To accurately measure crack lengths, we developed a real-time surface imaging method (SAW PA) combining an ultrasonic phased array (PA) with a surface acoustic wave (SAW). SAW PA using a Rayleigh wave with a high sensitivity to surface defects was implemented for contact testing using a wedge with the third critical angle that allows the Rayleigh wave to be generated. Here, to realize high sensitivity imaging, SAW PA was optimized in terms of the wedge and the imaging area. The improved SAW PA was experimentally demonstrated using a fatigue crack specimen made of an aluminum alloy. For further verification in more realistic specimens, SAW PA was applied to stainless-steel specimens with a fatigue crack and stress corrosion cracks (SCCs). The fatigue crack was visualized with a high signal-to-noise ratio (SNR) and its length was measured with a high accuracy of better than 1 mm. The SCCs generated in the heat-affected zones (HAZs) of a weld were successfully visualized with a satisfactory SNR, although responses at coarse grains appeared throughout the imaging area. The SCC lengths were accurately measured. The imaging results also precisely showed complicated distributions of SCCs, which were in excellent agreement with the optically observed distributions.

The accurate nondestructive measurement of both crack depth and crack length is essential for ensuring the safety and reliability of aging power plants. Here the crack depth is the size of a crack in the thickness direction of structures and the crack length is its size in the direction perpendicular to the crack depth. Thus far, measurement of the crack depth by ultrasonic testing (UT) with bulk waves has been widely carried out since ultrasound is scattered at the crack tip in the depth direction,1–13 where phased array with bulk waves has been intensively studied recently14–17 and most UT has been applied from the side opposite to the crack opening. On the other hand, accurate measurement of the crack length is not easy since the edge part of a crack in the length direction is generally shallow, resulting in the low sensitivity in UT with bulk waves.

To overcome this difficulty, the use of surface acoustic waves (SAWs) from the crack-opening side is promising. Specifically, a Rayleigh wave,18 in which most of the energy is within a wavelength in the depth direction, is highly sensitive even to shallow parts of cracks, such as their edges. To attain both a high spatial resolution and high sensitivity, the focusing of the Rayleigh wave is also useful. A focused contact transducer for SAWs has been discussed.19 Because a focused transducer has a fixed focal point that is determined by the curved geometry of the transducer, precise mechanical scanning of the transducer is required to measure crack lengths. On the other hand, the combination of an array transducer and a Rayleigh wave has been proposed.20,21 The array transducer was positioned on the top of a wedge whose angle is equal to the third critical angle that allows the Rayleigh wave to be generated on the surface of the specimen. A surface defect was detected with high sensitivity based on a time-reversal process. Furthermore, by exciting each element in the same configuration while following a delay law, the focusing of a Rayleigh wave at a focal point was demonstrated.22 However, imaging by electronic scanning has yet to be realized for practical use.

To realize such imaging, we previously proposed a real-time surface imaging method, namely, a surface-acoustic-wave phased array (SAW PA) method.23,24 SAW PA was first implemented for contact testing, where a wedge with an array transducer was placed in contact with a specimen through a liquid couplant.9 However, the sensitivity was low. Therefore, we proposed SAW PA for water immersion testing.10 We demonstrated that it can visualize cracks with high sensitivity and a high resolution. However, the in situ application of water immersion testing to large structures has been very limited.

Here, we markedly improved SAW PA for contact testing by optimizing the wedge and the imaging area. To demonstrate the high sensitivity in SAW PA imaging, the SAW PA method was applied to specimens with fatigue cracks and stress corrosion cracks (SCCs).

A schematic of SAW PA is shown in Fig. 1(a). A linear array transducer is placed on a wedge with the third critical angle for the generation of the Rayleigh wave. A longitudinal wave emitted from the array transducer is converted into a Rayleigh wave at the interface between the wedge and specimen. Strictly speaking, leaky Rayleigh and Rayleigh waves propagate beneath and outside the wedge, respectively. By exciting each element of the array transducer in accordance with a delay law,24 a Rayleigh wave is focused at an arbitrary point on the surface of the specimen. When a surface defect is in the vicinity of the focal point, the Rayleigh wave is scattered at the defect, and thereafter, the scattered waves are received by the array transducer through mode conversion from the Rayleigh wave into a longitudinal wave at the interface. Subsequently, the delay-and-sum processing of the received waves is carried out to extract only the waves scattered from the focal point. By repeating this process for multiple focal points by electronic scanning, a surface image that corresponds to a C scan with information on the defects in the vicinity of the surface is obtained in real time.

FIG. 1.

Schematic illustration of SAW PA. (a) Schematic cross-sectional view of SAW PA, (b) a bird’s-eye view of the use of SAW PA for imaging beneath a wedge.

FIG. 1.

Schematic illustration of SAW PA. (a) Schematic cross-sectional view of SAW PA, (b) a bird’s-eye view of the use of SAW PA for imaging beneath a wedge.

Close modal

In this study, we optimized the wedge, which is the most critical part in the measurement, using a Rayleigh wave. In the previous SAW PA for contact testing,23 acrylic resin was selected as the wedge material among several kinds of materials25–29 because of its low cost and high workability. However, it is difficult to use a frequency of, for example, 5 MHz, which has been frequently used in UT, because acrylic resin is a highly attenuative material. In this regard, polystyrene and polyimide are more suitable than acrylic resin. Another factor is the propagation distances of the longitudinal wave in the wedge and the leaky Rayleigh wave beneath the wedge. To reduce these lengths, the optimization of the wedge shape is required. The wedge shape is determined by the third critical angle θR for the generation of the Rayleigh wave, which is calculated from Snell’s law:

(1)

where VW and VR are the velocities of a longitudinal wave in the wedge and a Rayleigh wave in the specimen, respectively. As θR decreases, the propagation distances of both a longitudinal wave in the wedge and a leaky Rayleigh wave beneath the wedge decrease. The leaky Rayleigh wave emits longitudinal waves into the wedge. The attenuation coefficient αL of the leaky Rayleigh wave due to the leaky loss is given by30 

(2)

where f is the SAW frequency and ρW and ρS are the densities of the wedge and specimen, respectively. Equation (2) shows that αL increases as ω increases. Therefore, reducing the propagation distance of the leaky Rayleigh wave may allow the use of a high frequency. To reduce the propagation distance, we must select the wedge material to obtain a small θR. From Eq. (1), θR for a polystyrene wedge is smaller than that for a polyimide wedge because VW (2430 m/s) in polystyrene is lower than that (2710 m/s) in polyimide. Thus, polystyrene was selected as a promising wedge material.

In addition, we considered the optimal imaging area for SAW PA. The outside of the wedge was previously used as the imaging area in SAW PA for contact testing. This is reasonable because the main role of the wedge is to convert a longitudinal wave into a Rayleigh wave, and the Rayleigh wave can propagate a long distance away from the wedge. This usage is common also for the UT with a Rayleigh wave employing a monolithic transducer and a wedge.18 However, the Rayleigh wave undergoes transmission losses at the edge of the wedge when it propagates out of the wedge upon transmission and enters the wedge upon reception. Note that this loss is significant at a high frequency because a Rayleigh wave with a high frequency is more sensitive to a change in the upper boundary condition than one with a low frequency. To solve this problem, we propose the use of an imaging area beneath the wedge (Fig. 1(b)) to avoid the transmission losses. This can also reduce the propagation distance of a leaky Rayleigh wave beneath the wedge.

In our experiment, we fabricated a wedge made of polystyrene based on the above considerations. To show the improvement in imaging using the wedge, SAW PA using the polystyrene wedge was compared with the previous SAW PA using an acrylic resin wedge for the imaging of a fatigue crack specimen.3 The experimental configuration is illustrated in Fig. 2(a). The PZT array transducer (manufactured by Imasonic) used had 32 elements with 0.5 mm pitch and a center frequency of 5 MHz. The array transducer was driven by phased array hardware (produced by Krautkramer). The excitation voltage was a pulse wave with a voltage of 100 V. The sampling rate was 50 MS/s. Delay-and-sum processing was carried out with steps of 2 mm and 1 . The same imaging conditions were also used in the later experiments.

FIG. 2.

Experimental configurations and imaging results of a fatigue crack (A7075) obtained by SAW PA with different wedges and different imaging areas. A fatigue crack with 17 mm length was formed from a starting notch in an aluminum alloy specimen (A7075) by performing a three-point bending fatigue test. The fatigue conditions were a maximum stress intensity factor of 5.3 MPa•m1/2 and a minimum stress intensity factor of 4.3 MPa•m1/2. (a) Experimental configuration for imaging the area outside of the wedge, [(b) and (c)] SAW PA images obtained outside of an acrylic resin wedge and polystyrene wedge, respectively, using the configuration in (a). (d) Experimental configuration for imaging the area beneath the polystyrene wedge, (e) SAW PA image obtained beneath the polystyrene wedge using the configuration in (d).

FIG. 2.

Experimental configurations and imaging results of a fatigue crack (A7075) obtained by SAW PA with different wedges and different imaging areas. A fatigue crack with 17 mm length was formed from a starting notch in an aluminum alloy specimen (A7075) by performing a three-point bending fatigue test. The fatigue conditions were a maximum stress intensity factor of 5.3 MPa•m1/2 and a minimum stress intensity factor of 4.3 MPa•m1/2. (a) Experimental configuration for imaging the area outside of the wedge, [(b) and (c)] SAW PA images obtained outside of an acrylic resin wedge and polystyrene wedge, respectively, using the configuration in (a). (d) Experimental configuration for imaging the area beneath the polystyrene wedge, (e) SAW PA image obtained beneath the polystyrene wedge using the configuration in (d).

Close modal

Imaging results obtained under the same conditions except for the wedge are shown in Figs. 2(b) and 2(c). The crack was not visualized using the acrylic resin wedge (Fig. 2(b)), whereas it was clearly visualized using the polystyrene wedge (Fig. 2(c)). To quantify the improvement, the mean intensity in area A surrounded by a white square in Fig. 2(b) was compared with that in area A’ in Fig. 2(c). The mean intensity in the latter was increased by 17.4 dB, showing that the polystyrene wedge is useful for reducing the attenuation of the longitudinal wave in the wedge and the leaky Rayleigh wave beneath the wedge.

To demonstrate the imaging area beneath the wedge, the difference between the SAW PA imaging results obtained outside and beneath the polystyrene wedge was examined using the configurations in Figs. 2(a) and 2(d). The imaging results corresponding to Figs. 2(a) and 2(d) are shown in Figs. 2(c) and 2(e), respectively. In Fig. 2(c), the crack was visualized to the right of the edge of the wedge, which corresponds to the imaging area outside of the wedge. In Fig. 2(e), the crack was visualized to the left of the edge of the wedge, which corresponds to the imaging area beneath the wedge. Note that the intensity in the crack in Fig. 2(e) was higher than that in Fig. 2(c). To quantify the improvement, the mean intensity in area B surrounded by a white square in Fig. 2(c) was compared with that in area B’ in Fig. 2(e). The mean intensity in the latter was increased by 4.9 dB because the imaging area beneath the wedge avoided the transmission losses at the edge of the wedge and reduced the propagation distances as compared with those outside the wedge. Thus, a total increase in the crack response of more than 20 dB was achieved because of the optimization of the wedge and the imaging area.

For further demonstration in more realistic specimens used to simulate materials in nuclear power plants, SAW PA was applied to the two specimens shown in Fig. 3. For crack length measurement, we formed a fatigue crack with 20 mm length in an austenitic stainless steel (SUS316L) specimen with 50 mm width using a four-point bending fatigue test. Note that SUS316L has been used in the components of nuclear power plants. The fatigue conditions were a maximum load of 30 kN and a stress ratio of 0.1. The number of fatigue cycles was 105000. The shape of the specimen was convex to ensure the formation of a fatigue crack with 20 mm length (Fig. 3(a)). A starting notch was fabricated at the center on the top of the convex part. The fatigue crack initiated from the starting notch and propagated into the base portion of the convex part. Importantly, the crack length was limited to the width of the convex part. Therefore, by machining the convex part by electrical discharge machining (EDM), a specimen with a crack having the length of 20 mm was obtained as shown in Fig. 3(b).31 Since the length of crack was less than the width of the specimen, it was non-penetrating to the side wall of the specimen.

FIG. 3.

Crack specimens: (a) formation of a fatigue crack in a convex-shaped specimen made of an austenitic stainless steel (SUS316L) by a four-point bending fatigue test, (b) fabrication of fatigue crack with 20 mm length by cutting the convex part by electrical discharge machining and (c) SCC specimen made of an austenitic stainless steel (SUS304).

FIG. 3.

Crack specimens: (a) formation of a fatigue crack in a convex-shaped specimen made of an austenitic stainless steel (SUS316L) by a four-point bending fatigue test, (b) fabrication of fatigue crack with 20 mm length by cutting the convex part by electrical discharge machining and (c) SCC specimen made of an austenitic stainless steel (SUS304).

Close modal

The fatigue crack was imaged beneath the polystyrene wedge by SAW PA with the array transducer placed in turn at three positions A, B and C, as respectively illustrated in Figs. 4(a), 4(c) and 4(e), to visualize the entire fatigue crack. The imaging results are shown in Figs. 4(b), 4(d) and 4(f). The crack was explicitly visualized at each measurement position with a high signal-to-noise ratio (SNR). To measure the crack length, the obtained images were merged into a single image by taking into account the measurement positions. In the merged image (Fig. 4(g)), the crack length was measured to be 20 mm. This was in good agreement with the true crack length (Fig. 4(h)). This shows that SAW PA is useful for measuring the fatigue crack length with a high measurement accuracy of better than 1 mm.

FIG. 4.

Experimental configurations and corresponding images of the fatigue crack (SUS316L) obtained by SAW PA. (a) Schematic illustration of experimental configuration and (b) SAW PA image for position A. (c) Schematic illustration of experimental configuration and (d) SAW PA image for position B. (e) Schematic illustration of experimental configuration and (f) SAW PA image for position C. (g) Images obtained by merging (b), (d) and (f). (h) Photograph of the fatigue crack specimen. The dotted rectangle is the crack area.

FIG. 4.

Experimental configurations and corresponding images of the fatigue crack (SUS316L) obtained by SAW PA. (a) Schematic illustration of experimental configuration and (b) SAW PA image for position A. (c) Schematic illustration of experimental configuration and (d) SAW PA image for position B. (e) Schematic illustration of experimental configuration and (f) SAW PA image for position C. (g) Images obtained by merging (b), (d) and (f). (h) Photograph of the fatigue crack specimen. The dotted rectangle is the crack area.

Close modal

SAW PA was applied to an SCC specimen.32 The SCCs were formed in the heat-affected zones (HAZs) of a weld made of an austenitic stainless steel (SUS304) due to residual stress in high-temperature pressurized water at 280 °C, which is almost the same as the environment in actual nuclear reactors. As a consequence, two SCCs were introduced into both HAZs as illustrated in Fig. 3(c). Note that the specimen had coarse grains with a maximum size of approximately 150 μm. The array transducer was mechanically scanned with 5 mm steps at positions A and B shown in Fig. 5(a) to image the two SCCs. The images obtained at each position were merged into a single image.

FIG. 5.

Experimental configurations and imaging results of SCCs (SUS304) obtained by SAW PA. (a) Schematic illustration of experimental configuration. [(b) and (c)] SAW PA images obtained at positions A and B, respectively. The areas surrounded by white dotted rectangles show the branch parts of the SCC. (d) Photograph of the SCC specimen. The areas surrounded by white dotted curves show the SCC positions.

FIG. 5.

Experimental configurations and imaging results of SCCs (SUS304) obtained by SAW PA. (a) Schematic illustration of experimental configuration. [(b) and (c)] SAW PA images obtained at positions A and B, respectively. The areas surrounded by white dotted rectangles show the branch parts of the SCC. (d) Photograph of the SCC specimen. The areas surrounded by white dotted curves show the SCC positions.

Close modal

The images of the SCCs obtained at positions A and B are shown in Figs. 5(b) and 5(c), respectively. Two SCCs were clearly visualized with a satisfactory SNR at both positions, although responses at coarse grains appeared throughout the imaging area. At position A, the crack length was measured to be 30 mm in Fig. 5(b). At position B, the penetration of crack in the length direction was accurately measured in Fig. 5(c). These measured lengths were in good agreement with the true crack lengths obtained by optical observation in Fig. 5(d). Note that in contrast to the images of the fatigue crack in Fig. 4, Figs. 5(b) and 5(c) show the complicated distributions of SCCs. The distributions were in excellent agreement with the optically observed ones (Fig. 5(d)). Specifically, in addition to the overall geometries, the branches of the SCCs were precisely visualized in the areas surrounded by dotted rectangles in Figs. 5(b) and 5(c). These results show that SAW PA is very useful for accurately measuring crack lengths and for precisely visualizing complicated distributions of SCCs even in coarse-grained materials.

Optimized SAW PA was used as a linear ultrasonic method in this study because the crack opening displacement (COD) at the root of a crack is generally larger than that at the crack tip. On the other hand, the closure of the crack at the root due to the high-temperature environment, which can occur in thermal power plants has been explicitly observed in eddy current testing (ECT).32 In such a case, it is necessary to apply a nonlinear ultrasonic method to SAW PA,23,24 such as by using subharmonic waves3 and/or thermal stress.10 Furthermore, from a physical viewpoint, the comparison between UT and ECT is useful for clarifying the physical mechanism of crack closure.

The repeatability of the measurement by SAW PA is important for its practical application. We confirmed that the high repeatability can be obtained by keeping the coupling condition that the gap between the wedge for SAW PA and specimens is filled with a liquid couplant (glycerin). One of the advantages of using the polystyrene as a wedge material is that it can be visually confirmed that there is no air gap beneath the wedge because the polystyrene is transparent. Furthermore, the use of an imaging area beneath the wedge also improved the repeatability. When the outside of the wedge is used as an imaging area, the liquid couplant that overflows from the wedge changed the attenuation of the Rayleigh wave propagating through the edge of the wedge. On the other hand, the use of an imaging area beneath the wedge can avoid this issue. As a result, it turned out that the error was less than ±0.5 mm in the repeated crack length measurement in Fig. 4. Also note that the stable measurements were achieved during the mechanical scanning.

In nuclear power plants, the measurement of the crack length from the crack-opening side has mainly been carried out by penetration testing (PT) and ECT. However, these methods are time-consuming. On the other hand, SAW PA can provide imaging results in real time. Furthermore, the array transducers used in situ for crack depth measurement can also be used for SAW PA by preparing a wedge with the third critical angle for the generation of the Rayleigh wave and using the delay laws formulated for electronic scanning.24 This can markedly reduce the cost of introducing SAW PA to in situ inspection.

This work was partly supported by JSPS KAKENHI Grant Numbers 15H04139 and 26630340.

1.
J.
Blitz
and
G.
Simpson
,
Ultrasonic methods of non-destructive testing
(
Springer
,
London
,
1996
).
2.
D. E.
Bray
and
R. K.
Stanley
,
Nondestructive Evaluation: A Tool in Design, Manufacturing, and Service
(
CRC Press
,
New York
,
1997
).
3.
Y.
Ohara
,
T.
Mihara
,
R.
Sasaki
,
T.
Ogata
,
S.
Yamamoto
,
Y.
Kishimoto
, and
K.
Yamanaka
,
Appl. Phys. Lett.
90
,
011902
(
2007
).
4.
Y.
Ohara
,
S.
Yamamoto
,
T.
Mihara
, and
K.
Yamanaka
,
Jpn. J. Appl. Phys.
47
,
3908
(
2008
).
5.
Y.
Ohara
,
H.
Endo
,
T.
Mihara
, and
K.
Yamanaka
,
Jpn. J. Appl. Phys.
48
,
07GD01
(
2009
).
6.
Y.
Ohara
,
S.
Horinouchi
,
M.
Hashimoto
,
Y.
Shintaku
, and
K.
Yamanaka
,
Ultrasonics
51
,
661
(
2011
).
7.
S.
Horinouchi
,
M.
Ikeuchi
,
Y.
Shintaku
,
Y.
Ohara
, and
K.
Yamanaka
,
Jpn. J. Appl. Phys.
51
,
07GB15
(
2012
).
8.
Y.
Ohara
,
Y.
Shintaku
,
S.
Horinouchi
,
M.
Ikeuchi
, and
K.
Yamanaka
,
Jpn. J. Appl. Phys.
51
,
07GB18
(
2012
).
9.
M.
Ikeuchi
,
K.
Jinno
,
Y.
Ohara
, and
K.
Yamanaka
,
Jpn. J. Appl. Phys.
52
,
07HC08
(
2013
).
10.
Y.
Ohara
,
K.
Takahashi
,
S.
Murai
, and
K.
Yamanaka
,
Appl. Phys. Lett.
103
,
031917
(
2013
).
11.
J. N.
Potter
,
A. J.
Croxford
, and
P. D.
Wilcox
,
Phys. Rev. Lett.
113
,
144301
(
2014
).
12.
J.
Cheng
,
J. N.
Potter
,
A. J.
Croxford
, and
B. W.
Drinkwater
,
Smart Mater. Struct.
26
,
05506
(
2017
).
13.
S.
Haupert
,
G.
Renaud
, and
A.
Schumm
,
NDT & E Int.
87
,
1
(
2017
).
14.
T. L.
Szabo
,
Diagnostic Ultrasound Imaging: Inside Out
(
Academic
,
New York
,
2000
).
15.
C.
Holmes
,
B. W.
Drinkwater
, and
P. D.
Wilcox
,
NDT & E Int.
38
,
701
(
2005
).
16.
B. W.
Drinkwater
and
P. D.
Wilcox
,
NDT & E Int.
39
,
525
(
2006
).
17.
L. W.
Schmerr
, Jr.
,
Fundamentals of Ultrasonic Phased Arrays
(
London
,
Springer
,
2015
).
18.
I. A.
Victrov
,
Rayleigh and Lamb Waves
(
Plenum
,
New York
,
1967
), p.
1
.
19.
B. G.
Kim
,
J. O.
Lee
, and
S.
Lee
,
IEEE Trans. Ultrason. Ferroelectr. Freq. Control
40
,
162
(
1993
).
20.
R. K.
Ing
,
M.
Fink
, and
O.
Casula
,
Appl. Phys. Lett.
68
,
161
(
1996
).
21.
W. A. K.
Deutsch
,
A.
Cheng
, and
J. D.
Achenbach
,
Res. Nondestr. Eval.
9
,
81
(
1997
).
22.
W. A. K.
Deutsch
,
A.
Cheng
, and
J. D.
Achenbach
,
IEEE Trans. Ultrason. Ferroelectr. Freq. Control
46
,
333
(
1999
).
23.
A.
Ouchi
,
A.
Sugawara
,
Y.
Ohara
, and
K.
Yamanaka
,
Proc. 35th Symp. Ultrasonic Electronics
,
2014
, p.
259
.
24.
A.
Ouchi
,
A.
Sugawara
,
Y.
Ohara
, and
K.
Yamanaka
,
Jpn. J. Appl. Phys.
54
,
07HC05
(
2015
).
25.
J.-Y.
Kim
and
S. I.
Rokhlin
,
Int. J. Solids Struct.
39
,
1487
(
2002
).
26.
Z.
Ling
,
H.
Zhou
, and
H.
Zhang
,
IEEE Trans. Instrument. Meas.
58
,
1578
(
2009
).
27.
J.
Chang
,
K.
Ohira
,
O.
Takahashi
,
T.
Hayashi
,
M.
Murase
, and
K.
Kataoka
,
Proc. Symp. Ultras. Electron.
30
,
61
(
2009
).
28.
S. V.
Walker
,
J.-Y.
Kim
,
J.
Qu
, and
L. J.
Jacobs
,
NDT & E Int.
48
,
10
(
2012
).
29.
Z.
Ling
,
W.
Cai
,
C.
Li
, and
W.
Guo
,
Appl. Mech. Mater.
742
,
11
(
2015
).
30.
A. J.
Slobodnik
, Jr.
,
J. Appl. Phys.
43
,
2565
(
1972
).
31.
T.
Uchimoto
,
T.
Takagi
,
X.
Wu
,
H.
Feng
, and
R.
Urayama
,
Int. J. Appl. Electromagn. Mech.
52
,
677
(
2016
).
32.
A.
Sugawara
,
K.
Jinno
,
Y.
Ohara
, and
K.
Yamanaka
,
Jpn. J. Appl. Phys.
54
,
07HC08
(
2015
).