ULTRASONIC DEFECT DETECTION IN BILLET USING TIME- OF-FLIGHT OF BOTTOM ECHO Ryusuke Miyamoto Graduate School of Systems and Information Engineering, University of Tsukuba, Tsukuba, Ibaraki 305-8573 Japan email: miyamoto@aclab.esys.tsukuba.ac.jp Koichi Mizutani, Tadashi Ebihara and Naoto Wakatsuki Faculty of Engineering, Information and Systems, University of Tsukuba, Tsukuba, Ibaraki 305-8573, Japan A billet, which is a semi finished products of steel products, sometimes contains defects. For the defect detection inside billets, pulse echo method, by which defects are detected using ultrasonic echoes from defects, is generally used in billet production lines. However, it is hard to detect defects in high-attenuation billets because echoes from defects becomes feeble owing to ultrasonic diffusion and scattering at grain boundaries as the distance between the transducer and the defect increases. In this study, defect detection in billet using time-of-flight (TOF) of bottom echo were numerically simulated as an initial demonstration. By using the bottom echo, sufficient intensity of the received signal can be obtained even if the billet is made of high-attenuation material. Although the bottom echo is used as an indicator of the bottom position in pulse echo method, the echo is used to detect defects in this paper. From the bottom echo, defects are detected from the TOF deviation caused by diffraction around the defects. As a result of simulations, it was found that a defect can successfully be detected from the TOF deviation of bottom echo. 1. Introduction In our modern daily lives, steel is commonly used in various products such as automobiles, building products, and electric appliances. Steel products are made of billets that is manufactured by continuous casting [1]. In the casting process, defects are caused such as blowholes, porosities, and thin fractures [2]. In this study, we focused on blowholes and porosities, which is caused by inclusions of gas bubbles such as CO2 and H2, because these defects have larger effect on the final products than thin fractures. To detect these defects, ultrasonic inspection is studied and used as an effective nondestructive inspection method. As one of the ultrasonic inspection methods, pulse echo method is commonly used in steel production line [3]. Although this method is simple and fast, the echo from a defect which is small, slanted or in a deep position becomes feeble [4]. The echo may not be received because of its diffusion and grain boundary scattering as the grain size in billets becomes large. As an alternative inspection method, the authors have proposed ultrasonic computerized tomography (CT) using time-of-flight (TOF) and linear scanning using TOF profile, as shown in Fig. 1 [4,5,6,7]. These methods use transmission method and detect defects inside billets by deviation of TOF ( ) of transmitted longitudinal wave caused by diffraction at defects. caused by the defects was observed and the defects were visualized experimentally [4,5]. By employing transmission method, sufficient intensity of the received signal can be obtained regardless of position of defects, and billets with large grain size are effectively inspected [4]. Ultrasonic CT using TOF visualizes defects in billets, even though CT requires a huge number of measurement paths as shown in Fig. 1(a) and it takes a long measurement time [4,5,6]. Linear scanning using TOF profile reduces 1
Figure 1: Ultrasonic transmission methods using TOF. (a) CT using TOF and (b) linear scanning using TOF profile [7]. Figure 2: Outline of defect detection. (a) measurements on cross sections, measurement and reference planes, (b) difference in received waveforms between measurement and reference signals, and (c) linear scanning and TOF profile of bottom echo. measurement paths as many as that of pulse echo method as shown in Fig. 1(b) [7]. However, transmission method complicates measurement equipment because it requires both transmitter and receiver although the pulse echo method only requires single transducer. Therefore, we propose a defect detection method for billet inspection using TOF of bottom echo which can be performed by single transducer. Although bottom echo was used as an indicator of the bottom position in the conventional pulse echo method, it is expected that the bottom echo also has information of defects as well as transmission methods with TOF. The proposed method can be used for high attenuation billet which has large grain size compared with pulse echo method because intensity of bottom echo is larger than echo from defect. In this study, a defect detection in billet by TOF of bottom echo is numerically simulated, as an initial demonstration. The validity of the proposed method is evaluated by wave propagation simulation placing a defect at the center of a cross section and shifted slightly from the center. 2 ICSV23, Athens (Greece), 10-14 July 2016
2. Principle of defect detection Figure 2 shows scheme of defect detection by the proposed method. As shown in Fig. 2(a), a cross section of a billet is measured by linear scanning of a transducer. An ultrasonic signal is projected to a billet from a transducer and echoes are received by the same transducer. In this method, defect is detected by deviation of TOF of bottom echo owing to diffraction around the defect as shown in Fig. 2(a) below. If there is a defect on the ultrasonic propagation path, TOF of bottom echo deviates by. This deviation of TOF appears as a time shift between the received bottom echo signals r(t) and m(t). r(t) is measured at reference plane which contains no defect and m(t) is measured at measurement plane which may contain defects as shown in Fig. 2(b). To eliminate the effects of input signal and echo from defects, r(t) and m(t) are set to be only bottom echo by calculating arrival time of bottom echo from size of billet beforehand. If there are no defects at the measurement plane, r(t) and m(t) are same and becomes zero. can be obtained by calculating the time shift where the cross-correlation function between r(t) and m(t) takes maximum. Although the velocity difference could be caused by stress or transducer coupling, the effects on the defect detection can be suppressed by using reference plane when TOF is measured. In our previous study, deviation of TOF was obtained experimentally by transmission CT method. In the experiments, transducer coupling was not changed because the transducer and the billet were coupled in the water and do not directly touch each other. As shown in Fig. 2(c), a cross section of a billet is measured by linear scanning of a transducer, and TOF profile, which is relationship between the transducer position X and TOF deviation is obtained. From this TOF profile, defects can be detected. 3. Numerical simulation 3.1 Simulation condition To simulate wave propagation for defect detection by the proposed method, two-dimensional elastic finite-difference time-domain (FDTD) method was employed [8,9]. In the simulation, isotropic elastic material was assumed. Figure 3 shows the simulation condition. Tested billet was assumed to be steel which has cross section of 100 100 mm 2 with a mesh size of 0.1 mm, a density was 7,700 kg/m 3, and a velocities of longitudinal wave and shear wave were 5,950 and 3,240 (m/s), respectively. The surface and a defect of a billet was assumed to be a free boundary, in which stress is zero. The time step of the simulation was 1.12 ns. The input signal is two up-chirp signals using a Hann window, with frequency is 0.5-1.5 MHz with duration of 10 s. was determined by the peak position in the cross-correlation function between bottom echoes of received signals on the reference plane r(t) and measurement plane m(t). For the calculation of cross-correlation function considering only bottom echo, received signal between 33 and 45 s were used because propagation Figure 3: Simulation condition. ICSV23, Athens (Greece), 10-14 July 2016 3
time of bottom echo is about 33.6 s and duration of input signal is 10 s. The diameter of a defect was 2 mm and the defect was located at the center of a cross-section (i) or the position shifted by 25 mm from the center to the scanning direction of the transducer (ii) as shown in Fig. 3. The transducer whose aperture is 6 mm was scanned linearly in 0.5 mm steps. TOF profile was obtained as relationship between transducer position X and deviation of TOF. 3.2 Results and discussions Figure 4 shows the results of simulations. The left side of the figure is TOF profile when a defect was at (i) as shown in Fig. 3. increases when the transducer position X is close to the defect position. decreases when X is slightly different from the defect position. One of the reasons considered is an interference between the direct wave and the reflected wave from the defect. This tendency of deviation of can also be observed in the transmission method using TOF [7]. The right side of Fig. 4 is TOF profile when a defect was at (ii). increases when the transducer position X is close to the defect position in the same way as the left side of Fig. 4. However, the shape of the TOF profile when a defect was at (ii) is slightly different from that when a defect was at (ii). This is thought to be caused by reflected wave from the sides of a billet. These results suggest that a defect can be detected and position of the defect in scanning direction can be roughly estimated by the proposed method. Figure 4: TOF profiles. Figure 5: Received signals at X = 50. 4 ICSV23, Athens (Greece), 10-14 July 2016
Figure 5 shows the received signals at X is 50. Signals surrounded by a broken line is used for calculation of cross correlation. The amplitudes were normalized by that of maximum of the bottom echo at the reference plane. In Fig. 5, the amplitude of the bottom echo is larger than that of the echo from a defect. This feature suggests that the proposed method is more robust than the pulse echo method. Although high frequency signals which have small wavelength is required for receiving the echo from a defect in the pulse echo method, low frequency signals can be used in the proposed method which requires only the bottom echo. This tendency suggest that the proposed method can be efficient for inspection in a high attenuation billet which has large grain size. As seen above, a defect in billets can be detected by the proposed method which uses TOF of the bottom echo. The position of the defect in the scanning direction also can be estimated by the method. 4. Conclusions In this paper, a defect detection in billets by TOF of bottom echo was proposed and numerically simulated, as an initial demonstration. As a result, it was found that a defect can be detected and the position of defect in the scanning direction can be estimated by the proposed method. As future work, position and size of defect and noise in received signals should be considered. An experimental verification of the defect detection is also planned. REFERENCES 1 Campbell, F. C. Ed., Elements of Metallurgy and Engineering Alloys, ASM International, Materials Park, OH (2008). 2 Karamis, M. B. and Nair, F. Effects of reinforcement particle size in MMCs on extrusion die wear, Wear 265, 1741-1750 (2008). 3 Abe, M., Fujioka, T. and Nagata, Y. Location of a defect in a concrete block by a non-destructive technique, Acoustical Science and Technology 23, 308-312 (2002). 4 Norose, Y., Mizutani, K. and Wakatsuki, N. Application of ultrasonic computerized tomography using time-of-flight measured by transmission method to nondestructive inspection for high-attenuation billets, Japanese Journal of Applied Physics 53,07KC19 (2014). 5 Mitsui, H., Mizutani, K., Wakatsuki, N. and Norose, Y. Artifact Reduction in Tomographic Images for Nondestructive Testing of Square Billets Using Ultrasonic Computerized Tomography, Japanese Journal of Applied Physics 50,116601 (2011). 6 Kakuma, K., Norose, Y., Mizutani, K. and Wakatsuki, N. Interval of Observation Plane in Visualization of Region near Defects in Billets Using Ultrasonic Computerized Tomography Method, Japanese Journal of Applied Physics 52,07HC10 (2013). 7 Miyamoto, R., Mizutani, K. Ebihara, T. and Wakatsuki, N. Defect detection and size estimation in billet from profile of time-of-flight using ultrasonic transmission method with linear scanning, Japanese Journal of Applied Physics 54,07HC11 (2015). 8 Virieux, J. P-SV wave propagation in heterogeneous media; velocity-stress finite-difference method, Geophysics 51, 889-901 (1986). 9 Sato, M. Finite-Difference Time-Domain Numerical Analysis of Elastic Wave Fields Using both Elastic and Velocity Potential Variables, Japanese Journal of Applied Physics 45, 4453 (2006). ICSV23, Athens (Greece), 10-14 July 2016 5