Heat treatment analysis methods, heat treatment analysis equipment, heat treatment analysis programs

Abstract

To realize a heat treatment analysis precisely and quickly.SOLUTION: A heat treatment analysis method executes a three-dimensional heat treatment analysis by the finite element method to a three-dimensional model 100 simulating a metal member, extracts an outer surface or a section to be evaluated from the three-dimensional model 100 as evaluation surfaces 101, 104, and executes processing for estimating a two-dimensional heat treatment analysis result in the evaluation surfaces 101, 104 with a computer by inputting a two-dimensional heat treatment analysis condition in the evaluation surfaces 101, 104 to an estimator 33. The estimator 33 is subjected to learning processing using partial data of a three-dimensional heat treatment analysis by the finite element method as teacher data so as to output a two-dimensional heat treatment analysis result in the evaluation surfaces 101, 104 when a two-dimensional heat treatment analysis condition in the evaluation surfaces 101, 104 is inputted.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a heat treatment analysis method, a heat treatment analysis device, and a heat treatment analysis program.

Background Art

[0002] In the heat treatment of a metal member, the metal member is heated to a certain temperature or higher and then cooled to be modified. At this time, thermal contraction due to temperature difference or transformation expansion due to phase transformation occurs in the metal member, and deformation and stress occur due to such thermal contraction and transformation expansion. The deformation of the metal member can lead to an increase in the cutting cost after heat treatment, and the stress in the metal member can lead to cracking of the metal member. Therefore, in order to reduce the cutting cost and avoid cracking of the metal member, deformation prediction and stress prediction by simulation are utilized.

[0003] For example, Patent Document 1 discloses a simulation technique for predicting deformation and stress during heat treatment using the finite element method. Patent Document 2 discloses a technique for shortening the analysis time by reducing the number of times of magnetic field calculation in a part of the simulation in a special heat treatment such as high-frequency heat treatment using the finite element method.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0005] Simulations using the finite element method require a significant amount of computation time per run. Therefore, performing numerous calculations would require an enormous amount of computation time, making practical application difficult. Consequently, there is room for improvement in achieving high-precision and rapid heat treatment analysis, particularly in cases requiring many calculations, such as the search for optimal heat treatment conditions.

[0006] The objective of this invention is to achieve high-precision and rapid heat treatment analysis. [Means for solving the problem]

[0007] A first aspect of the present invention is, We performed a 3D heat treatment analysis using the finite element method on a 3D model that simulates a metal component. From the aforementioned 3D model, the outer surface or cross-section to be evaluated is extracted as the evaluation surface. The results of the two-dimensional heat treatment analysis on the evaluation surface are estimated by inputting the two-dimensional heat treatment analysis conditions on the evaluation surface into the estimator. A heat treatment analysis method that includes performing a process by computer, The estimator provides a heat treatment analysis method in which, when the two-dimensional heat treatment analysis conditions on the evaluation surface are input, a learning process is performed using partial data from a three-dimensional heat treatment analysis using the finite element method as training data to output the two-dimensional heat treatment analysis results on the evaluation surface.

[0008] This method allows for the acquisition of heat treatment analysis results at a lower computational cost compared to computationally expensive finite element method simulations, by utilizing an estimator, enabling faster heat treatment analysis. Furthermore, because it uses some data from 3D heat treatment analysis using the finite element method (data on the evaluation surface) as training data, it is possible to achieve estimation with the same high accuracy as the finite element method while reducing the size of the training data. In other words, by employing a technique called a surrogate model, which replaces finite element method simulations with machine learning, it is possible to achieve high-precision and rapid heat treatment analysis that enables the search for optimal heat treatment conditions. In addition, since the input and output of the estimator are data from 2D heat treatment analysis, the size of data handled can be reduced compared to 3D heat treatment analysis. For example, the above method can be used to search for optimal heat treatment conditions for avoiding quench cracking, reducing residual stress, or reducing heat treatment deformation.

[0009] The two-dimensional heat treatment analysis conditions may include at least one of the carbon component concentration distribution, quenching temperature, or heat transfer coefficient on the evaluation surface. The two-dimensional heat treatment analysis results may include at least one of the stress distribution, displacement distribution, or microstructure distribution after heat treatment on the evaluation surface.

[0010] This configuration allows for the rapid search for optimal heat treatment conditions for at least one of the factors that significantly influence the results of two-dimensional heat treatment analysis: carbon component concentration distribution, quenching temperature, or heat transfer coefficient.

[0011] The evaluation surface may be extracted to include the area where the stress or displacement is maximum in the three-dimensional heat treatment analysis.

[0012] With this configuration, since areas where stress or displacement is maximum in 3D heat treatment analysis are likely to become crack initiation points, it is useful to be able to perform heat treatment analysis on evaluation surfaces including these areas with high accuracy and in a short time.

[0013] A second aspect of the present invention is, A three-dimensional heat treatment analysis unit that performs three-dimensional heat treatment analysis by the finite element method on a three-dimensional model simulating a metal member, A cross-section extraction unit that extracts an outer surface or a cross-section to be evaluated from the three-dimensional model as an evaluation surface, An estimator that has been subjected to learning processing using part of the data of the three-dimensional heat treatment analysis by the finite element method so as to output the two-dimensional heat treatment analysis result on the evaluation surface when the two-dimensional heat treatment analysis conditions on the evaluation surface are input, A heat treatment analysis device comprising:

[0014] A third aspect of the present invention is Performing three-dimensional heat treatment analysis by the finite element method on a three-dimensional model simulating a metal member, Extracting an outer surface or a cross-section to be evaluated from the three-dimensional model as an evaluation surface, Inputting the two-dimensional heat treatment analysis conditions on the evaluation surface to the estimator and estimating the two-dimensional heat treatment analysis result on the evaluation surface, A heat treatment analysis program that causes a computer to execute a process including: The estimator has been subjected to learning processing using part of the data of the three-dimensional heat treatment analysis by the finite element method so as to output the two-dimensional heat treatment analysis result on the evaluation surface when the two-dimensional heat treatment analysis conditions on the evaluation surface are input. A heat treatment analysis program is provided.

Advantages of the Invention

[0015] According to the present invention, heat treatment analysis can be realized with high accuracy and in a short time.

Brief Description of the Drawings

[0016] [Figure 1] Schematic configuration diagram of a heat treatment analysis device according to an embodiment of the present invention. [Figure 2] Perspective view showing a three-dimensional model. [Figure 3] Perspective view showing the carbon component concentration distribution of a three-dimensional model. [Figure 4]Perspective view showing the carbon component concentration distribution of the three-dimensional model cut along the A-A line (first cross-section) of FIG. 2. [Figure 5] Perspective view showing the carbon component concentration distribution of the three-dimensional model cut along the B-B line (second cross-section) of FIG. 2. [Figure 6] Perspective view showing the X-direction stress distribution of the three-dimensional model. [Figure 7] Perspective view showing the X-direction stress distribution of the three-dimensional model cut at the first cross-section. [Figure 8] Perspective view showing the X-direction stress distribution of the three-dimensional model cut at the second cross-section. [Figure 9] Carbon component concentration distribution of the first cross-section. [Figure 10] Graph showing the X-direction stress by the finite element method and the estimator at the center point of the first cross-section. [Figure 11] Mises stress distribution by the finite element method of the first cross-section. [Figure 12] Mises stress distribution by the estimator of the first cross-section. [Figure 13] Z-direction displacement distribution by the finite element method of the first cross-section. [Figure 14] Z-direction displacement distribution by the estimator of the first cross-section. [Figure 15] Carbon component concentration distribution of the bottom surface. [Figure 16] Graph showing the X-direction stress by the finite element method and the estimator at the center point of the bottom surface. [Figure 17] Mises stress distribution by the finite element method of the bottom surface. [Figure 18] Mises stress distribution by the estimator of the bottom surface. [Figure 19] Z-direction displacement distribution by the finite element method of the bottom surface. [Figure 20] Z-direction displacement distribution by the estimator of the bottom surface. [Figure 21] Flowchart of the heat treatment analysis method according to an embodiment of the present invention.

Embodiments for Carrying Out the Invention

[0017] Embodiments of the present invention will be described below with reference to the attached drawings.

[0018] Referring to Figure 1, the heat treatment analysis apparatus 1 of this embodiment performs heat treatment analysis on a metal member. Heat treatment analysis broadly refers to analyses associated with heat treatment and may include, for example, heat transfer analysis, phase transformation analysis, and elastoplastic analysis. Heat transfer analysis analyzes the cooling behavior of the metal member, phase transformation analysis analyzes the transformation behavior of the metal member, and elastoplastic analysis analyzes the deformation behavior and stress behavior of the metal member.

[0019] The heat treatment analysis apparatus 1 of this embodiment is composed of a computer such as a desktop PC, laptop computer, workstation, or tablet terminal. The heat treatment analysis apparatus 1 includes an input / output interface unit 10, a storage unit 20, and a control unit 30.

[0020] The input / output interface unit 10 functions as an input device for receiving information from the user and an output device for outputting information to the user. The input / output interface unit 10 includes at least one human-machine interface. The human-machine interface includes, for example, an input device such as a keyboard, a pointing device (e.g., a mouse or trackball), or a touchpad, and an output device such as a display or speaker. The human-machine interface also includes an input / output device such as an in-cell touch panel display (e.g., a liquid crystal panel or an organic EL panel).

[0021] The storage unit 20 is a storage medium capable of storing various types of information. The storage unit 20 can be implemented as, for example, a memory such as DRAM, SRAM, or flash memory, an HDD, an SSD, or other storage devices, or a combination thereof as appropriate. The storage unit 20 stores programs for implementing various processes performed by the control unit 30, which will be described later. The storage unit 20 also stores various types of data, which will be described later, a model for heat treatment analysis, and other information necessary for heat treatment analysis. The programs executed by the control unit 30 may be provided by a communication unit that communicates according to a predetermined communication standard, or they may be stored in a portable storage medium.

[0022] The control unit 30 performs arithmetic processing and controls the entire device. The control unit 30 may consist of dedicated electronic circuits designed to realize predetermined functions, or hardware circuits such as reconfigurable electronic circuits. Alternatively, the control unit 30 may consist of various semiconductor integrated circuits. Such various semiconductor integrated circuits include, for example, CPUs, MPUs, microcomputers, DSPs (Digital Signal Processors), FPGAs (Field Programmable Gate Arrays), and ASICs (Application Specific Integrated Circuits).

[0023] The control unit 30 includes, functionally, a 3D heat treatment analysis unit 31, a cross-section extraction unit 32, and an estimator 33. These are realized through the cooperation of hardware resources such as the circuits and software resources such as programs.

[0024] Referring to Figure 2, the 3D heat treatment analysis unit 31 performs a 3D heat treatment analysis using the finite element method on a 3D model 100 that simulates a metal member. In the illustrated example, the 3D model 100 is a rectangular parallelepiped with dimensions of 80 × 80 × 30 [mm] in length (X direction) × width (Y direction) × height (Z direction), which is divided into multiple small cubic elements.

[0025] In 3D heat treatment analysis, various material properties are set for the 3D model 100. For example, these material properties include density, coefficient of linear expansion, and modulus of elasticity. Then, various boundary conditions corresponding to the desired heat treatment analysis are set. For example, these boundary conditions include the heat transfer coefficient. In addition, various initial conditions corresponding to the desired heat treatment analysis are set. For example, these initial conditions include the concentration distribution of various components (e.g., carbon component concentration distribution) and the quenching temperature. Hereafter, the various material properties, boundary conditions, and initial conditions in 3D heat treatment analysis will also be referred to as 3D heat treatment analysis conditions.

[0026] Referring to Figure 3, various carbon content concentrations can be set for the 3D model 100 depending on the part. Carbon content refers to the carbon content (mass percentage concentration) in the metallic material. In the illustrated example, the darker the color, the lower the carbon content concentration. Specifically, the darkest part has a carbon content concentration of 0.15%, and the lightest part has a carbon content concentration of 0.45%.

[0027] In the 3D model 100, the carbon content is set high in specific areas 110 located approximately one-quarter of the way from the end in the vertical (X direction) and in specific areas 120 located approximately three-quarters of the way from the end in the vertical (X direction). In other words, the carbon content is set low in the basic areas 130 other than these specific areas 110 and 120.

[0028] Figure 4 shows a cross-section 101 (hereinafter also referred to as the first cross-section 101) at a position one-quarter of the way from the vertical (X-direction) end of the 3D model 100. The first cross-section 101 corresponds to the cross-section along line AA in Figure 2. Figure 5 shows a cross-section 102 (hereinafter also referred to as the second cross-section 102) at a position three-quarters of the way from the vertical (X-direction) end of the 3D model 100. The second cross-section 102 corresponds to the cross-section along line BB in Figure 2. In the illustrated example, as shown in the first cross-section 101 and the second cross-section 102, the distribution of carbon component concentration is uniform in the Z direction. Therefore, the top surface 103 and the bottom surface 104 in Figure 3 show the same carbon component concentration distribution.

[0029] In the first cross-section 101 shown in Figure 4, the black area corresponding to the basic part 130 and the white area corresponding to the specific part 110 are mixed together, meaning that the carbon content varies greatly. Surfaces with such large variations in carbon content often experience problems such as cracking. In contrast, in the second cross-section 102 shown in Figure 5, the entire surface is composed of the black area corresponding to the basic part 130, and the carbon content is roughly uniform. Surfaces with such relatively uniform carbon content rarely experience problems such as cracking.

[0030] In this embodiment, the 3D heat treatment analysis unit 31 outputs the stress distribution in the 3D model 100 based on the 3D model 100 and the 3D heat treatment analysis conditions. Alternatively, the 3D heat treatment analysis unit 31 may output the displacement distribution, or the microstructure distribution after heat treatment, or a combination thereof. Hereafter, the output result of the 3D heat treatment analysis unit 31 will also be referred to as the 3D heat treatment analysis result.

[0031] Figures 6-8 show stress distributions as an example of 3D heat treatment analysis results. This was achieved by setting the carbon component concentration distribution shown in Figures 3-5 to the 3D model 100 shown in Figure 2, and setting the heat transfer coefficient to 500 [W / m²]. 2 This result was obtained by setting the temperature to [K] and the quenching temperature to 820 [°C].

[0032] Figure 6 shows the entire 3D model 100, similar to Figure 3. Figure 7 shows the 3D model 100 with the first cross-section 101 removed for visualization. Figure 8 shows the 3D model 100 with the second cross-section 102 removed for visualization. In these figures, the whiter the color, the greater the stress in the X direction. For example, the darkest colored area has an X-direction stress of -200 [MPa], while the whitest colored area has an X-direction stress of 450 [MPa].

[0033] Referring to Figure 6, high X-direction stress (white area) was not observed on the outer surface of the 3D model 100. Referring to Figure 7, high X-direction stress (white area) was observed in the first section 101, particularly as a maximum value. Referring to Figure 8, high X-direction stress (white area) was not observed in the second section 102. In the first section 101, where high X-direction stress was observed, problems such as cracking often occur, making the first section 101 a preferred target for evaluation. This is an example regarding stress, but the same applies to displacement; in outer surfaces or sections where high displacement is observed, deformation-related problems often occur, making such surfaces preferred for evaluation.

[0034] The cross-section extraction unit 32 extracts an outer surface or cross-section to be evaluated from the 3D model 100 as an evaluation surface. The evaluation surface may be a surface that has been manually set in advance, or it may be extracted automatically or manually based on the results of the 3D heat treatment analysis. In the latter case, for example, the evaluation surface may be extracted to include the part of the 3D model 100 where the stress is maximum (e.g., the first cross-section 101). Alternatively, the evaluation surface may be extracted to include the part of the 3D model 100 where the displacement is maximum. Furthermore, there may be one evaluation surface or multiple evaluation surfaces.

[0035] In this embodiment, the cross-section extraction unit 32 extracts the outer surface (bottom surface) 104, which has been manually set in advance, as an evaluation surface, and automatically or manually extracts the first cross-section 101 as an evaluation surface so as to include the area where the stress is maximum in the 3D model 100 based on the results of the 3D heat treatment analysis. That is, in this embodiment, two evaluation surfaces 101 and 104 are extracted.

[0036] The estimator 33 is trained using partial data from a 3D heat treatment analysis using the finite element method as training data, so that when 2D heat treatment analysis conditions for evaluation surfaces 101 and 104 of the 3D model 100 are input, it outputs the 2D heat treatment analysis results for evaluation surfaces 101 and 104. Specifically, the 3D heat treatment analysis conditions for evaluation surfaces 101 and 104 are used as training data for the 2D heat treatment analysis conditions, and the 3D heat treatment analysis results for evaluation surfaces 101 and 104 are used as training data for the 2D heat treatment analysis results. The training data is trained on a sufficient amount of various data to guarantee highly accurate estimation results. Possible training methods include decision trees, linear regression, random forests, support vector machines, Gaussian process regression, or convolutional neural networks. If necessary, an estimator may be prepared before or during training, and this estimator may be trained using the above training data.

[0037] In this embodiment, the two-dimensional heat treatment analysis conditions include at least one of the carbon component concentration distribution, quenching temperature, or heat transfer coefficient on the evaluation surfaces 101 and 104. The two-dimensional heat treatment analysis results also include at least one of the stress distribution, displacement distribution, and microstructure distribution after heat treatment on the evaluation surfaces 101 and 104.

[0038] In this embodiment, the results of the two-dimensional heat treatment analysis on the evaluation surfaces 101 and 104 are estimated by inputting the two-dimensional heat treatment analysis conditions on the evaluation surfaces 101 and 104 to the trained estimator 33.

[0039] Referring to Figures 9-14, the results of verifying the estimation accuracy by the estimator 33 on the evaluation surface (first cross-section) 101 will be explained. Specifically, the stress estimated by this embodiment was compared with the stress obtained by the finite element method simulation.

[0040] The analysis conditions are basically the same as those for the 3D heat treatment analysis described above, but since the data from the aforementioned 3D heat treatment analysis was used as training data, only the carbon component concentration was changed from the aforementioned 3D heat treatment analysis. Figure 9 shows the carbon component concentration distribution on the evaluation surface (first cross section) 101. When this change is confirmed in the figure, comparing the first cross section 101 in Figure 4 with the first cross section 101 in Figure 9, the black and white colors are swapped. That is, in Figure 9, contrary to the carbon component concentration in the first cross section 101 in Figure 4, the carbon component concentration of the specific part 110 is set higher than the carbon component concentration of the basic part 130.

[0041] Figure 10 is a graph showing the stress in the X direction at the center point of the evaluation surface (first cross-section) 101, calculated using the finite element method and the estimator 33. The horizontal axis represents time t [sec], and the vertical axis represents the stress in the X direction σx [MPa]. The solid line shows the result using the finite element method, and the consecutive black circles show the result using the estimator 33. The two results were found to be in good agreement, numerically confirming the high estimation accuracy of the estimator 33.

[0042] Figure 11 shows the von Mises stress distribution obtained by the finite element method on the evaluation surface (first section) 101, and Figure 12 shows the von Mises stress distribution obtained by the estimator 33 on the evaluation surface (first section) 101. In these figures, whiter areas indicate higher von Mises stress. For example, the black areas have a von Mises stress of 19 [MPa], while the white areas have a von Mises stress of 492 [MPa]. The two results were found to be in good agreement, visually confirming the high estimation accuracy of the estimator 33.

[0043] Figure 13 shows the Z-direction displacement distribution obtained by the finite element method on the evaluation surface (first section) 101, and Figure 14 shows the Z-direction displacement distribution obtained by the estimator 33 on the evaluation surface (first section) 101. In these figures, the darker the color, the greater the Z-direction displacement. For example, the black area has a Z-direction displacement of 0.27 [mm], and the white area has a Z-direction displacement of 0 [mm]. The Z-direction displacements shown here are absolute values. The two results were found to be in good agreement, visually confirming the high estimation accuracy of the estimator 33.

[0044] Referring to Figures 15-20, the results of verifying the estimation accuracy by the estimator 33 on the evaluation surface (bottom surface) 104 will be explained. Specifically, the stress estimated by this embodiment was compared with the stress obtained by the finite element method simulation.

[0045] The analysis conditions are basically the same as those for the 3D heat treatment analysis described above, but since the data from that 3D heat treatment analysis was used as training data, only the carbon component concentration was changed from that of the 3D heat treatment analysis. Figure 15 shows the carbon component concentration distribution on the evaluation surface (bottom surface) 104. When this change is confirmed in the figure, comparing the top surface 103 in Figure 3 with the bottom surface 104 in Figure 15, the black and white colors are swapped. That is, in Figure 15, contrary to the carbon component concentration on the top surface 103 in Figure 3, the carbon component concentrations in specific areas 110 and 120 are set higher than the carbon component concentration in the basic area 130.

[0046] Figure 16 is a graph showing the stress in the X direction at the center point of the evaluation surface (bottom surface) 104, calculated using the finite element method and the estimator 33. The horizontal axis represents time t [sec], and the vertical axis represents the stress in the X direction σx [MPa]. The solid line shows the result using the finite element method, and the consecutive black circles show the result using the estimator 33. The two results were found to be in good agreement, numerically confirming the high estimation accuracy of the estimator 33.

[0047] Figure 17 shows the von Mises stress distribution obtained by the finite element method on the evaluation surface (bottom surface) 104, and Figure 18 shows the von Mises stress distribution obtained by the estimator 33 on the evaluation surface (bottom surface) 104. In these figures, whiter areas indicate higher von Mises stress. For example, the black areas have a von Mises stress of 13 [MPa], while the white areas have a von Mises stress of 600 [MPa]. The two results were found to be in good agreement, visually confirming the high estimation accuracy of the estimator 33.

[0048] Figure 19 shows the Z-direction displacement distribution obtained by the finite element method on the evaluation surface (bottom surface) 104, and Figure 20 shows the Z-direction displacement distribution obtained by the estimator 33 on the evaluation surface (bottom surface) 104. In these figures, the darker the color, the greater the Z-direction displacement. For example, the black area has a Z-direction displacement of 0.03 [mm], and the white area has a Z-direction displacement of 0.008 [mm]. The Z-direction displacements shown here are absolute values. The two results were found to be in good agreement, visually confirming the high estimation accuracy of the estimator 33.

[0049] Referring to Figure 21, the heat treatment analysis method of this embodiment will be explained again.

[0050] In the heat treatment analysis method of this embodiment, first, a three-dimensional heat treatment analysis is performed on a three-dimensional model 100 (see Figure 2) that simulates the metal member to be subjected to heat treatment analysis (step S1). This three-dimensional heat treatment analysis contributes to the extraction of evaluation surfaces and the generation of training data.

[0051] Next, the outer surface or cross-section to be evaluated is extracted (step S2). For example, the evaluation surface may be extracted to include the area where the stress or displacement is maximum in the 3D heat treatment analysis.

[0052] Next, estimation is performed on the extracted evaluation surface using the estimator 33 (step S3). The estimator 33 may be input with at least one of the carbon component concentration distribution, quenching temperature, or heat transfer coefficient on the evaluation surface as a two-dimensional heat treatment analysis condition, and at least one of the stress distribution, displacement distribution, or microstructure distribution after heat treatment on the evaluation surface may be estimated as a two-dimensional heat treatment analysis result.

[0053] Comparing the calculation times for the verifications in Figures 9-14 and 15-20 using computers with roughly the same specifications, the finite element method simulation took several thousand seconds, while in this embodiment it was completed in just a few tens of seconds. Therefore, this embodiment confirms that heat treatment analysis can be achieved in an extremely short time.

[0054] According to this embodiment, the following effects and advantages are achieved.

[0055] Compared to simulations using the computationally expensive finite element method, the use of estimator 33 allows for obtaining heat treatment analysis results at a lower computational cost, enabling heat treatment analysis in a shorter time. Furthermore, because some data from 3D heat treatment analysis using the finite element method (data on the evaluation surface) is used as training data, it is possible to reduce the size of the training data while achieving estimation with the same high accuracy as the finite element method. In other words, by employing a method called a surrogate model, which replaces finite element method simulations with machine learning, it is possible to achieve high-precision and rapid heat treatment analysis that enables the search for optimal heat treatment conditions. In addition, since the input and output of estimator 33 are data from 2D heat treatment analysis, the size of the data handled can be reduced compared to 3D heat treatment analysis. For example, the above method can be used to search for optimal heat treatment conditions for avoiding quench cracking, reducing residual stress, or reducing heat treatment deformation.

[0056] Furthermore, it is possible to quickly search for the optimal heat treatment conditions for at least one of the factors that greatly influence the results of two-dimensional heat treatment analysis: carbon component concentration distribution, quenching temperature, or heat transfer coefficient.

[0057] Furthermore, since areas where stress or displacement is maximum in 3D heat treatment analysis are prone to crack initiation, it is useful to be able to perform heat treatment analysis on the evaluation surface 101, including these areas, with high accuracy and in a short time.

[0058] This embodiment may also be implemented by a heat treatment analysis program that causes a computer to execute each step constituting the heat treatment analysis method described above. By executing such a heat treatment analysis program, it is possible to obtain the effects and benefits of the heat treatment analysis method described above. In other words, it can be said that the heat treatment analysis method described above is being used.

[0059] Although specific embodiments of the present invention have been described above, the present invention is not limited to the above embodiments and can be implemented with various modifications within the scope of this invention. [Explanation of Symbols]

[0060] 1. Heat treatment analysis device 10 Input / Output Interface Section 20 Memory section 30 Control Unit 31 Three-Dimensional Heat Treatment Analysis Department 32 Cross section extraction part 33 Estimator 100 3D models 101 Cross-section (First cross-section) (Evaluation surface) 102 Cross section (2nd cross section) 103 Top surface (outer surface) 104 Bottom surface (outer surface) (evaluation surface) 110 Specific parts 120 Specific parts 130 Basic parts

Claims

1. We performed a 3D heat treatment analysis using the finite element method on a 3D model that simulates a metal component. From the aforementioned three-dimensional model, the outer surface or cross-section to be evaluated is extracted as the evaluation surface. The results of the two-dimensional heat treatment analysis on the evaluation surface are estimated by inputting the two-dimensional heat treatment analysis conditions on the evaluation surface to the estimator. A heat treatment analysis method that includes performing a process by computer, A heat treatment analysis method wherein the estimator undergoes a learning process using partial data from a three-dimensional heat treatment analysis as training data, so that when the two-dimensional heat treatment analysis conditions on the evaluation surface are input, it outputs the two-dimensional heat treatment analysis results on the evaluation surface.

2. The two-dimensional heat treatment analysis conditions include at least one of the carbon component concentration distribution, quenching temperature, or heat transfer coefficient on the evaluation surface. The heat treatment analysis method according to claim 1, wherein the two-dimensional heat treatment analysis result includes at least one of the stress distribution, displacement distribution, or microstructure distribution after heat treatment on the evaluation surface.

3. The heat treatment analysis method according to claim 1, wherein the evaluation surface is extracted to include the region where the stress or displacement is maximum in the three-dimensional heat treatment analysis.

4. A 3D heat treatment analysis unit performs 3D heat treatment analysis using the finite element method on a 3D model that simulates a metal component, A cross-section extraction unit extracts an outer surface or cross-section to be evaluated from the three-dimensional model as an evaluation surface, When the two-dimensional heat treatment analysis conditions for the evaluation surface are input, the estimator is trained using partial data from a three-dimensional heat treatment analysis using the finite element method as training data, so that it outputs the two-dimensional heat treatment analysis results for the evaluation surface. A heat treatment analysis device equipped with the following features.

5. We performed a 3D heat treatment analysis using the finite element method on a 3D model that simulates a metal component. From the aforementioned three-dimensional model, the outer surface or cross-section to be evaluated is extracted as the evaluation surface. The estimator is given the two-dimensional heat treatment analysis conditions for the evaluation surface and estimates the two-dimensional heat treatment analysis results for the evaluation surface. A heat treatment analysis program that causes a computer to perform a process including the following, The estimator is a heat treatment analysis program that has undergone a learning process using partial data from a three-dimensional heat treatment analysis using the finite element method as training data, so that when the two-dimensional heat treatment analysis conditions on the evaluation surface are input, it outputs the two-dimensional heat treatment analysis results on the evaluation surface.

Images