JPS6126857A - Ultrasonic flaw detecting method and defect deciding method of centrifugally cast iron pipe - Google Patents

Abstract

PURPOSE:To detect a flaw in a centrifugally cast iron pipe accurately by processing flaw detection data collected by an ultrasonic flaw detecting method as specified and setting a mean noise level, and regarding signals having large deviation from the noise level as defect signals. CONSTITUTION:A longitudinal ultrasonic wave incident from one probe 4A is caused to strike a defect 3 directly and its regularly reflected ultrasonic wave 5' is received by the other probe 4B. The longitudinal ultrasonic wave is used, so a wave containing defect information is received firstly as to the reflected wave which is reflected by the defect regularly and a ripple echo noise as a scattered wave from a crystal grain boundary, etc., or a lateral wave converted to reflection mode is received later. For the purpose, gating is set almost at the time obtained by dividing the beam path length by the speed of the longitudinal wave, thereby detecting the defect information selectively. Data obtained by performing flaw detection over a specific range are summed up and averaged to improve the S/N ratio, and the multiplication and division process covering the whole is performed so that the noise level is some constant output, detecting the defect signal securely.

Description

Detailed Description of the Invention (Industrial Field of Application) The present invention is directed to the prevention of internal defects, particularly defects in the tube circumferential direction, that occur due to aging in centrifugally cast tubes, typically heating tubes for steam catalytic reforming. This article relates to an ultrasonic flaw detection method that solves the difficulties of sonic flaw detection and its defect determination method.

This type of heating tube for steam catalytic reforming is used for HK-40 (0,4096C, 2596Or,
It is made by centrifugally casting austonite heat-resistant cast steel such as 20% Ni-based steel and welding a plurality of pipes to a predetermined length. Letter of use li! In step ii, raw material gas such as methane and high-pressure steam are fed under pressure into the ζ tube filled with catalyst and heated from outside the tube, so the inside of the tube is exposed to high temperature and high pressure. If it is used for a long time under this condition, creep fin shear due to 7-1 stress tends to develop radially toward the inner surface of the tube and the outer surface of the tube. There is a risk that defects may occur in the circumferential direction of the inner surface of the tube due to internal cooling.

Therefore, understanding the aging of the heating tube and estimating its remaining life is essential for stable 1-1t operation.

[Conventional technology]

In order to investigate the internal defects that may occur due to aging in the large number of heating tubes installed vertically in the steam catalytic reforming reactor, it is necessary to individually conduct non-destructive testing. In general, non-destructive inspections that can be applied for this purpose include radiographic inspection and ultrasonic flaw detection, but radiographic inspection is performed when the thickness exceeds a certain level in the direction of radiation propagation, generally about 196 mm or more of the plate thickness. It is detected that the defect does not have a thickness of
Other defects such as crack defects72
Unable to detect 18th.

Therefore, detection accuracy is poor.

On the other hand, ultrasonic flaw detection can be effectively used for nondestructive inspection of internal defects in relatively homogeneous steel materials, and specific methods include pulse reflection method, transmission method, resonance method, etc. Methods using pairs of one or two children, and
There are known methods for setting the direction of ultrasonic incidence on the object to be inspected, either vertically or at an oblique angle, and these methods are selected depending on the application conditions such as the shape of the object to be inspected and the type of defect.However, In the case of the above-mentioned centrifugally cast austenitic heat-resistant cast steel pipe, the tin is generally under pressure, and the crystals are coarse and forest-like echoes due to grain boundary reflection are likely to appear, making the received waveform complex, so it is often considered a defect report. Therefore, in the case of centrifugally cast pipes, all of the prior art ultrasonic flaw detection methods lack measures to improve the discrimination against roughness factors from the material aspect, resulting in poor accuracy. Therefore, reliable flaw detection was considered difficult.

(Problems to be Solved by the Invention) The present invention is inspired by the ultrasonic flaw detection of centrifugally cast tubes with coarse crystal grains and large attenuation of ultrasonic waves, and solves the above-mentioned difficulties of the prior art methods.
- Solve and detect internal defects.

The purpose of this paper is to provide an ultrasonic flaw detection method which is mainly effective for detecting defects in the circumferential direction, and a defect determination method thereof.

(Defect #1 of the ultrasonic flaw detection method for centrifugally cast pipes of the present invention) Defect #1 of the ultrasonic flaw detection method for centrifugally cast pipes of the present invention Flaw detection is performed by emitting longitudinal ultrasonic waves from one probe and receiving the specularly reflected wave from the defect with the other probe.The flaw detection signal in the pipe circumferential direction is averaged over a predetermined range and the level tJI4 is adjusted. The method is characterized in that a flaw detection signal with a large change is determined to be a defect signal by using the following.

In other words, due to the probe arrangement and flaw detection method of unexploded BAFi, the attenuation of ultrasonic waves is small, and the collected flaw detection data can be processed in a specific manner. Lv is established in the mutual relationship between setting the average noise level by performing the above steps, detecting a large differential value from that level, and determining it as a defective signal.

Hereinafter, the present invention will be specifically explained in detail with reference to the accompanying drawings to clarify its characteristics.

General! 1: In order to detect defects in a material by reflection of ultrasonic waves, the base material (1) of the object to be inspected and the welded part (2) shown in Fig. 7 (q) are
), the defect (3), the probe (4), and the ultrasonic beam (5). It usually receives sound waves.

In the case of two-time reflection of a longitudinal wave, as shown in the relationship between the two-time sound pressure reflectance on the vertical axis and the incident angle on the horizontal axis in Figure 7 (C), the reflectance for incident angles of 20'' to 7σ is constant. However, it is low at less than 1596.For transverse waves, there is a remarkable change in sound pressure reflectance twice due to the change in the angle of incidence 11: shown in Figure 7←).For longitudinal waves, there is a significant change in the sound pressure reflectance twice between the bottom surface and the defect. Energy loss due to reflection is large (, t* When reflected at the bottom surface, mode conversion occurs and part of the longitudinal wave becomes a transverse wave, causing energy loss.

Therefore, in the double reflection, L of the centrifugally cast tube which is the object of the present invention
It is difficult to detect defects in materials that already have high attenuation. The results when a longitudinal Mi sound wave is reflected once are shown in Fig. 7 (f), and when the incident angle is in the range of 60 to 50, the reflectance is approximately linearly proportional to the incident angle, and the 7th Figure (In the case of →, the machining is also large.

Utilizing this characteristic and the fact that the shape of the object is a tube, we can construct the final version of 8Ak in Figure 1 (a).
As shown in (b), a longitudinal ultrasonic wave (5 N) incident from one probe (4A) is directly applied to the defect (3), and the specularly reflected ultrasonic wave (5) is transferred to the other probe. (Probe arrangement configuration and flaw detection style for receiving with 4 phantoms.

In the present invention, the following various settings are made in this arrangement. In other words, first of all, the reason for using longitudinal ultrasound is that if there is a defect (3), the reflected wave that is specularly reflected by the defect and reaches the receiving probe (4B) is a wave that carries defect information. is first transmitted in the Q-column, and forest-like echo noise due to perturbed waves from grain boundaries or transverse waves whose mode has been converted upon reflection arrives later.

Therefore, by applying darts near the time obtained by dividing the beam path by the velocity of the longitudinal wave, it becomes possible to obtain an output that selectively includes defect information.

Another advantage is that, due to the anisotropy of the object, the degree to which the ultrasonic waves are bent is less with longitudinal waves than with transverse waves.

Next, it goes without saying that it is advantageous, from the viewpoint of reducing attenuation, to place the flaw detector in the flaw detection @ area in a position close to the assumed defective part where the beam path is as short as possible. However, there is currently a geometrical constraint in that there is a stepped part (taper S) formed by machining around the welded part of the test tube. If the probe gets too close to the defect position, the ultrasonic beam will approach the circumferential direction.
A part of the ultrasonic waves emitted from A) directly reaches the reception probe (4B) without undergoing specular reflection at the defect (as shown in Figure 5), thereby lowering the BIM of the reception data.

Therefore, in the present invention, the probe arrangement that can be provided needs to be within the range of reasonable incidence angles and outside the above-mentioned constraint conditions. According to the experiment, L+ = 15 to 40, Lm - 15 to 4 oa, LmI - + 10 = 50 to B□ with the symbols written in Figure 1 (0).
It was found that good S/H flaw detection data could be obtained in the range of a, w = 50 to 100.

@5iC%8/N is affected by the spread of the ultrasonic beam, so in order to reduce the generation of unidentified echoes reflected from tube walls etc., an acoustic lens is attached to the probe to reduce the beam width. It is effective to reduce the thickness.

In the present invention, in addition to performing flaw detection based on the above-mentioned probe arrangement and fc setting format, the flaw detection data obtained by performing flaw detection within a certain range (W fc) Based on this, two-stage signal processing 1M of averaging and level vj4 adjustment is performed to improve defect display characteristics.

FIG. 6 shows an example of flaw detection data obtained by performing F flaw detection using the above-mentioned probe arrangement and flaw detection method.

The horizontal axis represents the angular position of the tube to be inspected in the tube circumferential direction, and the vertical axis plots the voltage at the peak value in the darted range. The tube to be inspected Fi, an artificial defect (5A) t- as shown in Fig. 7 is used as a test piece to simplify the conditions.
The prepared centrifugal casting tube was tested. The wall thickness of the pipe is set to 1, and an artificial defect with a depth of 1/3t is placed at a position of 90° on the pipe circumference, and the pipe circumference is 270°.
There is an artificial defect with a depth of 1/2t at the position. The flaw detection waveform of the flaw detection data shown in Fig. 3 is an irregular waveform consisting of many peaks, including forest echo noise, but since a test piece was used, what appears to be a defect has been observed.

However, both the defect signal and the puncture noise have a large amplitude (approximately 8/H- = 1). Therefore, when testing an actual centrifugally cast tube, the waveform is complex and indicates an internal defect tl- It is difficult to point out clearly.

Therefore, in the present invention, Fi subcutaneous signal processing is performed. First, the first stage of processing is as follows.

Since the defect has a certain extent of expansion in the circumferential direction,
While scanning in the pipe circumferential direction, flaw detection data is collected in a range corresponding to the spread of the pipe circumference, and the output signal is summed in sync with the time of a specific part of the input signal as the origin, and the data is averaged. adopt. As a result, the defect information is multiplied by about M with M additions, whereas the noise is random, so it is only j high, and therefore S/ff is improved by 10 log M (dB) with M additions. . Fourth
The figure shows all the results obtained by adding such averaging processing when obtaining the flaw detection data shown in FIG. 3. The B/H is improved to 2 to 3, and the defect detection accuracy is improved.

As a result of research, it has been found that even if the first stage processing described above is performed, there is still a possibility that a defective signal may be misidentified. Therefore, in final completion #i, the second stage of the next level adjustment is performed.

Figure 5 shows another example of flaw detection data to be compared with Figure 6. The test piece of the pipe to be inspected in Figure 6 has relatively fine crystal grains, whereas the case in Figure 5 has relatively fine crystal grains. I tried a coarse one. The test piece in the case of Fig. 5 has a circumference of 180 mm.
An artificial defect of 1/2t is provided at position 11i. FIG. 6 shows the results compared with FIG. 4, which was obtained from the flaw detection data of FIG. 5 and subjected to the first stage processing. When considering the quality of the material of the pipe to be inspected using this t'L, we find that the fourth
The fine-grained one in the figure and the coarse-grained one in Figure 6 are
In this case, a gain difference of 75 dB is observed. Therefore, when evaluating internal defects with the same flaw detection gain, in actual centrifugally cast tubes, internal defects in coarse-grained tubes may be overlooked, or noise in fine-grained tubes may be overestimated and mistakenly interpreted as defect signals. There is fear. On the other hand, in the second stage of the present invention, the one with large attenuation attenuates the defective signal level and also reduces the noise level. The entire multiplication/division process is performed to obtain a certain constant output. To specifically describe the above example, in order to match the noise level of the 6th factor to the noise level of FIG. As another judgment method, the sum of the changed average value of the noise level and a constant value is set as a judgment level, and any signal exceeding the judgment level is judged to be a defective signal.

In the previous example, it was explained that the first stage signal processing was carried out using the No. A method of reducing the noise level by adding and averaging the flaw detection data obtained during the test, or a method of reducing the noise by adding and averaging the flaw detection data obtained by performing tandem scanning in which flaws are detected by shifting a small distance in the direction of the tube axis. Therefore, the averaging method can be selected as appropriate depending on the setting range of location and time changes in flaw detection.

The above is a description of a failed example for localized 72 defects, but if the defects are all around the circumference, the implementation will be done in a slightly different manner. In other words, in this case, the tube to be inspected is scanned in the circumferential direction at position A in the tube axis direction to collect flaw detection data at several points, and when comparing the flaw detection data at the position after gB, the L9 gain is significantly increased, for example by 6 dB. A case where fc changes is defined as a full-circumference defect.

The following is a consideration of the case where there are defects all around the circumference. In other words, in this case, the noise level increases, which is equivalent to k, and there is a fear that the determination method of the previous embodiment may miss the defect. However, fortunately, all of the flaw detection data is collected at several points on the tube to be inspected, rather than at one location, so it is possible to compare the noise level of the flaw detection results at the position after mJ. In reality, since only a small number of steels have defects all around, it is considered that there is less than a 99% probability that all the fault points to be inspected have defects all the way around. Therefore, when compared with the before and after flaw detection data.

If the level changes significantly (approximately 5 dB),
The ninth step is to determine if 'L is a defect all around.

(Effects of the Invention) As described above, the method Mj of the present invention eliminates internal defects that occur due to aging in centrifugal casting tubes, such as heating tubes for steam catalytic reforming, which have coarse grains and large attenuation of ultrasonic waves. When detecting flaws by non-destructive testing using ultrasonic flaw detection, it is possible to reliably detect defects with a defect value of 91 by reducing the effects of noise due to attenuation and scattering of ultrasonic waves and the difference in influence due to the material of the pipe. The number of inspection steps is small, and the reliability of the tube can be determined accurately due to the reliability of the flaw detection.

[Brief explanation of drawings]

Fig. 1 (a) is a schematic side view showing the arrangement of an ultrasonic probe with respect to a pipe to be inspected for carrying out the method of the present invention, Fig. 1 (b) is a schematic enlarged view thereof, and Fig. 2 is a schematic side view showing the arrangement of an ultrasonic probe with respect to a pipe to be inspected for carrying out the method of the present invention. A schematic side view of the case where the arrangement is inappropriate, Figure 6 shows the tube circumferential angle position on the horizontal axis and the vertical @
Figure 4 shows an example of flaw detection data by measuring the reflection intensity.
The figure is a chart of the averaging processing results derived from the flaw detection data in Figure 6 under the same vertical and horizontal axes.
: Chart showing felling of a2 flaw data, Figure 6 is a chart showing the results of the normalization process from the flaw detection data in Figure 5 on the same vertical and horizontal axes, Figure 7 is a diagram showing the general double reflection. Fig. 7 (B) is a diagram showing the beam path when the ultrasonic beam direction is reversed, and Fig. 7 (C) is a longitudinal IFI side view showing the method of ultrasonic flaw detection. (Figure 7) is a chart showing the relationship between the incident angle on the horizontal axis and the sound pressure reflectance of twice reflected on the vertical axis. A chart showing the relationship with sound pressure reflectance. FIG. 7(e) is a chart showing the relationship between the incident angle on the horizontal axis and the longitudinal wave reflectance on the vertical axis in the case of longitudinal ultrasound. (1)...Object to be inspected, (2) Fist φ weld, (3) C5
A) -Φ Ultrasonic beam, (θ)... Incident angle % (In) (
L Spirit) (W)・Fist distance, (T)・Thickness. Name of applicant's agent

Claims (4)

[Claims] (1) Perform flaw detection by placing two probes on the pipe to be inspected so that one probe emits longitudinal ultrasonic waves and the other probe receives specularly reflected waves from defects. A flaw determination method using ultrasonic flaw detection for centrifugally cast pipes, characterized in that the flaw detection signals in the circumferential direction are averaged over a predetermined range and the levels are adjusted to determine a flaw detection signal with a large change as a flaw signal. . (2) Average the flaw detection signals in the circumferential direction of the pipe using a synchronous averaging method over a predetermined time range, obtain a signal adjusted by multiplication/division of the entire circumference so that the noise level becomes constant after all-round flaw detection, and The defect determination method according to claim 1, wherein a signal exceeding a certain threshold is determined to be a local defect signal. (3) The flaw detection signals in the pipe circumferential direction are averaged using the synchronous averaging method over a predetermined time range, and after all-round flaw detection, the signal is adjusted by multiplication/division of the whole so that the noise level becomes a constant value, and the noise level is The defect determination method according to claim 1, wherein a signal exceeding the sum of an average value and a constant value is determined to be a local defect signal. (4) The pipe to be inspected is flaw-detected in the circumferential direction at different positions along the pipe axis, and if the gain changes by more than a certain value by comparison of the test signals before and after, it is determined that there is a defect all around the circumference. The defect determination method according to item 1.