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

net/publication/323650667

Experimental Study on Fine Titanium Wire Drawing with Two Ultrasonically

Oscillating Dies

Article  in  IEEE Access · March 2018

DOI: 10.1109/ACCESS.2018.2811714

CITATIONS READS

6 1,940

4 authors, including:

Shen Liu Yuancai Yang

Harbin Institute of Technology 6 PUBLICATIONS   111 CITATIONS   
9 PUBLICATIONS   63 CITATIONS   
SEE PROFILE
SEE PROFILE

Xiaobiao Shan
Harbin Institute of Technology
124 PUBLICATIONS   1,143 CITATIONS   

SEE PROFILE

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

Study on mechanism and method of piezoelectric energy harvesting from low-speed water based on flow-induced composite vibration View project

Research on key technologies of intelligent handling robots for the large glass substrate in the clean environment View project

All content following this page was uploaded by Xiaobiao Shan on 25 October 2018.

The user has requested enhancement of the downloaded file.

Received January 15, 2018, accepted February 14, 2018, date of publication March 8, 2018, date of current version April 18, 2018.
Digital Object Identifier 10.1109/ACCESS.2018.2811714

Experimental Study on Fine Titanium Wire

Drawing with Two Ultrasonically Oscillating Dies
SHEN LIU , YUANCAI YANG, (Student Member, IEEE), TAO XIE, AND XIAOBIAO SHAN
School of Mechatronics Engineering, Harbin Institute of Technology, Harbin 150001, China
Corresponding author: Xiaobiao Shan (shanxiaobiao@hit.edu.cn)
This work was supported in part by the National Natural Science Foundation of China under Grant 51575130 and in part by the
Postdoctoral Scientific Research Staring Foundation of Heilongjiang Province under Grant LBH-Q12092.

ABSTRACT This paper investigates the influences of ultrasonic vibration on the two-pass titanium wire
drawing process through experiments conducted at different drawing speeds (99–480 mm/s) and under
various ultrasonic amplitudes (5.63–13.97 µm). Ultrasonic vibrators were designed, analyzed, and manufac-
tured to impose ultrasonic vibrations, along opposite directions, on the dies. Their dynamic characteristics
were predicted by finite element analysis and subsequently measured using specific instruments. The
interaction between the two vibrators, when driven by the same ultrasonic generator, resulted in the deviation
of coupled resonant frequency (50–60 Hz) and the decrease in maximum output amplitudes (about 4 µm).
A wire drawing equipment was developed to record the drawing forces under different experimental
conditions, and the surface morphologies of the drawn wires were then inspected with a scanning electron
microscope. Results show that increased ultrasonic amplitude contributes to greater drawing force reduction,
whereas higher drawing speed brings about the reverse effect. At the lowest drawing speed and with the
largest ultrasonic amplitude, the drawing force declined by over 50%. Both factors have influences on the
surface finish of the drawn wires. However, higher drawing velocity (480 mm/s) with the moderate ultrasonic
amplitude (9.69 µm) turned out to be most effective in removing the surface defects of the wire products.
This paper provides a low-cost, high-efficiency, and eco-friendly approach for the industrial production of
fine titanium wires at room temperature.

INDEX TERMS Metal forming, surface morphology, titanium wire, ultrasonic transducer.

I. INTRODUCTION at room temperature is further degraded by the strain hard-

Titanium and its alloys are distinguished by lightweight, high
mechanical strength, good corrosion resistance, and excellent
biocompatibility. These superior properties make titanium
wires suitable for use in automobile, aerospace, military, and
biomedical industries. Cotton et al. [1] reported that, com-
pared with steel springs, titanium springs have better corro-
sion resistance and can save up to 70% weight. Elias et al. [2]
summarized the applications of titanium wires in medical
and dental fields as joints and screws and emphasized the
necessity of using titanium alloys as the material to make den-
tal plant devices. However, compared with traditional metal
wires, such as aluminum wires and steel wires, their usage is
limited by two factors: the high cost of the raw material, and
the complexity of the processing procedure, as Boyer [3] has
pointed out. Due to high mechanical resistance and high yield
stress to tensile strength (Y/T) ratio, titanium and its alloys
are classified as less metallic materials, and their workability
ening effect. Siegert and Möck [4] have presented that the
conventional cold wire drawing process requires large draw-
ing force, induces high contact and friction forces, leading
to increased energy consumption and rough surface finish,
and therefore needs to be improved. Elevated temperature,
good lubrication condition, and suitable surface treatments,
which contribute to reducing the flow stress and improv-
ing the surface quality of the processed wires, are usually
adopted during the titanium wire drawing process. Sever-
denko and Zhilkin [5] have reported that heat treatments,
primarily annealing, are performed between passes to restore
the plasticity of the drawn wires. Ishii et al. [6] patented
a method for beta-titanium wire production which involves
cold drawing, heat treatment, solution treatment, and coat-
ing. Useful as these measures are, they are time-consuming,
polluting, and will remarkably increase the cost of the wire
products.

2169-3536
2018 IEEE. Translations and content mining are permitted for academic research only.
16576 Personal use is also permitted, but republication/redistribution requires IEEE permission. VOLUME 6, 2018
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
S. Liu et al.: Experimental Study on Fine Titanium Wire Drawing with Two Ultrasonically Oscillating Dies
Ultrasonic-assisted drawing, which provides all the ben-
efits of above-mentioned solutions, is a green technology
about introducing ultrasonic vibrations to the wire draw-
ing process. Siegert and Möck [4] reported that ultra-
sonic vibration is beneficial for decreasing the drawing
force and improving the surface finish of the drawn wire.
Gallego-Juarez and Graff [7] summarized the benefits of
ultrasonic vibrations in wire and tube drawing process:
increased drawing speed, prolonged die life, greater area
reduction per pass, etc. Zhou et al. [8] studied the influence of
ultrasonic vibration on the forming process of aluminum and
titanium, revealing that ultrasonic vibrations help to increase
the plasticity and refine the grain size of titanium. Blaha and
Langenecker first introduced power ultrasonics to the static
tensile test of monocrystalline zinc. They observed an abrupt
decrease of 40% in drawing force when ultrasonic vibration
of 800 kHz was imposed, and the drawing force bounced
back to its original value once the vibration was removed [9].
In their follow-up reports, Langenecker interpreted the phe-
nomenon as a result of the activation and increased mobility
of the lattice dislocations, which reduced the obstruction
at barriers during the metal deformation process [10], [11].
Eaves and his group reviewed the various hypotheses and
experimental results about ultrasonic rod drawing and other
metal forming processes, concluding that all these observa-
tions could be explained by the superposition of an oscillatory
load on a steady state force [12], [13], which was later termed
as stress superposition theory. They also proposed a multi-die
arrangement method in order to achieve greater area reduc-
tion [14]. Until now, no consensus has been reached on the
mechanism of ultrasonic assisted metal forming processes,
however, studies on relevant industrial applications have con-
tinued up to the present. Murakawa and Jin reported the
experimental results of applying radial ultrasonic vibration,
15 kHz, 4 µm, to the drawing of 85.8 mm stainless steel wire.
The operation realized the skin-pass ultrasonic wire drawing
using the chlorine-free lubricants, and expanded the critical
speed range of ultrasonic wire drawing by about 10 times,
to 55 m/min [15]. Susan et al. [16], [17] investigated the
effects of ultrasonic vibration during steel tube and wire
drawing process, both theoretically and experimentally. They
derived the mathematical equations to calculate the drawing
force according to the friction reversion mechanism. It was
found that ultrasonic vibration not only is responsible for the
decreased drawing force and friction force but also helps to
weaken the work hardening effect. For the drawing of fine
wires, Haiqun et al. [18] examined the possibility of intro-
ducing longitudinal vibrations to the industrial production of
brass wire. The experimental results presented: a 17% rise in
drawing velocity from 980 mm/s, under conventional condi-
tions, to 1126 mm/s, with ultrasonic vibration; a 7% decrease
in drawing force at the lowest drawing speed of 676 mm/s
and with a driving voltage of 240 V; prolonged tension
regulation period from 0.5 s to about 1.5 s; and improved
surface finish of drawn wires. Yang et al. [19] analyzed the
influence of longitudinal-torsional composite vibration on the
VOLUME 6, 2018
process of titanium wire drawing. They found that, compared
with pure longitudinal vibration, the addition of torsional
vibration could further decrease the drawing force by 23%,
however, causing slightly increased friction force, and worse
surface finish. Liu et al. [20] compared the cold dry drawing
process of titanium wires under three circumstances: without
vibration, with large amplitude longitudinal vibration, and
with longitudinal-torsional hybrid vibration. Considerable
improvement has been achieved with respect to the surface
finish of the drawn wires, with cracks and fragments almost
completely disappeared. However, a drawing force reduction
of only 5% was achieved, which has little commercial value
if force reduction is the sole criterion.
Although inconsistencies, sometimes contradictions, exist
among previous reports, generally, it is possible to reach the
following conclusions. First, regarding the feasibility, typical
ultrasonic-related phenomena, basically acoustic softening,
anti-friction, and enhanced surface finish were observed in
the wire drawing of both thick wires (rods, tubes) and fine
wires, no matter for traditional metallic materials or for
less deformable metals. Second, in terms of the efficiency,
the influence of ultrasonic vibration seems less pronounced in
fine wire drawing occasions, particularly for less deformable
materials. In addition, previous studies tended to focus on
the single-pass ultrasonic wire drawing rather than on multi-
pass wire drawing which is adopted in practical industrial
production. In this paper, we carried out an experimental
study on the two-pass titanium wire drawing process with two
dies oscillating along opposite directions. The influences of
wire drawing speed and vibration amplitude of the dies on
the drawing force and the surface finish of the processed wire
were studied. Ultrasonic vibrators were designed, analyzed,
tuned, and tested to impose ultrasonic vibrations on the dies.
Due to unavoidable machining and assembly errors, discrep-
ancies between the manufactured vibrators are unlikely to be
eliminated. Therefore, the coupling effects when they work
together were also investigated.
FIGURE 1. Exploded diagram of the designed ultrasonic vibrator for wire
drawing.
II. ULTRASONIC VIBRATOR DESIGN
In the wire drawing experiment, ultrasonic vibrations were
applied to the dies via the ultrasonic vibrators, the exploded
view of which is displayed in Fig. 1. The system is com-
prised of the ultrasonic transducer, the ultrasonic horn, and
the wire drawing die. The ultrasonic transducer is employed
16577
 S. Liu et al.: Experimental Study on Fine Titanium Wire Drawing with Two Ultrasonically Oscillating Dies

TABLE 1. Material parameters for the components of the vibrator.

FIGURE 2. 2D cross-section structure of the designed ultrasonic horn.

to convert electrical signals into mechanical movements in

the form of high-frequency longitudinal vibrations at the and 820 mm, respectively. Therefore, the length could be
output end. The output amplitude of the piezoelectric trans- calculated as 125.5 mm, and the amplification can be deter-
ducer is further amplified by the ultrasonic horn to meet the mined to be 3.61. More details about the calculations and
demands of subsequent titanium wire drawing experiments. design of ultrasonic horns could be found in the ultrasonics
As indicated in Fig. 1, the Langevin-type transducer mainly handbook. The 2D (two-dimensional) cross-section view of
consists of four components: the front mass which is made the designed stepped ultrasonic horn is illustrated in Fig. 2.
of 2A12 aluminum alloy, six pieces of PZT (lead zirconate The performance of the horn was further evaluated by FE
titanate) discs, the steel back mass, and the clamping bolt. (finite element) analysis in ANSYS R
, using APDL (ANSYS
Beryllium bronze sheets (not shown Fig. 1) are sandwiched Parametric Design Language) codes. Modal analysis results
between the PZT rings to serve as electrodes. These compo- show the horn resonates at the frequency of 19.929 kHz, with
nents are compressed tightly using the clamping bolt with a magnification ratio of 3.516, which is in good agreement
a preload of 20-30 MPa on the PZT discs. Higher acoustic with the calculated values.
impedance material, 2Cr13 stainless steel, was chosen for
the back mass to encourage mechanical wave propagation
towards the front mass. PZT-8 is the selected material for the
PZT rings, as it provides high mechanical quality factor, high
permittivity, and low dielectric losses, which is preferable
than other PZT materials for high power applications. The rel-
evant parameters about above materials are listed in Table 1.
Furthermore, to guarantee the performance of the transducer,
the contacting surfaces of the front and back masses with
the PZT rings or the electrodes should be ground to enhance
their flatness and roughness. The stepped half-wavelength FIGURE 3. FE model of the designed ultrasonic vibration system.

ultrasonic horn is also made of 2A12 aluminum alloy and

was connected to the transducer through threaded joints. The In order to study the dynamic performance of the ultra-
wire drawing die, 810 mm in diameter and 5.5 mm in height, sonic vibration system, an integrated FE model combining
is composed of a steel case and a diamond core and was the transducer, the horn, and the die was established in
attached to the thin end of the horn using the shrink-fitting ANSYS R
, as shown in Fig. 3. In the model, the PZT rings
technique. A shoulder was designed at the nodal position were meshed with the 3D (three-dimensional) 10-node cou-
where the ultrasonic amplitude reaches the minimum value so pled field element SOLID227, whereas the other components
that the vibrator could be installed on and dismounted from were meshed using the SOLID185 structural element. All of
the wire drawing experimental platform conveniently. the PZT rings were polarized along the thickness direction
The ultrasonic horn is an essential part of ultrasonic wire (Z direction), however, each adjacent two pieces of them were
drawing system as a consequence of the insufficient output installed along reverse directions, i.e., along or opposite to
amplitude of the ultrasonic transducer, which is normally the longitudinal vibration direction. The vibrator, as a whole,
4–5 µm, far below the minimum requirement of 10 µm. was constrained at the nodal position, as illustrated in Fig. 3.
Most frequently used ultrasonic horns can be classified as The piezoelectric strain matrix [d], the compliance coeffi-
stepped, exponential, conical, and catenoid, according to their cient matrix [s], and the permittivity matrix [ε S ] about the
longitudinal cross-section shapes. For an ultrasonic horn of PZT-8 [21] were defined by (1) to (3).
a specific type, the amplification factor is determined by  
the area ratio between the input and output surfaces. In the 0 0 −140
 0 0 −140 
current work, stepped ultrasonic was chosen because it can  
 0 0 350 
achieve the largest amplification factor, which is in proportion [d] =   × 10−12 m/V (1)
to the area ratio between the input and output surfaces. The  0 580 0  
 580 0 0 
ultrasonic horn was designed with an operating frequency
of 20 kHz, and the input and output diameters of 838 mm 0 0 0

16578 VOLUME 6, 2018

S. Liu et al.: Experimental Study on Fine Titanium Wire Drawing with Two Ultrasonically Oscillating Dies

FIGURE 4. Modal analysis result of the established FE model.

  FIGURE 5. Sensitivity analysis of the longitudinal resonant frequency

13.3 −4.26 −4.26 0 0 0 versus the lengths of the input and output sections.
 −4.26 13.3 −4.26 0 0 0 
 
 −4.26 −4.26 17.05 0 0 0 
[s] = 
 0

 0 0 41.16 0 0  
 0 0 0 0 41.16 0 
0 0 0 0 0 35.11
× 10−12 m2 /N (2)
 
h i 1800 0 0
εS =  0 1800 0  (3)
0 0 1500
Fig. 4 presents the contour of the deformed ultrasonic FIGURE 6. Transient response of ultrasonic vibrator (a) within 25 ms,
vibrator at its second order longitudinal resonant frequency. (b) from 24.2 ms to 24.6 ms.
It can be seen that the system resonated at 19.769 kHz, which
is quite close to the intended 20 kHz. And there was almost
no deformation at the nodal position, as had been expected. within 15 ms. Fig. 6(b) indicates the vibrator finally output a
Furthermore, the sensitivity of the resonant frequency to the stable sinusoidal vibration with an amplitude of 20.9 µm.
lengths of the input and output sections of the ultrasonic horn
was analyzed, the result of which is plotted in Fig. 5. It can
be found that the longitudinal resonant frequency of the ultra-
sonic vibration system shows an upward trend as the length of
the ultrasonic horn decreases, for both the input section and
the output section. However, the resonant frequency is more
sensitive to the length of the later than that of the former,
as illustrated in Fig. 5. This provides a convenient way to
regulate the resonant frequencies of the ultrasonic vibrators
in practice.
The time-varying output displacement of the vibrator at
FIGURE 7. Assembled prototypes of the ultrasonic vibration system.
the second resonant frequency was determined through tran-
sient analysis. During the simulation, an alternating voltage
of 19.769 kHz, 800 Vp-p was applied across the PZT rings, III. ULTRASONIC VIBRATORS TESTING
as illustrated in Fig. 3. The time duration was set as 25 ms, According to above simulation results, ultrasonic transducers
which equals to 500 vibration cycles. And the minimum and horns were manufactured, the assembly of which is dis-
recommended substep number was specified as 16 for each played in Fig. 7. The vibrators were fitted into the specially
vibration period. The dynamic response of the vibrator is designed holders through the shoulders machined at their
shown in Fig. 6. It can be seen that the amplitude of the nodal positions for subsequent tests, installation, and experi-
vibrator fluctuates remarkably at the beginning, however, sta- ments, because the shoulders themselves are too narrow to be
bilizes gradually as it reaches the steady value of about 20 µm clamped conveniently.

VOLUME 6, 2018 16579

 S. Liu et al.: Experimental Study on Fine Titanium Wire Drawing with Two Ultrasonically Oscillating Dies
FIGURE 8. Impedance measurement results: (a) impedance of vibrator A,
(b) phase angle of vibrator A, (c) impedance of vibrator B, (d) phase angle
of vibrator B.
The impedance testing equipment Agilent HP 4294A was
employed to measure the electrical impedance of the vibra-
tors, the results of which were plotted in Fig. 8. By comparing
Fig. 8(a) and Fig. 8(c), it can be seen that the impedance of
the vibrators reaches their minimum values of 9.95  and
13.17  at the frequencies of 19.645 kHz and 19.656 kHz,
respectively. That means the vibrators resonant at these fre-
quencies which are quite close to the targeted frequency
of 20 kHz. The difference between the resonant frequencies
of the two vibrators is merely 11 Hz, which is negligible.
Also, from Fig. 8(b) and Fig. 8(d), we can find that the phase
angles of the vibrators are 5.8◦ and 3.4◦ , far below 10◦ . The
power factors of the vibrators can be subsequently calculated
as 0.9948 and 0.9982, respectively, which means the vibrators
themselves can efficiently transfer electrical energy and there
is no need to conduct the impedance matching procedure.
What is more, as the phase angle difference between the
vibrators is only 2.4◦ , they almost vibrate at the same pace
when driven by the identical electrical signals. Similarly,
the vibrators would oscillate along opposite directions when
the driving signals have a phase difference of 180◦ . There-
fore, the vibrators are qualified for the following two-pass
wire drawing experiments. In general, as both the vibrators
have high mechanical quality factors, their frequency bands
(about 200 Hz) are relatively narrow, as shown in Fig. 8(b)
and Fig. 8(d). Fortunately, the dynamic characteristics of the
vibrators are in good consistency.
The mode shapes of the vibrators, as well as their oscil-
lating velocities at different frequencies, were measured
using the Polytec PSV-400-M2 scanning laser vibrometer,
the results of which are demonstrated in Fig. 9. During the
test, chirp signals with a sweep frequency band of 18–20 kHz,
and a voltage of 10 Vp-p were applied to the vibra-
tors. From Fig. 9(a) and 9(b), it can be found that
16580
FIGURE 9. Vibration mode detection results: (a) mode shape of vibrator
A, (b) mode shape of vibrator B, (c) oscillating velocity of vibrator A at
different frequencies, (d) oscillating velocity of die B at different
frequencies.
the dies can effectively produce longitudinal vibrations.
In Fig. 9(c) and 9(d), we can see that the vibration veloc-
ity of the output surface was influenced by the frequency
of the input chirp signals. The vibration velocities reached
their maximum values when the sweep frequency matched
with the resonant frequencies of the vibrators. As a result,
the two vibrators resonated respectively at the frequency
of 19.672 kHz and 19.664 kHz, with a negligible difference
of 8 Hz. On the other hand, if the frequency of the driving
voltage deviates from the resonant frequency, the vibration
amplitude or velocity would drop significantly, considering
the narrow bandwidth of the ultrasonic vibrators. In Fig. 9(c)
and 9(d), when the frequency deviation exceeds 130 Hz,
the output amplitude and velocity are reduced by 50%, even
though the driving voltage remains unchanged. The rela-
tionship between the frequency deviation and the oscillating
amplitude (or velocity) reduction ratio is depicted in Fig.10.
Although the frequency difference between the vibrators
is rather insignificant, it is very difficult to be eliminated
entirely. When driven by alternating signals with the same
ultrasonic generator, the two vibrators would interfere with
each other. For this reason, the coupled vibration mode
of the vibrator when they work together was examined
under the same sweep signal, the result of which is shown
in Fig. 11. It could be seen that the output ends of the vibra-
tors arrived at their maximum displacement simultaneously.
Fig. 11(b) shows, however, that the maximum value of their
coupled oscillating velocity occurs at the coupled frequency
of 19.727 kHz, which is 55 Hz and 63 Hz higher than the reso-
nant frequencies of each vibrator. In other words, the coupling
effect between the vibrators whose dynamic performances are
in good agreement, would not exert remarkable influence on
their phase angle and coupled resonant frequency.
The real-time output velocities of the ultrasonic vibra-
tors were measured by the Doppler Vibrometer using the
VOLUME 6, 2018
S. Liu et al.: Experimental Study on Fine Titanium Wire Drawing with Two Ultrasonically Oscillating Dies
FIGURE 10. Variation of amplitude reduction ratio with frequency
deviation.
FIGURE 11. Vibration mode detection results when the vibrators work
simultaneously: (a) coupled-mode shape, (b) the relationship between
oscillating velocity and frequency.
FIGURE 12. Vibration measurement system with the Scanning Laser
Doppler Vibrometer.
single-point velocity determination technique, the construc-
tion of the testing equipment is shown in Fig. 12. During
the test, a point on the output surface of the vibrator was
selected as the target point. Then, the position of the scanning
head was adjusted so that the laser beam could focus on the
point. Next, 100 kHz was assigned as the sampling frequency
of the vibration testing instrument, and alternating voltages
VOLUME 6, 2018
FIGURE 13. Velocity curves of (a) die A, (b) die B.
of 800 Vp-p , 19.672 kHz and 19.664 kHz in frequency were
applied to the piezoelectric transducers. The changing curves
of the vibration velocities of die A and die B with time are
plotted in Fig. 13.
It can be seen that the output velocities of the two vibrators
vary with time periodically in the form of sinusoidal curves,
the amplitudes of which are 2.457 m/s and 2.288 m/s, respec-
tively. Based on this, the corresponding vibration amplitudes
of the individual vibrators, under the same driving voltage,
could be calculated as 19.88 µm and 18.51 µm, which are
quite close to the FE simulation result of 20.91 µm. Simi-
larly, the actual amplitudes of the vibrators working at their
resonant frequencies of 19.672 kHz and 19.664 kHz, as the
driving voltage increased from 100 Vp-p to 800 Vp-p , were
calculated, the results of which have been plotted in Fig. 14.
We can find that the changing curves of the vibrators are
alike. With the increment of the driving voltage, the output
amplitude rises almost linearly. However, the amplitude of
vibrator A is a little higher than that of vibrator B, when
excited by the same voltage. This can be attributed to the
relatively smaller impedance of vibrator A, as the energy
consumption of an ultrasonic vibrator is in reverse proportion
to its equivalent impedance.
FIGURE 14. Independent output amplitudes under different driving
voltages for (a) vibrator A, and (b) vibrator B.
When die A and die B work simultaneously in the same
circuit, the interaction between them has to be taken into
account. In this condition, the vibrators are working at their
coupled resonant frequency of 19.727 kHz, with the driving
voltages ranging from 100 Vp-p to 800 Vp-p . The changing
curves of their output amplitude with the increased driving
voltages are shown in Fig. 15. The maximum output ampli-
tudes of the vibrators are 15.77 µm and 13.97 µm, much
smaller than their corresponding values when working sep-
arately. Three reasons might be responsible for this. The first
16581
 S. Liu et al.: Experimental Study on Fine Titanium Wire Drawing with Two Ultrasonically Oscillating Dies
FIGURE 15. Coupled output amplitudes under different driving voltages
for (a) vibrator A, (b) vibrator B.
is the frequency deviation between the coupled frequency and
their independent resonant frequencies. As shown in Fig. 10,
a 55 Hz to 60 Hz gap in frequency would cause about 17%
reduction in their output amplitudes. The second reason is the
impedance difference between the vibrators, which should
account for the inferior performance of vibrator B. The third
reason might be the inadequate output power of the ultrasonic
generator. The energy consumption will be reduced for each
vibrator when they work together compared with when they
work alone. However, it is still reasonable to use the vibrators
for the subsequent wire drawing experiment, as their output
amplitudes are quite close and higher than 10 µm, which
is sufficient for experimental purpose according to previous
reports.
FIGURE 16. Established experimental equipment for two-pass ultrasonic
wire drawing.
IV. WIRE DRAWING AND INSPECTION EQUIPMENT
The arrangement of the developed experimental equip-
ment for double-die ultrasonic wire drawing is illustrated
in Fig. 16. The equipment is mainly composed of three
systems: the wire feeding system to collect and provide the
raw wire; the ultrasonic vibration system to impose ultrasonic
vibration on the dies; and the winding system which provides
the drawing force and collects the drawn wire. Specifically,
the ultrasonic vibration system or the vibrator consists of
a piezoelectric transducer, an ultrasonic horn, and the wire
drawing die, as aforementioned. The driving signals of the
R
vibrators, with a phase difference of 180◦ , are provided
by the identical ultrasonic generator so that the two dies
oscillate along the opposite directions. The frequencies of the
driving signals were set as the coupled resonant frequency
16582
FIGURE 17. The relationship between the wire drawing velocity and the
frequency of the modulator.
of 19.727 kHz and remained unchanged. Their voltages, how-
ever, could be adjusted to regulate the oscillating amplitudes
of the dies. The vibrators were clamped by the specially
designed T-shaped steel frame, which, as a whole, were sub-
sequently installed onto the Hiwin slide blocks. The slide
blocks can move smoothly along the slide guideways and
could be fixed at appropriate positions conveniently by the
lock bolts. The winding system is composed of an electromo-
tor, a reduction gearbox, a torque sensor, a coiler drum, and
the clutches between them. The wire drawing speed is deter-
mined by the output rotational velocity of the electromotor,
which is in turn regulated by the gearbox and the frequency
modulator. The relationship between the input frequency
of the modulator and the wire drawing velocity is plotted,
as shown in Fig. 17. It can be seen that the wire drawing
velocity increases linearly from 30 mm/s to 530 mm/s as the
input frequency changes from 10 Hz to 90 Hz. The two sides
of the torque sensor were respectively connected to the motor
and the coiler drum to measure the wire drawing force. The
torque values were acquired and shown in real time on the
screen of the digital display. During the experiment, the raw
titanium wire, with a diameter of 0.4 mm, was pulled through
two oscillating dies, with a diameter reduction of 0.02 mm
for each pass. The material of the wire is commercially
pure titanium (TA2), and graphite has been employed as the
lubricant to reduce wire break-off ratio.
Although it is convenient for us to observe the changing
trend of the drawing force by reading the digital display, its
refresh rate is inadequate for the average drawing force calcu-
lation. Therefore, a data acquisition system was established
in LabVIEW
, the block diagram of which is illustrated
in Fig. 18. The system is made up of the torque sensor,
the NI 9215 data acquisition (DAQ) card, and the computer,
as shown in Fig. 16. The output analog signal of the torque
sensor is transferred to the NI 9215 module and converted into
VOLUME 6, 2018
S. Liu et al.: Experimental Study on Fine Titanium Wire Drawing with Two Ultrasonically Oscillating Dies
FIGURE 18. Block diagram established in LabVIEW for wire drawing force
measurement.
the digital signal, which is subsequently transmitted to the
computer through LabVIEW 2016 interface. When subjected
to the torque induced by the wire drawing force, the frequency
of the output signal generated by the sensor would change
correspondingly. This change will be detected by the program
after Fourier transformation, and therefore, the torque or the
drawing force could be calculated based on the relationship
between the torque and the frequency deviation. The cal-
culated drawing force was displayed on the front panel of
the program in real time and would be written to assigned
local file path automatically. After the experiment, specimens
of the drawn wire were submerged into a jar of acetone
solution, which was later put into an ultrasonic cleaner to
remove the dirt and stains adhering to the surface of the drawn
wire. Then, Hitachi S4300 field-emission scanning electron
microscope (FESEM) was employed to inspect the surface
morphologies of the drawn wire samples, and the appearance
of the instrument is shown in Fig. 19. The image of the
morphological features of the drawn wire specimen when
magnified to 25 times is presented in the monitor display,
as shown in the right subfigure.
FIGURE 19. Appearance of the surface morphology inspection
instrument.
V. RESULTS
Fig. 20 shows the changing curves of wire drawing force
along with the drawing velocity when ultrasonic vibrations
of various amplitudes were applied to the dies. During the
VOLUME 6, 2018
FIGURE 20. Variation of drawing force with wire drawing velocity when
ultrasonic vibrations of different amplitudes were applied.
experiment, the frequency of the ultrasonic generator was
set as the coupled frequency of the vibrators, 19.727 kHz,
and remained unchanged. A phase difference of 180◦ existed
between the driving signals of the vibrators so that the two
dies oscillated oppositely to each other. The output volt-
age of the generator was adjusted to regulate the oscillating
amplitude of the dies which changes from 0 µm (without
vibration) to around 15 µm. The minor difference between
the output amplitudes of the two vibrators, as have been
analyzed in above sections, was ignored. Unless otherwise
stated, the amplitude mentioned below means the amplitude
of die B or vibrator B. The input diameter of the raw wire was
0.4 mm, which was separately reduced by 0.02 mm when the
wire was pulled through die A and die B.
Under traditional wire drawing conditions, where the
vibration amplitude is 0 µm, the value of the drawing force
stabilizes at around 19 N with slight fluctuations, although
the drawing speed increased significantly from 100 mm/s
to 500 mm/s. Under ultrasonic wire drawing conditions,
however, the drawing force shows an upward trend with the
increment of the drawing velocity, especially when ultrasonic
amplitude exceeds 7 µm. In addition, although the drawing
force tends to be higher at increased drawing speed, it never
surpasses the original wire drawing force without ultrasonic
vibrations. Moreover, comparing those changing curves cor-
responding to different vibration amplitudes, the wire draw-
ing force is smaller when ultrasonic vibrations of larger
amplitude were superimposed. But exceptions may occur
when the ultrasonic amplitude was relatively lower, below
7 µm for instance. At the drawing speed of 99mm/s and
298 mm/s, the drawing forces tend to be lower than sur-
rounding data points. This might be caused by the inferior
performance of the motor or the gearbox, as the offsets about
these points are almost the same. Therefore, the following
conclusion could be drawn: first, ultrasonic vibration con-
tributes to reducing drawing force; second, increased draw-
ing velocity weakens the drawing force decreasing effect;
16583
 S. Liu et al.: Experimental Study on Fine Titanium Wire Drawing with Two Ultrasonically Oscillating Dies
TABLE 2. Influence of drawing velocity on drawing force reduction when
ultrasonic amplitude is 10.93 mm.
third, larger ultrasonic amplitude boosts the drawing force
reduction phenomenon. To demonstrate the weakened force
reduction phenomenon under relative higher drawing veloc-
ity conditions more precisely, the drawing forces and the
relevant reduction rates at different drawing speeds, but with
the same ultrasonic amplitude of 10.93 µm, are summarized
in Table 2.
FIGURE 21. Variation of drawing force with ultrasonic amplitude at
different drawing velocities.
Fig. 21 presents the variation curves of wire drawing
force with the amplitude of the imposed ultrasonic vibration
at different drawing speeds. The experiment configuration
stayed unchanged as before. A remarkable decline in the
wire drawing force could be easily observed as the ultrasonic
amplitude increases from about 5 µm to roughly 14 µm.
In more detail, a very sharp decrease appears when the
ultrasonic amplitude is below 10 µm. However, this trend
is moderated when the ultrasonic amplitude is greater than
10 µm. Wire drawing velocity also exerts some influence
on these changing curves. For example, at the lowest draw-
ing speed of 99 mm/s, the corresponding changing curve
is relative irregular compared with curves at other drawing
speeds. Furthermore, the curves relating to higher drawing
velocities are generally above those for lower drawing veloc-
ities, although they interweave with each other occasionally.
16584
We can arrive at similar conclusions that greater vibration
amplitude and lower drawing speed are beneficial for cutting
down the drawing force. Within the variation range given
in this work, the vibration amplitude seems to have a more
significant influence on the drawing force than the drawing
speed. To illustrate the influence of the vibration amplitude
more explicitly, the reduction rate of the drawing force at the
same drawing speed of 298 mm/s, but with different vibration
amplitudes, were calculated, as listed in Table 3.
TABLE 3. Influence of vibration amplitude on drawing force reduction at
the drawing speed of 298 MM/S.
FIGURE 22. Surface morphologies of drawn wire specimens at the
drawing speed of 99 mm/s.
To further investigate how the vibration amplitude and
wire drawing speed affect the surface finish of the processed
titanium wires, specimens of the drawn wires produced at
the speeds of 99 mm/s, 298 mm/s, 480 mm/s, and with the
vibration amplitudes of 0 µm (without vibration), 5.63 µm,
9.69 µm, and 13.97 µm were examined using the FESEM
instrument, the results are shown in the following figures.
Fig. 22 presents the surface morphologies of the drawn wires
at the drawing speed of 99 mm/s with a magnification fac-
tor of 250. In the upper left subfigure, it is easy to find
that galling, spalling, checking, with tiny chips and debris,
distribute evenly on most regions of the processed wire
surface when conventional drawing method was adopted.
VOLUME 6, 2018
S. Liu et al.: Experimental Study on Fine Titanium Wire Drawing with Two Ultrasonically Oscillating Dies
In the other subfigures, with ultrasonic vibrations applied,
the chips and debris were removed, making their surfaces
much cleaner. In addition, the vibration amplitude also affects
the surface quality of the processed wires. When the applied
ultrasonic amplitude is insufficient, as shown in Fig. 22(b),
cracks and fragments still exist in the most areas of the
outer surface. However, these defections would disappear
with the increment of the vibration amplitude, as shown in the
third subfigure. But, when the amplitude surpasses a specific
value, to 13.97 µm for instance, flaws such as furrows and
fragments reappear, although this only occurs in very lim-
ited regions, as the lower right subfigure shows. Therefore,
it can be concluded that ultrasonic vibration is beneficial
for improving the surface finish of the drawn wire, but a
moderate vibration amplitude is more suitable at a relative
lower drawing speed.
FIGURE 23. Surface morphologies of drawn wire specimens at the
drawing speed of 298 mm/s.
Fig. 23 shows the surface appearance of the processed
wires drawn at the velocity of 298 mm/s when the dies
oscillate with different amplitudes. Compared with Fig. 22,
the surface smoothness of these drawn wire samples has been
improved noticeably when the experiment was performed at
a relatively higher speed. More specifically, chips and debris
no longer adhere to the surface of the drawn wire even under
the traditional drawing conditions. Surface flaws of spalling
have been eliminated, although furrows and cracks along the
wire drawing direction appear instead. For ultrasonic wire
drawing conditions, the surface quality improvement effect
is influenced by the amplitude of the imposed ultrasonic
vibration. When the dies oscillate with a smaller amplitude,
for instance, 5.63 µm, a higher drawing speed can remarkably
improve the surface finish of the drawn wire, comparing
Fig. 22(b) and 23(b). However, for the wires drawn with
R
a larger ultrasonic vibration amplitude, there is almost no
observable improvement in their surface quality, as shown in
VOLUME 6, 2018
Fig. 22(d) and 23(d). As for the samples drawn in moderate
vibration amplitude occasions, the surface quality has also
been apparently improved.
FIGURE 24. Surface morphologies of drawn wire specimens at the
drawing speed of 480 mm/s.
Fig. 24 compares the surface finish of the titanium wire
samples drawn at the velocity of 480 mm/s, with different
vibration amplitudes of the dies. Still, all the samples show a
better surface finish than their counterparts in Fig. 22 and 23,
especially for those samples drawn with the assistance with
ultrasonic vibrations. For the sample drawn without ultra-
sonic vibration, there seems no change in its surface appear-
ance. However, the area with the surface defects has narrowed
down. In ultrasonic wire drawing conditions, the influence
of a higher drawing speed is much more impressive. All the
three samples, in spite of the difference in oscillating ampli-
tudes, have achieved a much better surface finish, especially
for the sample drawn with the largest vibration amplitude
of 13.97 µm. Under this drawing speed, the surface mor-
phologies of the three ultrasonic drawing samples are alike,
except several shallow furrows appear in Fig. 24(b) and an
unexpected debris shown in Fig. 24(c). That means ultrasonic
vibration is still beneficial for enhancing the surface quality of
the drawn wires at a relative higher drawing speed. However,
its amplitude is no longer make a big difference in this aspect.
In other words, when the oscillating amplitude reaches a
specific value but the surface finish is still bad, increasing the
drawing speed would be a better choice to tackle the problem.
VI. CONCLUSION
In this paper, two-pass titanium wire drawing with the dies
oscillating along opposite directions were studied experi-
mentally. The ultrasonic vibration system was designed and
analyzed using an FE model established in ANSYS
. The
influences of the lengths of ultrasonic horns on the res-
onant frequency, as well as the dynamic response of the
16585
 S. Liu et al.: Experimental Study on Fine Titanium Wire Drawing with Two Ultrasonically Oscillating Dies
vibrator, were obtained by the numerical analysis. Based
on the results, two prototype vibrators were manufactured,
and their impedance characteristics and mode shapes were
measured using specific instruments. The performances of
the vibrators were in good agreement with the simulation
results, but inevitable deviations still existed between them
with respect to resonant frequency and impedance charac-
teristics. Therefore, the coupling effects of the two vibrators
were further investigated to make sure they could oscillate
synchronously. At last, a two-pass wire drawing equipment
was developed, and the wire drawing experiments were con-
ducted. The drawn wire samples were inspected with the
FESEM equipment, and the corresponding drawing forces
were recorded. Based on the result, we can arrive the follow-
ing conclusions:
1) The coupling effect between the vibrators should not
be neglected, as the coupled working frequency was
50–60 Hz higher than their own resonant frequen-
cies and their maximum output amplitudes were low-
ered by nearly 4 µm compared with when they work
independently.
2) Ultrasonic vibration is beneficial for decreasing the
drawing force; however, the effectiveness is determined
by both the vibration amplitude of the dies and the
wire drawing velocity. Larger amplitude contributes to
greater drawing force reduction, whereas higher draw-
ing speed exerts the opposite effect. At the drawing
velocity of 99 mm/s, with the ultrasonic amplitude
of 13.97 µm, the drawing force could be decreased by
more than 50%.
3) Increased drawing speed helps to improve the surface
finish of the drawn wire, no matter with or without the
assistance of ultrasonic vibrations.
4) Ultrasonic wire drawing is advantageous to enhance
the surface smoothness of the drawn wires, regardless
of at the lower, moderate, or higher drawing speeds.
But the optimal vibration amplitude is influenced by
the wire drawing velocity. When wire drawing was
conducted at or below the moderate speed of 298 mm/s,
a medium vibration amplitude of 9.69 µm is the best.
However, in higher drawing speed (480 mm/s) occa-
sions, the vibration amplitude could change within a
wider range (5.63–13.97 µm) without inducing unac-
ceptable surface morphologies.
REFERENCES
[1] J. D. Cotton et al., ‘‘State of the art in beta titanium alloys for airframe
applications,’’ JOM, vol. 67, no. 6, pp. 1281–1303, Jun. 2015.
[2] C. N. Elias, J. H. C. Lima, R. Valiev, and M. A. Meyers, ‘‘Biomedical
applications of titanium and its alloys,’’ JOM, vol. 60, no. 3, pp. 46–49,
Mar. 2008.
[3] R. R. Boyer, ‘‘An overview on the use of titanium in the aerospace indus-
try,’’ Mater. Sci. Eng. A, vol. 213, nos. 1–2, pp. 103–114, Aug. 1996.
[4] K. Siegert and A. Möck, ‘‘Wire drawing with ultrasonically oscillat-
ing dies,’’ J. Mater. Process. Technol., vol. 60, nos. 1–4, pp. 657–660,
Jun. 1996.
[5] V. P. Severdenko and V. Z. Zhilkin, ‘‘Heat treatment in drawing of titanium
wire,’’ Met. Sci. Heat Treat. Met., vol. 1, no. 10, pp. 21–24, Oct. 1959.
16586
[6] I. Naoki, K. Takashi, S. Shin, Y. Hideki, and N. Ichiro, ‘‘Beta titanium wire,
method for its production and medical devices using beta titanium wire,’’
European Patent 1 083 243 (A2), Mar. 14, 2001.
[7] J. Gallego-Juarez and K. F. Graff, ‘‘Ultrasonic metal forming: Process-
ing,’’ in Power Ultrasonics. Applications of High-Intensity Ultrasound.
Waltham, MA, USA: Elsevier, 2014, pp. 377–431.
[8] H. Zhou, H. Cui, and Q. H. Qin, ‘‘Influence of ultrasonic vibration on
the plasticity of metals during compression process,’’ J. Mater. Process.
Technol., vol. 251, pp. 146–159, Jan. 2018.
[9] F. Blaha and B. Langenecker, ‘‘Tensile deformation of zinc crystal under
ultrasonic vibration,’’ Naturwissenschaften, vol. 42, no. 20, p. 556, 1955.
[10] B. Langenecker, ‘‘Effects of ultrasound on deformation characteristics of
metals,’’ IEEE Trans. Sonics Ultrason., vol. 13, no. 1, pp. 1–8, Mar. 1966.
[11] B. Langenecker, ‘‘Dislocation damping in macrosonic fields,’’ Phys. Rev.,
vol. 145, no. 2, pp. 487–491, 1966.
[12] C. E. Winsper and D. H. Sansome, ‘‘The superposition of longitudinal
sonic oscillations on the wire drawing process,’’ in Proc. Inst. Mech. Eng.,
vol. 183. 1968, pp. 545–562.
[13] A. E. Eaves, A. W. Smith, W. J. Waterhouse, and D. H. Sansome, ‘‘Review
of the application of ultrasonic vibrations to deforming metals,’’ Ultrason-
ics, vol. 13, no. 4, pp. 162–170, Jul. 1975.
[14] C. E. Winsper and D. H. Sansome, ‘‘A study of the mechanics of wire
drawing with a superimposed ultrasonic stress,’’ in Proc. Adv. Mach. Tool
Design Res., 1970, pp. 553–565.
[15] M. Murakawa and M. Jin, ‘‘The utility of radially and ultrasonically
vibrated dies in the wire drawing process,’’ J. Mater. Process. Technol.,
vol. 113, nos. 1–3, pp. 81–86, Jun. 2001.
[16] M. Susan, L. G. Bujorenau, D. G. Gǎluşcǎ, C. Munteanu, and V. Iliescu,
‘‘Influence of the relative deformation rate on tube processing by ultrasonic
vibration drawing,’’ Rev. Metalurgia, vol. 40, no. 2, pp. 109–117, 2004.
[17] M. Susan, C. Munteanu, and M. Mantu, ‘‘On the drawing in ultrasonic field
of metallic wires with high mechanical resistance,’’ J. Optoelectron. Adv.
Mater., vol. 7, no. 2, pp. 637–645, Apr. 2005.
[18] H. Qi, X. Shan, and T. Xie, ‘‘Design and experiment of the high speed wire
drawing with ultrasound,’’ Chin. J. Mech. Eng., vol. 22, no. 4, pp. 580–586,
2009.
[19] C. Yang, X. Shan, and T. Xie, ‘‘Titanium wire drawing with longitudinal-
torsional composite ultrasonic vibration,’’ Int. J. Adv., vol. 83, nos. 1–4,
pp. 645–655, Mar. 2016.
[20] S. Liu, X. Shan, K. Guo, Y. Yang, and T. Xie, ‘‘Experimental study on
titanium wire drawing with ultrasonic vibration,’’ Ultrasonics, vol. 83,
pp. 60–67, Feb. 2017.
[21] C. H. Sherman and J. L. Butler, ‘‘Small-signal properties of piezoelectric
ceramics,’’ in Transducers and Arrays for Underwater Sound. New York,
NY, USA: Springer, 2007, pp. 551–552.
SHEN LIU was born in Yongcheng, Henan, China,
in 1989. He received the B.S. degree from the
School of Mechanical Engineering, Zhejiang Uni-
versity of Technology, in 2012, and the M.S.
degree from the School of Mechatronics Engi-
neering, Harbin Institute of Technology, China,
in 2014, where he is currently pursuing the Ph.D.
degree. His research interests include energy har-
vesting and ultrasonic-assistant wire drawing.
YUANCAI YANG was born in Zigong, Sichuan,
China, in 1993. He received the B.S. degree from
the School of Mechanical Engineering, Chongqing
University, in 2015, and the M.S. degree from
the School of Mechatronics Engineering, Harbin
Institute of Technology, China, in 2017, where
he is currently pursuing the Ph.D. degree. His
research interests include self-sensing actuator and
ultrasonic-assistant wire drawing.
VOLUME 6, 2018
S. Liu et al.: Experimental Study on Fine Titanium Wire Drawing with Two Ultrasonically Oscillating Dies
TAO XIE was born in 1965. He received the
B.S. degree from the Department of Precision
Instrumentation, the M.S. degree from the School
of Astronautics, and the Ph.D. degree from the
School of Mechatronics Engineering, Harbin Insti-
tute of Technology, in 1987, 1990, and 2003,
respectively. Since 2003, he has been a Profes-
sor with the School of Mechatronics Engineer-
ing, Harbin Institute of Technology. His research
interests include ultrasonic-assisted wire drawing,
piezoelectric energy harvesting, smart structures, and dynamic simulation.
VOLUME 6, 2018
View publication stats
XIAOBIAO SHAN was born in Fuzhou, Jiangxi,
China, in 1977. He received the B.S. degree from
the School of Automotive Engineering, Harbin
Institute of Technology, in 2001, the M.S. and
Ph.D. degrees from the School of Mechatron-
ics Engineering, Harbin Institute of Technology
in 2004 and 2008, respectively. He is currently an
Associate Professor with the School of Mechatron-
ics Engineering, Harbin Institute of Technology.
His research interests include energy harvesting,
ultrasonic-assisted wire drawing, nanogenerator, and dynamic simulation.
16587