19th World Conference on Non-Destructive Testing 2016

1 License: http://creativecommons.org/licenses/by-nd/3.0/

Fast Non-Contact Defect Imaging with Scanning Laser Source Technique

Takahiro HAYASHI 1, Misaki FUKUYAMA 1, Ken ISHIHARA 1 1 Kyoto University, Nishikyo-ku Kyoto, Japan

Contact e-mail: hayashi@kuaero.kyoto-u.ac.jp

Abstract. The authors have developed a defect imaging technique for plate-like structures such as pipes and bridges. The imaging technique uses a characteristic of a low frequency flexural mode of guided waves wherein the amplitude of waveforms varies with thickness in the vicinity of the laser spot when the flexural mode is generated with a pulsed laser and detected by receiving probes fixed on the object. Scanning the laser source with a galvano mirror scanner enabled us to measure waveforms at many points and obtain amplitude distributions corresponding to defect images. This study discusses topics related to developing a fast non-contact defect imaging system (elastic wave camera). Although replacing a receiving probe on a laser Doppler vibrometer yields full non-contact measurements in generating and receiving elastic waves, one encounters difficulties in measurements in large structures such as pipes and bridges due to the small signal to noise ratio. To overcome such difficulties, we generated burst waves by modulating high-repetition pulses emitted from a fiber laser. Taking frequency spectra of the burst signals improved the signal to noise ratio and gave clear defect images even with signals with small amplitude. Using the non-contact imaging system, the defect images in an aluminum alloy plate could be obtained within a few seconds in the experiments with a small number of laser emission points.

Introduction

Thin plates are used for various large structures such as pipes, bridges, wings and body of aircrafts, and automobile bodies. To maintain such plate-like structures, wide varieties of non-destructive evaluation (NDE) techniques have been developed. One of the most popular NDE techniques for large structures is visual inspection using one's own eyes as well as optical cameras and infrared cameras. Following the progress using cameras, strain distributions and temperature distributions of large areas have become measurable in real time. However, X-ray and ultrasonic (elastic) waves have to be incorporated to inspect inside plate-like structures for pipe thinning and delamination of a layered plate.

The authors have studied inspection techniques for plate-like structures using elastic waves [1, 2]. Impact hammer techniques and ultrasonic pulse echo methods have been widely used as inspection techniques using elastic waves. However, the conventional, less-efficient techniques require an enormous amount of time and money for inspecting the whole range of large structures. Recently, efficient inspection techniques using guided waves

More info about this article: http://ndt.net/?id=19323

2

have been developed [3-7]. However, the guided wave pulse echo inspection method sometimes lacks detectability due to the use of low frequency ultrasonic waves for long-range inspection [1-2].

This paper, therefore, introduces a defect imaging with scanning laser source (SLS) technique as a means of efficient, high-resolution, non-contact, and fast imaging [8-18]. First, the principle of the defect imaging is described. Then, a generation technique for narrowband tone-burst waves using a fiber laser is introduced, and defect images obtained with the non-contact fast measurement are shown.

1. Defect imaging with a scanning laser source technique

Elastic waves are generated with thermo-elastic effects or ablation when a laser pulse is emitted on material. Scanning the laser emission points with rotating mirrors enables elastic wave generation at many points. The measurement technique of the elastic waves generated with the scanning laser and received with devices at fixed points on the material, called scanning laser source (SLS) technique, was adopted for inspection and material evaluation. For example, Kromine et al. and Fomitchov et al. showed that flaws on material surfaces could be detected with high sensitivity using SLS [8-10]. Takatsubo et al. created an animation of elastic wave propagation from the waves detected with SLS and applied it to defect detection [11]. The authors proved that the amplitude distribution of waveforms measured with SLS roughly corresponds to the thickness distribution of the plate, and defect images can be obtained by SLS [12-18].

Fig. 1 is a schematic showing the principle of the defect imaging with SLS. An A0 mode of Lamb wave (flexural wave) is generated by laser pulses emitted onto the surface of a thin plate. When the low frequency components of the A0 mode below the cut-off frequency of the A1 mode are detected, smaller signals are obtained for laser emission on the thick intact area, while larger signals are detected for laser emission on the thin defected area. Using this characteristic of Lamb waves, signal distributions obtained with SLS correspond to thickness distributions and two-dimensional scanning yields a defect image in a plate.

The defect imaging with SLS has various advantages in practical use compared with conventional techniques. For example: (1) elastic waves can be generated with laser pulse emission with no difficulty; (2) fast defect imaging is possible because laser spots can be rastered very quickly using a galvano mirror scanner; (3) measurement is relatively easy because very low frequency A0 modes (a few kHz) are used in the defect imaging technique; and (4) defect images can be obtained with high resolution even in such a low-frequency range because the resolution of the defect imaging depends on laser spot size, and not on the wavelength used.

The authors took the opportunity to investigate applications to curved plates [14] and fully non-contact measurements [15]. This paper shows defect imaging with remote, non-contact, and fast measurements using a laser Doppler vibrometer (LDV) as a receiving device. Because the use of the LDV reduces the signal to noise ratio (SNR) significantly, elastic wave generation needs to be improved. Therefore, we developed a measurement system with a high-repetition fiber laser [17,18]. This paper discusses images of artificial defects with zigzag and circular patterns obtained with the non-contact fast measurement system.

3

Fig. 1. Schematic figure of the principle of defect imaging by SLS.

2. Narrowband tone-burst wave generation with a high-repetition fiber laser

Q-switched Nd:YAG and CO2 lasers have been used to generate elastic waves for NDE of solid materials. The pulse laser equipment generates broadband pulse waves in a material. Although broadband pulse waves are suitable for NDE due to their high resolution in time and space, broadband pulse waves cannot be easily separated from random electrical noise for low SNR. Moreover, excessive laser emission for improvement of SNR causes surface damage called laser spots.

The authors developed a measurement system using fiber laser equipment that generates narrowband tone-burst waves and enables defect imaging in low SNR [16-18]. Fig. 2 shows the principle of narrowband wave generation by a fiber laser. Fiber laser equipment can emit laser pulses with a high repetition rate on the order of kHz or MHz (B). Applying an external modulation signal as (C) to the high-repetition laser pulses results in the burst laser shots shown in (D). The burst laser shots controlled by the modulation signal generate elastic waves with the following frequency components: laser repetition rate and modulation signal. In our previous studies, we used low frequency narrowband tone-burst waves controlled by the modulation signal below 20 kHz [17, 18]. Even when the burst waves showed a small amplitude in the time domain, SNR was improved in the frequency domain. Moreover, because a fiber laser generally emits pulses with a small energy per shot and the laser spot moves slightly at every shot, we can avoid surface damage. [16].

Thick areaSmall signal

Thin areaLarge signal

Amplitude distribution= Thickness distribution

Laser = Elastic wave source

Thick areaSmall signal

4

Fig. 2. Principle of generation of narrow-band tone-burst wave by a fiber laser.

3. Experimental set-up and specimens used

Fig. 3 shows the experimental system used in this study. The galvano mirror controller outputs a TTL signal at specified angles (1), and the TTL signal triggers a record of waveforms in an AD converter and an output of the modulation signal in an FPGA (2). The square burst modulation signal, as shown in Fig. 2 (C), controls laser emission. When modulated laser pulses are emitted onto a specified point on the plate surface (3), elastic waves corresponding to the modulation signal are generated as shown in Fig. 2 (E), and waves travelling through various paths in the plate (4) are measured at the left edge with the LDV (5). After the measured waveforms are digitized in an AD converter (7), filtering, Fourier transforms, and imaging processes are carried out in a PC in nearly real time. Fig. 4 shows the test plates used in this study. Aluminum alloy plates measuring 500 mm ×

500 mm ×3 mm was cut half-way along its centerline. A laser emission area and a receiving point were located at the right and left regions of the plate, respectively. Direct paths from the laser emission area and the receiving point were divided by the thru notch, and only reflected waves from the plate edges and refracted waves were detected at the receiving point.

Using a similar test plate, the reference [17] proved that clearer images are obtained by frequency image averaging (FIA), in which defect images are obtained at multiple frequencies and the multiple images are averaged to reduce the resonant pattern in the plate. In this study, three defect images were obtained for tone-burst waves with three frequency components at 7, 9, and 11 kHz, and the averaged images were shown as defect images. Artificial defects of zigzag and circular patterns, as shown in Fig. 4, were engraved on the back surface of the laser emission area. This paper discusses the characteristics of the imaging technique through images of these defects.

Trigger signal (A)

Modulation signal

created by a FG (C)

Generated waves (E)

Possible Laser output (B)

Actual laser output (D)= (A)x(B)

540 kHz rep. rate

7,9,11 kHz

FPGA(A)

(C)

(D)

(E)

200 Hz

5

Fig. 3. Experimental system used in this study.

Fig. 4. Test plates used in this study.

4. Experimental results

Figs. 5 (a) and (b) are defect images obtained with FIA for a test plate with a zigzag pattern. Images (a) and (b) are defect images at 1-mm increments for a 50 mm × 80 mm range (the number of rastering points: 51×81 = 4131). Because the repetition rate of burst laser shots was 200 Hz, measurement times were about 22 s (= 4131/200). The images depict average values of three frequency peaks at 7, 9, and 11 kHz in the frequency spectrum of a waveform extracted at 10 ms intervals. Fig. 5 (a) shows resonant patterns in the test plate as spurious images. The FIA is an effective means to reduce such resonant patterns, as shown in reference [17].

To create defect images with higher speed, we need to increase the repetition frequency of the burst laser shots (namely, to increase scanning speed) or reduce the number of rastering points. This paper shows the results for reduced rastering points. Fig. 5 (b) is the

LDV controller

AD converter& FPGA

1.Trigger TTL2. Modulation signal3. Laser pulse train4. Elastic waves 5. Wave detection6. RF signal from LDV 7. Digitized RF signal data

Galvano mirrorcontroller

1.

2.

3.

4.

5.

7.PC 6.

Fiber laser unit

Laser emission area

50 mm x 80 mm for zigzag notch80 mm x 100 mm for circular holes

25

0 m

m

50

0 m

m

16 mm

t = 3 mm

20

0 m

m

Zigzag notch

Notch width: 2mmDepth: 1.5mm

40

mm

50

mm

Circular defects

Depth: 1.5 mm

500 mm

Aluminum alloy plate

Retroreflective tape(Laser spot for detection)

6

image at 3 mm increments for a 48 mm × 81 mm range (number of rastering points: 17×28 = 476). The measurement time was reduced to about 3 s (= 476/200). Although the image resolution decreased as expected, the zigzag pattern can still be visualized.

(a) 1 mm increment (b) 3 mm increment

for a 50 mm × 80 mm range for a 48 mm × 81 mm range Measurement time: 22 s 3 s

Fig. 5 Defect image for an artificial defect with zigzag pattern. Repetition rate: 200 Hz.

Next, images for circular defects are shown in Fig. 6. Fig. 6 (a) is an FIA image at 1

mm increments for the range of 50 mm × 100 mm. The number of rastering points is 5151 (= 51 × 101), and the repetition frequency is 200 Hz, which results in a measurement time of about 26 s (= 5151/200). Because the circular images resemble surrounding spurious images, defects cannot be as clearly recognized as compared to the zigzag pattern. Fig. 6 (b) is an FIA image at 3 mm increments for the range of 51 mm × 99 mm (the number of rastering points: 17 × 34 = 578), which results in a measurement time of about 3 s (= 578/200). Like in Fig. 5 (b), the resolution decreased. However, it became more difficult to recognize the circular defects because the defect images resemble the spurious images. Namely, it can be concluded that the defect images can be obtained with sufficient resolution in fast measurements by reducing the rastering points if the features of the expected defects are different from the spurious images caused by resonant patterns in a plate.

Because defect images could be obtained in about three seconds in Fig. 5 (b) and Fig. 6 (b) with the non-contact laser measurements, this imaging system will be used like an optical camera that can record images and movies of objects with non-contact measurements in nearly real time. The elastic camera can be applied in a wide variety of fields as a new easy-to-use technology that has a different operating principle from optical and infrared cameras.

7

(a) 1-mm increment (b) 3-mm increment

for a 50 mm × 100 mm range for a 51 mm × 99 mm range Measurement time: 26 s 3 s

Fig. 6 Defect image for artificial circular defects. Repetition rate: 200 Hz.

4. Conclusions

This paper introduced a non-contact defect imaging technique with a scanning laser source using LDV as a receiving device. To improve SNR, narrowband tone-burst waves were generated with a fiber laser. Fast imaging experiments were carried out for test plates with artificial defects with zigzag and circular patterns. Reducing rastering points enabled us to obtain defect images within a few seconds. However, it was difficult to recognize circular defects in a reduced resolution image because spurious images resemble the defect images. We can expect that this non-contact fast imaging technique will be applied to create a movie showing the state of a plate with elastic wave-like optical or infrared cameras.

This work was supported by JSPS KAKENHI Grant Number 26282094 and Chubu Electric Power Co., Inc.

References

[1] T. Hayashi, M. Murase, Defect imaging with guided waves in a pipe, J. Acoust. Soc. Am. 117 (2005) 2134. [2] T. Hayashi, M. Nagao, M. Murase, Defect imaging with guided waves in a long pipe, J. Solid Mech. Mater. Eng. 2 (2008) 888. [3] J. L. Rose, Ultrasonic waves in solid media, Cambridge University Press, 1999. [4] D. N. Alkyne, B. Pavlakovic, M. J. S. Lowe, P. Cawley, Rapid long-range inspection of chemical plant pipework using guided waves, Insight 43 (2001) 93. [5] P. J. Mudge, Field application of the Teletest (R) long-range ultrasonic testing technique, Insight 43 (2001) 74. [6] P. Cawley, M. J. S . Lowe, D. N. Alleyne, B. Pavlakovic, P. Wilcox, Practical long range guided wave testing: Applications to pipes and rails, Mater. Eval. 61 (2003) 66. [7] J. L. Rose, A baseline and vision of ultrasonic guided wave inspection potential, J. Press. Vessel Technol. 124 (2002) 273. [8] A. Kromine, P. Fomitchov, Sridhar Krishnaswamy, J. D. Achenbach, Laser ultrasonic detection of surface breaking discontinuities: scanning laser source technique, Mater. Eval. 58 (2) (2000) 173. [9] P. A. Fomitchov, A. K. Kromine, Y. Sohn, S. Krishnaswamy, J. D. Achenbach, Ultrasonic imaging of small surface-breaking defects using scanning laser source technique, in: D. Thompson, D. Chimenti (Eds.), Review of Progress in Quantitative Nondestructive Evaluation, American Institute of Physics (2002) 356.

8

[10] Y. Sohn, S. Krishnaswamy, Interaction of a scanning laser-generated ultrasonic line source with a surface-breaking flaw, J. Acoust. Soc. Am. 115 (1) (2004) 172. [11] J. Takatsubo, B. Wang, H. Tsuda, N. Tooyama, Generation laser scanning method for the visualization of ultrasounds propagating on a 3-D object with an arbitrary shape, J. Solid. Mech. Mater. Eng. 1 (12) (2007) 1405. [12] T. Hayashi, M. Murase, M. N. Salim, Rapid thickness measurements using guided waves from a scanning laser source, J. Acoust. Soc. Am. 126 (2009) 1101. [13] T. Hayashi, M. Murase, T. Kitayama, Frequency dependence of images in scanning laser source technique for a plate, Ultrasonics 52 (2012) 636. [14] M. N. Salim, T. Hayashi, M. Murase, T. Ito, S. Kamiya, Fast remaining thickness measurement using a laser source scanning technique, Mater. Trans. (2012) 1. [15] T. Hayashi, M. Murase, N. Ogura, T. Kitayama, Imaging defects in a plate with full non-contact scanning laser source technique, Mater. Trans. 55 (2014) 1045. [16] T. Hayashi, K. Yamaguchi, S. Biwa, Numerical analysis of narrow band ultrasonic wave generation with high repetition pulse laser and laser scanning, J. Phys. Conf. Ser. 520 (2014) 012024. [17] T. Hayashi, Imaging defects in a plate with complex geometries, Applied Physics Letters, accepted [18] T. Hayashi, K. Ishihara, Generation of narrowband elastic waves with a fiber laser and its application to the imaging of defects in a plate, Ultrasonics, under review.