You are on page 1of 147

FINITE ELEMENT SIMULATION OF ORTHOGONAL METAL CUTTING USING AN ALE APPROACH

by Abdulfatah Maftah

B.Sc.Eng. University of Seventh April, 1998

A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of

Master of Science in Engineering

in the Graduate Academic Unit of Mechanical Engineering

Supervisors:

Dr. H. A. Kishawy, Mechanical Engineering Department Dr. R. J. Rogers, Mechanical Engineering Department

Examining Board: Dr. A. Gerber, Mechanical Engineering Department, (Chair) Dr. Z. Chen, Mechanical Engineering Department Dr. A. Schriver, Civil Engineering Department

This thesis is accepted by the Dean of Graduate Studies

THE UNIVERSITY OF NEW BRUNSWICK April, 2008 Abdulfatah Maftah, 2008

1*1

Library and Archives Canada Published Heritage Branch 395 Wellington Street Ottawa ON K1A 0N4 Canada

Bibliotheque et Archives Canada Direction du Patrimoine de I'edition 395, rue Wellington Ottawa ON K1A 0N4 Canada
Your file Votm reference ISBN: 978-0-494-63711-1 Our file Notre reference ISBN: 978-0-494-63711-1

NOTICE: The author has granted a nonexclusive license allowing Library and Archives Canada to reproduce, publish, archive, preserve, conserve, communicate to the public by telecommunication or on the Internet, loan, distribute and sell theses worldwide, for commercial or noncommercial purposes, in microform, paper, electronic and/or any other formats. The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. In compliance with the Canadian Privacy Act some supporting forms may have been removed from this thesis. While these forms may be included in the document page count, their removal does not represent any loss of content from the thesis.

AVIS: L'auteur a accorde une licence non exclusive permettant a la Bibliotheque et Archives Canada de reproduire, publier, archiver, sauvegarder, conserver, transmettre au public par telecommunication ou par I'lnternet, prefer, distribuer et vendre des theses partout dans le monde, a des fins commerciales ou autres, sur support microforme, papier, electronique et/ou autres formats.

L'auteur conserve la propriete du droit d'auteur et des droits moraux qui protege cette these. Ni la these ni des extraits substantiels de celle-ci ne doivent etre imprimes ou autrement reproduits sans son autorisation.

Conformement a la loi canadienne sur la protection de la vie privee, quelques formulaires secondaires ont ete enleves de cette these. Bien que ces formulaires aient inclus dans la pagination, il n'y aura aucun contenu manquant.

Canada

ABSTRACT
Understanding of the fundamentals of metal cutting processes through the experimental studies has some limitations. Metal cutting modelling provides an alternative way for better understanding of machining processes under different cutting conditions. Using the capabilities of finite element models, it has recently become possible to deal with complicated conditions in metal cutting. Finite element modelling makes it possible to model several factors that are present during the chip formation including friction at the chip tool interface, temperature, stress, strain, and strain rate. The aim of improved understanding of metal cutting is to find ways to have high quality machined surfaces, while minimizing machining time and tooling cost. In this study, an Arbitrary Lagrangian Eulerian (ALE) finite element formulation is used to simulate the continuous chip formation process in orthogonal cutting. The ALE is an effective way to simulate the chip formation as it reduces element distortion that causes several numerical problems. Several ALE models are available in the open literature. Using an ALE approach one needs to understand the various options in order to reach the best results. The combination of Lagrangian and Eulerian formulations has been utilized in the current model to achieve the benefits of both formulations. The study involves the turning of AISI 4140 steel using a cutting tool made of carbide material. All the material properties are extracted from previously published work, including the Johnson-Cook parameters. The effect of initial chip geometry, feed rate of friction coefficient on cutting forces, stresses, strains, temperature, and formed chip geometry have been studied. Model solutions were obtained by using the commercially available finite element package ABAQUS/Explicit, version 6.7. The ii

model verification is accomplished by comparing the predicted results to published experimental results. The current study showed that the effect of the initial chip height does not have major effects on the results. The new formulation with no initial chip is shown to give reasonable prediction of cutting force, feed force and chip thickness. To date all simulations underestimate the chip contact length. Friction behaviour at the chip-tool interface is one of the complicated subjects in metal cutting that still needs a lot of work. Several models have been presented in the past with different assumptions. In the current model, the Coulomb friction model, which assumes a constant friction coefficient, is used to model the friction in order to simplify the model. The effect of the constant friction model is considered by analyzing the results for several friction coefficient values and comparing them to the previous work. The comparison illustrates some weak points in this model that need to have more study.

in

ACKNOWLEDGMENTS
I would like to express my feelings and gratitude to my great supervisors Dr. Hossam A. Kishawy and Dr. Robert J. Rogers for their support and guidance and thank them for offering me the chance to work under their supervision. I truly appreciate their time that is offered to me whenever I need their assistance and their assistance in writing my thesis. I would like to thank all members of my committee, Dr. Z. Chen, Dr. A. Gerber, and Dr. A. Schriver for their review and helpful suggestions. My appreciation goes to the Department of Mechanical Engineering, faculty and staff for their help. Also, I would like to thank my great friend Lei Pang for his help and suggestions. I do not forget to thank the Libyan Embassy for supporting me to finish my study. Finally, I would also like to thank my family for the support they provided me through my entire life and in particular, I must acknowledge my wife for her patience, and assistance. With out all, I would not have finished this thesis.

IV

I would like to dedicate this thesis to my great father, Mohamed Maftah, and my great mother, Amina Salm who are trying to offer me the best in all my life, to my wife Sara Elfatouri for her patience, and last but not least to my little boy whom I am excited to see soon.

TABLE OF CONTENTS
Abstract Acknowledgments List of Figures List of Tables Nomenclature
u

1V

X1

xv

XV1

1- INTRODUCTION
1-1 1-2 1-3 1 -4 Motivation and Background Scope of Work Thesis Objectives Thesis Outline 1 2 2 3

2- INTRODUCTION TO METAL CUTTING


2-1 2-2 2-3 2-4 Introduction Machining Geometry Orthogonal Cutting Forces in Metal Cutting 5 6 11 12

3- LITERATURE REVIEW
3-1 3-2 Introduction Finite Element Formulations Lagrangian formulation Eulerian formulation 15 16 16 18

3-2-1 3-2-2

vi

3-2-3 3-3

Arbitrary Lagrangian Eulerian formulation

19 21 23 27 29 31

Friction Models Friction characteristics Albrecht's Coulomb friction coefficient

3-3-1 3-3-2 3-4 3-5

Heat Generation and Deformation Zones Residual Stresses

4- GOVERNING EQUATIONS OF THE NUMERICAL MODEL


4-1 4-2 4-3 4-4 4-5 Introduction Coupled Thermal - Stress Analysis Equations of Motion Flow Stress Heat Generation Convection heat transfer Conduction heat transfer Friction energy and gap conductance definition 33 34 34 36 38 38 39 39 41 41 42 43 44 44

4-5-1 4-5-2 4-5-3 4-6

Friction Characteristics Simple Coulomb friction definition Shear limited friction factor

4-6-1 4-6-2 4-7

Contact Algorithms Kinematic algorithm Penalty Algorithm

4-7-1 4-7-2

vii

5- FINITE ELEMENT MODELLING


5-1 5-2 5-3 Introduction Model Definition and Assumptions Material Properties Tool material properties Workpiece material properties Cutting conditions 46 47 48 48 48 50 50 50 51 54 58

5-3-1 5-3-2 5-3-3 5-4

Modelling Description Elementtypes Geometry and boundary conditions Mesh and chip formation Contact algorithms

5-4-1 5-4-2 5-4-3 5-4-4

6- RESULTS AND DISCUSSION


6-1 6-2 Introduction Effect of the Initial Chip Geometry Chip thickness Contact length Cutting and feed forces 60 63 66 68 70 74 74 77 77

6-2-1 6-2-2 6-2-3 6-3

Results with No Initial Chip Chip formation Cutting and feed forces Stress and strain distribution

6-3-1 6-3-2 6-3-3

viii

6-3-3-1
r

Von-Mises stress distribution Shear stress distributions Normal and friction shear stress along the chip tool interface Equivalent of the plastic strain

78 __
/y

~- ~

6-3-J-z

6.3.3.3 6-3-3-4 6-3-4

81 82 82 83 86 gg g7 go g\ 93 95 97 99 103

Temperature distributions Temperature distribution in the chip Temperature distribution along the rake face

6-3-4-1 6-3-4-2 6-4

Effect of Friction Factor Contour stress, strain and temperature distribution Von-Mises stress distributions Distribution of shear stress Distribution of equivalent plastic strain Distribution of temperature

6.4. i

6-4-1-1 6-4-1-2 6-4-1-3 6-4-1-4 6.4-2 6-4-3 6-4-4 6-5

Chip thickness Contact length Cutting and feed forces

Effect of Mass Scaling

7- CONCLUSIONS
74 7_2 7.3 7.4 Summary Conclusions Contributions Recommendation for Future Work
105

107 108 108

ix

REFERENCES APPENDIX A
INPUT FILE FOR ABAQUS EXPLICIT Curriculum Vitae

LIST OF FIGURES
Figure 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 3.1 3.2 3.3 Orthogonal cutting geometry Oblique cutting geometry Schematic illustration of two-dimensional orthogonal cutting Thin shear plane model Thick shear plane model Pisspanen's shearing process Velocity diagram Forces acting on a cutting tool in two-dimensional cutting Lagrangian definition Eulerian definition Arbitrary Lagrangian Eulerian (a) undeformed shape, (b) deformed shape 3.4 Explanation of contact between two surfaces (a) Two bodies with friction after applying the load (b) Free body diagram for the block on a rough surface 3.5 3.6 3.7 3.8 3.9 3.10 4.1 Variation of the friction force between two bodies Distribution of normal and shear stress at chip-tool interface Force decomposition in the Albrecht's model Corresponding cutting for different feeds Definition of the critical feed rate Heat transfer definition in metal cutting and the deformation zones .... Gap conductance model 23 25 27 28 29 31 40 22 Page 7 7 8 9 10 11 12 13 17 19 20

XI

4.2 4.3 5.1 5.2 5.3 5.4 6.1 6.2

Coulomb friction model with a limiting shear stress Master and slave surfaces in contact Four-node solid element Two-node rigid element Boundary conditions and partition scheme for the previous model .... Boundary conditions and partition scheme for the new model Schematic of both models Chip formation of orthogonal machining at different times with initial chip (f = 0.2 mm, h = 0.5 mm, V = 200 m/min, \i = 0.23)...

43 45 51 51 52 54 63 65

6.3 6.4 6.5 6.6 6.7 6.8

Chip thickness and contact length measurement Chip thickness obtained by different models Percentage error ofthe chip thickness for different models Comparison between the models for the contact length Percentage error of the contact length Cutting forces versus time for all initial chip height cases (f = 0.2 mm, V = 200 m/min, [i = 0.6)

66 67 67 69 69 71

6.9 6.10 6.11

Comparison between the models for the cutting forces Percentage error of the cutting force Feed forces versus time for all initial chip height cases (f = 0.2 mm, V = 200 m/min, n = 0.6)

71 72 72

6.12 6.13 6.14

Comparison between the models for the feed forces Percentage error of the feed force Chip formation of orthogonal machining at different times with no initial chip (f =0.3 mm, V = 200 m/min, u. = 0.23)

73 74 76

6.15

Cutting and feed force versus time (f= 0.3, V = 200 m/min, u, = 0.23).

77

xii

6.16 6.17 6.18

Distribution of Von-Mises stresses in the chip and the workpiece (Pa). Distribution of shear stresses in the chip and the workpiece (Pa) Normal contact pressure and friction shear stress distribution over the rake face (f=0.3, V=200 m/min, u = 0.23)

78 79 80

6.19

Friction shear stress and normal contact pressure for the Coulomb friction model identification

80

6.20 6.21 6.22 6.23

Distribution of equivalent plastic strain in the chip and the workpiece.. Distribution of temperature in the chip and the workpiece Distribution of temperature in the tool Temperature distribution on the rake face (f = 0.3, V = 200 m/min, u = 0.23)

81 82 83 84

6.24 6.25 6.26

Location of the selected nodes in the rake face of the tool Rake face temperature versus cutting time Contour plots of Von-Mises stress for different coefficients of friction (f= 0.3 mm, V = 200 m/min)

85 86 88

6.27

Contour plots of shear stress for different coefficients of friction (f = 0.3 mm, V = 200 m/min)

90

6.28

Contour plots of equivalent plastic strain distribution for different coefficients of friction (f = 0.3 mm, V = 200 m/min)

92

6.29

Contour plots of temperature distribution for different coefficients of friction (f= 0.3 mm, V = 200 m/min)

94

6.30 6.31

Chip thickness obtained for the experimental and numerical models ... Percentage of error of the obtained chip thickness for numerical models

96 96

6.32

Contact length along the chip-tool interface obtained for the experimental [4] and numerical models

97

6.33

Percentage error for contact length for numerical models compared to published experimental values

98

Xlll

6.34 6.35 6.36

Measured and predi cted force values for di fferent feeds Percentage of error for the obtained cutting force of the numerical models Percentage of error for the obtained feed force of the numerical models

100 101 101

6.37 6.38 6.39 6.40

Cutting force vs. feed Feed force vs. feed 6.39 Chip thickness vs. feed 6.40 Contact length vs. feed

102 103 104 104

xiv

LIST OF TABLES
Table 5.1 5.2 5.3 5.4 Cemented carbide tool physical properties Workpiece steel AISI4140 physical properties Johnson Cook equation coefficients Cutting conditions Page 48 49 49 50

xv

NOMENCLATURE
General Symbols A B C d Del E F Fs Fc Ff Fn FN Yield stress constant Strain hardening coefficient Strain rate sensitivity Gap clearance Elastic matrix Effusivity Friction force Shear force Cutting force Feed force Force normal to the shear force Normal force Weighting factor for distribution of the heat between interacting surfaces Average of any predefined field variable External body force vector Reference film coefficient Hardness of the metal asperities Gap conductance Conductivity Contact length Mass Normal vector Strain hardening component or Normal force Surface pressure Rate of friction energy dissipation per unit area Heat flow rate per unit area Heat flow rate per unit volume Force between the tool rake face and the chip or the Resultant force Force between the workpiece and the chip along the shear plane Uncut chip thickness (feed) Chip thickness Cutting velocity Chip velocity Shear velocity

f
fr g h H k K
lc

m n N P Pfr q Qpl R R' t


tc

V Vc

vs

XVI

Greek Symbols a Rake angle Friction angle P Rate of plastic strain *P<

^0

Reference plastic strain rate Cauchy stress tensor Static yield stress Yield stress Normal stress Total true elastic stress Equivalent stress Yield stress at nonzero strain rate Shear angle or clearance angle Temperature Melting temperature Reference sink temperature Average temperature Nondimensional temperature Coefficient of friction Shear stress Fraction coefficient of energy converted into heat Slip rate Density

a a y

N
a
(7 eqv

el

<t> e
"melt

e
0 6 M
T

f
P

XVll

CHAPTER 1: INTRODUCTION

CHAPTER 1 INTRODUCTION
1-1 Motivation and Background
The machining process includes the effect of coupling the plastic deformation and the friction zone at the workpiece, chip, and cutting tool. Any study of metal cutting models by a finite element method should consider some parameters such as simulation geometry and material properties. During machining, the material will reach a high temperature and therefore the finite element method considers how the analysis includes the changes in temperature. The flow stress can be determined by a combination of the temperature, strain and strain rate. The Johnson-Cook model can include the aforementioned to calculate the flow stress. Most researchers usually make friction

CHAPTER 1: INTRODUCTION
assumptions based on the experimental data. Many researchers have come up with

different techniques to model the chip-tool interface. One of these studies has been done by Arrazola el al. [1, 2]. They divided the chip tool interface into two parts and were able to finally obtain good agreement of the cutting and feed forces, as well as the chip thickness. Haglund [3] continued this work by developing a finite element model with a range of friction models. In all cases, the simulations underestimated the chip-tool contact length. As well there was some concern that the choice of the initial chip geometry may have affected the results somewhat. These results provided motivation for the present work.

1-2 Scope of Work


The goal of this thesis is to evaluate the role of initial chip geometry and, if possible, to develop a finite element model where there is no initial geometry of the undeformed chip. This model is performed as a two-dimensional Arbitrary LagrangianEulerian finite element model using ABAQUS Explicit version 6.7. Another goal is to study the effect of the constant friction coefficient on cutting and feed forces, chip thickness, and contact length.

1-3 Thesis Objectives


The specific objectives for the thesis are as follows: To develop a finite element model by using more realistic initial chip geometries using the Arbitrary Lagrangian Eulerian technique. To compare the obtained results of the proposed finite element model with the previously published measured data during the cutting of AISI 4140 [1, 2]. The

CHAPTER 1: INTRODUCTION

comparison includes the feed force, cutting force, chip thickness, and contact length, as well as the shear stress, normal stress, and temperature distribution along the rake face. To study the contact mechanism at the chip-tool interfaces in order to obtain better understanding of the friction behaviour by considering a range of friction coefficients.

1-4 Thesis Outline The thesis consists of seven chapters which are listed below: First, a brief introduction of the principles of metal cutting processing is given in chapter 2. The introduction includes machining geometry, orthogonal cutting, and the force model. In the orthogonal cutting section, a short description of the shear plane and the velocity model is illustrated. In addition, the force model is presented with analysis of all forces that act in the shear plane and the friction plane. Chapter 3 is a literature review of numerical models. First, the finite element formulations are shown with different types of formulations and followed by the friction models, which includes definitions of sticking and sliding. Also, a brief description of the Albrecht assumptions [4] is introduced. Next, the heat generation at the primary and secondary deformation zones is presented. The final point in the literature review section considers the residual stress, which occurs on the product surfaces. Chapter 4 describes the governing equations of the finite element model. These are used in simulations with the Dynamic Temperature Explicit step. As well, the material model and the contact algorithms are explored.

CHAPTER 1: INTRODUCTION

Chapter 5 shows the finite element model in detail. All parameters that are contained in the simulation, such as the material properties, and model geometry are illustrated in this section. In addition, boundary conditions and applied loads are explained. Chapter 6 discusses several cases of results with and without an initial chip. The results are compared to the previously published experimental data. The main comparisons include the chip thickness, contact length, and the cutting and feed forces. The trends in the results for a range of friction coefficients are presented. Chapter 7 includes a brief summary of this work. It is followed by a list of conclusions, contributions and recommendations for future work.

CHAPTER 2: INTRODUCTION TO METAL CUTTING

CHAPTER 2 INTRODUCTION TO METAL CUTTING


2-1 Introduction Metal cutting is the process of removing unwanted material from the workpiece to obtain a part with high quality surfaces and accurate dimensions with acceptable tolerances. This process has represented a very large segment in industry since the last century. It is estimated that 15% of the value of all mechanical components manufactured worldwide is derived from machining operations [5]. The metal cutting process includes different forms of machining processes such as grinding, turning, milling, sawing, etc. For all these types of machining, the productions of chips have different forms and each process has unique chip morphology. Therefore, it is important to understand the mechanism of chip formation in order to understand the machining process.

CHAPTER 2: INTRODUCTION TO METAL CUTTING Many studies have been performed in the area of metal cutting. In the middle of the 19th century, the old (trial and error) experimental method was the earliest way to develop models of the metal cutting process. The simplified models were also presented and used based on the shear zone theory [6]. The chip formation was assumed to take place as the result of shear actions in the shear zone. Later, finite element analysis was utilized, trying to optimize metal cutting processes. This opened a new way to investigate the state of stresses, strains, temperatures, and feed and cutting forces in the deformation zones. These models provide a better understanding of metal cutting and provided ways to do detailed studies of the effect of different parameters where the magnitude of some parameters such as the temperature cannot be easily measured experimentally.

2-2 Machining Geometry Metal cutting processes can be divided into two basic categories: orthogonal and oblique metal cutting. In orthogonal metal cutting, the cutting edge is perpendicular to the relative cutting velocity and also normal to the feed direction, as shown in Figure 2.1. However, in oblique cutting, the cutting edge is inclined at an acute angle to the direction of the cutting velocity as shown in Figure 2.2. During the machining, the tool will be given a certain position to obtain the amount of feed that will be removed from the workpiece. In general, the cutting edge of the tool will engage into the workpiece; therefore, high pressure and high temperature will occur at the front of the tool.

CHAPTER 2: INTRODUCTION TO METAL CUTTING


Chi Tool

Cutting edge a

Motion of Workpiece Workpiece

Figure 2.1 Orthogonal cutting geometry


Tool Cutting edge axis

Cutting edge inclination

Motion of Workpiece Workpiece

Figure 2.2 Oblique cutting geometry The easiest way to present the fundamentals of the orthogonal metal cutting process is by the two dimensional metal cutting geometry as shown in Figure 2.3. As the workpiece starts moving, the cutting edge penetrates into the workpiece and forces the chip to grow up so that the chip will be formed and move along the rake face of the tool.

CHAPTER 2: INTRODUCTION TO METAL CUTTING This process causes high pressure and plastic deformation is expected to take place in front of the cutting edge. The shape of the formed chip will be affected by the cutting conditions (cutting speed, feed and depth of the cut), tool geometry and material properties.

Motion of chip

i
Chip ,--'<X Cutho toof

Or itfinal surface R M C face

Motion of Workpiece ,

_____

..'

, A

ti

:':''.'"'
"' "' " ' ""'""" *

J. '

Batik fees Produced surface

0*
*" '

>

Shear deformation

zone
Cutting edge

Figure 2.3 Schematic illustration of two-dimensional orthogonal cutting

The uncut chip thickness t is known as the feed while the deformed chip has a different chip thickness ^c. The tool will be defined by rake face angle a and flank angle/? . The rake angle is defined to be positive on the right side (clockwise from vertical) and negative on the left side (counter clockwise). The contact length lc is defined as the distance from the tip of the tool to the point where the chip loses contact with the tool on the rake face. The friction between the chip and the tool plays a significant role in

CHAPTER 2: INTRODUCTION TO METAL CUTTING the cutting process because of the heat energy that is transferred into the workpiece. It may be reduced by optimized tool geometry, tool material, cutting speed, rake angle, and cutting fluid. Because of the high pressure and temperature, a built up edge (BUE) may exist near the tool tip. As a result of the built up edge, welded material will become a part of the cutting tool and may lead to tool wear. The shear angle if) is affected by the welded material so the size of the welded material grows until it reaches a critical size. Then, it breaks and starts the new welded material. In orthogonal machining the shearing action takes place along the shear plane so the chip will start to flow over the rake face. The shearing zone has been modelled using either one of two assumptions. Merchant [7] developed an orthogonal cutting model by assuming the shear zone to be thin as shown in Figure 2.4. Once the material approaches the shear plane, the plastic deformation begins. A thin shear zone is usually created at high cutting speeds.

Figure 2.4 Thin shear plane model Some other researchers had different assumptions where the shear zone would be thick as shown in Figure 2.5. This kind of shear zone is more complicated and normally seen when using low cutting speeds.

CHAPTER 2: INTRODUCTION TO METAL CUTTING

Thick shear plane

Workpiece

Figure 2.5 Thick shear plane model Both models have been used to analyze metal cutting processes where the thin shear zone relates to the shear plane angle, cutting condition, material properties, and friction behavior, while the thick shear zone model is based on the slip-line theory [6]. Many researchers have focused on chip formation. One side of the chip is in contact with the rake face and as result of the relative motion and friction it forms what is called the secondary shear zone. On the other side of the chip, the free surface is mainly affected by the primary shear zone. Because of the high speed of machining, the primary zone will have high pressure and temperature. From the geometry shown in Figure 2.3, the cutting ratio can be calculated from this equation [6]:
r =L =

tc

ABsint AB cos(^ - a)

(21)

where t is uncut chip thickness, tc is the deformed chip thickness, a is the rake angle and <j) is the shear angle which can be determined such that [6]: rcosor tan^ = 1 - r sin a (2.2)

10

CHAPTER 2: INTRODUCTION TO METAL CUTTING 2-3 Orthogonal Cutting


The single shear plane model was proposed to explain the chip formation in the metal cutting process. From the earlier approaches the concept of shear plane has been developed analytically in several models such as Pisspanen's model [6]. He described the shearing process as a deck of cards such as when the first card slides forward, it will be followed by the second card and so on as far as the cutting process keeps going. See Figure 2.6.

Chip

1
Parallel shear cards

(J)

Tool

Shear plane

Figure 2.6 Pisspanen's shearing process [6] The main velocity components in metal cutting can be seen in Figure 2.7. The velocity of the cutting tool relative to the workpiece is known as cutting velocity (V ) . The velocity of the chip relative to the tool is known as shear velocity (V s ). The velocity of the chip relative to the workpiece is called the chip velocity (V c ).

11

CHAPTER 2: INTRODUCTION TO METAL CUTTING

90-(a-^)

^Is

Ir
V
Figure 2.7 Velocity diagram [6]

\T

The velocity diagram shows the summation of the cutting velocity and the chip velocity equals to the shear velocity. V^ and V~ are given by:

V, C

sin^ cos I m - a
COS Of

(2.3)

S =

COS

(4-a)

(2.4)

2-4 Forces in Metal Cutting


Knowing the forces that are acting in metal cutting is important for many reasons such as for the power requirement. Some parameters including the cutting speed, feed, and the depth of the cut influence the forces. Most likely, the forces can be reduced to two main forces in 2-D instead of three forces in 3-D. There are two main forces we can consider in orthogonal metal cutting. The force between the rake face and the chip (R),

12

CHAPTER 2: INTRODUCTION TO METAL CUTTING and the force along the shear plane (R'). The two forces R and R' are equal but opposite to each other. These forces are decomposed into three sets as illustrated in the free body diagram of the chip shown in Figure 2.8. The study of these forces can help us to estimate the power requirement, tool geometry and material properties of the tool. In general, the horizontal and vertical forces are called cutting force component (Fc) and feed force component (Fj), respectively. In the shear plane the force (R') can be resolved into the shear force (Fs) and the normal force to the shear force (Fn). On the rake face the force (R) can also be resolved into two components: friction force (F) and normal force (TV).

Chip

Figure 2.8 Forces acting on a cutting tool in two-dimensional cutting

13

CHAPTER 2: INTRODUCTION TO METAL CUTTING

In 1945, Merchant developed the most popular analytical model used in metal cutting. The relations among the shear and friction components of forces in terms of cutting and feed force can be obtained as follow [6]: Fs = Fc c o s (/> - Ff sin </> Fn = Ff c o s </)-Fc sin <f> F = Fcsina + Ffcosa N = Fccosa-Ff where Fc and Ff since (2.5) (2.6) (2.7) (2.8)

are the cutting and feed forces, respectively; Fs is the shear force, Fn the normal

is the normal force along the shear plane; F is the friction force and force along the rake face.

14

CHAPTER 3: LITERATURE REVIEW

CHAPTER 3 LITERATURE REVIEW


3-1 Introduction Finite Element Models (FEM) have been involved in manufacturing fields because of the advantages that can be achieved from modelling processes such as metal cutting. The FEM has the ability to solve complicated calculations of metal cutting by detailed modelling of parameters such as the material properties and friction characteristics. For example, the material properties can be incorporated by the JohnsonCook formula, which contains strain, strain rate, and temperature. Creating FEM with different formulations can provide better results. Researchers have utilized two formulations to model orthogonal metal cutting. The comparison shows some weak points in each model. The latest formulation was created by combining the

15

CHAPTER 3: LITERATURE REVIEW


two formulations, Lagrangian and Eulerian, and is called Arbitrary Lagrangian Eulerian. The advantages of FEM are explained in the following section. As explained above, complicated friction is involved in the FEM of orthogonal metal cutting to study the interactions between the surfaces of two different bodies, the tool and workpiece. The reality shows that friction behavior is hard to estimate in machining. Some other models present the friction coefficient in terms of shear limit and temperature dependence. Friction characteristics require more studies because the obtained results of all models show some weak points. The temperature effect is one of the most important parameters that might cause trouble to the product surface and the tool as well. The temperature changes result in thermal stresses in the workpiece so that many studies focus on the residual stresses where the stress remains in the product surface after the machining.

3-2 Finite Element Formulations


The specific mesh formulations used for models of orthogonal machining are Lagrangian, Eulerian, and Arbitrary Lagrangian Eulerian. The advantages and disadvantages of these formulations will be discussed in this section as follows.

3-2-1 Lagrangian formulation The Lagrangian or updated-Lagrangian method is often used in FEM. These formulations are similar. The only difference is that updated-Lagrangian uses an adaptive mesh technique to reduce mesh distortion. The updated-Lagrangian formulation was first used for machining model by Klamecki in 1973 [8].

16

CHAPTER 3: LITERATURE REVIEW

The basic concept of the Lagrangian definition is that the mesh will follow the material. The deformation can happen by increments in time. After each increment the reference domain is updated based on material coordinates. In this way, the history of the material is easily taken into account. This updated position situation is used as an initial condition for the next increment, so the FE mesh is connected with its material. However, the updated-Lagrangian method can be costly because of the mesh distortion during the large deformation in these calculations (see Figure 3.1).

I
2

=M
(a) Initial Mesh (b) Deformed Mesh Figure 3.1 Lagrangian definition Carroll et al. [5] formulated two models. The first one was Updated-Lagrangian and the second was Eulerian. The updated-Lagrangian model successfully determined the deformed chip, stress, and the temperature under the failure criteria to control the chip formation process. Shih and Yang [9] have developed an FEM for metal cutting based on Updated-Lagrangian which includes the effect of elasticity, visco-plasticity, temperature, strain rate and the effect of frictional force. Marusich and Ortiz [10] presented an interesting FEM analysis of machining, based on a Lagrangian formulation with adaptive mesh processing for modelling high speed machining work. Recent work with the

17

CHAPTER 3: LITERATURE REVIEW

Updated-Lagrangian formulation was done by Ozel in 2005 [11]. He used this formulation to simulate a continuous chip formation process in orthogonal cutting by presenting different friction models based on the experimental data. 3-2-2 Eulerian formulation Another FEM option is to use the Eulerian formulation. The simple definition of the Eulerian formulation is that the mesh will be fixed in space and the material will flow through the mesh. The advantage of the Eulerian formulation is that this formulation does not have any mesh distortion because the mesh is spatially fixed during the simulation (see Figure 3.2). However, the mesh does not connect to the material. It is difficult to obtain accurate data from free surfaces, which is an important result of the simulation of the forming process as seen in Figure 2.3. Some researchers [12, 13, 14] used the Eulerian formulation model in metal cutting because this model can use fewer elements and reduce the time of the analysis. Leopold et al. [27] has developed an Eulerian analysis of 3D oblique machining with a single cutting edge. Carroll et al. [5] compared the Eulerian model with Lagrangian model and they found that the advantage of the Eulerian model is that it does not need failure characteristics. Also, the mesh cannot have high distortion. Strenkowski an Moon [14] built an FEM of orthogonal metal cutting with an Eulerian formulation. The model predicted temperature distribution. He found that the shear stress occurred over a finite region in front of the tool. In the same year, Childs and Maekawa [13] used the Eulerian formulation to create an FEM to study the tool wear of cemented carbide tools in high speed machining. The results of the model were very good except there are small percentage of errors in the cutting forces.

18

CHAPTER 3: LITERATURE REVIEW

(a) Initial Mesh

(b) Deformed Mesh

Figure 3.2 Eulerian definition

3-2-3 Arbitrary Lagrangian Eulerian formulation Recently, researchers have been focusing on the Arbitrary Lagrangian Eulerian (ALE) formulation to combine the best features of both the Lagrangian and Eulerian formulations. The concept of ALE was first proposed lately. This formulation was called "the coupled Eulerian-Lagrangian method" and later on was changed to "the Arbitrary Lagrangian Eulerian Model." The ALE method was introduced into the finite element method by Belytschko and Kennedy [15]. It was applied to finite strain deformation problems in solid mechanics. The ALE formulation helps to solve problems with large deformation in solid mechanics The ALE adaptive meshing is a helpful feature that can smooth the deformation throughout the analysis by allowing the material to flow with the mesh. The mesh has some limitations during the analysis; however, ALE helps to avoid these limitations. The ALE formulations shown in Figure 3.3 uses re-meshing techniques to keep the analysis

19

CHAPTER 3: LITERATURE REVIEW

going. Other formulations such as the Lagrangian formulation have difficulty avoiding high distortion of the workpiece.

(a)

(b)

Figure 3.3 Arbitrary Lagrangian Eulerian (a) undeformed shape, (b) deformed shape Wang and Gadala [16] investigated the ability and the accuracy of mesh formulation. They ended with the conclusion that many problems occurred during extensive mesh distortion, load fluctuation, and inaccurate description at the boundary condition with a corner (tool tip). With ALE, most problems have been avoided; however, at the time, ALE was not developed carefully for solid mechanics problems. Olovsson et al. [17] developed FEM by using the ALE formulation so that the large strain that is caused from the high deformation in metal cutting does not affect the element distortion at the tool tip. Movahhedy et al. [18] presented that the arbitrary LagrangianEulerian (ALE) formulation offers the most efficient modelling approach. He included the features of an ALE analysis of the cutting process in his conclusion. Movahhedy et al. [19] focused on the chamfered cutting edge in their FEM simulation that utilized ALE. The results of the simulation show that the chamfer angle does not have a large effect on

20

CHAPTER 3: LITERATURE REVIEW


the chip removal process. The successful results of the ALE motivated many researchers, such Arrazola et al. [1] who studied the friction on the chip tool interface, to keep developing this model. After that, the majority of researchers have followed the ALE procedure to build new ideas into their research. In recent work, Kishawy et al. [20] in 2006 considered the effects of different cutting edge radii on the rake face of the tool by using the ALE formulation. They concluded that the cutting edge significantly affects the cutting forces, the chip thickness, chip contact and the temperature.

3-3 Friction Models


A general conception of friction can be considered as the tangential force generated between two surfaces. Friction can be represented as a resistance force acting on the surface to oppose slipping. Figure 3.4 (a) shows a simple example of friction where a block is pushed horizontally with mass m over rough horizontal surface. As showing in the free body diagram, Figure 3.4 (b), the body has distributions of both normal force N and horizontal force/along the contact surface. From the equilibrium, the normal force N acts to resist the weight force of the mass mg and the friction force/acts to resist the force F.

21

CHAPTER 3: LITERATURE REVIEW

mg

mg

(a)

(b)

Figure 3.4 Explanation of contact between two surfaces (a) Two bodies with friction after applying the load (b) Free body diagram for the block on a rough surface

Basically, there are two types of friction, which are static and kinetic as shown in Figure 3.5. By increasing the force F, friction force /increases too. The blocks cannot move until the force F reaches the maximum value. This is called the limiting static factional force. Increasing of the force F further will cause the block to begin to move. In the static portion, the limiting friction force can be expressed as:
F

a,ic=V,N

(3.1)

where jus is called the coefficient of static friction When the force F becomes greater than Fstaljc, the frictional force in the contact area drops slightly to a smaller value, which is called kinetic frictional force. Machining models generally just consider the kinetic friction coefficient which can be calculated by the following equation:

22

CHAPTER 3: LITERATURE REVIEW

J1 kinetic ~

Mk^

(3.2)

No motion

Motion Force required to start sliding Kinetic friction

o
a) o c
"55 w

-p

Applied force (F) Figure 3.5 Variation of the friction force between two bodies

3-3-1 Friction characteristics The contact region and the friction coefficient at chip-tool interface are affected by parameters such as feed rate, cutting speed, and rake angle. The reason for this effect is that high normal pressures act on the surface. Many researchers have tried to explain what happens at the chip-tool interface where different friction models were employed in finite element models [11, 20, 21]. Some of them investigated reliable predictive models based on experiments. Other researchers such as Johnson [22] summarized different models and applications to model dynamical contact problem with friction.

23

CHAPTER 3: LITERATURE REVIEW

Friction at the tool-chip interface is a complicated problem. It is hard to estimate the fiction coefficient at the chip-tool interface based on the relationship between the normal stress and the shear stress. The basic Coulomb friction model is stated as the relation between the friction force and the normal force. From the metal cutting geometry, the Coulomb friction coefficient can be calculated from the measured cutting and feed forces as an average value follows [6]: F r + Fc tan a Fc - Ff tan a where Ff is the feed force, Fc is the cutting force, and a is the rake angle. The best way to obtain the friction coefficient at the tool-chip interface is to directly measure the normal and shear stresses during the actual metal cutting process. Usui and Takeyama [23] studied the distribution of the normal and shear stresses along the tool face by using the photo elastic method at low speed. They found that the shear stress was constant over half of the contact length at the chip-tool interface (sticking region) and then decreased to zero (sliding region). Most popular FEM have been developed based on the basic Coulomb friction law: the friction force is proportional to the normal load. Merchant and Zlatin [6] defined the friction coefficient along the chip sticking and sliding regions. Over the chip-tool interface at the sticking region is used a constant shear stress x. Over the remaining sliding region, the shear stress can be calculated using the friction coefficient ju. The normal and shear stress distributions can be illustrated in the two regions. See Figure 3.6.

24

CHAPTER 3: LITERATURE REVIEW

n .

T
f,ti!t)

"X.
{^swT***

Figure 3.6 Distribution of normal and shear stress at chip-tool interface The values of the shear stress can be calculated in the sliding and sticking zones such that: T = jUa when //O" < Tmax (sliding)
T = r

(3.4) (3.5)

max when Ma - rmax

(sticking)

Previously, FEM of metal cutting used the friction coefficient as a constant based on Coulomb's law over the entire chip-tool interface [8, 24]. Some researchers assumed

the limit of shear stress to be 7max - j= where ov is the yield stress [24, 25]. The shear stress along the chip tool interface can be calculated from equations 3.4 and 3.5 but the problem is how the sticking and the sliding regions can be defined analytically.

25

CHAPTER 3: LITERATURE REVIEW

Other researchers developed Coulomb friction models based on dependent temperature [3, 26, 27]. The critical shear stress would be modelled as function of the temperature along the chip tool interface. Astakhov and Outeiro [28] presented a model to show the contact stress distribution at the chip tool interface. This study includes a comprehensive investigation of many attempts to define the stress distributions. After the comparison between the FEM and the experiment, the normal and shear stresses were not uniform. Studying the stresses along the chip tool interface helps to understand the behaviour of the friction coefficient. Some models have been formed based on variable shear friction along the chiptool interface. One of the models was simulated by Usui and Shirkashi [23] who derived the empirical stress characteristic equation as a variable friction model as follows:

- k 1-exp \ k )

(3.6)

where k is the shear flow stress, // is the friction coefficient obtained from experiments,
T

and &N are the shear and normal stresses, respectively. The idea of variable shear

friction was developed by several researchers such as Ozel et al. [11] who extended further modification to equation (3.6) so that the sticking stresses in that region can be different from the shear strength as follows:

f
T

\n wk J
(3.7)

= wk 1-exp

26

CHAPTER 3: LITERATURE REVIEW

where w and n are correction factors. The correction factors help to keep the friction stress less than the shear flow stress of the material. 3-3-2 Albrecht's Coulomb friction coefficient In order to define the Coulomb friction coefficient, Albrecht's analysis has been used to estimate the coefficient of friction along the chip-tool interface by eliminating the cutting edge effect [4]. Figure 3.7 illustrates the basic concept of Albrecht's model.

Undeformed chip thickness

Workpiece

Figure 3.7 Force decomposition in the Albrecht's model [4] The forces are resolved into two components where P is close to the cutting edge and Q is applied on the rake face. With the sharp cutting tool, the ploughing force P has insignificant value. But for the tool that is not sharp, the force P will affect significantly the force model. For uncut chip thickness greater than the critical uncut chip thickness, Albrecht assumes that the force P has a constant value; however, at feeds less than the critical uncut chip thickness, the force P will affect the feed force significantly. After passing the critical chip thickness, the force P slightly affects the feed force. Example

27

CHAPTER 3: LITERATURE REVIEW

feeds and chip thickness are shown in Figure 3.8. The sum of the two force components (cutting and feed) can be obtained by the sum of two vectors P and Q.

Figure 3.8 Corresponding cutting for different feeds Figure 3.9 illustrates the cutting force and the feed force relation at different uncut chip thicknesses. At the smallest feeds in Figure 3.9 (A and B sections), a non-linear relation will describe the behaviour of the cutting and feed forces. Below the critical point, the P force will cause a relatively large feed force. The section C where the relation takes a linear behaviour is used to approximate the value of the Coulomb friction coefficient. The friction coefficient along the chip tool interface can be defined by taking the slope of section C as tan (A - a) and then ju = tan X [ 1 ].

28

CHAPTER 3: LITERATURE REVIEW

Figure 3.9 Definition of the critical feed rate [3, 4] Arrazola et al. [1, 2] used the previously presented model by Albrecht [4]. In this study the friction was defined at a variety of feeds. Arrazola et al. [1, 2] applied the dual friction idea to the chip-tool interface. More recently, these studies were followed by different investigations. Some other researchers also used the idea of the dual friction model [3].

3-4 Heat Generation and Deformation Zones


During machining, high pressure and shear stress occur in the contact surface. Most of the plastic deformation energy is converted into heat, which is usually approximated at 90% [20, 21, 29]. The mechanical work that is done in the machining process in the primary deformation zone can be predicted analytically or measured experimentally. The higher temperature that exists in the secondary deformation zone is caused by the hard contact and friction. Because of raised temperature in metal cutting,

29

CHAPTER 3: LITERATURE REVIEW

the heat energy will influence the tool wear, tool life, and chip formation [11, 30]. Increasing the temperature of the workpiece at the primary deformation zone will soften the material. Also, the temperature at the secondary deformation zone will affect the contact process at the chip tool interface. Finally, the heat generated in the tertiary deformation zone will influence the produced surface as shown in Figure 3.10. Temperatures in the primary and secondary zones are mainly affected by the cutting conditions [6, 31]. Measuring temperature during metal cutting is extremely difficult but some researchers are trying to develop ways to measure the temperature of the models experimentally. Blok [32] was one of the researchers who developed energy partition analysis to study the heat sources in metal cutting. The measured temperature due to machining was obtained from a thermal imaging camera [32]. Basically, the assumption is that all the mechanical work done in the machining process is converted into heat. Increasing the temperature in the workpiece, chip, and tool will affect the product surface, since the temperature has an effect on the surface quality and the tool wear. Some of the heat will be removed from the primary and secondary zones by the chip (see Figure 3.10). The temperature at the tool will be raised in the secondary heat zone. Also, the heat that comes from the primary deformation zone will affect the temperature of the cutting tool, so that part of the heat generated at the shear plane will be carried by the chip through the rake face into the tool.

30

CHAPTER 3: LITERATURE REVIEW

Convection Boundary

Tool Holder

Primary deformation zone Tertiary deformation


Z O n e

,A,

Workpiece Figure 3.10 Heat transfer definition in metal cutting and the deformation zones [31]

3-5 Residual Stresses


Due to the thermal effect in orthogonal metal cutting, an important subject has been noticed inside the workpiece which is residual stress. Some of the heat energy comes from the plastic deformation zones and the thermal stress goes to the surface layer of the workpiece. Genzel [42] studied the residual stress by creating a numerical model and compared the obtained results with the experimental data obtained with X-rays. He found that the influence of the cutting speed and feed causes tensile residual stress over the productive surface of the workpiece. Liu and Guo [25] investigated the effect of sequential cuts and the friction at the chip tool interface on residual stresses in the machined layer. They found that the residual stress is sensitive to the friction at the chip tool interface. In their analysis, they performed more than one cut to predict the effect of sequential cuts on the residual stress. M'Saoubia et al. [32] considered the residual 31

CHAPTER 3: LITERATURE REVIEW

stresses in a wide range of cutting feeds including cutting speed, feed rate, tool geometry and tool coating. By taking consideration the hardness of the material, they found a high tensile stress on the workpiece surface. This tensile stress is caused by local thermal effects that are generated from heat energy. Some other influences could come from the mechanical and feed rate effects [33, 34]. Hua et al. [35] focused on the effects of workpiece hardness by using a newly proposed hardness flow stress model. Also, they included different cutting edge shapes such as sharp edge, honed, and chamfered and different cutting conditions. They ended with a brief presentation of residual stresses in the axial and circumferential directions of the machined surface. A more compressive residual stress was achieved at higher workpiece hardness. In addition, a larger hone radius or a chamfered edge generated more compressive residual stresses; however, the effect of the chamfered was less than the honed. Nasr et al. [36] in 2006 studied the effect of the tool edge using an Arbitrary Lagrangian Eulerian formulation. They found that the cutting edge geometry had a major effect on the cutting process. As a result, a higher tensile residual stress in the produced surface layer was caused by a high cutting edge radius. With a large cutting edge radius, a high compressive residual stress was generated beneath the surface. The maximum compressive residual stress occurred deeper into the workpiece surface. The development of non-sharp cutting edge caused a stagnation zone underneath the tool.

32

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL

CHAPTER 4 GOVERNING EQUATIONS OF THE NUMERICAL MODEL


4-1 Introduction
An understanding of the system solution is important to form a complete input file for ABAQUS Explicit. With its many of options, ABAQUS can model the very complicated conditions of metal cutting. It is important to understand how to use its tools which are available such as element types, material models, mesh density (coarse or fine), adaptive meshing techniques, boundary conditions, etc. Stress and thermal characteristics at the contact area are the greatest challenges with this type of model. The constitutive equations for the model are described below.

33

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL 4-2 Coupled Thermal - Stress Analysis In the metal cutting process, fully coupled thermal-stress is a requirement for the analysis because the stress depends on the temperature distribution and the temperature depends on the stress solution. The contact condition exists in the secondary deformation zone where the heat is conducted between the chip tool interfaces, depending on the pressure. The thermal and mechanical solutions can be solved simultaneously [3, 37, 38].

4-3 Equations of Motion


The explicit dynamic analysis procedure is based on using very small time steps. The dynamic equations are integrated using the explicit central difference integration method, which uses a diagonal mass matrix. The velocity estimated equation is integrated through the time as follows [37]:

At,. ,, + At,.,
"(1+1/2) ~
M

(i-l/2)

n.

(i)

V*A>

And the displacement is determined as


<.)="(0 + A V.)"( J ) + l/2) <4'2)

where the subscript (i) refers to the increment number, uN represents the displacement vector, and At represents the time increment. The central difference integration operator is explicit in that the kinematic state may be advanced using known values of u^_V2] and u^ from the previous increment.

34

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL The explicit integration rule provides high computation efficiency due to the use of diagonal element mass matrices so that the accelerations at the beginning of the increment may be computed by

^=(M-)-*(F(^)

(4.3)

where M ^ i s the diagonal lumped mass matrix, F J i s the applied load vector, and IJis the internal force vector. The explicit procedure integrates through time by using many small time increments. The explicit forward difference time integration method integrates for the current temperature using the value of integrated as $,)=<+AW&./2)
N

0{" from the last increment. The heat transfer is

(4 4)

"

where 9,., is the temperature at node Nand the subscript /refers to the increment number in an explicit step. The values of 9,", are calculated at the beginning of the increment by ^=(C-)"'(^-^) (4.5)

where CNJ is the lumped capacitance matrix, P,i is the applied nodal vector, and F,i is the internal flux vector. Since both the forward difference and central difference integrations are explicit, displacement and heat transfer solutions are obtained simultaneously so that no iteration or tangent stiffness matrix are required [37]. One of the options with ABAQUS is the time step increment. Chosen time step increments are based on the smallest element to

35

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL reach the stable time for the system. In addition, ABAQUS explicit will take the time step increment that is less than the estimated one with a factor from 1 / v 2 to 1 for two dimensions. Mass scaling is a feature that is used when the finite element model contains very small elements. Many finite element models contain complicated meshes that can have very small elements. These small elements affect the time increment of the system and can cause very small increments. By scaling the mass of these controlling elements at the beginning of the step, the time increment can be increased. In fact, applying the mass scaling should not be over the entire model because increasing the over all mass can influence the accuracy. ABAQUS Explicit has two types of mass scaling: fixed mass scaling and variable mass scaling. Fixed mass scaling is used to define the element masses that are assembled to the global mass matrix. The mass scaling can also be defined as a factor that can be used to reach a desired time increment. Variable mass scaling is used to scale the mass at specified solution times as an addition to any fixed mass scaling that is exists. Mass scaling factors will be calculated automatically and applied through the desired steps.

4-4 Flow Stress Workpiece material deforms extensively along the shear plane and the contact area where the plastic deformation exists with high values. Strain rate dependence has been assumed to model the plasticity region. Considering the strain rate based on the previous work where the parameters are available, the Johnson-Cook plasticity model will be explained in this section. The Johnson-Cook model is the perfect formula that can

36

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL be used to determine the flow stress. It is a function of strain, strain rate, and temperature. This function can be utilized for high strain rate deformation of materials [3,37]. In the plastic regime, flow stress can be expressed primarily as written below:

= *(spl,e) R s
\
.PI

(4.6)

where a is the yield stress at nonzero strain rate, s is the equivalent plastic strain rate,
-pi

fp'\

G\e

,0 is the static yield stress, and R


V

is the ratio of the yield stress at nonzero


J

strain rate. Johnson-Cook strain rate can be presented as [3, 37] 1


C

.pi

=0 exp

(*-l)

for R>\

(4.7)

where is the equivalent plastic strain rate, and s0 and C are the material parameters measured at or below the transition temperature, 9tmnsUion. The yield stress performed can be written as

,pi

(T = IA +

B(P1)") 1 + Cln

(l-#m)

(4.8)

v; where a is the material flow stress, sp the equivalent plastic strain, sQ the reference

plastic strain rate at 9tmnsition, A yield strength, B hardening modulus, C strain rate sensitivity, n hardening coefficient, and m thermal softening coefficient

37

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL


0
0 = \{0Vra^on )/(#/, " ^WoB )

for
M

e<0lr transition
6><rnnsilio <&< 9melt (4.9)

for

e>emeh

where #is the current temperature, 9melt\% the melting temperature, and 0tmmiljm\s the transition temperature defined as the one at or below which there is no temperature dependence on the expression of the yield stress. The material parameters must be measured at or below the transition temperature [37].

4-5 Heat Generation In metal cutting, the heat is generated due to the plastic work in the primary zone and by friction in the secondary shear zones. Due to high speed machining, heat generated does not have sufficient time to diffuse along both bodies, the tool and the workpiece. The plastic strain gives rise to a heat flow rate per unit volume, as shown below [37, 38]: Qpl=rioJs (4.10)

where Qpl is the heat flow rate per unit volume, 7 is the percentage of plastic work 7 transformed into heat which is approximated as 90%, a
.pi

is the equivalent stress, and

e is the plastic strain rate. 4-5-1 Convection heat transfer A convection heat transfer is applied on the free surface of the workpiece and the flank surface of the tool. The convection heat is expected to affect the production surface. q = h[e-B) (4.11)

38

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL where q is the heat flux across the surface, h is a reference film coefficient, 6 is the temperature at the surface, and 9 is the reference sink temperature [37].

4-5-2 Conduction heat transfer Because of the heat energy from the primary and secondary deformation zones, conduction heat transfer occurs inside the workpiece material. The conduction heat transfer can be calculated basically from the equation below q = k(OA-0B) (4.12)

where q is the heat flow per unit area crossing the interface from point A to point B, 6A and 6B are the temperatures at A and B, and k is the thermal conductivity.

4-5-3 Friction energy and gap conductance definition During the contact between the tool and the chip along the secondary shear zone, heat generation and gap conductance models are used at the friction surface. The rate of frictional energy rate per unit area can be calculated as: pfr=T-r (4.13)

where r is the frictional stress and y is the slip rate. The frictional thermal energy that goes into each surface is given as qA=flPfi qB={\-f)r?Pfl. ( 4 - 14 ) (4.15)

39

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL where / is the weight factor, qA is the heat flux into the slave surface (chip), and qB is the heat flux into the tool. The fraction of the heat energy conducted into the chip is given as[43]

f=
fjjy I Hd />

(4.16)

^C,T

yJ^-CjPc.T^CT

(4.17)

where the Kc

is the thermal conductivity, pc

is the density, and Cc T is the specific

heat capacity. The subscript symbol C indicates the chip, and T indicates the tool [3, 37, 43].

The gap conductance can be expressed as seen in Figure 4.1 in terms of the distance d. The maximum conductivity can be reached when the two surfaces are in the perfect contact so there is no gap and the distance d is equal to zero. When the gap appears, the magnitude of conductivity decreases.

<>->d
Figure 4.1 Gap conductance model [4]

40

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL


When the distance d reaches a certain value, the conductivity magnitude can be ignored because of the very small value. The lowest value of the conductivity can be controlled by a constant. The possible variables affects gap conductance are

k = k{0,d,pjr)

(4.18)

where 6 is the average temperature, d is the gap clearance, p is the surface pressure, and / is the average of any predefined field variable.

4-6 Friction Characteristics


Friction plays a very important role in metal cutting. Friction can change many properties because of the heat energy gained from the contact. These effects include the surface quality of the products and the rate of tool wear. Most researchers have used simplistic simulations to present their models with the basic Coulomb friction role [11, 36, 39]. Others choose more complicated models with variable coefficients where the shear limit or temperature dependency is developed.

4-6-1 Simple Coulomb friction definition Basically, the definition of the Coulomb friction model is illustrated as a ratio between the maximum shear force to the maximum normal force that act at the chip tool interface. In this definition, the shear stress can be written in terms of the local forces (shear and normal force) that exist along the chip-tool interface. The sliding shear stress can be simplified as r = MP 41 ( 4 - 19 )

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL

where the ju is the coefficient of the friction and p is the contact pressure at the chip tool interface. The value of the coefficient of friction will affect the friction behaviour between the contact surfaces at the chip-tool interface. In ABAQUS, the default model for the friction coefficient is defined as
jU = jU

yeq,P,6,f^
j

(4.20)

where

Yeq-\Y\^Y2

^s

tne

equivalent

slip rate, p

is the contact pressure, =(/"+ fg ) is

0 = (6A+0B)\s the average temperature at the contact point, and /

the average of a predefined field variable a at the contact point. 6A, 6B, f", and f% are the temperature and predefined field variables at points A and B on the surfaces.

4-6-2 Shear limited friction coefficient The limiting shear stress is one of the friction options available in ABAQUS. The contact pressure at the chip tool interface can be divided into two regions, sticking and sliding. In the sticking region, the shear stress will reach the maximum shear stress as shown in Figure 4.2. In the sliding region, the shear stress is less than the maximum shear stress. The shear stress limit is typically introduced in cases when the contact pressure stress may become very high (as can happen in some manufacturing processes such as metal cutting) causing the Coulomb theory to provide a critical shear stress at the interface that exceeds the yield stress in the material beneath the contact surface. The maximum shear stress is sensitive to the temperature as seen in the Figure 4.2. By

42

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL increasing the temperature, the maximum shear stress decreases, (Tx < T4) .The limiting of shear stress can be estimated as:
(4.25)

r
max

V3~

where av is the Mises yield stress [37]. w w

a>
i_
4

w
CD (D

-C CO

''max a

'1

max a ^

T2 T3

max 3 *

cf 7
^ / /

max

T4

Normal Stress Figure 4.2 Coulomb friction model with a limiting shear stress [3]

4-7 Contact Algorithms


There are two specific contact algorithms in ABAQUS Explicit: Kinematic and Penalty contact that can be used to simulate surface to surface contact. By default, ABAQUS Explicit uses the Kinematic contact algorithm. Penalty contact is the other option that might be used in more general cases of contact between the surfaces.

43

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL 4-7-1 Kinematic algorithm The Kinematic contact algorithm is defined by the forces in a pure master-slave contact. At each increment of the analysis, the algorithm determines which nodes from the slave surface will penetrate the master surface so that it can apply the resistance force (see Figure 4.3). During hard contact, the forces that are applied between the two surfaces caused the slave nodes to exactly contact with the master surface before the penetration occurs. The kinematic method will not affect the time increment as the Penalty method will do.

4-7-2 Penalty algorithm The Penalty contact algorithm is more general so that it is a common option. An additional element will be added to the model where the stiffness for this part is neglected when there is a gap between the two surfaces; however, the stiffness will have high values when contact exists. The stiffness can influence the stable time increment. The Penalty contact method considers one surface as a master and the other as a slave as shown in Figure 4.3. The Penalty contact algorithm tracks the slave nodes that may penetrate the master surface, so the contact applies forces to the slave nodes to prevent the penetration. A finer mesh is recommended on the slave surfaces to minimize the number of nodes of the master surfaces that will penetrate the slave surface. Sliding formulation can be included in contact with three options: finite sliding, small sliding, and infinitesimal sliding and rotation. The small sliding option can be used to linear and nonlinear contacts similar to the finite sliding case. The only difference exists that the slave nodes will interact small local area in the master nodes. The other option is infinitesimal small sliding which is unavailable in this analysis because the

44

CHAPTER 4: GOVERNING EQUATIONS OF THE NUMERICAL MODEL infinite small sliding cannot perform nonlinear geometry. For chip tool interface, using finite sliding is appropriate.

Slave nodes can not penetrate master segments

Penetration

Master surface (segments) *"

Gap

Master node can penetrate slave segment

Figure 4.3 Master and slave surfaces in contact [37]

45

CHAPTER 5: FINITE ELEMENT MODELLING

CHAPTER 5: FINITE ELEMENT MODELLING


5-1 Introduction
Finite element methods have been developed to solve many problems in many applications. In metal cutting, the finite element models have been used to investigate several aspects which include stresses and strain. In the current application, the finite element model is created to analyze a metal cutting problem in terms of dynamic displacement and temperature. After introducing all the governing equations in the previous chapter, the choice of the finite element parameters is explained in this chapter as follows.

46

CHAPTER 5: FINITE ELEMENT MODELLING

5-2 Model Definition and Assumptions A two-dimensional finite element model was developed under the plane strain assumption to simulate the orthogonal metal cutting of steel (AISI 4140) with a continuous chip. ABAQUS can simulate large deformation accompanied by elastic, plastic, thermal, and friction effects. The element type is bilinear (four-nodes) with reduced integration and hourglass control to deal with the large deformation. The dynamic displacement-temperature explicit method was used to analyze this model with a combination of some input parameters such as temperature-dependence and heat transfer. To simplify the analysis, a perfectly rigid cutting tool is assumed because of the significantly high elastic modulus of most tool materials. This should be an acceptable approach since the elastic properties of the cutting tool do not affect the large plastic deformation at the workpiece. To simplify the analysis, it is better to consider some parameters from the beginning so that the goal of the analysis will be achieved efficiently. The model assumptions are considered as follows: 1. The initial temperature of both the workpiece and tool is 25 C (room temperature) 2. The cutting tool is sharp with a 5 urn cutting edge. 3. Constant cutting velocity is equal to 200 m/min. 4. Tool wear is neglected in this study so that the running time will be reduced. 5. The machining model does not include any coolant. 6. Homogenous material is used to model the workpiece material.

47

CHAPTER 5: FINITE ELEMENT MODELLING

5-3 Material Properties This finite element model was developed with two materials: the cutting tool and the workpiece. The cutting tool is modelled with a specific material with high modules of elasticity; whereas, the workpiece material is simulated by the Johnson-Cook model. In the current work the tool is defined as a rigid body with thermal properties.

5-3-1 Tool material properties The tool was made of cemented carbide grade P10. The physical and mechanical properties of the cutting tool material are presented in Table 5.1 [1, 2, 3]:

Table 5.1 Cemented carbide tool physical properties [1, 2, 3].


Density kg.m" Young's modulus GPa Poisson's ratio Specific heat J.kg'.C"' Conductivity W m"1 C"' Expansion urn m"1 C~' 10600 520 0.22 200 25 7.2

5-3-2 Workpiece material properties The workpiece material used for the plane strain orthogonal metal cutting simulation is steel AISI 4140. The physical properties of the used material are given in Table 5.2 [1,2, 3].

48

CHAPTER 5: FINITE ELEMENT MODELLING Table 5.2 Workpiece steel AISI 4140 physical properties [1, 2, 3].
Density kg.m"J Young's modulus GPa Poisson's ratio Specific heat J.kg'.C" 1 7800 210 0.3 473 at 200 C. 519 at 400 C. 561 at 600 C. 1520 0.9 42.6 at 100 C. 42.3 at200 C. 37.7 at 400 C. 33 at 600 C. 12.2 at 20 C 13.7 at 250 C 14.6 at 500 C

Melting temperature C Inelastic heat fraction Conductivity W m"1 C '

Expansion Urn m"1 C"'

As discussed in section 4.4, the Johnson-Cook model is used to model the workpiece material under thermo-viscoplastic behaviour. This model has been used widely because it is suitable to obtain the flow stress as a function of strain, strain rate, and temperature. The Johnson-Cook model is represented in the following equation [1,2, 3,40]:

- = M*")'\

-pi

1 + Cln
V oJ

0-0,
V @melt

ref 0>ef J

(5.1)

And the Johnson-Cook parameters are written in Table 5.3. Table 5.3 Johnson Cook equation coefficients [1, 2, 3, 40]:

Material AISI4140

A(MPa) 598

B(MPa) 768

n 0.2092

C 0.0137

m 0.807

Z0 0.001

49

CHAPTER 5: FINITE ELEMENT MODELLING 5-3-3 Cutting conditions The current work is based on the continuous chip model for different feed rates and friction coefficients to study the effect of the friction coefficients and the initial geometry. Through this work, the cutting conditions are given in Table 5.4

Table 5.4 Cutting conditions [1, 2, 3].


Cutting Conditions Feeds (mm) Cutting Speed (m/min) Rake Angle (degree) Tool Geometry Clearance Angle (degree) Cutting Edge Radius (pm) 0.1,0.2,0.3 200 6 6 5

5-4 Modelling Description


In order to build any simulation with finite elements, the understanding of the mesh formulation, boundary conditions, and the initial geometry are very important. It can clearly be seen that the new model utilizes a combination of mesh formulations to achieve high deformation. 5-4-1 Element types Figure 5.1 shows the two-dimensional four nodes plane strain element, (CPE4RT), used to model both the workpiece and the tool with reduced integration and hourglass control elements [3, 37, 39].

50

CHAPTER 5: FINITE ELEMENT MODELLING

c~-~ Y

(2) 2

Figure 5.1 Four-node solid element The two-dimensional, two-node rigid link element (R2D2) is shown in Figure 5.2. The element is defined by two nodes with two degree of freedom in X and Y directions at each node. This type of element is used to present the tool shank where it is tied strongly to the tool insert as a part of the tool. There is no output associated with these elements [37, 39].

P
/

0
1 Figure 5.2 Two-node rigid element

5-4-2 Geometry and boundary conditions Figure 5.3 shows the geometry of the old model with the variable initial chip height (h) in the case of feed 0.2 mm. The mesh of the workpiece is constrained in space so the material will flow from the left side to the right side and over the rake face to form

51

CHAPTER 5: FINITE ELEMENT MODELLING the chip. The old model is constrained in the Y direction along the bottom of the workpiece. Also, the tool is fully fixed at the reference point (RP). The length of the workpiece is 1.5 mm with width of 0.6 mm. The workpiece has 59613 nodes and the tool has 1134 nodes. The element size is bigger if it is compared to the new model. The calculation time for the case of feed 0.2 mm is 42 hours in the case of a computer with 3 GB of RAM memory and 2.4 GHz. The old model shows a critical initial chip height. For a certain short value of h, the model will not complete the analysis because the upper element of the chip at the chip-tool interface will penetrate the tool surface. Also, at some high value of h, the model cannot complete the analysis because of the high deformation that occurs. RP

0.6 mm

m&mmm 0.4 mm

1.5 mm

Figure 5.3 Boundary conditions and partition scheme for the previous model

52

CHAPTER 5: FINITE ELEMENT MODELLING Figure 5.4 shows the basic geometry of the two-dimensional finite element model that is used in this analysis for 0.3 mm feed with no initial chip height. The material moves from the left to the right at cutting speed V, with boundary conditions fixed in Y direction on the bottom of the workpiece. The tool is fully fixed at the reference point (RP). Figure 5.4 shows that the workpiece material has been divided into two main parts, A and B. These two zones are used to define the ALE boundaries as will be explained later. The initial length of the workpiece was approximately 2.7 mm with a width of 0.6 mm and a feed equal to 0.2 mm. As the workpiece flows pass the tool, its length reduces. This is different from the previous work [3] where the workpiece length was constant. In the present model, the workpiece and the tool have 4153 and 1731 nodes, respectively. The element size is chosen carefully such that small elements are used where the highest deformation is expected. The calculation time for the case of feed 0.2 mm and p, = 0.23 is 9 hours in the same computer as mentioned above with mass scaling factor equal to 50. The new model is a good contribution as it can avoid the initial chip geometry and let the mesh perform the chip deformation. During the deformation, the upper element on the side of the contact surface may penetrate the tool surface so the model cannot complete the analysis. The idea that is used to avoid the penetration is ignore the friction effect at the upper element. This will not cause any trouble because the upper element is the first element that will lose contact with the tool face. The way to separate the first element from the mesh is to perform the partition before the mesh is completed and then make a small cut at the upper element with approximate size equal to the mesh seed of the element below.

53

CHAPTER 5: FINITE ELEMENT MODELLING

RP Tool Workpiece

A
" %mmf
:

,4mn

;,/ mni

Figure 5.4 Boundary conditions and partition scheme for the new model

5-4-3 Mesh and chip formation The Arbitrary Lagrangian Eulerian method is utilized in the proposed finite element model. This technique is used to solve large deformation problems [3, 24]. One of the critical issues in this type of modelling is the positions at which Lagrangian or Eulerian regions are applied. Nasr et al. [36] presented an ALE model where they divided the workpiece into different parts. In the present work, a similar idea has been applied to hard steel AISI 4140. The explicit method has been used here to simulate metal cutting with high deformation; however, some researchers have used the implicit method [29, 41]. The implicit method causes difficulty in convergence because the contact and the material models with a high number of iterations cause more cost. The explicit integration method

54

CHAPTER 5: FINITE ELEMENT MODELLING is more efficient than the implicit integration method for solving extremely discontinuous events or processes. Before starting to explain the geometry and the structure of the new model, a brief explanation of the adaptive mesh technique that is used in the new model is given. The selection of choices for the adaptive mesh can play a major role to enable the model to run perfectly by choosing the optimum options. An Arbitrary Lagrangian Eulerian (ALE) adaptive mesh domain can be applied with the following cases [37]: Can be used to analyze either Lagrangian or Eulerian problems. Can contain only first order, reduced integration, solid elements (4-node quadrilaterals, 3-node triangles, 8-node hexahedra, and 4-node tetrahedra). Have boundary regions and surfaces where the applied loads and boundary conditions can exist. The choice starts with a defined mesh domain. In the current model, the adaptive mesh domain was applied to the entire workpiece by selected region. The ALE Adaptive Mesh Control is used where one chooses a file that was created earlier. Other options are left in default. The ALE Adaptive Mesh Constraint is used to apply the mesh boundaries in region B by choosing the right direction boundaries. There are two types of boundary region edges: Lagrangian and sliding. Note that the Eulerian boundary is not available in the CAE files. The only way to apply the Eulerian boundary is to write it directly to the input file. In all cases, the mesh just constrains the displacement of the boundary. For the Lagrangian boundary definition, the nodes are allowed to move with the material and this is what is called in a CAE file as flow of underlying material. The flow of underlying

55

CHAPTER 5: FINITE ELEMENT MODELLING material is used in this analysis; it is altered to be Eulerian boundary later in the input file. The other option is a sliding boundary which is called independent of underlying material. A sliding boundary is similar to the Lagrangian boundary except that it has a sliding edge. The mesh is constrained to move with the material in the direction normal to the boundary region but it is completely unconstrained in the direction tangential to the boundary region. Both Lagrangian and sliding boundaries can be viewed with ABAQUS/CAE. Finally, the Eulerian boundary constrains the mesh in space and allows the material to flow through it. The last step with the adaptive mesh technique is mesh control. This is one of the keys that is important to perform the new model. There are a lot of options here but the most important one in the new model is the mesh constraint angle for boundary region smoothing. This option controls the smallest angle in an element with default value equal to 60. Currently, the angle can be changed to be less than 60 because the simulation can have smaller angles. It has been found that 20 works well. In the current model, region A is modelled as a Lagrangian region with an adaptive mesh domain. The mesh will follow the material while the re-meshing will prevent the elements from being distorted exceedingly. Because the mesh deforms with the underlying material, free surfaces can be modelled properly and the boundary conditions can be applied in a simple way. The Lagrangian surface is applied along the contact surface where the deformation is expected. A mapped mesh is applied to this region so that the productive elements will have a uniform shape to deal with the deformation. The initial mesh is made very small at the region, where the mesh expects to grow (chip), to allow the model to reach steady state geometry.

56

CHAPTER 5: FINITE ELEMENT MODELLING Region B is modelled as an Eulerian region. The simple definition to explain the Eulerian region is as a fixed net in space that allows the material to flow through. The mesh is constrained to make sure that there is no separation around the tool tip. The top surface in region B should be higher than the tool tip but lower than the expected contact length so that the material will flow around the tool edge as if it is a fluid. The mesh is finely formed around the tool edge in order to allow high stress gradients, even though region B in general has a coarse mesh. In region A, the area above region B is where the chip is expected to grow and from the final shape of the chip. The current model is running without any initial chip height. The chip now has the ability to grow automatically where the only limitation is the element distortion. Reducing the computational cost by using mass scaling can be done in ABAQUS Explicit by multiplying the density of the material by a factor/'. The mass scaling should not apply to the entire model because it will increase mass of the workpiece. Based on the previous study [3], the chosen factor is equal to 50. The mass scaling is applied just in the region B (Eulerian region) and the small area above region B where mesh is expected to grow. Because the smallest element occurs in that region, the scale factor is applied in this portion of the workpiece. With mass scaling the file was solved in 9 hours for feed of 0.2 mm and friction coefficient 0.23; however, with no mass scaling, the simulation took around 32 hours to complete running. Mass scaling results have a short running time that saves the cost of the simulation. Achieving the steady state temperature over the whole tool insert is not possible in the short time of running but at least the contact temperature along the chip-tool interface

57

CHAPTER 5: FINITE ELEMENT MODELLING reaches the steady state temperature at the contact area. Although, the previous models [1, 2, 3] had a longer running time (4 ms), the new model reaches the same maximum temperature with shorter time (0.6 ms). The temperature distribution on the rake face grows gradually to reach the maximum values at the contact surface. In this study, the temperature at the contact surface is studied to explain the heat energy in the friction zones. 5-4-4 Contact algorithms The interaction between two surfaces can be solved with ABAQUS by using the contact algorithms as explained in section 4-6. When the surfaces make imperfect contact the gap conductance applies between the surfaces as a gap exists. These surfaces may separate after contact, such as in the metal cutting example. As explained in section 4.6 about the contact algorithm, either the kinematic or penalty contact algorithm must be chosen. Both options can be used in this analysis if the tool and the workpiece are modeled as elastic parts; however, the current model used the penalty contact algorithm because the tool is model as a rigid body. The penalty contact algorithm is useful in more general cases of contact. A further difference between kinematic and penalty contact is that the critical time increment is unaffected by the kinematic contact but can be affected by penalty contact. During the hard contact between the two surfaces, tool and workpiece, the contact algorithm applies distributed forces to nodes of the slave surfaces. There are three approaches for the relative motion of the two contact surfaces which include finite sliding, small sliding, and infinite small sliding. Using the finite sliding allows the nodes of the slave surface either to separate or to slide when they come into contact anywhere

58

CHAPTER 5: FINITE ELEMENT MODELLING along the rake face of the tool. ABAQUS Explicit tracks the position of these nodes relative to the master nodes. The choice of finite sliding is used in the current work because it is useful for the nonlinear geometry that has large deformation. At high values of the friction coefficient, the maximum temperature will reach more than 1000 C in a very short time. This high temperature will considerably affect the properties of the workpiece and make the study more complicated. The friction along the chip-tool interface is present in the two regions that affect the surface, sticking and sliding. Different approaches have been considered to determine the shear stress in these zones at the contact surface [3]. These approaches include a constant shear stress along the entire chip-tool interface, a constant shear stress in the sticking region, Coulomb friction in the sliding region, variable Coulomb friction along the entire chip-tool interface, constant friction along the chip-tool interface. In the current model, the constant Coulomb friction model is selected to simplify the analysis. For the thermal boundary conditions, the workpiece and the tool are initially set at 25 C. The convection heat transfer coefficient has been calculated. It is 10 W/m2oC for the free surface of the workpiece and 157 W/m2oC along the flank face of the tool [3]. The gap conductance model along the chip tool interface has thermal conductance of 0.1GW/mC with zero gap distance and zero at gap distance of 0.01 jam along the chip tool interface. Using equations 4.17& 4.18, the fraction of heat energy conducted to the chip and the tool are 0.623 and 0.377, respectively.

59

CHAPTER 6: RESULTS AND DISCUSSION

CHAPTER 6: RESULTS AND DISCUSSION


6-1 Introduction In this chapter, the results of the finite element models that are used to investigate the effect of the friction coefficient and the initial chip geometry are presented for the AISI 4140 workpiece and carbide tool. The main results include the cutting and feed forces, chip thickness, and contact length. The behaviour of the friction at the chip-tool interface is also shown to affect some other parameters. In order to discuss the results of the FEM, the basic Coulomb friction coefficient was utilized with friction coefficient equal to 0.23 based on the previous work [1, 2, 3]. The experimental data were obtained from reference [1,2] with same material at the same cutting conditions from feeds 0.01 mm to 0.4 mm. The initial part of this chapter is

60

CHAPTER 6: RESULTS AND DISCUSSION

focused on the analysis of the obtained results with friction coefficient 0.23 at feed 0.3 mm. The results of reference [2] were obtained by scanning the charts and then measuring from CorelDraw by choosing the right scale. At the beginning of the analysis, different models will be presented based on the initial chip height. These models will be compared to a new model without initial chip height and the experiment results [2]. The analysis results include the chip formation, cutting and feed forces, stress, strain, and temperature. The chip formation is presented at six different time steps. The maximum deformation of the chip in this specific cutting condition (feed 0.2 mm, friction coefficients 0.23 and 0.6, and cutting velocity 3.33 m/s) is presented at time step 0.6 ms. By 0.6 ms, the chip has reached the final form and the interaction keeps acting steadily at the chip-tool interface. The results of the simulation show slightly different data with different initial chip heights. This study is verified by comparing the current model with the previous model for a variety of initial chip heights. The comparison includes the chip thickness, contact length, and cutting and feed forces. The results present some interesting problems especially the force results. With the new model, the chip starts forming from the beginning of the deformation with the correct chip thickness. One of the advantages of the new model is that the final chip thickness occurs sooner. The maximum deformation of the chip in this specific cutting condition (feed 0.3 mm, friction coefficient 0.23, and cutting velocity 3.33 m/s) is presented at time 0.6 ms. After 0.6 ms, the chip will have more curl and this curl will cause damage in the mesh (see Figure 6.14).

61

CHAPTER 6: RESULTS AND DISCUSSION

The obtained forces are presented first by plotting cutting and feed forces versus time. The plot shows that the stability exists early before 0.1 ms. The forces change as a result of parameters such as the friction coefficient, feed and velocity. The primary results had been chosen in specific cutting conditions. The next section shows the stresses acting in the chip and workpiece. The analysis of the contour plots illustrates the maximum and minimum stress locations. The next section is mainly focused on the temperature, which is one of the most important issues because of the tool life. The contour plots help to study the distribution of temperature in the workpiece and the chip. The maximum temperature occurs in the friction zone. Steady temperature is one of the goals that needs to be achieved for steady state at the chip-tool interface. Also, studying the contact surface shows the distribution of friction and normal stresses along the chip-tool interface. Since the friction behaviour along the chip-tool interface is not clear, many studies were focused on a variety of friction models that could be used in metal cutting analysis. In this study, the simple friction coefficient is utilized based on Coulomb's low. Basically, the magnitudes of different values of friction coefficient, which are used in this work, are chosen from force data of Arrazola et al. [1]. The friction coefficient is selected to be 0.23, 0.4, 0.5 and 0.6 according to Albrecht's definition where the plot of feed forces versus the cutting forces is presented.

62

CHAPTER 6: RESULTS AND DISCUSSION

6-2 Effect of the Initial Chip Geometry


The traditional method for the ALE model includes initial chip height and thickness as an assumption. Although the predicted results were successful, the assumed initial chip usually depends on the feed and the friction coefficient. In this work, an attempt has been made to model the cutting process without defining the chip height. Before the comparison starts, it is important to state the differences between both models. Figure 6.1 shows the schematic diagrams of the new model and the previous model. The previous model will be presented with several initial chip heights.

Chip thickness Contact length

(a) The previous nrxM

(b)Therewrrrxtel

Figure 6.1 Schematic of both models

The initial geometry has some effect on the results. The following graphs show the percentage error of the chip thickness, contact length, cutting force, and feed force with different models for a feed of 0.2 mm. In each case, different initial chip heights (L) are used. For each chip height, the results are presented for two different friction coefficients (u=0.23, u=0.6). This analysis has been done for both 0.23 and 0.6 friction coefficient with the same parameters (cutting conditions). The new model, which has no initial chip height, is represented with L = 0 and all others are presented with different initial chip 63

CHAPTER 6: RESULTS AND DISCUSSION

heights from 0.3 up to 0.8 mm with an initial chip thickness of 0.35 mm. In some cases, the new model gives the best agreement with experimental results as will be explained in the following sections. Figure 6.2 shows the deformation process for an initial chip height case (L = 0.5 mm). The time step starts from 0.1 ms up to 0.6 ms. The chip geometry changes until it reaches the steady state. The free surface of the chip shows the chip deforms gradually from the bottom left of the chip to upper. It is obvious that the mesh does not have any severe distortion due to adaptive meshing and the Eulerian region.

64

CHAPTER 6: RESULTS AND DISCUSSION

(a) Time = 0.1 ms

(b) Time = 0.2 ms

(c) Time = 0.3 ms

(d) Time = 0.4 ms

(e) Time = 0.5 ms

(f) Time = 0.6 ms

Figure 6.2 Chip formation of orthogonal machining at different times with initial chip (f = 0.2 mm, L = 0.5 mm, V = 200 m/min, \i = 0.23)

65

CHAPTER 6: RESULTS AND DISCUSSION

6-2-1 Chip thickness Before presenting the chip thickness results, it is important to illustrate how it is measured (see Figure 6.3). After picking point A, which is away from the contact surface, measuring the distance between that point and several points at the free surface like B, C, and D is done to figure out the shortest distance. Because of the chip is curled, the shortest distance is expected to be the distance perpendicular to the point A

'Contact length

Figure 6.3 Chip thickness and contact length measurement Figures 6.4 and 6.5 show the obtained results for the chip thickness. The left hand bars are the experimental results of Arrazola et al. [1, 2]. It can be seen in the figure that the results illustrate the effect of the friction coefficient where the percentage of error has higher values for the higher friction coefficient. The new model without any initial chip height presents the best result at lower friction coefficient with less than 3% of error; however, at the higher friction coefficient, the data do not show the best results (see Figure 6.5). The results agree with the above explanation that increasing the friction coefficient causes a thicker chip because of the increase in the shear angle.

66

CHAPTER 6: RESULTS AND DISCUSSION

B|j = 0.23 I p = 0.6 DExp

0.5 0.45 -g- 0.4 0.35

S 0.3
J= 0.25 | 0.2

.9- 0.15 O 0.1 0.05 0


EXP L=0 L=0.3 L=0.4 L=0.5 L=0.6 L=0.7 L=0.8 Initial chip height

Figure 6.4 Chip thickness obtained by different models

The maximum magnitude of the chip thickness at 0.23 friction coefficient occurs in 0.6 mm initial chip height with percentage of error equal to 10%. The percentage of error increases at the higher friction coefficient to reach the maximum value at the initial chip height of 0.5 mm with percentage of error over 25%.

g
in

L=0

L=0.3

L=0.4

L=0.5

L=0.6

L=0.7

L=0.8

Initial Chip Height (mm) P|j = 0.23 i p = 0 . 6

Figure 6.5 Percentage error of the chip thickness for different models

67

CHAPTER 6: RESULTS AND DISCUSSION

6-2-2 Contact length The results of contact length are obtained by measuring along the chip-tool interface (see Figure 6.3). Figures 6.6 and 6.7 display the contact length for all cases of initial

chip. The average percentage of error with the smaller friction coefficient is 43% and for the higher friction coefficient it is close to 27%. It is obvious that the contact length has a high percentage of error. The results of the contact length with different initial chip heights from 0.3 to 0.8 mm have close results to each other. Some improvement has been gained by using the new model as will be explained below. The new model presents the best results with contact length for both friction coefficients with 44% percentage of error at 0.23 friction coefficient and 23% percentage of error at 0.6 friction coefficient. The advantage of the new model shows the better results when the highest node of the chip interacts with the rake face and slips over the rake face in short distance during the deformation as illustrated in section 6.2 when the chip is deformed at different times. Generally, the figures demonstrate that friction coefficient has big effects on the contact length in all simulations. Also, using the new model avoids having to decide the initial chip geometry and simplifies initial meshes. Sometimes the node in the highest location will lead the simulation to stop because of the high deformation that might occur in the element that will separate first from contact with tool. The reason for that high deformation is some extra force will be applied at the highest element in the chip-tool interface. This deformation can exist even in some cases of the initial chip height. One of the keys that might help to build a numerical model is to avoid applying friction in the highest element at the chip-tool interface during the analysis at high friction coefficients.

68

CHAPTER 6: RESULTS AND DISCUSSION

H|J = 0.23 ^ M = 0.6 DExp

0.7 0.6 0.5


c

0.4 0.3

u
o o

c 0.2

0.1 0
EXP L=0 L=0.3 L=0.4 L=0.5 L=0.6 L=0.7 L=0.8 Initial Chip Height

Figure 6.6 Comparison between the models for the contact length

l|j=0.23 B|J=0.6

0
L+0 liil tills L=

-10
-20 o -30 -40 -50 -60

BH

L=P6

L= W

L=t>.8

Initial Chip Height (mm)

Figure 6.7 Percentage error of the contact length

69

CHAPTER 6: RESULTS AND DISCUSSION

6-2-3 Cutting and feed forces The force results are easy to obtain from the software because the reference point (RP) collects the entire force vectors that act in the simulation. There are several results presented in this section. The first part is the force time histories for six different initial chip heights. The average cutting and feed forces (over the last 0.5 ms) are then compared between these models and the experiment [1]. First, the cutting forces versus time for the cases for \i = 0.6 are illustrated as shown in Figure 6.8. Obviously, there is no big difference between these cases. Figure 6.9 shows the comparison of the average cutting forces for (i = 0.23 and \i = 0.6 for various initial chip heights. By increasing the friction coefficient, the cutting forces increases higher than the experiment value of the cutting force on the one hand, while using the lower friction coefficient gives smaller values of the cutting force than the experiment. Figure 6.10 shows the percentage differences for the cutting force. The cutting forces for all models have quite good agreement with the experiment for both friction coefficients. The maximum percentage error of around -10% at friction coefficient 0.23 for the model L = 0.4 mm and 15% at friction coefficient 0.6 for the model L = 0.5 (see Figure 6.10).

70

CHAPTER 6: RESULTS AND DISCUSSION

L-0.3 cnn -r~-

L-0.4

L-0.5

L=0.6

L-0.7

L-0.8

480 460 - fli^k 440 %Cw T ^ Wr' y 'CTy^iidw^.VsS^Q^ ^ % 420 8 400 380360 340 320 inn i i i 0 0.0005 0.001

srf

0.0015 0.002

0.0025 0.003

0.0035 0.004

0.0045

Time (s)

Figure 6.8 Cutting forces versus time for all initial chip height cases (f = 0.2 mm, V = 200 m/min, \x = 0.6)

Hu=0.23 Hu=0.6 DExp

EXP

L=0

L=0.3

L=0.4

L=0.5

L=0.6

L=0.7

L=0.8

Initial Chip Height

Figure 6.9 Comparison between the models for the cutting forces

71

CHAPTER 6: RESULTS AND DISCUSSION

Su=0.23 H|J=0.6 20

,2 10 9^1

3L
'-If
0

L=

L= . 1

L= ) 5

L=

L= ).7

L= 3.8

-10 -15
Initial Chip Height (mm)

Figure 6.10 Percentage error of the cutting force The feed forces versus time for the cases of the initial chip height are illustrated in Figure 6.11. Obviously, there are big differences with feed forces.
L-0.3
? i n -.

L-0.4

L-0.5

^-0.6

L-0.7

L-0.8

190 _
2^

170 o 110 Qf)

Force

In
i ^^^^^^^^^^^ ^^^t^ff^^^^u^r^^^f^Si
0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0.0045 Time (s)

o 0

Figure 6.11 Feed forces versus time for all initial chip height cases (f = 0.2 mm, 200 m/min, (j, = 0.6) 72

CHAPTER 6: RESULTS AND DISCUSSION

The results for the feed forces are shown in Figure 6.12 and 6.13. The feed forces for the low friction coefficient demonstrate very high percentage of error. The maximum percentage of error is 80%. It is apparent that the lower friction coefficient cannot model the feed force. Well, the feed forces have improved results with less than 30% as a maximum percentage of error when the friction coefficient is 0.6. The presented data shows that the friction coefficient affects the feed force significantly; the average percentage of error for the lower friction coefficient is 76%, while the average is 14% for the higher friction coefficient. Generally, the new model shows the best average results for both cases of friction coefficient (|a, = 0.23, u=0.6). The new model will be presented in more details in section 6.3. Results with no initial chip present some progress in the contact length and forces as well.

Figure 6.12 Comparison between the models for the feed forces

73

CHAPTER 6: RESULTS AND DISCUSSION

|j=0.23 B|j=0.6

0 -10

mo

L=

L= )4

p D.5

L*I

fe^m. LH

8 -20
-30

$ -40 o -50 -t -60 -UJ

5? -70 -80 -90 Initial Chip Height (mm)

Figure 6.13 Percentage error of the feed force

6-3 Results with No Initial Chip


The results of the new model will be emphasized in current section where there is no initial chip geometry. The study shows the new model can perform the metal cutting problems efficiently. The primary results are calculated at friction coefficient 0.23 with the same cutting conditions and tool geometry as reference [1]. The results of no initial chip have a good agreement with the published [1] and previous work [3] for the case of friction coefficient 0.23. The results will show several interesting points that can be considered in machining as follows:

6-3-1 Chip formation The chip formation process of orthogonal cutting is explained in the section (see Figure 6.14). Before the deformation starts, the tool is in perfect contact with the

74

CHAPTER 6: RESULTS AND DISCUSSION

workpiece to reduce the running time and make sure that the contact occurs from the beginning. As the tool starts to move into the workpiece, more plastic deformation of the workpiece material exists along the chip-tool interface and the primary deformation zone. The chip starts increasing gradually along the rake face of the tool. The deformed chip has a high density of elements so these elements can be stretched to the chip geometry; however, the elements outside the primary deformation zone from the workpiece side are larger because these elements shrink during the deformation. Figure 6.14 (b) demonstrates that the top of the chip starts separating from the contact area while the continuous chip is still growing. During machining, the simulation reachs the steady state in both deformation zones so the shear plane softens the material to flow as a chip. The resistance to the tool penetration decreases because of the effect of the high temperature. The chip thickness achieves steady state (chip thickness and contact length) before 0.2 ms as seen in Figures 6.14 a, b, c. Figures 6.14 d, e present the curl of the chip during the deformation later. The model stopped running when the supply of material stopped (see Figure 6.14 f).

75

CHAPTER 6: RESULTS AND DISCUSSION

(a) Time = 0.1 ms

(b) Time = 0.2 ms

(c) Time = 0.3 ms

(d) Time = 0.4 ms

L,

L.

(e) Time = 0.5 ms

(f) Time = 0.6 ms

Figure 6.14 Chip formation of orthogonal machining at different times with no initial chip (f = 0.3 mm, V = 200 m/min, \i = 0.23)

76

CHAPTER 6: RESULTS AND DISCUSSION

6-3-2 Cutting and feed forces The engagement of the tool with the workpiece causes high deformation zones at the shear zone and along the friction zone. As a result of machining, the contact pressure on the chip-tool interface increases at different time increments (see Figure 6.14). Increasing the contact pressure affects the cutting force and the feed force as well. Figure 6.15 shows the predicted feed and cutting forces during steady state at 0.3 mm feed and 200 m/min cutting speed. The forces reach steady state before 0.1 ms. The average cutting force was 528 N and the feed force was 55 N for friction coefficient 0.23.

-Cutting_ 600 500 400

Feed-F

300 200 ^ 100 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Time (ms)

Figure 6.15 Cutting and feed force versus time (f = 0.3 mm, V = 200 m/min, u = 0.23) 6-3-3 Stress and strain distributions In this section, the contour plots of Von-Mises stress, shear stress, contact stress and equivalent plastic strain at friction coefficient 0.23 (f = 0.3 mm and V = 200 m/min) are

77

CHAPTER 6: RESULTS AND DISCUSSION

shown in Figures, 6.16, 6.17 and 6.18. The distribution of these variables will be discussed to obtain better understanding of chip formations.

6-3-3-1 Von-Mises stress distribution One of the stresses that is considered in the analysis is the Von-Mises stress. The analysis shows the maximum stresses occur in the primary deformation zone due to high strain and strain rate (see Figure 6.16). The stress decreases gradually on both sides of the shear zone. Along the chip-tool interface, where the surfaces interact, smaller stress is noticed in the friction zone. The stresses are still decreased and reach low values when the chip is separated from the contact region. It is easy to observe that stress still occurs in the chip after the separation. The contour of Von-Mises stress is shown below with five interval stresses lines.

S, Mises 1 +1.166e+09 2- +9.382e+08 3- +7. !08e+08 4- +4 833e+08 5- +2 559e+08

Figure 6.16 Distribution of Von-Mises stresses in the chip and the workpiece (Pa) (f = 0.3 mm, V = 200 m/min, u. = 0.23)

78

CHAPTER 6: RESULTS AND DISCUSSION

6-3-3-2 Shear stress distribution One of the results that are presented in this section is the shear stresses (see Figure 6.17). The highest stress was located in the primary deformation zone in front of the cutting tool edge and the end of the primary shear zone (region 1). The significance of the shear stress study was to focus on the friction shear zone because of the high interaction effect between the tool surface and the chip surface. A smaller shear stress is noticed in the shear zone despite the high temperature (region 3).

SI T 1- +2 716e+08 2- +7.549e+07 i- -1 206e+08 4- -3.167e+08 5_ -5.128e+08

Figure 6.17 Distribution of shear stresses in the chip and the workpiece (Pa) (f = 0.3 mm, V = 200 m/min, (x = 0.23) 6-3-3-3 Normal and friction shear stress along the chip tool interface Figure 6.18 shows the normal and shear stress distributions along the chip tool interface. The contact pressure along the rake face demonstrates that the maximum pressure occurs at the edge of the tool. It seems that the contact pressure is almost constant with slight decrease while moving away from the tool edge. The same trend is 79

CHAPTER 6: RESULTS AND DISCUSSION

observed for the shear stress with smaller values. It has nearly constant values where the sticking region occurs and decreases along the sliding region.

1.

35 w ~i | * 3 $ i JS
c

eg

^ | 7 2 ^

^uuu k 1800 -, 1600 1400 1200 1000 800 600 400 1 200 n -) C 0.05

onnn _,

Contact Pressure Friction Shear Stress

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

Distance along Chip-Tool interface (mm)

Figure 6.18 Normal contact pressure and friction shear stress distribution over the rake face (f=0.3 mm, V=200 m/min, \i = 0.23)
450 _.^ 400ra S 350 $ 300 .-

5> 250 2 200

Friction Sh

o o
n <

()

200

400

600

800

1000

1200

1400

1600

1800

2000

Contact Pressure (MPa)

Figure 6.19 Friction shear stress and normal contact pressure for the Coulomb friction model identification

80

CHAPTER 6: RESULTS AND DISCUSSION

The relation between the contact pressure and the fiction shear stress is illustrated in Figure 6.19. The slope of the chart presents the Coulomb friction model due to sliding with friction coefficient equal to 0.23. 6-3-3-4 Equivalent plastic strain Plastic strain is observed after the flow material passes the shear plane. Once the material enters the shear zone, the magnitude of plastic strain increases rapidly with different values depending on the friction zone. Figure 6.20 shows the distribution of the plastic strain in the chip and workpiece. It is obvious that plastic strain decreases rapidly from the contact area (the right side of the chip) to the free surface at the left side of the chip. The maximum strain occurs in the contact area at the friction zone where the high temperature is experienced.

~^
i-+2.357e+00 2-+1.885e+00 i-+1.414e+00 4- +9 427e-01 5- +4.714e-01

/ >v
/ / / 5 ^ / / ^ / 4 ^S J>^ X / A f\ \3 \ \ \

u
Figure 6.20 Distribution of equivalent plastic strain in the chip and the workpiece (f = 0.3 mm, V = 200 m/min, \i = 0.23)

81

CHAPTER 6: RESULTS AND DISCUSSION

6-3-4 Temperature distributions The contour plots of temperature distribution of both the workpiece and the tool are shown during steady state for f = 0.3 mm, V = 200 m/min, and \i = 0.23. This study shows the location of the maximum temperature and where the temperature decreases along the rake face of the tool. Studies of temperature distribution helps to obtain the optimum cutting condition of the tools to avoid the tool wear. 6-3-4-1 Temperature distribution in the chip The contour plot in Figure 6.21 demonstrates the temperature distribution in the chip during steady state orthogonal machining. The undeformed workpiece was initiated to be at room temperature (25 C ). The temperature increased as a result of the heat generated by the plastic deformation and the friction at the secondary deformation zone. The highest temperature was located at the friction zone (region 1) with an approximate value equal to 940 C. The lowest temperature was located in the primary shear zone (region 5).

TEMP I- +7.776e+02 2- +6.27le+02 .1- +4,766e+02 4- +3.261e+02 5- + l.755e+02

Figure 6.21 Distribution of temperature in the chip and the workpiece (f = 0.3 mm, V = 200 m/min, u. = 0.23)

82

CHAPTER 6: RESULTS AND DISCUSSION

6-3-4-2 Temperature distribution along the rake face The temperature rise occurs at the chip-tool interface where some of the heat energy is held in the tool and other is carried out with the chip. The magnitude of the heat that is carried out with the chip is higher than what is held in the tool because of the flow of the chip. Besides, the properties of the tool and the workpiece such as the conductivity and the specific heat will significantly affect the heat energy which is transferred to both. Figure 6.22 shows that the maximum temperature exists almost in the middle of the rake face (region 1).

TEMP
1 +7.846e+02 2- +6.327e+02 3- +4.808e+02 4- +3.288e+02 5- +1.769e+02

Figure 6.22 Distribution of temperature in the tool (f = 0.3 mm, V = 200 m/min, \i = 0.23)

83

CHAPTER 6: RESULTS AND DISCUSSION

A tool's temperature plays a very important role in the tool's life. It is important to use the tool for a long time as long as the properties of the tool have not changed and the shape matches the perfect cutting geometry. Due to friction, the tool life will decrease, so the tool life should be managed by using the optimum cutting conditions such as the velocity and the feed [6]. Figure 6.23 shows the temperature distribution along the chip tool interface. It illustrates that a maximum temperature of 940 C will be reached at a distance away from the cutting edge. Away from the maximum temperature point, the temperature values start to decrease close to the zone where the chip physically leaves the tool at 0.4 mm.

1000 | 900 800 -

8 700T 600 1 500a E 400 *> Y 300 200 100 -

o-l)
C

0.1

0.2

0.3

0.4

0.5

0.6

0.7

08

Distance along the Rake Face ( m m )

Figure 6.23 Temperature distribution on the rake face (f = 0.3 mm, V = 200 m/min, \L = 0.23)

84

CHAPTER 6: RESULTS AND DISCUSSION

TKMP 1 - -7.846e-02 .?.- 6 . 3 2 7 C 0 2 3 - 4.808c-'-02 3.288e-:02 5 - -1.76<V -02

Node - 4

Node - 3

Node - 2

Node - 1

Figure 6.24 Location of the selected nodes in the rake face of the tool (f = 0.3 mm, V = 200 m/min, \L = 0.23) Figure 6.24 shows the location of selected nodes at the rake face. The nodes are selected to be in different zones to show the distribution of the temperature along the entire rake face. The steady state will be achieved at different times at these nodes. The maximum temperature is expected to be somewhere in region 1. In Figure 6.25, the maximum temperature seems to reach a reasonable steady state after 0.5 ms. After 0.5 ms the temperature still increases but with very small values.

85

CHAPTER 6: RESULTS AND DISCUSSION

Node-1

Node-2

Node-3

Node-4

1000 y 900 800 o 700 | 600 " < 500 5 E 400 300 200 100 00

0.0001

0.0002

0.0003

0.0004

0.0005

0.0006

0.0007

Time (s)

Figure 6.25 Rake face temperature versus cutting time (f = 0.3 mm, V = 200 m/min, u = 0.23)

6-4 Effect of Friction Factor


During machining, the interaction between the tool and the chip includes forces in the normal and tangential directions. In this study, the simple Coulomb friction model was applied along the chip-tool interface with different friction coefficients. The results show different responses when using different friction coefficient. The effect of the friction will be shown in the following sections.

6-4-1 Contour stress, strain and temperature distributions The results of the contour plots of stresses, equivalent strains, and temperature for different friction coefficients are shown in the following plots from Figures 6.26 to 6.29. The distribution of these variables at the workpiece and tool are explained to explore the effect of friction in orthogonal machining. These results can be used to gain good information that may be used in understanding chip formation.

86

CHAPTER 6: RESULTS AND DISCUSSION

6-4-1-1 Von-Mises stress distributions Figure 6.26 shows the steady state Von-Mises stress in the chip for four friction coefficients, 0.23, 0.4, 0.5, and 0.6 during machining. The stress was calculated based on the strain, strain rate, and temperature (Johnson-Cook formula). The maximum and minimum stress contour values were constrained in the four cases to simplify the analysis. The location of any stresses over the whole chip is defined by six lines as shown in the plots. The maximum stresses occur in the primary shear plane with approximated stress equal to 1.34 GPa. The only differences that can be noticed are the sizes of the stress regions. Lower stresses are shown near the contact surface at the friction deformation zone. As far as the friction coefficient increase, the secondary deformation zone will show clearly that the stresses will increase at the friction surface. In general, the plots look similar; however, the stress distributions show some differences. Also, the geometry of the chip is affected by the friction coefficient as it becomes thicker with higher friction coefficients. The stresses decrease after the separation at the chip tool interface and increase gradually after a while because of the curl. Some effect of the stresses will influence even the produced surface and the workpiece itself. The influence for the workpiece is important because high stress near the produced surface is what it is called residual stress.

87

CHAPTER 6: RESULTS AND DISCUSSION

S, Mises '- +1.340e+09 2- +1.117e+09 3- +8.945e+08 4- +6.718e+08 5- +4.490e+08 6- +2.263e+08

S, Mises 1- +1.340e+09 2- +1.117e+09 3- +8.945e+08 4- +6.718e+08 5- +4490e+08 6- +2263e+08

u
(a) n = 0.23

u
(b) p, = 0.4

S, Mises 1-+1.340e+09 2- +1.117e+09 3- +8.945e+08 4-+6.718e+08 5- +4.490e+08 6- +2.263e+08

(c) ji = 0.5

(d) \i = 0.6

Figure 6.26 Contour plots of Von-Mises stress for different coefficients of friction (Pa) (f = 0.3 mm, V = 200 m/min)

88

CHAPTER 6: RESULTS AND DISCUSSION

6-4-1-2 Distribution of shear stress The shear stress distribution in the chip and the workpiece during the steady state is presented in Figure 6.27 for different friction coefficient. The friction coefficient has some effects on the shear stress as seen from the plots. As a result of the interaction between the surfaces at the friction zone, the stress increases gradually by increasing the friction coefficient from 0.23 to 0.6 (regions 4, 5) (negative values). Increasing the coefficient of friction will generate higher temperatures at the chip-tool interface, which will be explained later. The shear stress acts at the friction zone by applying tangential loads over the chip contact surface to make the separation harder. High shear stresses occur just where the tip of the tool contacts the workpiece. The whole interior chip, regions 1, 2, and 3, has a positive shear stress; however, the regions, 4, 5 and 6 have negative shear stress. It can be seen that the stress diffuses gradually from negative shear stress at the primary and secondary zone to the positive shear stress. The stress will gradually affect some layers of the produced surface.

89

CHAPTER 6: RESULTS AND DISCUSSION

S.S12 1 - +3.540e+08 2- +1.781e+08 3- +2.200e+06 4- 1.737e+08 5- -3.496e+08 6- -5.255e+08

S.S12 1 - +3.540e+08 2 - +1.781e+08 3 - +2.200e+06 4 - -1.737e+08 5 - -3.496e+08 6 - -5.255e+08

Lx (a) p, = 0.23
S.S12 S.S12 123456+3.545e+08 +1 785e+08 +2.515e+06 -1.735e+08 -3.495e+08 -5.255e+08 1- +3.540e+08 2- +1.781e+08 3- +2.200e+06 4--1.737e+08 5- -3.496e+08 6- -5.255e+08

(b) n = 0.4

Lx
(c) n = 0.5

Lx (d) n = 0.6

Figure 6.27 Contour plots of shear stress for different coefficients of friction (Pa) (f = 0.3 mm, V = 200 m/min)

90

CHAPTER 6: RESULTS AND DISCUSSION

6-4-1-3 Distribution of equivalent plastic strain The effect of the friction coefficient on the distribution of equivalent plastic strain over the whole workpiece is illustrated in Figure 6.28. The pattern of the equivalent plastic strain looks similar; however, the magnitude is different as shown in the contour plots. The plastic strain starts after the workpiece material passes the shear plane. Extensive plastic strains occur in the secondary deformation zone. The maximum plastic strain occurs in the layer at the chip-tool interface. The magnitude of the equivalent plastic strain is significantly affected by the friction coefficient. The highest value of the equivalent plastic strain occurs in a thin layer near the chip tool interface in the sticking region close to the transition into the slip region (see Figure 6.28 a). In the Figures 6.28 b, c, d the starting point is not clear because of the use of 6 lines of equivalent plastic strain in the contour plots. The peak values of the plastic strain are 5, 18.6, 19.9, and 20.2 for coefficients of friction 0.23, 0.4, 0.5, and 0.6, respectively. A high plastic strain also appears in the tertiary deformation zone near the tool tip position. The magnitude of this strain in the tertiary is related to the friction coefficient, so by increasing the friction coefficient, the tertiary zone reaches a higher magnitude plastic strain.

91

CHAPTER 6: RESULTS AND DISCUSSION

PEEQ l-+5.000e+00 2- +4.167e+00 3 - +3.333e+00 4- +2.500e+00 5-+1.667e+00 6- +8.334e-01 7- +5.000e-05

PEEQ i-+1.866e+01 2- +5.000e+00 3-+4.167e+00 4 - +3.333e+00 s-+2.500e+00 6-+1.667e+00 7-+8.334e-01 s-+5.000e-05

U
(a) [i = 0.23 (b) \i = 0.4

PEEQ i-+1.994e+01 2- +5.000e+00 3- +4.167e+00 4- +3.333e+00 5- +2.500e+00 e-+1.667e+00 7- +8.3346-01 8- +5.000e-05

PEEQ i-+2.020e+01 2 - +5.000e+00 3 - +4.167e+00 4 - +3.333e+00 s - +2.500e+00 e-+1.667e+00 7- +8.334e-01 8- +5.000e-05

1-x (c) (x = 0.5

Lx (d) n = 0.6

Figure 6.28 Contour plots of equivalent plastic strain distribution for different coefficients of friction (f - 0.3 mm, V - 200 m/min)

92

CHAPTER 6: RESULTS AND DISCUSSION

6-4-1-4 Distribution of temperature Figure 6.29 presents temperature distributions in the chip during the steady state. The temperature rise is a result of converted heat energy from both the plastic deformation and friction effect. As explained above, most of the plastic work will transfer to heat (assumed 90%). In the primary deformation zone, the chip and the workpiece material are considered as one body; however, the chip and the tool material are two different bodies that should be emphasized because of the large influence at the contact surface. Although the temperature at the shear plane is high to soften the material, the temperature at the friction zone will be even higher because of the interaction between the tool surface and the chip surface. The contour plots show that the maximum

temperature occurs not at the tool tip but on the chip-tool interface in some distance over the tool tip. The friction coefficient clearly affects the maximum temperature. Figure 6.29 shows that the maximum temperature values along the chip-tool interface increase as the friction coefficient increases in a uniform distribution as seen in the contour plots. For the friction coefficient 0.23, 0.4, 0.5, and 0.6, the maximum temperatures are equal to 936.5, 1151, 1163, and 1184 Celsius, respectively. The temperature decreases gradually from the contact surface to the free surface of the chip. The maximum plastic strain also occurs at the maximum temperature region as seen in Figure 6.28 and 6.29. By increasing the temperature, the crystal structure of the workpiece may be more free to deform and result in large plastic strains.

93

CHAPTER 6: RESULTS AND DISCUSSION


NTH MaxT:9.365e+02 1 - +9.736e+02

2 - +7.839e+02 3- +5.942e+02 4- +4.044e+02


5 - +2.147e+02 6 - +25006+01

NT11 MaxT: 1.151e+03 1- +9.736e+02 2- +7.839e+02 3- +5.942e+02 4- +4.044e+02 5- +2.147e+02 6- +2.500e+01

L
(a) ji = 0.23

u
(b) n = 0.4

NT11 MaxT: 1.163e03 1- +9.736e+02 2- +7.839e+02 3- +5.942e+02 4- +4.044e+02 5- +2.147e+02 6- +2.500e+01

NT11 MaxT: +1.184e+03 1- +9.736e+02 2- +7.839e+02 3- +5.942e+02 4- +4.044e+02 5- +2.147e+02 6- +2.500e+01

L
(c) n = 0.5

L_x
(d) n = 0.6

Figure 6.29 Contour plots of temperature distribution for different coefficients of friction (f = 0.3 mm, V = 200 m/min) 94

CHAPTER 6: RESULTS AND DISCUSSION

6-4-2 Chip thickness The comparison between the measured chip thickness [1, 2], and the predicted values is shown in Figure 6.30 for six different feeds for four friction coefficients. The obtained chip thickness is measured as explained in section 6-2-1. As expected when the feed increases, the chip thickness increases linearly. The results of the current model are very close to the experiment done by Arrazola et al. [1, 2] and the simulation work done by Hagland [3]. In fact, increasing the friction coefficient alters the simulation results of the chip thickness, so as the friction coefficient increases the chip thickness as well. The results show that the simulations can have a good agreement with the experiment if the friction model uses the right assumptions. The presented models seem to have excellent results compared to the published experiment, especially for friction coefficient u=0.23 where the magnitude of percentage of error is less than 6% (see Figure 6.31). Figure 6.31 shows that maximum error occurs for the highest friction coefficient u=0.6 where the maximum magnitude error is located at feed 0.25 mm with value equal to 30%. As a result of increasing the friction coefficient, the friction force will be increased along the chip-tool interface; consequently, the shear angle decreases and causes the chip thickness to increase. The present model shows the steady state chip thickness from the beginning of the deformation, which makes the measurement of the chip thickness easier.

95

CHAPTER 6: RESULTS AND DISCUSSION

Figure 6.30 Chip thickness obtained for the experimental [1,2] and numerical models
B |J=0.23 |J=0.4 M=0.5 D |J=0.6 35 30

a c 25 o i - 20
Q.

O 15 *-

o 10
o k 5 0

Ill

3?

0.1

0.15

0.2

0.25 Feed (mm)

0.3

0.35

Figure 6.31 Percentage of error of the obtained chip thickness for numerical models

96

CHAPTER 6: RESULTS AND DISCUSSION


6-4-3 Contact length Figure 6.32 shows the contact length at different feeds with different friction coefficient. The results of the contact length have the same trend of the chip thickness as explained in section 6.4.2 where the obtained contact length is plotted against the feed and compared with the published experimental data [1, 2]. In all cases the contact length is under predicted. It can be observed that better contact length can be reached by increasing the friction coefficient. The influence of the friction coefficient is shown in Figure 6.32. As explained above in section 6-3-3-3, the contact shear stress along the chip-tool interface will have the highest magnitude at the tip of the tool and stay almost constant over the sticking region then the shear stress decreases gradually, so when the friction coefficient has a high value, the shear force will increase and prevent the chip from separating. While increasing the friction coefficient gives a better contact length, it will badly affect other parameters such as chip thickness.
# M=0.23 O|j=0.4 (J=0.5 A M :=0.6 A EXP
Q.

-I ^_..

0.9 J A A A A A 1
A

| ? 0.8 1
g> 0.7 5 8 0.6 -

tl
* a>

0.5 ll

i o
1

3 o "
t3 ,2 0.2 i
g
0.1 -

g " 0.4 > 3

9
I

,9
O

r >

u 0.05

' 0.1

0.15

0.2

' 0.25

0.3

0.35

0.4

Feed (mm)

Figure 6.32 Contact length along the chip-tool interface obtained for the experimental [1, 2] and numerical models 97

CHAPTER 6: RESULTS AND DISCUSSION

At feed 0.1 mm for example, the maximum percentage error is 50% at friction coefficient 0.23 and will drop to 22% for friction coefficient 0.6. The big jump of the percentage error really comes from the friction coefficient effect (see Figure 6.33). These results illustrate that the simulations do not have a good agreement with experiment in the contact length. The contact length still needs some study since the simulation shows poor results. Even the previous work in which dual friction was used [1, 2, 3], the results poorly presented the contact length. In the current work, the contact length shows some response to the friction coefficient. The reason for these poor results might be the friction model or the material model.
I |J=0.23 M=0.4 D |J=0.5 O (J=0.6

iJBiial O

t-10
c
0)

Ksj

Iffiffrl

i -20 o -30

-40 -50 -60 Feed (mm)

Figure 6.33 Percentage error for contact length for numerical models compared to published experimental values [1,2]

98

CHAPTER 6: RESULTS AND DISCUSSION

6-4-4 Cutting and feed forces In Figure 6.34, the average cutting and feed forces, obtained from the last 0.05 ms of the solution, are plotted to illustrate the effect of friction coefficient on these forces. Before starting the discussion, it is important to know that the constant Coulomb friction model has been utilized in this analysis. The obtained results at friction coefficient 0.23 at different feeds show the same trend if they are compared to the experimental results. The feed and cutting force show high sensitivity to the friction coefficient (see Figure 6.34). The results present less than 10% error in the cutting forces and over 70% error in the feed forces (see Figure 6.35, 6.36). The amount of error, especially in the feed force, is not acceptable. Increasing the friction coefficient, the feed force demonstrates a better response. For example, friction coefficient 0.4 gives less error in both cutting and feed force: the maximum percentage error is equal to less than 10% for the cutting force with feed 0.2 mm and less than 20% for the feed force at the same feed. At friction coefficient 0.5, the results show some improvment for the feed force; however, the cutting force increases the percentage of errors. The same trend will occur for the friction coefficient 0.6. In general, the increasing of the friction coefficient means that the feed forces will have better results but not in the all cases, as seen in Figure 6.33. The friction coefficient 0.4 gives the best obtained results in the feed forces generally. In some cases, friction coefficient 0.5 and 0.6 have better results than the friction coefficient 0.4. Also, friction coefficient 0.4 has good results but not the best for the cutting force.

99

CHAPTER 6: RESULTS AND DISCUSSION

Figure 6.34 Measured and predicted force values for different feeds It is obvious that a big jump for the feed force at higher friction coefficient at feed of 0.15 mm occurred. This may be what is called the critical feed. Below that feed, the tool tip force will have higher magnitude, which causes a rapid increase in feed force. The tip force vector will affect the feed force significantly. Below the critical feed, the feed force can include the tip force, but after that the tool tip vector will have insignificant values to affect the feed force so this may be the reason of higher values of the feed force. This explanation explores the assumptions of Albrecht [4] who explained the affect of two forces that act in the tip and the rake face of the tool, as discussed in section 3-3-2.

100

CHAPTER 6: RESULTS AND DISCUSSION

H|j=0.23 H|j=0.4 D|j=0.5 D|J=0.6 35 30

S 25
O)

20
15

.E

10
o
ill

Hi
0.1

0
-5
-10 -15

5-

S?

0.15

0.2

' '' J ! m
0.25

ill

0.3

0.35

Feed (mm)

Figure 6.35 Percentage of error for the obtained cutting force of the numerical models

l|j=0.23 B|j=0.4 D|j=0.5 D|j=0.6 60 40

S
a
0)

20 A
0

dl
^

< u "o
o
UJ

-20
-40

P15

0.25

0.3

0.35

g=

-60 -80

-100

Feed (mm)

Figure 6.36 Percentage of error for the obtained feed force of the numerical models

101

CHAPTER 6: RESULTS AND DISCUSSION

6-5 Effect of Mass Scaling


In order to explore the effect of mass scaling, the numerical results for friction coefficient 0.5 and 0.6 for the cases with and without mass scaling are shown in Figures 6.37 to 6.40. Figure 6.37 shows the cutting forces versus feed. All numerical results of the cutting force are very similar. The numerical results have larger values compared to the experiment because of the high values of friction coefficient. The maximum error of the cutting forces is less than 30% compared to the experiment. Figure 3.38 shows the feed forces versus feed. Again the effect of mass scaling is quite small. The trend of the feed force shows some noise especially for feeds 0.15 mm and 0.25 mm. The reason may be that the solution without mass scaling needs more running time. For example, at 0.15 mm feed and 0.6 friction coefficient with mass scaling, the maximum error jumps from 48% to 60 %.

p=0.5 --" |j=0.6 (j=0.5 + Mass scaling (j=0.6 + Mass scaling * EXP A 800 -r 700 - 600 "g 500 400 O)

j= 300 O 200 100 00 0.05 0.1 0.15 0.2 0.25 Feed (mm) 0.3 0.35 0.4

Figure 6.37 Cutting force vs. feed

102

CHAPTER 6: RESULTS AND DISCUSSION

M-0.5

# (j=0.6 -- (j=0.5 + Mass scaling (j=0.6 + Mass scaling *--EXP

ocn

200 Force(N)

j .

/ ^

^ ^ -*

^ "

g 100LL.

Y^
0.05 0.1 0.15 0.2 Feed (mm) 0.25 0.3 0.35 0.4

50 n 0

Figure 6.38 Feed force vs. feed

The chip thickness and the contact length for the cases of mass and non-mass scaling are shown in Figures 6.39 and 6.40. The numerical results are very similar with maximum error for the chip thickness 25% and for the contact length around 25% compared to the experimental results. It seems there is no big effect for the mass scaling for both chip thickness and contact length for the cases that are shown in Figures 6.39 and 6.40. The difference between the numerical results of mass and non-mass scaling is less than 10%.

103

CHAPTER 6: RESULTS AND DISCUSSION

-A|j=0.5 -~- |j=0.6 |j=0.5 + Mass Scaling p=0.6 + Mass scaling - - * --EXP
n o

0.7 <-. | 0.6-

~^?8l&

0.5 a>

5. 0.4-

o jE 0.3 a. !E 0.2 -

< > > ^ ^ ^

0.1 n 0 0.05 0.1 0.15 0.2 Feed (mm) 0.25 0.3 0.35 0.4

Figure 6.39 Chip thickness vs. feed

-4r~-|j=0.5

u=0.6 |j=0.5 + Mass scaling - u=0.6 + Mass scaling -as EXP

1
0.9

-g" 0.8
E. 0.7 o i re o O 0.5 0.4 0.3 0.2

0.1 -I 0 0

0.05

0.1

0.15

0.2 Feed (mm)

0.25

0.3

0.35

0.4

Figure 6.40 Contact length vs. feed

104

CHAPTER 7 CONCLUSIONS

CHAPTER 7 CONCLUSIONS
7-1 Summary An Arbitrary Lagrangian Eulerian approach (ALE) has been used to develop a finite element model of orthogonal metal cutting using ABAQUS in order to study the behaviour of the friction along the chip-tool interface. During this investigation, some parameters such as the geometry of the workpiece were included in the study to see the effect of the initial geometry. The finite element model includes nonlinear features such as the material properties. The Johnson-Cook material model is one of the options that is available in ABAQUS. It is easy to input this model after obtaining the right parameters, which are obtained experimentally [1,2]. The finite element method can provide detailed

105

CHAPTER 7 CONCLUSIONS results of different variables that cannot be obtained from experiments such as the distributions of contact stress and temperature. The current finite element model was developed by using no initial chip geometry. By allowing the chip to grow over the rake face of the tool, the mesh is deformed to the shape of the chip. The deformed chip is affected by the interaction at the chip tool interface, so increasing friction coefficient influences the results. The obtained results of no initial chip height were compared to the old models with initial chip geometry and published experimental results for AISI 4140 steel [1,2]. The comparisons include chip thickness, contact length, cutting force, and feed force for different models for a feed of 0.2 mm, and u=0.23, u=0.6. The model of no initial chip height showed improving results that give the best agreement with experimental results. Friction at the chip tool interface was studied to investigate the effect of friction coefficient. Constant friction coefficient was applied to the entire chip-tool interface to simplify the analysis. Albrecht theory [4] was discussed as a way to estimate the friction coefficient as the feed changed. The friction model was used with different coefficient values, 0.23, 0.4, 0.5, and 0.6, for the same cutting conditions and tool geometry as reference [1,2] for six different feeds from 0.1 mm up to 0.35 mm in order to investigate the mechanism of machining process. The effect of mass scaling was explored briefly. Although it considerably reduced the solution time, its effect on the results was quite small.

106

CHAPTER 7 CONCLUSIONS 7-2 Conclusions In this study, an ALE finite element simulation is used to simulate the continuous chip formation process in orthogonal cutting of steel AISI 4140. The conclusions of this work based on the obtained results can be drawn as follows: 1. The initial geometry of the workpiece has a little influence on the results; however, the new model, which is presented with no initial chip height, shows some improved results compared to the old models. These improved results include the contact length and the cutting forces. 1. The chip thickness has the best results with no initial chip model at the lowest friction coefficient compared to the other models; however, with friction coefficient 0.6, the result is not the best. 2. Using constant Coulomb friction coefficient over the chip-tool interface with the Albrecht definition (JJ, = 0.23) cannot give good results for the feed force compared to the experimental results. The reason for poor results of the feed force and contact length may be due to either the cutting edge or the friction model. 3. Raising the friction coefficient improves the prediction of the contact length; however, it changes the other results, cutting and feed force as well as the temperature. The contact length is always underestimated, compared to the experimental results.

107

CHAPTER 7 CONCLUSIONS 7-3 Contributions The contributions of this research include the following: 1. An ALE two-dimensional finite element model capable of simulating orthogonal metal cutting processes using a mechanical and thermal explicit solution was further developed. 2. The no initial chip model avoids issues related to defining initial chip geometry. The chip presented by the deformed mesh and material grows smoothly along the chip-tool interface. This model reaches steady state sooner than the old models with initial chips and so saves computer time. The comparison study focused on the effect of the chip height and showed that the no initial chip model often gives the best results for chip thickness, contact length, cutting force and feed force. 3. Constant friction coefficient was used along the chip-tool interface with four different values to obtain detailed analysis of the effect of the friction during machining. Six different feed values were simulated. 4. This work defined a way to create the right input file by using the interactive software ABAQUS.CAE. Nearly all the input data is entered to the simulation by CAE so that there is no way to make mistakes with nodes position and number of elements that are needed for the model.

7-4 Recommendations for Future Work


The following points are suggested for further work: 1. Review heat transfer coefficient values used on all surface and investigate need for radiation heat transfer.

108

CHAPTER 7 CONCLUSIONS 2. Study different friction models that can be applied along the chip-tool interface such as limiting shear stress model, temperature dependence, and variable friction model can be used. 1. Consider the residual stress that may affect the product surface because of the high temperature that is generated along the chip tool interface. The thermal stresses occur in a layer close to the machined surface, after the relaxation to the room temperature. 2. Perform more complicated models of metal cutting by using three-dimensional geometry. The cost of this kind of model will be high because of the long solution time. 3. Use different materials to represent the workpiece material for successful metal cutting model. Some materials have the Johnson-Cook parameters available in the published literature. 4. Study the effect of the workpiece hardness in the metal cutting model because this may influence the results of the cutting and feed forces and the shape of the chip. Study also the hardness effect on the residual stress.

109

References
[1] P.J. Arrazola, D. Ugarte and X. Dominguez, "A new approach for the friction identification during machining through the use of finite element modeling", International Journal of Machine Tools & Manufacture, Vol. 48, pp. 173-183, 2008.

[2] P. J. Arrazola, F. Meslin and S. Marya., "A technique for the identification of friction at tool/chip interface during maching". Proceedings 6th CIRP International Workshop on Modeling of Machining Operations, Hamilton, ON, Canada, May 2003.

[3] A.J. Haglund, "On friction modeling in orthogonal machining: An arbitrary Lagrangian Eulerian finite element model", M.Sc. Eng. thesis, University of New Brunswick, 2005.

[4] P. Albrecht. "New Developments in the Theory of the Metal-Cutting Progress, Part I. The Ploughing Process in Metal Cutting", Journal of Engineering for Industry, Vol. 82, pp. 348-358, 1960.

[5] V. P. Astakhov, "On the inadequacy of the single-shear plane model of chip formation", International Journal of Mechanical Sciences, Vol. 47, pp. 1649-1672, 2005.

[6] M. C. Shaw, Metal Cutting Principles, Oxford University Press, Oxford, 1996.

110

[7] M.E. Merchant, "Mechanics of the metal cutting process. I. Orthogonal cutting and type 2 chip", Journal of Applied Physics, Vol. 16, pp. 267-675, 1945.

[8] B. E. Klamecki, "Incipient chip formation in metal cutting-a three dimension finite element analysis", Ph.D. Dissertation, University of Illinois, Urbana-Champaign, 1973.

[9] A. J. M. Shih and H.T.Y. Yang, "Experimental and Finite Element Predictions of Residual Stresses due to Orthogonal Metal Cutting", International Journal for Numerical Methods in Engineering., Vol. 36, pp. 1487-1507, 1993.

[10] T. D. Marusich and M. Ortiz, "Modelling and Simulation of High-Speed Machining", International Journal for Numerical Methods in Engineering, Vol. 38, pp. 3675-3694, 1995.

[11] T. Ozel, "The influence of friction models on finite element simulations of machining", International Journal of Machine Tools & Manufacture, Vol. 46, pp. 518530, 2006.

[12] A. O. Tay and M. G. Stevenson, "Using the finite element method to determine temperature distributions in orthogonal machining", Proceedings of Institution for Mechanical Engineers, Vol. 188, pp. 627-638, 1974.

Ill

[13] T. H. C. Childs and K. Maekawa, "Computer-aided simulation and experimental studies of chip flow and tool wear in the turning of low alloy steels by cemented carbide tools", Wear, Vol. 139, pp. 235-250, 1990.

[14] J. S. Strenkowski and K. J. Moon, "Finite element prediction of chip geometry and tool/workpiece temperature distributions in orthogonal metal cutting", ASME Journal of Engineering for Industry, Vol. 112, pp. 313-318, 1990.

[15] T. Belytschko, J. M. Kennedy, "Computer models for subassembly simulation", Nuclear Engineering and Design, Vol. 49, pp. 17-38,1978.

[16] J. Wang and M. S. Gadala, "Formulation and survey of ALE methods in nonlinear solid mechanics", Finite Element in Analysis and Design, Vol. 4, pp. 253 1997.

[17] L. Olovsson, L. Nilsson and K. Simonsson, "An ALE formulation for the solution of two-dimensional metal cutting problems", Computers and Structures, Vol. 72, No. 4, pp. 497-507, 1999.

[18] M. Movahhedy, M.S. Gadala and Y. Altintas, "Simulation of the orthogonal metal cutting process using an Arbitrary Lagrangian Eulerian finite element method", Journal of Material Processing Technology, Vol. 103, pp. 267-275, 2000.

112

[19] M. R. Movahhedy, Y. Altintas, and M. S. Gadala, "Numerical Analysis of Metal Cutting With Chamfered and Blunt Tools", Journal of Manufacturing Science and Engineering, Vol. 124, No. 2, pp. 178-188, 2002.

[20] H. A. Kishawy, I. M. Deiab and A. J. Haglund, "Arbitrary Lagrangian Eulerian analysis on cutting with a honed tool", Journal of Engineering Manufacturing, Vol. 222, pp. 155-162,2008.

[21] A. J. Shih, "Finite element simulation of orthogonal metal cutting", Journal of Engineering for Industry, Vol. 117, pp. 84-93, 1995.

[22] K. L. Johnson, Contact Mechanics, Cambridge University Press, Cambridge, 1987.

[23] E. Usui and H. Takeyama. "A Photo elastic Analysis Stress", Journal of Engineering for Industry, 82A, 303.

[24] H. A. Kishawy, R. J. Rogers, and N. Balihodzic, "A Numerical Investigation of the Chip Tool Interface in Orthogonal Machining", Machining Science and Technology, Vol. 6, No. 3, pp. 397-414, 2002

[25] C. R. Liu, and Y. B. Guo, "Finite element analysis of the effect of sequential cuts and tool chip friction on residual stresses in a machined layer", International Journal of Mechanical Sciences, Vol. 42, pp. 1069-1086, 2000.

113

[26] K. Potdar and T. Zehnder, "Measurements and Simulations of Temperature and Deformation Fields in Transient Metal Cutting", Journal of Manufacturing Science and Engineering, Vol. 125, pp. 645-655, 2003.

[27] J. Leopold, U. Semmler, and K. Hoyer, "Applicability, robustness and stability of the finite element analysis in metal cutting operations", Proceedings of the 2nd CIRP International Workshop on Modelling of Machining Operations, Vol. 5. pp. 81-94, Nantes, France, 1999.

[28] V. P. Astakhov and J. C. Outeiro, "Modeling of the Contact Stress Distribution at the
T o o l - C h i p Interface", Machining Science and Technology, Vol. 9, No. 1, pp. 85-99, 2005.

[29] A. G. Mamalis, M. Horvath, A.S. Branis and D.E. Manolakos, "Finite element simulation of chip formation in orthogonal metal cutting", Journal of Materials Processing Technology, Vol. 110, pp. 19-27, 2001.

[30] Y. Takeuchi, M. Sakamoto and T. Sata, "Improvement in the working accuracy of an NC lathe by compensating for thermal expansion", Precision Eng., Vol. 4(1), pp. 19-24, 1982.

[31] N. A. Abukhshim, P. T. Mativenga and M.A. Sheikh, "Heat generation and temperature prediction in metal cutting: A review and implications for high speed

114

machining", International Journal of Machine Tools & Manufacture, Vol. 46, pp. 782800, 2005.

[32] H. Blok, "Theoretical study of Temperature Rise at Surfaces of Actual Contact Under Oiliness Lubricating Conditions", Proceedings of the General Discussion on Lubrication and Lubricants, The Institution of Mechanical Engineers, London, England, pp. 222-235, 1938.

[33] J. C. Outeiro, A. M. Dias, J. L. Lebrun, and V. P. Astakhov, "Machining Residual Stresses in AISI316L Steel and Their Correlation with the Cutting Paramters", Machining Science and Technology, Vol. 6:2, pp. 251 - 270, 2002.

[34] D.Y. Jang, T.R. Watkins, K.J. Kozaczek, CR. Hubbard and O.B. Cavin, "Surface residual stresses in machined austenitic stainless steel", Wear, Vol. 194, No. 1, pp. 168173, 1996.

[35] J. Hua, D. Umbrello, and R. Shivpuri, "Investigation of cutting conditions and cutting edge preparations for enhanced compressive subsurface residual stress in the hard turning of bearing steel", Journal of Material Processing Technology, Vol. 171, pp. 180187,2006.

115

[36] M. N. A. Nasr, E. G. Ng and M. A. Elbestawi "Modeling the effects of tool edge radius on residual stresses when orthogonal cutting AISI316L", International Journal of Machine Tools & Manufacture, Vol. 47, pp. 401-411, 2007.

[37] ABAQUS Theory Manual and ABAQUS/EXPLICIT User's Manual, Volume I & II, Version 6.7, Hibbitt, Karlsson & Sorensen, Inc., 2007.

[38] Y. B. Guo, C. R. Liu, "3D FEA Modeling of Hard Turning", Journal of Manufacturing Science and Engineering, Vol. 124, pp. 189-199, 2002.

[39] Nihad Balihodzic, "A numerical investigation of orthogonal machining", M.Sc.Eng. thesis, University of New Brunswick, 2000.

[40] V. Grolleau, Approche de la validation experimentale des simulations numeriques de la coupe avec prise en compte des phenomenes locaux a 1'arete de l'outil , Phd Thesis, E.C. Nantes, 1996.

[41] J. Lin, S. L. Lee and C.I. Weng, "Estimation of cutting temperature in high speed machining", Journal of Engineering Materials and Technology, Vol. 114, pp. 289-296, 1992. [42] C. Genzel, "A self-consistent method for X-ray diffraction analysis of multiaxial residual-stress fields in the near-surface region of polycrystalline materials. I. Theoretical concept", Journal of Applied Crystallography, Vol. 32, pp. 770-778, 1999.

116

[43] Ng. Eu-Gene, T. El-Wardany, M. Dumitrescu, and M. Elbestawi, "3D finite element analysis for the high speed machining of hardened steel", American Society of Mechanical Engineers, Manufacturing Engineering Division, MED, Vol.13, pp. 201-209, 2002.

117

APPENDIX A INPUT FILE FOR ABAQUS EXPLICIT


A finite element code is defined by using ABAQUS Explicit. The data of the model is entered to the input file by using ABAQUS CAE. The input file is made up of comment lines, keyword lines and data lines. Frequently, the block lines of the input file starts with common lines which have a description to the current common. Next, the keyword lines which have parameters. Finally, data lines are used to provide data that are more easily given in lists. Most options require one or more data lines which have the numeric entries.

*Heading ** Job name:


**

Model name:

*Preprint, echo=NO, model=NO, history=NO, contact=NO

** PARTS
**

*Part, name=SHANK *End Part


**

*Part, name=TOOL *End Part


**

*Part, name=WORKPIECE *End Part


**

** ASSEMBLY

118

*Assembly, name=Assembly
**

**

*Instance, name=TOOL-l, part=TOOL *Node

* Element, type=CPE4RT

*Nset, nset=TOOLL, instance=TOOL-l *Nset, nset=_PickedSet2, internal, generate

*Elset, elset=_PickedSet2, internal, generate

** Section: TOOL *Solid Section, elset=_PickedSet2, material=TOOL

*End Instance instance, name=WORKPIECE-l, part=WORKPIECE *Node * Element, type=CPE4RT

*Nset, nset= PickedSet2, internal, generate

=_PickedSet2:, internal, generate *Elset,, elset=

** Section: WP

119

*Solid Section, elset=_PickedSet2, material=WP *End Instance

Instance, name=SHANK-l, part=SHANK *Node *Element, type=T2D2

*Node

*Nset, nset=SHANK-l-RefPt_, internal *Nset, nset=_PickedSet4, internal, generate *Elset, elset=_PickedSet4, internal

** Section: SHANK *Solid Section, elset=_PickedSet4, material=TOOL *End Instance **********************rrQQT -ET-ESET- EXAMPLE ************************

*Nset, nset=ALL, instance=TOOL-l, generate

Elset, elset-ALL, instance=WORKPIECE-l, generate

*Elset, elset=ALL, instance=TOOL-l, generate

*Nset, nset=TOOLL, instance=TOOL-l

120

*******************WORKPIECE-NSET-ESET'EXAMPLE *******************

*Nset, nset=INFLOW, instance=WORKPIECE-l

*Elset, elset=INFLOW, instance-WORKPIECE-1

*Nset, nset=S-l, instance=WORKPIECE-l

*********************gjl^-^j^_^ggy_gggy. EXAMPLE **********************

*Nset, nset=SHANK, instance=SHANK-l, generate

*Elset, elset=SHANK, instance=SHANK-l

*Nset, nset=REF-P, instance=SHANK-l

**********DEFINED-SURFACES-IN-THE-WORKPIECE: EXAMPLE************

*Elset, elset=_0UT-S_S2, internal, instance=WORKPIECE-l

*Elset, elset=_0UT-S_S4, internal, instance=WORKPIECE-l

* Surface, type=ELEMENT, name=OUT-S

*Elset, elset=_EULERCHIP0UT_S3, internal, instance=WORKPIECE-l

* Surface, type=ELEMENT, name=EULERCHIPOUT

121

*Elset, elset=_FREE-S_S4, internal, instance=WORKPIECE-l

Elset, elset=_FREE-S_Sl, internal, instance=WORKPIECE-l Elset, elset=_FREE-S_S2, internal, instance=WORKPIECE-l

* Surface, type=ELEMENT, name=FREE-S

*Elset, elset=_SHANK-l_SNEG, internal, instance=SHANK-l

*Surface, type=ELEMENT, name=SHANK-l SHANK-1SNEG, SNEG *Elset, elset=_SHANK-2_SP0S, internal, instance=SHANK-l

*Surface, type=ELEMENT, name=SHANK-2 SHANK-2SPOS, SPOS Elset, elset=_TOOLR_S2, internal, instance=TOOL-l Elset, elset= TOOLRS1, internal, instance=TOOL-l

* Elset, elset= TOOLRS4, internal, instance=TOOL-l, generate

* Surface, type=ELEMENT, name=TOOLR *Elset, elset= TOOLRTS2,, internal, instance1=TOOL-l * Elset, elset= TOOLRTS3,, internal, instance:=TOOL-l, generate Elset, elset= TOOLRTSL , internal, instance1=TOOL-l

Elset, elset=_TOOLRT_S4, internal, instance=TOOL-l, generate

122

** Constraint: Rigid-Shank *Rigid Body, ref node=REF-P, elset=INSERT, tie nset=SHANK ** Constraint: Tie *Tie, name=Tie, adjust=yes, type=SURFACE TO SURFACE TOOLRT, SHANK-1 *End Assembly

*P 5JC 5j* *jZ m% *jC 5JC 5fC *[C J|C 5|C 5p 3} 5jC p 5JC *(C *J* 5(C 5|C 3f ( #p 5|C *f 3f* *JC 7j* *|C *f* *J( 5JC ^ | \ / | I J I %

I I I 1 1 \-i

**^ *1^ *t* *T* *t* *l* *** ^

M*

*?* *(* *i* *P ^ ^ T^ "P ^ T^ T^ ^ ^ ^ ^ ^ ^ ^

* Amplitude, name=RAMP, definition=SMOOTH STEP 0.,0., 1.5e-05, 1.


**

*********************MATERIAL-MODEL'EXAMPLE **********************

** MATERIALS
**

*Material, name=TOOL * Conductivity 25., *Density 10600., *Elastic 5.2e+l 1,0.22 * Expansion 7.2e-06, * Specific Heat 200., 123

*Material, name=WP * Conductivity 42.6,100. 42.2,200. 37.7,400. 33.,600. *Density 7800., ^Elastic 2.1e+ll, 0.3 * Expansion 1.22e-05,20. 1.37e-05,250. 1.46e-05,500. * Inelastic Heat Fraction 0.9, *Plastic, hardening=JOHNSON COOK 5.98e+08,7.68e+08, 0.2092, 0.0137,0.001 * Specific Heat 473.,200. 519.,350. 561.,550.
**

0.807,

1520.,

25.

*Rate Dependent, type=JOHNSON COOK

****************** INTERACTION PROPERTIES: EXAMPLE *****************

** INTERACTION PROPERTIES
**

* Surface Interaction, name=FRICTION 124

*Friction 0.23, *Gap Conductance le+09, 0. 0., le-08 *Gap Heat Generation 1., 0.632 * Surface Interaction, name=FRICTION-l * Friction 0.23, *Gap Conductance le+09, 0.

0., le-08 *Gap Heat Generation 1., 0.632


**

******************gQjjNi)ARY CONDITIONS' EXAMPLE ******************

** BOUNDARY CONDITIONS
**

** Name: BOTTOM Type: Displacement/Rotation *Boundary BOTTOM, 2, 2 ** Name: REF-P Type: Displacement/Rotation * Boundary REF-P, 1, 1 REF-P, 2, 2 REF-P, 6, 6
**

** PREDEFINED FIELDS

125

** Name: TEMPERATURE-1 Type: Temperature * Initial Conditions, type=TEMPERATURE ALLW, 25. ** Name: TEMPERATURE-2 Type: Temperature * Initial Conditions, type=TEMPERATURE INSERT, 25.

**

** STEP: STEP
**

*Step, name=STEP DYNAMIC, TEMP-DISP, EXPLICIT *Dynamic Temperature-displacement, Explicit , 0.0006 *Bulk Viscosity 0.06, 1.2

******************************JyT A C SCALING***************************

** Mass Scaling: Semi-Automatic **


**

ABC

*Fixed Mass Scaling, elset=ABC, factor=50.

###*******************#**goUNDARY CONDITIONS**********************
*
*i* J p 5]> 5J Jp *J* /(* J[* #|* #[* *|* J|C J* JJ5 5[*t* *t* *i* *i* *3+J| *p y^ *J *^ *^ ^ ^ ^ *^ ^ ^ ^ *(* 3J* 5 ^ 5fl ^> # ^ ^ ^^ ^ ^ *^ J *^ ^ *^ ^ ^ ^ ^ ^ ^ ^ ^ *^ ^ ^ t^ ^ t*|* *^ >^ *J *^ t^ ^ ^ ^ *^ ^ ^ ^ ^ ^ n* *>* ^* *i* *l* ? ^tf *fc ^1; ^If ^ ^* ^fc ^ *l* *i* *^ J ^ J ^ ^ *|> *fc ^* *^ * ^ ^ ^ ^f v|* *fc J *i* *T* *T* ^* *T* *fe 5[ *i* *t* *i* *i* *l# *l *l* *l* l# *i i *i *i* SJ ?J* Sf* *t* f* l# j* *l ^ ^ ^ *J* ^*

** BOUNDARY CONDITIONS
**

126

** Name: INFLOW Type: Velocity/Angular velocity *Boundary, amplitude=RAMP, type=VELOCITY INFLOW, 1, 1,3.33 * Adaptive Mesh Controls, name=ALE, geometric enhancement=YES, mesh constraint angle=20. l.,0.,0. * Adaptive Mesh, elset=ALLW, controls=ALE, initial mesh sweeps=5, mesh sweeps=5, op=NEW
**

** ADAPTIVE MESH CONSTRAINTS


**

** Name: OUTFLOW Type: Displacement/Rotation * Adaptive Mesh Constraint S-2, 2, 2, 0.0 S-l, 1, 1,0.0 OUTFLOW, 1, 1,0.0 *********************** JNTERACTIONS'EXAMPLE ***********************

** INTERACTIONS
**

** Interaction: FRICTION * Contact Pair, interaction=FRICTION, mechanical constraint=PENALTY, weight=l., cpset=FRICTION TOOLL, CHIPCONTACT ** Interaction: FRICTION-1 ^Contact Pair, interaction=FRICTION-l, mechanical constraint=PENALTY,

cpset=FRICTION-l RAKE-F, CHIPCONTACTUP ** Interaction: SFILM-1 *Sfilm

127

FREE-S, F, 25., 10. ** Interaction: SFILM-2 Sfilm Flank-F, F, 25., 140. ** Interaction: SFILM-3 * Sfilm OUT-S, F, 25., 10.
**

*******************Qjjypjjp REOUESTS' EXAMPLE ***********************

** OUTPUT REQUESTS
**

* Restart, write, number interval=l, time marks=NO


**

** FIELD OUTPUT: F-Output-1


**

*Output, field, number interval=400 *Node Output A, NT, RF, RFL, U, V *Element Output, directions=YES ER, HFL, LE, PE, PEEQ, S * Contact Output CFORCE, CSTRESS, FSLIP, FSLIPR
**

** HISTORY OUTPUT: H-Output-1


**

*Output, history, variable=PRESELECT, time interval=le-05 *End Step


************************************************************************

128

VITAE
Name: Abdulfatah Maftah EDUCATION Master of Science in Mechanical Engineering University of New Brunswick, Fredericton, NB (Jan. 2006 - April. 2008)

Thesis: Finite Element Simulation of Orthogonal Metal Cutting Using an ALE Approach Related Courses: Flow Induced Vibrations, Principle of Metal Cutting, Mechanics of Continua, Applied of Finite Elements Bachelor of Science in Mechanical Engineering Seventh April University, Sabratha, Libya Thesis: Design of Heat Exchangers using Heat Pipes (Sept. 1993 - Apr. 1998)

Conference Paper: A. Maftah, H. Kishawy, and R. Rogers "FINITE ELEMENT SIMULATION OF ORTHOGONAL METAL CUTTING WITH DIFFERENT INITIAL GEOMETRIES", Third Mechanical Engineering Graduate Students Conference University of New Brunswick, November 12, 2007.

You might also like