Composite Structures 213 (2019) 144–152

Contents lists available at ScienceDirect

Composite Structures
journal homepage: www.elsevier.com/locate/compstruct

Development of vacuum-assisted prepreg compression molding for T

production of automotive roof panels
⁎
Jeong-Min Leea, Byung-Min Kima, Dae-Cheol Kob,
a
Precision Manufacturing Systems Division, Pusan National University, 63, Busandaehak-ro, Geumjeong-gu, Busan 609-735, South Korea
b
Graduate School of Convergence Science, Pusan National University, 63, Busandaehak-ro, Geumjeong-gu, Busan 609-735, South Korea

A R T I C LE I N FO A B S T R A C T

Keywords: Prepreg compression molding (PCM) is well known for its high productivity among the manufacturing processes
Roof panel for carbon fiber reinforced plastic (CFRP) products. However, the disadvantages of CFRP products manufactured
Micro groove by the PCM process are related to defects involving micro grooves and voids. The purpose of this study is to
Vacuum Assisted Process (VAP) develop a vacuum-assisted prepreg compression molding (VA-PCM) process to prevent these defects. To verify
Vacuum Assisted Prepreg Compression
the application of the VA-PCM process, the mechanical properties of small-scaled specimens fabricated by the
Molding (VA-PCM)
VA-PCM process were compared with the mechanical properties of specimens fabricated by the PCM process.
The tests related to surface roughness, void content, tensile, and three-point bending were performed to evaluate
the mechanical properties, including elastic modulus, tensile strength, bending stiffness, and energy absorption.
Finally, automotive roof panels produced by each process were evaluated by a bending test to validate the VA-
PCM process.

1. Introduction associated with large autoclaves [11–13]. The process involves diffi-
culties in installing additional equipment, such as sealant tape, vacuum
Carbon fiber reinforced plastic (CFRP) represents a lightweight bag, porous peel ply, and breather. The high-pressure resin transfer
material with superior strength and stiffness when compared to metals molding (HP-RTM) process with shorter injection times was developed
and advanced materials. However, CFRP has been only applied to ad- to increase productivity. HP-RTM improves impregnation of preform
vanced products such as airplanes and sporting goods due to the high and leads to shorter cycle times when compared to the traditional RTM
cost of manufacturing. Given the steady increase in research and de- process [14]. However, HP-RTM equipment is very expensive because it
velopment efforts focusing on manufacturing low-priced carbon fiber, required high capacity injector for fast resin injection and high-preci-
CFRP is now mass produced for automotive parts. Generally, CFRP sion machining molds to prevent resin leakage.
automotive parts are manufactured by four different processes, namely The PCM process is an alternative process to resolve the dis-
injection molding (IM), resin transfer molding (RTM), carbon-sheet advantages of IM and the RTM process. The PCM process exhibits high
mold compound (C-SMC), and prepreg compression molding (PCM) productivity and low equipment cost and consists of a simple sequence
[1–4]. in which an uncured laminate with prepregs is pressed and cured be-
Recently, there is an increasing interest in out-of-autoclave (OOA) tween heated molds designed to maintain the curing temperature [15].
processing to improve mechanical properties and productivity. The IM Fig. 1 shows the schematic sequence of the conventional PCM process:
process is appropriate for high-rate production, although the dis- (a) The prepregs are laid up on the flat plate; (b) The laminate is pre-
continuous fiber used in the process cannot be expected to exhibit heated in a chamber; (c) The preform is fabricated by the cold forming
mechanical properties such as high stiffness and strength [5–7]. In stage; (d) Following the removal of the preform from the dies, the
contrast, the RTM process using continuous fiber is expected to exhibit cutting procedure is implemented based on the product line; (e)
high stiffness and strength, although its disadvantages include low Heating dies are used to manufacture the CFRP products; and, finally,
productivity and bad surface quality caused by the equipment problems (f) The CFRP product is completed by the machining of the final pro-
and invalid parameters [8–10]. In addition, the liquid resin infusion duct line. Typically, a cavity is generated owing to the thinning of the
(LRI) process is increasingly used for both aerospace and automotive laminate caused by friction between the mold and the laminate during
parts as it offers savings over the capital and operating expenses the formation. When the forming is completed, the gravity causes the

⁎
Corresponding author.
E-mail address: [email protected] (D.-C. Ko).

https://doi.org/10.1016/j.compstruct.2019.01.092
Received 23 April 2018; Received in revised form 5 December 2018; Accepted 28 January 2019
Available online 29 January 2019
0263-8223/ © 2019 Elsevier Ltd. All rights reserved.
J.-M. Lee et al. Composite Structures 213 (2019) 144–152

Upper tool
Friction
Cavity Laminat
e
Thinning Unfilled resin
Laminate Cavity
Friction Resin Lower tool

Upper tool Vacuum pressure

Cavity
Laminat
Control of e
resin flow
Vacuum Vacuum
Friction
gauge pump Laminate
Friction Lower tool

Fig. 1. Schematic drawing of PCM and VA-PCM processes.

laminate resin to flow into the cavity. This creates defects in the CFRP process to reduce defects caused by voids and unfilled resin.
products owing to unfilled resin and voids [16]. These defects degrade The purpose of this study is to investigate the feasibility of the VA-
the mechanical properties of the CFRP products, including strength and PCM process to manufacture automotive parts. First, after preparing
stiffness, and require an additional coating process. Therefore, it is specimens by the VA-PCM process, two simple tests which is used in
necessary to develop a new process to manufacture products without other researches, namely the tensile test and three-point bending tests,
unfilled resin defects. were performed to investigate their mechanical properties, including
Several studies focusing on the vacuum-assisted process (VAP) were elastic modulus, tensile strength, flexural modulus, and energy ab-
performed to remove the defects caused by voids and unfilled resin sorption [27–32]. The effects of void content and surface roughness on
[17–19]. Vacuum-assisted resin transfer molding (VA-RTM) is a process the fracture behavior of CFRP specimens were also evaluated, based on
in which resin is drawn through the preform in a one-sided mold using the results of the tests [22,31]. Finally, automotive roof panels were
an induced pressure gradient [20–22]. Yokozeki et al. studied the ap- manufactured by both the conventional PCM process and VA-PCM
plication of a porous mold process to the VA-RTM process [23]. The VA- process. Their quality and bending stiffness were compared by per-
RTM process with a porous mold led to flat specimens of CFRP with low forming an analysis of the cross-sectional image and bending test, re-
void content and good mechanical properties, when compared to the spectively.
conventional RTM process. In addition, Li et al. examined the VAP to
improve strength and reduce void content [24]. Several studies re- 2. Feasibility study of VA-PCM process
ported that adding a VAP reduced void content because pressure acting
uniformly throughout the material minimizes the voids generated by To verify the effectiveness of the VA-PCM process, the flat plates
the excessive resin flow [20,23,24]. However, there is a paucity of were fabricated by PCM and VA-PCM. The fabricated flat plates were
studies examining PCM process with VAP. used for four types of tests: surface roughness, void content, tensile, and
When the VAP is applied to the PCM process, the resin flow is three-point bending test. The depth of the micro grooves on the surface
controlled by the vacuum film, even if the cavities are generated by the was evaluated using a three-dimensional (3D) microscope. The voids
thinning laminate, as shown in Fig. 1. The VA-PCM process is identical content was measured by the acid leaching method. Finally, the effects
to the conventional PCM process until step (c) and placed prior to the of the micro grooves and voids were evaluated for the mechanical
cutting process. The VA-PCM process is briefly explained as follows. properties of the elastic modulus, tensile strength, flexural modulus,
After removal of the upper die, the vacuum equipment, including the and energy absorption.
release film, breather, and vacuum film, are placed on preform surfaces
(d-1). The vacuum is used to control the resin flow and remove the 2.1. Manufacturing CFRP testing plates using PCM and VA-PCM processes
defects related to voids and unfilled resin without causing excessive
resin flow during the pressing stage (d-2) of the cure cycle (d-3). The CFRP specimens were fabricated by conventional PCM and VA-PCM
following steps are the same as those in the conventional PCM process. processes to investigate whether the VA-PCM process was effective in
If resin content of prepreg is increased to remove the defects such as preventing defects, including void and unfilled resin. In the study, the
micro grooves and voids, the void content will increase owing to ex- CFRP laminates using plain weave were used to fabricate the specimens
cessive resin flow [25,26]. Thus, it is effective to apply the VA-PCM as shown in Fig. 2. The prepregs(SK Chemical, UGN200) with carbon

PCM process VA-PCM process Breather

Release film
CFRP prepreg CFRP prepreg

Stopper
Release film

Stopper
Sealing tape
Vacuum hose Vacuum film

Fig. 2. Processes of PCM and VA-PCM for manufacturing of test specimens.

145
J.-M. Lee et al. Composite Structures 213 (2019) 144–152

fiber and a thermosetting epoxy resin (bisphenol-A type properties of CFRP plates based on whether or not a micro groove is
epoxy + phenol novolac type epoxy) were cured for 120 min at 130 °C. formed.
The initial thickness of one prepreg ply was 0.3 mm prior to pressing.
The final thickness of the laminate with 10 plies was 2.0 mm, and the 2.3. Void contents of the CFRP plates
fiber volume fraction of the manufactured laminate was approximately
47%. The void content and fiber volume fraction of the CFRP plates were
The CFRP plates were fabricated by stacking CFRP prepregs that evaluated by an acid leaching method as per ASTM D3171 [36]. As
were cut into dimensions corresponding to 300 × 300 mm in the 0° shown in Fig. 3(a), the void content and fiber volume fraction were
direction. A release film was placed at the interface between the molds measured with a dimension of 25 × 25 mm in nine areas. First, the
and prepregs to prevent sticking. In the PCM process, pressing (20 kN) weight of the specimens was measured in water and air. The CFRP
and heating stages using upper and bottom molds were simultaneously matrix was removed in hot nitric acid (70%) for a duration of 360 min
applied to the prepregs without additional equipment. However, the at a temperature of 70 °C. After dissolving in hot nitric acid, the carbon
CFRP plates in the VA-PCM process were manufactured by using a fiber was subsequently filtered, washed, dried, cooled, and weighed.
breather and vacuum film for air intake, as shown in Fig. 2. The plates The fiber volume fraction and void content of the CFRP specimen
were cut by a water-jet to minimize dimensional errors and to avoid were summarized as shown in Tables 3 and 4, respectively. The void
delamination in the test specimens. The test specimens were cut contents of CFRP plates manufactured by the PCM process slightly ex-
250 × 250 mm from the flat plate center that approximately 25 mm of ceeded those of the CFRP plates manufactured by the VA-PCM process.
the outer area of the laminates was removed due to potential fiber Generally, the void in a composite is generated owing to the velocity
distortions which may have negative influence on the measured prop- difference of the resin flow by the non-uniform pressure [37]. The
erties. The thickness variation of the test specimens after water-jet conventional PCM process does not include any devices to prevent resin
cutting was ± 0.02 mm. flow toward the exterior and cavity; thus, the void was generated by the
velocity difference in the resin flow. However, the VA-PCM process
2.2. Surface roughness evaluation of the CFRP laminate reduces the velocity difference in the resin flow because the vacuum
film controls the resin flow toward the exterior and cavity. The void
To minimize the measurement error, a specimen with a dimension content in Table 4 was slightly different for the manufacturing pro-
of 100 × 100 mm was obtained from the center of the flat plate. The cesses owing to the simple shape of the CFRP plate; however, no sig-
surface roughness was measured at nine areas (A1–A9), as shown in nificant differences were found.
Fig. 3(a). The average values of surface roughness for the CFRP plate
manufactured by PCM and VA-PCM were measured to be 0.93 and 0.84, 2.4. Comparison of mechanical properties
respectively. It can be confirmed that the surface roughness of the CFRP
plate by VA-PCM was improved by 0.09 (See Table 1). Mechanical properties of the CFRP specimens including elastic
To obtain more detailed information on surface roughness, the modulus and tensile strength were obtained by performing a tensile
surface profiles of micro grooves were measured using a 3D microscope test, as shown in Fig. 5(a). The tensile specimens were fabricated in the
(HIROX digital microscope KH-8700). In contrast to the specimen 0° direction along the warp yarns with a dimension of 250 × 25 mm as
manufactured by the VA-PCM process, the results indicate that the per ASTM D3039 [38]. As shown in Table 5, the average elastic mod-
defects related to unfilled resin exist on the surface of CFRP plates ulus and ultimate tensile strength (UTS) of the CFRP specimens man-
manufactured by the conventional process. As shown in Fig. 3(b), the ufactured by the conventional PCM process were 45.92 GPa and
micro groove resembling a crack shape, because of the unfilled resin, 843.49 MPa, respectively. In the tensile test of the CFRP specimens
was observed over the entire surface. manufactured via the VA-PCM process, the elastic modulus exhibited a
Fig. 4 shows the surface profiles of the micro grooves for each similar trend with a value of 47.13 GPa, because the tension behavior of
process. The average micro groove depth of three specimens manu- three specimens was identical prior to the fracture, as shown in
factured by the VA-PCM process was 50.05 μm. The average micro Fig. 5(b). The UTS of specimens manufactured via the VA-PCM process
groove depth of approximately 126.52 μm was observed on the surface were 890.85 MPa, which exceeds that of specimens obtained by the
of three specimens manufactured by the conventional PCM process as conventional PCM process.
shown in Table 2. Several studies conducted fatigue tests with a micro As shown in Fig. 6(a), the delamination initially occurred at the
groove depth of 100 μm to observe crack initiation [33–35]. These types micro groove owing to the stress concentration by the tensile load. The
of micro groove depths that exceed 100 μm deteriorate the mechanical results indicated that delamination on the surface layer occurred owing
properties of the product because of crack initiation by stress con- to the crack at the micro groove, as shown in Fig. 6(b). In specimens
centrations. Therefore, it is necessary to evaluate the mechanical made by VA-PCM, the occurrence of the microgroove and the intra-

100mm
PCM plate surface Micro groove
25mm
z
25mm

A1 A4 A7
X
100mm

y
A2 A5 A8
Warp
X
Weft
A3 A6 A9 Micro groove Cross-section 200 m

(a) Dimension of CFRP plate specimen (b) Surface and cross-sectional image
Fig. 3. CFRP plate specimen and image of micro groove.

146
J.-M. Lee et al. Composite Structures 213 (2019) 144–152

Table 1
Surface roughness for the each sections.
Section 1 2 3 4 5 6 7 8 9 Avg. SD

PCM 1.21 0.81 0.74 0.81 0.85 0.77 1.02 0.96 1.21 0.93 0.17
VA-PCM 0.93 0.75 1.01 1.08 0.87 0.74 0.6 0.79 0.77 0.84 0.14

laminar void was prevented, so that delamination-induced fracture was Table 2

suppressed. Thus, the UTS of CFRP improved by approximately 5.32% Micro groove depth according to the manufacturing processes.
by preventing the micro grooves on the surface and intra-laminar voids. Process Conventional PCM VA-PCM
The three-point bending tests were conducted to assess the bending
properties of CFRP plates manufactured by each process. Fig. 7(a) Specimens #1 #2 #3 Avg. #1 #2 #3 Avg.
shows the equipment comprising the three-point bending test per-
Micro 142.78 128.33 108.45 126.52 63.21 29.52 57.42 50.05
formed on a universal test machine (MTS, 100 kN) with a cross-head groove
speed of 10 mm/min. The dimensions corresponding to specimen depth
length, width, and thickness are 120 × 25 × 2 mm, respectively. The (μm)
three-point bending tests were performed on the samples cut along 0°
direction along warp yarns from CFRP composite plate stacked by plain
weave [0,90]10. Fig. 7(b) shows the graph of the stress roller stroke fracture properties of the CFRP product because several micro grooves
versus bending load. In the CFRP specimen involving the PCM process, are generated.
the fracture stroke was shorter than that of the CFRP specimen invol-
ving the VA-PCM process. However, the graph of the bending test for 3. Manufacturing automotive roof panels
the two cases exhibits a similar gradient. To quantify these results, the
flexural modulus and energy absorption (EA) were measured, and the This section discusses application of the VA-PCM process to the
results are summarized in Table 6. manufacture of automotive roof panels. The design methods and com-
The flexural modulus of conventional PCM and VA-PCM were parison of the mechanical properties of the CFRP roof panels are as
51.98 GPa and 52.43 GPa, respectively. The results can be attributed to follows. First, the thickness of the product fulfilling the required
the same reason as the elastic modulus in the tensile tests. However, the bending stiffness of the existing steel product was determined by
EA indicated that the results of the VA-PCM process were better than structural analysis. Next, the roof panels were manufactured by PCM
those of the conventional PCM process. The EA of the VA-PCM process and VA-PCM. The roof panels were used for cross-sectional observations
improved up to 19% when compared with that of the conventional PCM and measurements of the void contents, and in bending tests to assess
process, where the crack occurred at the micro groove for the same the validity of the VA-PCM process. Finally, the effectiveness of the VA-
reasons as mentioned above. During the bending test, a cross-sectional PCM process was evaluated by comparing the energy absorption and
inspection of the specimen was conducted to investigate the crack the void contents of the CFRP roof panels produced by both processes.
growth under the same conditions as those in the tensile tests. The first
crack was detected at a stroke of approximately 9.5 mm, although mi-
croscopic damage was not observed in the case of the specimen invol- 3.1. Thickness determination using finite element analysis
ving the conventional PCM process. At a stroke of 10.2 mm, the second
crack was detected, and delamination on the bottom surface was ob- The thickness of the CFRP panel that most significantly affected the
served, as shown in Fig. 8(a). Specifically, delamination occurs more bending stiffness was determined to exceed the required bending stiff-
easily at the void, as shown in the initial state. The crack propagated ness target value of 20 N/mm, considering the stiffness of the steel
from the micro groove induced delamination and caused a fracture at product. The bending stiffness indicated the gradient of the load stroke
11.2 mm. In the case of the VA-PCM specimen, the results confirmed at 30 mm, which was the pure bending area before the tension acted in
that the first delamination occurred in the middle of the cross section at the transverse direction of the roof panel [28]. In the study, the prepreg
a stroke of 12.1 mm, as shown in Fig. 8(b). The fracture occurred at stacking angles were not considered in the design of the product be-
13.3 mm owing to crack growth in the middle of the cross section. cause the purpose of the study involved the evaluation of products
Consequently, the micro grooves formed on the CFRP surface caused manufactured by each process. A 3D finite element- (FE) analysis was
the fracture to occur more quickly. A CFRP product with a complicated conducted to determine the thickness of the roof panels. A dynamic
shape by the PCM process significantly influences degradation in the explicit model was developed using ABAQUS software to numerically
investigate the effect of thickness on the bending stiffness of the CFRP

PCF process #2
140 Avg. depth : 126.52ȝm 140 VA-PCM process #1

120 120 Avg. depth : 50.05ȝm VA-PCM process #2

Z-coordinates (ȝm)

Z-coordinates (ȝm)

VA-PCM process #3
100 100

80 80

60 PCM process #1 60

40 PCM process #2 40

20 PCM process #3 20

0 0
0 500 1000 1500 2000 0 400 800 1200 1600 2000
X-coordinates (ȝm)
X-coordinates (ȝm)

(a) Conventional PCM process (b) VA-PCM process

Fig. 4. Surface roughness of micro groove.

147
J.-M. Lee et al. Composite Structures 213 (2019) 144–152

Table 3
Fiber volume fraction according to the manufacturing processes.
Section 1 2 3 4 5 6 7 8 9 Avg. SD

PCM (%) 47.52 46.95 47.44 48.02 47.25 47.88 47.64 48.1 47.33 47.57 0.36
VA-PCM (%) 47.82 47.77 48.01 48.24 47.38 47.38 47.24 48.11 47.58 47.73 0.33

Table 4
Void contents for the each sections.
Section 1 2 3 4 5 6 7 8 9 Avg. SD

PCM (%) 2.72 2.67 2.45 2.4 2.63 2.59 2.54 2.41 2.67 2.56 0.11
VA-PCM (%) 2.30 2.20 2.30 2.21 2.29 2.22 2.21 2.28 2.30 2.26 0.04

1000
Upper grip PCM process #1
PCM process #2
PCM process #3
800
VA-PCM process #1
VA-PCM process #2

Stress(MPa)
VA-PCM process #3
600
Extension
meter

400
CFRP

200
2mm/min

0
0.0 0.5 1.0 1.5 2.0 2.5
Lower grip
Strain(%)
(a) Experimental equipment for the tensile tests (b) Stress-strain curve obtained from the tensile tests
Fig. 5. Experimental equipment and results of tensile tests.

Table 5 whereas the roof panels with a thickness of 2.0 mm with 10 layers of
Mechanical properties of CFRP for the tensile tests. plain weave satisfied the target value of 20 N/mm at a stroke of 30 mm.
Process Conventional PCM VA-PCM
3.2. CFRP roof panels manufactured using VA-PCM
Specimens #1 #2 #3 Avg. #1 #2 #3 Avg.

Elastic modulus 46.0 46.8 45.0 45.9 47.0 47.8 46.5 47.1 The CFRP roof panels were manufactured to validate VA-PCM
(GPa) process reliability. Fig. 10 shows the experimental equipment used in
UTS (MPa) 853.8 841.4 835.3 843.5 885.5 899.7 887.3 890.3 the VA-PCM process to manufacture the CFRP roof panel. The lay-up
conditions were similar to those of the CFRP specimen and material,
and the fabrication test was performed using heated molds. The tem-
roof panels, as shown in Fig. 9(a). The cross-head speed of the punch perature was controlled by using a cartridge heater inserted into the
was 50 mm/min, and the punch radius was 100 mm. Fixed boundary upper and lower molds. The following is a description of the VA-PCM
conditions were imposed at the exterior of the CFRP roof panels to more process for manufacturing the CFRP roof panels. First, 10 layers of
accurately model real products assembled with automotive parts. A prepregs cut into 1800 × 1200 mm shapes were laid up based on the
bending analysis of the roof panels modeled with four different thick- simple lay-up conditions of [0,90]10 and subsequently placed on the
nesses of 1.0 mm, 1.6 mm, 2.0 mm, and 2.6 mm was performed to in- lower die. To prevent the prepreg from sticking to the mold, a liquid
vestigate whether the target value of 20 N/mm was satisfied. Fig. 9(b) release agent for high temperature applications was applied on the
shows a graph of the punch stroke versus bending load. The bending upper and lower molds. Thereafter, the upper and lower molds were
analysis of the CFRP roof panels indicated that the stiffness for the preheated to 60 °C to soften the epoxy resin, which increases the
thickness of 1.0 mm with five layers of plain weave was 8 N/mm, formability of the CFRP laminate [39,40]. In the next step, preforming

Longitudinal
direction Tensile load
Transverse
direction
Tensile load acting on warp

Tensile load Delamination initiation

Delamination
Micro groove
Resin 1mm
Micro groove Weft Warp

Fig. 6. Delamination initiation and fracture area of the tensile specimen manufactured by PCM process.

148
J.-M. Lee et al. Composite Structures 213 (2019) 144–152

600
VA-PCM process #1
VA-PCM process #2
500 VA-PCM process #3
Stress roller PCM process #1
PCM process #2
400 PCM process #3

Load(N)
10mm/min
300

200
Support Support
roller roller
100

0
0 2 4 6 8 10 12 14 16
Stroke(mm)
(a) Experimental equipment for the bending tests (b) Stress-strain curve obtained from the bending tests
Fig. 7. Experimental equipment and results of bending tests.

Table 6 3.3. CFRP roof panel evaluation

Mechanical properties of CFRP for the bending tests.
Process Conventional PCM VA-PCM
To compare surface quality, the wall and corner sections of the
CFRP roof panels were evaluated, as shown in Fig. 12. In the CFRP roof
Specimens #1 #2 #3 Avg. #1 #2 #3 Avg. panels manufactured by the conventional PCM process, the defects
caused by unfilled resin were observed at the corner section. However,
Flexural modulus (GPa) 53.3 51.3 51.3 52.0 54.0 52.4 50.9 52.4
Energy absorption (kJ) 2.97 2.78 2.85 2.87 3.56 3.47 3.60 3.54
the defect caused by unfilled resin was not observed in the product
manufactured by the VA-PCM process. Three representative cross-sec-
tional images of A area in Fig. 12 were used to investigate the defect in
the CFRP roof panel in further detail, as shown in Fig. 13(a). The de-
Delamination fects, including unfilled resin, void, and delamination, were observed in
Void Fracture the cross-sectional images of (1), (2), and (3) in Fig. 13(a). Conversely,
with respect to the CFRP roof panels fabricated by the VA-PCM process,
Void
the defects were barely observed although the few voids were con-
Micro firmed in Fig. 13(b). The void content of the CFRP roof panels in the
Micro
groove area 1–7 of Fig. 12 was quantitatively evaluated by an acid leaching
0.1mm grooove 0.1mm 0.1mm
method as per ASTM D3171, as shown in Table 7. The void content of
Initial state Stroke : 10.2mm Stroke : 11.2mm
the CFRP roof panels manufactured by the VA-PCM process was re-
(a) PCM process duced by approximately 2.64% when compared to the products man-
ufactured by the conventional process. The result shows the same
Crack initaitaion Fracture
tendency as the result mentioned in Section 2.4.
at the neutral line
There are two main reasons for the occurrence of voids. First, va-
cuum pressure was low. It was necessary to use a high capacity vacuum
pump in the process to resolve the aforementioned problem.
0.5mm Alternatively, the installation of additional air intake holes could have
0.1mm 0.5mm
been used to easily remove the voids. Second, the voids were caused by
Initial state Stroke : 12.1mm Stroke : 13.3mm
the lack of resin content owing to resin leakage at the flange area.
(b) VA-PCM process Therefore, increasing the resin content of prepreg or sealing the flange
Fig. 8. Fracture images of the bending specimens manufactured by PCM and
line of the product improves the quality of the CFRP roof panels.
VA-PCM processes. CFRP roof panels manufactured by PCM and VA-PCM were eval-
uated by performing a bending test to observe the effects of defects on
their mechanical properties. Fig. 14(a) shows the bending test equip-
was performed prior to the vacuum bag process. In contrast with the
ment, including a universal test machine. The boundaries of the CFRP
conventional PCM process, the vacuum bag was immediately installed
roof panels were constrained under the same conditions as those used in
on the dies because the cutting process was omitted. To uniformly apply
the FE-analysis. The punch radius and cross-head speed were 100 mm
the vacuum pressure, a breather was laid up on the release film. A
and 50 mm/min, respectively.
vacuum film was installed on the bottom die to seal the laminated CFRP
Fig. 14(b) shows the result of the bending test for the CFRP roof
within an airtight envelope. Furthermore, the temperature of the CFRP
panels manufactured by each process based on the load per punch
laminate was maintained at 80 °C for a few minutes (5 min) such that
stroke. A comparison of the experimental results indicates that the
the resin flowed evenly during the compression molding. Finally, using
bending loads obtained using the CFRP roof panels manufactured by
an upper die, vacuum pressure and compression were applied during
the conventional PCM process were lower than those fabricated by the
the laminate curing process for 90 min at 130 °C (i.e., the curing tem-
VA-PCM process. The bending stiffness of the CFRP roof panels manu-
perature of epoxy resin). To prevent distortion owing to temperature
factured by the conventional PCM process was 13.61 N/mm at a stroke
differences, after the CFRP roof panels were cured, they were naturally
of 30 mm. However, the result for the CFRP roof panels fabricated by
cooled and removed from the dies when the temperature reached below
the VA-PCM process satisfied the target value of 20 N/mm.
80 °C, as shown in Fig. 11 [41].
Finally, the EA was evaluated up to a stroke of 30 mm. The EAs of
the CFRP roof panels manufactured by PCM and VA-PCM processes
were 6.764 kJ and 10.625 kJ, respectively. The EAs increased by

149
J.-M. Lee et al. Composite Structures 213 (2019) 144–152

1.0t
Punch 5000 1.6t

Roof panel
2.0t Pure Bending +
2.6t
4000 bending tention

Load (N)
3000
20 N/mm
2000
11
93

Gradient
.9m

1000
.9mm
1507
m

z
y
x 0
Boundary condition : x = y = z = Fix 0 10 20 30 40 50
Stroke (mm)
(a) FE-model for bending analysis (b) Load-stroke graph obtained from bending analysis
Fig. 9. FE-model and results for bending analysis of roof panel.

Hydraulic press : 500 ton

Punch
Vacuum film
Sealing tape
Punch

Breather
Release film

Die Die

Vacuum hole Breather

Vacuum Release film
Vacuum film
pump
Sealing tape
Cartridge heater

Fig. 10. Experimental equipment for manufacturing of roof panel by VA-PCM process.

140 Curing : 120 minute 1193.9mm

120 A area
7 (3)
Temperature(°C)

100 Cooling & Demolding

20 minute 6
80 CFRP 5
10mm
roof panel
10mm

60 1 2 3 4
Forming & Vacuuming 55mm
40 : 10 minute (2)
(1)
20
0 20 40 60 80 100 120 140 A area
Time (min)
PCF VA-PCF
Fig. 11. Processing time for manufacturing CFRP roof panel.

approximately 36% when compared to that of conventional roof panels.

The results indicated that the VA-PCM process can be applied to au-
tomotive parts manufacturing to improve mechanical properties and Unfilled
resin
reduce defects.

Fig. 12. Surface quality of roof panel corner.

4. Conclusions

composite.
The purpose of this study involved investigating the feasibility of
the VA-PCM process for manufacturing automotive parts. To verify its
1) The feasibility of the VA-PCM process was examined by manu-
feasibility, the VA-PCM process was applied to small-scaled CFRP
facturing real size automobile roof panels. Defects caused by the
plates, and the defects and mechanical properties were compared with
unfilled resin and voids were observed at the corner section of the
those of the CFRP plates manufactured by the conventional PCM pro-
roof panels manufactured by the conventional PCM process.
cess. The following results were obtained by applying the VA-PCM
However, with the VA-PCM process, the defects in the CFRP roof
process to automotive roof panels made of plain weave CFRP

150
J.-M. Lee et al. Composite Structures 213 (2019) 144–152

(1) (2) (3)

Unfilled resin Unfilled resin
Unfilled resin

Delamination
Void

10mm 10mm
X100 10mm X100 X100

(a) Roof panel manufactured by PCM process

(1) (2) (3)
Void

10mm 10mm 10mm

X100 X100 X100

(b) Roof panel manufactured by VA-PCM process

Fig. 13. Cross-sectional image of A area. (a) Roof panel manufactured by PCM process. (b) Roof panel manufactured by VA-PCM process.

Table 7 5.32% and 19%, respectively, compared with the conventional

Void contents for CFRP roof panel. CFRP specimens.
Section 1 2 3 4 5 6 7 Avg. SD 4) The bending test of the roof panel was performed to evaluate the EA.
The EA of the CFRP roof panels manufactured by the VA-PCM
PCM (%) 5.28 5.14 5.98 6.78 6.27 5.91 7.33 6.10 0.72 process increased approximately 36% when compared to that of the
VA-PCM (%) 2.46 3.76 3.14 4.10 3.36 4.63 2.79 3.46 0.70
conventional roof panels. The results indicate that the application of
the VA-PCM process to automotive parts manufacturing improves
mechanical properties and reduces defects.
panels were reduced owing to the VAP effects of resin flow control
5) Because the VA-PCM process uses vacuum film, breather, release
and uniform pressure.
film, and vacuum tape, the unit process cost is higher than the
2) Several micro grooves were observed on the surface of the CFRP
conventional PCM process. However, it is possible to reduce the
specimens cut from plates manufactured by the conventional PCM
defect rate of the CFRP products, and to complete the forming and
process; however, they were not found in the specimens fabricated
curing process with one mold. Hence, the VA-PCM process can be
by the VA-PCM process. The average micro groove depth of ap-
applied to produce automotive roof panels at reduced equipment
proximately 126.52 μm, which caused the crack, was observed on
and total process costs.
the surface of CFRP specimen manufactured by conventional PCM
process. The average micro groove depth of the CFRP specimens
manufactured by the VA-PCM process was 50.05 μm.
Acknowledgements
3) Tensile and bending tests were performed on CFRP plate specimens
to evaluate the effects of micro grooves and voids on the mechanical
This work was supported by the National Research Foundation of
properties. A crack propagated in the micro groove induced dela-
Korea(NRF) grant funded by the Korea government(MSIT) (No.
mination in both tests. The UTS and the EA of the CFRP specimens
2012R1A5A1048294)
cut from plates fabricated by the VA-PCM process were improved by

700
Roof panel manufactured by VA-PCM
Roof panel manufactured by PCM
Holder 600

Holder 500
Holder
Holder Punch
10.625 kJ
Load (N)

400
Roof panel
300

Holder 200 6.764 kJ

100

0
0 5 10 15 20 25 30
Stroke (mm)
(a) Experimental equipment for the bending tests (b) Load-stroke graph obtained from the bending tests
Fig. 14. Comparison of results for the bending tests of roof panels.

151
J.-M. Lee et al. Composite Structures 213 (2019) 144–152

References [21] Nash NH, Young TM, Stanley WF. The reversibility of Mode-I and-II interlaminar
fracture toughness after hydrothermal aging of Carbon/Benzoxazine composites
with a thermoplastic toughening interlayer. Compos Struct 2016;152:558–67.
[1] Pimenta SP, Silvestre T. Recycling carbon fibre reinforced polymers for structural [22] Nasir MA, Khan ZM, Nauman S, Anas S, Pasha A, Khalil S, et al. Tensile Strength
applications: technology review and market outlook. Waste Manage changeability of Multilayered Composites, fabricated through Optimized “VARTM”
2011;31(2):378–92. Technology, An experimental Approach. University of Engineering and Technology
[2] Han SH, Cho EJ, Lee HC, Jeong K, Kim SS. Study on high-speed RTM to reduce the Taxila. Technical J 2015;20(4):55.
impregnation time of carbon/epoxy composites. Compos Struct 2015;119:50–8. [23] Yokozeki T, Kobayashi Y, Aoki T, Yoshida D, Hirata T. VaRTM process of compo-
[3] Wulfsberg J, Herrmann A, Ziegmann G, Lonsdorfer G, Stöß N, Fette M. Combination sites using porous mold. Adv Compos Mater 2013;22(2):99–107.
of carbon fibre sheet moulding compound and prepreg compression moulding in [24] Li W, Krehl J, Gillespie JW, Heider D, Endrulat M, Hochrein K, et al. Process and
aerospace industry. Procedia Eng 2014;81:1601–7. performance evaluation of the vacuum-assisted process. J Compos Mater
[4] Verrey J, Wakeman MD, Michaud V, Månson JA. Manufacturing cost comparison of 2004;38(20):1803–14.
thermoplastic and thermoset RTM for an automotive floor pan. Compos A Appl Sci [25] Poodts E, Minak G, Dolcini E, Donati L. FE analysis and production experience of a
Manuf 2006;37(1):9–22. sandwich structure component manufactured by means of vacuum assisted resin
[5] Vu-Khanh T, Denault J, Habib P, Low A. The effects of injection molding on the infusion process. Compos B Eng 2013;53:179–86.
mechanical behavior of long-fiber reinforced PBT/PET blends. Compos Sci Technol [26] Dong CJ. Development of a process model for the vacuum assisted resin transfer
1991;40(4):423–35. molding simulation by the response surface method. Compos A Appl Sci Manuf
[6] Fu SY, Lauke B, Mäder E, Yue CY, Hu X. Tensile properties of short-glass-fiber-and 2006;37(9):1316–24.
short-carbon-fiber-reinforced polypropylene composites. Compos A Appl Sci Manuf [27] Bambach MR. Fibre composite strengthening of thin steel passenger vehicle roof
2000;31(10):1117–25. structures. Thin-Walled Structures 2014;74:1–11.
[7] Lee WJ, Lee SE, Kim CG. The mechanical properties of MWNT/PMMA nano- [28] Lee JM, Lee KH, Kim BM, Ko DC. Design of roof panel with required bending
composites fabricated by modified injection molding. Compos Struct stiffness using CFRP laminates. Int J Precis Eng Manuf 2016;17(4):479–85.
2006;76(4):406–10. [29] Velea MN, Wennhage P, Zenkert D. Multi-objective optimisation of vehicle bodies
[8] Taketa I, Yamaguchi K, Wadahara E, Yamasaki M, Sekido T, Kitano A. The CFRP made of FRP sandwich structures. Compos Struct 2014;111(1):75–84.
automobile body project in Japan. Proceedings of 12th US-Japan Conference [30] Youn BD, Choi KK, Yang RJ, Gu L. Reliability-based design optimization for
Computational Materials; 2006. p. 411–21. crashworthiness of vehicle side impact. Struct Multidiscip Optim
[9] Palardy G, Hubert P, Haider M, Lessard L. Optimization of RTM processing para- 2004;26(3–4):272–83.
meters for Class A surface finish. Compos B Eng 2008;39(7–8):1280–6. [31] Ben G, Sakata K. Fast fabrication method and evaluation of performance of hybrid
[10] Gebart BR. Permeability of unidirectional reinforcements for RTM. J Compos Mater FRTPs for applying them to automotive structural members. Compos Struct
1992;26(8):1100–33. 2015;133:1160–7.
[11] Comer A, Ray D, Obande W, Clancy G, Stanley W. OOA (Out-of-Autoclave) man- [32] Kobayashi S, Kitagawa J. Effect of fine particle incorporation into matrix on me-
ufacturing of benzoxazine resin-systems by LRI (Liquid-Resin-Infusion) for ambient chanical properties of plain woven carbon fiber reinforced plastics fabricated with
and high temperature aerospace applications. Paris: 35th International Technical vacuum assisted resin transfer molding. Compos B Eng 2016;85:31–40.
Conference & Forum, SAMPE Europe, Porte de Versaille; 2014. [33] Cheng L, Gui YT. Surface crack detection for carbon fiber reinforced plastic (CFRP)
[12] Wang P, Drapier S, Molimard J, Vautrin A, Minni JC. Characterization of Liquid materials using pulsed eddy current thermography. IEEE Sens J
Resin Infusion (LRI) filling by fringe pattern projection and in situ thermocouples. 2011;11(12):3261–8.
Compos A Appl Sci Manuf 2010;41(1):36–44. [34] Cheng L, Tian GY. An evaluation on the growth rate of small fatigue cracks in cast
[13] Kleineberg M, Herbeck L, Schöppinger C. Advanced liquid resin infusion-A new AM50 magnesium alloy at different temperatures in vacuum conditions. Int J
perspective for space structures. Toulouse, France: European Conference on Fatigue 2006;28(1):79–86.
Spacecraft Structures, Materials and Mechanical Testing; 2002. p. 173–80. [35] Morton J, Kellas S. Damage characteristics in notched carbon fiber composites
[14] Rosenberg P, Chaudhari R, Karcher M, Henning F, Elsner P. Investigating cavity subjected to fatigue loading – environmental effects. J Compos Mater
pressure behavior in high-pressure RTM process variants. AIP Conf Proc 1988;22(7):657–73.
2014;1593:463–6. [36] ASTM D3171. Standard Test Methods for Constituent Content of Composite
[15] Akiyama K. Development of PCM* technology. 11th-Annual SPE Automotive Materials. ASTM International; 2010. p. 1–6.
Composites Conference & Exhibition (ACCE). [37] Jinlian H, Liu Y, Shao X. Study on void formation in multi-layer woven fabrics.
[16] McGee SH. Curing characteristics of particulate-filled thermosets. Polym Eng Sci Compos A Appl Sci Manuf 2004;35(5):595–603.
1982;22(8):484–91. [38] ASTM D3039. Standard Test Method for Tensile Properties of Polymer Matrix
[17] Balzer BB, McNabb J. Significant effect of microwave curing on tensile strength of Composite Materials. ASTM International; 2005. p. 1–13.
carbon fiber composites. J Indust Technol 2008;24(3):2–8. [39] Cartwright B, de Luca P, Wang J, Stellbrink K, Paton R. Some proposed experi-
[18] Witik RA, Gaille F, Teuscher R, Ringwald H, Michaud V, Månson JAE. Economic mental tests for use in finite element simulation of composite forming. Paris: 12th
and environmental assessment of alternative production methods for composite International Conference on Composite Materials (ICCM-12); 1999. p. 12.
aircraft components. J Cleaner Prod 2012;29:91–102. [40] Poodts E, Minak G, Mazzocchetti L, Giorgini L. Fabrication, process simulation and
[19] Garschke C, Weimer C, Parlevliet PP, Fox BL. Out-of-autoclave cure cycle study of a testing of a thick CFRP component using the RTM process. Compos B Eng
resin film infusion process using in situ process monitoring. Compos A Appl Sci 2014;56:673–80.
Manuf 2012;43(6):935–44. [41] Miyano Y, Nakada M, McMurray MK, Muki R. Prediction of flexural fatigue strength
[20] Ortiz de Mendibil I, Aretxabaleta L, Sarrionandia M, Mateos M, Aurrekoetxea J. of CRFP composites under arbitrary frequency, stress ratio and temperature. J
Impact behaviour of glass fibre-reinforced epoxy/aluminium fibre metal laminate Compos Mater 1997;31(6):619–38.
manufactured by Vacuum Assisted Resin Transfer Moulding. Compos Struct
2016;140:118–24.

152