Ultrasonics 75 (2017) 80–90

Contents lists available at ScienceDirect

Ultrasonics
journal homepage: www.elsevier.com/locate/ultras

Feasibility study of ultrasonic elliptical vibration-assisted reaming of

carbon fiber reinforced plastics/titanium alloy stacks
Daxi Geng a, Deyuan Zhang a,⇑, Zhe Li a, Dapeng Liu b
a
School of Mechanical Engineering & Automation, Beihang University, No. 37 Xueyuan Road, Haidian District, Beijing 100191, China
b
Chengdu Aircraft Industrial (Group) Co., Ltd., Huangtianba, Qingyang District, Chengdu 610000, China

a r t i c l e i n f o a b s t r a c t

Article history: The production of high quality bolt holes, especially on the carbon fiber reinforced plastics/titanium alloy
Received 12 July 2016 (CFRP/Ti) stacks, is essential to the manufacturing process in order to facilitate part assembly and
Received in revised form 24 October 2016 improve the component mechanical integrity in aerospace industry. Reaming is widely used as a manda-
Accepted 17 November 2016
tory operation for bolt holes to meet the strict industry requirements. In this paper, the ultrasonic ellip-
Available online 18 November 2016
tical vibration-assisted reaming (UEVR) which is considered as a new method for finish machining of
CFRP/Ti stacked holes is studied. The paper outlines an analysis of tool performance and hole quality
Keywords:
in UEVR compared with that in conventional reaming (CR). Experimental results show that the quality
CFRP/Ti stacks
Reaming
of holes was significantly improved in UEVR. This is substantiated by monitoring cutting force, hole geo-
Cutting force metric precision and surface finish. The average thrust forces and torque in UEVR were decreased over
Surface finish 30% and 60% respectively. It is found that, during first 45 holes, better diameter tolerance (IT7 vs. IT8),
smaller diameter difference of CFRP and Ti holes (around 3 lm vs. 12 lm), better geometrical errors were
achieved in UEVR as compared to CR. As for surface finish, both of the average roughness and hole surface
topography in UEVR were obviously improved.
Ó 2016 Elsevier B.V. All rights reserved.

1. Introduction CFRP, continuous and high temperature chips passing through

Nowadays, the CFRP and metal alloy stacked multilateral struc-
tures, especially the CFRP/Ti stacks, are widely used for the wing
segments and fuselage of an aircraft due to their high strength to
weight ratio, and excellent corrosion and heat resistance [1]. A
lot of benefits could be brought, such as better maneuverability,
the decrease of fuel consumption and the increase of the payload.
In order to fasten the stacks, a large number of holes need to be
produced for various purposes such as bolting and riveting. Due
to the safety reasons, the requirements of hole tolerance and hole
surface integrity are much higher than other industries, especially
for the bolt holes which mainly distribute on the high mechanical
load bearing components such as aircraft wing and tail-plane [2].
Because of the advantage in productivity, it is usually desirable
for the CFRP/Ti stacks to be drilled in one pass with a single drill.
Meanwhile, many drawbacks could be brought during drilling of
CFRP/Ti stacks with a single drill [2–7]. The modulus of elasticity
of the materials causes different elastic deformations and therefore
results in varying tolerances along the entire stacked hole [3]. Fur-
thermore, when the titanium alloy is stacked at the bottom of
⇑ Corresponding author.
E-mail address: zhangdy@buaa.edu.cn (D. Zhang).
the CFRP section deteriorate the quality of the machined CFRP
holes. Even with the extensive amount of studies done to improve
the hole quality of CFRP/metal alloy stacks, drilling the stacks in
one pass with a single drill is still a challenging task which usually
results in oversized holes (around IT10) and drilling-induced dam-
ages characterized by poor hole geometric defects, thermal injuries
and physical imperfections [2,4,5]. Therefore, in order to meet the
strict industry requirements for stacked bolt holes, reaming pro-
cess is widely used as a mandatory finishing operation to achieve
narrow hole tolerance and remove drilling-induced damages
[2,6]. In aircraft assembly, the mandatory finishing operation for
bolt holes followed by rough drilling is a compromise in drilling
ductile titanium alloy and highly abrasive carbon fiber. For exam-
ple, reaming the stacks with carbide reamers followed by CFRP and
Ti rough drilling is still widely used in the CFRP/Ti stacks drilling in
Boeing Company [7]. Although conventional reaming process
under adequate lubrication conditions is widely used to drill metal
materials in other industries [8,9], it is still a challenge to obtain
high quality CFRP/Ti stacked bolt holes in terms of excellent geo-
metrical characteristics and surface finish under dry or near dry
condition which is required in aircraft assembly. In conventional
dry reaming of CFRP/Ti stacked holes, the typical imperfections
include unstable hole geometric accuracy (diameter tolerance,

http://dx.doi.org/10.1016/j.ultras.2016.11.011
0041-624X/Ó 2016 Elsevier B.V. All rights reserved.
 D. Geng et al. / Ultrasonics 75 (2017) 80–90
roundness and cylindricity), large diameter difference between
CFRP and Ti sections, unstable hole surface roughness and damages
at CFRP/Ti interface [10]. These imperfections of bolt holes are usu-
ally considered fatal and deteriorate the structure integrity of the
machined components [11–13]. According to the study by Liu
et al. [10], in order to meet the strict industry requirements for
large diameter CFRP/Ti bolt holes (>U7.5 mm), such as narrow hole
tolerance (IT8), small surface roughness and free delamination in
the laminates, a multi-step reaming operation in which both of
the machining allowance and feed rate are strictly controlled is
used as the mainstream finish drilling process for CFRP/Ti bolt
holes in current industrial practice. The typical drawback of this
multi-step reaming operation is low productivity. Therefore, in
order to improve the hole quality and process productivity, it is
essential to develop cost-effective finish drilling processes for
CFRP/Ti bolt holes under dry condition.
Recently, ultrasonic vibration-assisted machining has been
widely used for drilling of CFRP, titanium alloy and CFRP/Ti stacks.
As a nontraditional drilling method, rotary ultraosnic machining
(RUM) has been successfully used in drilling of Ti [14,15] and CFRP
[16–18] with low cutting force, long tool life, low surface rough-
ness, negligible Ti burr and CFRP delamination. Later, RUM was
extended to drill CFRP/Ti stacks [19]. The experiment results
showed that compared with other methods for drilling of CFRP/Ti
stacks, RUM resulted in lower cutting forces and CFRP surface
roughness, smaller groove depth and hole size variation, longer
tool life, no obvious Ti burr and CFRP entrance delamination, and
the same Ti surface roughness as other methods. In order to wash
away the swarf and prevent tool jamming, pumped coolant was
used in RUM. However, only dry or near dry condition is allowed
in aircraft assembly, which constraints the wide application of
RUM in aircraft assembly. In order to improve the hole quality
under dry condition, rotary ultraosnic elliptical machining (RUEM)
was introduced to drill CFRP in aircraft assembly with the advan-
tages in terms of reduction of cutting force, longer tool life, preven-
tion of tool jamming and improvement of chip removal rate
compared to conventional drilling [20,21]. However, the used
brazed diamond core drill in RUEM can result in poor hole preci-
sion and surface roughness. As a result, the finishing operation is
usually mandatory to meet the required hole tolerance and surface
roughness. From the previous works [22–26] on ultrasonic ellipti-
cal vibration-assisted cutting (UEVC), it suggests that the UEVC
method could offer a satisfied answer to avoid the typical disad-
vantages in conventional reaming of CFRP/Ti stacks under dry con-
dition. In [22], the reason of the machining accuracy improvement
in UEVC due to the reduction of cutting force was clarified and ver-
ified experimentally in comparison with conventional cutting and
cutting-directional vibration-assited cutting (CDVC). Besides the
improvement of machining accuracy, better surface integrity could
also be achieved in UEVC method compared to CDVC. Li et al. [23]
pointed that, the finish surface in CDVC could be destroyed due to
the collision and rub of the tool flank when the vibrating tool
leaves the cutting point of the workpiece. And he claimed that
the UEVC method could effectively avoid this collision problem.
Similar collision between tool flank and machined surface was also
found in axial-directional vibration-assisted reaming of difficult-
to-cut materials [24]. The investigations of Xu and Zhang [25,26]
showed that the cutting performance in UEVC of CFRP was more
effective than conventional cutting and CDVC in terms of better
chip removal, less tool wear and superior CFRP surface integrity.
However, the cutting speed in their experiment was controlled as
low as 1 m/min. The typical limitation for UEVC or CDVC is that
in order to achieve noticeable effect of ultrasonic vibration, the cut-
ting velocity should be controlled under the vibration velocity in
the cutting direction. This drawback could result in low efficiency.
In order to achieve the separation between the workpiece and the
81
tool with high cutting speed, researchers applied the vibration in
the feed direction in ultrasonic assisted turning (UAT) and also
achieved an improvement of the surface quality [27].
Various sources confirm that drilling CFRP/Ti stacked bolt holes
with high hole quality is still a challenge to manufacturing engi-
neers in aircraft assembly. To our best knowledge, insufficient
efforts have been made on the finish drilling of hybrid CFRP/Ti
structures and few studies have been reported on applying ultra-
sonic elliptical vibration assistance to machine CFRP/Ti stacked
holes in order to achieve more favorable cutting conditions. This
study proposes a novel finish drilling method for CFRP/Ti stacks
(i.e. ultrasonic elliptical vibration-assisted reaming, UEVR), and
aims to investigate the effects of ultrasonic elliptical vibration on
hole quality when reaming of the CFRP/Ti stacks. The results from
the reaming experiments are reported mainly on how the finish
hole quality of the CFRP/Ti stacks is changed due to the different
cutting mechanisms between UEVR and conventional reaming
(CR). Here, aspects such as cutting force, tool wear, hole geometry
precision, hole surface roughness as well as defects, CFRP hole
boundary quality and chip morphology will be studied to analyze
the cutting behaviors in UEVR.
2. Experiments
2.1. Apparatus of UEVR system
As Fig. 1 shows, the self-designed ultrasonic elliptical trans-
ducer which is a sandwiched piezoelectric structure consists of a
reamer, a front cylinder (made of aluminum alloy), two orthogonal
groups of piezoelectric plates (PZT’s), and a back cylinder (made of
stainless steel). An elliptical mode vibration in the plane of reamer
cross section at tool end could be generated by the synthesis of the
two bending vibrations incited by these two groups of PZT’s
[20,21]. The vibrator is supported by a fixture at the two nodal
points of vibration. In order to achieve resonance at the tool tip,
the tool overhang of 70 mm is adjusted, and the corresponding res-
onant frequency is 21.35 kHz. The frequency and amplitude are
controlled by an ultrasonic power supply with an maximum out-
put power of 600 W. The oscillating amplitude for the frequency
can be adjusted by changing the percentage of power from the
ultrasonic power supply. The maximum value of the oscillating
amplitude is 8 lm for the designed transducer in this experiment.
The processes of UEVR and CR are achieved by switching on and off
the ultrasonic generator, respectively.
2.2. Experimental set-up
Reaming experiments were carried out on a lathe machine
(HARDINGE HLV-V) with a UEVR system composed of an ultrasonic
transducer and a corresponding fixture, as shown in Fig. 2. The
workpiece was mounted on the lathe spindle and the UEVR system
supported by a 3D dynamometer was clamped on the machine
platform. The reaming process was achieved by feeding the reamer
to the rotating workpiece. The force measurement system con-
sisted of a 3D dynamometer, a multichannel charge amplifier, a
data acquisition system and a computer with software. The
dynamometer (Model 9272, Kistler Inc., Winterthur, Switzerland)
was used to measure the thrust force and torque. The measured
electrical signals were amplified by the multichannel charge
amplifier (Model 5070, Kistler). The amplified signals were trans-
formed by an A/D converter and then processed by the data acqui-
sition system (Model 5697, Kistler). The curves of cutting force
signal were eventually displayed on the computer with the help
of Dynoware software. The sampling rate was set to 1 kHz. The
82 D. Geng et al. / Ultrasonics 75 (2017) 80–90

Fig. 1. The schematic illustration of UEVR vibrator.

Fig. 2. Experimental set-up.

mean cutting forces were used to represent the cutting force dur- left hand flutes with a length of 36 mm and helix angle of 15° were
ing reaming. chosen as recommended by the tool supplier.
The workpieces were two layer stacks comprising of CFRP and Dry machining was employed in reaming process. Although the
Ti6Al4V plates. The CFRP laminates with a thickness of 10 mm near-dry MQL technique is assumed as a promising necessity for
were composed of 56 unidirectional plies with a thickness of hybrid stacks drilling in the future, dry reaming is still widely used
0.18 mm involving high tensile strength pre-pregs laid up accord- in aircraft assembly due to the process cost and productivity [4,10].
ing to an orientation of [45°/0°/
45°/90°/
45°/0°]S. The used CFRP The experimental conditions chosen in the experiments are based
specimen is of a fiber volume fraction of 56% giving a tensile on the industrial practice and other authors’ works [4,10], as
strength of 1200 MPa and Young’s modulus of 145 GPa. The shown in Table 1.
Ti6Al4V workpieces with a thickness of 10 mm were supplied in
the annealed condition with hardness of 41 HRC, yield strength
of 880 MPa and Young’s modulus of 114 GPa. The stacking
Table 1
sequence in the experiments was CFRP atop of the plate due to
Experimental conditions.
the typical sequence in the aircraft final assembly environment.
Thin strip workpiece samples (80  18  20 mm) were employed. Types Contents
The pilot holes were drilled by two-flute uncoated tungsten car- Drilling Drill bit Uncoated tungsten carbide drill
bide twist drills with a diameter of 7.50 ± 0.003 mm. The wet Cutting speed in CFRP/Ti 49.5/16.5
(m/min)
machining was employed in the single-shot drilling to reduce the
Feed rate (mm/min) 20
heat and mechanical damage to the pilot holes. The tolerances of Cutting condition Flood coolant
the pilot holes were evaluated after drilling and strictly con-
Reaming Tool Uncoated tungsten carbide reamer
strained between IT9 to IT10. After pilot hole drilling, the drill bit Cutting speed in CFRP/Ti 14.5/4.8
was replaced by a reamer for the following reaming operation. (m/min)
Two uncoated tungsten carbide reamers with a diameter of Feed rate (mm/min) 20
7.67 ± 0.0025 mm were employed in the experiments. In order to Cutting condition Dry cutting
Frequency, f (kHz) 21.35, 0
evacuate the chips effectively and avoid the CFRP hole erosion
Amplitude (radius) a = b = 3 lm (circle), 0
resulting from the titanium chip evacuation during reaming, six
 D. Geng et al. / Ultrasonics 75 (2017) 80–90
2.3. Experimental procedures
The hole quality parameters measured in this study include the
hole geometrical parameters (i.e. diameter, roundness and cylin-
dricity), inner hole surface defects, roughness parameters for both
CFRP and Ti holes, hole entrance and exit condition of CFRP holes.
In addition, cutting force, tool wear and chip morphology of Ti
were also analyzed, as shown in Fig. 3.
The hole diameter, roundness and cylindricity of both CFRP and
Ti holes were measured using a coordinate measuring machine
(CENTURY 977, Beijing Precision Engineering Institute, China) with
1 lm resolution. Each hole was measured at four levels, i.e. at 2, 4,
6, and 8 mm from the hole top surface. Twelve points equally dis-
tributed around the circumference were probed at each level of the
hole, giving 48 points probed on each hole. Every measurement
was repeated for three times and the final calculations of the three
geometrical parameters were averaged.
The SEM (scanning electron microscope) was used to observe
the surface defects of the inner holes. The surface roughness (Ra,
Rt) across the depth direction at hole middle position was mea-
sured by a roughness tester (TR200, Beijing Shidai Instrument
Co., China) with a traverse length of 5.6 mm and a cutoff length
of 0.8 mm. In order to evaluate the tool wear condition and CFRP
conditions at hole entry and exit, an OLYMPUS BX51 digital optical
microscope with a digital camera was used.
3. Results and discussions
3.1. Thrust force and torque
Fig. 4 shows the average thrust forces and torque of the first
three holes in both UEVR and CR, thus the influence of tool wear
on the cutting forces is negligible. Each bar represents the average
values of 3 holes with error bars which indicate ± standard devia-
tion (r) of the data set. It should be noted that all the error bars
represent the ± standard deviations of the data set in this paper. It
can be found that compared to CR, the average thrust forces in UEVR
were decreased 38% and 32% for CFRP and Ti respectively while the
average torque in UEVR were decreased 68% and 64% for CFRP and
Ti respectively. In UEVR, in order to improve the positive effect of
ultrasonic vibration on the cutting performance, the separating
cutting mode was achieved. So as to realize the separating cutting
mode, the nominal cutting velocity (m) should be controlled smaller
than the maximum vibration velocity in the cutting direction
(i.e. m < 2pfa, m = 14.5, 4.8 m/min, 2pfa = 24.1 m/min in UEVR)
[25–27].
In UEVR, the cutting edges performs the same cutting actions as
in elliptical vibration-assited cutting (EVC), thus the reduction of
Fig. 3. Schematics of the hole quality parameters.
83
thrust force in UEVR can be explained by the force reduction mech-
anism in EVC. In 1-D vibration-assisted cutting (VAC), it has been
previously asserted that the average cutting forces grows smaller
with decreasing value of duty cycle [18,28,29]. Similarly, the duty
cycle effect can also result in force reduction in EVC. Besides, each
tooltip vibration cycle in EVC contains an interval of a negative
thrust force, which can bring about a decrease of chip thickness
and a reduction of the average cutting force [22,30]. In one cycle
of the elliptical vibration cutting T, assuming the cutting begins
at tb, frictional directional direction between the rake face of the
tool and the chip is reversed at ti, and cutting ends at te. According
to [30], the average thrust cutting forces in conventional cutting
and EVC can be respectively expressed as
F mcc ¼ koc sap ð1Þ
t c1
t c2
F mev c ¼ kev c sap ð2Þ
T
where koc and kevc are the specific resistance in the thrust direction
in conventional cutting and EVC, respectively. ap is the depth of cut
and s is the feed rate. tc1 = ti
tb and tc2 = te
ti. Since kevc < koc [29],
and (tc1
tc2) < T, it indicates that smaller cutting forces can be
obtained in EVC. In UEVR, the performance of cutting edge is the
same with that in EVC, thus the reduction of thrust force in UEVR
can be explained by the force reduction mechanism in EVC. As the
torque is correlative to cutting force in reaming, the same explaina-
tion for the thrust force reduction in UEVR can be used for those of
torque [16].
The value of thrust force which is directly related to the delam-
ination defect is a key parameter in CFRP drilling. A number of
experimental, analytical, and numerical studies have been con-
ducted to propose the ‘‘critical thrust force” below which no
delamination appears [2,31–33]. For example, in [31], the delami-
nation factor increased from 1.15 to 1.55 when the thrust force
varied from 120 N to 250 N during drilling CFRP with twist drill
bit. The experimental critical thrust forces for CFRP were around
30–50 N for different drill bits including twist drill bit, slot drill
bit, brad point drill bit, core drill bit and step drill bit [32,33].
Therefore, negligible delamination at hole exit can be produced
because the values of the thrust force during reaming of CFRP were
below 25 N.
3.2. Dimensional and geometrical accuracy
Both dimension parameters (diameter), form parameters
(roundness and cylindricity) in both UEVR and CR were measured
using a CMM. Fig. 5 shows the hole diameters in both UEVR and CR
with respect to the hole number. The process reliability of the both
processes is verified by comparing the average hole diameters
84 D. Geng et al. / Ultrasonics 75 (2017) 80–90

Fig. 4. Thrust force and torque results.

Fig. 5. Hole diameter results in UEVR and CR.

during reaming. It can be observed that the hole tolerances at
around 5th, 25th and 45th holes for each process are within H8
(corresponding to a range from 0 to +22 lm, referred to the nom-
inal value of hole diameter). However, the hole diameter in UEVR
achieved a better tolerance limit (H7) after reaming 45 holes com-
pared with that in CR (H8). In addition, the smaller error bars rep-
resenting the smaller standard deviation of data set in UEVR also
indicates the cutting performance was more stable in UEVR. Dis-
crepancies in diameter within each hole over the CFRP and Ti sec-
tions are evident in both processes, which is considered due to the
different mechanical properties, such as the elastic modulus and
heat conduction of the two stack elements [3].
It can be also seen from Fig. 5 that the diameter difference of
stacked holes in UEVR is about 1.0–3.2 lm while the diameter dif-
ference in CR is about 6–12 lm. It indicates that closer tolerances
for CFRP and Ti layers were achieved in UEVR. In specific, the size
of CFRP holes in UEVR was a little larger than that in CR while the
size of Ti holes in UEVR was a little smaller than that in CR. In
reaming of CFRP/Ti stacks, the machined holes are usually over-
sized due to the tool displacement in the plane perpendicular to
the reamer axis and the specific workpiece mechanical properties.
There are two main factors responsible for the tool displacement in
UEVR, i.e. ultrasonic vibration of the tool tip in the hole radial
direction and the tool regenerative effect induced vibration in
the plane perpendicular to the reamer axis. The machined hole
would be theoretically enlarged a bit due to the ultrasonic vibra-
tion locus of cutting edges. This is the main reason for the slight
enlargement of CFRP hole size in UEVR compared to CR. In reaming,
an unbalanced directional force in the plane perpendicular to the
reamer axis could be induced by the slight eccentricity of the
reamer and pilot hole, which can cause lateral deflection of the tool
in this plane. The displacement of the tool in this plan can be cal-
 D. Geng et al. / Ultrasonics 75 (2017) 80–90
culated by the dynamic model of the machine tool [22,34]. In ultra-
sonic elliptical vibration cutting (UEVC), it is indicated that the tool
displacement almost linearly increases with the increase of unbal-
anced directional force and the smallest tool displacement was
obtained in UEVC compared to conventional cutting and 1D UVA
cutting [22,35]. From the analysis of cutting forces in UEVR and
CR in Section 3.1, it is known that smaller forces can be obtained
in UEVR. Therefore, the unbalanced directional force in UEVR on
the reamer which is the summation of all the unbalanced forces
on the cutting edges is much smaller than that in CR. As a result,
the corresponding tool displacement induced by the regenerative
vibration in UEVR is much smaller than that in CR, such that both
the dynamic stiffness of the reamer and machining accuracy of the
machined holes can be effectively improved in UEVR. It can be seen
from Fig. 4 that the cutting forces for Ti layer in reaming is much
larger than that for CFRP layer. Therefore, much larger tool dis-
placement induced by regenerative vibration was produced during
reaming Ti layer compared to reaming CFRP layer. So, the regener-
ative vibration induced displacement is the main cause for the sig-
nificant hole enlargement for Ti holes. The ultrasonic vibration in
UEVR can significantly reduce the regenerative effect induced tool
vibration and improve the hole accuracy. Thus, this is the main rea-
son for the slight reduction of Ti hole size in UEVR compared to CR.
Roundness provides the hole shape information which shows
the extent of the hole circular cross-section approximates to a
nominal circle while the cylindricity refers to the degree by which
the entire cylinder deviates. Fig. 6 shows the average roundness
and cylindricity errors of 3–5th and 43–45th holes. It is observed
that both of roundness and cylindricity decreased slightly as the
tests proceeded except for the CFRP holes in CR which increased
from 6.5 lm to 10.9 lm for roundness error and 8.4–11.4 lm for
cylindricity error. The form error of the 43–45th CFRP holes in CR
was increased sharply compared with the initial cutting stage
due to the worse surface quality (see Section 3.3). In general, the
average hole roundness errors over the entire stacked holes varied
between 5.4 and 6.7 lm in UEVR while the range was 6.8–10.9 lm
for CR. Similarly, the average hole cylindricity errors varied from
6.1 to 9.0 lm in UEVR and from 8.5 to 11.4 lm in CR. I can be found
that both of roundness and cylindricity errors in UEVR was smaller
and more stable than that in CR.
Fig. 6. Hole roundness and cylindricity results in UEVR and CR.
85
Thus, it can be concluded that the ultrasonic elliptical vibration
in UEVR allowed better machining performance in terms of dimen-
sional and geometrical accuracy than that in CR under dry reaming
condition.
3.3. Roughness and tool wear
In reaming, it is important not only to obtain a surface with low
roughness, but also a low roughness variation on different areas of
the hole. Thus, measuring profiles were recorded on the reamed
specimens at 4 different positions equally distributed around the
hole circumference during reaming. The roughness of Ra and Rt
which can provide an indication of the hole quality were measured
along the depth direction at hole middle position. Fig. 7 shows the
bar graphs of Ra and Rt. Each bar represents the average values of 4
positions in one specific hole and error bars indicates ± standard
deviation (r) of the 4-position data set.
It can be seen in Fig. 7(a) that the Ra values of CFRP section in
UEVR fluctuated down and up and were centered in 1.3 lm. In con-
trast, the Ra value of CFRP section in CR showed an increase with
the reaming progression and achieved at an average value of
1.63 lm after machining 45 holes. The Ra values of CFRP hole indi-
cate that the hoe surface is smooth, which is attributed to the
matrix smearing under dry condition (see Section 3.4). The phe-
nomenon of matrix smearing is also common in high speed drilling
CFRP even with coolant [6]. On the contrary, the Ra values of Ti sur-
face in both UEVR and CR were stable during reaming while the
values in UEVR were consistently lower than that in CR irrespective
of the number of holes (i.e. around 0.61 lm vs. 0.74 lm).
Somewhat different trends are observed when seeing the Rt val-
ues of CFRP surface in Fig. 7(b) which directly relate to the surface
defects such as fiber pullout, voids and matrix degradation. The Rt
values of CFRP surface in both UEVR and CR increased with the
increase of hole number. The value in CR eventually sharply
increased to 30.2 lm after reaming 45 holes while the holes made
by UEVR had a consistent Rt values around 15.7 lm. The different
trends of Rt values in the two processes might be related to the tool
wear condition. Usually, the large edges radius and fracture of
cutting edges can cause large deflection of fibers during reaming,
resulting in deeper subsurface damage, fiber pullout and the
86 D. Geng et al. / Ultrasonics 75 (2017) 80–90
Fig. 7. Hole Ra and Rt results.
fiber/matrix loss. On the contrary, the Rt values of Ti surfaces were
more stable in both processes although the value in CR also
showed an obvious increase after machining 45 holes. It should
be noted that almost all the error bars in both bar graphs of Ra
and Rt for UEVR, especially the CFRP section, are smaller compared
with the corresponding error bars in CR, indicating a consistent
surface roughness was obtained in UEVR.
The reduction and better stability of Ra and Rt values in UEVR is
might due to the following reasons. First of all, the extrusion of the
guide-rails on the machined surface which can produce surface
defects is significantly alleviated due to the smaller thrust force
and torque in UEVR. Besides, the stability of the tool-workpiece
system is improved in UEVR. In addition, the tool wear progression
is significantly alleviated in UEVR.
Fig. 8 shows the optical microscope view of typical worn reamer
flute after machining 45 holes both in UEVR and CR. It should be
noted that the worn flutes in Fig. 8 are the worst ones after com-
paring with the other flutes of the reamers, the rest of cutting
edges and guide-rails did not show significant changes in geome-
try. In CR, craters are noticeable on the tip of guide-rail and obvious
workpiece material adhesion are observed on the guide-rail and
rake face after machining 45 hole, as shown in Fig. 8(b). In contrast,
just slight abrasive wear at the tip of the flank can be observed in
UEVR. The craters usually located on the guide-rail tip which suf-
fered severe friction between the tool/workpiece and always
reached high temperature. The crater wear which might be origi-
nated by the diffusion and attrition is also a typical wear type in
drilling of Ti6V4V [36]. In reaming process, the craters on the
guide-rail tip can not only weaken the tool but also apparently
Fig. 8. Optical microscope view of reamer after machining 45 holes in (a) UEVR and (b) CR.
effect the finishing surface quality. No visible chip adhesion edges
in UEVR are mainly attributed to the separating forces at the inter-
face of tool tip and chip [37]. Most of the vibration energy at the
tool tip would deflect back rather than transmit to the chip due
to the sharp difference of acoustic resistance between the tool
and workpiece materials. The separating force at the tool-chip
interface resulted from the energy deflection can significantly
resist the workpiece chip adhesion.
3.4. Hole surface integrity
Fig. 9 shows the SEM images of the 5th CFRP hole surface topog-
raphy in experiments. Due to the poor heat dissipation under dry
conditions, the matrix degradation is obvious in both UEVR and
CR while the surface profile in UEVR is more uniform. When
observing the fiber fractures under high magnification, it can be
found that the fibers in UEVR was more integrated while the fibers
were significantly squeezed and covered with the smeared matrix
in CR. Fig. 10 shows the SEM images of the 45th CFRP hole surface
topography. It can be observed that both of the surface profile
quality in UEVR and CR became worse, particularly in the CR.
Besides the matrix degradation, typical defects such as serious
matrix degradation, the uncut fibers that protrude out into the
reamed surface and voids due to fiber/matrix loss between the
laminates were found in CR (see Fig. 10(b)). In contrast, surface
defects including the slighter matrix degradation and smaller voids
were found in UEVR. Fig. 11 shows the high magnification SEM
images of the Ti surface topography at the beginning of reaming
tests. It can be observed that there are clear tool feed marks and
 D. Geng et al. / Ultrasonics 75 (2017) 80–90 87

Fig. 9. SEM images of CFRP hole surface of the 5th hole in (a) UEVR and (b) CR.

Fig. 10. SEM images of CFRP hole surface of the 45th hole in (a) UEVR and (b) CR.

Fig. 11. SEM images of Ti hole surface of the 5th hole in (a) UEVR and (b) CR.

regular vibration marks on the Ti surface in UEVR while besides the
feed marks, obvious Ti chip smearing and adhesion are observed on
the Ti surface in CR. The chip adhesion and smearing is the primary
factor for higher surface roughness of Ti hole surface in CR.
The better surface topography and integrity of CFRP surface
achieved in UEVR are mainly based on the following facts. First,
the separating cutting mode was achieved in UEVR (see Sec-
tion 3.1). As a result, the separating cutting mode in RUEM could
break a feed step down to many intermittent cutting actions. So
the chipping and breakage of the carbon fiber were much more
localized which can facilitate the breakage of the carbon fibers
and eventually result in the improvement of the surface integrity.
Second, the cutting distance in each vibration cycle, D = m/
f = 11.3 lm (see Fig. 1) was the same level with the fiber diameter
which is approximate 7 lm for the used CFRP workpiece. In this
condition, the fiber and the matrix with different mechanical prop-
erties could be sheared separately. Hence, the fibers did not pre-
vent the shearing of the plastic matrix and, consequently, the
surface qualities could be improved [38]. Third, the tool wear pro-
gression was modified in UEVR (see Section 3.3).
3.5. Boundary quality of CFRP holes
Images of the reamed CFRP hole entrance and exit were carried
out by optical microscope during reaming to investigate the
mechanical damage at the CFRP hole boundaries. Clear hole bound-
aries were obtained for the first 30 holes in CR. However, after
that, damages in CR were observed including uncut fibers, fiber
pullout at the hole entrance, the overheat matrix and fuzzing as
shown in Fig. 12. On the contrary, no obvious damages were found
88 D. Geng et al. / Ultrasonics 75 (2017) 80–90
at the hole boundaries after reaming 45 holes in UEVR due to the
better cutting performance and less tool wear extent achieved in
this process. No obvious delamination was observed in both pro-
cesses due to the low values of thrust forces. This is consistent with
the analysis of the ‘‘critical thrust force” in Section 3.1.
The boundary damages at the hole entrance can easily be
removed by the countersinking operation for the bolt mounting
while the residual damages at the hole exit can have some negative
effects on the mechanical integrity of the stacks. Thus, it is more
important to control the damage at the interface between the CFRP
section and Ti section during reaming. As the reamer penetrates into
the CFRP/Ti interface, discontinuous tool-workpiece interaction
including the tool-CFRP and tool-Ti coupling takes place, as shown
in Fig. 13. The resulted bi-material interface machining makes the
region more vulnerable to severe damage formation, particularly
the CFRP section. In interface reaming period, the cutting edges
experience an dynamic contact transition and disparate thermal/
mechanical responses. As a result, the reamer will become unstable
due to the suffered intense force fluctuation and load vibration. The
instability of the tool-workpiece system is usually regarded as the
main trigger to the interface damage during stacks drilling [4].
There are usually two methods to alleviate the detrimental
effects arising from the interface reaming. First, reduce the inter-
face interaction time which is used as a direct solution in CFRP/Ti
drilling by increasing the reamer point angle (h) and the feed speed
(vf) [4]. Second, reduce the instability of the tool-workpiece system
during interface reaming by reducing the cutting force. The thrust
force and torque in either CFRP or Ti section in UEVR were signif-
icantly reduced in comparison with that in CR, as described in
Section 3.1. Thus, the tool would suffer less force fluctuation and
load vibration when entering the Ti section in interface cutting
and became more stable.
3.6. Chip analysis
Regarding the aspects related to tool performance, chips in
reaming are analyzed. Fig. 14 shows the chips obtained in both
UEVR and CR at the beginning of the tests. It can be seen that both
long and thick curled titanium swarf was produced in CR while the
spiral form chips with shorter length and slender width were gen-
erated in UEVR. In general, the volume of chips in UEVR is smaller
than that in CR because the high-frequency movement of the tool
Fig. 12. Images of 45th CFRP hole edge quality: (a) entrance in UEVR, (b) entrance in CR, (c) exit in UEVR and (d) exit in CR.
helps in breaking long chips into short length [36]. Fig. 15 shows
the SEM observations of the chip morphology. The typical serrated
chips were observed in both processes. The formation of the ser-
rated chips can be explained by the adiabatic shear theory. In pre-
vious studies, the role of cutting conditions on adiabatic shear
banding and chip serration has been widely investigated by com-
bining finite element calculations and analytical modeling
[39,40]. As for the chip surface morphology, the regular vibration
marks were found on the chip surface in UEVR while the consider-
able squeezed tracks and chip adhesion were observed on the chip
surface in CR. The chip formation in UEVR is influenced by the
ellipse geometry, the ratio of tool path vertical vibration amplitude
b and the nominal depth of cut ap, see Fig. 1. In this experiment, the
nominal depth of cut ap is much larger than the amplitude b, thus
the long continuous with regularly spaced thin segments from the
intermittent contact are visible. Slim helix chips with small volume
achieved in UEVR can result in smaller cutting force, less tool wear
and better surface integrity. According to the views proposed by
Liao [41] and Chang [42], the resistance of chip disposal in UEVR
can be reduced owing to smaller size chips in the tool flute which
in turn reduces the thrust force and alleviates tool wear. Reduced
tool wear means fewer particles are left in the cut to cause sec-
ondary damage by gouging the material surface [28]. Therefore,
better surface finish of the machined holes can be achieved in
UEVR due to the better chip evacuation.
For the purpose of CFRP/Ti assembly, the burr defect produced
in Ti phase is a key problem compared to other surface damage
since it usually leads to further disassembly, deburring and
reassembly of the stacks [4]. This means additional work which
can reach 30% of the total machining costs in precision operations
[43]. At present, there are usually four approaches to minimize the
exit burrs, i.e. optimization of process parameters and drill geom-
etry, development of finite element models and online monitoring
of burr formation [43]. In this experiment, the Ti exit burr defect
after reaming is usually inevitable under dry condition. In order
to minimize the burr height for the industrial requirement, work
on optimization of process parameters with minimal quantities
of lubricant (MQL) in UEVR is currently carried out. With the
development of intelligent spindle technology [44], an intelligent
spindle with the functions of ultrasonic elliptical vibration and
online monitoring of burr formation will be developed in the
future.
 D. Geng et al. / Ultrasonics 75 (2017) 80–90 89

Fig. 13. Scheme of the reamer involved in reaming of the interface of CFRP/Ti stacks.

Fig. 14. Chips collected in (a) UEVR and (b) CR.

Fig. 15. SEM images of chip in (a) UEVR and (b) CR.

4. Conclusion ventional dry reaming process in industrial practice to improve the

In this study, ultrasonic elliptical vibration-assisted reaming
(UEVR) is firstly introduced into finish drilling CFRP/Ti stacked bolt
holes to verify its feasibility. According to the investigation, it can
be concluded that the UEVR is an ideal finishing method for CFRP/
Ti stacks to obtain high hole quality under dry condition. Thus, the
UEVR process could be a promising alternative to the present con-
hole quality and process productivity during finish drilling large
diameter bolt holes. The following conclusions can be drawn from
the investigation:
1. Compared to CR, UEVR can result in much smaller average
thrust force and torque, which is the main cause for the
improvement of hole quality.
90 D. Geng et al. / Ultrasonics 75 (2017) 80–90
2. Tool wear condition in UEVR is significantly alleviated after
reaming 45 holes compared with that in CR and hence tool life
in UEVR can be prolonged.
3. Smaller diameter difference between CFRP and Ti holes, smaller
geometrical errors (roundness and cylindricity) of both CFRP
and Ti holes can be achieved in UEVR in comparison with that
in CR. Furthermore, there is less variation of hole size tolerance,
roundness and cylindricity in UEVR.
4. The values of average roughness (Ra and Rt) of CFRP and Ti
holes in UEVR is obviously smaller and more stable than that
in CR and the surface defects in UEVR is also significantly
reduced. These phenomena are particularly apparent at the
final stage of the experiments.
5. Spiral shaped continuous Ti swarf is prevalent in CR while both
short and slim helical Ti chips are found in UEVR, which indi-
cates the chip size is significantly reduced in UEVR.
Acknowledgements
The authors would like to thank Mrs. He in Chengdu Aircraft
Industrial (Group) Co., Ltd. for supplying the workpieces together
with technical support.
References
[1] M. Ramulu, T. Branson, A study on the drilling of composite and titanium
stacks, Compos. Struct. 54 (1) (2001) 67–77.
[2] D. Kim, A. Beal, P. Kwon, Effect of tool wear on hole quality in drilling of carbon
fiber reinforced plastic-titanium alloy stacks using tungsten carbide and
polycrystalline diamond tools, J. Manuf. Sci. Eng. 138 (3) (2016) 31006.
[3] E. Brinksmeier, R. Janssen, Drilling of multi-layer composite materials
consisting of carbon fiber reinforced plastics (CFRP), titanium and aluminum
alloys, CIRP Ann. – Manuf. Technol. 51 (1) (2002) 87–90.
[4] J. Xu, A. Mkaddem, M.E. Mansori, Recent advances in drilling hybrid FRP/Ti
composite: a state-of-the-art review, Compos. Struct. 135 (2016) 316–338.
[5] R. Zitoune, V. Krishnaraj, F. Collombet, Study of drilling of composite material
and aluminum stack, Compos. Struct. 92 (5) (2010) 1246–1255.
[6] I.S. Shyha, S.L. Soo, D.K. Aspinwall, et al., Hole quality assessment following
drilling of metallic-composite stacks, Int. J. Mach. Tools. Manuf. 51 (7) (2011)
569–578.
[7] W.L. Cong, Z.J. Pei, C. Treadwell, Preliminary study on rotary ultrasonic
machining of CFRP/Ti stacks, Ultrasonics 54 (6) (2014) 1594–1602.
[8] L.D. Chiffre, G. Tosello, M. Píška, P. Müller, Investigation on capability of the
reaming process using minimal quantity lubrication, CIRP J. Manuf. Sci.
Technol. 2 (1) (2009) 47–54.
[9] W. Belluco, L.D. Chiffre, Surface integrity and part accuracy in reaming and
tapping stainless steel with new vegetable based cutting oils, Tribol. Int. 12
(35) (2002) 865–870.
[10] J. Liu, Study on Ultrasonic Vibration Machining Processing Technology of
Aircraft Fastener Holes (in Chinese) PhD, Mechanical Engineering, Beihang
University, Beijing, 2013.
[11] A. Fink, P.P. Camanho, J.M. Andrés, et al., Hybrid CFRP/titanium bolted joints:
Performance assessment and application to a spacecraft payload adaptor,
Compos. Sci. Technol. 70 (2) (2010) 305–317.
[12] M. Saleem, R. Zitoune, E.I. Sawi, H. Bougherara, Role of the surface quality on
the mechanical behavior of CFRP bolted composite joints, Int. J. Fatigue 80
(2015) 246–256.
[13] M. Saleem, L. Toubal, R. Zitoune, H. Bougherara, Investigating the effect of
machining processes on the mechanical behavior of composite plates with
circular holes, Compos. Part A – Appl. Sci. Manuf. 55 (2013) 169–177.
[14] N.J. Churi, Z.C. Li, Z.J. Pei, C. Treadwell, Rotary ultrasonic machining of titanium
alloy: a feasibility study, in: Proceedings of the 2005 ASME International
Mechanical Engineering Congress and Exposition (IMECE 2005), Orlando, FL,
USA, November 5–11 2005, pp. 885–892.
[15] N.J. Churi, Z.J. Pei, C. Treadwell, Rotary ultrasonic machining of titanium alloy:
effects of machining variables, Mach. Sci. Technol. 10 (3) (2006) 301–321.
[16] F.D. Ning, W.L. Cong, Z.J. Pei, C. Treadwell, Rotary ultrasonic machining of
CFRP: a comparison with grinding, Ultrasonics 66 (2016) 125–132.
[17] W.L. Cong, Z.J. Pei, Q. Feng, T.W. Deines, C. Treadwell, Rotary ultrasonic
machining of carbon fiber reinforced plastic composites: using cutting fluid
versus cold air as coolant, J. Compos. Mater. 46 (14) (2012) 1747–1753.
[18] W.L. Cong, Z.J. Pei, X. Sun, C.L. Zhang, Rotary ultrasonic machining of CFRP: a
mechanistic predictive model for cutting force, Ultrasonics 54 (2) (2014) 663–
675.
[19] W.L. Cong, Z.J. Pei, T.W. Deines, D.F. Liu, C. Treadwell, Rotary ultrasonic
machining of CFRP/Ti stacks using variable federate, Compos. Part B – Eng. 52
(2013) 303–310.
[20] J. Liu, D. Zhang, L. Qin, L. Yan, Feasibility study of the rotary ultrasonic elliptical
machining of carbon fiber reinforced plastics (CFRP), Int. J. Mach. Tools
Manufact. 53 (1) (2012) 141–150.
[21] D. Geng, D. Zhang, Y. Xu, F. He, F. Liu, Comparison of drill wear mechanism
between rotary ultrasonic elliptical machining and conventional drilling of
CFRP, J. Reinf. Plast. Compos. 33 (9) (2014) 797–809.
[22] C.X. Ma, E. Shamoto, T. Moriwaki, et al., Study of machining accuracy in
ultrasonic elliptical vibration cutting, Int. J. Mach. Tool Manuf. 44 (2004)
1305–1310.
[23] X. Li, D. Zhang, Ultrasonic elliptical vibration transducer driven by single
actuator and its application in precision cutting, J. Mater. Process. Technol. 180
(1) (2006) 91–95.
[24] L. Xing, Study on Vibration Drilling and Vibration Reaming of Difficult-to-cut
Materials (in Chinese) PhD, Mechanical Engineering, Beihang University,
Beijing, 2006.
[25] W. Xu, L.C. Zhang, Y. Wu, Elliptic vibration-assisted cutting of fibre-reinforced
polymer composites: understanding the material removal mechanisms,
Compos. Sci. Technol. 92 (2014) 103–111.
[26] W. Xu, L.C. Zhang, Mechanics of fibre deformation and fracture in vibration-
assisted cutting of unidirectional fibre-reinforced polymer composites, Int. J.
Mach Tools Manufact. 103 (2016) 40–52.
[27] Ainhoa Celaya, Luis Norberto López de Lacalle, Francisco Javier Campa, Aitzol
Lamikiz, The effects of ultrasonic vibration parameters on machining
performance in turning of mild steels, Int. J. Mater. Prod. Technol. 37 (1–2)
(2010) 60–70.
[28] D.E. Brehl, T.A. Dow, Review of vibration-assisted machining, Precis. Eng. 32
(2008) 153–172.
[29] E. Shamoto, T. Moriwaki, Study on elliptical vibration cutting, CIRP Ann. 43
(1994) 35–38.
[30] C. Ma, E. Shamoto, T. Moriwaki, Study on the thrust cutting force in ultrasonic
elliptical vibration cutting, Mater. Sci. Forum 471–472 (2004) 396–400.
[31] N. Feito, J. López-Puente, C. Santiuste, M.H. Miguélez, Numerical prediction of
delamination in CFRP drilling, Compos. Struct. 108 (2014) 677–683.
[32] H. Hocheng, C.C. Tsao, Comprehensive analysis of delamination in drilling of
composite materials with various drill bits, J. Mater. Process. Technol. 140
(2003) 335–339.
[33] D.F. Liu, Y.J. Tang, W.L. Cong, A review of mechanical drilling for composite
laminates, Compos. Struct. 94 (2012) 1265–1279.
[34] J. Yang, V. Jaganathan, R. Du, A new dynamic model for drilling and reaming
processes, Int. J. Mach. Tools Manufact. 42 (2002) 299–311.
[35] Xu. Wei-Xing, Liang-Chi Zhang, Ultrasonic vibration-assisted machining:
principle, design and application, Adv. Manuf. 3 (2015) 173–192.
[36] J. Pujana, A. Rivero, A. Celaya, L.N. López de Lacalle, Analysis of ultrasonic-
assisted drilling of Ti6Al4V, Int. J. Mach. Tools Manufact. 49 (6) (2009) 500–
508.
[37] X. Li, Study on The Ultrasonic Elliptical Vibration Cutting Technology (in
Chinese) PhD, Mechanical Engineering, Beihang University, Beijing, 2007.
[38] D. Geng, D. Zhang, Y. Xu, et al., Rotary ultrasonic elliptical machining for side
milling of CFRP: tool performance and surface integrity, Ultrasonics 59 (2015)
128–137.
[39] A. Molinari, X. Soldani, M.H. Miguélez, Adiabatic shear banding and scaling
laws in chip formation with application to cutting of Ti-6Al-4V, J. Mech. Phys.
Solids 61 (2013) 2331–2359.
[40] M. Cotterell, G. Byrne, Characterisation of chip formation during orthogonal
cutting of titanium alloy Ti-6Al-4V, CIRP J. Manuf. Sci. Technol. 1 (2008) 81–85.
[41] Y.S. Liao, Y.C. Chen, H.M. Lin, Feasibility study of the ultrasonic vibration
assisted drilling of Inconel superalloy, Int. J. Mach. Tools Manufact. 47 (2007)
1988–1996.
[42] S.F. Chang, G.M. Bone, Burr size reduction in drilling by ultrasonic assistance,
Robotics Comput.-Integr. Manuf. 21 (2005) 442–450.
[43] Begoña Peña, Gorka Aramendi, Asunción Rivero, Luis N. López de Lacalle,
Monitoring of drilling for burr detection using spindle torque, Int. J. Mach.
Tools Manufact. 45 (2005) 1614–1621.
[44] Hongrui Cao, Xingwu Zhang, Xuefeng Chen, The concept and progress of
intelligent spindles: a review, Int. J. Machine Tools Manuf., http://dx.doi.org/
10.1016/j.ijmachtools.2016.10.005.