Ultrasonics 61 (2015) 52–61
Contents lists available at ScienceDirect
Ultrasonics journal homepage: www.elsevier.com/locate/ultras
Non-destructive evaluation of anchorage zones by ultrasonics techniques M. Kharrat ⇑, L. Gaillet Institut Français des Sciences et Technologies de Transport, de l’Aménagement et des Réseaux, Département Matériaux et Structures, Laboratoire Structures Métalliques et à Câbles, Route de Bouaye, CS4, 44344 Bouguenais Cedex, France
a r t i c l e
i n f o
Article history: Received 15 June 2014 Received in revised form 9 March 2015 Accepted 12 March 2015 Available online 23 March 2015 Keywords: NDT Anchorages Cables Acousto-Ultrasonics Ultrasonics
a b s t r a c t This work aims to evaluate the efficiency and reliability of two Non-Destructive Testing (NDT) methods for damage assessment in bridges’ anchorages. The Acousto-Ultrasonic (AU) technique is compared to classical Ultrasonic Testing (UT) in terms of defect detection and structural health classification. The AU technique is firstly used on single seven-wire strands damaged by artificial defects. The effect of growing defects on the waves traveling through the strands is evaluated. Thereafter, three specimens of anchorages with unknown defects are inspected by the AU and UT techniques. Damage assessment results from both techniques are then compared. The structural health conditions of the specimens can be then classified by a damage severity criterion. Finally, a damaged anchorage socket with mastered defects is controlled by the same techniques. The UT allows the detection and localization of damaged wires. The AU technique is used to bring out the effect of defects on acoustic features by comparing a healthy and damaged anchorage sockets. It is concluded that the UT method is suitable for local and crack-like defects, whereas the AU technique enables the assessment of the global structural health of the anchorage zones. Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction Upon commissioning, civil engineering infrastructures, like bridges, are subjected to multiple aggressive conditions. Generally, the older the bridge is, the higher the risk of disorder. Most degradation mechanisms are currently known and have been reported in several works [1–5]. Corrosion and fatigue are the main causes of damage and have often a harmful combination on the health condition of bridges. In the most critical cases, the damaged bridge’s components must be replaced. Among the major bridge’s components, cables and anchorage zones could be subjected to hard solicitations. Cables are primarily undergoing environmental aggressions. Run-off water added with deicing salts could penetrate in cables and gravitationally flow toward tightened areas (ties of suspension lines) or toward lower zones (such as anchorages). This confined humidity often engenders corrosion. Non-alloy steel used in cables is sensitive to corrosion by dissolution, so that a section loss more or less homogeneous can affect all or part of the wires constituting cables and can decrease their breaking strength [6,7]. Moreover, fretting⇑ Corresponding author at: Department of Applied Mechanics, FEMTO-ST Institute, Franche-Comté University, 25000 Besançon, France. Tel.: +33 682212604. E-mail address: [email protected] (M. Kharrat). http://dx.doi.org/10.1016/j.ultras.2015.03.007 0041-624X/Ó 2015 Elsevier B.V. All rights reserved.
fatigue mechanism produces small relative displacements between the wires of a single cable (inter-wires contact), between neighboring cables, or also between a cable and other fixed parts like anchorages. This can cause wear and/or cracking of the wires [8,9]. As the anchorage zone is an inaccessible part of the bridge and is located at the bottom of the retaining cables, it provides an ideal place for water accumulation and promotes corrosion development. The major problem here is the rupture of wires that cannot be anticipated by visual inspections. This bridge’s component requires the implementation of operational inspection techniques able to assess its structural health and prematurely detect damages before the failure of the whole structure [10,11]. Regarding Non-Destructive Testing (NDT) of civil engineering structures, an important distinction is commonly made between accessible and inaccessible parts. Accessible parts relate to components where the element to be inspected is directly visible and easily accessible; while inaccessible parts involve hidden components of the structure, so that require a prior intervention before inspections. If the current state-of-the-art of monitoring techniques of accessible parts is based on visual inspection [12], magnetic inspection [13], and acoustic monitoring [14,15], that concerning inaccessible parts is still missing. The inspection and monitoring of inaccessible areas are of major concern as they are often considered as the weakest parts where fatigue and corrosion
53
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
mechanisms can appear. However, many structural factors can greatly complicate their inspection. Key factors mainly include the influence of different material layers and multiple interfaces in protection systems of cables embedded in anchorages and suspension collars. Furthermore, various techniques have been studied among which ultrasonic guided waves [16,17], magnetic flux leakage inspection [18], micro-magnetic methods, and acoustic emission monitoring [19] seem to be suitable. The AcoustoUltrasonic (AU) is a relatively new technique in civil engineering. It has been initially used for inspection and characterization of graphite-reinforced plastics (GRP) [20–22]. Acousto-Ultrasonic parameters (AUPs) have shown a good correlation with the mechanical strength for both tension and compression test configurations [23] and even for detecting corrosion evolution in coatings of aircraft wings [24]. This technique has been also used for non-destructive quantitative characterization of residual impact strength of polyvinyl chloride (PVC) using the AUPs [25]. Detection of pre-machined defects on a metal plate has been studied using a broadband excitation through temporal, frequency and wavelet analyzes [26]. Correlation between the amount of artificial defects in bonded regions and AUPs in adhesively bonded joints of carbon-fiber reinforced plastic (CFRP) laminates and aluminum
Transmitter E1 Receiver R
Defect
Strand
265 mm
Transmitter E2
36 mm 695 mm
(a) Transmitter E1
5
6
1
2
3
4
Transmitter E2
(b) Fig. 3. Acouto-ultrasonic test performed on a single strand using two transmitters E1 and E2 and a receiver R. (a) Lateral view; (b) sectional view.
3000
E1R−1V3C E1R−3V3C E1R−5V3C E1R−7V3C E1R−10V3C
2500
Energy
2000 1500 1000 500 0
0
1
2
3
4
5
6
Number of cut wires
(a) 2500
Fig. 1. UT instruments: Krautkramer’s USM25 pulser/receiver device (on the left); and SMWB70-6 probe (on the right).
E2R−1V3C E2R−3V3C E2R−5V3C E2R−7V3C E2R−10V3C
Controller AE chain
Energy
2000 1500 1000 500
Digitizer
0
Pass band filter
Signal generator
0
1
2
3
4
5
6
Number of cut wires
(b) Amplifier
Fig. 4. Evolution of the average energy depending on the defect expansion while exciting by different waveforms. (a) Waves emitted by E1; (b) waves emitted by E2.
Pre-amplifier Transmitter
Receiver
Specimen Fig. 2. Acousto-Ultrasonics principle.
plates has been also investigated [27]. However, this technique has been rarely used for damage characterization in metallic cables. This work aims to test the reliability of two NDT methods for damage assessment in some bridge’s parts. Specifically, the Acousto-Ultrasonic technique is compared to classical Ultrasonic Testing (UT) technique in terms of defect detection and structural health classification.
54
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
100
Amplitude (dB)
95 90 85 80
E1R−1V3C E1R−3V3C E1R−5V3C E1R−7V3C E1R−10V3C
75 70
Fig. 7. Sectional view of a stranded cable having 8 layers of wires surrounding a core wire.
Cable
65
0
1
2
3
4
5
Transmitter Receiver
6
Number of cut wires
(a)
Transmitter positions
Anchorage
100 Fig. 8. Principle of the AU test performed on anchorages from ‘‘Le Teil’’ Bridge.
Amplitude (dB)
95 90 85
E2R−1V3C E2R−3V3C E2R−5V3C E2R−7V3C E2R−10V3C
80 75 70
0
1
2
Fig. 9. Photos of the inspected anchorages with the transducers during the AU test: the transmitter on the left and the receiver on the right.
3
4
5
6
Number of cut wires
(b) Fig. 5. Evolution of the average amplitude depending on the defect expansion while exciting by different waveforms. (a) Waves emitted by E1; (b) waves emitted by E2.
2. NDT techniques The UT technique was used in reflection mode to assess the presence of defects in anchorage zones. A Krautkramer’s USM25 ultrasonic pulser/receiver device and a SMWB70-6 probe, characterized by an angle of incidence of 70° in transmission and an operating frequency of 6 MHz, were employed (Fig. 1). A SLC 70 gel-type couplant was applied between the transducer and the test specimen to ensure a proper transmission of signals. The AU technique combines aspects of Acoustic Emission (AE) signal analysis with Ultrasonics assessment methods. The transmitter transducer generates a specific ultrasonic waveform that propagates through the tested specimen before being collected by a receiver transducer. Signals resulting from multiple reflections
1
2
Fig. 6. Potentially damaged anchorages collected from ‘‘Le Teil’’ bridge.
3
and interactions with the microstructure of the material are treated in the same way as AE signals. The principle of the AU technique is illustrated in Fig. 2. In our study, an ARB-1410 Arbitrary Waveform Generator Board (Mistras Group) was used to generate the ultrasonic signals. The waveforms were 3-cycle windowed sine functions at 150 kHz. Different amplitudes were used: 1, 3, 5, 7 and 10 V. These signals were sent repeatedly at a rate of 1 Hz over a recording duration of about 1 min. The acquisition system included a PCI-2 card having two input channels with a sampling rate of 40 MHz. The transducers employed for both emission and reception were of R15atype characterized by a resonance frequency of 150 kHz and an operating bandwidth of 50–400 kHz. Transducers’ bonding was performed by using a ‘‘HBM glue for gauges’’.
3. Experimental tests and results’ discussions This section presents the experimental tests performed on different types of specimens. Firstly, the AU technique was used on single strands where some defects were progressively created. The effect of damage growth on the collected signals was evaluated. Thereafter, three anchorages with unknown defects were controlled by AU and UT techniques. The damage assessment abilities of both techniques were compared The test specimens were also classified by a damage severity criterion. Finally, healthy and damaged test specimens with mastered defects were inspected by the same techniques. Each specimen consisted in a long cable anchored in an anchorage socket. A comparison of the efficiency of both techniques was addressed.
55
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
Specimen 2 Specimen 1 330 Specimen 3
Specimen 2 Specimen 1 330 Specimen 3
0 30
300
60
100
120 210
Specimen 2 Specimen 1 330 Specimen 3
60
300 90
240
30
300
200
270
0
600 800 200 400 90
270
240
150
120 210
150
180
180
(a)
(b)
Specimen 2 Specimen 1 330 Specimen 3
0 30
300
60 300 400 100 200 90
270
240
120 210
150
0 30
300
60 800 1000 400 600 200 90
270
240
120 210
150
180
180
(c)
(d)
Fig. 10. Comparison between average energies of the signals transmitted through the three anchorages. Excitation signals emitted from four circumferential positions around the cable at: (a) 10 V, 3 cycles; (b) 25 V, 3 cycles; (c) 10 V, 6 cycles; and (d) 25 V, 6 cycles.
3.1. Acousto-Ultrasonic tests on single strands The first tests were performed on single strands by the AU technique in order to study the influence of progressing defects on selected AU parameters. The tested T15.7-type strand was 1.39m-long and composed of seven wires, each having a diameter of 5.4 mm. As shown in Fig. 3(a), two transmitters (E1 and E2) were coaxially positioned at the middle of the strand; a receiver sensor (R) was mounted at one end. A growing defect was artificially created by cutting out progressively 36-mm-long segments from the wires. The order of cutting the wires’ segments is schematized in Fig. 3(b). The test was started by cutting segments from wires 1 and 2, whereas E1 was in contact with intact wires 5 and 6. Successively, after each cutting, a wave was firstly sent by the E1 transmitter, and then by E2. The transmitted wave was collected by the R sensor in each case. Energy and Amplitude (averaged values over 1 min of recording) are selected as AU parameters in order to assess the damage expansion during this test. The evolution of the average energy and amplitude depending on the defect expansion is illustrated in Figs. 4 and 5. Fig. 4(a) shows an increase in energy depending on the number of the cut wires until the fourth one. Then, energy decreases with the fifth and sixth cut-wires on which E1 was mounted. Fig. 4(b)
shows a global decrease in energy with the number of cut wires. Indeed, for the two transmitters, the incident wave is substantially guided by the wires on which the transmitters E1 and E2 are mounted. A part of the incident energy is also guided by the other wires. When a transmitter is mounted on the cut wires (case of E2), the decrease in the energy transmitted to the R sensor is due to the defect, which reflects most of the incident wave. The other wires that are still intact would guide the rest of the wave energy. Thus, larger the damage is, higher the wave is reflected. When a transmitter is located on the uncut wires (case of E1), energy received by the R sensor increases until reaching the wires on which E1 is mounted. However, we can note that amplitudes of the received signals remain almost constant for the first four wires (see Fig. 5), although energy is evolving. This could be produced when duration and/or counts of an AU signal evolve. To explain the phenomenon occurred in Fig. 4(b), we can mention that the incident wave, which is guided by undamaged wires, propagates to the R sensor, whilst the presence of a defect creates mode conversions and wave interactions [28,29]. A part of these waves could propagate in turn toward the R sensor, and thereafter could be superimposed with the main incident wave. Hence, this could increase duration and/or counts of the signal, leading to an increase in energy.
56
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
Specimen 2 Specimen 1 330 Specimen 3
Specimen 2 Specimen 1 330 Specimen 3
0 30
300
60
240
Specimen 2 Specimen 1 330 Specimen 3
60 60 80 20 40 90
270
120 210
30
300
60 80 20 40 90
270
0
240
150
120 210
150
180
180
(a)
(b) Specimen 2 Specimen 1 330 Specimen 3
0 30
300
60
240
60 80 100 20 40 60 90
270
120 210
30
300
60 80 20 40 90
270
0
240
150
120 210
150
180
180
(c)
(d)
Fig. 11. Comparison between average amplitudes of the signals transmitted through the three anchorages. Excitation signals emitted from four circumferential positions around the cable at: (a) 10 V, 3 cycles; (b) 25 V, 3 cycles; (c) 10 V, 6 cycles; and (d) 25 V, 6 cycles.
Table 1 Anomalies detected by UT in the anchorage N°1.
Fig. 12. Preparation of the wires’ surfaces.
3.2. Anchorages with unknown defects Real bridge’s components were inspected using the AU and UT techniques. The aim of this study was to compare the efficiency of both techniques in damage assessment. The inspected
Defect zones
Wires
d (1 mm)
Gain (2 dB)
1
1 2 3
51 44 48
18 51 42
2
1 2 3 4
33 36 46 42
18 27 27 34
structures, shown in Fig. 6, consisted in three potentially damaged anchorages collected from ‘‘Le Teil’’ bridge, which is a suspension bridge crossing the ‘‘Rhone’’ river in France. The external dimensions of each anchorage are 0:3 0:38 0:45 m. In each anchorage, a short-length cable with a diameter of 85 mm is anchored. The cable consists of a core wire surrounded by 8 layers of twisted round-wires each having a diameter of 5 mm (Fig. 7). The AU technique was primarily used to classify the anchorages according to
57
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61 Table 3 Defects detected by UT on the damaged anchorage socket. Defect zones
Wires
d (1 mm)
dm (mm)
Gain (2 dB)
Average gain (2 dB)
S1
1 2 3 4 5 6 7 8
16 17 17.5 20 19 22 24 18
20 4
30 37.5 75 75 37.5 53.6 75 75
57.32
S2
18 19 20
25.5 26.5 24
25 1
75 60 7.5
47.5
S3
29 30 31 32 33 34
13 11 14.5 21 20.5 18
16 5
71 71 35.5 60 60 30
54.58
Fig. 13. Location of defects in the damaged test specimen.
Table 2 Dimensions of the machined notches. Defects
Affected wires (0.5)
Size (2 mm)
D1 D2 D3
7 2 5
28 8 20
D1
D2
D3
Fig. 14. Notches machined on the anchorage socket.
S3: 6 wires
S1: 8 wires
Fig. 17. Mounting of the transducers during the AU test on the anchorage sockets: the transmitter on the left and the receiver on the right.
Cable
S2: 3 wires
Fig. 15. Positions of the defect zones detected on the cable using UT.
Socket anchorage Transmitter
Transmitter positions
Receiver
Cable
Fig. 16. Principle of Acousto-Ultrasonic tests on the anchorage sockets.
their damage severity. Then, a further test was performed by the UT technique. 3.2.1. AU tests A transmitter transducer was mounted on the cable at about 100 mm from the anchorage; a receiver transducer was located on the other side of the anchorage (see Figs. 8 and 9). The generated waves were guided by the cable, passed through the anchorage before being detected by the receiver. For a better scanning of the cable circumference and seeking for the influence of the angular position of the transmitter on the received waves, four positions around the cable were tested, as illustrated in Fig. 8. Sensors were glued using a ‘‘HBM glue for gauges’’. Instruments used for signals’ generation and acquisition were the same as previously. A detection threshold of 35 dB, allowing the elimination of the background noise, was fixed.
Likewise, AE energy is used to classify the inspected structures. Fig. 10 shows the transmitted energy through the three anchorages according to four circumferential positions of the transmitter for different excitation levels. For each test specimen, energy varies slightly depending on the circumferential position of the transmitter. This is probably due to the presence of discontinuities encountered on the path of the wave transmitted to the receiver. Moreover, the difference in energy levels obtained for each anchorage is clear, which allows their classification according to their damage’s level. Since energy is better transmitted through an intact structure than a damaged one, it can be concluded that the anchorage N°2 has the best structural health, the anchorage N°3 has the worst one, and the anchorage N°1 has an intermediate state. A second AUP is observed thereafter. Fig. 11 shows amplitudes (averaged values) of the received signals. Assuming that the healthier the inspected structure, the higher the amplitude of the transmitted signal, we then obtain the same classification as that ascertained using energy. However, by comparing the test specimens, the difference between the amplitude values is not as high as that obtained for energy. This can be explained by the fact that energy does not change linearly with amplitude. As energy is the measured area under the rectified signal envelope, it is not only amplitude-dependent. For instance, a change in the AU signal duration (due to interactions with damages’ reflections for example) may lead to an important evolution in energy without a great change in amplitude. This phenomenon has been found in previous studies [27,30]. In Figs. 10(a) and 11(a), we remark that the anchorage N°3 has zero energy and amplitude at the 180° and 270° positions. This could be engendered by the existence of sharp discontinuities preventing or impairing the waves transmission. Obviously, the higher the damage extent, the lower the transmitted wave, as it has been reported by some researchers [28,31].
58
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
Healthy Damaged 330
0
Healthy Damaged 330
30
300
60
240
60 300 500 400 200 100 90
270
120 210
30
300
80 100 20 40 60 90
270
0
240
150
120 210
150
180
180
(a)
(b) Healthy Damaged 330
0 30
300
60 300 100 200 90
270
240
120 210
150 180
(c) Fig. 18. Comparison between average energies of the signals transmitted through healthy and damaged anchorage sockets. Excitation signals emitted from four circumferential positions around the cable at: (a) 100 V, 1 cycle; (b) 120 V, 6 cycles; and (c) 150 V, 3 cycles.
3.2.2. UT tests UT requires a preparation of the surfaces of the wires to be inspected, especially where sensors will be mounted. The paint covering the wires was scraped over a length of about 100 mm from the anchorage as shown in Fig. 12. A light sanding was performed using a SiC polishing paper. After testing and analyzing results, no defects were detected on the anchorages N° 2 and 3, whereas two possible zones of defects were revealed on the anchorage N°1. Table 1 shows the detections found for the anchorage N°1, where ’d’ is the defect’s position inside the anchorage. Gain values provide information about the severity of damage. The first defect zone includes three damaged wires located on average at 47 1 mm inside the anchorage. The second defect zone includes four wires located on average at 39 1 mm inside the anchorage. As predicted, the location and the shape of rupture have a significant influence on the response intensity. Uniform corrosion provoking a section loss is difficult to detect by UT, whereas stress corrosion that produces cracking can be detected. This aspect has been observed in many researches [32–35]. They have demonstrated that UT is efficient for detecting crack-like defects, given the high frequencies at which the probes operate (in the order of MHz). Consequently, it can be concluded
that the UT technique is suitable for detecting defects inside anchorages, provided that some considerations on shapes and dimensions of the defects have to be taken into account.
3.3. Anchorage sockets with mastered defects The last tested specimens consisted in two anchorage sockets with rooted cables: one is healthy, the other is damaged [36]. Each cable has a diameter of 56 mm and is composed of a core wire surrounded by five layers of round wires and two additional layers of Z-shaped wires. The outer layer contains 39 wires each having a width of about 4 mm. The anchorage sockets are cylindrical with an outer diameter of 260 mm and a height of 326 mm. The damaged specimen contains three notches created on the cable inside the anchorage socket, as shown in Figs. 13 and 14. The notches have different sizes and affect one or two layers of wires. They are angularly spaced by about 120° around the cable circumference. Table 2 gives the number of affected wires and the dimensions of each defect. Hereinafter, the effectiveness of the AU and UT techniques in detecting the machined defects was evaluated.
59
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
3.3.1. UT tests After a surface preparation, the SMWB70-6 probe was used to inspect the wires one by one. A gel couplant was applied between the probe and each wire. The UT technique can be used to inspect solely the outer layer of wires. The defects detected in the damaged specimen are shown in Table 3. The wires of the outer layer were numbered from 1 to 39. Three defect zones were found: S1, S2 and S3. The first one includes eight damaged wires (N°1–8). The second defect zone has three damaged wires (N°18–20). Finally, the third defect zone contains six wires (N°29–34). In Table 3, ‘‘d’’ represents the position of defects on each wire inside the anchorage socket, and ‘‘dm’’ is the average position of each defect zone. Gain indicates the severity level of the detected damages. The higher the gain is, the larger the defect. The defect zones S1, S2 and S3 are localized as shown in Fig. 15. A good correlation in terms of positions and extents is found between the notches D1, D2 and D3, and the detected defects S1, S2 and S3, respectively. A slight error on the number of damaged wires is found, which could be justified because they were not entirely cut. The average gain is higher for zones with more defects than those having fewer damages. Small damages reflect the incident waves less strongly than the larger damages. In fact, the amplitude of the reflected wave increases
Healthy Damaged 330
with the defect size, as it has been reported in many studies [37,38]. 3.3.2. AU tests The structural health conditions of the two anchorage sockets were compared using the AU technique. The instruments used here were the same as previously. The transmitter transducer was mounted on the cable at about 35 mm from the anchorage socket. In order to better sweep the circumference of the cable and evaluate the influence of the angular position of the transmitter transducer, four positions around the cable’s circumference were tested. The receiver transducer was mounted directly on the resin and centered on the socket’s base (see Figs. 16 and 17). A detection threshold of 40 dB was used in this test. As previously, energies and amplitudes of the AE waveforms (averaged over one minute of recording) are used to compare the health conditions of the two test specimens. Fig. 18 shows the transmitted energy through the healthy and damaged anchorage sockets according to four circumferential positions for three different excitation signals. It can be noticed that energy varies with the circumferential position of the transmitter. This variation could be related to an irregular disposition of damages around the cable
0
Healthy Damaged 330
30
300
60
240
60 80 100 20 40 60 90
270
120 210
30
300
60 80 20 40 90
270
0
240
150
120 210
150
180
180
(a)
(b) Healthy Damaged 330
0 30
300
60 80 100 20 40 60 90
270
240
120 210
150 180
(c) Fig. 19. Comparison between average amplitudes of the signals transmitted through healthy and damaged anchorage sockets. Excitation signals emitted from four circumferential positions around the cable at: (a) 100 V, 1 cycle; (b) 120 V, 6 cycles; and (c) 150 V, 3 cycles.
60
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
Acknowledgments
Anchorage socket
2
This work was supported by the Franco-German project ANRFraunhofer FILAMENDT.
1 3
References Fig. 20. Schema of the tested receiver’s positions on the base of the anchorage socket.
Table 4 Effect of the receiver’s position on the transmitted wave. Receiver’s positions
Radii (mm)
Energy
Amplitude (dB)
1 2 3
0 60 90
22 16.5 10.64
60 57.64 57
inside the anchorage socket. Besides, a clear difference between the healthy and damaged specimens in terms of energy can be remarked. Energies transmitted through the healthy anchorage socket are higher than those transmitted through the damaged one. Indeed, incident waves are better transmitted through the healthy specimen, since damages represent discontinuities that can reduce the waves transmission. When observing the signal’s amplitudes in Fig. 19, same conclusions as previously can be drawn. Amplitudes are higher for the healthy anchorage socket than the damaged one. In order to check the better receiver’s position for the damage assessment, energies and amplitudes obtained for three locations, as shown in Fig. 20), were compared. A 3-cycle excitation signal with an amplitude of 10 V and a frequency of 150 kHz was used. Table 4 shows energies and amplitudes obtained for the three considered receiver’s positions. As can be seen, energy and amplitude are influenced by the receiver’s position. Particularly, energy is higher when the receiver is bonded in the center of the socket’s base. The geometric distribution of the wires embedded in the resin could focus the transmitted wave toward the center of the socket. Moreover, wave attenuation in the resin is higher than that in steel, as reported in many researches [39,40]. This could explain the decrease in energy when moving away from the socket’s center. 4. Conclusion In this study, the Acousto-Ultrasonic technique was compared to classical Ultrasonics Testing method for damage detection in anchorage zones. Several test specimens of different natures were controlled. Firstly, single strands were inspected by the AU technique in order to study the influence of progressing defects on selected AU parameters. Then, three anchorages with unknown defects were tested by both techniques. The AU technique allowed their classification according to their damage severity. The findings were confirmed by UT inspections. The last tests were performed on anchorage sockets with mastered damages. The effect of damage on AU parameters was emphasized by comparing healthy and damaged structures. UT tests allowed the detection of defect zones and revealed the damaged wires. The AU technique was found to be useful for evaluating the global structural health of the anchorage zones. Whereas the UT method would be rather suitable for detecting local and crack-like defects, and inefficient for large and diffuse damages. It should be mentioned that both techniques could be complementary. The global structural health of bridge’s components can be assessed using AU tests, followed by UT inspections of the zones potentially damaged.
[1] SETRA, TRL, LCPC, Post-tensioned concrete bridges, Anglo-French liaison report, 1999, pp. 164. [2] M. Perrin, L. Gaillet, C. Tessier, H. Idrissi, Hydrogen embrittlement of prestressing cables, Corros. Sci. 52 (6) (2010) 1915–1926. [3] D. Siegert, P. Brevet, Fatigue of stay cables inside end fittings: high frequencies of wind induced vibrations, OIPEEC Bull. 89 (2005) 43–51. [4] B. Cherry, S. Price, Pitting, crevice and stress corrosion cracking studies of cold drawn eutectoid steels, Corros. Sci. 20 (11) (1980) 1163–1183. [5] L. Dieng, J. Urvoy, D. Siegert, P. Brevet, V. Périer, C. Tessier, Assessment of lubrication and zinc coating on the high cycle fretting fatigue behaviour of high strength steel wires, OIPEEC, Johannesburg, 2007. pp. 85–97. [6] S. Watson, D. Stafford, Cables in trouble, Civil Eng. ASCE 58 (4) (1988) 38–41. [7] C. Chaplin, Failure mechanisms in wire ropes, Eng. Fail. Anal. 2 (1) (1995) 45– 57. [8] V. Périer, L. Dieng, L. Gaillet, C. Tessier, S. Fouvry, Fretting-fatigue behaviour of bridge engineering cables in a solution of sodium chloride, Wear 267 (1) (2009) 308–314. [9] R. Hobbs, M. Raoof, Mechanism of fretting fatigue in steel cables, Int. J. Fatigue 16 (4) (1994) 273–280. [10] R. Woodward, Detecting fractures in steel cables, Wire Ind. 56 (1989) 401–405. [11] H. Weischedel, The inspection of wire ropes in service: a critical review, Mater. Eval. 43 (1985) 1592–1605. [12] A. Potts, Non-destructive testing of steel wire ropes, in: Symposium of the British Institute of NonDestructive Testing, London, 1988. [13] K. Hanasaki, K. Tsukada, T. Moriya, A magnetic method for evaluation of the deterioration of large diameter wire ropes, in: WCNDT Symposium, Roma, 2000. [14] A. Nair, C. Cai, Acoustic emission monitoring of bridges: review and case studies, Eng. Struct. 32 (6) (2010) 1704–1714. [15] D. Cullington, D. MacNeil, P. Paulson, J. Elliott, Continuous acoustic monitoring of grouted post-tensioned concrete bridges, NDT and E Int. 34 (2) (2001) 95– 105. [16] B. Pavlakovic, M. Lowe, P. Cawley, The inspection of tendons in post-tensioned concrete using guided ultrasonic waves, Insight 41 (7) (1999) 446–448. [17] N. Suzuki, H. Takamatsu, S. Kawashima, K.-I. Sugii, M. Iwasaki, Ultrasonic detection method for wire breakage, Kobelco Technol. Rev. 4 (1988) 23–26. [18] A. Ghorbanpoor, Evaluation of prestressed concrete girders using magnetic flux leakage, in: Proceedings of the 1999 ASCE Structures Congress, Structural Engineering in the 21st Century, New Orleans, 1999, pp. 284–287. [19] H. Zejli, L. Gaillet, A. Laksimi, S. Benmedakhene, Detection of the presence of broken wires in cables by acoustic emission inspection, J. Bridge Eng. 17 (6) (2012) 921–927. [20] A. Vary, The Acousto-Ultrasonic approach, in: J.C. Duke Jr. (Ed.), AcoustoUltrasonics, Theory and Applications, Plenum Press, 1988. [21] A. Vary, Acousto-Ultrasonics, in: J. Summerscales (Ed.), Nondestructive Testing of Fibre-Reinforced Plastics Composites, Elsevier Science, Publishers, 1990. [22] A. Bartos, Acousto-Ultrasonic pattern analysis of impact-damaged composite panels, in: Proceedings of the 2nd International Conference On AcoustoUltrasonics, Alex Vary, ed., ASNT, 1993. [23] A. Bartos, J. Strycek, R. Gewalt, H. Loertscher, T. Chang, Ultrasonic and AcoustoUltrasonic inspection and characterization of titanium alloy structures, in: Nondestructive Characterization of Materials VIII, Springer, 1998, pp. 239– 244. [24] K. Shiloh, A. Bartos, A. Frain, E. Lindgren, Ultrasonic detection of corrosion between riveted plates, in: SPIE’s 5th Annual International Symposium on Nondestructive Evaluation and Health Monitoring of Aging Infrastructure, International Society for Optics and Photonics, 2000, pp. 70–79. [25] G. Molina, Y. Haddad, Acousto-ultrasonics approach to the characterization of impact properties of a class of engineering materials, Int. J. Press. Vess. Pip. 67 (3) (1996) 307–315. [26] C. Biemans, W. Staszewski, C. Boller, G. Tomlinson, Crack detection in metallic structures using broadband excitation of acousto-ultrasonics, J. Intell. Mater. Syst. Struct. 12 (8) (2001) 589–597. [27] O.-Y. Kwon, S.-H. Lee, Acousto-ultrasonic evaluation of adhesively bonded CFRP–aluminum joints, NDT and E Int. 32 (3) (1999) 153–160. [28] A. Baltazar, C.D. Hernandez-Salazar, B. Manzanares-Martinez, Study of wave propagation in a multiwire cable to determine structural damage, NDT and E Int. 43 (8) (2010) 726–732. [29] P. Rizzo, F.L. Di Scalea, Wave propagation in multi-wire strands by waveletbased laser ultrasound, Exp. Mech. 44 (4) (2004) 407–415. [30] S. Al-Dossary, R. Hamzah, D. Mba, Observations of changes in acoustic emission waveform for varying seeded defect sizes in a rolling element bearing, Appl. Acoust. 70 (1) (2009) 58–81. [31] A.L. Gyekenyesi, G.N. Morscher, L.M. Cosgriff, In-situ monitoring of damage in SiC/SiC composites using acousto-ultrasonics, Compos. Part B: Eng. 37 (1) (2006) 47–53.
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61 [32] W. Yeih, R. Huang, Detection of the corrosion damage in reinforced concrete members by ultrasonic testing, Cem. Concr. Res. 28 (7) (1998) 1071–1083. [33] A. Shah, Y. Ribakov, C. Zhang, Efficiency and sensitivity of linear and non-linear ultrasonics to identifying micro and macro-scale defects in concrete, Mater. Des. 50 (2013) 905–916. [34] O. Gericke, Determination of the geometry of hidden defects by ultrasonic pulse analysis testing, J. Acoust. Soc. Am. 35 (3) (2005) 364–368. [35] G.J. Gruber, Defect identification and sizing by the ultrasonic satellite-pulse technique, J. Nondestr. Eval. 1 (4) (1980) 263–276. [36] H. Zejli, Détection et localisation par Emission Acoustique de fils rompus dans les ancrages des câbles d’ouvrages d’art, Ph.D. thesis, Université de Technologie de Compiègne, 2007.
61
[37] D. Alleyne, M. Lowe, P. Cawley, The reflection of guided waves from circumferential notches in pipes, J. Appl. Mech. 65 (3) (1998) 635–641. [38] A. Demma, P. Cawley, M. Lowe, A. Roosenbrand, B. Pavlakovic, The reflection of guided waves from notches in pipes: a guide for interpreting corrosion measurements, NDT and E Int. 37 (3) (2004) 167–180. [39] S. Rokhlin, D. Lewis, K. Graff, L. Adler, Real-time study of frequency dependence of attenuation and velocity of ultrasonic waves during the curing reaction of epoxy resin, J. Acoust. Soc. Am. 79 (6) (1986) 1786– 1793. [40] A.M. Lindrose, Ultrasonic wave and moduli changes in a curing epoxy resin, Exp. Mech. 18 (6) (1978) 227–232.
Contents lists available at ScienceDirect
Ultrasonics journal homepage: www.elsevier.com/locate/ultras
Non-destructive evaluation of anchorage zones by ultrasonics techniques M. Kharrat ⇑, L. Gaillet Institut Français des Sciences et Technologies de Transport, de l’Aménagement et des Réseaux, Département Matériaux et Structures, Laboratoire Structures Métalliques et à Câbles, Route de Bouaye, CS4, 44344 Bouguenais Cedex, France
a r t i c l e
i n f o
Article history: Received 15 June 2014 Received in revised form 9 March 2015 Accepted 12 March 2015 Available online 23 March 2015 Keywords: NDT Anchorages Cables Acousto-Ultrasonics Ultrasonics
a b s t r a c t This work aims to evaluate the efficiency and reliability of two Non-Destructive Testing (NDT) methods for damage assessment in bridges’ anchorages. The Acousto-Ultrasonic (AU) technique is compared to classical Ultrasonic Testing (UT) in terms of defect detection and structural health classification. The AU technique is firstly used on single seven-wire strands damaged by artificial defects. The effect of growing defects on the waves traveling through the strands is evaluated. Thereafter, three specimens of anchorages with unknown defects are inspected by the AU and UT techniques. Damage assessment results from both techniques are then compared. The structural health conditions of the specimens can be then classified by a damage severity criterion. Finally, a damaged anchorage socket with mastered defects is controlled by the same techniques. The UT allows the detection and localization of damaged wires. The AU technique is used to bring out the effect of defects on acoustic features by comparing a healthy and damaged anchorage sockets. It is concluded that the UT method is suitable for local and crack-like defects, whereas the AU technique enables the assessment of the global structural health of the anchorage zones. Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction Upon commissioning, civil engineering infrastructures, like bridges, are subjected to multiple aggressive conditions. Generally, the older the bridge is, the higher the risk of disorder. Most degradation mechanisms are currently known and have been reported in several works [1–5]. Corrosion and fatigue are the main causes of damage and have often a harmful combination on the health condition of bridges. In the most critical cases, the damaged bridge’s components must be replaced. Among the major bridge’s components, cables and anchorage zones could be subjected to hard solicitations. Cables are primarily undergoing environmental aggressions. Run-off water added with deicing salts could penetrate in cables and gravitationally flow toward tightened areas (ties of suspension lines) or toward lower zones (such as anchorages). This confined humidity often engenders corrosion. Non-alloy steel used in cables is sensitive to corrosion by dissolution, so that a section loss more or less homogeneous can affect all or part of the wires constituting cables and can decrease their breaking strength [6,7]. Moreover, fretting⇑ Corresponding author at: Department of Applied Mechanics, FEMTO-ST Institute, Franche-Comté University, 25000 Besançon, France. Tel.: +33 682212604. E-mail address: [email protected] (M. Kharrat). http://dx.doi.org/10.1016/j.ultras.2015.03.007 0041-624X/Ó 2015 Elsevier B.V. All rights reserved.
fatigue mechanism produces small relative displacements between the wires of a single cable (inter-wires contact), between neighboring cables, or also between a cable and other fixed parts like anchorages. This can cause wear and/or cracking of the wires [8,9]. As the anchorage zone is an inaccessible part of the bridge and is located at the bottom of the retaining cables, it provides an ideal place for water accumulation and promotes corrosion development. The major problem here is the rupture of wires that cannot be anticipated by visual inspections. This bridge’s component requires the implementation of operational inspection techniques able to assess its structural health and prematurely detect damages before the failure of the whole structure [10,11]. Regarding Non-Destructive Testing (NDT) of civil engineering structures, an important distinction is commonly made between accessible and inaccessible parts. Accessible parts relate to components where the element to be inspected is directly visible and easily accessible; while inaccessible parts involve hidden components of the structure, so that require a prior intervention before inspections. If the current state-of-the-art of monitoring techniques of accessible parts is based on visual inspection [12], magnetic inspection [13], and acoustic monitoring [14,15], that concerning inaccessible parts is still missing. The inspection and monitoring of inaccessible areas are of major concern as they are often considered as the weakest parts where fatigue and corrosion
53
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
mechanisms can appear. However, many structural factors can greatly complicate their inspection. Key factors mainly include the influence of different material layers and multiple interfaces in protection systems of cables embedded in anchorages and suspension collars. Furthermore, various techniques have been studied among which ultrasonic guided waves [16,17], magnetic flux leakage inspection [18], micro-magnetic methods, and acoustic emission monitoring [19] seem to be suitable. The AcoustoUltrasonic (AU) is a relatively new technique in civil engineering. It has been initially used for inspection and characterization of graphite-reinforced plastics (GRP) [20–22]. Acousto-Ultrasonic parameters (AUPs) have shown a good correlation with the mechanical strength for both tension and compression test configurations [23] and even for detecting corrosion evolution in coatings of aircraft wings [24]. This technique has been also used for non-destructive quantitative characterization of residual impact strength of polyvinyl chloride (PVC) using the AUPs [25]. Detection of pre-machined defects on a metal plate has been studied using a broadband excitation through temporal, frequency and wavelet analyzes [26]. Correlation between the amount of artificial defects in bonded regions and AUPs in adhesively bonded joints of carbon-fiber reinforced plastic (CFRP) laminates and aluminum
Transmitter E1 Receiver R
Defect
Strand
265 mm
Transmitter E2
36 mm 695 mm
(a) Transmitter E1
5
6
1
2
3
4
Transmitter E2
(b) Fig. 3. Acouto-ultrasonic test performed on a single strand using two transmitters E1 and E2 and a receiver R. (a) Lateral view; (b) sectional view.
3000
E1R−1V3C E1R−3V3C E1R−5V3C E1R−7V3C E1R−10V3C
2500
Energy
2000 1500 1000 500 0
0
1
2
3
4
5
6
Number of cut wires
(a) 2500
Fig. 1. UT instruments: Krautkramer’s USM25 pulser/receiver device (on the left); and SMWB70-6 probe (on the right).
E2R−1V3C E2R−3V3C E2R−5V3C E2R−7V3C E2R−10V3C
Controller AE chain
Energy
2000 1500 1000 500
Digitizer
0
Pass band filter
Signal generator
0
1
2
3
4
5
6
Number of cut wires
(b) Amplifier
Fig. 4. Evolution of the average energy depending on the defect expansion while exciting by different waveforms. (a) Waves emitted by E1; (b) waves emitted by E2.
Pre-amplifier Transmitter
Receiver
Specimen Fig. 2. Acousto-Ultrasonics principle.
plates has been also investigated [27]. However, this technique has been rarely used for damage characterization in metallic cables. This work aims to test the reliability of two NDT methods for damage assessment in some bridge’s parts. Specifically, the Acousto-Ultrasonic technique is compared to classical Ultrasonic Testing (UT) technique in terms of defect detection and structural health classification.
54
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
100
Amplitude (dB)
95 90 85 80
E1R−1V3C E1R−3V3C E1R−5V3C E1R−7V3C E1R−10V3C
75 70
Fig. 7. Sectional view of a stranded cable having 8 layers of wires surrounding a core wire.
Cable
65
0
1
2
3
4
5
Transmitter Receiver
6
Number of cut wires
(a)
Transmitter positions
Anchorage
100 Fig. 8. Principle of the AU test performed on anchorages from ‘‘Le Teil’’ Bridge.
Amplitude (dB)
95 90 85
E2R−1V3C E2R−3V3C E2R−5V3C E2R−7V3C E2R−10V3C
80 75 70
0
1
2
Fig. 9. Photos of the inspected anchorages with the transducers during the AU test: the transmitter on the left and the receiver on the right.
3
4
5
6
Number of cut wires
(b) Fig. 5. Evolution of the average amplitude depending on the defect expansion while exciting by different waveforms. (a) Waves emitted by E1; (b) waves emitted by E2.
2. NDT techniques The UT technique was used in reflection mode to assess the presence of defects in anchorage zones. A Krautkramer’s USM25 ultrasonic pulser/receiver device and a SMWB70-6 probe, characterized by an angle of incidence of 70° in transmission and an operating frequency of 6 MHz, were employed (Fig. 1). A SLC 70 gel-type couplant was applied between the transducer and the test specimen to ensure a proper transmission of signals. The AU technique combines aspects of Acoustic Emission (AE) signal analysis with Ultrasonics assessment methods. The transmitter transducer generates a specific ultrasonic waveform that propagates through the tested specimen before being collected by a receiver transducer. Signals resulting from multiple reflections
1
2
Fig. 6. Potentially damaged anchorages collected from ‘‘Le Teil’’ bridge.
3
and interactions with the microstructure of the material are treated in the same way as AE signals. The principle of the AU technique is illustrated in Fig. 2. In our study, an ARB-1410 Arbitrary Waveform Generator Board (Mistras Group) was used to generate the ultrasonic signals. The waveforms were 3-cycle windowed sine functions at 150 kHz. Different amplitudes were used: 1, 3, 5, 7 and 10 V. These signals were sent repeatedly at a rate of 1 Hz over a recording duration of about 1 min. The acquisition system included a PCI-2 card having two input channels with a sampling rate of 40 MHz. The transducers employed for both emission and reception were of R15atype characterized by a resonance frequency of 150 kHz and an operating bandwidth of 50–400 kHz. Transducers’ bonding was performed by using a ‘‘HBM glue for gauges’’.
3. Experimental tests and results’ discussions This section presents the experimental tests performed on different types of specimens. Firstly, the AU technique was used on single strands where some defects were progressively created. The effect of damage growth on the collected signals was evaluated. Thereafter, three anchorages with unknown defects were controlled by AU and UT techniques. The damage assessment abilities of both techniques were compared The test specimens were also classified by a damage severity criterion. Finally, healthy and damaged test specimens with mastered defects were inspected by the same techniques. Each specimen consisted in a long cable anchored in an anchorage socket. A comparison of the efficiency of both techniques was addressed.
55
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
Specimen 2 Specimen 1 330 Specimen 3
Specimen 2 Specimen 1 330 Specimen 3
0 30
300
60
100
120 210
Specimen 2 Specimen 1 330 Specimen 3
60
300 90
240
30
300
200
270
0
600 800 200 400 90
270
240
150
120 210
150
180
180
(a)
(b)
Specimen 2 Specimen 1 330 Specimen 3
0 30
300
60 300 400 100 200 90
270
240
120 210
150
0 30
300
60 800 1000 400 600 200 90
270
240
120 210
150
180
180
(c)
(d)
Fig. 10. Comparison between average energies of the signals transmitted through the three anchorages. Excitation signals emitted from four circumferential positions around the cable at: (a) 10 V, 3 cycles; (b) 25 V, 3 cycles; (c) 10 V, 6 cycles; and (d) 25 V, 6 cycles.
3.1. Acousto-Ultrasonic tests on single strands The first tests were performed on single strands by the AU technique in order to study the influence of progressing defects on selected AU parameters. The tested T15.7-type strand was 1.39m-long and composed of seven wires, each having a diameter of 5.4 mm. As shown in Fig. 3(a), two transmitters (E1 and E2) were coaxially positioned at the middle of the strand; a receiver sensor (R) was mounted at one end. A growing defect was artificially created by cutting out progressively 36-mm-long segments from the wires. The order of cutting the wires’ segments is schematized in Fig. 3(b). The test was started by cutting segments from wires 1 and 2, whereas E1 was in contact with intact wires 5 and 6. Successively, after each cutting, a wave was firstly sent by the E1 transmitter, and then by E2. The transmitted wave was collected by the R sensor in each case. Energy and Amplitude (averaged values over 1 min of recording) are selected as AU parameters in order to assess the damage expansion during this test. The evolution of the average energy and amplitude depending on the defect expansion is illustrated in Figs. 4 and 5. Fig. 4(a) shows an increase in energy depending on the number of the cut wires until the fourth one. Then, energy decreases with the fifth and sixth cut-wires on which E1 was mounted. Fig. 4(b)
shows a global decrease in energy with the number of cut wires. Indeed, for the two transmitters, the incident wave is substantially guided by the wires on which the transmitters E1 and E2 are mounted. A part of the incident energy is also guided by the other wires. When a transmitter is mounted on the cut wires (case of E2), the decrease in the energy transmitted to the R sensor is due to the defect, which reflects most of the incident wave. The other wires that are still intact would guide the rest of the wave energy. Thus, larger the damage is, higher the wave is reflected. When a transmitter is located on the uncut wires (case of E1), energy received by the R sensor increases until reaching the wires on which E1 is mounted. However, we can note that amplitudes of the received signals remain almost constant for the first four wires (see Fig. 5), although energy is evolving. This could be produced when duration and/or counts of an AU signal evolve. To explain the phenomenon occurred in Fig. 4(b), we can mention that the incident wave, which is guided by undamaged wires, propagates to the R sensor, whilst the presence of a defect creates mode conversions and wave interactions [28,29]. A part of these waves could propagate in turn toward the R sensor, and thereafter could be superimposed with the main incident wave. Hence, this could increase duration and/or counts of the signal, leading to an increase in energy.
56
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
Specimen 2 Specimen 1 330 Specimen 3
Specimen 2 Specimen 1 330 Specimen 3
0 30
300
60
240
Specimen 2 Specimen 1 330 Specimen 3
60 60 80 20 40 90
270
120 210
30
300
60 80 20 40 90
270
0
240
150
120 210
150
180
180
(a)
(b) Specimen 2 Specimen 1 330 Specimen 3
0 30
300
60
240
60 80 100 20 40 60 90
270
120 210
30
300
60 80 20 40 90
270
0
240
150
120 210
150
180
180
(c)
(d)
Fig. 11. Comparison between average amplitudes of the signals transmitted through the three anchorages. Excitation signals emitted from four circumferential positions around the cable at: (a) 10 V, 3 cycles; (b) 25 V, 3 cycles; (c) 10 V, 6 cycles; and (d) 25 V, 6 cycles.
Table 1 Anomalies detected by UT in the anchorage N°1.
Fig. 12. Preparation of the wires’ surfaces.
3.2. Anchorages with unknown defects Real bridge’s components were inspected using the AU and UT techniques. The aim of this study was to compare the efficiency of both techniques in damage assessment. The inspected
Defect zones
Wires
d (1 mm)
Gain (2 dB)
1
1 2 3
51 44 48
18 51 42
2
1 2 3 4
33 36 46 42
18 27 27 34
structures, shown in Fig. 6, consisted in three potentially damaged anchorages collected from ‘‘Le Teil’’ bridge, which is a suspension bridge crossing the ‘‘Rhone’’ river in France. The external dimensions of each anchorage are 0:3 0:38 0:45 m. In each anchorage, a short-length cable with a diameter of 85 mm is anchored. The cable consists of a core wire surrounded by 8 layers of twisted round-wires each having a diameter of 5 mm (Fig. 7). The AU technique was primarily used to classify the anchorages according to
57
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61 Table 3 Defects detected by UT on the damaged anchorage socket. Defect zones
Wires
d (1 mm)
dm (mm)
Gain (2 dB)
Average gain (2 dB)
S1
1 2 3 4 5 6 7 8
16 17 17.5 20 19 22 24 18
20 4
30 37.5 75 75 37.5 53.6 75 75
57.32
S2
18 19 20
25.5 26.5 24
25 1
75 60 7.5
47.5
S3
29 30 31 32 33 34
13 11 14.5 21 20.5 18
16 5
71 71 35.5 60 60 30
54.58
Fig. 13. Location of defects in the damaged test specimen.
Table 2 Dimensions of the machined notches. Defects
Affected wires (0.5)
Size (2 mm)
D1 D2 D3
7 2 5
28 8 20
D1
D2
D3
Fig. 14. Notches machined on the anchorage socket.
S3: 6 wires
S1: 8 wires
Fig. 17. Mounting of the transducers during the AU test on the anchorage sockets: the transmitter on the left and the receiver on the right.
Cable
S2: 3 wires
Fig. 15. Positions of the defect zones detected on the cable using UT.
Socket anchorage Transmitter
Transmitter positions
Receiver
Cable
Fig. 16. Principle of Acousto-Ultrasonic tests on the anchorage sockets.
their damage severity. Then, a further test was performed by the UT technique. 3.2.1. AU tests A transmitter transducer was mounted on the cable at about 100 mm from the anchorage; a receiver transducer was located on the other side of the anchorage (see Figs. 8 and 9). The generated waves were guided by the cable, passed through the anchorage before being detected by the receiver. For a better scanning of the cable circumference and seeking for the influence of the angular position of the transmitter on the received waves, four positions around the cable were tested, as illustrated in Fig. 8. Sensors were glued using a ‘‘HBM glue for gauges’’. Instruments used for signals’ generation and acquisition were the same as previously. A detection threshold of 35 dB, allowing the elimination of the background noise, was fixed.
Likewise, AE energy is used to classify the inspected structures. Fig. 10 shows the transmitted energy through the three anchorages according to four circumferential positions of the transmitter for different excitation levels. For each test specimen, energy varies slightly depending on the circumferential position of the transmitter. This is probably due to the presence of discontinuities encountered on the path of the wave transmitted to the receiver. Moreover, the difference in energy levels obtained for each anchorage is clear, which allows their classification according to their damage’s level. Since energy is better transmitted through an intact structure than a damaged one, it can be concluded that the anchorage N°2 has the best structural health, the anchorage N°3 has the worst one, and the anchorage N°1 has an intermediate state. A second AUP is observed thereafter. Fig. 11 shows amplitudes (averaged values) of the received signals. Assuming that the healthier the inspected structure, the higher the amplitude of the transmitted signal, we then obtain the same classification as that ascertained using energy. However, by comparing the test specimens, the difference between the amplitude values is not as high as that obtained for energy. This can be explained by the fact that energy does not change linearly with amplitude. As energy is the measured area under the rectified signal envelope, it is not only amplitude-dependent. For instance, a change in the AU signal duration (due to interactions with damages’ reflections for example) may lead to an important evolution in energy without a great change in amplitude. This phenomenon has been found in previous studies [27,30]. In Figs. 10(a) and 11(a), we remark that the anchorage N°3 has zero energy and amplitude at the 180° and 270° positions. This could be engendered by the existence of sharp discontinuities preventing or impairing the waves transmission. Obviously, the higher the damage extent, the lower the transmitted wave, as it has been reported by some researchers [28,31].
58
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
Healthy Damaged 330
0
Healthy Damaged 330
30
300
60
240
60 300 500 400 200 100 90
270
120 210
30
300
80 100 20 40 60 90
270
0
240
150
120 210
150
180
180
(a)
(b) Healthy Damaged 330
0 30
300
60 300 100 200 90
270
240
120 210
150 180
(c) Fig. 18. Comparison between average energies of the signals transmitted through healthy and damaged anchorage sockets. Excitation signals emitted from four circumferential positions around the cable at: (a) 100 V, 1 cycle; (b) 120 V, 6 cycles; and (c) 150 V, 3 cycles.
3.2.2. UT tests UT requires a preparation of the surfaces of the wires to be inspected, especially where sensors will be mounted. The paint covering the wires was scraped over a length of about 100 mm from the anchorage as shown in Fig. 12. A light sanding was performed using a SiC polishing paper. After testing and analyzing results, no defects were detected on the anchorages N° 2 and 3, whereas two possible zones of defects were revealed on the anchorage N°1. Table 1 shows the detections found for the anchorage N°1, where ’d’ is the defect’s position inside the anchorage. Gain values provide information about the severity of damage. The first defect zone includes three damaged wires located on average at 47 1 mm inside the anchorage. The second defect zone includes four wires located on average at 39 1 mm inside the anchorage. As predicted, the location and the shape of rupture have a significant influence on the response intensity. Uniform corrosion provoking a section loss is difficult to detect by UT, whereas stress corrosion that produces cracking can be detected. This aspect has been observed in many researches [32–35]. They have demonstrated that UT is efficient for detecting crack-like defects, given the high frequencies at which the probes operate (in the order of MHz). Consequently, it can be concluded
that the UT technique is suitable for detecting defects inside anchorages, provided that some considerations on shapes and dimensions of the defects have to be taken into account.
3.3. Anchorage sockets with mastered defects The last tested specimens consisted in two anchorage sockets with rooted cables: one is healthy, the other is damaged [36]. Each cable has a diameter of 56 mm and is composed of a core wire surrounded by five layers of round wires and two additional layers of Z-shaped wires. The outer layer contains 39 wires each having a width of about 4 mm. The anchorage sockets are cylindrical with an outer diameter of 260 mm and a height of 326 mm. The damaged specimen contains three notches created on the cable inside the anchorage socket, as shown in Figs. 13 and 14. The notches have different sizes and affect one or two layers of wires. They are angularly spaced by about 120° around the cable circumference. Table 2 gives the number of affected wires and the dimensions of each defect. Hereinafter, the effectiveness of the AU and UT techniques in detecting the machined defects was evaluated.
59
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
3.3.1. UT tests After a surface preparation, the SMWB70-6 probe was used to inspect the wires one by one. A gel couplant was applied between the probe and each wire. The UT technique can be used to inspect solely the outer layer of wires. The defects detected in the damaged specimen are shown in Table 3. The wires of the outer layer were numbered from 1 to 39. Three defect zones were found: S1, S2 and S3. The first one includes eight damaged wires (N°1–8). The second defect zone has three damaged wires (N°18–20). Finally, the third defect zone contains six wires (N°29–34). In Table 3, ‘‘d’’ represents the position of defects on each wire inside the anchorage socket, and ‘‘dm’’ is the average position of each defect zone. Gain indicates the severity level of the detected damages. The higher the gain is, the larger the defect. The defect zones S1, S2 and S3 are localized as shown in Fig. 15. A good correlation in terms of positions and extents is found between the notches D1, D2 and D3, and the detected defects S1, S2 and S3, respectively. A slight error on the number of damaged wires is found, which could be justified because they were not entirely cut. The average gain is higher for zones with more defects than those having fewer damages. Small damages reflect the incident waves less strongly than the larger damages. In fact, the amplitude of the reflected wave increases
Healthy Damaged 330
with the defect size, as it has been reported in many studies [37,38]. 3.3.2. AU tests The structural health conditions of the two anchorage sockets were compared using the AU technique. The instruments used here were the same as previously. The transmitter transducer was mounted on the cable at about 35 mm from the anchorage socket. In order to better sweep the circumference of the cable and evaluate the influence of the angular position of the transmitter transducer, four positions around the cable’s circumference were tested. The receiver transducer was mounted directly on the resin and centered on the socket’s base (see Figs. 16 and 17). A detection threshold of 40 dB was used in this test. As previously, energies and amplitudes of the AE waveforms (averaged over one minute of recording) are used to compare the health conditions of the two test specimens. Fig. 18 shows the transmitted energy through the healthy and damaged anchorage sockets according to four circumferential positions for three different excitation signals. It can be noticed that energy varies with the circumferential position of the transmitter. This variation could be related to an irregular disposition of damages around the cable
0
Healthy Damaged 330
30
300
60
240
60 80 100 20 40 60 90
270
120 210
30
300
60 80 20 40 90
270
0
240
150
120 210
150
180
180
(a)
(b) Healthy Damaged 330
0 30
300
60 80 100 20 40 60 90
270
240
120 210
150 180
(c) Fig. 19. Comparison between average amplitudes of the signals transmitted through healthy and damaged anchorage sockets. Excitation signals emitted from four circumferential positions around the cable at: (a) 100 V, 1 cycle; (b) 120 V, 6 cycles; and (c) 150 V, 3 cycles.
60
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61
Acknowledgments
Anchorage socket
2
This work was supported by the Franco-German project ANRFraunhofer FILAMENDT.
1 3
References Fig. 20. Schema of the tested receiver’s positions on the base of the anchorage socket.
Table 4 Effect of the receiver’s position on the transmitted wave. Receiver’s positions
Radii (mm)
Energy
Amplitude (dB)
1 2 3
0 60 90
22 16.5 10.64
60 57.64 57
inside the anchorage socket. Besides, a clear difference between the healthy and damaged specimens in terms of energy can be remarked. Energies transmitted through the healthy anchorage socket are higher than those transmitted through the damaged one. Indeed, incident waves are better transmitted through the healthy specimen, since damages represent discontinuities that can reduce the waves transmission. When observing the signal’s amplitudes in Fig. 19, same conclusions as previously can be drawn. Amplitudes are higher for the healthy anchorage socket than the damaged one. In order to check the better receiver’s position for the damage assessment, energies and amplitudes obtained for three locations, as shown in Fig. 20), were compared. A 3-cycle excitation signal with an amplitude of 10 V and a frequency of 150 kHz was used. Table 4 shows energies and amplitudes obtained for the three considered receiver’s positions. As can be seen, energy and amplitude are influenced by the receiver’s position. Particularly, energy is higher when the receiver is bonded in the center of the socket’s base. The geometric distribution of the wires embedded in the resin could focus the transmitted wave toward the center of the socket. Moreover, wave attenuation in the resin is higher than that in steel, as reported in many researches [39,40]. This could explain the decrease in energy when moving away from the socket’s center. 4. Conclusion In this study, the Acousto-Ultrasonic technique was compared to classical Ultrasonics Testing method for damage detection in anchorage zones. Several test specimens of different natures were controlled. Firstly, single strands were inspected by the AU technique in order to study the influence of progressing defects on selected AU parameters. Then, three anchorages with unknown defects were tested by both techniques. The AU technique allowed their classification according to their damage severity. The findings were confirmed by UT inspections. The last tests were performed on anchorage sockets with mastered damages. The effect of damage on AU parameters was emphasized by comparing healthy and damaged structures. UT tests allowed the detection of defect zones and revealed the damaged wires. The AU technique was found to be useful for evaluating the global structural health of the anchorage zones. Whereas the UT method would be rather suitable for detecting local and crack-like defects, and inefficient for large and diffuse damages. It should be mentioned that both techniques could be complementary. The global structural health of bridge’s components can be assessed using AU tests, followed by UT inspections of the zones potentially damaged.
[1] SETRA, TRL, LCPC, Post-tensioned concrete bridges, Anglo-French liaison report, 1999, pp. 164. [2] M. Perrin, L. Gaillet, C. Tessier, H. Idrissi, Hydrogen embrittlement of prestressing cables, Corros. Sci. 52 (6) (2010) 1915–1926. [3] D. Siegert, P. Brevet, Fatigue of stay cables inside end fittings: high frequencies of wind induced vibrations, OIPEEC Bull. 89 (2005) 43–51. [4] B. Cherry, S. Price, Pitting, crevice and stress corrosion cracking studies of cold drawn eutectoid steels, Corros. Sci. 20 (11) (1980) 1163–1183. [5] L. Dieng, J. Urvoy, D. Siegert, P. Brevet, V. Périer, C. Tessier, Assessment of lubrication and zinc coating on the high cycle fretting fatigue behaviour of high strength steel wires, OIPEEC, Johannesburg, 2007. pp. 85–97. [6] S. Watson, D. Stafford, Cables in trouble, Civil Eng. ASCE 58 (4) (1988) 38–41. [7] C. Chaplin, Failure mechanisms in wire ropes, Eng. Fail. Anal. 2 (1) (1995) 45– 57. [8] V. Périer, L. Dieng, L. Gaillet, C. Tessier, S. Fouvry, Fretting-fatigue behaviour of bridge engineering cables in a solution of sodium chloride, Wear 267 (1) (2009) 308–314. [9] R. Hobbs, M. Raoof, Mechanism of fretting fatigue in steel cables, Int. J. Fatigue 16 (4) (1994) 273–280. [10] R. Woodward, Detecting fractures in steel cables, Wire Ind. 56 (1989) 401–405. [11] H. Weischedel, The inspection of wire ropes in service: a critical review, Mater. Eval. 43 (1985) 1592–1605. [12] A. Potts, Non-destructive testing of steel wire ropes, in: Symposium of the British Institute of NonDestructive Testing, London, 1988. [13] K. Hanasaki, K. Tsukada, T. Moriya, A magnetic method for evaluation of the deterioration of large diameter wire ropes, in: WCNDT Symposium, Roma, 2000. [14] A. Nair, C. Cai, Acoustic emission monitoring of bridges: review and case studies, Eng. Struct. 32 (6) (2010) 1704–1714. [15] D. Cullington, D. MacNeil, P. Paulson, J. Elliott, Continuous acoustic monitoring of grouted post-tensioned concrete bridges, NDT and E Int. 34 (2) (2001) 95– 105. [16] B. Pavlakovic, M. Lowe, P. Cawley, The inspection of tendons in post-tensioned concrete using guided ultrasonic waves, Insight 41 (7) (1999) 446–448. [17] N. Suzuki, H. Takamatsu, S. Kawashima, K.-I. Sugii, M. Iwasaki, Ultrasonic detection method for wire breakage, Kobelco Technol. Rev. 4 (1988) 23–26. [18] A. Ghorbanpoor, Evaluation of prestressed concrete girders using magnetic flux leakage, in: Proceedings of the 1999 ASCE Structures Congress, Structural Engineering in the 21st Century, New Orleans, 1999, pp. 284–287. [19] H. Zejli, L. Gaillet, A. Laksimi, S. Benmedakhene, Detection of the presence of broken wires in cables by acoustic emission inspection, J. Bridge Eng. 17 (6) (2012) 921–927. [20] A. Vary, The Acousto-Ultrasonic approach, in: J.C. Duke Jr. (Ed.), AcoustoUltrasonics, Theory and Applications, Plenum Press, 1988. [21] A. Vary, Acousto-Ultrasonics, in: J. Summerscales (Ed.), Nondestructive Testing of Fibre-Reinforced Plastics Composites, Elsevier Science, Publishers, 1990. [22] A. Bartos, Acousto-Ultrasonic pattern analysis of impact-damaged composite panels, in: Proceedings of the 2nd International Conference On AcoustoUltrasonics, Alex Vary, ed., ASNT, 1993. [23] A. Bartos, J. Strycek, R. Gewalt, H. Loertscher, T. Chang, Ultrasonic and AcoustoUltrasonic inspection and characterization of titanium alloy structures, in: Nondestructive Characterization of Materials VIII, Springer, 1998, pp. 239– 244. [24] K. Shiloh, A. Bartos, A. Frain, E. Lindgren, Ultrasonic detection of corrosion between riveted plates, in: SPIE’s 5th Annual International Symposium on Nondestructive Evaluation and Health Monitoring of Aging Infrastructure, International Society for Optics and Photonics, 2000, pp. 70–79. [25] G. Molina, Y. Haddad, Acousto-ultrasonics approach to the characterization of impact properties of a class of engineering materials, Int. J. Press. Vess. Pip. 67 (3) (1996) 307–315. [26] C. Biemans, W. Staszewski, C. Boller, G. Tomlinson, Crack detection in metallic structures using broadband excitation of acousto-ultrasonics, J. Intell. Mater. Syst. Struct. 12 (8) (2001) 589–597. [27] O.-Y. Kwon, S.-H. Lee, Acousto-ultrasonic evaluation of adhesively bonded CFRP–aluminum joints, NDT and E Int. 32 (3) (1999) 153–160. [28] A. Baltazar, C.D. Hernandez-Salazar, B. Manzanares-Martinez, Study of wave propagation in a multiwire cable to determine structural damage, NDT and E Int. 43 (8) (2010) 726–732. [29] P. Rizzo, F.L. Di Scalea, Wave propagation in multi-wire strands by waveletbased laser ultrasound, Exp. Mech. 44 (4) (2004) 407–415. [30] S. Al-Dossary, R. Hamzah, D. Mba, Observations of changes in acoustic emission waveform for varying seeded defect sizes in a rolling element bearing, Appl. Acoust. 70 (1) (2009) 58–81. [31] A.L. Gyekenyesi, G.N. Morscher, L.M. Cosgriff, In-situ monitoring of damage in SiC/SiC composites using acousto-ultrasonics, Compos. Part B: Eng. 37 (1) (2006) 47–53.
M. Kharrat, L. Gaillet / Ultrasonics 61 (2015) 52–61 [32] W. Yeih, R. Huang, Detection of the corrosion damage in reinforced concrete members by ultrasonic testing, Cem. Concr. Res. 28 (7) (1998) 1071–1083. [33] A. Shah, Y. Ribakov, C. Zhang, Efficiency and sensitivity of linear and non-linear ultrasonics to identifying micro and macro-scale defects in concrete, Mater. Des. 50 (2013) 905–916. [34] O. Gericke, Determination of the geometry of hidden defects by ultrasonic pulse analysis testing, J. Acoust. Soc. Am. 35 (3) (2005) 364–368. [35] G.J. Gruber, Defect identification and sizing by the ultrasonic satellite-pulse technique, J. Nondestr. Eval. 1 (4) (1980) 263–276. [36] H. Zejli, Détection et localisation par Emission Acoustique de fils rompus dans les ancrages des câbles d’ouvrages d’art, Ph.D. thesis, Université de Technologie de Compiègne, 2007.
61
[37] D. Alleyne, M. Lowe, P. Cawley, The reflection of guided waves from circumferential notches in pipes, J. Appl. Mech. 65 (3) (1998) 635–641. [38] A. Demma, P. Cawley, M. Lowe, A. Roosenbrand, B. Pavlakovic, The reflection of guided waves from notches in pipes: a guide for interpreting corrosion measurements, NDT and E Int. 37 (3) (2004) 167–180. [39] S. Rokhlin, D. Lewis, K. Graff, L. Adler, Real-time study of frequency dependence of attenuation and velocity of ultrasonic waves during the curing reaction of epoxy resin, J. Acoust. Soc. Am. 79 (6) (1986) 1786– 1793. [40] A.M. Lindrose, Ultrasonic wave and moduli changes in a curing epoxy resin, Exp. Mech. 18 (6) (1978) 227–232.
