International Journal of Machine Tools & Manufacture 39 (1999) 731–750

Thermo-mechanical modeling of orthogonal machining

process by finite element analysis
S. Lei, Y.C. Shin*, F.P. Incropera
School of Mechanical Engineering, Purdue University, West Lafayette, IN 47907, USA
Received 4 November 1997; received in revised form 5 June 1998

Abstract
A plane strain finite element method is used with a new material constitutive equation for 1020 steel to
simulate orthogonal machining with continuous chip formation. Deformation of the workpiece material is
treated as elastic–viscoplastic with isotropic strain hardening, and the numerical solution accounts for coup-
ling between plastic deformation and the temperature field, including treatment of temperature-dependent
material properties. To avoid numerical errors associated with large deformation of elements, automatic
remeshing is used, with at least 15 rezonings required to achieve a satisfactory solution. Effects of the
uncertainty in the constitutive model on the distributions of strain, stress and temperature around the shear
zone are presented, and the model is validated by comparing average values of the predicted stress, strain,
strain rate and temperature at the shear zone with experimental results. Parametric effects associated with
cutting speed and initial work temperature are considered in the simulations.  1998 Elsevier Science Ltd.
All rights reserved.
Keywords: Finite element modeling; Machining simulation; Orthogonal machining; Thermo-mechanical modeling

Nomenclature
b width of cut, mm
Cp specific heat, J/kg/°C
cMi⫹1 difference between current and the exact solution
f body force per unit volume, N/m3
fT transpose of f
FN force component conjugate to the Nth variable, N

* Corresponding author. Tel.: 765-494-9775; fax: 765-494-0539; e-mail: shin@ecn.purdue.edu

0890-6955/99/$—see front matter  1999 Elsevier Science Ltd. All rights reserved.
PII: S 0 8 9 0 - 6 9 5 5 ( 9 8 ) 0 0 0 5 9 - 5
732 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750

FNi FN after the ith iteration

i inclination angle, deg
k thermal conductivity, W/m/°C
K NP i Jacobian matrix
L crack length, m
m strain rate sensitivity in the constitutive model
n strain hardening index in the constitutive model; surface normal
NN interpolation functions
NTN transpose of NN
q̇pl heat generation rate due to plastic deformation, W/m3
qint heat flux due to friction at tool–chip interface, W/m2
S surface bounding volume V, m2
t1 uncut chip thickness, mm
T temperature, °C
T̄ average shear zone temperature, °C
Tin initial temperature, °C
Tm melting temperature, °C
T* ratio of T to Tm
ux velocity in x direction, m/s
uy velocity in y direction, m/s
U cutting speed, m/s
uM value of the Mth variable
uM i uM after the ith iteration
uM i⫹1 uM after the (i ⫹ 1)th iteration
V volume occupied by the material, m3
V0 reference volume, m3
x,y,z coordinate axis in FEM model
␣ rake angle, deg
␤N strain variation
␦D virtual strain rate, 1/s
␦u virtual velocity field, m/s
⑀ strain
⑀¯ average shear zone strain
⑀0 reference strain
⑀˙ strain rate, 1/s
⑀¯˙ average shear zone strain rate, 1/s
⑀˙ 0 reference strain rate, 1/s
⑀˙ pl rate of plastic straining, 1/s
⌫ surface traction per unit area, N/m2
⌫T transpose of ⌫
␩ fraction of mechanical work required for plastic deformation which is converted to
thermal energy
␳ density, kg/m3
␴ stress, MPa
 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750 733

␴¯ average shear zone stress, MPa

␴0 reference stress, MPa
␶c material stress, MPa

1. Introduction

Although analytical and numerical studies of metal cutting processes are extensive, the defor-
mation characteristics of machining processes are not fully understood and accurate predictive
models have yet to be developed. The common plane strain, orthogonal machining model which
has been frequently used with the assumption of continuous chip formation, but no built-up edge,
is shown in Fig. 1. Material first undergoes severe plastic shearing in the primary shear defor-
mation zone, which is the area bounded by OBCAD, and subsequently experiences additional
shearing in a very thin secondary deformation zone (ODE) at the tool–chip interface. The defor-
mation process is complex, with large strains and strain rates, as well as a wide range of tempera-
tures, creating special difficulties in developing reliable process models. Hence, many predictive
models are based on an idealized shear plane and rely on experimental inputs. However, the finite
element method (FEM) provides an effective means of treating the foregoing complexities and
is used in this study to determine distributions of strain, stress, strain rate and temperature in
1020 steel during orthogonal machining. The predictions are compared with results obtained from
tube turning experiments which approximate orthogonal machining.
Many investigators have adopted the finite element method (FEM) to gain a better understand-
ing of machining processes which involve deformation with large strains, strain rates and tempera-
tures [1–11]. Through FEM analysis, various quantities can be numerically calculated, including
spatial distributions of the strain, stress, temperature, and strain rate, each of which is difficult to
measure during machining experiments. However, to obtain meaningful results which reflect the
physical mechanisms of metal cutting processes, it is essential to have a satisfactory material
constitutive equation which describes the deformation behavior of the workpiece material. Differ-
ent material constitutive equations have been used in the literature, thus resulting in substantial
variation among the simulation results. Since none of the aforementioned studies has provided a

Fig. 1. Orthogonal machining model with continuous chip formation.

734 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750

thorough comparison of predicted quantities in the shear zone with experimental results, it consti-
tutes a main objective of this study.
In this study, a constitutive equation developed by the authors [12] for 1020 carbon steel is
used in implementing a plain strain FEM analysis. The constitutive equation was obtained from
orthogonal machining tests. Strains, strain rates, stresses and temperatures were calculated using
machining theories with measured cutting forces, chip thicknesses and grain elongation directions
in chips. The constitutive model was then formed by correlating the stress as a function of the
strain, strain rate and temperature over ranges typically encountered in machining processes.
Uncertainties in determining the flow stress are linked to uncertainties in obtaining strains, strain
rates and temperatures from the machining experiments, and their effect on the FEM analysis is
assessed. To validate the model, average values of these quantities obtained from the simulation
are compared with those obtained from the machining experiments at the shear deformation zone.

2. Finite element modeling

2.1. Finite element mesh

The initial workpiece mesh is shown in Fig. 2, where two-dimensional behavior is based on
assuming a plane strain condition. The assumption is appropriate if the width of the cut, b, is
much larger than the uncut chip thickness, t1 (b ⱖ 10t1). The workpiece is represented by four-
node bilinear displacement and temperature plane strain elements, and since the objective is to
obtain a steady-state solution, initial chip formation is prescribed to facilitate the analysis. The
mesh consists of 1819 nodes and 1231 elements, with a typical element dimension of 0.018 mm.
Elements located along the tool–chip interface and the machined surface are refined to account
for large deformation gradients.
In view of the large elastic modulus of the cutting tool relative to the work material, the cutting
tool is modeled as a rigid body. This assumption presumes a small elastic deformation of the tool
compared to the significant plastic deformation of the workpiece material. A segment of the
surface (rake face) of the rigid tool is defined by two-node, two-dimensional rigid elements. The
deforming workpiece is on the same side of the rake face, and the surface normal extends outwards
from the rake face. The surface normal is critical for establishing contact between the tool and
the workpiece. A reference node which has both translational and rotational freedom is assigned

Fig. 2. Finite element mesh.

 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750 735

to the rigid tool and acts as the master node for the tool. The motion of the rigid tool is entirely
determined by the reference node, which can therefore be used to prescribe loads and kinem-
atic constraints.

2.2. Modeling of workpiece–chip separation and tool–chip contact

A schematic showing the potential tool–chip and chip–workpiece surfaces is given in Fig. 3.
Surface EF represents the bottom side of the potential chip, with its outward normal pointing
downward, while surface EG represents the upper side of the machined surface, with its outward
normal pointing upward. The two surfaces form a contact pair, with EF and EG providing master
and slave surfaces, respectively. By definition, the nodes of the slave surface are constrained from
penetrating into the master surface. Initially, the corresponding nodes on EF and EG are perfectly
bonded and therefore have identical displacements in both x and y directions, as well as equivalent
temperatures. During the simulation, the nodes debond to form a chip surface and a machined
surface. This separation process is governed by a debonding criterion specified in the model.
Although various chip separation criteria have been used in applying finite element analysis to
metal cutting, a criterion based on controlled crack propagation is considered to be appropriate
for this study, since material is ultimately separated through microcracks which develop as the
tool moves into the workpiece. Crack propagation through the material is illustrated schematically
in Fig. 4. Selection of a crack length versus time relationship is based on movement of the cutting
tool, and a reference point is defined on the slave surface, which is the machined surface from
which crack length is measured. The length is calculated between the reference point and each
node on the contact surfaces that may debond. Based on these lengths and the given crack length–
time curve, the time increment is adjusted so that the debonding event occurs within the prescribed
time tolerance for the crack to reach that length. The master and slave surfaces of the tool–chip
contact pair consist of the rigid tool surface and surface EFH.
The interaction between the tool and chip surfaces is modeled by a small-sliding contact formu-
lation [13] and Fig. 5 shows the corresponding evolution of this interaction between a slave node
580 and the master tool surface. At time t ⫽ 0, it is prescribed that slave node 580 will interact
with the master surface formed by nodes 10002 and 10003, and a potential contact is thereby
enforced. For t > 0, node 580 cannot penetrate this surface, and the mechanical interaction at the

Fig. 3. Contact modeling.

736 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750

Fig. 4. Schematic for debonding of contact surface.

Fig. 5. Small-sliding contact with geometrical nonlinearity.

 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750 737

Fig. 6. Pressure–clearance relationship for contact interface.

tool–chip interface is governed by a contact pressure–clearance relationship, as shown in Fig. 6.

In the small-sliding contact formulation, any pressure stress can be transmitted between the sur-
faces when they are in contact. The surfaces separate if the contact pressure becomes zero, and
separated surfaces come into contact when there is no clearance between them.

2.3. Boundary conditions

Kinematic boundary conditions (Fig. 7) are prescribed by constraining the left (AEL), right
(BG) and bottom (AB) sides of the workpiece from any movement, while allowing all other
boundaries to move. These conditions are valid as long as side AEL is far from the deformation
zone and depth penetration on the machined surface is small. The cutting speed, U, is assigned
to the cutting tool by choosing the time interval and corresponding tool displacement in the
analysis. In this study, friction forces at the tool–chip interface are neglected, but a heat flux
associated with friction effects is calculated from the machining experiments and applied to the

Fig. 7. Kinematic boundary conditions.

738 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750

contact area (Fig. 8). The secondary shear zone at the tool–chip interface is neglected, since it is
localized within a very thin layer of the chip bottom and has little effect on the primary shear
zone, within which strain, stress, strain rate and temperature distributions are of primary concern.
As shown in Fig. 8, the top surface of the workpiece and exposed surfaces of the chip are
assumed to be adiabatic (∂T/∂n ⫽ 0), since they are in contact with air, to which heat transfer
may be neglected. The same assumption is made for the top and right boundaries of the machined
workpiece material. The left and bottom boundaries of the workpiece are assumed to remain at
the initial (room) temperature, since they are well removed from the deformation zone. Finally,
a uniform heat flux is applied to the chip at the tool–chip interface to account for heat generation
due to friction.

3. The constitutive relation and workpiece material properties

In this study, an elastic–viscoplastic material model, which assumes isotropic strain hardening,
is used to quantify deformation of the workpiece material. The material constitutive equation has
been developed from orthogonal machining experiments [12] and is given by:

␴ ⫽ ␴0 冉 冊冉 冊
⑀
⑀0
n
⑀˙
⑀˙ 0
m
F(T*) (1)

where, n, the strain hardening index, and m, the strain rate sensitivity, are equal to 0.34 and 0.08
respectively for the 1020 steel used in the present analysis. The function F(T*) describes the
temperature effect, where T* ⫽ T/Tm and Tm is the melting temperature of the material:
F(T*) ⫽ ⫺ 1.23T* ⫹ 1.27, for T ⱕ 600°C (2)

F(T*) ⫽ 72.3(T*)3 ⫺ 137.5(T*)2 ⫹ 80.4T* ⫺ 13.9, for 600 ⬍ T ⱕ 900°C (3)

Fig. 8. Thermal boundary conditions.

 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750 739

Reference values ␴0, ⑀0, and ⑀˙ 0 of stress, strain and strain rate are 780 MPa, 0.85, and 16,000 s−1,
respectively. This constitutive model characterizes deformation of 1020 steel for the large strains,
strain rates and temperatures which typically occur during metal cutting.
Workpiece material properties, including temperature dependence where appropriate, are
1. Thermal conductivity: k ⫽ 53.56 ⫺ 0.0266T, W/m/°C
2. Specific heat: Cp ⫽ 398.8 ⫹ 0.8273T ⫺ 0.00055T2, J/kg/°C
3. Thermal expansion coefficient: 1.6 ⫻ 10−5 m/m°C
4. Young’s modulus: 75 GPa
5. Poisson’s ratio: 0.3
6. Density: ␳ ⫽ 7882 kg/m3

4. Finite element formulation

The general purpose finite element package ABAQUS1 is used in this study to simulate the
orthogonal machining process with a compatible displacement formulation. The displacement
finite element method approximates the equilibrium requirement in the form of a virtual work
statement

冕
V
冕
␴:␦DdV ⫽ ⌫T·␦udS ⫹ fT·␦udV
S
冕V
(4)

where V denotes the volume occupied by the body, S the surface bounding this volume, ⌫ the
surface traction per unit area, f the body force per unit volume, ␴ the cauchy stress matrix, ␦D
the virtual strain rate, ␦u the virtual velocity field, and the superscript T indicates transpose of a
matrix. From Eq. (4), a set of nonlinear equilibrium equations can be derived:

冕 冕 冕
␤N:␶cdV0 ⫽ NTN·⌫dS ⫹ NTN·fdV (5)
V0 S V

where ␤N defines the matrix of strain variation, ␶c is the matrix of material stresses, and NN is a
matrix of interpolation functions. This set of equations is the basic finite element stiffness model,
and may be written as
FN(uM) ⫽ 0 (6)
where FN is the force component conjugate to the Nth variable and uM is the value of the Mth vari-
able.
The nonlinear equations are numerically solved using Newton’s method. Assuming that uM i is

1
ABAQUS is a trademark of Hibbitt, Karlsson & Sorensen, Inc.
740 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750

an approximation to the solution after the ith iteration and cM

i ⫹ 1 is the difference between this
solution and the exact solution, the set of equations may be expressed as
i ⫹ ci ⫹ 1) ⫽ 0
FN(uM M
(7)

Expanding the left-hand side of this result in a Taylor series about uM

i and eliminating higher
order terms with small cM
i⫹1 , it follows that
i ci ⫹ 1 ⫽ ⫺ Fi
K NP M N
(8)
where K NP
i is the Jacobian matrix. The difference cM
i ⫹ 1 is calculated from Eq. (8), and the next
approximation to the solution is then
i ⫹ 1 ⫽ ui ⫹ c i ⫹ 1
uM M M
(9)

Iteration continues until the solution converges. Convergence of the analysis is enforced by
ensuring that all entries in FNi and cM
i ⫹ 1 are sufficiently small. Convergence is satisfied when the
largest residuals are small compared with the average conjugate fluxes for all the fields, and the
largest correction to the solution for each field is also small compared to the largest incremental
change in the corresponding solution variable.

5. Simulation of orthogonal metal cutting

A nonlinear temperature–displacement solution procedure is used to determine the coupled

temperature and stress fields in the workpiece and the chip during material removal. Heat gener-
ation due to plastic deformation may be expressed as
q̇pl ⫽ ␩␴:⑀˙ pl (10)
where q̇pl is the volumetric rate of heat generation, ␩ is the fraction of the mechanical work
required for plastic deformation which is converted into thermal energy, ␴ is the stress, and ⑀˙ pl
is the rate of plastic straining.
The steady-state, two-dimensional form of the energy equation governing the orthogonal mach-
ining process is

k 冉
∂2T ∂2T
∂x2
∂y 冊
⫹ 2 ⫺ ␳Cp ux
∂T
∂x 冉
⫹ uy
∂T
∂y
⫹ q̇pl ⫽ 0冊 (11)

The temperatures are integrated using a backward difference scheme, and the coupled system
is solved using Newton’s method.
Nonlinearities in the numerical model are associated with material behavior, the geometry of
the elements, and the change in boundary contact conditions. To enhance convergence, minimum
and maximum time increments are specified, and within these limits actual increments are auto-
matically adjusted based on the difficulty or ease with which convergence is achieved.
 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750 741

With increasing strain, some of the elements near the cutting tool become so severely distorted
that they are no longer suitable for use. In such cases, rezoning is necessary to map the old
solution onto a new mesh and continue the analysis. The need for rezoning is established by
monitoring the extent of distortion of the elements from the deformed configuration plots, and
rezoning is implemented by extracting the current model geometry from data in the results file.
After forming a new mesh, the analysis is continued, using the previous results as new initial con-
ditions.
A two-step interpolation technique is used to transform the old solution to the new mesh. First,
all values of the variables are obtained at the nodes of the old mesh by extrapolating the data
from the integration points to the nodes of each element. The values are then averaged over the
elements adjacent to each node. Second, each integration point in the new mesh is located with
respect to the old mesh, and the variables are interpolated from the nodes of the element in the
old mesh to the locations in the new mesh. After all of the solution values are transformed, the
analysis proceeds on the new mesh. A typical simulation of the orthogonal machining process
requires many rezonings before achieving converged results. For all cases in this study, rezoning
had to be done after the tool moved approximately 0.05 mm into the workpiece.

6. FEM simulation results

6.1. Machining conditions and experimental results

A sample solution of the FEM analysis is provided for 1020 steel. The cutting conditions were
chosen on the basis of having achieved good repeatability of experiments performed for equivalent
conditions and because the experiments were characterized by uniform, continuous chip formation.
The tool geometry and machining parameters are summarized in Table 1, together with the exper-
imentally obtained average values of the strain, ⑀¯ , stress, ␴¯ , strain rate, ⑀¯˙ , and temperature, T̄, at
the shear zone. Machining was performed on a 1020 steel tube using a 7-hp engine lathe. The
width of cut, b, was 3 mm, which is more than 10 times the uncut chip thickness, t1, thus resulting
in negligible side flow. Cutting forces were measured by a three-component Kistler 9257B dyna-
mometer, and the chip thickness was obtained by examining the chip cross-section under an
optical microscope. The average strain, stress, temperature and strain rate were obtained by using
the measurements in conjunction with orthogonal machining theory [12].

Table 1
Machining conditions and results

Cutting tool: tungsten carbide, ␣ ⫽ 5°, i ⫽ 0°

U (m/s) t1 (mm) ⑀¯ ␴¯ (MPa) T̄ (°C) ⑀¯˙ (s−1)

3.0 0.195 0.72 725 246 23,000

742 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750

6.2. Simulation results

Each simulation starts as the tool moves towards the workpiece at a specified speed. The chip
separates from the workpiece as debonding occurs at the contact surfaces and establishes contact
with the tool rake face. To avoid instability of the numerical analysis caused by unbalanced forces
at the nodes of the separating surfaces, a force reduction curve is specified to allow the forces to
gradually reduce to zero. Due to severe distortion of elements close to the tool tip, rezoning had
to be performed to continue the analysis. In this study, an average of 15 rezonings was needed
for the simulation to reach steady state.
The displaced configuration is given in Fig. 9, at which point the cutting tool has moved about
1 mm from the starting position. The chip has separated from the workpiece; the machined surface
has formed; and tool–chip contact has been fully established with a part of the chip curling away
from the rake face.
From Fig. 10, which shows the strain distribution, it is evident that deformation begins ahead
of the shear zone, and strain increases as the material moves towards the shear zone, until it
reaches the upper boundary of the zone. The strain increases from 0 to its maximum value with
flow of the work material. Changes are abrupt near the tool tip and are more gradual at locations
removed from the tip. When viewed along the shear zone from the tool tip to the free surface,
strain decreases quickly at first, from a maximum value at the tool tip, and then remains almost
constant throughout the shear zone. Along central portions of the shear zone, the strain varies
from approximately 3.1 to 0.8. The work material undergoes plastic deformation in the shear
zone, which, consistent with experimental observations [14,15], is narrow around the tool tip and
wider at the free surface. Fig. 11 illustrates the strain rate distribution, which is also characterized

Fig. 9. Deformed mesh after 1 mm machining.

 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750 743

Fig. 10. Distribution of strain.

Fig. 11. Distribution of strain rate.

by a maximum value near the tool tip. The value decreases with increasing distance from the
tool tip, becoming almost constant from the midpoint to the free surface of the shear zone.
Stress distributions, which result from the combined effect of strain, strain rate and temperature,
are shown in Fig. 12. When material moves towards central portions of the shear zone, the stress

Fig. 12. Distribution of stress.

744 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750

increases due to strain hardening and an increasing strain rate, until it reaches a maximum at the
center of the shear zone. The stresses are nearly uniform along the central shear zone, with
lower values at the tool–chip interface where elevated temperatures act to soften the material.
The temperature field is shown in Fig. 13. Temperatures increase due to plastic deformation in
the material, as well as friction at the tool–chip interface. A high rate of deformation leads to
concentrated heating of material in the deformation zone, with little heat transfer occurring by
conduction from the zone. Temperatures begin to increase from the start of straining and continue
to increase until straining ceases in the chip. However, temperature in a thin layer of material
adjacent to the tool rake face also increases due to heat generated by friction. Therefore, the
highest temperature, which is about 800°C, is located at the tool–chip interface and close to the
chip–tool separation point. From the tool tip to the free surface along the center of the shear zone,
the temperature varies from approximately 170 to 320°C and is higher near the tool edge due to
the intensive interaction between the work material and the cutting tool.
To compare the predictions with average values of the shear plane strain, strain rate, stress and
temperature obtained experimentally from machining tests, the effective strain, strain rate, stress
and temperature were computed for each element along the central shear zone and the results are
plotted in Fig. 14, where the error bars are associated with ⫾ 15% uncertainties in the constitutive
equation used for the predictions. Average values of the predictions along the shear plane are
given in Table 2.
Contrasting the experimental results of Table 1 with the prediction of Table 2, it is evident
that predictions of the average strain and stress exceed the measured values. This behavior is
partially due to the shear plane model used to obtain the experimental results, which attributes
deformation exclusively to shear. Hence, only the shear strain and shear stress along plane are
used to calculate the effective strain and stress. In the simulation, however, all strain and stress
components are included in calculating the effective strain and stress. Uncertainties in the
measurements constitute another source of discrepancy between the two sets of results.
Predicted and measured values of the average strain rate are in good agreement with the result
based on an empirical correlation by Oxley [15]. A significant feature of the strain rate plot in
Fig. 14 is the steep gradient near the tool tip, beyond which ⑀˙ decreases gradually towards the
free surface.
The average shear zone temperature listed in Table 2 is lower than the result of Table 1, which

Fig. 13. Temperature distribution.

 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750 745

Fig. 14. Variations of strain, strain rate, stress and temperature along central shear zone (U ⫽ 3 m/s, t1 ⫽ 0.195 mm,
Tin ⫽ 25°C).

Table 2
Simulation results along the central shear zone

⑀¯ ␴¯ (MPa) T̄ (°C) ⑀¯˙ (s−1)

1.06 950 200 24,000

746 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750

is based on the Loewen and Shaw [16] shear plane temperature model. The temperature rise is
due to plastic deformation within the material and is determined by the product of strain rate and
stress. Since these quantities are distributed over a finite zone in the FEM analysis, the correspond-
ing heat generation is also distributed over a finite region. Hence, the Loewen and Shaw model
may overpredict the temperature because it assumes a plane heat source on the shear plane [17].
Another possible source of discrepancy is the error in the experimentally determined constitut-
ive equation.
To examine the effect of uncertainty in the material constitutive equation on the FEM predic-
tions, two cases were tested, with errors of ⫺ 15% and ⫹ 15% presumed for the stress. These
variations are believed to represent appropriate limits associated with uncertainties in strain, strain
rate and temperature, and their effect on the FEM predictions is designated by the error bands of
Fig. 14. Average values across the central shear zone are summarized in Table 3 for the three
simulations and the machining experiment. With respect to the strain and strain rate, results of
the three simulations are comparable, which is reasonable in view of the fact that the same tool
geometry and machining conditions are used. However, the average temperature and stress
increase monotonically with increasing values of the stress obtained from the constitutive model.
Since uncertainties in the constitutive model have little effect on the strain rate, the temperature
increases with increasing stress, due to increased thermal energy generation. The average tempera-
ture for simulation 3 is close to the shear plane temperature inferred from the machining experi-
ment.

6.3. Additional simulations

Two more cases were simulated to study the effects of cutting speed and preheating, and results
at the central shear zone are shown in Fig. 15 for U ⫽ 4 m/s (case 4), and in Fig. 16 for an
initial workpiece temperature of Tin ⫽ 400°C (case 5). Average values of stresses, strains, strain
rates and temperatures along the central shear zone are listed in Table 4, together with experi-
mental results for the two cases, and agreement between the predictions and measurements is
generally good. Compared with the nominal case, the larger cutting speed of case 4 results in a
slightly lower strain, while the strain rate increases nearly in proportion to the cutting speed,
which agrees with experimental observations [15]. Tool–chip interface temperatures increase with
cutting speed due to an increase in frictional heating at the interface, although there is little change
in the shear zone temperature, which is generally insensitive to cutting speed. There is little
difference in the stresses, although the increase in the strain rate tends to slightly increase the
stress, since the decrease of strain has an opposing effect on the stress.
Table 3
Comparison between experimental and simulation results

Case Error in stress ⑀¯ ␴¯ (MPa) T̄ (°C) ⑀¯˙ (s−1)

Machining test 0.72 725 246 23,000

Simulation 1 ⫺ 15% 0.9 807 161 21,000
Simulation 2 0 1.06 952 200 24,000
Simulation 3 ⫹ 15% 0.98 1026 220 20,000
 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750 747

Fig. 15. Variations of strain, strain rate, stress and temperature along central shear zone (U ⫽ 4 m/s, t1 ⫽ 0.195 mm,
Tin ⫽ 25°C).

The effect of work preheating on the process is revealed by case 5. In the experiment, the
workpiece was heated by a plasma arc, and the temperature was measured by an infrared imaging
system. In the simulation, the initial temperature was prescribed as a uniform temperature field.
A significant effect of preheating relates to reduction of the stress with temperature, a generally
known benefit of hot machining. Since the heat generation rate is proportional to the product of
stress and strain rate, a reduction in the stress reduces heat generation in the shear zone, and the
corresponding temperature rise is less than that for the nominal case.
748 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750

Fig. 16. Variations of strain, strain rate, stress and temperature along central shear zone (U ⫽ 3 m/s, t1 ⫽ 0.1 mm,
Tin ⫽ 400°C).

7. Conclusions

A plain strain, finite element simulation of an orthogonal machining process has been conduc-
ted, for which temperature and stress calculations are coupled with temperature-dependent material
properties. The effects of strain hardening, strain rate, and temperature are considered by using
a new material constitutive model developed for 1020 steel [12]. Parametric calculations to deter-
mine the effect of uncertainties in the constitutive model reveal a significant influence on the
stress and temperature distributions.
Due to large plastic deformation, multiple rezonings are needed to guarantee accurate results
 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750 749

Table 4
Comparison between average experimental and simulation results along the central shear zone

Case Method T̄ (°C) ⑀¯ ␴¯ (MPa) ⑀¯˙ (s−1)

4 Machining test 240 0.7 790 31,000

4 Simulation 190 0.78 950 30,000
5 Machining test 620 0.92 600 47,000
5 Simulation 508 1.05 611 45,000

Case 4: U ⫽ 4 m/s, t1 ⫽ 0.195 mm, Tin ⫽ 25°C.

Case 5: U ⫽ 3 m/s, t1 ⫽ 0.1 mm, Tin ⫽ 400°C.

through the entire simulation process, with continuity between solutions imposed as a necessary
criterion for successful rezoning. Although the effect of frictional heating along the tool–chip
interface is considered, the phenomenon has little effect on conditions in the primary shear zone.
Distributions of strain, strain rate, stress and temperature along the shear zone are presented for
machining conditions corresponding to those considered experimentally, and the corresponding
average values are in good agreement with those obtained from the measurement.

References

[1] E. Usui, A. Hiwta, M. Masuko, Analytical prediction of three dimensional cutting process, ASME Journal of
Engineering for Industry 100 (1978) 222–243.
[2] K. Iwata, K. Osakada, Y. Terasaka, Process modeling of orthogonal cutting by the rigid-plastic finite element
method, ASME Journal of Engineering Materials and Technology 106 (1984) 132–138.
[3] J.S. Strenkowski, J.T. Carroll III, A finite element model of orthogonal metal cutting, ASME Journal of Engineer-
ing for Industry 107 (1985) 349–354.
[4] J.T. Carroll III, J.S. Strenkowski, Finite element models of orthogonal cutting with application to single point
diamond turning, International Journal of Mechanical Science 30 (1988) 899–920.
[5] A.J. Shih, S. Chandrasekar, H.T. Yang, Finite element simulation of metal cutting process with strain-rate and
temperature effects, in: B.E. Klamecki, K.J. Weinmann (Eds), Fundamental Issues in Machining, ASME PED,
vol. 43, 1990, pp. 11–24.
[6] K. Komvopoulos, S.A. Erpenbeck, Finite element modeling of orthogonal metal cutting, ASME Journal of Engin-
eering for Industry 113 (1991) 253–267.
[7] V.S. Joshi, P.M. Dixit, V.K. Jain, Viscoplastic analysis of metal cutting by finite element method, International
Journal of Machine Tools Manufacture 34 (4) (1994) 553–571.
[8] T.D. Marusich, M. Ortiz, Finite element simulation of high-speed machining, Mechanics in Materials Processing
and Manufacturing AMD194 (1994) 137–149.
[9] A.J. Shih, Finite element simulation of orthogonal metal cutting, ASME Journal of Engineering for Industry 117
(1995) 84–92.
[10] K.W. Kim, H. Sin, Development of a thermo-viscoplastic cutting model using finite element method, International
Journal of Machine Tools Manufacture 36 (3) (1996) 379–397.
[11] J. Wu, O.W. Dillon, W. Lu, Thermo-viscoplastic modeling of machining process using a mixed finite element
method, ASME Journal of Manufacturing Science and Engineering 118 (1996) 471–482.
[12] S. Lei, Y.C., Shin, F.P. Incropera, Material constitutive modeling under high strain rates and temperatures through
orthogonal machining tests, in: Proceedings of Manufacturing Science and Engineering, ASME IMECE, MED-
vol. 6-2 (1997) 91–98.
[13] Hibbitt, Karlsson, and Sorenson, Inc. ABAQUS Theory and Users’ Manuals, Version 5.5, Providence, RI, 1995.
750 S. Lei et al. / International Journal of Machine Tools & Manufacture 39 (1999) 731–750

[14] K. Okushima, K. Hitomi, An analysis of the mechanism of orthogonal cutting and its application to discontinuous
chip formation, ASME Journal of Engineering for Industry 83 (1961) 545–556.
[15] P.L.B. Oxley, The Mechanics of Machining: An Analytical Approach to Assessing Machinability, Ellis Horwood
Limited, Chichester, UK, 1989.
[16] M.C. Shaw, Metal Cutting Principles, Oxford University Press, Oxford, 1984.
[17] D.A. Stephenson, An inverse method for investigating deformation zone temperatures in metal cutting, ASME
Journal of Engineering for Industry 113 (1991) 129–136.