Processing device, processing system, processing method, and storage medium

Abstract

A processing system of an embodiment is provided with a processing device. The processing device receives a detection result of a reflected wave from a detector including a plurality of detection elements arranged in a first direction and a second direction intersecting each other and performs a probe including transmission of an ultrasonic wave toward a welding object and detection of the reflected wave. The processing device performs a first determination of determining joint and non-joint at a plurality of points of the welding object along the first direction and the second direction based on the detection result. The processing device performs a second determination of determining whether a result of the first determination is appropriate based on the detection result or a result of the first determination.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to a processing device, a processing system, a processing method, and a storage medium. BACKGROUND

[0002] In welding, a part of two or more members is fused and joined to each other. The welded members are inspected whether the welded part (hereinafter, referred to as a welding portion) is properly joined. For example, in non-destructive inspection, a person (an inspector) who holds a detector brings the detector into contact with the welding portion. An ultrasonic wave is transmitted from the detector to the welding portion, and data related to the welding object is derived based on a reflected wave thereof. For non-destructive inspection, a technique capable of improving the accuracy of the data related to the welding object is required.

[0003] PRIOR ART DOCUMENTS

[0004] PATENT DOCUMENTS

[0005] Patent Document 1: Japanese Patent Application Publication No. 2019-90727

[0006] NON-PATENT DOCUMENTS

[0007] Non-Patent Document 1: Shigeo Ushijima, Shinsho Saito, and Makoto Matsumoto (2019) “Spot Welding Inspection Robot Contributing to Labor Saving and Reliability Improvement in Non-Destructive Inspection” Toshiba Review, vol. 74, No. 4, pp. 25-28 SUMMARY

[0008] PROBLEMS TO BE SOLVED BY THE INVENTION

[0009] The present application relates to a processing system, a processing method, a program, and a storage medium capable of improving the accuracy of data related to a welding object.

[0010] MEANS FOR SOLVING THE PROBLEMS

[0011] The processing system of the embodiment has a processing device. The processing device receives a detection result of a reflected wave from a detector including a plurality of detection elements arranged in a first direction and a second direction intersecting each other and performing a probe including transmission of an ultrasonic wave toward a welding object and detection of the reflected wave. The processing device performs a first determination of determining joining and non-joining at a plurality of points of the welding object along the first direction and the second direction based on the detection result. The processing device performs a second determination of determining whether a result of the first determination is appropriate based on the detection result or a result of the first determination. BRIEF DESCRIPTION OF DRAWINGS

[0012] Figure 1 is a block diagram showing the structure of the processing system of the embodiment.

[0013] Figure 2 is a schematic view showing a case of non-destructive inspection.

[0014] Figure 3 is a schematic view showing the internal structure of the front end of the detector.

[0015] Figure 4 is a schematic view for explaining the processing of the processing system of the embodiment.

[0016] Figure 5 is an example of an image obtained by the processing system of the embodiment.

[0017] Figure 6 is a graph for explaining the processing of the processing system of the embodiment.

[0018] Figure 7 is an example of an image showing the processing result of the processing system of the embodiment.

[0019] Figure 8 is a schematic view showing an image based on a detection result of a reflected wave.

[0020] Figure 9 is a schematic view showing an image based on a detection result of a reflected wave.

[0021] Figure 10 is a graph for explaining the second determination.

[0022] Figure 11 is a flowchart showing an inspection flow using the processing system of the embodiment.

[0023] Figure 12 is a flowchart showing an inspection flow using the processing system of the embodiment.

[0024] Figure 13is a graph illustrating an intensity distribution of a reflected wave in the Z direction in one cross section.

[0025] Figure 14 is a graph illustrating an intensity distribution of a reflected wave in the Z direction.

[0026] Figure 15 is a graph illustrating a result of filtering an intensity distribution of a reflected wave.

[0027] Figure 16 is a schematic view illustrating a detection result of a reflected wave.

[0028] Figure 17 is an example of an intensity distribution of a reflected wave in the X-Y plane.

[0029] Figure 18 is a schematic view illustrating a detection result of a reflected wave.

[0030] Figure 19 is a flowchart showing a flow of estimation of a range in a processing system of an embodiment.

[0031] Figure 20 is an image illustrating a detection result of a reflected wave.

[0032] Figure 21 is a graph for explaining a process of a processing system of an embodiment.

[0033] Figure 22 is an example of an image obtained by a processing system of an embodiment.

[0034] Figure 23 is a schematic view showing a result of classification based on a first model.

[0035] Figure 24 is a schematic view showing a structure of a processing system of a modification example of an embodiment.

[0036] Figure 25 is a perspective view showing a part of a processing system of a modification example of an embodiment.

[0037] Figure 26 is a flowchart showing an action of a processing system of a modification example of an embodiment.

[0038] Figure 27 is a block diagram showing a hardware configuration of a system. DETAILED DESCRIPTION

[0039] Hereinafter, each embodiment of the present application will be explained with reference to the attached drawings.

[0040] The drawings are schematic or conceptual; the proportions among the thickness of the respective portions, the ratios of the sizes of the portions, and so on are not necessarily the same as those in reality. Even when the same portions are represented, the sizes, the ratios, and so on can be represented differently from one another depending on the drawings.

[0041] In the present application specification and the respective drawings, the same reference numerals are applied to the same elements already described, and a detailed description is appropriately omitted.

[0042] Figure 1 is a block diagram showing the structure of the processing system of the embodiment.

[0043] As Figure 1 shown, the processing system 100 of the embodiment is provided with a processing device 110 and a storage device 120. The storage device 120 stores data related to welding inspection. The processing device 110 processes the data related to welding inspection.

[0044] Figure 1 The processing system 100 shown further includes a detector 130, an input device 140, and a display device 150. The detector 130 transmits an ultrasonic wave to an object and detects (receives) a reflected wave thereof. The detector 130 includes, for example, a probe. Hereinafter, the transmission of the ultrasonic wave and the detection of the reflected wave by the detector 130 are referred to as probing.

[0045] The processing device 110 performs various processes based on the detected reflected wave. The processing device 110 causes the display device 150 to display a user interface. The user can easily confirm the data obtained by the processes through the user interface displayed on the display device 150. The user can input data to the processing device 110 via the user interface using the input device 140.

[0046] The processing device 110 is connected to the storage device 120, the detector 130, the input device 140, and the display device 150 via wired communication, wireless communication, or a network.

[0047] Here, welding inspection is specifically described. In the welding inspection, non-destructive inspection of a welded portion is performed.

[0048] Figure 2 is a schematic view showing a case of non-destructive inspection.

[0049] The detector 130 includes a plurality of detection elements for inspecting a welded portion. The detector 130 includes, for example, as Figure 2As shown, the detector 130 has a shape that can be held by a hand of a person. A person holding the detector 130 brings the front end of the detector 130 into contact with the weld 13, and inspects the weld 13. Here, an example in which a person holds the detector 130 and performs the weld inspection will be described. Hereinafter, the person (for example, an inspector) who holds the detector 130 and performs the weld inspection will be referred to as a user.

[0050] Figure 3 is a schematic view showing the internal structure of the front end of the detector.

[0051] As shown in Figure 3 , an element array 131 including a plurality of detection elements 132 is provided inside the front end of the detector 130. The detection elements 132 are, for example, transducers. Each detection element 132 emits an ultrasonic wave of, for example, a frequency of 1 MHz or more and 100 MHz or less. The plurality of detection elements 132 are arranged in a first direction and a second direction that intersect each other. In the example shown in Figure 3 , the plurality of detection elements 132 are arranged in an X direction and a Y direction that are orthogonal to each other.

[0052] The element array 131 is covered by, for example, a hard propagation member 133. The hard propagation member 133 is located between the element array 131 and the weld 13 when the front end of the detector 130 is brought into contact with the weld 13. The hard propagation member 133 is made of a resin material or the like through which an ultrasonic wave easily propagates. By providing the hard propagation member 133 corresponding to the shape of the surface of the weld 13, the ultrasonic wave easily propagates into the inside of the weld 13. By the hard propagation member 133, deformation, damage, or the like of the element array 131 when the detector 130 is brought into contact with the weld 13 can be suppressed. In order to suppress the deformation, damage, or the like when the detector 130 is brought into contact with the weld 13, the hard propagation member 133 has sufficient hardness.

[0053] Figure 2 and Figure 3 show a situation in which the member 10 that is a subject of welding is inspected. The member 10 is made by spot welding the metal plate 11 (first member) and the metal plate 12 (second member) at the weld 13. As shown in Figure 3 , in the weld 13, a portion of the metal plate 11 and a portion of the metal plate 12 are melted, and a solidified portion 14 in which the melted portions are mixed and solidified is formed.

[0054] For example, in the inspection, it is investigated whether the weld 13 is formed. In the inspection, the diameter of the weld 13, whether the diameter is sufficient, or the like is investigated. At the time of the inspection, a coupling agent 15 is applied to the surface of the subject in such a manner that an ultrasonic wave easily propagates between the subject and the detector 130. Each detection element 132 transmits an ultrasonic wave US to the member 10 on which the coupling agent 15 is applied, and receives a reflected wave RW from the member 10.

[0055] Alternatively, instead of the coupling agent 15, a soft propagating component that easily transmits ultrasonic waves can be provided at the front end of the detector 130. This soft propagating component is softer than the hard propagating component 133. When in contact with the weld 13, the soft propagating component deforms to conform to the shape of the surface of the weld 13. The soft propagating component is, for example, made of a gel-like resin.

[0056] For example, such as Figure 3 As shown, a detection element 132 sends an ultrasonic wave US toward the welded part 13. A portion of the ultrasonic wave US is reflected by the upper or lower surface of the component 10. Multiple detection elements 132 respectively receive (detect) the reflected wave RW. Each detection element 132 sequentially sends an ultrasonic wave US, and each reflected wave RW is detected by the multiple detection elements 132.

[0057] Upon receiving the detection result of the reflected wave, the processing device 110 performs the following first determination and second determination. In the first determination, the processing device 110 determines whether each point of the welding object is joined based on the obtained detection result. In the second determination, the processing device 110 determines whether the result of the first determination is appropriate based on the detection result of the reflected wave or the result of the first determination.

[0058] The first and second determinations will be explained in detail below.

[0059] (First Judgment)

[0060] Figure 4 This is a schematic diagram illustrating the processing of the processing system used to explain the implementation method.

[0061] like Figure 4 As shown in (a), a portion of the ultrasonic wave US is reflected by the upper surface 11a of the metal plate 11 or the upper surface 13a of the weld 13. Another portion of the ultrasonic wave US is incident on the component 10 and reflected by the lower surface 11b of the metal plate 11 or the lower surface 13b of the weld 13.

[0062] The upper surface 11a, upper surface 13a, lower surface 11b, and lower surface 13b are all at different positions in the Z direction. That is, the distances in the Z direction between these surfaces and the detection element 132 are all different. When the detection element 132 receives reflected waves from these surfaces, it detects the peak values ​​of the reflected wave intensity. After transmitting the ultrasonic wave US, by calculating the time until each peak value is detected, it is possible to investigate on which surface the ultrasonic wave US is reflected.

[0063] Figure 4 (b) and Figure 4 (c) is a graph illustrating the relationship between the time after the transmission of the ultrasonic wave US and the intensity of the reflected wave RW. Figure 4 (b) andFigure 4 In (c), the vertical axis represents the elapsed time after the ultrasonic wave US is transmitted. The horizontal axis represents the intensity of the detected reflected wave RW. Here, the intensity of the reflected wave RW is expressed in absolute value. Figure 4 The graph in (b) illustrates the detection results of the reflected wave RW from the upper surface 11a and lower surface 11b of the metal plate 11. Figure 4 The graph in (c) illustrates the detection results of the reflected wave RW from the upper surface 13a and lower surface 13b of the welded part 13.

[0064] exist Figure 4 In the curve (b), the first peak Pe11 is based on the reflected wave RW from the upper surface 11a. The second peak Pe12 is based on the reflected wave RW from the lower surface 11b. The times when peaks Pe11 and Pe12 are detected correspond to the Z-direction positions of the upper surface 11a and lower surface 11b of the metal plate 11, respectively. The time difference TD1 between the times when peaks Pe11 and Pe12 are detected corresponds to the Z-direction distance Di1 between the upper surface 11a and the lower surface 11b.

[0065] Similarly, in Figure 4 In the curve (c), the first peak Pe13 is based on the reflected wave RW from the upper surface 13a. The second peak Pe14 is based on the reflected wave RW from the lower surface 13b. The times when peaks Pe13 and Pe14 are detected correspond to the Z-direction positions of the upper surface 13a and lower surface 13b of the welded part 13, respectively. The time difference TD2 between the times when peaks Pe13 and Pe14 are detected corresponds to the Z-direction distance Di2 between the upper surface 13a and the lower surface 13b.

[0066] The processing device 110 determines whether the time difference between peak values ​​corresponds to the thickness of the weld portion 13. If it determines that the time difference between peak values ​​corresponds to the thickness of the weld portion 13, it determines that the point is joined. For example, the processing device 110 compares the time difference between peak values ​​at each point in the XY plane with a preset threshold. When the time difference is above the threshold, the processing device 110 determines that the point is joined. When the time difference is less than the threshold, the processing device 110 determines that the point is not joined. The threshold is set based on the thickness of the weld portion 13. Alternatively, a range can be set instead of a threshold. The processing device 110 determines that the point is joined when the time difference is within this range.

[0067] Furthermore, the intensity of the reflected wave can be represented in any way. For example, the intensity of the reflected wave output from the detection element 132 may include positive and negative values ​​depending on the phase. Various processing can also be performed based on the reflected wave intensity including positive and negative values. The reflected wave intensity including positive and negative values ​​may also be converted to absolute values. The average value of the reflected wave intensity at each time step may also be subtracted from the reflected wave intensity at each time step. Alternatively, a weighted average, a weighted moving average, or the like may be subtracted from the reflected wave intensity at each time step. Even when using the results of these processing applied to the reflected wave intensity, the various processing described in this application can still be performed.

[0068] Figure 5 This is an example of an image obtained by the processing system of the implementation method.

[0069] Figure 5 It is an image depicted based on the detection results of reflected waves. In Figure 5 In the diagram, the brightness (white) of each point represents the intensity of the reflected wave at that point. For example... Figure 5 As shown, three-dimensional volume data is obtained as the detection result of the reflected wave. Based on the detection result of the reflected wave, the joint or non-joint status of each point of the welding object is determined.

[0070] Figure 6 This is a diagram illustrating the processing of the processing system in the implementation method.

[0071] Figure 6 (a) is a schematic top view showing the vicinity of weld 13. Based on the reflected wave detected by detector 130, a determination is made, for example... Figure 6 Whether the points in the detection area DA shown in (a) are joined.

[0072] Figure 6 (b) indicates Figure 6 An example of the detection results for each point on line segment Li1 shown in (a). Figure 6 In (b), the vertical axis represents the position in the Z direction, which is perpendicular to the X and Y directions. The horizontal axis represents the position in the X direction. Figure 6 In (b), ○ (white circle) indicates the position of the first reflecting surface of component 10 in the Z direction. The first reflecting surface is the upper surface 11a of the metal plate 11, the upper surface 13a of the welded part 13, etc. ● (black circle) indicates the position of the second reflecting surface of component 10 in the Z direction. The second reflecting surface is the lower surface 11b of the metal plate 11, the lower surface 13b of the welded part 13, etc. As described above, these positions are calculated based on the time from the transmission of the ultrasonic wave US to the detection of the peak value of the reflected wave RW. Figure 6 In (b), ◆ indicates the determination result of joining or not joining. Points determined to be joined are represented by a value of 1, and points determined to be not joined are represented by a value of 0.

[0073] Figure 7 is an example of an image representing a processing result of the processing system of the embodiment.

[0074] By the above-described method, the first determination of determining whether or not each point of the detection area DA is bonded is performed. The processing device 110 generates, for example, the image shown in Figure 7 in accordance with the result of the first determination. In Figure 7 , white color indicates that the point is bonded. Black color indicates that the point is not bonded. The first area R1 based on the set of white points corresponds to the weld portion 13. The second area R2 based on the set of black points corresponds to the portion of the member 10 around the weld portion 13.

[0075] The processing device 110 can determine the quality of the welding in the welding object based on the area of the first area R1. The processing device 110 can also calculate the area of the weld portion 13 based on the area of the first area R1. The processing device 110 can further calculate the diameter of the weld portion 13 based on the diameter of the first area R1. For example, the distance between the detection elements 132 is stored in the storage device 120 in advance. The processing device 110 calculates the area or the diameter of the weld portion 13 using the number of pixels of the first area R1 and the stored distance. For example, the processing device 110 calculates the major axis and the minor axis of the weld portion 13 as the diameter. The processing device 110 can also calculate the average of the major axis and the minor axis. The processing device 110 can also calculate the equivalent circle diameter of the first area R1 as the diameter of the weld portion 13. The equivalent circle diameter of the first area R1 has the area of the first area R1, and is obtained by calculating the diameter of an imaginary circle having the area. The processing device 110 determines the quality of the welding in the welding object by comparing the calculated arbitrary value with a threshold value set in advance.

[0076] The first area R1 can include only the set of white points, or can include a part of the black points. For example, the processing device 110 sets the set of white points and the black points surrounded by the set of white points as the first area R1. In a case where a plurality of sets of white points exist, the processing device 110 sets the sets of white points and the black points located between the sets of white points as the first area R1.

[0077] The upper surface 13a and the lower surface 13b of the weld portion 13 are sometimes inclined with respect to the upper surface 11a of the metal plate 11. This is based on the case where the weld portion 13 includes the solidified portion 14, the deformation of the shape during the welding process, or the like. In this case, it is preferable to transmit the ultrasonic wave US in a direction that is orthogonal to the upper surface 13a or the lower surface 13b on average. Thereby, the ultrasonic wave can be more strongly reflected at the upper surface 13a and the lower surface 13b, and the accuracy of the inspection can be improved.

[0078] (Second determination)

[0079] In the second determination, it is determined whether the result of the first determination is appropriate. That is, it is determined whether the determination result of the bonding or non-bonding at each point at which the first determination is performed is appropriate. In order to determine whether the result of the first determination is appropriate, the processing device 110 uses the detection result of the reflected wave or the result of the first determination.

[0080] For example, if the result of the first determination is determined to be appropriate, the processing device 110 performs the following first action. In the first action, the processing device 110 adopts the result of the first determination. For example, the processing device 110 adopts data derived based on the result of the first determination as the inspection result of the welding object. The data includes at least any one of the area and the diameter of the welded portion 13, for example. The data can also include the determination of the good or bad of the welding based on the area or the diameter of the welded portion 13. The data can be derived between the first determination and the second determination, or can be derived after the second determination. The processing device 110 can also derive data based on the result of the first determination only when the result of the first determination is determined to be appropriate. In the first action, the processing device 110 can also output at least any one of the image indicating the result of the first determination, the area of the welded portion 13, the diameter of the welded portion 13, and the determination result of the good or bad of the welding to the storage device 120 or the display device 150.

[0081] If it is determined that the result of the first determination is not appropriate, the processing device 110 performs the following second action. In the second action, the processing device 110 does not adopt the result of the first determination. For example, in a case where data such as the area of the welded portion 13, the diameter of the welded portion 13, the determination result of the good or bad of the welding, and the like is derived between the first determination and the second determination, the processing device 110 does not adopt these data as the inspection result of the welding object. In the second action, the processing device 110 can also output the determination result indicating that the result of the first determination is not appropriate to the storage device 120 or the display device 150. In the second action, the processing device 110 can also cause the user to perform the detection to the welded portion 13 again. The detector 130 can also automatically perform the detection to the welded portion 13 again when it is determined by the processing device 110 that the result of the first determination is not appropriate.

[0082] Effects of the embodiments will be described.

[0083] For example, there is a method in which, at the time of the inspection of the welding object, the bonding and the non-bonding are determined at a plurality of points of the welding object, and the good or bad of the welding is determined based on the area of the first region R1. Alternatively, there is a method in which the area or the diameter of the welded portion 13 is calculated based on the area or the diameter of the first region R1, and the good or bad of the welding is determined based on the area or the diameter of the welded portion 13. According to these methods, it is possible to determine the good or bad of the welding in the welding object with substantially high accuracy.

[0084] Further verification by the inventors revealed that, in the aforementioned method, it is difficult to determine whether the weld 13 is good or bad. Specifically, it was found that even when the tilt of the detector 130 relative to the weld 13 is sufficiently small and the detector 130 is in reliable contact with the weld 13, there are still cases where the weld 13, which is actually well welded, is judged to be bad.

[0085] Figure 8 (a)~ Figure 8 (c) and Figure 9 This is a schematic diagram representing the image of the detection results based on reflected waves.

[0086] Figure 8 (a)~ Figure 8 The image of (c) and Figure 7 Similarly, it indicates that the result of the joint is determined at multiple points of the welded object. Figure 8 (a)~ Figure 8 The determination result of the joint shown in (c) is based on the detection result obtained under the condition that the tilt angle of the detector 130 is sufficiently small relative to the same part of the same weld object. The inventors have found that, as Figure 8 (a)~ Figure 8 As shown in (c), there are cases where the shape and size of the first region R1 deviate significantly for each test result. The cause of the deviation is not yet clear, but it is believed to be affected by factors such as the tilt of the welded part 13 relative to the overall welded object and the material of the welded object.

[0087] exist Figure 8 In image (a), a portion of the first region R1 is highlighted. Figure 8 In (c), the entire first region R1 is bent. The actual shape of the weld 13 is difficult to produce as the shape of the first region R1 shown in these images. Figure 8 In image (b), the first region R1 is close to a circle compared to the other images. Figure 8 The first region R1 in (b) is closer to the shape of the actual weld 13 than the first region R1 in other images.

[0088] Figure 9 and Figure 8 The image shown in (a) is the same. When the major and minor diameters of weld 13 are calculated during inspection, as shown... Figure 9 As shown, it is possible to calculate the major axis L1 and minor axis L2 of the first region R1 based on the protruding portion. In this case, the major and minor axes of the welded portion 13 are calculated to be longer than they actually are. Regarding the area, the area of ​​the welded portion 13 is also calculated to be larger than it actually is.

[0089] exist Figure 8In the image shown in (c), the size of the weld 13 represented by the first region R1 is smaller than the actual size of the weld 13. With these images, it is possible that the major axis and the minor axis of the weld 13 are calculated to be shorter than the actual values.

[0090] If the area or the diameter of the weld 13 is calculated to be different from the actual value, it is possible that the quality of the welding based on the area or the diameter of the weld 13 is erroneously determined. For example, in a case where the area or the diameter of the weld 13 is calculated to be larger than the actual value, although the welding is actually poor, it is possible that the welding is determined to be good. In a case where the area or the diameter of the weld 13 is calculated to be smaller than the actual value, although the welding is actually good, it is possible that the welding is determined to be poor. In a case where the area or the diameter of the weld 13 is referred to in another process for quality management, it is possible that a problem occurs in the other process.

[0091] To solve the technical problem, in the processing system 100 of the embodiment, the processing device 110 performs a second determination on the basis of the first determination. In the second determination, it is determined whether the result of the first determination is appropriate. That is, it is determined whether the determination result of the joining and the non-joining at the plurality of points of the welding object is appropriate. For example, only when it is determined through the second determination that the result of the first determination is appropriate, more accurate data related to the welding object can be derived by using the result of the first determination. For example, more accurate data can be obtained regarding the area of the weld 13, the diameter of the weld 13, the quality of the welding, and the like.

[0092] By performing the second determination, the user does not need to determine whether the result of the first determination is appropriate on the basis of the images shown in (c). Thus, it is possible to determine whether the result of the first determination is appropriate without depending on the knowledge and the experience of the user. Even in a case where the user who is insufficient in experience performs the inspection, only more appropriate data can be used as the inspection result. Figure 7

[0093] Hereinafter, a plurality of methods for performing the second determination will be described.

[0094] (First Method)

[0095] The processing device 110 calculates a first evaluation value on the basis of the detection result of the reflected wave or the result of the first determination. The processing device 110 stores the calculated first evaluation value in the storage device 120. The processing device 110 refers to the past first evaluation value stored in the storage device 120. The processing device 110 determines whether the result of the first determination is appropriate using the first evaluation value and the past first evaluation value.

[0096] ​Specifically, the processing device 110 calculates a second evaluation value based on a previous first evaluation value. The processing device 110 determines whether the result of the first determination is appropriate by comparing the difference between the first evaluation value and the second evaluation value with a first threshold. For example, the first evaluation value from just now can be used as the second evaluation value. Preferably, the second evaluation value is set based on multiple past first evaluation values. For example, the average of a predetermined number of past first evaluation values ​​can be used as the second evaluation value.

[0097] The first threshold can be set by the user or based on past first evaluation values. For example, the processing device 110 calculates the variance or standard deviation of multiple past first evaluation values. The processing device 110 calculates the first threshold by multiplying the variance or standard deviation by a predetermined value.

[0098] The first evaluation value is set, for example, based on the diameter of the welded portion 13. As described above, the diameter of the welded portion 13 is calculated based on the result of the first determination. The processing device 110 may use the diameter of the welded portion 13 as the first evaluation value, or it may use a value calculated based on the diameter of the welded portion 13 as the first evaluation value.

[0099] Alternatively, the first evaluation value can also be the value output from the first model. For example, the first model is learned to output a larger value the more appropriate the first decision is. The learned first model is stored in storage device 120. During the learning of the first model, the following methods are used... Figure 7 Image data representing the determination results of joint and non-joint points of a welding object, as shown. Each image data point is assigned a label indicating whether it is appropriate or inappropriate. A first model is learned using multiple image data points and multiple labels. The processing device 110 inputs image data based on the results of the first determination into the learned first model and uses the value (activity) of the output layer indicating the appropriateness of the image data as a first evaluation value. Alternatively, the processing device 110 may use other values ​​calculated based on the output layer values ​​as the first evaluation value.

[0100] like Figure 8 (a)~ Figure 8 As shown in (c), the more appropriate the result of the first determination, the closer the shape of the obtained first region R1 is to a circle. Therefore, the circle similarity of the first region R1 can also be used as the first evaluation value. Circle similarity indicates the degree to which the shape of the first region is similar to a circle. For example, the larger the value representing circle similarity, the closer the shape of the first region is to a circle. Circularity, roundness, or ellipticity can be used as circle similarity. The processing device 110 uses the ratio of roundness, roundness, ellipticity, or diameter as the first evaluation value. The processing device 110 uses the value calculated based on the ratio of roundness, roundness, ellipticity, or diameter as the first evaluation value.

[0101] Circularity is calculated using the following method. A circle inwardly tangent to the outer edge of a first region and another circle outwardly tangent to the outer edge of the first region are defined. The centers of these two circles are located at the same position. The two circles are defined such that their interval decreases. The difference in radii between the two circles corresponds to the roundness. The method for defining the center of the circle is arbitrary. For example, the following four methods are used: In the first method, the center of an approximate circle based on least squares is used. In the second method, the center of the largest circle inwardly tangent to the outer edge is used. In the third method, the center of the smallest circle outwardly tangent to the outer edge is used. In the fourth method, the centers of the inwardly and outwardly tangent circles with the smallest radius difference are used. Circularity can be calculated according to JIS B 0621 (1984). JIS B 0621 (1984) corresponds to ISO 1101 (1983).

[0102] Circularity is determined using the area A of the first region and the length L of the outer perimeter of the first region, by 4πA / L. 2 Ellipticity is represented by the ratio of the major axis to the minor axis. For example, the major axis is the length of the longest line segment connecting any two points on the outer edge of the first region R1. The minor axis is the length of the line segment passing through the center of the major axis and perpendicular to it.

[0103] As a measure of circular similarity, the ratio of the diameter r2 of the first region to the equivalent circular diameter r1 of the first region R1 can also be used. The diameter r2 can be, for example, the average of the major and minor axes. Alternatively, the average length of the first region R1 in multiple directions can also be used as the diameter r2.

[0104] The processing device 110 may also use two or more selected from roundness, circularity, ellipticity, and the ratio of diameter to calculate the first evaluation value. For example, the processing device 110 calculates the average or sum of two or more selected from roundness, circularity, ellipticity, and the ratio of diameter as the first evaluation value. For example, as Figure 9 As shown, the major axis L1 and minor axis L2 of the first region R1 may be calculated to be significantly different from the major and minor axes of the actual welded portion 13. Consequently, the minor axis L2 of the first region R1 may become a value close to the major axis L1. Therefore, when using ellipticity in the calculation of the first evaluation value, although the shape of the first region R1 is significantly different from a circle, there is a possibility that the first region R1 may be judged as resembling a circle. Therefore, when using ellipticity as a measure of circle similarity, it is preferable to further use the ratio of roundness, circularity, or diameter.

[0105] The first evaluation value can also be calculated by comparing the detection result of the reflected wave with pre-prepared data, or by comparing the result of the first judgment with pre-prepared data. For example, pre-prepared data could be... Figure 7The image data that is the result of the first determination as appropriate. The processing device 110 calculates a degree of similarity between the image data that is the result of the first determination and the image data prepared in advance as the first evaluation value. Alternatively, the processing device 110 can calculate a degree of similarity between the volume data that is the result of the detection of the reflected wave and the data prepared in advance as the first evaluation value. The processing device 110 can use a value calculated on the basis of the degree of similarity as the first evaluation value.

[0106] The processing device 110 determines that the result of the first determination is appropriate when the difference between the first evaluation value and the second evaluation value is less than the first threshold value. If the result of the first determination is determined to be appropriate, the processing device 110 performs the first action described above. The processing device 110 determines that the result of the first determination is inappropriate when the difference between the first evaluation value and the second evaluation value is the first threshold value or more. If the result of the first determination is determined to be inappropriate, the processing device 110 performs the second action described above.

[0107] Figure 10 is a graph for explaining the second determination.

[0108] In Figure 10 , the horizontal axis represents time, and the vertical axis represents the first evaluation value calculated at each time. For example, the processing device 110 sets the second evaluation value and the first threshold value using a plurality of first evaluation values that are earlier than the time t when the first evaluation value at the time t is calculated. The processing device 110 determines whether the difference between the first evaluation value and the second evaluation value is the first threshold value or more. In Figure 10 , the broken line BR1 represents a value obtained by adding the first threshold value to the second evaluation value. The broken line BR2 represents a value obtained by subtracting the first threshold value from the second evaluation value. In this example, the first evaluation value at the time t is the value obtained by adding the first threshold value to the second evaluation value or more. In other words, the difference between the first evaluation value and the second evaluation value is the first threshold value or more. Therefore, the processing device 110 determines that the result of the first determination is inappropriate.

[0109] (Second Method)

[0110] The processing device 110 can determine that the result of the first determination is appropriate by inputting the result of the detection of the reflected wave or the result of the first determination to a first model. In the first model, the result of the detection of the reflected wave or the result of the first determination is distinguished (classified or clustered).

[0111] For example, the processing device 110 inputs the image data that represents the result of the first determination illustrated in Figure 7 to the first model. The first model learns by supervised learning and classifies the input image data. In the learning of the first model, the image data that represents the result of the first determination is used as a learning target. Figure 7Image data representing the results of determining the joining and non-joining of various points on a welded object, as shown. For example, for... Figure 8 (a) or Figure 8 As shown in (c), image data that protrudes from a portion of the first region R1, or image data that is curved in the first region R1, is given an inappropriate label. Figure 8 (a) or Figure 8 As shown in (c), when the first region R1 is close to a circle, an appropriate label is assigned. The first model learns using multiple image data and multiple labels. When image data representing an appropriate first category is classified, the processing device 110 determines the result of the first determination as appropriate. When image data representing an inappropriate second category is classified, the processing device 110 determines the result of the first determination as inappropriate.

[0112] The first model can also be learned through unsupervised learning. In this case, the input image data is clustered into one of several categories by the first model. The input image data is clustered into categories including... Figure 8 When the image data, as shown in (b), has a shape close to a circle, belonging to the first category, the processing device 110 determines the result of the first determination to be appropriate. When the input image data is clustered into categories including... Figure 8 Image data of a portion of the first region R1 shown in (a) or Figure 8 When the image data of the first region R1 shown in (c) is included in the second category, the processing device 110 determines the result of the first determination as inappropriate.

[0113] Alternatively, three-dimensional volume data representing the detection results of reflected waves can be input to the first model. In this case, supervised or unsupervised learning is performed using multiple three-dimensional data sets for the first model. When data representing an appropriate first category is input and classified, the processing device 110 also determines the result of the first determination based on that data to be appropriate. When data representing an inappropriate second category is input and classified, the processing device 110 determines the result of the first determination based on that data to be inappropriate as well.

[0114] As described above, in the second method, when the detection result of the reflected wave or the result of the first determination is classified into a first category by the first model, the processing device 110 determines the result of the first determination as appropriate. If the result of the first determination is determined to be appropriate, the processing device 110 performs the first action described above. When the detection result of the reflected wave or the result of the first determination is classified into a second category by the first model, the processing device 110 determines the result of the first determination as inappropriate. If the result of the first determination is determined to be inappropriate, the processing device 110 performs the second action described above.

[0115] The processing device 110 can also combine the first method and the second method to determine whether the result of the first determination is appropriate. For example, the processing device 110 finally determines that the result of the first determination is appropriate when the result of the first determination is determined to be appropriate by the first method and the result of the first determination is determined to be appropriate by the second method. Alternatively, the processing device 110 can finally determine that the result of the first determination is appropriate when the result of the first determination is determined to be appropriate by the first method or the result of the first determination is determined to be appropriate by the second method. By appropriately combining the first method and the second method, the accuracy of the second determination can be improved.

[0116] Figure 11 is a flowchart showing an inspection flow of a processing system using the embodiment.

[0117] The user brings the front end of the detector 130 into contact with the weld 13. The user performs detection by the detector 130 (step S1). For example, a button for performing detection is provided on the detector 130. The user can perform detection based on the detector 130 by operating the button. Alternatively, the user can perform detection based on the detector 130 through a user interface displayed on the display device 150. The detector 130 transmits the detection result of the reflected wave obtained by the detection to the processing device 110.

[0118] The processing device 110 performs the first determination of determining whether the joint and the unjoint are present at a plurality of points of the weld object when the detection result is received (step S2). The processing device 110 performs the second determination of determining whether the result of the first determination is appropriate (step S3). If the result of the first determination is determined to be appropriate, the processing device 110 performs the first action (step S4). If the result of the first determination is determined to be inappropriate, the processing device 110 performs the second action (step S5).

[0119] The processing device 110 can also perform estimation of the range of the weld 13 and calculation of the inclination with respect to the weld 13 based on the detection result of the reflected wave. Here, the angle between the normal direction of the surface of the weld 13 and the direction of the detector 130 is referred to as the inclination. The direction of the detector 130 corresponds to, for example, the Z direction perpendicular to the arrangement direction of the detection elements 132. When the detector 130 is in contact with the surface of the weld 13 perpendicularly, the inclination is zero.

[0120] The inclination of the detector 130 with respect to the weld object or the weld 13 can affect the inspection result. For example, if the first determination is performed in a state where the detector 130 is inclined with respect to the weld object, it can be determined that the joint is unjoint even though the joint is actually appropriately joined. Therefore, it is preferable to set the inclination of the detector 130 with respect to the weld object to be small before performing the first determination.

[0121] The inclination of the detector 130 is calculated using the detection results of the reflected waves from the weld portion 13. In the first determination, it is sufficient to determine the joining and non-joining of the reflected waves from the weld portion 13. By reducing the amount of calculation with respect to the detection results of the reflected waves from the areas other than the weld portion 13, the time required for each process can be shortened. Therefore, it is preferable to extract a part of the detection results including the reflected waves from the weld portion 13 before the calculation of the inclination and the execution of the first determination.

[0122] Figure 12 is a flowchart showing an inspection flow of a processing system using the embodiment.

[0123] Referring to Figure 12 The flow of the inspection in the case of the estimation of the execution range and the calculation of the inclination will be described. The user performs the probing of the detector 130 (step S1). When the probing is performed, the processing device 110 determines whether the estimation of the range corresponding to the reflected waves from the weld portion 13 is completed with respect to the weld object on which the probing is performed (step S11). When the range is not estimated, the processing device 110 estimates the range (step S12).

[0124] For example, as shown in (a) of Figure 4 and (b) of Figure 6 The ultrasonic waves are also reflected from the surface other than the weld portion 13. The processing device 110 estimates the range corresponding to the reflected waves from the weld portion 13, and performs the calculation of the inclination thereafter based on the reflected waves included in the range. Thereby, the amount of calculation necessary can be reduced. The accuracy of the calculated inclination can be improved.

[0125] The processing device 110 calculates the inclination of the detector 130 based on the detection results of the reflected waves in the estimated range (step S13). It is determined whether the calculated inclination is within the allowable range (step S14). The determination can be performed by the user or by the processing device 110. In the case where the processing device 110 performs the determination, the allowable range can be set by the user in advance or can be set based on the history of the past inspection results.

[0126] For example, when processing device 110 inspects weld 13, it determines the diameter of weld 13 based on the inspection results. If the tilt of detector 130 is too large, the diameter of weld 13 is calculated to be smaller than it actually is. As the tilt of detector 130 decreases, the calculated diameter of weld 13 increases. If the tilt of detector 130 is sufficiently small, the calculated diameter of weld 13 hardly changes. Such previously calculated relationships between the tilt of detector 130 and the diameter of weld 13 are stored in storage device 120. Based on the data stored in storage device 120, processing device 110 determines a boundary value for reducing the change in the diameter of weld 13 in response to changes in the tilt of detector 130. Processing device 110 sets the size of an allowable range based on this boundary value. For example, processing device 110 sets the boundary value as the size of the allowable range. Alternatively, to further improve the accuracy of the inspection, processing device 110 may also set a smaller value calculated based on the boundary value as the allowable range.

[0127] When the tilt is outside the allowable range, the user adjusts the tilt of detector 130 (step S15). If the processing device 110 executes step S14, it may also notify the user that the tilt is outside the allowable range. After step S15, step S1 is executed again with the adjusted tilt. When the tilt is within the allowable range, a first determination is performed (step S2). Preferably, in the first determination, points on the XY plane within the range estimated in step S12 are determined to be either joined or not joined. After step S2, [the process continues with...]. Figure 11 The flowchart shown is executed in the same way, following steps S3 to S5.

[0128] The following section explains a specific case regarding the estimation of the range, the calculation of the inclination, and the inspection.

[0129] (Range estimation)

[0130] Reference Figures 13-20 Please provide a detailed explanation of the estimated range.

[0131] For example, in Figure 5 In (b), the detection result of the reflected wave is represented in a two-dimensional manner. The detection result of the reflected wave can also be represented in a three-dimensional manner. For example, multiple voxels are set for component 10. Coordinates in the X, Y, and Z directions are set for each voxel. Based on the detection result of the reflected wave, a correlation is established between each voxel and the intensity of the reflected wave. The processing device 110 estimates the range (group of voxels) corresponding to the weld portion 13 for the multiple voxels. The number of voxels set and the size of each voxel can be determined automatically or set by the user through the user interface of the display device 150.

[0132] Figure 13 (a) andFigure 13 (b) is a graph that illustrates the intensity distribution of the reflected wave in the Z direction.

[0133] Figure 14 (b) is a graph that illustrates the intensity distribution of the reflected wave in the Z direction.

[0134] The processing device 110 generates the intensity distribution of the reflected wave in the Z direction based on the detection result of the reflected wave. Figure 13 (a) of (b) is an example thereof. In (a) of (b), Figure 13 (b) of (b) is an example thereof. In (b) of (b), Figure 13 (a) of (b) and Figure 13 (b) of (b), the horizontal axis indicates the position in the Z direction, and the vertical axis indicates the intensity of the reflected wave. Figure 13 (a) illustrates the intensity distribution of the reflected wave in the Z direction in one X-Z cross section. Figure 13 (b) illustrates the intensity distribution of the reflected wave in the Z direction in one Y-Z cross section. In (b) of (b), Figure 13 (a) of (b) and Figure 13 (b) of (b), the result of converting the intensity of the reflected wave into an absolute value is indicated.

[0135] Alternatively, the processing device 110 can also aggregate the intensity of the reflected wave on the X-Y plane at each point in the Z direction to generate the intensity distribution of the reflected wave in the Z direction. Figure 14 (b) is an example thereof. In (b) of (b), Figure 14 (b), the horizontal axis indicates the position in the Z direction, and the vertical axis indicates the intensity of the reflected wave. In (b) of (b), Figure 14 (b), the result of converting the intensity of the reflected wave into an absolute value and subtracting the average value of the intensity of the reflected wave from the intensity of the reflected wave at each point in the Z direction is shown.

[0136] The intensity distribution of the reflected wave in the Z direction includes components reflected on the upper surface 13a and the lower surface 13b of the weld portion 13 and components reflected on the upper surfaces and the lower surfaces of other portions. The processing device 110 extracts only the components reflected on the upper surface 13a and the lower surface 13b of the weld portion 13 from the intensity distribution of the reflected wave by filtering. For example, a value corresponding to an integral multiple of half the thickness (distance between the upper surface 13a and the lower surface 13b) of the weld portion 13 in the Z direction is set in advance. The processing device 110 refers to the value to extract only the periodic component of the value.

[0137] As the filtering, a band-pass filter, a zero-phase filter, a low-pass filter, a high-pass filter, or threshold determination for the filtered intensity, or the like can be used.

[0138] Figure 15 (b) is a graph that illustrates the result of filtering the intensity distribution of the reflected wave.

[0139] In (b) of (b), Figure 15In the drawing, the horizontal axis indicates the position in the Z direction, and the vertical axis indicates the intensity of the reflected wave. As shown in the drawing, the result of the filtering extracts only the components reflected from the upper surface and the lower surface of the weld. Figure 15

[0140] The processing device 110 estimates the range in the Z direction of the weld based on the extraction result. For example, the processing device 110 detects a peak included in the extraction result. The processing device 110 detects the position in the Z direction of the first peak and the position in the Z direction of the second peak. The processing device 110 estimates, for example, the range Ra1 shown in the drawing as the range in the Z direction of the weld based on these positions. Figure 15

[0141] According to the configuration of the weld, the structure of the element array 131, and the like, sometimes the sign (positive or negative) of the reflected wave intensity from the upper surface of the weld and the sign of the reflected wave intensity from the lower surface of the weld are reversed from each other. In this case, the processing device 110 can also detect a peak of one of the positive and negative and another peak of the other of the positive and negative. The processing device 110 estimates the range in the Z direction of the weld based on the positions of these peaks. According to the processing of the reflected wave intensity, sometimes the reflected wave intensity is represented only by one of the positive and negative values. In this case, the range in the Z direction of the weld can be estimated based on the positions of a plurality of peaks, can be estimated based on the positions of a peak and a valley, and can be estimated based on the positions of a plurality of valleys. That is, the processing device 110 estimates the range in the Z direction of the weld based on the positions of a plurality of extreme values with respect to the filtered reflected wave intensity.

[0142] In generating the intensity distribution of the reflected wave in each of the X-Z cross section and the Y-Z cross section, the range in the Z direction based on the intensity distribution in the X-Z cross section and the range in the Z direction based on the intensity distribution in the Y-Z cross section are estimated. For example, the processing device 110 calculates the average, the weighted average, the weighted moving average, and the like with respect to these plurality of estimation results, and estimates the calculation result as the range in the Z direction of the weld as a whole.

[0143] Alternatively, the processing device 110 can estimate the range in the Z direction of the weld based on the intensity distribution of the reflected wave in one of the X-Z cross section and the Y-Z cross section, and regard the estimation result as the range in the Z direction of the weld as a whole. The processing device 110 can estimate the range in the Z direction of the weld based on the intensity distribution of the reflected wave in a part of the X direction and a part of the Y direction, and regard the estimation result as the range in the Z direction of the weld as a whole. According to these processes, it is possible to reduce the amount of calculation required for the generation of the intensity distribution of the reflected wave.

[0144] In the drawing, the horizontal axis indicates the position in the Z direction, and the vertical axis indicates the intensity of the reflected wave. As shown in the drawing, the result of the filtering extracts only the components reflected from the upper surface and the lower surface of the weld. Figure 15 ​​In the example of FIG. 9, the position in the Z direction of the lower limit of the range Ra1 is set to a value obtained by subtracting a prescribed value from the position in the Z direction of the first peak. The position in the Z direction of the upper limit of the range Ra1 is set to a value obtained by adding a prescribed value to the position in the Z direction of the second peak. Thus, when the upper surface and the lower surface of the weld portion are inclined with respect to the arrangement direction of the detection elements 132, it is possible to suppress the second peak from deviating from the range in the Z direction at any point in the X-Y plane of the weld portion.

[0145] After estimating the range in the Z direction of the weld portion, the processing device 110 estimates the range in the X direction and the range in the Y direction of the weld portion.

[0146] Figure 16 and Figure 18 is a diagram illustrating the detection result of the reflected wave.

[0147] In Figure 16 and Figure 18 , the region R indicates a region of the entire detection result of the reflected wave obtained by the element array 131. In one cross section of the region R, the component of the reflected wave in the upper surface and the lower surface of the weld portion and the component of the reflected wave in the upper surface and the lower surface of the other portion are included.

[0148] The processing device 110 generates the intensity distribution of the reflected wave in the X-Y plane at each point in the Z direction. The processing device 110 can also generate the intensity distribution within a range in the Z direction set in advance. Thus, it is possible to reduce the amount of calculation. Alternatively, the processing device 110 can also generate the intensity distribution within the estimated range in the Z direction. Thus, it is possible to reduce the amount of calculation and to suppress the reflected wave from the lower surface of the weld portion from deviating when generating the intensity distribution of the reflected wave in the X-Y plane.

[0149] Figure 17 (a) to Figure 17 (c) of FIG. 10 is an example of the intensity distribution of the reflected wave in the X-Y plane. Figure 17 (a) of FIG. 10 indicates the intensity distribution of the reflected wave in the X-Y plane at the coordinate of Z = 1. Figure 17 (b) of FIG. 10 indicates the intensity distribution of the reflected wave in the X-Y plane at the coordinate of Z = 2. Figure 17 (c) of FIG. 10 indicates the intensity distribution of the reflected wave in the X-Y plane at the coordinate of Z = 350. In Figure 16 , Figure 17 (a) to Figure 17 (c) of FIG. 10, the intensity of the reflected wave is schematically binarized. Figure 18

[0150] ​The processing device 110 calculates the center-of-gravity position of the intensity distribution of the reflected wave in the X-Y plane at each point in the Z direction. Here, the center-of-gravity position of the intensity distribution is obtained by calculating the center-of-gravity position of the image representing the intensity distribution. For example, as shown in (a) to (c) of FIG. 10, the processing device 110 calculates the center-of-gravity positions Cl to C350 in each image. In (a) to (c) of FIG. 10, the line segments Li2 represent the result of connecting all the center-of-gravity positions from Z = 0 to Z = 350. Figure 17 Figure 17 Figure 18

[0151] The processing device 110 averages the center-of-gravity positions from Z = 0 to Z = 350. Thereby, the average position of the center of gravity in the X direction and the average position of the center of gravity in the Y direction can be obtained. In (a) to (c) of FIG. 11, the average position AP represents the average position of the center of gravity in the X direction and the average position of the center of gravity in the Y direction. The processing device 110 sets a prescribed range as the range Ra2 of the welding portion in the X direction and the range Ra3 of the welding portion in the Y direction in each of the X direction and the Y direction with the average position AP as the center. Figure 18

[0152] For example, in order to estimate the ranges Ra2 and Ra3, a value V representing the diameter of the detector 130 (element array 131) is set in advance. The processing device 110 sets the ranges Ra2 and Ra3 from AP - V / 2 to AP + V / 2 in the X direction and the Y direction, respectively. In this case, the estimated range of the X-Y plane is a quadrilateral shape. The estimated range of the X-Y plane is not limited to this example, and can be a polygon shape of five or more sides or a circular shape, or the like. The shape of the estimated range of the X-Y plane can be appropriately changed according to the shape of the welding portion.

[0153] Other values based on the value V can be used to determine the ranges Ra2 and Ra3. Instead of the value representing the diameter of the detector 130, a value representing the average diameter of the welding portion can be set in advance. This is because the diameter of the welding portion corresponds to the diameter of the detector 130. The value representing the diameter of the welding portion can be substantially regarded as the value representing the diameter of the detector 130.

[0154] By the above processing, the range Ra1 in the Z direction, the range Ra2 in the X direction, and the range Ra3 in the Y direction of the welding portion are estimated. After the ranges are estimated, the step S12 shown in FIG. 8 is performed based on the detection results of the reflected wave within the estimated ranges. Figure 12

[0155] Figure 19 is a flowchart of a flow of the estimation of the range in the processing system of the embodiment.

[0156] ​​​​​The processing device 110 generates an intensity distribution of the reflected wave in the Z direction based on the detection result of the reflected wave of the detector 130 (step S121). The processing device 110 filters the intensity distribution based on the value of the thickness of the weld (step S122). Thereby, only the reflected wave component in the weld 13 is extracted from the intensity distribution. The processing device 110 estimates the range in the Z direction of the weld based on the extraction result (step S123). The processing device 110 calculates the center-of-gravity position of the reflected wave intensity on the X-Y plane at each point in the Z direction (step S124). The processing device 110 calculates the average position by averaging the plurality of center-of-gravity positions calculated (step S125). The processing device 110 estimates each of the ranges in the X direction and the Y direction based on the average position and the diameter of the detector 130 (step S126).

[0157] Further, the estimation of the range in the Z direction can also be performed after the estimation of the ranges in the X direction and the Y direction. For example, in the flowchart shown in FIG. 12, the steps S121 to S123 can be performed after the steps S124 to S126. In this case, the processing device 110 can also calculate the intensity distribution of the reflected wave in the Z direction based on the estimated ranges in the X direction and the Y direction. Thereby, the amount of calculation can be reduced. Figure 19

[0158] (Calculation of the inclination)

[0159] Figure 20 is an image illustrating the detection result of the reflected wave.

[0160] In Figure 20 , the whiter the color, the greater the intensity of the reflected wave at the point. The processing device 110 performs the action shown in Figure 20 on the detection result. As a result, the range Ra is estimated. Figure 21

[0161] Hereinafter, a specific example of the calculation method of the inclination in the range Ra will be described.

[0162] Figure 21 is a graph for explaining the processing of the processing system of the embodiment.

[0163] Figure 22 is an example of the image obtained by the processing system of the embodiment.

[0164] Figure 22 (a) of Figure 5 indicates the surface of the weld 13 in the volume data shown in Figure 22 (b) of Figure 5 indicates the Y-Z cross section in the vicinity of the weld 13 in the volume data shown in Figure 22 (c) of Figure 5 ​​The XZ section near weld 13 in the volume data shown. Figure 22 (b) and Figure 22 In (c), the upper part represents the surface of the welded part, and the lower part represents the data in the depth direction. The brighter parts are those with high ultrasonic wave reflection intensity. Ultrasonic waves are strongly reflected on the bottom surface of the welded part 13, the surfaces between unjoined parts, etc.

[0165] The tilt of detector 130 corresponds to Figure 21 The angle shown is between the direction 13d, which is perpendicular to the welded part 13, and the direction 130a of the detector 130. This angle is represented by the angle θx about the X direction and the angle θy about the Y direction. The direction 130a of the detector 130 is perpendicular to the arrangement direction of the detection elements 132.

[0166] Angle θx Figure 22 As shown in (b), the angle θy is calculated based on the detection results in the YZ section. Figure 22 As shown in (c), the calculation is based on the detection results in the XZ section. The processing device 110 calculates the average of the three-dimensional brightness gradient for each section as angles θx and θy. The processing device 110 stores the calculated angles θx and θy as the tilt of the detector 130 in the storage device 120. The processing device 110 can also make the display device 150 display the calculated tilt.

[0167] By calculating the tilt angle of detector 130 and reducing it, the accuracy of determining whether points are aligned can be improved in the first determination. This enhances the accuracy of data based on the results of the first determination. By estimating the tilt angle calculation and the range of the first determination, the computational load required for each process can be reduced.

[0168] For example, if the result of the first determination is deemed inappropriate in the second determination, a second detection based on detector 130 is performed. In this case, the conditions for performing the second detection may differ from those for performing the previous detection. For example, the conditions for performing the second detection may change from those for performing the previous detection by performing at least one of recoating the coupling agent, readjusting the tilt, or changing the allowable range of the tilt.

[0169] For example, if the result of the first determination is deemed inappropriate, the user separates the detector 130 from the welding object and reapplies the coupling agent to the welding object. If no coupling agent was present between the welding object and the detector 130 during the previous detection, reapplication of the coupling agent may fill this gap. When reapplying the coupling agent, the coupling agent already applied to the welding object can also be removed.

[0170] In the re-probing, the inclination of the detector 130 with respect to the welding object can also be set to the same value as the inclination in the immediately preceding probing. Even if the inclination of the detector 130 is the same value as the immediately preceding inclination, the detection result of the reflected wave can be different from the immediately preceding detection result. By making the inclination of the detector 130 the same value as the immediately preceding inclination, the inclination is maintained within the allowable range.

[0171] Alternatively, the inclination of the detector 130 with respect to the welding object can also be set to a value different from the inclination in the immediately preceding probing. In this case, it is preferable that the amount of change in the inclination be smaller than the difference between the immediately preceding inclination and the critical value of the allowable range. Thus, even in the case where the inclination of the detector 130 is changed, it is possible to suppress the changed inclination from being outside the allowable range. By changing the inclination of the detector 130, the likelihood of obtaining a detection result different from the immediately preceding detection result of the reflected wave becomes higher.

[0172] The inclination of the detector 130 can also be set outside the allowable range. Thus, the inclination of the detector 130 is set to a value greatly different from the inclination in the immediately preceding probing. As a result, the likelihood of obtaining a detection result greatly different from the immediately preceding probing is increased. Based on the detection result greatly different from the immediately preceding probing, the inclination of the detector 130 is adjusted again, and thus it is possible to adjust the detector 130 to a state where a more appropriate detection result can be obtained.

[0173] In the re-probing, the allowable range of the inclination can also be changed. For example, the processing device 110 narrows the allowable range of the inclination. The processing device 110 adjusts the inclination so that the inclination of the detector 130 enters the narrow allowable range. The processing device 110 can also set the inclination of the detector 130 to a value different from the inclination in the immediately preceding probing when the allowable range is changed.

[0174] After the re-probing is performed under conditions different from those of the immediately preceding probing, the first determination is performed again based on the detection result of the reflected wave obtained by the probing. Alternatively, the estimation of the range can also be performed again before the first determination. In the first determination, the joining or non-joining is detected for each point of the range estimated again.

[0175] The condition at the time of the detection can also be changed according to the result of the determination of the soldering depending on whether the result of the first determination is determined to be inappropriate and the detection based on the detector 130 is performed again. For example, at least one of the reapplication of the coupling agent, the readjustment of the inclination, and the change of the allowable range of the inclination is performed when the soldering is determined to be good and the result of the first determination is determined to be inappropriate. Thus, the condition at the time of the detection is changed from the condition at the time of the detection immediately before to another condition (first condition). At least one of the reapplication of the coupling agent, the readjustment of the inclination, and the change of the allowable range of the inclination is performed when the soldering is determined to be poor and the result of the first determination is determined to be inappropriate. Thus, the condition at the time of the detection is further changed from the condition at the time of the detection immediately before to another condition (second condition). By setting the condition at the time of the detection according to the combination of the result of the first determination and the result of the determination of the soldering, a more appropriate detection result of the reflected wave is more easily obtained in the re-detection.

[0176] Figure 23 is a diagram indicating the result of the classification based on the first model.

[0177] In the second method of the second determination, the detection result of the reflected wave or the result of the first determination can also be classified into three or more categories by the first model. Figure 23 (a) to (c) of Figure 23 (c) of indicates an example in which the image data indicating the result of the first determination is classified into three or more categories by the first model. Figure 23 (a) to (c) of Figure 23 (c) of respectively indicates the image data classified into mutually different categories. Figure 23 The image data indicated in (a) of indicates that the result of the first determination is appropriate. Figure 23 The image data indicated in (b) of and Figure 23 The image data indicated in (c) of indicates that the result of the first determination is inappropriate.

[0178] Figure 23 The area ratio of the first region R1 of the image data of (b) of is larger than Figure 23 The area ratio of the first region R1 of the image data of (a) of is smaller than Figure 23 A part of the first region R1 of the image data of (b) of protrudes. Figure 23 The image data of (b) of is classified into a second category different from Figure 23 The image data of (a) of is classified into a first category. Figure 23 The area ratio of the first region R1 of the image data of (c) of is larger than Figure 23 The area ratio of the first region R1 of the image data of (a) of is smaller than Figure 23 The first region R1 of the image data of (c) of is bent as a whole. Figure 23The image data of (c) is classified into a third category different from the first and second categories. Figure 23 The image data of (b) and Figure 23 The shape of the first region R1 in the image data of (c) is further from a circle than Figure 24 The image data of (a).

[0179] Instead of the image data representing the result of the first determination, three-dimensional volume data as a detection result of the reflected wave can be input to the first model. The image data representing the result of the first determination is based on the detection result of the reflected wave. Therefore, with respect to the three-dimensional volume data, it is also possible to classify by the first model into one of a plurality of categories.

[0180] For example, if it is determined that the result of the first determination is inappropriate, re-probing based on the detector 130 is performed. At this time, the condition at the time of the probing can be changed depending on the category into which the data is classified. For example, when the data is classified into the second category, at least one of re-coating of the coupling agent, re-adjustment of the inclination, and change of the allowable range of the inclination is performed. Thereby, the condition at the time of the re-probing is changed from the condition at the time of the immediately preceding probing to another condition (first condition). When the data is classified into the third category, at least one of re-coating of the coupling agent, re-adjustment of the inclination, and change of the allowable range of the inclination is performed. Thereby, the condition at the time of the re-probing is further changed from the condition at the time of the immediately preceding probing to another condition (second condition). By setting the condition at the time of the probing depending on the category into which the data is classified, a more appropriate detection result of the reflected wave is more easily obtained in the re-probing.

[0181] (Modified Example)

[0182] The inspection of the weld described above can also be performed automatically by a robot.

[0183] Figure 25 is a schematic view showing the structure of a processing system of a modified example of the embodiment.

[0184] Figure 24 is a perspective view showing a part of a processing system of a modified example of the embodiment.

[0185] Figure 25 The processing system 100a shown has a processing device 110 and a robot 160. The robot 160 includes a detector 130, an imaging device 161, a coating device 162, an arm 163, and a control device 164.

[0186] The imaging device 161 captures the component after welding, and acquires an image. The imaging device 161 extracts a welding trace from the image, and detects the approximate position of the weld 13. The coating device 162 coats the coupling agent on the upper surface of the weld 13.

[0187] As shown in Figure 26 The detector 130, the camera 161, and the applicator 162 are provided at the tip of the arm 163. The arm 163 is, for example, a multi-joint robot. By driving the arm 163, the detector 130, the camera 161, and the applicator 162 are displaced. The control device 164 controls the operation of each of the components (the detector 130, the camera 161, the applicator 162, and the arm 163) of the robot 160.

[0188] Figure 12 is a flowchart showing the operation of the processing system of the modification example of the embodiment.

[0189] The processing device 110 transmits the coordinates of the weld 13 stored in the storage device 120 to the control device 164. The control device 164 drives the arm 163 so that the tip of the arm moves toward the received coordinates (step S21). If the detector 130 is moved to the vicinity of the received coordinates, the camera 161 captures the member 10, and the detailed position of the weld 13 is detected from the obtained image (step S22). The control device 164 drives the arm 163 so that the applicator 162 is moved to the vicinity of the detected position (step S23). The applicator 162 applies the coupling agent to the weld 13 (step S24). The control device 164 drives the arm 163 so that the detector 130 is moved so that the tip of the detector 130 comes into contact with the weld 13 to which the coupling agent is applied (step S25). Hereafter, S1 to S5 and S11 to S15 are executed similarly to the flowchart shown in Figure 27 S1 to S5 and S11 to S15 are executed similarly to the flowchart shown in

[0190] In the second operation of step S5, the detection of step S2 can be performed again. For example, when it is determined in step S3 that the result of the first determination is inappropriate, the processing device 110 transmits the determination result to the control device 164. The control device 164 performs the detection based on the detector 130 again upon receiving the determination result. At this time, the control device 164 can perform at least one of reapplication of the coupling agent, readjustment of the inclination, and change of the allowable range of the inclination.

[0191] In the detection again, the control device 164 can set the inclination of the detector 130 with respect to the welding object to the same value as the inclination in the detection immediately before. Even if the inclination of the detector 130 is the same value as the inclination immediately before, the detection result of the reflected wave can be different from the detection result immediately before. By setting the inclination of the detector 130 to the same value as the inclination immediately before, the inclination is maintained within the allowable range.

[0192] Alternatively, the control device 164 can also set the inclination of the detector 130 with respect to the welding object to a value different from the inclination in the immediately preceding detection. In this case, it is preferable that the amount of change in the inclination be smaller than the difference between the immediately preceding inclination and the critical value of the allowable range. Thus, even in the case where the inclination of the detector 130 is changed, it is possible to suppress the changed inclination from being outside the allowable range. By changing the inclination of the detector 130, the likelihood of obtaining a detection result different from that of the immediately preceding reflected wave increases.

[0193] In the case where the inclination of the detector 130 is set to a value different from the inclination in the immediately preceding detection, the control device 164 can also set the inclination of the detector 130 outside the allowable range. Thus, the inclination of the detector 130 is set to a value greatly different from the inclination in the immediately preceding detection. As a result, the likelihood of obtaining a detection result greatly different from that of the immediately preceding detection increases. Based on the detection result greatly different from that of the immediately preceding detection, the inclination of the detector 130 is adjusted again, and thus it is possible to control the detector 130 to a state where a more appropriate detection result can be obtained.

[0194] The control device 164 can also execute step S22 or S24 again. That is, the control device 164 detects the position of the welding portion 13 again. Thus, a more accurate position of the welding portion 13 is detected, and it is possible to obtain a more appropriate result of the first determination. Alternatively, the control device 164 applies the couplant to the welding portion 13 again. For example, when it is determined that the result of the first determination is inappropriate, the control device 164 drives the arm 163 to move the detector 130 away from the welding object. The control device 164 causes the application device 162 to apply the couplant to the welding object again. The control device 164 can clean the nozzle of the application device 162 before applying the couplant. The control device 164 can also test the emission of the couplant at a position away from the welding object before applying the couplant, and determine whether the couplant is emitted normally. When the couplant is not emitted normally, the control device 164 can clean the nozzle of the application device 162. In the case where there is an abnormality in the application device 162, there is a possibility that the abnormality in the application device 162 is eliminated by these actions. After the cleaning or the test emission, the control device 164 causes the welding object to be applied with the couplant again. By the reapplication of the couplant, there is a possibility that the welding object and the detector 130 are filled with the couplant. As a result, it is possible to obtain a more appropriate result of the first determination. After step S22 or S24 is executed, the subsequent steps are executed again.

[0195] The control device 164 can also change the conditions at the time of the probe when the second probe is performed, according to the result of the first determination and the result of the determination of the welding. For example, when the result of the first determination is determined to be inappropriate and the second probe of the detector 130 is performed, the control device 164 changes the conditions at the time of the probe, according to the result of the determination of the welding. For example, when the welding is determined to be good and the result of the first determination is determined to be inappropriate, the control device 164 performs at least one of the reapplication of the coupling agent, the readjustment of the inclination, and the change of the allowable range of the inclination. Thereby, the conditions at the time of the probe when the second probe is performed are changed from the conditions at the time of the probe immediately before to other conditions (first conditions). When the welding is determined to be poor and the result of the first determination is determined to be inappropriate, the control device 164 performs at least one of the reapplication of the coupling agent, the readjustment of the inclination, and the change of the allowable range of the inclination. Thereby, the conditions at the time of the probe when the second probe is performed are further changed from the conditions at the time of the probe immediately before to other conditions (second conditions). By setting the conditions at the time of the probe according to the combination of the result of the first determination and the result of the determination of the welding, a more appropriate detection result of the reflected wave is more easily obtained in the second probe.

[0196] When the second method is used in the second determination, the control device 164 can also change the conditions at the time of the probe, according to the category into which the data is classified. For example, when the data is classified into the second category, the control device 164 performs at least one of the reapplication of the coupling agent, the readjustment of the inclination, and the change of the allowable range of the inclination. Thereby, the conditions at the time of the probe when the second probe is performed are changed from the conditions at the time of the probe immediately before to other conditions (first conditions). When the data is classified into the third category, the control device 164 performs at least one of the reapplication of the coupling agent, the readjustment of the inclination, and the change of the allowable range of the inclination. Thereby, the conditions at the time of the probe when the second probe is performed are further changed from the conditions at the time of the probe immediately before to other conditions (second conditions). By setting the conditions at the time of the probe according to the category into which the data is classified, a more appropriate detection result of the reflected wave is more easily obtained in the second probe.

[0197] In the processing system 100a, by performing the second determination, more accurate data related to the welding object can be obtained. For example, in a case where the area or the diameter of the welded portion 13 is calculated based on the result of the first determination, by using the value based on the result of the first determination that is determined to be appropriate, a more accurate value can be obtained.

[0198] Figure 1 is a block diagram showing the hardware structure of the system.

[0199] The processing device 110 of the processing system 100 of the embodiment is, for example, a computer having a ROM (Read Only Memory) 111, a RAM (Random Access Memory) 112, a CPU (Central Processing Unit) 113, and a HDD (Hard Disk Drive) 114.

[0200] The ROM 111 stores a program for controlling the operation of the computer. In the ROM 111, a program necessary for the computer to implement each of the above processes is stored.

[0201] The RAM 112 functions as a storage area in which the program stored in the ROM 111 is expanded. The CPU 113 includes a processing circuit. The CPU 113 reads in the control program stored in the ROM 111 and controls the operation of the computer in accordance with the control program. The CPU 113 expands various data obtained through the operation of the computer in the RAM 112. The HDD 114 stores data necessary for reading and data obtained in the reading process. The HDD 114 functions as, for example, a storage device 120 shown in the drawing. Figure 27

[0202] The processing device 110 can also have an eMMC (embedded Multi Media Card), an SSD (Solid State Drive), an SSHD (Solid State Drive), or the like instead of the HDD 114.

[0203] The input device 140 includes at least one of a mouse, a keyboard, and a touch panel. The display device 150 includes at least one of a monitor and a projector. A device that functions as both the input device 140 and the display device 150, such as a touch panel, can also be used.

[0204] As for the control device 164, the hardware structure shown in the drawing can also be applied. Alternatively, Figure 27 one computer shown in the drawing can also function as the processing device 110 and the control device 164. Alternatively, the functions of the processing device 110 or the control device 164 can also be implemented through cooperation of a plurality of computers. ​

[0205] ​​In the above example, the case where the average shape of the weld 13 is circular is described. The above-described second determination can also be applied to the case where the shape of the weld 13 is other than circular. For example, the processing device 110 extracts the outer edge of the first region Rl after setting the first region Rl based on the result of the first determination. The processing device 110 can also determine whether the result of the first determination is appropriate based on the similarity of the outer edge of the first region Rl to a pre-set shape. For example, the processing device 110 calculates the degree of similarity between an image representing the outer edge of the first region Rl and an image including a pre-set shape. The degree of similarity is calculated based on feature points and the like of each image. The processing device 110 determines whether the result of the first determination is appropriate by comparing the degree of similarity to a pre-set threshold value.

[0206] By using the processing system and the processing method of the above-described embodiments, more accurate data related to the welding object can be obtained. By using the program for causing a computer to function as the processing system, the same effects can be obtained.

[0207] The processing of the above-described various data can also be recorded as a program that can cause a computer to execute, in a non-transitory computer-readable storage medium such as a magnetic disk (floppy disk and hard disk, etc.), an optical disk (CD-ROM, CD-R, CD-RW, DVD-ROM, DVD±R, DVD±RW, etc.), a semiconductor memory, and the like.

[0208] For example, the data recorded in the recording medium can be read by a computer (or an embedded system). The recording form (storage form) in the recording medium is arbitrary. For example, the computer reads out the program from the recording medium and causes the CPU to execute the instructions described in the program based on the program. In the computer, the acquisition (or reading out) of the program can also be performed through a network.

[0209] The above, several embodiments of the present application are illustrated, but these embodiments are suggested as examples, and are not intended to limit the scope of the application. These new embodiments can be implemented in other various ways, and various omissions, substitutions, changes, and the like can be made within the scope of the gist of the application. These embodiments and their modifications are included in the scope or gist of the application, and are included in the scope of the application and its equivalents described in the claims. Furthermore, each of the above-described embodiments can be implemented in combination with each other.

Claims

1. A processing apparatus that performs the following processing: The detector receives the detection result of the reflected wave from the detector, which includes a plurality of detection elements arranged in a first direction and a second direction that intersect each other, and performs detection including the transmission of ultrasonic waves toward the welding object and the detection of the reflected wave; Based on the detection results, a first determination is performed, in which joint and non-joint are determined at multiple points along the first and second directions of the welding object; as well as A second determination is performed, in which a first evaluation value is calculated based on the detection result or the result of the first determination, a second evaluation value is calculated based on multiple past first evaluation values, and the appropriateness of the result of the first determination is determined by comparing the difference between the first evaluation value and the second evaluation value with a first threshold.

2. The processing apparatus according to claim 1, wherein, In the second determination, When the detection result or the result of the first determination is classified into the first category by the first model, the result of the first determination is deemed appropriate. When the detection result or the result of the first determination is classified into the second category by the first model, the result of the first determination is deemed inappropriate.

3. A processing system, comprising: The processing apparatus according to claim 1; and The detector, When the processing device determines that the result of the first determination is inappropriate, the detector performs the detection on the welding object again.

4. The processing system according to claim 3, wherein, The processing device determines the quality of the weld on the object being welded based on the result of the first determination. When the processing device determines the weld to be good and the result of the first determination is unsuitable, the detector performs the detection on the weld object again under the first condition. When the processing device determines the weld to be defective and the result of the first determination is deemed inappropriate, the detector performs the detection on the weld object again under the second condition.

5. A processing system, comprising: The processing apparatus according to claim 1; and The detector, In the second determination, the processing device When the detection result or the result of the first determination is classified into the first category by the first model, the result of the first determination is deemed appropriate. When the detection result or the result of the first determination is classified into a second category or a third category by the first model, the result of the first determination is deemed inappropriate. When the detection result or the result of the first determination is classified into the second category by the first model, the detector performs the detection on the welding object again under the first condition. When the detection result or the result of the first determination is classified into the third category by the first model, the detector performs the detection on the welding object again under the second condition.

6. The processing system according to claim 3 or 5, wherein, The processing device derives data related to the welding object based on the result of the first determination. When the result of the first determination is deemed appropriate, the processing device uses the data.

7. The processing system according to claim 6, wherein, The data includes at least one of the following: the area of ​​the welded part, the diameter of the welded part, and the determination result of whether the weld on the welded object is good or bad.

8. The processing system according to claim 3 or 5, wherein, It also includes a coating device for applying coupling agent to the welding object. The detector comes into contact with the welding object that has been coated with the coupling agent.

9. The processing system according to claim 3 or 5, wherein, It also has: An arm, with the detector disposed at its front end; and The control device controls the detector and the arm. The control device drives the arm to bring the detector into contact with the object being welded.

10. The processing system according to claim 9, wherein, It also includes a camera device. The camera device captures an image of the object being welded. The control device detects the position of the welded part in the welded object based on the image, and causes the detector to contact the detected position.

11. A processing method, Perform the following processing: The detector receives the detection result of the reflected wave from the detector, which includes a plurality of detection elements arranged in a first direction and a second direction that intersect each other, and performs detection including the transmission of ultrasonic waves toward the welding object and the detection of the reflected wave; Based on the detection results, a first determination is performed, in which joint and non-joint are determined at multiple points along the first and second directions of the welding object; as well as A second determination is performed, in which a first evaluation value is calculated based on the detection result or the result of the first determination, a second evaluation value is calculated based on multiple past first evaluation values, and the appropriateness of the result of the first determination is determined by comparing the difference between the first evaluation value and the second evaluation value with a first threshold.

12. The processing method according to claim 11, wherein, In the second determination, when the detection result or the result of the first determination is classified into a first category by the first model, the result of the first determination is determined to be appropriate; when the detection result or the result of the first determination is classified into a second category by the first model, the result of the first determination is determined to be inappropriate.

13. The processing method according to claim 11 or 12, wherein, If the result of the first determination is deemed inappropriate, the detector is used to perform the detection on the welding object again.

14. A storage medium storing a program that causes a computer to perform the following processes: The detector receives the detection result of the reflected wave from the detector, which includes a plurality of detection elements arranged in a first direction and a second direction that intersect each other, and performs detection including the transmission of ultrasonic waves toward the welding object and the detection of the reflected wave; Based on the detection results, a first determination is performed, in which joint and non-joint are determined at multiple points along the first and second directions of the welding object; as well as A second determination is performed, in which a first evaluation value is calculated based on the detection result or the result of the first determination, a second evaluation value is calculated based on multiple past first evaluation values, and the appropriateness of the result of the first determination is determined by comparing the difference between the first evaluation value and the second evaluation value with a first threshold.

15. The storage medium according to claim 14, wherein, The program causes the computer to perform the following processing in the second determination: When the detection result or the result of the first determination is classified into the first category by the first model, the result of the first determination is deemed appropriate. When the detection result or the result of the first determination is classified into the second category by the first model, the result of the first determination is deemed inappropriate.

16. The storage medium according to claim 14 or 15, wherein, When the program determines that the result of the first determination is inappropriate, it causes the computer to re-execute the detection of the welding object using the detector.

Images