Calibration of acoustic emission sensors with laser generatedultrasonic

Size: px

Start display at page:

Download "Calibration of acoustic emission sensors with laser generatedultrasonic"

Marjorie Webb
5 years ago
Views:

1 J. Acoust. Soc. Jpn.(E) 13, 2 (1992) Calibration of acoustic emission sensors with laser generatedultrasonic wave Youichi Matsuda,*Hidetoshi Nakano,*Kazuo Muto,**and Satoshi Nagai* National Research Laboratory of Metrology, 1-1-4, Umezono, Tsukuba, Ibaraki, 305 Japan The Institution of Vocational Training, 1960, Aihara, Sagamihara, Kanagawa, 299 Japan (Received 30 May 1991) Laser-generated ultrasonic wave is employed as a standard acoustic emission (AE) source to calibrate sensors. Characteristics of ultrasonic wave concerning its genera tionefficiency and directivity were investigated. By covering a transfer block with water layer, a unipolar planar pulse with 80 ns duration was efficiently generated. The surface excitation was interrogated by an optical interferometer, and subsequently measured by AE sensors. Comparing these outputs, the sensitivity of AE sensors can be determined in the wide frequency range. Experimental results on five commercial sensors are demonstrated. Keywords:Laser-generated ultrasonic wave, Acoustic emission, Sensor, Calibration, Standard source PACS number: Ud, Le 1. INTRODUCTION Acoustic emission (AE) has been widely used to detect flaws in structural members and also to study fracture mechanisms of materials. Nowadays its frequency range covers from several tens khz to over 1 MHz in one of the current studies' where fast phenomena e.g. generation and propagation of microcrack are monitored. In these applications un calibratedae sensors are still employed. Thus, it results in the lack of quantitative interpretation and furthermore makes intercomparisons among differ entlaboratories difficult. In this respect, calibration of sensors in the wide range is one of key technolo giesin AE application. As a primary standard source, Breckenridge et al.2) used a fast step-force loading on a massive block by breaking a glass capillary on it. The obtained wave formcontains frequency components ranging from 100 khz to 1 MHz. Various methods are proposed as a secondary source, such as a helium gas jet,3) fracture of pencil leads,4) and spark plug.5) Though these techniques are easy to use, they are less repro ducibleand/or have the limited frequency range, as compared with the glass capillary method. Another approach is based on the reciprocity principle.6) It features no need to a standard source. Three sensors are calibrated in sets. Two of them are mounted at a time on a transfer block and elec tricaltransfer characteristic is measured, using one as a transmitter and the other as a receiver. Measure mentsare repeated for possible three pairs. It is reported that calibration results agreed well with those of the glass capillary method in the frequency range within 1 MHz.7) One disadvantage results from the fact that it gives only magnitude informa tionof the sensors without phase one. Recently laser-generated ultrasonic wave has been successfully applied to non-destructive testing and further characterization of materials.8) Hutchins et al.9) investigated waveforms generated by laser irradiation and showed that pulsed laser can produce 91

2 J. Acoust. Soc. Jpn. (E) 13, 2 (1992) different types of standard acoustic sources, which include a normal monopole force generating a longi tudinalwave. The obtained longitudinal pulse offers an advantage of short duration without any trailing edges and hence is suited as a broadband source. The pulse height increases as the increase of laser energy. High energies, however, cause surface damage of the block. Scruby et al.10) used a notched block and realized that a unipolar pulse in a thermo elasticregion could simulate the growth of a mode I crack in a half space. In the present paper we investigate acoustic be haviorof laser-generated ultrasonic wave, in par ticular,concerning its generation efficiency and directivity. We employ a flat transfer block with water coating and show that a monopole source can be obtained without causing any damage to the block. Calibrated results on five AE sensors are also demonstrated. The measured magnitude responses agreed quite well with those determined by the reci procitymethod. 2. EXPERIMENTAL PROCEDURE Figure 1 shows a schematic drawing of the appa ratus.a Q-switched Nd:YAG laser with 10ns duration is operated at the second harmonic wave lengthof 533 nm. A transfer block is a cylinder of an aluminium alloy with 240mm in diameter and 250 mm long. The surface is covered with water to in creaseof the generation efficiency of ultrasonic wave. The laser hits the center of the one surface of the aluminium block and then generates ultrasonic wave, which propagates inside and reaches to the other surface. The deformation due to elastic wave motion is measured by an optical interferometer. The interferometer is Michelson type and employs a 1 mw frequency-stabilized He-Ne laser as a light source and an avalanche diode as a detector. The rear surface measured is polished up to 0.2ƒÊm roughness to serve as a mirror in one arm of the interferometer. Ambient disturbances below 1 khz are compensated using a piezoelectric actuator attached to a reference mirror. After amplified by a low-noise wideband amplifier, the output of the photodiode is fed into a digitizing oscilloscope with a 10 bit resolution. The oscilloscope is triggered by sampling a portion of the incident beam. Sampling interval is varied between 2ns and 10ns to record waveforms precisely. A microcomputer logs digi tizeddata via IEEE-488 bus. The obtained signal is averaged 16 times and, if necessary, the moving average method is also applied in order to improve the signal-noise ratio. The absolute displacement can be easily calibrated by driving the piezo-actuator with the amplitude over a quarter wavelength and then by monitoring the output voltage of the inter ferometer.the sensitivity of the interferometer is found to be 2.0 ~108V/m and the displacement resolution attains about 0.05nm over 20MHz band width. Fig. 1 Ultrasonic wave source and the laser interferometer. 92

3 Y. MATSUDA et al.:calibration OF AE SENSORS In the process of sensor calibration the transient response on the rear surface is recorded by the opti calprobe. Then sensors are coupled to the block so that the sensor axis coincides with the light spot and thereby the sensor detects deformation signal. Both outputs are transformed into frequency spectra. The quotient of these two spectra gives the sensitivity of the sensor. 3. CHARACTERIZATION OF LASER-GENERATED ULTRASONIC WAVE 3.1 Generation Efficiency Figure 2 shows ultrasonic waveforms recorded at the epicentre. The specimen is a stainless steel disc with 20mm in diameter and 6mm thick. At low laser intensity, elastic wave is generated due to ther moelasticmechanism shown in Fig. 2(a). With increasing the intensity, ablation mechanism be comesdominant, which results in the rapid growth of unipolar longitudinal pulse and also eliminates shear wave, as seen from Fig. 2(b) to (e). At the ulti mateintensity shown in Fig. 2(c) the surface of the specimen was severely damaged. A method of liquid coating11) was adopted to ex citelongitudinal ultrasonic wave efficiently. Light oil or water applied on the surface evaporates by heat of the laser pulse, thus creating a normal drive force in favor of generating ultrasonic waves. In the present experiments the surface was covered with water layer. A water jacket is shown in Fig. 3. Laser beam is incident on water through a ground glass window. Since the glass scatters the beam randomly, light density becomes spatially uniform on the block surface over the range of 10mm in diameter. The thickness of water layer is chosen to be 8mm. The working time is 10,as, which is limited by the rever berationfrom the glass window. The thicker water layer becomes, the longer working time is obtained at the expense of the ultrasonic amplitude. The surface displacement measured with and without water is compared in Fig. 4. Enhancement in the generation efficiency is clearly observed. The dis placementhas increased approximately one order. For example, with surface treatment the threshold energy by which the block undergoes a tiny damage Fig. 3 Schematic view of the water jacket. The ground glass window is 30 mm in diameter. Fig. 2 Displacement waveforms at several intensities of laser beam. The first spike indicates the time when laser is triggered. The intensity increases gradually from (a) to (c). The letters L and S denote arrival of longitudinal and shear wave, respective ly. Fig. 4 Effect of surface treatment. œ: unmodified surface :water-covered sur face. 93

4 T. Acoust. Soc. Jpn. (E) 13, 2 (1992) is 70 mj. At this energy level, the pulse shows the maximum displacement 0.7 nm with 80 ns duration. In contrast the same energy can cause just 0.07 nm deformation in the case of free surface irradiation. The variations of the displacement and the pulse width remained on a shot-to-shot basis within 2 and 3 %, respectively. Referring to these results, we irradiated the water-covered block with a laser energy slightly lower than 70 mj. 3.2 Directivity Patterns Elastic wave was recorded on the rear surface at 3 different distances from the epicentre. These are shown in Fig. 5 Elastic wave shows a slight delay with increasing the distances. Thus, departure from the planar wavefront causes the lack of phase coher enceover sensor's face. We evaluate this effect by assuming that the wave is emanating from a virtual Fig. 5 Directivity patterns. The letter D denotes the distance from the epicentre to a probed point. source located at a depth b. Geometrical configura tionsare shown in Fig. 6. The average displacement D. sensed by a finite-sized sensor is represented as follows: (1) where w is the radius of the sensor, k and A(r) are the wavenumber and the amplitude of elastic wave. R= ãr2+b2 is distance from the source to a obser vationpoint. If the condition b âw is fulfilled, the amplitude A(r) keeps a constant value Ao and Eq. (1) is reduced to (2) From relationship between arrival-times and r, b is estimated to be about 400mm. Equation (2) indi catesthat the average displacement over 20mm diameter is only 2dB smaller at 7 MHz than that of a pure plane wave. Therefore the error due to sphericity is negligible and thereby we do not make any phase correction. The frequency spectra of the pulses are calculated with FFT algorithm. Obviously observed from Fig. 7 the generated pulses have enough frequency range to calibrate commercial AE, sensors. The frequency spectrum of the off-epi centralpulse coincides well with that of the epi centralone within 3dB up to 7MHz. Employing a numerical procedure of the inverse problem,12) the released force of the present monopole source is deduced to be approximately 45 N (peak height), which is one order larger compared to the breaking of a glass capillary.13) Fig. 6 Calculation of aperture effect. Fig. 7 Fourier transforms of the wave forms. The waveforms are detected at the distances of 0mm (solid line) and 10 mm (dotted line), respectively. 94

5 Y. MATSUDA et al.:calibration OF AE SENSORS 3.3 Calibration of AE Sensors Five sensors labelled from A to E were calibrated. We employed an AE-couplant (W-400, Nittetsu Technos), which showed the best reproducibility among several couplants we tried. The sensors were pressed to the block with contact pressure of 0.2 MPa. The transient response of the sensors to elastic waves are presented in Fig. 8, from which we can roughly classify the sensors A and B as a resonant type featuring high-q value, while the sensors C, D and E belong to broadband type with low-q value. The obtained frequency spectra were compared with Fig. 7. From these spectra, the sensitivity of the sensors were deduced as a function of frequency. As one example the magnitude and phase response of the sensor C versus the surface displacement is shown in Fig. 9. Magnitude is almost proportional to frequency up to 3MHz. Since this sensor is called as a velocity-type, response function to the surface velocity is also calculated, which are presented in Fig. 10(a). The velocity-sensitivity is about 40 db. For comparison, Fig. 10(b) shows a certificated value of a maker by the reciprocity method, where measurements are carried out with a steel block. A small difference between two results may be ex - Fig. 9 Displacement sensitivity of the sensor C as a function of frequency.(a) magnitude (b) phase. Fig. 8 Waveforms recorded by five AE sensors. The diameters of the sensors are 17.5mm (A), 12.8mm (B), 20mm (C and D) and 5mm (E). Fig. 10 Velocity sensitivity of the sensor C as a function of frequency.(a) present study (b) maker's result. 95

6 J. Acoust. Soc. Jpn. (E) 13, 2 (1992) plained by the discrepancy of acoustic impedances between steel and aluminium. The steel block trans mitsacoustic waves into sensors better than the aluminium block does. This leads to an apparent improvement of the sensitivity when the steel block is used. So it can be concluded that both results agree well quantitatively, at least below 6 MHz. The present sensor has two small resonant frequen ciesat 3.5 MHz and 6.5 MHz and around these frequencies phase value changes rapidly. 4. CONCLUSIONS It is shown that a pulsed laser can be used as a standard acoustic source for AE sensor calibration. Characteristics of ultrasonic wave are clarified. The generation efficiency is improved by surface treat ment,which provides a strong longitudinal pulse without any surface damage. The realized pulse is broadband, containing spectra up to 7 MHz and applicable to calibration of recently developed broadband AE sensors. Furthermore the present technique provides not only the magnitude but the phase information of sensor characteristics. The transient response of five AE sensors are presented. The measured amplitude of the sensor agreed well with that by the reciprocity method, thus demon stratingthe feasibility of our method. ACKNOWLEDGEMENTS The authors thank Dr. N. N. Hsu for kindly sup plyingthe program to solve numerically the inverse problem of acoustic sources. REFERENCES 1 ) M. Enoki and T. Kishi,"Theory and analysis of deformation moment tensor due to microcraking," Int. J. Fract. 38, (1988). 2 ) F. R. Breckenridge, C. E. Tschiegg, and M. Greenspan,"Acoustic emission:some applications of Lamb's problem," J. Acoust Soc. Am. 57, (1975). 3 ) S. McBride and T. Hutchison,"Helium gas jet spectral calibration of acoustic emission trans ducersand system," Can. J. Phys. 54, (1976). 4 ) N. N. Hsu, J. A. Simmons, and S. C. Hardy,"An approach to acoustic emission signal analysis-theory and experiment," Mater. Eval. 35, (1977). 5 ) C. Feng and R. M. Whittier,"Acoustic emission transducer calibration using transient surface waves and signal analysis," Dunegan/Endevco Tech. Rep. DE 79-1 (1979). 6 ) H. Hatano and E. Mori,"Acoustic emission trans ducerand its absolute calibration," J. Acoust. Soc. Am. 59, (1976). 7 ) F. R. Breckenridge, T. Watanabe, and H. Hatano, "Calibration of acoustic emission transducers: comparison of two methods," in Progress in Acoustic Emission 1, K. Yamaguchi et al., Eds.(The Japanese Society for Nondestructive Inspection, Tokyo, 1982), p ) For reviews see, for example, D. A. Hutchins, "Ultrasonic generation by pulsed lasers," in Physical Acoustics, Vol.18, P. W. Mason and R. N. Thurston, Eds. (Academic Press, New York, 1988), p ) D. A. Hutchins, R. J. Dewhurst, and S. B. Palmer, "Laser -generation as a standard acoustic source in metals," Apply. Phys. Lett. 38, (1981). 10 ) C. B. Scruby, H. N. G. Wadley, R. J. Dewhurst, D. A. Hutchins, and S. B. Palmer,"A laser generatedstandard acoustic emission source," Mater. Eval. 39, (1981). 11 ) D. A. Hutchins, R. J. Dewhurst, and S. B. Palmer, "Laser generated ultrasound at modified metal surfaces," Ultrasonics 19, (1981). 12 ) N. N. Hsu and D. G. Etizen,"Experimental deter minationof a point impact force-time function," Exp. Mech. (to be published). 13 ) F. R. Breckenridge, T. M. Proctor, N. N. Hsu, S. E. Fick, and D. G. Eitzen,"Transient sources for acoustic emission work," in Progress in Acoustic Emission 5, K. Yamaguchi et al., Eds.(The Japanese Society for Nondestructive Inspection, Tokyo, 1990), p

EWGAE 2010 Vienna, 8th to 10th September

EWGAE 2010 Vienna, 8th to 10th September Frequencies and Amplitudes of AE Signals in a Plate as a Function of Source Rise Time M. A. HAMSTAD University of Denver, Department of Mechanical and Materials