Estimation device, learning device, estimation method, learning method, and computer program

Abstract

An estimation device comprising: a calculation result acquisition unit that acquires magnetic feature quantities of a magnetic material to be estimated; and an estimation unit that estimates a magnetic domain refinement index for the magnetic material to be estimated by inputting the acquired magnetic feature quantities into an estimation model that has been trained to output a magnetic domain refinement index that indicates whether or not magnetic domain refinement has occurred in the magnetic material, or the degree of magnetic domain refinement, when the magnetic feature quantities of the magnetic material are input.

Description

Technical Field

[0001] The present invention relates to an estimation device, a learning device, an estimation method, a learning method, and a computer program. This application claims priority to Japanese Patent Application No. 2024-120185, filed in Japan on July 25, 2024, the content of which is incorporated herein by reference.

Background Art

[0002] Electromagnetic steel sheets are used for various purposes such as social infrastructure. It is important to ensure the material quality of electromagnetic steel sheets. Among electromagnetic steel sheets, especially in the case of grain-oriented electromagnetic steel sheets, there is a technique of reducing iron loss by irradiating a laser to subdivide magnetic domains. At the stage of quality assurance, it is necessary to confirm whether a laser has been irradiated on the electromagnetic steel sheet or whether the magnetic domains have been appropriately subdivided. Conventionally, this has been done by taking a magnetic domain image and observing the magnetic domain image (see Patent Document 1).

[0003] Note that Patent Document 2 discloses a device for measuring the magnetic properties of a ferromagnetic endless belt with high resolution. Further, Patent Document 3 discloses a device for calculating the surface hardness of a magnetic material based on the demagnetization and magnetization of the surface layer of the magnetic material and the electromagnetic characteristic values measured on the surface layer.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Patent Document 3

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in acquiring magnetic domain images using a magneto-optical element as described in Patent Document 1, it is necessary to bring the magneto-optical element and the steel plate very close together and maintain that distance. Patent Document 1 shows a mechanism in which the distance between the element and the steel plate is kept constant by pressing the steel plate with a roll and mounting the magneto-optical element on a jig connected to the roll shaft. However, with such a method, it is only possible to acquire a magnetic domain image of the central part of the plate, and it is difficult to acquire a magnetic domain image of the entire plate. Furthermore, specialized knowledge is required to determine from the magnetic domain image whether or not laser irradiation was performed appropriately. For this reason, there are limitations to increasing the number of inspections, such as 100% inspection, and inspections can only be performed on a limited number of samples. Therefore, with inspection by observing magnetic domain images, it is difficult to obtain appropriate inspection results in a short time, such as performing in-line inspection on an electrical steel plate manufacturing line.

[0006] This invention has been made in view of the above circumstances, and aims to provide an estimation device, learning device, estimation method, learning method, and computer program that can quickly estimate whether or not magnetic domain subdivision processing, such as laser irradiation or electron beam irradiation, has been performed on a magnetic material such as an electrical steel sheet. [Means for solving the problem]

[0007] One aspect of the present invention is an estimation device comprising: a calculation result acquisition unit that acquires magnetic feature quantities of a magnetic material to be estimated; and an estimation unit that estimates a magnetic domain refinement index in the magnetic material to be estimated by inputting the acquired magnetic feature quantities into an estimation model that has been trained to output a magnetic domain refinement index that indicates whether or not the magnetic domains in the magnetic material have been refined, or the degree of magnetic domain refinement, when the magnetic feature quantities of the magnetic material are input.

[0008] One aspect of the present invention is a learning device comprising: a data acquisition unit that acquires data indicating magnetic features of a magnetic material to be learned and a magnetic domain refinement index indicating whether or not the magnetic domains in the magnetic material have been refined or the degree of magnetic domain refinement; and a learning unit that generates an estimation model that takes the magnetic features of a magnetic material to be estimated as input and outputs an estimated magnetic domain refinement index by learning the relationship between the acquired magnetic features and the data indicating the magnetic domain refinement index.

[0009] One aspect of the present invention is an estimation method comprising: a calculation result acquisition step of acquiring magnetic feature quantities of a magnetic material to be estimated; and an estimation step of estimating a magnetic domain refinement index in the magnetic material to be estimated by inputting the acquired magnetic feature quantities into an estimation model that has been trained to output a magnetic domain refinement index that indicates whether or not magnetic domain refinement has occurred in the magnetic material, or the degree of magnetic domain refinement, when the magnetic feature quantities of the magnetic material are input.

[0010] One aspect of the present invention is a learning method comprising: a data acquisition step of acquiring data indicating magnetic features of a magnetic material to be learned and a magnetic domain refinement index indicating whether or not the magnetic domains in the magnetic material have been refined or the degree of magnetic domain refinement; and a learning step of generating an estimation model that takes the magnetic features of a magnetic material to be estimated as input and outputs an estimated magnetic domain refinement index by learning the relationship between the acquired magnetic features and the data indicating the magnetic domain refinement index. [Effects of the Invention]

[0011] According to the present invention, it is possible to quickly estimate whether or not magnetic domain subdivision processing has been performed. [Brief explanation of the drawing]

[0012] [Figure 1] This is a system configuration diagram showing an example of the configuration of the learning system according to the first embodiment. [Figure 2] This is a system configuration diagram showing an example of the configuration of the estimation system of the first embodiment. [Figure 3] It is a diagram showing a configuration example of a measuring device. [Figure 4] It is a first diagram showing a specific example of a magnetic feature quantity measured by a measuring device. [Figure 5] It is a second diagram showing a specific example of a magnetic feature quantity measured by a measuring device. [Figure 6] It is a diagram showing a configuration example of a learning device according to the first embodiment. [Figure 7] It is a flowchart showing an example of the learning flow of an estimation model in the learning system according to the first embodiment. [Figure 8] It is a diagram showing a configuration example of an estimation device according to the first embodiment. [Figure 9] It is a flowchart showing an example of the estimation flow in the estimation system according to the first embodiment. [Figure 10] It is a system configuration diagram showing a configuration example of a learning system according to the second embodiment. [Figure 11] It is a flowchart showing an example of the learning flow of an estimation model in the learning system according to the second embodiment. [Figure 12] It is a flowchart showing an example of the estimation flow in the estimation system according to the second embodiment.

Embodiments for Carrying Out the Invention

[0013] Hereinafter, embodiments of an estimation device, a learning device, an estimation method, a learning method, and a computer program according to the present invention will be described with reference to the drawings.

[0014] FIG. 1 is a diagram showing a configuration example of a learning system 100 according to the first embodiment. The learning system 100 is a system that learns the relationship between the magnetic characteristic quantity F of the sample steel material S (hereinafter referred to as the sample material S) and the magnetic domain subdivision index indicating whether or not magnetic domain subdivision processing has been performed, using the sample material S. The sample material S is a steel material of the same type as the steel material T (hereinafter referred to as the target material T) to be estimated by the estimation system 200 (see FIG. 2) described later, and is a steel material used as a learning sample. Hereinafter, the types of the sample material S and the target material T will be described as steel type A. The magnetic characteristic quantity F is a characteristic quantity calculated from the waveform of the current and / or the waveform of the voltage obtained by electromagnetic measurement of the steel material. Hereinafter, the waveform of the current and / or the waveform of the voltage will be referred to as the "measurement waveform". The inventors have intensively studied and found that when magnetic domain subdivision processing is performed on a magnetic material such as a steel material, the magnetic characteristic quantity F changes according to the degree of magnetic domain subdivision processing so as to indicate the state in which the magnetic domain subdivision processing has been performed, and it is possible to estimate whether or not magnetic domain subdivision processing has been performed based on one or more magnetic characteristic quantities F. Hereinafter, the case where magnetic domain subdivision processing is performed by laser irradiation will be described as an example, but magnetic domain subdivision processing may be performed by electron beam irradiation. In this case, the following laser irradiation shall be replaced with electron beam irradiation.

[0015] The magnetic domain subdivision index can be said to indicate whether or not the magnetic domains have been subdivided. Also, as described above, since magnetic domain subdivision processing is performed by laser irradiation, the magnetic domain subdivision index can be said to indicate whether or not laser irradiation has been performed.

[0016] As shown in FIG. 1, the learning system 100 includes a learning device 1. In the embodiment shown in FIG. 1, the learning system 100 includes a laser irradiation device (magnetic domain subdivision processing device) 10, a learning device 1, and a measurement device 4. The laser irradiation device 10 is a device for irradiating the sample material S with a laser in order to perform magnetic domain subdivision of the sample material S. By switching on and off the laser irradiation of the sample material S by the laser irradiation device 10, a sample material S irradiated with laser and a sample material S not irradiated with laser are created.

[0017] The measuring device 4 is a device for calculating the magnetic feature quantities F of the sample material S. The measuring device 4 magnetizes the sample material S and calculates various magnetic feature quantities F based on the measurement waveform of at least one of the current or voltage measured as a result of magnetization.

[0018] Furthermore, the learning device 1 uses the presence or absence of laser irradiation on the sample material S and the magnetic feature quantity F of the sample material S to perform a learning process that learns the relationship between the magnetic feature quantity F and the magnetic domain subdivision index for steel material of steel type A. More specifically, the learning device 1 learns the above relationship by using the set of magnetic feature quantity F and magnetic domain subdivision index acquired for each sample material S as training data.

[0019] Specifically, in the embodiment shown in Figure 1, the learning device 1 creates training data by associating and storing the magnetic domain subdivision index and the calculation result of the magnetic feature quantity F by the measuring device 4 for each sample material S. The magnetic domain subdivision index may be manually input into the learning device 1 by a person who knows whether or not each sample material S has been irradiated. Alternatively, it may be automatically transmitted to the learning device 1 by the laser irradiation device 10 and the learning device 1 being connected to each other for communication. Similarly, the magnetic feature quantity F for each sample material S may be input into the learning device 1 by a person who confirms the calculation result of the measuring device 40, or it may be automatically stored in the learning device 1 from the measuring device 40. The learning device 1 acquires the magnetic domain subdivision index and the calculation result of the magnetic feature quantity F for each sample material S, and creates training data by associating and storing this information for each sample material S. In this embodiment, the measuring device 4 is configured to output the calculation result of the magnetic feature quantity F externally, and in this embodiment, it is output to the learning device 1. However, the learning device 1 is not limited to this embodiment, and may acquire pre-created training data using another device or the like.

[0020] Through this learning process, the learning device 1 takes the calculation result of the magnetic feature quantity F of the target material T of steel type A as input and generates an estimation model M (trained model) that outputs a magnetic domain subdivision index of the target material T. The learning device 1 outputs the generated estimation model M to the estimation device 2 of the estimation system 200, which will be described next. The estimation device 2 stores the estimation model M in the memory unit 21.

[0021] Figure 2 shows an example configuration of the estimation system 200 according to the first embodiment. The estimation system 200 is a system that estimates the magnetic domain subdivision index of steel material T based on magnetic feature quantities F calculated from the target material T using the estimation model M described above. The estimation system 200 is intended to estimate the magnetic domain subdivision index inline for multiple target materials T. Note that the multiple target materials T may be a group that flows through the line in various processes from manufacturing to shipment.

[0022] As shown in Figure 2, the estimation system 200 includes an estimation device 2. In the embodiment shown in Figure 2, the estimation system 200 includes the estimation device 2 and a measurement device 4. The estimation device 2 estimates the magnetic domain subdivision index of the target material T by inputting the magnetic feature quantity F of the target material T obtained from the measurement device 4 or the like into the estimation model M. The measurement device 4 may be the same device as the learning system 100, and outputs the measurement result or the calculation result of the magnetic feature quantity F to the estimation device 2. The estimation model M is, for example, a program, and the estimation device 2 outputs a calculation result corresponding to the input by executing the program.

[0023] The magnetic domain subdivision index is a value configured to take one of two values ​​(e.g., 0 and 1), where one value indicates that laser irradiation occurred and the other indicates that laser irradiation did not occur. The magnetic domain subdivision index does not have to be expressed numerically; it may also be expressed by two different symbols or codes.

[0024] The magnetic domain subdivision index may be configured to take any value between 0 and 1, where 1 indicates that laser irradiation occurred and 0 indicates that laser irradiation did not occur. Here, a value between 0 and 1 indicates the possibility of laser irradiation occurring. The closer the magnetic domain subdivision index is to 1, the higher the probability of laser irradiation occurring, and the closer it is to 0, the higher the probability of no laser irradiation occurring. For example, the magnetic domain subdivision index can show the probability that each item falls under "laser irradiation present" and "laser irradiation absent" (e.g., laser irradiation present: 0.85, laser irradiation absent: 0.15). This probability indicates the degree of likelihood of being classified into each item, as calculated by the estimation model M based on the input information. The estimation model M may estimate that there was no laser irradiation when the magnetic domain subdivision index is between 0 and a predetermined value (e.g., 0.5), and estimate that there was laser irradiation when the magnetic domain subdivision index is between that predetermined value and 1.

[0025] If the magnetic domain subdivision index is a numerical value, the inspector may determine whether or not the target material T has been irradiated with a laser based on the numerical value output from the estimation model M.

[0026] In the example shown in Figure 2, the estimation device 2 and the measuring device 4 are configured as separate components, but the estimation device 2 and the measuring device 4 may be configured as a single unit. Furthermore, the learning device 1 of the learning system 100 and the estimation device 2 of the estimation system 200 may be configured as a single unit. When the learning device 1 and the estimation device 2 are configured as a single unit, the measurement device 4 may be the same device.

[0027] Below, we will describe the measurement device 4, the learning device 1, and the estimation device 2 in detail using Figures 3 to 9. First, the measuring device 4 will be described in detail using Figures 3 to 5.

[0028] Figure 3 shows an example of the configuration of the measuring device 4. As shown in Figure 3, the measuring device 4 comprises a main body 40, a magnetizer 41, and a detection coil 45. The main body 40 of the measuring device 4 comprises an oscillator 42, an excitation power supply 43, a magnetic field calculation unit 44, a magnetic flux density calculation unit 46, and a calculation result output unit 47. The magnetizer 41 is a device that non-contactively energizes the surface of the steel material S1 to be measured. Figure 3 shows a case in which the magnetizer 41 sequentially energizes the target area by moving the steel material S1 in the direction of the arrow, but the excitation of the steel material S1 may also be performed by moving the magnetizer 41.

[0029] The magnetizer 41 includes, for example, a yoke 411 and an excitation coil 412. The U-shaped yoke 411 has a body portion 411b and a pair of iron core portions 411a formed at both ends of the body portion 411b. The pair of iron core portions 411a are positioned with their magnetic pole ends facing the surface of the steel material S1 to be measured. The excitation coil 412 is wound around each of the pair of iron core portions 411a. With this configuration, when an alternating current flows through the excitation coil 412, the yoke 411 can generate a magnetic field of a strength corresponding to the magnitude of the alternating current on the surface of the steel material S1 positioned opposite the iron core portions 411a.

[0030] The oscillator 42 outputs a signal with a frequency corresponding to the frequency of the target AC signal. The excitation power supply 43 outputs an AC current to the excitation coil 412 corresponding to the frequency of the signal received from the oscillator 42. The excitation power supply 43 can also set the magnitude of the output AC current, i.e., the amplitude of the AC current. The magnetic field calculation unit 44 detects the magnitude of the AC current output from the excitation power supply 43 to the excitation coil 412, and calculates the strength of the magnetic field generated on the surface of the steel material S1 (magnetic field strength) from the detected magnitude of the AC current and the number of turns of the excitation coil 412 stored in advance. The magnetic field calculation unit 44 outputs the calculated magnetic field strength to the calculation result output unit 47.

[0031] The detection coil 45 is wound around at least one of the tip portions of a pair of iron core portions 411a, surrounding the tip surface that will become the magnetic pole. The magnetic flux Φ generated in the gap between the magnetic pole and the surface of the steel material S1 changes depending on the magnetic field generated by the magnetizer 41 and the state of the surface of the steel material S1. Then, a current is generated in the detection coil 45 by electromagnetic induction in response to the time change of this magnetic flux Φ. The magnetic flux density calculation unit 46 detects the voltage generated in the detection coil 45 and calculates the magnetic flux density from the detected voltage, the number of turns of the detection coil 45 which has been determined in advance, the cross-sectional area of ​​the detection coil 45, etc. The magnetic flux density calculation unit 46 outputs the calculated magnetic flux density to the calculation result output unit 47. Alternatively, the detection coil 45 may be omitted, and the voltage applied to the excitation coil 412 may be input to the magnetic flux density calculation unit 46.

[0032] The calculation result output unit 47 calculates the magnetic feature quantity F based on the magnetic field strength output from the magnetic field calculation unit 44 and the magnetic flux density output from the magnetic flux density calculation unit 46. The calculation of the magnetic feature quantity F may also be performed based on the detection signal obtained from the magnetizer 41 and the measurement conditions of the magnetizer 41. The measurement conditions of the magnetizer 41 may include, for example, controllable variable conditions such as the strength and frequency of the voltage applied to the magnetizer 41, and static conditions such as the cross-sectional area and number of turns of the coil of the magnetizer 41. The calculation result output unit 47 outputs the calculation result of the magnetic feature quantity F to the estimation device 2.

[0033] Figures 4 and 5 show specific examples of magnetic feature quantities F measured by the measuring device 4. Magnetic feature quantities F are physical quantities calculated from the measurement waveform in the measuring device 4. The voltage mentioned above may be the voltage applied to the magnetizer 41 (excitation coil 412). The current mentioned above may be the current generated in the magnetizer 41 by the application of voltage to the magnetizer 41. The voltage mentioned above may be the voltage induced in the detection coil 45 by the current flowing through the magnetizer 41. Magnetic feature quantities F include, for example, the BH feature quantity Fa, the eddy current feature quantity Fb, and the excitation waveform feature quantity Fc shown in Figure 4. The BH feature quantity Fa is a feature quantity obtained from BH hysteresis (magnetization characteristics) which shows the relationship between the magnetic field strength calculated from the current generated in the excitation coil 412 and the magnetic flux density calculated from the voltage generated in the detection coil 45 or the excitation coil 412. The eddy current feature quantity Fb is a feature quantity obtained from the impedance waveform showing eddy currents. The excitation waveform feature quantity Fc is a feature quantity obtained from the waveform of the current or voltage applied to the magnetizer 41 during excitation. The magnetic field strength and magnetic flux density generated in the detection coil 45, and the impedance of the detection coil 45, can be calculated based on the current or voltage generated in the detection coil 45 or the excitation coil 412.

[0034] For example, one example of a BH feature quantity Fa is a feature quantity relating to the waveform of change in magnetic permeability. The excitation waveform feature quantity Fc is a feature quantity of the waveform of the voltage or current applied to the steel material when measuring BH hysteresis, and the eddy current feature quantity Fb is a feature quantity of the impedance waveform when measuring BH hysteresis. That is, the eddy current feature quantity Fb and the excitation waveform feature quantity Fc can be measured simultaneously with the BH feature quantity Fa. These are classification categories of magnetic feature quantity F. That is, the BH feature quantity Fa, the eddy current feature quantity Fb, and the excitation waveform feature quantity Fc may each contain one or more types of features. At least one of these features quantities of BH feature quantity Fa, eddy current feature quantity Fb, and excitation waveform feature quantity Fc is measured by the measuring device 4.

[0035] The BH hysteresis to be measured may include both a major loop and a minor loop, and the excitation waveform in this case will be a two-frequency superimposed waveform as shown in Figure 5. The example in Figure 5 shows the superimposed waveform when the frequency for measuring the major loop is 100 Hz and the frequency for measuring the minor loop is 3 kHz. Note that the frequencies of the major loop and the minor loop may differ from those described above.

[0036] The measuring device 4 may be equipped with multiple small magnetizers 41 to locally magnetize the steel material S1 to be measured. In this case, the influence of magnetic anomalies occurring around the magnetizers 41 can be reduced. Furthermore, the waveform measurement function and the magnetic feature quantity F calculation function of the measuring device 4 may be distributed to different devices (housings). In other words, the calculation processing of the magnetic feature quantity F may be performed by a device other than the measuring device 4. For example, the measuring device 4 may output the measurement result to the learning device 1, and the learning device 1 may perform the calculation of the magnetic feature quantity F. In this case, the measuring device 4 may output the measurement result of the measurement waveform to the estimation device 2, and the estimation device 2 may perform the calculation of the magnetic feature quantity F. The measurement conditions of the magnetizer 41 shall be stored in advance in the measuring device 4 as setting information for the magnetizer 41.

[0037] Next, learning device 1 will be described in detail using Figures 6 and 7. Figure 6 shows an example configuration of the learning device 1 according to the first embodiment. The learning device 1 includes a processor such as a CPU (Central Processing Unit) connected by a bus, memory, auxiliary storage devices, etc., and executes a program. The learning device 1 functions as a device comprising a storage unit 11, an output unit 12, a data acquisition unit 13, and a learning unit 14 through program execution. Note that all or part of each function of the learning device 1 may be implemented using hardware such as an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), or an FPGA (Field Programmable Gate Array). The program may be recorded on a computer-readable recording medium. Computer-readable recording media include, for example, portable media such as flexible disks, magneto-optical disks, ROMs, CD-ROMs, and storage devices such as hard disks built into computer systems. The program may be transmitted via a telecommunications line.

[0038] The storage unit 11 is configured using a storage device such as a magnetic hard disk drive or a semiconductor storage device. The storage unit 11 stores for each sample material S the magnetic feature quantity F acquired from the measuring device 4 and whether or not laser irradiation by the laser irradiation device 10 has been performed. In addition to this information, the storage unit 11 may be used as a storage area for arbitrary information related to the operation of the learning device 1.

[0039] The output unit 12 outputs the generated estimated model M. For example, the output unit 12 may output and display the information on a display device such as a CRT (Cathode Ray Tube) display, a liquid crystal display, or an organic EL (Electro-Luminescence) display. The information to be output may be any information relating to the learning device 1. Note that displaying information on a display device is just one example of how information can be output, and is not limited to this. For example, the information may be output as sound, or transmitted to another device via communication.

[0040] The data acquisition unit 13 acquires the calculation result of the magnetic feature quantity F and the magnetic domain subdivision index of the sample material S. For example, the data acquisition unit 13 includes a communication interface and is configured to receive data from the measuring device 4 by communicating with the measuring device 4. Alternatively, the data acquisition unit 13 may include an input device such as a touch panel, mouse, or keyboard and be configured to accept input of the calculation result and magnetic domain subdivision index through these input devices. Alternatively, the data acquisition unit 13 may include a connection interface for a detachable recording medium and be configured to read the calculation result from the recording medium connected to it. The data acquisition unit 13 may acquire the magnetic domain subdivision index from the laser irradiation device 10. The communication interface included in the data acquisition unit 13 may acquire the magnetic domain subdivision index by communicating with the laser irradiation device 10. The data acquisition unit 13 may also acquire the calculation result of the magnetic feature quantity F by acquiring the measurement waveform data described above and calculating the magnetic feature quantity F.

[0041] The learning unit 14 has the function of constructing an estimation model M. More specifically, the learning unit 14 learns an estimation model M for estimating the magnetic domain subdivision index from magnetic features F by performing machine learning using training data stored in the memory unit 21. In this case, creating the training data may require mapping and labeling the calculation results of the magnetic features F to the magnetic domain subdivision index, etc., which may be done manually or automatically by the learning unit 14 according to predetermined rules. Furthermore, the machine learning algorithm used is not limited to a specific one, and any algorithm may be selected depending on the estimation target, such as support vector machines, linear regression, random forests, decision trees, k-nearest neighbors, neural networks, and deep learning. The learning unit 14 stores the estimation model M constructed by machine learning in the memory unit 21.

[0042] In training this estimation model M, the learning unit 14 uses the calculation result of magnetic feature F that includes BH hysteresis of at least one electrical angle period. This allows the estimation model M to be trained by calculating only the magnetic feature F that includes BH hysteresis of at least one electrical angle period, thus suppressing the high cost required for training. Furthermore, as a result, the learning device 1 learns the relationship with the magnetic domain subdivision index based on the magnetic feature F extracted from the entire hysteresis of BH hysteresis, so that the training cost is reduced while the estimation device 2 can estimate the magnetic domain subdivision index with sufficient accuracy.

[0043] Figure 7 is a flowchart illustrating an example of the learning flow of the estimated model M in the learning system 100. First, the laser irradiation device 10 irradiates the sample material S with a laser or not (step S101). Then, the laser irradiation device 10 outputs a magnetic domain subdivision index to the learning device 1 (step S102). Here, the magnetic domain subdivision index of the sample material S may be manually recorded in the learning device 1. Next, the measurement device 4 calculates the magnetic feature quantity F of the sample material S (S103). As mentioned above, it is preferable that the magnetic feature quantity F is calculated at a timing such that excitation for at least one period of electrical angle is completed for the magnetic feature quantity F. The calculation result of the magnetic feature quantity F obtained for the sample material S is recorded in the learning device 1 (S104). For example, the recording of the calculation result of the magnetic feature quantity F to the learning device 1 may be done manually, or it may be done automatically by connecting the measurement device 4 and the learning device 1 in a communicative manner.

[0044] Steps S101 to S104 described above are repeated for multiple sample materials S, thereby accumulating the correspondence between the calculation results of the magnetic feature quantity F for each sample material S and the magnetic domain subdivision index. Subsequently, the learning device 1 constructs an estimation model M by performing machine learning using the accumulated data of the above multiple correspondences as training data, and saves the learning results (S105). The above is the processing performed by the learning device 1.

[0045] Next, we will explain the estimation device 2 using Figures 8 and 9. Figure 8 shows an example configuration of the estimation device 2 according to the first embodiment. The estimation device 2 may be configured in the same way as the learning device 1 already described. That is, it includes a processor, memory, and auxiliary storage device connected by a bus, and executes a program. The estimation device 2 functions as a device comprising a storage unit 21, an output unit 22, a calculation result acquisition unit 23, and an estimation unit 25 by executing the program. Note that all or part of each function of the estimation device 2 may be implemented using various hardware. The program may be recorded on a computer-readable recording medium or transmitted via a telecommunications line.

[0046] The storage unit 21 is configured using a storage device such as a magnetic hard disk drive or a semiconductor storage device. The storage unit 21 stores magnetic feature quantities F and estimation models M acquired from the measurement device 4. In addition to this information, the storage unit 21 may be used as a storage area for arbitrary information related to the operation of the estimation device 2.

[0047] The output unit 22 outputs the estimation result from the estimation unit 25. The information may also be output to and displayed on a display device as described earlier. The information to be output may be any information relating to the estimation device 2. Note that displaying information on a display device is just one example of how information can be output, and is not limited to this. For example, the information may be output as sound, or transmitted to another device via communication.

[0048] The calculation result acquisition unit 23 acquires the calculation result of the magnetic feature quantity F of the target material T from the measuring device 4. For example, the calculation result acquisition unit 23 includes a communication interface and is configured to receive data from the measuring device 4 through communication with the measuring device 4. Alternatively, for example, the calculation result acquisition unit 23 may include an input device as already described and be configured to accept the input of calculation results through these input devices. Alternatively, for example, the calculation result acquisition unit 23 may include a connection interface for a detachable recording medium and be configured to read the calculation result from the recording medium connected to it. The calculation result acquisition unit 23 acquires the calculation result of the magnetic feature quantity F of the target material T from the measuring device 4 and supplies it to the estimation unit 25. Here, the calculation result acquisition unit 23 may output the calculation result directly to the estimation unit 25, or it may store the calculation result in the storage unit 21 and supply the calculation result to the estimation unit 25 via the storage unit 21.

[0049] The estimation unit 25 has the function of estimating whether or not the target material T has been irradiated with a laser. More specifically, the estimation unit 25 obtains the calculation result of the magnetic feature quantity F of the target material T from the calculation result acquisition unit 23, reads the estimation model M from the storage unit 21, and inputs the calculation result of the acquired magnetic feature quantity F into the read estimation model M, thereby obtaining the estimated result of the magnetic domain subdivision index of the target material T as an output. The estimation model M may be input with the calculation result of one type of magnetic feature quantity F, or it may be input with the calculation results of two or more types (multiple types) of magnetic feature quantities F. The estimation unit 25 displays the estimation result obtained in this way on the output unit 22 and stores it in the storage unit 21.

[0050] Figure 9 is a flowchart illustrating an example of the estimation flow in the estimation system 200 of the first embodiment. Using the example of execution on a manufacturing line, first, the magnetic feature quantity F of the target material T is calculated by the measuring device 4 upstream of the estimation device 2 on the manufacturing line (S201). In this calculation, the magnetic feature quantity F is preferably calculated based on results measured at multiple measurement points to a degree that allows for a comprehensive understanding of its periodic changes. These multiple measurement points can be defined as multiple measurement points sufficient to allow for a comprehensive understanding of the BH hysteresis curve drawn by excitation for at least one period of electrical angle. The calculation results of one or more magnetic feature quantities F obtained from one or more points acquired for the target material T are supplied to the estimation device 2 (S202). The measuring device 4 continuously calculates the magnetic feature quantity F for the estimation targets that arrive sequentially according to the flow of the line, and outputs the calculation results sequentially to the estimation device 2. Next, the estimation device 2 sequentially inputs the calculation results of the magnetic feature quantities F supplied sequentially from the measurement device 4 into the estimation model M constructed by the learning device 1, thereby sequentially obtaining estimation results of the magnetic domain subdivision index for the target material T (S203). By repeatedly executing S201 to S203 while the line is running, the estimation system 200 can perform in-line estimation of the magnetic domain subdivision index of the steel material flowing through the line, that is, estimation of whether or not laser irradiation has been performed on the steel material flowing through the line.

[0051] In this explanation, the estimation system 200 describes a case where the measurement device 4 continuously calculates the magnetic feature quantity F according to the flow of the line and outputs the calculation results sequentially to the estimation device 2. However, the calculation results are not output individually; multiple results may be output together. In this case, the estimation device 2 may input multiple calculation results output from the measurement device 4 together into the estimation model M, and the estimation model M may output the estimation results for each of the multiple calculation results together.

[0052] (Second embodiment) In the first embodiment described above, the magnetic domain subdivision index indicates whether or not subdivision of magnetic domains has occurred. However, in the second embodiment described below, the magnetic domain subdivision index is a magnetic domain subdivision value that indicates the degree of subdivision of magnetic domains. The construction of an estimation model for estimating the magnetic domain subdivision value and the estimation of the magnetic domain subdivision value based on magnetic features will be described below. Figure 10 is a system configuration diagram showing an example of the configuration of the learning system 100 of the second embodiment. The learning system 100 of the second embodiment includes an analysis device 6 in addition to the learning system 100 equipped with the learning device 1 of the first embodiment. The analysis device 6 is a device that has the function of measuring or calculating the magnetic domain subdivision value. The laser irradiation device 10 and the measuring device 4, which have the same configuration as in the first embodiment, will not be described in this document. The learning system 100 in the second embodiment uses a sample material S to learn the relationship between the magnetic features F of steel type A and the magnetic domain subdivision value that indicates the degree of subdivision of magnetic domains, and generates an estimation model M.

[0053] The magnetic domain subdivision value is configured to indicate the degree of subdivision of the magnetic domain, for example, by taking discrete values ​​within a predetermined range or continuous values ​​within a predetermined range. In this case, the magnetic domain subdivision value may be configured to show a correspondence between the magnetic domain subdivision value and the amount of change in magnetic properties, for example, by setting it based on the amount of reduction in iron loss obtained by measuring iron loss before and after laser irradiation. Magnetic properties are the magnetic properties exhibited by a magnetized magnetic material, and include typical properties such as iron loss, magnetic flux density, permeability, coercivity, and remanent magnetic flux density, which are used as indicators in material design.

[0054] The magnetic domain subdivision value may be artificially determined from the measurement results obtained by the analyzer 6. For example, the magnetic domain subdivision value may be determined based on the degree of subdivision of the magnetic domains, which is artificially determined based on the magnetic domain image of the steel material. The degree of subdivision of the magnetic domains is, for example, the magnetic domain width, and in this case, the magnetic domain subdivision value is a value determined in correspondence with the magnetic domain width.

[0055] The degree of domain refinement indicated by the magnetic domain refinement value may include the state where no domain refinement has occurred, i.e., where laser irradiation has not occurred. For example, the magnetic domain refinement value may be configured to take the values ​​of 0 or 1, or it may be configured to take any value between 0 and 1. In this case, for example, 0 indicates that no laser irradiation has occurred, and 1 indicates that laser irradiation has occurred.

[0056] In the embodiment shown in Figure 10, the analyzer 6 analyzes the sample material S and obtains the magnetic domain subdivision value of the sample material S. The analyzer 6 is a device that measures magnetic properties representing the magnetic properties of a magnetic material, such as the iron loss of the sample material S. It measures the magnetic properties of the sample material S before and after laser irradiation, and obtains the magnetic domain subdivision value by calculating it from the measurement results of the magnetic properties. The relationship between magnetic properties and magnetic domain subdivision value, such as the calculation formula for calculating the magnetic domain subdivision value from iron loss, is stored in the analyzer 6 in advance, and the analyzer 6 uses the above relationship, such as the calculation formula, to determine the magnetic domain subdivision value based on the magnetic properties. However, the present invention is not limited to this embodiment, and the analyzer 6 may be, for example, a device that takes magnetic domain images of steel, and obtain the magnetic domain subdivision value by taking magnetic domain images and calculating the reduction rate of magnetic domain width as the magnetic domain subdivision value from the magnetic domain images. The analyzer 6 may be composed of multiple devices, for example, the measurement by electromagnetic measurement and the calculation of the magnetic domain subdivision value may be functionally distributed among multiple devices. The analyzer 6 may be the same type of device as the measuring device 40.

[0057] Note that the magnetic feature quantity F calculated by the measuring device 4 does not include magnetic domain subdivision values ​​such as iron loss in steel materials. While the magnetic feature quantity F is a physical quantity calculated from the measurement waveform by the measuring device 4, the magnetic domain subdivision value cannot be calculated from the measurement waveform by the measuring device 4 and is calculated from magnetic properties that cannot be calculated from the measurement waveform, such as iron loss.

[0058] In the second embodiment, the memory unit 11 of the learning device 1 (see Figure 6) stores the magnetic feature quantity F obtained from the measuring device 4 and the magnetic domain subdivision value obtained from the analysis device 6 for each sample material S.

[0059] In the second embodiment, the information to be output by the output unit 12 of the learning device 1 is, for example, the estimated model M generated by the learning unit 14. In the second embodiment, the data acquisition unit 13 of the learning device 1 acquires the calculation results of the magnetic feature quantity F for each sample material S and the magnetic domain subdivision value data, and records them in the storage unit 11.

[0060] In the second embodiment, the learning unit 14 of the learning device 1 learns an estimation model M for estimating magnetic domain subdivision values ​​from magnetic feature F by performing machine learning using the calculation results of magnetic feature F and magnetic domain subdivision value data of the sample material S stored in the memory unit 11 as training data.

[0061] The estimation system 200 of the second embodiment has the same configuration as the estimation system 200 of the first embodiment. The estimation system 200 of the second embodiment estimates the magnetic domain subdivision value for the target material T using the relationship between the magnetic feature quantity F and the magnetic domain subdivision value. The estimation device 2 of the second embodiment obtains the calculation result of the magnetic feature quantity F of the target material T from the measurement device 4 and estimates the magnetic domain subdivision value of the target material T by inputting the measurement result into the estimation model M.

[0062] In the second embodiment, the storage unit 21 of the estimation device 2 (see Figure 6) stores the magnetic feature quantities F and estimation model M acquired from the measuring device 4.

[0063] In the second embodiment, the information to be output by the output unit 22 of the estimation device 2 is, for example, the estimation result of the magnetic domain subdivision value by the estimation unit 25. In the second embodiment, the calculation result acquisition unit 23 of the estimation device 2 acquires the calculation result of the magnetic feature quantity F of the target material T and records it in the storage unit 21.

[0064] In the second embodiment, the estimation unit 25 of the estimation device 2 has the function of estimating magnetic domain subdivision values ​​from the calculation results of the magnetic feature quantity F of the target material T.

[0065] Figure 11 is a flowchart showing an example of the learning flow of the estimation model M in the learning system 100 of the second embodiment. The measuring device 4 calculates the magnetic feature quantity F of the sample material S (S301). As described above, it is preferable that the magnetic feature quantity F is calculated at a timing such that excitation for at least one period of electrical angle is completed for the magnetic feature quantity F. The calculation result of the magnetic feature quantity F obtained for the sample material S is supplied to the estimation device 2 (S302). For example, the recording of the calculation result of the magnetic feature quantity F to the learning device 1 may be done manually, or it may be done automatically by connecting the measuring device 4 and the learning device 1 in a communicative manner. The analysis device 6 measures the magnetic domain subdivision value of the sample material S (S303). The data of the magnetic domain subdivision value obtained for the sample material S is supplied to the estimation device 2 (S304). For example, the magnetic domain subdivision value may be recorded manually to the learning device 1, or it may be recorded automatically by connecting the analysis device 6 and the learning device 1 in a communicative manner. Note that in Figure 11, steps S301 and S302 are performed before S303 and S304, but the order can be reversed. By repeating the above steps S301 to S304 for multiple sample materials S, the correspondence between the calculation results of the magnetic feature quantity F for each sample material S and the magnetic domain subdivision values ​​is accumulated.

[0066] Next, the learning device 1 constructs an estimation model M by performing machine learning using the accumulated data of the multiple correspondences described above as training data, and saves the learning results (S305).

[0067] Figure 12 is a flowchart illustrating an example of the estimation flow in the estimation system 200 of the second embodiment. Using the example of execution on a manufacturing line, first, upstream of the estimation device 2 on the manufacturing line, the measuring device 4 calculates the magnetic feature quantity F of the target material T (S401). In this calculation, it is preferable that measurements are taken at multiple measurement points to allow for a general understanding of the periodic change, thereby calculating the magnetic feature quantity F at multiple points. These multiple measurement points can be defined as multiple measurement points sufficient to allow for a general understanding of the BH hysteresis curve drawn by excitation for at least one electrical angle period. The calculation results of the magnetic feature quantity F obtained for the target material T are supplied to the estimation device 2 (S402). The measuring device 4 continuously calculates the magnetic feature quantity F according to the flow of the line and sequentially outputs the measurement results to the estimation device 2. Next, the estimation device 2 sequentially inputs the calculation results of the magnetic feature quantities F supplied sequentially from the measurement device 4 into the estimation model M constructed by the learning device 1, thereby sequentially obtaining estimation results of magnetic domain subdivision values ​​for the target material T (S403). The estimation system 200 can perform inline estimation of magnetic domain subdivision values ​​for steel materials flowing through the line by repeatedly executing S401 to S403 while the line is in operation.

[0068] The following describes the tests performed to validate the estimation model. In this validation test, an estimation model M was trained to estimate the classification problem of determining whether or not laser irradiation was performed on the measurement position on the steel surface by the measurement device 4, using 20 types of magnetic features F, which can be derived from BH hysteresis, as explanatory variables (inputs). In other words, the target variable (output) of the estimation model M in this case is a binary variable indicating whether or not laser irradiation was performed. A support vector machine was used as the machine learning algorithm.

[0069] In validation using training data, the system detected areas without laser irradiation with a 98.9% accuracy rate and a false positive rate of 1.0%. In validation using test data, the system detected areas without laser irradiation with a 97.8% accuracy rate and a false positive rate of 3.4%. The test data overfit rate is 3%, but since areas without laser irradiation are expected to appear in clusters, a rate of around 3% is not considered problematic. Furthermore, the overfit rate in the validation using the training data was 1.0%, indicating that overfitting did not occur. Thus, the learning system 100 of this embodiment enables the construction of an estimation model M with sufficient accuracy.

[0070] As described above, according to the estimation device, learning device, estimation method, learning method, and computer program of the embodiment, the magnetic feature quantity F of the target steel material (magnetic material) can be calculated, and the calculation result can be input into the estimation model to estimate whether or not laser irradiation has occurred or the magnetic domain subdivision value. This significantly reduces the time required to determine whether or not laser irradiation has occurred or the magnetic domain subdivision value of the steel material compared to observation of magnetic domain images. Therefore, rapid inspection is possible, and inline inspection can also be performed.

[0071] Furthermore, in the first embodiment, a magnetic domain refinement index is obtained during learning to indicate whether or not the magnetic domains were refined by the laser irradiation device 10. Therefore, even if the laser irradiation of the sample material S is insufficient, the calculation result of the magnetic feature quantity F obtained from the sample material S is treated as the calculation result when laser irradiation was performed. However, in the second embodiment, the magnetic domain subdivision value is obtained based on the sample material S after laser irradiation. As a result, the relationship between the calculation result of the magnetic feature quantity F and the magnetic domain subdivision value in the second embodiment is considered to have a higher correlation than the relationship between the calculation result of the magnetic feature quantity F and the magnetic domain subdivision index indicating whether or not magnetic domain subdivision has occurred in the first embodiment, and can be estimated more accurately.

[0072] More specifically, we developed a method using machine learning to estimate whether laser irradiation occurred or the magnetic domain subdivision value from magnetic feature quantities F. This allows us to estimate whether laser irradiation occurred or the magnetic domain subdivision value by utilizing a single calculation result (magnetic feature quantities F calculated based on measurements at multiple measurement points to the extent that the BH hysteresis curve drawn by at least one period of electrical angle excitation can be grasped as a whole).

[0073] While embodiments of this invention have been described in detail above with reference to the drawings, the specific configuration is not limited to these embodiments and includes designs and the like that do not depart from the spirit of this invention.

[0074] In the embodiment described above, steel is the material to be estimated, but any magnetic material that can have its magnetic feature quantity F calculated and whose magnetic feature quantity F changes upon laser irradiation can similarly be used as the material to be estimated.

[0075] In the embodiments described above, the types of sample material S and target material T were described as steel material A, but the invention is not limited to this. For example, the sample material S may contain multiple types of steel material, and the learning device 1 may perform a learning process to learn the relationship between magnetic feature quantity F and magnetic domain subdivision index for multiple types of steel material. In this case, the type of target material T is one of the types of steel material included in the sample material S.

[0076] In the embodiment described above, the estimation device 2 may perform learning and estimation based on the regions of the steel material. That is, the estimation device 2 switches whether or not to irradiate each region of the steel material with a laser, or measures the magnetic domain subdivision value for each region of the steel material. Furthermore, the estimation device 2 obtains a dataset for each region of the steel material by calculating the magnetic feature quantity F for each region of the steel material. The estimation device 2 performs learning based on this dataset and creates an estimation model M, but the estimation model M may or may not include data indicating the regions of the steel material as explanatory variables. In addition, the estimation device 2 may create different estimation models M for each region of the steel material.

[0077] Estimation device 2 calculates the magnetic feature quantity F of the target region of the steel material and inputs it into estimation model M to estimate whether or not laser irradiation occurred or the magnetic domain subdivision value. The target region is the region to be estimated within the target material T, and is at least a part of it. If estimation model M includes data indicating the region of the steel material as an explanatory variable, estimation device 2 estimates whether or not laser irradiation occurred or the magnetic domain subdivision value based on the magnetic feature quantity F and the data indicating the region of the steel material. If estimation device 2 creates a different estimation model M for each region of the steel material, estimation device 2 estimates whether or not laser irradiation occurred or the magnetic domain subdivision value using the estimation model M corresponding to the region for which the magnetic feature quantity F was calculated.

[0078] In the embodiments described above, we explained the case of determining whether or not a steel material has been irradiated with a laser, or learning and estimating the magnetic domain subdivision value of a steel material. However, the present invention is also applicable to cases where a metal other than steel, such as a magnetic material that can calculate magnetic features and measure magnetic domain subdivision values, has been irradiated with a laser, or where learning and estimating the magnetic domain subdivision value is required. [Industrial applicability]

[0079] According to the present invention, it is possible to quickly estimate whether or not magnetic domain subdivision processing has been performed. [Explanation of Symbols]

[0080] 100...Learning System 1…Learning device 10…Laser irradiation device (magnetic domain subdivision processing device) 11...Storage section 12…Output section 13...Data acquisition unit 14…Learning Department 200... Estimated System 2... Estimation device 21...Storage section 22…Output section 23…Calculation result acquisition unit 25…Estimation part 4… Measuring device 40...Main unit 41...Magnetizer 411... York 411a…Iron core part 411b... Torso 412... Excitation coil 42…Oscillator 43...Excitation power supply 44... Magnetic field calculation unit 45...Detection coil 46... Magnetic flux density calculation unit 47...Calculation result output section 6…Analyzer F... Magnetic features Fa…BH features Fb...Eddy current characteristic Fc... Excitation waveform feature quantity

Claims

1. A calculation result acquisition unit that acquires the magnetic feature quantities of the magnetic material to be estimated, An estimation unit estimates the magnetic domain refinement index in the target magnetic material by inputting the acquired magnetic features into an estimation model that has been trained to output a magnetic domain refinement index indicating whether or not magnetic domain refinement has occurred in the magnetic material, or the degree of magnetic domain refinement, when the magnetic features of the magnetic material are input. Equipped with, The aforementioned magnetic feature quantity includes the feature quantity obtained from the BH hysteresis of the magnetic material. Estimation device.

2. The aforementioned magnetic domain subdivision index indicates whether or not magnetic domain subdivision processing has been performed. The estimation device according to claim 1.

3. The aforementioned magnetic domain subdivision index is a magnetic domain subdivision value that indicates the degree of magnetic domain subdivision processing. The estimation device according to claim 1 or 2.

4. The magnetic feature quantity of the magnetic material is a magnetic feature quantity calculated based on measurement results of at least a portion of the magnetic material. The estimation device according to claim 1 or 2.

5. The magnetic material is steel. The estimation device according to claim 1 or 2.

6. The aforementioned magnetic feature quantity is a feature quantity determined based on the voltage or current at multiple measurement points within one period of the electrical angle in a magnetizer that magnetizes the object to be measured. The estimation device according to claim 1 or 2.

7. The magnetic feature quantity includes at least one of the following feature quantities calculated from the current and voltage: a BH feature quantity, an eddy current feature quantity, and an excitation waveform feature quantity, each of which includes one or more feature quantities. The estimation device according to claim 6.

8. The aforementioned magnetic feature quantity includes multiple types of magnetic feature quantities, The estimation model is a trained model that has learned the relationship between the multiple types of magnetic features and data indicating whether or not magnetic domain refinement processing has been performed on the magnetic material. The estimation device according to claim 1 or 2.

9. The magnetic feature quantities obtained by the calculation result acquisition unit and the estimation result by the estimation unit, which is whether or not magnetic domain subdivision processing has been performed or the magnetic domain subdivision value, are used to retrain the estimation model. The estimation device according to claim 1 or 2.

10. A data acquisition unit that acquires data indicating the magnetic features of the magnetic material to be studied, and a magnetic domain refinement index that indicates whether or not the magnetic domains in the magnetic material have been refined or the degree of magnetic domain refinement, The system includes a learning unit that generates an estimation model that takes the magnetic features of the target magnetic material as input and outputs an estimated magnetic domain subdivision index, by learning the relationship between the acquired magnetic features and the data indicating the magnetic domain subdivision index. The aforementioned magnetic feature quantity includes the feature quantity obtained from the BH hysteresis of the magnetic material. Learning device.

11. A calculation result acquisition step to obtain the magnetic features of the magnetic material to be estimated, An estimation step in which the magnetic domain refinement index in the target magnetic material is estimated by inputting the acquired magnetic features into an estimation model that has been trained to output a magnetic domain refinement index indicating whether or not magnetic domain refinement has occurred in the magnetic material, or the degree of magnetic domain refinement, when the magnetic features of the magnetic material are input, and thereby estimating the magnetic domain refinement index in the target magnetic material; It has, The aforementioned magnetic feature quantity includes the feature quantity obtained from the BH hysteresis of the magnetic material. Estimation method.

12. A data acquisition step to acquire data showing the magnetic features of the magnetic material to be studied, and a magnetic domain refinement index indicating whether or not the magnetic domains in the magnetic material have been refined or the degree of magnetic domain refinement, A learning step to generate an estimation model that takes the magnetic features of the target magnetic material as input and outputs an estimated magnetic domain subdivision index by learning the relationship between the acquired magnetic features and the data showing the magnetic domain subdivision index, It has, The aforementioned magnetic feature quantity includes the feature quantity obtained from the BH hysteresis of the magnetic material. Learning methods.

13. A program that causes a computer to perform the estimation method described in claim 11.

14. A program that causes a computer to execute the learning method described in claim 12.

15. The BH hysteresis includes a major loop and a minor loop, The estimation device according to claim 1 or 2.

Images