The final fracture of composites occurs due to the cumulative integration of microfractures. Therefore, the mechanical properties of textile composites can be known by studying the initiation of microfractures in composites.
Knee point, which is known as the transition from linear to nonlinear behavior, occurs in the stress-strain curve. Knee point appears as the consequence of the initial microfractures. Knee point is an important parameter to understand the initial fracture behavior of composites, because any fracture does not occur until the knee point. After the knee point, microfractures progressed and accumulated until final fracture.
The initial fracture behavior of woven fabric composites was investigated by several researches [1–4]. Osada et al.  investigated the initial fracture behavior of satin woven fabric composites and studied the effects of the changes in crimp ratio and aspect ratio on the initial fracture behavior of woven fabric composites.
The fracture behavior of composites can be investigated by acoustic emission (AE). AE is a phenomenon of stress wave radiation caused by a dynamic reconstruction of material’s structure that accompanies the processes of deformation and fracture. The modes of elastic wave propagation are longitudinal (dilatational, P-), shear (or transverse, or distortional, or equivolumal, S-), Rayleigh (or surface), lamb (or plate), stoneley (or interfacial), love, and creeping. In the thinnest plates, only lamb wave arrivals are visible. The gradual decrease in AE amplitude due to energy loss mechanisms from dispersion, diffraction, or scattering is called attenuation. Plastic deformation is a major source of AE .
AE occurs when stored energy is instantly released from localized sources within a material under external loadings such as static, dynamic, and impact . The sources of AE are fiber-matrix cracking and debonding, fiber fracture, and ply delamination in composite materials. To obtain information about fracture mechanisms of composite materials, it is important to identify the source of emission. If we can identify the source of emission, it is possible to predict their imminent fracture. One of the major issues in the use of AE technique is how to identify the AE signatures due to the different damage mechanisms .
Many researchers have already worked in this field [9–29]. Zhuang and Yan  studied damage mechanisms in self-reinforced polyethylene composites by AE and fabricated model specimens exhibiting a dominant failure mechanism. These authors established correlations between the observed damage growth mechanisms and the AE results in terms of the event amplitude and found that the AE technique was a viable and effective tool for identifying damage mechanisms such as fiber-matrix debonding, matrix cracking, fiber pullout, fiber breakage, and delamination in the ultra-high molecular weight polyethylene/high-density polyethylene composite materials. Woo and Goo  analyzed the bending fracture process for piezoelectric composite actuators (PCA) using dominant frequency bands by AE and examined the fracture process of a thin monolithic PZT and a plate-type PCA subjected to a three-point bending load with the aid of an AE monitoring. These authors found that the damage of a PCA under a bending load was initiated in the brittle PZT layer, which induced the interlaminar delamination between PZT layer and adjacent fiber composite layers. The authors successfully identified the fracture process and mechanisms of PCAs by AE characteristics based on the analysis of dominant frequency bands.
There is a relationship between the level of the frequency content of the AE signals and the damage types on the composites. An average frequency is in fact an apparent frequency equal to the number of counts divided by the duration of the signal . Some researchers used fast Fourier transform (FFT) to obtain the frequency content of the AE signals [5, 11, 15, 17, 19, 20, 24, 29–34]. Usually, matrix cracks occurred in the low-frequency range. Fiber failure has a high-frequency range, and fiber pullout has an intermediate-frequency range. Delaminations have a frequency range between fiber pullout and fiber fracture .
Woo and Choi  studied the fracture process in single-edge-notched (SEN) laminated composites based on high-amplitude AE events and analyzed the fracture process of SEN laminated composites with different lay-up configurations and different fiber composite systems based on the behavior of high-amplitude AE signals. These authors successfully summarized the AE characteristics for SEN laminated composites in association with the individual fracture process. Choi et al.  studied the effects of fiber orientation on the AE and fracture characteristics of composite laminates. Their study showed that AE signals with a high-intensity and high-frequency band were due to fiber fracture, whereas weak AE signals with a low-frequency band were due to matrix cracking.
Sause and Horn  studied the simulation of lamb wave excitation for different elastic properties and AE source geometries and demonstrated that the microscopic elastic properties of the AE source have significant influence on the excitation of distinct lamb wave modes. Gorman and Prosser  studied the AE source orientation by plate wave analysis and investigated the wave propagation behavior in thin composites.
Giordano et al.  studied an AE characterization of the failure modes in polymer composite materials and developed a methodology to prevent source misinterpretation. These authors did a spectral analysis by FFT to evaluate the frequency spectrum content of the signal. They also did tensile tests on single fiber and evaluated the ability of the resonant transducer to discriminate between fiber failure and electrical and mechanical noise. Bohse  studied the AE characteristics of microfailure processes in polymer blends and composites and did a qualitative correlation of the mechanical energy released from fiber/matrix debonding and fiber-fracture processes in single-fiber pullout experiments with the measured AE energy.
Some researchers showed the relationship between the value of the AE count and the source mechanism in the composites [7–9, 16, 17, 20, 21, 38]. Bussiba et al.  studied the fracture characterization of C/C composites under various stress modes by monitoring both mechanical and acoustic responses and showed a good relation between the mechanical properties and acoustic responses. Higher AE counts rate resulted in higher fracture stress. Amanda et al.  studied damage sensing of adhesively bonded hybrid composite/steel joints using carbon nanotubes and found a good relation among load, resistance, and AE response of the specimens.
The present study investigated the fracture process of woven fabric composites by analyzing AE events and cross-sectional observations. Although initial defects have a significant influence on the reliability and the integrity of laminated composites, analyses based on the various strain rates in bending tests relating to the fracture process of various types of laminated composites have been hardly reported in the literature. To give the same strain levels to the bent specimens, we developed a special metal holder. After that, the cross-sectional observations of specimens from three-point bending tests were performed. No study that has used a special device to investigate the cross-sections of specimens after bending tests has been found in the literature.
The present work concentrated on the investigation of the initial and final fracture behaviors of woven composites. Two types of woven composites, such as glass roving and glass cloth, were fabricated. Tensile tests were conducted on specimens to understand the mechanical performance of composites. To understand the initial and final fracture behaviors, three-point bending tests were conducted on composite specimens. During bending tests, the testing machine was periodically stopped. The observation of the initial and final fractures in warp fiber bundles using AE and optical microscopy was performed. The obtained results of AE and cross-sectional observations from three-point bending tests can be used to design new textile preforms during the development of different composite materials.
2 Materials and methods
2.1 Composite constituents
Two kinds of woven fabrics (glass roving and glass cloth; Nitto Boseki Co., Ltd., Japan) with 0.4 wt% acryl silane coupling agent were used as reinforcement materials. Figure 1A and B shows the woven reinforcement fabrics (glass roving and glass cloth). Table 1 shows the parameters of the woven fabrics. Unsaturated polyester resin (RIGORAC 150HRBQNTNW; Showa Highpolymer Co., Ltd., Japan) was used as matrix resin.
2.2 Fabrication method
Composite panels with eight plies glass roving and 14 plies glass cloth woven preforms were fabricated by hand lay-up method. The stacking sequence of 8 and 14 layers is written in a symmetric laminate code, such as [0/90/0/90]s and [0/90/0/90/0/90/0]s. 0° means the direction of warp yarns and 90° means the direction of weft yarns. The composite panels were cured at room temperature for 24 h followed by a 2 h postcure at 100°C. Fiber volume fractions were found out by performing burnout tests. Glass roving 8-layer composites had about 3.5 mm overall panel thickness and 50.3% overall fiber volume fraction. Glass cloth 14-layer composites had about 3 mm overall thickness and 37.6% overall fiber volume fraction.
2.3 Mechanical characterization
Tensile and three-point bending tests were conducted on specimens according to ASTM-D303 and ASTM-D790 standards. The measurements were performed using universal testing machine type 55R4206 (Instron) under displacement control with a speed of 1 mm/min. Figure 2A and B shows the geometry of the specimen from tensile and three-point bending tests. Lamina and aluminum thickness are shown with tC and tAl. The thickness of aluminum tabs was 0.5 mm. The composite coupons have a nominal dimension: (i) 200×20 mm for tensile test and (ii) 90×15 mm for bending test. The test span length was 60 mm in the three-point bending test. A 10.000 N load cell was used for tensile test and 500 N was used for the three-point bending test. The tensile and bending measurements were performed at ambient conditions of 23±2°C and 50±5% relative humidity. Three specimens for the tensile test and 10 specimens for the three-point bending test were tested.
2.4 AE events
AE events were monitored and recorded using an AE instrument system MISTRAS 2001 [Physical Acoustics Corporation (PAC)]. Figure 2C shows the test set-up of specimen for the three-point bending test using AE system. One sensor was placed on the outer surface of the specimen with a layer of silicone grease and secured using tape. The used type of sensor is Micro30, PAC, which had a bandwidth of 100–600 kHz with a dominant sensitivity of 265 kHz.
To improve the test precision, the sensitivity and consistency of the AE system are calibrated using the pencil lead break test before the experiments. After the calibration step, the AE signals were captured during mechanical testing. Signal descriptors (amplitude, duration, rise time, counts, and energy) were calculated using the software of the system (AEwin).
2.5 Metal holder
Figure 3 shows the photograph of a special device (metal holder), which was used to bend specimens from the three-point bending test. The metal holder was prepared at the laboratory of Kyoto Institute of Technology. Using this device, we were able to control the deflection and strain rates of the center of specimens. By turning the bolt screw, the removable pin was moved up and down and the specimens were bent until the target deflection or strain. The target strain rates of specimens (they were the same as the strain value from three-point bending tests) were given on specimens using the metal holder.
2.6 In situ macroscopic observation
In situ macroscopic observations using high-resolution digital video camera for the fracture behavior on specimen surfaces under bending loading were performed to help identify the damage mechanism. To observe the initiation, propagation, and pattern system of the surface cracks, the side surfaces of the bending specimen types were softly painted by a thin layer of a water-based red color. The observation of the cracks on the surface of the specimen was accomplished from the contrast between the white color of the composites and the red color of the side surface painting .
3 Results and discussion
3.1 Microstructural observations of as-fabricated composites
The cross-sectional photographs and schematic drawings of as-fabricated composites are shown in Figure 4A–D (Figure 4A and B for glass roving and Figure 4C and D for glass cloth). The cross-section of the specimens had different geometry and aspect ratio. To find the changes in the cross-sections, quantitative evaluation of the aspect ratio of the warp fiber bundles was performed. The aspect ratio of the cross-sections was calculated by dividing the short diameter into the long diameter (Figure 4B) . The aspect ratio of the glass roving specimen was 0.1, and that for the glass cloth was 0.35.
3.2 Tensile properties of composites
Figure 5A and B shows the tensile test results of the composites with glass roving and glass cloth reinforcements in the warp direction. Figure 5A shows that the tensile stress increases linearly with the increase in the strain and that was followed by a sudden drop in a stress value corresponding to the ultimate failure of the composite. The specimen with glass roving composite showed higher tensile strength and tensile ultimate strain than that with glass cloth composite. The ultimate strain with glass roving composite was about 1.4 times higher than that with glass cloth composite in the warp direction.
Figure 5B shows the tensile modulus and strength results of the composites. The highest tensile modulus and strength were obtained by glass roving composites in the warp direction (29.4 GPa and 435 MPa). The tensile modulus and strength of glass roving composites were 19% and 39% higher than glass cloth composites (23.8 GPa and 266 MPa).
3.3 Bending properties of composites
3.3.1 Results of the three-point bending test
Figure 6A and B shows the stress-strain and AE events of glass roving and glass cloth specimens with maximum strain rate (final fractured) from the three-point bending test in the warp direction. The specimen with glass roving composite showed higher bending strength and bending ultimate strain that that with glass cloth composite. The ultimate bending strain with glass roving composite was about 1.8 times higher than that with glass cloth composite in the warp direction.
The tensile modulus of composites (Figure 5B) was higher than the bending modulus of composites (Figure 6B) in the warp direction. A different failure mechanism could be responsible for this different trend. A measure of the resistance to deformation of the composite in bending is called bending modulus. The bending strength and stiffness are mainly controlled by the strength of reinforcement fibers. The highest bending modulus and strength were obtained by glass roving composites in the warp direction (15.8 GPa and 517 MPa). The bending modulus and strength of glass roving composites was 23% and 51% higher than that of glass cloth composites (12.2 GPa and 251 MPa). The possible reason for the obtained higher tensile and bending properties of the composites with glass roving fabric would be the higher fiber volume fraction (50.3%) than that of the composites with glass cloth fabric (37.6%).
3.3.2 Initial fracture and AE responses
The stress-strain and AE events of glass roving specimens with various strain rates from the three-point bending test in the warp direction are shown in Figure 7A–F. During bending test, the testing machine was periodically stopped. An observation of the initial fracture in the warp fiber bundles using AE was performed. The first AE counts were observed at a bending stress of 164 MPa with 0.010% strain rate (Figure 7A and B). The cumulative AE counts increased with increasing bending stress. The AE counts gradually increased with increasing bending stresses of 336 and 411 MPa with 0.022% and 0.028% strain rates (Figure 7C–F). The peak cumulative AE counts of the first event detected from each AE were 5.632, 31.380, and 54.082 (Figure 7A–F), respectively.
Figure 7G–L shows the stress-strain and AE events of glass cloth specimens with various strain rates from three-point bending tests in the warp direction. The first AE counts were observed at a bending stress of 148 MPa with 0.016% strain rate (Figure 7G and H). The AE counts gradually increased with increasing bending stress of 175 and 226 MPa with 0.019% and 0.030% strain rates (Figure 7I and L). The peak cumulative AE counts of the first event detected from each AE were 26, 98, and 3.196 (Figure 7G–L). The peak cumulative AE count of the events was significantly higher in glass roving composites than that in the glass cloth composites. The reason for this result would be a different failure mechanism and better resistance against the bending load of specimens with glass roving fabric than that of specimens with glass cloth fabric.
Any fracture does not occur until the knee point. The stress and strain at the knee point were defined as initial fracture stress and initial fracture strain. The initial fracture stress and strain for the composites with glass roving fabric were 164 MPa and 0.010% (Figure 7A and B). These were 148 MPa and 0.016% for the glass cloth fabric (Figure 7G and H).
3.3.3 Initial fracture and cross-sectional observations
Figures 8A–D and 9A–D show the cross-sectional photographs and schematic drawings of glass roving specimens with various strain rates during the three-point bending test. To give the same strain rates on specimens as the three-point bending test, the metal holder was used to bend the specimens. An observation of the initial fracture in the warp fiber bundles using an optical microscope was performed. Few transverse cracks appeared in the warp fiber bundles in the tension side of the specimen with 0.010% strain rate (Figures 8A and 9A). The number of the transverse cracks increased and then the cracks enlarged with increasing strain rates (0.022% and 0.028%; Figures 8B and C and 9B and C). The delamination and some of the additional transverse cracks appeared in the specimen with 0.034% strain in the compression side (Figures 8D and 9D).
The cross-sectional photographs and schematic drawings of glass cloth specimens with various strain rates during the three-point bending test are shown in Figures 8E–G and 9E–G. The delaminations appeared in the compression side of the specimen with 0.016% and 0.019% strain rates (Figures 8E and F and 9E and F). The delaminations enlarged in the compression side and then the transverse cracks appeared with increasing strain rate (0.030%) in the tension side of the specimen (Figures 8G and 9G).
3.3.4 Final fracture and AE responses
In situ macroscopic observations using high-resolution digital video camera for the fracture behavior on glass roving specimen surfaces under bending loading in the warp direction are shown in Figure 10A. There was no crack at the stress 420 MPa at first stage. Then, the stress increased and decreased. The matrix and fiber cracks and delamination appeared at 391 MPa in the tension side at the second stage. Then, the fiber cracks occurred at 303 MPa in the compression side at the third stage. The final fracture occurred by the fiber fracture accompanied with delaminations at the stress 59 MPa in the tension side at the final stage.
The stress-strain and AE events of the fractured glass roving specimens from the three-point bending test in the warp direction are shown in Figure 10B–D. The stress-strain and AE counts clearly described the bending fracture process during the test. We divided the bending fracture process of glass roving specimens into four stages according to the stress (MPa)-strain (%) curves and distribution of AE events. The linearity in the curves of AE counts could be seen in stage I. There were no detectable AE events, indicating that no damage occurred in glass roving composites. A change in the slope of the curves of AE counts and the appearance of detectable AE counts were seen in stage II. As explained before, the initial fractures of glass roving composites were detected in stage II, which had the strain levels from 0.010% to 0.034%. The AE events were also gradually increased in this stage. These events were related with the transverse cracks, which occurred in the warp fiber bundle of the composites. In the end of this stage, matrix and fiber cracks and delamination appeared at 391 MPa. A sharp increase in AE activity and a corresponding sharp decrease in the stress were evident, indicating the occurrence of critical damages, such as fiber breakages accompanied with delaminations in stage III. In stage IV, AE events gradually continued to increase. The peak cumulative AE count of the events was 120.508. There were some additional delaminations and fiber fractures that occurred in stage IV compared with stage III.
Figure 10E shows the in situ macroscopic observations using high-resolution digital video camera for the fracture behavior on glass cloth specimen surfaces under bending loading in the warp direction. There was no crack at the stress 220 MPa at the first stage. Then, the stress decreased. The first matrix crack and delamination occurred at stress 225 MPa in the compression side at second stage. We observed fiber fractures at stress 54 MPa in the tension side at the final stage.
The stress-strain and AE events of the fractured glass cloth specimens from the three-point bending test in the warp direction are shown in Figure 10F–H. We divided the bending fracture process of glass cloth specimens into three stages according to the stress (MPa)-strain (%) curves and distribution of AE events. There was linearity in the curves of AE counts in stage I and there were no detectable AE events, indicating that no damage occurred in glass cloth composites. A change in the slope of the curves of AE counts and the appearance of detectable AE counts were seen in stage II. The initial fractures of the glass cloth composites were detected in stage II. The strain levels of these fractures were from 0.016% to 0.030%. The AE events gradually increased in this stage. The highest cumulative AE count of the events was 16.034. These events were related with the matrix crack and delamination, which was explained before. AE events continued to increase in stage III. The peak cumulative AE count of the events was 26.494 (0.036% strain level). At the end of this stage, we observed fiber fractures at stress 54 MPa, which were the reason for the AE events.
A correspondence between the dominant modes of damage and strain rates is shown in Table 2. All these damage mechanism analyses with AE counts as well as cross-sectional photographs allow us to identify the main damage mechanisms of woven reinforcement composites.
3.3.5 Final fracture and cross-sectional observations
The cross-sectional photographs and schematic drawings of specimens with final fracture during the three-point bending test are shown in Figure 11A–D. The cross-section of these composites was detected under an optical microscope. Delaminations, transverse fractures, and fiber breakages were seen in the cross-section of the composites. The composites with glass roving fabric (Figure 11A and C) had higher transverse cracks compared with glass cloth composites (Figure 11B and D).
When we compared the final stage of damages from Figure 11A–D and Figure 10A and E, we can see that both fracture patterns in two different figures of glass roving and glass cloth (Figure 10A-final stage and Figure 11A and Figure 10E-final stage and Figure 11B) agreed well.
3.3.6 Observation of textile structure
The aspect ratio of glass roving composites (0.1) was smaller than that of glass cloth composites (0.35), which was explained in Section 3.1. The stress concentration in the smaller aspect ratio of fiber bundle was smaller than that in the larger aspect ratio . The stress concentration at the center of warp fiber bundle in glass roving composites decreased, and higher stress was needed at the onset of the transverse cracks so that higher knee point stress is obtained. The knee point stress was 164 MPa in glass roving composites (Figure 7A), and that for the glass cloth composites was 148 MPa (Figure 7D).
3.3.7 Relationship between cumulative AE counts and initial energy
The relationship between cumulative AE counts and initial energy from the three-point bending test is shown in Figure 12. The area under load-displacement curves gives the absorbed energy during the three-point bending test. Initiation energy was found to calculate the area under load-displacement curve until maximum load. This graphic showed that the initial energy from the three-point bending test increased with increasing cumulative AE counts from the three-point bending test. The results of cumulative AE counts and initial energies from glass roving composites were higher than that from glass cloth composites.
In this study, the initial and final fracture behaviors in glass roving and glass cloth composites were investigated. The initial fracture in glass roving composites was found as transverse cracks in the warp fiber bundle in the tension side, and that for glass cloth composite delaminations was found in the compression side. The final fractures were transverse cracks, delaminations, and fiber breakages in glass roving and glass cloth composites.
During the study, a new metal holder that was used to investigate the cross-sectional observations of specimens after the three-point bending test was developed. The cross-sectional observations of the fractured specimens after the three-point bending test showed that composites with glass roving fabric had higher transverse cracks compared with glass cloth composites. The stress concentration at the center of warp fiber bundle in glass roving composites decreased, and higher stress was needed at the onset of the transverse cracks so that higher knee point stress is obtained. The results of cumulative AE counts and initial energies from glass roving composites were higher than that from glass cloth composites during the three-point bending test. Our study showed that the AE monitoring and cross-sectional observation were applicable for identifying the fracture mechanisms, such as initial and final fractures, in glass roving and glass cloth composites.
The author thanks Prof. Hiroyuki Hamada and Dr. Eng. Mohamed S. Aly-Hassan (Kyoto Institute of Technology) for their help and support.
Osada T, Mizoguchi M, Kotaki M, Nakai A, Hamada H. 13th International Conference on Composite Materials (ICCM 13), ID 1083, Beijing, China, 2001.Google Scholar
Osada T, Nakai A, Hamada H. 4th Joint Canada-Japan Workshop on Composites, CRC Press: Florida, USA, 2002, p. 205–212.Google Scholar
Khondker OA, Fukui T, Nakai A, Hamada H. Compos. Part A 2004, 35, 1185–1194.Google Scholar
Kadoma Y, Odawara A, Okano M, Nakai A, Hamada H. 17th International Conference on Composite Materials (ICCM 17), 2009.Google Scholar
Osada T, Nakai A, Hamada H. Compos. Struct. 2003, 61, 333–339.Google Scholar
Dr. Boris Muravin. Acoustic Emission Wave Propagation and Source Location, www.muravin.com.
Woo SC, Goo NS. Compos. Sci. Technol. 2007, 67, 1499–1508.Google Scholar
Zhuang X, Yan X. Compos. Sci. Technol. 2006, 66, 444–449.Google Scholar
Majid MSA, Gibson AG, Assaleh TA, Rookus CAP, Hekman M, Afendy M, Haftirman AS, Yayuthi AR. 8th Asian-Australasian Conference on Composite Materials (ACCM-8), ID: O-TUB-171, Kuala Lumpur, Malaysia, 2012.Google Scholar
Groot PJD, Wijnen PAM, Janssen RBF. Compos. Sci. Technol. 1995, 55, 405–412.Google Scholar
Jimenez CRR, Papadakis N, Reynolds N, Gan TH, Purnell P, Pharaoh M. Compos. Sci. Technol. 2004, 64, 1819–1827.Google Scholar
Benmedakhene S, Kenane M, Benzeggagh ML. Compos. Sci. Technol. 1999, 59, 201–208.Google Scholar
Rosa IMD, Santulli C, Sarasini F. Compos. Part A 2009, 40, 1456–1469.Google Scholar
Woo SC, Choi NS, Cho N. Compos. Sci. Technol. 2008, 68, 1521–1530.Google Scholar
Godin N, Huguet S, Gaertner R, Salmon L. NDT&E Int. 2004, 37, 253–264.Google Scholar
Giordano M, Calabro A, Esposito C, D’Amore A, Nicolais L. Compos. Sci. Technol. 1998, 58, 1923–1928.Google Scholar
Bussiba A, Kupiec M, Ifergane S, Piat R, Böhlke T. Compos. Sci. Technol. 2008, 68, 1144–1155.Google Scholar
Huguet S, Godin N, Gaertner R, Salmon L, Villard D. Compos. Sci. Technol. 2002, 62, 1433–1444.Google Scholar
Scholey J, Wilcox P, Wisnom M, Friswell M. Compos. Part A 2010, 41, 612–623.Google Scholar
Gu J, Yoon H, Choi N. Compos. Part A 2012, 43, 2211–2220.Google Scholar
Moevus M, Godin N, R’Mili M, Rouby D, Reynaud P, Fantozzi G, Farizy G. Compos. Sci. Technol. 2008, 68, 1258–1265.Google Scholar
Moevus M, Rouby D, Godin N, R’Mili M, Reynaud P, Fantozzi G, Farizy G. Compos. Sci. Technol. 2008, 68, 1250–1257.Google Scholar
Czigany T. Compos. Sci. Technol. 2006, 66, 3210–3220.Google Scholar
Kostopoulos V, Tsotra P, Karapappas P, Tsantzalis S, Vavouliotis A, Loutas TH, Paipetis A, Friedrich K, Tanimoto T. Compos. Sci. Technol. 2007, 67, 822–828.Google Scholar
Loutas TH, Kostopoulos V. Compos. Sci. Technol. 2009, 69, 265–272.Google Scholar
Yu YH, Choi JH, Kweon JH, Kim DH. Compos. Struct. 2006, 75, 163–169.Google Scholar
Loutas TH, Kostopoulos V, Ramirez-Jimenez C, Pharaoh M. Compos. Sci. Technol. 2006, 66, 1366–1375.Google Scholar
Woo SC, Choi NS. Compos. Sci. Technol. 2007, 67, 1451–1458.Google Scholar
Sause MGR, Müller T, Horoschenkoff A, Horn S. Compos. Sci. Technol. 2012, 72, 167–174.Google Scholar
Sause MGR, Horn S. J. Acous. Emis. 2010, 28, 142–154.Google Scholar
Gorman MR, Prosser WH. J. Acous. Emis. 1990, 9, 283–288.Google Scholar
Bohse J. Compos. Sci. Technol. 2000, 60, 1213–1226.Google Scholar
Amanda SL, Zachary RM, Erik TT, Chou TW. Compos. Sci. Technol. 2011, 71, 1183–1189.Google Scholar
Aly-Hassan MS, Takai Y, Nakai A, Hamada H, Shinyama Y, Hatsukade Y, Tanaka S. 18th International Conference on Composite Materials (ICCM 18), 2011.Google Scholar
About the article
Published Online: 2014-11-11
Published in Print: 2016-03-01