WO2024154571A1 - Machining system and machinability determination system - Google Patents

Abstract

This machining system includes: a laser machining device capable of executing a machining step for emitting laser light on a material to be machined under machining emission conditions and cutting same, and a machinability determination step for emitting the laser light on the material to be machined under determination emission conditions that the material to be machined is melted but not penetrated, and determining the machinability of the material to be machined; a measurement device that measures an emission spectrum generated when the laser light is emitted on the material to be machined under the determination emission conditions; and a determination device that determines the machinability of the material to be machined on the basis of chronological data of the measured emission spectrum. First and second determination models are provided. Determination results of the machinability of the material to be machined are output on the basis of a combination of a surface quality evaluation and an internal quality evaluation of the material to be machined, obtained as a result of inputting first and second waveform information extracted from the chronological data of the emission spectrum into the first and second determination models.

Description

Machining system and machinability assessment system

The present invention relates to a processing system and a processability assessment system.

Conventionally, in laser processing devices such as laser processing machines, processing conditions are generally preset according to the material and thickness of the workpiece. Therefore, the operator of the laser processing device selects processing conditions that match the material and thickness of the workpiece, or performs laser processing under processing conditions instructed by the processing program.

However, even when the workpiece has the same material and thickness, good processing quality may not be obtained using the processing conditions pre-prepared in the laser processing device due to individual differences in the country of manufacture, manufacturer, production lot, storage conditions, etc., as well as differences between blast furnace and electric furnace materials. On the other hand, there is also a known technology in which processing conditions suitable for the material and thickness of the workpiece are selected, the actual material and thickness of the workpiece are measured, and the selected processing conditions are modified based on the measurement results (see Patent Document 1).

Patent No. 6754614

However, in the device disclosed in Patent Document 1, the processing conditions selected according to the material and thickness of the workpiece can be modified based on the actually measured material and thickness, but it is not envisaged that the device will be able to determine the level of processing quality that can be obtained if the selected processing conditions are not modified.

In other words, the above-mentioned conventional technology does not determine the workability (degree of suitability for cutting) of the workpiece based on the selected processing conditions before the actual cutting process, taking into account the individual differences (quality variations) of the actual workpiece that differ, for example, between manufacturers, between production lots, or between storage conditions.

As a result, there is a problem in that the operator cannot judge the workability based on the processing conditions of the workpiece before cutting, and then, based on the judgment result, decide whether to cut under the pre-set processing conditions or change the processing conditions to be more suitable for the workpiece without performing test processing (trial processing).

One aspect of the present invention is a processing system and a processability judgment system that judges the processability based on the processing conditions of the workpiece before cutting, and makes it easy to make a decision based on the judgment results as to whether cutting should be performed under preset processing conditions or whether the processing conditions should be changed to those suitable for the workpiece, thereby reducing processing defects.

A processing system according to one aspect of the present invention includes a laser processing device capable of executing a processing step of irradiating a workpiece with laser light under processing irradiation conditions to cut the workpiece, and a machinability judgment step of irradiating the workpiece with the laser light under judgment irradiation conditions that melt but do not penetrate the workpiece to judge the machinability of the workpiece, a measurement device that measures an emission spectrum generated when the workpiece is irradiated with the laser light under the judgment irradiation conditions, and a judgment device that judges the machinability of the workpiece based on time series data of the emission spectrum measured by the measurement device, and the judgment device judges the machinability of the workpiece based on a first judgment model and a second judgment model. The measuring device has a function to extract first waveform information in a first time domain and second waveform information in a second time domain from the time series data of the emission spectrum measured by the measuring device, input the extracted first waveform information as estimation data to the first judgment model to obtain a surface quality evaluation of the workpiece, input the extracted second waveform information as estimation data to the second judgment model to obtain an internal quality evaluation of the workpiece, and output a judgment result of the workability of the workpiece when cut and processed under preset processing conditions in the processing step based on a combination of the obtained surface quality evaluation and internal quality evaluation of the workpiece.

A machinability judgment system according to one embodiment of the present invention has a first judgment model and a second judgment model, and in a machinability judgment process for judging the machinability of a workpiece, first waveform information and second waveform information calculated from time series data of an emission spectrum generated when a laser beam is irradiated to the workpiece under judgment irradiation conditions that melt but do not penetrate the workpiece are input into the first judgment model and the second judgment model, respectively, and the system judges the machinability of the workpiece based on the first waveform information and the second waveform information, and a learning device that creates the first judgment model and the second judgment model, and the judgment device extracts first waveform information in a first time domain and second waveform information in a second time domain from the time series data of the emission spectrum, and uses the extracted first waveform information as estimation data to input the first waveform information and the second waveform information to the first judgment model and the second judgment model, respectively. The first waveform information and the surface quality evaluation result indicating the surface quality of the workpiece are input to the second judgment model as estimation data to obtain an internal quality evaluation of the workpiece, and based on the combination of the obtained surface quality evaluation and internal quality evaluation of the workpiece, a judgment result of the workability of the workpiece when cut and processed under preset processing conditions in a processing step in which the workpiece is cut by irradiating the workpiece with the laser light under processing irradiation conditions is output. The learning device inputs the first waveform information and the surface quality evaluation result indicating the surface quality of the workpiece as first teacher data to perform machine learning to create the first judgment model, and inputs the second waveform information and the internal quality evaluation result indicating the internal quality of the workpiece as second teacher data to perform machine learning to create the second judgment model.

In accordance with one aspect of the present invention, the processing system and machinability judgment system output a judgment result of the machinability of the workpiece when it is cut under preset processing conditions in the processing step based on a combination of the surface quality evaluation and internal quality evaluation of the workpiece obtained by inputting the first waveform information in the first time domain and the second waveform information in the second time domain extracted from the time series data of the emission spectrum measured by the measuring device when the workpiece is irradiated with laser light under judgment irradiation conditions that melt but do not penetrate the workpiece, into the first judgment model and the second judgment model as estimation data. This makes it possible to judge the machinability based on the processing conditions of the workpiece before cutting, making it easier to make a judgment based on the judgment result as to whether the cutting should be performed under the preset processing conditions during actual cutting, or whether the processing conditions should be changed to those suitable for the workpiece, and thus making it possible to reduce processing defects.

A processing system according to another aspect of the present invention includes a laser processing device capable of executing a processing step of irradiating a workpiece with laser light under processing irradiation conditions to cut the workpiece, and a workability judgment step of irradiating the workpiece with the laser light under first judgment irradiation conditions under which the laser light melts but does not penetrate the workpiece, and under second judgment irradiation conditions under which the laser light does not exceed the melting point of the material of the workpiece, to judge the workability of the workpiece; a first measurement unit that measures the emission spectrum generated when the laser light is irradiated to the workpiece under the first judgment irradiation conditions; and a second measurement unit that irradiates the workpiece with the laser light under the second judgment irradiation conditions. a measuring device including a second measuring unit that measures an infrared intensity of radiated light generated when light is irradiated; a first determining unit that determines the workability of the workpiece based on time series data of the emission spectrum measured by the first measuring unit of the measuring device; a second determining unit that determines the workability of the workpiece based on time series data of the infrared intensity measured by the second measuring unit of the measuring device; and a cutting process under preset processing conditions in the processing step based on a combination of a first determination result determined by the first determining unit and a second determination result determined by the second determining unit. and a third judgment unit that judges the workability of the workpiece when the workpiece is processed and outputs a third judgment result, the judgment device having a first judgment model and a second judgment model, the first judgment unit of the judgment device extracts first waveform information in a first time domain and second waveform information in a second time domain from time series data of the emission spectrum measured by the first measurement unit of the measurement device, inputs the extracted first waveform information as estimation data into the first judgment model to obtain a surface quality evaluation of the workpiece, and outputs the extracted second waveform information as estimation data to obtain a surface quality evaluation of the workpiece. The surface quality evaluation and internal quality evaluation of the workpiece are input into a fixed model to obtain an internal quality evaluation of the workpiece, and the first judgment result that judges the workability of the workpiece is output based on the combination of the obtained surface quality evaluation and internal quality evaluation of the workpiece, and the second judgment unit of the judgment device extracts feature information that indicates a temporal or positional change in the temperature of the workpiece based on the time series data of the infrared intensity measured by the second measurement unit of the measurement device, and outputs the second judgment result that judges the workability of the workpiece based on the extracted feature information and preregistered reference information for judging the workability of the workpiece.

A machinability judgment system according to another aspect of the present invention has a first judgment model and a second judgment model, and in a machinability judgment step for judging the machinability of a workpiece, a first judgment unit inputs first waveform information and second waveform information extracted from time series data of an emission spectrum generated when a laser beam is irradiated to the workpiece under a first judgment irradiation condition that melts but does not penetrate the workpiece, into the first judgment model and the second judgment model, respectively, and judges the machinability of the workpiece based on the first waveform information and the second waveform information; a second judgment unit that judges the workability of the workpiece based on feature amount information extracted from time-series data of infrared intensity of radiated light generated when the workpiece is irradiated under second judgment irradiation conditions not exceeding the melting point of the material, and pre-registered reference information for judging the workability of the workpiece; and a third judgment unit that judges the workability of the workpiece when cut under preset processing conditions in the processing step based on a combination of a first judgment result judged by the first judgment unit and a second judgment result judged by the second judgment unit, and outputs a third judgment result; and a learning device that creates a second judgment model, wherein the judgment device has a first judgment unit that extracts first waveform information in a first time domain and second waveform information in a second time domain from the time series data of the emission spectrum, inputs the extracted first waveform information as estimation data into the first judgment model to obtain a surface quality evaluation of the workpiece, inputs the extracted second waveform information as estimation data into the second judgment model to obtain an internal quality evaluation of the workpiece, and outputs the first judgment result that judges the workability of the workpiece based on a combination of the obtained surface quality evaluation and internal quality evaluation of the workpiece. The second judgment unit extracts feature information indicating the temporal or positional change in temperature of the workpiece based on the time series data of the infrared intensity, and outputs the second judgment result that judges the workability of the workpiece based on the extracted feature information and the reference information. The learning device inputs the first waveform information and the surface quality evaluation result indicating the surface quality of the workpiece as first teacher data to perform machine learning to create the first judgment model, and inputs the second waveform information and the internal quality evaluation result indicating the internal quality of the workpiece as second teacher data to perform machine learning to create the second judgment model.

According to another aspect of the present invention, the processing system and the workability judgment system judge the workability of the workpiece when cut under preset processing conditions in the processing step based on a combination of the surface quality evaluation and internal quality evaluation of the workpiece obtained by inputting the first waveform information and the second waveform information extracted from the time series data of the emission spectrum generated when the workpiece is irradiated with laser light under the first judgment irradiation condition, which melts but does not penetrate the workpiece, into the first judgment model and the second judgment model, and the second judgment result of the workability of the workpiece judged based on the feature information extracted from the time series data of the infrared intensity of the radiated light generated when the workpiece is irradiated with laser light under the second judgment irradiation condition, which does not exceed the melting point of the material of the workpiece, and the previously registered standard information for judging the workability of the workpiece, and a third judgment result is output. This allows the workability of the workpiece to be judged based on the processing conditions before cutting, making it easier to make decisions based on the judgment results, such as whether to perform the actual cutting process using preset processing conditions or to change to processing conditions suitable for the workpiece, thereby reducing processing defects.

According to one aspect of the present invention, the workability of the workpiece is judged based on the processing conditions before cutting, and a decision can be made based on the judgment results as to whether cutting should be performed under preset processing conditions or whether the processing conditions should be changed to those suitable for the workpiece, thereby reducing processing defects.

FIG. 1 is an explanatory diagram illustrating a basic configuration of a machining system according to a first embodiment of the present invention. FIG. 2 is a diagram showing an example of time-series data based on the measurement results obtained by measuring, with a spectroscope, the emission spectrum generated when the laser beam under the judgment irradiation conditions is irradiated onto a workpiece. FIG. 3 is a diagram showing an example of a first irradiation condition and a second irradiation condition of the laser beam. FIG. 4 is a diagram showing extraction points in the wavelength axis direction and the time axis direction for extracting the first wavelength information and the second waveform information from the time series data of the first emission spectrum and the time series data of the second emission spectrum. FIG. 5 is a diagram showing an example of a surface quality evaluation result that displays, in a comparative manner, the time waveform of the time-series data of region a, in which the surface state differs for each material of the processed material, and a surface image. FIG. 6 is a diagram showing an example of an internal quality evaluation result that represents, in a comparable manner, the time waveform of the time-series data of region b, in which the internal state differs for each material of the processed material. FIG. 7 is a schematic functional block diagram of the processing system. FIG. 8 is a diagram showing examples of standard conditions and difficult-to-work conditions as processing conditions according to the plate thickness of a workpiece of mild steel. FIG. 9 is a block diagram showing a schematic configuration of a workability determination system used in the machining system. FIG. 10 is an explanatory diagram showing a basic hardware configuration of the manufacturability determination unit, the learning device and/or the manufacturability determination system. FIG. 11 is a diagram illustrating an example of the decision matrix information. FIG. 12 is a flowchart showing an example of a processing flow of the machining system. FIG. 13 is a diagram showing a result table including the prediction results of the material state according to the first and second judgment models, the judgment results of the workability, and the results of the actual processing verification. FIG. 14 is an explanatory diagram showing the quality evaluation criteria for processing quality. FIG. 15 is a diagram showing the results of classifying the actual machining results in the determination matrix information. FIG. 16 is a diagram showing an example of details of the verification result of actual machining of a workpiece determined to be a standard material. FIG. 17 is a diagram showing an example of details of a verification result of actual machining of a workpiece determined to be a difficult-to-machine material. FIG. 18 is a diagram showing an example of details of the verification result of actual machining of a workpiece determined to be a difficult-to-machine material (recommended material for test cutting). FIG. 19 is an explanatory diagram illustrating a basic configuration of a processing system according to the second embodiment of the present invention. FIG. 20 is a schematic functional block diagram of the processing system. FIG. 21 is a result table showing the results of investigation of the thermal conductivity, laser cut surface, and cut surface roughness for each material of the processed material. FIG. 22 is a graph showing the relationship between the average roughness of the cut surface and the thermal conductivity. FIG. 23 is a graph showing the relationship between temperature and time when repeatedly heating and cooling a workpiece, and the relationship between the temperature reached by cooling and time. FIG. 24 is a graph showing the relationship between temperature and time when repeatedly heating and cooling a workpiece, and the relationship between the temperature reached by cooling and time. FIG. 25 is a diagram for explaining the results of the investigation into the surface condition of the type A workpiece. FIG. 26 is a diagram showing a cut surface image and surface roughness when the workpiece shown in FIG. 25 is cut. FIG. 27 is a diagram for explaining the results of the investigation into the surface condition of the type B workpiece. FIG. 28 is a diagram showing a cut surface image and surface roughness when the workpiece shown in FIG. 27 is cut. FIG. 29 is a diagram for explaining the results of investigating the surface condition of the type C workpiece. FIG. 30 is a diagram showing a cut surface image and surface roughness when the workpiece shown in FIG. 29 is cut. FIG. 31 is a flowchart showing an example of a process flow for determining workability using a radiation thermometer. FIG. 32 is a flowchart showing an example of a process flow for determining workability using a radiation thermometer. FIG. 33 is a diagram showing a result of classifying the judgment results using the spectroscope and the radiation thermometer and the cut quality evaluation results in the overall judgment matrix information. FIG. 34 is a table showing the results of the determination of the material state (material characteristics) of Samples 1 to 4 by the spectroscope and the radiation thermometer, and the evaluation results of the cut surfaces. FIG. 35 is a diagram showing the chemical components (mass %) other than iron contained in the materials of Samples 1 to 4. FIG. 36 is a graph showing the relationship between temperature and time when sample 1 was repeatedly heated and cooled. FIG. 37 is a graph showing the relationship between temperature and time when sample 3 was repeatedly heated and cooled. FIG. 38 is a graph showing the relationship between temperature and time when laser scanning was performed on Samples 1 and 2. FIG. 39 is a diagram showing a cut surface image and surface roughness when sample 1 is cut. FIG. 40 is a diagram showing a cut surface image and surface roughness when sample 2 is cut. FIG. 41 is a graph showing the relationship between temperature and time when laser scanning was performed on Samples 3 and 4. FIG. 42 is a diagram showing a cut surface image and surface roughness when sample 3 is cut. FIG. 43 is a diagram showing a cut surface image and surface roughness when sample 4 is cut. FIG. 44 is a flowchart showing an example of a process flow based on a comprehensive judgment of a machining system.

Below, a processing system and a machinability assessment system according to an embodiment of the present invention will be described in detail with reference to the attached drawings. However, the following embodiments do not limit the invention according to each claim, and not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention. Furthermore, in the following embodiments, identical or corresponding components are given the same reference numerals and duplicated explanations are omitted. Furthermore, in the embodiments, the arrangement, scale, dimensions, etc. of each component may be exaggerated or minimized and shown in a state that does not match the actual ones, and descriptions of some components may be omitted.

[First embodiment]
[Basic configuration of the machining system]
FIG. 1 is an explanatory diagram illustrating a basic configuration of a machining system according to a first embodiment of the present invention.
As shown in FIG. 1, the processing system 100 according to the first embodiment includes a laser processing device 10 capable of executing a processing step of irradiating a workpiece W with a laser beam LB under processing irradiation conditions to cut the workpiece W, and a workability judgment step of irradiating the workpiece W with the laser beam LB under judgment irradiation conditions to melt but not penetrate the workpiece W and judge the workability of the workpiece W, a spectrometer (measuring device) 30 that measures an emission spectrum generated when the workpiece W is irradiated with the laser beam LB under the judgment irradiation conditions, and a workability judgment unit (judgment device) 50 that judges the workability of the workpiece W based on time series data of the emission spectrum measured by the spectrometer (measuring device) 30. The workability judgment unit (judgment device) 50 has a first judgment model 5 (FIG. 9) and a second judgment model 6 (FIG. 9), and extracts the first waveform information 3 (FIG. 9) in the first time domain and the second waveform information 4 (FIG. 9) in the second time domain from the time series data of the emission spectrum measured by the spectrometer (measurement device) 30. The workability judgment unit (judgment device) 50 inputs the extracted first waveform information 3 as estimation data into the first judgment model 5 to obtain a surface quality evaluation of the workpiece W. The workability judgment unit (judgment device) 50 also inputs the extracted second waveform information 4 as estimation data into the second judgment model 6 to obtain an internal quality evaluation of the workpiece W. Then, the workability judgment unit (judgment device) 50 outputs a judgment result 9 (FIG. 9) of the workability of the workpiece W when cut and processed under preset processing conditions in the processing step based on a combination of the obtained surface quality evaluation and internal quality evaluation of the workpiece W. In addition, the above term "melting" means, for example, generating a molten pool, and "not penetrating" means, for example, not drilling a hole.

The first judgment model 5 is created by inputting the first waveform information 1 (Fig. 9) and the surface quality evaluation result 109a (Fig. 9) indicating the surface quality of the workpiece W as first teacher data and performing machine learning. The second judgment model 6 is created by inputting the second waveform information 2 (Fig. 9) and the internal quality evaluation result 109b (Fig. 9) indicating the internal quality of the workpiece W as second teacher data and performing machine learning.

The processing system 100 also includes an NC (Numerical Control) device 60 that controls the laser processing device 10, the spectrometer 30, and the machinability judgment unit 50, and a display 70 that can display various information. The machinability judgment unit 50 and the NC device 60 may be mounted so as to be included in the laser processing device 10.

If the workpiece W is steel, for example, it contains iron (Fe) as its main component, and if it is non-ferrous aluminum alloy steel, it contains aluminum (Al) as its main component. In addition to these main components, the workpiece W contains elements intentionally added by the manufacturer and elements mixed in as impurities. Note that impurities here refer to inclusions that are not the main inclusions of the material, and also include substances that induce physical phenomena such as bubbles when the workpiece W is melted.

The laser processing device 10 performs processing such as laser cutting and laser drilling of the material of the workpiece W (hereinafter, "cutting" and "drilling" are collectively referred to as "cutting processing"). The laser processing device 10 has a processing table 11 on which the workpiece W such as sheet metal is placed, an X-axis carriage 12 that moves in the X-axis direction of the processing table 11 in the figure relative to the processing table 11, a Y-axis carriage 13 that moves in the Y-axis direction on the X-axis carriage 12 in the figure, and a laser processing unit 20 that irradiates the workpiece W with laser light LB to perform cutting processing.

The laser processing unit 20 includes a laser oscillator 21 that generates and emits laser light LB, a laser processing head 22 that is mounted on a Y-axis carriage 13 and configured to be movable in the X-axis and Y-axis directions by the X-axis carriage 12 and the Y-axis carriage 13, and a process fiber 23 that transmits the laser light LB generated by the laser oscillator 21 to the laser processing head 22.

The laser processing device 10 also includes an assist gas supply device (not shown) that supplies assist gas. Note that the laser processing device 10 is not limited to a configuration in which the laser processing head 22 moves relative to the workpiece W, and a configuration in which the workpiece W moves relative to the laser processing head 22 may also be adopted.

The laser oscillator 21 may be of a type in which seed light emitted from a laser diode excites and amplifies Yb (ytterbium) or the like in a resonator to emit laser light LB of a specified wavelength, or a type that directly utilizes the laser light LB emitted from a laser diode.

The laser oscillator 21 emits laser light LB in the 1 μm band with a wavelength of 900 nm to 1100 nm. For example, a DDL (Direct Diode Laser) oscillator emits laser light LB with a wavelength of 910 nm to 950 nm, and a fiber laser oscillator emits laser light LB with a wavelength of 1060 nm to 1080 nm.

The blue semiconductor laser emits laser light LB with a wavelength of 400 nm to 460 nm. The green laser may be a fiber laser oscillator or a DDL oscillator that emits laser light LB with a wavelength of 500 nm to 540 nm, or may be a multi-wavelength resonator that combines the green laser with laser light LB in the 1 μm band.

In addition, although not shown, it may be possible to emit a guide light GB (e.g., wavelength 650 nm) that confirms the position on the workpiece W where the laser light LB is to be emitted.

The laser oscillator 21 emits laser light LB under processing irradiation conditions in the processing step, and emits laser light LB under judgment irradiation conditions in the workability judgment step for judging the workability of the workpiece W. The laser light LB emitted based on the judgment irradiation conditions melts the workpiece W but does not penetrate it, and is therefore not used for cutting the workpiece W by the laser processing device 10. The laser light LB under such judgment irradiation conditions may be irradiated by pulse oscillation.

The laser processing head 22 has a beam control unit 24. The beam control unit 24 has a function of controlling the laser light LB to a focusing diameter and divergence angle suitable for the material of the workpiece W. The beam control unit 24 has a collimator lens 24a that receives the laser light LB emitted from the output end of the process fiber 23 and converts it into a parallel beam, a folding mirror 24b that reflects the almost parallel beam of laser light LB emitted from the collimator lens 24a downward in the Z-axis direction perpendicular to the X-axis and Y-axis, and transmits light of a predetermined wavelength, and a processing condenser lens 24c that focuses the laser light LB reflected by the folding mirror 24b and irradiates it on the workpiece W. The folding mirror 24b is coated with a coating that reflects at least the wavelengths of the laser light LB and the guide light GB (e.g., 1080 nm, 650 nm).

The laser processing head 22 is provided with a nozzle 25 at its tip, which has a circular opening 25a for irradiating the workpiece W with the laser light LB. The nozzle 25 has a nozzle function of injecting a gas flow of a predetermined assist gas pressure supplied from an assist gas supply device onto the workpiece W together with the laser light LB in order to remove the molten workpiece W. The nozzle 25 is provided detachably on the laser processing head 22.

The collimator lens 24a, the folding mirror 24b, the condenser lens 24c, and the nozzle 25 are fixed in the laser processing head 22 with the optical axis adjusted in advance. In addition, a lens drive unit (not shown) is provided in the beam control unit 24 to drive the collimator lens 24a in a direction parallel to the optical axis (X-axis direction) to adjust the focusing position. In addition, in order to adjust the focusing position, the laser processing head 22 itself may be configured to be movable in the Z-axis direction perpendicular to the X-axis and Y-axis directions by a drive mechanism (not shown).

The spectroscope 30 is mounted, for example, on the upper housing of the laser processing head 22. The spectroscope 30 receives the light of the measurement object transmitted through the return mirror 24b from the processing side of the workpiece W toward the return mirror 24b at the transmission side of the return mirror 24b at the upper housing, and receives it at the light receiving unit 31. The spectroscope 30 disperses the light of the measurement object received at the light receiving unit 31 using a diffraction grating or the like, and detects the light intensity (spectrum) for each wavelength. In this way, the spectroscope 30 is configured to be able to measure the emission spectrum generated when the workpiece W is irradiated with the laser light LB. The spectroscope 30 outputs the detected light intensity (spectrum) for each wavelength as time series data. The spectroscope 30 may be provided on the side of the housing of the laser processing head 22.

Specifically, the spectrometer 30 outputs time series data of the spectrum of the light to be measured (first emission spectrum) generated in the workpiece W by the laser light LB irradiated to the workpiece W under the first irradiation condition to the machinability judgment unit 50. The spectrometer 30 also outputs time series data of the spectrum of the light to be measured (second emission spectrum) generated in the workpiece W by the laser light LB irradiated to the workpiece W under the second irradiation condition to the machinability judgment unit 50. Note that instead of the spectrometer 30, for example, a photodetector having a bandpass filter can be used as a measurement device.

Next, a process of determining workability in the machining system 100 configured as above will be described.
Fig. 2 is a diagram showing an example of time series data based on the measurement result of the emission spectrum generated when the laser light under the judgment irradiation condition is irradiated to the workpiece by a spectroscope. Fig. 3 is a diagram showing an example of the first irradiation condition and the second irradiation condition of the laser light. Fig. 4 is a diagram showing extraction points in the wavelength axis direction and the time axis direction for extracting the first waveform information and the second waveform information from the time series data of the first emission spectrum and the time series data of the second emission spectrum. Note that Fig. 2 shows the time series data of the 800 nm band in particular among the time series data of each wavelength, with the horizontal axis representing time (Time: milliseconds) and the vertical axis representing the amplitude (Amplitude) of the emission intensity in arbitrary units.

In the machinability judgment process, the laser processing device 10 irradiates the workpiece W with the laser light LB according to the judgment irradiation conditions. The judgment irradiation conditions may be constant during the period in which the laser light LB is irradiated, but in this embodiment, as shown in FIG. 2, the time from the start of irradiation of the laser light LB to the end of irradiation is divided into a first stage and a second stage, and as the judgment irradiation conditions, in the first stage, the laser light LB is irradiated to the workpiece W under the first irradiation condition 101 (FIG. 3), and in the second stage, the laser light LB is irradiated to the workpiece W under the second irradiation condition 102 (FIG. 3). Note that the laser processing device 10 may spot irradiate the laser light LB under the judgment irradiation conditions (irradiate without moving the laser light LB) in the machinability judgment process. The first irradiation condition 101 is a condition suitable for evaluating the surface quality of the workpiece W, and the second irradiation condition is a condition suitable for evaluating the internal quality (e.g., the proportion of impurities, etc.) of the workpiece W. The second irradiation condition 102 may be a condition for shortening the measurement time more than the first irradiation condition 101. The laser output of the first irradiation condition 101 and the second irradiation condition 102 of the laser light LB is smaller than the laser output of the processing irradiation condition in the processing step. Also, the laser output of the first irradiation condition 101 is smaller than the laser output of the second irradiation condition 102. Specifically, the laser output of the processing irradiation condition is 7000W to 8000W, whereas the laser output of the first irradiation condition 101 is 1000W and the laser output of the second irradiation condition 102 is 1200W to 1600W. Also, the laser irradiation time (first stage) of the first irradiation condition 101 of the laser light LB irradiated to the workpiece W is shorter than the laser irradiation time (second stage) of the second irradiation condition 102. The nozzle gap of the laser processing device 10 when irradiating the laser light LB under the judgment irradiation conditions (first irradiation conditions 101 and second irradiation conditions 102) is set to be larger than the nozzle gap of the laser processing device 10 when irradiating the laser light LB under the processing irradiation conditions.

More specifically, as shown in Figures 2 and 3, the first irradiation condition 101 of the laser light LB in the first stage is a laser output of 1000 W and a laser irradiation time of 0.5 seconds (500 ms) from the start of the laser (Laser ON). The second irradiation condition 102 of the laser light LB in the second stage is a laser output of 1400 W and a laser irradiation time of 2.0 seconds (2000 ms) from 500 ms to 2500 ms when the laser is ended (Laser OFF).

The output frequency of the laser light LB common to the first and second irradiation conditions 101, 102 is 1000 Hz, the duty is 100%, the gas type of the assist gas is air, the gas pressure is 0.1 MPa, and the nozzle gap is 50 mm. The material of the workpiece W irradiated with the laser light LB is, for example, mild steel (SS400) with a plate thickness of 19 mm (t19).

Next, we will explain the process of obtaining time series data of the emission spectrum measured by the spectrometer 30 as a result of irradiating the workpiece W with the laser light LB under the above-mentioned judgment irradiation conditions, and the first waveform information and the second waveform information used to judge workability from this time series data.
The time series data of the emission spectrum measured by the spectrometer 30 is time series data having the time series data shown in FIG. 2 for the wavelengths constituting the emission spectrum. The time series data is dimensionally compressed to extract first waveform information and second waveform information representing features suitable for machine learning. Various methods of dimensional compression are possible, but in this embodiment, data of a specific wavelength band at the start and end parts of the time series data is compressed. First waveform information 1 and 3 representing features of the surface state of the workpiece W are generated based on the time series data of the first time domain in the first stage. In addition, second waveform information 2 and 4 representing features of the internal state of the workpiece W are generated based on the time series data of the second time domain in the second stage.

In other words, during the irradiation time of the laser light LB under the first irradiation condition 101, for example, as shown in FIG. 2, the time series data of region a from the start of the laser to approximately 0.2 seconds (0 seconds to 0.23 seconds) mainly includes surface condition features that indicate the reaction between the laser light LB and the surface (material surface) of the workpiece W. In addition, the time series data from approximately 0.2 seconds onwards indicates the reaction between the laser light LB and the inside of the workpiece W (material interior).

Furthermore, within the irradiation time of the laser light LB under the second irradiation condition 102, for example, the time series data of region b between about 1.8 seconds and about 2.3 seconds (1.83 seconds to 2.35 seconds) contains a feature of the internal state that indicates the reaction between the laser light LB and the inside of the workpiece W. Such a feature varies depending on the material of the workpiece W (differences in material quality and individual differences, etc.).

Note that the first and second irradiation conditions 101, 102 of the laser light LB are significantly different from the processing irradiation conditions used in actual product processing. In general product processing, for example, the nozzle gap is about 0.3 mm to 1.5 mm, whereas in the first and second irradiation conditions 101, 102, a very wide gap of 50 mm is set.

This is because the oxide film on the material surface of the workpiece W is very thin, about 10 μm to 50 μm, and if the energy density of the laser light LB is too high, the physical phenomenon caused by the reaction accompanied by light emission upon irradiation will be fast, making it very difficult to capture this reaction with the spectroscope 30. Therefore, by irradiating the workpiece W with defocused laser light LB and ensuring a suppressed energy density and the size of the molten pool formed, the surface condition of the workpiece W can be captured.

According to the results of experiments conducted by the applicant, if the nozzle gap is 10 mm or less, the reaction cannot be detected by the spectroscope 30, the reaction can be seen from about 20 mm, and if the distance is too far, such as about 100 mm, it is not possible to obtain a sufficient amount of light due to emission. For this reason, it is considered that the appropriate nozzle gap range for the first and second irradiation conditions 101, 102 is about 30 mm to 70 mm. Since the energy density at this time is about 5 kW/ ^cm2 to 29 kW/ ^cm2 , it was set to 50 mm, which is in the middle of this appropriate range.

On the other hand, unlike when capturing the surface state, it is possible to increase the energy density when measuring the internal state. Therefore, in order to shorten the measurement time, the laser output is set higher in the second irradiation condition 102 than in the first irradiation condition 101. However, if the laser output is increased to, for example, about 1800 W, the excessive energy makes it difficult to distinguish the time waveforms of standard materials that can be cut well (materials that can be cut well) from difficult-to-process materials, so the appropriate laser output range is preferably 1200 W to 1600 W. Therefore, the laser output in the second irradiation condition 102 is set to 1400 W, which is in the middle of this laser output range.

As shown in FIG. 4, time series data of four wavelength components 105-108 set in the wavelength direction and time direction is extracted from the time series data of the first and second emission spectra obtained from the spectrometer 30. Specifically, these four wavelength components 105-108 are wavelength components 105, 106 in a first time region of 0 seconds to 0.23 seconds and wavelength components 107, 108 in a second time region of 1.83 seconds to 2.35 seconds, which are extracted from a first wavelength band (first wavelength band) 103 with a wavelength of 450 nm to 570 nm and a second wavelength band (second wavelength band) 104 with a wavelength of 720 nm to 850 nm, respectively.

Wavelength components 105 and 106 are each composed of spectral data in which samples are extracted from, for example, 66 rows and 100 rows in the wavelength direction and 200 columns in the time direction. Wavelength components 107 and 108 are each composed of spectral data in which samples are extracted from, for example, 66 rows and 100 rows in the wavelength direction and 450 columns in the time direction.

Then, the first waveform information 1, 3 and the second waveform information 2, 4 are obtained from the above wavelength components 105 to 108 as follows. That is, first, four time waveforms are calculated by averaging each of the wavelength components 105 to 108 in the wavelength direction. Next, the ratio between the time waveform of the central wavelength (e.g., 514 nm) of the first wavelength band 103 and the time waveform of the central wavelength (e.g., 811 nm) of the second wavelength band 104 is calculated. The ratio between these averaged time waveforms and the time waveform of the central wavelength contains the bandwidth information and time change information of the time series data.

Among the calculated values thus obtained, the time waveform obtained by averaging the time series data (wavelength components 105, 106) in the a region of the first emission spectrum is used as first waveform information 1, 3. In addition, the ratio of the time waveform obtained by averaging the time series data (wavelength components 107, 108) in the b region of the second emission spectrum to the time waveform of the central wavelength is used as second waveform information 2, 4. These first waveform information 1, 3 and second waveform information 2, 4 can be used in machine learning, which will be described later, as data that accurately represent the surface condition and internal condition of the workpiece W, respectively.

The first and second wavelength bands 103, 104 of the emission spectrum extracted as described above are preferably wavelength bands from the visible light range to the near infrared range excluding the wavelength (emission wavelength band) that is the reference for laser processing and the wavelength of the guide light GB, and more preferably, are wavelength bands from 450 nm to 850 nm excluding the wavelength of the guide light GB. This makes it possible for the first waveform information 1, 3 and the second waveform information 2, 4 used to determine the workability to be calculated using only the emission spectrum from the workpiece W without being affected by the reflected light of the laser light LB and the guide light GB.

FIG. 5 is a diagram showing an example of a surface quality evaluation result that comparatively represents the time waveform of the time series data of region a, where the surface state differs for each material of the workpiece, and the surface image. FIG. 6 is a diagram showing an example of an internal quality evaluation result that comparatively represents the time waveform of the time series data of region b, where the internal state differs for each material of the workpiece. Note that the surface quality evaluation result 109a and the internal quality evaluation result 109b include quality scores (point evaluations) that score the surface quality evaluation and internal quality evaluation of the workpiece W into multiple different numerical values.

As shown in FIG. 5, in the surface quality evaluation result 109a of the workpiece W made of mild steel with a plate thickness of 19 mm, the quality evaluations of the surface quality, "○", "△" and "×", are defined by, for example, actually checking and evaluating the material surface of the workpiece W from the viewpoint of workability, and in this order represent "good", "poor" and "unacceptable", and are assigned quality scores (point evaluations) of "2", "1" and "0", respectively. In the case of the material (material A) of the workpiece W made by manufacturer A, for example, the adhesion of the oxide film is good and the surface roughness is low, so the quality evaluation is "○" and the quality score is "2". Also, the time waveform of the time series data of region a has a peak at the rising edge and then stabilizes.

In addition, in the case of the material W of manufacturer B (material B) and the material W of manufacturer C (material C), for example, the oxide film peels off easily and the surface is rough, so the quality evaluation is "△" and the quality score is "1." In addition, the time waveform of the time series data of region a, for example, compared to material A with a quality evaluation of "○," the rising peak occurs at a lower level like material B or a higher level like material C, and then the waveform becomes unstable.

Furthermore, in the case of material W (material D) of manufacturer D, for example, the oxide film is extremely thin, rust has occurred, and the surface is rough, so the quality evaluation is "x" and the quality score is "0." Also, the time waveform of the time series data of region a is a waveform that does not have a peak at the rising edge. In this way, according to the surface quality evaluation result 109a, it can be seen that the time waveform of the time series data of region a differs depending on the surface condition of the material.

On the other hand, as shown in FIG. 6, in the internal quality evaluation result 109b of the workpiece W made of mild steel with a plate thickness of 19 mm, the internal quality evaluations of "○", "△" and "×" are defined by actually checking and evaluating the quality from the viewpoints of, for example, the intensity and magnitude of fluctuation of the obtained time waveform and the degree of removal of impurities during laser irradiation, and in this order represent "good", "poor" and "unacceptable", and are assigned quality scores (score evaluations) of "2", "1" and "0", respectively.

In the case of material W (material E) made by manufacturer E, compared to material W made by manufacturer F (material F) and material W made by manufacturer G (material G), for example, the intensity of the time waveform in the time series data of region b is lower and the fluctuation is smaller. Also, the amount of impurities extracted from the molten pool during laser irradiation is small. For this reason, the quality evaluation is "○" and the quality score is "2".

In the case of material G, for example, the intensity of the time waveform of the time series data in region b is high and the fluctuation is large compared to materials E and F. Also, a large amount of impurities were removed from the molten pool during laser irradiation. For this reason, the quality rating is "X" and the quality score is "0". In the case of material F, the intensity and fluctuation of the time waveform of the time series data in region b is somewhere between those of materials E and G, and it was observed that impurities were removed from the molten pool during laser irradiation to a certain extent more than material E and to a certain extent less than material G. For this reason, the quality rating is "△" and the quality score is "1".

In this way, according to the internal quality evaluation result 109b, it can be seen that the degree of removal of impurities during laser irradiation varies depending on the internal state of the material, and that as the amount of removal increases, the intensity and fluctuation of the time waveform of the time series data in region b increases. It was also found that as these intensity and fluctuation increase, cutting processing such as laser cutting by the laser processing device 10 becomes more difficult.

The surface condition and internal condition of the workpiece W indicated by the above quality evaluation and quality score may be divided to allow for more detailed understanding. For example, a quality evaluation of surface quality "△" may be changed to "△´" for those with high peaks in the time waveform and to "△" for those with low peaks, or may be subdivided based on the frequency components of the time waveform.

In this way, for example, in cutting processes such as oxygen cutting of mild steel using a laser processing device 10, both the surface condition and internal condition of the workpiece W are important factors. The applicant has discovered that it is possible to appropriately grasp both the surface condition and internal condition of the material based on the time waveform of the time series data in region a and the time waveform of the time series data in region b obtained from the spectroscope 30 shown in surface quality evaluation result 109a and internal quality evaluation result 109b in Figures 5 and 6.

The processing system 100 is configured to determine the workability of the workpiece W based on a judgment model created by machine learning using the first waveform information 1 and the second waveform information 2 calculated from the time series data of the emission spectrum including these time waveforms as explanatory variables and the quality score as the objective variable. Note that, because different physical phenomena occur on the surface side and the interior side when the workpiece W is irradiated with the laser light LB under the first and second irradiation conditions 101 and 102, it was decided to create separate judgment models for each side.

Here, a description will be given of the machinability determining unit 50 and the NC device 60 of the machining system 100. Fig. 7 is a schematic functional block diagram of the machining system.
As shown in FIG. 7 , the machinability judgment unit 50 includes a judgment model storage unit 51 , a machinability calculation unit 52 , and a machinability judgment unit 53 .

The judgment model storage unit 51 readably stores a first judgment model 5 (FIG. 9) for evaluating machining conditions according to the surface state of the workpiece W, and a second judgment model 6 (FIG. 9) for evaluating machining conditions according to the internal state of the workpiece W. For example, the first judgment model 5 and the second judgment model 6 are created for each thickness of the workpiece W. The machining conditions to be evaluated are stored in advance, for example, in the machining condition storage unit 62 of the NC device 60, which will be described later. The first judgment model 5 and the second judgment model 6 may be stored in an external server or the like, rather than being stored inside the judgment model storage unit 51 of the machinability judgment unit 50.

The workability calculation unit 52 assigns a quality score to the surface condition and a quality score to the internal condition as a workability evaluation of the workpiece W based on, for example, the plate thickness information of the workpiece W, the first and second judgment models 5, 6 read from the judgment model storage unit 51, and the first waveform information and the second waveform information acquired by the spectrometer 30, and generates judgment matrix information as a combination of the surface quality evaluation and the internal quality evaluation. The judgment matrix information can be stored in the judgment model storage unit 51.

The judgment matrix information is used, for example, to predict the workability of the workpiece W based on a combination of the quality of the surface and internal conditions before cutting the workpiece W to be processed, and is information that makes it possible to determine the recommended processing conditions according to the predicted workability. This judgment matrix information will be described later.

The machinability determination unit 53 determines and decides the machinability of the workpiece W (standard material that can be cut well, difficult-to-cut material) and the recommended machining conditions (standard conditions, difficult-to-cut material conditions) based on the judgment matrix information generated by the machinability calculation unit 52. The machining conditions include, for example, various machining conditions for laser cutting according to the material (mild steel, stainless steel, blast furnace material, electric furnace material, etc.) and plate thickness (19 mm, 22 mm, etc.) of the workpiece W (laser output (peak power, repetition frequency, pulse width), machining velocity, focus position, etc.). In addition, the energy density per unit time applied to the workpiece can be determined by parameters such as the above-mentioned processing conditions (laser output (average laser output), focal position (focus diameter), and processing speed).

The NC device 60 is equipped with a control unit 61 that outputs control instructions for various operations to the laser processing device 10, the spectrometer 30, and the machinability judgment unit 50, and can output display instructions for various information showing the judgment result on the display 70 based on the judgment result of the machinability judgment unit 53, and a processing condition storage unit 62 that stores various processing conditions for the workpiece W in a storage device or the like.

The machining conditions may be prepared in advance in the NC device 60 as standard conditions for standard specifications and machining conditions for difficult-to-machine materials (hereinafter referred to as "difficult-to-machine material conditions") according to the material and thickness of the workpiece W, for example, and may be stored in the machining condition storage unit 62. The control unit 61 can also output various information notification commands, such as audio output and lighting, to a speaker or lamp via an output I/F (Interface) described later.

FIG. 8 is a diagram illustrating standard conditions and difficult-to-machine conditions as machining conditions according to the thickness of the workpiece of mild steel. Machining conditions 110a shown in FIG. 8(a) show the standard conditions and difficult-to-machine conditions when the workpiece W has a thickness of 19 mm (t 19 mm), and machining conditions 110b shown in FIG. 8(b) show the standard conditions and difficult-to-machine conditions when the workpiece W has a thickness of 22 mm (t 22 mm).

As shown in Figures 8(a) and 8(b), the standard conditions and difficult-to-machine material conditions for machining conditions 110a for a plate thickness of 19 mm and machining conditions 110b for a plate thickness of 22 mm each include the following parameters: nozzle, machining speed (mm/min), laser output (peak output) (W), (pulse) frequency (Hz), (pulse) duty (pulse width) (%), gas type, gas pressure (MPa), nozzle gap (mm), and ACL (function to change the focused diameter). Note that ACL is a parameter related to the beam diameter when the laser light LB is collimated, and the larger the value, the larger the beam diameter of the collimated light, and the larger the beam diameter of the collimated light, the smaller the beam spot diameter.

In the standard conditions of the processing conditions 110a, the above-mentioned parameters are set to type-A, 1000, 7000, 1000, 75, O ₂ , 0.06, 1, and 70. In the difficult-to-machine material conditions of the processing conditions 110a, the above-mentioned parameters are set to the same items except for the nozzle and ACL, but the nozzle is type-B and the ACL is set to 80.

In the standard conditions of the processing conditions 110b, the above-mentioned parameters are set to type-A, 900, 8000, 1000, 75, _O2 , 0.06, 1, and 70. In the difficult-to-machine material conditions of the processing conditions 110b, the above-mentioned parameters are set to the same except for the nozzle, gas pressure, and ACL, but the nozzle is type-B, the gas pressure is set to 0.07, and the ACL is set to 80.

FIG. 9 is a block diagram showing a schematic configuration of a workability determination system used in the machining system.
As shown in Figure 9, a workability judgment system 90 according to one embodiment has a first judgment model 5 and a second judgment model 6, and in a workability judgment process for judging the workability of a workpiece W, first waveform information 3 and second waveform information 4 calculated from time series data of the emission spectrum generated when laser light LB is irradiated onto the workpiece W under judgment irradiation conditions that melt but do not penetrate the workpiece W are input into the first judgment model 5 and the second judgment model 6, respectively, and the workability judgment unit (judgment device) 50 judges the workability of the workpiece W based on the first waveform information 3 and the second waveform information 4, and a learning device 80 creates the first judgment model 5 and the second judgment model 6. The machinability judgment unit 50 extracts the first waveform information 3 in the first time domain and the second waveform information 4 in the second time domain from the time series data of the emission spectrum, inputs the extracted first waveform information 3 as estimation data to a first judgment model 5 to obtain a surface quality evaluation of the workpiece W, and inputs the extracted second waveform information 4 as estimation data to a second judgment model 6 to obtain an internal quality evaluation of the workpiece W. In addition, based on a combination of the obtained surface quality evaluation and internal quality evaluation of the workpiece W, the machinability judgment unit 50 outputs a judgment result of the machinability of the workpiece W when cut under preset processing conditions in a processing step in which the workpiece W is cut by irradiating the workpiece W with a laser beam LB under processing irradiation conditions. The learning device 80 inputs the first waveform information 1 and the surface quality evaluation result 109a indicating the surface quality of the workpiece W as first teacher data and performs machine learning to create a first judgment model 5, and inputs the second waveform information 2 and the internal quality evaluation result 109b indicating the internal quality of the workpiece W as second teacher data and performs machine learning to create a second judgment model 6.

In other words, the machinability judgment system 90 is configured to include the learning device 80 described above and the machinability judgment unit 50 included in the processing system 100. The learning device 80 may be provided inside the processing system 100 or outside the processing system 100.

The learning device 80 has a first judgment model learning unit 81 and a second judgment model learning unit 82. Two pieces of teacher data are input to the first judgment model learning unit 81 during the learning process. The first piece of data is the first waveform information 1 as an explanatory variable based on the time series data of the a region of the first emission spectrum measured by the spectrometer 30. The second piece of data is the quality score of the surface quality as a target variable included in the surface quality evaluation result 109a. The first judgment model learning unit 81 inputs the first waveform information 1 and the surface quality evaluation result 109a as first teacher data, performs machine learning based on the first teacher data, and creates and outputs a first judgment model 5 for predicting the surface condition of the workpiece W.

Two pieces of training data are input to the second judgment model learning unit 82 during the learning process. The first is the second waveform information 2 as an explanatory variable based on the time series data of the b region of the second emission spectrum measured by the spectrometer 30. The second is the quality score of the internal quality as a target variable included in the internal quality evaluation result 109b. The second judgment model learning unit 82 inputs the second waveform information 2 and the internal quality evaluation result 109b as second training data, performs machine learning based on the second training data, and creates and outputs a second judgment model 6 for predicting the internal state of the workpiece W.

Meanwhile, two pieces of estimation data are input to the workability calculation section 52 of the workability judgment unit 50 during the estimation process. The first is first waveform information 3 based on time series data in the a region of the first emission spectrum measured by the spectrometer 30 before cutting the workpiece W to be newly cut. The second is second waveform information 4 based on time series data in the b region of the second emission spectrum measured by the spectrometer 30 before cutting the workpiece W to be newly cut.

The machinability calculation unit 52 inputs the first and second waveform information 3, 4 as estimation data, and generates judgment matrix information based on the first and second judgment models 5, 6 created by the learning device 80. The machinability judgment unit 53 outputs a judgment result 9 that can notify the machinability of the workpiece W in the cutting process that will be performed by the laser processing device 10 and the recommended processing conditions based on the judgment matrix information. Note that the explanatory variables and objective variables used in the machinability judgment system 90 are not limited to those exemplified above.

The machinability judgment unit (judgment device) 50 may include a display (notification section) 70 that notifies the judgment result 9 in a confirmable manner, and the judgment result 9 may include a machinability evaluation that can express the adaptability to the workpiece W according to the processing conditions in the combination of the surface quality evaluation and the internal quality evaluation. The display (notification section) 70 may be configured to notify at least one of information informing the processing conditions set in the laser processing device 10 according to the material of the workpiece W based on the machinability evaluation, information encouraging test processing by the laser processing device 10, and information encouraging the execution of any one of adjustment, change, and setting of the processing conditions. The machinability judgment unit (judgment device) 50 may also include a setting section that accepts input information selected and input by the operator based on the information notified by the display (notification section) 70, and sets the processing conditions in the laser processing device 10 according to the accepted input information.

In addition, for machine learning and judgment in the learning device 80 and the machinability judgment unit 50, Random Forest (RF), Regression Analysis (RA), Principal Component Analysis (PCA), Singular Value Decomposition (SVD), Linear Discriminant Analysis (LDA), Independent Component Analysis (ICA), Gaussian Process Latent Variable Model (Gaussian Process Latent Variable Model (GPV)), and other methods are used. Various algorithms can be used, including, but not limited to, GPLVM, Logistic Regression (LR), Support Vector Machine (SVM), Discriminant Analysis (DA), Ranking Support Vector Machine (RSVM), Gradient Boosting (GB), Naive Bayes (NB), K-Nearest Neighbor Algorithm (K-NN), and Neural Network (NN).

[Hardware configuration]
FIG. 10 is an explanatory diagram showing a basic hardware configuration of the manufacturability determination unit, the learning device and/or the manufacturability determination system.
As shown in FIG. 10 , the processability judgment unit 50, the learning device 80 and/or the processability judgment system 90 are realized by hardware including, for example, a GPU (Graphics Processing Unit) 212, a CPU (Central Processing Unit) 201, a RAM (Random Access Memory) 202, a ROM (Read Only Memory) 203, a HDD (Hard Disk Drive) 204, a SSD (Solid State Drive) 205, and a memory card 206.

The machinability determination unit 50, the learning device 80 and/or the machinability determination system 90 further include, for example, an input I/F (interface) 207, an output I/F (interface) 208 and a communication I/F (interface) 209. Each of the hardware components 201 to 209 is connected to each other by a bus 200.

The input I/F 207 is connected to input equipment 211 including various input devices such as a keyboard, trackball, joystick, mouse, and touch panel, a measuring device such as a spectrometer 30, and various sensors such as a temperature sensor, an optical sensor, an acoustic sensor, an image sensor, and a spectroscopic sensor. The output I/F 208 is connected to output equipment 210 including a display 70 that functions as an alarm unit, and a speaker and a lamp (not shown). The communication I/F 209 communicates with an external device 214 such as a server via a network 213 such as the Internet. Each component of the above-mentioned machinability judgment unit 50, learning device 80, and/or machinability judgment system 90 may be configured by such hardware.

FIG. 11 is a diagram illustrating an example of the decision matrix information.
As shown in Figure 11, the judgment matrix information 110 is composed of data in which, for example, the vertical axis represents the quality score of the quality evaluation of the surface condition of the workpiece W and the horizontal axis represents the quality score of the quality evaluation of the internal condition of the workpiece W, and these are combined in a matrix table.

In FIG. 11, "S_score" represents the quality score of the surface state, and "B_score" represents the quality score of the internal state. These quality scores are classified according to quality evaluations of "○", "△", and "×" based on, for example, a predetermined threshold value. For example, the quality evaluation of "○" is a surface state quality score of 1.4 or more (S_score≧1.4) and an internal state quality score of 1.4 or more (B_score≧1.4). The quality evaluation of "△" is a surface state quality score of 0.6 or more and less than 1.4 (0.6≦S_score<1.4) and an internal state quality score of 0.6 or more and less than 1.4 (0.6≦B_score<1.4). The quality evaluation of "×" is a surface state quality score less than 0.6 (S_score<0.6) and an internal state quality score less than 0.6 (B_score<0.6).

The workpiece W for which the quality evaluation based on the quality scores of the surface condition and internal condition scored by the machinability calculation unit 52 is both "○" is classified into the area 111 of the first recommended conditions (Reco_cond=1), where the variable of the recommended machining conditions (Reco_cond) in the judgment matrix information 110 is "1".

In addition, the workpiece W whose quality evaluation based on the quality score of the surface condition scored by the machinability calculation unit 52 is either "△" or "X" and whose quality evaluation based on the quality score of the internal condition is "X" is classified into the area 113 of the third recommended conditions (Reco_cond = 3) in which the variable of the recommended machining conditions (Reco_cond) in the judgment matrix information 110 is "3".

Furthermore, any workpiece W whose quality evaluation based on the quality scores of the surface condition and internal condition scored by the machinability calculation unit 52 is any other than those mentioned above is classified into the area 112 of the second recommended conditions (Reco_cond=2) in which the variable of the recommended machining conditions (Reco_cond) in the judgment matrix information 110 is "2".

The workpiece W classified in the region 111 of the first recommended conditions (Reco_cond = 1) is determined to be a standard material that can be laser cut (cutting process) well under standard processing conditions. The workpiece W classified in the region 112 of the second recommended conditions (Reco_cond = 2) is determined to be a difficult-to-process material that can be laser cut (cutting process) well under difficult-to-process material processing conditions. The workpiece W classified in the region 113 of the third recommended conditions (Reco_cond = 3) is determined to be a difficult-to-process material (test cut recommended material) that is difficult to laser cut (cut process) well even when difficult-to-process material conditions are used as processing conditions. In other words, difficult-to-process materials include test cut recommended materials that cannot be cut well under either standard conditions or difficult-to-process material conditions.

The machinability determination unit 53 uses such determination matrix information 110 to determine the machinability of the workpiece W, and outputs a determination result 9 that can be notified on the display 70, etc., including the machining conditions recommended under the first to third recommended conditions. Specifically, the machining conditions recommended under the first recommended conditions (Reco_cond = 1) are standard conditions, and a determination result 9 is output that can notify the operator that the workpiece W can be machined under the standard conditions. In addition, the machining conditions recommended under the second recommended conditions (Reco_cond = 2) are difficult-to-machine material conditions, and a determination result 9 is output that can notify the operator that the workpiece W should be machined under difficult-to-machine material conditions (including a change to the difficult-to-machine material conditions). In addition, the processing conditions recommended under the third recommended condition (Reco_cond = 3) are difficult-to-process material conditions, but a judgment result 9 is output that can inform the operator that test processing (test cut) including surface treatment processing, including material surface modification processing, is recommended for the workpiece material W (including the fact that adjustment, change, or resetting of processing conditions is required due to the material being recommended for test cut).

[Processing system process flow]
FIG. 12 is a flowchart showing an example of a processing flow of the machining system.
As shown in Figure 12, before cutting the workpiece W, the NC device 60 acquires, for example, thickness information of the workpiece W on the machining table 11 (step S100) and transmits it to the judgment model storage section 51 of the workability judgment unit 50.

Next, based on the transmitted plate thickness information, the judgment model storage unit 51 selects first and second judgment models 5, 6 corresponding to the plate thickness of the workpiece W (step S101) and transmits them to the machinability calculation unit 52. In addition, in the laser processing unit 20, the workpiece W is irradiated with laser light LB under first and second irradiation conditions (step S102), and the emission spectrum is measured by the spectrometer 30 (step S103). Then, of the emission spectra measured by the spectrometer 30, first waveform information 3 calculated from the time series data of the first emission spectrum and second waveform information 4 calculated from the time series data of the second emission spectrum are transmitted to the machinability calculation unit 52.

The machinability calculation unit 52 inputs the transmitted first waveform information 3 and second waveform information 4 into the selected first judgment model 5 and second judgment model 6, respectively, calculates a quality score indicating the quality evaluation (surface quality evaluation and internal quality evaluation) of the surface condition and internal condition of the workpiece W (step S104), and generates judgment matrix information 110 (step S105).

The machinability determination unit 53 determines the machinability of the workpiece W based on the quality score and the determination matrix information 110 calculated and generated by the machinability calculation unit 52. In determining the machinability, it first determines whether the quality score of the surface condition of the workpiece W is 1.4 or more (S_score ≧ 1.4) and the quality score of the internal condition is 1.4 or more (B_score ≧ 1.4), i.e., whether the quality evaluations of both the surface condition and the internal condition are "○" (step S106).

If the quality score satisfies this judgment condition (T: True in step S106), the workpiece W can be classified into the area 111 of the first recommended condition (Reco_cond = 1) in the judgment matrix information 110, and the judgment result 9 in which the first recommended condition (Reco_cond = 1) is flagged (step S107) is sent to the control unit 61 of the NC device 60.

The control unit 61 of the NC device 60 notifies the operator based on the transmitted judgment result 9, for example, by displaying on the display 70, etc., information about the judgment result 9 including machinability evaluation such as the numerical value of the quality score of the surface condition (S_score) and the numerical value of the quality score of the internal condition (B-score) (step S108). At this time, for example, the standard conditions may be called up from the processing condition storage unit 62 in the background.

On the other hand, if the quality score does not satisfy the judgment condition in step S106 (F: False in step S106), it is judged whether the quality score of the surface condition of the workpiece W is less than 1.4 (S_score<1.4) and the quality score of the internal condition is less than 0.6 (B_score<0.6), i.e., whether the quality evaluation of the surface condition is "△" or "X" and the quality evaluation of the internal condition is "X" (step S109).

If the quality score does not satisfy this judgment condition (F: False in step S109), the workpiece W can be classified into the area 112 of the second recommended condition (Reco_cond = 2) in the judgment matrix information 110, and therefore a judgment result 9 in which the second recommended condition (Reco_cond = 2) is flagged (step S110) is sent to the control unit 61 of the NC device 60.

The control unit 61 of the NC device 60 notifies the operator based on the transmitted judgment result 9, for example, by displaying on the display 70, etc., information about the judgment result 9 including a machinability evaluation such as the numerical value of the quality score of the surface condition (S_score) and the numerical value of the quality score of the internal condition (B-score) and the like (step S112), that the workpiece W can be machined under machining conditions suitable for difficult-to-machine materials (difficult-to-machine material conditions), a recommendation to change to the difficult-to-machine material conditions, etc. At this time, for example, the difficult-to-machine material conditions may be called up from the machining condition storage unit 62 in the background.

If the quality score satisfies the judgment condition in step S109 (T: True in step S109), this means that the quality evaluation of the surface condition of the workpiece W is "△" or "X" and the quality evaluation of the internal condition is "X", so the workpiece W can be classified into the area 113 of the third recommended condition (Reco_cond = 3) in the judgment matrix information 110. Therefore, the judgment result 9 in which the third recommended condition (Reco_cond = 3) is flagged (step S113) is sent to the control unit 61 of the NC device 60.

The control unit 61 of the NC device 60 notifies the operator based on the transmitted judgment result 9, for example, by displaying on the display 70 or the like information about the judgment result 9 including a machinability evaluation such as the numerical value of the quality score of the surface condition (S_score) and the numerical value of the quality score of the internal condition (B-score) and the like, that the workpiece W is a difficult-to-machine material (recommended test cut material) that is difficult to machine even due to the difficult-to-machine material conditions, that the implementation of the above-mentioned test machining is recommended, and so on (step S114). At this time, for example, the difficult-to-machine material conditions may be called up from the machining condition storage unit 62 in the background.

Even if the workpiece W is determined to be difficult to cut, the control unit 61 may adjust the parameters of each adjustable item included in the difficult-to-cut material conditions, such as the laser output, processing speed, and focal position (focus position), to improve the processing quality or even make cutting possible. For this reason, after step S114, the control unit 61 performs test processing based on the difficult-to-cut material conditions, based on the input information received by the operator via the input device 211, such as a touch panel (step S115).

Then, the condition of the cut surface of the workpiece W from the test machining is checked, and it is determined whether the machining quality is OK (step S116). If it is determined that the machining quality is not OK (F: False in step S116), the operator inputs, for example, to adjust each parameter item of the machining conditions, focusing on the conditions for difficult-to-machine materials (step S117), and the test machining is carried out again (step S115), and the subsequent processes are repeated.

If the processing quality is OK (T: True in step S116) and information on the judgment result 9 is displayed (steps S108, S112), processing conditions that match the material state (surface state and internal state) of the workpiece W whose workability has been judged are set in the laser processing device 10, for example, based on an operation input by an operator (step S118), product processing of the workpiece W is performed based on the set processing conditions (step S119), and the series of processes according to this flowchart is terminated.

In this way, when the operator processes a product, the processing system 100 can determine the processability based on the processing conditions of the workpiece W before cutting, so that the processing conditions can be set according to the material state of the workpiece W, and the operator can easily grasp the judgment result 9 of whether cutting should be performed under the processing conditions set in advance or whether the processing conditions should be changed to those suitable for the workpiece. Note that the setting process of the above step S118 can be manually set by the operator through input operation, or the control unit 61 can automatically set according to the judgment result 9.

[Example]
Fig. 13 is a diagram showing a result table including the prediction results of the material state by the first and second judgment models, the judgment results of the workability, and the actual processing verification results. Fig. 14 is an explanatory diagram showing the quality evaluation criteria of the processing quality.

In order to verify the judgment of workability in the processing system 100 and the workability judgment system 90, the applicant prepared multiple samples for each of 27 types of processed materials W, each having a plate thickness of, for example, 19 mm, and differing in material quality, production lot, and production process (electric furnace/blast furnace). Then, for each sample, a first judgment model 5 was created by machine learning based on the first waveform information 1 and the surface quality evaluation result 109a, and a second judgment model 6 was created by machine learning based on the second waveform information 2 and the internal quality evaluation result 109b.

Using the machinability judgment unit 50 having the first and second judgment models 5, 6 created in this way, 17 types of workpieces W (t19 mm: 12 types, t22 mm: 5 types) that were not used for machine learning were irradiated with laser light LB under the first and second irradiation conditions, the first and second waveform information 3, 4 were measured to predict the material state, judgment matrix information 110 was created, and machinability was judged. Note that the first and second waveform information 3, 4 were measured at five different points for one workpiece W, and quality evaluations were performed on the surface state and internal state at each measurement point, and quality scores were calculated and the average value was calculated and used as the representative value.

In the result table 120 shown in FIG. 13, the prediction results 121 show the predicted results of the quality evaluation ("○", "△" and "×") of the surface condition and internal condition of each workpiece W of material No. (No: Number) 1 to 17, and the judgment results 122 show the judgment results of the workability (good cutting (standard material), difficult-to-cut material, difficult-to-cut material for which test cutting is recommended (test cut recommended material)) for each workpiece W based on the prediction results 121. In addition, the actual processing verification results 123 show the verification results ("○", "△" and "×") of the standard conditions and difficult-to-cut material conditions when each workpiece W was actually processed under each of the processing conditions 110a and 110b shown in FIG. 9.

The verification results shown by "○", "△" and "×" for the standard conditions and the difficult-to-machine material conditions in the actual machining verification results 123 are the results of evaluations compiled based on the quality evaluations performed by the operator at the various focal positions based on the quality evaluation criteria 114 for machining quality as shown in FIG. 14, in order to take into account the margin of focus position. The quality evaluation criteria 114 are used to evaluate the cutting quality at each focal position by changing the focal position several times. The quality evaluation criteria 114 are used to evaluate the cutting quality at each focal position. The quality evaluation criteria include, for example, the height of the dross, which is the molten metal adhering to the back surface of the machining part, the depth of the notches, which is the roughness that suddenly increases on the cutting surface, and the number of notches. The margin of focus position refers to the margin of deviation of the focal position in the machining conditions with respect to the cutting quality when the cutting focus is changed from the standard value to a value away from the condenser lens 24c and/or the opposite side, and also includes the margin of adaptability of the machining conditions to the material.

In other words, as shown in FIG. 14, a quality rating of "◎" indicates a high-quality cut, and a quality rating of "○" indicates a good cut. A quality rating of "○´" indicates a cut of slightly poor quality, and a quality rating of "○△" indicates a cut of poor quality. Furthermore, a quality rating of "△" indicates a cut that is capable of being severed, and a quality rating of "×" indicates that cutting is not possible. Note that the classification of quality ratings is not limited to these six evaluation ranges, and may be classified to allow for more detailed evaluation.

If the quality evaluation has two or more "◎"s or three or more "○"s, it is judged as "Good cuttable: ○". If the condition for "Good cuttable: ○" is not met and there are two or more "○´"s or higher in the quality evaluation, it is judged as "Cuttable: △". Furthermore, if there are less than two "○´"s or higher in the quality evaluation, it is judged as "Difficult to cut: ×".

FIG. 15 is a diagram showing the results of classifying the actual machining results in the determination matrix information.
As shown in Fig. 15, among the workpieces W No. 1 to No. 17, the workpieces W (material No.: No. 3, No. 8, No. 10, No. 13, No. 14) for which the quality evaluations of both the surface condition and the internal condition of the material were judged to be "○" were judged to have a machinability judgment result of "good cutting (standard material)" and were classified into the region 111 of the first recommended condition (Reco_cond = 1). These workpieces W were actually verified to have a "○" verification result, which indicates that good cutting can be obtained under the standard condition, and it was confirmed that the machinability was correctly judged.

Note that for the workpiece W with material number No. 6 classified in region 112 of the second recommended conditions (Reco_cond = 2), the surface condition was rated as "○" and the internal condition was rated as "×", so the machinability was judged to be "difficult to cut". However, the verification result was actually "○", which means that good cutting was obtained under standard conditions. However, since the verification result was "○" even under the difficult to cut conditions, it can be determined that the machinability was correctly judged.

The workpieces W (material numbers: No2, No4, No7, No9, No12, No16, No17) whose surface condition was evaluated as "△", whose internal condition was evaluated as "×", and whose machinability was evaluated as "recommended for test cutting" are classified into region 113 of the third recommended condition (Reco_cond = 3). Of these workpieces W, the workpieces W with material numbers No2, No4, No7, and No16 were actually evaluated as "×" for difficulty in cutting under standard conditions, but were verified as "○" for satisfactory cutting under difficult-to-cut conditions. On the other hand, the workpieces W (materials No. 9, No. 12, and No. 17) other than material No. 2, No. 4, No. 7, and No. 16 were actually rated "X" for difficulty in cutting under standard conditions, and were rated "X" or "△" for poor cutting even under difficult-to-cut conditions. In other words, it can be confirmed that the workpieces W classified in this region 113 may require test cuts with new adjustments to the processing conditions, and it can be determined that the workability has been correctly assessed.

The workpieces W other than those mentioned above (material numbers: No1, No5, No11, No15) were judged to be "difficult to cut" and were classified into region 112 of the second recommended conditions (Reco_cond = 2). These workpieces W were actually rated "X" for difficult cutting under standard conditions or "△" for poor cutting, but were verified to be "○" for good cutting under difficult-to-cut conditions, confirming that the workability was judged correctly.

Fig. 16 is a diagram showing an example of details of an actual machining verification result of a workpiece determined to be a standard material. Fig. 17 is a diagram showing an example of details of an actual machining verification result of a workpiece determined to be a difficult-to-machine material. Fig. 18 is a diagram showing an example of details of an actual machining verification result of a workpiece determined to be a difficult-to-machine material (recommended material for test cut).
16, for example, the workpiece W, material number 13, whose workability has been determined to be the standard material, has a cut surface quality evaluation of "◎" in the five-level range of focal positions from 2.5 mm to 6.5 mm under standard conditions of machining conditions 110b with a plate thickness of 22 mm. Therefore, it was confirmed that the workability determination result 9 using the first and second determination models 5 and 6 and the determination matrix information 110 was accurate.

Also, as shown in FIG. 17, for example, for the workpiece W with material No. 15, whose workability was judged to be difficult to work, under standard conditions of processing conditions 110a with a plate thickness of 19 mm, the quality evaluation of the cut surface was all "○△" in the five-level range of focal positions from 2.5 mm to 6.5 mm. In contrast, under the difficult-to-work material conditions, the quality evaluation of the cut surface was "◎" or "○" in the five-level range of focal positions from 3.0 mm to 7.0 mm. This shows that the use of difficult-to-work material conditions improves the processing quality, and it was confirmed that the workability judgment result 9 using the first and second judgment models 5 and 6 and the judgment matrix information 110 was accurate.

Also, as shown in FIG. 18, for example, for the workpiece W with material No. 17, whose workability was judged to be difficult to work (recommended material for test cutting), under standard conditions of processing conditions 110a with a plate thickness of 19 mm, the quality evaluation of the cut surface was all "△" in the five-level range of focal positions from 2.5 mm to 6.5 mm. Even under the difficult-to-work material conditions, the quality evaluation of the cut surface was "○", "○´", or "○△" in the five-level range of focal positions from 3.0 mm to 7.0 mm. This indicates that high-quality cutting (good cutting) is not achieved even when using difficult-to-work material conditions, and highlights the possibility that test processing may be required. Therefore, it was confirmed that the workability judgment result 9 using the first and second judgment models 5 and 6 and the judgment matrix information 110 was accurate.

As described above, the processing system 100 and the machinability judgment system 90 of the first embodiment judge the machinability based on the processing conditions of the workpiece W before cutting, and easily determine the measures to be taken in accordance with the contents of the judgment result 9 (the material state of the workpiece W and the recommended processing conditions) (for example, whether cutting should be performed under the recommended processing conditions, whether the processing conditions should be changed to those suitable for the workpiece W, whether the recommended processing conditions should be adjusted, whether test processing should be performed, etc.), and the processing quality can be quickly improved, making it possible to reduce processing defects.

Second Embodiment
Next, a second embodiment will be described. The processing system and the machinability judgment system according to the second embodiment are configured to add a configuration for measuring the infrared intensity of the radiation light generated when the laser light LB is irradiated by a radiation thermometer (infrared sensor) to the processing system and the machinability judgment system according to the first embodiment, which measures the emission spectrum generated when the laser light LB is irradiated by the spectrometer 30 described above, and to combine the judgment result of the machinability of the workpiece W by the spectrometer 30 (first judgment result) and the judgment result of the machinability of the workpiece W by the radiation thermometer (second judgment result) to output a judgment result of the machinability of the workpiece W when it is cut under preset processing conditions in the processing step (third judgment result).

For this reason, all of the basic configuration, schematic configuration, functional configuration, hardware configuration, etc. described in the first embodiment can be included, and all of the effects can be realized. Therefore, in the following, unless otherwise specified, the same reference numerals are used for parts that overlap with the first embodiment, and descriptions that overlap with parts that have already been described will be omitted.

As described above, oxygen cutting of a workpiece W of mild steel is affected by the surface condition (state of the oxide film) and internal condition of the workpiece W. Even with the same steel type and plate thickness, the cutting quality differs under the same laser cutting processing conditions due to individual differences such as differences in manufacturer and production lot. The inventors have constructed the processing system and machinability assessment system of the first embodiment in order to find the material property factors that have a strong influence on the cutting quality and to measure and analyze them.

In other words, the oxide film on the material surface (raw material surface) of the workpiece W affects the laser absorption rate and oxidation reaction at the surface. The laser absorption rate differs depending on the type of oxide film; for example, black magnetite has a high laser absorption rate and oxide film adhesion, and the cutting quality is stable. On the other hand, hematite has a low laser absorption rate and poor oxide film adhesion. For this reason, the oxide film is easily peeled off during laser cutting, and notches are easily created starting from the peeled part.

In addition, areas where the oxide film is thin or peeled off are prone to oxidation reactions on the material surface, making excessive combustion more likely to occur on the material surface and resulting in a rough cut surface. From this perspective, the inventors have discovered that the adhesion of the oxide film can be determined by the emission spectrum generated immediately after laser irradiation when the surface of the workpiece W is melted with laser light LB.

For example, the difference between a workpiece W having a highly adhesive oxide film and a workpiece W having a poorly adhesive oxide film, or a thin or peeled oxide film, can be determined by the change over time in the emission spectrum that occurs until the oxide film on the surface (material surface) of the workpiece W disappears.

In addition, to evaluate the distribution of the oxide film, it is possible to irradiate the material surface with laser light LB without melting it, and evaluate the adhesion of the oxide film from the thermal stress generated on the material surface. When the oxide film peels off due to thermal stress, a flash is generated, which can be detected by infrared spectroscopy. In addition, the type and roughness (unevenness) of the oxide film on the material surface cause temperature fluctuations due to differences in laser absorption rate, which can be captured as fluctuations in the infrared spectrum.

Furthermore, it has been found that the internal characteristics of the workpiece W are affected by the stability of the melting behavior due to the internal components of elements other than iron contained in the material, and the ease of heat transfer (thermal conductivity). Regarding the influence of elements other than iron, those with a particularly high carbon content are prone to component segregation, resulting in an uneven carbon concentration distribution and an uneven melting temperature, which makes the cut surface prone to becoming rough due to uneven melting. Next, manganese is also an element that increases component segregation, and uneven melting can easily make the melting behavior unstable.

Therefore, the stability of the melting behavior affects the time-dependent change in the emission spectrum that occurs during melting. For example, a high carbon content leads to greater component segregation within the steel sheet, resulting in non-uniformity between the low-melting point and high-melting point areas, and greater fluctuation in the time-dependent change in the emission spectrum.
In addition, as the material contains more elements other than iron, its thermal conductivity decreases, so it is more likely to become hot at the point of laser cutting, which in turn makes it more likely to melt excessively and results in rough cut surfaces.

Taking the above points into consideration, the processing system and workability judgment system of the first embodiment have been configured. In order to further improve the judgment accuracy, the processing system and workability judgment system of the second embodiment have the following configuration.
The thermal conductivity of the internal components can be evaluated as the ease of heat transfer from the magnitude of the temperature rise by repeatedly heating and cooling under laser irradiation conditions that do not melt the material. Note that, since it is not possible to measure temperatures below about 600°C in the visible light region, the infrared intensity in the wavelength band above 1600 nm was measured to measure the temperature on the low-temperature side of the temperature rise.

Therefore, in the processing system and machinability judgment system of the second embodiment, the judgment of the machinability of the workpiece W with visible light is performed using a judgment model that utilizes machine learning as described above, and the judgment of the machinability of the workpiece W with infrared light is performed by comparing information (feature information) in which the temporal or positional change in infrared intensity is used as a feature value with a threshold value.

[Basic configuration of the machining system]
FIG. 19 is an explanatory diagram illustrating a basic configuration of a processing system according to the second embodiment of the present invention.
As shown in Fig. 19, the processing system 100A according to the second embodiment includes a laser processing device 10A capable of executing a processing step of cutting the workpiece W by irradiating the workpiece W with the laser light LB under processing irradiation conditions, and a processability judgment step of irradiating the workpiece W with the laser light LB under a first judgment irradiation condition (the "judgment irradiation condition" in the first embodiment) under which the workpiece W is melted but not penetrated, and under a second judgment irradiation condition under which the melting point of the material of the workpiece W is not exceeded, to judge the processability of the workpiece W. The processing system 100A also includes a laser processing unit 20A (measurement device) including a spectrometer 30 (first measurement unit) that measures an emission spectrum generated when the workpiece W is irradiated with the laser light LB under the first judgment irradiation condition, and a radiation thermometer 30A (second measurement unit) that measures the infrared intensity of the radiation generated when the workpiece W is irradiated with the laser light LB under the second judgment irradiation condition. The processing system 100A is provided with a judgment device (PC 50A and NC device 60) including a PC (Personal Computer) 50A (first judgment unit) that judges the workability of the workpiece W based on time series data of the emission spectrum measured by the spectroscope 30 of the laser processing unit 20A, an NC device 60 (second judgment unit) that judges the workability of the workpiece W based on time series data of the infrared intensity measured by the radiation thermometer 30A of the laser processing unit 20A, and an NC device 60 (third judgment unit) that judges the workability of the workpiece W when cut and processed under preset processing conditions in the processing step based on a combination of the first judgment result judged by the PC 50A and the second judgment result judged by the NC device 60 and outputs a third judgment result. The judgment device has a first judgment model 5 (FIG. 9) and a second judgment model 6 (FIG. 9). The PC 50A of the judgment device extracts the first waveform information 3 (FIG. 9) in the first time domain and the second waveform information 4 (FIG. 9) in the second time domain from the time series data of the emission spectrum measured by the spectroscope 30 of the laser processing unit 20A. The PC 50A of the judgment device inputs the extracted first waveform information 3 as estimation data into the first judgment model 5 to obtain a surface quality evaluation of the workpiece W. The PC 50A of the judgment device also inputs the extracted second waveform information 4 as estimation data into the second judgment model 6 to obtain an internal quality evaluation of the workpiece W. The PC 50A of the judgment device then outputs the first judgment result that judges the workability of the workpiece W based on a combination of the obtained surface quality evaluation and internal quality evaluation of the workpiece W. In addition, the NC device 60 of the judgment device extracts feature information indicating temporal or positional changes in the temperature of the workpiece W based on the time series data of infrared intensity measured by the radiation thermometer 30A of the laser processing unit 20A, and outputs a second judgment result in which the workability of the workpiece W is judged based on the extracted feature information and reference information for judging the workability of the workpiece W that has been registered in advance in a storage device or the like not shown.

The laser processing unit 20A of the laser processing device 10A differs from the laser processing unit 20 of the first embodiment in that it includes a radiation thermometer 30A and a dichroic mirror 24d. The radiation thermometer 30A is provided, for example, on the upper part of the laser processing head 22. The PC 50A may be included in the NC device 60, and the NC device 60 may be mounted so as to be included in the laser processing device 10A.

The radiation thermometer 30A captures the infrared radiation, including the infrared radiation emitted by thermal radiation, from the electromagnetic waves with a wide band of wavelengths emitted from the workpiece W heated by the laser light LB, and converts it into an electrical signal. The radiation thermometer 30A measures the infrared intensity in the wavelength band of 1600 nm or more. That is, the radiation thermometer 30A has, for example, a wavelength filter in the front stage that transmits only wavelengths in a specific band (1600 nm or more and 2500 nm or less), and uses, for example, a photodiode using InGaAs (indium gallium arsenide) as a photoelectric conversion element in the rear stage.

Dichroic mirror 24d is provided above folding mirror 24b. Dichroic mirror 24d transmits infrared light (infrared rays) of a specific wavelength that is transmitted through folding mirror 24b, out of the radiated light contained in the light traveling from the processing side of workpiece W toward folding mirror 24b, to radiation thermometer 30A, and reflects visible light to light receiving section 31 of spectrometer 30.

Here, the PC 50A and the NC device 60 of the machining system 100A will be described.
FIG. 20 is a schematic functional block diagram of the processing system.
20, like the machinability judgment unit 50 of the first embodiment, the PC 50A includes a judgment model storage unit 51, a machinability calculation unit 52, and a machinability judgment unit 53. The PC 50A includes a hardware configuration (CPU, RAM, ROM, etc.) similar to that of a general PC.

The machinability determination unit 53 of the PC 50A can output to the NC device 60 a first determination result that determines the machinability of the workpiece W (standard material that can be cut well, difficult-to-machine material) based on the determination matrix information generated by the machinability calculation unit 52. The functions and actions of the other units 51 to 53 are the same as those described in the first embodiment, so a description thereof will be omitted here.

In addition to controlling the laser processing unit 20A that performs the cutting process, the NC device 60 functionally comprises a calculation processing unit 63, a workability judgment unit 66, a comprehensive judgment unit 65, a control unit 61, and a processing condition storage unit 62. The calculation processing unit 63 is made up of a hardware or middleware calculation processing device that can perform specific calculation processes at high speed.

The calculation processing unit 63 converts the instantaneous voltage value of the output signal of the radiation thermometer 30A into infrared light intensity and can obtain time series data of this light intensity. The calculation processing unit 63 extracts feature information that represents the temporal transition of the temperature characteristics based on the obtained time series data of the light intensity. The calculation processing unit 63 also functions as, for example, a signal processing unit.

In other words, the output signal of the current output captured by the radiation thermometer 30A and photoelectrically converted is transmitted to the calculation processing unit 63. The output signal transferred as a current is converted into a voltage signal by a current-voltage conversion circuit (not shown). The output signal that has been converted into a voltage signal is then converted into a digital signal by the A/D converter 40, and this digital signal is input to the calculation processing unit 63.

The machinability determination unit 66 outputs a second determination result that determines the machinability of the workpiece W based on the processing conditions of the upcoming processing (the machinability of the workpiece W when cut under processing conditions previously set in the laser processing device 10A) based on the feature information extracted by the calculation processing unit 63 and reference information stored in a storage device (not shown), for example. The reference information is information that is selected and determined, for example, through experiments, and is registered in advance.

The overall judgment unit 65 uses the information obtained by the visible light measurement using the spectrometer 30 (such as a quality score of material characteristics) and the information obtained by the infrared measurement using the radiation thermometer 30A (such as a quality evaluation of material characteristics) based on a combination of the first judgment result from the workability judgment unit 53 of the PC 50A and the second judgment result from the workability judgment unit 66 to judge the workability of the workpiece W when it is cut and processed under the preset processing conditions in the processing step, and outputs an overall judgment result (third judgment result).

The control unit 61 controls the operation of the laser processing unit 20A (laser processing head 22, beam control unit 24, etc.) via various I/Fs (interfaces). The control unit 61 also outputs control instructions for various operations to the laser processing device 10A and the spectrometer 30, etc., and outputs display instructions for various information representing the judgment result, for example, to the display 70, based on the judgment result (third judgment result) of the overall judgment unit 65. Note that other functions and actions of the control unit 61, as well as the processing condition storage unit 62, are the same as in the first embodiment, and therefore will not be described here.

The display 70 can display a setting input screen for inputting various information such as the cutting conditions, a display screen for displaying various information such as the judgment results so that the operator can confirm them, and the like. The display 70 also functions as a notification unit that notifies the operator of the third judgment result determined by the overall judgment unit 65 of the NC device 60 so that the third judgment result can be confirmed. The display 70 can be configured with, for example, a touch panel that functions as an input unit. When the display 70 is equipped with a touch panel, the operator can input various information related to the workpiece W to the control unit 61 of the NC device 60, for example, by operating the display 70.

The machinability judgment system (PC 50A and NC device 60) according to the second embodiment has a first judgment model 5 and a second judgment model 6, and in a machinability judgment process for judging the machinability of the workpiece W, first waveform information 3 and second waveform information 4 extracted (calculated) from time series data of the emission spectrum generated when the laser light LB is irradiated to the workpiece W under first judgment irradiation conditions (first irradiation condition, second irradiation condition) that melt the workpiece W but do not penetrate the workpiece W are input into the first judgment model 5 and the second judgment model 6, respectively, and the machinability of the workpiece W is judged based on the first waveform information 3 and the second waveform information 4. The machinability judgment system (PC 50A and NC device 60) includes a PC 50A (first judgment unit) including a learning device 80, and in the machinability judgment process, the laser light LB is irradiated to the workpiece W under second judgment irradiation conditions (third irradiation condition, fourth irradiation condition) that do not exceed the melting point of the material of the workpiece W. the NC device 60 (second judgment unit) that judges the machinability of the workpiece W based on feature amount information extracted from time-series data on the infrared intensity of radiated light generated when the workpiece W is irradiated with infrared light under preset processing conditions in the processing step, and pre-registered reference information for judging the machinability of the workpiece W; the NC device 60 (third judgment unit) that judges the machinability of the workpiece W when cut under preset processing conditions in the processing step, based on a combination of the judgment results (first judgment result) of the surface condition and internal condition of the workpiece W judged by the PC 50A (first judgment unit) and the judgment results (second judgment result) of the surface texture and internal characteristics of the workpiece W judged by the NC device 60 (second judgment unit); and the PC 50A (learning device) that creates a first judgment model 5 and a second judgment model 6. In the judgment device, the PC 50A (first judgment unit) extracts the first waveform information 3 in the first time domain and the second waveform information 4 in the second time domain from the time series data of the emission spectrum, inputs the extracted first waveform information 3 as estimation data to a first judgment model 5 to obtain a surface quality evaluation of the workpiece W, inputs the extracted second waveform information 4 as estimation data to a second judgment model 6 to obtain an internal quality evaluation of the workpiece W, and outputs a first judgment result that judges the workability of the workpiece W based on a combination of the obtained surface quality evaluation and internal quality evaluation of the workpiece W. Also, the NC device 60 (second judgment unit) extracts feature amount information (first feature amount information, second feature amount information) indicating a temporal or positional change in temperature of the workpiece W based on the time series data of the infrared intensity, and outputs a second judgment result that judges the workability of the workpiece W based on the extracted feature amount information and reference information (first judgment criterion, second judgment criterion). The PC 50A (learning device 80) inputs the first waveform information 1 and the surface quality evaluation result 109a indicating the surface quality of the workpiece W as first teacher data, performs machine learning to create a first judgment model 5, and inputs the second waveform information 2 and the internal quality evaluation result 109b indicating the internal quality of the workpiece W as second teacher data, performs machine learning to create a second judgment model 6.

In other words, the machinability determination system of the second embodiment can be configured by the above-mentioned machining system 100A equipped with a PC 50A including the machinability determination system 90 (machinability determination unit 50 and learning device 80) of the first embodiment shown in FIG. 9, and an NC device 60.

[Processability assessment method]
Next, a method for determining workability by the machining system 100A will be described.
As described above, the judgment irradiation conditions (first irradiation conditions, second irradiation conditions) of the laser light LB when the emission spectrum is measured by the spectroscope 30 to judge the surface condition and internal condition of the workpiece W are described below as the first judgment irradiation conditions, but the overall description of the judgment of the workpiece W by the spectroscope 30 on the PC 50A side is the same as that of the first embodiment. Therefore, while it is assumed that the judgment result of the workability of the workpiece W by the spectroscope 30 (first judgment result) is also used in the overall judgment, the description will first focus on measuring the infrared spectrum (infrared intensity) by the radiation thermometer 30A to judge the material properties of the surface condition and internal characteristics of the workpiece W.

In the second embodiment, in the process of determining the machinability, the laser processing device 10A spot-irradiates the workpiece W with the laser light LB under the first determination irradiation condition, and under the second determination irradiation condition, irradiates the surface of the workpiece W with the laser light LB while moving the irradiation position, and repeatedly irradiates the interior of the workpiece W with the laser light LB.

Here, as described above, the first judgment irradiation condition includes the first irradiation condition and the second irradiation condition. When the laser processing apparatus 10A irradiates the laser light LB under the first judgment irradiation condition in the machinability judgment process, the time from the start of irradiation of the laser light LB to the end of irradiation is divided into a first stage and a second stage, and as the first judgment irradiation condition, the laser light LB is irradiated to the workpiece W under the first irradiation condition 101 (FIG. 3) in the first stage, and the laser light LB is irradiated to the workpiece W under the second irradiation condition 102 (FIG. 3) in the second stage. The spectroscope 30 of the laser processing unit 20A measures a first emission spectrum generated when the laser light LB is irradiated to the workpiece W under the first irradiation condition, and a second emission spectrum generated when the laser light LB is irradiated to the workpiece W under the second irradiation condition. The PC 50A of the judgment device extracts first waveform information 1, 3 (Figure 9) in the first time domain from the time series data of the first emission spectrum measured by the spectrometer 30 of the laser processing unit 20A, and extracts second waveform information 2, 4 (Figure 9) in the second time domain from the time series data of the second emission spectrum measured by the spectrometer 30 of the laser processing unit 20A.

On the other hand, when measuring with the radiation thermometer 30A, the laser light LB is irradiated under a second judgment irradiation condition that does not exceed the melting point of the material of the workpiece W. In other words, the laser is irradiated within a range that does not melt the material of the workpiece W, and the material characteristics are judged from the temporal or positional changes in the infrared spectrum. The inventors cut (laser cut) various workpieces W with different internal components (internal components) and surface properties (surface properties) under standard processing conditions of the NC device 60, and observed the cut surfaces of the workpieces W. As a result, it was found that the workability (cuttability) of the workpieces W is greatly affected by the surface properties and internal component properties (internal properties) of the workpieces W.

In other words, the surface quality of the workpiece W specifically refers to the state of the oxide film on the surface layer of the workpiece W and its adhesion to the material surface (hereinafter referred to as "state and adhesion of the oxide film"). Here, the surface layer refers to the part corresponding to the oxide film layer formed on the material surface, and if there is no oxide film on the material surface, this surface layer part will also not exist. It has been found that the workability of the workpiece W is primarily affected by the state and adhesion of the oxide film on the material surface of the workpiece W. Hereinafter, this factor will be referred to as "surface-related". When the surface-related factor is not good, the cut surface has large irregularities in the striations on the laser incident surface side. In particular, in the parts where the oxide film has peeled off, the striations are irregular and notches are likely to occur.

The internal characteristics of the workpiece W specifically refer to the magnitude of the thermal conductivity due to the internal components of the workpiece W. Secondly, it has been found that the workability of the workpiece W is greatly affected by the magnitude of the thermal conductivity due to the internal components of the workpiece W. Hereinafter, this factor will be referred to as "internal causes." For example, in the case of a thick plate of mild steel, oxygen is used as the assist gas, and melt cutting is performed using the heat of the laser light and the heat of the oxidation reaction as the heat source. At that time, if there are many elements with low thermal conductivity, the cut surface is likely to be rough. This is thought to be due to the fact that if there are many elements with low thermal conductivity, heat is difficult to diffuse, which tends to promote a rise in temperature on the cut surface, and low thermal conductivity makes it easier for excessive combustion to occur.

Therefore, we decided to evaluate the thermal conductivity in the solid state and the distribution of the oxide film on the surface of the workpiece W as the material properties. As the judgment irradiation condition (second judgment irradiation condition) for the laser light LB, we decided to measure the wavelength spectrum intensity (infrared intensity) in the infrared region of wavelengths above 1600 nm, because the visible light region is weak on the low temperature side of 600°C or less, which is the temperature range that does not melt the material (does not exceed the melting point).

First, the first study was conducted to investigate the influence of internal factors. In the first study, in order to eliminate the influence of the surface condition of the workpiece W, the oxide film on the surface of the workpiece W was removed to make the surface condition uniform, and the laser cutting was performed in that state. The oxide film was removed by polishing the surface so that the average surface roughness was 0.8 or less. The laser cutting was performed under the following processing conditions: laser output 3 (kW), processing speed 630 (mm/min), and assist gas oxygen ( _O2 ).

Thermal conductivity was measured using the laser flash method. For example, rectangular test pieces 10 mm square and 4 mm thick were prepared for samples a to c, graphite was applied to the surface of the material, and laser light LB was flash-irradiated from one side (the surface side of the material). The thermal diffusivity was measured from the temperature rise on the opposite side (the back side of the material) to determine the thermal conductivity.

Figure 21 is a result table showing the results of investigations into the thermal conductivity of each material of the processed material, the laser cut surface, and the roughness of the cut surface. Figure 22 is a graph showing the relationship between the average roughness of the cut surface and the thermal conductivity. Note that the quality evaluations of each item in the result table 300, "○", "△", and "×", are defined by actually observing and evaluating the obtained numerical values and the cut surfaces, etc., from the viewpoint of quality, and correspond to "good", "fair (poor)", and "unacceptable", respectively.

As shown in FIG. 21, a result table 300 shows results 301, 302, and 303 of samples a, b, and c.
Sample a had a thermal conductivity of 56.7 W/(m·K), and the cut surface was rated as "△" because the streaks were partially deep, the pitch was large, and the cut surface was somewhat rough. It can be seen that the roughness of the cut surface (streaking) when the laser beam LB was incident to a depth of 1 mm showed some variation in depth.

Sample b had a thermal conductivity of 60.3 W/(m·K), few notches on the cut surface, and the shallowest streak depth, so it was rated as "○". The cut surface roughness (streak change) when the laser light LB was incident to a depth of 1 mm showed little variation in depth and was the shallowest of all the samples, so it can be confirmed that the results were generally in line with the cut surface evaluation.

Sample c had a thermal conductivity of 49.5 W/(m·K) and the cut surface had the most notches of all the samples, so it was rated as "X". The cut surface roughness (streak change) when the laser light LB was incident to a depth of 1 mm was deep with a lot of variation in depth, so it can be confirmed that the results were generally in line with the cut surface evaluation.

Then, the average roughness of the cut surfaces of Sample a, Sample b, and Sample c was calculated, and the calculation results were plotted on a graph 304, as shown in Figure 22, in which the vertical axis represents the average roughness of the cut surface (μm) and the horizontal axis represents the magnitude of thermal conductivity (W/(m·K)), to obtain a regression line 305. As is clear from Figure 22, it was found that the roughness of the cut surface increases in the order of the magnitude of the thermal conductivity. Therefore, it was proven that there is a correlation between the cuttability of the material of the workpiece W and the magnitude of the thermal conductivity.

In order to evaluate the above-mentioned thermal conductivity before laser cutting, a second study was conducted to classify the internal material and thermal conductivity of the workpiece W. In the second study, the difference in thermal conductivity was investigated by repeatedly heating and cooling the inside of the workpiece W with laser light LB in the laser output range that does not melt the material of the workpiece W (the laser output range under the second judgment irradiation conditions).

FIG. 23 is a graph showing the relationship between temperature and time when repeatedly heating and cooling a workpiece, and the relationship between the temperature reached by cooling and time.
Fig. 23(a) shows a temperature waveform 310 obtained by measuring sample b with radiation thermometer 30A and converting the time series data of infrared intensity output from radiation thermometer 30A into time series data of temperature. In Fig. 23(a), the vertical axis represents temperature (°C) and the horizontal axis represents time (milliseconds: ms).

In the second investigation, the second judgment irradiation conditions were set to 425 W laser output, 10 Hz frequency, and 50% duty for the laser light LB irradiated to the inside of the workpiece W having a plate thickness of 19 mm, and heating was repeated 10 times per second. Figure 23(b) shows a temperature waveform 311 in which two periods of the temperature waveform 310 in Figure 23(a) are shown enlarged on the time axis in ms units.

As shown in Figures 23(a) and (b), during the first 10 seconds of irradiation with the laser beam LB, the temperature waveforms 310, 311 of the workpiece W fluctuate finely between approximately 300°C and approximately 750°C (repeated generation and release of heat) over time due to multiple cycles of laser irradiation.

23(c) is a waveform diagram showing the lower envelope of the temperature waveform 310 by repeated heating and cooling shown in FIG. 23(a). This waveform diagram is obtained by extracting the temporal transition of the cooling temperature for each cycle of the waveform diagram in FIG. 23(a) as feature information (second feature information) 312. This feature information 312 indicates the degree of temperature rise caused by thermal conduction of the workpiece W, and can be used for subsequent judgment evaluations. Based on this feature information 312, the cooling temperature reached 5 seconds after the start of laser irradiation was set as the judgment criterion (second judgment criterion) for judging the internal characteristics of the material of the workpiece W. That is, if the cooling temperature reached 5 seconds after the start of laser irradiation is 550°C or higher, the thermal conduction is evaluated as "x", if it is 470°C or higher but less than 550°C, the thermal conduction is evaluated as "△", and if it is less than 470°C, the thermal conduction is evaluated as "○".

As shown by the arrow in Figure 23 (c), for sample b, the cooling temperature reached 5 seconds after the start of laser irradiation was 421°C, so the thermal conductivity was rated as "○". This indicates that the heat generated at the processing point is easily diffused and the material is not prone to temperature increases. It can also be seen from result 302 for sample b shown in result table 300 in Figure 21 that the thermal conductivity is relatively large, which was found to match the temperature analysis results of repeated heating and cooling by laser irradiation.

FIG. 24 is a graph showing the relationship between temperature and time when repeatedly heating and cooling a workpiece, and the relationship between the temperature reached by cooling and time.
24A shows a temperature waveform 315 obtained by measuring sample c with the radiation thermometer 30A and converting the time series data of infrared intensity output from the radiation thermometer 30A into time series data of temperature. The processing conditions and the second judgment irradiation conditions are the same as those of sample b. The temperature waveform 316 and feature amount information (second feature amount information) 317 are also the same as those of FIG. 23.

As shown by the arrow in Figure 24 (c), for sample c, the cooling temperature reached 5 seconds after the start of laser irradiation was 551°C, so the thermal conductivity was rated as "x". This indicates that the heat generated at the processing point does not easily escape, and that the material has low thermal conductivity. It can be seen from result 303 of sample c shown in result table 300 in Figure 21 that the thermal conductivity is relatively low, which is consistent with the temperature analysis results of repeated heating and cooling by laser irradiation.

Next, a third investigation was carried out to investigate surface causes. In the third investigation, we first investigated the relationship between the state of the oxide film on the surface of the workpiece W and the irregularities of the streaks.
For example, the type A workpiece W shown in FIG. 25 corresponds to a type in which the oxide film is peeled off (in a spotted form) in the raw material state, and the peeling rate of the oxide film is relatively large. The type B workpiece W shown in FIG. 27 corresponds to a type in which the oxide film is not peeled off (the oxide film is attached) in the raw material state, but the oxide film is peeled off by laser irradiation. In both types A and B workpiece W, when the oxide film on the surface of the material peels off, the laser absorption rate and the ease of oxidation reaction on the surface change. For this reason, the cut surface has uneven streaks in the area where the oxide film has peeled off. Therefore, it is considered that the irregularity of the streaks on the cut surface is strongly correlated with the adhesion of the oxide film on the surface of the material. This is also clear from the fact that in the images of the cut surfaces of the type A and type B workpiece W examined as above, the streaks were deep and notched, starting from the peeled part of the oxide film.

In light of the above, it was found that the workpiece W can be classified into three types in terms of surface properties: Type A and Type B, as well as Type C, shown in Figure 29, in which the oxide film on the material surface does not peel off at all even when irradiated with a laser. In the third investigation, in order to investigate the adhesion (distribution state of the oxide film) of the oxide film on the material surface of the workpiece W, for example, as the second judgment irradiation condition, the laser output of the laser light LB irradiated on the material surface of the workpiece W with a plate thickness of 19 mm was set to 125 W, the speed to 240 mm/min, the frequency to 100 Hz, and the duty to 100%, and laser irradiation (laser scanning) was performed with a scanning distance of 80 mm. By investigating the change in temperature over time at that time, it is possible to investigate the positional adhesion state of the oxide film. Thermal stress is generated on the surface of the workpiece W by heating and cooling due to laser scanning, but if the adhesion of the oxide film is weak, the oxide film will peel off, which appears as a temperature change.

Fig. 25 is a diagram for explaining the results of an investigation into the surface condition of the workpiece of type A. Fig. 26 is a diagram showing a cut surface image and surface roughness when the workpiece of Fig. 25 is cut.
As shown in Fig. 25(a) and Fig. 25(b), the type A workpiece W has some areas where the oxide film is peeled off in the raw state, as can be seen from the surface image 330 and the enlarged image 330a of the material surface. Fig. 25(c) includes a waveform diagram showing the positional change in the detected temperature when the type A workpiece W is subjected to laser scanning. This waveform diagram is obtained by extracting the positional change in temperature as feature amount information (first feature amount information) 331. Since the type A workpiece W has some areas where the oxide film is peeled off in the raw state, when the type A workpiece W is subjected to laser scanning as shown in Fig. 25(c), the feature amount information (first feature amount information) 331 showing the positional change in temperature (°C) shown on the vertical axis at the distance (mm) shown on the horizontal axis of the waveform diagram is expressed as shown. The feature amount information (first feature amount information) 331 represents a temperature pattern measured by the radiation thermometer 30A with the infrared intensity of 1.95 μm to 2.5 μm as emissivity 1 (the same applies below). A surface image 332 below the waveform diagram represents the surface condition after the laser scan corresponding to the waveform diagram.

In other words, the temperature is low in areas on the material surface where the oxide film has peeled off before the laser light LB is irradiated. In areas where an oxide film is present (attached) and the adhesion of the oxide film is weak, the oxide film peels off when irradiated with the laser light LB. In areas where the oxide film has peeled off due to such laser irradiation (whitish areas in surface image 332), the oxide film emits light and generates heat, causing the temperature to rise. As a result, the feature information (first feature information) 331 shows many peaks on the high temperature side, and there is a lot of waveform distortion on both the high temperature and low temperature sides.

The cut surface image of type A workpiece W was as shown in image 333 in Figure 26(a), and the surface roughness at a depth of 1 mm from the surface (line indicated by I) and at a depth of 2 mm (line indicated by II) was as shown in graph 334 in Figure 26(b). For type A workpiece W, the surface roughness was unstable and varied overall at both a depth of 1 mm and a depth of 2 mm from the surface, with many irregularities in the striations on the cut surface.

Fig. 27 is a diagram for explaining the results of investigating the surface condition of the workpiece of type B. Fig. 28 is a diagram showing a cut surface image and surface roughness when the workpiece of Fig. 27 is cut.
As shown in Fig. 27(a) and Fig. 27(b), the type B workpiece W has no peeling of the oxide film in the raw state (the oxide film is attached) as can be seen from the surface image 340 and the enlarged image 340a of the material surface. Fig. 27(c) includes a waveform diagram showing the positional change in the detected temperature when the type B workpiece W is subjected to laser scanning. This waveform diagram is obtained by extracting the positional change in temperature as feature amount information (first feature amount information) 341. Although there is no peeling of the oxide film in the raw state, as shown in Fig. 27(c), when the type B workpiece W is subjected to laser scanning, the feature amount information (first feature amount information) 341 showing the positional change in temperature (°C) shown on the vertical axis at the distance (mm) shown on the horizontal axis of the waveform diagram is shown as shown, and the surface state after laser scanning becomes as shown in the surface image 342 below the waveform diagram.

In other words, since there is a tendency for the adhesion of the oxide film on the material surface to be weak and for the oxide film to peel off due to laser irradiation to increase (the whitish areas in surface image 342), the feature amount information (first feature amount information) 341 is more stable on the low temperature side compared to that of type A, but there are still many peaks on the high temperature side, and some distortion of the waveform can be seen on both the high temperature side and the low temperature side.

The cut surface image of type B workpiece W was as shown in image 343 in FIG. 28(a), and the surface roughness at a depth of 1 mm from the surface (line indicated by I) and at a depth of 2 mm (line indicated by II) was as shown in graph 344 in FIG. 28(b). For type B workpiece W, the surface roughness at a depth of 2 mm from the surface was more stable than the surface roughness at a depth of 1 mm from the surface, and there was less overall variation than with type A, but some irregular streaks appeared on the cut surface.

Fig. 29 is a diagram for explaining the results of an investigation into the surface condition of the workpiece of type C. Fig. 30 is a diagram showing a cut surface image and surface roughness when the workpiece of Fig. 29 is cut.
As shown in Fig. 29(a) and Fig. 29(b), as can be seen from the surface image 350 and the enlarged image 350a of the material surface of the type C workpiece W, the oxide film on the material surface is not peeled off at all in the raw state (the oxide film is firmly attached). Fig. 29(c) includes a waveform diagram showing the positional change in the detected temperature when the type C workpiece W is subjected to laser scanning. This waveform diagram is obtained by extracting the positional change in temperature as feature amount information (first feature amount information) 351. Since the oxide film is not peeled off at all in the raw state, when the type C workpiece W is subjected to laser scanning as shown in Fig. 29(c), the feature amount information (first feature amount information) 351 showing the positional change in temperature (°C) shown on the vertical axis at the distance (mm) shown on the horizontal axis of the waveform diagram is shown as shown. The surface image 352 below the waveform diagram shows the surface state after the laser scanning corresponding to the waveform diagram.

In other words, there is almost no peeling of the oxide film on the surface of the material, and although a very small portion of the oxide film is peeled off by laser irradiation (the whitish area in surface image 352), the feature information (first feature information) 351 shows a slight peak on the high temperature side, but there is almost no distortion of the waveform on the low temperature side, resulting in a stable result.

The cut surface image of type C workpiece W is shown in image 353 in Figure 30(a), and the surface roughness at a depth of 1 mm from the surface (line indicated by I) and at a depth of 2 mm (line indicated by II) is shown in graph 354 in Figure 30(b). For type C workpiece W, the surface roughness at a depth of 1 mm from the surface is more stable than the surface roughness at a depth of 2 mm from the surface, but there is less overall variation than types A and B, and no irregular streaks appear on the cut surface.

From the above perspective, it has been proven that the cuttability (machinability) of the workpiece W of mild steel is strongly dependent on the distribution state of the oxide film on the material surface and the magnitude of thermal conductivity (internal characteristics) due to internal components, etc., and can be determined by the time or positional change in temperature based on infrared intensity. Therefore, the applicant has taken into consideration the results of the above-mentioned investigations conducted using the laser processing unit 20A and the radiation thermometer 30A, and set standard information for determining the workability of the workpiece W.

In other words, the reference information includes a first judgment criterion (threshold value) for judging the surface properties of the workpiece W and a second judgment criterion (threshold value) for judging the internal properties of the workpiece W. The first judgment criterion includes information on at least one of the reference temperature range and temperature variation of the workpiece W, and the second judgment criterion includes information indicating the degree of temperature rise of the workpiece W.

In the judgment of the workability by the radiation thermometer 30A in the NC device 60, specifically, in the processing system 100A, the laser processing device 10A irradiates the material surface of the workpiece W with the laser light LB under the second judgment irradiation condition while moving the irradiation position in the workability judgment process. The calculation processing unit 63 extracts the positional change in the temperature of the workpiece W as feature information based on the time series data of the infrared intensity measured by the radiation thermometer 30A. For example, the reference information stored in a storage device or the like includes a judgment criterion based on at least one of the reference temperature range and the temperature variation of the workpiece W. The workability judgment unit 66 judges the workability of the workpiece W depending on whether the positional change in the temperature of the workpiece W included in the feature information stored in a storage device or the like meets the judgment criterion. In the processing system 100A, the laser processing device 10A repeatedly irradiates the laser light LB into the inside of the material of the workpiece W under the second judgment irradiation condition in the workability judgment process. The calculation processing unit 63 extracts the temporal change in temperature of the workpiece W as feature information based on the time series data of the infrared intensity measured by the radiation thermometer 30A. For example, the reference information stored in a storage device or the like includes a judgment criterion that indicates the degree of temperature rise of the workpiece W. The workability judgment unit 66 judges the workability of the workpiece W based on whether the temporal change in temperature of the workpiece W included in the feature information stored in a storage device or the like is in a state of temperature rise that falls outside the judgment criterion.

Furthermore, in the processing system 100A, the second judgment irradiation condition includes a third irradiation condition and a fourth irradiation condition, and the laser processing apparatus 10A irradiates the laser light LB onto the material surface of the workpiece W while moving the irradiation position under the third irradiation condition in the process of judging the workability, and repeatedly irradiates the laser light LB into the material interior of the workpiece W under the fourth irradiation condition. In addition, the radiation thermometer 30A (second measurement unit) measures a first infrared intensity of the infrared light (radiated light) generated when the laser light LB is irradiated onto the workpiece W under the third irradiation condition, and a second infrared intensity of the infrared light (radiated light) generated when the laser light LB is irradiated onto the workpiece W under the fourth irradiation condition. The calculation processing unit 63 extracts first feature information indicating a positional change in temperature of the workpiece W based on the time series data of the first infrared intensity measured by the radiation thermometer 30A, and extracts second feature information indicating a time change in temperature of the workpiece W based on the time series data of the second infrared intensity measured by the radiation thermometer 30A. The workability judgment unit 66 compares the first feature information with a first judgment criterion (threshold) included in the reference information to judge whether the positional change in temperature of the workpiece W included in the first feature information satisfies the first judgment criterion (threshold) based on at least one of the reference temperature range and the temperature variation of the workpiece W included in the reference information. At the same time, the workability judgment unit 66 compares the second feature information with a second judgment criterion (threshold) included in the reference information to judge whether the time change in temperature of the workpiece W included in the second feature information is a temperature rise state that falls outside the second judgment criterion (threshold) indicating the degree of temperature rise of the workpiece W included in the reference information, and judges the workability of the workpiece W based on the combination of the results. The storage device etc. stores, for example, the first feature amount information and the second feature amount information extracted by the calculation processing unit 63 in a readable and writable manner.

In order to accurately determine the internal characteristics of the workpiece W, the laser processing device 10A preferably irradiates the surface of the workpiece W with the laser beam LB under the fifth irradiation condition before irradiating the interior of the workpiece W with the laser beam LB under the fourth irradiation condition, thereby removing the oxide film (surface modification) on the surface of the workpiece W. Here, the oxide film on the surface of the workpiece W is removed by, for example, increasing the gap with a high laser output of 2500 W and a nozzle gap of 70 mm under the fifth irradiation condition, thereby securing an area in which to peel off the oxide film. As a result, a crater with a depth of about 1.0 mm to about 1.5 mm and a depression diameter of about 6 mm is generated on the surface of the workpiece W.

Therefore, the laser output of the laser light LB irradiated by the laser processing device 10A under the third, fourth, and fifth irradiation conditions is set to be smaller than the laser output included in the cutting processing conditions, and the laser output under the third irradiation condition is set to be smaller than the laser output under the fourth irradiation condition.

Specifically, the standard processing conditions (standard conditions) for laser cutting are, for example, set as parameters according to the thickness of the workpiece W, such as a laser output of 3000 (W), a processing speed of 630 (mm/min), a (pulse) frequency of 2000 (Hz), a (pulse) duty (pulse width) of 100 (%), a gas type of _O2 , a gas pressure of 0.06 (MPa), a nozzle gap of 1 (mm), and an ACL of 62.

On the other hand, the third irradiation condition of the laser light LB is, for example, a laser output of 125 (W), a processing speed of 500 (mm/min), a (pulse) frequency of 100 (Hz), a (pulse) duty (pulse width) of 100 (%), a gas pressure of 0.04 (MPa), a nozzle gap of 50 (mm), a nozzle of D2.5W, a gas type of _O2 , a lens focal length of 190 (mm), an ACL of 120, a B-axis of 25 (mm), and a laser irradiation diameter of 5.9 (Φ), and is a condition for performing laser scanning over a distance of, for example, 80 mm at the above processing speed and determining the surface condition of the material of the workpiece W (adhesion or peeling state of the oxide film) based on the temperature characteristics.

The fourth irradiation condition of the laser light LB is, for example, a laser output of 425 (W), a processing speed of 0 (mm/min), a (pulse) frequency of 10 (Hz), a (pulse) duty (pulse width) of 50 (%), a gas pressure of 0.01 (MPa), a nozzle gap of 35 (mm), a nozzle of D7.0AL, a gas type of air (Air), a lens focal length of 190 (mm), an ACL of 140, a B-axis of 25 (mm), and a laser irradiation diameter of 4.5 (Φ), and these are conditions for, for example, performing laser scanning by spot (fixed point) irradiation and determining the internal state (thermal conductivity) of the material of the workpiece W based on its temperature characteristics.

The fifth irradiation condition of the laser light LB is, for example, a condition in which the focal position is located on the material surface, the parameters are set as follows: laser output 2500 (W), processing speed 0 (mm/min), (pulse) frequency 100 (Hz), (pulse) duty (pulse width) 100 (%), gas pressure 0.4 (MPa), nozzle gap 70 (mm), nozzle D7.0AL, gas type air, and lens focal length 190 (mm), and is a condition for performing surface modification (removal of oxide film) of the material of the workpiece W under a laser output higher than those under the third and fourth irradiation conditions.

The energy density of the laser beam LB irradiated under the third irradiation condition is 1 W/mm ² or more and less than 5 W/mm ² , and the energy density of the laser beam LB irradiated under the fourth irradiation condition is 5 W/mm ² or more and 10 W/mm ² or less. The reason for determining the surface texture of the workpiece W by setting the third irradiation condition as described above is as follows.

That is, the above-mentioned first to third investigations have revealed that an oxide film with weak adhesion is likely to peel off from the material surface during laser cutting, and a notch is likely to be formed on the cut surface. For this reason, it was decided to predict the adhesion from the behavior of temperature change due to laser scanning. In order to accurately evaluate the state of the material surface of the workpiece W, the thermal stress is generated by heating and cooling the oxide film on the material surface of the workpiece W while minimizing the penetration depth of heat by the laser. If the energy density of the irradiated laser light LB is too strong, the influence of the internal components of the material also appears, so it was assumed that an energy density of 1 W/mm2 or ^more and less than 5 W/ ^mm2 is sufficient for the purpose of loading thermal stress. And, when the oxide film peels off due to laser irradiation, it becomes sputtered, emits light, and becomes hot, so conversely, the part of the material surface where the oxide film has already peeled off in the raw state has a low laser absorption rate and therefore a low temperature. According to the knowledge of the present applicant, it has been found that if the oxide film has good adhesion to the material surface, the temperature is stable within a certain temperature range. If the irradiation area of the laser beam LB (laser irradiation diameter) is large, the state change of the oxide film is averaged by the laser scanning, so the behavior of the temperature change becomes small and the difference in the surface state becomes difficult to see. In addition, if the irradiation area is small, the laser output must be weakened at the same time, so the laser output becomes unstable. For these reasons, when judging the surface quality of the workpiece W, for example, the laser output is set to 125 (W) and the irradiation area (laser irradiation diameter) is set to about 5.9 (Φ).

The reason why the internal characteristics of the workpiece W are determined by setting the fourth irradiation condition as described above is as follows.
That is, the first and second investigations have revealed that the quality of the material workability (cuttability) of the workpiece W can be determined by evaluating the thermal conductivity of the material from the behavior of temperature change. For example, iron (Fe) is a material that undergoes magnetic transformation at a temperature of about 800°C, phase transformation at a temperature of about 900°C, and melts at a temperature of about 1500°C, so that the thermal conductivity and specific heat change at these temperatures. Therefore, in order to analyze the thermal conductivity inside the material of the workpiece W, it is evaluated by the change in temperature over time until the material melts. For this reason, the energy density of the laser irradiation is set to be 5 W/mm ² or more and 10 W/mm ² or less. In addition, this energy density is sufficiently weak compared to the energy density of the laser irradiation required for laser cutting, so the laser output (power) is likely to become unstable. Therefore, the irradiation area (laser irradiation diameter) is set to be large and the laser output (power) is adjusted.

It has been found that, in order to evaluate the thermal conductivity of a material, it is better to evaluate the cooling temperature when the laser irradiation is turned off than to evaluate the temperature reached by heating with laser irradiation, and a result that strongly reflects the thermal conductivity of the material is obtained. In other words, the behavior of the temperature change caused by heating with laser irradiation is affected by, for example, the laser absorption rate of the material surface in addition to the thermal conductivity, and is therefore unsuitable for evaluating thermal conductivity. On the other hand, with regard to cooling, even if the energy of the laser irradiation is small, the thermal conductivity of steel is relatively large, so it is cooled from a high temperature to near room temperature in about one cycle (100 ms). Therefore, it was assumed that a cooling time of about 50 ms is appropriate for evaluating the magnitude of the cooling temperature reached by the material. In addition, in order to accurately evaluate the behavior of the temperature change of the cooling temperature due to repeated heating and cooling, the (pulse) frequency was set to 10 (Hz) and the (pulse) duty (pulse width) was set to 50 (%). In addition, if the gas pressure of the assist gas is too strong, the cooling speed will be too fast, so the gas pressure was set to 0.01 (MPa) for the purpose of preventing spatter scattering on the protective glass. Since the heating temperature itself by laser irradiation is low and therefore the amount of oxidation heat generated is small, the type of assist gas may be either oxygen (O ₂ ) or nitrogen (N ₂ ).

From the above viewpoint, the applicant has set the first judgment criterion (threshold) for judging the surface quality of the workpiece W to, for example, whether the data components in the temperature range of 450°C to 800°C are 40% or less in the feature amount information 331, 341, 351 acquired by laser scanning at a predetermined distance, or whether the standard deviation in the temperature pattern is 100°C or more. If the surface quality is 40% or less in the temperature range, or if the standard deviation is 100°C or more, the condition of the oxide film is poor and the quality is poor. In addition, the second judgment criterion (threshold) for judging the internal characteristics of the workpiece W is set to, for example, whether the cooling temperature reached 5 seconds after the start of laser irradiation by repeated heating and cooling is less than 470°C. If the cooling temperature reached 5 seconds after the start of laser irradiation is less than 470°C, the internal characteristics are considered to have a high thermal conductivity and good quality. Note that the first and second judgment criteria are not limited to these.

For example, the first criterion may be a comparison of the total accumulated time during which the measured temperature is in the range of 450°C to 800°C with an accumulation time of 1 ms when laser scanning is performed for 10 seconds with a laser output of 150 (W), an irradiation area (laser irradiation diameter) of 6.5 (Φ), and a processing speed of 100 (mm/s). If this total accumulated time is shorter than the first criterion, it can be determined that the condition of the oxide film is poor, and if it is longer, the adhesion of the oxide film is good. Also, for example, the standard deviation of the measured temperature data with an accumulation time of 1 ms may be compared. If this standard deviation is larger than the first criterion, it can be determined that the condition of the oxide film is poor, and if it is smaller, the adhesion of the oxide film is good.

Furthermore, for example, the second criterion may be a comparison of the time taken from the start of cooling to reach 300°C after spot heating by irradiating the laser up to a temperature of 800°C and then turning off the laser irradiation and starting cooling. If this time is shorter than the second criterion, it can be determined that the internal characteristics are good, and if it is longer, the cut surface will be rough. Also, the cooling temperature 50 ms after turning off the laser irradiation can be compared. If this cooling temperature is lower than the second criterion, it can be determined that the internal characteristics are good, and if it is higher, the cutting quality will be poor. Note that the criterion information (first criterion, second criterion) is information that can be changed or set as appropriate, for example, by the operator.

[Processing ability assessment process flow]
31 and 32 are flowcharts showing an example of a process flow for determining workability using a radiation thermometer.
As shown in Fig. 31, when the processability judgment process using the radiation thermometer 30A starts in the processing system 100A, first, in the NC device 60, a processing program required for the control unit 61 is selected and read from a processing program DB (Database) 390 such as a storage device (step S120). Next, the processing conditions selected by the processing program are read from a processing condition DB (Database) 391 such as a storage device (step S121), and the internal characteristic judgment threshold value (second judgment criterion) and the surface quality judgment threshold value (first judgment criterion) of the workpiece W to which the processing conditions are applied are loaded from an internal characteristic judgment threshold value DB (Database) 392 and a surface quality judgment threshold value DB (Database) 393 such as a storage device, and set in the control unit 61 and the arithmetic processing unit 63, respectively.

Then, the calculation processing unit 63 determines the threshold value for each judgment (step S122). In the internal (characteristic) judgment process, the laser light LB is irradiated to the workpiece W under the fifth irradiation condition to perform the surface modification as described above, and then the laser light LB is irradiated to the workpiece W under the fourth irradiation condition to the surface modification section (step S123), and the temperature based on the infrared light emitted during laser irradiation is measured by the radiation thermometer 30A (step S124). In the surface (property) judgment process, the laser light LB is irradiated to the workpiece W under the third irradiation condition (step S125), and the temperature based on the infrared light emitted during laser irradiation is measured by the radiation thermometer 30A (step S126). Note that the internal judgment process and the surface judgment process may be performed in parallel, or one may be performed first and the other may be performed later.

As shown in FIG. 32, the calculation processing unit 63 inputs the time series data of the temperature measured in step S124 and extracts feature amount information (second feature amount information) indicating the temporal change in temperature of the workpiece W. The workability judgment unit 66 compares the extracted feature amount information (second feature amount information) with a set threshold value for internal property judgment to, for example, judge whether the internal property of the material is equal to or lower than the threshold value (step S127). The calculation processing unit 63 also inputs the time series data of the temperature measured in step S126 and extracts feature amount information (first feature amount information) indicating the positional change in temperature of the workpiece W. The workability judgment unit 66 compares the extracted feature amount information (second feature amount information) with a set threshold value for surface property judgment to, for example, judge whether the surface property of the material is equal to or lower than the threshold value (step S130).

If it is determined in step S127 that the internal characteristic is equal to or less than the threshold (YES in step S127), it is determined that the internal characteristic of the material of the workpiece W is good (step S128). On the other hand, if it is determined in step S127 that the internal characteristic is not equal to or less than the threshold (NO in step S127), it is determined that the internal characteristic of the material of the workpiece W is poor (step S129).

If it is determined in step S130 that the surface texture is equal to or less than the threshold (YES in step S130), the surface texture of the material of the workpiece W is determined to be good (step S131). On the other hand, if it is determined in step S130 that the surface texture is not equal to or less than the threshold (NO in step S130), the surface texture of the material of the workpiece W is determined to be poor (step S132).

Then, the machinability determination unit 66 performs a determination result process of the machinability of the workpiece W based on the results of either of the steps S128 and S129 and either of the steps S131 and S132 (step S133). That is, in the case of the combination of the steps S128 and S131, the internal characteristics (magnitude of thermal conductivity due to internal components) are good and the surface properties (state and adhesion of the oxide film) are good, so for example, a determination result (second determination result) is obtained that the machinability of the workpiece W is good. In addition, in the case of the other combinations of the steps S128 and S132, the internal characteristics are good but the surface properties are poor, in the case of the combination of the steps S129 and S131, the internal characteristics are poor but the surface properties are good, and in the case of the combination of the steps S129 and S132, both the internal characteristics and the surface properties are poor, so in these cases a determination result (second determination result) is obtained that the machinability of the workpiece W is poor.

The judgment result of the workability of the workpiece W thus obtained (second judgment result) is output from the workability judgment unit 66 to the overall judgment unit 65, and is also notified to the operator as the judgment result of the workability by the radiation thermometer 30A, for example, by being displayed on the display 70 (step S134), and the series of processes according to this flowchart is terminated. In this manner, the workability judgment process by the radiation thermometer 30A in the NC device 60 can be performed.

[Comprehensive judgment]
Next, a description will be given of the comprehensive judgment using the spectroscope 30 and the radiation thermometer 30A by the comprehensive judgment unit 65 of the NC device 60. In the comprehensive judgment, the comprehensive judgment unit 65 makes a comprehensive judgment on the workability of the workpiece W based on a combination of the judgment result of the workpiece W by the spectroscope 30 (first judgment result) and the judgment result of the workpiece W by the radiation thermometer 30A (second judgment result).
FIG. 33 is a diagram showing the result of classifying the judgment results using the spectroscope and the radiation thermometer and the cutting quality (cuttability) evaluation results obtained by actual cutting in the comprehensive judgment matrix information.

As shown in FIG. 33, the overall judgment matrix information 400 is made up of data in which the vertical axis represents the quality evaluations "○", "△", and "×" as the judgment results of the surface judgment of the surface condition and surface properties of the workpiece W by the spectroscope 30 and the radiation thermometer 30A, and the horizontal axis represents the quality evaluations "○", "△", and "×" as the judgment results of the internal judgment of the internal condition and internal properties of the workpiece W by the spectroscope 30 and the radiation thermometer 30A, and these are combined in a matrix table.

The matrix area 404 of the overall judgment matrix information 400 is divided into a category 1 area 401 surrounded by a solid line in the figure, a category 2 area 402 surrounded by a dashed line in the figure, and a category 3 area 403 surrounded by a two-dot chain line in the figure. The category 1 area 401 corresponds to all combinations where the quality evaluations of the surface judgment and the interior judgment by the radiation thermometer 30A are both "○" and the quality evaluations of the surface judgment and the interior judgment by the spectroscope 30 are "○" to "×".

In addition, the area 401 in category 1 corresponds to a combination in which the quality evaluations of the surface judgment and the internal judgment by the radiation thermometer 30A are both "○", or the quality evaluations of the surface judgment and the internal judgment by the spectroscope 30 are both "○".

On the other hand, area 402 in category 2 corresponds to combinations that do not belong to category 1, where the surface quality assessment by radiation thermometer 30A is "△" or "X" and the interior quality assessment is "○", or the surface quality assessment by spectrometer 30 is "△" or "X" and the interior quality assessment is "○".

Furthermore, area 403 of category 3 is a combination that does not belong to categories 1 and 2, and corresponds to a combination in which the radiation thermometer 30A gives a surface quality evaluation of "○" to "×" and the interior quality evaluation of "△" or "×", or the spectrometer 30 gives a surface quality evaluation of "○" and the interior quality evaluation of "△" or "×".

The above categories 1 to 3 do not simply indicate the quality of the workability (cuttability) of the workpiece W, but are classified to more accurately predict the workability based on a combination of the surface and internal judgment results of the spectrometer 30 and the radiation thermometer 30A before cutting the workpiece W. Therefore, by using the overall judgment matrix information 400, the processing conditions to be recommended according to the predicted workability can be more appropriately determined. The overall judgment unit 65 judges the workability (cuttability) of the workpiece W classified into the area 401 of category 1 as a workpiece (standard material) that can be laser cut (cutting) well under standard conditions as the processing conditions recommended in the overall judgment matrix information 400. The overall judgment unit 65 judges the workability of the workpiece W classified into the area 402 of category 2 as a workpiece (material that requires test cutting) that can be laser cut (cut) under standard conditions as the processing conditions recommended in the overall judgment matrix information 400, but that the standard conditions need to be adjusted. Furthermore, the overall judgment unit 65 judges that the workpiece W classified in the area 403 of category 3 is a workpiece (test cut recommended material) that is difficult to cut with a laser (cutting process) using the standard conditions as the recommended processing conditions in the overall judgment matrix information 400. In other words, the workpiece W includes standard materials, materials requiring test cuts that can be cut under standard conditions but cannot be cut well, and test cut recommended materials that are difficult to cut under standard conditions.

In addition, when evaluating the cutting quality of the workpiece W, 25 types (samples 1 to 4, No. 1 to 21) of mild steel workpiece W with a plate thickness of 25 mm and different materials, production lots, and production processes, etc., which were not used in the machine learning of the first and second judgment models 5 and 6 described above, were used. These workpieces W were irradiated with laser light LB under the first judgment irradiation condition and measured using a spectrometer 30 to obtain the above-mentioned judgment results of the surface condition and internal condition of the workpiece W (first judgment result), and were irradiated with laser light LB under the second judgment irradiation condition and measured using a radiation thermometer 30A to obtain the above-mentioned judgment results of the surface properties and internal characteristics of the workpiece W (second judgment result), and comprehensive judgment matrix information 400 was created to comprehensively judge the workability (cuttability).

In addition, when evaluating the cutting quality (cuttability) in Figure 33, the cutting conditions (processing conditions) for laser cutting were set to a laser output of 9000 (W), processing speed of 850 (mm/min), (pulse) frequency of 2000 (Hz), (pulse) duty (pulse width) of 65 (%), gas pressure of 0.09 (MPa), nozzle gap of 0.5 (mm), nozzle DG of 2.5, gas type of oxygen ( _O2 ), lens focal length of 190 (mm), and ACL of 70, and cutting processing was performed at multiple focal positions, and the cutting quality was evaluated by observation evaluation.

In FIG. 33, the cutting quality (cuttability) was evaluated as follows: good cutting quality (good cuttability) was assigned a quality score of "2", and slightly poor cutting quality (poor cuttability) was assigned a quality score of "1". Also, poor cutting quality or cutting impossible (uncuttable) was assigned a quality score of "0". The type of each workpiece W (samples 1-4, No. 1-21) was classified and displayed in the matrix area 404 of the overall judgment matrix information 400 along with the cut quality evaluation.

According to the overall judgment matrix information 400 shown in FIG. 33, it can be seen that the judgment results (quality evaluation) of the material state (material characteristics) of each workpiece W (samples 1 to 4, No. 1 to No. 21) may differ between the judgment results (first judgment result) by the spectroscope 30 and the judgment results (second judgment result) by the radiation thermometer 30A.

As shown in FIG. 33, for example, in the case of the workpiece W with material numbers (material number: No) No. 13 and No. 14, the surface judgment (surface condition and surface properties) judgment results are both rated as "○" and "good" by the spectrometer 30 and the radiation thermometer 30A, but the internal judgment (internal condition and internal characteristics) judgment results are different. That is, the internal characteristic judgment result by the radiation thermometer 30A is rated as "△" and "fair", indicating slightly poor thermal conductivity. On the other hand, the internal condition judgment result by the spectrometer 30 is rated as "○" and "good", indicating that the internal condition is good. Also, according to the cutting processing results of the workpiece W with No. 13 and No. 14, the cutting quality is scored as "2", indicating good cuttability.

In other words, in the above case, it can be seen that the spectrometer 30 captures physical phenomena that cannot be captured by, for example, the radiation thermometer 30A, and makes an internal judgment. On the other hand, while the spectrometer 30 may not be able to capture the material state (surface state and internal state) and may make an erroneous judgment on cuttability, the radiation thermometer 30A may capture the material state (surface properties and internal characteristics) and be able to accurately judge cuttability. In order to investigate this, the fourth study was conducted.

FIG. 34 is a table showing the results of the determination of the material state (material characteristics) of Samples 1 to 4 by the spectroscope and the radiation thermometer, and the evaluation results of the cut surfaces.
Samples 1, 2 and samples 3, 4 shown in the results table 405 of Figure 34 are processed materials W made of two types of steel material with different internal components that are manufactured in the same lot, with samples 1 and 3 being in their raw material state and samples 2 and 4 being in a surface-modified state.

The surface modification was performed, for example, by removing the oxide film on the material surface by polishing to reduce the average surface roughness to 0.8 or less, and then heating and holding the material in a water vapor atmosphere at 500°C for 2 hours to generate an oxide (magnetite) with high adhesion on the material surface. The cutting conditions for the laser cutting were set to a laser output of 3 (kW), a processing speed of 630 (mm/min), and oxygen ( _O2 ) as the assist gas, and the cutting process was performed, and the cut surface was evaluated by observation.

In the fourth investigation, as shown in Figure 34, first, when judged by the spectroscope 30, the surface quality and internal quality of samples 1 to 4 were the same, but the quality evaluation of the cut surface was different (samples 1 and 3 were "x", sample 2 was "○", and sample 4 was "△"). In other words, when judged by the spectroscope 30, the surface quality (oxide film type and thickness) of samples 1 to 4 was judged to be "x", and the internal quality (stability of melting behavior) was judged to be "△". The reason for this will be considered.

Fig. 35 is a diagram showing chemical components (mass%) other than iron contained in the materials of Samples 1 to 4. The component table 406 in Fig. 35 shows the contents (mass percentages (mass%)) of carbon (C), silicon (Si), manganese (Mn), phosphorus (P), sulfur (S), chromium (Cr), nickel (Ni), molybdenum (Mo), and copper (Cu).
As shown in the composition table 406 of FIG. 35, samples 1 and 2 are materials characterized by a high carbon (C) content, while samples 3 and 4 are materials characterized by a low carbon content and a high manganese (Mn) content.

Carbon is an element that tends to concentrate in the final solidification part of the material, and the parts with concentrated carbon melt easily, while parts with less carbon melt less easily. If there is a lot of carbon, uneven melting is more likely to occur inside the material. Manganese also shows a similar tendency, although it is not as pronounced as carbon. Samples 1 and 2 have a large amount of carbon, while samples 3 and 4 have a small amount of carbon but a large amount of manganese. Therefore, it is thought that samples 1 and 2 and samples 3 and 4 have almost the same results in terms of the temporal fluctuations of the emission spectra measured by the spectrometer 30 in the molten state.

On the other hand, the evaluation using radiation thermometer 30A, which evaluates thermal conductivity in a solid state, is strongly influenced by silicon and manganese in addition to carbon, so it is believed that samples 3 and 4, which contain large amounts of alloy elements other than iron such as manganese, were judged to have low thermal conductivity when evaluated internally. The thermal conductivity was actually measured using the laser flash method, and the thermal conductivity of samples 3 and 4 was 48.9 W/(m·K), which was lower than the thermal conductivity of samples 1 and 2, which was 57.7 W/(m·K), and this is believed to reflect the accuracy of the internal evaluation using radiation thermometer 30A.

Fig. 36 is a graph showing the relationship between temperature and time when repeated heating and cooling were performed on sample 1. Fig. 36 shows a temperature waveform 410 obtained by measuring sample 1 with radiation thermometer 30A and converting the time series data of infrared intensity output from radiation thermometer 30A into time series data of temperature.
According to the temperature waveform 410 shown in FIG. 36, the cooling temperature reached 5 seconds after the start of laser irradiation was 420° C., as indicated by the arrow in the figure, which was less than the second judgment criterion (threshold value) of 470° C., and therefore the quality was evaluated as “○”.

Fig. 37 is a graph showing the relationship between temperature and time when repeated heating and cooling were performed on sample 3. Fig. 37 shows a temperature waveform 411 obtained by measuring sample 3 with radiation thermometer 30A and converting the time series data of infrared intensity output from radiation thermometer 30A into time series data of temperature.
37, the cooling temperature reached 5 seconds after the start of laser irradiation was 558° C., as indicated by the arrow in the figure, which was higher than the second judgment criterion (threshold value) of 470° C., and therefore the quality evaluation was “x.” In this way, it was confirmed that the internal judgment using the radiation thermometer 30A made it possible to judge the difference between Samples 1 and 2 and Samples 3 and 4.

Figure 38 is a graph showing the relationship between temperature and time when laser scanning was performed on samples 1 and 2. Figure 38(a) shows a temperature waveform 412 obtained by radiation thermometer 30A by performing laser scanning on the material surface of sample 1, and Figure 38(b) shows a temperature waveform 413 obtained by radiation thermometer 30A by performing laser scanning on the material surface after surface modification for sample 2. Figure 39 shows a cut surface image and surface roughness when sample 1 is cut. Figure 40 shows a cut surface image and surface roughness when sample 2 is cut.

As described above, when samples 1 and 2 were judged by the spectrometer 30, the surface and internal quality were judged to be the same (surface quality was "x" and internal quality was "△"), but when judged by the radiation thermometer 30A, the surface quality was different (surface quality was "x" and "○"). Note that when judged by the radiation thermometer 30A, the internal quality was the same (internal quality was "○").

As shown in the temperature waveform 412 in FIG. 38(a), sample 1 has large and frequent temperature fluctuations during laser irradiation, and is therefore judged to have poor adhesion of the oxide film and significant peeling. The cut surface image of sample 1 is shown in image 414 in FIG. 39(a), and the surface roughness at a depth of 1 mm (line indicated by I) from the laser irradiation surface and at a depth of 15 mm (line indicated by II) are shown in graphs 415a in FIG. 39(b) and 415b in FIG. 39(c), respectively. For sample 1, the surface roughness at a depth of 1 mm from the laser irradiation surface is stable with little variation, but the surface roughness at a depth of 15 mm from the laser irradiation surface is unstable with a lot of variation, resulting in some irregular streaks appearing on the cut surface.

In contrast, as shown in the temperature waveform 413 in FIG. 38(b), sample 2 has a small temperature fluctuation during laser irradiation, and therefore it is determined that the oxide film is uniformly adhered. The cut surface image of sample 2 is as shown in image 416 in FIG. 40(a), and the surface roughness at a depth of 1 mm (line indicated by I) from the laser irradiation surface and at a depth of 15 mm (line indicated by II) are as shown in graphs 417a in FIG. 40(b) and 417b in FIG. 40(c), respectively. For sample 2, the surface roughness at a depth of 1 mm from the laser irradiation surface is more stable than the surface roughness at a depth of 15 mm from the laser irradiation surface, but overall there is little variation and it is stable, resulting in no irregular streaks appearing on the cut surface.

The quality evaluation of the cut surface by laser cutting is different, with sample 1 receiving an "X" and sample 2 receiving an "O" despite the fact that they both have the same thermal conductivity of 57.7 W/(m·K). This indicates that the cut surface of sample 2, which has been surface-modified, has less roughness and the cutting quality has been greatly improved. This is thought to be because the surface modification increases the adhesion and uniformity of the oxide film on the material surface, suppressing the oxidation reaction on the material surface and reducing the effect of the oxidation reaction on the cutting ability.

Figure 41 is a graph showing the relationship between temperature and time when laser scanning was performed on samples 3 and 4. Figure 41(a) shows a temperature waveform 418 obtained by radiation thermometer 30A by performing laser scanning on the material surface of sample 3, and Figure 41(b) shows a temperature waveform 419 obtained by radiation thermometer 30A by performing laser scanning on the material surface after surface modification for sample 4. Figure 42 shows a cut surface image and surface roughness when sample 3 is cut. Figure 43 shows a cut surface image and surface roughness when sample 4 is cut.

Similar to samples 1 and 2 above, samples 3 and 4 were judged by the spectrometer 30 to have the same surface and internal quality (surface quality: "x" and internal quality: "△"), but the surface quality was different when judged by the radiation thermometer 30A (surface quality: "x" and "o"). The internal quality was judged by the radiation thermometer 30A to have the same quality (internal quality: "x").

As shown in the temperature waveform 418 in FIG. 41(a), sample 3 has large and frequent temperature fluctuations during laser irradiation, and is therefore judged to have poor adhesion of the oxide film and significant peeling. The cut surface image of sample 3 is shown in image 420 in FIG. 42(a), and the surface roughness at a depth of 1 mm (line indicated by I) from the laser irradiation surface and at a depth of 15 mm (line indicated by II) are shown in graphs 421a in FIG. 42(b) and 421b in FIG. 42(c), respectively. For sample 3, the surface roughness at a depth of 1 mm from the laser irradiation surface is stable with little variation, but the surface roughness at a depth of 15 mm from the laser irradiation surface is quite unstable with a lot of variation, resulting in many irregular streaks being seen on the cut surface.

In contrast, as shown by the temperature waveform 419 in FIG. 41(b), sample 4 has less temperature fluctuation during laser irradiation, and therefore it is judged that the uniformity of the oxide film is improved compared to sample 3. The cut surface image of sample 4 is as shown in image 422 in FIG. 43(a), and the surface roughness at a depth of 1 mm (line indicated by I) from the laser irradiation surface and at a depth of 15 mm (line indicated by II) are as shown by graphs 423a in FIG. 43(b) and 423b in FIG. 43(c), respectively. Although the surface roughness of sample 4 is stable at a depth of 1 mm from the laser irradiation surface, the surface roughness at a depth of 15 mm from the laser irradiation surface is unstable and varies, and many irregular striations are observed on the cut surface.

The quality evaluation of the cut surface by laser cutting is different, with sample 3 receiving an "X" and sample 4 receiving a "△" despite the fact that they both have the same thermal conductivity of 48.9 W/(mK). This indicates that the roughness of the cut surface of sample 4, which has been surface modified, is reduced, improving the cutting quality. The reason for this is as stated in the discussion of samples 1 and 2. However, samples 3 and 4 have a lower thermal conductivity than samples 1 and 2, so the cut surface is more likely to become rough and the roughness of the cut surface is somewhat greater, resulting in inferior cutting quality.

As described above, it has been proven that the cutting ability (machinability) of the mild steel workpiece W is strongly correlated with the state of the oxide film on the surface of the workpiece W and the internal characteristics, and that the type and adhesion of the oxide film, as well as the stability of the melting behavior inside the material, can be analyzed based on the temporal change (time series data) in the emission spectrum intensity generated when the material surface is melted, and that the oxide film distribution state and the magnitude of heat conduction inside the material can be analyzed based on the temporal or positional change (time series data) in temperature at a laser output that does not melt the material. These analysis results are included in the first and second judgment results described above.

Therefore, in the comprehensive judgment by the processing system 100A, the judgment result of the workability of the workpiece W by the spectrometer 30 (first judgment result) and the judgment result of the workability of the workpiece W by the radiation thermometer 30A (second judgment result) are used (combined) to comprehensively judge the workability of the workpiece W. This makes it possible to mutually complement weak points in the judgment result by the spectrometer 30 and the judgment result by the radiation thermometer 30A, and further improves the judgment accuracy of the workability (cuttability).

In addition, when the radiation thermometer 30A judges the surface of the workpiece W to have a large temperature variation, the result of the judgment is combined with the result of the spectrometer 30 to determine the type and thickness of the oxide film by combining the judgment results, and it is possible to set optimal processing conditions (cutting conditions).

For example, if the oxide film is mainly composed of hematite (Fe ₂ O ₃ ), the adhesion of the oxide film is weak (small) and it is easy to peel off, so it is effective to adjust the parameters of the processing conditions, such as narrowing the focus diameter of the laser light LB on the material surface. Also, if part of the oxide film is composed of magnetite (Fe ₃ O ₄ ), the adhesion of the oxide film is somewhat high, but for those that peel off during laser cutting, it is considered effective to take measures such as marking the cutting path before laser cutting or performing a pre-burning process to make the surface condition uniform. Such a processed material W can be classified into the category 2 area 402 in the matrix area 404 of the overall judgment matrix information 400 in FIG. 23.

Furthermore, in the internal judgment of the workpiece W, for the workpiece W whose internal condition is judged to be poor by the spectroscope 30, it is effective to take measures such as increasing the nozzle diameter of the nozzle 25 of the laser processing head 22 to suppress the flow rate of the assist gas in order to suppress the combustion reaction. For the workpiece W whose internal characteristics are judged to be poor by the radiation thermometer 30A, it is considered effective to take measures such as increasing the amount of water used for cooling, lowering the frequency or duty in the processing conditions, and correspondingly lowering the cutting speed in order to improve the cooling effect. Such workpiece W can be classified into the category 3 area 403 in the matrix area 404 of the overall judgment matrix information 400 in FIG. 33.

[Processing flow based on overall judgment]
FIG. 44 is a flowchart showing an example of a process flow based on a comprehensive judgment of a machining system.
As shown in FIG. 44, in the processing system 100A, before cutting the workpiece W, the NC device 60 acquires, for example, thickness information of the workpiece W on the processing table 11 (step S150) and transmits it to the judgment model storage unit 51 of the PC 50A.

Next, based on the transmitted plate thickness information, the first and second judgment models 5, 6 corresponding to the plate thickness are selected in the judgment model storage unit 51 as described above and transmitted to the machinability calculation unit 52. Then, in the laser processing unit 20A, the probe light (laser light LB) for the spectrometer 30 is irradiated to the workpiece W under the first judgment irradiation conditions (first irradiation conditions, second irradiation conditions) (step S151), the spectrometer 30 acquires spectral data based on the time series data of the emission spectrum (step S152), and the first waveform information 3 and the second waveform information 4 are calculated and transmitted as described above.

The machinability calculation unit 52 calculates a quality score indicating a quality evaluation of the surface condition and internal condition of the workpiece W based on the first waveform information 3 and second waveform information 4 and the first judgment model 5 and second judgment model 6 transmitted as described above, and generates judgment matrix information 110 (step S153).

Then, the machinability determination unit 53 determines the quality of the surface condition and internal condition of the workpiece W (surface determination and internal determination) based on the calculated quality score and the generated determination matrix information 110 (step S154), and transmits the first determination result of the machinability of the workpiece W to the overall determination unit 65.

In addition, in the laser processing unit 20A, the probe light (laser light LB) for the radiation thermometer 30A is irradiated to the workpiece W under the second judgment irradiation conditions (third irradiation condition to fifth irradiation condition) (step S155), the radiation thermometer 30A acquires temperature data based on the time series data of the infrared intensity (step S156), and the calculation processing unit 63 extracts feature information (first feature information, second feature information).

The machinability judgment unit 66 judges the quality of the surface texture and internal characteristics of the workpiece W (surface judgment and internal judgment) based on the extracted feature information and judgment criteria (first judgment criterion, second judgment criterion) (step S157), and outputs the second judgment result of the machinability of the workpiece W to the overall judgment unit 65.

Then, the overall judgment unit 65 judges the overall quality of the material state of the workpiece W from the first judgment result by the spectrometer 30 and the second judgment result by the radiation thermometer 30A using the overall judgment matrix information 400 (step S158), and obtains an overall judgment result (third judgment result) that comprehensively judges the workability of the workpiece W.

If the workpiece W is classified into the category 1 area 401 in the matrix area 404 of the overall judgment matrix information 400 based on the overall judgment result, for example, condition A is displayed on the display 70 (step S159). Condition A includes, for example, information regarding the third judgment result indicating that the workpiece W is a standard material, information notifying that the workpiece W can be processed under standard conditions, etc.

Then, processing conditions that correspond to the material state of the workpiece W whose workability has been determined by the overall determination unit 65 are set in the laser processing device 10A based on, for example, an input from an operator (step S165), and product processing of the workpiece W is carried out based on the set processing conditions (step S166), thus completing the series of processes according to this flowchart.

In addition, if the workpiece W is classified into the category 2 area 402 in the matrix area 404 of the overall judgment matrix information 400, for example, condition B is displayed on the display 70 (step S160). Condition B includes, for example, information on the third judgment result indicating that the workpiece W is a material requiring test cutting, and information informing the user that the processing conditions for the workpiece W need to be changed or adjusted from the standard conditions, for example, by narrowing the focusing diameter as described above.

In addition, if the workpiece W is classified into the category 3 area 403 in the matrix area 404 of the overall judgment matrix information 400, for example, condition C is displayed on the display 70 (step S161). Condition C includes, for example, information on the third judgment result indicating the workability of the workpiece W as a material recommended for test cutting, information notifying that the processing conditions of the workpiece W need to be changed or adjusted from the standard conditions or the difficult-to-process material conditions, for example, by reducing the flow rate of the assist gas to suppress the combustion reaction as described above, and the like.

Furthermore, after condition B is displayed in step S160 or condition C is displayed in step S161, the processing conditions are changed or adjusted based on, for example, an operator's input, and then test processing is performed (step S162). The quality (processing quality) of the state of the cut surface of the workpiece W obtained by the test processing is checked, and it is determined whether the processing quality is OK (step S163).

If it is determined that the processing quality is not OK (F: False in step S163), the parameters of the processing conditions at the time of the test cut are adjusted, for example, by the operator's input (step S164), test processing is performed again (step S162), and the subsequent processes are repeated.

On the other hand, if it is determined that the processing quality is OK (T: True in step S163), the process proceeds to step S165 as described above, where processing conditions suited to the material state of the workpiece W are set in the laser processing device 10A (step S165), product processing is carried out based on this (step S166), and the series of processes according to this flowchart ends. Note that the order of operations may be reversed between the processes on the spectrometer 30 side in steps S150 to S154 and the processes on the radiation thermometer 30A side in steps S155 to S157.

As described above, the processing system (machinability judgment system) 100A of the second embodiment judges the processability based on the processing conditions of the workpiece W before cutting based on the measurement by the spectrometer 30 and based on the measurement by the radiation thermometer 30A, and makes a comprehensive judgment based on a combination of these judgment results, which can further improve the judgment accuracy, and the presentation of the measures to be taken according to the contents of the judgment results (the material state of the workpiece W and the recommended processing conditions) (for example, various measures such as whether cutting should be performed under the recommended processing conditions, whether the processing conditions should be changed or adjusted to be suitable for the workpiece W, whether the recommended processing conditions should be further adjusted, and whether test processing should be performed) becomes more accurate and easier to judge. As a result, it is possible to accurately and quickly improve the processing quality, so that the skills of an expert are not required, the occurrence of product defects and the effort required for condition adjustment can be reduced, and processing defects can be further reduced.

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

1. First waveform information (teaching data of explanatory variables)
2. Second waveform information (teaching data of explanatory variables)
3. First waveform information (data for estimating explanatory variables)
4. Second waveform information (data for estimating explanatory variables)
5 First judgment model 6 Second judgment model 9 Judgment result 10, 10A Laser processing device 20, 20A Laser processing unit 30 Spectrometer 50 Machinability judgment unit 51 Judgment model storage unit 52 Machinability calculation unit 53, 66 Machinability judgment unit 60 NC device 61 Control unit 62 Machining condition storage unit 63 Calculation processing unit 65 Overall judgment unit 70 Display 80 Learning device 81 First judgment model learning unit 82 Second judgment model learning unit 90 Machinability judgment system 100, 100A Machining system 109a Surface quality evaluation result (teaching data of objective variable)
109b Internal quality evaluation result (teaching data of objective variable)

Claims (23)

a laser processing device capable of executing a processing step of cutting a workpiece by irradiating the workpiece with a laser beam under processing irradiation conditions, and a workability judgment step of irradiating the workpiece with the laser beam under judgment irradiation conditions that melt the workpiece but do not penetrate the workpiece, and judging the workability of the workpiece;
a measuring device for measuring an emission spectrum generated when the workpiece is irradiated with the laser light under the judgment irradiation conditions;
a determination device for determining the workability of the workpiece based on the time series data of the emission spectrum measured by the measurement device;
Equipped with
The determination device includes:
A first judgment model and a second judgment model are included,
extracting first waveform information in a first time domain and second waveform information in a second time domain from the time series data of the emission spectrum measured by the measurement device;
inputting the extracted first waveform information as estimation data into the first judgment model to obtain a surface quality evaluation of the workpiece;
inputting the extracted second waveform information as estimation data into the second judgment model to obtain an internal quality evaluation of the workpiece;
a processing system that outputs a judgment result of the workability of the workpiece when cut and processed under preset processing conditions in the processing step, based on a combination of the obtained surface quality evaluation and internal quality evaluation of the workpiece. The processing system according to claim 1 , wherein the laser processing device spot irradiates the laser light under the judgment irradiation condition in the processability judgment step. The first judgment model is created by performing machine learning using the first waveform information and a surface quality evaluation result indicating a surface quality of the workpiece as first teacher data,
The processing system according to claim 1 or 2, wherein the second judgment model is created by performing machine learning by inputting the second waveform information and an internal quality evaluation result indicating the internal quality of the workpiece as second training data. The laser processing apparatus divides a time from the start of irradiation of the laser light to the end of irradiation in the processability judgment step into a first stage and a second stage, and irradiates the workpiece with the laser light under a first irradiation condition in the first stage and irradiates the workpiece with the laser light under a second irradiation condition in the second stage as the judgment irradiation condition,
the measuring device measures a first emission spectrum generated when the laser beam is irradiated onto the workpiece under the first irradiation condition, and a second emission spectrum generated when the laser beam is irradiated onto the workpiece under the second irradiation condition;
3. The machining system according to claim 1 or 2, wherein the determination device extracts first waveform information in the first time domain from time series data of the first emission spectrum measured by the measurement device, and extracts second waveform information in the second time domain from time series data of the second emission spectrum measured by the measurement device. 3. The processing system according to claim 1 or 2, wherein the determination device extracts the first waveform information in a first time domain and in a first wavelength band and a second wavelength band, and the second waveform information in a second time domain and in the first wavelength band and the second wavelength band, from the time series data of the emission spectrum measured by the measurement device. The processing system according to claim 5 , wherein the first wavelength band and the second wavelength band are different from an emission wavelength band of the laser light. The processing system according to claim 1 or 2, wherein the laser output of the judgment irradiation condition is smaller than the laser output of the processing irradiation condition. The processing system according to claim 4 , wherein a laser output under the first irradiation condition is smaller than a laser output under the second irradiation condition. The processing system of claim 4 , wherein the first stage is shorter than the second stage. The processing system according to claim 1 or 2, wherein a nozzle gap of the laser processing device while irradiating the laser light under the judgment irradiation condition is larger than a nozzle gap of the laser processing device while irradiating the laser light under the processing irradiation condition. the determination device includes a notification unit that notifies the determination result in a confirmable manner,
The machining system according to claim 1 or 2, wherein the judgment result includes a processability evaluation capable of expressing the adaptability of the machining conditions to the workpiece in a combination of the surface quality evaluation and the internal quality evaluation. The processing system according to claim 11, wherein the notification unit notifies at least one of information informing the user of the processing conditions to be set in the laser processing device according to the workpiece based on the processability evaluation, information encouraging test processing by the laser processing device, and information encouraging the user to adjust, change, or set the processing conditions. The processing system according to claim 12, wherein the determination device includes a setting unit that receives input information selected and input by an operator based on the information notified by the notification unit, and sets processing conditions corresponding to the received input information to the laser processing device. The processing system according to claim 3 , wherein the surface quality evaluation result and the internal quality evaluation result include a score evaluation obtained by scoring the surface quality evaluation and the internal quality evaluation of the workpiece into a plurality of different numerical values. The processing system according to claim 1 or 2, wherein the first judgment model and the second judgment model are created for each thickness of the workpiece. a judgment device having a first judgment model and a second judgment model, in a workability judgment step for judging the workability of a workpiece, inputting first waveform information and second waveform information calculated from time series data of an emission spectrum generated when a laser beam is irradiated onto the workpiece under judgment irradiation conditions that melt the workpiece but do not penetrate the workpiece, into the first judgment model and the second judgment model, respectively, and judging the workability of the workpiece based on the first waveform information and the second waveform information;
a learning device that creates the first determination model and the second determination model;
Equipped with
The determination device includes:
extracting first waveform information in a first time domain and second waveform information in a second time domain from the time series data of the emission spectrum;
inputting the extracted first waveform information as estimation data into the first judgment model to obtain a surface quality evaluation of the workpiece;
inputting the extracted second waveform information as estimation data into the second judgment model to obtain an internal quality evaluation of the workpiece;
Based on the obtained combination of the surface quality evaluation and internal quality evaluation of the workpiece, a judgment result of the workability of the workpiece when cut under preset processing conditions in a processing step of irradiating the workpiece with the laser light under processing irradiation conditions to cut the workpiece is outputted;
The learning device includes:
The first waveform information and a surface quality evaluation result indicating the surface quality of the workpiece are input as first teacher data, and machine learning is performed to create the first judgment model;
A workability judgment system that inputs the second waveform information and an internal quality evaluation result indicating the internal quality of the workpiece as second training data and performs machine learning to create the second judgment model. a laser processing device capable of executing a processing step of irradiating a workpiece with a laser beam under processing irradiation conditions to cut the workpiece, and a workability judgment step of irradiating the workpiece with the laser beam under first judgment irradiation conditions under which the laser beam melts the workpiece but does not penetrate it, and under second judgment irradiation conditions under which the laser beam does not exceed the melting point of the material of the workpiece, to judge the workability of the workpiece;
a measuring device including a first measuring unit that measures an emission spectrum generated when the workpiece is irradiated with the laser light under the first judgment irradiation condition, and a second measuring unit that measures an infrared intensity of radiant light generated when the workpiece is irradiated with the laser light under the second judgment irradiation condition;
a first judgment unit that judges the workability of the workpiece based on time series data of the emission spectrum measured by the first measurement unit of the measuring device, a second judgment unit that judges the workability of the workpiece based on time series data of the infrared intensity measured by the second measurement unit of the measuring device, and a third judgment unit that judges the workability of the workpiece when cut under preset processing conditions in the processing step based on a combination of a first judgment result judged by the first judgment unit and a second judgment result judged by the second judgment unit, and outputs a third judgment result;
Equipped with
the determination device has a first determination model and a second determination model;
The first determination unit of the determination device
extracting first waveform information in a first time domain and second waveform information in a second time domain from time series data of the emission spectrum measured by the first measurement unit of the measuring device, inputting the extracted first waveform information as estimation data into the first judgment model to obtain a surface quality evaluation of the workpiece, inputting the extracted second waveform information as estimation data into the second judgment model to obtain an internal quality evaluation of the workpiece, and outputting the first judgment result that judges the workability of the workpiece based on a combination of the obtained surface quality evaluation and internal quality evaluation of the workpiece,
The second determination unit of the determination device
A processing system which extracts feature information indicating a temporal or positional change in temperature of the workpiece based on time series data of the infrared intensity measured by the second measurement unit of the measuring device, and outputs the second judgment result which judges the workability of the workpiece based on the extracted feature information and pre-registered reference information for judging the workability of the workpiece. 18. The processing system according to claim 17, wherein, in the processability judgment step, the laser processing device spot irradiates the workpiece with the laser light under the first judgment irradiation condition, irradiates the material surface of the workpiece with the laser light while moving an irradiation position under the second judgment irradiation condition, and repeatedly irradiates the laser light into the inside of the workpiece. the first determination irradiation condition includes a first irradiation condition and a second irradiation condition,
In the laser processing apparatus, when the laser beam is irradiated under the first judgment irradiation condition in the processability judgment step, a time from the start of irradiation of the laser beam to the end of irradiation is divided into a first stage and a second stage, the laser beam is irradiated to the workpiece under the first irradiation condition in the first stage, and the laser beam is irradiated to the workpiece under the second irradiation condition in the second stage;
the first measurement unit of the measuring device measures a first emission spectrum generated when the laser light is irradiated onto the workpiece under the first irradiation condition, and a second emission spectrum generated when the laser light is irradiated onto the workpiece under the second irradiation condition;
19. The machining system according to claim 17 or 18, wherein the first determination unit of the determination device extracts first waveform information of the first time domain from time series data of the first emission spectrum measured by the first measurement unit of the measurement device, and extracts second waveform information of the second time domain from time series data of the second emission spectrum measured by the first measurement unit of the measurement device. The first judgment model is created by performing machine learning using the first waveform information and a surface quality evaluation result indicating a surface quality of the workpiece as first teacher data,
The second judgment model is created by inputting the second waveform information and an internal quality evaluation result indicating the internal quality of the workpiece as second teacher data and performing machine learning.
19. The processing system according to claim 17 or 18. the second determination irradiation condition includes a third irradiation condition and a fourth irradiation condition,
In the laser processing apparatus, in the processability judgment step, when the laser light is irradiated under the second judgment irradiation condition, the laser light is irradiated onto the surface of the workpiece while moving the irradiation position under the third irradiation condition, and the laser light is repeatedly irradiated into the inside of the workpiece under the fourth irradiation condition;
the second measuring unit of the measuring device measures a first infrared intensity of radiated light generated when the laser beam is irradiated onto the workpiece under the third irradiation condition, and a second infrared intensity of radiated light generated when the laser beam is irradiated onto the workpiece under the fourth irradiation condition;
19. The machining system according to claim 17 or 18, wherein the second judgment section of the judgment device extracts first feature information indicating a positional change in temperature of the workpiece based on the time series data of the first infrared intensity measured by the second measurement section of the measuring device, extracts second feature information indicating a temporal change in temperature of the workpiece based on the time series data of the second infrared intensity measured by the second measurement section of the measuring device, judges whether the positional change in temperature of the workpiece included in the first feature information satisfies a first judgment criterion based on at least one of a reference temperature range and a temperature variation of the workpiece included in the reference information, and judges whether the temporal change in temperature of the workpiece included in the second feature information is a temperature rise state that falls outside a second judgment criterion indicating a degree of temperature rise of the workpiece included in the reference information, and outputs the second judgment result that judges the workability of the workpiece based on a combination of the results. The machining system according to claim 17 or 18, wherein the determination device includes a notification unit that notifies the third determination result determined by the third determination unit in a verifiable manner. a first judgment unit having a first judgment model and a second judgment model, which in a workability judgment step of judging the workability of a workpiece, inputs first waveform information and second waveform information extracted from time series data of an emission spectrum generated when a laser beam is irradiated to the workpiece under a first judgment irradiation condition where the workpiece is melted but not penetrated, into the first judgment model and the second judgment model, respectively, and judges the workability of the workpiece based on the first waveform information and the second waveform information;
a second determination unit that determines the workability of the workpiece based on feature amount information extracted from time series data of infrared intensity of radiated light generated when the laser light is irradiated to the workpiece under second determination irradiation conditions that do not exceed the melting point of the material of the workpiece in the workability determination step, and preregistered reference information for determining the workability of the workpiece;
a third judgment unit that judges the workability of the workpiece when it is cut and processed under preset processing conditions in the processing step based on a combination of a first judgment result judged by the first judgment unit and a second judgment result judged by the second judgment unit, and outputs a third judgment result;
a learning device that creates the first determination model and the second determination model;
Equipped with
The determination device includes:
The first determination unit,
extracting first waveform information in a first time domain and second waveform information in a second time domain from the time series data of the emission spectrum, inputting the extracted first waveform information as estimation data into the first judgment model to obtain a surface quality evaluation of the workpiece, inputting the extracted second waveform information as estimation data into the second judgment model to obtain an internal quality evaluation of the workpiece, and outputting the first judgment result that judges the workability of the workpiece based on a combination of the obtained surface quality evaluation and internal quality evaluation of the workpiece;
The second determination unit,
extracting feature amount information indicating a change in temperature of the workpiece over time or position based on the time-series data of the infrared intensity, and outputting the second judgment result that judges the workability of the workpiece based on the extracted feature amount information and the reference information;
The learning device includes:
The first waveform information and a surface quality evaluation result indicating the surface quality of the workpiece are input as first teacher data, and machine learning is performed to create the first judgment model;
the second waveform information and an internal quality evaluation result indicating the internal quality of the workpiece are input as second teacher data, and machine learning is performed to create the second judgment model;
Machinability assessment system.

Images