ARTICLE IN PRESS

International Journal of Machine Tools & Manufacture 48 (2008) 644–655

www.elsevier.com/locate/ijmactool

A numerical simulator to predict the dynamical behavior of the

self-vibratory drilling head
Nicolas Guiberta, Henri Parisa,, Joël Rechb
a
G-SCOP Laboratory, University of Grenoble, 46 avenue Felix Viallet, 38031 Grenoble Cedex 1, France
b
LTDS, ENISE, 58 rue Jean Parot, 42000 Saint Etienne, France
Received 6 July 2007; received in revised form 6 November 2007; accepted 9 November 2007
Available online 22 November 2007

Abstract

The manufacturing of deep holes has to face problems to evacuate chips, especially for small diameters. Such problems induce frequent
tool breakage and poor surface quality. The vibratory drilling enables the chip to be split into small elements thanks to the axial
vibrations of the drill, self-maintained by the cutting energy. Thus, chips are easily evacuated from the hole. A specific tool holder with an
adapted axial stiffness has been developed in order to investigate this drilling process. The cutting conditions are predetermined in order
to lead to axial vibrations with a stable frequency and amplitude. During a period, the amplitude of the vibrations is higher than the feed
per revolution, which enables the cutting edges to jump out of the work material. The vibrations are self-maintained and remain stable if
some disturbances are absent or very limited such as the friction of the drill against the work material along its margins, the ploughing
force induced by the chisel edge, the ploughing force induced by clearance face, etc. The objectives of this paper are (i) to model the
dynamical behavior of the self-vibrating drilling head, the cutting and ploughing forces, and the material removal, (ii) to foresee with a
numerical simulator the cutting conditions which generate good vibrations, (iii) to validate the numerical simulator with a experimental
round of test. This work has also shown that the productivity of the drilling is improved by the use of the vibratory drilling. Deep hole
(ratio deep/drill diameter 420) can be drilled with this new technology without any coolant and any retreat cycle with the same quality as
a conventional drilling operation.
r 2007 Elsevier Ltd. All rights reserved.

Keywords: Drilling; Self-excited vibration; Ploughing; Chatter; Numerical simulation

1. Introduction
The deep hole drilling operation remains an unsatisfac-
tory technology for industry since its productivity is rather
limited. The main limit for the increase of productivity is
directly related to the poor chip evacuation. One solution
consists in the application of internal high pressure
lubrication to push chips out of the hole. This solution
consumes a lot of energy and necessitates an additional
equipment that is very costly. Moreover, the use of
lubricant is very detrimental from the environmental point
of view. Another solution, so called ‘‘retreat cycles’’, leads
to stopping the drilling operation and to remove fast the
drill from the cutting area and to return back to the drilling
Corresponding author. Tel.: +33 4 76 82 89 31; fax: +33 4 76 82 70 43.
E-mail address: henri.paris@ujf-grenoble.fr (H. Paris).
area and so periodically when drill’s flutes are full. This
second solution is very detrimental to the productivity.
New techniques of machining have emerged over the last
few years to improve chip evacuation. One of these
techniques is based on chip fragmentation due to an axial
vibration of the drill. Various solutions have been
developed in this area, depending on the technical solutions
used to generate the vibrations and depending on the
frequency/amplitude targeted. It is possible to distinguish
the self-vibratory drilling technology [1] and the vibration-
assisted drilling. Among the vibration-assisted drilling
technologies, some use low frequency (up to 1 kHz)
[2–10] and others use ultrasonic-assisted drilling (from 1
to 40 kHz) [11–15].
Among the vibration-assisted technologies, several tech-
nical solutions exist. Gouskov [2] evaluated the amplitude
and the frequency of the axial movement of a twist drill to

0890-6955/$ - see front matter r 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijmachtools.2007.11.003
 ARTICLE IN PRESS
N. Guibert et al. / International Journal of Machine Tools & Manufacture 48 (2008) 644–655
Nomenclature
a feed for one revolution (m)
Ac chip area ðm2 Þ
Ap area of ploughing ðm2 Þ
c damping
d coefficient
F thrust force (N)
Fz axial force created by the material removal in
zone 1 (N)
F 1c ; F 2c ; F 3c cutting force in zones 1, 2 and 3 (N)
Fp axial force created by the ploughing (N)
h instantaneous uncut chip thickness (m)
k stiffness (N/m)
K 1t specific cutting coefficient for zone 1 ðN=m2 Þ
2
Kt specific cutting coefficient for zone 2 ðN=m2 Þ
3
Kt specific cutting coefficient for the indentation
zone ðN=m2 Þ
split the chip. Suciu [3] proposed a tool holder in which the
axial vibrations were generated by variation of an oil film
between two trays, one of them being in rotation. Chabra
[4] suggested creating these axial vibrations by means of
linear engines, which equip more and more machine tools
today. Toews [7] uses piezolectric system to induce the axial
movement. However, these devices require an external
power supply on the tool holder that is animated by the
movement of rotation. Such systems are difficult to install
on a standard machine tool involved in production among
other cutting technologies.
Another approach consists in using the cutting energy in
order to create natural axial vibrations necessary to split
the chip [1,16]. This technology does not need an external
adjunction of energy. The regenerative vibrations naturally
appear at certain revolution frequencies. The challenge is
to stabilize them at a suitable frequency and magnitude for
a good chip fragmentation. Indeed, the axial vibrations at
low frequency have an amplitude larger than the feed rate.
This enables the interruption of the cutting process, leading
to small chips. The centrifugal force, combined with the
helical flute, enables to evacuate chips easily without any
internal coolant.
This technology requires the development of a self-
vibrating drilling head (SVDH). The SVDH (Fig. 1) is
based on the natural vibration of a solid (m: drill-holder)
combined with a spring (k) submitted to the thrust force of
the drill. Vibrations appear for adequate cutting para-
meters by using the low stiffness of the spring located
between the body and the drill-holder. This low rigidity
part creates conditions for controlled axial regenerative
vibrations. To ensure a good quality of the vibratory
drilling, the variation of thrust force must be controlled.
Thrust force depends on the geometry of the drill and
especially of its chisel edge. The principal limit of this
technology is mainly related to the damping of the
645
Kp specific cutting coefficient for ploughing
ðN=m3 Þ
Kr coefficient
Kv coefficient
N number of sub-division of the cutting edge
Nc number of cutting lip on the drill
q1 ; q2 ; q3 coefficient for the cutting model
R radius of the drill (m)
R2 radius of zone 2 (m)
R3 radius of indentation zone (m)
V cutting speed in a determined point (m/min)
V zmax feed speed of the drill (m/min)
X 0; Y 0; Z0 estimation of the axial position, speed and
acceleration of the drill
g rake angle (degree)
a clearance angle (degree)
d point angle (degree)
vibrations (c) due to several parameters such as ploughing
effect and friction of drill’s margins on the hole.
A numerical simulation of the system makes it possible
to predict the vibratory domain of the drilling operation.
The global model requires three types of input model: a
vibratory model of the SVDH, a thrust force model
bringing the energy to the system, and a spatiotemporal
model predicting the position of the drill with regard to the
hole at any time.
Section 1 will describe the dynamic model of the SVDH.
Concerning the thrust force model, it depends on the
tool design (geometry, material), on the work material, on
the instantaneous uncut chip thickness and on the cutting
speed. Standard thrust force models coming from the
literature are not relevant for such an application, since
they do not consider the ploughing effect or the extrusion
of the work material near to the chisel edge, which are
strong damping phenomena. They basically have a macro-
scopic view of the thrust force. The development of a more
comprehensive thrust force model considering these
elements is necessary. It is important to keep in mind that,
inside the theoretical instable domain, vibrations may not
Fig. 1. The self-vibrating drilling head.
 ARTICLE IN PRESS
646 N. Guibert et al. / International Journal of Machine Tools & Manufacture 48 (2008) 644–655
exist in practice. Indeed, if the amplitude of the vibrations
is smaller than the feed rate, the tool will not jump out of
the material and the vibrations will be damped. Moreover,
if the clearance face comes in contact with the work
material, vibrations will also be damped. So a spatiotem-
poral model is necessary to predict the position of the drill
with regard to the hole at any time. All these models will be
described in Section 2 and will be introduced on a
numerical simulator specially developed for this applica-
tion developed in Section 3.
Finally in Section 4, a practical application of this global
model on an industrial context will show the validation of
the simulator and the efficiency of this self-vibratory
drilling technology.
2. Models of the self-vibrating drilling head
The global model of the vibratory drilling operation
requires three types of sub-model:
 A dynamic model of the SVDH.
 A model of the thrust forces F .
 A spatiotemporal model of the periodical material
removal process.
2.1. Dynamic model for the self-vibrating drilling head
The SVDH (Fig. 1) is composed of three main elements:
a drill-holder maintaining the drill, a body having an
interface with the spindle and a spring in between. The
drill-holder has got a mass m which can be modified easily
by adding metal rings with a calibrate mass. The spring is
obtained with a stacking of Belleville rings. Different
thicknesses of Belleville rings and different types of
mounted of the ring (series or parallel) between the body
and the drill-holder give the possibility to adjust the
stiffness of the SVDH. The SVDH can be adapted for
any kind of machine by changing the standard of the
interface.
For adequate cutting parameters, the SVDH generates
chatter due to the low stiffness of the spring located
between the body and the drill-holder. Elements constitut-
ing the machining system (machine-tool, work piece,
fixture and drill) have stiffness higher than the axial
stiffness of the SVDH. As a consequence, it is possible to
assume these elements as rigid. Furthermore, the spring of
the SVDH has a very big rigidity in twisting and a low
rigidity in compression. The dynamic behavior of the
SVDH can be modelled (Fig. 2) by Eq. (1) [17]. Two
parameters can be modified (the mass m and the stiffness of
the spring k) in order to change the vibratory conditions.
On the contrary, the damping coefficient c should be as low
as possible. As a consequence, the design of the SVDH and
its material should lead to low friction at the interfaces
between the drill-holder and the body.
m  x€ þ c  x_ þ k  x ¼ F ,
Fig. 2. Model of the SVDH.
with m: mass of the solid (drill-holder) in movement, k:
stiffness of the spring, c: damping. x corresponds to the
axial position of the drill due to the vibration motion, x_
corresponds to the axial speed and x€ corresponds to the
acceleration of the vibration’s motion.
2.2. Thrust force model
The thrust force induced by the drilling operation on the
SVDH provides the energy to create the vibrations. In
order to predetermine accurately the vibratory domain, the
thrust force model (F) should be identified carefully. This
model depends on several parameters: the work material,
the drill geometry and its material. Concerning its
geometry, it is well known that the cutting phenomena
are not identical along the cutting edge. Thus, the thrust
force model has to distinguish different zones.
2.2.1. The different cutting phenomenon
Drills are composed of two main parts: the principal
cutting edge and the chisel edge. It can be divided into three
different zones (Fig. 3) having different kinds of working
conditions [18]:
 The primary cutting edge, zone 1, which realizes the
majority of the chip formation by a material removal
phenomenon. In this zone, the local cutting angles
depend on the radius considered along the edge. The
variations of the angles have a high magnitude as
described in [19]. As a consequence, the cutting edge has
to be divided into small sub-divisions in order to predict
local forces. Globally, rake angles are positive in this
area.
 Two zones can be distinguished on the chisel edge:
 Zone 2: The secondary cutting edge which is able to
cut the work material. However, rake angles are
negative in this area, which leads to high thurst face
values.
 Zone 3: The center of the drill. The cutting edge is
almost absent and the local tangential velocity is close
(1) to zero. As a consequence, the work material is
 ARTICLE IN PRESS
N. Guibert et al. / International Journal of Machine Tools & Manufacture 48 (2008) 644–655 647

Fig. 3. Description of the various zones of a drill.

strongly plastically deformed by indentation and by

rotation.

For each zone, a cutting model is identified (F 1c , F 2c , F 3c ).

The total thrust force F represents the cumulative effect of
all forces:
F ¼ F 1c þ F 2c þ F 3c . (2)

2.2.2. Thrust model for the principal cutting edge

In this zone, the thrust force F 1c is the result of two
mechanical phenomena: the material removal phenomenon
which generates a cutting force F z and the ploughing effect
which generates a force F p :
Fig. 4. Representation of the chip area.
F 1c ¼ F z þ F p . (3)

2.2.2.1. Model for the cutting forces F z . The chip forma-

tion takes place in zone 1. The cutting edge penetrates into
the material and causes a plastic deformation leading to the
shearing of the metal and the formation of the chip. This
phenomenon produces the axial force F z . Its amplitude is a
function of the area Ac depending on the drill geometry
and on the uncut chip thickness h (Fig. 4). Yang [20]
modelled this component as
1
F z ¼ N c  K 1t  ðh  ðR  R2 ÞÞq , (4)
with K 1t : specific
cutting coefficient, N c : number of cutting
1
edges, q : a coefficient depending on the pair ‘‘tool-work
material’’.
K 1t depends on the work material, on the cutting speed
Fig. 5. Sub-division of the primary cutting zone.
V, on the tool geometry (rake angle g ) and on the uncut
chip thickness h. The cutting speed is a linear function of
the radius and of the rotary speed. According to
Bandyopadhyay [21], K 1t can be determined as follows: vary along the cutting edge. As a consequence, it is
necessary to divide the cutting zone 1 into N small
K 1t ¼ K v  ha  V b  ð1  sinðgÞÞd , (5)
subdivisions of width DRi (for example three divisions
with K v , a and b: coefficients depending on the pair ‘‘tool- in Fig. 5) in which these parameters can be considered
work material’’. as constant (Eq. (6)). One specific cutting coefficient
The coefficient is a constant for a defined cutting speed V K 1ti is associated to the corresponding sub-division, so
and a defined rake angle g. In drilling, these parameters that the cutting force for one lip in a sub-division can
 ARTICLE IN PRESS
648 N. Guibert et al. / International Journal of Machine Tools & Manufacture 48 (2008) 644–655
be expressed as:
1 ploughing forces of each sub-division:
DF zi ¼ K 1ti  ðh  DRi Þq .
The main cutting force F z can be directly calculated as:
X
N i¼1
F z ¼ Nc  DF zi .
i¼1
2.2.2.2. Model for the ploughing force F p . The ploughing
effect is a major problem in all machining operations. It
happens when the clearance face is in contact with the
machined surface. In conventional drilling, this problem is
rather limited, since it may happen only for high feed rates
and for drills having a low clearance angle. In vibratory
drilling, the axial vibration induces a higher instantaneous
axial speed. If the penetration angle is higher than the
orientation of the surface machined during the previous
revolution, then the contact between the clearance face and
the machined surface is possible. In vibratory drilling, the
ploughing effect can be obtained easily, which leads to an
immediate damping.
Due to the conical geometry of zone 1, the ploughing
does not take place on the entire cutting lip. Ploughing is a
function of the geometry of the generated surface (Fig. 6).
The geometry of the generated surface is the consequence
of the trajectory of the cutting lip. The profile of the surface
is the function of the radius. The lead angle of the surface is
more important for weak radius ðR2 Þ than for big radius
(R). So ploughing can take place only in some parts of the
clearance face. The ploughing force can be dissociated into
various components (width of the components: DRi ) along
the cutting edge as previously for the thrust force (Eq. (6)).
Wu [22] modelled the ploughing force F p as a force
proportional to the area of the theoretical displaced
material Ap on each sub-division (the area is function of
the radius). The utilization of the discretization taken
previously gives the expression of the ploughing force for
one subdivision:
DF pi ¼ K p  Api  DRi .
Fig. 6. Representation of the ploughing area.
The total ploughing force will then be the sum of all the
(6)
X
N
F p ¼ Nc  DF pi . (9)
(7) 2.2.3. Thrust model for the secondary cutting edge
The cutting in zone 2 differs from the cutting in zone 1,
because of the low tangential speed. Moreover, the rake
angle is usually negative instead of strongly positive as
shown by Claudin [19]. Since Yang [20], the thrust force
can be modelled by Eq. (10) with a couple of coefficients K 2t
and q2 . Due to the small size of this zone (about 101
of the
radius and function of the type of grinding of the tool), the
variation along the edge can be neglected. So, it is not
necessary to split this edge into several sub-divisions:
2
F 2c ¼ N c  K 2t  ðh  ðR2  R3 ÞÞq . (10)
The ploughing exists also in this zone and does not vary
along this zone.
2.2.4. Thrust model for the indentation zone
In zone 3, the cutting speed can be considered as null. In
that case, the cutting phenomenon is not any more a chip
formation phenomenon but an extrusion phenomenon.
The cutting model for this zone cannot be the same as the
one in zone 1 or 2. For Yang [20], the extrusion does not
take place on a constant zone. It appears on an area which
has a variable radius R3 , function of the uncut chip
thickness and the point angle d of the tool (Eq. (11)) and
the drill sharpening. Based on Yang [20] a model of thrust
force for this phenomenon of extrusion can be established
(Eq. (12)):
 
Kr  h
R3 ¼ , (11)
2 tanððp  dÞ=2Þ
with K r : coefficient function of the drill sharpening;
 q3
K r  h2
(8) F 3c ¼ K 3t  , (12)
2 tanððp  dÞ=2Þ
 ARTICLE IN PRESS
N. Guibert et al. / International Journal of Machine Tools & Manufacture 48 (2008) 644–655
with K 3t : specific cutting coefficient for a pair ‘‘tool-work
material’’.
2.3. Spatiotemporal model
In a conventional drilling operation, the position of a
cutting edge can easily be determined at any time and the
uncut chip thickness is constant. On the contrary, in
vibratory drilling, the axial vibration superimposed to the
feed of the drill necessitates to calculate this position. The
position depends on the interaction between the thrust
force model and the dynamic model. At any time, the
evolution of the position depends on its previous revolu-
tion and on the previous revolution of the second cutting
edge.
In order to model the machined surface, one tool
revolution can be divided into P steps. For the model,
the cutting angle is considered as null because it has no
influence on the geometry of the surface if the twisting is
not taken into account. The evaluation of the uncut chip
thickness is geometrical (Figs. 7 and 8). At each step P, the
distance between the cutting lip (point A) and the previous
machined surface (surface machined by the previous
cutting lip, point B) is evaluted. The uncut chip thickness
is function of the relative positions of points A and B. For
each step P, two situations are possible:
Case 1: (Fig. 7): the current position of the cutting edge
(point A) is under the position of the previous machined
surface (point B). In that case, the tool is on the material
and the uncut chip thickness h is equal to the distance D
between the lip and the surface (distance between A and B).
Fig. 7. Material removal model, case 1: the drill removes material.
Fig. 8. Material removal model, case 2: the drill is outside the material.
649
After the machining, the new machined surface will pass
through point A.
Case 2: (Fig. 8): the current position of the cutting edge
is over the position of the previous machined surface. The
tool does not enter into the material, the uncut chip
thickness h is zero. The position of the machined surface
remains unchanged (point B).
3. Simulation of the vibratory drilling
3.1. Integration of the dynamical equation
The main difficulty of the vibratory drilling is to foresee
the cutting conditions that will generate regular vibrations
able to induce interrupted cutting. A simulator has been
designed on Matlab in order to predict the shape of the
machined surface, the instantaneous uncut chip thickness,
the instantaneous thrust force, and the amplitude of
vibration. It should also predict the existence of the
ploughing effect leading to damping and to a restriction
of the acceptable cutting conditions. The simulator is based
on the calculation of the position of the tool, at each step
P. The position is determined by the resolution of the
differential equation (1) at each step.
To resolve the equation, an integration method of
Newmark [23] is applied (Fig. 9):
 For each increment, a position of the drill X 0, its axial
speed V 0 (expressed by x_ in Eq. (1)) and its axial
acceleration Z0 (expressed by x€ in Eq. (1)) are predicted
Fig. 9. Newmark’s integration algorithm.
 ARTICLE IN PRESS
650 N. Guibert et al. / International Journal of Machine Tools & Manufacture 48 (2008) 644–655
according to the linear acceleration method of New-
mark.
 The uncut chip thickness is evaluated with the material
removal model and the resulting cutting force is
calculated with the models of the previous section.
 Because of the non-linearity of the cutting force
equation, a Newton algorithm ensures the evaluation
of an optimized solution of the non-linear equation (1)
at each increment.
All the models described in this article are integrated in the
simulator. The simulator accepts a sub-division of zone 1
of the drill from 1 to 10 parts (Fig. 5). A specific cutting
force model with its specific coefficients can be associated
to each part.
3.2. Example of application
As an example, the drill considered in this paper is a
carbide drill with a 5 mm diameter with a split point chisel
edge geometry (Fig. 10). The radius of the secondary
cutting zone has been measured at R2 ¼ 0:7 mm. The
primary cutting zone has been divided into two elements.
The first element goes from the end of the secondary
cutting zone R2 to the radius R21 ¼ 1:4 mm. The second
element goes from R21 to R11 ¼ 2:5 mm. To determine the
different coefficients of the models, an experimental round
of test has been made. These experiments are based on the
drilling of pre-bored holes. Two pre-bored holes of
diameter R2 and R21 have been made. The recording of
thrust force during the drilling of the pre-bored holes and
the recording of a plain drilling make it possible to identify
the different coefficients reported in Table 1. This
Fig. 10. Geometry of the carbide drill.
Table 1
Coefficient for the cutting models
Zone 1 (two subdivisions)
K 1tð1Þ ðN=m2 Þ R11 (m) K 1tð2Þ ðN=m2 Þ R21 (m) q1
7:0E þ 17 2.5E-3 6:0E þ 14 1.4E-3 2.11
methodology does not allow to find the coefficients a, b,
d and K v of Eq. (5). They are included in the two
parameters K 1tð1Þ and K 1tð2Þ . Some researches are carried out
in the laboratory to analyze the evolution of the thrust
force along the cutting lip by recording the thrust force
during the penetration of the drill on the material. This
methodology will permit to identify all the coefficients of
the models.
The different zones and theirs coefficients are reported in
Table 1.
The following configuration of vibratory drilling has
been tested on the simulator with the following cutting
parameters: drill diameter of 5 mm, workplace: crankshaft
steel 35MnV7 (280 Hb), cutting speed of 95 m/min, feed
rate of 0.05 mm/rev and the vibratory drilling head
parameters: SVDH mobile mass: 2.45 kg and SVDH
stiffness of 260 N/mm. The simulator provides some results
about the amplitude of the vibrations, the uncut chip
thickness, the total thrust force, and the thrust force
generated by the different zones of the drill (Fig. 11). The
vibrations are stable during the entire drilling. The split of
the chip is very good (the uncut chip thickness goes to zero
several times in each turn). The thrust forces vary on the
different zones. The chisel edge (zones 2 and 3) is
responsible for more than 70% of the total thrust force
which is coherent with the scientific literature [19]. The
influence of the ploughing effect on the total thrust force is
quite important (above 15% of the total thrust force). The
ploughing effect makes it possible to stabilize the vibra-
tions. Without any ploughing the vibration’s amplitude
would increase and become divergent. On the contrary, a
too large ploughing effect can lead to the total damping of
the vibrations [24]. The ploughing effect is different on each
zone of the drill (above 50 N in zone 1 and 25 N on zone 2)
because of the variation of the clearance angle and the
shape of the profile (Fig. 6).
4. Experimental approach of the vibratory drilling
4.1. Experimental protocol
The experiments have been performed on a CMC milling
machine (Fig. 12). The workpiece is a crankshaft steel
(35MnV7, 280 Hb). The thrust forces are recorded by a
3-component piezoelectric sensor. A sensor measures the
axial oscillation between the mobile part and its body. The
experimental conditions are the same as the one taken on
Zone 2 Zone 3
K p ðN=m3 Þ K 2t ðN=m2 Þ q2 R2 (m) K 3t ðN=m2 Þ q3
5:0E þ 8 1:0E þ 7 0.63 7.0E-4 9:0E þ 7 0.75
 ARTICLE IN PRESS
N. Guibert et al. / International Journal of Machine Tools & Manufacture 48 (2008) 644–655 651

Fig. 11. Simulation for a drilling of 5 mm of diameter, cutting speed: 95 m/min, feed rate: 0.05 mm/rev, mass: 2.45 kg, stiffness: 260 N/mm, zone 1:
principal cutting edge, zone 2: secondary cutting zone ðR2 ¼ 0:7 mmÞ, zone 3: indentation zone (variable radius (Eq. (11))).

Fig. 12. Experimental setup.

 ARTICLE IN PRESS
652 N. Guibert et al. / International Journal of Machine Tools & Manufacture 48 (2008) 644–655
the preceding simulation (Section 3). The drill used is a
carbide drill of 5 mm diameter with a split point grinding.
The hole is 90 mm deep. No coolant and no retreat cycle
have been used. The cutting speed and the feed rate are the
same as earlier on simulation (95 m/min and 0.05 mm/rev).
Fig. 13. Recording of the displacement of the vibratory drilling head
measured by the Eddy current sensor.
Fig. 14. Comparison of the experimental thrust force measured by the Kistler dynamometer and the simulated thrust force. (a) Experimental thrust force
and (b) simulated thrust force.
The drilling head parameters are also constant m ¼ 2:45 kg
and k ¼ 260 N=mm.
4.2. Experimental results and validation of the simulator
Fig. 13 presented the axial displacement between the
mobile part and the body of the SVDH. During the first
step, the spring is being compressed by the thrust force due
to the penetration of the drill in the material. After this
step, the cutting energy is enough to generate vibrations.
The variations are very regular in terms of amplitude and
frequency during the entire drilling.
In the stable phase, the amplitude of vibrations is about
120 mm, which is approximatively more than twice the feed
rate. The split of the chip is then possible. The amplitude of
the predicted vibrations (Fig. 11) was around 130 mm which
is very close.
A comparison of the simulation and the experiments is
made in Fig. 14. The first 20 revolutions of the drill are very
different in practice compared to the simulated results since
the simulator does not consider the entrance of the conical
geometry in the work material. The simulator assumes that
the drill starts immediately with a full contact. After this
transient period, the experimental results can be compared
to the simulated ones.
 ARTICLE IN PRESS
N. Guibert et al. / International Journal of Machine Tools & Manufacture 48 (2008) 644–655
The value of the measured thrust force (Fig. 14a)
oscillates between 0 (the drill is not in contact with the
material) and 750 N when the instantaneous uncut chip
thickness is maximum. The period is 5 ms corresponding
to 1.7 oscillations during one revolution. Chips are
then split into small pieces ð1:7 vibration per lip  2 lips ¼
3:4 chips per revolutionÞ and can be easily evacuated. The
vibrations of the tool are very stable, which confirms that
the evacuation of the chip is correct during the entire
operation. The simulated thrust force is represented in
Fig. 14b. The amplitude of the experimental thrust force is
something smaller (640 N) but the frequency is very close:
1.65 vibrations for one revolution.
A last comparison can be made on the cutting rate
parameter. The cutting rate is the ratio between the time
Fig. 15. Chips obtained in vibratory drilling during the experimental
round of test.
Fig. 16. Evolution of the measured and the simulated uncut chip thickness.
653
the lip removes material and the time of machining. For the
experiment, the cutting rate is 0.53. For the simulation, the
cutting rate is very close and is equal to 0.5. From the value
of the cutting rate and the vibration’s frequency, it is
possible to evaluate the length of the chip. In the
experiment, the vibration’s frequency is 1.7 vibrations per
revolution which leads to the formation of 3.4 chips per
revolution. With a cutting rate of 0.53, the chip represents a
material’s portion of 56 (360 for one revolution  the
cutting rate of 0.53/3.4 chips per revolution). Fig. 15 shows
the experimental chip obtained. The chip is spilt into small
pieces with an angle near from 56 (the angle cannot be
measured precisely because of the shape of the chip).
With the information of the axial motion of the tool
given by the displacement sensor, a reconstruction of the
trajectory in function of time of one point of the lip (for
example at the outer corner of the drill) is possible. The
trajectory of the first lip during three revolutions is
represented by the black curve (Fig. 16) for a point of
the cutting edge at radius R ¼ 2:5 mm. With the hypothesis
of a perfect tool, the trajectory of the second lip is identical
to the trajectory of the first lip but translated to 12
revolution. The previous trajectory of the first lip is
obtained by another translation of 12 revolution. The
superposition of the three curves allows finding
the instantaneous uncut chip thickness generated by the
machining (grey zone in Fig. 16). The experimental
maximum uncut chip thickness measured is 0.12 mm. The
results of the simulation are similar: the amplitude
simulated is equal to 0.11 mm (visible in Fig. 11).
 ARTICLE IN PRESS
654 N. Guibert et al. / International Journal of Machine Tools & Manufacture 48 (2008) 644–655
Table 2
Comparison between experimental results and simulations
Cutting speed Feed rate (mm/ Experimental thrust
(rev/min) rev) force (N)
8000 0.075 800
9000 0.075 750
7000 0.05 750
The agreement between the experimental and the
predicted thrust forces, the amplitude of the vibrations
and the uncut chip thickness is very good. The simulator is
able to predetermine the vibration for one configuration
(one mass, one stiffness, one tool, one work material, one
rotary speed, one feed). It is now necessary to investigate
several configurations in order to check if this model is still
efficient. Table 2 presents the comparison between experi-
ments and simulations for different cutting parameters,
with the following drilling head parameter and with the
same drill and the same work material: SVDH mobile
mass: 3.25 kg and SVDH stiffness: 550 N/mm. The error
made by the simulator is very weak (o15%).
5. Conclusion
The use of vibratory drilling for small diameter deep
holes allows improving the productivity by eliminating the
retreat cycles and the use of coolant. The main difficulty is
to choose the cutting parameters and the cutting head
parameters to obtain chatter compatible with the tool life
and the split of the chip. For that a behavior model of the
SVDH has been taken. In order to have good models, an
analysis of the different forces generated during the
operation is necessary. After an identification of different
coefficients of the model, these are included in a numerical
simulator. The simulator gives some precious information
about the trajectory of the lip, the uncut chip thickness and
the thrust forces on the different zones of the tool. An
experimental round of test validates the models taken. The
amplitude of the thrust force, the vibration, the uncut chip
thickness and the vibration’s frequency are very close
between the simulation and the experimentation. The
utilization of the numerical simulator to foresee the
vibratory drilling domain of a drill is possible.
The vibratory drilling leads to a very good fragmenta-
tion of the chip due to the discontinuity of the cutting
phenomenon. The drill is outside the material for more
than 50% of the time machining. The warming of the tool
is then less important than in traditional drilling. This
advantage combined with the good evacuation of the chip
increases the drilling productivity. It is then possible to
make deep holes (ratio deep/drill diameter 420) without
any coolant and any retreat cycle at a cutting speed larger
than the cutting speed recommended by the manufacturers.
With this technology, the time of machining for a 90 mm
deep hole is less than 8 s.
Simulated thrust Experimental Simulated amplitude
force (N) amplitude ðmmÞ ðmmÞ
925 0.15 0.16
700 0.12 0.13
650 0.13 0.11
The quality of the drill (diameter, deviation and rough-
ness) is similar to a standard drilling. A tool life experiment
has been made with this tool on vibratory drilling without
coolant and retreat cycle. The total length machining with
this tool and with the previous cutting and drilling head
parameters is over 25 m. This technology looks very
competitive. An optimization of the parameters of the
drilling head and the geometry of the tool should give
better results.
Acknowledgment
Thanks are due to CETIM who financed the experi-
mental part of this work.
References
[1] G. Peigne, E. Kamnev, D. Brissaud, A. Gouskov, Self-excited
vibratory drilling: a dimensionless parameter approach for guiding
experiments, Proceedings of the Institution of Mechanical Engineers
Part B—Journal of Engineering Manufacture 219 (2005) 73–85.
[2] A.M. Gouskov, S.A. Voronov, S. A Batzer, Chatter synchronization
in vibratory drilling, American Society of Mechanical Engineers 108
(68) (2000) 263–270.
[3] C. Suciu, Analytical investigation of an active-control squeeze film
damper, Rotrib’96, Editura Tehnica.
[4] P.N. Chabra, B. Ackroyd, W.D. Compton, Low frequency modula-
tion-assisted drilling using linear drives, Proceedings of the Institu-
tion of Mechanical Engineers Part B—Journal of Engineering
Manufacture 216 (2002) 321–330.
[5] L. Deyuan, W. Lijang, Investigation of chip in vibration drilling,
International Journal of Machine Tools & Manufacture 38 (3) (1999)
165–176.
[6] K. Okamura, H. Sasahara, T. Segawa, M. Tustsumi, Low frequency
vibration drilling of titanium alloy, JSME International Journal,
Series C 49 (1) (2006).
[7] H.G. Toews, W.D. Compton, S. Chandrasekar, A study of the
influence of superimposed low frequency modulation on the drilling
process, Precision Engineering 22 (1998) 1–9.
[8] X. Wang, L.J. Wang, J.P. Tao, On thrust in vibration drilling of fiber
reinforced plastics, Journal of Materials Processing Technology 148
(2004) 239–244.
[9] S. Aoki, T. Nishimura, Prevention of delimation during drilling of
composite material using vibration, Key Engineering Materials 261
(2004) 381–386.
[10] K. Adachi, N. Arai, S. Harada, K. Okita, S. Wakisada, A study on
burr in low frequency vibratory drilling–drilling of aluminium,
Bulletin of JSPE 21 (4) (1987) 258–264.
[11] K.I. Ishikawa, H. Suwabe, T. Hishide, M. Uneda, A study on
combined vibration drilling by ultrasonic, and low frequency
vibrations for hard and brittle materials, Precision Engineering 22
(1998) 196–205.
 ARTICLE IN PRESS
N. Guibert et al. / International Journal of Machine Tools & Manufacture 48 (2008) 644–655
[12] B. Azarhoushang, J. Akbari, Ultrasonic-assisted drilling of Incon-
el738-LC, Internatinal Journal of Machine Tools & Manufacture 47
(2006) 1027–1033.
[13] S.S.F. Chang, G.M. Bone, Burr size reduction in drilling by
ultrasonic assistance, Robotics and Computer-Integrated Manufac-
turing 21 (2005) 442–450.
[14] K. Onihura, O. Ohnishi, J.H. Feng, T. Morita, U. Bopp, Effects of
ultrasonic vibration on machining accuracy in microdrilling, Inter-
national Journal of JSPE 30 (3) (1996) 210–216.
[15] H. Takeyama, S. Kato, Burrless drilling by means of ultrasonic
vibration, Annals of CIRP 40 (1) (1991) 83–86.
[16] S. Tichkiewitch, G. Moraru, D. Brun-Picard, A. Gouskov, Self-
excited vibration drilling models and experiments, Annals of the
CIRP 51 (1) (2002) 311–314.
[17] H. Paris, S. Tichkiewitch, G. Peigne, Modelling the vibratory drilling
to foresee cutting parameters, Annals of the CIRP 54 (1) (2004).
[18] P. Anish, S.G. Kapoor, R.E. DeVor, A chisel edge model for
arbitrary drill point geometry, Journal of Manufacturing Science and
Engineering 127 (2005) 23–34.
655
[19] C. Claudin, G. Poulachon, M. Lambertin, J.J. Janosch, B. Vecchi, F.
Gig, Influence of minimum quantity of lubrification on drilling
process, in: 5th International Conference on High Speed Machining,
Metz, 2006.
[20] J.A. Yang, J. Venkatraman, D. Ruxu, A new dynamic model for
drilling and reaming process, Journal of Machine Tools &
Manufacture 42 (2001) 299–311.
[21] P. Bandyopadhyay, D. A Stephenson, Process-independent force
characterization for metal-cutting simulation, Transactions of the
ASME, Journal of Engineering for Industry 119 (1997) 86–94.
[22] D.W. Wu, A new approach of formulating the transfer function for
dynamic cutting process, ASME Journal of Engineering for Industry
111 (1989) 37–47.
[23] N.M. Newmark, A method of computation for structural dyna-
mics, ASCE Journal of Engineering Mechanics Division 85 (1959)
67–94.
[24] N. Guibert, H. Paris, Influence of the ploughing effect on the
vibratory drilling behavior, in: 5th International Conference on High
Speed Machining, Metz, 2006.