Ultrasonic inspection method for round bar material

Abstract

To provide an ultrasonic flaw detection method for a round bar material enabling a surface flaw and a surface layer flaw to be promptly and surely discriminated.SOLUTION: A lateral wave ultrasonic beam Ub is scanned in a direction along a peripheral surface of a round bar material M while transmitting the lateral wave ultrasonic beam Ub to a round bar material M.A reflected ultrasonic wave reflected and returned by a surface layer flaw M1 of the round bar material M or a surface layer flaw M2 generated within the round bar material M immediately below a surface is received and denoted as a flaw echo signal. A flaw echo image is created by using, as axes of a two-dimensional image, each scanning position of the oblique angle flaw detection ultrasonic wave (Ub) and a time from transmitting the lateral wave ultrasonic beam Ub at the scanning position until obtaining the flaw echo signal. A required number of flaw echo images are given to a neural network 3 as learning data for learning, and a new flaw echo image is given to the neural network 3 which has learnt the learning data, to discriminate whether a flaw corresponding to the flaw echo image is a surface flaw M1 or a surface layer flaw M2.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an ultrasonic flaw detection method for a round bar material, and more particularly to an ultrasonic flaw detection method capable of easily distinguishing and detecting surface flaws and surface layer flaws occurring inside the round bar material just below the surface. [Background technology]

[0002] When detecting flaws near the surface of a round bar material, the shear waves of the ultrasonic beam Ub for flaw detection are used with the refraction angle (sector scan angle) set to about 45 degrees, as shown in Figure 7. However, with this method, it is often difficult to distinguish between the flaw echo signals (Figures 8(1) and 8(2)) from the surface flaw M1 (Figure 7(1)) that opens on the surface of the round bar material M and the surface flaw M2 (Figure 7(2)) that occurs inside just below the surface of the round bar material M, because these signals are nearly the same size and appear at the same time.

[0003] Therefore, for example, Patent Document 1 discloses a flaw detection method in which different sector scan angles are set and when the magnitude of the flaw echo signal obtained at each sector scan angle exceeds a predetermined threshold, it is determined that there is a surface flaw or a subsurface flaw. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent Publication No. 2012-225887 Summary of the Invention [Problem to be solved by the invention]

[0005] However, the conventional method described above had the problem that it was time-consuming to carry out flaw detection because it was necessary to change the sector scan angle and repeat the same flaw detection, making it impossible to quickly detect flaws in the round bar material traveling through the line.

[0006] Therefore, the present invention is devised to solve such problems, and has an object to provide an ultrasonic flaw detection method for round bar material that can quickly and reliably distinguish between surface flaws and subsurface flaws. [Means for solving the problem]

[0007] In order to achieve the above object, in the first invention, an ultrasonic wave for angle beam inspection (Ub) is transmitted toward a round bar (M) while scanning the round bar (M) in a direction along the peripheral surface thereof, and reflected ultrasonic waves reflected and returned from a surface flaw (M1) of the round bar (M) or a surface flaw (M2) occurring inside the round bar (M) just below the surface are received as a flaw echo signal, and a flaw echo image is created from each scanning position of the ultrasonic wave for angle beam inspection (Ub) and the time change in the signal intensity of the flaw echo signal at the scanning position, and a required number of the flaw echo images are used as learning data images. and feeding the flaw echo image to a convolutional neural network (CNN) (3) for learning, and feeding a new flaw echo image to the trained convolutional neural network (3) to determine whether the flaw corresponding to the flaw echo image is the surface flaw (M1) or the surface flaw (M2). In the flaw echo image, a predetermined area (X) including a flaw echo signal is cut out for each of the surface flaw (M1) and the surface flaw (M2), and the predetermined area (X) is extracted at a scanning position that is less than all scanning positions in the predetermined area (X). Image Let be the learning data image.

[0008] In the first invention, a defect echo image is created, and a new defect echo image is provided to a neural network that has been trained in advance using the defect echo image to determine whether the defect corresponding to the defect image is a surface defect or a subsurface defect, making it possible to quickly and reliably distinguish between surface defects and subsurface defects. Furthermore, by appropriately changing the number of extracted flaw echo signals, the accuracy of discrimination and the discrimination speed, which are in a trade-off relationship with each other, can be appropriately selected according to the object to be detected.

[0009] In the second invention, a phased array probe (1) is used for transmitting and receiving the ultrasonic waves for angle beam inspection.

[0010] According to the second aspect of the present invention, the phased array probe can transmit and scan ultrasonic waves for angle beam inspection, and receive reflected ultrasonic waves in a simple and compact manner.

[0013] The third invention In the flaw echo image, a portion where the signal intensity of the flaw echo signal is high is displayed with high brightness, and a portion where the signal intensity of the flaw echo signal is low is displayed with low brightness.

[0014] The fourth invention In the flaw echo image, a portion where the signal intensity of the flaw echo signal is high is displayed with low brightness, and a portion where the signal intensity of the flaw echo signal is low is displayed with high brightness.

[0015] The fifth invention In the above, the round bar material (M) is a black-skin material.

[0016] The reference characters in parentheses indicate, for reference, the correspondence with specific means described in the embodiments described later. Effect of the Invention

[0017] As described above, the ultrasonic inspection method for round bar material of the present invention can quickly and reliably distinguish between surface defects and surface layer defects. More specifically, the method can be applied to round bar material that has been through a rolling process, a heat treatment process, and a bending straightening process, and whose surface is covered with black rust, which is a metal oxide. Since black skin material has black rust on the outermost surface layer, it is particularly difficult to distinguish between surface defects and surface layer defects, but by applying the method of the present invention, it becomes possible to quickly and reliably distinguish between surface defects and surface layer defects. [Brief description of the drawings]

[0018] [Figure 1] FIG. 1 is a diagram showing the configuration of an apparatus for carrying out the method of the present invention. [Diagram 2] FIG. 13 is a diagram showing an example of a flaw echo image. [Diagram 3] FIG. 2 is a diagram showing an image obtained by cutting out a predetermined region including a flaw echo signal from a flaw echo image. [Figure 4]FIG. 1 is a diagram showing a schematic configuration of a convolutional neural network. [Diagram 5] FIG. 13 is a diagram showing another example of a flaw echo image. [Figure 6] FIG. 13 is a diagram showing yet another example of a flaw echo image. [Figure 7] FIG. 1 is a schematic cross-sectional view showing a conventional flaw detection method. [Figure 8] FIG. 1 is a diagram showing a flaw echo signal in a conventional flaw detection method. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] It should be noted that the embodiment described below is merely an example, and various design improvements made by those skilled in the art without departing from the gist of the present invention are also included in the scope of the present invention.

[0020] FIG. 1 shows the configuration of an apparatus for carrying out the ultrasonic flaw detection method of the present invention. In FIG. 1, a phased array probe 1 is provided facing the outer circumferential surface of a metal round bar M. In the phased array probe 1, a number of ultrasonic transducers (not shown) are adjacently arranged in a known structure to form a transmission / reception surface 1a that curves along the outer circumferential surface of the round bar M. In this embodiment, a shear wave ultrasonic beam Ub, which is an ultrasonic beam for angle flaw detection, is generated by vibrating a predetermined number of adjacent ultrasonic transducers with an excitation signal having a predetermined time difference output from a control device 2 having a built-in computer, and has a sector scan angle of about 45 degrees and converges on the surface of the round bar and the surface layer immediately below it. Then, by sequentially moving the range of the predetermined number of ultrasonic transducers to be vibrated, the shear wave ultrasonic beam Ub is scanned in an angle range D of about 90 degrees, and the reflected ultrasonic waves from the surface flaw M1 and the surface flaw M2 in this range are received again by the phased array probe 1 and output to the control device 2 as a flaw echo signal.

[0021] In addition, when detecting surface defects M1 and surface defects M2 around the entire circumference of the round bar M, multiple (four) phased array probes 1 of similar configuration are provided along the entire circumference of the round bar M, or the round bar M is rotated or the phased array probe 1 is rotated around the round bar M.

[0022] Fig. 2 shows a flaw echo image (B-scope) drawn in the control device 2 by scanning with the ultrasonic beam Ub, with the horizontal axis representing the scanning position and the vertical axis representing the time change in the signal intensity of the flaw echo signal. Here, the white line portion in the area surrounded by the dashed line in Fig. 2 corresponds to the flaw echo signal. Note that the flaw echo image in this case was obtained at 97 lines (scanning positions) / 45 degree scanning position intervals.

[0023] Here, in the flaw echo signals shown in Figures 8(1) and (2) above, the reflection intensity waveforms are converted into absolute values, so they do not represent both positive and negative waveforms, and the resolution is low, which affects the accuracy of flaw determination.In contrast, the flaw echo image (B-scope) shown in Figure 2 is not converted into absolute values, and both positive and negative flaw echo signals are reflected in the signal intensity, so the resolution is twice as high as that of the flaw echo signals in Figures 8(1) and (2), making it clear that it has high resolution.

[0024] Furthermore, the flaw echo image has a high resolution because the parts of the flaw echo signal with a high signal strength are displayed at a high brightness and the parts of the flaw echo signal with a low signal strength are displayed at a low brightness, and thus can be suitably used as a learning data image, which will be described later. Note that the flaw echo image can also have a high resolution even if the parts of the flaw echo signal with a high signal strength are displayed at a low brightness and the parts of the flaw echo signal with a low signal strength are displayed at a high brightness, and thus can be suitably used as a learning data image, as will be described later.

[0025] Of the predetermined number of obtained flaw echo images, a predetermined region X (Fig. 2) containing a flaw echo signal was cut out for the surface flaw M1 and the surface flaw M2 as shown in Fig. 3(1) and (2), respectively, and an appropriate number of learning data images and test data images were allocated for each. As an example, 309 learning data images and 1236 test data images were obtained for the surface flaw M1, and 260 learning data images and 1040 test data images were obtained for the surface flaw M2.

[0026] Then, training data images (309 images + 260 images) were given to a convolutional neural network (CNN) of a known configuration having an appropriate number of convolution layers 31, pooling layers 32, and fully connected layers 33 as shown in Fig. 4, which was constructed in the control device 2, to learn the image data, and then test data images (1236 images + 1040 images) were given to obtain a test accuracy rate. This is shown in Table 1 (1).

[0027] As shown in Table 1(1), the test accuracy rate for both surface defects M1 and subsurface defects M2 was 100%, which is an extremely good result. However, the data transfer speed, which is a problem when considering use on a production line, is relatively slow, and the processing time for generating the learning model is also relatively long. Note that the data transfer speed ratio and the processing time ratio for generating the learning model in Table 1 are ratios when those in Table 1(3) below are set to 1.0.

[0028] Therefore, the above-mentioned predetermined area X (FIG. 2) including the flaw echo signal is extracted only at the scanning position of three lines. Image (Fig. 5) was obtained, from which 372 training data images and 1125 test data images were obtained for surface flaw M1, and 299 training data images and 900 test data images were obtained for superficial flaw M2. Next, the above training data images (372 + 299) were fed to the CNN for learning, and then the test data images (1125 + 900) were fed to obtain the test accuracy rate. This is shown in Table 1 (2). According to this, the test accuracy rate for surface flaw M1 was 100%, and the test accuracy rate for superficial flaw M2 was 99.4%. Both the data transfer speed and the processing time for generating the learning model have been greatly improved.

[0029] Furthermore, the above-mentioned predetermined area X (FIG. 2) including the flaw echo signal is extracted only at the scanning position of one line. Image(Fig. 6), 298 training data images and 1203 test data images were obtained for surface flaw M1, and 246 training data images and 992 test data images were obtained for superficial flaw M2. Next, the above training data images (298 + 246) were fed to the CNN for learning, and then the test data images (1203 + 992) were fed to obtain the test accuracy rate. This is shown in Table 1 (3). According to this, the test accuracy rate for surface flaw M1 was 98.3%, and the test accuracy rate for superficial flaw M2 was 98.0%. Both the data transfer speed and the processing time for generating the learning model have been further improved.

[0030] In this way, increasing the number of defect echo signals extracted to obtain the defect echo image to be provided to the CNN (increasing the resolution of the defect echo image) improves the CNN test accuracy rate (judgment result); however, on the other hand, the data transfer speed is slow and the processing time for generating the learning model is long, so it is best to optimally select the number of defect echo signals extracted to obtain the defect echo image to be provided to the CCN depending on the application.

[0031] [Table 1] [Explanation of symbols]

[0032] 1...phased array probe, 2...control device, 3...convolutional neural network, M...round bar material, M1...surface defect, M2...surface defect, Ub...shear wave ultrasonic beam (ultrasonic for angle beam inspection).

Claims

1. While transmitting ultrasonic waves for angle beam inspection toward a round bar material, the ultrasonic waves are scanned in a direction along the circumferential surface of the round bar material, and reflected ultrasonic waves reflected and returned from surface defects of the round bar material or surface defects occurring inside the round bar material just below the surface are received as defect echo signals. A defect echo image is created from each scanning position of the ultrasonic waves for angle beam inspection and the time change in signal intensity of the defect echo signal at the scanning position, and a required number of the defect echo images are given to a convolutional neural network (CNN) as learning data images to learn the defect echo image. a new flaw echo image is provided to the trained convolutional neural network to determine whether the flaw corresponding to the flaw echo image is the surface flaw or the subsurface flaw, wherein a predetermined area including a flaw echo signal is cut out from the flaw echo image for each of the surface flaw and the subsurface flaw, and an image extracted from the predetermined area at a scanning position that is fewer than all scanning positions in the predetermined area is used as the training data image.

2. 2. The ultrasonic inspection method for a round bar material according to claim 1, wherein a phased array probe is used for transmitting and receiving the ultrasonic waves for the angle beam inspection.

3. 3. The ultrasonic inspection method for a round bar material according to claim 1 or 2, wherein the defect echo image is displayed with high brightness in parts of the defect echo signal where the signal strength is high, and with low brightness in parts of the defect echo signal where the signal strength is low.

4. 3. The ultrasonic inspection method for a round bar material according to claim 1 or 2, wherein the defect echo image is displayed with low brightness in parts of the defect echo signal where the signal strength is high, and with high brightness in parts of the defect echo signal where the signal strength is low.

5. 5. The ultrasonic flaw detection method for a round bar material according to claim 1, wherein the round bar material is a black-skinned material.

Images