary shear zone. The material in the chip underside deforms plastically again
Chip Formation Simulation Technology 50
under the pressure and friction of the cutting tool face. This results in higher plastic
strain formed in the chip underside than in the other part of the chip.


(a) Chip growth analysis step, t=0.09ms


(b) Mesh modification, t=1ms (c) Model regenerating, t=1ms
Fig. 3.13 Equivalent plastic strain distribution

In Fig. 3.13(a), the top of the chip has no plastic strain because only the geometry of
this part is imported from the initial chip formation analysis step, not including the
created plastic strain. In the steady-state analysis, the two model methods create
similar plastic strain field, and only the maximum values are different which is caused
by the different contact condition due to element size in the workpiece model and the
difference between deformable cutting tool and rigid body cutting tool.

3.2.5.3 Strain Rate

Fig. 3.14 shows the distribution of strain rate, which is defined as solution SDV9 by
material subroutine. Under the example cutting condition, the maximum strain rates
Chip Formation Simulation Technology 51
distribute along the primary shear zone, especially in the areas close to the cutting
tool edge and the free surface of the workpiece, and reach up to 10
5
, which is
assumed as typical maximum strain rate in conventional machining [Arnd-73].


(a) Chip growth step, t=0.09ms

(b) Mesh modification, t=1ms (c) Modal regenerating, t=1ms
Fig. 3.14 Strain rate distribution

3.2.5.4 Temperature Analysis

In Fig. 3.15(b), within the cutting process of 1ms cutting temperatures at most of the
tool face nodes in the tool/chip interface, i.e. the highlighted nodes in Fig. 3.15(a), is
reaching steady values, while at the tool face nodes inside the cutting insert, the
highlighted nodes in Fig. 3.15(c), the temperature is still climbing, as shown in Fig.
3.15(d). This means that thermal steady state is not realized in the whole cutting tool.

Chip Formation Simulation Technology 52

(a) Position of monitored tool face nodes (b) Temperature progress

Node 429
Node 497
Node 588
Node 492
Node 507
(b) Position of nodes inside the tool (d) Temperature history
Fig. 3.15 Temperature history of tool nodes at steady-state chip formation analysis
step

Fig. 3.16 shows the temperature distribution at 1ms. The highest temperature is at
the rake/chip interface and most part of the tool is still at room temperature.

Chip Formation Simulation Technology 53

Fig. 3.16 Temperature distribution at t=1ms of steady-state chip formation analysis
step

3.2.5.5 Verification With Experimental Data

By adding the reaction force component in the same direction at all constrained
nodes of the cutting tool and then taking the negative value, the cutting force
components F
c
and F
t
in the continuous steady-state chip formation step are
obtained. Fig. 3.17 shows that the cutting force components change within a very
narrow range from 0.7ms, and it is deemed that the mechanical steady state is
realized.


Fig. 3.17 Cutting force progress (under cutting condition: v
c
=300m/min, a
p
=2mm,
f=0.145mm/r)
0
100
200
300
400
500
600
700
0,0000 0,0004 0,0008 0,0012
Time [s]
C
u
t
t
i
n
g

a
n
d

t
h
r
u
s
t

f
o
r
c
e

[
N
]
Fc
Ft
Chip Formation Simulation Technology 54

In Fig. 3.18, the cutting force values obtained from the simulation with
ABAQUS/Explicit code are compared with the experiment data [Feve-01] and the
result from other software including Third Wave, Deform2D and Oxcut-F [Söhn-03].
The results from ABAQUS include ABQ-f030I, ABQ-f030S, ABQ-f048I and ABQ-
f048S. I means the result is obtained from initial chip formation step, while S from
continuous steady state chip formation step. Two different frictional coefficient values
are used. One is 0.30, another is calculated according to Eq. 1.3, and the value is
0.48.
Fig. 3.18 shows that the cutting force components obtained from initial chip formation
step and continuous steady state chip formation step have no great difference. The
cutting force components created in the chip formation analysis with the frictional
coefficient of 0.48 gives the better result than other FEM code, the prediction error of
cutting force F
c
is about 2%, and the error of thrust force F
c
is about 5%.

2%
5%
2%
5%
2% 2%
5% 5%











Fig. 3.18 Comparison of cutting force (under cutting condition: v
c
=300m/min,
a
p
=2mm, f=0.145mm/r)

3.3 Chip Formation Simulation For Milling Operation

In milling operation, cutting action is discontinuous and the chip produced is
discontinuous. The modelling method developed for the continuous chip formation is
not suitable to simulate chip formation process in milling operation. Therefore a
different modelling method is introduced in the following part.
Chip Formation Simulation Technology 55

3.3.1 Chip Separation

In every milling cycle, the produced chip will separate with the newly produced
workpiece surface without any connection when the cutting tool disengages from the
workpiece. Hence the adaptive meshing technique in ABAQUS/Explicit cannot be
used as a chip separation tool any more. In this method, chip separation is realized
by defining shear failure criterion.

3.3.1.1 Shear Failure Criterion

The shear failure model is based on the value of the equivalent plastic strain at
element integration points; when the equivalent plastic strain reaches the strain at
failure
pl
f
ε , then the damage parameter exceeds 1, material failure takes place. If
at all the integration point material failure takes place, the element is removed from
the mesh. The damage parameter, , is defined as
w
w


|
|
.
|

\
|

=
pl
f
pl
w
ε
ε
, (3.8)

where
pl
ε ∆ is an increment of the equivalent plastic strain. The summation is
performed over all increments in the analysis.
There are two methods to define the strain at failure. For Johnson-cook plasticity
model, the stain at failure is given according to Eq. 3.9.

( θ
ε
ε
ε
ˆ
1 1 exp
5
0
4 3 2 1
d In d
q
p
d d d
pl
pl
f
+
(
¸
(

¸

|
|
.
|

\
|
+
(
¸
(

¸

|
|
.
|

\
|
+ =
&
&
) (3.9)

where strain at failure,
pl
f
ε , is dependent on a nondimensional plastic strain rate,
0
ε ε &
&
pl
; a dimensional pressure-stress ratio, q p (where p is the pressure stress and
is the Mises stress); and a nondimensional temperature, (defined as 0, q θ
ˆ
( ) ( )
transition melt transition
θ θ θ θ − − , or 1 depending on the temperature range). Stain at
failure is defined by giving the failure parameters .
5 1
d d −
Chip Formation Simulation Technology 56
For Mises plasticity model, strain at failure or the dependencies of strain at failure on
strain rate, pressure/stress ratio and temperature are given directly in tabular form in
the data line.

3.3.1.2 A Numerical Method To Determine Strain At Failure

Normally, equivalent plastic strain at failure,
pl
f
ε , is obtained by using experimental
methods. For example, Bacaria et al determined failure parameters d by
performing tensile and torsion tests [Baca-00]. Ng et al integrated orthogonal tests
with some analytical equations in metal cutting theory to define the dependency of
the equivalent plastic strain at failure
5 1
d −
pl
f
ε on the plastic strain rate
pl
ε
&
; the
hydrostatic stress p and temperature [Ng-02b].
By employing the continuous chip formation analysing methods, it is possible to
determine stain at failure without making any experiment.
Observing the movement of material points on the chip underside and the machined
workpiece surface in steady-state chip formation process, we can find a separation
area of the workpiece material. For example, in Fig. 3.19, the separation area is
between Node 13 and Node 16. The material above the separation area moves
upwards into the chip and the material below the separation area moves downwards
to join in the machined surface.

Node 13
Node 16
Node 15
Node 14
Node 13
Node 16
Node 15
Node 14

Fig. 3.19 Velocity of material points at workpiece nodes on the chip underside and
the machined surface (the arrows shows the velocity vectors)

Chip Formation Simulation Technology 57
According to the sliding velocities of the workpiece material points along tool-chip
interface, a more exact position of the separation area can be defined in Fig. 3.20.
The directions of sliding velocities of the material points in the area between Node 15
and Node 16 change. It can be assumed that material failure is taking place in this
area. The equivalent plastic strain between Node 15 and Node 16 gives a value
range from 2.25 to 2.7 to strain at failure.

Node16
Node15
Node16
Node15
Node16
Node15

(a) Monitored points
Node16
Node15
Node16
Node15
Node16
Node15

Node16
Node15
Node16
Node15
Node16
Node15

(b) Sliding velocity of monitored points (c) Equivalent plastic strain of monitored
points
Fig. 3.20 Determination of strain at failure

By varying cutting parameters or tool geometry, the dependency of strain at failure on
temperature, strain rate, pressure, etc can be studied.
In the following part of this chapter, strain at failure for mild carbon steel is set to 2.5.

Chip Formation Simulation Technology 58
3.3.2 Chip Formation Modeling

This chip formation modeling method is explained by taking a milling case as an
example in which an uncoated carbide tool is used to machine mild steel CK45. The
cutting condition is given in Table 3.2.

Table 3.2 Cutting condition
Cutting type Orthogonal cutting, milling operation, dry cutting
Work material Mild carbon steel CK45
Tool material Uncoated carbide WC-Co
Tool geometry
° = 7
o
γ , ° = 7
o
α
Cutting parameters
min / 600m v
c
= , , a , mm a
e
2 = mm
p
1 = r mm f
z
/ 2 . 0 =

The diameter of the milling tool is 125mm. In order to reduce the calculation time,
only a small part of the workpiece and the cutting insert is included in the model. Fig.
3.21 shows the initial geometry, mesh and assembly of the workpiece and the cutting
insert.











0.4
0.2
Rotation center
W
o
r
k
p
i
e
c
e
Insert
6
2
.
5
2
0.4
0.2
Rotation center
W
o
r
k
p
i
e
c
e
Insert
6
2
.
5
2
0.4
0.2
Rotation center
W
o
r
k
p
i
e
c
e
Insert
6
2
.
5
2
0.4
0.2
Rotation center
W
o
r
k
p
i
e
c
e
Insert
6
2
.
5
2
Fig. 3.21 Initial geometry, mesh and assembly of the tool and the workpiece in chip
formation analysis

The workpiece is simplified as a small segment of a ring; whose outside radius is
62.7mm and inside radius 62.3mm. The centre of the ring is positioned at the rotation
centre of the cutting insert. The workpiece is 2mm high. The extension of its upper
Chip Formation Simulation Technology 59
surface passes through the center of the ring and the lower surface is parallel to the
upper surface. The workpiece is discretized with a mesh composed of CPE4RT
elements, and local fine mesh is given along the moving path of the cutting edge
because of very high gradients of solutions in this area, such as temperature, stress,
etc.
The cutting insert in the model includes only the part near the cutting edge, which is
discretized with CPE4RT elements. The cutting insert is modelled as a deformable
body in order to obtain all the necessary cutting process variables for the latter study
on tool wear.
The chip formation process is treated as a Lagrangian problem. Every boundary
segment of workpiece is defined as a Lagrangian boundary region.
There are different ways to assign shear failure criterion to form different shape of
chips. Ng et al designed two different kinds of shear failure criteria, one criterion is
assigned to a line of element along the moving path of the cutting edge to separate
the chip from the workpiece; another criterion is assigned to part of the chip material
to generate cracks in order to simulate serrated chips [Ng-02b]. Bacaria defined only
one material shear failure model for the whole workpiece material [Baca-00]. In the
model the shear failure criterion is integrated with a material model designed
specially for the workpiece material CK45 and assigned to the whole workpiece.
One milling cycle takes 39.27ms. In each milling cycle, cutting phase takes 0.2ms
and cooling phase takes 39.07ms. The chip formation analysis is performed for
0.5ms, covering the whole cutting phase and 0.3ms of the later cooling phase.

3.3.3 Result & Discussion

3.3.3.1 Stress Analysis

At the beginning, the cutting insert is at the bottom of the workpiece, and there is no
contact with the workpiece. With the tool rotating in clock-wise direction, the cutting
insert engages in the fixed workpiece. A small chip is formed, and the contact
between the chip and the cutting insert concentrates in a small area near the cutting
edge, which results in a high stress in this area, as shown in Fig. 3.22(a).
Fig. 3.22(b) shows that the primary deformation zone has the maximum stress in the
workpiece.
Chip Formation Simulation Technology 60
In Fig. 3.22(c), the cutting insert is disengaging the workpiece. The workpiece
material to be cut away deforms seriously under the pressure of the cutting insert and
protrudes from the original top surface, which provides a possibility for burr formation.
But after crack generates, it propagates along the direction of maximum stress
deeper and deeper into the workpiece material, instead of along the moving path of
the cutting edge.
During the entire cutting phase, the cutting edge is bearing higher stress than other
part of the insert because of positive rake angle and very sharp tool edge.


(a) t= 0.025ms (b) t= 0.100ms

(c) t=0.175ms (d) t=0.200ms
Fig. 3.22 Stress field (Mpa) in the chip formation analysis

Chip Formation Simulation Technology 61

3.3.3.2 Cutting Temperature

The predicted temperatures generated during chip formation process are shown in
Fig. 3.23. The heat is generated mainly in two zones, the shear zone and the chip
underside sliding along the tool face. The obvious temperature increment take place
in shear zone, then the chip underside is heated again to a higher temperature by the
friction with the tool face, as shown in Fig. 3.23(b), 3.23(c) and 3.23(d). In addition,
when chip breakage takes place in Fig. 3.23(a), local high temperature is formed.










(a) t= 0.025ms (b) t= 0.100ms


(c) t=0.175ms (d) t=0.200ms
Fig. 3.23 Temperature distribution (in Kelvin) in the chip formation analysis
Chip Formation Simulation Technology 62

3.3.3.3 Cutting Force Analysis

In order to make the cutting insert rotate as a deformable body, its bottom is pinned
on and rotates with the rotation center point. Therefore cutting force is exerted on the
rotation center point. Fig. 3.24 shows the cutting force progress during the cutting
process. Because the cutting insert has exited from the workpiece and no contact
with the workpiece any more after 0.2ms, cutting force components in x-direction and
y-direction are reducing to zero. The ‘noise’ of the cutting force signal is caused by
the removal of the elements; they reach the shear failure criterion and then stresses
in these elements are set to zero, which result in the fluctuation of cutting force. This
is different from ‘noise’ observed in continuous chip formation analysis, which always
appear when the element of workpiece is coarser than tool element, and contact
problem results in ‘noise’ of cutting force signal.

C
u
t
t
i
n
g

f
o
r
c
e


[
N
]
C
u
t
t
i
n
g

f
o
r
c
e


[
N
]

Fig. 3.24 Cutting force progress during the cutting process
Chip Formation Simulation Technology 63
3.4 Summaries & Conclusion

Two different chip formation modeling methods are designed to simulate the chip
formation process in milling operation and turning operation.
Chip formation model for turning operation is designed to simulate the whole cutting
process including initial chip formation, chip growth and steady state. No experiment
is required to get material failure parameters or steady chip geometry. Chip
separation is formed automatically by using ALE technique supplied by
ABAQUS/Explicit. In order to get good contact between the chip and the tool face
even when a serious crater wear is formed on rake face, model regeneration method
is suggested to update and refine the mesh of the workpiece, especially at the tool-
chip interface.
With this complete model, the normal tool geometry such as blunted, chamfered and
worn cutting tool can be used in the chip formation model.
The calculation time to reach steady state is relative short comparing with the chip
formation model in which the chip formation is taken as a pure Lagrangian problem.
Especially when it is used in tool wear estimation, the total calculation time to reach
tool reshape criterion is reduced sharply because except the initial chip formation and
chip growth are run only one time and then with the tool wear increasing, only steady
state analysis step is necessary.
Chip formation in milling operation is modeled by introducing the shear failure
criterion because of the intermittent cutting process. The shear failure criterion is
used to the entire workpiece. This model is expected to have a wider application
because it can model various chip type, such as serrated chip, when the suitable
material constitutive and material failure model are provided.


Heat Transfer Analysis In Metal Cutting 64
Chapter 4 Heat Transfer Analysis In Metal Cutting

4.1 Introduction

When the cutting process is simulated using chip formation analysis, the cutting time
is normally limited to a short time, because coupled thermal mechanical analysis is
too expensive. For example, in the former chapter, at the end of the chip formation
analysis in turning operation, temperatures at nodes inside the cutting tool are still
climbing while those at tool-chip interface nodes approach steady state. It is
concerned how the temperature distributes in the cutting tool finally. For milling
operation, the chip formation analysis is only carried out in the first milling cycle. It is
important for the correct calculation of tool wear how the tool temperature changes in
the further milling cycles.
Therefore pure heat transfer analysis is performed after chip formation analysis for
the further cutting process in order to get such knowledge at a low calculation cost.

4.2 General Considerations

4.2.1 Geometry And Mesh

In the heat transfer analysis, only a single object is considered, for example only the
cutting tool or the workpiece. Otherwise, the simulation will become complex because
of the interaction between the cutting tool and the workpiece.
ABAQUS uses some Eulerian elements, diffusive elements, which have only
temperature degrees of freedom, to model convective heat transfer. Diffusive
elements are provided in one, two or three dimension. Interpolation can be first-order
and second-order [ABA-01b]. Two-dimensional first-order four node diffusive
element, DC2D4, is chosen to discretize the geometry of the studied object in the
heat transfer analysis because quantities of DC2D4 are integrated at nodes and this
simplifies the design of heat flux subroutine by importing heat flux at nodes of chip
formation model directly into integration points of heat transfer model as basic data
for the calculation of the current heat flux. In addition, the error caused by the
conversion from nodal value to integration point value is avoided during the
importation of temperature data. These advantages are based on the conservation of
Heat Transfer Analysis In Metal Cutting 65
the node label, element label and element connectivity of the chip formation analysis
model.

4.2.2 Heat Flux

In the cutting phase the cutting tool is heated by the heat flux acted on the tool-chip
and tool-workpiece interface. The total heat flux is composed of frictional heat flux
and conductive heat flux . Frictional heat flux is created due to the sliding friction
between the workpiece material and the tool face. The amount of frictional heat flux
into the cutting tool is calculated by Eq. 4.1.
f
q
c
q

( )
s
r
f q ν ητ ⋅ − = 1 (4.1)

where
τ is the frictional stress;
s
ν is the sliding velocity;
η specifies the fraction of mechanical energy converted into thermal energy;
f gives the fraction of the generated heat flowing into the workpiece.
Therefore frictional heat flux is influenced by chip form, sliding condition and contact
with the tool face.
Conductive heat flux is caused by the temperature difference of tool-chip and tool-
workpiece at the interface. It is governed by Eq. 4.2.

(
B A
c
k q θ θ − = ) (4.2)

where
c
q is the conductive heat flux crossing the interface from point A on the workpiece to
point B on the cutting tool;
k is the gap conductance;
θ is the nodal temperature on the surface.
Therefore conductive heat flux is temperature dependent.
Both heat flux components are varying from node to node and the basic nodal heat
flux data can be obtained from the chip formation analysis.
Heat Transfer Analysis In Metal Cutting 66

4.3 In Turning Operation

4.3.1 Modeling

In order to study on the temperature distribution of the cutting tool at steady state,
heat transfer analysis is performed after the chip formation analysis finishes.
Because in the orthogonal cutting experiment [Schm-02], the part of the cutting edge
engaged in the cutting is located in the center part circled in Fig. 4.1, the geometry
model of the cutting tool used in the two-dimensional heat transfer analysis is the
section created by intersecting the center area with a surface perpendicular to the
edge. It includes the part, which is surrounded by the rake face, flank face, bottom
face, and the surface of the central hole. The part of the tool used in the former chip
formation analysis steps is only the highlighted part, and the element label, node
label and element connectivity of this part in chip formation analysis steps remain
unchanged.









Heat convection
Heat radiation
Frictional heat
Conductive heat
Hole surface
Bottom face
Flank face
Rake face
Hole surface
Flank face
Rake face
Bottom face
area
R
Heat convection Heat convection
Heat radiation Heat radiation
Frictional heat
Conductive heat
Hole surface
Bottom face
Flank face
Rake face
Hole surface
Bottom face
Flank face
Rake face
Hole surface Hole surface
Bottom face Bottom face
Flank face Flank face
Rake face Rake face
Hole surface
Flank face
Rake face
Bottom face
area
Hole surface
Flank face
Rake face
Bottom face
Hole surface
Flank face
Rake face
Bottom face
area area area
RRR
A
A
A
A
A
A
A
A
Cutting
oom temperature
Cutting Cutting Cutting Cutting
oom temperature oom temperature oom temperature
Fig. 4.1 Geometry and mesh of the cutting tool used in heat transfer analysis, the
circled part is the part of the edge engaged in the cutting

Temperature data at the end of the steady-state chip formation analysis step is
imported and used as the initial temperature definition of the nodes in the highlighted
part. At other nodes the initial temperature is set to room temperature.
At the nodes on tool/chip interface heat flux is defined. In steady-state chip formation
analysis step, it is found that the total heat flux (heat passing through the tool-chip
and tool-workpiece interface per second) is changing as the cutting process
continues, as shown in Fig. 4.2. One component of the total heat flux, the total
Heat Transfer Analysis In Metal Cutting 67
frictional heat flux, reaches steady state in a short time once the chip gets into
contact with the tool face and restores to the steady sliding velocity, while another
component, the total conductive heat flux, is always decreasing within the entire
analysis period. At the end of analysis, as the temperature of cutting tool and
workpiece become steady, the decreasing rate is becoming lower and lower and
approaching a steady value.











-600000
-400000
-200000
0
200000
400000
600000
800000
0,0000 0,0003 0,0006 0,0009 0,0012
Time [s]
H
e
a
t

f
l
u
x

[
m
J
/
s
]
Total conductive heat f lux
Total f rictional heat f lux
Total Heat f lux
Fig. 4.2 Heat flux at tool-chip and tool-workpiece interface in steady-state chip
formation analysis step

Observation of nodal temperature of workpiece nodes at tool-chip interface shows
that the variation of nodal temperature at the end of the analysis is very small, as
shown in Fig. 4.3. Therefore thermal steady state is assumed in workpiece material
at the interface. Then the nodal conductive heat flux can be converted from Eq. 4.1
into Eq. 4.3

( ) ( )
B t B t B t A
c
s s s
k k q θ θ θ θ − + − =
) , ( ) , ( ) , (
(4.3)

where
s
t
(
is the time point at the end of steady-state chip formation analysis. Because of the
above assumption about steady nodal temperature of workpiece node at interface,
) ,
s
t A
θ is equal to and replaces
A
θ . The first part ( )
) , ( ) , (
s s
t B t A
k θ θ − is nodal conductive
heat flux, which can be obtained from the end of steady-state chip formation analysis.
The second part is the variation of nodal conductive heat flux and it is dependent on
Heat Transfer Analysis In Metal Cutting 68
the difference in temperature between current nodal temperature and the
temperature at the end of steady-state chip formation analysis.


N134
N128
N129
N133
N127









Fig. 4.3 Nodal temperature at selected workpiece nodes

Therefore nodal total heat flux can be expressed by Eq. 4.4.

( )
B t B
c
t
r
t
t
s s s
k q q q θ θ − + + =
) , (
(4.4)

where
t
q is the total nodal heat flux;
s
t
c
q stands for ( )
) , ( ) , (
s s
t B t A
k θ θ − ;
s
t
r
q is the nodal frictional heat flux;
s
t
c
q and do not change after importation. Based on Eq. 4.4, a temperature-
dependent heat flux subroutine is developed for the heat transfer analysis.
s
t
r
q
In addition, the tool makes heat transfer with the environment through rake face and
flank face by heat convection and radiation.
The nodes on bottom face and hole surface always keep room temperature because
of their contact with the tool holder and the screw.

4.3.2 Results & Discussion

When maximum temperature change of 10K between two times of incrementation of
heat transfer analysis is defined as steady state criterion, steady state is reached in
Heat Transfer Analysis In Metal Cutting 69
the whole tool in 2.2s.s Fig. 4.4 shows the progress of temperature at four selected
nodes in the tool.

Node 48
Node 340
Node 436
Node 13
(a) Position of the selected nodes (b) Temperature history
Fig. 4.4 Temperature history of nodes in the tool



(a) t=0s (b) t=2.2s
Fig. 4.5 Temperature field (Kelvin) change of the tool in heat transfer analysis

Fig. 4.5 shows that at the beginning of heat transfer analysis the high temperature
region concentrates in a small area near the cutting edge, and after 2.2s this region
extends to nearly one-third of the tool.
Heat Transfer Analysis In Metal Cutting 70

4.4 In Milling Operation

Heat transfer analysis is performed for both the workpiece and the cutting tool. In the
cooling phase of the first milling cycle, whether the workpiece can restore to room
temperature affects the chip formation, temperature distribution and heat transfer with
the cutting tool in the second milling cycle. Study on the development of temperature
distribution in the cutting tool in multi milling cycles is important for the
implementation of tool wear estimation.

4.4.1 On Workpiece

4.4.1.1 Modelling

Generally, the actual workpiece is very huge comparing with the small part of
workpiece used in the chip formation analysis. Therefore the workpiece geometry is
extended in the heat transfer analysis. It is composed of two parts, as shown in Fig.
4.6. Part 1 is the remaining workpiece geometry after the chip is cut away in the chip
formation analysis. Part 2 is some additional workpiece material; it is attached to the
bottom of part 1.









= +
part 1 part 2 workpiece
Fig. 4.6 Geometry and mesh of the workpiece in heat transfer analysis

The initial temperature of part 2 is set to room temperature. Nodal temperature at the
end of previous chip formation analysis step is imported and defined as initial
temperature of part 1. Fig. 4.7(a) shows the temperature distribution at the beginning
of heat transfer analysis.
Heat Transfer Analysis In Metal Cutting 71


i
Heat emitted to
environment due
to convec t ion
and r ad ia t ion
Heat conduction
n the material
(a) t=0ms (b) t=38.77ms
Fig. 4.7 Temperature field (in Celsuis) change of the workpiece in heat transfer
analysis

During the cooling phase, in addition to the heat conduction from cutting area to the
whole workpiece bulk, the workpiece cools down due to heat convection and
radiation through boundary.
Heat flux due to convection is calculated by

( )
0
θ θ − − = h q
v
; (4.5)

where
v
q is the heat flux due to convection;
h is a reference film coefficient, unit C m s J ° . .
2
;
θ is the temperature at a point on the surface;
0
θ is the sink temperature, i.e. room temperature.
Heat flux due to radiation to the environment is governed by

( ) ( )
(
¸
(

¸

− − − =
4
0
4
z z r
q θ θ θ θ εσ , (4.6)

where
r
q is the heat flux due to radiation on a surface;
ε is the emissivity of the surface;
σ is the Stefan-Boltzmann constant;
Heat Transfer Analysis In Metal Cutting 72
θ is the temperature at a point on the surface;
0
θ is the ambient temperature;
z
θ is the value of absolute zero on the temperature scale being used.

4.4.1.2 Results & Discussion

The heat transfer analysis covers the rest period of the first milling cycle after the chip
formation analysis ends. Fig. 4.7(b) shows that after 38.77ms of cooling, the entire
workpiece restores nearly to room temperature. The heat in the workpiece is emitted
to the environment.
Observation on the temperature progress at several selected nodes shows that after
the first milling cycle the temperature increment is smaller than 10K, as shown in Fig.
4.8. It is assumed that temperature increment of the workpiece in the real cutting
experiment is smaller than the predicted because the huge workpiece used in reality
supplies a higher heat capacity and a bigger boundary surface to emit the heat.


Node 839
Node 2140
Node 464
Node 550
Node 638
Node 501
(a) Monitored nodes (b) Temperature progress at monitored nodes
Fig. 4.8 Temperature history of workpiece nodes in heat transfer analysis step

It is assumed that in the second milling cycle this small temperature variation in the
workpiece has no big influence on material deformation, heat generation,
temperature distribution in the workpiece and tool-chip and tool-workpiece contact. It
is expected that if the influence of the temperature variation of the cutting tool is not
considered, chip formation analysis result in the second milling cycle can be
assumed similar to that in the first milling cycle. Solutions obtained from the chip
Heat Transfer Analysis In Metal Cutting 73
formation analysis in the first milling cycle can be used to the second milling cycle.
According to the same reasoning, they are useable in the third, fourth, and further
milling cycles. Therefore solutions obtained from the chip formation analysis of the
first milling cycle are used in the heat transfer analysis of the cutting tool in multi
milling cycles and tool wear estimation.

4.4.2 On Tool

4.4.2.1 Modeling

In the cooling phase of milling operation, if the heat in the tool is not emitted
completely to the environment by heat convection and heat radiation, the
temperature of the tool will get an increment in the later milling cycle due to the
remaining heat. This part will try to analyse the temperature variation of the tool with
the accumulation of heat. The heat transfer analysis is performed in 8 milling cycles.
The tool geometry and mesh in chip formation analysis are inherited and used in the
heat transfer analysis. Although the real cutting tool moves continuously with the
rotation of the shaft, in the simulation it is fixed spatial because the degree of
freedom in the heat transfer analysis is limited only to temperature.
Heat transfer analysis starts from the time when the chip formation analysis ends.
The temperature distribution at the end of the chip formation analysis is imported into
the heat transfer analysis as initial conditions.
In every milling cycle, the tool is heated in the cutting phase by the heat flux at the
tool-chip and tool-workpiece interface. The two components of the total heat flux,
frictional and conductive heat flux are time-dependent varying. Frictional heat flux
changes because of varying shear stress and sliding velocity caused by the change
of chip thickness in milling operation. Conductive heat flux changes with the varying
of difference in temperature between the tool and the workpiece at contact interface.
Although the nodal total heat flux in the cutting phase is changing continuously from
time to time, it is possible to obtain the basic values of nodal frictional heat flux and
nodal conductive heat flux from the chip formation analysis of the first milling cycle
only at some discrete time points 0, t
1
, …, t
j
, …, t
n
. These data are written in the heat
flux and temperature files. The heat flux value at other time point is obtained by
performing interpolation.
Heat Transfer Analysis In Metal Cutting 74
In addition, conductive heat flux is temperature dependent. When the nodal
temperature of the cutting tool in the later milling cycles is higher than that in the first
milling cycle due to the accumulation of heat, the nodal conductive heat flux value will
change due to the varying of the difference in temperature between the cutting tool
and the chip or the workpiece.
Based on these analyses, a heat flux subroutine DFLUX is designed to create time-
and temperature-dependent nodal heat flux data.
Every time when the subroutine DFLUX is called, the time, element number, face
number and integration point are entered as input variables. The subroutine first finds
out the corresponding nodal label because only nodal label is used in the heat flux
and temperature files. Then the basic nodal total heat flux and temperature values at
all time points are read from the heat flux and temperature files. By finding out the
remainder of the current time divided by the period of one milling cycle and
comparing this remainder with the time points, the interval and the two time points at
the end of the interval is determined. If the cutting tool is in the cooling phase, the
current nodal total heat flux is set to zero. Otherwise when the tool is located
between the time point j-1 and j, the nodal total heat flux is calculated by

( ) ( ( k q q q
b
j i
b
j i i
t
j i
t
j i
c
i
× + × − − + × =
− − ) 1 , ( ) , ( ) 1 , ( ) , (
5 . 0 5 . 0 θ θ θ )) (4.7)

where
c
q is the current nodal total heat flux;
t
q is the basic nodal total heat flux (the sum of the nodal frictional heat flux and the
nodal conductional heat flux);
θ is the current nodal temperature;
b
θ is the basic nodal temperature;
i is the nodal label;
j is the time point number;
k is the gap conductance.
In addition, in order to improve the convergence rate during the solution of non-linear
equations in an increment, the rate of change of the current nodal total heat flux with
respect to the temperature,
i
c
i
d dq θ is given the value –k.
Heat Transfer Analysis In Metal Cutting 75
The gap conductance in the heat transfer analysis is 10000 when the tool face node
is in contact with the chip and the workpiece. It is similar to the value used in the chip
formation analysis. When the tool face node has no contact with the chip and the
workpiece, the gap conductance is set to zero. The contact status of the tool face
node at a time point is derived from the value of normal pressure at the
corresponding time point in chip formation analysis of the first milling cycle.
Considering the movement of the tool, a high reference film coefficient is defined in
the model.

4.4.2.2 Results & Discussion

Fig. 4.9 shows the temperature distribution of the cutting tool when it cuts out of the
workpiece 0.3ms in the first, forth and eighth milling cycle. The high temperature
region is widening as the milling process continues.







(a) At the end of chip formation analysis







(b) After 4 milling cycles
Heat Transfer Analysis In Metal Cutting 76







(c) After 8 milling cycles
Fig. 4.9 Temperature field (in Kelvin) progress of the tool in heat transfer analysis

Fig. 4.10 and Fig. 4.11 show the variation of nodal temperature in the cutting tool in
more than 8 milling cycles, including both chip formation analysis step and heat
transfer analysis step.











Node57
Node42
Node39
Node35
Node4
(a) Position of the selected nodes (b) The temperature history of the nodes
Fig. 4.10 Progress of nodal temperature on the tool face

In Fig. 4.10 the nodal temperature at the tool face nodes on the too-chip and tool-
workpiece contact interface increases in cutting phase and decreases in cooling
phase. The peak value of temperature in every milling cycle appears when the cutting
tool is cutting out of the workpiece. The valley value appears when the cutting tool is
going to enter into the workpiece. Both the peak and valley value are increasing with
the cutting process continuing, but the increments are decreasing and the increment
of the peak value is smaller than 1K after every milling cycle while the increment of
Heat Transfer Analysis In Metal Cutting 77
the valley value is much great. In the last milling cycle, the peak value can be
assumed to become steady because the increment is smaller than 0.5K, whereas the
increment of the valley value is still greater than 5K.
Inside the cutting tool a different progress tendency of nodal temperature is
observed. At the nodes close to the cutting area, nodal temperature increases in
cutting phase and decreases in cooling phase. At the nodes far from the cutting area,
nodal temperature is always increasing during the entire milling cycles, for examples
N210 and N3 in Fig. 4.11.


N210
N127
N135
N13
N3








(a) Position of the selected nodes (b) The temperature history of the nodes
Fig. 4.11 Progress of nodal temperature inside the tool

Both the nodal temperature of nodes inside the tool and on the tool face shows that
cyclical thermal balance state is not attained in the first 9 milling cycle and heat gain
is greater than heat loss in every milling cycle. Higher temperature is expected in the
further milling cycles. It is very difficult to analyse the cyclical thermal balance by only
manually adding more milling cycles in heat transfer model file because the number
of milling cycles to reach cyclical thermal balance state is unknown.

4.4.2.3 Application Of Preheated Cutting Tool

According to the analysis above, tool temperature increases due to accumulation of
remaining heat. Heat loss increases with the tool temperature. When heat loss
becomes equal to heat gain, cyclical thermal balance state is attained. In order to
reduce the number of milling cycles to reach cyclical thermal balance state and
speed up the calculation process, the whole cutting tool is preheated beforehand by
Heat Transfer Analysis In Metal Cutting 78
defining a high initial temperature, then it is used in the milling process. Because at
present only tool temperature is concerned, milling process is analysed by performing
only heat transfer analysis. Only 8 milling cycles are included in the heat transfer
analysis.

Analysis 1: Preheated to 600K

In this analysis, the cutting tool is preheated to 600K. Nodal temperature at the same
tool nodes as in the former heat transfer analysis is observed. In Fig. 4.12, nodal
temperature on the tool face and inside the tool is increasing after every milling cycle,
but the increment is smaller than that in the former analysis.










(a) (b)
Fig. 4.12 Preheated to 600K (a) The temperature history of the nodes inside the tool
(b) The temperature history of the nodes on the tool face

At the nodes on tool-chip interface, for example N4 and N42, valley value of nodal
temperature appears at the time when the tool comes into contact with the workpiece
instead of before the contact takes place. This is explained by the heat conduction
between the tool with higher temperature and the workpiece with room temperature.
At the farthest nodes away from the cutting area, for example, node 3 and node 210,
temperature decreases first and then increases again. The explanation is that at the
beginning of the analysis the entire workpiece has a same temperature value, no
heat conduction takes place in the vicinity of these nodes, these nodes are located at
the boundary, heat convection to the environment makes the temperature decrease.
After a time of cutting, the workpiece material in the vicinity is heated by the heat
Heat Transfer Analysis In Metal Cutting 79
generated in the cutting process and these nodes are heated because of heat
conduction.

Analysis 2: Preheated to 700K

According to analysis 1, when the tool preheated to 600K is used in milling operation,
tool temperature increases still and no cyclical thermal balance state is attained.
Higher tool temperature is expected in cyclical thermal balance state. Therefore, in
this analysis, the cutting tool is preheated to 700K.
Nodal temperature at the same tool nodes is observed. In Fig. 4.13, nodal
temperature on the tool face and inside the tool is decreasing after every milling cycle
and the decreasing rate is comparable to the increasing rate in analysis 1. At node 3
and node 210, temperature decreases in the entire 8 milling cycles.
The analysis shows that cyclical thermal balance state is not attained in the 8 milling
cycles because the workpiece is heated too high and heat loss is greater than heat
gain in every milling cycle.










(a) (b)
Fig. 4.13 Preheated to 700K (a) The temperature history of the nodes inside the tool
(b) The temperature history of the nodes on the tool face

Analysis 3: Preheated to 650K

According to analysis 1 and analysis 2, when the cutting tool is heated to a
temperature between 700k and 600K, cyclical thermal balance state is expected to
Heat Transfer Analysis In Metal Cutting 80
realize in the first 8 milling cycle. Therefore, in this analysis, the cutting tool is
preheated to 650K.









(a) (b)
(a) (b)
Fig. 4.14 Preheated to 650K (a) The temperature history of the nodes inside the tool
(b) The temperature history of the nodes on the tool face

Nodal temperature is monitored at the same tool nodes. In Fig. 4.14, nodal
temperature on the tool face and inside the tool is approaching cyclical thermal
balance state. When the cutting tool move to the same position in the 7
th
and the 8
th

milling cycle, the maximum difference in temperature at the same node is smaller
than 0.1K. Therefore, it can be assumed that cyclical thermal balance state is
realized in the 8
th
milling cycle.
Heat Transfer Analysis In Metal Cutting 81

4.5 Summaries & Conclusion

ABAQUS/Standard is effective in heat transfer analysis. By introducing the heat flux
and temperature distribution from the chip formation analysis output file and using
user-developed heat flux subroutines, thermal steady state in the turning operation
and cyclical thermal balance state are analysed.
In turning operation, the heat transfer analysis is performed for the cutting tool until
the thermal steady state is reached.
In milling operation, the heat transfer analysis is performed for the cutting tool for
several milling cycle. By using preheated cutting tool in the heat transfer analysis, the
cyclical thermal balance state is analysed.

Estimation Of Tool Wear In Turning Operation 82
Chapter 5 Estimation Of Tool Wear In Turning Operation



5.1 Introduction

The following two chapters will describe the modelling of progressive tool wear in
turning and milling operation by developing user program with programming
language Python and integrating it with commercial FEM code ABAQUS/Explicit and
ABAQUS/Standard. In this chapter, the study will focus on the modelling of tool wear
in turning operation. Base on the obtained experience, the more complex modelling
problem, tool wear in milling operation, will be studied in the next chapter. All the tool
wear simulation models will be developed for two-dimension. Solution to the
problems met in 2D modelling will be helpful for the implementation of 3D modelling
in the future.
Tuning operation is characterized by continuous cutting process; the entrance and
exit of cutting tool takes place infrequently and takes only a short time. In continuous
cutting process, if the effects of tool wear and uneven distributions of workpiece
material are neglected, cutting thickness, chip shape, and various cutting process
variables will have no great change and steady state can be assumed. Tool wear
calculation can be simplified by assuming that tool wear is created completely by the
steady state cutting process and neglecting the effect of entrance and exit phase.
By integrating tool wear mathematical model with the finite element steady-state
cutting analysis, tool wear estimation is implemented. It is performed with a tool wear
estimation program. The program controls the submission of chip formation and heat
transfer analysis jobs, monitors their analysis process, accesses the created result
and output database files once the analysis jobs are finished, performs tool wear
calculation and modifies the related model files according to the calculated tool wear.

5.2 Tool Wear Calculation Program Design

Fig. 5.1 shows the flow chart of the tool wear calculation program. The program is
designed to perform tool wear calculation automatically cycle by cycle until a tool
reshape criterion is reached. In every calculation cycle, chip formation and heat
transfer analysis jobs are submitted to analyse the steady-state cutting process and
obtain the cutting process variable values necessary for the calculation of wear rate
Estimation Of Tool Wear In Turning Operation 83
at steady state. Nodal wear rate is calculated by using the tool wear mathematical
model. Based on the calculated nodal wear rate, a suitable cutting time increment is
searched by program according to a user-specified VB increment value. Then the
nodal displacement due to wear in the cutting time increment is calculated at every
tool face node, and the tool geometry is updated according to the calculated nodal
displacement. If the produced flank wear VB is smaller than the user-defined tool
reshape criterion VB
max
, a second tool wear calculation cycle starts with the updated
tool geometry.

















Yes
No
VB>=VB
max
?
N
e
x
t

c
a
l
c
u
l
a
t
i
o
n

c
y
c
l
e

Nodal wear rate
calculation
Cutting time increment
calculation
Nodal displacement
calculateion
Heat transfer analysis
Chip formation analysis
End
Tool geometry updating
Start

Fig.5.1 Flow chart of tool wear calculation program

5.3 Modeling Procedure

During the explanation of the entire modelling procedure, the tool wear under the
same cutting condition as in the chip formation simulation of turning operation is
estimated.

Estimation Of Tool Wear In Turning Operation 84
5.3.1 Chip Formation And Heat Transfer Analysis

Chip formation analysis provides the mechanical variables at steady state and the
thermal variables at steady state are predicted by heat transfer analysis.

5.3.1.1 Normal Pressure

Fig. 5.2 shows the variation of normal pressure at the tool face nodes along the tool-
chip interface at the end of chip formation analysis.
At the tool tip area, the normal pressure has the maximum value. Then the
distribution exhibits a plateau of high stress near the tool. Beyond the feed distance,
the normal pressure drop off sharply, as observed by Childs and Mahdi [Chil-89]
when turning mild steel. At the distance of about 0.35, the chip loses contact with the
tool face, therefore the normal pressure drops to zero.


Fig.5.2 Normal pressure of the tool face nodes at tool-chip interface at steady state

5.3.1.2 Sliding Velocity

In Fig. 5.3(a) workpiece nodes in the tool tip area have negative relative sliding
velocities. This means that they are flowing into machined surface. The nodes that
are more than 0.02mm away from the tool tip are moving out of cutting area with the
chip. There is no contact between the chip and the tool face in the area beyond the
distance of 0.35mm. Therefore the sliding velocity becomes zero, as designed by
ABAQUS [ABA-01a].

Estimation Of Tool Wear In Turning Operation 85

(a) Sliding velocity at the workpiece nodes at steady state

1 1 + + i i i i
( )
( )
s
i
s
i
i i
i j s
i
s
j
v v
y y
y y
v v −


+ =
+
+
1
1
( )
( )
s
i
s
i
i i
i j s
i
s
j
v v
x x
x x
v v −


+ =
+
+
1
1
Workpiece
Tool
If :




Else:

− ≥ − x x y y
Tool face node j (x
j
, y
j
)
Workpiece node i+1 (x
i+1
, y
i+1
)
Workpiece node i (x
i
, y
i
)










(b) Calculation of sliding velocity at the position of tool face nodes
Fig. 5.3 Calculation of sliding velocity at the position of tool face nodes

Only sliding velocities at the position of workpiece nodes can be obtained directly
from the simulation. But when calculating nodal wear rate, it is necessary to know the
sliding velocity value of workpiece material at the position of tool face nodes. Not all
the tool face nodes and workpiece nodes are in contact. First all the tool face nodes
and workpiece nodes in contact are found out depending on whether the absolute
values of their normal pressure are greater than a critical value, for example, 1e-
6Mpa. Then they are arranged in counter-clockwise order. Every tool face nodes in
contact has two neighbouring workpiece nodes before and after it. The calculation is
performed based on their position relationship, as shown in Fig. 5.3(b).
Estimation Of Tool Wear In Turning Operation 86

5.3.1.3 Tool Temperature

Temperature of the tool face nodes at thermal steady state is obtained from heat
transfer analysis, as shown in Fig. 5.4. High temperature forms at the tool tip and a
distance from the tool tip on rake face.


Fig. 5.4 Temperature of tool face node at steady state

5.3.2 Wear Rate Calculation

After the cutting process variables, sliding velocity of the workpiece material, tool
temperature and normal pressure at every tool face node are obtained, wear rate at
the position of every tool face node is calculated by using wear mathematical model.
Because at low cutting speed, the flank wear and crater wear are assumed to be
created mainly by abrasive wear and adhesive wear, Usui’s model, described in
Chapter 1, is employed in the calculation. The constants in Usui’s equation for the
combination of carbide cutting tool and mild steel are shown in Table 1.2.

5.3.3 Nodal Move Direction

Tool wear expression in geometry can be realized with two approaches: element
deletion and nodal movement. The latter one is adopted in this paper. The nodal
move direction is calculated at every tool face node.

Estimation Of Tool Wear In Turning Operation 87
5.3.3.1 Dividing Node

Before calculating the nodal move direction, a dividing node, the circled node in Fig.
5.5 is searched by the program. It divides the entire tool face into flank face and rake
face. It has the minimum y-coordinate.
Before searching, all the tool face nodes are found out and arranged in counter
clock-wise order in a list. The first tool face node in the list is given to the dividing
node. Then the current dividing node is compared with all the tool face nodes. Any
node, whose y coordinate is smaller than the current dividing node by 2e-4mm, will
become the new dividing node. After one searching cycle, the real dividing node is
found out. It is saved as the tool edge position for the latter calculation of flank wear
land width.
On flank face and rake face nodal move direction is calculated with different
methods.

5.3.3.2 On Rake Face

In the rake face part nodal move direction is assumed to be perpendicular to the
relative sliding velocity of the workpiece material and pointed into the tool body.
Every tool face node is attached with two tool face segments. On every tool face
segment the sliding velocity is along the tangential direction. Points on the face
segment should move along the negative normal direction. Tool face node belongs to
two face segments. It moves along the negative direction of the average unit normal
vector of the two face segments or the negative direction of their resultant vector, as
shown in Fig. 5.5.
Estimation Of Tool Wear In Turning Operation 88























i
j
n
r
1 + j
n
r
( ) ( ) ( ) ( ) ( ) ( )
|
|
.
|

\
|
− + −

− + −

=
+ +
+
+ +
+
2 1
2
1
2
1
1
2 1
2
1
2
1
1
,
i i i i
i i
i i i i
i i
j
y y x x
x x
y y x x
y y
n
r
( ) ( ) ( ) ( ) ( ) ( )
|
|
.
|

\
|
− + −

− + −

=
+ + + +
+ +
+ + + +
+ +
+ 2 1
2
1 2
2
1 2
2 1
2 1
2
1 2
2
1 2
1 2
1
,
i i i i
i i
i i i i
i i
j
y y x x
x x
y y x x
y y
n
r
1 +
+ =
j j
r
i
n n n
r r r
r
i
r
i i
n n D
r r
r
− =
i
r
( ) 1 , 0 =
k
r
n
2

On flank face, nodal move direction D
On rake face, nodal move direction is calculated by
D
Node k
Flank face part
R
a
k
e

f
a
c
e

p
a
r
t


Dividing node
Face segment j+1
Face segment j
Node i+1 (x
i+1
,y
i+1
)
Node i+1 (x
+1
,y
i+1
)
Node i (x
i
,y
i
)

Fig. 5.5 Nodal move directions (thick arrows) of tool face nodes

5.3.3.3 On Flank Face

In the flank face part the relative sliding velocity can be assumed to be in the cutting
speed direction when the elastic recovery of workpiece material is neglected.
Therefore all the nodes in this part have the same nodal move direction. It is in y-
direction and pointed upwards.
Every nodal move direction is normalized to unit vector
) , ( j i
D
r
, where subscript i is
nodal label, j is the calculation cycle number.

Estimation Of Tool Wear In Turning Operation 89
5.3.4 Cutting Time Increment Calculation

In metal cutting experiment cutting time increment means the duration of cutting time
between two successive measurements of tool wear. In the simulation the calculation
of tool wear and the tool geometry updating are based on a certain cutting time
increment. Within the cutting time increment an unchanged nodal wear rate value is
used to calculate the tool wear. Therefore, if a big cutting time increment is specified,
a big error will be created during the calculation of tool wear. But if the cutting time
increment is too small, only small tool wear increment is produced in every
calculation cycle. In order to reach tool reshape criterion, many calculation cycles
have to be performed. The chip formation analysis is carried out in every calculation
cycle; it is very time-consuming. Therefore a suitable cutting time increment should
be given. But when there is no knowledge about the tool wear in the simulated
cutting conditions, for example, when novel workpiece material is machined, it is
difficult to define a suitable value. But it is easier for the user to specify a tool wear
increment. Since the nodal wear rate is already known, the cutting time increment, in
which the specified tool wear increment is produced, can be searched by program.
Therefore a searching module is designed to carry out the searching work. While the
suitable cutting time increment is being searched, a flank wear calculation subroutine,
, is called frequently. ) , ( wearrate t Flankwear ∆

∆ ∆
0
t ∆
1
t t
2
t ∆
VB [mm]
0.05
VB
2

VB
0

) , ( wearrate t Flankwear ∆
Aimed VB
value range
Example of a wear curve
VB
1

0.10
VB
a b
c d
a
1
b
1
c
1
VB
Edge
position
Time [s]
(a) Flank wear calculation (b) Cutting time increment searching process
Fig. 5.6 Flank wear calculation and cutting time increment searching process

Estimation Of Tool Wear In Turning Operation 90
5.3.4.1 Flank Wear Calculation Subroutine

The flank wear calculation subroutine calculates flank wear
land width VB. VB is the distance from the cutting edge position (which has been
saved) to the last moved tool face node. For example, in Fig. 5.6(a), node a is the
last tool face node with non-zero wear rate. In cutting time increment , it should
move to point a
) , ( wearrate t Flankwear ∆
t ∆
1
, then node b and c will have smaller y-values than point a
1
, and a
bulge will be formed on the flank face. But in practice the wear process is continuous.
Once node b or c comes into contact with workpiece material due to wear of the
cutting edge, they are also worn away and no bulge is formed. Therefore node b and
c should move to point b
1
and c
1
in order to have the same y-value with point a
1
. VB
is calculated from the cutting edge position to node c, because it is the last moved
tool face node.

5.3.4.2 Cutting Time Increment Searching Procedure

The cutting time increment searching procedure can be described by Fig. 5.7.
At the beginning the aimed VB median value VB
m
is calculated according to the user-
specified VB increment value. For example, in Fig. 5.6(b), the tool gets a flank wear
land width of 0.05mm from the previous tool wear calculation cycle. is
specified by the user. Therefore in this tool wear calculation cycle, VB
mm VB 05 . 0 = ∆
m
is 0.1mm. In
order to save the searching time, the aimed VB value should be given a permitted
error range, e.g. the dotted range in Fig. 5.6(b). In addition, a positive initial cutting
time increment value is given arbitrarily. Then the searching process starts.
During the searching process, the searching lower limit ∆t
0
t ∆
1
and the searching upper
limit ∆t
2
are changing until the calculated tool wear VB value under the cutting time
increment falls into the aimed VB value range. t ∆







Estimation Of Tool Wear In Turning Operation 91























∆t=(∆t
1
+∆t
2
)/2
error<-δ?
∆t
2
=∆t

∆t
1
=∆t

Current cutting
time increment ∆t
is output
|error|<=δ?
No
Yes
No
Yes
No
∆t=2∆t

∆t
2
>∆t
1
?
End
Start
Yes
VB
m
, initial value
∆t
0
, permitted error
δ


Call subroutine Flankwear(∆t,wearrate) to
calculated the current flank wear VB, calculate
the current error, error=VB- VB
m

Cutting time increment ∆t,
the searching lower limit ∆t
1
and the searching upper
limit ∆t
2
are set to ∆t
0
Fig. 5.7 Flow chart of cutting time increment searching procedure

5.3.5 Nodal Displacement

Nodal displacement due to wear is calculated at every tool face node by

) , ( ) , ( ) , ( j i j j i j i
D t w w
r
&
r
⋅ ∆ ⋅ = (5.1)

where
w
r
is the nodal displacement vector;
i is nodal label;
Estimation Of Tool Wear In Turning Operation 92
j is tool wear calculation cycle number.
In addition, some nodes on flank face must be moved in order to avoid forming bulge
on flank face, as mentioned above.

5.3.6 Tool Geometry Updating

In order to visualize the tool wear profile and prepare tool geometry model for the
next tool wear calculation cycle, tool geometry updating is performed. It is
accomplished with two steps. In these two steps nodes on the tool bottom surface,
marked with small triangles in Fig. 5.8, are fixed spatial.

5.3.6.1 Step 1: Initial Tool Wear Profile

In the first step the tool face nodes, including the nodes on rake face and flank face,
the circled nodes in Fig. 5.8, are moved according to the calculated nodal
displacement. The entire movement is accomplished several times. Every time the
tool face node is moved a very small distance. Then the mesh inside the tool is
remeshed with one of the smoothing methods: volume smoothing, Laplacian
smoothing and equipotential smoothing, or their combination. In the following part,
volume smoothing is employed because of the robustness. Remeshing improves
mesh distortion and enables additional nodal movement of the tool face nodes in the
next times. After the first step, an initial tool wear profile appears on the cutting tool,
as shown in Fig. 5.9(b).
Estimation Of Tool Wear In Turning Operation 93


Tool bottom
Rake face
Flank face
Tool bottom
Tool bottom
Rake face
Flank face
Tool bottom
Rake face
Flank face
Tool bottom

Fig. 5.8 Boundary conditions in step 1 of tool geometry updating model

5.3.6.2 Step 2: Adjustment

Because of the contact problem on the tool-chip and tool-workpiece interface caused
by the coarsened workpiece mesh in chip formation analysis, the predicted
distributions of cutting process variables along the tool face often contain ‘vibration’.
These results in zigzags of the initial tool wear profile, e.g, zone A in Fig. 5.9(b).
In addition, the mesh inside the cutting tool has been remeshed many times in step 1.
But the tool face nodes are moved according to the calculated nodal displacement
without any additional adjustment of nodal position. Sometime very fine mesh is
formed in the cutting edge area, for example, zone B in Fig. 5.9(b); they make tool
geometry updating in the next calculation cycle difficult because negative element
areas may be created by the nodal displacement due to additional produced tool
wear.
Fig. 5.9(c) shows that in the second step, zigzags of the crater wear profile are
smoothened and the mesh near the cutting edge is coarsened. The final tool wear
profile and tool geometry is produced by step 2. The tool geometry model file is
updated according to the produced result in step 2.
Estimation Of Tool Wear In Turning Operation 94









Zone B
Z
o
n
e

A

(a) (b) (c)
Fig. 5.9 Changes of the mesh during tool updating steps (a) The tool geometry and
mesh at the beginning of step 1 (b) At the end of step 1, nodes on the tool face are
moved according to the calculated nodal displacement, crater wear and flank wear
are formed (c) At the end of step 2, zigzags of the crater wear are smoothened.

5.4 Results & Discussion

5.4.1 Tool Wear

With this tool wear estimation program, tool wear progress under the same turning
cutting conditions as described in Table 3.1 is calculated. Tool reshape criterion is set
to 0.15mm, and tool wear increment is specified by user, permitted
error δ in the cutting time increment searching process is set to 0.02mm. The tool
wear estimation process is accomplished with three tool wear calculation cycles.
After the first calculation cycle, the new tool in Fig. 5.10(b) is updated to the worn tool
in Fig. 5.10(c). After the second calculation cycle, increased crater wear and flank
wear can be found on the updated tool in Fig. 5.10(d).
mm VB 05 . 0 = ∆

Estimation Of Tool Wear In Turning Operation 95

t=0s
t=5s
t=46s
t=187s
(a) (b) t=0s (c) t=5s

(d) t=46s (e) t=187s
Fig. 5.10 Tool wear profile progress










0
0,05
0,1
0,15
0,2
0 50 100 150 200
t [s]
F
l
a
n
k

w
e
a
r

w
i
d
t
h

V
B

[
m
m
]
Measured
Est imat ed
0
0,02
0,04
0,06
0,08
0,1
0 50 100 150 200
t [s]
C
r
a
t
e
r

w
e
a
r

d
e
p
t
h

[
m
m
]
Measured
Est imat ed
(a) Flank wear (b) Crater wear
Fig. 5.11 Comparison between estimated and experimental progress curves for tool
wear (under cutting condition: v
c
=300m/min, a
p
=2mm, f=0.145mm/r)

The solid line in Fig. 5.11 shows the wear progress curves of flank wear and crater
wear obtained from experiment [Schm-02] under the same cutting condition. The dot
Estimation Of Tool Wear In Turning Operation 96
lines are predicted tool wear curves. It is found that the estimated flank wear and
crater wear are smaller than experimental ones. In experiment, after 20s of cutting,
the flank wear has exceeded 0.15mm and crater wear 0.06mm, while after 187s, the
estimated flank wear just arrives at 0.14mm and crater wear 0.08mm.
The discrepancy may be caused by:
• the simplified and low coefficient of friction. In the chip formation analysis,
Coulomb’s friction model is adopted and a constant coefficient of friction 0.3 is
used in the whole tool wear estimation process. According to the verification of
chip formation analysis in continuous chip formation, the predicted cutting
force and thrust force are smaller than the experimental data by about 15%
and 35% when the coefficient of friction is set to 0.3. This maybe means that
the predicted variables for the calculation of tool wear have error as well.
Therefore chip formation modeling is very important for the accuracy of tool
wear estimation. In order to improve the prediction, it is expected that in the
later tool wear estimation, the coefficient of friction should be calculated
according to the cutting force and tool geometry or with a more reliable
method.
• inconsistentness of material combination. Because the characteristic equation
of tool wear and the tool wear data come from different literatures and
researchers, it is unavoidable that difference exist in these tool and workpiece
material’s chemical composition and structure. It was tested by Kitagawa et al
that the content and size of abrasive particle dispersed in workpiece material
and chemical composition of tool material could be correlated with change in
the constants of the wear characteristic equation both in higher and lower
temperature ranges [Kita-88].
• contact problem between flank wear and the workpiece. From Fig. 5.11, after
a certain tool wear is formed, both wear rate on flank face and on rake face
are decreasing, the wear rate on flank face decreases more than that on rake
face. It is observed that the temperature on flank wear drops off to a low value.
This may be caused by the poor contact between flank wear and the
workpiece. In order to improve the contact, maybe on the flank wear face,
some nodes should be adjusted to form a small negative flank angle.
Estimation Of Tool Wear In Turning Operation 97

5.5 Summaries & Conclusion

In this chapter 2-D tool wear estimation in orthogonal cutting of turning operation is
implemented by integrating ABAQUS/Explicit and ABAQUS/Standard with Python
user-program. The main findings of this study are as follows:
(1) Python user program launches chip formation and heat transfer analysis job
automatically every time the new value about cutting process variables at
steady state are needed. Then displacement of every tool face node due to
wear is calculated by calculating nodal wear rate at steady state, searching a
suitable cutting time increment by program and nodal displacement
calculation. Finally tool geometry is updated according to the calculated nodal
displacements and one calculation cycle is finished.
(2) The Python user program runs automatically until a tool reshape criterion is
reached. The number of calculation cycles carried on before Python user
program stop is defined by dividing tool reshape criterion by the specified
wear increment. Because of the huge calculation time and cost of chip
formation analysis, a bigger wear increment is preferred in order to reduce the
calculation cycle number, which certainly will bring bigger errors in estimated
result. A trade-off value should be found.
(3) In order to improve the estimate result and realize tool wear estimation in
quantity, more efforts should be made in several aspects: more reasonable
frictional modelling, further mesh control, regenerate workpiece model when
chip shape has a great change due to tool geometry change caused by
serious tool wear and some modification of the flank wear shape in order to
improve the contact between flank wear and workpiece material.

Estimation Of Tool Wear In Milling Operation 98

Chapter 6 Estimation Of Tool Wear In Milling Operation



6.1 Introduction

Cutting action in milling operation is different from turning operation. With the cutting
tool rotating, workpiece moves in feed direction. In every rotation/milling cycle, the
cutting insert cuts away a layer of workpiece material and then cools down in the
environment. Therefore every milling cycle comprises cutting phase and cooling
phase. In the cutting phase, cutting thickness varies with tool engage angle. For
example, in the down milling operation in Fig. 6.1, cutting thickness has the maximum
value when the cutting insert advances into the workpiece. Then the cutting thickness
decreases continuously. At the exit the cutting thickness becomes zero. According to
metal cutting theory, nearly all the cutting process variables or solutions, stress,
strain, temperature, etc are related with the cutting thickness. Therefore they change
with the tool engage angle and so does nodal wear rate.

t
k
tt
kk+1 +1
t
k
zz
vvff
Cutting phase
Cooling phase
f
w
Tool
Workpiece
Node i











Fig. 6.1 Feature of milling operation

Although milling operation has no steady state, the cutting process possesses
periodicity. If the effect of tool wear, uneven distribution of the workpiece material, etc
are neglected, mechanical variables in one milling cycle can be assumed to be
repeated in other milling cycles because of the same cutting path of the cutting insert
Estimation Of Tool Wear In Milling Operation 99
and the same change of undeformed chip thickness. Mechanical cutting process
variables, normal pressure on tool face and relative sliding velocity of workpiece
material on the tool face, obtained from the first milling cycle can stand for those from
all other milling cycles. As tool temperature is concerned, the tool temperature
obtained in one milling cycle can stand for that in other milling cycle only when the
cyclical thermal balance state (heat loss is equal to heat gain per cycle) is attained.
Before the cyclical thermal balance state is reached, heat loss is smaller than heat
gain per milling cycle and the tool temperature increased after each milling cycle. The
tool temperature obtained in one of these milling cycles is always lower than that in
the milling cycle of the cyclical thermal balance state. In the milling operation with a
long continuous milling path, the cyclical thermal balance state dominates the entire
cutting process. The tool wear is mainly decided by the cyclical thermal balance
state. If the tool wear is calculated according to the tool temperature in a milling cycle
before the thermal balance state is reached, a lower estimated value of the tool wear
is expected.
Periodicity of cutting action and the existence of the cyclical thermal balance state
enable the implementation of tool wear estimation in milling operation. Once cyclical
thermal balance state is attained, tool wear and nodal average wear rate per cycle do
not change from cycle to cycle. Therefore the implementation of tool wear estimation
in milling operation can be simplified by calculating nodal average wear rate per cycle
in one milling cycle of cyclical thermal balance state and then using it in other milling
cycles.

6.2 Tool Wear Calculation Program Design

Fig. 6.2 shows the flow chart of the tool wear calculation program.









Estimation Of Tool Wear In Milling Operation 100

















Yes
No
VB>=VB
max
?
Nodal average wear
rate per cycle
Cutting time increment
calculation
Nodal displacement
calculation
Heat transfer analysis
Chip formation analysis
End
Tool geometry updating
Start
N
e
x
t

c
a
l
c
u
l
a
t
i
o
n

c
y
c
l
e

Fig. 6.2 Flow chart of the tool wear calculation program

The tool wear calculation program is designed to perform tool wear calculation
automatically cycle by cycle until a tool reshape criterion is reached. In every
calculation cycle, chip formation and heat transfer analysis are performed to predict
the cutting process variables, which are necessary for the wear rate calculation.
According to the above discussion on milling features, mechanical variables can be
obtained from the first milling cycle and the heat transfer analysis helps to decide
from which milling cycle tool temperature is read for the calculation of nodal wear
rate. The nodal wear rate is time-dependent and calculated at some selected time
points of one milling cycle, then the nodal average wear rate is calculated according
to these nodal wear rate at the selected time points. Based on the calculated nodal
average wear rate, a suitable cutting time increment value is searched according to a
user-specified VB increment value. Then nodal displacement due to wear produced
in the cutting time increment is calculated and the tool geometry updating aims at
forming the tool wear profile on the tool face. If the produced flank wear VB is smaller
Estimation Of Tool Wear In Milling Operation 101
than the user-defined tool reshape criterion VB
max
, a second calculation cycle will
start with the updated tool geometry.

6.3 Modeling Procedure

During the explanation of the entire modelling procedure, the tool wear under the
same cutting condition as described in Table 3.2 is estimated.

6.3.1 Chip Formation Analysis

Chip formation analysis aims at obtaining the mechanical variables for the calculation
of nodal wear rate. Because tool wear takes place only in cutting phase and there is
no tool wear created in cooling phase, chip formation analysis covering the entire
cutting phase is necessary and enough for the calculation of wear rate.
During the cutting phase, mechanical variables are varying from time to time. But with
explicit method, finite element analysis of chip formation process is performed by
advancing the time with small time increments. Hence the time for outputting the
variables is discontinuous. Furthermore, high frequency of variables outputting will
result in a large output database file and increase the amount of calculation in the
tool wear estimation. Therefore the frequency of mechanical variables output should
be decided by making a compromise between calculation accuracy and calculation
cost.
The mechanical variables, sliding velocity of workpiece material and normal pressure
on tool face, are required during the calculation of wear rate. Sliding velocity is
available at the workpiece node. For example, Fig. 6.3(b) shows the sliding velocity
at the time when the cutting tool engaging into the workpiece 0.1ms. At the tool tip
sliding velocity is very small, even some nodes flow towards the machined surface
before material failure takes place. Other nodes are flowing out with the chip at
increasing sliding velocity.
According to the sliding velocity at the position of workpiece nodes, sliding velocity at
the position of tool face nodes are calculated at the time point when the calculation of
nodal wear rate is required.
Estimation Of Tool Wear In Milling Operation 102

(a) The cutting at 0.100ms (b) At t=0.100ms
Fig. 6.3 Relative sliding velocity of workpiece material on the tool face at the time of
0.100ms

Normal pressure at the position of tool face node can be obtained directly. For
example, Fig. 6.4 shows the normal pressure at the time of 0.1ms. From tool tip to
the separation point of the chip and tool normal pressure is decreasing. No plateau is
observed in the entire tool-chip contact area as in turning operation.


Fig. 6.4 Normal pressure on the tool face at the time of 0.100ms


Estimation Of Tool Wear In Milling Operation 103
6.3.2 Heat Transfer Analysis

According to the heat transfer analysis, the tool temperature at the tool face nodes
increase after every milling cycle until cyclical thermal balance state is reached. Fig.
6.5 shows the nodal temperature on the tool-chip interface at the time when the tool
is engaging into the workpiece 0.1ms in three selected milling cycles. The selected
milling cycle are the first, the ninth and cyclical thermal balance cycle, which is
obtained from the last milling cycle of heat transfer analysis of the tool preheated to
650K. It is found that from the first to the ninth milling cycle, nodal temperature at tool
face has a jump, while from the ninth to cyclical thermal balance cycle; the
temperature has a relative small increment.
The tool temperature in the cyclical thermal balance state is read for the calculation
of tool wear.


Fig. 6.5 Tool temperature at the tool face nodes after the cutting insert advancing into
the workpiece 0.1ms

6.3.3 Nodal Average Wear Rate Calculation

Usui’s model is employed in the calculation of nodal wear rate at a certain time. The
wear characteristic constants in Usui’s equation for the combination of carbide cutting
tool and mild steel are shown in Table 1.2.

6.3.3.1 Discussion About The Calculation Method Of Nodal Average Wear Rate

Estimation Of Tool Wear In Milling Operation 104
Nodal wear rate varies with the cutting time. In the cutting phase, tool wear takes
place under the contact of the tool with the workpiece. In cooling phase, nodal wear
rate is equal to zero and no wear produced. Nodal average wear rate is calculated
by

( )
Ζ
=

Ζ +
0
0
) , (
) , (
t
t
j i
j i
dt t w
w
&
& , (6.1)

where
w& is the nodal average wear rate;
( ) t w& is nodal wear rate;
Ζ is the time span of one milling cycle;
i is the nodal label;
j is the milling cycle number.
At present it is very difficult to get the function of nodal wear rate . But nodal
wear rate values at some discrete time points can be obtained by sampling cutting
process variables during chip formation and heat transfer analysis and then
calculating the individual nodal wear rate values, as shown in Fig. 6.6. Based on
these nodal wear rate values, an approximate nodal average wear rate can be
calculated by the following equation.
( ) t w
j i ) , (
&

( ) ( )
Ζ
⋅ − ⋅ +
=
∑ + +
n
k k k j i k j i
j i
t t w w
w
1
1 ) 1 , , ( ) , , (
) , (
2
1
& &
& (6.2)

where
n means that the entire milling cycle is divided into n-1 small portions by n evenly
spaced time points;
k is the time point number; nodal wear rate is calculated at every time point.
In the real calculation, sampling of cutting process variables and the calculation of
nodal wear rate are not performed in the entire milling cycle because no wear takes
place in cooling phase. For example, the whole milling cycle may take about
39.27ms, but the cutting phase only takes place in the first 0.2ms of every milling
Estimation Of Tool Wear In Milling Operation 105
cycle. According to Chapter 4, the chip formation analysis includes the whole cutting
phase and 0.3ms of cooling phase in the first milling cycle. During the calculation of
nodal average wear rate, the mechanical variables sliding velocity and normal
pressure are read only at sampling time points 0.025, 0.05, … and 0.5ms in the chip
formation analysis. Tool temperature values are read at the corresponding time
points 0.025, 0.05, … 0.5ms of the selected milling cycle of heat transfer analysis.
Average nodal wear rate calculation is performed only at these time points. The part
of cooling phase from 0.5ms to 39.27ms is not considered.












Cutting phase
Cooling phase
Cooling phase
t
k
) , , ( k j i
w&
t
k+1
) 1 , , ( + k j i
w&
( ) t w
j i ) , (
&
N
o
d
a
l

w
e
a
r

r
a
t
e

o
f

n
o
d
e

i
Cutting time t [s]
Cutting phase
Cooling phase
Cooling phase
t
k
) , , ( k j i
w&
t
k+1
) 1 , , ( + k j i
w&
( ) t w
j i ) , (
&
N
o
d
a
l

w
e
a
r

r
a
t
e

o
f

n
o
d
e

i
Cutting time t [s]
Cutting phase Cutting phase
Cooling phase
Cooling phase
Cooling phase
Cooling phase
t
k
) , , ( k j i
w&
t
k+1
) 1 , , ( + k j i
w&
( ) t w
j i ) , (
&
N
o
d
a
l

w
e
a
r

r
a
t
e

o
f

n
o
d
e

i
Cutting time t [s]
N
o
d
a
l

w
e
a
r

r
a
t
e

o
f

n
o
d
e

i
Cutting time t [s] Cutting time t [s]
Fig. 6.6 Calculation of nodal average wear rate

Then in Eq. 6.2, the meaning of n becomes the number of time points dividing the
period of chip formation analysis.

6.3.3.2 Classification Of Workpiece Node

The sliding velocity of workpiece material at the position of tool face nodes are
calculated using the same method explained in Chapter 5.
Chip formation modelling in turning operation is different from milling operation. In
turning operation, workpiece nodes, which have possibility of getting contact with the
tool face, are fixed on several nodes on the chip underside. Only these nodes are
considered during the calculation.
Since in milling operation shear failure criterion is defined in the whole workpiece,
some nodes not on the moving path of the cutting edge may be exposed due to
Estimation Of Tool Wear In Milling Operation 106
element removal and get contact with the tool face. Therefore a large number of
workpiece nodes have to be considered.
In order to calculate the sliding velocity correctly and efficiently, workpiece nodes
considered are classified into three types. The first type, called inner node, includes
the nodes still inside the workpiece, for example, Node 411 in Fig. 6.7. Node 343
belongs to the second type, surface node, including the nodes exposed on the
surface. Node 893 is included in the third type, free node, which consists of the
nodes lose connection with the workpiece body because all the attached elements
are removal. Because inner nodes have no contact with the tool face, they only
increase the calculation time. When free nodes get contact with the tool face, they
will introduce calculation error. Hence only surface nodes join in the calculation of
relative sliding velocity.



Node 411
Node 893
Node 343
Fig. 6.7 Classification of workpiece node

Node type is judged by considering the number of the attached elements and the
deleted elements of a node. Every node is attached to several elements. The number
of the attached elements is denoted as N
attached
. The attached element is deleted or
removed as it reaches the shear failure criterion. The number of the deleted elements
is denoted as N
delete
. The type of a node is decided by:


, (6.3)
S
¦
´
¦
¦
att
N
N

att
N
=
ached

>
ached
deleted
=

de
N
de
N
0
¹
Inner node
urface node
0
0
de
N
de
N
,
leted
,
leted

leted

leted
Free node

Estimation Of Tool Wear In Milling Operation 107
6.3.4 Nodal Move Direction

In milling operation, position of every tool face segment and its normal direction are
varying with the rotation of the cutting tool. The calculation of nodal move directions
and tool geometry updating should be performed at the same rotation position of the
cutting tool, for example at the beginning of one milling cycle.

6.3.4.1 Dividing Node

Similar to the calculation of nodal move direction in turning operation, at the
beginning a dividing node that divides the entire tool face into rake face and flank
face is searched.
Instead of comparing the y-coordinate of every tool face node, the dividing node is
defined according to the distance between tool face nodes and the rotation centre. At
the beginning of the search, the first tool face node (the tool face nodes are arranged
in counter-clockwise order in advance) is given to the dividing node. Then every tool
face node is compared with the current dividing node one by one in counter-
clockwise order. Any tool face node whose distance to the rotation center is greater
than that of the current dividing node by 2e-4mm will update the record of dividing
node.
Nodal move direction is calculated with different methods on the rake face and flank
face, as shown in Fig. 6.8.










v
s

v
s

v
s

v
s

v
s

v
s

n
F
l
a
n
k

f
a
c
e

p
a
r
t

Rake face part
Dividing
node
v
s

Fig. 6.8 Calculation of nodal move direction

Estimation Of Tool Wear In Milling Operation 108
6.3.4.2 On Rake Face

The calculation method of nodal move direction on rake face used in turning
operation, explained in Chapter 5, is applied here as well.

6.3.4.3 On Flank Face

In the flank face part, relative sliding velocity of the workpiece material at each flank
face node is assumed along the tangential direction of the moving path of the flank
face node, when the elastic recovery of workpiece material is neglected. Nodal move
direction at the flank face node is perpendicular to the relative sliding velocity, i.e.
pointed from the flank face node to the rotation centre.
Every nodal move direction is normalized to unit vector
) , ( j i
D
r
, where subscript i is
nodal label, j is the calculation cycle number.

6.3.5 Cutting Time Increment Calculation

Cutting time increment is searched by the program according to a user-specified
flank wear increment value ∆ and a permitted error VB δ . Because of the particularity
of milling operation, the searching procedure and the flank wear calculation
subroutine have some difference compared with those in turning operation.


Edge position
a
b
c
a
1
b
1

c
1








V
B

d
Fig. 6.9 Flank wear calculation

6.3.5.1 Flank Wear Calculation Subroutine

Flank wear land width VB is calculated by a flank wear calculation subroutine
. ) , ( wearrate t Flankwear ∆
Estimation Of Tool Wear In Milling Operation 109
VB is the distance from the edge position to the last moved tool face node. The last
moved tool face node is searched by considering the movement of the cutting insert.
In milling operation the cutting insert rotates around a rotation centre. Any tool
material point which becomes the farthest point to the rotation centre will get contact
with the workpiece and then be worn away. For example, in Fig. 6.9 node a is the last
tool face node with non-zero nodal wear rate. According to the calculation, it should
move to point a
1
in the cutting time increment t ∆
.
Then node b becomes the farthest
point to the rotation center. It will be worn away. Because in this calculation cycle the
comprehensive information about the average wear rate of node b in the entire
cutting time increment cannot be obtained, its displacement is decided according to
node a. Node b will be move to point b
1
and it will have the same distance to the
rotation centre as node a. In the same way, node c will be moved to point c
1
. VB is
calculated from edge position to node c, because it is the last moved tool face node.

6.3.5.2 Cutting Time Increment Searching Procedure

Because milling operation is intermittent cutting, the existence of cutting phase and
cooling phase in milling cycles complexes the cutting time increment searching
procedure. In order to simplify the problem, the cutting time increment is increased in
step of whole milling cycles. The relationship between the milling cycle number N
cycle

and the cutting time increment ∆ is given by t

(6.4) Z N t
cycle
× = ∆

where
N
cycle
is positive integer;
Z is the time span of one milling cycle.
The cutting time increment searching procedure is described as follows:
(1) At the beginning, the aimed VB median value VB
m
is calculated according to
the user-specified VB increment value and the tool wear obtained in the
previous calculation cycle. According to the permitted error δ , the aimed VB
value range ( δ −
m
VB , δ +
m
VB ) is determined.
Estimation Of Tool Wear In Milling Operation 110
Then an initial cycle number N
cycle0
is given a positive integer value arbitrary,
the cycle number N
cycle
and the cycle number lower limit N
cycle1
are set to
N
cycle0
.
(2) Calculate the cutting time increment ∆ corresponding to N t
wearrate
cycle
with Eq. 6.4,
and call the subroutine to calculate the flank wear land
width, let
) , ( t Flankwear ∆
m
VB wearrate t Flankwear error − ∆ = ) , ( .
If ≤ error δ , then the searching procedure will end, and the present value
will be output as the result.
t ∆
(3) Otherwise, if error < -δ , then the cycle number lower limit N
cycle1
=N
cycle
, and
, repeat step 2 until erro
cycle cycle
N N ⋅ = 2 ≥ r -δ is satisfied.
Else, if error > δ , then the cycle number lower limit . 1
1
=
cycle
N
(4) The cycle number upper limit
cycle cycle
N N =
2
(5) N
cycle
takes the integer part of ( ) 2
2 1 cycle cycel
N N + . If the value of N
cycle
is equal
to N
cycle1
or N
cycle2
, then calculate the cutting time increment . The present
value will be output as result and the searching procedure will end.
t ∆
t ∆
(6) Calculate the cutting time increment corresponding to N t ∆
cycle
, and call the
subroutine to calculate the flank wear land width, let ) , ( wearrate t Flankwear ∆
m
VB wearrate t Flankwear error − ∆ = ) , ( .
If ≤ error δ , then the searching procedure will end, and the present value
will be output as the result. Otherwise, if error < -
t ∆
δ , then the cycle number
lower limit , else, the cycle number upper limit ,
repeat step 5 until
cycle cycle
N N =
1 cycle cycle
N N =
2
≤ error δ is satisfied.

6.3.6 Nodal Displacement Calculation

Nodal displacement is calculated at every tool face node by

) , ( ) , ( ) , ( j i j j i j i
D t w w
r
&
r
⋅ ∆ ⋅ = (6.5)

where
Estimation Of Tool Wear In Milling Operation 111
w
r
is the nodal displacement vector;
i is nodal label;
j is tool wear calculation cycle number.
In addition, some nodes on flank face, e.g. node b and c in Fig. 6.9, are moved as
well as explained above.

6.3.6 Tool Geometry Updating

Tool geometry updating is performed with the same procedure as explained in
turning operation.

6.4 Results & Discussion

With this tool wear program, tool wear under the same cutting condition as described
in Table 3.2 is estimated. ∆ is specified by the user, permitted error is
set to 0.01mm. After the cutting time of about 603s, the new tool in Fig. 6.10(a) is
updated to the worn tool in Fig. 6.10(b), which has a flank wear width of 0.06mm and
crater wear on the rake face.
mm VB 05 . 0 =









VB=0.06mm
Crater wear
Flank wear

VB=0.06mm

VB=0.06mm VB=0.06mm
Crater wear
Flank wear
Crater wear
Flank wear
(a) t=0s (b) t=603s
Fig. 6.10 Tool wear profile

A discrepancy between the estimated tool wear from the program and the expected
tool wear from experiment is unavoidable mainly because of the following reasons:
(1) Complex tool wear in milling operation. In this study, only the abrasive and
adhesive wear are considered, whereas the main tool wear in high-speed-
milling results from chipping, thermal crack, etc. They reduce the strength of
the tool edge and accelerate tool wear.
Estimation Of Tool Wear In Milling Operation 112
(2) Chip formation analysis modeling. It is verified by the test that the chip
thickness and tool/chip contact are sensitive to the element size and given
value of shear failure criterion, when shear failure criterion is used as the chip
separation method. A multi variable dependent shear failure criterion may
provide a better chip formation simulation result and improve the predicted
tool wear profile and tool wear value.
Estimation Of Tool Wear In Milling Operation 113

6.5 Summaries & Conclusion

In this chapter, a tool wear estimation model is implemented for the milling operation.
The main findings of this study are as follows:
• Milling operation is an intermittent cutting process. By sampling the cutting
process variables in the chip formation analysis and heat transfer analysis at
the corresponding time points and calculating the nodal average wear rate per
cycle, tool wear estimation modelling can be implemented.
• Because of the temperature difference in the cyclical thermal balance state
and in the milling cycle before cycle thermal balance state is reached, tool
temperature in the cyclical thermal balance state should be used in tool wear
estimation. Otherwise, a very slow tool wear process is expected because in
Usui’s tool wear equation, wear rate and tool temperature has an exponent
relationship.
• With the developed tool wear program, tool wear under a cutting condition with
high cutting speed is calculated, both crater wear and flank wear are formed
on the tool face.









Summary And Prospect 114
Chapter 7 Summary And Prospect

7.1 Summaries

In this study the methodologies to numerical implementation of tool wear estimation
in turning and milling operation are discussed. Based on the researches of tool wear
mechanism, which show that wear rate of cutting tool is dependent on some cutting
process variables such as tool temperature, sliding velocity of workpiece material and
normal pressure on tool face, the preliminary qualitative tool wear estimation models
are developed.
Tool wear estimation in turning operation is based on the study of finite element
simulation of steady-state cutting process.
A new chip formation modeling method is developed to simulate the entire process
from initial chip formation, chip growth to steady state. Chip separation is formed
automatically by solution-dependent mesh adaptivity instead of material failure
criterion. It is not necessary to get material failure parameters or chip geometry from
experiment. Instead it provides an alternative method to decide material failure
parameters. In addition, no separation path is preset in advance. No obvious crack is
formed in front of the cutting edge. This chip formation model is verified by
experimental data. When the coefficient of friction calculated according to cutting
force in experiment is used, the error of two cutting force components is smaller than
5% compared to the experimental data.
Pure heat transfer analysis of only the cutting tool is carried out to save the
calculation time to reach thermal steady state. Temperature dependent heat flux at
tool-chip interface and heat convective and radiation of tool face are considered
during modeling. After only several minutes of calculation the cutting tool gets a
steady temperature distribution.
Then the problems about calculating nodal wear rate at steady state according to
Usui’s tool wear equation, cutting time increment searching, nodal displacement
calculation and geometry updating are discussed. A tool wear estimation program is
developed. It can calculate the tool wear until the tool reshape criterion is reached.
The estimated tool wear is verified by experimental data. It is assumed that the error
is created by the low coefficient of friction in chip formation analysis.
Because milling operation is intermittent cutting process, the chip formation, heat
transfer and tool wear estimation modeling are different from turning operation.
Summary And Prospect 115
The chip formation simulation is realized by introducing shear failure criterion. The
strain at failure in shear failure criterion is defined according to the former chip
formation modeling method. Shear failure criterion is applied to the entire workpiece.
Pure heat transfer analysis of only the workpiece shows that the workpiece cool
down to room temperature in the cooling phase of one milling cycle if the cutting
speed is not tool high. Accordingly, the chip formation process in every milling cycle
is assumed similar because of the negligible temperature increment in the workpiece.
Pure heat transfer analysis of only the cutting tool is carried out for 8 milling cycles.
Temperature is observed after every milling cycle. No cyclical thermal balance state
is realized. In order to speed up the realization process of cyclical thermal balance
state, different preheated cutting tool is used in the cutting process. It is found that
the temperature in cyclical thermal balance state is higher than the first several
milling cycle.
Tool wear estimation in milling operation is performed by calculation nodal average
wear rate in one milling cycle and use it to other milling cycles. Because in milling
cycle the cutting tool is rotating instead of the workpiece as in turning operation, all
the problems about cutting time increment searching, nodal displacement calculation
and geometry updating are different from turning operation and they are discussed.
Then a tool wear estimation program for milling operation is developed. Using this
program, tool wear in one milling case is calculated. Both crater wear and flank wear
are formed.
During the study multi aspects of cutting process simulation modelling in turning and
milling operations including chip formation analysis, heat transfer analysis, and tool
wear estimation are studied. In order to fulfil the purpose of tool wear study, multi-
programming tools including commercial FE code ABAQUS/Explicit,
ABAQUS/Standard, Fortran, Python are employed and integrated. This lays a ground
for the study on more complex problem and the extension of functionality of FEM in
the future.
Tool wear estimation with the help of finite element method can predict not only tool
life, but also wear profile of both crater wear and flank wear, and relate tool wear with
some wear mechanisms. This tool wear estimation method will relate the geometry
appearance to physical basic of tool wear and bridge the gap between macro and
micro studies of tool wear. This is very meaningful for the scientific research and
education. For tool designer, it is very helpful to optimise tool geometry and structure
Summary And Prospect 116
knowing wear profile and wear mechanism; for material engineer, it is useful to
improve tool material according to the determined main wear mechanism. In this tool
wear estimation method, tool wear is related to wear mechanism, once tool wear
mathematical model for a combination of tool-workpiece material is determined, it is
possible to estimate tool wear by program without doing any experiment. In addition,
this estimation method is helpful to reduce the size of various cutting database by
replacing tool life equation with tool wear mathematical model, which is applicable to
wider cutting range.

7.2 Prospect

The tool wear estimation models should be improved in several aspects:
After a certain tool wear is formed, the chip formation analysis of steady state
sometimes produces relatively low tool temperature on flank wear. This may be
caused by the contact problem between the flank wear and the workpiece. In order to
produce good contact in this area a negative flank angle designed on the flank wear
may be a good solution.
In tool geometry updating, the formed wear profile is not smooth even after the
second updating step and some nodes have to be adjusted manually. A special
smoothing algorithm should be designed to solve this problem.
Friction has big influence on the chip formation analysis and tool wear. When using
the coefficient of friction calculated according to the cutting force from experiment,
the result of tool wear estimation is maybe improved.
A further improvement of tool wear estimation may be realized by introducing multi
wear mechanism. The tool wear is calculated according to their combination. It will be
possible to study on the contribution of every wear mechanism under different cutting
conditions.
Because in the tool wear estimation modeling, flank wear is produced by moving
nodes individually according to the nodal wear rate instead of according to an
average value. It provides a method to produce the complex tool wear in 3D, such as
wear notch.
In order to spread the application of this method in industry practices, except
improving the precision of tool wear estimation, it is necessary to develop tool wear
mathematical model for most common used materials, develop tool wear estimation
Summary And Prospect 117
model for coated carbide tool, CBN cutting tool, ceramics cutting tool, etc, and
research on 3D tool wear estimation model in the future.
References 118
References

[ABA-01a] ABAQUS/Explicit User’s Manual Version 6.2, HSK, Inc., U.S.A., 2001
[ABA-01b] ABAQUS/Standard User’s Manual Version 6.2, HSK, Inc., U.S.A.,
2001
[ABA-01c] ABAQUS Scripting Manual Version 6.2, HSK, Inc., U.S.A., 2001
[Ahma-89] Ahmad, M. M., Draper, W. A., and Derricott, R. T., An application of the
finite element method to the prediction of cutting tool performance, Int.
J. Mach. Tools Manufact. 29 (2) (1989) 197-206.
[Amon-99] Amontons, Historie de l’Academie Royale des Sciences avec les
Memoires de Mathematique et de Physique, 1699.
[Arnd-73] Arndt, G., Ultra-high-speed machining: a review and an analysis of
cutting forces, Applied Mechanics Group, Proc. Instn. Mech. Engrs.
187 (1973) 625-634.
[Arol-02] Arola, D., Sultan, M. B., Ramulu, M., Finite element modeling of edge
trimming fiber reinforced plastics, Journal of Manufacturing Science
and Engineering 124 (2) (2002) 32-41.
[Arra-02] Arrazola, P. J., Meslin, F., Hamann, J-C. and Le Maitre, F., Numerical
cutting modeling with Abaqus/Explicit 6.1, 2002 ABAQUS Users’s
Conference, Newport, Rbode Island, May 29-31, 2002.
[Atha-98] Athavale, S. M., Strenkowski, J. S., Finite element modeling of
machining: from proof-of-concept to engineering applications, in:
Proceedings of the CIRP International Workpshop on Modeling of
Machining Operations, Atlanta, USA, 1998, pp. 203-216.
[Baca-00] Bacaria, J. L., Dalverny, O., Pantale, O., Rakotomalala, R., and
Caperaa, S., 2D and 3D numerical models of metal cutting with
damage effects, European Congress on Computational Methods in
Applied and Engineering, Barcelona, Sep. 11-14, 2000.
[Bäke-00] Bäker, M., Rösler, J., Siemers, C., High speed chip formation of
Ti6Al4V, Proceedings of Materials Week, Munich, 2000
[Bäke-02] Bäker, M., Rösler, J., Siemers, C., Finite element simulation of
segmented chip formation of Ti6Al4V, Journal of Manufacturing
Science and Engineering 124 (2002) 485-488.
[Batz-02] Batzer, S. A., Subhash, G., Olson, W., and Sutherland, J. W., Large
References 119
strain constitutive model development for application to orthogonal
machining, in: Proc. 5
th
CIRP International Workshop on Modeling of
Machining Operations, West Lafayette, USA, 2002, pp. 31-40.
[Behr-98a] Behrens, A., Westhoff, B., Anwendung von MARC auf
Zerspanprozesse, MARC Benutzertreffen 1998, München, 1998.
[Behr-98b] Behrens, A., Westhoff, B., Probleme und Herausforderungen bei der
Simulation von Zerspanprozessen, XXV. FEM-Kongress, Baden-
Baden, 1998.
[Behr-99] Behrens, A., Westhoff, B., Finite element modeling of high speed
machining processes, 2
nd
International German and French
Conference on High Speed Machining, Darmstadt, 1999, pp. 185-190.
[Boot-89] Boothroyo, G., Knight, W. A., Fundamentals of machining and machine
tools, Marcel Dekker, Inc. USA, 1989
[Chil-89] Childs, T., Mahdi, M., On the stress distribution between the chip and
tool during metal turning, Ann. CIRP 38 (1) (1989) 55-58.
[Cere-96] Ceretti, E., Fallböhmer, p., Wu, W.-T., Altan, T., Application of 2D FEM
to chip formation in orthogonal utting, Journal of Materials Processing
Technology 59 (1996) 169-180.
[Cere-99] Ceretti, E., Lucchi, M., Altan, T., FEM simulation of orthogonal cutting:
serrated chip formation, Journal of Materials Processing Technology
95 (1999) 17-26.
[Cere-00] Ceretti, E., Lazzaroni, C., Menegardo, L., Altan, T., Turning simulations
using a three-dimensional FEM code, Journal of Materials Processing
Technology 98 (2000) 99-103.
[Chil-98] Childs, T. H. C., Dirikolu, M. H., and Maekawa, K., Modelling of friction
in the simulation of metal machining, in: 24
th
Leeds-Lyon Symposium
on Tribology. Amsterdam: Elsevier Tribology Series No. 34, 1998,
pp.337-46.
[Chuz-03a] Chuzhoy, L., Devor, R. E., and Kapoor, S. G., Machining simulation of
ductile iron and its constituents, part 1: estimation of material model
parameters and their validation, Journal of Manufacturing Science and
Engineering 125 (5) (2003) 181-191.
[Chuz-03b] Chuzhoy, L., Devor, R. E., and Kapoor, S. G., Machining simulation of
ductile iron and its constituents, part 2: numerical simulation and
References 120
experimental validation of machining, Journal of Manufacturing
Science and Engineering 125 (5) (2003) 192-201.
[Coul-85] Coulomb, C. A., Memoires de Mathematique et de Physique de
l’Academie Royale des Sciences, Paris, 1785.
[Didj-97] Didjanin, L. and Kovac, P., Fracture mechanisms in chip formation
processes, Materials Science and Technology 13 (1997) 439-444.
[Dill-00] Dillon, O. W., Zhang, H., An analysis of cutting using a grooved tool,
International Journal of Forming Processes 3 (1-2) (2000) 115-130.
[Feve-01] Feve, S., Yen, Y-C., Altan, T., Tool wear in orthogonal turning of AISI
1045 steel with uncoated tungsten carbide tool, ERC/NSM Report No.
HPM/ERC/NSM-01-R-13, The Ohio State University, 2001.
[Finn-56] Finnie, I., Shaw, M. C., The friction process in metal cutting,
Transactions of the ASME 78 (1956) 1649-57.
[Fran-02] Frank, P., Improvement of the FEM-based Predictive Model of Tool
Wear, Diplomarbeit, Universität Karlsruhe (TH), 2002.
[Gu-02] Gu, L. Z., Wang, d., Xing, L., Chang, J., Chen, G. J., Computer
simulation and optimization of metal cutting process for mild carbon
steels, Journal of Materials Processing Technology 129 (2002) 60-65.
[Guo-00] Guo, Y. B., Dornfeld, D. A., Finite element modeling of burr formation
process in drilling 304 stainless steel, Journal of Manufacturing
Science and Engineering 122 (2000) 612-619.
[Guo-02] Guo, Y. B., Liu, C. R., 3D FEA modeling of hard turning, Journal of
Manufacturing Science and Engineering 124 (2002) 189-199.
[Hast-79] Hastings, W.F., Mathew, P., Oxley, L.B., Taylor, J., Proc. 20
th
Int.
M.T.D.R. Conf., 1979, pp.313
[Huan-96] Huang, J. M., Black, J. T., An evalution of chip separation criteria for
the FEM simulation of machining, Journal of Manufacturing Science
and Engineering 118 (1996) 545-554.
[Iwat-84] Iwata, K., Osakada, K., Terasaka, Y., Process modeling of orthogonal
cutting by the rigid-plastic finite element method, Journal of
Engineering Materials and Technology 106 (1984) 132-138.
[John-01] Johnson, D., Why cutting tools fail, CNC Machining Magazine 5 (18)
(2001).
[Kalh-01] Kalhori, V., Modelling and Simulation of Mechanical Cutting, Doctoral
References 121
thesis, Institutionen för Maskinteknik, Avdelningen för Datorstödd
maskinkonstruktion, Lulea Tekniska Universitet, 2001.
[Kato-72] Kato, S., Yamaguchi, K., Yamada, M., Stress distribution at the
interface between tool and chip in machining, Journal of Engineering
for Industry 94 (1972) 683-689.
[Kim-99] Kim, K. W., Lee, W. Y., Sin, H. C., A finite-element analysis of
machining with the tool edge considered, Journal of Materials
Processing Technology 86 (1999) 45-55.
[Kita-88] Kitagawa, T., Maekawa, K., Shirakashi, T. and Usui, E., Analytical
prediction of flank wear of carbide tools in turning plain carbon steels,
part 1: characteristic equation of flank wear, Bull. Japan Soc. of Prec.
Engg. 22 (4) (1988) 263-269.
[Kita-89] Kitagawa, T., Maekawa, M., Shirakashi, T. and Usui, E., Analytical
prediction of flank wear of carbide tools in turning plain carbon steels,
part 2: prediction of flank wear, Bull. Japan Soc. of Prec. Engg. 23 (2)
(1989) 126-133.
[Klam-73] Klamecki, B. E., Incipient chip formation in metal cutting - a three
dimension finite element analysis, Ph. D. dissertation, University of
Illinois, 1973.
[Kopp-01] Koppka, F., Sartkulvanich, P. and Altan, T., Experimental
Determination of Flow Stress Data for FEM Simulation of Machining
Operations, ERC/NSM Report No. HPM/ERC/NSM-01-R-63, The Ohio
State University, 2001.
[Köni-84] König, W., Fertigungsverfaren / Manufacturing Methods, VDI-Verlag,
Dusserdolf, Germany, 1984.
[Kwon-00] Kwon, P., Predictive models for flank wear on coated inserts, Journal
of Tribology 122 (1) (2000) 340-347.
[Leop-98] Leopold Jürgen, FEM modeling and simulation of 3-D chip formation,
in: Proceedings of the CIRP International Workpshop on Modeling of
Machining Operations, Atlanta, USA, 1998, pp. 235-245.
[Li-02] Li, K., Gao, X.-L., Sutherland, J. W., Finite element simulation of the
orthogonal metal cutting process for qualitative understanding of the
effects of crater wear on the chip formation process, Journal of
Materials Processing Technology 127 (2002) 309-324.
References 122
[Lim-01] Lim, C. Y. H., Lau, P. P. T., Lim, S. C., The effects of work material on
tool wear, Wear 250 (2001) 344-348.
[Lin-99] Lin, L.-C., Lin, Y.-Y., A study of an oblique cutting model, Journal of
Materials Processing Technology 86 (1999) 119-130.
[Lin-00] Lin, Z.-C., Zheng, Y.-L., Three-dimensional thermo-elastic-plastic finite
element analysis with different cutting speeds, International Journal of
Computer Applications in Technology 13 (2000) 210-220.
[Lin-01b] Lin, Z.-C., Lin, Y.-Y., A study of oblique cutting for different low cutting
speeds, Journal of Materials Processing Technology 115 (2001) 313-
325.
[Liu-00] Liu, C. R., Guo, Y. B., Finite element analysis of the effect of
sequential cuts and tool-chip friction on residual stress in a machined
layer, International Journal of Mechanical Science 42 (2000) 1069-
1086.
[Lutt-98] van Luttervelt, C. A., Childs, T. H. C., Jawahir, I. S., Klocke, F.,
Venuvinod, P. K., Keynote papers: present situation and future trends
in modeling of machining operations progress report of the CIRP
Working Group ‘Modelling of Machining Operations’, Annals CIRP 47
(2) (1998) 587-626.
[Maru-95] Marusich, T. D., Oritz, M., Modelling and simulation of High-Speed
Machining, International Journal for Numerical Methods in Engineering
38 (1995) 3675-3694.
[Maru-02] Marusich, T. D., Brand, C. J., and Thiele, J. D., A methodology for
simulation of chip breakage in turning processes using an orthogonal
finite element model, in: Proc. 5
th
CIRP International Workshop on
Modeling of Machining Operations, West Lafayette, USA, 2002, pp.
139-148.
[McCl-02] McClain, B., Batzer, S. A., Maldonado, G. I., A numeric investigation of
the rake face stress distribution in orthogonal machining, Journal of
Materials Processing Technology 123 (2002) 114-119.
[Merc-98] Merchant, E., in Proceedings of the CIRP International Workshop on
Modeling of Machining Operations, Atalantic, USA, 1998
[Moli-00] Molinari, J. F., Ortiz, M., Radovitzky, R., and Repetto, E. A., Finite-
element modeling of dry sliding wear in metals, Engineering
References 123
Computations 18 (3/4) (2001) 592-609.
[Mona-99] Monaghan, J. and MacGinley, T., Modelling the orthogonal machining
process using coated carbide cutting tools, Computational Materials
Science 16 (1999) 275-284.
[Mova-02] Movahhedy, M. R., Altintas, Y., Gadala, M. S., Numerical analysis of
metal cutting with chamfered and blunt tools, Journal of Manufacturing
Science and Engineering 124 (2002) 178-187.
[Ng-99] Ng, E.-G., Aspinwall, D. K., Brazil, D., and Monaghan, J., Modelling of
temperature and forces when orthogonally machining hardened steel.,
International Journal of Machine Tools & Manufacture 39 (1999) 885-
903
[Ng-02a] Ng, E.-G. and Aspinwall, D. K., Modelling of hard part machining,
Journal of Materials Processing Technology 127(2002) 222-229.
[Ng-02b] Ng, E.-G., EI-Wardany, T., Dumitrescu, M., and Elbestawi, M. A., Proc.
5
th
CIRP International Workshop on Modeling of Machining Operations,
West Lafayette, IN, USA, May 20-21, 2002, pp. 1-19.
[Ohbu-03] Ohbuchi, Y., Obikawa, T., Finite element modeling of chip formation in
the domain of negative rake anlge cutting, Journal of Engineering
Materials and Technology 125 (2003) 324-332.
[Okus-71] Okushima, K., Kakino, Y., The residual stress produced by metal
cutting, Ann. CIRP 20 (1) (1971) 13-14.
[Owen-99] Owen, D. R. J., M. Vaz Jr., Computational techniques applied to high-
speed machining under adiabatic strain localizetion conditions,
Computer methods in applied mechanics and engineering 171 (1999)
445-446.
[Özel-00a] Özel, T., Altan, T., Process simulation using finite element method –
prediction of cutting force, tool stress and temperatures in high-speed
flat end milling, International Journal of Machine Tools & Manufacture
40 (2000) 713-738.
[Özel-00b] Özel, T., Altan, T., Determination of workpiece flow stress and friction
at the chip.tool contact for high-speed cutting, International Journal of
Machine Tools & Manufacture 40 (2000) 133-152.
[Özel-02] Özel, T., Modeling of hard part machining: effect of insert edge
preparation in CBN cutting tools, Journal of Materials Processing
References 124
Technology (2002).
[Rice-60] Rice, W. B., Salmon, R., Syniuta, W. D., Photoelastic determination of
cutting tool stress, Transactions of the E. I. C. 4 (1) (1960) 20-23.
[Sand-98] Sandstrom, D. R., Modeling the physics of metal cutting in High-Speed
machining, in: Proceedings of the CIRP International Workpshop on
Modeling of Machining Operations, Atlanta, USA, 1998, pp. 217-224.
[Schm-02] Schmidt, C., Development of a FEM-based Tool Wear Model to
Estimate Tool Wear and Tool Life in Metal Cutting, Diplomarbeit,
Universität Karlsruhe (TH), 2002.
[Schu-00] Schulze, V. and Vöhringer, O., Influence of alloying elements on the
strain rate and temperature dependence of the flow stress of steels,
Metallurgical and Materials Transactions A 31A (2000).
[Shat-00] Shatla, M., Yen, Y.-C., Altan, T., Tool-workpiece interface in orthogonal
cutting-application of FEM modeling, Transactions of NAMRI/SME
2000, XXVIII, 177-178.
[Shat-01a] Shatla, M., Kerk, C., Altan, T., Process modeling in machining. part I:
determination of flow stress data, International Journal of Machine
Tools & Manufacture 41 (2001) 1511-1534a.
[Shat-01b] Shatla, M., Kerk, C., Altan, T., Process modeling in machining. part II:
validation and applications of the determined flow stress data,
International Journal of Machine Tools & Manufacture 41 (2001) 1659-
1680.
[Shet-00] Shet, C., Deng, X., Finite element analysis of the orthogonal metal
cutting process, Journal of Material Processing technology 105 (2000)
95-109.
[Shet-03] Shet, C., Deng, X., Residual stress and strains in orthogonal metal
cutting, International Journal of Machine Tools & Manufacture 43
(2003) 573-587.
[Shi-02] Shi, G. Q., Deng, X. M., Shet, C., A finite element study of the effect of
friction in orthogonal metal cutting, Finite Elements in Analysis and
Design 38 (2002) 863-883.
[Shih-91] Shih, A., Yang, H. T. Y., Experimental and finite element simulation
methods for rate-dependent metal forming processes, International
Journal for Numerical Methods in Engineering 31 (1991) 345-367.
References 125
[Shih-93] Shih, A. J., Yang, H. T. Y., Experimental and finite element predictions
of residual stresses due to orthogonal metal cutting, International
Journal for Numerical Methods in Engineering 36 (1993) 1487-1507.
[Shih-95] Shih, A. J., Finite element simulation of orthogonal metal cutting,
Journal of Engineering for Industry 117 (1995) 84-93.
[Shih-96] Shih, A. J., Finite element analysis of the rake angle effects in
orthogonal metal cutting, International Journal of Mechanical Sciences
38 (1) (1996) 1-17.
[Shir-93] Shirakashi, T., New trend on machining theory-way to computational
machining, Int. J. Jpn. Soc. Prec. Eng. 27 (1993) 299-302.
[Söhn-01a] Söhner, J. and Altan, T., Material Database for Manufacturing
Simulation (MADAMS): Summary of the Activities and Flow Stress
Database, ERC/NSM Report No. HPM/ERC/NSM-01-R-76, ERC for
Net Shape Manufacturing, Ohio State University, 2001
[Söhn-01b] Söhner, J. and Yen, Y.-C., Estimation of tool wear in metal cutting with
the finite element method –a progress report-, ERC/NSM Report No.
HPM/ERC/NSM–01–R-67, ERC for Net Shape Manufacturing, Ohio
State University, 2001
[Söhn-03] Söhner, J., Beitrag zur Simulation zerspanungstechnologischer
Vorgäng mit Hilfe der Finite-Element-Methode, Dissertation,
Universität Karlsruhe (TH), 2003
[Stre-93] Strenkowski, J. S., Carroll III, J. T., A finite element model of
orthogonal metal cutting, Journal of Engineering for Industry 107
(1985) 347-354.
[Stre-02] Strenkowski, J. S., Shih, A. J., Lin, J.-C., An analytical finite element
model for predicting three-dimensional tool faces and chip flow,
International Journal of Machine Tools & Manufacture 42 (2002) 723-
731.
[Tay-74] Tay, A. O., Stevenson, M. G., David, G. V., Using the finite element
method to determine temperature distributions in orthogonal
machining, Proc. Inst. Mech. Eng. 188 (1974) 627-638.
[Tayl-07] Taylor, F.W., Trans. Am. Soc. Mech. Engrs 28,31 (1907)
[Tren-77] Trent, E.M., Metal Cutting, Butterworths, England, 1977
[Usui-78a] Usui, E., Hirota, A., Analytical prediction of three dimensional cutting
References 126
process, part 2: chip formation and cutting force with conventional
single-point tool, Journal of Engineering for Industry 100 (5) (1978)
229-235.
[Usui-78b] Usui, E., Hirota, A., Masuko, M., Analytical prediction of three
dimensional cutting process, part 1: basic cutting model and energy
approach, Journal of Engineering for Industry 100 (5) (1978) 222-228.
[Usui-78c] Usui, E., Shirakashi, T. and Kitagawa, T., Analytical prediction of three
dimensional cutting process, part 3: cutting temperature and crater
wear of carbide tool, Journal of Engineering for Industry 100 (5) (1978)
236-243.
[Usui-84] Usui, E., Obikawa, T., Shirakashi, T., Study on chip segmentation in
machining Titanium alloy, in: Proceedings of the 5
th
International
Conference on Production Engineering, Tokyo, Japan, 1984, pp. 233-
239.
[Walt-98] Walter, U., Einfluss des Kühlschmierstoffes beim Fräsen, Dissertation
Universität Karlsruhe (TH), Karlsruhe, Germany, 1998
[West-01] Westhoff, B., Modellierungsgrundlagen zur FE-Analyse von HSC-
Prozessen, Dissertation, Institut für Konstrucktions- und
Fertigungstechnik, Laboratorium Fertigungstechnik, Universität der
Bundeswehr, 2001.
[Wu-96] Wu, J. S., Dillon, O. W., Lu, W.-Y., Thermo-viscoplastic modeling of
machining process using a mixed finite element method, Journal of
Manufacturing Science and Engineering 118 (1996) 470-482.
[Yang-02] Yang, X.-P., Liu, C. R., A new stress-based model of friction behavior
in machining and its significant impact on residual stresses computed
by finite element method, International Journal of Mechanical Sciences
44 (2002) 703-723.
[Yen-02] Yen, Y. C., Söhner, J., Weule, H., Schmidt, J. and Altan, T., Estimation
of tool wear of carbide tool in orthogonal cutting using FEM simulation,
in: Proceedings of the 5th CIRP International Workshop on Modeling of
Machining Operations, West Lafayette, USA, 2002, pp.149-160
[Zhan-94] Zhang, B., Bagchi, A., Finite element simulation of chip formation and
comparison with machining experiment, Journal of Engineering for
Industry 116 (8) (1994) 289-297.
References 127
[Zore-63] Zorev, N.N., Inter-Relationship between Shear Processes occurring
along Tool Face and Shear Plane in Metal Cutting, International
Research in Production Engineering (1963) 42-49.


128

Resume


Name: Lijing Xie
Nationality: China
Date and place of birth: June. 17, 1971 in China
Family Status: Married with Dan Zheng since May. 18, 1998


Education:

1978-1983 Tangshan Nanxindao Elementary School, Hebei, China
1983-1986 The Eighth Junior High School, Tangshan, China
1986-1989: The First Senior High School, Tangshan, China
1989-1993 B.S., Department of Mechanical Engineering, Tangshan
Institute of Technology, Hebei, China
1993-1996 M.S., Department of Mechanical Engineering, Beijing
Institute of Technology, Beijing, China


Employment History:

1996-2000 Lecturer at Beijing Institute of Technology, Beijing, China
Since Nov. 2000 Scientist at Werkzeugmaschinen und Betriebstechnik
(WBK), Uni-Karlsruhe (TH), Germany



Forschungsberichte aus dem wbk Institut für Produktionstechnik Universität Karlsruhe (TH) Hrsg.: Prof. Dr.-Ing. Jürgen Fleischer Prof. Dr.-Ing. Hartmut Weule

Lijing Xie Estimation Of Two-dimension Tool Wear Based On Finite Element Method

ISSN 0724-4967 Band 120

© wbk Institut für Produktionstechnik Universität Karlsruhe (TH) alle Rechte vorbehalten Druck: Schnelldruck Ernst Grässer. Karlsruhe Tel: 0721/61 50 50 ISSN 0724-4967 .

-Ing. In diesem Sinne soll im Rahmen dieser Schriftenreihe in zwangloser Folge über aktuelle Forschungsergebnisse des Instituts für Werkzeugmaschinen und Betriebstechnik der Universität Karlsruhe berichtet werden. Dr. Prof. die durch Nutzung informationsverarbeitender Systeme eine Verbesserung der Leistungsfähigkeit fertigungstechnischer Einrichtungen und deren informationstechnischorganisatorische Einbindung in automatisierte Produktionssysteme ermöglichen. Maschinenkomponenten und Fertigungseinrichtungen insbesondere Aufgabenstellungen.Vorwort des Herausgebers Der rasche Fortschritt der Produktionstechnik und der weltweite Wettbewerb um technisch-wirtschaftliche Spitzenpositionen machen einen intensiven Austausch von Wissen und Erfahrung zwischen Universitäten und der Industrie erforderlich. Die Forschungsaktivitäten des Instituts umfassen neben der Untersuchung und Optimierung von Bearbeitungsverfahren.-Ing. Jürgen Fleischer Prof. Dr. Hartmut Weule .

.

-Ing. Dr. 2004 Hauptreferent: Korreferent: Prof. 02. Lijing Xie aus China Tag der mündlichen Prüfung: 05.Estimation Of Two-dimension Tool Wear Based on Finite Element Method Zur Erlangung des akademischen Grades eines Doktors der Ingenieurwissenschaften von der Fakultät für Maschinenbau der Universität Karlsruhe (TH) genehmigte Dissertation von M. Prof. Dieter Spath . Dr. Jürgen Schmidt o. Sc.-Ing.

.

Anurag Jain for the helpful suggestion and discussion in the research and help in personal living. Frank Biesinger for kindly offering the developed material subroutine. friendly. Ivan Tzitzelkov for solving many problems in my simulation work. Mr. Dr.-Ing.-Ing. In this period. especially Dr. for their support. Dipl. Dr. Jürgen Fleischer for their kindly concern in my living and work. Mr. intense research work and interesting activities. and their highly appreciated instruction. Thanks to all the members in group FT. special thanks are given to Dipl. Especially.-Ing. Jörg Söhner.-Ing. Especially. o. Mr. Prof. and active persons.-Ing. Dr. Jürgen Schmidt and o. The last three years has been a precious experience for me. 2000. I was touched by their friendship. Dieter Spath. I feel very fortunate to have an opportunity to concentrate on the interesting research field of manufacturing industry and get to know so intelligent. Hartmut Weule and o. with excellent learning. Michael Heinz for the warm-hearted assist and patient instruction in my experiment work.-Ing. Sc. Prof. M. Thanks to all the members in the institute for the unforgettable happy time in the past three years. I want to express my deepest appreciation and thanks to my supervisors. Xibin Wang for their constantly encouragement and help.-Ing.-Ing. Carsten Schmidt and Mr. Dr. I would like to express my thanks to Prof. their careful reviews of my papers and dissertation. I get uncountable unselfish help from them. This thesis is finished under the cooperation with scientists in Institut für Werkstoffe I.Acknowledgement The present research work was carried out at Institut für Produktionstechnik (WBK) in University of Karlsruhe (TH) since Nov. Dr. Klaus Simon for offering instruction about . Siqin Pang and Prof. Thanks to Prof. Prof. Mr.

in December 2003 Lijing Xie . I want to thanks my husband. Margarethe Schüßler for teaching me Deutsch language voluntarily. At last. Thanks to Mrs. Dan.measuring basic knowledge and helping me to look for the best measuring method. and my family and Dan’s for their love and support. Thomas Hildenbrand for preparing experimental condition and troubleshooting in the turning experiment. Mr. Karlsruhe. They give me the strength over all the problems in my research.

...............1...................................... 18 1.................................. 24 1..............2 Mechanical Aspects .....................................................................................................................2 Stability Limit .......2 Tool Wear Estimation With FEM............................................................................2 Approach .1......2.......1...................1.....2..1 Dynamic Analysis Procedure ............. 19 1..1.................................................... 8 1................ 27 1....................................................................................................................................................... 34 3..................2....................................1.......... 31 2.............................................2 Continuous Chip Formation Simulation...... 16 1............1...........3........................1 Comparison Between FEM Method And Empirical Method ............................................2 Technical Background About Tool Wear.....2.................2 Thermal Analysis Procedure ..............1..................................2 Material Constitutive Model ........ 24 1..................................................................................................................... 35 3........................... 34 3...... 37 .......................... 32 Chapter 3 Chip Formation Simulation Technology.1 Objectives.....2.....................................2 State Of Art: Numerical Implementation Of Tool Wear Estimation.................2 Mesh Adaptivity .....................3 Research Of Tool Wear With Finite Element Methods ......1.3...............2................................................................................................................................................................................1 State Of Art: Finite Element Simulation Of Cutting Process ...............................1 Wear Types In Metal Cutting ........... 36 3............................... 22 1..................... 13 1............ 5 1.......................................................................................................................Table of Contents I Table of Contents Chapter 1 Introduction.......2..3 Chip Separation.1 Contact And Friction ...3 Summary Of Literature ..............1........................................ 17 1........................................................1 Explicit Algorithm In Chip Formation Simulation...................................... 8 1.. 5 1...........2 Wear Mechanism...........................................1...1 Numerical Aspects.......................................................................................................................................... 1 1................................................................................... 22 1.........2............. 3 1..............1...........................1 Introduction ..1 Approach ..................1.........3...................................................1.....3..3 Tool Wear Model ...............................1 Tool Wear Estimation With The Combination Of Analytical Method And FDM ........... 31 2.. 6 1........................ 29 Chapter 2 Objective And Approach................2............ 34 3....1...........................3........................ 11 1.. 34 3...

.........3.................................................3 Result & Discussion ...........1 Boundary Region Types.....3..............................1................................................2....................................... 64 4............................................. 55 3..2..... 45 3....2.......... 39 3...5............ 43 3..............................1 Stress Analysis ......1 Limitation Of The Existing Chip Formation Models............................................2.................. 51 3...2 Cutting Temperature..... 53 3.2 Plastic Strain Analysis ..4 Temperature Analysis.............3.................2.......................................................... 63 Chapter 4 Heat Transfer Analysis In Metal Cutting ..........3........5 Verification With Experimental Data........................................3....5......3 Cutting Force Analysis.....5.............................1 Introduction ............... 49 3...................... 48 3..........................2.......................................3....................................................................... 64 4........2 Geometry Features ......2.............2.. 64 4...................................2 General Considerations ...........................2 Heat Flux.................3 Continuous Steady-state Chip Formation ......3 Adaptive Meshing Technique In ABAQUS/Explicit ...................4....................3....................3..............1 Geometry And Mesh ............. 55 3.......................... 54 3........................ 64 4.............3......... 40 3............................2..... 41 3....2........... 40 3......................2.........4....................................................................... 59 3.........................................................................................................................2 Chip Growth ........... 42 3.............................1 Initial Chip Formation........................................2........3 Curvature Refinement ......5.........Table of Contents II 3........................................................2 Advantages Of The New-developed Chip Formation Model .....................................................................................................................2................................................. 41 3.4........................ 45 3...................2............1 Shear Failure Criterion ......................................................3..........4 Analysis Steps................. 58 3.........................................................3..................2 Chip Formation Modeling ..2... 61 3......................................3 Strain Rate....2.................................................................................................... 62 3................................................5 Results & Discussion ................................. 65 4................................... 66 .................................................1 Modelling...........................1 Stress Analysis .......... 66 4....................................................................................................3.................3................................. 59 3...................... 56 3.......................................4 Summaries & Conclusion ......3 Chip Formation Simulation For Milling Operation..................1 Chip Separation ...................................................1...2..................3........ 48 3..... 37 3............................................................2 A Numerical Method To Determine Strain At Failure ..........5.............................2........3...... 50 3..................................................................................3 In Turning Operation.....

..................... 70 4.......................... 84 5.................................................................1 Tool Wear .......................2 Results & Discussion ............................................................. 94 5.................... 70 4.....2.........................................4............................... 68 4... 86 5.............4.........................................................................................................................................3......... 75 4...3.............. 93 5.......... 77 4........................1 Dividing Node...........................2............... 87 5.................................................................5 Summaries & Conclusion ................................................................................Table of Contents III 4................2 Results & Discussion .................................... 86 5.......... 84 5.................................2 Cutting Time Increment Searching Procedure ....1.............................................3.................2 Wear Rate Calculation....................1 Modelling .............1 Normal Pressure.........................2 On Rake Face......................................1 Step 1: Initial Tool Wear Profile ......................4................................. 92 5....................................1...4............ 81 Chapter 5 Estimation Of Tool Wear In Turning Operation........................ 82 5......................................................................... 73 4.....3 On Flank Face ................ 73 4....................3......................... 70 4.......................4 Cutting Time Increment Calculation .........3........1 Flank Wear Calculation Subroutine............3 Application Of Preheated Cutting Tool ........................................................................................................................................................................ 87 5...... 88 5......... 86 5......2 On Tool........................................................................................ 92 5...........1 On Workpiece....6................................................................................................ 72 4...............................................1 Chip Formation And Heat Transfer Analysis...........6.................. 84 5.... 90 5......5 Nodal Displacement ...........................3...................2 Sliding Velocity .......................... 91 5.........3.............2 Results & Discussion .................... 97 ...................................4.......................................................5 Summaries & Conclusion .......................1.........1 Introduction ...3................................. 89 5...............3.......................................2 Step 2: Adjustment...................................4 Results & Discussion.....................3...........2....................................3........................4............................................................................4...........................................................3....................................................1 Modelling .......3.......3 Modelling Procedure...............................3....................... 82 5............................................................2 Tool Wear Calculation Program Design.......................4.......3......................................4 In Milling Operation.......................3.....................4....................................... 82 5...........3 Nodal Move Direction................... 83 5....4........3..............3..................... 94 5.................3 Tool Temperature.............................................3.................1...............1.... 90 5.......6 Tool Geometry Updating ........................3...

............3.................................3 Modelling Procedure........................................ 111 6....................4.................. 98 6.....5 Cutting Time Increment Calculation ...............2 On Rake Face...3 Nodal Average Wear Rate Calculation ................ 108 6.......................................1 Dividing Node............5....................1 Chip Formation Analysis........................1 Flank Wear Calculation Subroutine...........................3..... 114 7....................1 Discussion About The Calculation Method Of Nodal Average Wear Rate .............3 On Flank Face ................................................................................................................................3...............3................. 99 6.....................................4 Results & Discussion...............................................................3................................................................................................................ 103 6...............2 Classification Of Workpiece Node ...............................................................................................3................................................................1 Summaries ...3.. 107 6................................................ 111 6.......................... 109 6...... 103 6......................................3......................................... 105 6......3.................................................3........................................................................... 113 Chapter 7 Summary And Prospect..4...................2 Prospect ................................4 Nodal Move Direction............................................................................5 Summaries & Conclusion ...........6 Tool Geometry Updating ..............................................5.................... 107 6................................ 118 .......... 116 References .................2 Heat Transfer Analysis ... 101 6..2 Tool Wear Calculation Program Design.........................................1 Introduction ..... 103 6.........................................................................3......................................Table of Contents IV Chapter 6 Estimation Of Tool Wear In Milling Operation...................... 108 6.......................3...3................... 108 6.......................................2 Cutting Time Increment Searching Procedure .... 114 7.....................................3.......................................... 108 6.............................4...................... 98 6..............................................3......................................... 101 6.

FEM HSC KT VB VC VN Artificial Intelligence Arbitrary Lagrangian Eulerian Cubic Boron Nitride Finite Difference Method Finite Element Method High Speed Cutting Depth of crater wear Width of flank wear (mean) Maximum wear of nose radius Notch wear .Nomenclature I Abbreviation AI ALE CBN FDM FE.

vibration and cutting temperature cause surface integrity deteriorated and dimension error greater than tolerance. cutting speed. such as Usui’s tool wear equation. Some researchers concentrate on the study of wear mechanism and investigate the mathematical relationship between wear due to various wear mechanisms and some cutting process variables such as relative sliding velocity of workpiece material along tool face. e. Because of its great economic and technical importance. when tool geometry is changed. Then the cutting tool must be replaced or ground and the cutting process is interrupted. The cost and time for tool replacement and adjusting machine tool increase cost and decrease productivity. Hence tool wear relates to the economic of machining and prediction of tool wear is of great significance for the optimization of cutting process. For the researcher and tool manufacturer tool wear progress and tool wear profile are also concerned. When tool wear reaches a certain value.Introduction 1 Chapter 1 Introduction Machining operations comprise a substantial portion of the world’s manufacturing infrastructure. The life of the cutting tool comes to an end. increasing productivity and lowering cost. a large quantity of research has been carried out in order to optimize cutting process in terms of improving quality. tool life and machining cost. They create about 15% of the value of all mechanical components manufactured worldwide [Merc-98]. new equation must be established by making experiment. machining quality. Some tool wear equation related to one or several wear mechanisms are developed. cutting temperature of tool face and normal pressure on tool face. and is very easy to use. such tool life equations are valid under very limited cutting conditions.g. For example. the prediction of tool wear is performed by calculating tool life according to experiment and empirical tool life equations such as Taylor’s equation or its extension versions. But capability of predicting the contributions of various wear mechanism is very helpful for the design of cutting tool material and geometry. increasing cutting force. Tool life equation gives no information about the wear mechanism. At present. it gives only the information about tool life. . Although Taylor’s equation gives the simple relationship between tool life and a certain cutting parameters. In addition. Tool wear influences cutting power.

stress. Various variables in the cutting process such as cutting force. Therefore a new tool wear prediction method may be developed by integrating FEM simulation of cutting process with tool wear model. cutting temperature. numerical methods such as finite element method (FEM).Introduction 2 In the recent decades. Among them. with the emergency of more and more powerful computer and the development of numerical technique. FEM has become a powerful tool in the simulation of cutting process. strain. finite difference method (FDM) and artificial Intelligence (AI) are widely used in machining industry. including those very difficult to detect by experimental method. etc can be predicted by performing chip formation and heat transfer analysis in metal cutting. . strain rate.

According to a comprehensive survey conducted by the CIRP Working Group on Modelling of Machining Operations during 1996-1997 [Lutt-98]. strain rate and temperature. FLUENTTM . built-up-edge. the local stress. Compared to empirical and analytical methods. and Tay et al [Tay-74] since the early 1970s. burr. Okushima [Okus-71]. chatter.Introduction 3 1. Non-linear geometric boundaries such as the free surface of the chip can be represented and used. temperature distributions. ABAQUS/StandardTM [Shi-02]. segmental chip formation is modelled to simulate high speed cutting [Bäke-00] [Bäke02] [West-01]. DEFORM 2DTM [Özel00b] [Cere-99]. application of finite element in metal cutting develops rapidly because of its advantages and the development of powerful computer [Atha-98][Sand-98]. FORGE 2DTM [Ng-99] [Mona-99]. In addition to the global variables such as cutting force. 32% in analytical modelling and 18% in numerical modelling in which finite element modelling techniques are used as the dominant tool. cutting temperature. MARCTM [Behr-98a]. more commercial FE codes are used in chip formation simulation. In recent years. 43% were active in empirical modelling. including: NIKE2TM [Stre-85]. Great progress has been made in this research field: Lagrangian approach is used to simulate the cutting process including incipient chip formation state [Shet-00]. model limitations and computational difficulty have been overcome to some extent. Finite element method has been used to simulate machining by Klamecki [Klam-73]. ABAQUS/ExplicitTM [Baca-00] and LS DYNATM [McCl-02]. tool wear. among the 55 major research groups active in modelling. With the development of faster processor with larger memory. ALGORTM. It is the basic of the research on physical phenomena-cutting force. feed force and chip geometry. hard-turning [Guo-02] [Usui-84] or large negative rake angle [Ohbu- .1 State Of Art: Finite Element Simulation Of Cutting Process Chip formation is the essential phenomenon in the cutting process. The interaction between chip and tool can be modelled as sticking and sliding. chip curling and chip breakage. etc can also be obtained. finite element methods used in the analysis of chip formation has advantages in several aspects [Zhan-94]: • • • • Material properties can be handled as functions of strain. In addition.

The effect of tool geometry on the chip formation process is studied. cast iron. composite [Arol-02]. mainly including varing rake angle [Shih-96] and tool geometry. Fig. the modelled cutting tool materials include uncoated carbide [Lin-01b]. ceramic cutting tool and diamond [Ohbu-03]. The modelled workpiece materials include carbon steel [Behr98b] [Gu-02]. ductile iron [Chuz-03a] [Chuz-03b]. microscopic cutting of single abrasive grain in grinding [Ohbu-03]. chamfered [Shat-01b] [Mova-02] and round edge [Ozel-02] [Kim-99]. CBN [Özel-02]. drilling. etc. cermet. The mainly simulated cutting types include tuning [Behr-99]. etc. . chip breaker [Dill-00]. coated carbide [Mona-99]. 3D simulation is performed to analyse oblique cutting [Leop-98] [Klam-73] [Lin00] [Cere-00] [Guo-02].Introduction 4 03]. Orthogonal cutting is the most frequently simulated cutting type [Stre-93]. milling [Özel-00a]. The studied tool geometries include sharp. 1. For example. and worn cutting tool [Li-02] [Shih-93].1 Modelling research trends [Ng-02a] A diversity of cutting tool and workpiece materials is used in the simulation of cutting process. high alloy steel [Ng-02a].

1. tool performance [Ahma-89]. 1. the influences of sequential cutting [Liu-00] and microstructure of workpiece material [Chuz-03a] [Chuz. etc are investigated as well. But experimental work is often necessary in order to determine the chip geometry and shear angle. varying cutting thickness in milling operation or serrated chip in high-speed-cutting because it is unable to simulate free surface conditions.03b] on chip formation are studied. Except the normally discussed variables cutting force. not including the transition from initial to steady state cutting process. residual stress [Yang-02] [Shih-93].1. But in order to extend the cutting time until steady state.1 Numerical Aspects The implementation of cutting process simulation is based on numerical theory and technique. Cutting process analysis with Eulerian approach requires less calculation time because the workpiece model consists of fewer elements.1 Approach Several approaches are supplied for numerical modelling: Lagrangian. the mesh follows the material. a long . Their development is helpful to improve the capability of the simulation. chip breakage [Maru-02]. Because the deformation of the free surface of the chip can be automatically treated by elastic-plastic material deformation.1.Introduction 5 In addition. Eulerian approach is suitable to analyse the steady state of cutting process. which is an unavoidable part of geometry modelling.1. cutting temperature and stress. chip flow angle [Stre-02]. burr formation [Guo-00]. tool wear [Söhn-01b] [Yen-02]. Eulerian Approach In Eulerian approach. Lagrangian Approach In Lagrangian approach. the mesh is fixed spatially and the material flows through the mesh. Eulerian and Arbitrary Lagrangian Eulerian (ALE). Lagrangian approach can be used to simulate from initial to steady state of cutting process. That is the reason why before 1995 the applications of Eulerian approach in chip formation analysis overrun those of Lagrangian approach.

Introduction

6

workpiece is needed in geometry modelling, which increases the calculation time. In order to perform chip separation, chip separation criteria and realization method are necessary. Arbitrary Lagrangian Eulerian Approach (ALE) ALE approach combines the features of pure Lagrangian and Eulerian approach, in which the mesh is allowed to move independently of the material. It is an effective tool for improving mesh quality in the analysis of large deformation problem. Many commercial FE codes introduce ALE approach by adjusting mesh based on different mesh adaptivity. The adaptive meshing technique in ABAQUS/Explicit belongs to ALE approach. It can be used to analyse not only Lagrangian problem but also Eulerian problem. By giving suitable mesh control parameters, the whole process from initial to steady state can be simulated without the need of chip separation criterion or any chip geometry data from experiment. Furthermore, it is not necessary to extend the size of workpiece model. Hence the calculation time is not increased. 1.1.1.2 Mesh Adaptivity Three types of mesh adaptivity are designed to create a new spatial discretisation and improve mesh quality: h-adaptivity, p-adaptivity and r-adaptivity [Kalh-01]. • • • H-adaptivity changes the size of the mesh. The new mesh has different number of elements and the connectivity of the nodes is changed. In p-adaptivity the degree of the interpolating polynomial is changed. R-adaptivity is based on relocation of the nodes, without altering the topology (elements and connectivity) of the mesh. For example, adaptive meshing technique in ABAQUS/Explicit is accomplished by using R-adaptivity. During meshing nodes are moved to more favourable positions to improve mesh distortion. In addition, solution-dependent meshing is supplied to concentrate mesh towards the developing boundary concave, e.g., chip separation area in the vicinity of the cutting edge, and produce local mesh refinement in this area. But it is found that only the application of r-adaptivity is not sufficient to maintain the mesh quality. Therefore some FE codes, e.g. Deform-2D and AdvantEdge employ

Introduction

7

the combination of r- and h-adaptivity. Mesh is refined where great difference in the gradients of a certain solution is detected between elements. For example, Marusich et al propose to refine mesh according to plastic work rate in each element [Maru-95]; Owen et al use an error estimator based on the rate of fracture indicator to produce a fine mesh in high plastic deformation area and the regions where material failure is going to take place [Owen-99]. Chip separation is produced during meshing and mesh refining. In addition, the contact at tool-chip interface can be improved as well.

Introduction 1.1.2 Mechanical Aspects

8

The development of metal cutting theory helps people get more and more correct understanding in mechanical aspects of cutting process including contact and friction, material property, chip separation, etc. The modelling of these aspects influences the accuracy of cutting process simulation. 1.1.2.1 Contact And Friction Friction behaviour on the tool face determines the cutting power, machining quality and tool wear. It plays an important role in metal cutting. Development Of Friction Model In Metal Cutting The nature of friction between two dry sliding surfaces was described by Amontoms in 1699 [Amon-99]. He put forward that the coefficient of friction µ is independent of apparent area of contact A and applied normal load Fn . In 1785, Coulomb [Coul-85] approved and developed these laws by proposing that the coefficient of friction is substantially independent of the sliding velocity. Accordingly a constant coefficient of friction is expected on the tool face in metal cutting process. Ff Fn

µ=

= const

(1.1)

where F f is the friction force. However in metal cutting process, it is generally observed that the mean coefficient of friction on the tool face varies considerably with the change in cutting speed, rake angle and so on. This results from the extreme conditions of metal cutting area where the normal pressure at tool-chip interface is very high. According to Eq. 1.2 proposed by Finne and Shaw [Finn-56], the ratio of the real area of contact Ar to the apparent area of contact A approaches or reaches 1 under cutting conditions, which is different from the application conditions of Coulomb’s assumption.

g. Ar is less than unity. e. Some advanced testing technologies.2 [Zore-63]. adjacent to the cutting edge.2 Stress distribution on tool-chip interface Plenty of evidence from worn tools. believed constant. 1. and shear stress is believed equal to shear strength of the workpiece material. In sliding region. Based on the assumption of shearing action within the workpiece material. The chip-tool interface is divided into sticking and sliding regions. from quick-stop sections and from chips showed the coexistence of sticking and sliding at tool/chip interface under many cutting conditions [Tren-77].2) where N is normal force. Zorev proposed the distribution of shear and normal stress on the rake face as shown in Fig. and the coefficient of friction is A Fig. are used in experiments to discover the form of stress distribution on the rake face. . Ar approaches unity under very high A normal stress. In sticking region. photoelastic measurements [Rice-60] or split tool dynamometers [Kato-72] [Chil-98].Introduction Ar = 1 − e − BN A 9 (1.1. But these techniques are limited when the stresses very close to the cutting edge are determined.

µ= Ft + FC tan α FC − Ft tan α (1. by using Regression analysis [Ng-02b].5. rake angle α .Introduction 10 Applied Friction Models In Cutting Process Simulation In the finite element analyses of metal cutting.5) τ s is the shear flow stress of the chip material. the division of the two regions is determined by two methods: one is to prescribe the length of each region [Shih-95] [Wu-96] [Shat-00]. various approaches are used in the modelling of friction. When Zorev’s sliding-sticking friction model is employed in the simulation. .1. cutting speed vc .4. 1. and rake angle α . 1. thrust force Ft . and feed f .034 − 0.3) Ng and his co-operators performed orthogonal cutting tests under different cutting conditions to establish a linear relation between the coefficient of friction µ . given by Eq. σ is normal stress. µ = 1.4) Liu et al [Liu-00] determined the coefficient of friction by performing simulation using different values and carrying out the sensitivity study on the coefficient of friction. given by Eq. τ = min( µσ . Normally the coefficient of friction µ is calculated by using Eq.00446α − 3.τ s ) where (1. the other is to determine the sliding and sticking region automatically by program according to a criterion [Zhan94] [Guo-00]. τ is friction stress.3 according to the cutting force Fc . Constant coefficient of friction based on Coulomb’s friction law is used in most cases.0002vc (1.888 f − 0.

A frictional shear factor is introduced into the relationship in order to make the calculated results agree with those of experiment. as shown in Fig. n=4 n =0 τ = ∑ anσ n (1. 1.7) a0 . are affected by the strain rate and temperature during material forming process with plastic deformation. Influence Factors Of Material Property Experiments shows that material properties. the stress is higher at higher strain rate due to the viscous effect during plastic deformation and lower at higher temperature due to material softening. p is contact pressure in MPa. given by Eq.07  V   τ = (1. a2 .2.07 µp   HV   Mpa  tanh   H  0.Introduction 11 Iwata et al [Iwat-84] proposed the relationship given by Eq. strain and strain rate are very high. 1.2 Material Constitutive Model The accuracy of the finite element analysis is severely dependent on the accuracy of the material mechanical properties. In metal cutting process. This overstress effect by strain rate is more pronounced as the temperature increases [Shih-91]. e. a3 and a4 are determined by fitting experimental stress curve on rake face. stress-strain relationship. 1.6 after put forward a method to test friction between newly created surfaces and tool material. a1 . Thermal- .6) where HV is the Vickers hardness of the workpiece material. 1.g.7. For the same value of strain. temperature.1.3.  0. Yang and Liu [Yang-02] proposed a stress-based polynomial model of friction.

3 Material property curve Many researchers are making efforts to establish such material constitutive models for different workpiece materials through experimental [Kopp-01]. • For CK45 The material constitutive model developed by O. 1. 1.8) with T0 = ∆G 0 & ε0 kIn ( pl ) & ε (1. analytical or simulation methods [Shat-01a] [Özel-00b] [Batz-02]. Vöringer is used. m  T & σ (T .Introduction 12 viscoplastic material constitutive model is necessary for the finite element analysis of metal cutting.9. a material model database has been developed by Söhner et al [Söhn-01a]. ε ) = σ 1 −    T0   ∗ v ∗ 0     n     (1. 1. Based on their supports. which is described by Eq.9) . Material Constitutive Model For Mild Carbon Steel The main workpiece materials used in the following research are mild carbon steel CK45 and AISI1045. Stress [N/mm2] Stress [N/mm2] Temperature [K] (a) overstress effect Strain [%] (b) material softening Strain [%] Fig.8 and Eq.

097.168.00005(T −700 )      (1. 1. e. a user material subroutine based on this material constitutive model is employed. a distance value.Introduction 13 & where the constants for CK45 are: m=1. k is Boltzmann constant and T is temperature in Kelvin [Schu- 00].2.29 × 10 5 s −1 . • For AISI1045 To describe the material property of AISI1045.78. By chip separation. n=0. C=0. Chip Separation Criterion The chip separation criteria used by researchers can be categorized as two types: geometrical and physical. σ is the effective stress in MPa. Normally it includes two aspects of consideration: chip separation criterion and model realization. ε 0 = 7.1. Tmelt=1480°C [Kopp-01]. ∆G0 = 0. The unwanted part forms the chip. In the simulation of cutting process.53. the workpiece material is separated into two parts. a new workpiece surface is formed on the created part. When the distance between the nearest workpiece node on the moving path of the cutting edge . and T is temperature in °C.g. with the cutting tool advancing into the workpiece. a=0. the Johnson-cook constitutive equation is used. The realization of chip separation is one of the main problems in the simulation of chip formation process. σ = (Bε n )1 + C ln    &  ε    Tmelt − T    1000    Tmelt − Troom    2    + ae −0.3 Chip Separation In the cutting process. ∗ and σ 0 = 1352MPa . n=0. Geometrical criteria define geometric parameters.275.58ev .1.10) where B=996.

Model Realization There are several methods to model chip separation in finite element mesh.Introduction 14 and the cutting edge is equal to or smaller than this given distance value. material failure takes place. although sometimes it is defined at random.and strain ratedependent strain at failure will provide a better simulation result. When such physical parameter reaches a critical value. for example. The most reliable critical value is obtained by performing experiments. the magnitude designated for these criteria did have a major effect on mesh distortion together with the value of maximum shear stress. material failure happens and the element carries no stress any more as if they do not exist. Fig. and the effective stress in the machined surface [Ng-02a]. distribution of shear stress.4 Element removal [Behr-98b] . strain energy density [Lin-99] or effective plastic strain [Shir-93]. effective stress or effective plastic strain in the chip and in the machined surface. 1. normally physical criterion. They are based on physical parameters such as stress [Iwat-84]. According to the investigation on both types of criteria made by Huang and Black [Huan-96]. However. is reached. They are related with the applied software. neither had a substantial effect on chip geometry. Physical criteria is related to some physical meaning of chip separation. chip separation takes place [Shih-95]. • Element removal [Cere-96] When chip separation criterion. temperature. A critical value considering multi-influencing factors. Such element can be removed and does not display.

6 Node splitting [Behr-98b] • Mesh adaptivity [Arra-02] Chip separation is performed by mesh refinement in the separation zone by increasing the number of elements or relocation of the nodes. Fig. the two elements move in different direction and lose contact. for example. . They are perfectly bonded together through some pair of nodes along the prospective parting line. two nodes overlap together and connect to two different elements.5 Node debond • Node splitting [Shih-95] Chip separation is realized by element separation in front of cutting edge. physical or their combination. Fig. When chip separation criterion is reached. When the separation criterion is met.Introduction • Node debond [Shi-02] [Shet-00] [Shet-03] 15 The chip and the workpiece are two separated parts. debond of the node pair takes place and the two nodes move in different direction. The two neighbouring elements have common node before separation. a node is very close to the cutting edge. The chip separation can be geometrical. Through the further movement of the cutting tool. Element separation takes place and a new node is created at the same position. 1. 1.

interface and machine tool: Tool wear economy. process security material structure texture material properties interface friction cooling lubricant cutting param. hardness. Any element changing contact conditions in cutting area affects tool wear. or finishing). 1. coolants are used to decrease cutting temperatures and likely . microstructure.2 Technical Background About Tool Wear 16 Prediction of tool wear is complex because of the complexity of machining system. feed rate. tool coatings and tool geometric design (edge preparation. which determine cutting force and energy for the applied cutting conditions. Tool wear in cutting process is produced by the contact and relative sliding between the cutting tool and the workpiece and between the cutting tool and the chip under the extreme conditions of cutting area. contact tool cutting material coating geometry machine dynamics design Fig. etc) need to be appropriately chosen for different operations (roughing. workpiece quality. The optimal performance of a cutting tool requires a right combination of the above tool parameters and cutting conditions (cutting speed. temperature at the cutting edge can exceed 1800°F and pressure is greater than 2.7 Influencing elements of tool wear [Söhn-01b] • Workpiece: It includes the workpiece material and its physical properties (mechanical and thermal properties. semi-roughing. tool. In 80% of the industrial cutting applications.Introduction 1. etc) • Interface: It involves the interface conditions. rake angle. These elements come from the whole machining system comprising workpiece. depth of cut. • Tool: Tool material.000psi [John-01]. etc).

thermal crack. flank wear. 1. Instable cutting processes with large vibrations (chatters) result in a fluctuating overload on the cutting tool and often lead to the premature failure of the cutting edge by tool chipping and excessive tool wear. 1. At high cutting speed. • Dynamic: The dynamic characteristic of the machine tool. fatigue crack.1 Wear Types In Metal Cutting Under high temperature. brittle crack. Fig. Increasingly new technologies.8 Wear types [Lim-01] • Crater wear: In continuous cutting. have been developed to reduce the cost of coolant that makes up to 16% of the total machining costs [Walt-98].Introduction 17 reduce tool wear. such as the minimum liquid lubrication. affected by the machine tool structure and all the components taking part in the cutting process.8 are the most common wear types. cutting tool has normally complex wear appearance. crater wear normally forms on rake face.2. plays an important role for a successful cutting. e. It conforms to the shape of the chip underside and reaches the maximum depth at a distance away from the cutting edge where highest temperature occurs. The dominating basic wear types vary with the change of cutting conditions. turning operation. insert breakage. plastic deformation and build-up edge. which consists of some basic wear types such as crater wear. Crater wear is improved by . crater wear is often the factor that determines the life of the cutting tool: the tool edge is weakened by the severe cratering and eventually fractures. Crater wear and flank wear shown in Fig. 1. high pressure.g. high sliding velocity and mechanical or thermal shock in cutting area.

2. • Oxidation wear: A slight oxidation of tool face is helpful to reduce the tool wear. diffusion wear. CoO. a decrease in the dimension accuracy of the tool and an increase in cutting force. It reduces adhesion. It is found that tool wear is not formed by a unique tool wear mechanism but a combination of several tool wear mechanisms. abrasive wear. e.e.Introduction 18 selecting suitable cutting parameters and using coated tool or ultra-hard material tool. This results in a rapid tool material loss. diffusion wear and oxidation wear are very important. solution wear.2 Wear Mechanism In order to find out suitable way to slow down the wear process. Among them. TiO2. • • Adhesive wear: Adhesive wear is caused by the formation and fracture of welded asperity junctions between the cutting tool and the workpiece. etc are formed rapidly. WO3. It is responsible for a poor surface finish. • Flank wear: Flank wear is caused by the friction between the newly machined workpiece surface and the tool flank face. Diffusion wear: Diffusion wear takes place when atoms move from the tool material to the workpiece material because of the concentration difference. oxidation wear. Tool wear mechanisms in metal cutting include abrasive wear. then taken away by the chip and the workpiece. electrochemical wear. Co3O4. oxidation wear.. etc. • Abrasive wear: Tool material is removed away by the mechanical action of hard particles in the contact interface passing over the tool face. But at high temperature soft oxide layers. These hard particles may be hard constituents in the work material. fragments of the hard tool material removed in some way or highly strain-hardened fragments of an unstable built-up edge [Boot-89]. Hence the width of the flank wear land VB is usually taken as a measure of the amount of wear and a threshold value of the width is defined as tool reshape criterion. adhesive wear. i. adhesive wear. 1. many research works are carried out to analyze the wear mechanism in metal cutting. . temperature and vibration. delamination wear.g. diffusion and current by isolating the tool and the workpiece. The rate of diffusion increases exponentially with the increase of temperature.

1. For example. but flank wear mainly dominated by abrasive wear due to hard second phase in the workpiece material. see Table 1.Introduction 19 Under different cutting conditions dominating wear mechanisms are different.9. e. as shown in Fig. Taylor’s tool life equation and its extension versions under different cutting conditions appear most frequently. In various sizes of cutting database. etc) Oxidizing vc Workpiece Fig.1. The constants n. 1. it is assumed that crater wear is mainly caused by abrasive wear. reveals the exponential relationship between tool life and cutting speed. the existing tool life equations need . A and B are defined by doing a lot of experiments with cutting speed changing and fitting the experimental data with the equation. the dominating wear mechanisms vary with cutting temperature. They can be categorized into two types: tool life models and tool wear rate models. CT. Taylor’s tool life equation [Tayl-07]. and Hastings tool life equation describes the great effect of cutting temperature on tool life [Hast-79].9 Wear mechanism [Köni-84] 1. According to the temperature distribution on the tool face. As the new machining technologies. Chip Oxidizing wear Tool Wear Diffusion wear Adhesive wear Abrasive wear Diffusion Abrasion Adhesion Cutting temperature (cutting speed. For a certain combination of cutting tool and workpiece. diffusion wear and oxidation wear. • Tool life models: This type of wear models gives the relationship between tool life and cutting parameters or variables. high-speed-cutting or dry cutting. are getting spread in manufacturing industry.3 Tool Wear Model Many mathematical models are developed to describe tool wear in quantity. feed.g. Tool life equations are suitable to very limited range of cutting conditions.2. It is very convenient to predict tool life by using this equation.

c f. λ (A. T = normal stress and temperature θ . Usui’s equation includes three variables: sliding velocity between the chip and the cutting tool. which are obtained from literatures. Usui’s model is derived from Shaw’s equation of adhesive wear [Usui-78c]. Table 1. the wear growth rate.σt. tool wear profile or tool wear mechanisms that are sometimes important for tool designers.T n = CT v c Ln T (n. One part shows that abrasive wear is influenced by the cutting speed and feed. λ . are related to several cutting process variables that have to be decided by experiment or using some methods [Kwon-00].(A. i. Another part including universal gas constant and tool temperature describes diffusive wear.Introduction 20 to be updated with new constants and a lot of experimental work has to be done.T = A “Differential” Tool Wear Rate Models Takeyama & Murata’s model. Takeyama & Murata’s model is developed by considering the combination action of abrasive wear and diffusive wear. D = constants) Usui’s model. In these modes. Therefore the equation sums two parts up.e. except that tool life can be predicted by these equations.dW/dt = rate of volume loss per unit contact area per unit time (mm/min) . 1979): B L = A θTB ⋅. B = wear characteristic constants A. the right column shows two tool wear rate models. sσexpS −λ θ ) dt = = t t V ( exp(-B/T) . • Tool wear rate models: These models are derived from one or several wear mechanisms.C. which was derived from equation of adhesive wear [Usui et al. 1978]: dWdW/dtCσAv.1 Tool wear models Empirical Tool Life Models Taylor’s tool life equation: Vc ⋅. CT = constants) Hastings tool life equation (Hastings et al. the rate of volume loss at the tool face (rake or flank) per unit contact area per unit time (mm/min). exp(-E/RT) dW dt == G(V. considering abrasive wear and diffusive wear (1963): dW/dt G (v c f) D ⋅ . They provide the information about wear growth rate due to some wear mechanisms. ) ++ Dexp (− E Rθ ) (G. B = constants) C.1. Except the constants A and B. tool temperature and normal . In addition. B = constants) vcc = Cutting speed V L T = Tool life E = Process activation energy T θ = Cutting temperature f = Feed V vsS = Sliding velocity R = Universal gas constant In Table 1. it is difficult to get further information about the tool wear progress..

Therefore Usui’s equation is very practical for the implementation of tool wear estimation by using FEM or by using the combination of FDM and analytical method. All the points for flank wear and crater wear defined by experiment distribute along two characteristic lines with different gradients.8 ×10 −9 θ f ≥ 1150 K 2. crater wear on rake face was assumed mainly caused by adhesive wear. When tungsten carbide tools are used to machine carbon steels. The constants in tool wear rate models are depending on the combination of workpiece and cutting tool material. These variables can be predicted by FEM simulation of cutting process or combining analytical method and FDM. Table 1.195 × 10 4 θ f < 1150 K 5.2 : Characteristic constants for carbon steels [Kita-89] C [m2/MN] θ f ≥ 1150 K λ [K] θ f < 1150 K 7. They are introduced in the later tool wear estimation models. which intersect at the critical temperature of around 1.302 ×10 3 1. which mainly results from abrasive wear [Kita-88]. Usui determined the constants for such cutting conditions and validated this model by the prediction of crater wear.150K. According to cutting experiment.Introduction 21 pressure on tool face. The latter study showed that this equation is able to describe flank wear as well. The experimental points for crater wear usually lie on the line in the higher temperature range.198 ×10 −2 .2 shows the charateristic constants in Usui’s equation for the combination of carbon steel and carbide tool that obtained from literature [Kita-89]. Table 1. whereas those for flank wear are usually distributed around the line in the lower temperature range.

Introduction
1.3 Research Of Tool Wear With Finite Element Methods 1.3.1 Comparison Between FEM Method And Empirical Method

22

Based on tool wear rate models, the estimation of tool wear profile progress with the cutting process can be implemented by predicting cutting process variables using finite element method. Its advantages and disadvantages are shown by the comparison with the empirical method in Table 1.3.

Table 1.3 Comparison of FEM method and empirical method Compared aspects Environment requirement Empirical method Special workpiece, cutting tests The procedure machine, personnel FEM method tool, Powerful code computer, tool

for wear rate model and FEM

of Cutting tests and regressive Obtaining tool wear rate analysis model by experiment or from literature, running the program with tool wear rate models under new cutting conditions

calculating tool wear

Application new conditions

under New experiments have to If only tool wear rate model cutting be carried out to update the is updated according to new constants of tool life models cutting conditions, the program can be used again

Time

The development of new The time for developing the tool life models is time entire program is relative consuming; long. Whereas the prediction of The time for calculating the tool wear with the tool life tool wear with the program model is very efficient depends on the performance of computer

Wear mechanism

Wear mechanism is not Yes, even the contributions

Introduction
considered

23
of the main mechanisms can be calculated

Workpiece material

Uneven material distribution Homogeneous material result from impurity, heat model, thermal visco-plastic treatment, work hardening material material

Tool material

Uneven material properties Homogeneous treatment, etc

result from impurity, heat model, ideal elastic material

Medium

Sensitive method,

to

the

cooling The types of heat emission type, through workpiece and considered tool face and under FEM surface their

coolant

cooling effect, etc

various cooling conditions implementation have to be

Vibration of machine- The constants are sensitive Not considered at present tool-workpiece system Predicted parameters to the vibration of the system wear Very limited information can Comprehensive information be obtained, for example about tool wear including only tool life is predicted crater wear profile, flank with Taylor’s tool life wear profile, VB, KT, VC (for 3D), VN (for 3D), etc can be predicted Cutting type Tool life models under At present, only tool wear milling studied. For different cutting types, the tool wear program may need adjusting according to the characteristic of relative operations are equation

various cutting type can be prediction in turning and developed

Introduction

24
motion of cutting tool and workpiece

Requirement on the No special requirement user

At present, except basic knowledge about metal cutting theory, user needs the basic knowledge about FEM chip formation, heat transfer analysis

Application present Quality prediction of

at Used in the real production

For research and education

the Quantitative

Qualitative

1.3.2 State Of Art: Numerical Implementation Of Tool Wear Estimation Tool wear estimation with Finite Element Method is developed from tool wear estimation with the combination of analytical method and Finite Difference Method (FDM). 1.3.2.1 Tool Wear Estimation With The Combination Of Analytical Method And FDM Usui’s Research-Prediction Of Crater Wear The earliest reported research work on tool wear estimation with the combination of analytical energy method and FDM was performed by E. Usui et al in 1978. He first derived a characteristic equation of crater wear theoretically by combining M.C.Shaw’s adhesive wear, temperature-dependent material asperity hardness and temperature-dependent Holm’s probability, given by

dw = Cσ t v s exp(− λ θ ) dt
Then he verified the equation experimentally.

(1.11)

The temperature distribution within the chip and the tool at steady state is obtained with FDM by considering the heat source on the shear plane and on rake face. such as round cutting edge. and the predicted distribution of the stress and the temperature. orthogonal cutting data about shear stress on shear plane. By using the predicted cutting force and tool-chip contact length together with an assumption of an exponential normal stress distribution and a triangle or trapezoidal frictional stress distribution on the tool face. the prediction of crater wear cannot be carried out without making experiment in advance. The effect of cutting edge preparation.) Limitations • When using the energy method to predict the chip formation and cutting force. sliding velocity of the chip and cutting force are predicted through energy method proposed in previous papers [Usui-78a] [Usui-78b] [Usui-78c]. Then computer calculation of crater wear is carried out by using the characteristic equation.Introduction a. or rounded cutting edge due to wear on the tool wear cannot be considered. The characteristic constants of the equation for the combination of carbon steel and P20 are determined with the aid of the predicted temperature. • The energy method is developed based on single shear plane for the cutting tool with sharp cutting edge. the frictional stress is calculated. friction angle and shear angle are needed. stress on tool face and the measured wear by curve fitting.) Implementation Procedure 25 The chip formation. b. c.) Result The predicted crater wear was reported in good agreement with the measured in experiment in depth and contour except some discrepancy in the location of the deepest portion. .

On the flank wear. which intersect at the critical temperature of around 1.150K. Eq. arbitrary set normal stress and sliding velocity. Kitagawa finds that flank wear can be described by the same characteristic equation.) Implementation Procedure In the prediction. the frictional stress is calculated. . the sliding velocity of workpiece material on the flank wear land is assumed equal to the cutting speed.Introduction Kitagawa’s Research-Prediction Of Flank Wear 26 By analysing the flank wear characteristics of tungsten carbide tools in turning plain carbon steels at steady-state cutting without a built-up edge experimentally. feed and workpiece material. and frictional stress on the other sites is arbitrary set. for crater wear. whereas those for flank wear are usually distributed around the line in the lower temperature range. Tool wear consists of two characteristic lines with different gradient. By prescribing a triangle distribution of frictional stress along the tool-chip contact length with maximum value at the cutting edge and neglecting the contribution of stress on flank face to the cutting force and thrust force.) Result It was reported that the predicted tool life. rake face and in the shear plane using FDM. The wear rate on the flank wear is calculated according to the predicted temperature. 1. The values of cutting force.11. The experimental points for crater wear usually lie on the line in the higher temperature range. the frictional stress at the cutting edge is set equal to the maximum value on rake face. a. temperature and mean stresses on the flank wear land are in reasonable agreement with experiment even with changing cutting speed. Normal stress on flank wear is set equal to frictional stress. Then the temperature on flank wear land is predicted by considering the heat generated on the flank wear. Normal stress on flank wear is adjusted continuously until a uniformly distributed wear rate is achieved everywhere on the flank wear land. thrust force and chip contact length obtained from orthogonal experiment must be given beforehand. b.

Introduction c. Hence the earliest reported research of tool wear estimation in quantity with FEM was done by Y. Söhner et al since 2001. J.) Limitations 27 • • • • The prediction method is developed under the assumption of no crater wear formed on the rake face. According to the paper in 2002 [Yen-02] and the dissertation of Söhner [Söhn-03]. This limits its application to low cutting speed range.2 Tool Wear Estimation With FEM Yen And Söhner’s Research (FEM) Although in a paper in 1999.3. thrust force and chip contact length obtained from orthogonal experiment. The assumption of uniform wear rate on flank wear excludes the formation of rounded edge due to wear that is often observed in experiments. a. the numerical implementation of the integration of tool wear rate models with FEM calculations to predict the evolution of the tool wear was performed by using commercial FE code DEFORM-2D. MacGinley claimed that they performed tool wear analysis based on a wear function related to normal stress and sliding velocity by predicting stress distribution within coated and uncoated carbide tool with and without chip breaker using commercial FEM code-FORGE2 [Mona-99]. C. It is suspected that only the tendency and possibility of tool wear distribution were analysed qualitatively. 1. Whether the predicted crater wear is sensitive to the frequency of measuring these values during the development of flank wear is very important for its application perspective. no implementation procedure. clear predicted tool wear profile and wear value are described or provided.) Implementation Procedure Usui’s wear model is used to calculate the wear rate of the uncoated carbide tool in cutting carbon steel. Monaghan and T. These values vary with the development of flank wear. Yen and J.2. . The prediction method is not applicable to the cutting tool with any edge preparation because of the assumption of stress distribution on rake face The prediction is based on cutting force.

In the second phase. Then new tool geometry accounting for tool wear is calculated based on the user input for a cutting time increment. normal stress and sliding velocity under steady-state cutting condition provided by the first two phases. which can prolong the cutting simulation to a sufficient long cutting time. ‘Konti-cut’. With the values of nodal temperature. The selection of a suitable cutting time increment is very difficult to perform for a user without doing experiment in advance. b. the tool geometry model is updated by moving nodes. When a sharp tool is used. In the last phase. the nodal wear rate is calculated in the third phase. instead of being performed automatically according to a certain algorithm.) Result Simulation study was made with worn tool initially including a pre-defined wear land of 0.Introduction 28 The complete procedure includes four phases. . In the first phase.) Limitations • • The tool geometry was updated manually. c. This problem was improved by using a new tool wear model especially developed for the simulated cutting condition [Fran-02]. the location of the maximum wear rate and the low wear rate close to tool radius are consistent with the experimental result.06mm on the flank face. the predicted wear rate on the flank face is one order of magnitude smaller than that on the rake face. a coupled thermalviscoplastic Lagrangian cutting simulation combined with an introduced special simulation module. pure heat transfer analysis for the tool is performed to attain thermal steady state in the tool. The wear rates of flank wear and crater wear are of the same order. Both the chip formation and heat transfer analyses are performed with commercial FE code DEFORM-2D. while crater wear and flank wear occur simultaneously at a similar wear rate in the experiment. is used to perform chip formation analysis until mechanical steady state is reached.

is different from turning operation because of the lack of steady state. as shown in Table 1. Tool wear estimation in intermittent cutting. • At present.3. for example. the simulation of cutting process is assumed more suitable to be performed with explicit method because of the large deformation. • Because of the short history of the research on tool wear estimation with FEM. milling operation. However. The end of tool life in intermittent cutting. • Only the commercial FE code DEFORM-2D is used in tool wear estimation.3 Summary Of Literature According to the above literature analysis. The study should be carried out with some FE code using explicit method and providing good development platform as well. only 2D tool wear of uncoated carbide tool cutting carbon steel workpiece AISI1045 was studied. numerical implementation of tool wear estimation is only developed for the cutting process with steady state. ABAQUS.4. The cutting type is limited turning operation and orthogonal cutting. is mainly caused by progressive tool wear. . impact and complex contact problem.2. Therefore the estimation of tool wear should be studied by developing new simulation procedure. for example. some conclusions can be obtained: • The advantages of tool wear estimation with FEM over tool wear estimation with the combination of analytical method and FDM are considered in several aspects.Introduction 29 1.

Tool wear rate model Predicted wear value Tool Only crater wear or only flank Crater wear and flank wear wear simultaneously FEM heat transfer analysis. tool included in tool geometry Experimental data Applicable conditions Prospective Yes. Assumption and simplification of Tool wear rate model With FEM For crater wear estimation. e. cutting force. method. the cutting condition. energy FEM chip formation analysis. etc Conventional cutting speed Conventional HSC Limited A necessary supplement to the empirical method cutting and . without crater wear. flank wear and without flank wear. Edge preparations are not considered edge preparation can be model For flank wear estimation.Introduction 30 Table 1. tool Crater wear.g.4 Comparisons between tool wear estimation with FEM and tool wear estimation with the combination of analytical method and FDM Compared aspects Realization With the combination of analytical method and FDM Analytical method. tool-chip No contact length.

Several modeling tools are used in order to accomplish the entire research project. 2. This tool wear estimation method is performed by predicting tool temperature. turning and milling operation Fig. The study is not limited to turning operation. TURNING OPERATION MODELING TOOL MILLING OPERATION Continuous chip formation analysis model ABAQUS/Explicit Fortran General chip formation analysis model Thermal steady state analysis ABAQUS/Standard Fortran Cyclical thermal balance state Analysis model Tool wear estimation program Objective-oriented programming language: Python Tool wear estimation program Objective: 2D. FEM simulation of turning and milling process are studied at first. The implementation of tool wear estimation is relatively easier and studied first.1 Objective and modeling tool Turning operation is a steady-state process when continuous chip is formed.1 Objectives The objective of this research is to develop methodology to predict tool wear and tool life in cutting process using finite element simulations. uncoated carbide tool.Objective And Approach 31 Chapter 2 Objective And Approach 2. sliding velocity of chip and normal stress on tool face with FEM simulation of cutting process. Therefore to achieve the objective. Based on . Because of the complexity of tool wear mechanisms and forms in real cutting process. the prediction of tool wear in milling operation is considered as well. including chip formation analysis and pure heat transfer analysis. the study at present will be concentrated on two-dimension tool wear estimation of uncoated carbide tool in dry cutting mild carbon steel.

wear calculation and tool geometry updating. 2. j ) ⋅ ∆t j ⋅ D(i . j ) Tool temperature θ Time increment ∆ t (specified by user or searched by program) ∆t Fig. as shown in Fig. Sliding velocity vc Width of flank wear VB [mm] Tool geometry updating Tool reshape criterion (eg. j ) = w(i . one is for turning operation. another for milling operation. Two different tool wear estimation models are developed. 2. heat transfer analysis. 0.2 Approach Although the tool wear estimation models for turning and milling operations are different. It can simulate the entire chip formation process from initial chip .2mm) tj-1 Heat flux Temperature Heat tranfer analysis Nodal displacement ∆t 1 ∆t 2 t0 t1 t2 ∆t 3 t3 ∆t 4 Cutting time t [min] t4 Wear calculation & w = Cσ t vs exp (− λ θ ) Nodal displacement r r & w(i .Objective And Approach 32 the obtained experience in turning operation.2. the methodology of tool wear estimation in milling operation is discussed by analyzing the feature of milling operation.2 Approach and procedure of tool wear estimation The study process for turning operation includes: Stage 1: chip formation analysis A new chip formation modeling method for continuous steady state chip formation is developed. Updated tool node file Chip formation analysis The jth calculation cycle Normal pressure σ t . the calculation procedure are similar and mainly composed of chip formation analysis. 2.

sliding velocity. Then the tool wear estimation modeling is studied. normal stress. the searching method of a suitable cutting time increment. Stage 3: tool wear estimation modeling Through the previous stages. Stage 2: heat transfer analysis In order to analyze the variation of tool temperature in the further milling cycles.Objective And Approach 33 formation. Stage 2: heat transfer analysis In order to save the calculation time. then pure heat transfer analysis of the tool is performed for several milling cycles. Then the tool wear estimation modeling is studied. The study process for milling operation includes: Stage 1: chip formation analysis The chip formation modeling method in milling operation is studied. It simulates the chip formation process in the first milling cycle. . The concerned modeling problem is discussed. the calculation of nodal displacement due to wear and tool geometry updating. the calculation of nodal displacement due to wear and tool geometry updating. the cooling of the workpiece is studied. It includes the calculation of wear rate at steady state. the temperature distribution in the cutting tool at thermal steady state is studied by performing pure heat transfer analysis. chip growth to steady state by making use of Arbitrary Lagrangian Eulerian technique in ABAQUS/Explicit. and tool temperature can be obtained. the searching method of a suitable cutting time increment. sliding velocity and tool temperature at steady state can be obtained. It includes the calculation of average wear rate in one selected milling cycle. Stage 3: tool wear estimation modeling Through the previous two stages. normal stress.

chip formation process is simulated using commercial FEM code.1 Explicit Algorithm In Chip Formation Simulation The chip formation simulation is performed using explicit method.1 Introduction Optimisation of the cutting process requires comprehensive knowledge about the relation between cutting process and the combination of cutting parameters. In this chapter. 3. M is the diagonal or lump mass matrix. dynamic and thermal analysis procedures are based on the implementation of an explicit integration rule. cutting tool and workpiece. Then the accelerations are integrated through time using the central differential rule. (∆t( i +1) + ∆t( i ) ) 2 & u 1 (i + ) 2 & =u 1 (i − ) 2 + && u( i ) (3.1.1.Chip Formation Simulation Technology 34 Chapter 3 Chip Formation Simulation Technology 3. 3.1 Dynamic Analysis Procedure Dynamic analysis procedure is performed with the following algorithm. In the simulation the entire cutting process is discretized into many small time increments. ABAQUS/Explicit.1.1) && where u(i ) is the acceleration at the beginning of the increment i. • Nodal calculation Accelerations are calculated by satisfying the dynamic equilibrium at the beginning of the increment: && u(i ) = M −1 ( P( i ) − I ( i ) ) (3. In every small time increment. P(i ) is externally applied load.2) . and I (i ) is internal load.

3) 1 (i + ) 2 • Element calculations & Element strain increment. & The values of θ (N) are computed at the beginning of the increment by i (3.1.4) In the chip formation analysis. dε . & σ = f (ε . P J is the applied nodal source vector.5) θ&(N) = (C NJ ) −1 ( P(iJ) − F(iJ) ) .6) C NJ is the lumped capacitance matrix. ε . i where (3. Then stress. The heat transfer . is computed from the material constitutive equation. F J is the internal flux vector.Chip Formation Simulation Technology 35 The velocities are integrated through time.2 Thermal Analysis Procedure (3. & u(i +1) = u( i ) + ∆t(i +1)u (3. ε . θ (N+1) = θ (N) + ∆t (i +1)θ&(N) i i i where θ N is the temperature at node N.1. θ ) 3. heat transfer equations are integrated using the explicit forward difference time integration rule. Fully coupled thermal-stress analysis is employed. The explicit integration rules are realized in both dynamic and thermal analysis procedures by using lumped mass matrix and capacitance matrix. is computed from the strain rate. σ . the stress analysis is dependent on the temperature distribution and the temperature distribution depends on the stress solution. In the analysis. which is decided according to the velocity of nodes.

7) where wmax is the highest frequency in the system of equations of the mechanical solution response and λ max is the largest eigenvalue in the system of equations of the thermal solution response. For coupled thermal-stress analysis.Chip Formation Simulation Technology 36 and mechanical solutions are obtained simultaneously by an explicit coupling. In order to produce accurate result.1. Therefore no iterations or tangent stiffness matrices are required. ) wmax λ max (3. the time increment must be quite small so that the integrated variables are nearly constant during an increment. The time increment must be smaller than a stability limit otherwise the solution becomes numerically unstable. the stability limit can be calculated by ∆t ≤ min( 2 2 .2 Stability Limit The central difference and forward difference integrate constant accelerations. The ABAQUS/Explicit solver supplies the default time incrementation scheme. which is fully automatic and requires no user intervention. 3. . velocities and temperature increments per unit time.

a longer workpiece has to be used in the simulation. are machined. A small crack is always created before tool edge. the length of the workpiece is often very small. But evidences from cutting experiments show that for the quasicontinuous chip formation that takes place in machining ductile materials under favourable cutting conditions. crack occurs along the shear direction [Didj-97]. crack is found ahead of the cutting edge. and aluminium.2 Continuous Chip Formation Simulation 37 Continuous chip is very common when most ductile materials. only enough to produce a steady chip shape.2.1. 3. 3. cutting force and temperature. mild steel. Only for discontinuous chip formation and chip formation with buildup edge. Cutting under these conditions is steady-state process with steady chip shape. • In most chip formation models. If analysis of the further cutting process is required. copper. 3. (a) Node debond [McCl-02] (b) Element removal [Ng-02a] Fig.Chip Formation Simulation Technology 3.1 Cracks formed before tool edge in the simulations . chip separation is realized by element removal or node debond. which increases calculation time direct proportionally. They are based on different approaches: Lagrangian or Eulerian. Many chip formation models were developed for these cutting conditions with different FE codes. as shown in Fig. such as wrought iron. Under the consideration of reducing calculation time.1 Limitation Of The Existing Chip Formation Models This continuous chip formation process cannot obtain very satisfactory simulation result because of the limitations of the existing models using ABAQUS FE code: • Most models take chip formation as a Lagrangian problem.

Although many material constitutive models are provided for the commonly used material by literatures. the chip swells up from the initial geometry . They should be determined experimentally instead of being given at random.2. When the given initial chip geometry is not suitable. In addition no shear failure criterion is required. The failure parameters are material dependent and different parameters are required for Johnson-cook and other material models. the position of the predetermined chip separation path has influence on the volume of material to be cut away. When round edge cutting tools are used. as shown in Fig. the conflict between the cutting time to reach steady state and the length of the workpiece model limited by calculation time is solved. Whether the cutting force components. as reported by Arrazola et al [Arra-02]. • A chip separation criterion is necessary. Failure parameters relate to the successful implementation of chip formation simulation.Chip Formation Simulation Technology 38 • The chip separation path is often predetermined instead of formed automatically by the deformation of workpiece material under the cutting action. • When the steady-state chip formation process is modelled as a Eulerian problem with ABAQUS/Explicit. etc change with the position of the separation path or not needs analysing as well. residual stress. This limits the usage of many material constitutive models. 3. (a) When the given initial chip is thinner than in experiment. chip geometry cannot adapt itself when the chip tends to swell up or to shrink too much from the given initial geometry. chip separation criterion is given by defining material failure criterion. the failure parameters are seldom given as well. But an initial chip geometry must be given according to experiment or experience in machining. In ABAQUS. Steady-state analysis is performed by prolonging the cutting time without increasing the length of the workpiece.

chip growth. No initial chip geometry is required. • Chip separation is realized with adaptive meshing technique supplied by ABAQUS/Explicit.Chip Formation Simulation Technology 39 (b) When the given initial chip is thicker than in experiment. User can prolong the cutting time without changing the size of the control area.2. No obvious crack is formed. no shear failure criterion or material failure parameters are required. the chip shrinks from the initial geometry Fig. 3. This model has the following advantages: • Workpiece geometry in the model only stands for a control area. Chip formation simulation includes the entire process from initial chip formation.2. Workpiece material is unlimited and flowing through this control area continuously. It combines the features of pure Lagrangian and Eurerian analysis and can be used . Most of the material constitutive models can be used in this model. to steady state. a new continuous chip formation model is developed with ABAQUS/Eplicit. 3.2 Problems in the chip formation analysis with an unsuitable initial chip geometry [Arra-02] 3. • • • • Chip separation is performed by the deformation of the workpiece material. Good contact is maintained in the cutting tool edge area.2 Advantages Of The New-developed Chip Formation Model Due to the limitations mentioned above.3 Adaptive Meshing Technique In ABAQUS/Explicit This model is developed based on adaptive meshing technique in ABAQUS/Explicit. instead of forming crack along a predetermined path. This model is especially suitable to simulate the cutting process with round edge tool or chamfered tool.

steadystate chip formation. .2. 3. Chip Tool E Boundary region edge L L E: Eulerian boundary region S E S S: Sliding boundary region L: Lagrangian boundary region Material flow direction Workpiece Fig. but it is completely unconstrained in the directions tangential to the boundary region. as in a fluid flow problem. and Eulerian problems. Mesh on the Eulerian boundary regions are fixed in space using spatial mesh constrains. and material flow velocity across the boundary is defined by boundary conditions. the material is constrained to move with the material in the directions normal to the boundary region. the mesh is constrained to move with the material in the direction normal to the surface of the boundary region and in the directions perpendicular to the boundary region edge. In Lagrangian boundary region. sliding or Eulerian. Eulerian boundary regions can be defined only on the exterior of a geometry model and the material flows across the boundary.Chip Formation Simulation Technology 40 to both Langrangian.3 Boundary regions in chip formation model The boundary regions of the adaptive meshing domain can be either Lagrangian.g. 3. e. In sliding boundary region. only sliding and Eulerian boundary regions can be defined. 3. When the adaptive meshing domain is Eulerian type.1 Boundary Region Types Adaptive meshing is performed in adaptive meshing domains.3. for example the workpiece in Fig. which can be either Lagrangian or Eulerian. e.g. initial chip formation.3.

Therefore it is sometimes necessary to deactivate the geometry features by defining a greater initial geometry angle.2. solution-dependent meshing is used to focus mesh gradation toward these areas automatically by defining the curvature refinement weight αc a high value. .4 Geometry Features (a) Geometry feature is detected.2 Geometry Features 41 On boundary regions where the angle θ between the normals on adjacent element faces is greater than an initial geometric feature angle θI given by the user. no mesh flow is permitted past the corner (b) No geometry feature is detected. 3.2. Having sufficient mesh refinement near highly curved boundaries is very important to model the detail of the chip shape near the chip separation area.Chip Formation Simulation Technology 3.3 Curvature Refinement During adaptive meshing. α c = 1 . mesh flow is permitted 3. geometry features are detected initially. mesh-smoothing algorithms based on minimizing element distortion tend to reduce the mesh refinement in area of concave curvature.3. n θ > θI n n θ n θ ≤ θI (a) (b) Fig. Adaptive meshing cannot be performed well across such geometry features because the nodes cannot move across the geometry features unless they flatten. To prevent the natural reduction in mesh refinement of areas near evolving concave curvature.3. for example. especially as the curvature evolves.

145mm / r In the finite element model. the cutting tool is defined as a rigid body.2mm. Moreover in the first two steps. During all the chip formation steps. which consists of three analysis steps.Chip Formation Simulation Technology 42 (a) αc=0 (b) αc=1 Fig. rε = 0.4 Analysis Steps The entire continuous chip formation process is performed with a complete modeling procedure from initial chip formation to the realization of steady state. turning operation. The simulated cutting condition is given in Table 3. whereas in the last analysis step the cutting tool is modelled as a deformable body in order to obtain better analysis result and more comprehensive analysis data. the workpiece has a size of 0.1.0245mm vc = 300m / min . chip growth. including initial chip formation. which is meshed with 4-node bilinear coupled temperature-displacement plain strain elements CPE4RT.2. f = 0. . coupled thermo-stress analyses are performed.6×3. In order to save calculation time. and steady-state chip formation as described in detail in the following parts. a p = 2mm . dry cutting Mild carbon steel AISI1045 Uncoated carbide WC-Co γ o = −7° .5 Effect of curvature refinement 3.1: Cutting condition Cutting type Work material Tool material Tool geometry Cutting parameters Orthogonal cutting. only the part of the cutting tool near the cutting edge is included in the chip formation modelling. The first two analysis steps supply steady chip geometry for the steady-state chip formation analysis step. 3. α o = 7° . Table 3.

then taking an angle value between this maximum angle value and 90 deg (because the angles at point A. According to this value. Adaptive meshing can be performed on any boundary regions except point A.4. C and D are about 90 deg). 3.2. the initial geometry feature angle should be given a suitable value. This chip formation process is modelled as a Lagrangian problem. the Lagrangian boundary region on the top surface of the workpiece traces the chip material continuously and forms the shape of the chip. At the cutting edge. B. 3. chip material separates with workpiece material. The value is defined by calculating the angle θ i between the normals on adjacent element faces in chip separation area. the four corner points A. 3. C and D in Fig.6 Initial chip formation analysis The initial workpiece geometry is designed to have a concave at the top-right corner under the consideration of seeding more nodes along the concave surface (see Fig. The boundary of the workpiece consists of only Lagrangian boundary regions.6(a) can be detected as geometry features but corner point E and the workpiece nodes on the surface ED should not be taken as geometry features.6(a)).7. finding out the maximum angle value.Chip Formation Simulation Technology 43 3. During the initial chip formation process. In order to maintain the mesh refinement in this area while the chip formation process continues.18ms Fig.1 Initial Chip Formation This analysis step aims to form initial chip geometry. . B. 3. Only very fine mesh can show exactly the shape of this area. vc A E D B (a) Initial geometry and mesh C (b) Initial chip geometry formed at t=0. C and D. B. as shown in Fig.

At the beginning the cutting tool is at the right side of the workpiece.18ms. With the cutting tool advancing into the workpiece. in this simulation the initial state information is read from the former analysis step at 0. This analysis step aims at forming steady chip geometry. when a desirable initial chip shape is produced. After 0.6(b) shows the mesh after the initial chip is formed. initial temperature input file. Then it writes them into the model files of the chip growth analysis step. 3.2. In this step the chip formation process is treated as a Eulerian problem. Fig. 3. chip growth analysis step is performed. The workpiece is fixed and the cutting tool is moving in the negative x-direction1. elements along the concave surface extend and compose the outside surface of the chip. A user program is developed with Python language. nodal temperature. including node input file.Chip Formation Simulation Technology 44 Tool θi Workpie Fig. It reads the variables about nodal coordinate. etc.7 Determination of initial geometry feature angle In addition solution-dependent meshing is used to focus mesh toward the chip separation areas automatically by setting the curvature refinement weight αc to unity.18ms an initial chip is formed. Therefore the initial state of the workpiece and the cutting tool in this step remains the state at the selected time point of the former analysis step. . x-direction is pointed to the right side and y-direction to the top of the page. 1 In all figures of this paper. For example.4. 3.2 Chip Growth After the initial chip is formed. etc of the workpiece and the cutting tool from the selected time point of the initial chip formation analysis step.

According to the definition of .09ms Fig. The left and right boundary of the control area are defined as Eulerian boundary regions. 3.8(a). The curvature refinement weight αc is set to unity.8 Chip growth analysis 3.3 Continuous Steady-state Chip Formation In the second analysis step. The movement of the mesh on the bottom boundary is constrained in y-direction. as indicated with the small arrows in Fig. Fig.8(b) shows that the chip is growing with the material flowing into the control area. The top and bottom boundary are sliding boundary regions. whose mesh is fixed in x-direction by using mesh constrain definition. the mesh of the chip extends too much in the direction of chip growth so that adaptive meshing cannot solve mesh distortion problem any more.8(a).Chip Formation Simulation Technology 45 In this step. Initial geometry feature angle is defined in the same way as explained above. the mesh will adjust itself to fit in with the developing chip geometry. The cutting tool is fixed in space. In this step. The steady-state chip formation analysis step is designed for simulating the further cutting process. 3. 3.4. The workpiece mesh in Fig. but material flows into the control area continuously from the left boundary at the cutting speed and flows out of from the right boundary. 3. with the chip growing to a certain extent. indicated with small triangle in Fig. But the movement of the mesh on the top boundary is not constrained. (a) Initial geometry and mesh (b) Chip growth at t= 0. the cutting process is treated as a Eulerian problem as well. the relative movement between the cutting tool and the workpiece is performed by the movement of workpiece material. 3.8 represents only a control area.2.

its position and size can adjust with the chip automatically in y-direction. as shown in Fig. (a) Initial geometry and mesh Fig.09ms. In order to allow the chip to flow out of the mesh and grow unlimitedly instead of grow visually with the mesh. The mesh of the boundary is not constrained in y-direction.Chip Formation Simulation Technology 46 the workpiece mesh. 3. Fig. the mesh in Fig. see Fig. Method 1: Mesh Modification During the growth of the chip in the second analysis step. 3. two methods can be used in the continuous steady-state chip formation step: mesh modification and model regenerating. 3. 3. The state of the workpiece and the cutting tool at this time point is written into the model files of steady-state cutting analysis step.9(b) shows the formed mesh at 1ms. Because the cutting tool is a deformable body.9(a). 3. its movement is fixed by defining constraint in x-direction at the right boundary and in y-direction at the top boundary. The coordinates of the nodes on this boundary are adjusted to locate these nodes along a vertical line in order to facilitate the definition of mesh movement constrain in x-direction.9(a) is read from chip growth analysis step at t=0.8(b). The definition of boundary regions and conditions for the other part of the workpiece control area are similar to those in the second analysis step. Initial geometry feature angle and curvature refinement weight αc are defined in the same way as explained above.9 Mesh modification (b) Mesh at t=1ms . For example. the mesh at the top boundary of the chip is defined as a Eulerian boundary region. the chip geometry near the chip root becomes steady since a certain time point.

The chip is 0. 3. The regenerated workpiece model in Fig. When the mesh concentrates in the cutting edge area according to solution-dependent meshing rule. it will complicate the definition of mesh constraint. e. (a) Initial geometry and mesh (b) Mesh at t=1ms Fig. crater wear. the mesh in other area becomes coarse. initial chip formation and chip growth analysis step.Chip Formation Simulation Technology 47 Method 2: Model Regenerating This method is especially important when the cutting tool has some special geometry.10 Model regenerating . ABAQUS/Explicit supplies only r-adaptivity. They can be obtained from the former two analysis steps. 3.3mm thick. The information necessary for model regenerating includes chip thickness and toolchip contact length. in which a steady chip shape is formed. Model regenerating supplies an approach to improve the contact problem. The mesh movement is constrained in y-direction. The mesh position and size in xdirection will be adjusted with the chip automatically.5mm long. The mesh at the top boundary of the chip is defined as a Eulerian boundary region. and 0.10(a) has a chip connected to the workpiece.g. and good contact between the workpiece and the cutting tool in these areas is desired. It is larger than the tool-chip contact length. The length of the chip should be determined carefully. But when the chip is too long. But sometimes fine mesh along the whole tool-workpiece and tool-chip interface is required.

3.11 Contact status along tool-chip interface 3. Maximum stress is located in the primary shear zone. the workpiece material undergoes serious shear plastic deformation in primary shear zone and becomes chip. Contact problem (a) Contact problem is created by using mesh modification (b) Contact is improved by using model regenerating Fig.3.1 Stress Analysis Fig. the workpiece and the chip throughout the entire steady-state analysis process. This ensures good contact between the cutting tool. as shown in Fig.2.12 shows the stress distribution in the three analysis steps. 3. .11.Chip Formation Simulation Technology 48 Very fine mesh is given along the entire tool-chip interface.2.5 Results & Discussion 3.5.

3ms (c) Steady state analysis. 3.12 Stress distribution (MPa) In steady state analysis step. The newly formed machined surface has contact and friction with the round edge and sometimes a small part of the flank face. t=0. 3.2 Plastic Strain Analysis Fig.18m (b) Chip growth analysis. t=1ms Fig.2. This results in a high stress in the workpiece material beneath the cutting tool edge. very high stress is observed in the small part of the cutting tool directly under the tool-chip contact area. The material in the chip underside deforms plastically again .13 shows that the workpiece material undergoes serious plastic deformation in primary shear zone. 3. the cutting tool is modelled as a deformable body.Chip Formation Simulation Technology 49 Further the underside of the chip undergoes high stress because of the contact and friction with the tool face when sliding away.5. t=0. (a) Initial chip formation analysis.

13(a). 3. t=1ms (c) Model regenerating. the two model methods create similar plastic strain field. This results in higher plastic strain formed in the chip underside than in the other part of the chip. 3. Under the example cutting condition. not including the created plastic strain. t=1ms Fig.09ms (b) Mesh modification.2. In the steady-state analysis. t=0.14 shows the distribution of strain rate. the top of the chip has no plastic strain because only the geometry of this part is imported from the initial chip formation analysis step. 3.3 Strain Rate Fig.5. and only the maximum values are different which is caused by the different contact condition due to element size in the workpiece model and the difference between deformable cutting tool and rigid body cutting tool.Chip Formation Simulation Technology 50 under the pressure and friction of the cutting tool face. which is defined as solution SDV9 by material subroutine. the maximum strain rates . 3.13 Equivalent plastic strain distribution In Fig. (a) Chip growth analysis step.

within the cutting process of 1ms cutting temperatures at most of the tool face nodes in the tool/chip interface.15(b).5. the highlighted nodes in Fig. especially in the areas close to the cutting tool edge and the free surface of the workpiece. This means that thermal steady state is not realized in the whole cutting tool. t=1ms (c) Modal regenerating. (a) Chip growth step. i.e. the highlighted nodes in Fig.14 Strain rate distribution 3. 3. and reach up to 105.15(c). the temperature is still climbing.4 Temperature Analysis In Fig. as shown in Fig. 3.09ms (b) Mesh modification. while at the tool face nodes inside the cutting insert. t=0. 3. which is assumed as typical maximum strain rate in conventional machining [Arnd-73]. 3.15(a).2.Chip Formation Simulation Technology 51 distribute along the primary shear zone. 3.15(d). t=1ms Fig. . is reaching steady values.

Chip Formation Simulation Technology 52 (a) Position of monitored tool face nodes (b) Temperature progress Node 429 Node 497 Node 588 Node 507 Node 492 (b) Position of nodes inside the tool step (d) Temperature history Fig. 3. . The highest temperature is at the rake/chip interface and most part of the tool is still at room temperature. 3.15 Temperature history of tool nodes at steady-state chip formation analysis Fig.16 shows the temperature distribution at 1ms.

3.0004 0.5 Verification With Experimental Data By adding the reaction force component in the same direction at all constrained nodes of the cutting tool and then taking the negative value. ap=2mm.0012 Time [s] Fig.Chip Formation Simulation Technology 53 Fig.2.5.0000 Fc Ft 0. f=0.17 shows that the cutting force components change within a very narrow range from 0. and it is deemed that the mechanical steady state is realized.0008 0. the cutting force components Fc and Ft in the continuous steady-state chip formation step are obtained. 3. 3. Fig.145mm/r) .16 Temperature distribution at t=1ms of steady-state chip formation analysis step 3.17 Cutting force progress (under cutting condition: vc=300m/min. 700 Cutting and thrust force [N] 600 500 400 300 200 100 0 0.7ms.

The modelling method developed for the continuous chip formation is not suitable to simulate chip formation process in milling operation.30. Deform2D and Oxcut-F [Söhn-03]. while S from continuous steady state chip formation step.18.48 gives the better result than other FEM code. the prediction error of cutting force Fc is about 2%. cutting action is discontinuous and the chip produced is discontinuous. 3.3 Chip Formation Simulation For Milling Operation In milling operation. and the value is 0. 3. another is calculated according to Eq. 2% 5% Fig. f=0. and the error of thrust force Fc is about 5%. ABQ-f048I and ABQf048S.3. the cutting force values obtained from the simulation with ABAQUS/Explicit code are compared with the experiment data [Feve-01] and the result from other software including Third Wave. Fig.145mm/r) 3.18 Comparison of cutting force (under cutting condition: vc=300m/min. ap=2mm.48. Therefore a different modelling method is introduced in the following part. The results from ABAQUS include ABQ-f030I.18 shows that the cutting force components obtained from initial chip formation step and continuous steady state chip formation step have no great difference. ABQ-f030S. Two different frictional coefficient values are used.Chip Formation Simulation Technology 54 In Fig. 3. 1. I means the result is obtained from initial chip formation step. . One is 0. The cutting force components created in the chip formation analysis with the frictional coefficient of 0.

3. is dependent on a nondimensional plastic strain rate. when the equivalent plastic strain reaches the strain at failure ε fpl . 3. or 1 depending on the temperature range).3.9. and a nondimensional temperature.Chip Formation Simulation Technology 55 3. There are two methods to define the strain at failure. θ (defined as 0. Stain at failure is defined by giving the failure parameters d1 − d 5 .   ε fpl = d1 + d 2 exp d 3    &  ε pl p    1 + d 4 In ε q     &0  ˆ  1 + d 5θ   ( ) (3. The damage parameter. the element is removed from the mesh. . is defined as  ∆ε pl w = ∑  pl  ε  f where ∆ε  . (θ − θ transition ) (θ melt − θ transition ) . For Johnson-cook plasticity model. chip separation is realized by defining shear failure criterion. ε fpl . In this method.1.1 Chip Separation In every milling cycle. w . Hence the adaptive meshing technique in ABAQUS/Explicit cannot be used as a chip separation tool any more. a dimensional pressure-stress ratio.1 Shear Failure Criterion The shear failure model is based on the value of the equivalent plastic strain at element integration points.   (3. the stain at failure is given according to Eq. The summation is performed over all increments in the analysis. & & ε pl ε 0 . material failure takes place.9) where strain at failure. then the damage parameter w exceeds 1. p q (where p is the pressure stress and ˆ q is the Mises stress).3. If at all the integration point material failure takes place.8) pl is an increment of the equivalent plastic strain. the produced chip will separate with the newly produced workpiece surface without any connection when the cutting tool disengages from the workpiece.

19 Velocity of material points at workpiece nodes on the chip underside and the machined surface (the arrows shows the velocity vectors) . The material above the separation area moves upwards into the chip and the material below the separation area moves downwards to join in the machined surface. the hydrostatic stress p and temperature [Ng-02b].2 A Numerical Method To Determine Strain At Failure Normally. pressure/stress ratio and temperature are given directly in tabular form in the data line. Observing the movement of material points on the chip underside and the machined workpiece surface in steady-state chip formation process. it is possible to determine stain at failure without making any experiment. 3. 3.Chip Formation Simulation Technology 56 For Mises plasticity model. equivalent plastic strain at failure. By employing the continuous chip formation analysing methods. is obtained by using experimental methods. ε fpl . strain at failure or the dependencies of strain at failure on strain rate. we can find a separation area of the workpiece material.1. Node 13 Node 14 Node 15 Node 16 Fig.19. Ng et al integrated orthogonal tests with some analytical equations in metal cutting theory to define the dependency of & the equivalent plastic strain at failure ε fpl on the plastic strain rate ε pl . 3. For example. For example.3. Bacaria et al determined failure parameters d1 − d 5 by performing tensile and torsion tests [Baca-00]. in Fig. the separation area is between Node 13 and Node 16.

3. the dependency of strain at failure on temperature. strain rate. The directions of sliding velocities of the material points in the area between Node 15 and Node 16 change.20. Node15 Node16 (a) Monitored points Node15 Node15 Node16 Node16 (b) Sliding velocity of monitored points (c) Equivalent plastic strain of monitored points Fig.7 to strain at failure.5. The equivalent plastic strain between Node 15 and Node 16 gives a value range from 2. etc can be studied. 3. In the following part of this chapter. pressure. strain at failure for mild carbon steel is set to 2.Chip Formation Simulation Technology 57 According to the sliding velocities of the workpiece material points along tool-chip interface. .25 to 2.20 Determination of strain at failure By varying cutting parameters or tool geometry. It can be assumed that material failure is taking place in this area. a more exact position of the separation area can be defined in Fig.

3. only a small part of the workpiece and the cutting insert is included in the model. The workpiece is 2mm high.2 Cutting condition Cutting type Work material Tool material Tool geometry Cutting parameters Orthogonal cutting.21 Initial geometry.7mm and inside radius 62. The extension of its upper . a p = 1mm .2 Chip Formation Modeling 58 This chip formation modeling method is explained by taking a milling case as an example in which an uncoated carbide tool is used to machine mild steel CK45. f z = 0. whose outside radius is 62. In order to reduce the calculation time. milling operation. Fig. dry cutting Mild carbon steel CK45 Uncoated carbide WC-Co γ o = 7° . 3.2 Insert Fig.2. The centre of the ring is positioned at the rotation centre of the cutting insert.Chip Formation Simulation Technology 3. Table 3. mesh and assembly of the tool and the workpiece in chip formation analysis The workpiece is simplified as a small segment of a ring.2mm / r The diameter of the milling tool is 125mm. 0.3mm. mesh and assembly of the workpiece and the cutting insert. α o = 7° vc = 600m / min . 0. a e = 2mm .21 shows the initial geometry. The cutting condition is given in Table 3.4 Rotation center Workpiece 2 5 62.3.

Every boundary segment of workpiece is defined as a Lagrangian boundary region. The chip formation analysis is performed for 0.27ms. one criterion is assigned to a line of element along the moving path of the cutting edge to separate the chip from the workpiece.Chip Formation Simulation Technology 59 surface passes through the center of the ring and the lower surface is parallel to the upper surface. In the model the shear failure criterion is integrated with a material model designed specially for the workpiece material CK45 and assigned to the whole workpiece.3 Result & Discussion 3. covering the whole cutting phase and 0. With the tool rotating in clock-wise direction. One milling cycle takes 39. The cutting insert in the model includes only the part near the cutting edge.3ms of the later cooling phase. which is discretized with CPE4RT elements. The cutting insert is modelled as a deformable body in order to obtain all the necessary cutting process variables for the latter study on tool wear.3. 3.22(b) shows that the primary deformation zone has the maximum stress in the workpiece.22(a). The chip formation process is treated as a Lagrangian problem.07ms. A small chip is formed. the cutting insert engages in the fixed workpiece. which results in a high stress in this area. and local fine mesh is given along the moving path of the cutting edge because of very high gradients of solutions in this area. 3. and there is no contact with the workpiece. as shown in Fig. etc. . 3. The workpiece is discretized with a mesh composed of CPE4RT elements.3.5ms. stress. There are different ways to assign shear failure criterion to form different shape of chips. In each milling cycle. cutting phase takes 0.2ms and cooling phase takes 39. another criterion is assigned to part of the chip material to generate cracks in order to simulate serrated chips [Ng-02b]. and the contact between the chip and the cutting insert concentrates in a small area near the cutting edge.3. such as temperature. Fig. the cutting insert is at the bottom of the workpiece. Ng et al designed two different kinds of shear failure criteria.1 Stress Analysis At the beginning. Bacaria defined only one material shear failure model for the whole workpiece material [Baca-00].

But after crack generates.Chip Formation Simulation Technology 60 In Fig. instead of along the moving path of the cutting edge. (a) t= 0.22(c).200ms Fig. which provides a possibility for burr formation.175ms (d) t=0.100ms (c) t=0. The workpiece material to be cut away deforms seriously under the pressure of the cutting insert and protrudes from the original top surface. the cutting edge is bearing higher stress than other part of the insert because of positive rake angle and very sharp tool edge. the cutting insert is disengaging the workpiece.22 Stress field (Mpa) in the chip formation analysis . 3.025ms (b) t= 0. it propagates along the direction of maximum stress deeper and deeper into the workpiece material. 3. During the entire cutting phase.

23(d). the shear zone and the chip underside sliding along the tool face. 3.Chip Formation Simulation Technology 61 3.175ms (d) t=0. 3.23.3. 3.23(a).200ms Fig. The heat is generated mainly in two zones. 3.23(b). local high temperature is formed.3. The obvious temperature increment take place in shear zone. In addition. when chip breakage takes place in Fig. then the chip underside is heated again to a higher temperature by the friction with the tool face. as shown in Fig.100ms (c) t=0.23 Temperature distribution (in Kelvin) in the chip formation analysis .23(c) and 3.025ms (b) t= 0.2 Cutting Temperature The predicted temperatures generated during chip formation process are shown in Fig. (a) t= 0. 3.

3. Therefore cutting force is exerted on the rotation center point. which always appear when the element of workpiece is coarser than tool element.24 Cutting force progress during the cutting process Cutting force [N] . Fig. Fig.3. they reach the shear failure criterion and then stresses in these elements are set to zero. 3.Chip Formation Simulation Technology 62 3. Because the cutting insert has exited from the workpiece and no contact with the workpiece any more after 0.2ms. This is different from ‘noise’ observed in continuous chip formation analysis.24 shows the cutting force progress during the cutting process.3. and contact problem results in ‘noise’ of cutting force signal. cutting force components in x-direction and y-direction are reducing to zero.3 Cutting Force Analysis In order to make the cutting insert rotate as a deformable body. its bottom is pinned on and rotates with the rotation center point. The ‘noise’ of the cutting force signal is caused by the removal of the elements. which result in the fluctuation of cutting force.

The shear failure criterion is used to the entire workpiece. Especially when it is used in tool wear estimation. Chip separation is formed automatically by using ALE technique supplied by ABAQUS/Explicit. chip growth and steady state. such as serrated chip. only steady state analysis step is necessary.4 Summaries & Conclusion 63 Two different chip formation modeling methods are designed to simulate the chip formation process in milling operation and turning operation. the normal tool geometry such as blunted. In order to get good contact between the chip and the tool face even when a serious crater wear is formed on rake face. With this complete model. chamfered and worn cutting tool can be used in the chip formation model.Chip Formation Simulation Technology 3. Chip formation model for turning operation is designed to simulate the whole cutting process including initial chip formation. . The calculation time to reach steady state is relative short comparing with the chip formation model in which the chip formation is taken as a pure Lagrangian problem. Chip formation in milling operation is modeled by introducing the shear failure criterion because of the intermittent cutting process. model regeneration method is suggested to update and refine the mesh of the workpiece. when the suitable material constitutive and material failure model are provided. This model is expected to have a wider application because it can model various chip type. No experiment is required to get material failure parameters or steady chip geometry. especially at the toolchip interface. the total calculation time to reach tool reshape criterion is reduced sharply because except the initial chip formation and chip growth are run only one time and then with the tool wear increasing.

Two-dimensional first-order four node diffusive element. DC2D4. the simulation will become complex because of the interaction between the cutting tool and the workpiece.2. because coupled thermal mechanical analysis is too expensive. 4. Therefore pure heat transfer analysis is performed after chip formation analysis for the further cutting process in order to get such knowledge at a low calculation cost. Diffusive elements are provided in one. only a single object is considered. temperatures at nodes inside the cutting tool are still climbing while those at tool-chip interface nodes approach steady state. diffusive elements.2 General Considerations 4.1 Introduction When the cutting process is simulated using chip formation analysis. to model convective heat transfer. which have only temperature degrees of freedom. For milling operation. Interpolation can be first-order and second-order [ABA-01b]. at the end of the chip formation analysis in turning operation. two or three dimension. is chosen to discretize the geometry of the studied object in the heat transfer analysis because quantities of DC2D4 are integrated at nodes and this simplifies the design of heat flux subroutine by importing heat flux at nodes of chip formation model directly into integration points of heat transfer model as basic data for the calculation of the current heat flux. the error caused by the conversion from nodal value to integration point value is avoided during the importation of temperature data. in the former chapter. It is concerned how the temperature distributes in the cutting tool finally. for example only the cutting tool or the workpiece.Heat Transfer Analysis In Metal Cutting 64 Chapter 4 Heat Transfer Analysis In Metal Cutting 4. For example. In addition.1 Geometry And Mesh In the heat transfer analysis. the chip formation analysis is only carried out in the first milling cycle. Otherwise. the cutting time is normally limited to a short time. ABAQUS uses some Eulerian elements. It is important for the correct calculation of tool wear how the tool temperature changes in the further milling cycles. These advantages are based on the conservation of .

Therefore conductive heat flux is temperature dependent. Both heat flux components are varying from node to node and the basic nodal heat flux data can be obtained from the chip formation analysis. 4. f gives the fraction of the generated heat flowing into the workpiece.2. Conductive heat flux is caused by the temperature difference of tool-chip and toolworkpiece at the interface. k is the gap conductance. q c = k (θ A − θ B ) where (4. 4. element label and element connectivity of the chip formation analysis model. . The total heat flux is composed of frictional heat flux q f and conductive heat flux q c .2) q c is the conductive heat flux crossing the interface from point A on the workpiece to point B on the cutting tool. ν s is the sliding velocity. Frictional heat flux is created due to the sliding friction between the workpiece material and the tool face. η specifies the fraction of mechanical energy converted into thermal energy. The amount of frictional heat flux into the cutting tool is calculated by Eq. sliding condition and contact with the tool face.2 Heat Flux In the cutting phase the cutting tool is heated by the heat flux acted on the tool-chip and tool-workpiece interface.1) τ is the frictional stress. 4. Therefore frictional heat flux is influenced by chip form.1. q r = (1 − f )ητ ⋅ν s where (4.Heat Transfer Analysis In Metal Cutting 65 the node label.2. θ is the nodal temperature on the surface. It is governed by Eq.

4. At other nodes the initial temperature is set to room temperature.3.1 Modeling In order to study on the temperature distribution of the cutting tool at steady state.3 In Turning Operation 4. node label and element connectivity of this part in chip formation analysis steps remain unchanged. the total . which is surrounded by the rake face. flank face. it is found that the total heat flux (heat passing through the tool-chip and tool-workpiece interface per second) is changing as the cutting process continues. bottom face. the part of the cutting edge engaged in the cutting is located in the center part circled in Fig. It includes the part. and the element label. 4.1. heat transfer analysis is performed after the chip formation analysis finishes.Heat Transfer Analysis In Metal Cutting 66 4. One component of the total heat flux. the circled part is the part of the edge engaged in the cutting Temperature data at the end of the steady-state chip formation analysis step is imported and used as the initial temperature definition of the nodes in the highlighted part. the geometry model of the cutting tool used in the two-dimensional heat transfer analysis is the section created by intersecting the center area with a surface perpendicular to the edge. The part of the tool used in the former chip formation analysis steps is only the highlighted part.2.1 Geometry and mesh of the cutting tool used in heat transfer analysis. as shown in Fig. 4. In steady-state chip formation analysis step. Hole surface Heat convection Heat radiation Frictional heat Conductive heat Room temperature A Hole surface Rake face Bottom face Bottom face Rake face Flank face A Cutting area Flank face Fig. and the surface of the central hole. At the nodes on tool/chip interface heat flux is defined. Because in the orthogonal cutting experiment [Schm-02].

Heat Transfer Analysis In Metal Cutting

67

frictional heat flux, reaches steady state in a short time once the chip gets into contact with the tool face and restores to the steady sliding velocity, while another component, the total conductive heat flux, is always decreasing within the entire analysis period. At the end of analysis, as the temperature of cutting tool and workpiece become steady, the decreasing rate is becoming lower and lower and approaching a steady value.
800000 600000 400000 200000 0 -200000 -400000 -600000 0,0000 0,0003 0,0006 0,0009 0,0012
Total conductive heat flux Total frictional heat flux Total Heat flux

Heat flux [mJ/s]

Time [s]

Fig. 4.2 Heat flux at tool-chip and tool-workpiece interface in steady-state chip formation analysis step Observation of nodal temperature of workpiece nodes at tool-chip interface shows that the variation of nodal temperature at the end of the analysis is very small, as shown in Fig. 4.3. Therefore thermal steady state is assumed in workpiece material at the interface. Then the nodal conductive heat flux can be converted from Eq. 4.1 into Eq. 4.3
q c = k θ ( A , t s ) − θ ( B ,t s ) + k θ ( B ,t s ) − θ B

(

) (

)

(4.3)

where

t s is the time point at the end of steady-state chip formation analysis. Because of the
above assumption about steady nodal temperature of workpiece node at interface,

θ ( A,t ) is equal to and replaces θ A . The first part k (θ ( A,t ) − θ ( B ,t ) ) is nodal conductive
s s s

heat flux, which can be obtained from the end of steady-state chip formation analysis. The second part is the variation of nodal conductive heat flux and it is dependent on

Heat Transfer Analysis In Metal Cutting

68

the difference in temperature between current nodal temperature and the temperature at the end of steady-state chip formation analysis.

N134 N133 N129 N128

N127

Fig. 4.3 Nodal temperature at selected workpiece nodes Therefore nodal total heat flux can be expressed by Eq. 4.4.
q t = qtrs + qtcs + k θ ( B ,t s ) − θ B

(

)

(4.4)

where

q t is the total nodal heat flux; q ct s stands for k θ ( A,t s ) − θ ( B ,t s ) ; q r t s is the nodal frictional heat flux; q ct s and q r t s do not change after importation. Based on Eq. 4.4, a temperaturedependent heat flux subroutine is developed for the heat transfer analysis. In addition, the tool makes heat transfer with the environment through rake face and flank face by heat convection and radiation. The nodes on bottom face and hole surface always keep room temperature because of their contact with the tool holder and the screw. 4.3.2 Results & Discussion When maximum temperature change of 10K between two times of incrementation of heat transfer analysis is defined as steady state criterion, steady state is reached in

(

)

Heat Transfer Analysis In Metal Cutting

69

the whole tool in 2.2s.s Fig. 4.4 shows the progress of temperature at four selected nodes in the tool.

Node 48 Node 340

Node 436 Node 13

(a) Position of the selected nodes

(b) Temperature history

Fig. 4.4 Temperature history of nodes in the tool

(a) t=0s

(b) t=2.2s

Fig. 4.5 Temperature field (Kelvin) change of the tool in heat transfer analysis Fig. 4.5 shows that at the beginning of heat transfer analysis the high temperature region concentrates in a small area near the cutting edge, and after 2.2s this region extends to nearly one-third of the tool.

4. 4. Therefore the workpiece geometry is extended in the heat transfer analysis. . Part 1 is the remaining workpiece geometry after the chip is cut away in the chip formation analysis. In the cooling phase of the first milling cycle.1 On Workpiece 4.1. + part 1 part 2 = workpiece Fig.7(a) shows the temperature distribution at the beginning of heat transfer analysis. Part 2 is some additional workpiece material. it is attached to the bottom of part 1. the actual workpiece is very huge comparing with the small part of workpiece used in the chip formation analysis.6 Geometry and mesh of the workpiece in heat transfer analysis The initial temperature of part 2 is set to room temperature. Fig.6.1 Modelling Generally. Study on the development of temperature distribution in the cutting tool in multi milling cycles is important for the implementation of tool wear estimation. 4. Nodal temperature at the end of previous chip formation analysis step is imported and defined as initial temperature of part 1. as shown in Fig. temperature distribution and heat transfer with the cutting tool in the second milling cycle. It is composed of two parts.4. 4.Heat Transfer Analysis In Metal Cutting 70 4.4 In Milling Operation Heat transfer analysis is performed for both the workpiece and the cutting tool. whether the workpiece can restore to room temperature affects the chip formation.4.

5) q v is the heat flux due to convection.7 Temperature field (in Celsuis) change of the workpiece in heat transfer During the cooling phase.m 2 . θ 0 is the sink temperature. unit J s. . 4. i. Heat flux due to convection is calculated by q v = −h θ − θ 0 . the workpiece cools down due to heat convection and radiation through boundary. h is a reference film coefficient.   ) (4.6) q r is the heat flux due to radiation on a surface. θ is the temperature at a point on the surface.e. ε is the emissivity of the surface. room temperature.77ms Fig. σ is the Stefan-Boltzmann constant.Heat Transfer Analysis In Metal Cutting 71 Heat emitted to environment due to co nvec tio n and r ad ia t io n Heat conduction in the material (a) t=0ms analysis (b) t=38.°C . in addition to the heat conduction from cutting area to the whole workpiece bulk. where ( ) (4. Heat flux due to radiation to the environment is governed by q r = εσ  θ − θ z   where ( ) − (θ 4 0 4 −θ z  .

The heat in the workpiece is emitted to the environment. Solutions obtained from the chip . It is assumed that temperature increment of the workpiece in the real cutting experiment is smaller than the predicted because the huge workpiece used in reality supplies a higher heat capacity and a bigger boundary surface to emit the heat. θ 0 is the ambient temperature. temperature distribution in the workpiece and tool-chip and tool-workpiece contact. θ z is the value of absolute zero on the temperature scale being used. 4.4. 4.77ms of cooling. It is expected that if the influence of the temperature variation of the cutting tool is not considered. Observation on the temperature progress at several selected nodes shows that after the first milling cycle the temperature increment is smaller than 10K.8 Temperature history of workpiece nodes in heat transfer analysis step It is assumed that in the second milling cycle this small temperature variation in the workpiece has no big influence on material deformation.Heat Transfer Analysis In Metal Cutting 72 θ is the temperature at a point on the surface.2 Results & Discussion The heat transfer analysis covers the rest period of the first milling cycle after the chip formation analysis ends. Node 839 Node 638 Node 501 Node 550 Node 464 Node 2140 (a) Monitored nodes (b) Temperature progress at monitored nodes Fig. Fig. chip formation analysis result in the second milling cycle can be assumed similar to that in the first milling cycle. heat generation. 4. as shown in Fig.1.8.7(b) shows that after 38. 4. the entire workpiece restores nearly to room temperature.

they are useable in the third.Heat Transfer Analysis In Metal Cutting 73 formation analysis in the first milling cycle can be used to the second milling cycle. Although the real cutting tool moves continuously with the rotation of the shaft. Heat transfer analysis starts from the time when the chip formation analysis ends. Therefore solutions obtained from the chip formation analysis of the first milling cycle are used in the heat transfer analysis of the cutting tool in multi milling cycles and tool wear estimation. it is possible to obtain the basic values of nodal frictional heat flux and nodal conductive heat flux from the chip formation analysis of the first milling cycle only at some discrete time points 0.2 On Tool 4. fourth. the tool is heated in the cutting phase by the heat flux at the tool-chip and tool-workpiece interface. 4.1 Modeling In the cooling phase of milling operation. The tool geometry and mesh in chip formation analysis are inherited and used in the heat transfer analysis. tn. tj.2. if the heat in the tool is not emitted completely to the environment by heat convection and heat radiation. The heat transfer analysis is performed in 8 milling cycles. This part will try to analyse the temperature variation of the tool with the accumulation of heat. …. In every milling cycle. These data are written in the heat flux and temperature files. The two components of the total heat flux. The temperature distribution at the end of the chip formation analysis is imported into the heat transfer analysis as initial conditions. the temperature of the tool will get an increment in the later milling cycle due to the remaining heat. Although the nodal total heat flux in the cutting phase is changing continuously from time to time. The heat flux value at other time point is obtained by performing interpolation. . According to the same reasoning.4.4. and further milling cycles. Frictional heat flux changes because of varying shear stress and sliding velocity caused by the change of chip thickness in milling operation. Conductive heat flux changes with the varying of difference in temperature between the tool and the workpiece at contact interface. frictional and conductive heat flux are time-dependent varying. in the simulation it is fixed spatial because the degree of freedom in the heat transfer analysis is limited only to temperature. …. t1.

k is the gap conductance. conductive heat flux is temperature dependent.5 × θ (bi . the time. dqic dθ i is given the value –k. j ) + q (ti . j −1) − θ i − 0. a heat flux subroutine DFLUX is designed to create timeand temperature-dependent nodal heat flux data.Heat Transfer Analysis In Metal Cutting 74 In addition. By finding out the remainder of the current time divided by the period of one milling cycle and comparing this remainder with the time points. i is the nodal label. j −1) × k ( ) ( ( )) (4. The subroutine first finds out the corresponding nodal label because only nodal label is used in the heat flux and temperature files. Every time when the subroutine DFLUX is called. element number. q t is the basic nodal total heat flux (the sum of the nodal frictional heat flux and the nodal conductional heat flux). the current nodal total heat flux is set to zero. . the rate of change of the current nodal total heat flux with respect to the temperature. face number and integration point are entered as input variables. the nodal total heat flux is calculated by qic = 0. θ is the current nodal temperature. the nodal conductive heat flux value will change due to the varying of the difference in temperature between the cutting tool and the chip or the workpiece. If the cutting tool is in the cooling phase. When the nodal temperature of the cutting tool in the later milling cycles is higher than that in the first milling cycle due to the accumulation of heat.5 × q (ti . j is the time point number. Based on these analyses. in order to improve the convergence rate during the solution of non-linear equations in an increment.7) where q c is the current nodal total heat flux. In addition. j ) + θ (bi . Otherwise when the tool is located between the time point j-1 and j. θ b is the basic nodal temperature. Then the basic nodal total heat flux and temperature values at all time points are read from the heat flux and temperature files. the interval and the two time points at the end of the interval is determined.

Considering the movement of the tool. forth and eighth milling cycle. When the tool face node has no contact with the chip and the workpiece. (a) At the end of chip formation analysis (b) After 4 milling cycles . the gap conductance is set to zero.2 Results & Discussion Fig. The high temperature region is widening as the milling process continues.9 shows the temperature distribution of the cutting tool when it cuts out of the workpiece 0. 4. a high reference film coefficient is defined in the model.3ms in the first. It is similar to the value used in the chip formation analysis. The contact status of the tool face node at a time point is derived from the value of normal pressure at the corresponding time point in chip formation analysis of the first milling cycle.2.Heat Transfer Analysis In Metal Cutting 75 The gap conductance in the heat transfer analysis is 10000 when the tool face node is in contact with the chip and the workpiece.4. 4.

4. Both the peak and valley value are increasing with the cutting process continuing.10 and Fig. 4.10 the nodal temperature at the tool face nodes on the too-chip and toolworkpiece contact interface increases in cutting phase and decreases in cooling phase. 4.Heat Transfer Analysis In Metal Cutting 76 (c) After 8 milling cycles Fig. including both chip formation analysis step and heat transfer analysis step. Node4 Node42 Node35 Node39 Node57 (a) Position of the selected nodes (b) The temperature history of the nodes Fig.11 show the variation of nodal temperature in the cutting tool in more than 8 milling cycles. 4.9 Temperature field (in Kelvin) progress of the tool in heat transfer analysis Fig. The peak value of temperature in every milling cycle appears when the cutting tool is cutting out of the workpiece.10 Progress of nodal temperature on the tool face In Fig. The valley value appears when the cutting tool is going to enter into the workpiece. but the increments are decreasing and the increment of the peak value is smaller than 1K after every milling cycle while the increment of . 4.

Higher temperature is expected in the further milling cycles. When heat loss becomes equal to heat gain. 4. At the nodes close to the cutting area.11. 4.11 Progress of nodal temperature inside the tool Both the nodal temperature of nodes inside the tool and on the tool face shows that cyclical thermal balance state is not attained in the first 9 milling cycle and heat gain is greater than heat loss in every milling cycle.2.5K. the peak value can be assumed to become steady because the increment is smaller than 0. N127 N135 N210 N13 N3 (a) Position of the selected nodes (b) The temperature history of the nodes Fig. In the last milling cycle. whereas the increment of the valley value is still greater than 5K.3 Application Of Preheated Cutting Tool According to the analysis above. Inside the cutting tool a different progress tendency of nodal temperature is observed. nodal temperature is always increasing during the entire milling cycles. It is very difficult to analyse the cyclical thermal balance by only manually adding more milling cycles in heat transfer model file because the number of milling cycles to reach cyclical thermal balance state is unknown. In order to reduce the number of milling cycles to reach cyclical thermal balance state and speed up the calculation process. the whole cutting tool is preheated beforehand by . At the nodes far from the cutting area.Heat Transfer Analysis In Metal Cutting 77 the valley value is much great. for examples N210 and N3 in Fig. nodal temperature increases in cutting phase and decreases in cooling phase. tool temperature increases due to accumulation of remaining heat. 4.4. Heat loss increases with the tool temperature. cyclical thermal balance state is attained.

After a time of cutting. nodal temperature on the tool face and inside the tool is increasing after every milling cycle. The explanation is that at the beginning of the analysis the entire workpiece has a same temperature value. but the increment is smaller than that in the former analysis. Only 8 milling cycles are included in the heat transfer analysis. no heat conduction takes place in the vicinity of these nodes. (a) (b) Fig. In Fig. milling process is analysed by performing only heat transfer analysis. for example N4 and N42. heat convection to the environment makes the temperature decrease. node 3 and node 210.12.Heat Transfer Analysis In Metal Cutting 78 defining a high initial temperature. temperature decreases first and then increases again. Analysis 1: Preheated to 600K In this analysis. the cutting tool is preheated to 600K. At the farthest nodes away from the cutting area. This is explained by the heat conduction between the tool with higher temperature and the workpiece with room temperature. valley value of nodal temperature appears at the time when the tool comes into contact with the workpiece instead of before the contact takes place. Nodal temperature at the same tool nodes as in the former heat transfer analysis is observed. for example. the workpiece material in the vicinity is heated by the heat . 4.12 Preheated to 600K (a) The temperature history of the nodes inside the tool (b) The temperature history of the nodes on the tool face At the nodes on tool-chip interface. these nodes are located at the boundary. 4. then it is used in the milling process. Because at present only tool temperature is concerned.

13 Preheated to 700K (a) The temperature history of the nodes inside the tool (b) The temperature history of the nodes on the tool face Analysis 3: Preheated to 650K According to analysis 1 and analysis 2. the cutting tool is preheated to 700K. Therefore.13. temperature decreases in the entire 8 milling cycles.Heat Transfer Analysis In Metal Cutting 79 generated in the cutting process and these nodes are heated because of heat conduction. In Fig. when the tool preheated to 600K is used in milling operation. The analysis shows that cyclical thermal balance state is not attained in the 8 milling cycles because the workpiece is heated too high and heat loss is greater than heat gain in every milling cycle. Higher tool temperature is expected in cyclical thermal balance state. 4. 4. in this analysis. (a) (b) Fig. nodal temperature on the tool face and inside the tool is decreasing after every milling cycle and the decreasing rate is comparable to the increasing rate in analysis 1. tool temperature increases still and no cyclical thermal balance state is attained. At node 3 and node 210. cyclical thermal balance state is expected to . Nodal temperature at the same tool nodes is observed. when the cutting tool is heated to a temperature between 700k and 600K. Analysis 2: Preheated to 700K According to analysis 1.

nodal temperature on the tool face and inside the tool is approaching cyclical thermal balance state. When the cutting tool move to the same position in the 7th and the 8th milling cycle. in this analysis.14 Preheated to 650K (a) The temperature history of the nodes inside the tool (b) The temperature history of the nodes on the tool face Nodal temperature is monitored at the same tool nodes. it can be assumed that cyclical thermal balance state is realized in the 8th milling cycle.Heat Transfer Analysis In Metal Cutting 80 realize in the first 8 milling cycle. (a) (a) (b) (b) Fig. the maximum difference in temperature at the same node is smaller than 0. In Fig. the cutting tool is preheated to 650K. 4. 4.14. .1K. Therefore. Therefore.

5 Summaries & Conclusion ABAQUS/Standard is effective in heat transfer analysis.Heat Transfer Analysis In Metal Cutting 81 4. the heat transfer analysis is performed for the cutting tool until the thermal steady state is reached. . thermal steady state in the turning operation and cyclical thermal balance state are analysed. In milling operation. By using preheated cutting tool in the heat transfer analysis. the heat transfer analysis is performed for the cutting tool for several milling cycle. the cyclical thermal balance state is analysed. In turning operation. By introducing the heat flux and temperature distribution from the chip formation analysis output file and using user-developed heat flux subroutines.

In this chapter. will be studied in the next chapter. and various cutting process variables will have no great change and steady state can be assumed. Base on the obtained experience. accesses the created result and output database files once the analysis jobs are finished. the entrance and exit of cutting tool takes place infrequently and takes only a short time. the study will focus on the modelling of tool wear in turning operation. if the effects of tool wear and uneven distributions of workpiece material are neglected. Tool wear calculation can be simplified by assuming that tool wear is created completely by the steady state cutting process and neglecting the effect of entrance and exit phase. tool wear estimation is implemented. The program controls the submission of chip formation and heat transfer analysis jobs. 5.1 Introduction The following two chapters will describe the modelling of progressive tool wear in turning and milling operation by developing user program with programming language Python and integrating it with commercial FEM code ABAQUS/Explicit and ABAQUS/Standard. the more complex modelling problem. tool wear in milling operation. All the tool wear simulation models will be developed for two-dimension. cutting thickness.2 Tool Wear Calculation Program Design Fig. Solution to the problems met in 2D modelling will be helpful for the implementation of 3D modelling in the future. 5. The program is designed to perform tool wear calculation automatically cycle by cycle until a tool reshape criterion is reached. chip shape. performs tool wear calculation and modifies the related model files according to the calculated tool wear. In continuous cutting process. In every calculation cycle. It is performed with a tool wear estimation program. By integrating tool wear mathematical model with the finite element steady-state cutting analysis.1 shows the flow chart of the tool wear calculation program. Tuning operation is characterized by continuous cutting process. monitors their analysis process.Estimation Of Tool Wear In Turning Operation 82 Chapter 5 Estimation Of Tool Wear In Turning Operation 5. chip formation and heat transfer analysis jobs are submitted to analyse the steady-state cutting process and obtain the cutting process variable values necessary for the calculation of wear rate .

Start Chip formation analysis Heat transfer analysis Next calculation cycle Nodal wear rate calculation Nodal displacement calculateion Cutting time increment calculation Tool geometry updating No VB>=VBmax? Yes End Fig. Then the nodal displacement due to wear in the cutting time increment is calculated at every tool face node.3 Modeling Procedure During the explanation of the entire modelling procedure. .Estimation Of Tool Wear In Turning Operation 83 at steady state. Based on the calculated nodal wear rate. Nodal wear rate is calculated by using the tool wear mathematical model.5. If the produced flank wear VB is smaller than the user-defined tool reshape criterion VBmax. a second tool wear calculation cycle starts with the updated tool geometry. a suitable cutting time increment is searched by program according to a user-specified VB increment value.1 Flow chart of tool wear calculation program 5. the tool wear under the same cutting condition as in the chip formation simulation of turning operation is estimated. and the tool geometry is updated according to the calculated nodal displacement.

as observed by Childs and Mahdi [Chil-89] when turning mild steel.02mm away from the tool tip are moving out of cutting area with the chip.3.1.3. Fig. There is no contact between the chip and the tool face in the area beyond the distance of 0. Then the distribution exhibits a plateau of high stress near the tool. 5. the normal pressure drop off sharply. 5. Therefore the sliding velocity becomes zero. therefore the normal pressure drops to zero. This means that they are flowing into machined surface.1. . 5.1 Normal Pressure Fig.35mm.3.2 Normal pressure of the tool face nodes at tool-chip interface at steady state 5.2 Sliding Velocity In Fig.2 shows the variation of normal pressure at the tool face nodes along the toolchip interface at the end of chip formation analysis. the chip loses contact with the tool face. At the tool tip area.3(a) workpiece nodes in the tool tip area have negative relative sliding velocities. as designed by ABAQUS [ABA-01a]. At the distance of about 0. The nodes that are more than 0.35.5.1 Chip Formation And Heat Transfer Analysis 84 Chip formation analysis provides the mechanical variables at steady state and the thermal variables at steady state are predicted by heat transfer analysis. Beyond the feed distance. the normal pressure has the maximum value.Estimation Of Tool Wear In Turning Operation 5.

Not all the tool face nodes and workpiece nodes are in contact. 1e6Mpa. as shown in Fig. yj) Tool : (y (x j − yi ) y i +1 − y i − xi ) (v (v s i +1 − vis ) ) Workpiece node i+1 (xi+1. 5.Estimation Of Tool Wear In Turning Operation 85 (a) Sliding velocity at the workpiece nodes at steady state Workpiece node i (xi. Then they are arranged in counter-clockwise order. yi+1) Else: v sj = vis + j x i +1 − x i s i +1 − vis Workpiece (b) Calculation of sliding velocity at the position of tool face nodes Fig. Every tool face nodes in contact has two neighbouring workpiece nodes before and after it. for example. But when calculating nodal wear rate. it is necessary to know the sliding velocity value of workpiece material at the position of tool face nodes.3 Calculation of sliding velocity at the position of tool face nodes Only sliding velocities at the position of workpiece nodes can be obtained directly from the simulation. . yi) If y i − y i +1 ≥ xi − xi +1 v sj = vis + Tool face node j (xj. First all the tool face nodes and workpiece nodes in contact are found out depending on whether the absolute values of their normal pressure are greater than a critical value. 5.3(b). The calculation is performed based on their position relationship.

as shown in Fig. The constants in Usui’s equation for the combination of carbide cutting tool and mild steel are shown in Table 1. High temperature forms at the tool tip and a distance from the tool tip on rake face. is employed in the calculation.2.3. Because at low cutting speed. described in Chapter 1.2 Wear Rate Calculation After the cutting process variables. Fig. wear rate at the position of every tool face node is calculated by using wear mathematical model. The nodal move direction is calculated at every tool face node.3.1. Usui’s model.4 Temperature of tool face node at steady state 5.4.Estimation Of Tool Wear In Turning Operation 86 5. tool temperature and normal pressure at every tool face node are obtained. The latter one is adopted in this paper. the flank wear and crater wear are assumed to be created mainly by abrasive wear and adhesive wear. . 5. sliding velocity of the workpiece material.3 Tool Temperature Temperature of the tool face nodes at thermal steady state is obtained from heat transfer analysis.3. 5. 5.3 Nodal Move Direction Tool wear expression in geometry can be realized with two approaches: element deletion and nodal movement.

It moves along the negative direction of the average unit normal vector of the two face segments or the negative direction of their resultant vector. Tool face node belongs to two face segments. will become the new dividing node.3. the real dividing node is found out. After one searching cycle.5 is searched by the program. Points on the face segment should move along the negative normal direction. It is saved as the tool edge position for the latter calculation of flank wear land width.2 On Rake Face In the rake face part nodal move direction is assumed to be perpendicular to the relative sliding velocity of the workpiece material and pointed into the tool body. Then the current dividing node is compared with all the tool face nodes. On every tool face segment the sliding velocity is along the tangential direction. as shown in Fig. It divides the entire tool face into flank face and rake face.Estimation Of Tool Wear In Turning Operation 5. all the tool face nodes are found out and arranged in counter clock-wise order in a list. 5. a dividing node. The first tool face node in the list is given to the dividing node.3. It has the minimum y-coordinate.3.1 Dividing Node 87 Before calculating the nodal move direction. the circled node in Fig. 5. 5.5. Every tool face node is attached with two tool face segments. . Before searching. whose y coordinate is smaller than the current dividing node by 2e-4mm.3. On flank face and rake face nodal move direction is calculated with different methods. Any node.

3.Estimation Of Tool Wear In Turning Operation 88 Rake face part Node i (xi. 12 i +1 − xi ) + ( yi +1 − yi ) 2 xi − xi +1 2 )   12    r yi + 2 − yi +1 n j +1 =   ( x − x )2 + ( y − y )2 i+2 i +1  i + 2 i +1 ( ) ((x .1) Fig. nodal move direction Di is calculated by r  yi +1 − yi nj =   ( x − x )2 + ( y − y )2 i +1 i  i +1 i ( ) ((x . nodal move direction Dk = (0. where subscript i is nodal label. It is in ydirection and pointed upwards. 5.yi+1) r nj r n j +1 Node k n2 Face segment j+1 Node i+1 (xi+1.3. . 12 i+2 − xi +1 ) + ( yi + 2 − yi +1 ) 2 xi +1 − xi + 2 2 )   12   r r r nir = n j + n j +1 r r r Di = − nir nir r On flank face.3 On Flank Face In the flank face part the relative sliding velocity can be assumed to be in the cutting speed direction when the elastic recovery of workpiece material is neglected. j ) .5 Nodal move directions (thick arrows) of tool face nodes 5. j is the calculation cycle number. r Every nodal move direction is normalized to unit vector D( i .yi+1) Dividing node Flank face part r On rake face. Therefore all the nodes in this part have the same nodal move direction.yi) Face segment j Node i+1 (xi+1.

for example. Since the nodal wear rate is already known.6 Flank wear calculation and cutting time increment searching process . Therefore a suitable cutting time increment should be given.05 Aimed VB value range Example of a wear curve a1 b1 c1 Edge a b c position d ∆t 0 ∆t1 ∆t ∆t 2 Time [s] (a) Flank wear calculation (b) Cutting time increment searching process Fig. in which the specified tool wear increment is produced. wearrate) VB2 VB 0. Flankwear (∆t . But it is easier for the user to specify a tool wear increment. wearrate) . In order to reach tool reshape criterion. VB [mm] Flankwear (∆t .Estimation Of Tool Wear In Turning Operation 5. can be searched by program. Therefore.3. only small tool wear increment is produced in every calculation cycle. In the simulation the calculation of tool wear and the tool geometry updating are based on a certain cutting time increment. it is very time-consuming. Therefore a searching module is designed to carry out the searching work. when novel workpiece material is machined.4 Cutting Time Increment Calculation 89 In metal cutting experiment cutting time increment means the duration of cutting time between two successive measurements of tool wear. it is difficult to define a suitable value. many calculation cycles have to be performed. a flank wear calculation subroutine. if a big cutting time increment is specified. is called frequently. The chip formation analysis is carried out in every calculation cycle. a big error will be created during the calculation of tool wear.10 VB1 VB0 VB 0. the cutting time increment. Within the cutting time increment an unchanged nodal wear rate value is used to calculate the tool wear. 5. While the suitable cutting time increment is being searched. But when there is no knowledge about the tool wear in the simulated cutting conditions. But if the cutting time increment is too small.

6(b).05mm is specified by the user. e. 5.3.1 Flank Wear Calculation Subroutine 90 The flank wear calculation subroutine Flankwear (∆t . because it is the last moved tool face node. Then the searching process starts.6(a). in Fig. For example. Therefore node b and c should move to point b1 and c1 in order to have the same y-value with point a1. In order to save the searching time. the aimed VB value should be given a permitted error range. VBm is 0.4. During the searching process. 5. For example. wearrate) calculates flank wear land width VB. it should move to point a1. then node b and c will have smaller y-values than point a1. the tool gets a flank wear land width of 0. VB is the distance from the cutting edge position (which has been saved) to the last moved tool face node.6(b). ∆VB = 0.3.05mm from the previous tool wear calculation cycle. the searching lower limit ∆t1 and the searching upper limit ∆t2 are changing until the calculated tool wear VB value under the cutting time increment ∆t falls into the aimed VB value range. 5. At the beginning the aimed VB median value VBm is calculated according to the userspecified VB increment value.1mm. in Fig. In cutting time increment ∆t . VB is calculated from the cutting edge position to node c. But in practice the wear process is continuous.Estimation Of Tool Wear In Turning Operation 5. they are also worn away and no bulge is formed. Therefore in this tool wear calculation cycle. Once node b or c comes into contact with workpiece material due to wear of the cutting edge.4. node a is the last tool face node with non-zero wear rate. 5. In addition. and a bulge will be formed on the flank face.g.2 Cutting Time Increment Searching Procedure The cutting time increment searching procedure can be described by Fig.7. a positive initial cutting time increment value ∆t 0 is given arbitrarily. 5. the dotted range in Fig. .

i is nodal label.7 Flow chart of cutting time increment searching procedure 5.1) .3. error=VB.5 Nodal Displacement Nodal displacement due to wear is calculated at every tool face node by r r & w( i .VBm ∆t=2∆t No |error|<=δ? Yes ∆t2>∆t 1? ∆t2=∆t No No error<-δ? Yes Yes ∆t1=∆t Current cutting time increment ∆t is output End Fig. calculate the current error. j ) = w(i . (5. 5. initial value ∆t0. permitted error δ Cutting time increment ∆t. the searching lower limit ∆t1 and the searching upper limit ∆t2 are set to ∆t 0 ∆t=(∆t1+∆t2)/2 Call subroutine Flankwear(∆t. j ) ⋅ ∆t j ⋅ D(i .wearrate) to calculated the current flank wear VB.Estimation Of Tool Wear In Turning Operation 91 Start VBm. j ) where r w is the nodal displacement vector.

Remeshing improves mesh distortion and enables additional nodal movement of the tool face nodes in the next times. 92 In addition. Then the mesh inside the tool is remeshed with one of the smoothing methods: volume smoothing.8. It is accomplished with two steps. an initial tool wear profile appears on the cutting tool. 5. 5.1 Step 1: Initial Tool Wear Profile In the first step the tool face nodes.3. In these two steps nodes on the tool bottom surface.6. 5. marked with small triangles in Fig. 5. Laplacian smoothing and equipotential smoothing. In the following part.Estimation Of Tool Wear In Turning Operation j is tool wear calculation cycle number.6 Tool Geometry Updating In order to visualize the tool wear profile and prepare tool geometry model for the next tool wear calculation cycle. The entire movement is accomplished several times. the circled nodes in Fig. Every time the tool face node is moved a very small distance.9(b). some nodes on flank face must be moved in order to avoid forming bulge on flank face. After the first step.8. volume smoothing is employed because of the robustness. are moved according to the calculated nodal displacement. including the nodes on rake face and flank face.3. tool geometry updating is performed. as shown in Fig. 5. . as mentioned above. or their combination. are fixed spatial.

zone B in Fig. .3. the predicted distributions of cutting process variables along the tool face often contain ‘vibration’. the mesh inside the cutting tool has been remeshed many times in step 1. But the tool face nodes are moved according to the calculated nodal displacement without any additional adjustment of nodal position.9(b). e. In addition. Fig. Sometime very fine mesh is formed in the cutting edge area.g. The final tool wear profile and tool geometry is produced by step 2. 5. 5.9(b).9(c) shows that in the second step. zone A in Fig.2 Step 2: Adjustment Because of the contact problem on the tool-chip and tool-workpiece interface caused by the coarsened workpiece mesh in chip formation analysis. 5. These results in zigzags of the initial tool wear profile.6. The tool geometry model file is updated according to the produced result in step 2.8 Boundary conditions in step 1 of tool geometry updating model 5. zigzags of the crater wear profile are smoothened and the mesh near the cutting edge is coarsened. for example. they make tool geometry updating in the next calculation cycle difficult because negative element areas may be created by the nodal displacement due to additional produced tool wear. 5.Estimation Of Tool Wear In Turning Operation 93 Tool bottom Rake face Tool bottom Flank face Fig.

5. the new tool in Fig.10(d). After the second calculation cycle. Tool reshape criterion is set to 0.4.05mm is specified by user. After the first calculation cycle. 5. 5. 5. 5. The tool wear estimation process is accomplished with three tool wear calculation cycles.02mm.1 is calculated.9 Changes of the mesh during tool updating steps (a) The tool geometry and mesh at the beginning of step 1 (b) At the end of step 1. permitted error δ in the cutting time increment searching process is set to 0.1 Tool Wear With this tool wear estimation program.4 Results & Discussion 5. . and tool wear increment ∆VB = 0.Estimation Of Tool Wear In Turning Operation 94 (a) Zone A Zone B (b) (c) Fig. zigzags of the crater wear are smoothened.10(b) is updated to the worn tool in Fig. nodes on the tool face are moved according to the calculated nodal displacement. tool wear progress under the same turning cutting conditions as described in Table 3.10(c). crater wear and flank wear are formed (c) At the end of step 2. increased crater wear and flank wear can be found on the updated tool in Fig.15mm.

1 0.02 0 200 Flank wear width VB [mm] 0.04 0.05 M easured Crater wear depth [mm] M easured Est imat ed Est imat ed 0 0 50 100 t [s] 150 0 50 100 t [s] 150 200 (a) Flank wear (b) Crater wear Fig.10 Tool wear profile progress 0.1 0. The dot . 5.11 shows the wear progress curves of flank wear and crater wear obtained from experiment [Schm-02] under the same cutting condition. 5.08 0.11 Comparison between estimated and experimental progress curves for tool wear (under cutting condition: vc=300m/min.15 0.06 0.145mm/r) The solid line in Fig.2 0. ap=2mm.Estimation Of Tool Wear In Turning Operation 95 t=187s t=46s t=0s t=5s (a) (b) t=0s (c) t=5s (d) t=46s (e) t=187s Fig. 5. f=0.

Estimation Of Tool Wear In Turning Operation 96 lines are predicted tool wear curves. In the chip formation analysis. Coulomb’s friction model is adopted and a constant coefficient of friction 0. It is found that the estimated flank wear and crater wear are smaller than experimental ones. the estimated flank wear just arrives at 0. • inconsistentness of material combination. This maybe means that the predicted variables for the calculation of tool wear have error as well. It was tested by Kitagawa et al that the content and size of abrasive particle dispersed in workpiece material and chemical composition of tool material could be correlated with change in the constants of the wear characteristic equation both in higher and lower temperature ranges [Kita-88]. the predicted cutting force and thrust force are smaller than the experimental data by about 15% and 35% when the coefficient of friction is set to 0. Because the characteristic equation of tool wear and the tool wear data come from different literatures and researchers. it is expected that in the later tool wear estimation.3 is used in the whole tool wear estimation process. This may be caused by the poor contact between flank wear and the workpiece. It is observed that the temperature on flank wear drops off to a low value. From Fig. after a certain tool wear is formed.08mm. .15mm and crater wear 0. In order to improve the contact. after 20s of cutting.14mm and crater wear 0. the wear rate on flank face decreases more than that on rake face. some nodes should be adjusted to form a small negative flank angle. the coefficient of friction should be calculated according to the cutting force and tool geometry or with a more reliable method. Therefore chip formation modeling is very important for the accuracy of tool wear estimation. The discrepancy may be caused by: • the simplified and low coefficient of friction. 5. while after 187s.3. In order to improve the prediction. it is unavoidable that difference exist in these tool and workpiece material’s chemical composition and structure. maybe on the flank wear face.06mm. both wear rate on flank face and on rake face are decreasing. According to the verification of chip formation analysis in continuous chip formation. In experiment.11. the flank wear has exceeded 0. • contact problem between flank wear and the workpiece.

Then displacement of every tool face node due to wear is calculated by calculating nodal wear rate at steady state. Because of the huge calculation time and cost of chip formation analysis. A trade-off value should be found.Estimation Of Tool Wear In Turning Operation 97 5. The number of calculation cycles carried on before Python user program stop is defined by dividing tool reshape criterion by the specified wear increment. The main findings of this study are as follows: (1) Python user program launches chip formation and heat transfer analysis job automatically every time the new value about cutting process variables at steady state are needed. . Finally tool geometry is updated according to the calculated nodal displacements and one calculation cycle is finished. regenerate workpiece model when chip shape has a great change due to tool geometry change caused by serious tool wear and some modification of the flank wear shape in order to improve the contact between flank wear and workpiece material. a bigger wear increment is preferred in order to reduce the calculation cycle number. which certainly will bring bigger errors in estimated result. more efforts should be made in several aspects: more reasonable frictional modelling. (2) The Python user program runs automatically until a tool reshape criterion is reached. searching a suitable cutting time increment by program and nodal displacement calculation.5 Summaries & Conclusion In this chapter 2-D tool wear estimation in orthogonal cutting of turning operation is implemented by integrating ABAQUS/Explicit and ABAQUS/Standard with Python user-program. (3) In order to improve the estimate result and realize tool wear estimation in quantity. further mesh control.

mechanical variables in one milling cycle can be assumed to be repeated in other milling cycles because of the same cutting path of the cutting insert .1. strain. In the cutting phase. the cutting process possesses periodicity. workpiece moves in feed direction.1 Feature of milling operation Although milling operation has no steady state. For example. 6. Therefore every milling cycle comprises cutting phase and cooling phase. At the exit the cutting thickness becomes zero.1 Introduction Cutting action in milling operation is different from turning operation. cutting thickness has the maximum value when the cutting insert advances into the workpiece. the cutting insert cuts away a layer of workpiece material and then cools down in the environment. stress. Then the cutting thickness decreases continuously. nearly all the cutting process variables or solutions.Estimation Of Tool Wear In Milling Operation 98 Chapter 6 Estimation Of Tool Wear In Milling Operation 6. In every rotation/milling cycle. uneven distribution of the workpiece material. etc are neglected. With the cutting tool rotating. If the effect of tool wear. fz Cooling phase w Tool Node i vf Workpiece Cutting phase tk+1 tk Fig. cutting thickness varies with tool engage angle. temperature. etc are related with the cutting thickness. According to metal cutting theory. in the down milling operation in Fig. 6. Therefore they change with the tool engage angle and so does nodal wear rate.

Periodicity of cutting action and the existence of the cyclical thermal balance state enable the implementation of tool wear estimation in milling operation. tool wear and nodal average wear rate per cycle do not change from cycle to cycle.2 Tool Wear Calculation Program Design Fig. Therefore the implementation of tool wear estimation in milling operation can be simplified by calculating nodal average wear rate per cycle in one milling cycle of cyclical thermal balance state and then using it in other milling cycles. In the milling operation with a long continuous milling path. Before the cyclical thermal balance state is reached. the tool temperature obtained in one milling cycle can stand for that in other milling cycle only when the cyclical thermal balance state (heat loss is equal to heat gain per cycle) is attained. Once cyclical thermal balance state is attained. The tool temperature obtained in one of these milling cycles is always lower than that in the milling cycle of the cyclical thermal balance state. heat loss is smaller than heat gain per milling cycle and the tool temperature increased after each milling cycle. Mechanical cutting process variables. a lower estimated value of the tool wear is expected.2 shows the flow chart of the tool wear calculation program. 6. the cyclical thermal balance state dominates the entire cutting process. As tool temperature is concerned. .Estimation Of Tool Wear In Milling Operation 99 and the same change of undeformed chip thickness. normal pressure on tool face and relative sliding velocity of workpiece material on the tool face. The tool wear is mainly decided by the cyclical thermal balance state. If the tool wear is calculated according to the tool temperature in a milling cycle before the thermal balance state is reached. 6. obtained from the first milling cycle can stand for those from all other milling cycles.

a suitable cutting time increment value is searched according to a user-specified VB increment value. then the nodal average wear rate is calculated according to these nodal wear rate at the selected time points.Estimation Of Tool Wear In Milling Operation 100 Start Chip formation analysis Heat transfer analysis Next calculation cycle Nodal average wear rate per cycle Nodal displacement calculation Cutting time increment calculation Tool geometry updating No VB>=VBmax? Yes End Fig.2 Flow chart of the tool wear calculation program The tool wear calculation program is designed to perform tool wear calculation automatically cycle by cycle until a tool reshape criterion is reached. The nodal wear rate is time-dependent and calculated at some selected time points of one milling cycle. According to the above discussion on milling features. 6. In every calculation cycle. Then nodal displacement due to wear produced in the cutting time increment is calculated and the tool geometry updating aims at forming the tool wear profile on the tool face. chip formation and heat transfer analysis are performed to predict the cutting process variables. which are necessary for the wear rate calculation. mechanical variables can be obtained from the first milling cycle and the heat transfer analysis helps to decide from which milling cycle tool temperature is read for the calculation of nodal wear rate. Based on the calculated nodal average wear rate. If the produced flank wear VB is smaller .

sliding velocity at the position of tool face nodes are calculated at the time point when the calculation of nodal wear rate is required. Hence the time for outputting the variables is discontinuous. chip formation analysis covering the entire cutting phase is necessary and enough for the calculation of wear rate. 6. At the tool tip sliding velocity is very small. Sliding velocity is available at the workpiece node. Other nodes are flowing out with the chip at increasing sliding velocity. During the cutting phase.Estimation Of Tool Wear In Milling Operation 101 than the user-defined tool reshape criterion VBmax. . high frequency of variables outputting will result in a large output database file and increase the amount of calculation in the tool wear estimation. According to the sliding velocity at the position of workpiece nodes.2 is estimated. sliding velocity of workpiece material and normal pressure on tool face. 6. the tool wear under the same cutting condition as described in Table 3.1ms. The mechanical variables. finite element analysis of chip formation process is performed by advancing the time with small time increments. Because tool wear takes place only in cutting phase and there is no tool wear created in cooling phase. a second calculation cycle will start with the updated tool geometry. are required during the calculation of wear rate.1 Chip Formation Analysis Chip formation analysis aims at obtaining the mechanical variables for the calculation of nodal wear rate. Furthermore. even some nodes flow towards the machined surface before material failure takes place.3 Modeling Procedure During the explanation of the entire modelling procedure. Fig. 6. But with explicit method. For example. mechanical variables are varying from time to time.3(b) shows the sliding velocity at the time when the cutting tool engaging into the workpiece 0.3. Therefore the frequency of mechanical variables output should be decided by making a compromise between calculation accuracy and calculation cost.

No plateau is observed in the entire tool-chip contact area as in turning operation.100ms . For example.Estimation Of Tool Wear In Milling Operation 102 (a) The cutting at 0. Fig. 6.4 Normal pressure on the tool face at the time of 0.100ms (b) At t=0. 6.4 shows the normal pressure at the time of 0. From tool tip to the separation point of the chip and tool normal pressure is decreasing.3 Relative sliding velocity of workpiece material on the tool face at the time of Normal pressure at the position of tool face node can be obtained directly.100ms 0.1ms. Fig.100ms Fig. 6.

6.3. the tool temperature at the tool face nodes increase after every milling cycle until cyclical thermal balance state is reached.Estimation Of Tool Wear In Milling Operation 6. The tool temperature in the cyclical thermal balance state is read for the calculation of tool wear.1ms in three selected milling cycles. nodal temperature at tool face has a jump. The selected milling cycle are the first.3. which is obtained from the last milling cycle of heat transfer analysis of the tool preheated to 650K. the temperature has a relative small increment.3. 6.2 Heat Transfer Analysis 103 According to the heat transfer analysis.1ms 6.3 Nodal Average Wear Rate Calculation Usui’s model is employed in the calculation of nodal wear rate at a certain time.2. 6.1 Discussion About The Calculation Method Of Nodal Average Wear Rate . The wear characteristic constants in Usui’s equation for the combination of carbide cutting tool and mild steel are shown in Table 1. It is found that from the first to the ninth milling cycle. while from the ninth to cyclical thermal balance cycle.5 shows the nodal temperature on the tool-chip interface at the time when the tool is engaging into the workpiece 0.3.5 Tool temperature at the tool face nodes after the cutting insert advancing into the workpiece 0. Fig. Fig. the ninth and cyclical thermal balance cycle.

& w( i .2ms of every milling .Estimation Of Tool Wear In Milling Operation 104 Nodal wear rate varies with the cutting time. & At present it is very difficult to get the function of nodal wear rate w( i . j . j ) = t0 & ∫ w (t )dt (i . Ζ is the time span of one milling cycle. 6. j .27ms. & w(t ) is nodal wear rate. j ) (t ) . In cooling phase. as shown in Fig. Nodal average wear rate is calculated by t0 + Ζ & w( i .2) where n means that the entire milling cycle is divided into n-1 small portions by n evenly spaced time points. tool wear takes place under the contact of the tool with the workpiece. j ) Ζ . For example. sampling of cutting process variables and the calculation of nodal wear rate are not performed in the entire milling cycle because no wear takes place in cooling phase. In the cutting phase. (6. but the cutting phase only takes place in the first 0. nodal wear rate is equal to zero and no wear produced. Based on these nodal wear rate values. an approximate nodal average wear rate can be calculated by the following equation.1) where & w is the nodal average wear rate. the whole milling cycle may take about 39. nodal wear rate is calculated at every time point.k +1) ) ⋅ (t k +1 − t k ) ⋅ Ζ 1 2 (6. In the real calculation.k ) & + w( i . j is the milling cycle number. k is the time point number. j ) = & ∑ (w n 1 (i . But nodal wear rate values at some discrete time points can be obtained by sampling cutting process variables during chip formation and heat transfer analysis and then calculating the individual nodal wear rate values.6. i is the nodal label.

3. which have possibility of getting contact with the tool face.Estimation Of Tool Wear In Milling Operation 105 cycle. During the calculation of nodal average wear rate. 6. Tool temperature values are read at the corresponding time points 0. Chip formation modelling in turning operation is different from milling operation.5ms in the chip formation analysis. According to Chapter 4. … 0. the mechanical variables sliding velocity and normal pressure are read only at sampling time points 0.2 Classification Of Workpiece Node The sliding velocity of workpiece material at the position of tool face nodes are calculated using the same method explained in Chapter 5.k +1) & w(i . Since in milling operation shear failure criterion is defined in the whole workpiece. 0. are fixed on several nodes on the chip underside. j . In turning operation. … and 0.2.05. workpiece nodes. the meaning of n becomes the number of time points dividing the period of chip formation analysis. some nodes not on the moving path of the cutting edge may be exposed due to . The part of cooling phase from 0.025. the chip formation analysis includes the whole cutting phase and 0. j .3ms of cooling phase in the first milling cycle.025.5ms of the selected milling cycle of heat transfer analysis.6 Calculation of nodal average wear rate Then in Eq. Only these nodes are considered during the calculation. 6. j ) (t ) tk tk+1 Cutting time t [s] Cutting phase Cooling phase Cooling phase Fig.5ms to 39. Average nodal wear rate calculation is performed only at these time points. 6.3.27ms is not considered.05.k ) & w(i . Nodal wear rate of node i & w(i . 0.

which consists of the nodes lose connection with the workpiece body because all the attached elements are removal. The attached element is deleted or removed as it reaches the shear failure criterion. 6. Because inner nodes have no contact with the tool face. Free node (6. The number of the deleted elements is denoted as Ndelete. N deleted ≠ 0 N  attached = N deleted . The first type. they will introduce calculation error.3) . workpiece nodes considered are classified into three types. surface node. Hence only surface nodes join in the calculation of relative sliding velocity. including the nodes exposed on the surface. N deleted ≠ 0 Inner node Surface node .Estimation Of Tool Wear In Milling Operation 106 element removal and get contact with the tool face. Therefore a large number of workpiece nodes have to be considered. for example. Every node is attached to several elements.7. Node 343 belongs to the second type. Node 411 in Fig. free node. The number of the attached elements is denoted as Nattached. they only increase the calculation time. 6. called inner node. In order to calculate the sliding velocity correctly and efficiently. Node 411 Node 893 Node 343 Fig. includes the nodes still inside the workpiece. When free nodes get contact with the tool face. The type of a node is decided by:  N deleted = 0   N attached > N deleted .7 Classification of workpiece node Node type is judged by considering the number of the attached elements and the deleted elements of a node. Node 893 is included in the third type.

The calculation of nodal move directions and tool geometry updating should be performed at the same rotation position of the cutting tool. Nodal move direction is calculated with different methods on the rake face and flank face. as shown in Fig.3.1 Dividing Node Similar to the calculation of nodal move direction in turning operation.4. Instead of comparing the y-coordinate of every tool face node. Any tool face node whose distance to the rotation center is greater than that of the current dividing node by 2e-4mm will update the record of dividing node. Dividing node vs Flank face part Rake face part vs vs n vs vs vs vs Fig.4 Nodal Move Direction 107 In milling operation. 6. the dividing node is defined according to the distance between tool face nodes and the rotation centre. at the beginning a dividing node that divides the entire tool face into rake face and flank face is searched. the first tool face node (the tool face nodes are arranged in counter-clockwise order in advance) is given to the dividing node. Then every tool face node is compared with the current dividing node one by one in counterclockwise order.8 Calculation of nodal move direction . position of every tool face segment and its normal direction are varying with the rotation of the cutting tool. 6.3.Estimation Of Tool Wear In Milling Operation 6. At the beginning of the search.8. 6. for example at the beginning of one milling cycle.

where subscript i is nodal label.3. 6. Nodal move direction at the flank face node is perpendicular to the relative sliding velocity. j ) . when the elastic recovery of workpiece material is neglected. Edge position a a1 b b1 c c1 d VB Fig.3 On Flank Face In the flank face part.3.4. wearrate) .9 Flank wear calculation 6. the searching procedure and the flank wear calculation subroutine have some difference compared with those in turning operation. 6. relative sliding velocity of the workpiece material at each flank face node is assumed along the tangential direction of the moving path of the flank face node. r Every nodal move direction is normalized to unit vector D( i .3.5.e. is applied here as well. i.1 Flank Wear Calculation Subroutine Flank wear land width VB is calculated by a flank wear calculation subroutine Flankwear (∆t .Estimation Of Tool Wear In Milling Operation 6. j is the calculation cycle number. explained in Chapter 5. 6.3.2 On Rake Face 108 The calculation method of nodal move direction on rake face used in turning operation.4.5 Cutting Time Increment Calculation Cutting time increment is searched by the program according to a user-specified flank wear increment value ∆VB and a permitted error δ . Because of the particularity of milling operation. pointed from the flank face node to the rotation centre. .

the aimed VB median value VBm is calculated according to the user-specified VB increment value and the tool wear obtained in the previous calculation cycle. In the same way. According to the permitted error δ . 6. VB is calculated from edge position to node c.3. . In milling operation the cutting insert rotates around a rotation centre. because it is the last moved tool face node.2 Cutting Time Increment Searching Procedure Because milling operation is intermittent cutting. Then node b becomes the farthest point to the rotation center. The cutting time increment searching procedure is described as follows: (1) At the beginning. node c will be moved to point c1. Any tool material point which becomes the farthest point to the rotation centre will get contact with the workpiece and then be worn away. in Fig. Because in this calculation cycle the comprehensive information about the average wear rate of node b in the entire cutting time increment cannot be obtained. the cutting time increment is increased in step of whole milling cycles. Z is the time span of one milling cycle. its displacement is decided according to node a. For example. 6. Node b will be move to point b1 and it will have the same distance to the rotation centre as node a. The last moved tool face node is searched by considering the movement of the cutting insert. it should move to point a1 in the cutting time increment ∆t .5. VBm + δ ) is determined.Estimation Of Tool Wear In Milling Operation 109 VB is the distance from the edge position to the last moved tool face node. the existence of cutting phase and cooling phase in milling cycles complexes the cutting time increment searching procedure. According to the calculation. The relationship between the milling cycle number Ncycle and the cutting time increment ∆t is given by ∆t = N cycle × Z (6. the aimed VB value range ( VBm − δ .9 node a is the last tool face node with non-zero nodal wear rate.4) where Ncycle is positive integer. It will be worn away. In order to simplify the problem.

and call the subroutine Flankwear (∆t . then the cycle number lower limit N cycle1 = 1 . The present ∆t value will be output as result and the searching procedure will end. (4) The cycle number upper limit N cycle 2 = N cycle (5) Ncycle takes the integer part of (N cycel1 + N cycle 2 ) 2 . wearrate) − VBm . if error > δ . Otherwise. else. 6. j ) where (6.6 Nodal Displacement Calculation Nodal displacement is calculated at every tool face node by r r & w( i .5) . then the cycle number lower limit Ncycle1=Ncycle. and N cycle = 2 ⋅ N cycle . and call the subroutine Flankwear (∆t . If error ≤ δ . (3) Otherwise.Estimation Of Tool Wear In Milling Operation 110 Then an initial cycle number Ncycle0 is given a positive integer value arbitrary. If error ≤ δ . let error = Flankwear (∆t . (2) Calculate the cutting time increment ∆t corresponding to Ncycle with Eq. let error = Flankwear (∆t .δ is satisfied. (6) Calculate the cutting time increment ∆t corresponding to Ncycle. wearrate) to calculate the flank wear land width.4. if error < . j ) = w(i . the cycle number upper limit N cycle 2 = N cycle . then calculate the cutting time increment ∆t . then the searching procedure will end. and the present ∆t value will be output as the result. repeat step 5 until error ≤ δ is satisfied.δ . 6. If the value of Ncycle is equal to Ncycle1 or Ncycle2. Else. if error < . repeat step 2 until error ≥ . the cycle number Ncycle and the cycle number lower limit Ncycle1 are set to Ncycle0. and the present ∆t value will be output as the result. j ) ⋅ ∆t j ⋅ D(i . wearrate) − VBm . then the searching procedure will end. wearrate) to calculate the flank wear land width.3.δ . then the cycle number lower limit N cycle1 = N cycle .

tool wear under the same cutting condition as described in Table 3.2 is estimated. 6. 6. whereas the main tool wear in high-speedmilling results from chipping. permitted error is set to 0. 111 i is nodal label.10(a) is updated to the worn tool in Fig. j is tool wear calculation cycle number. 6.05mm is specified by the user. 6. They reduce the strength of the tool edge and accelerate tool wear. the new tool in Fig. Flank wear Crater wear VB=0. which has a flank wear width of 0.9. etc.10(b).01mm.10 Tool wear profile (b) t=603s A discrepancy between the estimated tool wear from the program and the expected tool wear from experiment is unavoidable mainly because of the following reasons: (1) Complex tool wear in milling operation. only the abrasive and adhesive wear are considered. .06mm and crater wear on the rake face.4 Results & Discussion With this tool wear program. After the cutting time of about 603s. In addition. node b and c in Fig. 6.6 Tool Geometry Updating Tool geometry updating is performed with the same procedure as explained in turning operation. ∆VB = 0. are moved as well as explained above.Estimation Of Tool Wear In Milling Operation r w is the nodal displacement vector. thermal crack. e. some nodes on flank face.06mm (a) t=0s Fig. 6.3. In this study.g.

Estimation Of Tool Wear In Milling Operation 112 (2) Chip formation analysis modeling. A multi variable dependent shear failure criterion may provide a better chip formation simulation result and improve the predicted tool wear profile and tool wear value. when shear failure criterion is used as the chip separation method. . It is verified by the test that the chip thickness and tool/chip contact are sensitive to the element size and given value of shear failure criterion.

The main findings of this study are as follows: • Milling operation is an intermittent cutting process. tool wear under a cutting condition with high cutting speed is calculated. By sampling the cutting process variables in the chip formation analysis and heat transfer analysis at the corresponding time points and calculating the nodal average wear rate per cycle. tool temperature in the cyclical thermal balance state should be used in tool wear estimation. a very slow tool wear process is expected because in Usui’s tool wear equation. • With the developed tool wear program. wear rate and tool temperature has an exponent relationship. a tool wear estimation model is implemented for the milling operation.Estimation Of Tool Wear In Milling Operation 113 6. • Because of the temperature difference in the cyclical thermal balance state and in the milling cycle before cycle thermal balance state is reached. Otherwise. both crater wear and flank wear are formed on the tool face. tool wear estimation modelling can be implemented.5 Summaries & Conclusion In this chapter. .

Tool wear estimation in turning operation is based on the study of finite element simulation of steady-state cutting process. A tool wear estimation program is developed. Because milling operation is intermittent cutting process. chip growth to steady state. A new chip formation modeling method is developed to simulate the entire process from initial chip formation. which show that wear rate of cutting tool is dependent on some cutting process variables such as tool temperature. In addition. No obvious crack is formed in front of the cutting edge. heat transfer and tool wear estimation modeling are different from turning operation. nodal displacement calculation and geometry updating are discussed. Pure heat transfer analysis of only the cutting tool is carried out to save the calculation time to reach thermal steady state. When the coefficient of friction calculated according to cutting force in experiment is used. The estimated tool wear is verified by experimental data.1 Summaries In this study the methodologies to numerical implementation of tool wear estimation in turning and milling operation are discussed. . the chip formation. Temperature dependent heat flux at tool-chip interface and heat convective and radiation of tool face are considered during modeling. This chip formation model is verified by experimental data. It is assumed that the error is created by the low coefficient of friction in chip formation analysis. the preliminary qualitative tool wear estimation models are developed. It can calculate the tool wear until the tool reshape criterion is reached. Based on the researches of tool wear mechanism. It is not necessary to get material failure parameters or chip geometry from experiment. cutting time increment searching. no separation path is preset in advance. Then the problems about calculating nodal wear rate at steady state according to Usui’s tool wear equation.Summary And Prospect 114 Chapter 7 Summary And Prospect 7. sliding velocity of workpiece material and normal pressure on tool face. Instead it provides an alternative method to decide material failure parameters. Chip separation is formed automatically by solution-dependent mesh adaptivity instead of material failure criterion. the error of two cutting force components is smaller than 5% compared to the experimental data. After only several minutes of calculation the cutting tool gets a steady temperature distribution.

Fortran. it is very helpful to optimise tool geometry and structure . It is found that the temperature in cyclical thermal balance state is higher than the first several milling cycle. No cyclical thermal balance state is realized. In order to speed up the realization process of cyclical thermal balance state. Pure heat transfer analysis of only the workpiece shows that the workpiece cool down to room temperature in the cooling phase of one milling cycle if the cutting speed is not tool high. and relate tool wear with some wear mechanisms. Both crater wear and flank wear are formed. Tool wear estimation with the help of finite element method can predict not only tool life. This tool wear estimation method will relate the geometry appearance to physical basic of tool wear and bridge the gap between macro and micro studies of tool wear. the chip formation process in every milling cycle is assumed similar because of the negligible temperature increment in the workpiece.Summary And Prospect 115 The chip formation simulation is realized by introducing shear failure criterion. During the study multi aspects of cutting process simulation modelling in turning and milling operations including chip formation analysis. In order to fulfil the purpose of tool wear study. multiprogramming tools including commercial FE code ABAQUS/Explicit. This is very meaningful for the scientific research and education. Using this program. The strain at failure in shear failure criterion is defined according to the former chip formation modeling method. Shear failure criterion is applied to the entire workpiece. Tool wear estimation in milling operation is performed by calculation nodal average wear rate in one milling cycle and use it to other milling cycles. all the problems about cutting time increment searching. nodal displacement calculation and geometry updating are different from turning operation and they are discussed. but also wear profile of both crater wear and flank wear. different preheated cutting tool is used in the cutting process. tool wear in one milling case is calculated. Because in milling cycle the cutting tool is rotating instead of the workpiece as in turning operation. Temperature is observed after every milling cycle. Python are employed and integrated. Pure heat transfer analysis of only the cutting tool is carried out for 8 milling cycles. Accordingly. For tool designer. ABAQUS/Standard. This lays a ground for the study on more complex problem and the extension of functionality of FEM in the future. Then a tool wear estimation program for milling operation is developed. heat transfer analysis. and tool wear estimation are studied.

flank wear is produced by moving nodes individually according to the nodal wear rate instead of according to an average value. It provides a method to produce the complex tool wear in 3D. 7. which is applicable to wider cutting range. A further improvement of tool wear estimation may be realized by introducing multi wear mechanism. the result of tool wear estimation is maybe improved. It will be possible to study on the contribution of every wear mechanism under different cutting conditions. for material engineer. it is useful to improve tool material according to the determined main wear mechanism. the formed wear profile is not smooth even after the second updating step and some nodes have to be adjusted manually. develop tool wear estimation . The tool wear is calculated according to their combination. once tool wear mathematical model for a combination of tool-workpiece material is determined. it is possible to estimate tool wear by program without doing any experiment. In this tool wear estimation method. this estimation method is helpful to reduce the size of various cutting database by replacing tool life equation with tool wear mathematical model. In order to spread the application of this method in industry practices. A special smoothing algorithm should be designed to solve this problem. Friction has big influence on the chip formation analysis and tool wear. In tool geometry updating. In order to produce good contact in this area a negative flank angle designed on the flank wear may be a good solution.Summary And Prospect 116 knowing wear profile and wear mechanism. Because in the tool wear estimation modeling. This may be caused by the contact problem between the flank wear and the workpiece. the chip formation analysis of steady state sometimes produces relatively low tool temperature on flank wear. it is necessary to develop tool wear mathematical model for most common used materials. such as wear notch. When using the coefficient of friction calculated according to the cutting force from experiment. tool wear is related to wear mechanism. except improving the precision of tool wear estimation. In addition.2 Prospect The tool wear estimation models should be improved in several aspects: After a certain tool wear is formed.

. ceramics cutting tool.Summary And Prospect 117 model for coated carbide tool. and research on 3D tool wear estimation model in the future. etc. CBN cutting tool.

J-C. 2D and 3D numerical models of metal cutting with damage effects. D. 2000. 11-14. O.2. 2001 Ahmad.References References [ABA-01a] [ABA-01b] [ABA-01c] [Ahma-89] 118 ABAQUS/Explicit User’s Manual Version 6. G.S. B. U..A. Engrs. C. Draper. HSK. 1998.. Atlanta. [Arra-02] Arrazola. Meslin. Olson.. M. Newport. High speed chip formation of Ti6Al4V. S. [Baca-00] Bacaria. A. Applied Mechanics Group. 2001 ABAQUS/Standard User’s Manual Version 6... Inc. U. M. Instn. J.2. Mach. Inc. Dalverny. and Caperaa. Journal of Manufacturing Science and Engineering 124 (2) (2002) 32-41.. F. J. [Bäke-00] [Bäke-02] Bäker. Sultan.A.A. May 29-31... R... HSK. Rbode Island.2.. Finite element simulation of segmented chip formation of Ti6Al4V. 2002. Tools Manufact. Finite element modeling of edge trimming fiber reinforced plastics. [Arol-02] Arola. Int. O. European Congress on Computational Methods in Applied and Engineering. USA. S. Proc. Rösler. Large . [Batz-02] Batzer. Finite element modeling of machining: from proof-of-concept to engineering applications. An application of the finite element method to the prediction of cutting tool performance. Siemers. T.. W. pp. Munich.... 29 (2) (1989) 197-206. J. U. Inc. [Atha-98] Athavale.. and Le Maitre. M.1. Mech. R. Arndt. J.... Proceedings of Materials Week. P. J. J. HSK.. 203-216. Rösler. L. Ultra-high-speed machining: a review and an analysis of cutting forces. W. S... M.S. Sep. F. Journal of Manufacturing Science and Engineering 124 (2002) 485-488... 2002 ABAQUS Users’s Conference. M.. 187 (1973) 625-634. M.. 2001 ABAQUS Scripting Manual Version 6. Hamann. and Sutherland. Ramulu.. Siemers. G. Strenkowski.. Subhash. Historie de l’Academie Royale des Sciences avec les Memoires de Mathematique et de Physique.. A. J. 1699. in: Proceedings of the CIRP International Workpshop on Modeling of Machining Operations. W.S. Rakotomalala. and Derricott. [Amon-99] [Arnd-73] Amontons. S. M. C. 2000 Bäker. Numerical cutting modeling with Abaqus/Explicit 6.. Barcelona... Pantale..

.. Journal of Materials Processing Technology 98 (2000) 99-103... Wu. Westhoff. Application of 2D FEM to chip formation in orthogonal utting. Devor. Finite element modeling of high speed machining [Boot-89] [Chil-89] [Cere-96] processes. [Cere-99] Ceretti. A. 31-40. Mahdi. E. CIRP 38 (1) (1989) 55-58. T. Dirikolu. München. T. W. A. Darmstadt. Ceretti.. L. 1998... 1998. Modelling of friction in the simulation of metal machining. Fallböhmer.-T. R. 2nd International German and French Conference on High Speed Machining. Machining simulation of ductile iron and its constituents.. BadenBaden. 1999. [Chil-98] Childs. Altan. T. T. [Behr-98a] [Behr-98b] Behrens. Westhoff. Altan. Journal of Materials Processing Technology 59 (1996) 169-180. H. E.. H.. Lazzaroni.. L. part 2: numerical simulation and . Anwendung von MARC auf Zerspanprozesse. in: 24th Leeds-Lyon Symposium on Tribology.. Inc. M. 34.. XXV. A. C.. and Kapoor... 1998. Turning simulations using a three-dimensional FEM code. Fundamentals of machining and machine tools. USA... On the stress distribution between the chip and tool during metal turning. pp.. W. E. Journal of Materials Processing Technology 95 (1999) 17-26... S.. MARC Benutzertreffen 1998. B. Machining simulation of ductile iron and its constituents. E. and Kapoor. [Behr-99] Behrens. FEM-Kongress. [Chuz-03b] Chuzhoy. USA. E... Amsterdam: Elsevier Tribology Series No..337-46.. S. part 1: estimation of material model parameters and their validation.References 119 strain constitutive model development for application to orthogonal machining. Devor. Westhoff. M. L. 5th CIRP International Workshop on Modeling of Machining Operations. p. West Lafayette. B. Altan. R. 185-190.. G. T.. Behrens. Journal of Manufacturing Science and Engineering 125 (5) (2003) 181-191. and Maekawa. B. Menegardo.. 1989 Childs. M. C.. G. pp. [Cere-00] Ceretti. in: Proc. Boothroyo. Lucchi. A. G. Knight. [Chuz-03a] Chuzhoy. Ann. 2002. K. Marcel Dekker. FEM simulation of orthogonal cutting: serrated chip formation. pp. Probleme und Herausforderungen bei der Simulation von Zerspanprozessen.

S..313 Huang..B. 3D FEA modeling of hard turning. D.. 2001.. Fracture mechanisms in chip formation processes. B. Diplomarbeit. d. L.. Didjanin. Journal of Manufacturing Science and Engineering 125 (5) (2003) 192-201.. 1979. Journal of Manufacturing Science and Engineering 118 (1996) 545-554. Universität Karlsruhe (TH). Tool wear in orthogonal turning of AISI 1045 steel with uncoated tungsten carbide tool. J.. Computer simulation and optimization of metal cutting process for mild carbon steels. Why cutting tools fail.. Y. [Guo-00] Guo. Xing. Y. Frank. HPM/ERC/NSM-01-R-13. R. Gu. Chang. Improvement of the FEM-based Predictive Model of Tool Wear. G. B. ERC/NSM Report No. Memoires de Mathematique et de Physique de l’Academie Royale des Sciences.. Black.. D. Materials Science and Technology 13 (1997) 439-444. A.. Proc. Oxley. V.. [Coul-85] [Didj-97] [Dill-00] [Feve-01] Coulomb. A. P. P. [Iwat-84] Iwata... pp. T.. T. Taylor... Feve. Yen. Terasaka. L. Y-C. Dornfeld. [Guo-02] [Hast-79] [Huan-96] Guo. P. M. Z. The Ohio State University.References 120 experimental validation of machining.. Transactions of the ASME 78 (1956) 1649-57. K. Journal of Engineering Materials and Technology 106 (1984) 132-138. M.R. C. J. Dillon.D. Altan... W. [Finn-56] [Fran-02] [Gu-02] Finnie. H. Finite element modeling of burr formation process in drilling 304 stainless steel. L.. J. Wang. K. Journal of Manufacturing Science and Engineering 122 (2000) 612-619. 2002. 1785.F... O. CNC Machining Magazine 5 (18) (2001). J. Modelling and Simulation of Mechanical Cutting.. M.. C. Y.. Chen. Mathew. Zhang. Hastings. Journal of Materials Processing Technology 129 (2002) 60-65. and Kovac. Liu. Shaw. W. Journal of Manufacturing Science and Engineering 124 (2002) 189-199. Osakada.T. Kalhori. Doctoral .. J. Process modeling of orthogonal cutting by the rigid-plastic finite element method. The friction process in metal cutting. L.. An analysis of cutting using a grooved tool. C. I.. Conf. International Journal of Forming Processes 3 (1-2) (2000) 115-130.. [John-01] [Kalh-01] Johnson. An evalution of chip separation criteria for the FEM simulation of machining.. 20th Int. Paris.

Japan Soc. The Ohio State University. [Klam-73] Klamecki.a three dimension finite element analysis. Maekawa. T. E. of Prec. Predictive models for flank wear on coated inserts. 2001. and Usui. part 1: characteristic equation of flank wear. Leopold Jürgen. T. Kwon. part 2: prediction of flank wear. 1998.. Analytical prediction of flank wear of carbide tools in turning plain carbon steels. Analytical prediction of flank wear of carbide tools in turning plain carbon steels... FEM modeling and simulation of 3-D chip formation. M. University of Illinois. Stress distribution at the interface between tool and chip in machining. A finite-element analysis of machining with the tool edge considered. Sutherland. 23 (2) (1989) 126-133. Institutionen för Maskinteknik. T. Journal of Tribology 122 (1) (2000) 340-347. [Köni-84] [Kwon-00] [Leop-98] König.References 121 thesis.. D. Lee. VDI-Verlag. Bull.. W. K. S.. X. Sin. Journal of Materials Processing Technology 127 (2002) 309-324. T. Bull. pp. Maekawa. [Kita-89] Kitagawa.... T. [Kita-88] Kitagawa. Journal of Materials Processing Technology 86 (1999) 45-55. Ph. K. Engg. Finite element simulation of the orthogonal metal cutting process for qualitative understanding of the effects of crater wear on the chip formation process.. dissertation. Experimental Determination of Flow Stress Data for FEM Simulation of Machining Operations. Sartkulvanich. E. Shirakashi.. [Kato-72] Kato. M. Yamaguchi. [Kopp-01] Koppka. B. Japan Soc. W. Yamada. [Li-02] Li...-L. Fertigungsverfaren / Manufacturing Methods. K. W. . J... Germany. USA. Atlanta. P. Gao. and Altan. 22 (4) (1988) 263-269. in: Proceedings of the CIRP International Workpshop on Modeling of Machining Operations. of Prec.. C. 235-245. H.. Engg. HPM/ERC/NSM-01-R-63. Journal of Engineering for Industry 94 (1972) 683-689. Incipient chip formation in metal cutting . K. F. 1973. 2001. and Usui. Avdelningen för Datorstödd maskinkonstruktion. ERC/NSM Report No. Shirakashi. 1984.. P.. W. Y. [Kim-99] Kim. Lulea Tekniska Universitet.. Dusserdolf. E.

Engineering . A methodology for simulation of chip breakage in turning processes using an orthogonal finite element model.. Maldonado. pp. Finite element analysis of the effect of sequential cuts and tool-chip friction on residual stress in a machined layer. Lin.-C.-L. A study of an oblique cutting model. T.-Y.-C.References [Lim-01] [Lin-99] [Lin-00] 122 Lim. Lin. C. Journal of Materials Processing Technology 86 (1999) 119-130. D. Y. E. and Thiele. in: Proc. International Journal of Mechanical Science 42 (2000) 10691086.. D.. Wear 250 (2001) 344-348. I. P.. Annals CIRP 47 (2) (1998) 587-626. [Maru-95] Marusich. Y. Brand. R. D. H. 5th CIRP International Workshop on Modeling of Machining Operations. M. Klocke. 139-148. Y.. [Lin-01b] Lin. Three-dimensional thermo-elastic-plastic finite element analysis with different cutting speeds. B... Childs. C. M.-C. C. [Liu-00] Liu.. J. [McCl-02] McClain. [Lutt-98] van Luttervelt. Venuvinod. Finiteelement modeling of dry sliding wear in metals. S. Z. Y. J. 1998 Molinari. A. C.. West Lafayette.. Journal of Materials Processing Technology 123 (2002) 114-119... I. T. Lin... P. A. Ortiz. The effects of work material on tool wear.. C. P.-Y. Journal of Materials Processing Technology 115 (2001) 313325. E. Guo.. Z. International Journal for Numerical Methods in Engineering 38 (1995) 3675-3694. T.... Radovitzky. A study of oblique cutting for different low cutting speeds. S. C. B.. L.. Keynote papers: present situation and future trends in modeling of machining operations progress report of the CIRP Working Group ‘Modelling of Machining Operations’. Zheng. Atalantic.. Lim. and Repetto.. F.. T. Lau. K.. Y. 2002. Batzer. International Journal of Computer Applications in Technology 13 (2000) 210-220. R.. [Merc-98] [Moli-00] Merchant. USA. Lin. A. F. A numeric investigation of the rake face stress distribution in orthogonal machining. Modelling and simulation of High-Speed Machining.. G. Jawahir. H. [Maru-02] Marusich. Oritz.. USA. S. J. in Proceedings of the CIRP International Workshop on Modeling of Machining Operations..

T. Process simulation using finite element method – prediction of cutting force.. M.. International Journal of Machine Tools & Manufacture 40 (2000) 713-738. USA. Proc.....-G. West Lafayette. T. J. T.References Computations 18 (3/4) (2001) 592-609. The residual stress produced by metal cutting.. Altan. Computational Materials Science 16 (1999) 275-284. Kakino. E. [Özel-00b] Özel. J. Modelling of temperature and forces when orthogonally machining hardened steel. R. May 20-21. Altan. IN. T. Ng.tool contact for high-speed cutting.. [Okus-71] [Owen-99] Okushima. [Ng-99] Ng. CIRP 20 (1) (1971) 13-14. M... D. Altintas. pp. Y. and Elbestawi. [Ohbu-03] Ohbuchi. J. and Monaghan. Y. tool stress and temperatures in high-speed flat end milling. M. D. Modeling of hard part machining: effect of insert edge preparation in CBN cutting tools. [Mona-99] 123 Monaghan. 5th CIRP International Workshop on Modeling of Machining Operations.. Brazil. T. and MacGinley. M. D. Determination of workpiece flow stress and friction at the chip. Computational techniques applied to highspeed machining under adiabatic strain localizetion conditions.. R. Aspinwall... Computer methods in applied mechanics and engineering 171 (1999) 445-446. K. International Journal of Machine Tools & Manufacture 39 (1999) 885903 [Ng-02a] [Ng-02b] Ng. International Journal of Machine Tools & Manufacture 40 (2000) 133-152. Vaz Jr.. [Özel-00a] Özel. [Mova-02] Movahhedy... E.. Owen.. Journal of Manufacturing Science and Engineering 124 (2002) 178-187. Finite element modeling of chip formation in the domain of negative rake anlge cutting. Y.. A..-G.-G. Modelling of hard part machining. and Aspinwall. S. 2002.. 1-19. Dumitrescu. Journal of Materials Processing . T.. Obikawa. M. Ann... Journal of Materials Processing Technology 127(2002) 222-229. [Özel-02] Özel. EI-Wardany. Modelling the orthogonal machining process using coated carbide cutting tools. E. K. D. Journal of Engineering Materials and Technology 125 (2003) 324-332. Numerical analysis of metal cutting with chamfered and blunt tools. T. K. Gadala. T.

Kerk.. . T. C.. I. Transactions of the E... Altan. 177-178.. Altan. Y. X. X. part II: validation and applications of the determined flow stress data.. Residual stress and strains in orthogonal metal cutting.. Modeling the physics of metal cutting in High-Speed machining. [Shet-03] Shet. M. Deng. M. Salmon. Journal of Material Processing technology 105 (2000) 95-109. C. V. in: Proceedings of the CIRP International Workpshop on Modeling of Machining Operations.. C. G. pp. M. [Shat-00] Shatla. Diplomarbeit. [Schu-00] Schulze. Deng. A finite element study of the effect of friction in orthogonal metal cutting. Finite Elements in Analysis and Design 38 (2002) 863-883.. International Journal of Machine Tools & Manufacture 41 (2001) 1511-1534a. Photoelastic determination of cutting tool stress. 217-224. W.. USA. T. C. Deng. [Shat-01b] Shatla. Kerk. Atlanta. Altan. International Journal of Machine Tools & Manufacture 43 (2003) 573-587. Influence of alloying elements on the strain rate and temperature dependence of the flow stress of steels.. and Vöhringer. R. X. Y. Experimental and finite element simulation methods for rate-dependent metal forming processes.. 4 (1) (1960) 20-23. Transactions of NAMRI/SME 2000. D. Yang. [Rice-60] [Sand-98] 124 Rice. Process modeling in machining. Metallurgical and Materials Transactions A 31A (2000). Process modeling in machining. [Shat-01a] Shatla.. B. Shet.. T. Yen. Finite element analysis of the orthogonal metal cutting process.. Tool-workpiece interface in orthogonal cutting-application of FEM modeling.-C. R. A. C. part I: determination of flow stress data.. O. W. H. International Journal for Numerical Methods in Engineering 31 (1991) 345-367. C. XXVIII. [Shet-00] Shet.. 2002.References Technology (2002). 1998. T.. M. [Shih-91] Shih.. Development of a FEM-based Tool Wear Model to Estimate Tool Wear and Tool Life in Metal Cutting.. International Journal of Machine Tools & Manufacture 41 (2001) 16591680. Syniuta... [Schm-02] Schmidt. D... Universität Karlsruhe (TH). C. [Shi-02] Shi. Sandstrom. Q.

Y. G. Vorgäng J. Int.31 (1907) Trent. 2003 [Stre-93] Strenkowski. Stevenson. 1977 Usui. HPM/ERC/NSM-01-R-76. Experimental and finite element predictions of residual stresses due to orthogonal metal cutting.... ERC for Net Shape Manufacturing... Finite element analysis of the rake angle effects in orthogonal metal cutting. Universität Karlsruhe (TH). 2001 [Söhn-01b] Söhner. Shih. and Yen. F. A. Analytical prediction of three dimensional cutting .M.References [Shih-93] 125 Shih. Material Database for Manufacturing Simulation (MADAMS): Summary of the Activities and Flow Stress Database. Mech. Eng.. O. Metal Cutting. Ohio State University.. Using the finite element method [Tayl-07] [Tren-77] [Usui-78a] to determine temperature distributions in orthogonal machining. David. Shih. 188 (1974) 627-638. J. J.. Ohio State University. Eng. [Shih-95] [Shih-96] Shih. A finite element model of orthogonal metal cutting. J. Estimation of tool wear in metal cutting with the finite element method –a progress report-. Taylor. and Altan. [Shir-93] [Söhn-01a] Shirakashi. A.. mit Beitrag Hilfe zur Simulation zerspanungstechnologischer der Finite-Element-Methode. J.-C.. International Journal of Machine Tools & Manufacture 42 (2002) 723731. Inst. An analytical finite element model for predicting three-dimensional tool faces and chip flow. J.. Finite element simulation of orthogonal metal cutting. Carroll III. ERC/NSM Report No. Am. [Stre-02] Strenkowski. International Journal of Mechanical Sciences 38 (1) (1996) 1-17.... M. HPM/ERC/NSM–01–R-67. S. Journal of Engineering for Industry 107 (1985) 347-354.. Proc.W. J. T. A. Söhner. T. J. J. Jpn. Lin. J. 27 (1993) 299-302. Prec. England.. Y. G. A. S. T.-C... ERC for Net Shape Manufacturing. E. International Journal for Numerical Methods in Engineering 36 (1993) 1487-1507. Soc. V.. E. ERC/NSM Report No. [Tay-74] Tay. Mech. A. Journal of Engineering for Industry 117 (1995) 84-93. A. Soc. Butterworths.. T. Hirota. J. Engrs 28. J. Trans. H. Yang. New trend on machining theory-way to computational machining. Dissertation. 2001 [Söhn-03] Söhner.

E. [Yang-02] Yang. 233239. Journal of Engineering for Industry 100 (5) (1978) 222-228. A. part 3: cutting temperature and crater wear of carbide tool. 2001. Finite element simulation of chip formation and comparison with machining experiment. Analytical prediction of three dimensional cutting process. T. M. W. Study on chip segmentation in machining Titanium alloy. Institut für Konstrucktionsund Fertigungstechnik. Dissertation. X. Japan. [Usui-78b] Usui.References 126 process. O. and Altan. Obikawa.. part 2: chip formation and cutting force with conventional single-point tool. Universität der Bundeswehr. [Wu-96] Wu. Schmidt.149-160 [Zhan-94] Zhang. Dillon.. H. Tokyo. in: Proceedings of the 5th International Conference on Production Engineering. in: Proceedings of the 5th CIRP International Workshop on Modeling of Machining Operations.. . E. West Lafayette. Hirota. pp. Söhner. T. J.-Y. Modellierungsgrundlagen zur FE-Analyse von HSCProzessen. Masuko. E. 2002.. Shirakashi.. [Usui-84] Usui.-P.. Shirakashi.. T. 1984. T. Analytical prediction of three dimensional cutting process. Thermo-viscoplastic modeling of machining process using a mixed finite element method. Journal of Engineering for Industry 116 (8) (1994) 289-297. B. Karlsruhe. R.. Y. Estimation of tool wear of carbide tool in orthogonal cutting using FEM simulation. pp.. Dissertation Universität Karlsruhe (TH).. J. Bagchi. A new stress-based model of friction behavior in machining and its significant impact on residual stresses computed by finite element method. [Walt-98] [West-01] Walter. [Yen-02] Yen.. USA. 1998 Westhoff. W.. T.. J. Weule. Journal of Engineering for Industry 100 (5) (1978) 229-235... Lu. C.... and Kitagawa. part 1: basic cutting model and energy approach. S. B. Laboratorium Fertigungstechnik. A. International Journal of Mechanical Sciences 44 (2002) 703-723. Journal of Manufacturing Science and Engineering 118 (1996) 470-482. Einfluss des Kühlschmierstoffes beim Fräsen. U. [Usui-78c] Usui. C.. Journal of Engineering for Industry 100 (5) (1978) 236-243.. Liu.. Germany.

International Research in Production Engineering (1963) 42-49.N.References [Zore-63] 127 Zorev. N. . Inter-Relationship between Shear Processes occurring along Tool Face and Shear Plane in Metal Cutting..

128 Resume Name: Nationality: Date and place of birth: Family Status: Lijing Xie China June. China Scientist at Werkzeugmaschinen und Betriebstechnik (WBK). Department of Mechanical Engineering. Hebei. 2000 Lecturer at Beijing Institute of Technology. 1971 in China Married with Dan Zheng since May. Hebei. Tangshan.. Germany . 1998 Education: 1978-1983 1983-1986 1986-1989: 1989-1993 1993-1996 Tangshan Nanxindao Elementary School. 18. Tangshan. China B. Uni-Karlsruhe (TH). Beijing Institute of Technology.S. 17. Beijing. China Employment History: 1996-2000 Since Nov. Tangshan Institute of Technology. China The Eighth Junior High School. China M. China The First Senior High School.S. Department of Mechanical Engineering.. Beijing.

Sign up to vote on this title
UsefulNot useful