Method for determining a laser processing state and determination device

Abstract

The processing state determination method includes the following steps: using a light sensor to detect at least one of thermal radiation light, visible light, and reflected light generated by a welding portion formed on the surface of a workpiece by laser irradiation to the workpiece; obtaining a signal from the light sensor representing the change of the one of the thermal radiation, visible light, and reflected light in a time interval corresponding to the welding time of each workpiece; inputting a feature quantity including the signal intensity of the signal to a determination model for determining the processing state, determining the position and number of the welding area having a molten length and a molten width of the molten shape anomaly generated in the presence of foreign matter on the superimposed surface of the workpiece as the processing state; and outputting the determination result. The determination model is constructed based on training data including the feature quantity calculated under the condition of generating the molten shape anomaly and the processing state under the condition of generating the molten shape anomaly.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to a determination method of a processing state in laser processing for a stack welding and a determination device. BACKGROUND

[0002] A welding state determination method of laser welding and the like is disclosed in Patent Literature 1, which is applied to a laser welding method of welding a work by irradiating the work with laser light generated in a pulse shape, and is used to determine a welding state such as good / bad of welding in the work. In the method of Patent Literature 1, intensity of plasma light and reflected light emitted from the work at the time of laser welding is detected as a detection light intensity, and a characteristic value of each pulse is extracted from the detection light intensity in a predetermined extraction interval among one cycle of the detection light intensity corresponding to one pulse of the laser light for each pulse of the laser light. As the characteristic value of each pulse, an average value of the detection light intensity, a variation amount of differential processing, and an amplitude of differential processing, and the like are calculated. In the method of Patent Literature 1, a lower limit value or an upper limit value of the characteristic value of each pulse is obtained as an extreme value, the extreme value is compared with a given threshold value, and generation of a welding defect is determined as the welding state of each work.

[0003] PRIOR ART DOCUMENTS

[0004] PATENT LITERATURE

[0005] Patent Literature 1: JP Patent Publication No. 2000-153379 SUMMARY

[0006] According to one embodiment of the present disclosure, a determination method of a processing state in laser processing for a stack welding is provided. The method includes the steps of: detecting at least one of thermal radiation light, visible light, and reflected light generated in a welding portion formed on a surface of a work by irradiating the work with laser light; acquiring a signal indicating a change in the at least one of the thermal radiation light, the visible light, and the reflected light in a time interval corresponding to a welding time of each work from a light sensor; inputting a feature amount including a signal intensity of the signal based on the signal to a determination model for determining the processing state, to determine a position and a number of a welding area having a fusion length and a fusion width of a fusion shape abnormality generated when a foreign matter exists on a stack surface of the work as the processing state; and outputting the determined position and the number of the fusion shape abnormality as a determination result. The determination model is constructed based on training data including a feature amount calculated in a condition in which the fusion shape abnormality is generated and a processing state in the condition in which the fusion shape abnormality is generated.

[0007] According to one embodiment of the present disclosure, a determination device for a processing state in laser processing of a stacked weld is provided. The determination device includes an arithmetic circuit and a communication circuit. The communication circuit receives a signal generated by a light sensor detecting at least one of thermal radiation light, visible light, and reflected light generated by a weld portion formed on a surface of a workpiece by laser irradiation to the workpiece. The signal is a signal indicating a change in the at least one of the thermal radiation light, the visible light, and the reflected light in a time interval corresponding to a welding time of each workpiece. The arithmetic circuit acquires the signal via the communication circuit, inputs a feature quantity including a signal intensity of the signal based on the signal to a determination model for determining a processing state, determines a position and a number of a welding region having a fusion length and a fusion width, which are abnormal in a fusion shape due to a foreign object existing on a stacked surface of the workpiece, as the processing state, and outputs the determined position and number of the abnormal fusion shape as a determination result via the communication circuit. The determination model is constructed based on training data including a feature quantity calculated in a condition in which the abnormal fusion shape is generated and the processing state in the condition in which the abnormal fusion shape is generated. BRIEF DESCRIPTION OF DRAWINGS

[0008] Figure 1 is a diagram showing an outline of a determination system according to Embodiment 1 of the present disclosure.

[0009] Figure 2 is a diagram showing a structure of a laser processing device in the determination system.

[0010] Figure 3 is a diagram showing a structure of a light splitting device in the determination system.

[0011] Figure 4 is a block diagram showing a structure of a determination device in the determination system.

[0012] Figure 5 is a flowchart showing a determination process in the determination device.

[0013] Figure 6 is a diagram for explaining a signal acquired in the determination device.

[0014] Figure 7 is a diagram for explaining a process of calculating a feature quantity in the determination device.

[0015] Figure 8 is a diagram for explaining a process of the determination model in the determination device.

[0016] Figure 9 is a flowchart showing a training process of the determination model.

[0017] Figure 10 is a diagram for explaining a signal generated at the time of generation of an abnormal fusion shape. DETAILED DESCRIPTION

[0018] In laser welding, for example, in a case where a stain or a foreign matter is present in a workpiece, at the time of laser irradiation, sometimes an abnormality in a molten shape such as a hole is generated in a welded portion. In a method of determining generation of a welding defect based on a threshold value, it is possible to determine the presence or absence of such an abnormality, but it is difficult to determine detailed processing states such as the number and the position of the molten shape abnormality.

[0019] The present disclosure provides a determination method and a determination device capable of determining a processing state in laser processing for stack welding in detail.

[0020] Embodiments are described in detail below with appropriate reference to the accompanying drawings. However, sometimes a detailed description is omitted as necessary. For example, sometimes a detailed description for matters that are already widely known, a repeated description of substantially the same structure is omitted. This is to avoid the following description from becoming unnecessarily long, making it easy for those skilled in the art to understand. In addition, the inventors provide the drawings and the following description in order for those skilled in the art to fully understand the present disclosure, and the subject matter recited in the claims is not limited by these.

[0021] (Embodiment 1)

[0022] In Embodiment 1, as an example of using the determination method and the determination device related to the present disclosure, a determination system that detects a component of light generated in laser processing for stack welding, acquires a signal based on the detected component, and determines a processing state is described.

[0023] 1. Structure

[0024] The determination system related to Embodiment 1 is described using Figure 1 is a diagram that shows an outline of the determination system 100 related to the present embodiment. Figure 1

[0025] 1-1. Outline of the system

[0026] The determination system 100 is provided with: a laser processing device 30 that performs laser processing for stack welding; a light splitting device 40 that detects a component of light; and a determination device 50. The determination device 50 is an example of the determination device related to the present disclosure. The workpiece 70 for stack welding contains, for example, a metal, and if laser light 6 is irradiated, heat radiation light in the near-infrared region (also referred to as "heat radiation") caused by temperature rise, and metal-inherent luminescence or plasma luminescence mainly in the visible light region are generated. In addition, a part of the laser light 6 that does not contribute to processing is reflected as return light. In this way, if the laser light 6 is irradiated from the laser processing device 30 to the workpiece 70, heat radiation, visible light, and reflected light are generated in a molten portion 27 formed in the workpiece 70. The molten portion 27 is an example of a welded portion in the present embodiment.​

[0027] When the laser light 6 is irradiated, for example, if there is a foreign matter 80 containing a carbon-based material such as resin or oil between the two members 70a, 70b constituting the workpiece 70, a hole or a molten shape abnormality in which a protrusion is formed can occur in the welding region. The surface of the member 70a on the laser processing device 30 side after processing remains as a trace of the molten portion 27, and is a region having a molten length in the length direction of the welding processing and a molten width in the direction perpendicular to the length direction of the welding processing. By the melting of the foreign matter 80 present on the stacked surface of the workpiece 70, light emission can also occur in the molten portion 27.

[0028] The light emitted in the molten portion 27 is condensed by the laser processing device 30, transmitted to the light splitting device 40 through the optical fiber 13 connecting the laser processing device 30 and the light splitting device 40, and reaches the light splitting device 40. The light reaching the light splitting device 40 is split into heat radiation, visible light, and reflected light, and detected by the light sensor 22 of the light splitting device 40 and converted into a signal. The determination device 50 of the present embodiment determines the position and the number of the molten shape abnormality found in the form of a hole or the like, and further determines the size of the molten shape abnormality as the processing state, and outputs the determination result, if the signal is received from the light splitting device 40.

[0029] 1-2. Structure of the laser processing device

[0030] Figure 2 is a diagram illustrating the structure of the laser processing device 30 of the present embodiment. The laser processing device 30 is provided with a laser oscillator 1, a laser light transmission optical fiber 2, a lens barrel 3, a collimator lens 4, a condenser lens 5, 11, a first mirror 7, and a second mirror 8.

[0031] The laser oscillator 1 supplies light for generating laser light 6 in the form of a pulse having a wavelength of about 1070 nanometers (nm), for example. The light supplied from the laser oscillator 1 is amplified during transmission through the laser light transmission optical fiber 2, passes through the collimator lens 4 for obtaining a parallel light beam, forms the laser light 6, and travels straight in the lens barrel 3. The lens barrel 3 constitutes a processing head in the laser processing device 30.

[0032] The laser light 6 is reflected except for a portion that is transmitted through the first mirror 7, condensed by the condenser lens 5, and irradiated to the workpiece 70 fixed by the pressing jig 26 on a scanning stage (not shown), for example. Thereby, laser processing for stacked welding of the workpiece 70 is performed. In addition, the wavelength of the laser light 6 is not particularly limited to 1070 nm, and a wavelength at which the absorption rate of the material is high is preferably used.

[0033] If the laser light 6 is irradiated, heat radiation from the workpiece 70, visible light based on plasma light emission, and reflected light of the laser light 6 are generated at the molten portion 27. These lights pass through the first mirror 7, are reflected at the second mirror 8, are condensed by the condensing lenses 11, and are transmitted to the light splitting device 40 through the optical fiber 13. In addition, the light partially transmitted at the second mirror 8 can be detected by a camera or a sensor.

[0034] 1-3. Configuration of light splitting device

[0035] Figure 3 Fig. 1 is a diagram illustrating a configuration of the light splitting device 40 according to the present embodiment. The light splitting device 40 includes a collimator lens 15, a third mirror 16, a fourth mirror 17, a fifth mirror 18, condensing lenses 19, 20, 21, a light sensor 22, a transmission cable 23, and a controller 24 in the inside of a housing 28. The housing 28 prevents stray light from entering the inside from the outside of the light splitting device 40 and prevents light from leaking from the inside.

[0036] The collimator lens 15 makes the light transmitted from the laser processing device 30 through the optical fiber 13 into parallel light. The third mirror 16 transmits visible light of, for example, a wavelength of 400 nm to 700 nm and reflects components other than the visible light. The fourth mirror 17 reflects reflected light of the laser light 6 of, for example, a wavelength of about 1070 nm and transmits components other than the reflected light. The fifth mirror 18 reflects heat radiation of, for example, a wavelength of 1300 nm to 1550 nm.

[0037] The light that has passed through the collimator lens 15 is split into visible light, reflected light, and heat radiation by the third mirror 16, the fourth mirror 17, and the fifth mirror 18 and is condensed by the condensing lenses 19 to 21, respectively. In addition, a band pass filter can be provided in the light paths after the third mirror 16, the fourth mirror 17, and the fifth mirror 18 to select the wavelength to be passed, respectively.

[0038] The light sensor 22 includes, for example, light sensors 22a, 22b, 22c each having high sensitivity to a different wavelength. The light sensors 22a, 22b, 22c detect the visible light, the reflected light, and the heat radiation condensed by the condensing lenses 19 to 21, respectively, and generate an electric signal corresponding to the intensity of the detected light. In addition, the light sensor 22 can be constituted by one light sensor that can detect the intensity of each wavelength.

[0039] The electric signal generated by the light sensor 22 is transmitted to the controller 24 via the transmission cable 23. The controller 24 is a hardware controller that comprehensively controls the operation of the entire light splitting device 40. The controller 24 includes a CPU and a communication circuit and the like, and transmits the electric signal received from the light sensor 22 to the determination device 50. The controller 24 has, for example, an A / D converter that converts the analog electric signal into a digital signal (also simply referred to as "signal"). In addition, regarding the sampling period at the time of conversion into a digital signal, in the determination of the machining state, from the viewpoint of securing a sufficient number of samples in order to capture the tendency of the values of the characteristics of the machining process and the physical quantities, for example, it is preferable to perform the output control of the laser 6 at a time of 1 / 100 or less.

[0040] 1-4. Structure of determination device

[0041] Figure 4 is a block diagram illustrating the structure of the determination device 50 of the present embodiment. The determination device 50 is constituted by an information processing device such as a computer, for example. The determination device 50 has a CPU 51 that performs processing of operation, a communication circuit 52 for communicating with other devices, and a storage device 53 that stores data and computer programs.

[0042] The CPU 51 is an example of an operation circuit of the determination device in the present embodiment. The CPU 51 realizes given functions including training and execution of the determination model 57 by executing the control program 56 stored in the storage device 53. The determination device 50 realizes the functions as the determination device in the present embodiment by the CPU 51 executing the control program 56. In addition, the operation circuit that constitutes the CPU 51 in the present embodiment can be realized by various processors such as an MPU or a GPU, and can be constituted by one or a plurality of processors.

[0043] The communication circuit 52 is, for example, a communication circuit that communicates in compliance with standards such as IEEE 802.11, 4G, or 5G. The communication circuit 52 can perform wired communication in compliance with standards such as Ethernet (registered trademark), for example. The communication circuit 52 can be connected to a communication network such as the Internet. Furthermore, the determination device 50 can directly communicate with other devices via the communication circuit 52, and can communicate via an access point. In addition, the communication circuit 52 can be constituted so as to be able to communicate with other devices without going through a communication network. For example, the communication circuit 52 can include connection terminals such as a USB (registered trademark) terminal and an HDMI (registered trademark) terminal.

[0044] The storage device 53 is a storage medium that stores a computer program and data required to realize the functions of the determination system 100, stores the control program 56 executed in the CPU 51 and various data. The storage device 53 stores the determination model 57 after the determination model 57 is constructed. The determination model 57 is constructed based on training data including a feature amount calculated from a signal under a condition in which a molten shape abnormality occurs and a processing state at the time when a molten shape abnormality has occurred. Details of the determination model 57 are described later.

[0045] The storage device 53 is constituted by, for example, a magnetic storage device such as a hard disk drive (HDD), an optical storage device such as an optical disk drive, or a semiconductor storage device such as an SSD. The storage device 53 can have, for example, a temporary storage element constituted by a RAM such as a DRAM or an SRAM, and functions as an internal memory of the CPU 51.

[0046] 2. Action

[0047] In the determination system 100 constituted as described above, for example, as shown in FIG. 1, the light splitting device 40 detects heat radiation, visible light, and reflected light generated at the molten portion 27 by irradiation of the laser light 6 with the light sensor 22. The light splitting device 40 transmits a signal corresponding to the intensity of the detected heat radiation, visible light, and reflected light to the determination device 50. The action of the determination device 50 in the present system 100 is described below. Figure 1

[0048] 2-1. Determination processing

[0049] Hereinafter, determination processing in which the determination device 50 determines the position, number, and size of a molten shape abnormality is described using FIG. 2. Figures 5 to 8

[0050] Figure 5 is a flowchart illustrating the determination processing in the determination device 50 of the present embodiment. Each processing shown in the present flowchart is executed by, for example, the CPU 51 of the determination device 50. The present flowchart is started, for example, by inputting a given operation for starting the determination processing from an input device connected via the communication circuit 52 by a user or the like of the determination system 100.

[0051] First, the CPU 51 acquires a signal corresponding to heat radiation, visible light, and reflected light detected in the light sensor 22 of the light splitting device 40 via the communication circuit 52 (S1).

[0052] Figure 6 is a graph for explaining the signal acquired in the determination device 50. Figure 6 (A) of illustrates a signal waveform of a signal corresponding to any one of heat radiation, visible light, and reflected light in a case where a molten shape abnormality has occurred at the time of processing. Figure 6 ​​(B) illustrates the signal waveforms of any of the thermal radiation, visible light, and reflected light in the absence of a molten shape. Figure 6 (C) represents the output of the laser 6 irradiating the workpiece 70. Figure 6 The signals (A) and (B) and their transmission Figure 6 It corresponds to any of the thermal radiation, visible light, and reflected light generated by the laser output of (C).

[0053] exist Figure 6 In (A) to (C), the horizontal axis represents time, and the vertical axis represents signal strength. Figure 6 (A), (B)) or laser output ( Figure 6 (C)). Furthermore, time T1 represents the time interval corresponding to one pulse of laser 6, and time T2 represents the time interval of the peak output of the laser output excluding the rising and falling edges. Here, in the laser processing apparatus 30 of this embodiment, welding of each workpiece 70 is performed within time T1. In step S1, the CPU 51 acquires signals representing changes in thermal radiation, visible light, and reflected light within time T1 corresponding to the welding time of each workpiece 70.

[0054] like Figure 6 As shown in (A), if an abnormal melt shape occurs, then... Figure 6 Compared with the normal state of (B), a signal with a peak waveform that temporarily increases the signal strength is obtained. The peak signal when a melt shape abnormality occurs is caused, for example, by the luminescence caused by the foreign object 80 that causes the abnormality. Furthermore, when a melt shape abnormality occurs, the luminescence sometimes momentarily attenuates due to the foreign object 80, thus temporarily generating an attenuation peak. In this case, a signal with a peak waveform that temporarily decreases the signal strength is obtained. This also applies to the following description. Figure 5 The flowchart illustrates the process of extracting local minimum values ​​and calculating the integral value of the signal strength over the interval Tp by subtracting the average value Sa. The following flowchart serves as an example to illustrate the process for a waveform that generates a temporary increase in signal strength peaks.

[0055] exist Figure 5 In the flowchart, next, the CPU 51 calculates the feature quantity input to the decision model 57 based on the acquired signal (S2). In this embodiment, in addition to calculating the signal strength after preprocessing such as normalization, the CPU 51 also calculates the intensity value based on the signal strength in the peak (hereinafter referred to as the "peak intensity value") as the feature quantity.

[0056] Figure 7 This is a diagram illustrating the process (S2) for calculating the characteristic quantity in the determination device 50. Figure 7 (A) andFigure 6 (A) also indicates a time change of signal intensity of a signal corresponding to heat radiation, visible light, or reflected light in a case where a molten shape abnormality occurs. The signal is used Figure 7 to explain Figure 5 the processing of the characteristic quantity in which the peak intensity value is calculated in step S2.

[0057] The CPU 51 first performs processing of detecting a peak of the acquired signal. The CPU 51, for example, performs an operation of comparing values of signal intensity for each sampling period, and extracts a point having a value larger than that of points adjacent in time as a local maximum value. At this time, a threshold value can also be set from the viewpoint of limiting the value extracted as the local maximum value to a given signal intensity or more. The CPU 51, for example, extracts a local minimum value of signal intensity as well as the local maximum value, detects a peak as a region of a signal waveform in an interval Tp sandwiched by 2 points adjacent to the local maximum value. The interval Tp corresponds to a time of occurrence of the peak. Figure 7 (B) indicates an example in which the peak of the interval Tp is detected in the signal of (A). Figure 7 (A) of (B) indicates an example in which the peak of the interval Tp is detected in the signal of (A).

[0058] After the detection of the peak, the CPU 51 calculates an average value Sa of signal intensity other than the peak. The average value Sa is, for example, calculated as an average value of signal intensity in a time (T2-Tp) in which the time T2 from the peak output in 1 pulse of the laser 6 is removed, from the signal intensity. Figure 7 (C) indicates an example in which the average value Sa is calculated in the example of (B). Figure 7 (C) indicates an example in which the average value Sa is calculated in the example of (B).

[0059] Next, the CPU 51 calculates an integral value calculated for the interval Tp corresponding to the time of occurrence of the peak, as a peak intensity value, in a value obtained by subtracting the average value Sa of signal intensity other than the peak from the signal intensity of the interval Tp. Figure 7 (D) indicates an example in which the integral value is calculated in the example of (C). The integral value corresponds to an area of the region Rp indicated in (D). Figure 7 (D) indicates an example in which the integral value is calculated in the example of (C). The integral value corresponds to an area of the region Rp indicated in (D). Figure 7 (D) indicates an example in which the integral value is calculated in the example of (C). The integral value corresponds to an area of the region Rp indicated in (D).

[0060] After the calculation of the characteristic quantity as described above (S2), the CPU 51 inputs the characteristic quantity to the determination model 57, and performs processing of the determination model of determining a position, a number, and a size of a molten shape abnormality (S3). The characteristic quantity of signal intensity is, for example, input to the determination model 57 as an amplitude of a signal waveform for each sampling period in A / D conversion.

[0061] Figure 8 is a diagram for explaining the processing of the determination model (S3). Figure 8 (A) and Figure 6Similarly, (A) represents the signal waveform when an abnormal melt shape is generated. Figure 8 (B) schematically illustrates the generation Figure 8 The appearance of component 70a on the side of the laser processing apparatus 30 of the processed workpiece 70 when the (A) signal is received. Figure 8 In (B), a hole 85 is generated in the weld area 270 with a melt length Wx and a melt width Wy as an example of an abnormal melt shape.

[0062] In this embodiment, the laser processing apparatus 30 performs welding across a melting length Wx for each workpiece 70 within a time T1 equivalent to one pulse. Figure 8 In example (A), in laser processing apparatus 30 Figure 8 When machining is performed in the positive direction of the x-axis in (B), a hole 85 is generated, and correspondingly, a peak value of the interval Tp is generated, which is detected in step S2.

[0063] exist Figure 8 In the example, during the processing of the decision model (S3), CPU51 will determine based on... Figure 8 The signal strength and peak intensity value calculated from the (A) signal are input into the judgment model 57 to make a judgment. Figure 8 The location, number, and size of holes 85 in (B). For example, in an orthogonal coordinate system with the welding start point on component 70a as the origin, the location is determined by the coordinates of the centroid of hole 85. The size is determined, for example, by the area of ​​hole 85. Figure 8 In (B), since there is no abnormal melt shape except for hole 85, the quantity is determined to be "1".

[0064] Back Figure 5 The CPU 51 outputs the determination results of the location, quantity, and size of the molten shape abnormality, such as the hole 85, through the communication circuit 52 (S4). The determination results can be received and displayed by an external information processing device or display device, for example. In addition, the determination device 50 may also be equipped with a display device (e.g., a monitor) that can communicate with the CPU 51, so that the display device displays the determination results.

[0065] After that, CPU51 ended. Figure 5 The flowchart. Figure 5 The flowchart, for example, is repeated whenever welding is performed on each workpiece 70.

[0066] According to the above determination processing, the determination device 50 of the present embodiment acquires the signal generated by the light sensor 22 of the light splitting device 40 (S1), calculates the feature quantity from the signal (S2), and determines the position, number, and size of the molten shape abnormality based on the feature quantity by the determination model 57 (S3). Thus, the determination device 50 can determine the processing state related to the molten shape abnormality in detail in the laser processing for the overlay welding.

[0067] In addition, in Figure 5 Step S2, the feature quantity can be calculated for all of the thermal radiation, the visible light, and the reflected light, or can be calculated for only any one of the thermal radiation, the visible light, and the reflected light. Further, in the processing of the determination model (S3), the determination model 57 can determine only the position and number of the molten shape abnormality, for example.

[0068] Further, in Step S2, the signal intensity integral value can be detected and calculated for the above-mentioned decay peak value as well. In this case, the value with respect to the decay peak value becomes negative, on the other hand, the peak intensity value calculated for the increase peak value described in the example becomes positive. In this way, the change in the luminescence caused by the foreign matter 80 can be reflected in the feature quantity, distinguishing between the peak caused by the decay of the signal intensity and the peak caused by the increase. In addition, even in the case of detecting the decay peak value, the integral value of the signal intensity with respect to the peak is not limited to this, and for example, only the presence and size of the peak can be focused on, and the absolute value thereof can be used in the feature quantity. Figure 7

[0069] 2-2. Training processing

[0070] The training processing for constructing the determination model 57 will be described below using Figure 9 and Figure 10 .

[0071] Figure 9 is a flowchart illustrating the training processing of the determination model 57. Each processing of the present flowchart is executed by the CPU 51 of the determination device 50, for example.

[0072] First, the CPU 51 acquires the training data stored in the storage device 53 in advance (S11), for example.

[0073] ​The training data is data in which a feature amount such as a signal intensity and a peak intensity value of thermal radiation, visible light, and reflected light and a position, a number, and a size of a molten shape abnormality as a processing state are associated. The training data is constructed by associating and recording a feature amount calculated from a signal of thermal radiation, visible light, and reflected light detected based on laser processing under a plurality of conditions in which a processing state changes and a processing state determined by appearance measurement of the welding region 270 after processing. The appearance measurement is performed, for example, by observation of the welding region 270 by an optical microscope or measurement in an image of the welding region 270 captured by a camera, but is not particularly limited thereto.

[0074] Figure 10 is a graph for explaining a signal generated at the time of generation of a molten shape abnormality. In the construction of the training data, a feature amount of a signal of each of various waveform patterns such as Figure 10 A feature amount of a signal of each of various waveform patterns exemplified above and a corresponding processing state are associated.

[0075] In (A) of FIG. 22, Figure 10 In (A) of FIG. 22, a peak corresponding to one molten shape abnormality is detected in all of the signals Lt, Lv, and Lr generated respectively corresponding to the intensities of thermal radiation, visible light, and reflected light. In (B) of FIG. 22, Figure 10 In (B) of FIG. 22, peaks of one molten shape abnormality are detected in the two signals Lt, Lv of thermal radiation and visible light. In (C) of FIG. 22, Figure 10 In (C) of FIG. 22, peaks of one molten shape abnormality are detected only in the signal Lr of reflected light. In (D) of FIG. 22, Figure 10 In (D) of FIG. 22, two peaks corresponding to two molten shape abnormalities are detected in each of the signals Lt, Lv, and Lr of thermal radiation, visible light, and reflected light. As shown in (A) and (D) of FIG. 22, in the signal Lr of reflected light, there is a tendency that peaks are generated at an earlier time than the signals Lt, Lv of thermal radiation and visible light. Figure 10

[0076] By including a feature amount of a signal based on a detection pattern of such various peaks and a corresponding processing state in the training data, a determination model 57 that can determine a processing state in detail even when conditions such as light, time, and number of detection of peaks change can be generated by the processing described later. In the present embodiment, by using a feature amount based on all of thermal radiation, visible light, and reflected light, even in a case where peaks are detected only in a signal of one or two of thermal radiation, visible light, and reflected light, a tendency of generation of a molten shape abnormality can be reflected in the determination model 57. In the training data, data including two or fewer peaks as a number of peaks assumed at the time of actual processing is used, but is not particularly limited thereto, and data including three or more peaks can also be used. Furthermore, an interval of time in which one peak is considered can also be set in advance.​

[0077] The CPU 51 generates the determination model 57 by performing machine learning using the training data if the training data is acquired (S1). The determination model 57 is, for example, a regression model based on a random forest or a neural network.

[0078] According to the above training processing, the determination model 57 can be generated as a learned model that determines the position, the number, and the size of the molten shape abnormality based on the feature amounts based on the signals corresponding to the thermal radiation, the visible light, and the reflected light detected in the laser processing.

[0079] In addition, the training processing of the determination model 57 can be performed in an information processing device different from the determination device 50. The determination device 50 can acquire the constructed determination model via the communication circuit 52 through a communication network, for example.

[0080] Further, the feature amount of the case where the molten shape abnormality does not occur and the processing state of the case where the molten shape abnormality does not occur can also be included in the training data of the determination model 57. The feature amount of the case where the molten shape abnormality does not occur can be, for example, a peak intensity value "0". The processing state of the case where the molten shape abnormality does not occur can be, for example, a position "0", a number "0", and a size "0" of the molten shape abnormality.

[0081] 3. Effects and the like

[0082] As described above, in the present embodiment, the determination processing (S1 to S4) provides a determination method of the processing state in the laser processing for the stacked welding. The present method includes the following steps: detecting at least one of the thermal radiation (thermal radiation light), the visible light, and the reflected light generated in the molten portion 27 (an example of a welding portion) formed on the surface of the workpiece 70 by irradiating the laser 6 to the workpiece 70 using the optical sensor 22; a step (S1) of acquiring, from the optical sensor 22, a signal representing a change in the thermal radiation, the visible light, and the reflected light in a time T1 (a time interval) corresponding to the welding time of each workpiece 70; inputting a feature amount including a signal intensity of the signal based on the signal to the determination model 57 that determines the processing state to determine the position and the number of the molten shape abnormality in the welding region 270 having a molten length Wx and a molten width Wy of the molten shape abnormality existing in the case where the foreign matter 80 exists in the stacking surface of the workpiece 70 as the processing state (S2, S3); and a step (S4) of outputting the determined position and the number of the molten shape abnormality as a determination result. The determination model 57 is constructed based on the training data including the feature amount calculated in the case where the molten shape abnormality occurs and the processing state in the case where the molten shape abnormality occurs.

[0083] According to the above method, a signal (S1) of one or more of thermal radiation, visible light, and reflected light generated and detected by irradiation with the laser 6 is acquired, a characteristic quantity including a signal intensity is calculated, and a position and a number of the molten shape abnormality are determined as the processing state (S2, S3). Thus, the processing state related to the molten shape abnormality can be determined in detail based on the signal intensity of at least one of thermal radiation, visible light, and reflected light detected in laser processing for the overlay welding.

[0084] In the present embodiment, the determination process (S2, S3) further includes detecting a peak value of the signal and further determining a size of the molten shape abnormality as the processing state. The output process (S4) includes further outputting the determined size of the molten shape abnormality as the determination result. The characteristic quantity includes an example of the intensity value based on the signal intensity of the signal at the peak value, i.e., a peak intensity value. Thus, the processing state can be determined in more detail including the size of the molten shape abnormality based on the peak intensity value.

[0085] In the present embodiment, the intensity value is an integral value obtained by integrating a value of the signal intensity of the signal at the peak value with respect to the interval Tp (time of generation of the peak value) with respect to the average value Sa of the signal intensities of the signals other than the peak value (refer to Figure 7 Thus, the intensity of the luminescence associated with the generation of the molten shape abnormality caused by the foreign matter 80 can be reflected in the characteristic quantity, and thus the processing state such as the size of the molten shape abnormality can be determined in detail.

[0086] In the present embodiment, the determination model 57 includes a learned model generated by machine learning using training data that correlates a characteristic quantity calculated from a signal of at least one of thermal radiation, visible light, and reflected light detected based on laser processing under each of a plurality of conditions in which the processing state changes and the processing state determined by appearance measurement of the weld region 270. Thus, the determination model 57 that determines the processing state can be obtained from the characteristic quantity based on at least one of thermal radiation, visible light, and reflected light.

[0087] In the determination system 100 of this embodiment, the determination device 50 is an example of a determination device for determining the processing state in laser processing for superimposed welding. The determination device 50 includes a CPU 51, which is an example of an arithmetic circuit, and a communication circuit 52. The communication circuit 52 receives a signal generated by the light sensor 22 detecting at least one of thermal radiation (thermal radiation light), visible light, and reflected light generated from the molten portion 27 (an example of a welded portion) formed on the surface of the workpiece 70 by irradiating the workpiece 70 with the laser 6. The signal is a signal representing the change of at least one of thermal radiation, visible light, and reflected light in time T1, which is an example of a time interval corresponding to the welding time of each workpiece 70. CPU 51 obtains a signal via communication circuit 52 (S1), and inputs feature quantities, including the signal strength based on the signal, into the determination model 57 for determining the processing state. It then determines the position and number of welding areas 270 with melt length Wx and melt width Wy that exhibit molten shape abnormalities when foreign matter 80 is present on the superimposed surface of the workpiece 70, using these as the processing state (S2, S3). The determined position and number of molten shape abnormalities are then output as the determination result via communication circuit 52 (S4). The determination model 57 is constructed based on training data including feature quantities calculated under conditions of molten shape abnormalities and processing states under such conditions.

[0088] Based on the determination device 50 described above, the determination method described above can be executed to determine in detail the processing status in laser processing used for superimposed welding.

[0089] (Other implementation methods)

[0090] As described above, the embodiments described herein are examples of the technology disclosed in this application. However, the technology in this disclosure is not limited to this and can be applied to embodiments that are suitable for modification, substitution, addition, omission, etc. Furthermore, new embodiments can be made by combining the constituent elements described in the above embodiments.

[0091] In the above-described embodiment 1, the determination device 50 calculates characteristic quantities of the signal strength and peak intensity value during the determination process. Figure 5 (S2). In this embodiment, in step S2, the peak intensity value may not be specifically calculated, and only the signal intensity may be used as a characteristic quantity.

[0092] In the above-described embodiment 1, the determination device 50 acquires signals corresponding to the thermal radiation, the visible light, and the reflected light detected in the light sensor 22 of the light splitting device 40 (S1). In the present embodiment, the determination device 50 can also acquire signals with respect to only one or two of the thermal radiation, the visible light, and the reflected light. In this case, in steps S2 to S3, the characteristic quantity is calculated from the signals of only one or two of the thermal radiation, the visible light, and the reflected light, and is input to the determination model 57. Further, in the present embodiment, the determination model 57 can also be constructed with the characteristic quantity based on the signals of only one or two of the thermal radiation, the visible light, and the reflected light and the processing state as training data.

[0093] In the above-described embodiment 1, the determination model 57 is constructed with the characteristic quantity such as the signal intensity and the position, the number, and the size of the molten shape abnormality as training data (S11 to S12). In the present embodiment, the determination model 57 can also be constructed with the characteristic quantity and the position and the number of the molten shape abnormality as training data. In this case, the determination device 50 determines the position and the number of the molten shape abnormality as the processing state in the determination process (S1 to S4).

[0094] According to the determination method and the determination device in the present disclosure, in laser processing for lap welding, the processing state can be determined in detail, particularly with respect to the molten shape abnormality generated in the welding region.

[0095] The present disclosure is not limited to the above-described embodiments, and various modifications can be made. That is, embodiments obtained by combining technical means appropriately modified by those skilled in the art are also within the scope of the present disclosure.

[0096] Industrial Applicability

[0097] The present disclosure can be applied to a determination system of a processing state in laser processing for lap welding, particularly to a method and a device for determining a molten shape abnormality of a welded portion.

[0098] Explanation of Reference Signs

[0099] 1 Laser oscillator

[0100] 2 Laser transmission optical fiber

[0101] 3 Mirror barrel

[0102] 4 Collimator lens

[0103] 5, 11 Condenser lens

[0104] 6 Laser light

[0105] 7 First mirror

[0106] 8 Second mirror

[0107] 13 optical fiber

[0108] 15 collimator lens

[0109] 16 3rd mirror

[0110] 17 4th mirror

[0111] 18 5th mirror

[0112] 19, 20, 21 condenser lens

[0113] 22 light sensor

[0114] 23 transmission cable

[0115] 24 controller

[0116] 26 pressing jig

[0117] 27 fusion portion

[0118] 30 laser processing device

[0119] 40 light splitting device

[0120] 50 determination device

[0121] 51 CPU

[0122] 52 communication circuit

[0123] 53 storage device

[0124] 56 control program

[0125] 57 determination model

[0126] 70 workpiece

[0127] 70a, 70b member

[0128] 85 hole

[0129] 100 determination system

[0130] 270 welding region

Claims

1. A determination method for determining a processing state in laser processing of stack welding, the determination method comprising the following steps: detecting at least one of thermal radiation light, visible light, and reflected light generated at a weld portion formed on a surface of a work by irradiating laser light to the work using a light sensor; acquiring, from the light sensor, a signal indicating a change in the at least one of the thermal radiation light, the visible light, and the reflected light in a time interval corresponding to a welding time of each of the works; inputting, to a determination model for determining the processing state, a feature amount including a signal intensity of the signal based on the signal to determine a position and a number of a welding area having a molten length and a molten width of a molten shape abnormality generated when a foreign matter is present on a stack surface of the work as the processing state; and outputting the determined position and the number of the molten shape abnormality as a determination result, the determination model is constructed based on training data including the feature amount calculated in a condition where the molten shape abnormality is generated and the processing state in the condition where the molten shape abnormality is generated, the determination step includes detecting a peak value of the signal, and further determining a size of the molten shape abnormality as the processing state, the output step includes further outputting the determined size of the molten shape abnormality as the determination result, the feature amount includes an intensity value based on a signal intensity of the signal at the peak value.

2. The determination method according to claim 1, wherein the intensity value is an integrated value obtained by integrating a value obtained by subtracting an average value of signal intensities of the signal except for the peak value from a signal intensity of the peak value with respect to a generation time of the peak value.

3. The determination method according to claim 1 or 2, wherein the determination model includes a learned model generated by machine learning using training data in which a feature amount calculated from a signal of the at least one of the thermal radiation light, the visible light, and the reflected light detected in the laser processing under each of a plurality of conditions in which the processing state changes and the processing state determined by appearance measurement of the welding area are associated.

4. A determination device for determining a processing state in laser processing of stack welding, the determination device comprising: an arithmetic circuit; and a communication circuit that receives a signal generated by a light sensor detecting at least one of thermal radiation light, visible light, and reflected light generated at a weld portion formed on a surface of a work by irradiating laser light to the work, the signal is a signal indicating a change in the at least one of the thermal radiation light, the visible light, and the reflected light in a time interval corresponding to a welding time of each of the works, the arithmetic circuit performs the following processing: acquiring the signal through the communication circuit; a determination model that determines the processing state inputs a feature amount including a signal intensity of the signal based on the signal to determine a position and a number of a welding region having a fusion length and a fusion width of a fusion shape abnormality generated in a case where a foreign object exists on a superimposition surface of the workpiece, as the processing state; and the position and the number of the fusion shape abnormality determined are output as a determination result through the communication circuit, the determination model is constructed based on training data including the feature amount calculated in a condition where the fusion shape abnormality is generated and the processing state in the condition where the fusion shape abnormality is generated, the operation circuit performs the following processing: a peak value of the signal is detected, and a size of the fusion shape abnormality is further determined as the processing state; and the size of the fusion shape abnormality determined is further output as a determination result through the communication circuit, the feature amount includes an intensity value based on a signal intensity of the signal at the peak value.

5. The determination device according to claim 4, wherein the intensity value is an integrated value obtained by integrating a value obtained by subtracting an average value of signal intensities of the signal other than the peak value from a signal intensity of the peak value with respect to a generation time of the peak value.

6. The determination device according to claim 4 or 5, wherein the determination model includes a learned model generated by machine learning using training data in which a feature amount calculated from the at least one of the thermal radiation light, the visible light, and the reflected light detected according to the laser processing under each of a plurality of conditions in which the processing state changes and the processing state determined by appearance measurement of the welding region are associated.

Images