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

net/publication/338048022

Kerf Taper Defect Minimization Based on Abrasive Waterjet Machining of Low

Thickness Thermoplastic Carbon Fiber Composites C/TPU

Article  in  Materials · December 2019

DOI: 10.3390/ma12244192

CITATIONS READS

7 183

5 authors, including:

Alejandro Sambruno Fermín Bañón

Universidad de Cádiz Universidad de Cádiz
12 PUBLICATIONS   36 CITATIONS    12 PUBLICATIONS   29 CITATIONS   

SEE PROFILE SEE PROFILE

Moises Batista Ponce

Universidad de Cádiz
91 PUBLICATIONS   524 CITATIONS   

SEE PROFILE

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

Minerva View project

All content following this page was uploaded by Fermín Bañón on 23 December 2019.

The user has requested enhancement of the downloaded file.

 materials
Article
Kerf Taper Defect Minimization Based on Abrasive
Waterjet Machining of Low Thickness Thermoplastic
Carbon Fiber Composites C/TPU
Alejandro Sambruno 1, * , Fermin Bañon 1 , Jorge Salguero 1 , Bartolome Simonet 2 and
Moises Batista 1
1 Department of Mechanical Engineering and Industrial Design, Faculty of Engineering, University of Cadiz,
Av. Universidad de Cadiz 10, E-11519 Puerto Real, Cadiz, Spain; [email protected] (F.B.);
[email protected] (J.S.); [email protected] (M.B.)
2 Nanotures SL, C. Inteligencia 19, Tecnoparque Agroalimentario, E-11591 Jerez de la Frontera, Cadiz, Spain;
[email protected]
* Correspondence: [email protected]




Received: 27 October 2019; Accepted: 10 December 2019; Published: 13 December 2019


Abstract: Carbon fiber-reinforced thermoplastics (CFRTPs) are materials of great interest in industry.
Like thermosets composite materials, they have an excellent weight/mechanical properties ratio and a
high degree of automation in their manufacture and recyclability. However, these materials present
difficulties in their machining due to their nature. Their anisotropy, together with their low glass
transition temperature, can produce important defects in their machining. A process able to machine
these materials correctly by producing very small thermal defects is abrasive waterjet machining.
However, the dispersion of the waterjet produces a reduction in kinetic energy, which decreases its
cutting capacity. This results in an inherent defect called a kerf taper. Also, machining these materials
with reduced thicknesses can increase this defect due to the formation of a damage zone at the
beginning of cut due to the abrasive particles. This paper studies the influence of cutting parameters
on the kerf taper generated during waterjet machining of a thin-walled thermoplastic composite
material (carbon/polyurethane, C/TPU). This influence was studied by means of an ANOVA statistical
analysis, and a mathematical model was obtained by means of a response surface methodology
(RSM). Kerf taper defect was evaluated using a new image processing methodology, where the initial
and final damage zone was separated from the kerf taper defect. Finally, a combination of a hydraulic
pressure of 3400 bar with a feed rate of 100 mm/min and an abrasive mass flow of 170 g/min produces
the minimum kerf taper angle.

Keywords: AWJM (Abrasive waterjet machining); CFRTP (Carbon fiber-reinforced thermoplastics);

kerf taper; RSM (response surface methodology); ANOVA (Analysis of variance); C/TPU
(carbon/polyurethane)

1. Introduction
The use of composite materials in industry has generated a large number of publications and
research. Within their wide classification, carbon fiber–reinforced (CFRP) or glass fiber–reinforced
(GFRP) polymer matrix composites are the most interesting [1–3]. These materials have an excellent
weight-to-mechanical-properties ratio and have been of great importance in recent years, especially in
the aerospace and automotive sectors, although others, such as sport, wind energy, and construction,
also make use of these composites [4].
However, the type of polymeric matrix used in these materials is thermoset. This generates a
series of drawbacks in production and application of these materials, especially in terms of recyclability

Materials 2019, 12, 4192; doi:10.3390/ma12244192 www.mdpi.com/journal/materials

Materials 2019, 12, 4192 2 of 17

and processing times [5]. For this reason, there is an alternative to this kind of matrix. In recent years,
carbon fibers have been combined with a thermoplastic polymer (CFRTP) to replace thermosets [6,7].
Due to their chemical composition, these polymers have a major advantage over thermosets, as they
can be reshaped after curing. In addition, within the wide range of thermoplastic polymers, there are
high-performance polymers. These are able to reach large service temperature ranges and achieve
excellent impact resistance [8,9]. Also, compared to thermosets, CFRTPs have a high degree of
automation in their manufacturing and recyclability. This makes these materials strategic for various
industries, such as automotive, aeronautics, civil, or sports [10].
Thermoplastic polyurethane (TPU) stands out. It is an elastic elastomer that can be manufactured
by various methods and subsequently machined. According to Biron et al. [11], these polymers have
a high performance and a current consumption level. Due to this, they can provide a considerable
number of combinations of physical properties that make them extremely flexible materials and
adaptable to a multitude of uses [12].
However, due to their complex nature, they are highly difficult to machine in order to obtain their
final geometry. The main disadvantage is due to their reduced glass transition temperature. In most
conventional technologies, such as milling or drilling, the temperatures generated exceed the glass
transition temperature of CFRTP, giving rise to defectology.
Masek et al. [13] carried out a study on milling CFRTPs with different cutting geometries.
The thermoplastic matrix used was polyphenylene sulfide (PPS). When undergoing dry machining, i.e.,
in absence of cooling liquids, the temperatures reached softened the thermoplastic matrix. This boosted
the excess removal of the thermoplastic matrix, leaving the reinforcement free in the form of a burr
or fraying. This issue is also generated in the machining of composite materials with a thermoset
matrix [14]. This, together with the abrasive and adhesive wear generated on the tool, makes
conventional processes ineffective when machining CFRTPs.
However, there are non-conventional technologies that make it easier to machine materials that
are difficult to machine with conventional technology. Abrasive waterjet machining (AWJM) is now
of great interest for the machining of thermoset and thermoplastic composites due to its excellent
flexibility and high material removal performance [15,16]. Its main advantage compared to other
processes is the reduction of thermal defects during machining. In this sense, the temperatures reached
are very2019,
Materials low,12,and CFRTPs
x FOR and CFRPs can be machined without matrix removal [17–19].
PEER REVIEW 3 of 17
However, AWJM presents defectology due to the inherent nature of the process. The jet, when
hitting
composite.the material,
The matrix generates a reduction
with which in kinetic
the composite energy.
is created willThis produces
influence, to aagreater
decrease in its cutting
or lesser extent,
capacity, givingof
the generation rise to a conical
defects geometry known
after machining. Authorsas asuch
taperas[20,21].
Masek This
et al.is[15]
usually
and associated withhave
Rao et al. [14] two
geometric factors (Figure 1)—Upper width (Wt) and lower width (Wb), giving rise
similar conclusions after machining composites with different matrices. For that reason, due to the to the angles α (◦ )
and
fact that ◦
β ( ).CFRTP
The sum of these angles
composition is closegives the taper
to CFRP, angle orcan
similarities the be
kerf taper (KT,in
established Φ).their results.

Figure 1. Detail of the cut section where it is appreciated: The initial damage region (IDR), the smooth
Figure 1. Detail of the cut section where it is appreciated: the initial damage region (IDR), the smooth
cutting region (SMC), and the rough cutting region (RCR).
cutting region (SMC), and the rough cutting region (RCR).

Dhanawade et al. [27] carried out a study on abrasive waterjet cutting in a thermoset polymeric
matrix composite material. The CFRP used was 26 mm thick. In this study, a response surface
methodology was carried out in which an ANOVA statistical analysis was performed to determine
the influence of the cutting parameters. Dhanawade established that the most influential parameter
in the taper angle was hydraulic pressure. An increase in the kinetic energy of the waterjet is
produced by increasing this parameter. This increases your cutting capacity and reduces the conicity
Materials 2019, 12, 4192 3 of 17

Nevertheless, the methodology for evaluation of conicity used in most articles can generate
failures in their measurements. The formation of the zone known as the initial damage region (IDR) is
due to the erosive action of abrasive particles, which produce a combination between this defect and the
taper angle. In this way, the evaluation of upper width (Wt) can be increased due to this erosive action,
giving rise to a dispersion in the evaluated taper, as observed in results shown in references [22–24].
Other researchers [20,24–26], however, have taken into account three regions with different
defectology that may appear in the cut section (Figure 1). These are the upper region, called the initial
damage region (IDR), where a greater erosive effect is produced due to the impact of the waterjet and
abrasive particles on the surface of material; the central region, called the smooth cutting region (SCR),
which presents a more homogeneous cut due to the stabilization of the waterjet on the cutting slot;
and the lower region, called the rough cutting region (RCR), where the waterjet disperses again, losing
a great deal of kinetic energy.
The taper defect is of great importance because it occurs in any material and can result in final
geometries that do not have the required dimensions; it especially can result in reduced thicknesses.
Machining a composite material using AWJM creates a series of defects regardless of the matrix the
composite is made of. The jet is conditioned by the change of materials that compose the composite. The
matrix with which the composite is created will influence, to a greater or lesser extent, the generation of
defects after machining. Authors such as Masek et al. [15] and Rao et al. [14] have similar conclusions
after machining composites with different matrices. For that reason, due to the fact that CFRTP
composition is close to CFRP, similarities can be established in their results.
Dhanawade et al. [27] carried out a study on abrasive waterjet cutting in a thermoset polymeric
matrix composite material. The CFRP used was 26 mm thick. In this study, a response surface
methodology was carried out in which an ANOVA statistical analysis was performed to determine the
influence of the cutting parameters. Dhanawade established that the most influential parameter in
the taper angle was hydraulic pressure. An increase in the kinetic energy of the waterjet is produced
by increasing this parameter. This increases your cutting capacity and reduces the conicity of the cut.
In this way, the kerf taper is directly influenced by the amount of impact of abrasive particles and
kinetic energy given by the waterjet.
In addition, an increase in the feed rate of the machine decreases the overlap of abrasive particle
impacts, reducing its cutting capacity and thus increasing the taper angle. Kinetic energy is also
reduced with an increase in the distance between jet and the material—called the stand-off distance
(SOD)—Because it generates a greater dispersion of the jet at exit.
The abrasive mass flow also has a high influence on the conicity generated during cutting. A small
increase in this parameter decreases the conicity obtained due to the greater cutting capacity of the jet.
However, an excessive increase in the amount of abrasive particles produces a collision between them,
rounding their edges and reducing their cutting capacity, which generates a greater angle of conicity.
Similar results were obtained by El-Hofy [17]. In this study, the minimum conicity that can
be obtained was indicated by applying high hydraulic pressure combined with a small stand-off
distance. The conicity obtained is reduced by increasing the feed rate, contrary to the findings of
Dhanawade et al. [27]. This is because, at high pressures, an increase in feed rate generates a smaller
upper width (Wt), producing a more constant cut.
The distance between the focusing tube and the surface is a very important parameter for obtaining
the proper conicity. This is mainly due to a loss of kinetic energy in the form of dispersion of the
jet when leaving the focusing tube. Most studies indicate that a recommended distance is usually
2–3 mm [20,28,29].
Popan et al. [30] studied the influence of the variation of stand-off distances for a thickness of
6 mm. In this study, a reduction in this parameter of up to 0.5 mm reduced the upper cutting width
(Wt), thus decreasing the taper. In addition, a reduction in stand-off distance produces a decrease in
the radius of zone affected by erosion (IDR) due to the initial impacts of abrasive particles.
Materials 2019, 12, 4192 4 of 17

Also, the thickness of the material has a fundamental role in the conicity generated by machining.
A reduction in material thickness enhances the influence of parameters considered less decisive in
large thicknesses. Wong et al. [31] studied waterjet cutting in a thermoset composite material with a 3
mm thickness. In this study, hydraulic pressure and abrasive mass flow (AMF) take second place. In
this way, the main parameter that affects the conicity of the cut is the combination of stand-off distance
and feed rate. The combination of a high distance with a high feed produces maximum conicity.
In view of the above, more information is required on the influence of cutting parameters on
conicity and on the reduction of this defect. In addition, as there is no literature focusing on waterjet
machining with abrasive CFRTPs, it is necessary to determine the influence on matrix change [13].
Also, methodology established in most articles, gives rise to considerable errors in the assessment of
this defect, by not separating it from the area affected by abrasive particles.
For these reasons, this article proposes the evaluation of taper angle using a new methodology
based on image processing. In addition, influence of cutting parameters will be determined by means
of an ANOVA statistical analysis, in order to discuss the results obtained in machining of thermoset
composite materials.
Finally, by means of a response surface methodology (RSM), a mathematical model will be
obtained that predicts the conicity generated in abrasive waterjet cutting of low-thickness thermoplastic
matrix composite materials.

2. Materials and Methods

In this article, carbon fiber (Twill 200 g/m2 ) was used as reinforcement, and thermoplastic
resin (TPU, polyurethane) was used to manufacture the CFRTP composite. This composite was
manufactured by a thermoforming process. Table 1 shows its main characteristics, as well as the fiber
and matrix, respectively.

Table 1. Fiber, matrix, and carbon fiber–reinforced thermoplastic (CFRTP) characteristics.

Fiber type Twill 200 g/m2

Reinforcement
Fiber thickness 0.25 mm
Resin type TPU (Polyurethane)
Matrix
Melting Temperature 145 ◦ C
Ply Orientation [0◦ /90◦ ]s
CFRTP Number of plies 7
Composite thickness (t) 2.08 mm
Fiber/Matrix (%) 74.2/25.8

A three-axis water-jet machine (TCI Cutting, BP-C 3020, Valencia, Spain) was used for the
experimentation. The AWJM machine was equipped with an ultra-high capacity pump (KMT, Streamline
PRO-2 60, Bd Nauheim, Germany). The water orifice of the machine had a diameter of 0.30 mm. The
diameter and length of the focusing tube were 0.8 mm and 94.7 mm, respectively. All trials were
carried out by a 120 mesh Indian garnet abrasive material.
In order to carry out the experimental design, a response surface methodology (RSM) was set up.
This kind of methodology has already been employed by some authors in several experimental studies
of the same order [27,31]. A face-centered composite design (FCD) with a total of 20 trials (8 factorial
points–23 , 6 axial points–2 × 3, and 6 center points) was established and carried out using Minitab®
18 software (18.1, Minitab, LLC, State College, PA, USA).
A complementary experimental design was carried out. Because of this design, experimentally
obtained data can be matched with data predicted by Minitab analytical software (18.1, Minitab, LLC,
Materials 2019, 12, 4192 5 of 17

State College, PA, USA). This comparison makes it possible to obtain the error generated between these
values and confirms the accuracy of the response surface model.
Three main parameters, which include hydraulic pressure, feed rate, and abrasive mass flow, were
employed to determine their influence of the kerf taper generated. These parameters were designated
based on the limitations of the CNC machine used, as well as the levels most employed in reviewed
literature [17,22,24,31,32]. Also, they were converted into three different levels (–1, 0, 1) that represent
minimum,
Materials 2019,central,
12, x FOR and
PEERmaximum
REVIEW values, respectively (Table 2). In order to establish a stand-off
5 of 17
distance in accordance with the results of other authors [20,28,29], 2.5 mm was the stand-off distance
used in alldistance
stand-off the tests.in accordance with the results of other authors [20,28,29], 2.5 mm was the stand-
off distance used in all the tests.
Table 2. Cutting parameters set.
Table 2. Cutting parametersLevel
set.
Parameter Symbol Units Level 0 Level 1
−1
Parameter Symbol Units Level –1 Level 0 Level 1
Hydraulic
Hydraulic Pressure P P
Pressure barbar 1200 1200 2500 2500 3400 3400
Feed Rate
Feed Rate FR FR mm/min
mm/min 100 100 300 300 500 500
Abrasive Mass Flow AMF g/min 170 225 340
Abrasive Mass Flow AMF g/min 170 225 340
Stand-off
Stand-off distance SODSOD
distance mm mm Fixed at 2.5
Fixed at 2.5

Figure 2 shows the distribution of 20 test. They are machined in a 170 × 25 mm specimen with
Figure 2 shows the distribution of 20 test. They are machined in a 170 × 25 mm specimen with
an 8 mm gap to optimize material consumption. Before machining, a horizontal cut was made at
an 8 mm gap to optimize material consumption. Before machining, a horizontal cut was made at
coordinate 0.0 in order to ensure the perpendicularity of each cut with the final machined part. A
coordinate 0.0 in order to ensure the perpendicularity of each cut with the final machined part. A
cutting length of 15 mm was fixed for each trial. Furthermore, each single cut starts 10 mm before the
cutting length of 15 mm was fixed for each trial. Furthermore, each single cut starts 10 mm before the
material side to achieve a constant flow of water and abrasive. Machining was carried out on three
material side to achieve a constant flow of water and abrasive. Machining was carried out on three
specimens (KT1, KT2, KT3) of the same CFRTP in order to obtain reproducibility of the results
specimens (KT1, KT2, KT3) of the same CFRTP in order to obtain reproducibility of the results achieved.
achieved.

Figure
Figure 2. Experimental design
2. Experimental design scheme
scheme of
of CFRTP
CFRTP machining
machining carried
carried out.
out.

On the basis of this methodology, an ANOVA analysis was developed in order to obtain the
On the basis of this methodology, an ANOVA analysis was developed in order to obtain the
statistical influence of input parameters on output variables. Pressure, feed rate, and abrasive mass
statistical influence of input parameters on output variables. Pressure, feed rate, and abrasive mass
flow were changed in accordance with the fact that the experiment was conducted according to
flow were changed in accordance with the fact that the experiment was conducted according to Box-
Box-Behnken design (three-level). In addition, RSM allows us to generate different contour diagrams
Behnken design (three-level). In addition, RSM allows us to generate different contour diagrams or
or response surfaces from a second order polynomial Equation (1). There are several articles that have
response surfaces from a second order polynomial (Equation 1). There are several articles that have
implemented this type of equation in order to develop the results obtained in the experiments carried
implemented this type of equation in order to develop the results obtained in the experiments carried
out [31,33,34].
out [31,33,34]. k k k
X X X
Y = C0 + Ci xi + Cii x2i + Cij xi x j + ε (1)
𝑌 = 𝐶 +i=1 𝐶 𝑥 +i=1 𝐶 𝑥 +i< j 𝐶 𝑥 𝑥 + ɛ (1)
Y corresponds to the expected response, in this case the kerf taper generated (KT), xi are the
Y corresponds
parameters to the
used in the expected
study AMF), C0 ,inCithis
(P, FR, response, , Cii ,case
Cij are
thethe
kerf taper generated
regression 𝑥 εare
(KT),and
coefficients, is the
parameters
random used
error the study (P, FR, AMF), 𝐶 , 𝐶 , 𝐶 , 𝐶 are the regression coefficients, and ɛ is
in model.
of the
the random error of the model.
A stereoscopical microscope (Nikon, SMZ 800, Tokyo, Japan) was employed in order to obtain
macrographies of each slot (Figure 3a). Image processing software made it possible to define the
contour of the smooth cutting region (SCR) generated (Figure 3b). After machining, a first analysis
was carried out to establish the range that delimits the three separate zones in section of cut. Later,
the region obtained was split into 100 points. A trend line that adjust to those points was created. The
Materials 2019, 12, 4192 6 of 17

A stereoscopical microscope (Nikon, SMZ 800, Tokyo, Japan) was employed in order to obtain
macrographies of each slot (Figure 3a). Image processing software made it possible to define the
contour of the smooth cutting region (SCR) generated (Figure 3b). After machining, a first analysis
was carried out to establish the range that delimits the three separate zones in section of cut. Later,
the region obtained was split into 100 points. A trend line that adjust to those points was created.
The intersection of this trend line with the value 0 mm (minimum thickness) and 2.08 mm (maximum
Materials 2019,
thickness) was12,forced
x FOR PEER REVIEW
(Figure 3c). 6 of 17

Figure
Figure Macrographies
3. 3. Macrographiesobtaining
obtainingprocedure:
procedure: (a) positioningin
(a) positioning instereoscopical
stereoscopicalmicroscope;
microscope;(b)(b) image
image
of aofslot withwith
a slot the smooth cutting
the smooth regionregion
cutting (SCR) (SCR)
pointed out; (c)out;
pointed graph
(c)ofgraph
the SCR
of points
the SCRwith intersection
points with
at 0.0 mm and 2.08 mm.
intersection at 0.0 mm and 2.08 mm.

These intersection
These intersectionpoints
pointsare
areused
used to
to obtain
obtain the top kerf
the top kerf width
width(Wt,
(Wt,mm)
mm)and
andthe
thebottom
bottom kerf
kerf
width (Wb, mm). Finally, the kerf taper defect (KT) is defined as shown in Equation (2), where t ist is
width (Wb, mm). Finally, the kerf taper defect (KT) is defined as shown in Equation (2), where
thickness of of
thickness CFRTP
CFRTPininmillimeters.
millimeters.
 𝑊 t − b 𝑊
 W −W 
 2 2 

◦
KT ( ) = 2∗atan
𝐾𝑇 (°) = 2 ∗ atan t  (2) (2)
𝑡


3. Results
3. Results
The kerf taper values obtained in the three specimens (KT1, KT2, KT3), as well as the average kerf
taper andTheitskerf taper values
standard obtained
deviation (average in the three
KT), arespecimens
shown in (KT1,Table KT2,
3. KT3), as well as the average
kerf
All the results of the mean values given in Table 3 are positive. 3.
taper and its standard deviation (average KT), are shown in Table This means that, according to
Equation All(2),
thetheresults
upperof the
widthmean of values
the cutgiven in Table
is always 3 arethan
greater positive. This means
the lower width.that, according
It follows thattothe
Equation (2), the upper width of the cut is always greater than the lower width. It follows that the
geometric shape obtained in all machined slots has a “V” shape. This geometric effect is produced by
geometric shape obtained in all machined slots has a “V” shape. This geometric effect is produced by
the material thickness together with the transverse feed rate and energy amount (pressure and focusing
the material thickness together with the transverse feed rate and energy amount (pressure and
tube diameter). These variables affect the shape distortion of the slot [35]. It should be pointed out
focusing tube diameter). These variables affect the shape distortion of the slot [35]. It should be
that the kerf taper defect generated in waterjet machining is independent of the thickness of material.
pointed out that the kerf taper defect generated in waterjet machining is independent of the thickness
This defect will be greater or smaller according to this thickness, but it will always happen, even in
of material. This defect will be greater or smaller according to this thickness, but it will always
materials
happen,ofevensmall in thicknesses, such thicknesses,
materials of small as the CFRTP suchused in this
as the article.
CFRTP used in this article.
Average kerf taper values between 2.15 ◦ and 7.79◦ were obtained. It is necessary to remember
Average kerf taper values between 2.15° and 7.79° were obtained. It is necessary to remember
that there
that thereis is
currently
currentlyno noarticle
articlethat that analyzes
analyzes the the taper defect in
taper defect in abrasive
abrasivewaterjet
waterjetmachining
machining of of
thermoplastic
thermoplastic composite
compositematerials.
materials. Some Some of of reviewed literature[22,24,27]
reviewed literature [22,24,27]shows
showsvalues
values close
close to to
thethe
range
range obtained in this experiment, although some specific kerf taper values are different. It must must
obtained in this experiment, although some specific kerf taper values are different. It be
be taken
takeninto
intoaccount
account that
that thethe composites
composites usedused
in the inother
the other articles
articles are madeare of
made of a thermoset
a thermoset matrix.
matrix. Most
Most of them
of them contain
contain a higher
a higher glass
glass transition
transition temperature
temperature than
than thethe thermoplastic
thermoplastic resin
resin usedusedin in this
this
experiment.InInaddition,
experiment. addition,fixed
fixedvariables
variables such such as the focusing
focusingtubetubediameter
diameterand andabrasive
abrasive grain
grainsize
size
change
change during
during thethe mixingprocess
mixing processand andcould
couldgenerate
generate different results.
results.
OnOn thethe other
other hand,
hand, notnotallallofofthetheexisting
existingliterature
literaturetakestakesinto
intoaccount
accountthethethree
three areas
areas generated
generated in
in cutting
cutting slot—the
slot—The initial
initial damage damage region region
(IDR),(IDR), the smooth
the smooth cutting
cutting region region
(SCR), (SCR), andrough
and the the rough
cutting
cutting region (RCR). The emphasis in this article is on the independent
region (RCR). The emphasis in this article is on the independent treatment of such areas. Several treatment of such areas.
Several authors [17,22,24,27] calculate the kerf taper defect without taking this indication into
account. This could cause the kerf taper values to be altered by the rounding radius generated on the
top surface of machined material. This radius occurs when the waterjet hits the surface to be
machined. Abrasive particles, in a first instant of contact, meet a wall that they must pass through.
Not all of them are able to do it, so they disperse along the upper surface of the material, producing
Materials 2019, 12, 4192 7 of 17

Materials 2019, 12, x FOR PEER REVIEW 7 of 17

authors [17,22,24,27] calculate the kerf taper defect without taking this indication into account. This
could cause the kerf taper values to be
Table altered
3. Kerf taperby the rounding
values radius
obtained in generated on the top surface of
all tests.
machined material. This radius occurs when the waterjet hits the surface to be machined. Abrasive
particles, Pressure Feed Rate Abrasive KT 1 KT 2 KT 3 Average KT [°
Test in a first instant of contact, meet a wall that they must pass through. Not all of them are able
to do it, so they disperse along the upper surface of the material, producing a rounding at± the
[bar] [mm/min] [g/min] [°] [°] [°] (°)]
input
1
of cut. 1200 100 170 4.71 3.23 3.92 3.95 ± 0.60
2 1200 100 340 4.09 5.25 4.67 4.67 ± 0.47
Table 3. Kerf taper values obtained in all tests.
3 1200 300 225 6.14 4.92 6.25 5.77 ± 0.60
4 Pressure
1200 Feed
500Rate Abrasive
170 6.511 [◦ ] 7.182 [◦ ] Average KT
Test KT KT KT9.58
3 [◦ ] 7.76 ± 1.32
[bar] [mm/min] [g/min] [◦ ± (◦ )]
5 1200 500 340 7.53 7.80 7.73 7.69 ± 0.11
1 1200 100 170 4.71 3.23 3.92 3.95 ± 0.60
6 2500 100 225 1.39 3.34 2.91 2.55 ± 0.84
2 1200 100 340 4.09 5.25 4.67 4.67 ± 0.47
37 2500
1200 300
300 170
225 1.96
6.14 3.62
4.92 3.41
6.25 3.00 ±± 0.60
5.77 0.74
48 1200
2500 500
300 170
225 6.51
2.43 7.18
3.43 9.58
3.13 7.76 ± 1.32
3.00 ± 0.42
5 1200 500 340 7.53 7.80 7.73 7.69 ± 0.11
69 2500
2500 300
100 225
225 5.12
1.39 2.16
3.34 6.39
2.91 4.55 ±± 0.84
2.55 1.77
710 2500
2500 300
300 225
170 4.22
1.96 2.03
3.62 5.98
3.41 4.07 ±± 0.74
3.00 1.62
811 2500
2500 300
300 225
225 2.43
4.77 3.43
2.26 3.13
3.25 3.00 ± 0.42
3.43 ± 1.03
9 2500 300 225 5.12 2.16 6.39 4.55 ± 1.77
1012 2500
2500 300
300 225
225 5.12
4.22 2.85
2.03 3.16
5.98 3.71 ±± 1.62
4.07 1.00
1113 2500
2500 300
300 225
225 4.77
4.54 2.26
4.40 3.25
3.26 4.07 ±± 1.03
3.43 0.57
12 2500 300 225 5.12 2.85 3.16 3.71 ± 1.00
14 2500 300 340 3.39 5.59 4.70 4.56 ± 0.90
13 2500 300 225 4.54 4.40 3.26 4.07 ± 0.57
1415 2500
2500 500
300 225
340 5.88
3.39 6.10
5.59 4.94
4.70 5.64 ±± 0.90
4.56 0.50
1516 2500
3400 500
100 225
170 5.88
1.61 6.10
2.11 4.94
3.08 5.64 ± 0.50
2.27 ± 0.61
16 3400 100 170 1.61 2.11 3.08 2.27 ± 0.61
1717 3400
3400 100
100 340
340 0.75
0.75 2.91
2.91 4.84
4.84 2.83 ±± 1.67
2.83 1.67
1818 3400
3400 300
300 225
225 2.16
2.16 1.48
1.48 2.81 2.15 ±± 0.54
2.15
19 3400 500 170 1.52 2.69 3.31 2.51 ± 0.74
19 3400 500 170 1.52 2.69 3.31 2.51 ± 0.74
20 3400 500 340 1.68 2.16 3.68 2.51 ± 0.85
20 3400 500 340 1.68 2.16 3.68 2.51 ± 0.85

Furthermore, taking into account the RCR zone means that there may be an error when obtaining
Furthermore, taking into account the RCR zone means that there may be an error when obtaining
the values. In this area, the waterjet comes out of machined material, resulting in a new opening.
the values. In this area, the waterjet comes out of machined material, resulting in a new opening. In
In composite materials, carbon fibers and the matrix can become detached by creating loose yarns or
composite materials, carbon fibers and the matrix can become detached by creating loose yarns or
cavities. Therefore, 10% of the cutting slot as was chosen as the RCR zone.
cavities. Therefore, 10% of the cutting slot as was chosen as the RCR zone.
Two macrographies of the cutting section of two machined slots with different parameter
Two macrographies of the cutting section of two machined slots with different parameter
combinations are shown in Figure 4. A rounded radius produced in the upper face of the cutting slot
combinations are shown in Figure 4. A rounded radius produced in the upper face of the cutting slot
is shown in both figures. This defect is usually caused by the collision of the waterjet and abrasive
is shown in both figures. This defect is usually caused by the collision of the waterjet and abrasive
particles with the top face of material. The radius obtained for Figure 4a is 0.063 mm, while the radius
particles with the top face of material. The radius obtained for Figure 4a is 0.063 mm, while the radius
generated in Figure 4b is 0.037 mm.
generated in Figure 4b is 0.037 mm.

a b

Radius Radius
0.063 (mm) 0.037 (mm)

P = 1200 (bar) P = 2500 (bar) IDR

IDR SCR
FR = 100 (mm/min) FR = 300 (mm/min) 87.25 % of
SCR
AMF = 340 (g/min) 80.50 % of AMF = 340 (g/min) total thickness
total thickness

Figure4.4.Cutting
Figure Cuttingpath
pathmacrograph
macrographforformachining
machiningunder
underconditions:
conditions: (a).
(a). P 1200 bar,
bar, FR
FR 100
100 mm/min,
mm/min,
AMF 340 g/min; (b). P 2500 bar, FR 300 mm/min, AMF 340 g/min.
AMF 340 g/min; (b). P 2500 bar, FR 300 mm/min, AMF 340 g/min.

This confirms the randomness that can be obtained in two slots machined on the same material.
In addition, the point that separates IDR zone from SCR is found at different heights. Figure 4a shows
Materials 2019, 12, 4192 8 of 17

This confirms the randomness that can be obtained in two slots machined on the same material.
Materials 2019, 12, x FOR PEER REVIEW 8 of 17
In addition, the point that separates IDR zone from SCR is found at different heights. Figure 4a shows
the point at a height equivalent to 80.50% of total thickness (2.08 mm), and Figure 4b shows it at 87.25%
the point at a height equivalent to 80.50% of total thickness (2.08 mm), and Figure 4b shows it at
of that thickness.
87.25% of that thickness.
To sum up, taking into account the thickness of material used in this article, cutting section was
To sum up, taking into account the thickness of material used in this article, cutting section was
split according to 0–10% of the thickness for the rough cutting region, 10–75% for the smooth cutting
split according to 0–10% of the thickness for the rough cutting region, 10–75% for the smooth cutting
region and, 75–100% for the initial damage region. This article considers the kerf taper defect in the
region and, 75–100% for the initial damage region. This article considers the kerf taper defect in the
SCR (Figure 5).
SCR (Figure 5).

1 mm

Figure
Figure 5. Detail5.of
Detail
a realofcut
a real cut section
section where itwhere
can beitseen
can be
theseen theRough
areas: areas: cutting
rough cutting
region region (RCR) 0–10%,
(RCR) 0–10%,
smoothsmooth
cutting cutting region10–75%,
region (SCR) (SCR) 10–75%, initial damage
initial damage region75–100%.
region (IDR) (IDR) 75–100%.

3.1. Statistical Analysis

3.1. Statistical Analysis
An ANOVAAn ANOVAanalysis of the obtained
analysis kerf taper
of the obtained kerfvalues is shown
taper values in the Table
is shown in the4.Table
As discussed in
4. As discussed in
the methodology section,section,
the methodology a second order quadratic
a second model model
order quadratic was employed.
was employed.This model allowsallows
This model us to us to
relate the input
relate theparameters used inused
input parameters experiment (pressure,
in experiment feed, rate
(pressure, and
feed, abrasive
rate mass flow)
and abrasive masswith
flow)the
with the
results results
obtainedobtained after machining. The ratio between both variables makes it possible to greater
after machining. The ratio between both variables makes it possible to create create greater
or lesser
oraccuracy between
lesser accuracy them. The
between p-value
them. of this of
The p-value model is less than
this model is less0.05,
thanwhich indicates
0.05, which that it is
indicates that it is
statistically significant. In addition, a high F-value will make variable more relevant in the analysis.
statistically significant. In addition, a high F-value will make variable more relevant in the analysis.
As can be seen in Table 4, pressure and feed rate are the most significant parameters in taper
Table 4. ANOVA analysis of the kerf taper values obtained.
defect. This is deduced because p-value is 0, i.e., the significance is maximum. It is observed that the
AMF parameterSourcehas a p-value DF of 0.15,
Adjwhich
SS implies
Adj MSthat F-Value
this variablep-Value
has less influence on the
generation ofModel
the taper angle. In9
addition,
48.2278
it can be seen
5.3586
that the F-value
16.95
for pressure variable triples
0.00
to that obtained
Pressurefor feed rate. This
(bar) 1 allows
30.6617us to deduce
30.6617that, although
97.00 both0.00
variables are significant,
hydraulic
Feedpressure has greater1weight9.8658
Rate (mm/min) in the generation
9.8658 of the 31.21
kerf taper. 0.00
Abrasive
These (g/min)
statistical 1 consistent
results are 0.7668with the 0.7668 2.43 to the topic
articles related 0.15 studied. Therefore, it
Error 10 3.1610 0.3161
can be concluded that, for generation of a taper defect, the order of parameters according to their
Lack-of-Fit 5 1.6498 0.3300 1.09 0.46
influence isPure
hydraulic
Error
pressure,5
feed rate,
1.5112
and abrasive
0.3022
mass flow.
Total 19 51.3888
Table 4. ANOVA analysis of the kerf taper values obtained.

As can be seen in TableSource

4, pressure andDF feedAdj rateSSare Adj MS significant
the most F-Value parameters
P-Valuein taper
defect. This is deduced because
Model p-value is 0,
9 i.e., the
48.2278significance
5.3586 is maximum.
16.95 It is observed
0.00 that
the AMF parameter has a p-value of 0.15, which implies that this variable has less influence on the
Pressure (bar) 1 30.6617 30.6617 97.00 0.00
generation of the taper angle. In addition, it can be seen that the F-value for pressure variable triples
Feed
to that obtained for feed RateThis
rate. (mm/min)
allows us 1to deduce9.8658that,9.8658
although both31.21 0.00
variables are significant,
hydraulic pressure has greater
Abrasiveweight
(g/min)in the generation
1 0.7668of the kerf taper. 2.43
0.7668 0.15
These statistical results Error
are consistent with10 the articles0.3161
3.1610 related to the topic studied. Therefore,
it can be concluded that, for generation of a taper defect, the order of parameters according to their
Lack-of-Fit 5 1.6498 0.3300 1.09 0.46
influence is hydraulic pressure, feed rate, and abrasive mass flow.
Pure Error 5 1.5112 0.3022
Total 19 51.3888

Equation (3) shows mathematical model obtained for analyzed response surface methodology.
This equation presents an R2 of 93.85%, which implies a value very close to 100%. In later sections, a
verification of Equation (3) will be performed from input variables shown in Table 3.
 Materials 2019, 12, x FOR PEER REVIEW 9 of 17

Materials 2019, 12,𝐾𝑇 =

4192 1.00 + (0.7𝑃 + 124.8𝐹𝑅 + 153𝐴𝑀𝐹 + 0.08𝐹𝑅 − 0.17𝐴𝑀𝐹 − 0.04𝑃 ∗ 𝐹𝑅 9 of 17
(3)
−0.13𝐹𝑅 ∗ 𝐴𝑀𝐹) ∗ 10
Equation (3) shows mathematical model obtained for analyzed response surface methodology.
This equation
3.2. Effect of Hydraulic an R2 of
presentsPressure on93.85%, which implies a value very close to 100%. In later sections,
Kerf Taper
a verification of Equation (3) will be performed from input variables shown in Table 3.
The effect of hydraulic pressure on variation of the kerf taper is shown in Figure 6. To discuss
these
KT =data,
1.00 +feed rate
(0.7P +and abrasive
124.8FR mass flow
+ 153AMF variables
+ 0.08FR 2 have been
− 0.17AMF 2 kept fixed and at their intermediate
− 0.04P ∗ FR − 0.13FR ∗ AMF) ∗ 10−4 (3)
level (FR = 300 mm/min; AMF = 225 g/min). Three machined specimens have been classified with
different
3.2. colors.
Effect of Thus,Pressure
Hydraulic red color
on corresponds
Kerf Taper to test piece 1, yellow to test piece 2. and green to test
piece 3. In addition, each figure shows the mean value of taper defect obtained for each test. This
The effect of hydraulic pressure on variation of the kerf taper is shown in Figure 6. To discuss
value was represented with a discontinuous line followed by mean value and resulting deviation.
these data, feed rate and abrasive mass flow variables have been kept fixed and at their intermediate
Pressure is one of the most influential variables in machining, therefore Figure 6 contains images that
level (FR = 300 mm/min; AMF = 225 g/min). Three machined specimens have been classified with
facilitate visual compression.
different colors. Thus, red color corresponds to test piece 1, yellow to test piece 2. and green to test
At first sight, the taper decreases as pressure increases. By taking a look at the mean values
piece 3. In addition, each figure shows the mean value of taper defect obtained for each test. This value
obtained and pressures involved, it can be seen how the resulting data adjust to a large extent to a
was represented with a discontinuous line followed by mean value and resulting deviation. Pressure is
linear regression. Higher pressure means an increase in kinetic energy generated by the waterjet. This
one of theinmost
increase influential
kinetic variables
energy leads in machining,
to a higher materialtherefore
removal. Figure 6 contains
Consequently, theimages
walls ofthat
the facilitate
slot are
visual compression.
subjected to a greater force in removing the material that causes a more vertical final state.

KT 1 KT 2 KT 3 Average KT
7.00

6.00 5.77 ± 0.60

5.00

4.00 3.80 ± 1.07

KT (°)

3.00
2.15 ± 0.54
2.00

1.00

0.00
1200 2500 3400
Pressure (bar)

0.5 mm 0.5 mm 0.5 mm

P - Wb

Figure6.6. Variation
Figure Variation of
of the
the kerf
kerf taper
taper in
in three
three specimens as aa function
specimens as functionof
ofpressure (FR==300
pressure(FR 300mm/min;
mm/min;
AMF = 225 g/min).
AMF = 225 g/min).

At first sight, the taper decreases as pressure increases. By taking a look at the mean values
By looking at the graph with values of each specimen independently, it can be seen how
obtained
specimenand 2 inpressures involved,
all cases has a taperitvalue
can be seen how
smaller thanthe resulting
1 and 3. Thisdata
mayadjust to a large
have been affectedextent to a
by the
linear regression. Higher pressure means an increase in kinetic energy generated
nature of the composite. The material is composed by long carbon fibers and thermoplastic matrix by the waterjet. This
increase
sheets. Thein kinetic energyof
arrangement leads
thesetoelements
a higheralong
material
the removal.
compositeConsequently, the out
is crucial to carry wallstheofmachining.
the slot are
subjected to a greater force in removing the material that causes a more vertical final
In this way, a fiber yarn displaced at the time of manufacturing or an irregular consolidation of matrix state.
alongBythelooking at could
surface the graph
alterwith values of each
its homogeneity specimen
(Figure independently,
7). Nevertheless, the itdecreasing
can be seen how specimen
tendency of the
2taper
in alldefect
cases as
has
hydraulic pressure increases is reflected in three specimens studied. Inbythis
a taper value smaller than 1 and 3. This may have been affected thecase,
naturea
of the composite. The material is composed by long carbon fibers and thermoplastic
value of 3400 bar generates the lowest taper, being 2.15°. It should be noted that the thickness matrix sheets.
of
The arrangement
composite of these
employed elements along
is considered the composite
thin, which could makeis crucial to carry
it easier out theenergy
for kinetic machining. In this
created by
way, a fiber yarn displaced at the time of
pressure to generate a homogeneous wall in slot. manufacturing or an irregular consolidation of matrix along
the surface could alter its homogeneity (Figure 7). Nevertheless, the decreasing tendency of the taper
defect as hydraulic pressure increases is reflected in three specimens studied. In this case, a value of
 Materials 2019, 12, x FOR PEER REVIEW 10 of 17

A study carried out by Dhanawade et al. [27] shows a similar trend to that generated in this
research.
Materials 2019, 12,The
4192material
used in its study is a composite formed by carbon fibers and thermostable
10 of 17
epoxy resin. They agree that an increase in pressure generates a decrease in taper defect. In addition,
Dhanawade et al. use a hydraulic pressure that reaches 4000 bar, obtaining a taper angle that oscillates
3400 bar generates the lowest taper, being 2.15◦ . It should be noted that the thickness of composite
from 2.2° to 3.8°. It can be deduced that these results agree with those obtained in this article.
employed is considered thin, which could make it easier for kinetic energy created by pressure to
Therefore, if the pressure parameter is analyzed in isolation, it seems that the kind of resin used in
generate a homogeneous wall in slot.
composite material does not greatly influence the kerf taper generated.

CFRTP CFRP
(Polyurethane) (Epoxy)
x 100 x 100

a. b.

Figure
Figure 7. Macrographs
7. Macrographs of a composite
of a composite cross-section
cross-section with
with (a). (a). thermoplastic
thermoplastic matrix;matrix; (b). thermoset
(b). thermoset matrix.
matrix.
A study carried out by Dhanawade et al. [27] shows a similar trend to that generated in this
research.Ruiz-Garcia
The material used
et al. [20]incarried
its study outisa astudy
composite
on the formed
analysisby of carbon
the kerffibers
taper anddefect thermostable
in the abrasive
epoxy resin. machining
waterjet They agreeof that an increase
CFRP/UNS in pressure
A97075 stack.generates a decrease
In that paper, the kerfin taper
taperdefect.
results In addition,
obtained were
Dhanawade
found ineta al. rangeuse of
a hydraulic
1°. Ruiz-Garciapressure et that reaches
al. used lower 4000 bar,rates
feed obtaining a taperused
than those angle inthat
thisoscillates
article. This
2.2◦ to 3.8
fromresulted in ◦a. greater
It can behomogeneity
deduced thatof these
taper results
defectsagree withas
as well those obtained
a smaller in this article.
variation of them.Therefore,
In addition,
if the pressure parameter is analyzed in isolation, it seems that
the matrix used by Ruiz-Garcia et al. was epoxy resin, which has a higher vitreous transitionthe kind of resin used in composite
material does notthan
temperature greatly
the influence
thermoplastic the kerfresin taper
usedgenerated.
in this experiment (TPU—polyurethane, 145 °C). The
Ruiz-Garcia
use of a thermosetet al. resin
[20] carried out a the
could allow study on the analysis
composite to achieve of the kerf mechanical
greater taper defectproperties
in the abrasive
than the
waterjet machining of CFRP/UNS
CFRTP employed in this experiment. A97075 stack. In that paper, the kerf taper results obtained were
found in a range of 1◦ . Ruiz-Garcia et al. used lower feed rates than those used in this article.
This3.3. Effect ofinFeed
resulted Rate onhomogeneity
a greater Kerf Taper of taper defects as well as a smaller variation of them. In
addition, the matrix used by Ruiz-Garcia et al. was epoxy resin, which has a higher vitreous transition
Figure 8 shows the effect of feed rate on the kerf taper defect. The distribution of elements in
temperature than the thermoplastic resin used in this experiment (TPU—polyurethane, 145 ◦ C). The
graph is similar to the one shown in Figure 6. As for previous section, this figure contains 3 images
use of a thermoset resin could allow the composite to achieve greater mechanical properties than the
that make it easier to understand. It can be seen how, unlike for pressure, average values of kerf taper
CFRTP employed in this experiment.
rise as feed rate increases.
3.3. EffectAofslower
Feed Rate cutonis Kerf
made when a slot is made with a minimum speed of 100 mm/min, causing a
Taper
more homogeneous fracture along the surface of material. In this case, waterjet is able to pass through
Figure 8and
material shows the effect
remove each of feed
layer ofrate
carbonon thefiberkerf
and taper defect.
matrix in anThe distribution
orderly way. This of elements in
effect, together
graphwithis similar
abrasiveto particles,
the one shown reduces in Figure 6. As for
the influence ofprevious
cohesionsection,
nature thisof afigure contains
composite 3 images
material on the
thatgeneration
make it easier to understand. It can be seen how, unlike for pressure, average
of kerf taper. Therefore, it follows that applying a low feed rate, kerf taper achieved will values of kerf taper
rise be
as smaller,
feed rateor increases.
in other words, upper and lower width of slot are values closer to each other.
A slower cut
On the other is made
hand,when whena slot is made
a high feedwith
ratea is minimum
applied,speed a lossofof100 mm/min,
kinetic energy causing a more in
is generated
homogeneous fracture along the surface of material. In this case,
cutting channel. These could produce a decrease in material removal and a decrease in abrasive waterjet is able to pass through
material and affecting
particles remove each the layer
surface of carbon fiber and
of material. On matrix in an orderly
this occasion, way. This
the waterjet effect,susceptible
is more together with to the
abrasive particles, reduces the influence of cohesion nature of a composite
nature of the material. Different layers of carbon fiber and matrix take advantage of the reduced material on the generation
of kerf taper.
energy of Therefore,
the jet to makeit follows that applying
it difficult a low feed
to pass through therate, kerf taper
composite. Thisachieved
can leadwill be smaller,
to the slot wallsor not
in other
beingwords, upper andresulting
perpendicular, lower width in theofupperslot are values
width closer
of slot to each
being other.
greater than the lower.
On the other hand, when a high feed rate is applied, a
Thus, the lowest taper value will be produced for a feed rate of 100 mm/min,loss of kinetic energy is generated in cutting
being 2.55°, and
channel. These value
the highest could will
produce a decreasewith
be produced in material
a feed rate removal and a decrease
500 mm/min, resultingin abrasive
in 5.64°.particles
The lower
affecting the surface
thickness of material.
of the material On thiscombined
employed, occasion, the withwaterjet
a low feedis more
rate,susceptible
are parameters to thethatnature
helpoftothe
create
material. Different layers
a homogeneous cut by AWJM. of carbon fiber and matrix take advantage of the reduced energy of the jet to
make it difficult to pass through the composite. This can lead to the slot walls not being perpendicular,
resulting in the upper width of slot being greater than the lower.
Thus, the lowest taper value will be produced for a feed rate of 100 mm/min, being 2.55◦ , and
the highest value will be produced with a feed rate 500 mm/min, resulting in 5.64◦ . The lower
thickness of the material employed, combined with a low feed rate, are parameters that help to create a
homogeneous cut by AWJM.
Materials2019,
Materials 12,x4192
2019,12, FOR PEER REVIEW 11of
11 of17
17

KT 1 KT 2 KT 3 Average KT
7.00

6.00 5.64 ± 0.50

5.00

4.00 3.80 ± 1.07

KT (°)

3.00
2.55 ± 0.84

2.00

1.00

0.00
100 300 500
Feed Rate (mm/min)

0.5 mm 0.5 mm 0.5 mm

FR - Wb

Variationofofkerf
Figure8.8.Variation
Figure kerftaper
taperininthree
threespecimens
specimens
asas a function
a function of of feed
feed raterate = 2500
(P =(P2500 bar;bar;
AMF = 225=
AMF
225 g/min).
g/min).

Everything discussed above is consistent with Wong et al. [31]. In this paper, the authors
Everything discussed above is consistent with Wong et al. [31]. In this paper, the authors analyze
analyze the taper defects in a carbon fiber thermostable matrix (epoxy). The FR used oscillated in
the taper defects in a carbon fiber thermostable matrix (epoxy). The FR used oscillated in 1000–2500
1000–2500 mm/min, which is a little bit higher than that carried out in this article. However, the authors
mm/min, which is a little bit higher than that carried out in this article. However, the authors conclude
conclude that a greater feed rate implies a lower amount of abrasive particles affecting the material.
that a greater feed rate implies a lower amount of abrasive particles affecting the material. The
The decrease in the amount of these particles caused a dirtier and more random cut.
decrease in the amount of these particles caused a dirtier and more random cut.
As the cutting speed increases, the upper and lower widths of the slot decrease. However, the
As the cutting speed increases, the upper and lower widths of the slot decrease. However, the
width of the bottom surface has a greater decreasing tendency than the top surface. This is consistent
width of the bottom surface has a greater decreasing tendency than the top surface. This is consistent
with articles that study the taper defect in other kinds of materials [36,37].
with articles that study the taper defect in other kinds of materials [36,37].
3.4. Effect of Abrasive Flow Rate on Kerf Taper
3.4. Effect of Abrasive Flow Rate on Kerf Taper
The influence of the abrasive mass flow (AMF) on generation of taper defect is shown in Figure 9.
The influence of the abrasive mass flow (AMF) on generation of taper defect is shown in Figure
In Section 3.1, it was concluded that this parameter is the least influential of the three used in this
9. In Section 3.1, it was concluded that this parameter is the least influential of the three used in this
article. However, this fact does not imply that the amount of abrasive employed does not affect the
article. However, this fact does not imply that the amount of abrasive employed does not affect the
results. In fact, by looking at the trend of the mean values plotted in Figure 9, it can be seen how it
results. In fact, by looking at the trend of the mean values plotted in Figure 9, it can be seen how it
ascends as the amount of abrasive increases. The upward trend caused by AMF is less than that caused
ascends as the amount of abrasive increases. The upward trend caused by AMF is less than that
by pressure and feed rate. Therefore, this parameter is the least influential of those used.
caused by pressure and feed rate. Therefore, this parameter is the least influential of those used.
It should be noted that an excessive increase in AMF can lead to a loss of kinetic energy. In this
It should be noted that an excessive increase in AMF can lead to a loss of kinetic energy. In this
case, abrasive particles are more likely to collide with each other, causing a more disturbed cut, which
case, abrasive particles are more likely to collide with each other, causing a more disturbed cut, which
translates into greater erosion at the input of slot, a greater difference between the upper and lower
translates into greater erosion at the input of slot, a greater difference between the upper and lower
widths of the slot, and greater fraying at the outlet surface of the material [38].
widths of the slot, and greater fraying at the outlet surface of the material [38].
In this case, a smallest kerf taper will be produced with a small amount of abrasive, 170 g/min,
In this case, a smallest kerf taper will be produced with a small amount of abrasive, 170 g/min,
resulting in an angle of 3.00◦ . For the highest amount of abrasive, an angle of 4.56◦ is obtained.
resulting in an angle of 3.00°. For the highest amount of abrasive, an angle of 4.56° is obtained.
Ruiz-Garcia et al. [20] achieved an upward trend similar to that obtained in this experiment.
Ruiz-Garcia et al. [20] achieved an upward trend similar to that obtained in this experiment. The
The use of a higher abrasive flow means a greater difference between the upper width of the slot and
use of a higher abrasive flow means a greater difference between the upper width of the slot and the
the lower width. Due to that, it can be noted that the abrasive flow rate used in AWJM is not linked to
lower width. Due to that, it can be noted that the abrasive flow rate used in AWJM is not linked to
the kind of resin employed in the manufacture of composite.
the kind of resin employed in the manufacture of composite.
Materials 2019, 12, x FOR PEER REVIEW 12 of 17

KT 1 KT 2 KT 3 Average KT
7.00

6.00
Materials 2019, 12, 4192 12 of 17
Materials 2019, 12, x FOR PEER REVIEW 12 of 17

5.00
4.56 ± 0.90
KT 1 KT 2 KT 3 Average KT
4.00 7.00 3.80 ± 1.07
KT (°)

3.00 ± 0.74
3.00 6.00

2.00 5.00
4.56 ± 0.90

1.00 4.00 3.80 ± 1.07

KT (°)

0.00 3.00 3.00 ± 0.74

170 225 340
Abrasive Mass Flow (g/min)
2.00

Figure 9. Variation of kerf taper in three specimens as a function of abrasive mass flow rate (P = 2500
1.00
bar; FR = 300 mm/min).
0.00
170 225 340
3.5. Response Surface
Abrasive Mass Flow (g/min)
A response surface allows two input parameters to interact with an output variable, keeping all
Figure 9.
Figure Variation of
9. Variation kerf
kerftaper
taperininthree specimens asas
a function of abrasive mass flowflow (P =(P
raterate 2500 bar;
other parameters constant.ofThe ANOVA three specimens
analysis carried aout
function of abrasive
in Section mass
3.1 gave hydraulic = 2500
pressure
FR =FR300 mm/min).
and feedbar; = 300
rate the mm/min).
greatest significance on the machining process. Therefore, FR and P have been
3.5. Response
represented Surface
together with the kerf taper in Figure 10, keeping AMF = 225 g/min fixed.
3.5. Response Surface
As can be seen in this figure, both parameters seem very significant. A value close to 3400 bar
A response surface allows two input parameters to interact with an output variable, keeping all
combined A response surface
with various feedallows two input parameters
rate combinations offer the to interact
lowest with
kerf an values
taper output (Figure
variable,11).
keeping
As a all
other parameters constant. The ANOVA analysis carried out in Section 3.1 gave hydraulic pressure
other of
function parameters constant.
P-value, both P andThe ANOVA
FR were analysiswhile
significant, carried out in Section
according 3.1 gave
to F-value, hydraulic
pressure tripledpressure
the
and feed rate the greatest significance on the machining process. Therefore, FR and P have been
and feed
influence rate rate.
of feed the greatest
This shows significance on the
that a single machining
pressure resultsprocess.
in a rangeTherefore, FR and
of kerf taper P have
values been
(< 3°)
represented together with the kerf taper in Figure 10, keeping AMF = 225 g/min fixed.
represented
for the together
whole range with
of feed theused
rates kerf taper
in thisinexperiment.
Figure 10, keeping AMF = 225 g/min fixed.
As can be seen in 500 this figure, both parameters seem very significant. A value close to 3400 bar
combined with various feed rate combinations offer the lowest kerf Average
taper KT (°) (Figure 11). As a
values
0 < 3
function of P-value, both P and FR were significant, while according to F-value, pressure tripled the
3–4
influence of feed rate. This400
shows that a single pressure results in a range 4–5
of kerf taper values (< 3°)
for the whole range of feed rates used in this experiment. 5–6
FR (mm/min)

6–7
500
0>7
Average KT (°)
300
0<3
Hold Value
3–4
400 AMF 225 (g/min)
4–5
200 5–6
FR (mm/min)

6–7
0>7
300
100 Hold Value
1500 2000 2500 3000
AMF 225 (g/min)
P (bar)
200
Figure 10. Contour plot of kerf taper defect as a function of most influential variables: Pressure and
Figure 10. Contour plot of kerf taper defect as a function of most influential variables: pressure and
feed rate.
feed rate.
100
As can be seen in this figure, both parameters
1500 2000 seem
2500 very significant. A value close to 3400 bar
3000
On the other hand, a combination of 1200 bar of pressure and 500 mm/min of feed rate seems to
P (bar)
combined with various feed rate combinations offer the lowest kerf taper values (Figure 11). As a
be the most unfavorable in the study of kerf taper, and it seems to be the element that greatly increases
function of P-value, both P and FR were significant, while according to F-value, pressure tripled the
the upward trend.
Figure As speed increases, cutting capacity decreases due to the loss of kinetic energy in
influence of10. Contour
feed plot
rate. This of kerf taper
shows that adefect
singleaspressure
a function of most
results in influential
a range ofvariables:
kerf taperpressure
values and
(<3◦ ) for
the waterjet. To
feed rate. this effect, it would be necessary to add the reduction of the number of abrasive
the whole range of feed rates used in this experiment.
particles that affect the material.
On the other hand, a combination of 1200 bar of pressure and 500 mm/min of feed rate seems to
be the most unfavorable in the study of kerf taper, and it seems to be the element that greatly increases
the upward trend. As speed increases, cutting capacity decreases due to the loss of kinetic energy in
the waterjet. To this effect, it would be necessary to add the reduction of the number of abrasive
particles that affect the material.
Abrasive waterjet cutting is a technology that generates a lower temperature than other conventional
machining technologies, such as milling or drilling [14]. This process makes it easier to machine
temperature-sensitive materials. High temperatures could lead to thermoplastic matrix softening,
generating defects in the final quality of the slot. The kerf taper values achieved are similar to those
Materials 2019, 12, 4192 13 of 17
studied by Wong et al. [31]. This could mean that the effect of temperature on the thermoplastic
matrix is not enough to alter the kerf taper.

P = 3400 bar
FR = 100 mm/min
AMF = 340 g/min

KT = 0.75

Figure 11. Combination

Figure Combinationthat minimizes
that kerf kerf
minimizes tapertaper
defectdefect (P =
(P = 3400 bar; FR =bar;
3400 FR = 100AMF
100 mm/min; = 340
mm/min;
= 340 g/min).
g/min).
AMF

On the other
According tohand, a combination
our mathematical of 1200
model, bar of pressure
a combination of Pand 500 mm/min
= 3400 bar, FR = of feed mm/min,
172.73 rate seemsandto
be
AMFthe =most
170 unfavorable
g/min should in result
the study of kerf
in the taper,
lowest kerfand it seems
taper valuetoofbe the element
1.79°. The degreethat ofgreatly increases
desirability of
the upward trend. As speed increases, cutting capacity decreases due to
the result achieved is also obtained. Individual and compound desirability evaluates how well a the loss of kinetic energy
in the waterjet.
combination of To this effect,
parameters it would
satisfies thebe necessary
objectives to addfor
defined theoutput
reduction of the number of abrasive
variables.
particles that affect
Individual the material.
desirability (d) evaluates how adjustments optimize a single response. The
In addition,
compound (D), onthethe
polyurethane
other hand,matrix looks atapplied in this article
how adjustments has a melting
optimize a set oftemperature 145 ◦ C.
responses inofgeneral.
Abrasive waterjet
This variable has acutting
range of is 0a to
technology
1, 1 being that generates
the ideal a lower
case, and temperature
0 meaning that somethanof other conventional
answers obtained
machining technologies, such as milling or drilling [14]. This process makes
are outside acceptable limits. In this case, a maximum desirability was generated, with a value of 1, it easier to machine
temperature-sensitive
which implies that the materials.
combination High temperatures
of cutting parameterscould lead tooffers
selected thermoplastic
a desirablematrix
result.softening,
generating defects in the final quality of the slot. The kerf taper values achieved are similar to those
studied by Wong et
3.6. Mathematical al. [31].
Model This could mean that the effect of temperature on the thermoplastic matrix
Validation
is not enough to alter the kerf taper.
Table 5 shows the taper values obtained from the complementary experimental design together
According to our mathematical model, a combination of P = 3400 bar, FR = 172.73 mm/min, and
with the same values predicted by Equation (3).
AMF = 170 g/min should result in the lowest kerf taper value of 1.79◦ . The degree of desirability
of the result achieved is also obtained. Individual and compound desirability evaluates how well a
Table 5. Complementary DOE for validation of the mathematical model.
combination of parameters satisfies the objectives defined for output variables.
Individual Pressure
desirability (d) Feed Rate
evaluates howAbrasive KT
adjustments optimize a Predicted KT
single response. Error
The compound
Test
[bar] [mm/min] [g/min] [°] [°]
(D), on the other hand, looks at how adjustments optimize a set of responses in general. This variable [%]
has a range1 of 0 to 1200
1, 1 being the100 ideal case, and225 0 meaning 3.73 4.13
that some of answers obtained 10.81
are outside
acceptable 2 limits.1200
In this case, a 300 maximum desirability
170 was generated,
6.52 with5.46
a value of 1, which
16.21 implies
that the combination
3 1200 of cutting 300 parameters selected
340 offers a5.99 desirable result.
5.98 0.12
4 1200 500 225 6.77 7.95 17.47
3.6. Mathematical Model Validation
5 2500 100 170 2.23 2.75 23.19
Table
6
5 shows the taper values
2500 100
obtained from
340
the complementary
4.01
experimental
3.76
design
6.39
together
with the same values predicted by Equation (3).
7 2500 500 170 4.68 4.89 4.37
8 2500 500
Table 5. Complementary 340validation of
DOE for 4.17 5.00 model.
the mathematical 19.85
9 3400 100 225 1.60 2.28 42.33
Pressure Feed Rate Abrasive ◦ Predicted
10Test 3400 300 170 KT
1.40[ ] 1.91 Error36.27
[%]
[bar] [mm/min] [g/min] KT [◦ ]
11 3400 300 340 1.67 2.50 49.84
1 1200 100 225 3.73 4.13 10.81
12 2 3400 1200 500300 225
170 2.32
6.52 2.76
5.46 19.07
16.21
3 1200 300 340 5.99 5.98 0.12
In order
4 to evaluate
1200 the mathematical
500 model
225 carried out in this7.95
6.77 experiment, a comparison
17.47
between combinations
5 of parameters
2500 100 not used 170
in original DOE
2.23 and predicted
2.75 values23.19
was carried out
(Figure 12).6 2500 100 340 4.01 3.76 6.39
7 2500 500 170 4.68 4.89 4.37
8 2500 500 340 4.17 5.00 19.85
9 3400 100 225 1.60 2.28 42.33
10 3400 300 170 1.40 1.91 36.27
11 3400 300 340 1.67 2.50 49.84
12 3400 500 225 2.32 2.76 19.07
 It should be noted that, in most cases, experimental values are below those predicted at a similar
distance. When the errors tabulated in Table 5 are observed, it can be seen that, for pressures of 1200
and 2500 bar, errors oscillate in a range of 0–20%, while highest pressure of 3400 bar generates high
errors with values of up to 49.84%.
The
Materials average
2019, 12, 4192 error obtained is 20.49%. Experimental values follow the trend of the predicted 14 of 17
values but maintain an almost constant difference between them. This difference may be due to the
anisotropy and nature of material, as seen in Figure 7. This model could be used to predict the trend
that In
willorder to evaluate
follow the kerf the mathematical
taper as a functionmodel carried outused,
of parameters in this experiment,
although a comparison
the error obtainedbetween
must be
combinations of
taken into account. parameters not used in original DOE and predicted values was carried out (Figure 12).

Experimental Predicted
8.00

7.00

6.00
Average KT (°)

5.00

4.00

3.00

2.00

1.00

0.00
225 170 340 225 170 340 170 340 225 170 340 225 AMF (g/min)
100 300 300 500 100 100 500 500 100 300 300 500 FR (mm/min)
1200 2500 3400 P (bar)
Test
Figure 12. Kerf taper values (experimental and predicted).
Figure 12. Kerf taper values (experimental and predicted).
It should be noted that, in most cases, experimental values are below those predicted at a similar
4. Conclusions
distance. When the errors tabulated in Table 5 are observed, it can be seen that, for pressures of 1200
An experimental
and 2500 study on
bar, errors oscillate in athe influence
range of cutting
of 0–20%, while parameters on theofgeneration
highest pressure of a geometric
3400 bar generates high
defect,with
errors called a kerf
values of uptaper, focused on the machining of carbon fiber composite materials with
to 49.84%.
thermoplastic
The average matrix,
errorwas developed.
obtained A face-centered
is 20.49%. Experimental composite experimental
values follow the trenddesign (FCD)
of the based
predicted
on a response
values surface
but maintain anmethodology
almost constant (RSM) was development.
difference between them. ThisThis
typedifference
of experimental
may be design has
due to the
given rise to a second order polynomial equation that relates input parameters
anisotropy and nature of material, as seen in Figure 7. This model could be used to predict the trend to output variable
(kerfwill
that taper).
follow the kerf taper as a function of parameters used, although the error obtained must be
takenThe intoliterature
account. review allowed for the selection of pressure, feed rate, and abrasive mass flow as
the most influential input parameters in the machining performed while keeping other parameters,
4. Conclusions
such as stand-off distance or size of abrasive particles, fixed.
An second
A experimental
experimental study ondesign was used
the influence to verify
of cutting the second
parameters on theorder polynomial
generation equation
of a geometric
generated
defect, by the
called tapertaper,
a kerf obtained, which
focused oncontrasts combinations
the machining of parameters
of carbon not usedmaterials
fiber composite in the original
with
DOE with values predicted by model. An error of 20.49% was obtained,
thermoplastic matrix, was developed. A face-centered composite experimental design (FCD) based on which can be considered
asmall
responseif thesurface
anisotropy and nature
methodology of a was
(RSM) composite materialThis
development. are taken
type ofinto account. design has given
experimental
rise toThree
a secondzones along
order thicknessequation
polynomial of materialthat have
relates been
input identified—
parametersthe rough variable
to output cutting region at 0–
(kerf taper).
10%,The the literature
smooth cutting region at 10–75%, and the initial damage region at 75–100%.
review allowed for the selection of pressure, feed rate, and abrasive mass flow as In addition,
onlymost
the the Smooth
influential Cutting
inputRegion
parameterswas taken
in theinto accountperformed
machining for the development
while keeping of this experiment.
other parameters,
The kerf taper defect was studied in
such as stand-off distance or size of abrasive particles, fixed.three specimens of the same thermoplastic composite
material in order to obtain an average value and its respective deviation.
A second experimental design was used to verify the second order polynomial equation generatedMarginal graphs show how
hydraulic
by the taper pressure
obtained, causes
which a decrease
contrastsincombinations
taper generated, and feed rate
of parameters notand
used abrasive mas flow
in the original produce
DOE with
an increase in the same.
values predicted by model. An error of 20.49% was obtained, which can be considered small if the
ANOVA
anisotropy andanalysis
nature ofhas indicated material
a composite that hydraulic
are taken pressure and feed rate are the most influential
into account.
parameters
Three zones in abrasive waterjet of
along thickness machining.
material have The been
slot walls become more
identified—The rough vertical
cuttingatregion
high pressures
at 0–10%,
the smooth cutting region at 10–75%, and the initial damage region at 75–100%. In addition, onlytothe
and low feed rates. This is due to a higher concentration of energy impacting the composite be
Smooth Cutting Region was taken into account for the development of this experiment.
The kerf taper defect was studied in three specimens of the same thermoplastic composite material
in order to obtain an average value and its respective deviation. Marginal graphs show how hydraulic
pressure causes a decrease in taper generated, and feed rate and abrasive mas flow produce an increase
in the same.
Materials 2019, 12, 4192 15 of 17

ANOVA analysis has indicated that hydraulic pressure and feed rate are the most influential
parameters in abrasive waterjet machining. The slot walls become more vertical at high pressures
and low feed rates. This is due to a higher concentration of energy impacting the composite to be
machined, which translates into higher material removal. Also, the mathematical model obtained for
analyzed response surface methodology has presented an R2 of 93.85%.
The effect of temperature does not seem to influence the quality of the results obtained by AWJM.
After a concise literature review, it seems that the results obtained in taper defect agree with those
obtained by other scientific authors.
Finally, a combination of cutting parameters that minimizes kerf taper defect was found, resulting
in a pressure of 3400 bar, a feed rate of 100 mm/min, and an abrasive mass flow of 340 g/min, producing
an upper-lower width ratio close to 1, i.e., 0.75◦ . This small kerf taper defect means that for specific
applications of AWJM could be considered as a high precision process.

Author Contributions: A.S. and F.B. developed machining tests. M.B. and J.S. developed data treatment. F.B.,
A.S., M.B., B.S., and J.S. analyzed the influence of the parameters involved. F.B. and A.S. collaborated in preparing
figures and tables and F.B., A.S., M.B., B.S., and J.S. wrote the paper.
Funding: This work was developed with the support of a pre-doctoral industrial fellowship financed by
NANOTURES SL, the Mechanical Engineering and Industrial Design Department and Vice-Rectorate of
Transference and Technological Innovation of the University of Cadiz.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Sathish, P.; Kesavan, R.; Mahaviradhan, N. Hemp Fiber Reinforced Composites: A Review. Int. J. Res. Appl.
Sci. Eng. Technol. 2017, 5, 594–595. [CrossRef]
2. Altin Karatas, M.; Gökkaya, H. A review on machinability of carbon fiber reinforced polymer (CFRP) and
glass fiber reinforced polymer (GFRP) composite materials. Def. Technol. 2018, 14, 318–326. [CrossRef]
3. Pramanik, A.; Basak, A.K.; Dong, Y.; Sarker, P.K.; Uddin, M.S.; Littlefair, G.; Dixit, A.R.; Chattopadhyaya, S.
Joining of carbon fibre reinforced polymer (CFRP) composites and aluminium alloys-A review. Compos. Part
A Appl. Sci. Manuf. 2017, 101, 1–29. [CrossRef]
4. Witten, E.; Mathes, V.; Sauer, M.; Kühnel, M. Composites Market Report 2018—Market Developments, Trends,
Outlooks and Challenges; AVK Federation of Reinforced Plastics: Frankfurt, Germany, 2018.
5. Dauguet, M.; Mantaux, O.; Perry, N.; Zhao, Y.F. Recycling of CFRP for high value applications: Effect of
sizing removal and environmental analysis of the SuperCritical Fluid Solvolysis. Procedia CIRP. 2015, 29,
734–739. [CrossRef]
6. Masek, P.; Zeman, P.; Kolar, P. Development of a cutting tool for composites with thermoplastic matrix.
Mod. Mach. Sci. J. 2013, 3, 423–427. [CrossRef]
7. Biron, M. Outline of the actual situation of plastics compared to conventional materials. In Industrial
Applications of Renewable Plastics; William Andrew: Norwich, NY, USA, 2017; ISBN 9780323480659.
8. Goto, K.; Imai, K.; Arai, M.; Ishikawa, T. Shear and tensile joint strengths of carbon fiber-reinforced
thermoplastics using ultrasonic welding. Compos. Part A 2019, 116, 126–137. [CrossRef]
9. Christmann, M.; Medina, L.; Mitschang, P. Effect of inhomogeneous temperature distribution on the
impregnation process of the continuous compression molding technology. J. Thermoplast. Compos. Mater.
2016, 30, 1–18. [CrossRef]
10. Ishikawa, T.; Amaoka, K.; Masubuchi, Y.; Yamamoto, T.; Yamanaka, A.; Arai, M.; Takahashi, J. Overview
of automotive structural composites technology developments in Japan. Compos. Sci. Technol. 2018, 155,
221–246. [CrossRef]
11. Biron, M. Thermoplastics and thermoplastic composites, 3rd ed.; William Andrew: Norwich, NY, USA, 2018;
ISBN 9781856174787.
12. Olabisi, O.; Adewale, K. Handbook of thermoplastics, 2nd ed.; Taylor & Francis Group: New York, NY, USA,
2016; ISBN 9781466577237.
13. Masek, P.; Zeman, P.; Kolar, P.; Holesovsky, F. Edge trimming of C/PPS plates. Int. J. Adv. Manuf. Technol.
2019, 101, 157–170. [CrossRef]
Materials 2019, 12, 4192 16 of 17

14. Fu, R.; Jia, Z.; Wang, F.; Jin, Y.; Sun, D.; Yang, L.; Cheng, D. Drill-exit temperature characteristics in drilling of
UD and MD CFRP composites based on infrared thermography. Int. J. Mach. Tools Manuf. 2018, 135, 24–37.
[CrossRef]
15. Masek, P.; Zeman, P.; Kolar, P. Technology optimization of PPS/C composite milling using Taguchi method.
In Proceedings of the 9th International Conference on Machine Tools Automation, Technology and Robotics,
Prague, Czech Republic, 12–14 September 2012.
16. Kakinuma, Y.; Ishida, T.; Koike, R.; Klemme, H.; Denkena, B.; Aoyama, T. Ultrafast feed drilling of carbon
fiber-reinforced thermoplastics. Procedia CIRP. 2015, 35, 91–95. [CrossRef]
17. El-Hofy, M.; Helmy, M.O.; Escobar-Palafox, G.; Kerrigan, K.; Scaife, R.; El-Hofy, H. Abrasive water jet
machining of multidirectional CFRP laminates. Procedia CIRP. 2018, 68, 535–540. [CrossRef]
18. Ramulu, M.; Isvilanonda, V.; Pahuja, R.; Hashish, M. Experimental investigation of abrasive waterjet
machining of titanium graphite laminates. Int. J. Autom. Technol. 2016, 10, 392–400. [CrossRef]
19. Melentiev, R.; Fang, F. Recent advances and challenges of abrasive jet machining. CIRP J. Manuf. Sci. Technol.
2018, 470, 1–20. [CrossRef]
20. Ruiz-Garcia, R.; Ares, P.F.M.; Vazquez-Martinez, J.M.; Gómez, J.S. Influence of abrasive waterjet parameters
on the cutting and drilling of CFRP/UNS A97075 and UNS A97075/CFRP stacks. Materials 2018, 12, 107.
[CrossRef] [PubMed]
21. Vigneshwaran, S.; Uthayakumar, M.; Arumugaprabu, V. Abrasive water jet machining of fiber-reinforced
composite materials. J. Reinf. Plast. Compos. 2017, 37, 230–237. [CrossRef]
22. Pahuja, R.; Ramulu, M.; Hashish, M. Surface quality and kerf width prediction in abrasive water jet machining
of metal-composite stacks. Compos. Part B. 2019, 175, 107134. [CrossRef]
23. Pahuja, R.; Ramulu, M. Abrasive water jet machining of Titanium (Ti6Al4V)–CFRP stacks – A semi-analytical
modeling approach in the prediction of kerf geometry. J. Manuf. Process. 2019, 39, 327–337. [CrossRef]
24. Li, M.; Huang, M.; Chen, Y.; Gong, P.; Yang, X. Effects of processing parameters on kerf characteristics and
surface integrity following abrasive waterjet slotting of Ti6Al4V/CFRP stacks. J. Manuf. Process. 2019, 42,
82–95. [CrossRef]
25. Ramulu, M.; Pahuja, R.; Hashish, M.; Isvilonanda, V. Abrasive waterjet machining effects on kerf quality in
thin fiber metal laminate. In Proceedings of the WJTA-IMCA Conference and Expo, New Orleans, LA, USA,
2–4 November 2015.
26. Pahuja, R.; Ramulu, M.; Hashish, M. Abrasive waterjet profile cutting of thick Titanium/Graphite fiber metal
laminate. In Proceedings of the ASME 2016 International Mechanical Engineering Congress and Exposition,
Phoenix, AZ, USA, 11–17 November 2016.
27. Dhanawade, A.; Kumar, S. Experimental study of delamination and kerf geometry of carbon epoxy composite
machined by abrasive water jet. J. Compos. Mater. 2017, 51, 3373–3390. [CrossRef]
28. Chen, M.; Jiang, C.; Xu, Z.; Zhan, K.; Ji, V. Experimental study on macro- and microstress state, microstructural
evolution of austenitic and ferritic steel processed by shot peening. Surf. Coat. Technol. 2019, 359, 511–519.
[CrossRef]
29. Mayuet, P.F.; Girot, F.; Lamíkiz, A.; Fernández-vidal, S.R.; Salguero, J.; Marcos, M. SOM/SEM based
characterization of internal delaminations of CFRP samples machined by AWJM. Procedia Eng. 2015, 132,
693–700. [CrossRef]
30. Popan, I.A.; Contiu, G.; Campbell, I. Investigation on standoff distance influence on kerf characteristics in
abrasive water jet cutting of composite materials. MATEC Web Conf. 2017, 137. [CrossRef]
31. Wong, M.M.I.; Azmi, A.; Lee, C.; Mansor, A. Kerf taper and delamination damage minimization of FRP
hybrid composites under abrasive water-jet machining. Int. J. Adv. Manuf. Technol. 2018, 94, 1727–1744.
32. Gupta, V.; Pandey, P.M.; Garg, M.P.; Khanna, R.; Batra, N.K. Minimization of kerf taper angle and kerf width
using Taguchi’s method in abrasive water jet machining of marble. Procedia Mater. Sci. 2014, 6, 140–149.
[CrossRef]
33. Dumbhare, P.A.; Dubey, S.; Deshpande, Y.V.; Andhare, A.B.; Barve, P.S. Modelling and multi-objective
optimization of surface roughness and kerf taper angle in abrasive water jet machining of steel. J. Brazilian
Soc. Mech. Sci. Eng. 2018, 40, 1–13. [CrossRef]
34. Hlaváč, L.M.; Krajcarz, D.; Hlaváčová, I.M.; Spadło, S. Precision comparison of analytical and
statistical-regression models for AWJ cutting. Precis. Eng. 2017, 50, 148–159. [CrossRef]
 Materials 2019, 12, 4192 17 of 17

35. Hlaváč, L.M.; Hlaváčová, I.M.; Plančár, S.; Krenický, T.; Geryk, V. Deformation of products cut on AWJ x-y
tables and its suppression. IOP Conf. Ser. Mater. Sci. Eng. 2018, 307, 1–10. [CrossRef]
36. Hlaváč, L.M.; Hlaváčová, I.M.; Geryk, V.; Plančár, Š. Investigation of the taper of kerfs cut in steels by AWJ.
Int. J. Adv. Manuf. Technol. 2015, 77, 1811–1818. [CrossRef]
37. Hlaváč, L.M.; Hlaváčová, I.M.; Geryk, V. Taper of kerfs made in rocks by abrasive water jet (AWJ). Int. J. Adv.
Manuf. Technol. 2017, 88, 443–449. [CrossRef]
38. Azmir, M.A.; Ahsan, A.K. A study of abrasive water jet machining process on glass/epoxy composite laminate.
J. Mater. Process. Technol. 2009, 209, 6168–6173. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).

View publication stats