See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/257351589

Finite Element Modeling of Chip Formation in Orthogonal Machining

Chapter · February 2012

DOI: 10.1007/978-3-642-25859-6_3

CITATIONS READS

14 6,501

3 authors:

Amrita Priyadarshini Surjya K Pal

BITS Pilani, Hyderabad Indian Institute of Technology Kharagpur
16 PUBLICATIONS   37 CITATIONS    202 PUBLICATIONS   3,156 CITATIONS   

SEE PROFILE SEE PROFILE

Arun Kumar Samantaray

Indian Institute of Technology Kharagpur
197 PUBLICATIONS   2,774 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Fault Tolerant Control and Reconfiguration of Walking Robots View project

Lagrangian Mechanics View project

All content following this page was uploaded by Arun Kumar Samantaray on 18 November 2016.

The user has requested enhancement of the downloaded file.

3

Finite Element Modeling of Chip Formation

in Orthogonal Machining

Amrita Priyadarshini, Surjya K. Pal, and Arun K. Samantaray

Department of Mechanical Engineering, Indian Institute of Technology,

Kharagpur, Kharagpur- 721 302, West Bengal, India
[email protected],
[email protected],
[email protected]

Finite element method has gained immense popularity in the area of metal cutting for
providing detailed insight in to the chip formation process. This chapter presents an
overview of the application of finite element method in the study of metal cutting
process. The basics of both metal cutting and finite element methods, being the fore-
most in understanding the applicability of finite element method in metal cutting,
have been discussed in brief. Few of the critical issues related to finite element mod-
eling of orthogonal machining have been cited through various case studies. This
would prove very helpful for the readers not simply because it provides basic steps
for formulating FE model for machining but also focuses on the issues that should be
taken care of in order to come up with accurate and reliable FE simulations.

3.1 Introduction
Metal cutting or machining is considered as one of the most important and versa-
tile processes for imparting final shape to the preformed blocks and various manu-
factured products obtained from either casting or forging. Major portion of the
components manufactured worldwide necessarily require machining to convert
them into finished product. This is the only process in which the final shape of the
product is achieved through removal of excess material in the form of chips from
the given work material with the help of a cutting tool. Basic chip formation
processes include turning, shaping, milling, drilling, etc., the phenomenon of chip
formation in all the cases being similar at the point where the cutting edge meets
the work material. During cutting, the chip is formed by deforming the work ma-
terial on the surface of the job using a cutting tool. The technique by which the
metal is cut or removed is complex not simply because it involves high straining
and heating but it is found that conditions of operation are most varied in the
102 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

machining process as compared to other manufacturing processes. Although nu-

merous researches are being carried out in the area of metal cutting both for its
obvious technical and economical importance, machining still is sometimes re-
ferred to as one of the least understood manufacturing processes because of the
complexities associated with the process. Efforts are being made continuously to
understand the complex mechanism of cutting in a simple and effective way.
The knowledge of basic mechanism of chip formation helps understanding the
physics governing the chip formation process. This enables one to solve the prob-
lems associated with metal cutting so that the metal removal process can be car-
ried out more efficiently and economically. A significant improvement in process
efficiency may be obtained by process parameter optimization that leads to desired
outputs with acceptable variations ensuring a lower cost of manufacturing. Re-
searchers have developed various techniques and the associated studies that have
contributed remarkably in the area of metal cutting. The age old technique of trial
and error experiment has contributed greatly and is still widely used in metal cut-
ting research. Various mathematical models have also been developed which form
the core of the metal cutting theory. These models range from simplified analyti-
cal models to complex, computer-based models using Finite Element Method
(FEM).
Metal cutting tests have been carried out for at least 150 years using a variety of
cutting conditions in tremendously increasingly volume [1]. Frederick W. Taylor
is regarded as the great historical figure in the field of metal cutting whose works
published at the end of the nineteenth century had a great impact on manufactur-
ing industry. In industries, the conventional way to optimize metal cutting process
is to conduct experimental tests with varying cutting conditions and then analyze
its effect on various outputs, e.g., cutting forces, temperatures, tool wear and sur-
face finish based on optimization techniques such as Design of Experiment
(DOE), Taguchi or Response Surface Methodology (RSM). The experimental
study is as important as the theoretical study. Unfortunately, the experimental op-
timization approach requires large number of experiments to achieve desired accu-
racy, is very time consuming and expensive.
For more than half a century the analysis of the machining problems has been
carried out by researchers using different analytical models such as shear plane so-
lution and slip line field method [2]. Pioneering works to determine cutting tem-
perature using analytical approach began in 1951 by Hahn [3] followed by Chao
and Trigger [4], Leone [5], Loewan and Shaw [6], Weiner [7] and Rapier [8]. In
analytical modeling, especially in earlier works, uniform plane heat source and ve-
locity discontinuity were often assumed by considering that the chip is formed in-
stantaneously at the shear plane. The secondary deformation zone was usually
neglected and the tool chip frictional force was taken as uniform. Such simplified
assumptions modify the original problem to some extent. Besides, various other
important machining features such as strain hardening, temperature dependence,
chip curling, etc., are neglected owing to the complexity associated with the ma-
chining process.
In case of metal cutting process, taking all the boundary conditions into consid-
eration renders the mathematical equations that make the process so complicated
Finite Element Modeling of Chip Formation in Orthogonal Machining 103

that solutions to such problem seldom exist. One prospect, as done in case of analyt-
ical approach, is to make simplifying assumptions in order to ignore the difficulties
and reduce the problem to one that can be solved. This, however, modifies the actual
problem to great extent and leads to serious inaccuracies. Now that more and more
powerful computers have emerged and are being widely used, a more viable alterna-
tive is to obtain approximate numerical solutions rather than exact form solutions.
This enables one to retain the complexity of the problem on one hand, and the de-
sired accuracy on the other. The most popular numerical technique that has evolved
in recent decades for the analysis of metal cutting is FEM. FEM allows the coupled
simulation of plastic and thermal process and is capable of considering numerous
dependencies of physical constants on each other.
The advantages of FEM over empirical and analytical methods in machining
process can be summed up as follows [9]:
• some of the difficult to measure variables, namely, stress, strain, strain
rate, temperature, etc., can be obtained quantitatively, in addition to cut-
ting forces and chip geometry.
• non-linear geometric boundaries such as the free surface of the chip can
be represented and used.
• material properties can be handled as functions of strain, strain rate and
temperature.
• the interaction between the chip and the tool can be modeled as sticking
and sliding conditions.
Thus, FEM-based analysis provides detailed qualitative and quantitative insight
in to the chip formation process that is very much required for profound under-
standing of the influence of machining parameters. While experimental tests and
analytical models serve as the foundation of metal cutting, FEM leads to the ad-
vancement and further refinement of knowledge in the area of metal cutting.

3.2 Basics of Machining

Metal cutting process involves various independent and dependent variables. In-
dependent variables are the input variables over which the machinist has direct
control [10]. These include type of workpiece material and tool material, shape
and size of workpiece material, cutting tool geometry, type of machining process,
cutting parameters (speed, feed and depth of cut) and cutting fluids. The type of
input parameters selected during machining process decides much about the de-
pendent variables. The important dependent or output variables are cutting force
and power, surface finish, tool wear and tool failure, size and properties of the fi-
nished product. A small change in input variables, say, cutting parameters, tool
geometry and workpiece or tool material may alter the forces to great extent. Sur-
face finish is again a function of tool geometry, tool material, workpiece material
and cutting parameters. Tool wear is also a crucial aspect of machining pertaining
to the economics of machining since longer tool life leads to higher productivity.
Moreover, as the tool wear takes place, it changes in both geometry and size. Such
a change can result in increased cutting forces which in turn increase deflection in
104 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

the workpiece and may create a chatter condition. Again due to increased power
consumption, there can be increased heat generation, thus accelerating the wear
rate. The enormous variety in input variables leads to infinite combinations and
understanding the interrelationship between these input variables and output va-
riables again becomes an arduous task.

3.2.1 Orthogonal Cutting Model

During cutting as the tool touches the work material, severe plastic deformation
occurs within the workpiece material and the excess of material starts flowing
over the rake face of the cutting tool in the form of chip. The zone of plastic de-
formation known as shear plane lies between the chip and the undeformed work
material. Based on the position of cutting edge, the cutting process can be classi-
fied as orthogonal cutting and oblique cutting. In orthogonal cutting, also known
as two-dimensional cutting, the cutting edge is perpendicular to the cutting veloci-
ty vector. On the other hand, in oblique cutting, the edge is inclined with the cut-
ting velocity by a certain angle called the inclination angle. Oblique cutting is a
common type of three-dimensional cutting used in machining process. Orthogonal
cutting is only a particular case of oblique cutting such that any analysis of ortho-
gonal cutting can be applied to oblique cutting [11]. As far as practical require-
ments of rake and other angles are concerned, the ideal conditions of orthogonal
cutting are rarely achieved. But analysis of oblique cutting is much more difficult,
so focus has been mainly given on the orthogonal cutting with very few excep-
tions dealing with the mechanics of oblique cutting. The important theories based
on orthogonal cutting model which have paved the way for the analysis of chip
formation process include Merchant’s model [12], Lee and Shaffer’s model [13],
Oxley’s model [14], just to name a few.
The significant geometrical parameters involved in chip formation (as shown in
Fig. 3.1) are described as follows:

Chip

tc
bc
bo
γo Contact length

to β

Workpiece Vc

Fig. 3.1. Orthogonal cutting model

Finite Element Modeling of Chip Formation in Orthogonal Machining 105

β = Shear plane angle which is defined as the angle between the shear plane
(plane of separation of chip from the undeformed work material) and the cutting
velocity vector

γ0 = Rake angle of the cutting tool

φp = Principal cutting edge angle (shown in Fig. 3.2)

λ = Inclination angle
d = Depth of cut (mm)
s = Feed (mm/rev)
Vc = Cutting velocity (m/min)
to , bo , lo are thickness, width and length (mm) of the uncut chip thickness, re-
spectively, such that

s sin φ p
t0 =
cos λ
(3.1)
d
b0 =
sin φ p

For pure orthogonal cutting case, λ = 0° and φ p = 90° giving t0 = s and

b0 = d .

tc , bc , lc are thickness, width and length (mm) of the deformed chip thickness, re-
spectively.
ζ , Chip reduction coefficient is defined as the index of the degree of deformation
involved in the chip formation such that
tc
ζ = (3.2)
t0
The degree of chip thickening can also be expressed in terms of cutting ratio
⎛ 1⎞
r⎜= ⎟ .
⎝ ζ ⎠
Since tc > t0 , ζ is generally greater than one. Larger value of ζ signifies high-
er value of cutting forces that are required to carry out the machining process. The
chip reduction coefficient is affected by the tool rake angle and chip-tool interfacial
106 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

friction coefficient ( μ ). This is clear from the relation obtained from velocity anal-
ysis based on D’Alemberts’ principle applied to chip motion, given as [15]:

ζ =e
(
μ Π 2 −γ 0 ) (3.3)

The ζ can be reduced considerably by using cutting tools with larger positive
rake angles and by using cutting fluids.
Shear angle can be expressed in terms of ζ :

cos γ 0
tan β = (3.4)
ζ − sin γ 0
It is evident from Eq. (3.4) that decrease in ζ and increase in rake angle tend to
increase the shear angle. This suggests that increasing values of shear angle re-
quires lesser forces for cutting resulting in favorable machining.

3.2.2 Cutting Forces

The first scientific studies of metal cutting were conducted in about 1850, which
were aimed at establishing the power requirements for various operations. A con-
siderable amount of investigations has been directed toward the prediction and
measurement of cutting forces. This is because cutting force is a result of the ex-
treme conditions at the tool-workpiece interface and this interaction can be direct-
ly related to many other output variables such as generation of heat and conse-
quently tool wear and quality of machined surface as well as chip formation
process and the chip morphology [16, 17]. Measurement of forces becomes man-
datory at certain cases say, adequate equations are not available, evaluation of ef-
fect of machining parameters cannot be done analytically and theoretical models
have to be verified. Several works are available in the literature that makes use of
different types of dynamometers to measure the forces. The dynamometers being
commonly used nowadays for measuring forces are either strain gauge or piezoe-
lectric type. Though piezoelectric dynamometer is highly expensive, this has al-
most become standard for recent experimental investigations in metal cutting due
to high accuracy, reliability and consistency.
Estimation of forces acting between tool and work material is one of the vital
aspects of mechanics of cutting process since it is essential for:
• Determination of the cutting power consumption
• Structural design of machine-fixture-tool system
• Study of the effect of various cutting parameters on cutting forces
• Condition monitoring of both the cutting tools and machine tools
Analysis of force during machining includes magnitude of cutting forces and
their components and location of action of those forces as well as the pattern of
the forces, say, static or dynamic.
Finite Element Modeling of Chip Formation in Orthogonal Machining 107

A single cutting force is known to act in case of a single point cutting tool be-
ing used in turning operation which can be resolved into three components along
three orthogonal directions i.e. X , Y and Z , as shown in Fig. 3.2.

Workpiece

FX
FY FT

Φp

Cutting tool

Feed

Fig. 3.2. Forces acting on a cutting tool in turning

The resultant cutting force R is resolved as:

R = Fc + FT (3.5)

FT = FX + FY (3.6)

such that FX = FT sin φ p and FY = FT cos φ p .

The force components are described as follows:

Fc – This is the main or tangential cutting force acting in the direction of cutting
velocity. This when multiplied with cutting velocity gives the value of cutting
power consumption.

FX – It is feed force acting in the direction of feed.

FY – This is radial force acting in the direction perpendicular to the feed force.
FT – This is known as thrust force.
The above-mentioned forces act on the cutting tool during machining process and
can be measured directly using tool force dynamometers. But there exist another
set of forces acting on the chip both from the tool side and the workpiece side that
108 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

cannot be measured directly. Merchant’s circle diagram, one of the oldest

techniques, is widely used for analytical estimation of these kind of forces for
orthogonal cutting process.
Merchant’s circle diagram is shown in Fig. 3.3 with schematic representation of
forces acting on a portion of continuous chip passing through the shear plane at a
constant speed which is in a state of equilibrium.

Chip

Shear plane
Fs γo Tool
Fc Machined surface
β

FT Ns R
F
η
N
Workpiece

Fig. 3.3. Merchant’s circle diagram for 2D orthogonal cutting

Fs and N s are called shear force and normal force, respectively, that act on
the chip from workpiece side i.e. in the shear plane. F and N are friction force
at chip-tool interface and force normal rake face, respectively, that act on the chip
from the tool side i.e. in the chip-tool interface. These forces can be determined as
follows [10]:

Fs = Fc cos β − FT sin β
(3.7)
N s = FT cos β + Fc sin β
F = FT cos γ 0 + Fc sin γ 0
(3.8)
N = Fc cos γ 0 − FT sin γ 0
The average coefficient of friction between chip and tool ( μ ) can be deduced
either in terms of friction angle ( η ) or F and N , as given:
F
μ = tan η = (3.9)
N
Finite Element Modeling of Chip Formation in Orthogonal Machining 109

3.2.3 Cutting Temperature

Heat has a critical influence on machining process. It is found that almost all of
the mechanical energy is converted into heat during chip formation process that
tends to increase the temperature to very high values in the cutting zone. The heat-
ing action is known to pose many of the economic and technical problems of ma-
chining. During machining heat is generated in the cutting zone from three distinct
zones: Primary shear deformation zone (PSDZ), secondary shear deformation
zone (SSDZ) and tertiary shear deformation zone (TSDZ) (see Fig. 3.4). Major
part of the heat is generated due to severe plastic deformation in a very narrow
zone called PSDZ. Further heating is generated in the SSDZ due to friction be-
tween tool and chip as well as shearing at the chip-tool interface. It is considered
that both these zones together account for almost 99% of the total energy con-
verted into heat in the cutting process [18]. A very small fraction of heat is gener-
ated at the TSDZ due to rubbing between flank of the tool and machined surface.
The effect of this zone comes into play only when the flank wear develops. The
TSDZ is therefore neglected when studying unworn tools [19]. The heat so gener-
ated is shared by the chip, cutting tool and the workpiece, the major portion of
heat being carried away by the flowing chip. As chip is meant for disposal only,
focus should be on the chip to take as much heat as possible with it so that small
amount of heat gets retained in the cutting tool.

Chip SSDZ
PSDZ
Tool

TSDZ
Workpiece

Fig. 3.4. Heat generation zones during metal cutting

Elevated temperature in the cutting zone adversely affects the strength, hard-
ness and wear resistance of the cutting tool, thereby inducing rapid tool wear and
reduced tool life. Temperature rise in workpiece material may cause dimensional
inaccuracy of the machined surface and can also damage the surface properties of
the machined component by oxidation, rapid corrosion, burning, etc. Thus, estima-
tion of cutting temperature is a crucial aspect in the study of metal cutting.
Cutting temperatures are more difficult to measure accurately than cutting
forces. No simple analog to the cutting force dynamometer exists for measuring
the cutting temperatures; rather numerous methods have been found in the litera-
ture to experimentally measure the machining temperature [20]. These methods
are thermocouples, radiation methods, metallographic techniques and application
of thermal paints, fine powders and Physical Vapor Deposition (PVD) coatings
[21]. A particular method generally gives only limited information on the com-
plete temperature distribution.
110 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

3.2.4 Chip Morphology

Chip formation and its morphology are the key areas in the study of machining
process that provide significant information on the cutting process itself. The
process variables such as cutting force, temperature, tool wear, machining power,
friction between tool-chip interface and surface finish are greatly affected by the
chip formation process and chip morphology. Chip is formed due to deformation
of the metal lying ahead of the cutting tool tip by shearing process. The extent of
deformation that the work piece material undergoes determines the nature or type
of chip produced. The extent of deformation of chips again depends upon cutting
tool geometry (positive or negative rake angle), workpiece material (brittle or duc-
tile), cutting conditions (speed, feed and depth cut), machining environment (dry
or wet machining) [10].
The main chip morphologies observed in cutting process are described briefly
as follows:
Discontinuous chip: These chips are small segments that adhere loosely to each
other. The phenomenon can be attributed to the repeated fracturing that limits the
amount of deformation the chip undergoes. Hard and brittle materials like gray
cast iron, bronze, brass when machined with larger feed and negative rake cutting
tool in the absence of cutting fluid produce discontinuous chips.
Continuous chip: Continuous chips are in the form of long coil. These are
formed by the continuous plastic deformation of material without fracture ahead
of the cutting edge of the tool resulting in a smooth flow of chip up the tool face.
Ductile material when machined under low feed and high cutting speed, generally,
in the presence of cutting fluid produces such kind of chips.
Continuous chip with built-up edge: This kind of chip is not as smooth as that of
continuous chip and affects the surface finish adversely. Built-up edge is mostly
found when ductile materials are machined under larger feed and lower cutting
speed with inadequate or no cutting fluid.
Cyclic or serrated chips: These chips are continuous, but, possess a saw-tooth
appearance that is produced by a cyclical chip formation of alternating high shear
strain followed by low shear strain. Machining of certain difficult to machine
metals such as titanium alloys, nickel-base super alloys and austenitic stainless
steels are generally characterized by formation of segmented chip formation [22].

3.3 Basics of FEM

The considerable rise in computer processing speeds makes it possible to incorpo-
rate the FEM for analyzing various aspects of metal cutting process, thus over-
coming most of the restrictive assumptions associated with analytical models.
Many investigators have adopted FEM to gain better understanding of machining
process. One of the first FE models developed for metal cutting process was by
Klamecki in the year 1973 [23]. Since then many works have been found to use
FEM for gaining better understanding of the machining process. Numerous FE
Finite Element Modeling of Chip Formation in Orthogonal Machining 111

codes such as DEFORM, FORGE2, ABAQUS/Standard, ABAQUS/Explicit, An-

sys/LS-DYNA, MSC Marc and Thirdwave AdvantEdge have come up that are be-
ing used by the researchers, thus giving results closer to the experimental ones.
However, it should be noted that the accuracy of FE modeling is determined by
how adequately the characterization of selected input parameters reflects the de-
formation behavior undergoing during the chip formation in the actual practice.

3.3.1 Generalized Steps in FEM

The finite element method is a numerical method seeking an approximated solu-
tion to a wide variety of engineering problems. The engineering problems are bas-
ically boundary value problems (or field problems) containing one or more depen-
dent variables that must satisfy a differential equation everywhere within a known
domain and satisfy specific conditions on the boundary of the domain. The field is
the domain of interest, often representing a physical structure while the dependent
variables of interest are called field variables. The specified values of the field va-
riables on the boundaries of the field are called boundary conditions. The field va-
riables may include displacement in solid mechanics problem, temperature in heat
flow problem, velocity in fluid flow problem and so on. In formulating an FE
model for a physical problem, few basic steps are required to be followed that are
common to all kinds of analysis whether structural, heat flow or fluid flow prob-
lems. The basic concepts, procedures and formulations of FEM can be found in
many available books [24–30].
Discretization: The idea of FEM is to approximate the problem domain by replac-
ing it with a finite number of distinct elements and solving them separately. This
is done by subdividing the domain into smaller units of simpler geometry called
finite elements. Such a process is called discretization of domain [25]. The number
of elements used to mesh the domain (mesh density) depends upon the type of
element chosen, accuracy deserved and computer resource available. The elements
selected must be small enough to give practical results and yet large enough to
reduce computational effort.
Interpolation functions: The interpolation function or shape function is in the
form of polynomial which is defined within an element using the values of field
variables at the nodes. For problems involving curved boundaries, the construction
of shape functions that satisfy consistency requirements for higher order elements
becomes increasingly complicated. Moreover, exact evaluation of the integrals
that appear in the expressions of the element stiffness matrix is not always feasible
because of the algebraic complexity. These two obstacles can be surmounted
through the concepts of isoparametric elements and numerical integration, respec-
tively. The numerical integration or numerical quadrature, the most popular of
which is Gaussian (Gauss–Legendre) quadrature involves approximation of the in-
tegrand by a polynomial of sufficient degree because the integral of a polynomial
can be evaluated exactly [30].
112 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

Derivation of element stiffness equation: Derivation of basic equations for a fi-

nite element is based on governing equations and constitutive laws. This prefera-
bly involves weak form of system equation which is often an integral form and re-
quires a weaker requirement on the field variables and the integral operation. A
formulation based on weak form usually leads to a set of algebraic system equa-
tions when the problem domain, no matter how complex, is discretized into small-
er elements [29]. There are three different methods for obtaining weak form,
namely, direct method (for simple problems), variational method and weighted re-
sidual method [30]. The variational method is generally based on either principle
of virtual work, principle of minimum potential energy or Castigliano’s theorem to
derive the element equation [25]. Whereas, method of weighted residual (MWR)
uses trial functions to minimize the error over the problem domain [28]. MWR is
particularly useful when functional such as potential energy is not readily availa-
ble. Galerkin method is one such technique which is widely used for its simplicity
and easy adaptability in FEM. The constitutive equations are employed which are
nothing but the stress-strain laws giving much of the information about the materi-
al behavior. Finally, the element stiffness matrix is derived by substituting the in-
terpolation functions into the weak form.
Assembly of elements: The individual element equations are combined together
using a method of superposition called direct stiffness method to obtain the global
stiffness matrix.
Boundary conditions: After the global equations are ready, necessary boundary
conditions are imposed to the global equations which lead to a set of simultaneous
equations
Solution of system equations: The system equations are often solved by using direct
method such as Guass elimination method and LU decomposition method for small
systems. While iterative methods namely, Gauss–Jacobi method, Gauss-Siedel me-
thod, generalized conjugate residual method, etc. are generally employed for large
systems. For transient problems (highly dynamic and time-dependent problems),
another loop is required meant for time stepping. The widely used method is called
direct integration method such as explicit and implicit methods [29].

3.3.2 Modeling Techniques

With the emergence of more and more potential computers as well as powerful
general purpose FE software packages, FEM is becoming a popular tool in the
analysis of various kinds of manufacturing processes. It is true that there have
been remarkable developments in both computer hardware and software but prop-
er background in FEM is a must otherwise using FEM packages for analysis
would simply appear as a black box. It is the task of the researcher to make use of
such resources tactfully and judiciously to come up with favorable and meaningful
results. Thus, in order to develop a good FE model, few critical issues which may
be referred as modeling techniques must be taken care of. This not only improves
the computational efficiency but also ensures the reliability and accuracy of the
obtained results [29].
Finite Element Modeling of Chip Formation in Orthogonal Machining 113

3.3.2.1 Type of Approach

The selection of an appropriate formulation or approach is of prime importance

especially for problems involving nonlinearity and large deformations. There are
mainly three ways to formulate the problems involving the motion of deformable
materials, given as:
• Lagrangian approach
• Eulerian approach
• Arbitrary Lagrangian Eulerian (ALE) approach
The formulations are distinguished from each other by mesh descriptions, kinetics
and kinematics description [31].
Lagrangian and Eulerian formulations are the two classical approaches for
description of motion in continuum mechanics that differ mostly in terms of the
behavior of nodes. In Lagrangian meshes, the nodes and elements move with
the material i.e. the nodes are coincident with material points. But if the mesh is
Eulerian, the Eulerian coordinates of nodes are fixed, i.e. the nodes are coincident
with spatial points.
Lagrangian approach: It can be seen that in Lagrangian approach, both the ma-
terial points and nodes change position as the body deforms such that position of
the material points relative to the nodes remains fixed. Boundary nodes tend to
remain along the element edges throughout the evolution of the problem which in
turn simplifies the imposition of boundary conditions and interface conditions.
Quadrature points also remain coincident with the material points which signify
constitutive equations are always evaluated at the same material points, thus lead-
ing to a straightforward treatment of constitutive equations. For these reasons, La-
grangian approach is widely used in solid mechanics. However, this approach may
encounter severe mesh distortion because the mesh deforms with the material.
This becomes highly critical when the problems involve large deformation as this
may affect the performance of FE simulations greatly.
Lagrangian formulation can be again of two types, namely, total Lagrangian
formulations and updated Lagrangian formulations [31]. In the total Lagrangian
formulation, the weak form involves integrals over the initial (reference) configu-
ration and derivatives are taken with respect to the material coordinates. In the up-
dated Lagrangian formulation, the weak form involves integrals over the current
i.e. deformed configuration and the derivatives are with respect to the spatial (Eu-
lerian) coordinates. It has been suggested that the updated and total Lagrangian
formulations are two representations of the same mechanical behavior and each
can be transformed to the other, thus, implying that the choice of formulation is
purely a matter of convenience.
Many of the researchers have used Lagrangian formulation in metal cutting
models since it is easy to implement and is computationally efficient because of
easy tracking of the material boundaries. But difficulties arise when elements get
highly distorted as the mesh deforms with the deformation of the material in front
of the tool tip. Since element distortion degrades its element accuracy, the degree
114 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

of deformation that can be simulated with a Lagrangian mesh is restricted. In order

to overcome this, either chip separation criteria or remeshing techniques have been
used in the literature to simulate the cutting action at the cutting zone [32].

Eulerian approach: In Eulerian approach, the reference frame is fixed in space

that allows for the material to flow through the grid [33]. As the mesh is fixed in
space, the numerical difficulties associated with the distortion of elements are
eliminated. This approach permits simulation of machining process without the
use of any mesh separation criterion. The main drawback of Eulerian formulation
is that it is unable to model free boundaries since boundary nodes may not be
coincident with the element edge. This may only be used when boundaries of the
deformed material are known a priori. Hence, dimension of the chip must be spe-
cified in advance to produce a predictive model for chip formation in an Eulerian
framework [34]. Furthermore, handling and updating of constitutive equations
become complex because of the fact that material flows through the mesh i.e.
element quadrature points may not be coincident with material points. These
drawbacks often limit the application of Eulerian formulation in modeling the
metal cutting process.

Arbitrary Lagrangian Eulerian approach: It is known that Eulerian and La-

grangian approaches have their own advantages and disadvantages. As the name
suggests, in ALE formulation the nodes can be programmed to move arbitrarily in
such a way that advantages of both Lagrangian and Eulerian formulations can be
exploited and combined into one while minimizing their disadvantages. The idea
is that the boundary nodes are moved to remain on the material boundaries, while
the interior nodes are moved to minimize mesh distortion. While simulating metal
cutting process in ALE framework, ALE reduces to a Lagrangian form on free
boundaries while maintains an Eulerian form at locations where significant defor-
mations occur, as found during the deformation of material in front of the tool tip.
This approach not only enables to reduce mesh distortion without the need of re-
meshing but also eliminates the need of a priori assumption about the shape of the
chip [35].

3.3.2.2 Geometry Modeling

Creating the geometry of the problem domain is the first and foremost step in any
analysis. The actual geometries are usually complex. The aim should not be simp-
ly to model the exact geometry as that of the actual one, instead focus should be
made on how and where to reduce the complexity of the geometry to manageable
one such that the problem can be solved and analyzed efficiently without affecting
the nature of problem and accuracy of results much. Hence, proper understanding
of the mechanics of the problem is certainly required to analyze the problem and
examine the geometry of the problem domain. It is generally aimed to make use of
2 D elements rather than 3 D elements since this can drastically reduce the number
of degrees of freedom (DOFs).
Finite Element Modeling of Chip Formation in Orthogonal Machining 115

3.3.2.3 Computation Time

Computation time which is nothing but the time that CPU takes for solving finite
element equation is affected markedly by the total number of DOFs in the FE equ-
ation as shown [29]:

tcomp ∝ DOF β (3.10)

Here, β is a constant which generally lies in the range of 2-3 depending on the
type of solver used and the structure or bandwidth of the stiffness matrix. A small-
er bandwidth yields smaller value of β resulting in faster computation. The Eq.
(3.10) suggests that a finer mesh with larger number of DOFs increases the com-
putation time exponentially. Thus, it is always preferred to create FE model with
elements possessing lower dimensions so that number of DOFs are reduced as far
as possible. In addition, meshing should be done in such a way that critical areas
possess finer meshing while others possess coarse meshing. This is one way of re-
ducing the computation time without hampering the accuracy of results.

3.3.2.4 Mesh Attributes

Choice of element type: Based on the shape, elements can be classified as one
dimensional (line or beam), two-dimensional or plane (triangular and quadrilater-
al) and three-dimensional (tetrahedral and hexahedral) elements. Each of these
elements can again be either in their simplest forms i.e. linear or higher order
forms such as quadratic or cubic depending upon the number of nodes an element
possess. As discussed earlier, 2D elements are generally preferred over 3D ele-
ments as far as cost of computation is concerned. The linear triangular element
was the first type of element developed for defining 2D geometries, the formula-
tion of which is simplest of all. However, quadrilateral elements are mostly pre-
ferred nowadays for 2D solids [36]. This is not simply because quadrilateral ele-
ment contains one node more than that of triangular elements but also gradients of
quadrilateral elements are linear functions of the coordinate directions compared
to the gradient being constant in triangular elements. Besides, the number of ele-
ments is reduced to half that of a mesh consisting of triangular elements for the
same number of nodes. In nearly all instances, a mesh consisting of quadrilateral
elements is sufficient and usually more accurate than that of triangular elements.
Besides for higher order elements, more complex representations are achieved that
are increasingly accurate from an approximation point of view. However, care
must be taken to evaluate the benefits of increased accuracy against the computa-
tional cost associated with more sophisticated elements.
Mesh density: A mesh of varying density is generally preferred. The critical areas
need to be finely meshed while rest of the areas may contain coarse mesh. The
control of mesh density can be performed by placing the nodes according to the
given geometry and required degree of accuracy in FEM packages.
116 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

Element distortion: It is not always possible to have regularly shaped elements

for irregular geometries. Irregular or distorted elements are acceptable in the FEM,
but there are limitations, and one needs to control the degree of element distortion
in the process of mesh generation. If these are overlooked then there are chances
of termination of simulation program because of element distortion. Few possible
forms of distortion error of elements are aspect ratio distortion, angular distortion,
curvature distortion and so on. Aspect ratio error arises when the elongation of
element occurs beyond a particular limit disturbing the aspect ratio of the element.
Angular distortion occurs when the included angle between edges of the element
approaches either 0° or 180°. Similarly, when the straight edges of the element are
distorted into curves, curvature distortion is most likely to occur [29].

3.3.2.5 Locking and Hourglassing

Approximating the problem domain using simpler elements is relatively easy, but,
very often formulation of these elements can give inaccurate results due to some
abnormalities namely, shear locking, volumetric locking and hourglassing. Lock-
ing, in general, is the phenomenon of exhibiting an undesirable stiff response to
deformation in finite elements. This kind of problem may arise when the element
interpolation functions are unable to approximate accurately the strain distribution
in the solid or when the interpolation functions for the displacement and their de-
rivatives are not consistent. Shear locking predominantly occurs in linear elements
with full integration and results in underestimated displacements due to undesira-
ble stiffness in the deformation behavior. This problem can be alleviated by using
reduced integration scheme. Volumetric locking is exhibited by incompressible
materials or materials showing nearly incompressible behavior resulting in an
overly stiff response [37].
It is known that the stress comprises of two components: deviatoric (distortion-
al) and volumetric (dilatational). The volumetric component is a function of bulk
modulus and volumetric strain. The bulk modulus is given by:

E
K= (3.11)
3 − 6ν
where E is Young’s modulus and ν is Poisson’s ratio. It is understood that
when ν = 0 , there is no volumetric locking at all. But in the incompressibility
limit, when ν → 0.5 then lim K = ∞ resulting in overly stiff response. In
ν →0.5
this limit, the finite element displacements tend to zero and thus, the volumetric
locking prevails [37]. Hybrid Elements are often used to overcome volumetric
locking [38]. The idea is to include the hydrostatic stress distribution as an addi-
tional unknown variable, which must be computed at the same time as the dis-
placement field. This allows the stiff terms to be removed from the system of fi-
nite element equations. Further details on hybrid elements can be found in existing
literatures [39, 40].
Finite Element Modeling of Chip Formation in Orthogonal Machining 117

It has been also observed that fully integrated elements have higher tendency of
volumetric locking. This is because at each of the integration points the volume
remains nearly constant and when the integration points are more, as in case of
full integration scheme, it results in overconstraining of the kinematically admiss-
ible displacement field. Reduced integration scheme is a possible way to avoid
locking wherein the element stiffness is integrated using fewer integration points
than that are required for full integration scheme [24]. Reduced integration is an
effective measure for resolving locking in elements such as quadratic quadrilateral
and brick effectively but not in elements such as 4 noded quadrilateral or 8 noded
brick elements. The error occurs because the stiffness matrix is nearly singu-
lar which implies that the system of equations includes a weakly constrained de-
formation mode. This phenomenon is known as hourglassing which results in
wildly varying displacement field but correcting stress and strain fields. This can
be cured either by employing selectively reduced integration or by adding an ar-
tificial stiffness to the element that acts to constrain the hourglass mode (Reduced
Integration with Hourglass Control) [37].

3.3.2.6 Solution Methods

Time integration methods are employed for time stepping to solve the transient
dynamic system of equations. There are two main types of time integration me-
thods [29, 31]:
• Implicit
• Explicit
Both the procedures have their own benefits and limitations from solution point of
view. Selection of an appropriate solution method should be done carefully ac-
cording to the nature of the problem. The implicit method is generally suitable for
linear transient problems. In an implicit dynamic analysis, a global stiffness matrix
is assembled and the integration operator matrix must be inverted and a set of non-
linear equilibrium equations must be solved at each time increment. Suitable im-
plicit integrator, which is unconditionally stable, is employed for non-linear for-
mulations and the time increment is adjusted as the solution progresses in order to
obtain a stable, yet time-efficient solution. However, there are situations where
implicit method encounters problems finding a converged solution. In analyses
such as machining which involves complex contact problems, this algorithm is
less efficient due to the simultaneous solving of the equation system in every in-
crement. In such situations, explicit time integration method proves to be robust.
The simulation program seldom aborts due to failure of the numerical algorithm
since the global mass and stiffness matrices need not be formed and inverted.
However, the method is conditionally stable i.e. if the time step exceeds critical
time step, the solution may grow unboundedly and give erroneous results. The
critical time step for a mesh is given by [31]:

Δtcrit ≤ min ( Le ce ) (3.12)

118 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

where Le is the characteristic length of the element and ce is the current wave
speed of an element.

3.3.2.7 Mesh Adaptivity

Generally speaking, mesh adaptivity can be described as the capability of the

mesh to adapt according to the nature of problem in order to maintain a high quali-
ty mesh throughout the analysis. The need of such technique becomes more evi-
dent while simulating problems that involve heavy mesh distortion such as in the
case of cutting process undergoing large deformations. This is well illustrated by
citing an example which considers cutting process as a purely large deformation
problem comprising a rectangular block as workpiece and a portion of cutting tool
(see Fig. 3.5). Figure 3.5 (a) shows the initial mesh configuration of the workpiece
while Fig. 3.5 (b) shows the deformed mesh configuration as tool comes in contact
with the workpiece. Since no mesh adaptivity technique is employed, elements get
highly distorted near the tool tip as highlighted in Fig. 3.6 (b) that leads to termi-
nation of the simulation program.

(a) (b)
Fig. 3.5. (a) Initial mesh configuration and (b) Deformed mesh configuration

Figures 3.6 (a) and 3.6 (b) show two possible mesh adaptivity techniques that
are implemented to control the mesh distortion near the tool tip.

(a) (b)
Fig. 3.6. (a) Mesh refinement and (b) Relocation of nodes near the tool tip
Finite Element Modeling of Chip Formation in Orthogonal Machining 119

In the first case (Fig. 3.6 (a)), it is noted that refinement of mesh takes place at
critical areas i.e. finer mesh in the region closer to the tool tip wherein the size of
elements is adjusted based on selected error indicators and loading history. Here, the
number of elements and connectivity of the nodes are changed. This type of tech-
nique is often known as h–adaptivity [41]. In Fig. 3.6 (b), mesh distortion is con-
trolled by relocation of nodes without altering the number of elements and connec-
tivity of nodes. This approach is often termed as r–adaptivity [41]. Since the number
of elements remains the same, this approach is computationally less expensive
as compared to h–adaptivity. However, both h-refinement and r–refinement are
widely implemented in simulating metal cutting process, the former being used in
Lagrangian framework while the latter being employed in ALE framework.

3.4 Brief History of FEM in Machining

Pioneering works in FEM include works by Klamecki [23], Tay et al. [42], Mura-
ka et al. [43]. Klamecki [23] was the first one to apply FEM in the analysis of
metal cutting process, whereas Tay et al. [42] were the first to calculate the tem-
perature field distribution in orthogonal machining by the application of FEM.
Muraka et al. [43] investigated the effect of process variables such as cutting
speeds, flank wear, coolant, etc., on the temperature distribution using finite ele-
ment method. Morikawa et al. [44] developed a rigid-plastic finite element model
to examine the effect of the tool edge radius to depth of cut ratio on the micro ma-
chining process. Kim and Sin [45] developed a thermo-visco-plastic cutting model
by using finite element method to analyze the mechanics of steady state orthogon-
al cutting process. Liu and Guo [46], on the other hand, proposed a thermoelastic-
viscoplastic FE model to investigate the friction at the tool chip interface. Cerreti
et al. [47] used DEFORM 2D to estimate the chip and tool temperatures for conti-
nuous and segmented chip. Li et al. [48] employed Johnson–Cook’s model as con-
stitutive equation of the workpiece material for qualitative understanding of crater-
wear from the calculated temperature distribution by using commercial FE code
ABAQUS. Arazzola et al. [49] incorporated ALE formulation along with Johnson
Cook model as the material model for the workpiece using ABAQUS/Explicit. In
many of the works, researchers have attempted to make comparative study of the
effect of various material models such as Johnson-Cook model, Zerilli–Armstrong
model, Power law rate-dependent model, Oxley model, Litonski–Batra and many
more on the predicted results [50–52]. Adibi–Sedeh et al. [51] found that John-
son–Cook model predicts chip thickness better than that of other models when
compared with the experimental ones. In few of the works, comparison of some
selected friction models has been presented showing their effect on FE simulations
for orthogonal cutting of various work materials [53–55].
Mabrouki et al. [56] compared the results obtained from ABAQUS/Explicit and
Thirdwave Systems’ AdvantEdge with those obtained experimentally. Coelho et
al. [57] developed an ALE–FEM model to help understand the wear results found
in case of coated PCBN tools for turning hardened AISI 4340. Lorentzon and
Jarvstrat [58] investigated the impact of different wear models and friction models
on the parameters affecting the wear process for cemented carbide tools while
120 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

machining alloy 718, using the commercial software MSC Marc. Davim et al. [59]
made a comparative study between the performances of PCD (polycrystalline di-
amond) and K10 (cemented carbide) tools while machining aluminum alloys using
AdvantEdge software. Attanasio et al. [60] presented 3D numerical predictions of
tool wear based on modified Takeyama and Murata wear model using SFTC De-
form 3D software.

3.5 Formulation of Two-Dimensional FE Model for Machining

This section attempts to make understand the basic methodology and the related
theory for developing the FE model of orthogonal cutting process. Furthermore, a
brief description of various modules contained in ABAQUS program has been
presented so that readers become conversant with ABAQUS. It is agreed that
ABAQUS software is provided with extensive and excellent manuals [61–63], but
our aim is to abridge those details in a very lucid manner that may serve as a quick
guide to a beginner. This would facilitate the readers to have a much clearer un-
derstanding of the finite element concepts being implemented in the finite element
software for simulating machining operation.

3.5.1 Formulation Steps

The present study focuses on developing a 2D FE model of chip formation process
based on orthogonal cutting conditions. Although two–dimensional analysis is a re-
strictive approach from practical point of view, it reduces the computational time
considerably and provides satisfactory results regarding the details of the chip for-
mation. A plane strain condition was assumed because the feed value is generally
very less as compared to the depth of cut. A fully coupled thermal-stress analysis
module of finite element software ABAQUS/Explicit version 6.7–4 has been em-
ployed to perform the study of chip formation process. This makes use of elements
possessing both temperature and displacement degrees of freedom. It is necessary
because metal cutting is considered as a coupled thermo-mechanical process, where-
in, the mechanical and thermal solutions affect each other strongly.

3.5.1.1 Geometric Model, Meshing and Boundary Conditions

The 2D model comprises a rectangular block representing the workpiece and only
a portion of cutting tool which participates in the cutting. The nose radius is neg-
lected for the simplicity of the problem. Moreover, there is hardly any effect of
nose radius once a steady state is reached in cutting. The cutting tool includes the
following geometrical angles: inclination angle χ = 90º , rake angle γ = −6º
and flank angle α = 5º . An intermediate layer of elements known as damage
zone (highlighted region in Fig. 3.7) has been considered in the workpiece block
such that width of the upper surface is equal to the undeformed chip thickness or
in other words feed in case of orthogonal cutting process. As mechanical boundary
conditions, bottom of work piece is fixed in Y direction and left vertical edge of
Finite Element Modeling of Chip Formation in Orthogonal Machining 121

workpiece is fixed in X direction. The former not only constrains the movement
of workpiece in the Y direction but also aids in calculating feed force during ma-
chining while the latter, not only constrains the movement of workpiece in the
X direction but also aids in calculating cutting force during machining. The reac-
tion force components when added at all the constrained nodes of the left vertical
edge of the workpiece in X direction give the cutting force values while at bot-
tom edge of the workpiece in Y direction give the feed force. Tool is given the
cutting velocity in negative X direction and top edge of the tool is constrained in
Y direction.
Similarly, as thermal boundary conditions the tool and the workpiece are in-
itially considered at the room temperature. Heat transfer from the chip surface to
cutting tool is allowed by defining the conductive heat transfer coefficient (h),
given as:

∂T
−k = h (To − T ) (3.13)
∂n
where k is thermal conductivity and To is the ambient temperature.
The geometric model taken into consideration is shown in Fig. 3.7.

Tool

Y
Damage zone
Z X

Workpiece

Fig. 3.7. 2D geometric model

Damage zone is a sacrificial layer of elements that defines the path of separation
between chip surface and the machined surface which is going to take place as the
tool progresses. In actual practice, these two surfaces should be same but such as-
sumption has been taken only for the modeling purpose as a chip separation crite-
rion. The choice of the height of the designated damage zone is purely based on
computational efficiency. Generally, a very small value (say, 10–30 µm) which is
computationally acceptable has been taken as the width of the damage zone.
Four-node plane strain bilinear quadrilateral (CPE4RT) elements with reduced
integration scheme and hourglass control are used for the discretization for both
122 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

the workpiece and the cutting tool. The workpiece is meshed with CPE4RT-type
elements by unstructured grid generation which utilizes advancing front algorithm
in ABAQUS/Explicit.

3.5.1.2 Governing Equations

The governing equations of a body undergoing deformation consists of two sets of

equations, namely, the conservations laws of physics and the constitutive equa-
tions. The conservation laws can be applied to body of any material. But to distin-
guish between different materials undergoing deformation of varied degrees, one
needs the constitutive equations.
Conservation laws: The mass, momentum and energy equations governing the
continuum are given as follows:
ρ + ρ div v = 0 (3.14)

ρ v = f + div σ (3.15)

ρ e = σ : D − div q + r (3.16)

where ρ is the mass density, v material velocity, f body forces, σ Cauchy

stress tensor, e specific internal energy, D strain rate tensor, q heat flux vector
and r is body heat generation rate. The superposed ‘ • ’ denotes material derivative
in a Lagrangian description and ‘:’ denotes contraction of a pair of repeated indic-
es which appear in the same order such as, A:B = Aij Bij .
The basic idea of using FEM is to discretize the above equations and then to
seek a solution to the momentum equation.
Workpiece material constitutive equations: Constitutive equations describe the
thermo-mechanical properties of a material undergoing deformation. Based on the
simplicity or complexity of the material behavior, there could be one constitutive
equation or a set of constitutive equations that relate the deformation in the body
to the stresses. The elastic response can be described by Hook’s law. Whereas, the
constitutive equations of plasticity deal with yield criterion, flow rule and strain
hardening rule. The yield criterion describes the stress state when yielding occurs,
the flow rule defines increment of plastic strain when yielding occurs, and the har-
dening rule describes how the material is strain hardened as the plastic strain in-
creases. For large deformation problems, as in case of machining, plasticity mod-
els based on von Mises yield criterion and Prandtl–Reuss flow rule are generally
used to define the isotropic yielding and hardening [64].
It is known that during cutting process, the workpiece material is generally
subjected to high levels of, strain, strain rate and temperature which significantly
influences flow stress. Thus, accurate and reliable rate-dependent constitutive
models are required that can reflect such phenomenon effectively. Johnson–Cook
constitutive equation is one such model that considers the flow stress behavior of
Finite Element Modeling of Chip Formation in Orthogonal Machining 123

the work materials as multiplicative effects of strain, strain rate and temperature,
given as follows [65]:

⎛ ⎛ T −T ⎞
( )
m
⎞
σ = A + B (ε )
p n
(1 + C lnε ) ⎜⎜1 − ⎜ T − Troom ⎟ ⎟⎟
p∗
(3.17)
⎝ ⎝ m room ⎠ ⎠
εp ⎛ T − Troom ⎞
ε = p for ε 0 p = 1 s-1 and T * = ⎜
∗
⎟ (3.18)
ε0 ⎝ Tm − Troom ⎠
where Troom is the room temperature taken as 25 ºC, Tmelt is the melting temper-
ature of the workpiece, A is the initial yield stress (MPa), B the hardening mod-
ulus, n the work-hardening exponent, C the strain rate dependency coefficient
(MPa), and m the thermal softening coefficient. A, B, C, n and m used in the mod-
el are actually the empirical material constants that can be found from different
mechanical tests. Johnson-Cook model has been found to be one of the most suit-
able one for representing the flow stress behavior of work material undergoing
cutting. Besides, it is also considered numerically robust as most of the variables
are readily acceptable to the computer codes. This has been widely used in model-
ing of machining process by various researchers. [52, 66–68].

3.5.1.3 Chip Separation Criterion

A damage model should be incorporated in the damage zone along with the ma-
terial as a chip separation criterion in order to simulate the movement of the cut-
ting tool into workpiece without any mesh distortion near the tool tip. Specifica-
tion of damage model includes a material response (undamaged), damage
initiation criterion, damage evolution and choice of element deletion.
Damage initiation criterion is referred to as the material state at the onset of
damage. In the present case, Johnson–Cook damage initiation criterion has been
employed. This model makes use of the damage parameter ωD defined as the
sum of the ratio of the increments in the equivalent plastic strain Δε to the frac-
p

ture strain ε , given as follows [61].

f

Δε p
ωD = ∑ (3.19)
εf
The fracture strain εf is of the form as follows [69]:

ε f = ⎡⎣ D1 + D2exp ( D3σ ∗ ) ⎤⎦ × ⎡⎣1 + D4 lnε p∗ ⎤⎦ × ⎡⎣1 + D5T * ⎤⎦ (3.20)

−P
σ∗ = (3.21)
σ
124 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

where P is the hydrostatic pressure and D1 to D5 are failure parameters deter-

mined experimentally. The damage constants can be determined by performing
several experiments such as tensile test, torsion test and Hopkinson bar test. On
the basis of results obtained, a graph is generally plotted between equivalent plas-
tic strain at fracture and pressure stress ratio such that the constants related to
pressure i.e. D1, D2, and D3 can be found out at T * = 0 and ε p = 0 . Similarly,
*
D4 and D5 can be obtained by plotting another graph between T and ratio of frac-
ture strain at different strain rates. The damage initiation criterion is met when
ωD (Eq. 3.20) reaches one [61].
Once the element satisfies the damage initiation criterion, it is assumed that
progressive degradation of the material stiffness occurs, leading to material failure
based on the damage evolution. At any given time during the analysis, the stress
tensor in the material is given by:

σ = (1 − D ) σ (3.22)

where σ is the effective (undamaged) stress tensor computed in the current in-
crement. When overall damage variable D reaches a value 1, it indicates that the
material has lost its load carrying capacity completely. At this point, failure occurs
and the concerned elements are removed from the computation.
p
The effective plastic displacement ( u ), after the damage initiation criterion is
met can be defined with the evolution law as follows:

u p = Leε p
(3.23)

where Le is the characteristic length of the element and ε p equivalent plastic

strain.
When a linear evolution of the damage variable with plastic displacement is as-
sumed, the variable increases as per the following [61]:

Leε p u p
D= = p
uf p uf
(3.24)
2G f
uf =p

σ y0

where u f p is the plastic displacement at failure, G f is the fracture energy per

unit area and σ y 0 is the value of the yield stress at the time when the failure crite-
rion is reached.
Finite Element Modeling of Chip Formation in Orthogonal Machining 125

The model ensures that the energy dissipated during the damage evolution
process is equal to G f only if the effective response of the material is perfectly
plastic (constant yield stress) beyond the onset of damage. In this study, G f is
provided as an input parameter which is a function of fracture toughness KC,
Young's modulus E and Poisson's ratio ν as given in the equation for the plane
strain condition [70]:

⎛ 1 −ν 2 ⎞ 2
Gf = ⎜ ⎟ KC . (3.25)
⎝ E ⎠
The ELEMENT DELETION = YES module along with the Johnson Cook damage
model of the software enables to delete the elements that fail. This produces the
chip separation and allows the cutting tool to penetrate further into the workpiece
through a predefined path (damage zone).

3.5.1.4 Chip-Tool Interface Model and Heat Generation

A contact is defined between the rake surface and nodes of the workpiece materi-
al. Coulomb’s law has been assumed in the present study to model the frictional
conditions as the chip flows over the rake surface.
During the machining process, heat is generated in the primary shear deforma-
tion zone due to severe plastic deformation and in the secondary deformation zone
due to both plastic deformation and the friction in the tool–chip interface. The
steady state two-dimensional form of the energy equation governing the orthogon-
al machining process is given as:

⎛ ∂ 2T ∂ 2T ⎞ ⎛ ∂T ∂T ⎞
k⎜ 2 + 2 ⎟ − ρ Cp ⎜ u x + vy ⎟+q = 0 (3.26)
⎝ ∂x ∂y ⎠ ⎝ ∂x ∂y ⎠

q = qp + qf (3.27)

qp = η pσε p (3.28)

qf = ηf Jτγ (3.29)

where q p is the heat generation rate due to plastic deformation, ηp the fraction of
the inelastic heat, qf is the volumetric heat flux due to frictional work, γ the slip
rate, ηf the frictional work conversion factor considered as 1.0, J the fraction of
the thermal energy conducted into the chip, and τ is the frictional shear stress.
The value of J may vary within a range, say, 0.35 to 1 for carbide cutting tool
[71]. In the present work, 0.5 (default value in ABAQUS) has been taken for all
126 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

the cases. The fraction of the heat generated due to plastic deformation remaining
in the chip, ηp , is taken to be 0.9 [56].

3.5.1.5 Solution Method

An ALE approach is incorporated to conduct the FEM simulation. This avoids se-
vere element distortion and entanglement in the cutting zone without the use of
remeshing criterion.
ALE formulation: In the ALE approach, the grid points are not constrained to re-
main fixed in space (as in Eulerian description) or to move with the material
points (as in Lagrangian description) and hence have their own motion governing
equations. ALE description is given as follows:
∼ •
( ) = ( )+ c ∇ ( ) (3.30)

c = v − vˆ (3.31)

where superposed ‘ ∼ ’ is defined in the ALE description, c is the convective ve-

locity, v is the material velocity, v̂ is the grid velocity and ∇ is the gradient op-
erator. The Eqs. (3.14) to (3.16) can now be rewritten in the ALE framework, as
given below:
ρ + c ∇ρ + ρ div v = 0 (3.32)

ρ v + ρ c ∇v = f + div σ (3.33)

ρ e + ρ c ∇e = σ : D − div q + r (3.34)

Explicit dynamic analysis: The explicit dynamic ALE formulation is primarily

used to solve highly non-linear problems involving large deformations and chang-
ing contact, as observed in the case of machining. The explicit dynamic procedure
in ABAQUS/Explicit is based upon the implementation of an explicit integration
scheme for time dicretization of Eqs. (3.32) to (3.34). This analysis calculates the
state of a system at a later time from the state of system at the current time.
The explicit formulation advances the solution in time with the central differ-
ence method [31]:

(
vi = M -1 f ext − f int
(i ) (i)

) (3.35)

Δt (i +1) − Δt i i
v (i +1/ 2) = v ( i −1/ 2) + v (3.36)
2
Δt (i +1) − Δt i (i +1/ 2)
x (i +1) = xi + v (3.37)
2
Finite Element Modeling of Chip Formation in Orthogonal Machining 127

where x is the displacement, v is the velocity and v is the acceleration, M is the

ext int
lumped mass matrix, f is the external force vector and f is the internal
force. The subscript (i) refers to the increment number (i–1/2) and (i+1/2) and
refers to mid-increment values. In this algorithm, the accelerations are first inte-
grated to obtain the velocities, integration of which is broken into two half-steps
so that the velocities are available at an integer step in the computation of the
energy balance. The displacements are then computed in each time step by inte-
grating the velocities.

3.5.2 ABAQUS Platform

The type of software package chosen for the FE analysis of metal cutting
process is equally important in determining the quality and scope of analysis
that can be performed. There are currently large number of commercial software
packages available for solving a wide range of engineering problems that might
be static, dynamic, linear or non-linear. Some of the dominant general purpose
FE software packages include ABAQUS, ANSYS, MSC/NASTRAN, SRDC-
IDEAS, etc. It is obvious that different packages would possess different
capabilities. This makes it critical to select the suitable software package with
appropriate features required for performing a given analysis successfully. The
present study selects ABAQUS as a platform to explore the capabilities of finite
element method in analyzing various aspects of metal cutting process. ABAQUS
is known to be powerful general purpose FE software that can solve problems
ranging from relatively simple linear analyses to the highly complex non-linear
simulations. This software does not have any separate module for machining as
in the case of Deform or AdvantEdge. As a result, the user has to explicitly de-
fine the tool and the workpiece, process parameters, boundary conditions, mesh
geometry and simulation controls. This may certainly require more skill, effort
and time to set up simulations as no preset controls and assumptions are availa-
ble. But this is the feature that not only ensures very high level of details from
modeling point of view but also a thorough analysis by allowing a precise
control on boundary conditions, mesh attribute, element type, solver type and
so on.
A complete ABAQUS program can be subdivided into three distinct modules,
namely, ABAQUS/CAE, ABAQUS/Standard or ABAQUS/Explicit and ABAQUS/
Viewer as shown in Fig. 3.8. These modules are linked with each other by input and
output files. ABAQUS/Standard and ABAQUS/Explicit are the two main types of
solvers that are available for performing analysis, ABAQUS/Explicit being mainly
used for explicit dynamic analysis. It is said that the strength of ABAQUS program
greatly lies in the capabilities of these two solvers.
128 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

ABAQUS/CAE
(Preprocessing)

Input file
(*.inp)

ABAQUS/Standard or ABAQUS/Explicit
(Analysis)

Output file
(*.odb)

ABAQUS/CAE or ABAQUS/Viewer
(Post-processing)

Fig. 3.8. Work flow of ABAQUS program

The model of the physical problem is created in the pre-processing stage, de-
tails of which such as discretized geometry, material data, boundary conditions,
element type, analysis type and output request are contained in the input file.
ABAQUS/CAE is divided into functional units called modules with the help of
which the FE model can be created, input file can be generated and results can be
extracted from the output file. Each module has been designed to serve a specific
portion of the modeling task. The subsequent subsections would discuss about
various modules of ABAQUS/CAE in brief.
Part module: Individual parts are created in the part module either by sketching
their geometry directly in Abaqus/CAE or by importing their geometry from other
geometric modeling programs. Depending upon the analysis the parts can be 2D
or 3D deformable, discrete rigid, analytical rigid or Eulerian parts. In the present
study, both the cutting tool and the workpiece are considered as 2D deformable
bodies. The part tools contained in this module allow editing and manipulating the
existing parts defined in the current model.
Property module: Property module allows assigning sections to a part instance or
region of a part instance to which various material properties are defined. A ma-
terial definition specifies all the property data relevant to a particular analysis. For
a coupled temperature displacement analysis, both the mechanical strength proper-
ties (elastic moduli, yield stress, etc.) and heat transfer properties (conductivity,
specific heat) must be given as inputs. Various plasticity models and damage
models are also contained in the property module. As an input the material model
Finite Element Modeling of Chip Formation in Orthogonal Machining 129

constants of the selected plastic model and damage model such as Johnson–Cook
material model and Johnson–Cook damage model, respectively, are defined in a
tabular form.
Assembly module: The individual parts that are created in part module exist in
their own coordinate system. It is in the assembly module that these parts are as-
sembled by relative positioning with respect to each other in a global coordinate
system.
Step module: The step module is used to create and configure analysis steps, as in
the present case is coupled temperature displacement explicit dynamic analysis.
The associated output requests can also be created. The sequence of steps provides
a convenient way to capture changes that may occur in the model during the
course of the analysis such as changes in the loading and boundary conditions of
the model or changes in the interaction, etc. In addition, steps allow you to change
the analysis procedure, the data output and various controls. An output request
contains information regarding which variables will be output during an analysis
step, from which region of the model they will be output, and at what rate they
will be output. The general solution controls and solver controls can also be cus-
tomized. Furthermore, adaptive mesh regions and the controls for adaptive mesh-
ing in selected regions can be specified in this module. It is noted that implemen-
tation of suitable mesh adaptivity (remeshing in Lagrangian framework or
repositioning of nodes in ALE framework) depends upon the type of analysis taken
into consideration i.e. implicit (ABAQUS/Standard) or explicit (ABAQUS/Explicit).
Adaptive remeshing is available in ABAQUS/Standard while ALE adaptive
meshing is available in ABAQUS/Explicit.
Interaction module: Interaction module allows to specify mechanical and thermal
interactions between regions of a model or between a region of a model and its
surroundings. Surface-to-surface contact interaction has been used to describe
contact between tool and workpiece in the present study. The interaction between
contacting bodies is defined by assigning a contact property model to a contact in-
teraction which defines tangential behavior (friction) and normal behavior. Tan-
gential behavior includes a friction model that defines the force resisting the rela-
tive tangential motion of the surfaces. While normal behavior includes a definition
for the contact pressure–overclosure relationship that governs the motion of the
surfaces. In addition, a contact property can contain information about thermal
conductance, thermal radiation and heat generation due to friction. ABAQUS/ Ex-
plicit uses two different methods to enforce contact constraints, namely, kinematic
contact algorithm and penalty contact algorithm. The former uses a kinematic pre-
dictor/corrector contact algorithm to strictly enforce contact constraints such that
no penetrations are allowed, while the latter has a weaker enforcement of contact
constraints. In this study, a kinematic contact algorithm has been used to enforce
contact constraints in master–slave contact pair, where rake surface of the cutting
tool is defined as the master surface and chip as the slave (node-based) surface.
Both the frictional conditions and the friction-generated heat are included in the
kinematic contact algorithm through TANGENTIAL BEHAVIOUR and GAP
HEAT GENERATION modules of the software.
130 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

Load module: The Load module is used to specify loads, boundary conditions and
predefined fields.
Mesh module: The Mesh module allows generating meshes on parts and assem-
blies created within ABAQUS/CAE as well as allows selecting the correct ele-
ment depending on the type of analysis performed (as in the present case is
CPE4RT) for discretization. Variety of mesh controls are available that help in se-
lecting the element shape (tri, quad or quad dominated), meshing technique (struc-
tured or unstructured) or meshing algorithm (medial axis or advancing front). The
structured meshing technique generates structured meshes using simple predefined
mesh topologies and is more efficient for meshing regular shapes. Free meshing,
however, allows more flexibility than structured meshing. Two commonly used
free mesh algorithms while meshing with quadrilateral elements are medial axis
and advancing front. The advancing front is generally preferable because it gene-
rates elements of more uniform size (area–wise) with more consistent aspect ratio.
Since in ABAQUS/Explicit small elements control the size of the time step,
avoidance of large differences in element size reduces solution stiffness, i.e.,
makes the numerical procedure more efficient.
Job: The Job module allows to create a job, to submit it to ABAQUS/Standard or
ABAQUS/Explicit for analysis and to monitor its progress.
Visualization: The Visualization module provides graphical display of finite ele-
ment models and results. It extracts the model and result information from the
output database.

3.6 Case Studies in ABAQUS Platform

This section demonstrates the efficiency of finite element models to replicate the
actual phenomena occurring during the cutting process under varied conditions as
well as to understand basic mechanism of chip formation process in terms of vari-
ous numerical results. Two different work materials are considered, namely AISI
1050 and Ti6Al4V, the former producing continuous chips and the latter produc-
ing segmented chips. In the first case study, not only the simulation results for ma-
chining AISI 1050 are presented but the predicted results are confirmed with the
experimental ones. In the second case study, Ti6Al4V is considered for the study
and the predicted results are compared with those of AISI 1050 under similar con-
ditions, thus showing the effect of work material, being one of the machining in-
puts, on various output variables. Apart from machining inputs, being a numerical
model, it is obvious that various FE inputs such as mesh size, material models,
friction models and so on also have a typical impact on the output results. Hence,
effect of FE inputs, namely, mesh size and Johnson-Cook material model con-
stants have also been studied while simulating segmented chip formation during
machining Ti6Al4V.
Finite Element Modeling of Chip Formation in Orthogonal Machining 131

3.6.1 Case Study I: FE Simulation of Continuous Chip

Formation While Machining AISI 1050 and Its
Experimental Validation
AISI 1050 is a standard grade carbon steel. This particular grade of steel is both
relatively easy to machine as well simulate. Therefore, AISI 1050 was chosen as a
reference material in the present study. The basic aim of the present section is to
develop the FE model for simulating the continuous chip formation process during
machining of AISI 1050 as well as to explain the basic mechanism governing the
continuous chip formation process in terms of numerical results viz. distributions
stress, strain and temperature. The simulated results are then validated with the
experimental ones at varied cutting conditions. Table 3.1 and Table 3.2 show the
thermo-mechanical properties of the workpiece material as well as the values of
Johnson-Cook material and damage constants used as input, respectively. The
tool-chip interaction is based on the commonly used Coulomb’s law with the
mean coefficient of friction 0.2.

Table 3.1. Thermo-mechanical properties of tungsten carbide tool and AISI 1050 [72, 73]

Properties Carbide tool AISI 1050

Density, ρ (Kg/m3) 11900 8030
Elastic modulus, E (GPa) 534 210
Poission’s ratio,μ 0.22 0.3
Specific heat, Cp (J/KgoC) 400 472
Thermal conductivity, λ (W/moC) 50 51.9
Expansion, (μm/moC) - 13.7
Tmelt (oC) - 1460
Troom (oC) - 25
Fracture toughness,Kc (MPa m½) - 60
Conductive heat transfer coefficient h, (kW m–2 K–1) - 10

Table 3.2. Johnson-Cook material and damage constants for AISI 1050 [72]

Material Constants

A (MPa) B (MPa) n C m

880 500 0.234 0.013 1.00

Damage Constants

D1 D2 D3 D4 D5

0.06 3.31 –1.96 0.018 0.58

132 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

3.6.1.1 Simulation Results

Figure 3.9 shows the distributions of stress, strain and temperature while machin-
ing of AISI 1050 using tungsten carbide tool for a cutting speed of 120 m/min and
feed of 0.2 mm/rev.

(a) (b) (c)

Fig. 3.9. Distribution of (a) stress (b) strain and (c) temperature while machining of AISI
1050 at Vc =120 m/min and f =0.2 mm

As the tool touches the workpiece, compression occurs within the workpiece.
With the further advancement of the tool into the workpiece, stresses start develop-
ing near the tool tip and attaining high localized values in a confined region called
primary shear deformation zone or shear plane as shown in Fig. 3.9 (a). The stresses
in these regions are as high as 1.3 GPa. Consequently, such high values of stresses
cause high strains to occur in the shear zone. This allows the workpiece material to
deform plastically and shear continuously in the form of chip that flows over tool
rake face. The type of chip, thus, formed depends largely on the distribution of strain
and temperature within the chip surface. As the cutting continues the effective
strains (especially, around the tool tip) increase and spread over a wider area of the
chip surface with a maximum value not exceeding 1.7 in the shear plane. Conse-
quently, this phenomenon tends to make the temperature distribution uniform in the
chip region, thereby resulting in a steady state continuous chip having unvarying
chip thickness. It is interesting to note that the maximum equivalent plastic strain
and temperature are found along the tool–chip interface.
This can be attributed to the fact that the chip which is entering into the second-
ary shear deformation zone already possesses accumulated plastic strain and heat.
The instant it begins to flow over the rake surface, further plastic straining and lo-
cal heating occur because of severe contact and friction in the contact zone [74],
thus, attaining higher temperature values at the tool-chip interface, specifically in
the sliding region.

3.6.1.2 Model Verification

In order to validate the developed model, both the cutting speed and feed or uncut
chip thickness values are varied and their effect on predicted cutting forces is
Finite Element Modeling of Chip Formation in Orthogonal Machining 133

studied and compared with the experimental ones. The cutting speed is varied in
the range of 72–164 m/min for feed values of 0.1 and 0.2 mm.
It is known that the cutting force and thrust force increase with increasing feed
rate almost linearly [75] where as decrease with the increasing cutting velocity.
The reason for this can be explained from the expressions of the cutting force and
thrust forces which are given as follows:

Cutting force, Fc = tsτ s (ζ − tan γ + 1) (3.38)

Thrust force, Ft = tsτ s (ζ − tan γ − 1) (3.39)

where t is the depth of cut (mm), s is the feed rate (mm/rev), τ s is the dynamic
shear strength of the workpiece, γ is the rake angle and ζ is the chip reduction
coefficient, i.e., the ratio of deformed chip thickness to undeformed chip thick-
ness. From the equations, correlation between feed rate and forces is straightfor-
ward. As far as the variation of cutting velocity is concerned, as the former in-
creases, temperature of the shear zone increases. This causes softening of the work
piece, which means the value of τ s decreases and thereby reduces the value of
cutting and thrust force.
Predicted cutting forces also show similar kind of trend with variation of speed
and feed as shown in Fig. 3.10. Cutting forces measured during experimental tests
under conditions similar to that of simulations have also been presented in Fig. 3.10
for comparison. It is noted that at lower feed, predicted results closely match with
experimental ones. At higher feed, a flatter curve is observed for predicted values of
forces implying that there is, no doubt, a decrease in cutting force values with the in-
creasing speed but not as pronounced as in case of the experimental one. However,
the maximum deviation between predicted and measured values of cutting forces
remains within an acceptable range of 15%.

1400
Fc(exp)
Fc(sim)
1200

Feed=0.2 mm
1000
Force (N)

800

600 Feed=0.1 mm

400

60 80 100 120 140 160 180

Cutting velocity (m/min)

Fig. 3.10. Predicted and measured forces while machining AISI 1050
134 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

3.6.2 Case Study II: FE Simulation of Segmented Chip

Formation While Machining Ti6Al4V
The present study deals with the simulation of chip formation while machining
Ti6Al4V. It is a known fact that machining of Ti6Al4V yields segmented chips.
Formation of segmented chips makes not only the machining of Ti6Al4V chal-
lenging but also simulating the chip formation process of the same very difficult.
In this study, simulation of segmented chip formation is performed based on adia-
batic shearing phenomenon. Thermal softening becomes highly pronounced when
the rate of heat generation exceeds the rate of heat dissipation which comes into
picture when thermal conductivity of the workpiece is very low and cutting speed
is high. During cutting process, high stresses and strains along the primary shear
zone accumulate and plastic deformation occurs, thereby increasing cutting tem-
perature. As the tool proceeds further, temperature increases in a highly localized
area in the shear zone in the form of bands called adiabatic shear bands that extend
from tool tip toward the chip free surface. It has been found that temperature at the
back of the chip reaches such a high value that thermal softening occurs in a high-
ly localized region on the free side of the chip surface. This induces wave like ir-
regularities on the back of the chip.
Firstly, an attempt is made to simulate the desired chip by fine tuning two of
the FE inputs that seem to be critical as far as reliability of the model is concerned.
Then a comparative study is done between the predicted results of AISI 1050 and
Ti6Al4V to explain the significance of simulating the right kind of chip formation
and its correlation with other output variables.

3.6.2.1 Role of FE Inputs

Meshing: Selection of a suitable mesh size is a critical factor from both accuracy
and computational time points of view [76]. As discussed earlier, finer mesh leads
to greater accuracy but at the cost of higher computational time. It is important to
mesh the model in such a way that the model gives results closer to the experi-
mental ones on one hand and consumes fairly less time on the other hand. Moreo-
ver, it is illogical to start with a very fine mesh, instead a mesh refinement study
should be performed wherein the degree of meshing is gradually changed from
coarser mesh to a finer mesh and the corresponding results are compared among
each other. There exists a limit beyond which if the mesh is refined further, the
CPU time would, no doubt, increase but there would not be any significant
changes in the numerical results. It is the duty of the analyst to consider an opti-
mum meshing that comes as a fair compromise between accuracy and computa-
tional time. This shows the need of mesh refinement study to prove the reliability
of the developed model.
Researchers have pointed that with the decrease in element size, the tempera-
ture at the integration point increases. Since this study primarily deals with the
mechanism of adiabatic shearing during the formation of segmented chip, it is
Finite Element Modeling of Chip Formation in Orthogonal Machining 135

logical to carry out a mesh refinement study. The meshing is mainly varied in the
chip region by considering 5, 10 and 15 elements (on an average) in the chip
thickness direction as shown in Fig. 3.11. The average element size for 5, 10 and
15 elements are: 50x50 µm, 20x20 µm and 12x12 µm, respectively. Since advanc-
ing front algorithm (free mesh algorithm) is employed for meshing the workpiece,
slight skewness has been observed at certain mesh regions of the workpiece with
finer meshing at the right edge of the workpiece.

(a) (b) (c)

Fig. 3.11. Mesh configurations for (a) 5 (b) 10 and (c) 15 elements

Figure 3.12 shows the predicted chip morphology and the distribution of tem-
perature within the chip surface for all the three cases for a cutting speed of 210
m/min and uncut chip thickness 0.2 mm.

(a) (b) (c)

Fig. 3.12. Temperature distributions for (a) 5 (b) 10 and (c) 15 elements

It is interesting to note that there is a gross difference in the predicted chip

morphologies for the first two cases. The difference can be attributed to the varia-
tion in stress, strain and temperature distributions within the chip region. Though
the maximum value of stress in both the cases is nearly same but the distribution
136 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

pattern varies. As the meshing gets finer, the stresses become more concentrated
along the shear plane, as a result of which the strain and temperature not only
get localized in a very narrow zone but also attain very high values. Since an in-
crease in the maximum temperature is found with finer meshing, the tendency to
invoke chip segmentation due to thermal softening by adiabatic shearing be-
comes higher, thus producing segmented chips in the second case. However,
with further refinement of mesh as in the case of 15 elements, not much varia-
tion in chip morphology is observed when compared with that of 10 elements.
The adiabatic shear band, no doubt, appears relatively more distinct (as very thin
strip) in the former case, but the average temperature values in the shear band
( Tadia ) increases not even by 1% (see Table 3.3). This implies that further mesh
refinement may not be required as far as consistency of the results is concerned.
Moreover, there is a limit to the reduction of element size from the software
point of view. Since the time step in ABAQUS/Explicit is controlled by the size
of the smallest element, reducing the element size beyond this may not allow the
simulation to run properly.

Table 3.3. Mesh refinement study

Mesh type Stress (GPa) Strain Tadia (°C)

5 elements 1.48 1.32 570

10 elements 1.50 2.0 956
15 elements 1.54 2.5 963

It is observed that finer meshing increases CPU time considerably. Therefore,

fixed mass scaling is used to expedite the analysis in the last two cases. Scaling
the material density of the small elements throughout the analysis proves to be
very useful as it increases the stable time increment allowing the analysis to take
less time, while retaining the accuracy to a great extent [77]. Table 3.4 presents a
comparative study of the actual CPU time with respect to simulation time for 5, 10
and 15 elements.

Table 3.4. CPU time for 5, 10 and 15 elements

Simulation time CPU time (h:min:s)

(ms) 5 elements 10 elements 15 elements
0.4 0:09:53 1:05:20 1:50:02
0.6 0:21:37 1:44:55 ---
1 0:35:07 2:55:52 ---
Finite Element Modeling of Chip Formation in Orthogonal Machining 137

From Table 3.4, it can be inferred that the model considering 10 elements along
the uncut chip thickness is the best compromise between accuracy and computa-
tion time. In case of 15 elements, though adiabatic shearing is very prominent, it
was very difficult to run the simulation beyond 0.4 ms.
Johnson–Cook material model constants: Several classical plasticity models have
been widely employed that represent, with varying degrees of accuracy, the material
flow stresses as a function of strain, strain rate and temperature. These models
include the Litonski-Batra model [78, 79], Usui model [80], Maekawa model [81],
the Johnson–Cook model [65], Zerilli–Armstrong model [82], Oxley model [83],
Mechanical Threshold Stress (MTS) model [84] etc. It is very important to carefully
select the appropriate material model that satisfactorily predicts the desired chip
morphology and other output variables. Johnson-Cook model, being most widely
used, is employed to describe the flow stress property of workpiece material
Ti6Al4V in this study (see Eq. 3.17). This material model defines the flow stress as
a function of strain, strain-rate and temperature such that, it not only considers the
strain rates over a large range but also temperature changes due to thermal softening
by large plastic deformation. The work material constants contained in this constitu-
tive equation have been found by various researchers by applications of several me-
thods, thus producing different values of data sets for a specific material. As a result,
selection of suitable data sets along with appropriate material model becomes equal-
ly important [85]. The present study selects two sets of Johnson-Cook material con-
stants from the available literature namely, M1 and M2 as listed in Table 3.4 [86,
87]. Lee and Lin [87] obtained the Johnson–Cook material constants from a high
strain rate mechanical testing using Split Hopkinson Pressure Bar (SHPB) method
under a constant strain rate of 2000 s-1 within the temperature range of 700–1000 ºC
and maximum true plastic strain of 0.3. While Meyer and Kleponis [86] obtained the
material constants by considering strain rate levels of 0.0001, 0.1 and 2150 s-1and a
maximum plastic strain of 0.57.

Table 3.5. Johnson-Cook material parameters for Ti6Al4V [86, 87]

Johnson-Cook constants A (MPa) B (MPa) C n m

M1 model 862.5 331.2 0.012 0.34 0.8
M2 model 782.7 498.4 0.028 0.28 1.0

Besides the material model constants, all other parameters are kept constant
so that the results can be compared on the same conditions. Meshing in both the
cases is similar to the one that was considered as optimum in the previous case
study.
Figures 3.13 and 3.14 show the distribution of stress, strain and temperature for
the models M1 and M2 at a cutting speed of 210 m/min and uncut chip thickness
of 0.2 mm.
138 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

(a) (b) (c)

Fig. 3.13. Distribution of (a) Stress (Pa) (b) Strain and (c) Temperature (ºC) for M1 model

(a) (b) (c)

Fig. 3.14. Distribution of (a) Stress (Pa) (b) Strain and (c) Temperature (ºC) for M2 model

It is observed that model M1 is unable to simulate the formation of segmented

chips. The distribution of stress, strain and temperature when compared with that
of model M2 explains much about the variation in chip morphology. This suggests
that material model constants used in the constitutive equations also affect the
predicted results. Thermal softening coefficient m and work hardening exponent n
are often considered as the critical parameters that influence the segmentation of
chip [88]. Baker [88] suggested increase in m leads to increase in chip segmenta-
tion while increase in n causes lesser tendency to undergo chip segmentation. In
the present study, similar kind of observation has been found. Greater value of m
and lower value of n can be attributed for the formation of well-defined segmented
chips in case of model M2.

3.6.2.2 Significance of FE Outputs

Various difficult to measure output variables such as stresses, strains, temperatures

and many more can be determined easily from the FE simulations of chip forma-
tion process. These predicted outputs greatly help in understanding the basic me-
chanism of chip formation under different cutting conditions. Chip morphology,
considered as an important index in study of machining, is dependent on various
input variables like workpiece material, cutting conditions, tool material and so
on. The type of chip produced affects the cutting force, the residual stresses and
the cutting temperatures to a great extent. This section, thus, demonstrates the
Finite Element Modeling of Chip Formation in Orthogonal Machining 139

same by a comparative study of predicted forces and cutting tool temperature

while machining AISI 1050 and Ti6Al4V.

Cutting forces: Figure 3.15 shows the variation of predicted cutting force with
time while machining AISI 1050 and Ti6Al4V.

500
Cutting force (N)

400

300

200
Ti6Al4V
AISI 1050
100

0.0000 0.0002 0.0004 0.0006 0.0008 0.0010 0.0012

Time (s)

Fig. 3.15. Time signature of cutting forces for Ti6Al4V and AISI 1050

The force signatures obtained for both the materials vary remarkably not only
in magnitude but also in nature from each other. As expected, the cutting force
profile closely resembles the predicted chip morphology i.e. continuous for AISI
1050 and wavy profile for Ti6Al4V. While machining Ti6Al4V, chip formation
process begins with the bulging of the workpiece material in front of the tool as a
result of which the forces increase gradually but drop suddenly when the shear
band begins to form due to thermal softening. Obviously, magnitude of the wavi-
ness (difference between the peak and lower values) affects the surface finish
quality and thermal load that the cutting tool undergoes during cutting process.

Cutting temperature: In cutting operation, cutting tool is subjected to highly adverse

conditions such as high stresses and high temperatures due to which tool wears out.
As the tool wear occurs, tool geometry changes affecting residual stresses, surface
finish and dimensional accuracy of the machined surface. Hence, cutting tempera-
ture is a critical parameter which is again a function of various machining inputs
such as cutting conditions, workpiece material, tool material, tool geometry, etc.
Tool condition monitoring and optimization of process parameters are of paramount
importance that can lead to higher productivity and accuracy.
Figures 3.16 (a) and (b) present the temperature distributions over the cutting tool
that has developed during machining of AISI 1050 and Ti6Al4V, respectively.
140 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

(a) (b)
Fig. 3.16. Temperature contours on cutting tool while machining (a) AISI 1050 and
(b) Ti6Al4V

While machining Ti6Al4V the maximum temperature attained by carbide tool

is as high as 1000ºC as compared to that of AISI 1050 with maximum value not
even exceeding 400ºC. It can be seen that not only the maximum value of temper-
ature attained by cutting tool is different but also the distribution pattern over the
cutting tool varies widely depending upon the type of worpiece material ma-
chined. This gives a basic idea of how the tool wear progress takes place. Fur-
thermore, the local variables such as temperature, stress, strain, etc., at different
nodes that are predicted prove helpful in developing a tool wear prediction method
by integrating FEM simulation of the cutting process with a suitable tool wear rate
model.
In general, it can be concluded that FEM can prove to be a very effective tool in
analyzing various aspects of metal cutting process, if used with caution. The
stresses, strains, temperatures and many other outputs that are obtained from FE
simulations greatly help in understanding the basic mechanism of chip formation
under different cutting conditions. This consequently helps in the cutting tool de-
sign, optimization of cutting parameters for higher tool life and good surface
finish, prediction of tool wear growth, etc., thus leading to higher productivity.

References
[1] Astakhov, V.P.: Tribology of Metal Cutting. Elsevier (2006) ISBN: 978-0-444-52881-0
[2] Shaw, M.C.: Metal Cutting principles, 2nd edn. Oxford University Press, Oxford
(2005)
[3] Hahn, R.S.: On the temperature developed at the shear plane in the metal cutting
process. In: Proceedings of First U.S. National Congress Appl. Mech. ASME 661
(1951)
[4] Chao, B.T., Trigger, K.J.: An analytical evaluation of metal cutting temperature.
Trans. ASME 73, 57–68 (1951)
[5] Leone, W.C.: Distribution of shear-zone heat in metal cutting. Trans. ASME 76, 121–
125 (1954)
Finite Element Modeling of Chip Formation in Orthogonal Machining 141

[6] Loewen, E.G., Shaw, M.C.: On the analysis of cutting tool temperatures. Transactions
of the ASME 71, 217–231 (1954)
[7] Weiner, J.H.: Shear plane temperature distribution in orthogonal machining. Trans.
ASME 77, 1331–1341 (1955)
[8] Rapier, A.C.: A theoretical investigation of the temperature distribution in the metal
cutting process. Br. J. Appl. Phys. 5, 400–405 (1954)
[9] Bagci, E.: 3-D numerical analysis of orthogonal cutting process via mesh-free me-
thod. Int. J. the Physical Sciences 6(6), 1267–1282 (2011)
[10] ASM Handbook, Volume 16-Machining, ASM International Handbook Committee,
ASM International, Electronic (1989) ISBN: 978-1-61503-145-0
[11] Juneja, B.L., Sekhon, G.S., Seth, N.: Fundamentals of metal cutting and machine
tools, 2nd edn. New Age International Publishers, New Delhi (2003)
[12] Merchant, M.E.: Mechanics of the metal cutting process. J. Appl. Phys. 16, 318–324
(1945)
[13] Lee, E.H., Shaffer, B.W.: The theory of plasticity applied to a problem of machining.
Trans. ASME, J. Appl. Mech. 18, 405–413 (1951)
[14] Oxley, P.L.B.: Shear angle solutions in orthogonal machining. Int. J. Mach. Tool.
Des. Res. 2, 219–229 (1962)
[15] Bhattacharya, A.: Metal cutting theory and practice. Central book publishers, Kolkata
(1984)
[16] Lacale, L.N., Guttierrez, A., Llorente, J.I., Sanchez, J.A., Aboniga, J.: Using high
pressure coolant in the drilling and turning of low machinability alloys. Int. J. of Adv.
Tech. 16, 85–91 (2000)
[17] Tobias, S.A.: Machine tool Vibration. Blackie and Sons Ltd, Scotland (1965)
[18] Ghosh, A., Mallik, A.K.: Manufacturing science (1985) ISBN: 81-85095-85-X
[19] Jacobson, S., Wallen, P.: A new classification system for dead zones in metal cutting.
Int. J. Mach. Tool. Manufact. 28, 529–538 (1988)
[20] Trent, E.M., Wright, P.K.: Metal cutting, 4th edn. Butterworth-Heinemann (2000)
[21] Komanduri, R., Hou, Z.B.: A review of the experimental techniques for the measure-
ment of heat and temperatures generated in some manufacturing processes and tribol-
ogy. Tribol. Int. 34, 653–682 (2001)
[22] Groover, M.P.: Fundamentals of modern manufacturing: materials processes, and sys-
tems, 2nd edn. Wiley, India (2002)
[23] Klamecki, B.E.: Incipient chip formation in metal cutting- A three dimensional finite
element analysis. Ph.D. Thesis. University of Illinois, Urbana (1973)
[24] Hughes, J.R.T.: The Finite Element Method. Prentice-Hall International, Inc. (1987)
[25] Reddy, J.N.: An introduction to the finite element method, 2nd edn. McGraw-Hill
Inc. (1993)
[26] Bathe, K.J.: Finite Element Procedures. Prentice Hall, Englewood Cliffs (1996)
[27] Rao, S.S.: The Finite Element in Engineering, 3rd edn. Butterworth-Heinemann
(1999)
[28] Zienkiewicz, O.C., Taylor, R.L.: The Finite Element Method, 5th edn. Butterworth-
Heinemann (2000)
[29] Liu, G.R., Quek, S.S.: The Finite Element Method: A Practical Course. Butterworth
Hienemann (2003)
[30] Hutton, D.V.: Fundamentals of finite element analysis, 1st edn. Mc Graw Hill (2004)
[31] Belytschko, T., Liu, W.K., Moran, B.: Nonlinear Finite Elements for Continua and
Structures. John Wiley and Sons, New York (2000)
142 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

[32] Strenkowski, J.S., Moon, K.-J.: Finite element prediction of chip geometry and
tool/workpiece temperature distributions in orthogonal metal cutting. ASME J. Eng.
Ind. 112, 313–318 (1990)
[33] Raczy, A., Elmadagli, M., Altenhof, W.J., Alpas, A.T.: An Eulerian finite-element
model for determination of deformation state of a copper subjected to orthogonal cut-
ting. Metall Mater. Trans. 35A, 2393–2400 (2004)
[34] Mackerle, J.: Finite element methods and material processing technology, an adden-
dum (1994–1996). Eng. Comp. 15, 616–690 (1962)
[35] Rakotomalala, R., Joyot, P., Touratier, M.: Arbitrary Lagrangian-Eulerian thermome-
chanical finite element model of material cutting. Comm. Numer. Meth. Eng. 9, 975–
987 (1993)
[36] Pepper, D.W., Heinrich, J.C.: The Finite Element Method: Basic Concepts and Appli-
cations. Hemisphere Publishing Corporation, United States of America (1992)
[37] Bower, A.F.: Applied mechanics of solid. CRC Press, Taylor and Francis Group,
New York (2010)
[38] Dhondt, G.: The finite element method for three-dimensional thermomechanical ap-
plications. John Wiley and Sons Inc., Germany (2004)
[39] Pian, T.H.H.: Derivation of element stiffness matrices by assumed stress distribu-
tions. AIAA J. 2, 1333–1336 (1964)
[40] Zienkiewicz, O.C., Taylor, R.L.: The Finite Element Method. Basic formulations and
linear problems, vol. 1. McGraw-Hill, London (1989)
[41] Liapis, S.: A review of error estimation and adaptivity in the boundary element me-
thod. Eng. Anal. Bound. Elem. 14, 315–323 (1994)
[42] Tay, A.O., Stevenson, M.G., de Vahl Davis, G.: Using the finite element method to
determine temperature distribution in orthogonal machining. Proc. Inst. Mech.
Eng. 188(55), 627–638 (1974)
[43] Muraka, P.D., Barrow, G., Hinduja, S.: Influence of the process variables on the tem-
perature distribution in orthogonal machining using the finite element method. Int. J.
Mech. Sci. 21(8), 445–456 (1979)
[44] Moriwaki, T., Sugimura, N., Luan, S.: Combined stress, material flow and heat analy-
sis of orthogonal micromachining of copper. CIRP Annals - Manufact. Tech. 42(1),
75–78 (1993)
[45] Kim, K.W., Sin, H.C.: Development of a thermo-viscoplastic cutting model using fi-
nite element method. Int. J. Mach. Tool Manufact. 36(3), 379–397 (1996)
[46] Liu, C.R., Guo, Y.B.: Finite element analysis of the effect of sequential cuts and tool-
chip friction on residual stresses in a machined layer. Int. J. Mech. Sci. 42(6), 1069–
1086 (2000)
[47] Ceretti, E., Falbohmer, P., Wu, W.T., Altan, T.: Application of 2D FEM to chip for-
mation in orthogonal cutting. J. Mater Process Tech. 59, 169–180 (1996)
[48] Li, K., Gao, X.-L., Sutherland, J.W.: Finite element simulation of the orthogonal met-
al cutting process for qualitative understanding of the effects of crater wear on the
chip formation. J. Mater Process Tech. 127, 309–324 (2002)
[49] Arrazola, P.J., Ugarte, D., Montoya, J., Villar, A., Marya, S.: Finite element modeling
of chip formation process with abaqus/explicit. VII Int. Conference Comp., Barcelona
(2005)
[50] Davies, M.A., Cao, Q., Cooke, A.L., Ivester, R.: On the measurement and prediction
of temperature fields in machining AISI 1045 steel. Annals of the CIRP 52, 77–80
(2003)
Finite Element Modeling of Chip Formation in Orthogonal Machining 143

[51] Adibi-Sedeh, A.H., Vaziri, M., Pednekar, V., Madhavan, V., Ivester, R.: Investigation
of the effect of using different material models on finite element simulations of metal
cutting. In: 8th CIRP Int Workshop Modeling Mach Operations, Chemnitz, Germany
(2005)
[52] Shi, J., Liu, C.R.: The influence of material models on finite element simulation of
machining. J. Manufact. Sci. Eng. 126, 849–857 (2004)
[53] Ozel, T.: Influence of Friction Models on Finite Element Simulations of Machining.
Int. J. Mach. Tool Manufact. 46(5), 518–530 (2006)
[54] Filice, L., Micari, F., Rizzuti, S., Umbrello, D.: A critical analysis on the friction
modeling in orthogonal machining. International Journal of Machine Tools and Man-
ufacture 47, 709–714 (2007)
[55] Haglund, A.J., Kishawy, H.A., Rogers, R.J.: An exploration of friction models for the
chip-tool interface using an Arbitrary Lagrangian-Eulerian finite element model.
Wear 265(3-4), 452–460 (2008)
[56] Mabrouki, T., Deshayes, L., Ivester, R., Regal, J.-F., Jurrens, K.: Material modeling
and experimental study of serrated chip morphology. In: Proceedings of 7th CIRP Int
Workshop Model. Machin, France, April 4-5 (2004)
[57] Coelho, R.T., Ng, E.-G., Elbestawi, M.A.: Tool wear when turning AISI 4340 with
coated PCBN tools using finishing cutting conditions. J. Mach. Tool Manufact. 47,
263–272 (2006)
[58] Lorentzon, J., Jarvstrat, N.: Modelling tool wear in cemented carbide machining alloy
718. J. Mach. Tool Manufact. 48, 1072–1080 (2008)
[59] Davim, J.P., Maranhao, C., Jackson, M.J., Cabral, G., Gracio, J.: FEM analysis in
high speed machining of aluminium alloy (Al7075-0) using polycrystalline diamond
(PCD) and cemented carbide (K10) cutting tools. Int. J. Adv. Manufact. Tech. 39,
1093–1100 (2008)
[60] Attanasio, A., Cerretti, E., Rizzuti, S., Umbrello, D., Micari, F.: 3D finite element
analysis of tool wear in machining. CIRP Annals – Manufact. Tech. 57, 61–64 (2008)
[61] ABAQUS Analysis User’s manual. Version 6.7-4 Hibbitt, Karlsson & Sorensen, Inc.
(2007)
[62] ABAQUS Theory manual, Version 6.7-4 Hibbitt, Karlsson & Sorenson, Inc. (2007)
[63] ABAQUS/CAE User’s manual. Version 6.7-4 Hibbitt, Karlsson & Sorensen, Inc.
(2007)
[64] Wu, H.-C.: Continuum Mechanics and Plasticity. Chapman and Hall/CRC (2004)
[65] Johnson, G.R., Cook, W.H.: A constitutive model and data for metals subjected to
large strains, high strain rates and high temperatures. In: Proceedings of 7th Int Symp
Ballistics, the Hague, The Netherlands, pp. 541–547 (1983)
[66] Umbrello, D., M’Saoubi, R., Outeiro, J.C.: The influence of Johnson–Cook material
constants on finite element simulation of machining of AISI 316L steel. Int. J. Mach.
Tool Manufact. 47, 462–470 (2007)
[67] Davim, J.P., Maranhao, C.: A study of plastic strain and plastic strain rate in machin-
ing of steel AISI 1045 using FEM analysis. Mater Des. 30, 160–165 (2009)
[68] Vaziri, M.R., Salimi, M., Mashayekhi, M.: A new calibration method for ductile frac-
ture models as chip separation criteria in machining. Simulat Model Pract. Theor. 18,
1286–1296 (2010)
[69] Johnson, G.R., Cook, W.H.: Fracture characteristics of three metals subjected to vari-
ous strains, strains rates, temperatures and pressures. Eng. Fract. Mech. 21(1), 31–48
(1985)
144 A. Priyadarshini, S.K. Pal, and A.K. Samantaray

[70] Mabrouki, T., Girardin, F., Asad, M., Regal, J.-F.: Numerical and experimental study
of dry cutting for an aeronautic aluminium alloy. Int. J. Mach. Tool Manufact. 48,
1187–1197 (2008)
[71] Mabrouki, T., Rigal, J.: -F A contribution to a qualitative understanding of thermo-
mechanical effects during chip formation in hard turning. J. Mater. Process
Tech. 176, 214–221 (2006)
[72] Duan, C.Z., Dou, T., Cai, Y.J., Li, Y.Y.: Finite element simulation & experiment of
chip formation process during high speed machining of AISI 1045 hardened steel. Int.
J. Recent Trend Eng. 1(5), 46–50 (2009)
[73] Priyadarshini, A., Pal, S.K., Samantaray, A.K.: A Finite Element Study of Chip For-
mation Process in Orthogonal Machining. Int . J. Manufact., Mater. Mech. Eng. IGI
Global( accepted, in Press, 2011)
[74] Shi, G., Deng, X., Shet, C.: A finite element study of the effect of friction in ortho-
gonal metal cutting. Finite Elem. Anal. Des. 38, 863–883 (2002)
[75] Lima, J.G., Avila, R.F., Abrao, A.M., Faustino, M., Davim, J.P.: Hard turning: AISI
4340 high strength alloy steel and AISI D2 cold work tool steel. J. Mater. Process.
Tech. 169, 388–395 (2005)
[76] Priyadarshini, A., Pal, S.K., Samantaray, A.K.: Finite element study of serrated chip
formation and temperature distribution in orthogonal machining. J. Mechatron Intell.
Manufact. 2(1-2), 53–72 (2010)
[77] Wang, M., Yang, H., Sun, Z.-C., Guo, L.-G.: Dynamic explicit FE modeling of hot
ring rolling process. Trans. Nonferrous Met. Soc. China 16(6), 1274–1280 (2006)
[78] Litonski, J.: Plastic flow of a tube under adiabatic torsion. Bulletin of Academy of
Pol. Science, Ser. Sci. Tech. XXV, 7 (1977)
[79] Batra, R.C.: Steady state penetration of thermo-visoplastic targets. Comput Mech. 3,
1–12 (1988)
[80] Usui, E., Shirakashi, T.: Mechanics of machining–from descriptive to predictive
theory: On the art of cutting metals-75 Years Later. ASME PED 7, 13–55 (1982)
[81] Maekawa, K., Shirakashi, T., Usui, E.: Flow stress of low carbon steel at high tem-
perature and strain rate (Part 2)–Flow stress under variable temperature and variable
strain rate. Bulletin Japan Soc Precision Eng 17, 167–172 (1983)
[82] Zerilli, F.J., Armstrong, R.W.: Dislocation-mechanics-based constitutive relations for
material dynamics calculations. J. Appl. Phys. 61, 1816–1825 (1987)
[83] Oxley, P.L.B.: The mechanics of machining: An analytical approach to assessing ma-
chinability. Ellis Horwood Limited, Chichester (1989)
[84] Banerjee, B.: The mechanical threshold stress model for various tempers of AISI
4340 steel. Int. J. Solid Struct. 44, 834–859 (2007)
[85] Priyadarshini, A., Pal, S.K., Samantaray, A.K.: On the Influence of the Material and
Friction Models on Simulation of Chip Formation Process. J. Mach. Forming Tech.
Nova Science (accepted, 2011)
[86] Meyer, H.W., Kleponis, D.S.: Modeling the high strain rate behavior of titanium un-
dergoing ballistic impact and penetration. Int. J. Impact Eng. 26(1-10), 509–521
(2001)
[87] Lee, W.S., Lin, C.F.: High temperature deformation behaviour of Ti6Al4V alloy ev-
luated by high strain rate compression tests. J. Mater. Process. Tech. 75, 127–136
(1998)
[88] Baker, M.: The influence of plastic properties on chip formation. Comp. Mater.
Sci. 28, 556–562 (2003)

View publication stats