Laser processing equipment

Abstract

The laser processing apparatus (50) includes: a drive unit (5) that changes the relative position of a processing head (2) and a workpiece (W), the processing head (2) including a focusing optical system that focuses a laser emitted from a laser oscillator (1) to irradiate the workpiece (W); a control unit (3) that controls the laser oscillator (1), the processing head (2) and the drive unit (5) based on numerical parameters related to laser processing, i.e., processing parameters, to perform processing; a processing status observation unit (52) that detects the light intensity of a predetermined multiple wavelength bands of interest of the light emitted from the workpiece (W) by the laser irradiation, i.e., the processing light (8), as multiple photosensitive sensor signals; a feature extraction unit (53) that extracts at least one of the correlation indicators between the multiple photosensitive sensor signals and a feature quantity that can be obtained from a photosensitive sensor signal; and a correction calculation unit (55) that, based on the feature quantity, determines the processing parameters used to perform correction as correction parameters and determines the correction amount of the correction parameters.

Description

Technical Field

[0001] This invention relates to a laser processing apparatus for processing workpieces by irradiation with a laser. Background Technology

[0002] In sheet metal laser processing, sometimes the process starts well, but defects can occur during continuous processing due to heat accumulation in the processing head components and the workpiece. Sheet metal laser processing involves multiple processing parameters such as focal position, cutting speed, gas pressure, and laser output. The processing results, such as the amount of deposits and the surface roughness, also involve multiple factors, requiring considerable processing time for adjustments.

[0003] The laser processing machine disclosed in Patent Document 1 includes: a detection unit that detects the return light from the processing point side irradiated by the accompanying laser toward the laser processing head; and a monitoring unit that monitors the processing state of the laser processing by selecting the level of a specific wavelength band of light corresponding to the processing conditions in the return light detected by the detection unit according to a time sequence.

[0004] Patent Document 1: Japanese Patent Application Publication No. 2019-166543 Summary of the Invention

[0005] The laser processing apparatus disclosed in Patent Document 1 monitors light levels selected according to a time sequence, resulting in low accuracy in detecting the processing status. Furthermore, the laser processing apparatus disclosed in Patent Document 1 only determines whether the processing is successful, making it difficult to adjust the processing conditions.

[0006] The present invention was made in view of the above circumstances, and its purpose is to obtain a laser processing apparatus that can detect the processing status at high speed or with high precision and adjust the processing conditions accordingly.

[0007] To address the aforementioned issues and achieve the objectives, the laser processing apparatus of the present invention comprises: a drive unit that changes the relative position of a processing head and a workpiece, the processing head including a focusing optical system for converging laser light emitted from a laser oscillator to irradiate the workpiece, and a processing gas supply unit for supplying processing gas toward the workpiece; a control unit that controls the laser oscillator, processing head, and drive unit to perform processing based on numerical parameters related to laser processing, i.e., processing parameters; a processing state observation unit that detects the light intensity of multiple predetermined wavelength bands of interest of the light emitted from the workpiece by laser irradiation, i.e., processing light, as multiple photosensitive sensor signals; a feature extraction unit that extracts at least one of the correlation indicators between the multiple photosensitive sensor signals and a feature quantity that can be obtained from one photosensitive sensor signal; and a correction calculation unit that, based on the feature quantity, determines the processing parameters used to perform correction as correction parameters and determines the correction amount of the correction parameters.

[0008] The effects of the invention

[0009] The laser processing apparatus of the present invention has the following effect: it can detect the processing status at high speed or with high precision and adjust the processing conditions accordingly. Attached Figure Description

[0010] Figure 1 This is a diagram showing the structure of the laser processing apparatus according to Embodiment 1.

[0011] Figure 2 This is a flowchart illustrating an example of the sequence of actions related to the adjustment of processing parameters performed by the laser processing apparatus according to Embodiment 1.

[0012] Figure 3 This is a diagram showing the structure of the processing status observation section of the laser processing apparatus according to Embodiment 1.

[0013] Figure 4 This diagram illustrates an example of the wavelength bands of light received by the first, second, and third optical sensors in the laser processing apparatus according to Embodiment 1.

[0014] Figure 5 This is a diagram showing the structure of the laser processing apparatus according to the first variation of Embodiment 1.

[0015] Figure 6 This is a diagram showing the structure of the laser processing apparatus involved in the second variation of Embodiment 1.

[0016] Figure 7 This is a diagram showing the structure of the laser processing apparatus involved in the third variation of Embodiment 1.

[0017] Figure 8 This is a diagram showing the structure of the processing state resolver of the laser processing apparatus according to the fourth variation of Embodiment 1.

[0018] Figure 9 This is a diagram showing the structure of the processing state resolver of the laser processing apparatus according to Embodiment 2.

[0019] Figure 10 This is a diagram showing the structure of the neural network model involved in Implementation Method 2.

[0020] Figure 11 This is a diagram showing the structure of the laser processing apparatus involved in Embodiment 3.

[0021] Figure 12 This diagram illustrates a processor in which at least a portion of the control unit, drive unit, condenser lens position change drive unit, processing status observation unit, feature quantity extraction unit, evaluation unit, and correction quantity calculation unit of the laser processing apparatus according to Embodiment 1 are implemented by a processor.

[0022] Figure 13 This is a diagram showing the processing circuit in which at least a portion of the control unit, drive unit, condenser lens position change drive unit, processing status observation unit, feature quantity extraction unit, evaluation unit, and correction quantity calculation unit of the laser processing apparatus according to Embodiment 1 are implemented by the processing circuit. Detailed Implementation

[0023] The laser processing apparatus according to the embodiments will now be described in detail based on the accompanying drawings.

[0024] Implementation method 1.

[0025] Figure 1 This diagram illustrates the structure of the laser processing apparatus 50 according to Embodiment 1. The laser processing apparatus 50 includes a laser oscillator 1, a processing head 2, a drive unit 5, and a control unit 3. The laser processing apparatus 50 includes a processing state resolver 51 comprising the control unit 3. The laser oscillator 1 oscillates and emits a laser beam L. The wavelength of the laser beam L is selected considering the absorption rate and reflectivity of the laser beam L towards the workpiece. For example, the wavelength of the laser beam L is any wavelength from 0.193 μm to 11 μm. The workpiece is a workpiece W. The laser beam L emitted from the laser oscillator 1 is supplied to the processing head 2 via an optical path.

[0026] The processing head 2 includes: a focusing optical system that focuses the laser emitted from the laser oscillator 1 to irradiate the workpiece W; and a processing gas supply unit that supplies processing gas toward the workpiece W. Figure 1The machining gas supply unit is not shown. When machining gas is supplied to the interior of the machining head 2 and the laser beam L irradiates the workpiece W, the machining gas is injected into the workpiece W through the machining gas supply unit. The machining head 2 has a collimating lens 4 and a condensing lens 7, which consist of multiple lens groups. The collimating lens 4 and the condensing lens 7 are examples of a focusing optical system. The laser beam L emitted from the laser oscillator 1 is collimated by the collimating lens 4 and then focused by the condensing lens 7. The focused laser beam L irradiates the workpiece W. The machining head 2 focuses the laser beam L and irradiates the workpiece W, thereby cutting the workpiece W.

[0027] The machining head 2 has a nozzle (not shown). The nozzle opening is located in the optical path of the laser beam L between the condenser lens 7 and the workpiece W, through which the laser beam L and the machining gas pass. Generally, a motor (not shown) and a motor drive unit are provided on the shaft that sets the machining head 2, or on a machining table that positions the workpiece W.

[0028] The drive unit 5 changes the relative position of the processing head 2 and the workpiece W. The control unit 3 controls the laser oscillator 1, the processing head 2, and the drive unit 5 based on numerical parameters related to laser processing, i.e., processing parameters, to perform processing. Specifically, the control unit 3 controls the motor drive unit, which in turn controls the motor. The drive unit 5 operates in tandem with the motor, changing the relative position of the processing head 2 and the workpiece W. The processing head 2 has a focusing lens position changing drive unit 6 that changes the positional relationship between the focal point of the focusing optical system of the laser beam L and the workpiece W.

[0029] The type of laser oscillator 1 is not limited. An example of laser oscillator 1 is a fiber laser oscillator. Laser oscillator 1 can also be a direct diode laser, a carbon dioxide laser, a copper vapor laser, various ion lasers, or a solid-state laser. An example of a solid-state laser is a laser using YAG (Yttrium Aluminum Garnet) crystal as the excitation medium. The laser processing apparatus 50 may include a wavelength conversion unit that converts the wavelength of the laser light generated by the laser oscillator 1.

[0030] The control unit 3 controls the laser oscillator 1, the motor drive unit, and the focusing lens position change drive unit 6 according to the processing program and processing parameters indicating processing conditions, so that the laser beam L scans the processing path on the workpiece W. Examples of processing parameters related to the control performed by the control unit 3 include laser output, processing gas pressure, processing speed, focal position of the focusing optical system, focusing diameter of the focusing optical system, laser pulse frequency, laser pulse duty cycle, magnification of the focusing optical system, nozzle diameter, distance between the workpiece W and the nozzle, type of laser beam mode, and positional relationship between the center of the nozzle orifice and the laser beam L. The processing parameters are not limited to the examples described above. The processing parameters can also be determined based on one or both of the type of laser used and the functions of the laser oscillator 1.

[0031] The processing parameters used by the control unit 3 can be changed in correspondence with the correction amount calculated by the processing state resolver 51, as described later. That is, the processing parameters can be corrected by the processing state resolver 51. The processing parameters before correction by the processing state resolver 51 are predetermined, for example, in correspondence with the processing content. The laser processing apparatus 50 may have an input unit that accepts input from the operator, and the processing parameters before correction by the processing state resolver 51 can be changed through input from the operator. The processing parameters before correction by the processing state resolver 51 can be sent to the laser processing apparatus 50 by a device not shown in the figure. An example of such a device is a computer.

[0032] The laser beam L emitted from the laser oscillator 1 is collimated by the collimating lens 4 and focused by the condenser lens 7. The focused laser beam L irradiates the workpiece W. Within the workpiece W, irradiation by the laser beam L causes phenomena such as evaporation and melting, emitting a processing light 8. The emitted processing light 8 enters the interior of the processing head 2.

[0033] The laser processing apparatus 50 also includes a reflector 9. The processing light 8 passes through the reflector 9 via a focusing lens 7. The reflector 9 has the property of allowing light of wavelengths other than the wavelength of the laser beam L to pass through. The processing light 8 passing through the reflector 9 is converted into a time-series signal by a processing state observation unit 52. The processing state observation unit 52 is included in the processing state resolver 51. The processing state observation unit 52 detects the light intensity of multiple predetermined wavelength bands of interest emitted from the workpiece W, i.e., the processing light 8, by laser irradiation, and uses these as multiple light sensor signals.

[0034] The processing status analyzer 51 further includes: a feature extraction unit 53, which extracts feature quantities that are correlation indicators between multiple photosensitive signals; an evaluation unit 54, which determines whether the processing is qualified or not based on the feature quantities for at least one of multiple processing defects, and obtains a judgment result; and a correction calculation unit 55, which determines the processing parameters used to perform correction as correction parameters based on the feature quantities, and determines the correction amount of the correction parameters. Specifically, the correction calculation unit 55 determines the correction parameters that should be corrected and the correction amount of the correction parameters based on the above judgment result. The multiple processing defects include at least one of the following: roughness of the cut surface quality, scraping, slag, and oxide film peeling. Because the multiple processing defects include at least one of the following: roughness of the cut surface quality, scraping, slag, and oxide film peeling, the laser processing apparatus 50 can perform significant correction of the processing parameters. The feature extraction unit 53 extracts at least one of the correlation indicators between multiple photosensitive signals and the feature quantities that can be obtained from one photosensitive signal.

[0035] Furthermore, the correction calculation unit 55 can determine the processing parameters to be corrected and the correction amount of the processing parameters based on at least one of the cutting speed, focal position, focusing diameter, gas pressure, and laser output. When the processing parameters are at least one of the aforementioned cutting speed, focal position, focusing diameter, gas pressure, and laser output, and the processing state is defined as unfavorable, the laser processing apparatus 50 can restore the processing state from the unfavorable state to the favorable state earlier.

[0036] The time-series signal obtained from the processing status observation unit 52 is transformed into feature quantities by the feature quantity extraction unit 53. The evaluation unit 54 determines the processing status, such as whether the processing result is qualified, the degree of processing defects, the deviation from the good processing result, and signs of processing defects. Based on the determination result obtained by the evaluation unit 54, the correction quantity calculation unit 55 sends a command to the control unit 3 to change the processing parameters. Processing is performed continuously by changing the processing parameters when processing is actually carried out according to the command. The evaluation unit 54 may be included in the correction quantity calculation unit 55.

[0037] Next, the operation of Implementation Method 1 will be explained. Figure 2 This is a flowchart illustrating an example of the sequence of operations related to the adjustment of processing parameters performed by the laser processing apparatus 50 according to Embodiment 1. First, the laser processing apparatus 50 performs a cutting process (S1). Next, the processing status observation unit 52 acquires a processing light signal from the processing light 8 emitted by the laser processing (S2). The feature extraction unit 53 extracts feature quantities from the time-series signal obtained by the processing status observation unit 52 (S3).

[0038] The evaluation unit 54 determines the pass / fail status of the processing result based on the extracted feature quantities (S4). If the evaluation unit 54 determines the processing result to be good (S4 is Yes), the operation of the laser processing apparatus 50 jumps to step S2, and processing continues without changing the processing parameters. If the evaluation unit 54 determines the processing result to be bad (S4 is No), the correction amount calculation unit 55 determines the processing parameters that should be changed and calculates the correction amount for the processing parameters that should be changed (S5). The correction amount calculation unit 55 outputs the calculated correction amount to the control unit 3. The laser processing apparatus 50 performs processing based on the correction amount. Figure 2 The timing of the actions shown is not limited to the middle of the production process.

[0039] The details of the processing status observation section 52 are explained. Figure 3 This diagram illustrates the structure of the processing status observation section 52 included in the laser processing apparatus 50 according to Embodiment 1. Figure 3 Feature extraction unit 53 is also shown. Processing status observation unit 52 includes a beam splitter 10, multiple wavelength filters 11, multiple imaging lenses 12, and multiple light sensors 13.

[0040] Figure 1 The processing light 8 that passes through the reflector 9 is split by the beam splitter 10. Multiple wavelength filters 11 each allow the processing light 8 of its corresponding wavelength band to pass through. The processing light 8 transmitted by each of the multiple wavelength filters 11 is received by the corresponding optical sensor 13 among the multiple optical sensors 13 through the imaging lens 12. Each of the multiple optical sensors 13 outputs the light intensity of the processing light 8 as a time-series signal. The signals output from the multiple optical sensors 13 are sent to the feature extraction unit 53.

[0041] The characteristics of the processing light 8 are explained. The processing light 8 is mainly generated by the thermal radiation of the workpiece W. The light generated by thermal radiation is light with a peak wavelength that depends on the temperature of the molten metal, and the wavelength distribution of this light is determined only by temperature. If the temperature is high, the peak wavelength shifts towards the shorter wavelength side. The amount of processing light 8 varies depending on the processing state formed by processing the workpiece W, such as the cutting width shape and the cutting front shape. The amount of processing light 8 injected into the interior of the processing head 2 also varies depending on the shape of the nozzle used. For example, when the processing speed of thin plates is high, the inclination angle of the cutting front shape is large, the area of ​​the laser beam L hitting the cutting front is large, so the temperature of the molten metal is high, and the amount of processing light 8 returning into the interior of the processing head 2 is large.

[0042] To observe the detailed processing status, the processing status observation unit 52 divides the processing light 8. The processing status observation unit 52 has multiple wavelength filters 11. Each of the multiple wavelength filters 11 allows light of a wavelength different from the wavelength of light transmitted by the other wavelength filters 11 to pass through. The processing light 8 transmitted by each of the multiple wavelength filters 11 is directed to any one of the multiple light sensors 13.

[0043] Imagine multiple optical sensors 13, namely, a first optical sensor 13a, a second optical sensor 13b, and a third optical sensor 13c. Figure 4 This diagram illustrates an example of the wavelength bands of light received by the first optical sensor 13a, the second optical sensor 13b, and the third optical sensor 13c of the laser processing apparatus 50 according to Embodiment 1. For example, the first optical sensor 13a receives processing light 8 in the short wavelength band, the third optical sensor 13c receives processing light 8 in the long wavelength band, and the second optical sensor 13b receives processing light 8 in the middle wavelength band of the light received by the first optical sensor 13a and the third optical sensor 13c.

[0044] The first optical sensor 13a, the second optical sensor 13b, and the third optical sensor 13c do not need to receive the processing light 8 across the entire wavelength range; they can receive processing light 8 within a certain wavelength range. The processing status observation unit 52 can observe the shift in wavelength distribution based on the intensity ratio of each wavelength band received by the first optical sensor 13a, the second optical sensor 13b, and the third optical sensor 13c, and the ratio of the intensity of the light received by each of the first optical sensor 13a, the second optical sensor 13b, and the third optical sensor 13c to the total intensity. The total intensity is the intensity of all the light received by the first optical sensor 13a, the second optical sensor 13b, and the third optical sensor 13c.

[0045] The processing status observation unit 52 can observe the changes in the time-series signal and the amount of processing light 8 as the sum of the light intensities received by the first light sensor 13a, the second light sensor 13b, and the third light sensor 13c. A light sensor 13 may be provided that blocks only light of the wavelength of the laser beam L, while receiving processing light 8 of wavelengths other than that wavelength.

[0046] The processing status observation unit 52 can change the wavelength of the light incident on the light sensor 13 by combining the beam splitter 10 and the wavelength filter 11. The processing status observation unit 52, as shown... Figure 5 As shown, it can be replaced with a processing status observation unit 52A having a diffraction grating 10a. Figure 5This is a diagram showing the structure of the laser processing apparatus 50A according to the first modification of Embodiment 1. The laser processing apparatus 50A has a processing status observation section 52A that uses a diffraction grating 10a for beam splitting. The processing status observation section 52 is as follows: Figure 6 As shown, it can be replaced with a machining status observation section 52B having a prism 10b. Figure 6 This is a diagram showing the structure of the laser processing apparatus 50B according to the second modification of Embodiment 1. The laser processing apparatus 50B has a processing status observation section 52B that uses a prism 10b to perform beam splitting.

[0047] The light sensor 13 included in the processing status observation unit 52 can be a Si (Silicon) photodiode sensitive to light with wavelengths from 400 nm to 1100 nm, or an InGaAs (Indium Gallium Arsenide) photodiode sensitive to light with wavelengths longer than the near-infrared range. One of the plurality of wavelength filters 11 can be a short-pass filter that allows light with wavelengths shorter than the first wavelength to pass through, another of the plurality of wavelength filters 11 can be a long-pass filter that allows light with wavelengths longer than the first wavelength and longer than the second wavelength to pass through, and yet another of the plurality of wavelength filters 11 can be a band-pass filter that allows light with wavelengths longer than the first wavelength and shorter than the second wavelength to pass through.

[0048] To obtain processing light 8 in a more suitable wavelength band, the wavelength filter 11 can be a bandpass filter obtained by combining a short-pass filter and a long-pass filter. For example, a short-pass filter that allows light with wavelengths less than 500 nm to pass through, a bandpass filter that allows light with wavelengths greater than 500 nm and less than 700 nm to pass through, and a high-pass filter that allows light with wavelengths greater than 700 nm to pass through can be combined.

[0049] One of the multiple wavelength filters 11 may be a first wavelength filter that allows light with a wavelength shorter than 525 nm to pass through, another of the multiple wavelength filters 11 may be a second wavelength filter that allows light with a wavelength longer than 700 nm to pass through, and yet another of the multiple wavelength filters 11 may be a third wavelength filter that allows light with a wavelength greater than 530 nm and less than 700 nm to pass through.

[0050] One of the multiple wavelength filters 11 may be a wavelength filter that allows light with wavelengths above 475 nm and below 525 nm to pass through, another of the multiple wavelength filters 11 may be a wavelength filter that allows light with wavelengths above 575 nm and below 625 nm to pass through, and yet another of the multiple wavelength filters 11 may be a wavelength filter that allows light with wavelengths above 675 nm and below 725 nm to pass through.

[0051] One of the multiple wavelength filters 11 can be a wavelength filter that allows light with wavelengths above 400 nm and below 800 nm to pass through; another of the multiple wavelength filters 11 can be a wavelength filter that allows light with wavelengths above 475 nm and below 525 nm to pass through; and yet another of the multiple wavelength filters 11 can be a wavelength filter that allows light with wavelengths above 675 nm and below 725 nm to pass through. The processing status observation unit 52 has the aforementioned multiple wavelength filters 11, thereby enabling the processing status analyzer 51 to more effectively correct processing parameters and to perform detailed detection of processing defects.

[0052] One of the multiple optical sensors 13 can be positioned either in the direction in which the laser beam L emitted from the laser oscillator 1 irradiates the processing point, or in a direction different from the direction in which the laser beam L irradiates the processing point. By configuring the optical sensor 13 in two positions, the changes in the intensity ratio and wavelength distribution of the processing light 8 caused by the positional difference can be compared. By comparing the intensity ratio caused by the positional difference, the tilt of the incident light toward the processing head 2 can be determined. That is, by configuring the optical sensor 13 in two positions, the laser processing apparatus 50 can correct the processing parameters with higher precision.

[0053] Figure 7 This diagram illustrates the structure of the laser processing apparatus 50C according to the third modification of Embodiment 1. The laser processing apparatus 50C includes all the structural elements of the laser processing apparatus 50, a collimating lens 14 connected to the processing head 2, and an optical fiber 15 connecting the collimating lens 14 to the processing status observation unit 52. Figure 7 As shown, the processing light 8 can be transmitted from the processing head 2 to the processing status observation unit 52 via the optical fiber 15. Figure 7 The block representing the processing status resolver 51 is not shown, but the processing status observation unit 52 is included within the processing status resolver 51. In the laser processing apparatus 50C, the processing status resolver 51, which includes the processing status observation unit 52, is disposed outside the processing head 2, thus achieving the effect of making the processing head 2 smaller and lighter. The laser processing apparatus 50C determines the amount of correction for processing parameters during processing, or judges the pass or fail of processing related to processing defects that should be corrected, based on the processing light 8 transmitted from the processing head 2 via the optical fiber 15.

[0054] When the laser oscillator 1 is a fiber laser or a laser oscillator capable of fiber transmission, the processing light 8 returning in the fiber can be used for resolution, and thus the processing state resolver 51 can be configured inside the laser oscillator 1.

[0055] The feature extraction unit 53 transforms the time-series signal output from the processing state observation unit 52 into feature quantities. Various methods exist for creating feature quantities. The feature extraction unit 53 can perform calculations such as average value calculation, calculation of statistical quantities like standard deviation, frequency analysis, filter analysis, or wavelet transform on the time-series signal obtained from the processing state observation unit 52, and set the combination of values ​​obtained by analyzing the time-series signal as feature quantities.

[0056] The above-described method for creating feature quantities is an example. The feature quantity extraction unit 53 can create feature quantities using general time-series signal analysis methods. The number of feature quantities output by the feature quantity extraction unit 53 can be one or more. The feature quantity extraction unit 53 stores the feature quantities at the start of processing and their positions within the feature quantity space; these feature quantities and changes in their positions can also be considered feature quantities. Therefore, the laser processing apparatus 50 can also determine changes in feature quantities from the initial processing state and detect signs of processing defects.

[0057] The feature extracted by the feature extraction unit 53 can be a feature that reflects the output value of each of the multiple optical sensors 13, or it can be a feature that combines the output values.

[0058] The evaluation unit 54 determines whether the current processing is qualified or not based on the feature quantities extracted by the feature extraction unit 53. The evaluation unit 54 may output only the qualification or qualification result of the processing, or it may output the evaluation value of the processing. Alternatively, the evaluation unit 54 may not perform a binary qualification or qualification determination, but instead calculate a value close to 0 if the probability of a good result is high, and close to 1 if the probability of a bad result is high. This value can be any number in a continuous sequence. For example, the evaluation unit 54 may calculate an evaluation value where the probability of a good result is 90% and the probability of a bad result is 10%.

[0059] If the evaluation unit 54 determines that the processing is not satisfactory, it can output the presence or absence of items detailing the symptoms of processing defects. Examples of these items include the adhesion of molten metal to the cut surface during laser cutting, the formation of slag at the lower end of the cut surface, or the periodic roughness generated at the upper part of the cut surface. If roughness occurs, the depth of the streak's recess is deeper compared to the case where no roughness occurs. The evaluation unit 54 can also detect the presence or absence of oxide film peeling occurring on the cut surface. Oxide film peeling occurs when the processing gas used in cutting is oxygen.

[0060] The defects in processing are not limited to the examples mentioned above. For example, the evaluation unit 54 can also determine other defects such as discoloration of the workpiece W or the presence or absence of a vibrating surface. The evaluation unit 54 can, for example, change the processing parameters to be determined in relation to the laser output, processing speed, processing plate thickness, or type of processing gas.

[0061] For example, when the processing gas is oxygen, an oxide film is formed on the cut surface, so it is necessary to determine whether oxide film has been removed. However, when the processing gas is nitrogen, no oxide film is formed on the cut surface, so it is not necessary to determine whether oxide film has been removed. Therefore, the evaluation unit 54 may not need to determine oxide film removal when the processing gas is nitrogen.

[0062] The evaluation unit 54 can comprehensively consider the presence or absence of each defective item and output the result of whether the processing is qualified or not. The evaluation unit 54 can also analyze the symptoms of the defective items if the processing result is determined to be unqualified only based on the presence or absence of defects.

[0063] The evaluation unit 54 can display the judgment result on a display unit inside or outside the laser processing apparatus 50. The evaluation unit 54 can only display the judgment result on a display unit inside or outside the laser processing apparatus 50 if the judgment result of whether the cutting process is qualified or not is not. The display unit is not shown.

[0064] The evaluation unit 54 can use not only the feature quantities output from the feature extraction unit 53, but also other information to determine whether a work is qualified or not. Examples of other information include processing parameters related to the processing being performed, the temperature of the optical system contained inside the processing head 2, the temperature variation of the optical system contained inside the processing head 2, the thickness of the processing plate, and some or all of the information about the processing material. The processing plate thickness is the thickness of the workpiece W in the laser incident direction, and the processing material is the material of the workpiece W.

[0065] If the judgment result output from the evaluation unit 54 is unfavorable, the correction amount calculation unit 55 calculates the correction amount for the processing parameters based on the judgment result output from the evaluation unit 54. The correction amount calculation unit 55 outputs the calculated correction amount to the control unit 3. The correction amount calculation unit 55 can obtain the processing parameters set in the control unit 3 and can calculate the correction amount based on the judgment result output from the evaluation unit 54 and the currently set processing parameters.

[0066] The control unit 3 corrects the processing parameters based on the correction amount received from the correction amount calculation unit 55, thereby performing processing. As described above, if the determination result obtained by the evaluation unit 54 is negative, the laser processing apparatus 50 performs processing under the condition of correcting the processing parameters. The correction of the processing parameters is repeated until the determination result output from the evaluation unit 54 becomes good.

[0067] Next, the calculation of the correction values ​​for the processing parameters will be explained in detail. Examples of processing parameters that should be corrected include laser output, processing gas pressure, processing speed, focal position of the focusing optical system, focusing diameter of the focusing optical system, laser pulse frequency, laser pulse duty cycle, magnification of the focusing optical system, nozzle diameter, distance between the workpiece W and the nozzle, type of laser beam L mode, and positional relationship between the center of the nozzle orifice and the laser beam L.

[0068] When the correction calculation unit 55 outputs the judgment results of each defective item as evaluation value from the evaluation unit 54, it can determine the processing parameter to be corrected and the correction amount of the processing parameter based on the combination mode of the pass / fail judgment results related to each defective item. For example, the combination mode is set as good when the evaluation value is 1 and as bad when the evaluation value is 0. When the evaluation unit 54 outputs the evaluation values ​​corresponding to roughness judgment, oxide film peeling judgment and slag judgment respectively, it is, for example, a combination of three values ​​such as 0, 0 and 1.

[0069] For example, when the value corresponding only to the slag determination is 1 and all other values ​​are 0, the correction calculation unit 55 uses the laser output and processing gas pressure among the processing parameters as the objects of the correction calculation, and determines the correction amount by increasing the laser output and decreasing the processing gas pressure. As described above, it is possible to determine the processing parameters to be corrected for each combination mode and the correction amount of those processing parameters.

[0070] When the evaluation unit 54 outputs the judgment result for each defective item as an evaluation value, and the pass / fail judgment result of each defective item is output as a value indicating the degree of defect, the correction calculation unit 55 can assign a weight to the correction amount of the processing parameter that should be corrected for each defective item and change the correction amount, or it can change the processing parameter itself to be corrected in correspondence with the evaluation value of each defective item.

[0071] For example, the evaluation unit 54 assumes that for each defective processing item, it outputs an evaluation value from 0 to 1, corresponding to any value from three or more stages. For example, the evaluation value for slag determination is defined through four stages: 0, 0.3, 0.6, and 1.0. Corresponding to the evaluation value for slag determination, the correction amounts for laser output and processing gas pressure are determined. In a specific example, when the evaluation value for slag is 0.3, the correction amount for laser output is set to +0.2 [kW], and the correction amount for processing gas pressure is set to -0.01 [MPa]. When the evaluation value is 0.6, the correction amount is set to +0.5 [kW], and the correction amount for processing gas pressure is set to -0.02 [MPa].

[0072] The correction calculation unit 55 calculates the correction amount according to the correspondence between the evaluation value and the correction amount determined as described above. Thus, when the evaluation value of the molten slag is 0.3, the laser processing apparatus 50 increases the laser output by 0.2 [kW] and decreases the processing gas pressure by 0.01 [MPa]; when the evaluation value is 0.6, the laser processing apparatus 50 increases the laser output by 0.5 [kW] and decreases the processing gas pressure by 0.02 [MPa]. The above correction amount is just one example; the correction amount only needs to be determined in correspondence with the evaluation value. The correction amount can be determined as a value dependent on the value of the processing parameters before correction. The method for determining the correction amount is not limited to the example described above.

[0073] When any number from a series of consecutive evaluation values ​​is output from the evaluation unit 54 for each defective item, the correction calculation unit 55 can use a table showing the correspondence between the evaluation value and the correction amount to calculate the correction amount for each processing parameter by extrapolation or interpolation. The extrapolation method can be a polynomial curve method, or a trigonometric function or conic section method.

[0074] In the examples above, poor processing quality is cited as an example of a defective processing item. Sometimes, depending on the operator, the priorities for improvement items such as processing quality, productivity, or processing stability differ. Even with good processing quality, it may not be suitable when processing speed is extremely slow. Therefore, the processing status resolver 51 may have an input unit that receives input from the operator regarding the priority of each improvement item.

[0075] The correction calculation unit 55 can calculate the correction amount for the processing parameters based on the priority of each improvement item. The correction calculation unit 55 can determine the correction amount for the processing parameters based on the priority of multiple improvement items, including productivity, combination mode, and processing stability. For example, consider making the positive and negative signs of the correction amounts for the same processing parameter reversed through improvement items. In the case described above, the correction calculation unit 55 calculates the correction amount corresponding to the priority work item.

[0076] The correction calculation unit 55 can calculate the correction amount by performing a weighted calculation corresponding to the priority. For example, for each improvement item, a weight related to the correction amount of each processing parameter can be predetermined. The correction calculation unit 55 multiplies the weight corresponding to the priority of the improvement item by the correction amount, calculates the total correction amount after multiplying by the weight, and thus determines the output correction amount. If the weight is determined in such a way that the higher the priority of the item, the greater the weight value, then the higher the priority, the greater the contribution to the output correction amount. As described above, the correction calculation unit 55 can calculate the correction amount by performing a weighted calculation corresponding to the priority.

[0077] When the operator wishes to detect signs of processing defects, the laser processing apparatus 50 can determine these signs based on the values ​​output by the evaluation unit 54. For example, if the evaluation value output by the evaluation unit 54 is any value between 0 and 1, a value between 0 and 0.4 can be defined as good processing, a value between 0.4 and 0.7 as a sign of processing defects, and a value above 0.7 as processing defects. The correction calculation unit 55 can correct the processing parameters when the evaluation value is 0.4 or higher.

[0078] The processing state resolver 51 can determine the correction amount based on past test results. In this case, the processing state resolver 51 needs to store more than one set of processing parameters and evaluation values ​​related to past tests. Figure 8 This diagram illustrates the structure of the processing state resolver 56 included in the laser processing apparatus according to the fourth modification of Embodiment 1. The laser processing apparatus according to the fourth modification possesses all the structural elements of the laser processing apparatus 50, and further includes a processing condition storage unit 57. An example of the processing condition storage unit 57 is a semiconductor memory. The processing condition storage unit 57 is included in the processing state resolver 56. The processing state resolver 56 determines the correction amount based on the results of multiple tests. The processing state resolver 56 also includes a control unit 3, a processing state observation unit 52, a feature extraction unit 53, an evaluation unit 54, and a correction amount calculation unit 55.

[0079] In the processing state analyzer 56, one or more sets consisting of evaluation results output from the evaluation unit 54 in previous or past tests and processing parameters corresponding to those evaluation results are stored in the processing condition storage unit 57. The correction amount calculation unit 55 calculates the correction amount for the processing parameters based on the evaluation results output from the evaluation unit 54 and the past evaluation results and processing parameters stored in the processing condition storage unit 57.

[0080] As described above, the correction calculation unit 55 uses not only current information but also past information to calculate the correction amount, thereby improving the accuracy of the correction amount calculation. For example, the correction calculation unit 55 can use the evaluation results of multiple quantities and the processing parameters as discrete states in a Markov chain to calculate the correction amount. In the actual adjustment of processing conditions, a combination of multiple correction conditions is considered.

[0081] The correction calculation unit 55 selects a combination and determines the correction amount by considering how the defect pattern changes in the next trial processing, thereby enabling the calculation of a more accurate correction amount. For example, the correction calculation unit 55 calculates the correction amount by lowering the focal position, which is a processing parameter, and the laser processing apparatus 50 performs the cutting process based on the calculated correction amount.

[0082] For example, the processing condition storage unit 57 stores the processing parameters set during processing and the evaluation results corresponding to the results of the cutting process. If the evaluation result output from the evaluation unit 54 is unsatisfactory, the laser processing apparatus involved in the fourth modification lowers the focal position and performs a laser processing test. If the correction amount calculation unit 55 determines that one of the processing defects, namely slag, has not been improved through two tests, it can calculate a correction amount based on the set of processing parameters and evaluation values ​​stored in the processing condition storage unit 57, starting from the point where the focal position was lowered in the two tests, by increasing the focal position.

[0083] The processing status analyzer 51 has an input unit that can receive a threshold input from the operator. This threshold is used to determine the stages used in the evaluation value corresponding to each defective item, or in the evaluation value composed of two judgment results: pass or fail. The evaluation unit 54 uses the input threshold to determine the evaluation value. When the processing status analyzer 51 has an input unit, the laser processing apparatus 50 sets the evaluation stage related to each defective item to be fine or coarse for each operator, corresponding to the threshold input by the operator. When the processing status analyzer 51 has an input unit, the operator can set the benchmark for the evaluation value to be strict or lenient.

[0084] In laser processing, even when the process starts well, defects can sometimes occur due to changes in the condition of the processing head 2 or minor variations in the material of the workpiece W. Therefore, operators have traditionally performed continuous processing at a slower speed than actually capable of processing. In other words, operators reduce their productivity compared to their original capabilities.

[0085] To address the aforementioned issues, this invention divides the wavelength of the processing light 8 emitted during processing into multiple bands and detects the processing light 8. During processing, characteristic quantities such as changes in wavelength distribution are detected in detail, thereby enabling the detection of processing results, processing defects, or signs of processing defects. If a processing defect occurs and a sign of a processing defect is detected, the correction calculation unit 55 changes the processing parameters. This prevents processing defects from occurring and maintains a constant productivity during processing. Even if a processing defect occurs, it can be automatically restored to a good processing state.

[0086] Instead of correcting processing parameters after detecting defects, defective items, or signs of defects, the correction calculation unit 55 directly receives feature quantities extracted by the feature quantity extraction unit 53 and performs the correction. Therefore, while defects and signs of defects are not detected, the workload related to the calculations of the correction calculation unit 55 is reduced because processing parameter correction is performed without detection processing. The feature quantities used for defect detection and processing parameter correction can be the same or different.

[0087] Furthermore, the evaluation unit 54 can determine, based on characteristic quantities, the boundary value between the range of processing parameters for which the judgment result becomes good (i.e., the good processing range) and the range of processing parameters for which the judgment result becomes bad (i.e., the bad processing range) for at least one of multiple processing defects. The correction amount calculation unit 55 can determine the difference between the corrected processing parameters based on the correction amount and the boundary value (i.e., the deviation) if the deviation exceeds the boundary value. If so, the correction amount is determined and corrected during processing. Thus, the laser processing apparatus 50 can detect signs of processing defects with relatively high accuracy.

[0088] As described above, the laser processing apparatus 50 according to Embodiment 1 includes: a processing state observation unit 52, which detects the light intensity of a predetermined plurality of wavelengths of interest of the processing light 8 emitted from the workpiece W by laser irradiation and uses them as plurality of photosensitive sensor signals; a feature quantity extraction unit 53, which extracts feature quantities that serve as correlation indicators between the plurality of photosensitive sensor signals; and a correction quantity calculation unit 55, which determines the processing parameters used for correction as correction parameters based on the feature quantities, and determines the correction quantity of the correction parameters. Because the laser processing apparatus 50 uses the aforementioned feature quantities, it can obtain more information and detect the processing state at high speed or with high precision, and adjust the processing conditions, compared to observing the light in multiple wavelengths individually. The feature quantity extraction unit 53 extracts at least one of the correlation indicators between the plurality of photosensitive sensor signals and at least one of the feature quantities that can be obtained from a single photosensitive sensor signal.

[0089] In Embodiment 1, the evaluation unit 54 determines whether the processing is qualified or not based on at least one of the multiple processing defects, and obtains a determination result. For example, the correction calculation unit 55 determines the correction parameters to be corrected and the correction amount of the correction parameters based on the above determination result. In this case, the laser processing apparatus 50 according to Embodiment 1 can change the processing conditions with high precision and high speed, and as a result, stable continuous processing can be performed.

[0090] Implementation method 2.

[0091] In the laser processing apparatus according to Embodiment 2, instead of the processing state resolver 51 included in the laser processing apparatus 50 according to Embodiment 1, there is instead a... Figure 9 The processing status parser 58 is shown. Figure 9 This diagram illustrates the structure of the processing state resolver 58 included in the laser processing apparatus according to Embodiment 2. The laser processing apparatus according to Embodiment 2 differs from laser processing apparatus 50 in that it possesses the processing state resolver 58, which is not present in laser processing apparatus 50. In Embodiment 2, structural elements having the same functions as in Embodiment 1 are labeled with the same reference numerals as in Embodiment 1, and repeated descriptions are omitted. In Embodiment 2, the differences from Embodiment 1 will be primarily explained.

[0092] The processing status analyzer 58 includes: a processing status observation unit 52; a feature extraction unit 53; a machine learning unit 59, which learns the relationship between feature quantities and evaluation values ​​of processing defects related to processing parameters that should be corrected; an evaluation unit 54; and a correction amount calculation unit 55. The machine learning unit 59 learns by associating the feature quantities extracted by the feature extraction unit 53 with evaluation values ​​created by the operator. The evaluation values ​​created by the operator are the operator's evaluation values. The operator's evaluation values ​​can be input, for example, from an input unit not shown, or received by the machine learning unit 59 after being output from another device. The machine learning unit 59 can perform computational processing based on the feature quantities, thereby outputting the correction amount of the processing parameters.

[0093] The machine learning unit 59 includes a learning unit 60 and a data acquisition unit 61. The learning unit 60 learns from the input and result datasets using machine learning. The machine learning algorithm used by the learning unit 60 can be any algorithm. For example, the machine learning algorithm used by the learning unit 60 can be an algorithm learned by a teacher. The data acquisition unit 61, as input to the learning unit 60, acquires features from the feature extraction unit 53 and outputs the acquired features to the learning unit 60. The evaluation unit 54 may include both the feature extraction unit 53 and the learning unit 60.

[0094] The learning unit 60 also inputs the operator's evaluation value. The operator's evaluation value is the result of judging whether the processing result is satisfactory or not based on each defective item. Similar to the evaluation value obtained by the evaluation unit 54 in Embodiment 1, it can represent any value among multiple stages, or any value among consecutive numbers. That is, the operator's evaluation value is an evaluation value corresponding to the combination mode of Embodiment 1, determined by the operator. The data acquisition unit 61 can acquire the time-series signal of light intensity output from the light sensor 13 and processed by the feature extraction unit 53, and use it as input to the learning unit 60.

[0095] As described above, the data acquisition unit 61 acquires time-series data of light intensity or feature quantities output from the feature extraction unit 53 as state variables, and assigns the acquired state variables to the learning unit 60. The learning unit 60 uses the dataset consisting of the state variables and evaluation values ​​to perform machine learning on the pass / fail status of the processing results. The dataset is data that associates the state variables and evaluation data.

[0096] The learning unit 60 uses a trained model obtained through machine learning to output evaluation values ​​corresponding to the feature quantities, thereby enabling the correction calculation unit 55 to correct the processing parameters with higher accuracy. The learning unit 60 has both the function of machine learning to determine the pass / fail status of the processing result and the function of serving as a trained model; however, the inference unit that uses the trained model to output the evaluation values ​​can also be set up independently of the learning unit 60. That is, the processing state resolver 58 can have an inference unit that uses the trained model learned by the learning unit 60 to calculate the combination pattern of information from the time-series data of light intensity.

[0097] exist Figure 9 In this example, the machine learning unit 59 is located inside the processing state resolver 58, but the machine learning unit 59 can also be located outside the processing state resolver 58. In this case, for example, the processing state resolver 58 and the machine learning unit 59 are connected via a network. The machine learning unit 59 can reside on a cloud server.

[0098] The processing state parser 58 includes the evaluation unit 54 described in Embodiment 1, and has the function of learning using the determination result determined by the evaluation unit 54. For example, the processing state parser 58 can use the dataset described above. After learning to a certain extent, the processing state parser 58 corrects the determination result obtained by the evaluation unit 54, and the learning unit 60 learns the corrected determination result.

[0099] Learning Department 60, for example, uses a neural network model to learn from time-series data of light intensity and the evaluation results of the processing outcomes through so-called teacher-led learning. Teacher-led learning is machine learning that learns features from multiple datasets (i.e., multiple datasets) about a given input and result, and infers the results based on the input. The results in the dataset are labels.

[0100] A neural network consists of an input layer composed of multiple neurons, an intermediate layer composed of multiple neurons (also known as a hidden layer), and an output layer composed of multiple neurons. There can be one or more intermediate layers.

[0101] Figure 10 This is a diagram illustrating the structure of the neural network model according to Embodiment 2. X1, X2, and X3 are neurons in the input layer, Y1 and Y2 are neurons in the intermediate layer, and Z1, Z2, and Z3 are neurons in the output layer. Figure 10 In the 3-layer neural network model shown, if the three input values ​​are each input to any one of X1, X2, and X3, then each input value is multiplied by any one of the corresponding weights w11 to w16 and input to the neuron in the middle layer, namely Y1 or Y2.

[0102] The output values ​​from Y1 and Y2 are multiplied by any of the corresponding weights from w21 to w26, and then input to the neurons in the output layer, namely Z1, Z2, or Z3. The output layer adds the input values ​​and outputs the sum as the result. For example, the outputs from Z1, Z2, and Z3 can be correlated with the evaluation results for each poorly processed item. The output results change according to the values ​​of weights from w11 to w16 and weights from w21 to w26.

[0103] In Implementation 2, using the dataset described above, the values ​​of weights w11 to w16 and weights w21 to w26 are adjusted and learned in such a way that the output of the neural network is close to the evaluation result of whether the processing of the correct solution is qualified. Figure 10 As an example, the number of layers in a neural network model and the number of neurons in each layer are not limited to a certain number. Figure 10 Examples.

[0104] The Learning Department 60 uses a neural network model to learn how to assess whether a process is satisfactory or not through so-called teacherless learning. Teacherless learning involves learning how to distribute input data based solely on a large amount of input data, without using corresponding teacher output data. This learning can be applied to the input data using methods such as compression, classification, or shaping. For example, in teacherless learning, datasets with similar features can be clustered together. Furthermore, by setting a benchmark to optimize the clustering results, assessment outcomes can be assigned based on the clustering, thereby enabling prediction of assessment results.

[0105] As an intermediate problem setting between unsupervised learning and supervised learning, it is sometimes referred to as semi-supervised learning. In semi-supervised learning, only a portion of the data sets containing input and output exist; for the remainder, only the input data exists. Learning unit 60 can perform machine learning through semi-supervised learning.

[0106] The machine learning unit 59 can obtain datasets from multiple processing state parsers 58 and learn from the evaluation results of whether the processing results are qualified or not. Each of the multiple processing state parsers 58 can be the processing state parser 58 of Embodiment 2 or the processing state parser 51 of Embodiment 1. The multiple processing state parsers 58 can be both processing state parser 58 and processing state parser 51.

[0107] The machine learning unit 59 can obtain datasets from multiple processing state parsers 58 used in the same location, or from processing state parsers 58 operating independently in different locations. It can add or remove the processing state parser 58 that is the source of the dataset midway through the process. The machine learning unit 59 can be set up independently of the processing state parser 58. In this case, the machine learning unit 59 can learn from a dataset obtained from a certain processing state parser 58, then connect to other processing state parsers 58 to obtain datasets from other processing state parsers 58 for further learning.

[0108] As described above, the machine learning unit 59 learns the relationship between the time-series data of light intensity output from the light sensor 13 or the feature quantity output from the feature extraction unit 53 and the evaluation result of whether the processing result is qualified or not. The machine learning unit 59 can learn the relationship between the time-series data of light intensity output from the light sensor 13 or the feature quantity output from the feature extraction unit 53 and the correction amount of the processing parameters. In this case, the data acquisition unit 61 acquires the time-series data of light intensity output from the light sensor 13 or the feature quantity output from the feature extraction unit 53 and the correction amount output from the correction amount calculation unit 55. After learning, the machine learning unit 59 can calculate and output the correction amount of each processing parameter based on the time-series data of light intensity output from the light sensor 13 or the feature quantity output from the feature extraction unit 53. When the trained model is prepared independently of the machine learning unit 59, the processing state analyzer 58 has an inference unit that uses the trained model learned by the learning unit 60 to calculate the correction amount of the processing parameters based on the processing qualification result.

[0109] The data acquisition unit 61, as input to the learning unit 60, acquires not only time-series data of light intensity output from the light sensor 13 or feature quantities output from the feature extraction unit 53, but also one or both of the plate thickness and material of the workpiece W. The learning unit 60 can use deep learning algorithms to learn by extracting the feature quantities themselves. The learning unit 60 can perform machine learning using discriminant analysis based on other known methods such as genetic programming, functional logic programming, support vector machines, Fisher discriminant analysis, partial space methods, or Mahalanobis space.

[0110] The learning algorithm used by the learning unit 60 can be a decision tree, random forest, logistic regression, k-nearest neighbor method, partial space method, CLAFIC (CLAss-Featuring Information Compression method), Isolation Forest, LOF (Local Outlier Factor), boosting method, AdaBoost, LogitBoost, One-Class SVM (Support Vector Machine), or Gaussian Mixture Model. For example, in the case of learning to extract features from images, such as deep learning or convolutional neural networks, a feature extraction unit 53 may not be required. The machine learning unit 59 can be set up for each poorly processed item, or one machine learning unit 59 can correspond to multiple poorly processed items.

[0111] As described above, the laser processing apparatus according to Embodiment 2 uses time-series data of light intensity output from the light sensor 13 or feature quantities output from the feature extraction unit 53 and evaluation results of whether the processing result is qualified or not to perform machine learning on the determination results of whether the processing is qualified or not. As a result, the laser processing apparatus according to Embodiment 2 achieves the same effect as the laser processing apparatus 50 according to Embodiment 1, and can determine the correction amount of the processing parameters with higher accuracy than the laser processing apparatus 50.

[0112] Furthermore, the machine learning unit 59 can learn the relationship between feature quantities and evaluation values ​​related to whether a process is satisfactory or not, which are relevant to the defective items that should be evaluated. The machine learning unit 59 can perform computational processing based on the feature quantities, thereby outputting the evaluation value of the defective items. In this case, the laser processing apparatus can perform more accurate evaluations of defective items.

[0113] Implementation method 3.

[0114] Figure 11 This diagram illustrates the structure of the laser processing apparatus 50D according to Embodiment 3. The laser processing apparatus 50D has all the structural elements of the laser processing apparatus 50 according to Embodiment 1, and also includes a temperature sensor 17 and a focusing position estimation unit 62. The focusing position estimation unit 62 estimates the focusing position, i.e., the position where the laser converges in the workpiece W, as the estimated focusing position. In Embodiment 3, structural elements having the same function as in Embodiment 1 are labeled with the same reference numerals as in Embodiment 1, and repeated descriptions are omitted. In Embodiment 3, the differences from Embodiment 1 are mainly explained.

[0115] The processing head 2 has internal optical components that allow laser light directed towards the workpiece W to pass through or be reflected. An example of the optical component is a condenser lens 7. A focusing position estimation unit 62 detects temperature changes in the optical component and estimates the focusing position based on the temperature of the optical component. A correction amount calculation unit 55, based on the determination result and the estimated focusing position, determines and corrects the processing parameters that should be corrected and the correction amount for the processing parameters during processing.

[0116] If a material absorbs laser light and is heated, the density and refractive index of the heated portion change. For transmissive optical components, an anti-reflective coating made of a material optimally suited to the laser wavelength is applied. Almost all light passes through the optical component, but a portion of the laser light is absorbed and converted into heat. This heat creates a difference in refractive index between the optical component and its periphery, resulting in a lens-like effect within the optical component. This lens-like effect due to heat is called the thermal lensing effect. Similarly, reflective optical components also undergo a high-reflectivity coating, but a portion of the laser light is absorbed and converted into heat, resulting in a thermal lensing effect.

[0117] The laser processing apparatus 50D uses a temperature sensor 17 to measure the thermal lensing effect. The change in focal length is estimated based on the value output from the temperature sensor 17, the output of the currently output laser beam L, and the irradiation diameter towards the lens. The temperature sensor 17 can be a heat flux sensor that measures the heat flux of optical components. The output of the laser beam L and the irradiation diameter towards the lens are read by the control unit 3.

[0118] The correction calculation unit 55 adjusts the focal length based on the change in focal length. Therefore, the laser processing apparatus 50D according to Embodiment 3 can adjust the focal length not only using time-series data of the processing light but also using other characteristic quantities. As a result, the laser processing apparatus 50D can adjust processing conditions with higher precision. Furthermore, the laser processing apparatus 50D can evaluate the reliability of the accuracy of the value output by the evaluation unit 54. Moreover, the laser processing apparatus 50D uses not only information obtained from the light sensor 13 but also temperature information of the optical components, thus enabling more precise adjustment of the focal position.

[0119] Figure 12This diagram illustrates a processor 91 in the case where at least a portion of the control unit 3, drive unit 5, condenser lens position change drive unit 6, processing status observation unit 52, feature quantity extraction unit 53, evaluation unit 54, and correction quantity calculation unit 55 of the laser processing apparatus 50 according to Embodiment 1 are implemented by the processor 91. That is, the functions of at least a portion of the control unit 3, drive unit 5, condenser lens position change drive unit 6, processing status observation unit 52, feature quantity extraction unit 53, evaluation unit 54, and correction quantity calculation unit 55 can be implemented by the processor 91 executing a program stored in the memory 92. The processor 91 is a CPU (Central Processing Unit), processing device, arithmetic unit, microprocessor, or DSP (Digital Signal Processor). Figure 12 Memory 92 is also shown.

[0120] When at least some of the functions of the control unit 3, drive unit 5, condenser lens position change drive unit 6, processing status observation unit 52, feature quantity extraction unit 53, evaluation unit 54, and correction quantity calculation unit 55 are implemented by the processor 91, these functions are implemented by the processor 91 and software, firmware, or a combination of software and firmware. The software or firmware is described as a program and stored in the memory 92. The processor 91 reads the program stored in the memory 92 and executes it, thereby implementing at least some of the functions of the control unit 3, drive unit 5, condenser lens position change drive unit 6, processing status observation unit 52, feature quantity extraction unit 53, evaluation unit 54, and correction quantity calculation unit 55.

[0121] When at least a portion of the functions of the control unit 3, drive unit 5, condenser lens position change drive unit 6, processing status observation unit 52, feature quantity extraction unit 53, evaluation unit 54, and correction amount calculation unit 55 are implemented by the processor 91, the laser processing apparatus 50 has a memory 92. This memory 92 stores a program that ultimately executes at least a portion of the program steps executed by the control unit 3, drive unit 5, condenser lens position change drive unit 6, processing status observation unit 52, feature quantity extraction unit 53, evaluation unit 54, and correction amount calculation unit 55. The program stored in the memory 92 can be considered as at least a portion of the sequence or method by which the computer executes the functions of the control unit 3, drive unit 5, condenser lens position change drive unit 6, processing status observation unit 52, feature quantity extraction unit 53, evaluation unit 54, and correction amount calculation unit 55.

[0122] Memory 92 includes, for example, non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), disks, floppy disks, optical disks, compact disks, mini disks, or DVDs (Digital Versatile Disk).

[0123] Figure 13 This diagram illustrates the processing circuit 93 in the case where at least a portion of the control unit 3, drive unit 5, condenser lens position change drive unit 6, processing state observation unit 52, feature quantity extraction unit 53, evaluation unit 54, and correction quantity calculation unit 55 of the laser processing apparatus 50 according to Embodiment 1 are implemented by the processing circuit 93. That is, at least a portion of the control unit 3, drive unit 5, condenser lens position change drive unit 6, processing state observation unit 52, feature quantity extraction unit 53, evaluation unit 54, and correction quantity calculation unit 55 can be implemented by the processing circuit 93.

[0124] The processing circuit 93 is dedicated hardware. The processing circuit 93 may be, for example, a single circuit, a composite circuit, a programmable processor, a parallel programmable processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a combination thereof.

[0125] Part of the control unit 3, drive unit 5, condenser lens position change drive unit 6, processing status observation unit 52, feature quantity extraction unit 53, evaluation unit 54, and correction quantity calculation unit 55 can be implemented by dedicated hardware different from the remaining units.

[0126] Regarding the multiple functions of the control unit 3, drive unit 5, condenser lens position change drive unit 6, processing status observation unit 52, feature quantity extraction unit 53, evaluation unit 54, and correction quantity calculation unit 55, some of these multiple functions may be implemented by software or firmware, while the remaining multiple functions may be implemented by dedicated hardware. As described above, the multiple functions of the control unit 3, drive unit 5, condenser lens position change drive unit 6, processing status observation unit 52, feature quantity extraction unit 53, evaluation unit 54, and correction quantity calculation unit 55 can be implemented by hardware, software, firmware, or a combination thereof.

[0127] At least a portion of the functions of the processing state analysis unit 52, feature extraction unit 53, machine learning unit 59, evaluation unit 54, and correction calculation unit 55 in the laser processing apparatus processing state analyzer 58 according to Embodiment 2 can be implemented by a processor executing a program stored in a memory. This memory is the same as memory 92, and the processor is the same as processor 91. At least a portion of the aforementioned processing state analysis unit 52, feature extraction unit 53, machine learning unit 59, evaluation unit 54, and correction calculation unit 55 can be implemented by a processing circuit. This processing circuit is the same as processing circuit 93.

[0128] At least a portion of the functions of the processing state observation unit 52, feature quantity extraction unit 53, evaluation unit 54, correction quantity calculation unit 55, and focus position estimation unit 62 in the laser processing apparatus 50D according to Embodiment 3 can be implemented by a processor executing a program stored in a memory. This memory is the same as memory 92, and the processor is the same as processor 91. At least a portion of the aforementioned processing state observation unit 52, feature quantity extraction unit 53, evaluation unit 54, correction quantity calculation unit 55, and focus position estimation unit 62 can be implemented by a processing circuit. This processing circuit is the same as processing circuit 93.

[0129] The structure shown in the above embodiments is an example, and it can also be combined with other known technologies, and the embodiments can be combined with each other. Without departing from the spirit of the subject, some parts of the structure can be omitted or changed.

[0130] Explanation of the label

[0131] 1 Laser oscillator, 2 Processing head, 3 Control unit, 4, 14 Collimating lenses, 5 Drive unit, 6 Condenser lens position change drive unit, 7 Condenser lens, 8 Processing beam, 9 Reflector, 10 Beam splitter, 10a Diffraction grating, 10b Prism, 11 Wavelength filter, 12 Imaging lens, 13 Optical sensor, 13a First optical sensor, 13b Second optical sensor, 13c Third optical sensor, 15 Optical fiber, 17 Temperature sensor, 50, 50A, 50B, 50C, 50D Laser processing device, 51, 56, 58 Processing state resolver, 52, 52A, 52B Processing state observation unit, 53 Feature extraction unit, 54 Evaluation unit, 55 Correction calculation unit, 57 Processing condition storage unit, 59 Machine learning unit, 60 Learning unit, 61 Data acquisition unit, 62 Condenser position estimation unit, 91 Processor, 92 Memory, 93 Processing circuit.

Claims

1. A laser processing apparatus, characterized in that, have: The drive unit changes the relative position of the processing head and the workpiece. The processing head includes a focusing optical system that focuses a laser emitted from a laser oscillator to irradiate the workpiece, and a processing gas supply unit that supplies processing gas toward the workpiece. The control unit controls the laser oscillator, the processing head, and the drive unit to perform processing based on numerical parameters related to laser processing, namely processing parameters; The processing status observation unit detects the light intensity of multiple predetermined wavelengths of interest of the light emitted from the workpiece by the laser irradiation, i.e., the processing light, as multiple light sensor signals. The feature extraction unit extracts at least one of the correlation indicators among the plurality of optical sensor signals and at least one of the feature quantities that can be obtained from one optical sensor signal; and A correction amount calculation unit, based on the characteristic quantity, determines the processing parameters used to perform the correction as correction parameters, and determines the correction amount of the correction parameters; and The machine learning department learns the relationship between the feature quantities and the evaluation values ​​for whether the processing is satisfactory or not, related to the poorly processed items that should be evaluated. The correction calculation unit includes an evaluation unit that, based on the characteristic quantity, determines whether the processing is qualified or not for at least one of a plurality of defective items, thereby obtaining a judgment result. Based on the determination result, the correction calculation unit determines the correction parameter that should be corrected and the correction amount of the correction parameter. The machine learning unit performs computational processing based on the feature quantity, thereby outputting an evaluation value for the poorly processed item.

2. A laser processing apparatus, characterized in that, have: The drive unit changes the relative position of the processing head and the workpiece. The processing head includes a focusing optical system that focuses a laser emitted from a laser oscillator to irradiate the workpiece, and a processing gas supply unit that supplies processing gas toward the workpiece. The control unit controls the laser oscillator, the processing head, and the drive unit to perform processing based on numerical parameters related to laser processing, namely processing parameters; The processing status observation unit detects the light intensity of multiple predetermined wavelengths of interest of the light emitted from the workpiece by the laser irradiation, i.e., the processing light, as multiple light sensor signals. The feature extraction unit extracts at least one of the correlation indicators among the plurality of optical sensor signals and at least one of the feature quantities that can be obtained from one optical sensor signal; and The correction calculation unit, based on the characteristic quantity, determines the processing parameters used to perform the correction as correction parameters, and determines the correction amount of the correction parameters. The correction calculation unit includes an evaluation unit that, based on the characteristic quantity, determines whether the processing is qualified or not for at least one of a plurality of defective items, thereby obtaining a judgment result. Based on the determination result, the correction calculation unit determines the correction parameter that should be corrected and the correction amount of the correction parameter. Based on the feature quantity, the evaluation unit determines, with respect to at least one of the plurality of defective processing items, the boundary value between the range of processing parameters for which the judgment result becomes good (i.e., the good processing range) and the range of processing parameters for which the judgment result becomes poor (i.e., the defective processing range). If the machining parameters corrected based on the correction amount are within the machining defect range, the correction amount calculation unit determines the difference between the machining parameters corrected based on the correction amount and the boundary value, i.e., the deviation. If the deviation exceeds the boundary value, the correction amount of the machining parameters is determined and corrected during machining.

3. The laser processing apparatus according to claim 1 or 2, characterized in that, It also includes a focusing position estimation unit, which estimates the laser focusing position in the workpiece as the estimated focusing position. The processing head has internal optical components for the laser to pass through or be reflected toward the object being processed. The focusing position estimation unit detects the temperature change of the optical component and estimates the focusing position based on the temperature of the optical component, thus establishing the estimated focusing position. Based on the determination result and the estimated focusing position, the correction calculation unit determines and corrects the processing parameters that should be corrected and the correction amount of the processing parameters during processing.

4. The laser processing apparatus according to claim 1 or 2, characterized in that, The processing status observation unit includes: a first optical sensor positioned in the direction in which the laser emitted from the laser oscillator irradiates the processing point; and a second optical sensor positioned in a direction different from the direction in which the laser emitted from the laser oscillator irradiates the processing point.

5. The laser processing apparatus according to claim 1 or 2, characterized in that, The plurality of defects include at least one of the following: roughness of the cut surface, scraping, slag removal, and oxide film peeling.

6. The laser processing apparatus according to claim 1 or 2, characterized in that, The correction calculation unit determines the processing parameters to be corrected and the correction amount of the processing parameters based on at least one of the cutting speed, focal position, focusing diameter, gas pressure, and laser output that should be corrected.

7. The laser processing apparatus according to claim 1 or 2, characterized in that, The processing status observation unit has a short-pass filter that allows light of wavelengths less than or equal to a first wavelength to pass through, a long-pass filter that allows light of wavelengths greater than or equal to a second wavelength to pass through, and a band-pass filter that allows light of wavelengths longer than the first wavelength and shorter than the second wavelength to pass through.

8. The laser processing apparatus according to claim 1 or 2, characterized in that, The processing status observation unit has a first wavelength filter that allows light with a wavelength shorter than 525 nm to pass through, a second wavelength filter that allows light with a wavelength longer than 700 nm to pass through, and a third wavelength filter that allows light with a wavelength of 530 nm or more and 700 nm or less to pass through.

9. The laser processing apparatus according to claim 1 or 2, characterized in that, The processing status observation unit has a wavelength filter that allows light with a wavelength of 475 nm or more and 525 nm or less to pass through, a wavelength filter that allows light with a wavelength of 575 nm or more and 625 nm or less to pass through, and a wavelength filter that allows light with a wavelength of 675 nm or more and 725 nm or less to pass through.

10. The laser processing apparatus according to claim 1 or 2, characterized in that, The processing status observation unit has a wavelength filter that allows light with wavelengths of 400 nm or more and 800 nm or less to pass through, a wavelength filter that allows light with wavelengths of 475 nm or more and 525 nm or less to pass through, and a wavelength filter that allows light with wavelengths of 675 nm or more and 725 nm or less to pass through.

11. The laser processing apparatus according to claim 1 or 2, characterized in that, Based on the processing light transmitted from the processing head via optical fiber, the amount of correction for the processing parameters is determined during processing, or the pass / fail status of the processing is judged in relation to processing defects that should be corrected.

12. The laser processing apparatus according to claim 1, characterized in that, The evaluation unit has: Feature extraction unit, which extracts the feature quantities; and The learning unit learns the relationship between the feature quantity and the determination result, and determines the determination result based on the learning result.

Images