Change in B-value by ae propagation length in CFRP.
In this study, changes in b-values with several fracture modes (matrix cracking, de-bonding and fiber breakage) of FRPs were investigated. When AE sources contain single fracture mode, b-value does not change by propagation length. On the other hand, when AE sources contain several fracture modes, b-value changes with propagation length. The change in b-value for multiple fracture modes may occur by change in the proportion of each frequency components after the propagation due to frequency dependence of the attenuation.
Keywords: b-value, Carbon Fiber Reinforced Plastic (FRP), Frequency dependence of Attenuation.
Acoustic Emission Testing (AT) is one of the non-destructive testing methods which detect damages in mechanical structures by using AE signals caused from damages. Severity of damages can be estimated by evaluating AE parameters of detected AE signals. Amplitude distributions of AE signals are characterized by b-value and the value is also used for the severity evaluation. Shiotani et. al.  and Sammonds et.al. showed that the b-value for evaluating damages of ground slopes and rocks. b-value is now widely used in the fields of civil engineering. On the other hands, Carbon fiber reinforced plastic (CFRP) is a hybrid material that contains carbon fiber as reinforcements and epoxy resin as matrixes, and is widely used in the chemical and mechanical industries. It is known that FRP (fiber reinforced plastic) have several fracture modes (Matrix cracking, De-bonding and Fiber breakage) and each mode cause AE with different frequency range (see Figure 1). It is also known that attenuation of AE changes with the frequency of AE. In this study effects of propagation length and frequency of AE on b-value are discussed.
[FIGURE 1 OMITTED]
2. b-value analysis
The b-value is usually calculated using the cumulative frequency-magnitude distribution data and applying the Gutenberg-Richter relationship, which is widely used in earthquake seismology . The relationship can be expressed by a following empirical equation (2)
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)
[log.sub.10]N(M) = a-bM (2)
Where M = earthquake magnitude, n(M) = number of the earthquake of magnitude M, N(M) = total number greater than magnitude M, a and b = empirical constants. As shown in equation (2), the magnitude is proportional to the logarithm of the maximum amplitude [log.sub.10]N(M). Magnitudes of earthquakes are changed with distance from the source to sensor due to attenuation, although, b-value is constant when seismic wave contain single frequency or attenuation has no frequency dependence. From the equation (2), the b-value is the negative gradient of the log-linear AE hits-magnitude plot and therefore it represents the slope of the amplitude distribution (see Figure 2). In terms of AE technique, the Gutenberg-Ritcher formula can be modified as following equation.
[log.sub.10]N = [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)
Both the frequency and amplitude distribution of AE signals generated from source are changed by characteristics of fractures. For example, micro-cracks generate a large number of small AE. On the other hand, macro-cracks generate small number of large AE. Though, micro-cracks lead b-value for high value and macro-cracks for relatively low value. When damage level of structures reaches close to final fracture, fracture behaviour is changed from micro-crack to macro-crack and then b-value becomes small. This is the reason why b-value is used to evaluate damage severity of structures.
[FIGURE 2 OMITTED]
3. AE testing of CFRP plates with Center Holl
The specimens used in the experiments are CFRP laminates made from unidirectional prepreg sheet (Mitsubishi Rayon Japan, PYROFIL#380). The stacking sequences are [.sub.8] , [.sub.16], [.sub.16] and cross-ply[[0/90].sub8]. Plates of 250 mm length and 150 mm width were made using an autoclave. The pre-cure condition was 85[degrees]C x 2 hours, and the main curing condition was 135[degrees]C x 3 hours under 0.7 MPa pressure. From these CFRP plates, rectangular plate specimens of 200 mm length as shown in Figure 3 and table 1 were prepared. The GFRP-tabs with 2mm in thickness were attached to the specimens of both ends for the tensile test to avoid causing any damage by a jig. Three AE sensors (Physical Acoustics, Type: PICO) were mounted on the specimen with different distance from the holes (S1-S3 in Fig. 3). The cross head speed of tensile tester was controlled as 0.1mm/min. Due to the stress concentration, the specimens broke at near the center hole (See lower of Fig. 3). Fracture strain and stress are shown in Table 1. Matrix cracking was dominant for the 9[.sub.16] specimen, both the matrix cracking and de-bonding were observed for [.sub.16] specimen. All fracture modes shown in Fig. 1 were observed for [.sub.16] and [[0/90].sub.8] specimens. During the test, AE signals larger than 35dB were detected.
[FIGURE 3 OMITTED]
4. b-value analysis for CFRP
Figure 4 shows b-value calculated by using detected AE signals by the s1, s2 and s3 sensors in Fig. 3 (upper) for cross-ply [[0/90].sub.8] specimens. Calculated b-values from detected AE by three AE sensors are fluctuated during the test in some range, although, the average of b-values are increased with propagation length. Figure 5 shows centroid frequency of detected AE signals.
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
AE with 50-650 kHz are dominant for s1 (propagation length 10mm), although dominant frequencies for s3 (propagation length 30mm) are changed as 50-550 kHz. Change in dominant frequency with propagation length may change the b-value. Then, we conducted detailed discussion about the cause of b-value change.
Figure 6 (a, b, c) shows histogram of centroid frequencies of AE for [.sub.8], [.sub.16] and [.sub.8] detected by the S1 sensor respectively. Numbers of hits with logarithm scale are overwrapped in the graph with dotted line. AE in Fig. 6 (a) is mainly caused by matrix cracking. AE in Fig. 6(b) is caused by matrix cracking and de-bonding. On the other hand, AE in Fig. 6(c) is caused by all fracture modes described in Fig. 1. By comparing each results, dominant frequency for matrix cracking, de-bonding and fiber breakage were confirmed as100~250 kHz, 450~540 kHz and 550~660 kHz respectively.
As we confirmed dominant frequencies for each fracture modes, AE from matrix cracking, de-bonding and fiber breakage caused for cross-ply [0/90] 8 were separated by band-pass frequency filters of 100~250kHz, 450~550kHz and 550kHz~650kHz. Fig. 7 (a,b,c) are results of separation for AEs detected for cross-ply [[0/90].sub.8]. The symbols "[DELTA]", "O" and "X" indicate AEs that detected at different propagation length of 10, 20 and 30 mm. As show in the figure, slope of the each line (negative of b-value) dose not change by propagation length. It is found that b-value is not changed by propagation legth when detected AE signals are caused by certain fracture mode (When frequency range of detected AE signals is narrow) as not like Fig. 4. In order to discuss about the cause of changing b-value by propagation length in Fig. 4, the cumulative amplitude distributions of each frequency component according to fracture modes were investigated.
[FIGURE 6 OMITTED]
[FIGURE 7 OMITTED]
[FIGURE 8 OMITTED]
Figure 8 shows cumulative amplitude distribution for each fracture modes ([??]:matrix cracking, [DELTA] :de-bonding, X : fiber breakage) with different AE propagation length ((a):10 mm, (b):20mm, (c):30mm). AE distribution for total of each fracture modes is also shown in the graph (symbol [OMICRON]). Slope of the distribution of all AE sources ([OMICRON]) which propagation length 10mm (Fig. (a)) is composed with all frequency components of 100~250kHz, 450~550kHz and 550kHz~650kHz. The negative slope of the line (b-value) becomes 0.06 and the value is different from that of each frequency sorce (100-250kHz: 0.05, 450-550kHz: 0.07 and 550-650kHz: 0.049). On the other hands, due to the frequency dependence of attenuation (high attenuation factor for high frequency components) , number of high frequency AE due to fiber breakage (550-650kHz) is drastically reduced for the results of 20 and 30 mm propagation lentgh (Fig. (b) and (c)). Though, slope of the distribution of all AE sources ([OMICRON]) which propagation length 20 and 30 mm are almost decided by 100-250 kHz component (matrix cracking) and 450-550 kHz (de-bonding) components only. From these result, the change in b-value may be caused by change in the proportion of each frequency components after the propagation due to frequency dependence of the AE attenuation.
5. Failure mechanism of CFRP estimated by b-value analysis
As propagation length affects b-value for CFRP, we focused on certain propagation length of 10mm and discussed about failure mechanism of CFRP based on b-value analysis. Figure 9 is time-series data of b-value of total AE (O: 50kHz~650kHz) during tensile test (Extracted symbol  data from Fig. 4). Change in b-value for each frequency components (each fracture modes) are overwrapped in the graph (matrix cracking ([??]: 100~250kHz), de-bonding (?: 450~550kHz), fiber breakage (X: 550kHz~650kHz)). Dotted points indicate the stress during the test. Both matrix cracking and de-bonding are dominant fracture modes until 270sec., although, fiber breakage is added after 270sec.. When damage level of CFRP specimens reaches close to final fracture, fracture behaviour is changed from matrix cracking and de-bonding to fiber breakage and then each b-value changes drastically (matrix cracking: 0.003, de-bonding: 0.058, fiber breakage: 0.041 and all AE source: 0.048).
[FIGURE 9 OMITTED]
In this study, changes in b-values with fractures of FRPs were investigated. We conducted the tensile test with AE analysis for [.sub.8], [.sub.16], [.sub.16] and cross-ply[[0/90].sub.8] tensile specimens. Dominant fracture modes (matrix cracking, de-bonding and fiber breakage) are changed with stacking sequence of CFRP. The AE signals due to each fracture modes can be separated by using frequency filter of 100~250kHz (matrix cracking), 450~550kHz (de-bonding) and 550kHz~650kHz (fiber breakage). When AE sources contain single fracture mode, b-value does not chang by propagation length. On the other hand, when AE sources contain several fracture modes (AEs with different frequency range), b-value changes with propagation length. The change in b-value for multiple fracture modes may occur by change in the proportion of each frequency components after the propagation due to frequency dependence of the attenuation. Additionally, it was found that when damage level of cross-ply[[0/90].sub.8] specimens reaches close to final fracture, fracture behaviour is changed from matrix cracking and de-bonding to fiber breakage and then b-value for total AE is changed.
This work was supported by JSPS KAKENHI Grant Number 16J12078.
[1.] Shiotani, T., Fujii, K., Aoki, T., and Amou, K., "Evaluation of progressive failure using AE sourced and improved b-value on slope model tests.", Prog. Acoustic Emission. VII, 7, pp.529-534, 1994.
[2.] Colombo, Ing S., I. G. Main, and M. C. Forde. "Assessing damage of reinforced concrete beam using "b-value" analysis of acoustic emission signals." Journal of materials in civil engineering 15.3, pp 280-286, 2003.
[3.] Gutenberg, Beno, and Charles Francis Richter. "Magnitude and energy of earthquakes." Annals of Geophysics 9.1, 1-15, 1956.
Doyun JUNG (1), Yoshihiro MIZUTANI (1), Akira TDOROKI (1), Yoshiro SUZUKI (1)
(1) Tokyo Institute of Technology; 2-12-1, Ookayama, Meguro-ku, Tokyo, Japan
Phone: +81 3 5734 3178, Fax: +81 3 5734 3178; e-mail: firstname.lastname@example.org,
email@example.com, firstname.lastname@example.org, email@example.com
Table 1 Specifications of test specimen Test Layer Width Length C [l.sub.0] (mm) (mm) (mm) (mm) # 1 [.sub.16] 16 20 200 4 100 # 2 [.sub.16] 16 20 200 4 100 # 3 [.sub.8] 8 16 200 6 100 # 4 [[0/90].sub.8] 8 16 200 6 100 Test [sigma.sub.failure] [epsilon.sub.failure] (MPa) (%) # 1 [.sub.16] 11 0.35 # 2 [.sub.16] 18 0.58 # 3 [.sub.8] 805 4.1 # 4 [[0/90].sub.8] 840 4
|Printer friendly Cite/link Email Feedback|
|Author:||Jung, Doyun; Mizutani, Yoshihiro; Tdoroki, Akira; Suzuki, Yoshiro|
|Publication:||Journal of Acoustic Emission|
|Date:||Jan 1, 2016|
|Previous Article:||High sensitivity detection of ae events in noisy environment using stream recording and parallel computation.|
|Next Article:||Two-dimensional Q-value AE-Tomography and its verification on numerical invesitigations.|