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

net/publication/358080815

Design optimization of ultrasonic vibration cutting tool to generate well-

decoupled elliptical trajectory

Article in The International Journal of Advanced Manufacturing Technology · April 2022

DOI: 10.1007/s00170-022-08745-9

CITATIONS READS

6 437

4 authors:

Jinchuan Yang Pingfa Feng

Tsinghua University Tsinghua University
3 PUBLICATIONS 7 CITATIONS 253 PUBLICATIONS 3,200 CITATIONS

SEE PROFILE SEE PROFILE

Jianfu Zhang Jianjian Wang

Tsinghua University Tsinghua University
187 PUBLICATIONS 2,705 CITATIONS 85 PUBLICATIONS 1,383 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Jinchuan Yang on 22 April 2022.

The user has requested enhancement of the downloaded file.

The International Journal of Advanced Manufacturing Technology
https://doi.org/10.1007/s00170-022-08745-9

ORIGINAL ARTICLE

Design optimization of ultrasonic vibration cutting tool to generate

well‑decoupled elliptical trajectory
Jinchuan Yang1 · Pingfa Feng1,2 · Jianfu Zhang1,3 · Jianjian Wang1,3

Received: 30 October 2021 / Accepted: 13 January 2022

© The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2022

Abstract
Ultrasonic elliptical vibration cutting (UEVC) is a superior machining method for difficult-to-cut materials. The shape of
the elliptical tool trajectory crucially affects the integrity of the machined surface. However, the existing designs of UEVC
tool typically suffer from difficult motion decoupling, resulting in the low controllability of elliptical trajectory. This study
presents a new optimization design method of UEVC tool with dual longitudinal generators to create well-decoupled elliptical
trajectory. A theoretical model which establishes the effects of key design variables on the tool resonant characteristics was
adopted to optimize the configuration angle (the included angle of two longitudinal generators) as 90°. The structural param-
eters of the connection blocks of the two generators were also optimized to achieve resonance matching in both the normal
(depth-of-cut) and tangential (cutting) directions. The optimized tool has a stable resonant frequency of 18 kHz, a working
space of 5.1 μm × 5.3 μm, and a small trajectory error of less than 0.75 μm. The vibration measurement results showed that
the optimized design can enable the high controllability of tool trajectory. Grooving experiments were conducted to verify
the improved cutting performance of the designed tool due to the excellent control of tool trajectory.

Keywords Ultrasonic elliptical vibration · Horn · Design · Dual excitations

1 Introduction range [7]. The amplitude and phase difference of elliptical

vibration trajectory have different effects on the surface
Ultrasonic elliptical vibration cutting (UEVC) is a special microstructure and roughness of workpiece separately [8,
machining technology that applies ultrasonic elliptical vibra- 9]. Therefore, the adjustability of elliptical vibration trajec-
tion to the tool to enhance the machining performance, as tory has become one of the important indexes for the per-
shown in Fig. 1a [1, 2]. The elliptical trajectory is a crucial formance evaluation of ultrasonic elliptical vibration cutting
processing parameter in UEVC [3–5]. As shown in Fig. 1b, device. The accuracy of trajectory adjustment is also condu-
the orientation of the elliptical trajectory plays a critical role cive to predict and analyze the surface formation mechanism
in generating a high-quality surface of brittle materials [6]. and surface integrity.
There exists an apparent correlation between the pose angle Early types of ultrasonic elliptical vibration generators
of elliptical vibration trajectory and machining quality. The have been developed for UEVC [10, 11]. However, they all
cutting resistance can be reduced under a certain pose angle have their own limitations in vibration trajectory adjust-
ment. The novel vibration generators with single excita-
tion are often used to generate elliptical vibration trajec-
* Jianjian Wang tory relying on special structures. When the structure is
[email protected]
finalized, under the action of single excitation vibration,
1
Department of Mechanical Engineering, Tsinghua the shape of elliptical trajectory is determined finally, but
University, Beijing 100084, China its pose and inclination angle cannot be adjusted [12–14].
2
Division of Advanced Manufacturing, Tsinghua Shenzhen The UEVC advices in 3-DOF vibration state can produce
International Graduate School, Tsinghua University, elliptical vibration in three-dimensional space. But their
Shenzhen 518029, China adjustment ranges of vibration trajectory are small and their
3
State Key Laboratory of Tribology, Department vibration decoupling processes are complex and trajectory
of Mechanical Engineering, Tsinghua University, adjustments are also difficult in multi degree of freedom
Beijing 100084, China

13
Vol.:(0123456789)
 The International Journal of Advanced Manufacturing Technology

Fig. 1  Ultrasonic elliptical

vibration cutting for (a) surface
machining and (b) the influence
of the elliptical trajectory on
surface state after cutting [6]

state [15, 16]. In the classical types of UEVC tools, two The complex structures and dual-excitation modes com-
ultrasonic power sources with a phase difference are used to plicate the design of the vibration system and increase the
excite ultrasonic vibrations for creating the elliptical vibra- difficulty of vibration decoupling [25–27]. At present, the
tion trajectory based on different vibration modes [17–19]. configuration angle θ of reliable UEVC tools with double
One of the most widely dual-excitation ultrasonic elliptical longitudinal vibrations is less than or equal to 90° and lacks
vibration cutting devices is developed by Guo et al. using a accurate decoupling analysis [28–31]. Some studies illus-
certain configuration angle of 60° (see in Fig. 2). Its ellipse trate that the vibration decoupling and elliptical shapes of
shape can be adjusted in a certain range and has been widely different vibration modes are affected by the configuration
used for various applications, including micro-groove turn- structure and vibration transmission paths of the tools [32,
ing and structural coloration of metallic surfaces [20–23]. 33]. The equivalent stiffness and damping of the system in
However, it also does not have a function to accurately the connection transition region increase the difficulty of
adjust the ellipse pose and inclination angle [24]. decoupling analysis [34, 35]. In addition, the resonant modes

Fig. 2  Widely applied ultrasonic elliptical vibration cutting tool with (a) a configuration angle of 60° and (b) the ellipse shapes at different phase
inputs [20]

13
The International Journal of Advanced Manufacturing Technology

of the two vibration generators are coupled, and the overall

resonant frequency of the coupled system is significantly
affected. This phenomenon can result in a large discrepancy
between the resonant frequencies of the two coupled vibra-
tion modes [36, 37]. These factors make it difficult to find
suitable design variables for the optimization of a tool with
two compound vibration modes and ensure an optimal ellip-
tical vibration trajectory. Therefore, it is still challenging
to achieve accurate vibration decoupling and realize target
vibration performance in the design of the UEVC device.
In this study, a new decoupling method of elliptical vibra-
tion under dual-excitation vibration state is established. This
decoupling process can characterize the mapping relation- Fig. 3  Design of the ultrasonic elliptical vibration cutting tool with an
ship between vibration coupling and configuration angle arbitrary configuration angle
accurately. In addition, a new design principle for resonant
frequency matching of dual excitation generators is also pro- 2-degree-of-freedom system in the physical coordinate system
posed. The appropriate design variables and configuration of the tool, the equivalent vibration system is shown in Fig. 4.
angle are optimized. Finally, cutting tests are conducted to O is the starting position of the vibration in the fixed nodal
verify the adjusting performance of elliptical vibration locus. plane; P1/2 is the vibration output end of the connection block;
and Q is the vibration output end of the end effector. k and c
2 Vibration decoupling model are the equivalent stiffness and damping of the connection col-
of the dual‑excitation elliptical vibration umn section, respectively; k1 and c1 are the equivalent stiffness
tool and damping of the circular flexure hinge, respectively; kX and
cX are the equivalent stiffness and damping of the interaction
2.1 Modeling the dynamic characteristics of horn 1 and horn 2 in the X-direction, respectively; and kY
of the vibration tool and cY are the equivalent stiffness and damping of the inter-
action of horn 1 and horn 2 in the Y-direction, respectively.
A vibration model of the tool with an arbitrary configura- Moreover, kX/Y and cX/Y change with the angle θ and are set as
tion angle is established for the universal coupling analysis. functions of angle θ. As shown in Fig. 4, the mass m is equiva-
Two piezoelectric transducers are located on the outside lent to the total mass of the insert and the end effector, and the
of the fixed nodal plane, and the connection block and the equivalent mass M consists of the mass M1 of the connection
connected column sections of the horns are installed on column section and the mass M2 of the connection block.
the inside of the fixed nodal plane. The angle between the The steady-state responses of the longitudinal vibration
two central axes of the two generators is θ. The composite F1(t) and F2(t) at point P1/2 can be derived as:
normal vibration direction is the same as the cutting depth {
direction, and the composite tangential vibration direction is X1 (t) = A1 sin(𝜔t + 𝜑 + 𝛼)
𝜃1 (t) = A1 sin(𝜔t + 𝛼) (1)
the same as the cutting direction in the workpiece coordinate
system. X and Y are the axial and vertical directions of the
physical coordinate system and are used for vibration cou-
pling and motion decomposition analysis, respectively. ­XC
and ­YD are the cutting direction and cutting depth direction,
respectively, in the workpiece coordinate system and are
used for analyzing the motion trajectory, as shown in Fig. 3.
The vibration modeling can be performed by decomposing
the vibration excitation of horn 2, whereas horn 1 represents
the reference. The vibration modeling in the physical coordi-
nate system only requires vibration decomposition and cou-
pling analysis in the X- and Y-directions for the transmission
of vibration input 2. The vibration input 1 does not have to be
decomposed because it is equivalent to the longitudinal vibra-
tion in the X-direction. In the workpiece coordinate system,
it is necessary to analyze the change in the elliptical vibration Fig. 4  Model of the vibration system with the arbitrary configuration
trajectory at the tooltip. According to the vibration theory of a angle θ

13
 The International Journal of Advanced Manufacturing Technology

where
√
A[ (k)2 + (c𝜔)2 ]/k
A1 = �
� �2 � �2
2
1 − ( 𝜔𝜔 ) + 2𝜉( 𝜔𝜔 )
n n

and √α =  − arctan(− cω/k) + arctan((− c 1 ω)/(k − Mω 2 )),

ωn = k∕M , ξ = c/cn, cn = 2Mωn.
When X1(t) and θ1(t) are transferred to the output end Q of
the end effector, the two equivalent vibration excitations in the
Fig. 5  The two ideal resonant modes and their elliptical vibration locus
X- and Y-directions can be defined as [38]:

⎧ F (t) = k [X (t) − x] + c [Ẋ (t) − x] where

̇ + k1 [𝜃1 (t)cos𝜃XY − x] √
⎪ X 1 1
̇
1 1
⎨ +c [𝜃
1 1 (t)cos𝜃XY − x]
̇
(k1 A1 )2 + (k1 A1 𝜔)2 sin𝜃XY /(2k1 + kY )
⎪ FY (t) = k1 [𝜃1 (t)sin𝜃XY − y] + c1 [𝜃̇ 1 (t)sin𝜃XY − y]
̇
⎩ D= √[ ] [ ]
2 2
(2) 1 − ( 𝜔𝜔 )
2
+ 2𝜉Y ( 𝜔𝜔 )
nY nY
where tanθ = y/x. In Eq. (2), the vibration excitation FX(t) √
is defined as: and ωnY = kY ∕m , ξX = (c1 + cY)/cnY, cnY = 2mωnY, and
FX (t) = k1 A1 sin(𝜔t + 𝜑 + 𝛼) + c1 A1 𝜔cos(𝜔t + 𝜑 + 𝛼) 𝜔 𝜔 2
+k1 A1 sin(𝜔t + 𝛼)cos𝜃XY + c1 A1 𝜔cos(𝜔t + 𝛼)cos𝜃XY (3) Δ𝜑2 = arctan[2𝜉Y ( )∕1 − ( )]
𝜔nY 𝜔nY

where θXY is the vibration position of the output end of the According to the relationship between the workpiece
end effector and is related to the phase difference φ and the coordinate system and the physical coordinate system of the
angle θ. Equation (3) can be set as: tool in Fig. 3, the steady-state vibration response of the tool
𝜑 in the workpiece coordinate system can also be written as:
FX (t) = A3 sin(𝜔t + + Δ𝛽) (4) {
2
Xf (t) = Y(t) ⋅ cos 𝜃2 + X(t) ⋅ sin 𝜃2
where (7)
Yc (t) = Y(t) ⋅ sin 𝜃2 + X(t) ⋅ cos 𝜃2
�
⎧ If the influences of the phase parameters Δα, Δβ, Δϕ1, and
A
⎪ 3 = [(k1 A1 )2 + (k1 A1 𝜔)2 ][1 + cos2 𝜃XY + 2cos𝜃XY cos𝜑]
⎨ Δ𝛽 = 𝛼 + arctan(c1 𝜔∕k1 ) + arctan[(A2 − A2 cos𝜃XY )tan 𝜑 Δϕ2 on the vibration response and the phase delays are
⎪ 2
ignored, the elliptic equations can be obtained by eliminating
⎩ ∕(A2 + A2 cos𝜃XY )]
the time variable t in Eqs. (5) and (6) and can be derived as:
The steady-state responses of FX(t) at the output end Q of 2
X Y Y 2
the end effector are expressed as follows: ( − cos𝜑) = [1 − ( ) ]sin2 𝜑 (8)
C D D
X(t) = C ⋅ sin(𝜔t + 𝜑 + Δ𝛽 − Δ𝜑1 ) (5)
In the workpiece coordinate system, Eq. (10) can also be
where written as:

A3 /(2k1 + kX )
C= √
[ ]2 [ ]2
2
1 − ( 𝜔𝜔 ) + 2𝜉X ( 𝜔𝜔 )
nX nX

√
and ωnX = kX ∕m , ξX = (2c1 + cX)/cnX, cnX = 2mωnX, and
𝜔 𝜔 2
Δ𝜑1 = arctan[2𝜉X ( )∕1 − ( )]
𝜔nX 𝜔nX

Similarly, the steady-state responses of FY(t) at the output

end Q of the end effector are expressed as follows:
Y(t) = D ⋅ sin(𝜔t + Δ𝛼 − Δ𝜑2 ) (6)
Fig. 6  The ideal elliptical vibration locus for different angles φ

13
The International Journal of Advanced Manufacturing Technology

Fig. 9  UEVC tool with dual longitudinal vibration excitations

Fig. 7  Maximum amplitude ratio in the X- and Y-directions for differ- and the vibration differential equations in the θ-direction
ent angles θ
are as follows:

⎧ M ẍ 1𝜃 = k[F2 (t) − x1𝜃 ] + c[Ḟ 2 (t) − ẋ 1𝜃 ] − k1 (x1𝜃 − x𝜃 )

2
XC cos 𝜃2 + YD sin 𝜃2 XC sin 𝜃2 + YD cos 𝜃2
( − cos𝜑) ⎪ −c1 (ẋ 1𝜃 − ẋ 𝜃 )
C D ⎨ m̈x = k (x − x ) + c (ẋ − ẋ ) − [k x + c ẋ ]sin𝜃
2 ⎪ 𝜃 1 1𝜃 𝜃 1 1𝜃 𝜃 X 𝜃 X 𝜃 XY
XC sin 𝜃2 + YD cos 𝜃2 ⎩ −[kX x𝜃 + cX ẋ 𝜃 ]cos𝜃XY
= [1 − ( ) ]sin2 𝜑 (9) (11)
D
The resonant frequencies of the integrated system, includ-
The elliptical vibration trajectory of the tool tip in the reso-
ing M and m in the X- and θ-directions, can be determined
nant state is directly affected by the angle θ and the phase
by Eqs. (10) and (11), respectively. The phase difference φ
difference φ. Furthermore, based on Fig. 4 and Eq. (2), M
determines the ellipticity of the elliptical vibration locus.
and m are combined to analyze the dynamics of the system
When the phase difference φ = 0° or 180°, the vibration locus
in the X- and θ-directions. The vibration differential equa-
is approximately a straight line; when φ ∈ (0°, 180°), the
tions in the X-direction are as follows:
vibration locus is approximately an ellipse. If the input ampli-
⎧ M ẍ = k[F (t) − x ] + c[Ḟ (t) − ẋ ] − k (x − x) − c (ẋ − x) tude and frequency are the same, the angle θ influences the
̇
⎪ 1 1 1 1 1 1 1 1 1 ­ C- and Y
amplitude of the vibration locus in the X ­ D-directions
⎨ m̈x = k1 (x1 − x) + c1 (ẋ 1 − x)
̇ − kX x − cX ẋ + [k1 (x1𝜃 − x𝜃 )
by affecting the equivalent output amplitudes of point P and
⎪ +c1 (ẋ 1𝜃 − ẋ 𝜃 )]cos𝜃XY
⎩ Q and the related phases of the vibration responses.
(10)
2.2 Theoretical analysis of the effects
of the configuration angle

The normal resonance frequency and the tangential reso-

nance frequency should be the same. So their phase differ-
ences between the two longitudinal excitations in the two

Fig. 8  Relative ratio of the resonance frequencies in the X- and

Y-directions for different angles θ Fig. 10  Normal and tangential resonance state of the integrated horn

13
 The International Journal of Advanced Manufacturing Technology

Fig. 11  Equivalent structure of

the normal and tangential lon-
gitudinal vibrations. (a) Normal
resonance in the Y­ D-direction.
(b) Tangential resonance in the
­XC-direction

resonance modes should be 0° and 180° in a standard UEVC ratio C/D between the X-direction and Y-direction is 1
tool. Under the two resonance modes, the tool can gener- theoretically (Fig. 7). The reason is that only single longi-
ate the ultrasonic elliptical vibration locus in the resonance tudinal vibration is transmitted in the X- and Y-directions,
state, as shown in Fig. 5. whereas only one bending vibration mode is transmitted
According to the vibration response of vibration coupling in the vertical direction. In this state it does not inter-
described in “Sect. 2” the angle θ and phase difference φ are fere with the longitudinal vibration mode. Besides, the
two initial factors affecting the elliptical vibration trajectory maximum amplitude ratio between two resonance states
at the output of the end effector. When the phase delay error with different phase differences is equal when the angle
is not considered, the elliptical vibration trajectories of the θ = 90°. Similarly, we can also obtain the relative ratio of
tool tip for different phase differences φ and angles θ are the resonance frequencies in the X- and Y-directions for
shown in Fig. 6. any angle θ by using the natural frequencies in the X- and
The poses and ellipticity of elliptical vibration trajecto- Y-directions with θ is 90° as the reference, as shown in
ries change with the phase difference φ. Besides, according Fig. 8.
to vibration coupling, the resonance frequencies are also When θ = 90°, there is a minimum natural frequency
affected by θ and φ. According to Eqs. (5), (6), and Fig. 6, in the X-direction and a maximum natural frequency in
we can obtain the functional relationship between the ratio the Y-direction. It is easy to obtain a relatively consist-
of the X-direction amplitude to the Y-direction amplitude of ent resonant frequency with a minimum frequency differ-
the resonance and the angle θ for φ = 0° or 180°, as shown ence between the two directions. At this time, the system
in Fig. 7. reaches the optimal design layout of the UEVC tool with
When the input amplitudes of the two transducers are double excitations. Therefore, the layout state of θ = 90° and
the same and the configuration angle is 90°, the amplitude φ = 0°/180° facilitates the vibration coupling analysis and
improves the system’s performance.

2.3 Configuration of a well‑decoupled ultrasonic

elliptical vibration cutting tool

When the angle θ is less than 90°, the computational com-

plexity of calculating the vibration decoupling mode and
response becomes very high. This design can decrease

Fig. 12  Natural frequency trend of a horn with mass M Fig. 13  Mass distribution and structure of a single horn

13
The International Journal of Advanced Manufacturing Technology

3 Optimization of the structural parameters

A general model is established to clarify the relationship

between the design parameters and the vibration response.
We evaluate the influences of the parameters on the reso-
nance performance of the UEVC, including the trajectory
control, amplitude output, analytical model results, and
frequency estimation. The configuration can be optimized
by choosing appropriate design parameters and vibration
input parameters.

3.1 Effects of the structural parameters of the horn

and connection blocks

The prerequisite for achieving a stable elliptical vibration of

the system is that the coupled generator produces an ultra-
Fig. 14  Vibration transmission state of the integrated horn
sonic resonance in the ­XC- and ­YD-directions, as shown
in Fig. 10. The resonant frequencies of the system in both
directions are same or similar.
the accuracy of calculation and complicate the design
The two horns have the same resonance load on each
process. The main reason is that when included angle
other and they are equivalent to two spring damping system
θ is less than 90°, it can result in a bending vibration
with different phase differences. The equivalent vibration
mode and another additional longitudinal vibration mode
structure of the system is shown in Fig. 11.
in the vibration direction. The complex vibration modes
kXA = kYA and cXA = cYA denote the normal equivalent
increase the difficulty of vibration decoupling and reduce
stiffness and damping of the connecting section of the con-
the accuracy of the steady-state response and frequency
nection block and the end effector, respectively. In the tan-
calculation. When θ is 90°, the two longitudinal ultra-
gential resonance, kXT = kYT and cXT = cYT are the equivalent
sonic transducers and their horns are perpendicular to
stiffness and damping in the ­XC- and Y ­ D-directions of the
each other, pointing in the X- and Y-directions of the
connecting section between the connection block and the end
physical coordinate system of the tool. This layout facili-
effector, respectively. If there are differences in the equiva-
tates the analysis of the steady-state response and reso-
lent stiffness and damping in the X ­ C- and Y
­ D-direction, their
nant frequency and improves the accuracy of design and
resonant frequencies are different.
calculation. A vertical design of the UEVC with double
For the vibration transmission in the X-direction, the
longitudinal vibrations coupling is proposed to enable the
natural frequency ωXA of the system with phase difference
adjustability of the elliptical vibration trajectory. This
0° can be calculated:
design is helpful for decoupling of the double longitudi-
[ ]
nal vibration, as shown in Fig. 9. (k + kXA )M + (k1 + kXA )m
The two vibrations with phase difference produced by
2
c𝜔XA =
2Mm
the two ultrasonic transducers are transmitted to tooltip √[ ]2 (12)
and are combined into an elliptical vibration locus [39]. (k + kXA )M − (k1 + kXA )m + 4k12 Mm
When the tool works normally, the shape and pose of ±
2Mm
the elliptical vibration locus can be adjusted by changing
the phase difference between the two ultrasonic power Similarly, the natural frequency ωYA in Y-direction can be
signals. obtained and ωA = ωXA = ωYA, while the natural frequency

Table 1  Main material Material Density Elastic Poisson’s ratio

parameters of the horns and tool (g/cm3) modulus
(GPa)

Horns Stainless steel 7.93 196 0.25

Connection block Aluminum alloy 2.78 73 0.33
Tool Cemented carbide 13.1 750 0.18

13
 The International Journal of Advanced Manufacturing Technology

Table 2  Main structural parameters of the horns and tool

Column l1 l2 Column
diameter (mm) (mm) side
(mm) length
(mm)

Horns 18 18 46 –
Connection block – – – 10

ωXT in X-direction can also be obtained when φ = 180° and

ωT = ωXT = ωYT.
When the equivalent stiffnesses, damping, and pitch posi-
tions in each connection position are constant, the qualities
of the system’s parts are the main factors affecting the reso- Fig. 16  The relationship between the resonance frequencies and the
column sizes l1 and l2
nance. The modal superposition method is used to solve the
equation of motion. The steady-state vibration response of
the tooltip has the largest unidirectional longitudinal vibra-
tion amplitude and it can be written as: the natural frequency ωi, the amplitude of the steady-state
forced vibration of the horn in the ith order canonical
x2 (t) =U 12 ||H1 (𝜔)||Bsin(𝜔t + 𝜑 − 𝜑1 )
coordinate is very large, i.e., the steady-state vibration
+ U 22 ||H2 (𝜔)||Bsin(𝜔t + 𝜑 − 𝜑2 ) response of the tool tip has the largest unidirectional lon-
(13)
gitudinal vibration amplitude and it can be written as:
where U12 and U22 are the corresponding vectors in the
regular mode matrix U; |H(ω)| is the magnification factor x 2 (t) =U � 12 ||H � 1 (𝜔)||Bsin(𝜔t + 𝜑 − 𝜑� 1 + 𝜋)
��

corresponding to the regular coordinates; φ1 and φ2 are the + U � 22 ||H2 (𝜔)||Bsin(𝜔t + 𝜑 − 𝜑� 2 + 𝜋) (15)
phase angles corresponding to the regular coordinates; and B
is an amplitude function determined by stiffness k, damping The steady-state vibration response of the system consists
c, and frequency ω. of the steady-state vibration responses of the two horns. The
According to Fig. 11a, the steady-state vibration response synthetic steady-state vibration response of the integrated
of the system consists of the steady-state vibration responses horn in the ­YD-direction can be obtained as:
of the two horns. The synthetic steady-state vibration √
response of the integrated horn in the ­XC-direction can be (16)
��
YD (t) = 2 ⋅ x 2 (t)
obtained as:
√ Similar to the composite normal longitudinal vibration
XC (t) = 2 ⋅ x2 (t) (14) mode, the longitudinal vibration mode has two natural fre-
quencies. They are composite responses generated by the
According to Fig. 11b, when the frequency ω of the input superposition of the two external excitation vibrations with
external excitation vibrations F1(t) and F2(t) coincide with the same frequency and different phases.

Fig. 15  Simulated resonance modes of the integrated horn: (a) normal; (b) tangential

13
The International Journal of Advanced Manufacturing Technology

Table 3  Main structural parameters of the horns and tool in the opti-
mal design

Column l1 l2 Column
diameter (mm) (mm) side
(mm) length
(mm)

Horns 18 13 43 –
Connection block – – – 10

adjustment can change the design parameters of the inte-

grated horn for different vibration modes. For a specific
design and material, the parameter relationships are:
Fig. 17  The relationship between the differences between normal and
tangential resonance frequencies and the column sizes of the inte- ⎧ k1 = 𝛼k
grated horn ⎪ kXA = 𝛽k
⎨ k = 𝛾k (18)
⎪ XT
⎩ m = 𝜂M
The integrated horn and the single horn should generate
the same longitudinal resonance frequencies in the ­XC- and where α, β, and γ are the scale coefficients of the stiffness of
­YD-directions under ideal conditions, i.e., the two exter- each part, and η is mass ratio of m and M.
nal excitation vibrations can create the elliptical trajectory The normal and tangential resonance frequencies of the
with the same frequency. The same or similar longitudinal integrated horn are:
resonance frequency when the phase difference is 0° or
√
180° must be also produced. Thus, the elliptical trajectory ⎧ 2 k k [(1+𝛽)+(𝛼+𝛽)𝜂] [(1+𝛽)−(𝛼+𝛽)𝜂]2 +4𝛼 2 𝜂
can be formed when phase difference φ ∈ (0, π). Therefore, ⎪ 𝜔A = A1 M = M { ± }
⎨
2𝜂 √ 2𝜂
the resonant frequencies of the integrated horn in the ideal 2
⎪ 𝜔2 = A k = k { [(1+𝛾)+(𝛼+𝛾)𝜂] ± [(1+𝛾)−(𝛼+𝛾)𝜂] +4𝛼 𝜂 }
2

state in the normal and tangential directions should be ⎩ T 2M M 2𝜂 2𝜂

equal, namely: (19)

We use the mass√ratio of m and M as an independent vari-
𝜔A = 𝜔T (17)
able. Then ω0 = k∕M = 1 and ωA can be set as an example
If the equivalent k and c of the two resonant modes are equal, to illustrate the trend of difference in the natural frequen-
the normal resonance frequency of the integrated horn is cies of the composite longitudinal vibration of the integrated
equal to the tangential resonance frequency. horn, as shown in Fig. 12.
The resonance frequency of the integrated horn is The main mass M is composed of the column section M1
related to the mass M, the mass m of the end effector, of a single horn and the column section M2 of the connec-
and the equivalent stiffness and damping of each part. tion block, as shown in Fig. 13. The trend of the frequency
Therefore, the only approach to obtain optimal design and of the tangential longitudinal vibration resonance with the
matched resonant frequency is to adjust the mass M. This main mass M is consistent with the normal longitudinal

Fig. 18  Simulated normal and tangential resonance modes of the integrated horn after optimization: (a) normal; (b) tangential

13
 The International Journal of Advanced Manufacturing Technology

Table 4  Results of the normal and tangential resonance frequencies

(Hz)
Normal vibration Tangential Frequency
vibration difference

Simulation results 17,926 18,194 268

Impedance analysis 18,117 – –
Experimental results 18,380 18,530 150

3.2 Model verification using FEM simulation

The horn consists of stainless steel and the connection block

Fig. 19  Prototype of the ultrasonic elliptical vibration cutting tool and end effector are made of aluminum alloy whose vibra-
tion transmission state can be seen in Fig. 14. The diameter
vibration. The main mass M has the most significant impact of the rear column section of the horn is d = 40 mm, and the
on the resonance frequency. When α and β remain constant, column section length of the connection block is l = 20 mm.
the resonant frequencies can be adjusted by optimizing the The insert is a TPGH110302 PCD CNC turning tool, and
mass ratio m/M. the material parameters are shown in Table 1.
When α and β remain constant, Eq. (19) is redefined as: Since integrating horns can be equivalent to mechanical
{ 2 impedance, it can lead to great reduction of the resonant
k
𝜔A = (M +M
2)
A1 frequency of the UEVC. So the initial design frequency can
2
1
k (20) be set to 20 kHz, greater than the target resonance frequency
𝜔T = (M +M ) A2
1 2
of 18 kHz. Using the calculation principle of longitudinal
According to Fig. 5 and Eq. (19), a difference exists between vibration 1/4 wavelength, it can be concluded that the size
the two resonance frequencies in the ­XC- and ­YD-directions. of the column sections in front and rear nodal surfaces of the
The reason is that the equivalent stiffness and damping of the cylindrical horn is 72 mm. Then the data of initial size and
two vibration modes are different. Therefore, it is necessary material is induced into Eqs. (19) and (20) for calculation
to compare and analyze the normal and tangential resonance and the range of mass ratio of m to M is 0.01 ~ 0.02. Accord-
frequencies, namely: ing to Fig. 12, the range of ratio of ωA to ω0 of the system
with frequency reduction is 0.5 ~ 0.6 and the range of ωA is
Δf = min(||𝜔A − 𝜔T ||) (21) 10 ~ 12 kHz. Therefore, the size design frequency of 20 kHz
is not suitable and the sizes of column sections on both sides
The two frequencies satisfying Eq. (21) can be selected to
of the nodal surface need to be adjusted again.
optimize the design parameters of the integrated horn for
According to the range of ratio of ωA to ω0 and the tar-
matching the two resonances. After m and M2 have been
get resonance frequency, it is estimated that the range of size
determined, M1 can be adjusted by changing l1 and keep-
design frequency based on longitudinal vibration can set to be
ing M2 unchanged. In addition, l1 needs to be optimized
30 ~ 45 kHz. Taking the sizes of end effector and connection
to ensure mass symmetry at both sides of the nodal sur-
column section into calculation, the size parameters of the
face and obtain the same normal and tangential resonance
horns with design frequency of 30 kHz are shown in Table 2.
frequencies.

Fig. 20  Experimental setup of

the ultrasonic elliptical vibra-
tion system

13
The International Journal of Advanced Manufacturing Technology

Fig. 21  Vibration trajectory

of the tool output end for dif-
ferent phase differences (μm):
(a) 0° ± 3°; (b) 45° ± 3°; (c)
90° ± 3°; (d) 135° ± 3°; (e)
180° ± 3°

In the simulation of ABAQUS, it should be noted that It indicates that the normal and tangential resonant fre-
threaded connection is equivalent to the binding state quencies cannot match. Therefore, the parameters of the
between the two contact surfaces of end effector and horn. integrated horn need to be optimized to reduce the dif-
When the position of vibration nodal plane is determined, ference again. The relationships between the resonance
the thickness of the fixed flange on nodal plane and circle frequencies and l1 and l2 for the two resonance modes of
flexure hinge (as shown in Fig. 9) also affect resonant the integrated horn are shown in Fig. 16.
state of system. Therefore, the thickness can be uniformly For the same column size l1, the resonant frequencies
set to 1 mm and the minimum thickness of circle flexure decrease with an increase in l2; for the same l2, the reso-
hinge can be set to 3 mm. In addition, all DOFs of the two nance frequencies decrease with an increase in l1. Increas-
fixed nodal surfaces of the fixed flanges towards the end ing l2 is equivalent to increasing the overall size of the
effector are set as the constrained state and the grid cell integrated horn, causing a reduction in the resonance fre-
type of model is C3D10. The external excitation vibration quency. The relationship between the resonance frequency
is simulated at the input end to replace piezoelectric trans- difference, l1 and l2 are shown in Fig. 17. For the same
ducer under different resonant frequencies. This omis- l1, the difference between the two resonance frequencies
sion does not affect the vibration simulation results of the decreases with an increase in l2; for the same l2, the reso-
resonant frequencies and relative amplitudes. The external nance frequency of the integrated horn increases with an
excitation vibration is input on the contact surface of the increase in l1.
horn. The end face of the front cover of the transducer
is used for the modal and harmonic response vibration 3.3 Results of design optimization
analysis in FEM. The normal and tangential resonance
modes of the integrated horn are shown in Fig. 15. When a range of the difference between the normal and
The integrated horn produces two composite resonance tangential resonance frequencies is set as 0–300 Hz, the
modes at 16,874 Hz and 17,360 Hz, respectively, and optimum ranges of l1 and l2 are 12–14 mm and 41–46 mm,
the frequency difference Δf between the two is 486 Hz. respectively. The ideal operating frequency of a standard

Table 5  Parameters of elliptical Phase difference 30° 45° 90° 135° 150°
vibration cutting
Included angle (see Fig. 22) 75° 65° 90° 10° 0°
Amplitude Ac in XC (μm) 1.6 1.8 2.6 2.9 3.1
Velocity coefficient K 0.1 0.1 0.1 0.1 0.1
Cutting speed VC in XC (mm/s) 18.096 20.358 29.405 32.798 35.060
Cross-feed (μm) 50 50 50 50 50
Cutting depth (μm) 4 4 4 4 4

13
 The International Journal of Advanced Manufacturing Technology

Fig. 22  Position and orientation

in cutting coordinate system
of elliptical cutting trajectories
corresponding to different phase
differences: (a) 30°; (b) 45°; c
90°; (d) 135°; (e) 150°

ultrasonic transducer is 18000 ± 200 Hz. So the optimum 4.2 Measurement results of the resonance
ranges of l1 and l2 can be set as 12–14 mm and 42–45 mm, frequency
respectively. The main dimensional parameters of the inte-
grated horn with target frequency of 18 kHz after design The resonant frequency is mainly measured by the imped-
optimization are shown in Table 3. ance analyzer and it is compared with the simulation. In
The remaining dimensional parameters are consistent impedance analysis, the positive poles of the two trans-
with the preset design. The optimized design model is simu- ducers are connected together and the negative poles are
lated to verify the final design, as shown in Fig. 18. connected together. Then the wire with two positive poles
The normal resonant frequency is 17926 Hz and the is connected with the positive pole of the impedance ana-
tangential resonant frequency is 18194 Hz in simulation. lyzer, and the wire with two negative poles is connected
The difference Δf between the two resonant frequencies is with the negative pole of the impedance analyzer. At this
268 Hz. Relative to the target preset frequency of 18 kHz, time, the phase difference of measured resonant frequency
their frequency deviation error rates are 0.41% and 1.07%, is 0°, that is, the input signals of the two transducers have
respectively. The simulation result demonstrates the accu- the same phase. The impedance analyzer PV70A directly
racy of the optimization method. measures the resonant frequencies in the normal direction
(phase difference is 0). The comparison of resonance fre-
4 Experimental verification of the vibration quencies of the ultrasonic system is obtained by the imped-
performance ance analyzer, the simulation, and vibration experiments,
as shown in Table 4.
4.1 Experimental setup The simulation results of the normal and tangential res-
onance frequencies after the optimization are consistent
A prototype of the optimized UEVC tool is developed for with the impedance analysis and experimental measure-
experimental verification, as shown in Fig. 19. The experi- ment. The frequency differences meet the requirements
mental setup consists of the UEVC tool, the ultrasonic of the design, and the system achieves optimal frequency
power supply, and the measurement devices. The sche- matching.
matic diagram of experimental setup is shown in Fig. 20.
The power supply with dual power outputs can be used to 4.3 Measurement results of the elliptical vibration
adjust the phase difference from 0 to 180°. The measure- trajectory
ment devices consist of a signal collector, a computer and
two laser displacement sensors (KEYENCE LK-H008) with The measurement of elliptical vibration trajectory under
measurement accuracy of 0.005 µm and sampling period different phase differences is conducted to verify the pro-
of 2.55 µs. posed design principle. Resonant test frequencies of sys-
The vibration experiments are conducted with varying tem with different phase differences can be first set as the
excitation frequencies in a range of 16 ~ 20 kHz and phase resonant frequency of system with phase difference 0°.
differences between the two input electrical signals are set Then the phase difference between the two output elec-
as 0° and 180°, respectively. trical signals is adjusted, and two displacement sensors

13
The International Journal of Advanced Manufacturing Technology

Fig. 23  Machined surfaces of different elliptical vibration trajectories for different phase differences: (a) 30° ± 3°; (b) 45° ± 3°; (c) 90° ± 3°; (d)
135° ± 3°; (e) 150° ± 3°

13
 The International Journal of Advanced Manufacturing Technology

Fig. 24  Measurement method of surface topography. (a) Measurement of actual residual height. (b) Calculation of theoretical residual height. (c)
Measurement results of surface topography

are set at the output of end effector. When the phase dif- Surface fluctuation and residual height are adopted as
ference state is selected and the vibration output of end indices to characterize the quality of machined surface. The
effector can obtain approximate theoretical trajectory, the surface topography of these groove bottom in cutting direc-
system realizes resonance under this phase difference. tion is measured by white light interferometer (see Fig. 23).
Therefore, the consistency of the resonant frequency The measurement content consists of data of two param-
under different phase differences and the effectiveness eters: HF and HA. HF is the surface fluctuation in the range
of design can be confirmed by measuring elliptical vibra- of 100 μm and HA is the elliptical cutting residual height in
tion trajectory directly. the range of 10 μm (see Fig. 24a). The theoretical residual
The elliptical trajectory can be flexibly changed in a height HT can be calculated from the kinematic trajectory
working space of 5.1 μm × 5.3 μm by adjusting the phase of tool (see Fig. 24b), where the blunt circle radius of the
difference after filtering. The elliptical trajectories for phase tool is not considered.
differences of 0°, 45°, 90°, 135°, and 180° are shown in The measured and predicted results of surface fluctua-
Fig. 21. The peak-to-peak voltage of output 1 is 320 V, that tion and residual height are shown in Fig. 24c. The results
of output 2 is 280 V, and the current is 0.75 A. The designed demonstrate that the shape and pose of elliptical vibration
UEVC tool reaches ultrasonic elliptical vibration states, trajectory significantly affect surface quality. When cut-
indicating that the optimization design is feasible. ting included angle Ф = 0°, the minimum residual height
of elliptical cutting on the surface in the range of 10 μm
is 67 nm and the minimum surface fluctuation on the sur-
face in the range of 100 μm is 97 nm. When the cutting
4.4 Cutting experiments with different elliptical included angle Ф is 75°, the maximum difference between
vibration trajectories residual height of elliptical cutting and the surface fluctua-
tion appears on the surface.
Cutting experiments are carried out with different elliptical The experimental and predicted theoretical results of
vibration trajectories by adjusting different phase differ- residual height have a same variation tendency with the
ences to further verify the performance of designed UEVC increase of phase difference φ. There are two main rea-
tool. The process parameters for the plane machining are sons for the difference between the actual residual height
shown in Table 5. The poses of ellipse are characterized by HA and the theoretical residual height HT. (1) In the cal-
the included angle Ф between the long diameter of ellipse culation, H A is characterized by the arithmetic mean of
and the cutting direction. The elliptical cutting paths and the difference between the surface height line and the
their shape feature in the cutting coordinate system are contour bottom envelope in 10 μm cutting path. H F can
shown in Fig. 22. amplify HA to some certain extent. (2) Only a few of the

13
The International Journal of Advanced Manufacturing Technology

larger and smaller values of the contour coincide with the Code availability There is no program or software related code setting
envelope. So, the actual residual height HA on the actual in this study.
contour line is greater than HT. In addition, the influence
of nose radius of tool tip cannot be ignored, which can Declarations
reduce the surface residual height to a certain extent [40].
Ethics approval Authors promise that this manuscript is original, and
In summary, the results of cutting experiments verify it has not been published in whole or in part, nor is it being considered
that the designed tool has the capacity to optimize the for publication or submission elsewhere.
UEVC process by flexibly adjusting the tool vibration
trajectory. Consent to participate We make sure the author group, the correspond-
ing author, and the order of authors are all correct at submission.

5 Conclusions Consent for publication The authors have reviewed the present version
of the manuscript and approved it for the final publication.
This study proposed an optimized design of an ultrasonic
elliptical vibration cutting (UEVC) tool with dual longi- Conflict of interest The authors declare no competing interests.
tudinal vibration excitations to achieve high adjustability
of vibration trajectory. The key design parameters of the
integrated horn were identified and optimized. Vibration References
experiments and cutting tests were conducted to verify per-
formance of the optimized UEVC tool. The following con- 1. Ma C, Hu D (2003) Ultrasonic elliptical vibration cutting. Chin J
clusions can be drawn. Mech Eng 39(12):67–70
2. Kim G, Loh B (2007) An ultrasonic elliptical vibration cutting
tool for micro V-groove machining: kinematical analysis and
(1) The vibration decoupling of the UEVC system with micro V-groove machining characteristics. J Mater Process Tech-
a configuration angle θ of 90° (the included angle of nol 190(1–3):181–188
two longitudinal generators) is better than that with an 3. Kim D, Cha K, Sung I, Bryan J (2002) Design of surface micro-
structures for friction control in micro-systems applications. CIRP
acute angle due to the conciseness of kinematic equa- Ann-Manuf Technol 51(1):495–498
tion. This layout state cannot introduce additional inter- 4. Brandner J, Anurjew E, Bohn L, Hansjosten E, Henning T, Schygulla
ference vibration modes in decoupling. U, Wenka A, Schubert K (2006) Concepts and realization of micro-
(2) The main mass M of a single horn is the primary param- structure heat exchangers for enhanced heat transfer. Exp Thermal
Fluid Sci 30(8):801–809
eter affecting the normal and tangential resonance fre- 5. Fang F, Zhang X, Weckenmann A, Zhang G, Evans C (2013)
quencies of the integrated horn. The frequency devia- Manufacturing and measurement of freeform optics. CIRP Ann-
tion of the integrated horn between the normal and Manuf Technol 62(2):823–846
tangential directions can be reduced by adjusting the 6. Wang J, Liao W, Guo P (2019) Modulated ultrasonic elliptical
vibration cutting for ductile-regime texturing of brittle materials
structural parameters of the horn beam. with 2-D combined resonant and non-resonant vibrations. Int J
(3) The designed UEVC tool can generate standard ellip- Mech Sci 170(15):105347
tical vibration trajectories with adjustable shapes in a 7. Kim G, Loh B (2013) Cutting force variation with respect to tilt
working space of 5.1 μm × 5.3 μm when the phase dif- angle of trajectory in elliptical vibration V-grooving. Int J Precis
Eng Manuf 14(10):1861–1864
ference of two input voltages changes from 0 to 180°. 8. Huo B, Zhao B, Yin L, Guo X, Wang X (2021) Effect of double-
(4) The orientation of elliptical vibration trajectories excitation ultrasonic elliptical vibration turning trajectory on sur-
critically affects the morphology of machined surface. face morphology. Int J Adv Manuf Technol 113:1401–1414
When the included angle between the long axis of the 9. Zhang J, Suzuki N, Shamoto E (2013) Investigation on machining
performance of amplitude control sculpturing method in elliptical
ellipse and the cutting direction is 0°, the minimum sur- vibration cutting. Procedia Cirp 8(Complete):328–333
face roughness and surface fluctuation were obtained. 10. Brehl D, Dow T (2008) Review of vibration assisted machining.
Precis Eng 32(3):153–172
11. Shamoto E, Suzuki N, Naoi Y (2002) Development of ultrasonic
Author contribution Authors carry equal contributions. elliptical vibration controller for elliptical vibration cutting. CIRP
Ann-Manuf Technol 51:327–330
12. Yin Z, Dai C, Cao Z, Li W, Chen Z, Li C (2020) Modal analysis
Funding The authors gratefully acknowledge the financial sup-
and moving performance of a single-mode linear ultrasonic motor.
port for this research provided by the Shenzhen Basic Research
Ultrasonics 108:106216
Discipline Layout Project of China (Subject Layout) (Grant No.
13. Zhang S, Guo Y, Chen Z, Li G, Guo Q, Wu Y, Zeng J (2021) Pro-
JCYJ20160428181916222) and National Natural Science Foundation
posal for a novel elliptical ultrasonic aspirator and its fundamental
of China (Grant No. 51875311 and 52105458).
performance in cartilage removal. Ultrasonics 109:106259
14. Li X, Zhang D (2006) Ultrasonic elliptical vibration transducer
Data availability The data used or analyzed in the study are available driven by single actuator and its application in precision cutting.
from the corresponding author on reasonable request. J Mater Process Technol 180(1–3):91–95

13
 The International Journal of Advanced Manufacturing Technology

15. Lin J, Han J, Lu M, Zhou J, Gu Y, Jing X, Feng D (2017) Design elliptical-vibration texturing tool for surface texturing. J Mech
and performance testing of a novel three-dimensional elliptical Sci Technol 31(7):3465–3473
vibration turning device. Micromachines 8(10):305 29. Moriwaki T (2010) Development of 2-DOF ultrasonic vibration
16. Kurniawan R, JoKo T, Kumaran S, Ahmed F (2021) 3-DOF cutting device for ultra-precision elliptical vibration cutting.
ultrasonic elliptical vibration tool holder based on coupled Key Eng Mater 447–448:164–168
resonance modes for manufacturing micro-groove. Precis Eng 30. Lin J, Han J, Lu M, Yu B, Gu Y (2017) Design, analysis and
67:212–231 testing of a new piezoelectric tool actuator for elliptical vibra-
17. Shamoto E, Moriwaki T (1999) Ultaprecision diamond cutting tion turning. Smart Mater Struct 26(8):085008
of hardened steel by applying elliptical vibration cutting. CIRP 31. Tan R, Zhao X, Zou X, Sun T (2018) A novel ultrasonic ellipti-
Ann-Manuf Technol 48(1):441–444 cal vibration cutting tool based on a sandwiched and symmetri-
18. Zhou M, Hu L (2015) Development of an innovative tool for cal structure. Int J Adv Manuf Technol 97(1):1397–1406
ultrasonic elliptical vibration cutting. Ultrasonics 60:76–81 32. Yang Y, Gao S, Chen K, Pan Y, Guo P (2017) Vibration analysis
19. Amini S, Khosrojerdi M, Nosouhi R (2017) Elliptical ultrasonic– and development of an ultrasonic elliptical vibration tool based
assisted turning tool with longitudinal and bending vibra- on a portal frame structure. Precis Eng 50:421–432
tion modes. Proc Institution Mech Eng Part B J Eng Manuf 33. Lin S, Kong X, Wang C, Yun Z, Liang Y (2018) A novel three-
231(8):1389–1395 dimensional elliptical vibration cutting tool based on the free-
20. Guo P, Ehmann K (2013) Development of a tertiary motion dom and constraint topologies theory. Proc Institution Mech
generator for elliptical vibration texturing. Precis Eng Eng Part B J Eng Manuf 233(3):1–11
37(2):364–371 34. Richardson A, Walsh K, Abdullah M (2013) Closed-form equa-
21. Guo P, Ehmann K (2013) An analysis of the surface generation tions for coupling linear structures using stiffness and damping
mechanics of the elliptical vibration texturing process. Int J elements. Struct Control Health Monit 20(3):259–281
Mach Tools Manuf 64(1):85–95 35. Shamoto E, Hashimoto Y, Shinagawa M, Sencer B (2014) Ana-
22. Zhang C, Guo P, Ehmann K, Li Y (2016) Effects of ultrasonic lytical prediction of contact stiffness and friction damping in
vibrations in micro-groove turning. Ultrasonics 67:30–40 bolted connection. CIRP Ann-Manuf Technol 63(1):353–356
23. Yang Y, Pan Y, Guo P (2017) Structural coloration of metallic 36. Suzuki N, Yokoi H, Shamoto E (2008) Improvement of load
surfaces with micro/nano-structures induced by elliptical vibra- characteristics of 3 DOF ultrasonic elliptical vibration tool. Proc
tion texturing. Appl Surf Sci 402(APR.30):400–409 JSPE Semestrial Meeting 2008A:B63
24. Guo P, Lu Y, Ehmann K, Cao J (2014) Generation of hierarchi- 37. Deng Y, Chen T, Liu S, Ye M (2018) Design and optimization
cal micro-structures for anisotropic wetting by elliptical vibra- of a double-excitation ultrasonic elliptical vibration grinding
tion cutting. CIRP Ann-Manuf Technol 63(1):553–556 device. Piezoelectrics & Acoustooptics 42(2):182–186
25. Shamoto E, Suzuki N, Tsuchiya E, Hori Y, Inagaki H, Yoshino 38. Benaroya H (2004) Mechanical vibration: analysis, uncertain-
K (2005) Development of 3 DOF ultrasonic vibration tool for ties, and control. CRC Press/Taylor & Francis
elliptical vibration cutting of sculptured surfaces. CIRP Ann- 39. Moriwaki T, Shamoto E (1999) Ultrasonic elliptical vibration
Manuf Technol 54(1):321–324 cutting. CIRP Ann-Manuf Technol 44(1):31–34
26. Huang W, Yu D, Zhang M, Ye F, Yao J (2017) Analytical design 40. Zhang X, Kumar A, Rahman M, Liu K (2013) Modeling of the
method of a tool for ultrasonic elliptical vibration cutting. J effect of tool edge radius on surface generation in elliptical vibra-
Acoust Soc Am 141(2):1238 tion cutting. Int J Adv Manuf Technol 65(1–4):35–42
27. Zhang C, Song Y, Ehmann K (2019) Design and kinematic
analysis of a novel decoupled 3D ultrasonic elliptical vibration Publisher's Note Springer Nature remains neutral with regard to
assisted cutting mechanism. Ultrasonics 95:79–94 jurisdictional claims in published maps and institutional affiliations.
28. Kurniawan R, JoKo T, Ping L, Kumaran S, Kiswanto G,
Guo P, Ehmann K (2017) Development of a two-frequency,

13

View publication stats