JP7181356B1 - Inspection equipment for pipe welds - Google Patents

Abstract

An object of the present invention is to improve the accuracy of inspection for determining the presence or absence of weld defects in pipe welds. A pipe weld inspection apparatus according to an embodiment includes a camera, a laser beam measurement device, a mounting device, and an inspection unit. The camera photographs the welded portion of the first pipe and the second pipe that are connected by welding. A laser beam measuring device creates an image by irradiating a laser beam onto a welded portion and photographing the laser beam reflected by the welded portion. The mounting device includes a rail stand arranged around the first pipe or the second pipe, a ring rail gear that moves around the weld along the rail stand, and a camera and a laser light measuring device attached to the ring rail gear. and a mounting base for fixing the The inspection unit determines whether or not there is a weld defect in the weld based on the images captured by the camera and the laser beam measuring device. [Selection drawing] Fig. 2

Description

The present invention relates to an inspection apparatus for pipe welds.

A large number of pipes exist in plant facilities, and a huge amount of pipes are welded in each construction/modification work. As a non-destructive test of these pipe welds, visual inspection to determine the presence or absence of weld defects by directly observing the surface of the weld, penetration of inspection liquid into the weld, development of the permeated inspection liquid, and image showing the defective part There is a penetrant test that determines the presence or absence of welding defects by displaying . In addition to this, the non-destructive test includes a magnetic particle flaw detection test that determines the presence or absence of weld defects from the image formed by magnetizing the welded part and attaching magnetic particles to it, and irradiating the welded part with ultrasonic waves. There are ultrasonic testing that determines the presence or absence of welding defects by imaging reflected ultrasonic waves, and radiographic testing that determines the presence or absence of welding defects by imaging the amount of radiation leakage.

It takes an enormous amount of work time to verify a huge amount of inspection images acquired by these tests and find welding defects. Also, in order to determine the presence or absence of welding defects from the images obtained in each test, a skilled person's excellent inspection technique is required. In order to solve this problem, in recent years, the development of inspection technology using artificial intelligence (AI) is underway.

Japanese Patent Application Laid-Open No. 2020-024201

For example, in piping through which combustible gas passes, even a minute welding defect can cause a leak of combustible gas and cause a serious accident. Therefore, determination of welding defects requires an enormous amount of images captured by moving the imaged portion of the welded portion at a fine pitch. For example, it is preferable that the number of inspection images used to determine a weld defect in a weld connecting two pipes with a diameter of 1 m is 1,000 or more. A huge amount of work time is required to capture 1,000 or more still images. The operator takes a video of the surroundings of the welded part with a digital camera, and generates more than 1,000 still images from the taken video, thereby reducing the image creation time. However, when an operator shoots a moving image around the welded part, it is difficult to keep constant shooting conditions such as the moving speed of the camera and the shooting angle with respect to the welded part. If the presence or absence of a weld defect in a pipe welded portion is determined using an image that has been captured under unconstant photographing conditions, there is a risk that the accuracy of the inspection for determining the presence or absence of a weld defect will be degraded.

SUMMARY OF THE INVENTION It is an object of the present invention to improve the accuracy of inspection for determining the presence or absence of weld defects in pipe welds.

An inspection apparatus for a pipe welded portion according to an embodiment for solving the above problems includes a camera, a laser light measurement device, a mounting device, and an inspection section. The camera photographs the welded portion of the first pipe and the second pipe that are connected by welding. A laser beam measuring device creates an image by irradiating a laser beam onto a welded portion and photographing the laser beam reflected by the welded portion. The mounting device includes a rail stand arranged around the first pipe or the second pipe, a ring rail gear that moves around the weld along the rail stand, and a camera and a laser light measuring device attached to the ring rail gear. a mounting base for fixing the The inspection unit determines whether or not there is a weld defect in the weld based on the images captured by the camera and the laser beam measuring device.

1 is a block diagram of an inspection device for a pipe welded portion according to an embodiment; FIG. It is a top view of the attachment device concerning an embodiment. 1 is a front view of a mounting device according to an embodiment; FIG. It is a figure for demonstrating the ring rail gear which concerns on embodiment. 6 is a flow chart for explaining inspection processing using the pipe weld inspection apparatus according to the embodiment.

Hereinafter, an inspection apparatus for a pipe welded portion according to an embodiment will be described with reference to the drawings. In the description, an XYZ coordinate system consisting of mutually orthogonal X, Y, and Z axes is appropriately used. The pipe weld inspection device is an inspection device for determining the presence or absence of a weld defect in a weld between a first pipe and a second pipe that are connected by welding.

FIG. 1 is a block diagram of an inspection apparatus 1 for pipe welds according to an embodiment. The inspection device 1 has a camera 11 , a laser light measuring device 12 and an inspection section 20 . The camera 11 and the laser beam measurement device 12 are attached to an attachment device which will be described later.

The camera 11 is a photographing device that photographs a welded portion where the first pipe and the second pipe are connected by welding. The camera 11 transmits image information of the captured image to the inspection unit 20 . The camera 11 has an adjustment mechanism that automatically adjusts focus. A plurality of LEDs are provided around the lens of the camera 11 . The plurality of LEDs are arranged so that the light irradiation direction of the LEDs coincides with the photographing direction of the camera 11 (optical axis of the camera).

The laser beam measuring device 12 has an output portion 12a that outputs a laser beam toward the welded portion and a receiver portion 12b that receives the laser beam reflected by the welded portion. The receiver 12b is a camera that creates an image by capturing the laser beam reflected by the weld. The same camera as the camera 11 can be used for the receiving unit 12b. However, no LED is arranged in the camera used for the receiving section 12b. The output part 12a is arranged so that the laser beam is irradiated at a predetermined angle (for example, 45 degrees) with respect to the welded part. The predetermined angle is adjusted using a mounting base 56, which will be described later. The receiving unit 12b is arranged so that the imaging direction faces the center of the pipe. By setting an angle difference between the irradiation angle of the laser beam to the welded part and the angle of reception of the laser beam reflected by the welded part, the image formed at the receiving part will vary depending on the undercut and reinforcement height of the welded part. shadows are formed. The presence or absence of a weld defect can be determined based on the shape and size of this shadow. By creating a three-dimensional image from images captured by changing the angle difference between the irradiation angle and the light receiving angle of the laser light, it is possible to improve the accuracy of the inspection for determining the presence or absence of welding defects.

The inspection unit 20 is physically a computer having a CPU, a storage unit, an input unit and an output unit. The storage unit stores inspection application software (analysis program) using artificial intelligence (AI) and supervised data used for analysis using artificial intelligence. The inspection unit 20 determines the presence or absence of weld defects in the weld based on inspection application software using artificial intelligence and images of the weld taken by various measuring devices.

Commercially available software can also be used as inspection application software using artificial intelligence. For example, it is possible to use software using algorithms such as machine learning and deep learning configured with neural networks.

For supervised data, for each of the multiple images of the weld zone taken in each test described later, information about no welding defect is added to the image without welding defect, and information about the presence of welding defect is added to the image with welding defect. Data. Specifically, the supervised data adds information on whether there are welding defects or not to each of the multiple images with subtle differences such as weld undercuts, reinforcement heights, leg lengths, and very small pits. Created by Judgment of presence or absence of weld defects is made by a skilled engineer based on various inspection images and test results. The more supervised data, the better.

The inspection unit 20 includes an input unit 21, a camera image inspection unit 22, a laser image inspection unit 23, a penetrant test image inspection unit 24, a magnetic particle test image inspection unit 25, an ultrasonic flaw test image inspection unit 26, and a radiation penetrant test image. It has an inspection unit 27 and an output unit 28 .

The input unit 21 includes a network interface, a USB interface, and the like. The input unit 21 inputs image information of the image of the welded portion captured by the camera 11 and the image of the welded portion captured by the laser light measuring device 12 . Further, the input unit 21 inputs image information of the image of the weld obtained by penetrant testing, magnetic particle testing, ultrasonic testing, and radiation penetrating testing.

In the penetrant testing, a test liquid is permeated into the weld, and the permeated developer is developed. If there is a defect in the weld, capillary action allows the test liquid to penetrate the weld. Since the permeated test liquid spreads around, even in the case of a minute hole in which the weld defect cannot be visually recognized, the weld defect can be magnified into a visible image. Therefore, the weld defect can be displayed as a visible image.

In magnetic particle testing, a magnet is placed inside the pipe. If there is a defect on the surface of the piping, the magnetic flux leaks from the defect. When magnetic powder is applied to the surface of the pipe, if there is a defect in the pipe, the magnetic powder adheres to the leakage point due to the leaked magnetic flux, and the weld defect can be displayed as a visible image.

In the ultrasonic flaw detection test, ultrasonic waves are applied to the pipe, and the intensity distribution of the reflected ultrasonic waves is created as an image. If the intensity distribution of reflected waves is not uniform, there is a high possibility that there is a weld defect.

In the radiation penetration test, a radiation source is placed inside the pipe and an image is created by exposing the film to radiation that has penetrated through the joint. If there is a weld defect, the intensity of radiation penetrating through the weld increases. The wider the exposure range of the film, the more likely there is a weld defect.

The camera image inspection unit 22 inspects the welded portion with the camera 11 based on supervised data in which information on the presence or absence of welding defects is added to the image of the welded portion photographed by the camera 11, and inspection application software using artificial intelligence. It is determined whether each of the plurality of images to be inspected photographed in 11 belongs to "with welding defect" or "without welding defect".

The laser image inspection unit 23 is based on supervised data in which information on the presence or absence of welding defects is added to the image of the weld taken by the laser light measurement device 12, and inspection application software using artificial intelligence. It is determined whether each of the plurality of images of the welded portion to be inspected photographed by the optical measurement device 12 belongs to "with welding defect" or "without welding defect".

The penetrant test image inspection unit 24 is based on supervised data in which information on the presence or absence of weld defects is added to the image of the weld created by the penetrant test, and inspection application software using artificial intelligence. It is determined whether each of the plurality of images of the weld to be inspected belongs to "with weld defects" or "without weld defects".

The magnetic particle inspection image inspection unit 25 is based on supervised data in which information on the presence or absence of welding defects is added to the image of the weld created by the magnetic particle inspection, and inspection application software using artificial intelligence. It is determined whether each of the plurality of images of the weld to be inspected belongs to "with weld defects" or "without weld defects".

The ultrasonic flaw detection image inspection unit 26 is based on supervised data in which information on the presence or absence of welding defects is added to the image of the weld created by the ultrasonic flaw detection test, and inspection application software using artificial intelligence. Then, it is determined whether each of the plurality of images of the welded portion to be inspected belongs to "with weld defect" or "without weld defect".

The radiation penetration test image inspection unit 27 is based on supervised data in which information on the presence or absence of weld defects is added to the image of the weld created by the radiation penetration test, and inspection application software using artificial intelligence. It is determined whether each of the plurality of images of the weld to be inspected belongs to "with weld defects" or "without weld defects".

The output unit 28 is composed of a liquid crystal display, a printer, and the like. The output unit 28 includes the camera image inspection unit 22, the laser image inspection unit 23, the penetrant test image inspection unit 24, the magnetic particle test image inspection unit 25, the ultrasonic test image inspection unit 26, and the radiation penetrant test image inspection unit 27. When at least one determines that "there is a welding defect", the inspection result of "there is a welding defect" is output for the corresponding welded portion.

Next, the mounting device 50 will be described. FIG. 2 is a plan view of the mounting device 50. FIG. 3 is a front view of the mounting device 50. FIG. In FIG. 3, the first pipe 101, the second pipe 102, and the welded portion are referred to as the welded portion W. As shown in FIG. Here, descriptions of parts that are shown in the drawings but are not related to the present invention will not be given. The mounting device 50 moves the camera 11 and the laser light measuring device 12 along the periphery of the welded portion, thereby enabling the camera 11 and the laser light measuring device 12 to create a three-dimensional image of the welded portion. is.

The mounting device 50 includes a rail base 51, a ring rail gear 52, a punch lock 53, a motor/encoder 54, a mounting base 55 for mounting the camera 11 on the ring rail gear 52, and a mounting base for mounting the laser beam measuring device 12 on the ring rail gear 52. 56.

The rail base 51 is a ring-shaped member made of metal such as iron or aluminum. The rail base 51 includes a rail base body 511 , a plurality of rail guides 512 holding the ring rail gear 52 , and a fixed shaft 513 rotatably holding the rail guides 512 on the rail base body 511 .

The rail base main body 511 is composed of rail base main bodies 511a and 511b obtained by dividing a ring-shaped member into two parts. In FIGS. 2 and 3, the portion depicted on the left side of the center line is the rail base body 511a, and the portion depicted on the right side is the rail base body 511b. The rail base main body 511 becomes a ring-shaped member by connecting the rail base main bodies 511 a and 511 b with the punch lock 53 .

The rail guide 512 is a member made of metal or resin. The rail guide 512 has a groove 512a near the center in the Z-axis direction, as shown in FIG. This groove 512 a is a groove for guiding the ring rail gear 52 . As shown in FIG. 2 , the rail guides 512 are arranged at multiple locations along the outer and inner circumferences of the ring rail gear 52 . The rail guide 512 is held by a fixed shaft 513 arranged so as to protrude from the rail base body 511 in the -Z direction, and rotates around the fixed shaft 513 as a rotation axis.

The rail base 51 has four device holding legs 58 (see FIG. 2). The device holding leg 58 has a slide portion 581 that can slide toward the center of the pipe and a fixed portion 582 that forms a contact surface with the side surface of the pipe (see FIG. 3). The slide portion 581 can be fixed at any position according to the thickness of the pipe. The rail base 51 is fixed to the first pipe 101 by fixing the slide portion 581 while the fixing portions 582 of the four device holding legs 58 are in contact with the pipe. Since the device holding leg 58 has the slide portion 581, the rail base 51 can be fixed to piping of any thickness.

FIG. 4 is a diagram for explaining the ring rail gear 52. As shown in FIG. As shown in FIG. 4, the ring rail gear 52 is composed of a first member 521 and a second member 522 formed in a ring shape. The inner radii of the first member 521 and the second member 522 are the same. The outer radius of the second member 522 is larger than the outer radius of the first member 521 . The first member 521 is coupled to the −Z side surface of the second member 522 so that the inner peripheral surface of the first member 521 and the inner peripheral surface of the second member 522 are aligned. A gear is formed on the outer circumference of the first member 521 . 2 and 3, a gear is formed on the inner periphery of the first member 521, but in FIG. 4, a gear is formed on the outer periphery of the first member 521 in order to make it easier to fit with the gear described later. The figure being formed is described. Forming a gear on the inner circumference of the first member 521 can make the device more compact. However, the gears may be provided on either the inner circumference or the outer circumference of the first member 521 because they do not affect the measurement accuracy. The plate thickness (thickness in the Z-axis direction) of the second member 522 is formed so as to decrease toward the outer periphery. A portion of the outer periphery of the second member 522 where the plate thickness is reduced forms a protrusion 522a.

The ring rail gear 52 is arranged by fitting the protrusions 522 a of the second member 522 into the grooves 512 a of the rail guides 512 . 4, illustration of the rail guide 512 arranged on the inner peripheral side of the ring rail gear 52 is omitted. The ring rail gear 52 is also composed of two members like the rail base main bodies 511a and 511b, but illustration of the connecting portion of the two members is omitted in FIG.

Returning to FIGS. 2 and 3, motor encoder 54 is fixed to rail platform 51 . The motor/encoder 54 includes a DC motor 541 and a gear 542 operated by a storage battery (the DC motor 541 is omitted in FIG. 3). The motor encoder 54 rotates the DC motor 541 based on instructions from measurement application software installed in a control unit (not shown). A rotating shaft of the DC motor 541 is connected to a gear 542 . The gear 542 is arranged in engagement with a gear formed on the outer periphery of the first member 521 of the ring rail gear 52 (see FIG. 4). The motor encoder 54 rotates the gear 542 to rotate the ring rail gear 52 along the rail guide 512 of the rail base 51 . A plurality of motor encoders 54 may be provided.

Returning to FIGS. 2 and 3, the mounting base 55 is a member for mounting the camera 11 on the ring rail gear 52. As shown in FIG. The mount 55 is fixed to the ring rail gear 52 . The mount 55 has an angle adjustment mechanism that adjusts the imaging angle of the camera with respect to the welded portion. The shooting angle can be adjusted according to welding conditions such as butt welding and socket welding. The mount 55 also has a position adjustment mechanism that adjusts the distance between the welded portion and the camera 11 .

A mounting base 56 for mounting the laser beam measuring device 12 on the ring rail gear 52 is fixed to the ring rail gear 52 at a position 180 degrees away from the installation position of the mounting base 55 across the center line of the pipe. The mount 56 is an angle adjuster that adjusts the irradiation angle of the laser beam with respect to the weld (the angle between the irradiation direction and the light receiving direction of the laser beam on the XY plane, and the angle between the irradiation direction of the laser beam and the Z axis). It has a mechanism 561 . The output section 12 a of the laser beam measurement device 12 is attached to the angle adjustment mechanism 561 . The receiving unit 12b of the laser beam measuring device 12 is installed so that the photographing direction faces the center of the pipe. The mount 56 can also be configured by adding an angle adjustment mechanism 561 to the mount 55 .

Next, referring to FIG. 5, an inspection process using the inspection apparatus 1 for a pipe welded portion according to the embodiment will be described. Prior to inspection, supervised data is created and stored in the storage unit of the inspection apparatus 1. - 特許庁Specifically, supervised data is created for the image captured by the camera 11 and the image captured by the laser beam measuring device 12, and stored in the storage unit. In addition, supervised data is created for images obtained by penetrant testing, magnetic particle testing, ultrasonic testing, and radiation penetrating testing, and stored in the storage unit. The more supervised data, the better. Here, penetrant testing, magnetic particle testing, ultrasonic testing, and radiation penetrating testing are performed separately, and the case where the image information of the image of each test result is input to the input unit 21 of the inspection unit 20 will be described. do.

First, the mounting device 50 is fixed near the welded portion of the pipe (step S11). Specifically, the rail base 51 divided into two and the ring rail gear 52 divided into two are combined with a punch lock 53 with a pipe interposed therebetween. By adjusting the device holding legs 58, the mounting device 50 is fixed in the vicinity of the welded portion of the pipe. The projection 522 a of the second member 522 of the ring rail gear 52 is fitted into the groove 512 a of the rail guide 512 to fix the rail guide 512 to the rail base body 511 . Also, the gear 542 of the motor/encoder 54 is positioned so as to engage with the gear formed on the outer periphery of the first member 521 of the ring rail gear 52 .

Next, the camera 11 is fixed to the mount 55, and the photographing angle of the camera 11 with respect to the welded portion is adjusted to a predetermined angle. Further, the laser beam measuring device 12 is fixed to the mount 56 , and the laser irradiation angle and the laser light receiving angle with respect to the welded portion are adjusted to predetermined angles (step S12).

The ring rail gear 52 is then rotated around the pipe by the motor/encoder 54 as instructed by the measurement application software. The camera 11 and the laser light measuring device 12 capture images of the welded portion at a speed corresponding to the moving speed of the ring rail gear 52 based on instructions from the measurement application software. The camera 11 and the laser beam measurement device 12 take images of the welded portion so that one image can be acquired each time the ring rail gear 52 moves a predetermined distance (for example, 0.1 mm) (step S13). The predetermined distance is determined according to the required inspection accuracy.

Since the installation positions of the camera 11 and the laser light measuring device 12 (for example, positions separated by 180 degrees) are known, an image of an arbitrary position of the welded part taken by the camera 11 and the same position by the laser light measuring device 12 can be obtained. It can be associated with the photographed image. Specifically, based on the relationship between the installation positions of the camera 11 and the laser beam measuring device 12, the amount of rotation of the DC motor 541 of the motor/encoder 54, and the imaging timing of the camera 11 and the laser beam measuring device 12, the An image of an arbitrary portion photographed by a camera 11 and an image of the same portion photographed by a laser beam measuring device 12 are made to correspond. The image captured by the camera 11 and the image captured by the laser light measurement device 12 capturing the same portion are associated with each other and stored as image information in the storage unit (step S14). Next, if necessary, the imaging angle of the camera 11 and the laser irradiation angle of the laser light measuring device 12 are changed, and the process of step S14 is repeated. An image captured by changing the imaging angle of the camera 11 and the laser irradiation angle of the laser beam measuring device 12 and an image of the same portion of the weld are linked and stored as image information in the storage unit. Each photographed image is added with coordinate information that can specify the position of the image in the welded portion, and is stored in the storage unit.

Next, if there is an image obtained by penetrant testing, magnetic particle testing, ultrasonic testing, or radiation penetrating testing, the image information is input to the inspection unit 20 (step S15).

The inspection unit 20 analyzes the input image information for weld defects based on supervised data and inspection application software using artificial intelligence (step S16). Specifically, the camera image inspection unit 22 determines whether each of the plurality of images of the weld to be inspected photographed by the camera 11 belongs to "with welding defect" or "without welding defect". The laser image inspection unit 23 determines whether each of the plurality of images of the welded portion to be inspected photographed by the laser beam measuring device 12 belongs to "with welding defect" or "without welding defect". Since the image of the object to be inspected taken by the camera 11 and the image of the welded part to be inspected taken by the laser beam measurement device 12 are linked, the camera image inspection unit 22 and the laser image inspection unit 23 The presence or absence of welding defects can be determined based on the image of the same part of the part.

Further, the penetrant testing image inspection unit 24 determines whether each of the plurality of input images of the welded portion to be inspected belongs to "with welding defects" or "without welding defects". The magnetic particle inspection image inspection unit 25 determines whether each of the input images of the welded portion to be inspected belongs to "with welding defects" or "without welding defects". The ultrasonic testing image inspection unit 26 determines whether each of the plurality of input images of the welded portion to be inspected belongs to "with welding defects" or "without welding defects". The radiation penetration test image inspection unit 27 determines whether each of the plurality of inputted images of the welded portion to be inspected belongs to "with welding defect" or "without welding defect".

The inspection unit 20 includes a camera image inspection unit 22, a laser image inspection unit 23, a penetrant test image inspection unit 24, a magnetic particle test image inspection unit 25, an ultrasonic test image inspection unit 26, and a radiation penetrant test image inspection unit. 27 determines that "there is a welding defect" for any of the plurality of images to be inspected, it outputs to the output unit 28 that there is a welding defect in the corresponding weld (step S17). In addition, the inspection unit 20 outputs to the output unit 28 an image that serves as a basis for determining that there is a weld defect, together with coordinate information that enables the position of the image in the welded portion to be specified.

The inspection unit 20 receives image information of an image captured by the camera 11, image information of an image captured by the laser light measuring device 12, image information acquired by penetrant testing, image information acquired by magnetic particle penetrant testing, and ultrasonic flaw detection. Information on the determination result of the presence or absence of welding defects in the welded portion performed in step S16 or step S17 is added to each of the image information obtained by the test, the image information obtained by the ultrasonic flaw detection test, and the image information obtained by the radiographic test. The obtained information is added to the supervised data stored in the storage unit as new supervised data and stored (step S18). By accumulating supervised data in this way, it is possible to improve the accuracy of subsequent inspections for determining the presence or absence of weld defects in welded portions.

As described above, the pipe weld inspection apparatus 1 according to the present embodiment has the mounting device 50 that mounts the camera 11 and the laser beam measurement device 12 and rotates around the weld. The mounting device 50 includes a rail base 51 arranged around the first pipe or the second pipe, a ring rail gear 52 that moves around the weld along the rail base 51, and a camera attached to the ring rail gear 52. 11 and mounting bases 55 and 56 for fixing the laser beam measurement device 12 . By using this mounting device 50, the camera 11 and the laser light measurement device 12 can take images of the weld while maintaining a constant angle and distance and a constant spacing with respect to the weld. Therefore, the inspection apparatus 1 for pipe welds according to the present embodiment can improve the quality of images used for inspecting weld defects in pipe welds. Inspection accuracy can be improved.

In addition, since the pipe weld inspection apparatus 1 according to the present embodiment detects weld defects in the weld based on an analysis program using artificial intelligence, inspection when verifying a huge amount of inspection images can save time. In addition, the judgment of the presence or absence of welding defects does not require an expert's excellent inspection technique.

In addition, the inspection apparatus 1 for pipe welds according to the present embodiment creates new supervised data by adding information on the determination result of the presence or absence of welding defects in the welds to the image information of each inspection image, The newly created supervised data is added to the supervised data stored in the storage unit. Then, including the added supervised data, the presence or absence of welding defects in the subsequent welds is determined. By increasing the amount of supervised data in this way, it is possible to improve the accuracy of the inspection for determining the presence or absence of weld defects in welds.

In addition, the inspection apparatus 1 for pipe welds according to the present embodiment includes image information of an image taken by the camera 11, image information of an image taken by the laser light measurement device 12, image information of an image created by a penetrant inspection test, Image information of images created by multiple tests such as image information of images created by magnetic particle penetrant testing, image information of images created by ultrasonic testing, image information of images created by radiographic testing, etc. Based on an analysis program using artificial intelligence, the presence or absence of welding defects in welds is determined. As a result, it is possible to improve the accuracy of the inspection for determining the presence or absence of weld defects in the welded portion.

In the above description, the inspection unit 20 includes the camera image inspection unit 22, the laser image inspection unit 23, the penetrant inspection image inspection unit 24, the magnetic particle inspection image inspection unit 25, the ultrasonic inspection image inspection unit 26, the radiation Although the case where the penetration test image inspection unit 27 is provided has been described, one of the inspection units can be omitted.

Further, in the above description, in the description of creating supervised data and in the description of making the inspection device 1 determine the presence or absence of welding defects, the welding conditions for butt welding, socket welding, etc., arc welding, gas welding, etc. The case where the welding conditions, the material of the pipe, the size and thickness of the pipe, and other conditions are not specified has been described. These conditions are linked to create supervised data, and when the inspection device 1 determines the presence or absence of welding defects, these conditions are specified to determine the presence or absence of welding defects, thereby improving inspection accuracy. can be expected.

Further, in the above description, the camera image inspection unit 22, the laser image inspection unit 23, the penetrant test image inspection unit 24, the magnetic particle test image inspection unit 25, the ultrasonic test image inspection unit 26, and the radiation penetrant test image inspection unit 27 has determined that "there is a welding defect", the case where the inspection result "there is a welding defect" is output for the corresponding weld portion has been described. Other embodiments include a camera image inspection unit 22, a laser image inspection unit 23, a penetrant inspection image inspection unit 24, a magnetic particle inspection image inspection unit 25, an ultrasonic inspection image inspection unit 26, and a radiation penetration test image inspection unit. If two or more inspection units 27 determine that "there is a welding defect", the inspection result of "there is a welding defect" may be output for the corresponding weld.

Also, in the above description, the case where supervised data is stored in advance in the storage unit of the inspection unit 20 has been described. However, the analysis program using artificial intelligence used for inspection is not limited to this. For example, in the case of artificial intelligence using a deep learning algorithm, as learning progresses, the parameters of formulas used for internal judgment are optimized. When using artificial intelligence with parameters in a state where the accuracy rate for determining the presence or absence of welding defects is sufficiently high (for example, the accuracy rate is 99.9% or more), welding can be performed without using supervised data. It is possible to determine the presence or absence of defects.

In the above description, the case where the mounting device 50 includes the mount 55 for the camera 11 and the mount for the laser beam measurement device 12 is described. The measurement device 12 may be attached to the mounting base 55 and the measurements may be performed in two steps. Moreover, the mounting device 50 may include a mounting base for a measuring device used for an ultrasonic flaw detection test or the like.

In the above description, the rail guide 512 is rotatably held on the rail base body 511 by the fixed shaft 513, but the rail guide 512 is fixed to the rail base body 511 without rotating. good too.

(Modification 1)
In the above description, the case where supervised data is created by adding information indicating whether there is a welding defect or not to the image captured by the camera 11 and the image captured by the laser light measuring device 12 is described. As another embodiment, an image of a designated portion of the welded portion photographed by the camera 11 and a laser beam reflected from the welded portion by irradiating the same portion of the welded portion photographed by the camera 11 with a laser beam are photographed. The supervised data may be created by adding information indicating whether there is a welding defect or not to the information associated with the image obtained by the welding. The camera 11 and the laser beam measurement device 12 are arranged at predetermined positions (for example, positions separated by 180 degrees). Therefore, based on the relationship between the installation positions of the camera 11 and the laser light measuring device 12, the amount of rotation of the DC motor 541 of the motor encoder 54, and the imaging timing of the camera 11 and the laser light measuring device 12, an arbitrary part of the welded part can be associated with an image captured by the camera 11 and an image of the same portion captured by the laser light measuring device 12. - 特許庁When inspecting the presence or absence of welding defects, the presence or absence of welding defects is determined based on the information obtained by linking the image of the same portion of the welded portion photographed by the camera 11 and the image photographed by the laser beam measuring device 12. . In this case, the camera image inspection unit 22 and the laser image inspection unit 23 become one inspection unit.

The shape of the welded portion of the pipe varies from welded portion to welded portion. In addition, it may be discolored due to adhesion of dust or the like. Even if there is discoloration in the weld, it cannot be said to be a weld defect. Therefore, if the presence or absence of a welding defect is judged only by the image photographed by the camera 11, it is difficult to increase the accuracy of the judgment. On the other hand, if the presence or absence of welding defects is determined only by the image captured by the laser beam measuring device 12, it may not be possible to extract the welding defects due to the relationship with the irradiation angle of the laser beam, and it is difficult to increase the accuracy rate of determination. be. As described in the above embodiment, it is also possible to determine whether or not there is a weld defect by ORing the results of both determinations made by artificial intelligence. When the image captured by the camera 11 and the image captured by the laser light measurement device 12 are not linked, for example, the camera image inspection unit 22 determines that the image captured by the camera 11 has a welding defect, When the image inspection unit 23 determines that "there is no welding defect" for the image at the same position, the inspection apparatus 1 determines that "there is a welding defect". If there is discoloration in the weld, even if there is no weld defect, there is a high possibility of such an erroneous judgment.

On the other hand, when the image captured by the camera 11 and the image captured by the laser beam measurement device 12 are linked, the determination result of the inspection device 1 is likely to change. There is a high possibility that the feature amount extracted from the image will change depending on whether the image captured by the camera 11 and the image captured by the laser beam measurement device 12 are associated with each other or not. In addition, when the image captured by the camera 11 and the image captured by the laser light measurement device 12 are linked, artificial intelligence using a deep learning algorithm determines the value of the parameter of the intermediate layer in deep learning to determine the welding defect. is likely to change to a value specific to Therefore, supervised data is created by linking the image captured by the camera 11 and the image captured by the laser light measurement device 12, and the image captured by the camera 11 and the image captured by the laser light measurement device 12 are linked. By judging the presence or absence of welding defects in the obtained data, it is possible to expect an improvement in inspection accuracy.

(Modification 2)
In the explanation of Modified Example 1, an image of the same part of the welded portion photographed by the camera 11 and an image photographed by the laser light measuring device 12 are linked, and the linked image is determined whether there is a welding defect or not. We explained the case of creating supervised data by adding the information of As other embodiments, image information of an image captured by the camera 11, image information of an image captured by the laser beam measuring device 12, image information of an image created by a penetrant test, and image of an image created by a magnetic particle penetrant test information, image information of images created by ultrasonic testing, image information of images created by radiographic testing, and image information of images created by multiple tests, linked in any combination to information of welding defects It is also possible to create supervised data by adding information on the presence/absence.

(Modification 3)
In the inventions according to modifications 1 and 2, conditions such as welding conditions such as butt welding and socket welding, welding conditions such as arc welding and gas welding, conditions such as pipe material and pipe size and thickness, etc. to create supervised data. Then, when the inspection apparatus 1 determines whether or not there is a welding defect, these conditions are specified to determine whether or not there is a welding defect. In this way, by specifying welding conditions for inspection, it is possible to expect an improvement in inspection accuracy.

Although embodiments of the present invention have been described above, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and modifications can be made without departing from the scope of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are included in the scope of the invention described in the claims and its equivalents.

DESCRIPTION OF SYMBOLS 1... Inspection apparatus 11... Camera 12... Laser light measuring device 12a... Output part 12b... Receiving part 20... Inspection part 21... Input part 22... Camera image inspection part 23... Laser image inspection part 24... Penetrant flaw test image inspection part 25 Magnetic particle test image inspection unit 26 Ultrasonic test image inspection unit 27 Radiation penetration test image inspection unit 28 Output unit 50 Mounting device 51 Rail base 511, 511a, 511b Rail base body 512 Rail guide 512a Groove 513 Fixed shaft 52 Ring rail gear 521 First member 522 Second member 53 Punch lock 54 Motor/encoder 541 DC motor 542 Gear 55 Mounting base 56 Mounting base 561 Angle adjusting mechanism DESCRIPTION OF SYMBOLS 58... Device holding leg 581... Slide part 582... Fixed part 101... 1st piping 102... 2nd piping W... Welding part

Claims (9)

A camera for photographing the welded portion of the first pipe and the second pipe connected by welding, and irradiating the welded portion with a laser beam and photographing the laser beam reflected by the welded portion to create an image. a laser light measurement device;
A rail base arranged around the first pipe or the second pipe, a ring rail gear moving along the rail base around the welded portion, and the camera and the laser mounted on the ring rail gear. a mounting device having a mounting base for fixing the optical measurement device;
an inspection unit that determines the presence or absence of welding defects in the weld based on images captured by the camera and the laser beam measurement device;
Inspection device for pipe welds. The inspection unit receives supervised data obtained by adding information indicating whether there is a welding defect or not to image information indicating the image of the weld taken by the camera, and the image of the weld taken by the laser beam measuring device. Using supervised data obtained by adding information on whether there is a welding defect or not to the image information shown, and based on an analysis program using artificial intelligence, the presence or absence of a welding defect in the weld is determined.
The apparatus for inspecting welded portions of pipes according to claim 1 . The camera and the laser light measurement device are fixed at a predetermined position on the ring rail gear, and the laser light measurement device creates an image of the same position as the position of the welded part photographed by the camera.
3. The apparatus for inspecting welded portions of pipes according to claim 1 or 2. The inspection unit includes image information indicating an image of the designated portion of the welded portion captured by the camera, and the laser light measurement device irradiating the designated portion of the welded portion with a laser beam to perform the welding. Based on an analysis program using artificial intelligence, using supervised data that adds information on whether there is a welding defect or not to the image information that shows the image of the laser light reflected by the part and the information that is linked. to determine the presence or absence of welding defects in the weld,
4. The apparatus for inspecting welded portions of pipes according to claim 3. The inspection unit adds the image information indicating the image of the welded portion photographed by the camera and the image information indicating the image of the welded portion photographed by the laser light measurement device to each of the image information indicating the welding of the welded portion with respect to the image information. adding information to the supervised data to which information on the determination result of the presence or absence of a defect is added, and determining the presence or absence of a welding defect in subsequent images of the welded portion;
5. The apparatus for inspecting welded portions of pipes according to claim 2 or 4. The inspection unit
Image information showing an image of the weld zone developed with a test solution that has permeated the weld zone by a penetrant test, and image information showing an image of the weld zone represented by magnetic particles adhering to the weld zone by a magnetic particle penetrant test. , image information indicating an image of the welded portion by ultrasonic waves reflected by the welded portion measured by an ultrasonic flaw detection test, and image information indicating an image of the welded portion by a photosensitive film obtained by a radiographic test. has an input unit for
The inspection unit
Image information indicating an image of the weld taken by the camera, image information indicating the image of the weld taken by the laser beam measurement device, image information indicating the image of the weld acquired by a penetrant test, magnetic powder Image information indicating an image of the weld obtained by penetrant testing, image information indicating an image of the weld obtained by ultrasonic testing, and image information indicating an image of the weld obtained by radiographic testing. Using at least any two pieces of information as supervised data to determine the presence or absence of welding defects in the weld based on an analysis program using artificial intelligence;
The apparatus for inspecting a pipe welded portion according to claim 2. the ring rail gear is driven by a motor-encoder to move along the rail platform and around the weld;
The apparatus for inspecting welded portions of pipes according to claim 1 . The mounting base has an adjustment mechanism that adjusts an imaging angle of the camera and the laser light measurement device with respect to the welded part,
The apparatus for inspecting welded portions of pipes according to claim 1 . The supervised data is created by adding information indicating whether there is a welding defect or not to each image information showing a plurality of images of the welded portion having different undercuts, reinforcement heights, and leg lengths. to be
5. The apparatus for inspecting welded portions of pipes according to claim 2 or 4 .

Images