Crack detection device
The crack detection device uses a pointed member and accelerometer to quantify crack presence by measuring acceleration, overcoming inspector variability and improving crack detection reliability through machine-learned threshold values.
Patent Information
- Application Number
- JP2024005447
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-01-17
- Publication Date
- 2025-07-30
AI Technical Summary
Existing crack detection methods are inadequate for accurately and quantitatively evaluating the presence of minute cracks on object surfaces, leading to inconsistent results due to inspector skill variations and the inability to differentiate cracks from other surface defects.
A crack detection device equipped with a pointed member, a moving device, a biasing member, and an accelerometer, which measures acceleration to detect cracks by the pointed member's response to surface irregularities, utilizing machine learning to set threshold values for accurate crack detection.
Enables quantitative and reliable crack detection by automatically determining crack presence and position based on acceleration measurements, reducing user subjectivity and enhancing detection accuracy.
Smart Images

Figure 2025111190000001_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a crack detection device.
Background Art
[0002] For example, when performing welding such as build-up welding on a press die during the production or design change of the press die, minute cracks may remain at the welding boundary (the boundary between the welded part and the die) after removing the build-up part after welding. Cracks of this size do not cause much problem when using the press die that has been welded as it is, but when the press die is plated to improve the hardness of the forming surface, there is a risk that the above cracks will spread due to the heat treatment as part of the plating process. The spread of cracks not only reduces the strength of the die, but also there is a risk that marks reflecting the shape of the cracks that have spread to the surface of the press-formed product will remain. Therefore, this type of crack should be avoided as much as possible from the viewpoint of appearance quality.
[0003] As means for detecting the presence or absence of cracks occurring in the welded part, various methods have been proposed. For example, Patent Document 1 proposes a crack detection device capable of detecting the occurrence or progression of cracks by detecting elastic waves generated in a bridge structure due to the occurrence or progression of cracks in the welded part of the bridge structure.
[0004] Alternatively, Patent Document 2 discloses a method for detecting the presence or absence of cracks in a welded part of a structure such as a ship constructed by welding metal plates, based on the magnitude of a numerical value obtained by performing a predetermined process on a signal obtained by scanning an eddy current sensor along a weld bead.
Prior Art Documents
Patent Documents
[0005]
Patent Document 1
Patent Document 2
SUMMARY OF THE INVENTION
PROBLEMS TO BE SOLVED BY THE INVENTION
[0006] As described in Patent Document 1, a method of detecting the presence or absence of cracks by detecting a phenomenon (generation of elastic waves) occurring during the generation or progression of cracks is not suitable for detecting the presence or absence of existing cracks. Further, as described in Patent Document 2, a method of detecting the presence or absence of cracks based on the reaction when a detection signal such as an eddy current is emitted toward a location where the presence of a crack is suspected detects not only cracks opening on the die surface but also defects (such as voids) existing inside the die, and thus is not suitable as a means for detecting only cracks on the die surface.
[0007] Therefore, currently, the presence or absence of minute cracks is checked by a method of directly touching them by hand (whether or not a fingernail gets caught), but in this case, problems such as differences in inspection results depending on the skill and sense of the inspector may occur. Although crack detection is also performed using a crack detection means called a flaw detection method, with this method, since all gaps into which paint penetrates are colored, it is difficult to accurately detect cracks at the welding boundary.
[0008] The above-described problems are not limited to welded parts and can occur in all objects for which it is necessary to detect the presence or absence of cracks occurring on the object surface.
[0009] In view of the above circumstances, in this specification, the technical problem to be solved is to enable quantitative and accurate evaluation of the presence or absence of cracks on the object surface.
MEANS FOR SOLVING THE PROBLEM
[0010] The solution to the above problem is achieved by the crack detection device according to the present invention. That is, this evaluation device is a device for detecting cracks in an object, and is characterized by including a pointed member having a pointed shape toward the tip, a moving device capable of moving the pointed member along the surface of the object, a biasing member capable of biasing the pointed member toward the surface of the object, and an accelerometer capable of measuring the acceleration in the direction of the biasing of the pointed member.
[0011] When the crack detection device according to the above configuration is used to detect cracks, the following operational effects can be obtained. That is, when the pointed member is moved by the moving device with the tip of the pointed member in contact with the surface of the object, the pointed member moves in a state of being biased against the surface of the object by the biasing member. Therefore, when there is no crack on the trajectory through which the pointed member passes, the pointed member moves along the surface while moving up and down following the surface of the object. On the other hand, when the pointed member passes over a crack, since there is no object receiving the biasing force acting on the pointed member directly above the crack, the pointed member accelerates in the biasing direction, and the tip of the pointed member fits into the crack. Even when passing through, for example, a concave or convex portion on the surface, the pointed member moves up and down, but during that time, the tip of the pointed member is always in contact with the surface of the object. Therefore, the possibility that the acceleration of the pointed member fluctuates greatly is extremely low. From the above, according to the present invention, the presence or absence of cracks can be quantitatively and accurately evaluated, thereby enhancing the reliability of the crack detection result.
[0012] Further, the crack detection device according to the present invention may further include a determination unit that determines the presence or absence of a crack at the point where the acceleration is measured based on the value of the acceleration of the pointed member measured by the accelerometer.
[0013] According to the crack detection device of the present invention, based on the value of the acceleration of the pointed member, the presence and position of the crack can be quantitatively evaluated. Therefore, by providing a determination unit that determines the presence or absence of a crack based on the value of the acceleration, it becomes possible to automatically detect the presence or absence of a crack and its position. Thereby, it becomes possible to further enhance the reliability of the crack detection result.
[0014] In addition, when the crack detection device includes a determination unit, in the crack detection device according to the present invention, the determination unit learns, by machine learning, the magnitude of the acceleration of the pointed member measured by the accelerometer when a crack exists on the surface of the object, obtains an acceleration threshold value when it is determined that a crack exists, and is configured to determine that a crack exists when the value of the acceleration measured by the accelerometer exceeds the threshold value.
[0015] In this way, if the determination unit obtains the acceleration threshold value when it is determined that a crack exists by machine learning, and determines the presence or absence of a crack based on this threshold value and the measured value of the acceleration, even if the type of the object to be determined changes, the acceleration threshold value can be automatically obtained by machine learning each time, and the presence or absence of a crack in the object can be automatically and continuously determined using the obtained threshold value. As a result, it is possible to further eliminate the room for the user's subjectivity to enter, and it is possible to more universally evaluate the presence or absence of a crack.
Advantages of the Invention
[0016] As described above, according to the crack detection device of the present invention, it is possible to quantitatively and accurately evaluate the presence or absence of a crack on the surface of the object.
Brief Description of the Drawings
[0017]
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Best Mode for Carrying Out the Invention
[0018] Hereinafter, the content of the crack detection device according to an embodiment of the present invention will be described with reference to the drawings.
[0019] FIG. 1 shows the overall configuration of a crack detection device 10 according to an embodiment of the present invention. This crack detection device 10 is for detecting a crack 2 (see FIGS. 3 and the like described later) present on the surface 1a of the object 1, and includes a pointed member 11, a moving device 12, a biasing member 13, an accelerometer 14, and a determination unit 15. In the present embodiment, the crack detection device 10 further includes a fixing portion 16 to the object 1.
[0020] The pointed member 11 has a shape that tapers toward the tip. In the present embodiment, as a whole, it has a columnar shape and its tip portion 11a has a conical shape. Further, the angle of the outer peripheral surface of the tip portion 11a of the pointed member 11 (the angle with respect to the center line of the generatrix forming the conical surface in the case of a conical shape) can be set without particular limitation. However, if it is too large, the fitting according to the size (depth) of the crack 2 will be hindered, and if it is too small, there is a risk of hindering the pointed member 11 from coming out of the crack 2 after fitting into the crack 2. Therefore, in practice, it is preferable to set the angle of the outer peripheral surface of the tip portion 11a so as to satisfy the above two conditions. As an example, the angle of the outer peripheral surface of the tip portion 11a is set within the range of 10° to 30°.
[0021] The moving device 12 enables the pointed member 11 to move along the surface 1a of the object 1. In the present embodiment, it includes a rotational drive unit 17, a linear motion conversion mechanism 18 that converts the rotational driving force generated by the rotational drive unit 17 into a linear motion, and a support portion 19 that is attached to the linear portion 18a of the linear motion conversion mechanism 18 and supports the pointed member 11 so as to be movable in a direction orthogonal to the linear direction Y of the linear portion 18a.
[0022] Here, the rotational drive unit 17 is, for example, a motor, and is rotationally driven by electric power supplied from a power source (not shown). Also, the linear conversion mechanism 18 is, for example, a ball screw mechanism, and a linear part 18a including a nut part attached to the outer periphery of a screw part 18b is configured to be slidable in the longitudinal direction of the screw part 18b as the screw part 18b rotates about its axis.
[0023] The support part 19 is disposed at a position shifted in the planar direction from the screw part 18b of the linear conversion mechanism 18 (see FIG. 2). Also, in the present embodiment, the support part 19 is generally cylindrical, and includes a first small-diameter part 19a that can insert and support the pointed member 11 at the base end side (the upper side in FIG. 1), a second small-diameter part 19b that can insert and support the pointed member 11 at the tip end side, and a large-diameter part 19c that has a large inner peripheral surface with respect to the first and second small-diameter parts 19a and 19b and can accommodate the compression spring 20 as the biasing member 13 between it and the pointed member 11, integrally. In this case, a flange part 11b is integrally provided on the outer periphery of the pointed member 11, and the compression spring 20 is disposed between this flange part 11b and the first small-diameter part 19a. In this case, by disposing the pointed member 11 on the surface 1a of the object 1 (in the present embodiment, by fixing the pointed member 11 and the moving device 12 to a predetermined height position with respect to the surface 1a of the object 1 by the fixing part 16), the tip end part 11a of the pointed member 11 is always pressed against the surface 1a of the object 1 so that the compression spring 20 is always in a compressed state (see FIG. 3(a)). Note that, in the present embodiment, when the tip end part 11a of the pointed member 11 is not in contact with the surface 1a of the object 1, the natural length of the compression spring 20 is adjusted to a predetermined length so that the flange part 11b of the pointed member 11 is in contact with the second small-diameter part 19b in the axial direction as shown in FIG. 1.
[0024] The accelerometer 14 is fixed to the pointed member 11 and is configured to be movable integrally with the pointed member 11. Here, as the accelerometer 14, any type of accelerometer can be applied as long as it can measure at least the acceleration in the direction along the longitudinal direction of the pointed member 11 (more precisely, the acceleration a in the direction along the biasing direction).
[0025] The determination unit 15 is electrically connected to at least the accelerometer 14, and is configured to be able to determine the presence or absence of the crack 2 at the point where the acceleration is measured based on the value of the acceleration a in the longitudinal direction of the pointed member 11 measured by the accelerometer 14.
[0026] In the present embodiment, the determination unit 15 performs machine learning on the relationship between the acceleration a of the pointed member 11 and the crack 2 (machine learning step S1), sets a threshold value Ta of the acceleration a related to the occurrence of the crack 2 using the result of the machine learning (threshold value setting step S2), and is configured to determine the presence or absence of the crack 2 using the set threshold value Ta (determination step S3).
[0027] First, in the machine learning step S1, machine learning is performed to learn the magnitude of the acceleration a of the pointed member 11 measured by the accelerometer 14 when the crack 2 exists on the surface 1a of the object 1. In this case, by learning the acceleration data when the crack 2 exists (especially the data related to the maximum value of the acceleration a) and the acceleration data when the crack 2 does not exist, the magnitude of the acceleration a of the pointed member 11 when the crack 2 exists is learned.
[0028] Then, in the threshold value setting step S2, based on the magnitude of the acceleration a when the crack 2 exists learned in the machine learning step S1, a threshold value Ta of the acceleration a when it is determined that the crack 2 exists is set. Here, the "case where the crack 2 exists" means "there is a crack 2 of a size to be detected". Therefore, for example, even if a crack 2 of an order size that does not pose a problem for the object 1 (such as a press die) is detected from the value of the acceleration a, the crack 2 is not considered when setting the threshold value Ta. It is preferable to set the threshold value Ta of the acceleration a based on the value of the acceleration a measured when there is a crack 2 of a size to be detected.
[0029] After setting the threshold value Ta of the acceleration a in this way, in the determination step S3, based on the value of the acceleration a measured when the pointed member 11 moves along the surface 1a of the object 1 to be actually evaluated, it is determined whether or not a crack 2 exists in the movement range of the pointed member 11 (on the locus of the contact point between the tip 11a and the surface 1a). Specifically, first, the value of the acceleration a of the pointed member 11 measured by the accelerometer 14 is compared with the threshold value Ta. And when the measured value of the acceleration a exceeds the threshold value Ta, it is determined that a crack 2 exists at the point where the acceleration a was measured. Alternatively, when the measured value of the acceleration a is equal to or less than the threshold value Ta, it is determined that no crack 2 exists at the point where the acceleration a was measured.
[0030] The fixing part 16 is for fixing at least the part of the crack detection device 10 that is mechanically connected to the pointed member 11 (in this embodiment, the moving device 12, the biasing member 13, and the accelerometer 14) to the object 1, and is composed of a magnet in this embodiment. Thereby, it becomes possible to set the starting position and the height position of the movement of the pointed member 11 with respect to the surface 1a of the object 1 (see Fig. 3(a)). Of course, the magnet is just an example. For example, when the object 1 is not a magnetic material such as a steel material, an adsorption mechanism may be adopted, or when the object 1 is plate-shaped, a clamp mechanism may be adopted, etc. The fixing part 16 may be configured according to the shape and material of the object 1.
[0031] Next, an example of the crack detection method using the crack detection device 10 with the above configuration will be mainly described with reference to Figs. 3 to 5.
[0032] First, the crack detection device 10 is fixed to the object 1 by the fixing part 16. At this time, the tip 11a of the pointed member 11 abuts on the surface 1a of the object 1, and the crack detection device 10 is fixed so that the pointed member 11 is positioned above compared to the abutted state. Thereby, a downward biasing force f (elastic restoring force) is generated in the compression spring 20, and the pointed member 11 is in a state of being biased toward the surface 1a (see Fig. 3(a)).
[0033] With the biasing force f acting on the pointed member 11 toward the surface 1a of the object 1 in this manner, the pointed member 11 is moved along the surface 1a by the moving device 12. In the present embodiment, a rotational driving force is transmitted to the screw portion 18b of the rectilinear conversion mechanism 18 by the driving of the rotational driving portion 17, and the transmitted rotational driving force is transmitted from the screw portion 18b to the rectilinear portion 18a in a state where it is converted into a rectilinear force. As a result, the rectilinear portion 18a moves straight in a predetermined direction, and the pointed member 11 moves together with the rectilinear portion 18a via a support portion 19 (see FIG. 2) provided on the rectilinear portion 18a.
[0034] At this time, since the pointed member 11 is in a state where a downward biasing force f is applied by the biasing member 13 (the compressed spring 20 in a compressed state), as it moves in the straight-ahead direction Y (see FIG. 1), the tip portion 11a of the pointed member 11 moves in the straight-ahead direction Y while maintaining the state of being in contact with the surface 1a of the object 1. Therefore, the pointed member 11 moves in the straight-ahead direction Y while moving up and down following the shape of the surface 1a (see FIG. 3(a)).
[0035] Here, when the pointed member 11 passes over the crack 2 present on the surface 1a, on the crack 2, the tip portion 11a of the pointed member 11 is in a state of not being in contact with any object, so the biasing force f from the compressed spring 20 cannot be received by the object 1, and as a result, the pointed member 11 fits into the crack 2 while accelerating downward (see FIG. 4). At this time, the acceleration a of the pointed member 11 measured by the accelerometer 14 increases rapidly (see the corresponding part in FIG. 3(b)).
[0036] On the other hand, when the pointed member 11 passes over the recess 3 present on the surface 1a, since the pointed member 11 is always biased downward, the pointed member 11 moves downward as it advances over the recess 3, but during that time, the contact state between the tip portion 11a and the surface of the recess 3 is maintained (see FIG. 5). Therefore, while passing over the recess 3, the acceleration a of the pointed member 11 does not increase rapidly (hardly fluctuates).
[0037] As described above, the determination unit 15 compares the value of the acceleration a of the pointed member 11 measured by the accelerometer 14 with the threshold value Ta set in advance in the threshold setting step S2. When the value of the acceleration a exceeds the threshold value Ta, it is determined that a crack 2 exists at the point (position P1) where the acceleration a was measured (see the left - hand parts of FIGS. 3(a) and (b)). Further, when the measured value of the acceleration a is less than the threshold value Ta, it is determined that no crack 2 exists at the point (positions P2, P3) where the acceleration a was measured, or even if a crack 2 exists, it is only an extremely minute crack 2(2′) that does not pose a problem (see the right - hand parts of FIGS. 3(a) and (b)). In this way, the detection of the crack 2 on the surface 1a of the object 1 is automatically performed.
[0038] As described above, according to the crack detection device 10 according to the present embodiment, the following operational effects can be obtained. That is, when the pointed member 11 is moved by the moving device 12 with the tip 11a of the pointed member 11 in contact with the surface 1a of the object 1, the pointed member 11 moves while maintaining the state of being biased against the surface 1a of the object 1 by the biasing member 13 (compression spring 20). Therefore, when there is no crack 2 on the trajectory through which the pointed member 11 passes, the pointed member 11 moves along the surface 1a while moving up and down following the surface 1a of the object 1 (see FIG. 3(a)). On the other hand, when the pointed member 11 passes over the crack 2, since there is no object receiving the biasing force f acting on the pointed member 11 directly above the crack 2, the pointed member 11 instantaneously accelerates in the biasing direction, and the tip 11a of the pointed member 11 fits into the crack 2 (see FIG. 4). When passing through, for example, the concave portion 3 on the surface 1a, the pointed member 11 also moves up and down, but during that time, the tip 11a of the pointed member 11 is always in contact with the surface 1a of the object 1 (see FIG. 5). Therefore, the acceleration a of the pointed member 11 does not fluctuate significantly. As described above, according to the crack detection device 10 according to the present embodiment, based on the value of the acceleration a of the pointed member 11, it is possible to quantitatively and accurately evaluate the presence or absence of the crack 2.
[0039] In addition, in the present embodiment, a determination unit 15 is further provided that automatically determines the presence or absence of the crack 2 based on the measured value of the acceleration a. The determination unit 15 sets a threshold value Ta of the acceleration a when it is determined by machine learning that there is a crack 2 on the surface 1a of the object 1. When the measured value of the acceleration a exceeds the threshold value Ta, it is determined that there is a crack 2 at the point where the acceleration a is measured. In this way, if the determination unit 15 sets the threshold value Ta of the acceleration a when it is determined by machine learning that there is a crack 2, and determines the presence or absence of the crack 2 based on the threshold value Ta and the measured value of the acceleration a, then even if the type of the object to be determined (object 1) changes, the threshold value Ta of the acceleration can be automatically obtained by machine learning each time, and the presence or absence of the crack 2 in the object 1 can be automatically and continuously determined using the obtained threshold value Ta. As a result, it is possible to further eliminate the room for the user's subjectivity to enter, and it is possible to more generally evaluate the presence or absence of the crack 2.
[0040] As described above, one embodiment of the present invention has been described. However, the crack detection device according to the present invention can also adopt configurations other than the above within the scope not departing from the gist of the present invention.
[0041] For example, in the above embodiment, the case where the moving device 12 is composed of a rotation driving unit 17, a ball screw mechanism as a linear motion conversion mechanism 18, and a support unit 19 is illustrated. However, of course, other configurations can also be adopted. For example, although not shown in the figure, the moving device 12 may be configured such that the support unit 19 can be linearly moved along a predetermined direction by a cylinder. However, in any case, it is important to control the moving speed in the linear direction Y to such an extent that the tip 11a of the pointed member 11 in the energized state can maintain the contact state with the surface 1a of the object 1.
Explanation of Reference Numerals
[0042] 1 Object 1a Surface 2 Crack 3 Recess 10 Crack detection device 11 Pointed member 11a Tip 11b flange portion 12 moving device 13 biasing member 14 accelerometer 15 determination unit 16 fixing portion 17 rotational drive unit 18 linear conversion mechanism 18a linear portion 18b screw portion 19 support portion 20 compression spring a acceleration of the pointed member Ta acceleration threshold value f biasing force Y linear direction
Claims
【Claim 1】 An apparatus for detecting cracks in an object, comprising: a pointed member having a pointed shape towards the tip; a moving device configured to move the pointed member along the surface of the object; a biasing member configured to bias the pointed member towards the surface of the object; and an accelerometer configured to measure an acceleration in a direction along the biasing direction of the pointed member, the crack detection device.
Citation Information
Patent Citations
Method and apparatus for crack flaw detection of welded part
JP2000111530A
Detection device, detection system, detection method, and information processing device
JP2020112396A