ARTICLE IN PRESS

International Journal of Machine Tools & Manufacture 50 (2010) 765774

Contents lists available at ScienceDirect

International Journal of Machine Tools & Manufacture

journal homepage: www.elsevier.com/locate/ijmactool

A method for the identication of the specic force coefcients for

mechanistic milling simulation
Oscar Gonzalo a,b,n, Jokin Beristain a,b, Haritz Jauregi a,b, Carmen Sanz a,b
a
b

n Tekniker - IK4 Avenida Otaola 20, 20600 Eibar, Guipu

zcoa, Spain
Fundacio
zcoa, Spain
CIC marGUNE, Poligono Azitain 3, 20600 Eibar, Guipu

a r t i c l e in fo

abstract

Article history:
Received 2 October 2009
Received in revised form
12 February 2010
Accepted 21 May 2010
Available online 27 May 2010

This paper presents a new method to obtain the specic cutting coefcients needed to predict the
milling forces using a mechanistic model of the process. The specic coefcients depend on the tool
material couple, the cutting conditions and the geometry of the tool, being usually calculated applying
the force model in an inverse way. The most used inverse method is based on the calculation of the
average cutting force per revolution values measured in a series of slot machining tests at different feed
rates. In this research work, the inverse method is applied using the instantaneous cutting force values,
solving the equations system by a constrained least squares tting method. Furthermore, the cutting
force and specic cutting coefcients relation with rake angle and chip thickness is analysed. The
results are validated by the comparison of the simulations and experiments in orthogonal cutting test,
showing the advantages of using the new method.
& 2010 Elsevier Ltd. All rights reserved.

Keywords:
Modelling
Milling
Mechanistic
Simulation
Force

1. Introduction
End milling is one of the most common activities in the
manufacturing of die and mould, aeronautical, aerospace or
biomedical sectors. The cutting forces affect the quality and the
precision of the nal component; therefore precise prediction of
milling forces becomes an important factor to improve machining
performance. Moreover, reliable quantitative prediction of cutting
forces is essential for further prediction of the necessary power
and torque, machine tool vibrations, workpiece surface quality,
geometrical accuracy and process stability [1].
The calculation of cutting forces in interrupted machining is an
important research topic in which three main approaches can be
found nowadays. First, the analytical models which try to
establish mathematical relations between the cutting forces and
several mechanical aspects like friction, geometry and mechanical
behaviour of the materials. For instance, Merchant and Armarego
[2] are two of the main contributors to this kind of models by the
development of orthogonal and oblique cutting mechanics.
Secondly, the numerical methods [2,3] that are focused on the
tool tip zone and the study of the interaction between the tool and
the work piece. This nite element method was introduced in the
early 1970s [4] and it has become a relevant research eld in the

n
Corresponding author at: Fundacion Tekniker - IK4 Avenida Otaola 20, 20600
Eibar, Guipuzcoa, Spain. Tel.: +34 943 20 67 44; fax: +34 943 20 27 57.
E-mail addresses: [email protected] (O. Gonzalo), [email protected]
(J. Beristain), [email protected] (H. Jauregi), [email protected] (C. Sanz).

0890-6955/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijmachtools.2010.05.009

last decade, the twelve issues of the CIRP International Workshop

on Modeling of Machining Operations are an indicator of the
interest in this research eld.
Thirdly, the mechanistic methods [5,6] that use a series of
relations between some process variables and the cutting forces,
assuming that the cutting forces are proportional to the crosssectional area of the uncut chip [7]. These relations take into
account geometrical characteristics of the process and empirical
specic cutting coefcients obtained from experimental work that
characterize the toolworkpiece couple. The main advantage of
this type of models is the short calculation time, being the most
commonly used force modelling method nowadays.
Different types of mechanistic cutting force models are found
in the machining literature, among them, the research works done
by Armarego [8] and Kienzle [9] using linear and potential
approximation, respectively, can be highlighted. These models
have been further developed and both are used to simulate the
milling process [5,10]. For example, the potential model is a
function of the area of the uncut chip that is divided in two terms:
rst one is linear that takes into account the shearing mechanism
and the second one is a potential that considers friction, rubbing
and ploughing effects.
On the other hand, in the linear model the shearing and
friction+ rubbing+ ploughing mechanisms are modelled separately by two linear coefcients; the shearing coefcient multiplying the area of the uncut chip and the friction coefcient
multiplying the cutting edge length.
There are different options to obtain the specic cutting
coefcients from the experimental results, among them, the most

ARTICLE IN PRESS
766

O. Gonzalo et al. / International Journal of Machine Tools & Manufacture 50 (2010) 765774

extended are based on average forces measurements per revolution in

slot milling tests [5,11] or arbitrary radial immersion tests [1], while
others use the mechanics of cutting to obtain the coefcients [12].
Usually, the mechanistic models are used in an inverse way to
identify the specic cutting coefcients. The basis is to use the
cutting force measurement data of a series of characterization tests.
The method requires a set of slot milling tests at different feed rates
but constant radial and axial immersions. Assuming linear dependence of the average force on feed, the coefcients are obtained by
equating the analytical expressions for the average cutting and edge
forces to their measured counterparts [5,13,14].
This paper presents a new method for calculating the specic
cutting force coefcients using an inverse method in a linear
mechanistic model by the identication of the simulated and
measured forces in the time domain. The system of equations is
solved by a constrained least squares tting method. Furthermore,
the proposed method allows the use of selected parts of the
measured signal in order to avoid the problems associated to the
linear model that are also identied.
The specic cutting force coefcient results obtained are
improved using the proposed approach compared to the classical
average force based methods in addition to reduce the number of
required milling experiments to a single test. Furthermore, the
proposed method does not depend on the radial immersion
conditions used in the milling experimental test, so slot milling
tests are not necessary.
Finally, the inuence of chip thickness and cutting tool rake
angle on the cutting forces is analysed, identifying the main
drawbacks and limitations of the linear force model.

2. Proposed approach
As a rst approach, this research work obtains the specic
cutting coefcients from turning tests, which can be considered as
true coefcients. Later, the results obtained by using the
average force based [5] and instantaneous force based (the
proposed new method in this work) methods are compared.
The obtained data result in three groups of coefcients as
follows:

 Turning coefcients.
 Average force based coefcients from milling.
 Instantaneous force based coefcients from milling.
3. Process modelling. Mechanistic model
3.1. Milling process
The developed model allows the estimation of cutting forces in
milling processes, the calculating sequence follows three different

steps, see Fig. 1. In the rst step, the position of the end mill
cutting edges is calculated by cutting edge discretization [15,16].
Within the second step, the chip thickness for each element is
obtained for the forces calculation using the linear force model
[5]. Finally, in the last step, the resultant cutting force is estimated
by numerical integration.
In this mechanistic model, the elemental cutting forces model
is applied to each differential discrete element of the tool
(P discretization planes in each tooth, being Z the number of
teeth) and thus, the three force components: tangential, radial
and axial [5] are determined:
dFt Kte dS Ktc hn db
Linear model dFr Kre dS Krc hn db
dFa Kae dS Kac hn db

where dS is the edge length of each discrete element, hn is the

undeformed chip thickness, db is the chip width in each cutting
edge discrete element.
The proposed linear model is based on the expressions presented
by Altintas [5] which uses six empirical specic cutting coefcients
obtained from experiments: Ktc, Krc, Kac, Kte, Kre and Kae.
The basic equation of the force model includes two actions. On
one hand, the force component related to the material shearing
action (proportional to undeformed chip section: hn  db), given by
the Kic coefcients. On the other hand, the component related to
the friction, rubbing and ploughing actions (proportional to
elementary length of the cutting edge: dS), given by the Kie
coefcients. Where i means tangential, radial or axial.
The tangential, radial and axial components are transformed to
the X (feed), Y (normal) and Z (axial) directions by the following
transformations [5]:
8
9 2
9
38
cosy sinysinkr sinycoskr >
>
< dFX >
=
< dFt >
=
cosysinkr cosycoskr 7
dFY 6
4 siny
5 dFr
>
>
>
: dF ;
: dF >
;
sinkr
0
coskr
a
Z
2
where y is the tool rotation angular position and kr is the position
angle for each cutting edge discrete element. The total force is
calculated by adding all the contributions for each discrete
element of each tooth:
8 9
8
9
>
Z X
P >
< FX >
= X
< dFX >
=
FY
dFY
3
>
>
>
>
:F ;
1
1 : dF ;
Z
Z
The calculation of the tooth position, and the related chip
thickness, is based on the real trochoidal paths described by each
tool tooth [17]. The uncut chip thickness (hn) of the current tooth
is the distance between the current and previous tooth paths
along the line segment connecting the tool center to the current
tooths cutting edge, see Fig. 2A. This approach is used instead of
the lower computational cost approximation proposed by

Discretization
Geometry
Modelling

Force Mod.
Ft, Fr, Fa

Integration
P-planes
Z-teeth

Fig. 1. Mechanistic model scheme.

Cutting forces
Fx, Fy, Fz

ARTICLE IN PRESS
O. Gonzalo et al. / International Journal of Machine Tools & Manufacture 50 (2010) 765774

767

Fig. 2. Uncut chip thickness calculation: (A) trochoidal method; (B) Martellotis approximation and (C) comparison of both methods with details of the tool entrance and
exit zones.

Martellotti [18]: fzsin(y), see Fig. 2B. Martellottis approximation

states that the chip thickness at the tooth entrance (y 01) and
exit (y 1801) near the normal direction (Y) are zero, while the
chip thickness calculation based on the trochoidal path allows a
better modelling of these situations where the chip thickness is
not null (see Fig. 2C), improving the results of the simulation.
3.2. Method for obtaining specic cutting coefcients in milling
processes
The new method for calculating the specic cutting coefcients presented in this research work is based on the inverse
identication of the simulated and measured forces in the time
domain. Eq. (3) represents the force at any discrete time (t), and it
can be written as
8
9
Ktc >
>
>
>
>
>
>
>
>
Kte >
8 9simulated
>
>
>
>
>
>
F
>
> X=
>
>
<
<

 Krc =
FY
At, y, kr ,hn ,S
4
>
> ;
> Kre >
>
:
>
>
>
FZ t
>
>
>
> Kac >
>
>
>
>
>
>
:K >
;
ae

Equating the simulated and the experimental forces at several

discrete times, ti, the systems becomes
3
28 9
experimental
8
9
> FX >
<
=
7
6
Ktc >
>
7
6 FY
>
>
>
>
7
6> >
>
>
Kte >
7 2
6: ;
>
>
3>
>
>
7
>
>
6 FZ t1
At, y, kr ,hn ,St1 >
< Krc >
=
7
6
7
6
6
7
^
^
74
6
5
>
6 8 9experimental 7
K
re >
>
>
7
>
6 > FX >
At, y, kr ,hn ,Stn >
>
> Kac >
>
7
6< =
>
>
>
>
7
6
>
>
>
>
7
6 FY
:
;
5
4>
K
>
ae
:F ;
Z

tn

The specic force coefcients are determined by solving the

inverse problem presented in Eq. (5) applying a constrained
least squares tting method following the steps presented by
Tarantola et al. [19]. Furthermore, in order to improve the results,
the resolution of the system of equations include different
constraint and weight functions to reduce the search of the
local minimum to an interval of logical values for the specic
cutting coefcients obtained from bibliography [20] or user
experience, eliminating the mathematically consistent unrealistic
solutions.

ARTICLE IN PRESS
768

O. Gonzalo et al. / International Journal of Machine Tools & Manufacture 50 (2010) 765774

The objective is to minimize the following equation:

X
2
Jyi fi t,Ktc ,Krc ,Kac ,Kte ,Kre ,Kae J2D 99mmprior 99M

where yi are the X, Y and Z instantaneous components of the

measured force, and the function fi represents the force obtained
from the simulation model presented in the previous section. The
second part of the sums represents the constraining part, where m
is the iterative solution vector of specic cutting coefcients, and
mprior represents the vector of the initial estimation of logical
specic cutting coefcients values, given in order to constrain the
solution interval. D and M represent weighting factors.
The size of the system of equations resulting from a milling
experiment depends on the sampling frequency during the
experimental test, and the simulated points must match the
corresponding experimental points, i.e. the same origin for
the tool rotation angular position (y) must be selected. This is
done by an initial iterative process.
The milling cutting forces are periodic, and generally, the
modelling of the peak forces becomes one of the most important
values, because these can be used to establish suitable dimensions
in the machinetool or in the cutting tool design [12]. The cutting
forces are proportional to chip thickness, but the relationship is
not linear as will be demonstrated later in Section 5.2. So, it makes
sense to focus the cutting forces study and specic cutting
coefcients obtaining in specic selected intervals.
The tting process can be done using all the experimental
recorded forces or selected intervals of them. For example, zones
with maximum chip thickness as shown in Fig. 3, being the
objective to obtain local force models.
3.3. Method for obtaining specic cutting coefcients in turning
processes
The specic cutting coefcients are determined from cutting
force measurement under orthogonal cutting test conditions.

The experiments are conducted setting constants the depth of cut

and the cutting speed and varying the feed rate.
The corresponding specic cutting coefcients are obtained by
taking the average cutting force values and solving the inverse
identication using the least squares tting method as explained
in literature [21].

4. Experimental work
4.1. Experimental set-up
The work piece material analysed is AISI 1045 low carbon steel
with a hardness of 199 HBN. On one hand, the turning tests are
performed in a CNC CMZ TL-15M lathe (5000 rpm, 14 kW),
whereas the milling tests are carried out on a high speed milling
machine MAXIM (20 000 RPM, 15 kW). Both tests, milling and
turning, are performed with the same cylindrical end milling tool
with one removable insert made of uncoated hard metal, of
16 mm diameter and 01 helix angle, under orthogonal cutting
conditions. The tool is mounted in a HSK63A tool holder and a
taylor made xture when performing milling and orthogonal
cutting tests respectively. The turning tests consist in doing
orthogonal tube facing test, and the milling tests in slot milling
tests. The set-up for both processes can be seen in Fig. 4.
The cutting force measurements are done using a threecomponent dynamometer connected to a DAQ system.
In order to compare the results obtained from the turning
tests with those of the milling tests, the same insert geometry is
used: AEMW150304ER UTI20T. The tool holder reference is
BAE500R161S16, taylor made to obtain a 01 helix angle in the
orthogonal cutting test.
The experiments are carried out with two different rake
angles:  101 (nominal of the insert mounted in the tool holder)
and 61 (ultrasonically ground to obtain a positive rake angle). Both
inserts can be seen in the Fig. 5.

Fig. 3. Selected zones to obtain local force models in the region of high chip thickness in an orthogonal slot milling.

ARTICLE IN PRESS
O. Gonzalo et al. / International Journal of Machine Tools & Manufacture 50 (2010) 765774

769

Fig. 4. (A) Milling and (B) turning tests.

Fig. 5. Cutting insert and tool holder: (A)  101 rake angle; (B) +61 rake angle.

Tool ank wear (VB) is controlled with a contact microscope in

order to assure that it does not inuence on the cutting force
signicantly.
4.2. Experimental cutting conditions
Tables 1 and 2 show the cutting conditions for the turning and
milling test, respectively.
From the turning test, the specic cutting coefcients are
obtained, called turning coefcients in this work. Then, these
values of coefcients are compared to those obtained from milling
test, using both methods: average force based [5] and the new
proposed instantaneous force based method.

5. Results and discussion

In this section the inuence of the cutting speed and chip
thickness on the cutting forces is analysed. The suitability of the
linear force model [8] is assessed based on the orthogonal turning
experimental results, identifying the differences between positive
and negative rake angle tools.
Finally the specic cutting coefcients obtained by the three
different methods (turning, average force based and instantaneous force based) are compared and discussed.

Table 1
Cutting conditions for turning test.
Condition

Range

Cutting speed (vc)

Depth of cut (ap)
Feed rate (fn)

60; 80; 90; 100; 110; 120 m/min

2 mm
0.003; 0.004; 0,006; 0.008; 0.01; 0.015; 0.02;0.03;
0.04; 0.05; 0.06; 0.075 0.09; 0.105 mm/tooth
01
Carbide
16 mm
6  101

Obliquity/helix angle (i)

Tool material
Tool diameter
Rake angle (g)

Table 2
Cutting conditions for milling tests.
Condition

Range

Cutting speed (vc)

Axial depth of cut (ap)
Radial depth of cut (ae)
Feed rate (fz)

60; 80; 90; 100; 110; 120 m/min

2 mm
16 mm
0.05; 0.06; 0.075; 0.090;
0.105 mm/tooth
01
Carbide
16 mm
6  101

Obliquity/helix angle (i)

Tool material
Tool diameter
Rake angle (g)

5.1. Cutting speed dependencies

Fig. 6 shows that when the cutting speed increases from 60 to
90 m/min the cutting forces increase when machining AISI 1045.
However, the cutting forces decrease when the cutting speed
increases up to 120 m/min. This trend is observed for positive as
well as negative rake angle tools.

It is believed that this behaviour comes from metallurgical and

thermo-visco-plastic strength changes in the material at such
cutting speeds.
It has to be pointed out that the milling cutting forces
measured at 120 m/min have high residual vibrations in the

ARTICLE IN PRESS
770

O. Gonzalo et al. / International Journal of Machine Tools & Manufacture 50 (2010) 765774

Orthogonal Turning Cutting Forces

Orthogonal Turning Cutting Forces

300
Radial Force
Tangential Force

250

250

F t = 2168.7*f + 6.354

200

Force/ap (N)

Force/ap (N)

200

150
100
50

150
F r = 1160.7*f + 3.0625
100

50

0
60 m/min

80 m/min

90 m/min

100 m/min

0
0.000

120 m/min

Cutting speed

5.3. Specic cutting coefcients

After analyzing the basic behaviour of the proposed linear
model in turning, this section shows the specic cutting

50

0.060

0.080

0.100

0.120

Radial Force
Tangential Force
Linear

F t = 2230.4*f + 4.1743

40
Force/ap (N)

This section presents an analysis of the suitability of the

proposed linear force model for modelling of the cutting forces in
the feed range considered in this work. This analysis is done using
the data from the orthogonal turning experiments, being the feed
value equal to the uncut chip thickness.
Fig. 7A shows the tting of the cutting forces obtained
experimentally under different uncut chip thicknesses using
positive rake angle tools. This global t is acceptable although it
can be observed in Fig. 7B and C that a local t improves the
results obtaining lower errors. For instance, at low feeds, using the
global t the Ktc coefcient obtained is 2168.7 N/mm2, whereas
using the local t the coefcient is 2230.4 N/mm2. On the other
hand, at higher feeds the Ktc coefcient 1746.3 N/mm2.
Analyzing the case of negative rake angle tools, Fig. 8A shows
the tting of the cutting forces obtained experimentally under
different uncut chip thicknesses. This global t is not as
acceptable as in the case of positive rake angles. The reason of
this is depicted in Fig. 8B, showing that the evolution seems to be
more a complex function than a linear function. At high feed
values (see Fig. 8C), the Ktc coefcient obtained is 1979.3 N/mm2
while using global t is 2875.7 N/mm2.
Using negative rake angle, it is noticeable that the difference
between both force components, radial and tangential, is lower
than those observed in the case of positive rake angle. For
negative rake angle the forces take higher values, probably due to
the higher push energy necessary to deform the chip instead of
simply shearing it.
On balance, the results show that it is often difcult to
establish a tting model with a good behaviour over the whole
feed range, being the best option to make local models for
different feed ranges. In this analysis, the local linear models t
better the force at high feed values (f40.06 mm/rev) than the
global models, see Figs. 7C and 8C.

0.040

Orthogonal Turning Cutting Forces

30
20
F r = 1239.6*f + 0.6407
10
0
0.003 0.005 0.007 0.009 0.011 0.013 0.015 0.017 0.019
Feed (mm/rev)
Orthogonal Turning Cutting Forces
225
F t = 1746.3*f + 41.765

Force/ap (N)

5.2. Cutting force model

0.020

Feed (mm/rev)

Fig. 6. Cutting forces variation with cutting speed when machining AISI 1045
under orthogonal cutting conditions: g 61; fz 0.105 mm/tooth and ap 2 mm.

tooth entrance and exit zones, thus driving to slightly worse

recorded signal than those up to 100 m/min. For this reason in the
analysis of the results the case of the cutting speed of 120 m/min
is excluded.

Radial Force
Tangential Force
Linear fit

150
F r = 730.3*f + 39.324

75
Radial Force
Tangential Force
Linear

0
0.060 0.065 0.070 0.075 0.080 0.085 0.090 0.095 0.100 0.105
Feed (mm/rev)
Fig. 7. (A) Linear tting for a tool with positive rake angle: g 61; Vc 100 m/min;
ap 2 mm; (B) detail for local model at low feed values and (C) detail for local
model at high feed values.

coefcients obtained using the different approaches in milling

processes. As the cutting tool is the same for both operations,
turning and milling, it is assumed that the behaviour in milling
should be similar to that observed in orthogonal turning in the
previous section.
The objective is to simulate the peak forces in milling, being
these associated with high chip thickness, thus a good simulation
of these forces would require specic cutting coefcients obtained
from a local model in the zone of high chip thickness.

ARTICLE IN PRESS
O. Gonzalo et al. / International Journal of Machine Tools & Manufacture 50 (2010) 765774

771

Table 3
Values of the linear specic cutting coefcients for g 61 and vc 100 m/min.

Orthogonal Turning Cutting Forces

300
F t = 2875.7*f + 0.1575

Turning

Instantaneous
force based

Average force
based [5]

Turning

Milling

Milling

1746
41
730
39

1730
4
723
26

1639
14.3
852
15

Force/ap (N/mm)

250
200
2

150
F r = 2475.1*f - 8.343

Ktc (N/mm )
Kte (N/mm)
Krc (N/mm2)
Kre (N/mm)

100
Radial Force
Tangential Force
Linear fit

50
0
0.000

0.020

0.040

0.060

0.080

Table 4
Values of the linear specic cutting coefcients for g  101 and vc 100 m/min.

0.100

Feed (mm/rev)
Orthogonal Turning Cutting Forces
300

Ktc (N/mm )
Kte (N/mm)
Krc (N/mm2)
Kre (N/mm)

Force/ap (N)

250

Instantaneous
force based

Average force
based [5]

Turning

Milling

Milling

1979
79
1078
116

1987
35
1093
60

1512
54
1141
68

200
150
100
50
0
0.000

Radial Force
Tangential Force

0.020

0.040

0.060

0.080

0.100

Feed (mm/rev)
Orthogonal Turning Cutting Forces
300
F t = 1979.3*f + 79.405
250
Force/ap (N/mm)

Turning

200
F r = 1078.7*f + 116.79
150
100
Radial Force
Tangential Force
Linear

50

0
0.060 0.065 0.070 0.075 0.080 0.085 0.090 0.095 0.100 0.105
Feed (mm/rev)
Fig. 8. (A) Linear tting for a tool with negative rake angle: g  101; Vc 100 m/min;
ap 2 mm; (B) example of complex t function and (C) detail for local model at
high feed values.

The coefcients obtained using the instantaneous force

based method are calculated for feed values higher than
0.06 mm/rev while the average force based method calculates
the coefcients using the average force per revolution, involving
the whole range of chip thickness that results in a slotting milling
test [5].
Table 3 shows the values of the linear model coefcients
obtained using the different approaches for a tool with positive
rake angle at the cutting speed of 100 m/min.

The results show that there are minor differences between the
coefcients obtained with the instantaneous force based
method and the turning coefcients. The differences are  16
(  0.9%) in Ktc and  7 N/mm2 (  0.9%) in Krc.
The coefcients obtained with the classical average force based
method are different from those obtained by turning. The
differences are 107 (6%) in Ktc and +122 N/mm2 (+16.7%) in Krc.
Observing the results of Table 3, the instantaneous force
based method ts better than those from the classical average
force based method, both compared to the turning coefcients.
For the material used in this work, the friction coefcients are
less important than the shearing coefcients as can be noticed for
the absolute values. The differences observed in the friction
coefcients Kie, compared to the turning coefcients, could be a
consequence of the different chip type formed in milling and
turning. This could be explained by differences in the energy
necessary for the deformation of the chip (longer chips in
turning), and it is believed that this does not affect the shearing
during chip formation.
Table 4 shows the values of the linear model coefcients
obtained using the different approaches for a tool with negative
rake angle at the cutting speed of 100 m/min.
The results show that there are minor differences between
the Kic coefcients obtained with the instantaneous force
based method and the turning coefcients. The difference is
+8 N/mm2 ( +0.4%)in Ktc and +15 N/mm ( + 1.39%) in Krc. While the
 44 N/mm2
Kie coefcients shows higher differences,
(  55.70%)in Kte and 56 N/mm ( 48.28%) in Kre.
The coefcients obtained with the classical average force
based method are different from those obtained by turning.
Regarding the Kic coefcients, the differences are  467 N/mm2
(  23.6%) in Ktc and + 63 N/mm (+ 5.84%) in Krc. While the Kie
coefcients shows differences similar to those obtained with the
new method,  25 N/mm2 ( 31.65%) in Kte and  48 N/mm
(  41.38%) in Kre.
Observing the results of Table 4, the instantaneous force based
method also ts better than those from the classical average force
based method, both compared to the turning coefcients.
As in the positive rake angle tools, the signicant differences
observed in the friction coefcients Kie could be a consequence of
the different chip type formed in milling and turning.

ARTICLE IN PRESS
772

O. Gonzalo et al. / International Journal of Machine Tools & Manufacture 50 (2010) 765774

The good correlation of the shearing specic cutting coefcients (Kic) obtained using the instantaneous force based
method indicates that the shear action is correctly captured by
the new method, i.e. the actions depending on the undeformed
chip section (hn  b). This also reinforces the assumption that the
friction specic cutting coefcients (Kie) includes more effects
than the friction, rubbing and ploughing in the tool, such as the
difference in the chip shape and process kinematics.

5.4. Milling force simulation

In this section, the comparison of the milling forces for several
slot milling operations is presented. The analysis includes the
simulation of the cutting forces using different specic cutting
coefcients:

 Instantaneous force based with linear model.

 Average force based with linear model.

The results are compared to the experimental cutting forces

recorded in the milling test.
Figs. 9 and 10 show the simulation forces compared to the
experimental measured values for a tool with positive rake angle
at 100 m/min, at 0.060 and 0.090 mm/tooth, respectively.
The results show that any of the coefcients groups obtained
t well the experimental forces, with peak forces slightly higher
when using the coefcients obtained from the average force
based method. However, the difference is below 5%.
Figs. 11 and 12 show the simulation forces compared to the
experimental measured values for a tool with negative rake angle at
100 m/min, at 0.060 and 0.090 mm/tooth, respectively. As the same
way as in the positive angle case, the coefcients obtained t well
the experimental forces in addition to determining slightly higher
peak force values compared to average force based method.
As can be noted in Figs. 11 and 12, the simulated forces do not
match the experimental forces through all the cut, especially at
low chip thickness cutting conditions (entrance and exit of the
tool). These deviations are the result of the mist of the linear
force model over the whole feed range as explained in Section 5.2.

Fig. 9. Experimental validation of the different coefcients groups and force models. g 61; Vc 100 m/min; fz 0.060 mm/tooth and ap 2 mm.

Fig. 10. Experimental validation of the different coefcients groups and force models. g 61; Vc 100 m/min; fz 0.090 mm/tooth and ap 2 mm.

ARTICLE IN PRESS
O. Gonzalo et al. / International Journal of Machine Tools & Manufacture 50 (2010) 765774

773

Fig. 11. Experimental validation of the different coefcients groups and force models. g  101; Vc 100 m/min; fz 0.060 mm/tooth and ap 2 mm.

Fig. 12. Experimental validation of the different coefcients groups and force models. g  101; Vc 100 m/min; fz 0.090 mm/tooth and ap 2 mm.

However, since the coefcients obtained by the instantaneous

force based method is calculated to simulate the peak force, the
shape of the simulated cutting forces t properly the experimental forces obtained under high uncut chip thickness values.
On the other hand, the coefcients obtained from average
force based method ts the experimental forces similar to the
instantaneous force based method despite the differences in
the values of the coefcients between turning and average
force based, depicted in Tables 3 and 4. Therefore, it can be
said that using average force based coefcients, a good
mathematical t is achieved. But the physical explanation of the
t is not clear since the coefcients differ considerably from
the turning ones.
On balance, it is observed that there is a better correlation of
the obtained shearing turning cutting coefcients (Kic) with
those obtained by the instantaneous force based method than
those determined by average force based method.

6. Conclusions
In general terms, it is assumed that the shear coefcients (Kic)
cover the plastic deformation taking place in the shear plane during
the chip formation; while the friction coefcients (Kie) cover other
effects such as friction, rubbing, ploughing, chip deformation [8].
Thus the force behaviour is similar to the one showed in Fig. 13.
The specic cutting coefcients, obtained from turning and
instantaneous force based method in milling, show similar Kic
and different Kie values. Therefore, taking into account that the
same tool, cutting conditions and tool material are used, the
differences come mainly from process kinematics and chip type,
affecting the value of Kie. As pointed before, this behaviour is
believed to happen due to the differences in the chip shape and
deformation between turning and milling processes.
The instantaneous force based method proposed for obtaining
the specic cutting coefcients proves to be more efcient than the

ARTICLE IN PRESS
774

O. Gonzalo et al. / International Journal of Machine Tools & Manufacture 50 (2010) 765774

The authors thank Jon Etxarri and Ion Bengoetxea their help in
the experimental tests in the workshop, and to Jon Iturbe for
helping to set up the least squares tting method.

References

Fig. 13. Mechanisms contributing to the cutting forces.

classical average force based method, especially in order to

determine the Kic shear coefcients. Besides, it can reproduce in a
more physical way the variation of the cutting forces with the chip
thickness.
The biggest advantage of the instantaneous force based
method is noticed for tools with negative rake angle, improving
clearly the results obtained with the average force based
method in the calculation of specic cutting coefcients.
The classical average force based method drives to acceptable results in the simulated milling forces, but the suitability of
the obtained coefcients is difcult to explain attending to the
behaviour observed in the turning process. Therefore, the
average force based method leads to a mathematical t more
than to results with physical explanation.
The linear cutting force model is too simple to effectively
simulate the real behaviour of the cutting forces as a function of
the chip thickness.
The main advantages of the instantaneous force based
method compared to the average force based method are:

 The obtained shear specic cutting coefcients (Kic) are in

agreement with those obtained from turning.

 A single test is enough to obtain the specic cutting coefcients.

 Any type of milling operation can be used to obtain the specic
cutting coefcients, not only slot milling operations.
As a future work to complete this ongoing research, an attempt
will be made to extend the analysis to a multiple teeth tool with
helix angle including the run-out.

Acknowledgements
The authors thank the Basque Country Government for the
nancial support to the CIC MARGUNE project developed under
the ETORTEK program.

[1] J. Gradisek, M. Kalveram, K. Weinert, Mechanistic identication of specic

force coefcients for a general end mill, International Journal of Machine
Tools and Manufacture 44 (2004) 401414.
[2] C.A. Van Luttervelt, T.H.C. Childs, I.S. Jawahir, F. Klocke, P.K. Venuvinod,
Present situation and future trends in modeling of machining operations,
Annals of the CIRP 47 (1998) 587626.
[3] J. Mackerle, Finite-element analysis and simulation of machining: a
bibliography (19761996), Journal of Materials Processing Technology 86
(1999) 1744.
[4] B.E. Klamecki, Incipient chip formation in metal cuttinga three dimension
nite element analysis, Ph.D. Dissertation, University of Illinois at UrbanChampaign, 1973.
[5] Y. Altintas, in: Manufacturing Automation: Metal Cutting Mechanics Machine
Tool Vibrations and CNC DesignCambridge University Press, United Kingdom,
2000.
[6] A. Lamikiz, L.N. Lopez de Lacalle, J.A. Sanchez, M.A. Salgado, Cutting force
estimation in sculptured surface milling, International Journal of Machine
Tools and Manufacture 44 (2004) 15111526.
[7] A.J.P. Sabberwal, Chip section and cutting force during milling operation,
Annals of the CIRP 10/1 (1961) (1961) 197203.
[8] E.J.A. Armarego, R.H. Brown, in: The Machining of MetalsPrentice-Hall, New
Jersey, 1969.
[9] O. Kienzle, H. Victor, Spezische schnittkrafte bei der metallbearbeitung,
Werkstoffteechnik und Machinenbau 47/5 (1957) (1957) 224225.
[10] E. Budak, Y. Altintas, Modeling and avoidance of static form errors in
peripheral milling of plates, International Journal of Machine Tools and
Manufacture 35 (1995) 459476.
[11] J.-J.J. Wang, C.M. Zheng, Identication of shearing and ploughing cutting
constants from average forces in ball-end milling, International Journal of
Machine Tools and Manufacture 42 (2002) 695705.
[12] E. Budak, Y. Altintas, E.J.A. Armarego, Prediction of milling force coefcients
from orthogonal cutting data, Journal of Manufacturing Science and
Engineering 118 (1996) 216224.
[13] A. Lamikiz, L.N. Lopez de Lacalle, J.A. Sanchez, U. Bravo, Calculation
of the specic cutting coefcients and geometrical aspects in sculptured
surface machining, Machining Science and Technology 9 (2005)
411436.
[14] A. Azeem, H. Feng, L. Wang, Simplied and efcient calibration of a
mechanistic cutting force model for ball-end milling, International Journal
of Machine Tools and Manufacture 44 (2004) 291298.
[15] W.A. Kline, R.E. DeVor, J.R. Lindberg, The prediction of cutting forces in end
milling with application to cornering cuts, International Journal of Machine
Tool Design and Research 22 (1982) 722.
[16] R.E. DeVor, W.A. Kline, W.J. Zdeblick, A mechanistic model for the force
system in milling with application to machining airframe structure,
Manufacturing Engineering Transactions 8 (1980) 297303.
[17] L.M. Kumanchik, T.L. Schmitz, Improved analytical chip thickness model for
milling, Precision Engineering 31 (2007) 317324.
[18] M.E. Martelloti, An analysis of the milling process, Transactions of the ASME
63 (1941) 677700.
[19] A. Tarantola, in: Inverse Problem Theory and Methods for Model Parameter
EstimationSIAM, Philadelphia, 2005.
[20] CUTDATATM, Techsolve Inc., USA, 2009.
[21] O. Gonzalo, H. Jauregi, L.G. Uriarte, L.N. Lopez de Lacalle, Prediction of specic
force coefcients from a FEM cutting model, International Journal of
Advanced Manufacturing Technology 43 (2009) 348356.