Manufacturing condition determination method, manufacturing method of metallic materials and manufacturing condition determination device

Abstract

The manufacturing condition determination method includes: a defect prediction step, which predicts the presence or absence of defects in each specified range by inputting manufacturing conditions for each specified range into the defect prediction model; a characteristic prediction step, which predicts the characteristics of each specified range by inputting manufacturing conditions for each specified range into the characteristic prediction model; and a manufacturing condition determination step, which extracts the manufacturing conditions predicted as defect-free by the defect prediction model from multiple manufacturing conditions input into the defect prediction model and the characteristic prediction model, and determines the manufacturing conditions of the metal material as the manufacturing conditions of the extracted manufacturing conditions whose characteristics are predicted to be the maximum or minimum by the characteristic prediction model.

Description

[Technical Field]

[0001] This invention relates to a method for determining manufacturing conditions, a method for manufacturing metallic materials, and an apparatus for determining manufacturing conditions. [Previous Technology]

[0002] Patent Document 1 discloses a method for predicting the quality of a metallic material within a specified range by inputting manufacturing conditions for each specified range of the metallic material in each step into a quality prediction model. [Prior Art Documents] [Patent Documents]

[0003] Patent Document 1: International Publication No. 2021 / 014804 [Summary of the Invention]

[0004] [Problem to be solved by the invention] The method disclosed in Patent Document 1 does not mention improving the properties of the metallic material within a range that does not produce defects or within a range that allows for the production of defects, taking into account the occurrence of product defects. There is room for improvement.

[0005] This invention has been made in view of the above circumstances, and its object is to provide a method for determining manufacturing conditions, a method for manufacturing a metallic material, and an apparatus for determining manufacturing conditions that can determine optimal manufacturing conditions for the properties of a metallic material while suppressing the generation of defects. [Means for Solving the Problem]

[0006] To solve the aforementioned problems and achieve the objective, the manufacturing condition determination method of the present invention is used to improve the quality of metal materials manufactured through one or more steps. The manufacturing condition determination method includes: a defect prediction step, which predicts the presence or absence of defects in each specified range by inputting manufacturing conditions of each specified range into a defect prediction model, wherein the defect prediction model is generated by taking the manufacturing conditions of each specified range of the metal material in each step as input variables and the presence or absence of defects in each specified range as output variables; a characteristic prediction step, which predicts the characteristics of each specified range by inputting manufacturing conditions of each specified range into a characteristic prediction model, wherein the characteristic prediction model is generated by taking the manufacturing conditions of each specified range of the metal material in each step as input variables and the characteristics of each specified range as output variables; and a manufacturing condition determination step, which extracts manufacturing conditions predicted as defect-free by the defect prediction model from a plurality of manufacturing conditions input into the defect prediction model and the characteristic prediction model, and determines the manufacturing conditions of the metal material as the manufacturing conditions of the metal material by the manufacturing conditions in which the characteristics predicted by the characteristic prediction model are the maximum or minimum.

[0007] Furthermore, in the manufacturing condition determination method of the present invention, the manufacturing condition determination step, when setting the same manufacturing conditions for the specified range covering the entire length of the metal material, extracts manufacturing conditions from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model whose proportion of the specified range predicted to be defective covering the entire length of the metal material is within an acceptable range, and determines the manufacturing condition of the metal material as the manufacturing condition of the extracted manufacturing conditions whose statistical value of the characteristic predicted by the characteristic prediction model is the largest or smallest.

[0008] Furthermore, in the manufacturing condition determination method of the present invention, the manufacturing condition determination step, when setting the same manufacturing conditions for the specified range covering the entire length of the metal material, extracts manufacturing conditions from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model whose proportion of the specified range predicted to be defective covering the entire length of the metal material is within an acceptable range, extracts manufacturing conditions from the extracted manufacturing conditions that are predicted to be defect-free within the specified range, and determines the manufacturing condition of the metal material as the manufacturing condition of the manufacturing condition whose statistical value of the characteristic predicted by the characteristic prediction model is the largest or smallest.

[0009] In addition, in the manufacturing condition determination method of the present invention, the manufacturing condition determination step, when setting the same manufacturing conditions for the entire length of the metal material, extracts the manufacturing conditions that are predicted to be defect-free from the multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model within the specified range, and determines the manufacturing condition of the metal material as the manufacturing condition of the extracted manufacturing conditions whose statistical value of the characteristic predicted by the characteristic prediction model is the largest or smallest.

[0010] In addition, in the manufacturing condition determination method of the present invention, when the same manufacturing conditions are set for the specified range covering the entire length of the metal material, the manufacturing condition that is predicted to be defect-free and with characteristics within the allowable range among the multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model is determined as the manufacturing condition of the metal material.

[0011] In order to solve the aforementioned problems and achieve the objective, the method for manufacturing metal materials of the present invention manufactures metal materials by means of manufacturing conditions determined by the manufacturing condition determination method.

[0012] To solve the aforementioned problem and achieve the objective, the manufacturing condition determining apparatus of the present invention is used to improve the quality of a metal material manufactured through one or more steps. The manufacturing condition determining apparatus includes: a defect prediction unit, which predicts the presence or absence of defects in each specified range by inputting manufacturing conditions within each specified range into a defect prediction model, wherein the defect prediction model is generated by taking the manufacturing conditions within each specified range of the metal material in each step as input variables and the presence or absence of defects in each specified range as output variables; a characteristic prediction unit, which predicts the characteristics of each specified range by inputting manufacturing conditions within each specified range into a characteristic prediction model, wherein the characteristic prediction model is generated by taking the manufacturing conditions within each specified range of the manufacturing conditions in each step as input variables and the characteristics of each specified range as output variables; and a manufacturing condition determining unit, which extracts manufacturing conditions predicted as defect-free by the defect prediction model from among multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model, and determines the manufacturing conditions for the metal material as those whose characteristics are predicted to be maximum or minimum from the extracted manufacturing conditions. [Effects of the Invention]

[0013] By means of the present invention, the manufacturing conditions of the metal material can be determined in order to suppress the occurrence of defects in the product.

Implementation Method

[0015] The manufacturing condition determination method, the manufacturing method of metallic materials, and the manufacturing condition determination apparatus of the present invention will be described with reference to the drawings.

[0016] (Manufacturing Condition Determining Device) The structure of the manufacturing condition determining device of this embodiment will be described with reference to FIG1. ​​The manufacturing condition determining device is a device used to determine the manufacturing conditions for improving the quality of a metallic material manufactured through one or more steps (processes). Hereinafter, examples of determining the manufacturing conditions for metallic materials, especially steel materials (e.g., slabs, steel plates, etc.), will be described, but the manufacturing condition determining device is not limited to metallic materials and can be used to determine the manufacturing conditions for all other materials.

[0017] Specifically, the manufacturing condition determination device 1 is implemented using a general-purpose information processing device such as a personal computer or workstation. In addition, the manufacturing condition determination device 1 uses, for example, a processor including a central processing unit (CPU) and memory (main storage) including random access memory (RAM) or read-only memory (ROM) as its main components.

[0018] As shown in FIG1, the manufacturing condition determination device 1 includes a manufacturing performance collection unit 11, a manufacturing performance editing unit 12, a multi-step continuous performance editing unit 13, a performance database 14, a defect prediction model generation unit 15, and a defect prediction unit 16. Furthermore, as shown in FIG1, the manufacturing condition determination device 1 includes a characteristic prediction model generation unit 17, a characteristic prediction unit 18, and a manufacturing condition determination unit 19. Moreover, the manufacturing condition determination model generation device is composed of the components in the manufacturing condition determination device 1 other than the defect prediction unit 16, the characteristic prediction unit 18, and the manufacturing condition determination unit 19. Hereinafter, the manufacturing condition determination model generation device will also be described in the description of the manufacturing condition determination device 1.

[0019] A sensor (not shown) is connected to the manufacturing performance collection unit 11. The manufacturing performance collection unit 11 collects the manufacturing performance of each step according to the measurement cycle of the sensor and outputs it to the manufacturing performance editing unit 12.

[0020] As for the "manufacturing results", examples may include the manufacturing conditions of each step, the presence or absence of defects in the metal material manufactured through each step (hereinafter referred to as "defect presence or absence"), and the characteristics of the metal material manufactured through each step. In addition, as for the "manufacturing conditions", examples may include the composition of the metal material, temperature, pressure, plate thickness, and plate throughput speed in each step.

[0021] Furthermore, as for the "presence or absence of defects in the metallic material," information related to whether a corresponding part of the metallic material contains defects, and the probability that a corresponding part of the metallic material contains defects (defect occurrence probability), can be listed, for example. Furthermore, as for "defects in the metallic material," surface damage and poor crystallization of the metallic material can be listed, for example. Furthermore, the "characteristics of the metallic material" show, for example, values ​​representing the material properties of the metallic material. As for the characteristics of the metallic material, iron loss and tensile strength of the metallic material can be listed, for example.

[0022] Furthermore, the manufacturing conditions collected by the manufacturing performance collection unit 11 for each step include not only the measured values ​​of the manufacturing conditions measured by the sensors, but also the preset values ​​of the manufacturing conditions. That is, depending on the step, sometimes the sensors are not installed, so in this case, the preset values ​​are collected as manufacturing performance instead of the measured values.

[0023] The manufacturing performance collection unit 11 collects the manufacturing conditions of each step for each specified range of the predetermined metal material. Furthermore, the manufacturing performance collection unit 11 evaluates and collects the presence or absence of defects in the metal material manufactured through each step for each specified range. Additionally, the manufacturing performance collection unit 11 evaluates and collects the characteristics of the metal material manufactured through each step for each specified range.

[0024] The term "specified range" refers to a certain range along the long side of the metal material, for example, when the metal material is a slab or steel plate. The specified range is determined based on the distance (passing speed) of the metal material moving in each step corresponding to the conveying direction. The specific processing details based on the manufacturing performance collection unit 11 will be described later (see Figure 2).

[0025] Furthermore, in this embodiment, the following situation is explained: a model (defect prediction model, characteristic prediction model) is created to evaluate each specified range of the metal material, and the manufacturing conditions for each specified range are determined by predicting the presence or absence of defects and characteristics. However, the metal material as a whole can also be evaluated. That is, the manufacturing condition determination device 1 can also create a model by collecting manufacturing data for each metal material (e.g., each slab) and predicting the presence or absence of defects and characteristics, thereby determining the manufacturing conditions for each metal material.

[0026] Here, in the structure shown in FIG1, it is assumed that only one manufacturing performance collection unit 11 is provided, which collects the manufacturing performance data (hereinafter referred to as "performance data") of each step. However, for example, multiple manufacturing performance collection units 11 may be provided according to the number of each step, and the performance data of each step may be collected by different manufacturing performance collection units 11 respectively.

[0027] The manufacturing performance editing unit 12 edits the performance data of each step input from the manufacturing performance collection unit 11. That is, the manufacturing performance editing unit 12 edits the performance data collected by the manufacturing performance collection unit 11 in time units into performance data in length units of the metal material, and outputs it to the multi-step continuous performance editing unit 13. The specific processing based on the manufacturing performance editing unit 12 will be described later (see Figure 2).

[0028] The multi-step continuous performance editing unit 13 edits the performance data input from the manufacturing performance editing unit 12. The multi-step continuous performance editing unit 13 stores the manufacturing conditions of each step, the presence or absence of defects in the metal material manufactured under the manufacturing conditions, and the characteristics of the metal material manufactured under the manufacturing conditions in a related manner in the performance database 14 for each specified range.

[0029] The defect prediction model generation unit 15 generates a defect prediction model that predicts the presence or absence of defects in the metal material manufactured under the specified manufacturing conditions for each step, based on the manufacturing conditions within each specified range stored in the performance database 14. For example, if the output relative to the manufacturing conditions of each step input to the defect prediction model is the probability of defect occurrence, and the probability of defect occurrence exceeds a preset threshold, the defect prediction unit 16 predicts that a defect exists. The defect prediction model generation unit 15 sets the threshold value in such a way that no undetected defects exist, and generates the defect prediction model.

[0030] The defect prediction model generation unit 15 uses a combination of linear regression (partial least squares (PLS) regression, which emphasizes generalization, and random forest, which may also consider nonlinearity) as a machine learning method. In addition, various methods such as linear regression, local regression, principal component regression, PLS regression, neural networks, regression trees, and random forests can be used.

[0031] The defect prediction unit 16 uses a defect prediction model generated by the defect prediction model generation unit 15 to predict the presence or absence of defects in a metal material manufactured under arbitrary manufacturing conditions. The defect prediction unit 16 predicts the presence or absence of defects in each specified range by inputting manufacturing conditions within each specified range into the defect prediction model. The defect prediction model is generated by using the manufacturing conditions within each specified range of the metal material in each step as input variables and the presence or absence of defects in each specified range as output variables. For example, the output relative to the manufacturing conditions of each step input into the defect prediction model is the probability of defect occurrence, and if the probability of defect occurrence exceeds a preset threshold, it is predicted as defective; if the probability of defect occurrence does not exceed the threshold, it is predicted as defect-free.

[0032] The characteristic prediction model generation unit 17 generates a characteristic prediction model that predicts the characteristics of the metal material manufactured under the manufacturing conditions based on the manufacturing conditions within each specified range of each step stored in the performance database 14.

[0033] The feature prediction model generation unit 17 uses a combination of linear regression (PLS regression, which emphasizes generalization, and random forest, which can also consider nonlinearity) as a machine learning method. In addition, various methods such as linear regression, local regression, principal component regression, PLS regression, neural networks, regression trees, and random forests can be used.

[0034] The characteristic prediction unit 18 uses the characteristic prediction model generated by the characteristic prediction model generation unit 17 to predict the characteristics of the metal material manufactured under arbitrary manufacturing conditions. The characteristic prediction unit 18 predicts the characteristics of each specified range by inputting each specified range of manufacturing conditions into the characteristic prediction model, which is generated by taking each specified range of manufacturing conditions in each step as input variables and the characteristics of each specified range as output variables.

[0035] The manufacturing condition determination unit 19 determines the manufacturing conditions of the metal material based on the prediction results of the defect prediction unit 16 and the characteristic prediction unit 18. The specific processing based on the manufacturing condition determination unit 19 will be described later (see Figure 2).

[0036] (Manufacturing Condition Determination Method) The manufacturing condition determination method of this embodiment will be described with reference to Figures 2 to 8. The manufacturing condition determination method of this embodiment performs the processes shown in Figure 2, steps S1 to S8.

[0037] First, the manufacturing performance collection unit 11 collects performance data related to the manufacturing conditions, defects, and characteristics of the metal materials in each step (step S1). The manufacturing performance collection unit 11 collects performance data on the manufacturing conditions, defects, and characteristics of each metal material and each step.

[0038] The performance data collected by the manufacturing performance collection unit 11, as shown in the table in Figure 3, is data containing performance values ​​(or setting values) of multiple manufacturing conditions arranged by time. The performance data shown in Figure 3 includes items such as time t1, t2...; the speed of the metal material (pass-through speed) v1, v2... at that time; and multiple manufacturing conditions x11, x12..., x21, x22... measured by sensors at that time.

[0039] Furthermore, in the final step of the multiple steps, the performance data collected includes, in addition to the items shown in Figure 3, items related to the presence and characteristics of defects in the metallic material. At this time, the presence and characteristics of defects in the metallic material may also include information related to the presence or absence of multiple different defects and information related to multiple different characteristics.

[0040] Next, the manufacturing performance editing unit 12 converts the performance data collected by the manufacturing performance collection unit 11 into length units of metal materials (step S2). That is, the manufacturing performance editing unit 12 converts the performance data collected in time units as shown in FIG3 into performance data in length units of metal materials as shown in FIG4. Hereinafter, the method of converting the performance data of FIG3 into the performance data of FIG4 will be explained.

[0041] First, the manufacturing performance editing unit 12 calculates the position (position coordinates) of the metal material at each time point in Figure 3 using the property that multiplying time by speed (passing speed) results in distance. Next, the manufacturing performance editing unit 12 detects the front and rear ends of the metal material using the property that performance data is recorded when the metal material passes through the sensors installed in each step and a defect value is recorded when the metal material fails to pass. Then, except for the case where the metal material fails to pass through the sensors, the manufacturing performance editing unit 12 generates performance data corresponding to the position from the front end to the rear end of the metal material.

[0042] Furthermore, in this state, although the data is in units of length of the metal material, it is not data with a constant period. Therefore, the manufacturing record editing unit 12 performs, for example, linear interpolation, to convert it into data in units of length of the metal material with a constant period. That is, in each step, when the metal material passes through the plate at a slow speed, the data that can be collected is finer, and when the metal material passes through the plate at a fast speed, the data that can be collected is coarser. Therefore, the manufacturing record editing unit 12 performs the aforementioned interpolation to ensure that the coarseness of the data is consistent. By performing the above-described processing, the manufacturing record editing unit 12 produces data in units of length of the metal material as shown in FIG. 4.

[0043] Subsequently, the multi-step continuous performance editing unit 13 uses the length unit of the metal material to make the performance data of all steps consistent and combined (step S3). The multi-step continuous performance editing unit 13 uses the length unit of the metal material produced by the manufacturing performance editing unit 12 to make the performance data of the length unit of the metal material produced by the manufacturing performance editing unit 12 consistent and combined.

[0044] Thus, the multi-step continuous performance editing unit 13 associates the manufacturing conditions of the metal material in each step, the presence or absence of defects in the metal material manufactured under the said manufacturing conditions, and its characteristics for each specified range along the length direction of the metal material, and stores them in the performance database 14. An example of the processing based on the multi-step continuous performance editing unit 13 will be described below.

[0045] For example, as shown in Figure 5, consider the case where a metal material is manufactured via steps 1, 2, and 3. Steps 1 to 3 are, for example, rolling steps, in which the length of the material increases along its long side each time a step is performed. Furthermore, as shown in Figure 5, when moving from step 1 to step 2, material A is divided into material A1 and material A2, and when moving from step 2 to step 3, material A1 is divided into material A11 and material A12.

[0046] The multi-step continuous performance editing unit 13 takes into account the performance data in each step, and combines the performance data of all steps collected in detail along the long side direction by a sensor (not shown) with the length unit of the metal material in the final step. At this time, as shown in FIG5, the material lengths of steps 2 and 1 are scaled according to the material length of step 3, which is the final step (see the dashed line in FIG5).

[0047] Furthermore, the multi-step continuous performance editing unit 13 determines the location obtained by collecting each metal material, and within each specified range of the metal material in the final step, associates the presence or absence of defects within the specified range, the characteristics of the specified range, and the manufacturing conditions of all steps within the specified range, and saves them in the performance database 14. For example, in Figure 5, the shaded area obtained by collecting material A11 in step 3, which is the final step, is traced back to material A1 in step 2 and material A in step 1. By repeatedly performing this process on all metal materials, as shown in Figure 6, performance data is created that combines and aligns the performance data of multiple manufacturing conditions, presence or absence of defects, and characteristics of the metal materials in all steps, using the length unit of the metal material. Hereinafter, returning to Figure 2, we continue the explanation after step S4.

[0048] The defect prediction model generation unit 15 generates a defect prediction model that predicts the presence or absence of defects in each specified range of the metal material based on the manufacturing conditions of each specified range of the metal material in each step (step S4). Then, the defect prediction unit 16 uses the defect prediction model generated by the defect prediction model generation unit 15 to predict the presence or absence of defects in each specified range of the metal material manufactured under arbitrary manufacturing conditions (step S5).

[0049] In step S5, the defect prediction unit 16 predicts the presence or absence of defects for each manufacturing condition by inputting multiple manufacturing conditions of the same type into the defect prediction model. For example, if the manufacturing condition is temperature, multiple manufacturing conditions (e.g., α℃, α+10℃, α+20℃, etc.) after adding (or subtracting) a predetermined value from the temperature as a reference are input into the defect prediction model. Furthermore, when multiple manufacturing conditions (e.g., temperature, pressure, etc.) are input, multiple manufacturing conditions after adding (or subtracting) a predetermined value from each manufacturing condition are similarly input into the defect prediction model.

[0050] Next, the characteristic prediction model generation unit 17 generates a characteristic prediction model that predicts the characteristics of the metal material for each specified range based on the manufacturing conditions for each specified range of the metal material in each step (step S6). Next, the characteristic prediction unit 18 uses the characteristic prediction model generated by the characteristic prediction model generation unit 17 to predict the characteristics of the metal material manufactured under arbitrary manufacturing conditions for each specified range (step S7).

[0051] In step S7, the characteristic prediction unit 18 predicts the characteristics of each manufacturing condition by inputting multiple manufacturing conditions of the same type into the characteristic prediction model. For example, if the manufacturing condition is temperature, multiple manufacturing conditions (e.g., α℃, α+10℃, α+20℃, etc.) after adding (or subtracting) a predetermined value from the temperature as a reference are input into the characteristic prediction model. Furthermore, when multiple manufacturing conditions (e.g., temperature, pressure, etc.) are input, multiple manufacturing conditions after adding (or subtracting) a predetermined value from each manufacturing condition are similarly input into the characteristic prediction model.

[0052] Next, the manufacturing condition determination unit 19 determines the manufacturing condition for the metal material from among multiple manufacturing conditions that does not produce defects and has optimal (or within acceptable limits) characteristics (step S8). In step S8, the manufacturing condition determination unit 19 determines the manufacturing condition for the metal material from among multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model that is predicted by the defect prediction model to be defect-free and by the characteristic prediction model to have specified characteristics. In step S8, the manufacturing condition for the metal material can be determined by various methods. Each method will be explained below.

[0053] <First Method> In the first method, it is assumed that different manufacturing conditions are set for different ranges of metallic materials. In this method, firstly, manufacturing conditions predicted as defect-free by the defect prediction model are extracted from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model. Then, the manufacturing conditions whose characteristics are predicted to be maximum or minimum by the characteristic prediction model are determined as the manufacturing conditions for the metallic material. Furthermore, the "maximum or minimum" also includes values ​​close to the maximum and values ​​close to the minimum, respectively.

[0054] For example, in the case of a characteristic such as tensile strength that is judged to be better the larger the value, the manufacturing conditions for the metallic material will be determined by the manufacturing conditions under which the characteristic predicted by the characteristic prediction model is maximized. Conversely, in the case of a characteristic such as iron loss that is judged to be better the smaller the value, the manufacturing conditions for the metallic material will be determined by the manufacturing conditions under which the characteristic predicted by the characteristic prediction model is minimized.

[0055] Figure 7 shows an example of a mapping of defect prediction results with a specified range (20 segments) covering the entire length of the metal material as the horizontal axis and 5 levels of manufacturing conditions as the vertical axis. That is, Figure 7 is the result of inputting five manufacturing conditions (α℃, α+10℃, α+20℃, α+30℃, α+40℃) into the defect prediction model and predicting the presence or absence of defects. In Figure 7, blocks indicated by white (hereinafter referred to as "white blocks") show the areas predicted to be defect-free (specified range), and blocks indicated by black (dotted shading) show the areas predicted to be defective.

[0056] In this method, as shown in FIG7, manufacturing conditions predicted to be defect-free (e.g., α℃, α+10℃, α+20℃, α+30℃) are extracted from a defined range (e.g., range number 1) of the metallic material. Then, the characteristics predicted for the defined range are compared, and the manufacturing conditions corresponding to the range containing the characteristics predicted to have the best values ​​are determined as the manufacturing conditions for range number 1. Then, the same process is performed on the remaining range numbers 2 to 20 to determine the manufacturing conditions for each range.

[0057] Furthermore, in this method, the characteristics are predicted only for the specified range that is predicted to be defect-free. That is, the characteristic prediction model is only input with the manufacturing conditions corresponding to the specified range that is predicted to be defect-free by the defect prediction model (e.g., "α℃", "α+10℃", "α+20℃", "α+30℃" in Figure 7).

[0058] <Second Method> In the second method, it is assumed that the same manufacturing conditions are set for a specified range covering the entire length of the metal material. In this method, firstly, manufacturing conditions that are predicted to have a specified range of defects covering the entire length of the metal material within an acceptable range (e.g., within 20%) are extracted from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model. Then, the manufacturing conditions with the largest or smallest statistical value of the characteristic predicted by the characteristic prediction model are determined as the manufacturing conditions of the metal material. Furthermore, as "statistical value", for example, the average, maximum, minimum, mode, etc. can be listed. In addition, the "maximum or minimum" also includes values ​​close to the maximum and values ​​close to the minimum, respectively.

[0059] For example, in the case of a characteristic such as tensile strength that is judged to be better the larger the value, the manufacturing condition for the metallic material is determined to be the manufacturing condition where the statistical value of the characteristic predicted by the characteristic prediction model is the largest. Conversely, in the case of a characteristic such as iron loss that is judged to be better the smaller the value, the manufacturing condition for the metallic material is determined to be the manufacturing condition where the statistical value of the characteristic predicted by the characteristic prediction model is the smallest.

[0060] Figure 8 shows an example of a mapping of defect prediction results with a specified range (20 segments) covering the entire length of the metal material as the horizontal axis and 5 levels of manufacturing conditions as the vertical axis. That is, Figure 8 is the result of inputting five manufacturing conditions (α℃, α+10℃, α+20℃, α+30℃, α+40℃) into the defect prediction model and predicting the presence or absence of defects. In Figure 8, blocks indicated by white (hereinafter referred to as "white blocks") show the areas predicted to be defect-free (specified range), and blocks indicated by black (dotted shading) show the areas predicted to be defective.

[0061] In this method, as shown in FIG8, manufacturing conditions with a defect rate within an acceptable range (e.g., within 20%) are extracted throughout the entire length of the metal material. In FIG8, as shown in the thick box, two manufacturing conditions, "α℃" and "α+10℃", are extracted. Then, statistical values ​​of the predicted characteristics for each specified range (all blocks) of "α℃" are calculated. Similarly, statistical values ​​of the predicted characteristics for each specified range (all blocks) of "α+10℃" are calculated. Then, the statistical values ​​of the characteristics of "α℃" and "α+10℃" are compared, and the manufacturing conditions corresponding to the statistical value of the characteristic with the best value are determined as the manufacturing conditions throughout the entire length of the metal material.

[0062] Furthermore, in this method, only manufacturing conditions that are predicted to be defective within a specified range and whose proportion is within the allowable range (e.g., within 20%) are used to predict the characteristics. That is, only manufacturing conditions that are predicted to be defective within a specified range and whose proportion is within the allowable range (e.g., all blocks of "α℃" and all blocks of "α+10℃" in Figure 8) are input into the characteristic prediction model.

[0063] <Third Method> In the third method, it is assumed that the same manufacturing conditions are set for a specified range covering the entire length of the metal material. In this method, firstly, among multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model, the manufacturing conditions that are predicted to have a specified range of defects covering the entire length of the metal material within an acceptable range (e.g., within 20%) are extracted. Then, the manufacturing conditions that are predicted to have a specified range without defects are extracted. Next, the manufacturing conditions whose statistical values ​​(e.g., average, maximum, minimum, mode, etc.) of the characteristics predicted by the characteristic prediction model are the maximum or minimum among the extracted manufacturing conditions are determined as the manufacturing conditions of the metal material. Furthermore, the "maximum or minimum" also includes values ​​close to the maximum and values ​​close to the minimum, respectively.

[0064] For example, in the case of a characteristic such as tensile strength that is judged to be better the larger the value, the manufacturing condition that maximizes the statistical value of the characteristic predicted by the characteristic prediction model is determined as the manufacturing condition for the metallic material. Conversely, in the case of a characteristic such as iron loss that is judged to be better the smaller the value, the manufacturing condition that minimizes the statistical value of the characteristic predicted by the characteristic prediction model is determined as the manufacturing condition for the metallic material.

[0065] In this method, as shown in FIG8, manufacturing conditions with a defect rate within an acceptable range (e.g., within 20%) are extracted throughout the entire length of the metal material. In FIG8, as shown in the thick box, two manufacturing conditions, "α℃" and "α+10℃", are extracted. Then, manufacturing conditions that are predicted to be defect-free (white blocks) within each specified range (all blocks) of "α℃" are extracted, and statistical values ​​of the predicted characteristics for the extracted specified range are calculated. Similarly, manufacturing conditions that are predicted to be defect-free (white blocks) within each specified range (all blocks) of "α+10℃" are extracted, and statistical values ​​of the predicted characteristics for the extracted specified range are calculated. Then, the statistical values ​​of the characteristics of "α℃" are compared with the statistical values ​​of the characteristics of "α+10℃", and the manufacturing conditions corresponding to the statistical value of the characteristic with the best value are determined as the manufacturing conditions throughout the entire length of the metal material.

[0066] Furthermore, in this method, only manufacturing conditions that are predicted to be defective within a specified range and within an acceptable range (e.g., less than 20%) and are predicted to be defect-free are used to predict the characteristics. That is, only manufacturing conditions that are predicted to be defective within a specified range and within an acceptable range by the defect prediction model (e.g., the white patch "α℃" and the white patch "α+10℃" in Figure 8) are input into the characteristic prediction model.

[0067] <Fourth Method> In the fourth method, it is assumed that the same manufacturing conditions are set for a specified range covering the entire length of the metallic material. In this method, firstly, manufacturing conditions that are predicted to be defect-free within a specified range are extracted from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model. Then, the manufacturing conditions whose statistical values ​​(e.g., average, maximum, minimum, mode, etc.) of the characteristics predicted by the characteristic prediction model are the maximum or minimum are determined as the manufacturing conditions for the metallic material. Furthermore, the "maximum or minimum" also includes values ​​close to the maximum and values ​​close to the minimum, respectively.

[0068] In this method, as shown in Figure 7, manufacturing conditions for the specified range (white block) predicted to be defect-free within each specified range (all blocks) of "α℃" are extracted, and statistical values ​​of the predicted characteristics for the extracted specified range are calculated. Similarly, for "α+10℃", "α+20℃", "α+30℃", and "α+40℃", manufacturing conditions for the specified range predicted to be defect-free (white block) are extracted, and statistical values ​​of the predicted characteristics for the extracted specified range are calculated. Then, the statistical values ​​of the characteristics of "α℃", "α+10℃", "α+20℃", "α+30℃", and "α+40℃" are compared respectively. Finally, the manufacturing conditions corresponding to the statistical value of the characteristic with the best value are determined as the manufacturing conditions covering the entire length of the metal material.

[0069] <Fifth Method> In the fifth method, it is assumed that the same manufacturing conditions are set for a specified range covering the entire length of the metallic material. In this method, among the multiple manufacturing conditions input into the defect prediction model and the characteristic prediction model, the manufacturing condition that predicts the largest number of specified ranges with no defects and characteristics within the allowable range is determined as the manufacturing condition of the metallic material.

[0070] Furthermore, it is also assumed that the entire specified range is predicted to be "defective" by the defect prediction model. In this case, the characteristics predicted by the characteristic prediction model are not considered, and only the probability of defect occurrence calculated by the defect prediction model is used for evaluation. That is, the manufacturing conditions that minimize the probability of defect occurrence for each specified range predicted to be defective are determined as the manufacturing conditions for the metallic material.

[0071] (Method for manufacturing metallic materials) The manufacturing condition determination method of the embodiment can also be applied to the method for manufacturing metallic materials. In this case, in the method for manufacturing metallic materials, the manufacturing conditions are determined by the manufacturing condition determination method of the embodiment, and the metallic material is manufactured by the determined manufacturing conditions.

[0072] (Example) An embodiment of the manufacturing condition determination method will be described with reference to FIGS. 9 to 11. In this embodiment, the manufacturing condition determination method of the embodiment was used to perform offline verification when making a directional electromagnetic steel sheet from a metal material.

[0073] In this embodiment, the target variable for defect prediction is whether the manufactured directional electromagnetic steel sheet exhibits poor secondary recrystallization. Furthermore, the explanatory variables (manufacturing conditions) for defect prediction include the chemical composition of the metallic material in the steelmaking step, the temperature of the slab cut after casting in the casting step, and the extraction temperature of the slab in the slab heating step in the heating furnace. Moreover, the explanatory variables (manufacturing conditions) for defect prediction include the temperature of the steel sheet in the hot rolling step (the surface temperature of the sheet at the entry point of the precision press, the middle of the press, and the exit point of the press), the temperature of the steel sheet in the cooling step, the temperature of the steel sheet in the cold rolling step, and the temperature of the metallic material in the annealing step. Here, for stable manufacturing of directional electromagnetic steel sheets, it is considered important to control secondary recrystallization.

[0074] The target variable for characteristic prediction in this embodiment is the iron loss of the manufactured directional electromagnetic steel sheet. Furthermore, the explanatory variables (manufacturing conditions) for characteristic prediction include the chemical composition of the metal material in the steelmaking step, the temperature of the slab cut after casting in the casting step, and the extraction temperature of the slab in the slab heating step in the heating furnace. Moreover, the explanatory variables (manufacturing conditions) for characteristic prediction include the temperature of the steel sheet in the hot rolling step (the surface temperature of the sheet at the fine pressing mill inlet, the middle of the mill, and the pressing mill outlet), the temperature of the steel sheet in the cooling step, the temperature of the steel sheet in the cold rolling step, and the temperature of the metal material in the annealing step, etc. Additionally, the iron loss is set as the value of W17 / 50 of the power loss per unit weight when energized at a maximum magnetic flux density of 1.7 T at a commercial frequency of 50 Hz.

[0075] The defect and characteristic prediction method uses a combination of linear regression and random forest. Only the operation determined in this embodiment, namely the temperature of the annealing step, is set as a linear term; all other terms are set as nonlinear terms. The number of training data for the defect prediction model and the characteristic prediction model is 2000 pieces, and the number of validation data is 700 pieces.

[0076] Figure 9 shows the learning results of defect prediction. Figure 9 is the frequency distribution of the probability of defect occurrence in the case of defects and the case of no defects. In this embodiment, in order to avoid undetected defects, the threshold value used to determine the presence or absence of defects is set to "0.35". Under the above premise, the temperature of the annealing step is oscillated from α℃ to α+40℃ in increments of 10℃, and the temperature that does not produce defects and has the best characteristics is determined from these five temperatures.

[0077] As Example 1, the first method of the manufacturing condition determination method of the embodiment was used to verify the manufacturing conditions that do not produce defects and minimize iron loss within each specified range of the electromagnetic steel sheet (verification quantity: 700 pieces). Figure 10 shows the verification results of Example 1, and illustrates the frequency distribution of the actual iron loss value and the calculated iron loss value of the electromagnetic steel sheet. Furthermore, the iron loss value was standardized using the average value of the actual iron loss. As shown in Figure 10, the average value of the standardized iron loss corresponding to the prior art without considering the generation of defects is 100, and the average value of the standardized iron loss corresponding to the present invention is 93. Therefore, by applying the present invention, it is expected that the iron loss value will be improved [7].

[0078] As Example 2, using the second method of the manufacturing condition determination method of the embodiment, the proportion of defects covering the entire length of the electromagnetic steel sheet was set to "less than 5%" to determine whether it was within the allowable range, and the manufacturing condition with minimal iron loss was verified (700 pieces were verified). Figure 11 shows the verification results of Example 2, and shows the frequency distribution of the actual and calculated iron loss values ​​of the electromagnetic steel sheet. Furthermore, the iron loss value was standardized using the average value of the actual iron loss. As shown in Figure 11, the average value of the standardized iron loss corresponding to the prior art is 100, and the average value of the standardized iron loss corresponding to the present invention is 95. Therefore, by applying the present invention, it is expected that the iron loss value will be improved by "5".

[0079] In the manufacturing condition determination method, metal material manufacturing method, and manufacturing condition determination apparatus described above, a defect prediction model and a characteristic prediction model are generated that correlate the manufacturing conditions of each step with the presence and characteristics of defects in the metal material manufactured under said manufacturing conditions within a specified range. Then, the manufacturing conditions of the metal material are determined based on the prediction results generated by the generated defect prediction model and characteristic prediction model, thereby allowing for characteristic improvement within a range that does not produce defects, taking into account the occurrence of product defects.

[0080] Furthermore, in the manufacturing condition determination method, the metal material manufacturing method, and the manufacturing condition determination apparatus of the embodiments, the presence or absence of defects and characteristics of the product manufactured under the determined manufacturing conditions are predicted based on multiple manufacturing conditions within a defined range for the metal material. Then, by comparing the characteristics (characteristic prediction values) corresponding to the manufacturing conditions predicted to be defect-free, manufacturing conditions within a defined range that do not produce defects and have optimal characteristics can be determined.

[0081] Furthermore, in the manufacturing condition determination method, the method for manufacturing a metallic material, and the manufacturing condition determination apparatus of the embodiments, within a predetermined range covering the entire length of the metallic material, the presence or absence of defects and characteristics of the product manufactured under the stated manufacturing conditions are predicted based on multiple manufacturing conditions within each predetermined range. Then, manufacturing conditions with a defective proportion covering the entire length of the metallic material within an acceptable range are extracted, and statistical values ​​of the predicted characteristic values ​​for each predetermined range are calculated and compared. In this way, within a predetermined range covering the entire length of the metallic material, manufacturing conditions with a defect occurrence proportion below the acceptable value and optimal characteristics can be determined.

[0082] In addition, in the manufacturing condition determination method, the manufacturing method of the metal material and the manufacturing condition determination device of the embodiment, in a specified range covering the entire length of the metal material, the presence or absence of defects and characteristics after inputting multiple manufacturing conditions for each specified range, the corresponding defect-free specified range characteristic prediction value is the manufacturing condition with the maximum number of specified ranges within the allowable range.

[0083] Here, in the manufacturing condition determination method of the embodiment, the metal material is divided into a specified range and a model (defect prediction model, characteristic prediction model) is made to determine the manufacturing conditions for each specified range, but it is also possible to determine the manufacturing conditions without dividing it into a specified range.

[0084] In this case, for example, the defect prediction unit 16 predicts the presence or absence of defects in the metal material by inputting manufacturing conditions into a defect prediction model, which is generated by taking the manufacturing conditions of the metal material in each step as input variables and the presence or absence of defects in the metal material as output variables. Additionally, the characteristic prediction unit 18 predicts the characteristics of the metal material by inputting manufacturing conditions into a characteristic prediction model, which is generated by taking the manufacturing conditions of the metal material in each step as input variables and the characteristics of the metal material as output variables. Furthermore, the manufacturing condition determination unit 19 extracts multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model, and determines the manufacturing conditions that are predicted to be defect-free by the defect prediction model. Then, the manufacturing condition determination unit 19 determines the manufacturing conditions for the metal material based on the manufacturing conditions from the extracted manufacturing conditions whose characteristics are predicted to be the maximum or minimum by the characteristic prediction model. Moreover, the "maximum or minimum" also includes values ​​close to the maximum and values ​​close to the minimum, respectively.

[0085] Thus, in the manufacturing condition determination method of the embodiment, even if the manufacturing conditions are determined without dividing the metal material into a specified range, the characteristics can be improved within a range without causing defects, taking into account the occurrence of product defects.

[0086] The manufacturing condition determination method, the method for manufacturing metallic materials, and the manufacturing condition determination apparatus of the present invention have been specifically described above through forms and embodiments for carrying out the invention. However, the spirit of the present invention is not limited to these descriptions and must be interpreted broadly based on the descriptions in the claims. In addition, various modifications and alterations made based on these descriptions are naturally included in the spirit of the present invention. [Simplified Explanation of the Diagram]

[0014] FIG1 is a block diagram showing an example of the structure of the manufacturing condition determination apparatus according to an embodiment of the present invention. FIG2 is a flowchart showing an example of the process of the manufacturing condition determination method according to an embodiment of the present invention. FIG3 is a diagram showing an example of the performance data collected by the manufacturing performance collection unit in the manufacturing condition determination method according to an embodiment of the present invention. FIG4 is a diagram showing an example of the performance data edited by the manufacturing performance editing unit in the manufacturing condition determination method according to an embodiment of the present invention. FIG5 is a diagram showing an example of manufacturing a metal material through multiple steps in the manufacturing condition determination method according to an embodiment of the present invention. FIG6 is a diagram showing an example of the performance data edited by the multi-step continuous performance editing unit in the manufacturing condition determination method according to an embodiment of the present invention. FIG7 is a diagram showing the prediction results of the presence or absence of defects when multiple manufacturing conditions are input into the defect prediction model over a predetermined range covering the entire length of the metal material in the manufacturing condition determination method according to an embodiment of the present invention. Figure 8 is a diagram illustrating the prediction results of the presence or absence of defects when multiple manufacturing conditions are input into a defect prediction model over a specified range covering the entire length of the metal material, according to an embodiment of the manufacturing condition determination method of the present invention. The proportion of the specified range predicted as defective is shown in bold as an acceptable range for the manufacturing conditions. Figure 9 is a diagram showing the learning results of defect prediction in an embodiment of the manufacturing condition determination method of the present invention. Figure 10 is a diagram showing the verification results in Embodiment 1 of the manufacturing condition determination method of the present invention. Figure 11 is a diagram showing the verification results in Embodiment 2 of the manufacturing condition determination method of the present invention.

Claims

1. A method for determining manufacturing conditions to improve the quality of a metallic material manufactured through one or more steps, the method comprising: The defect prediction step predicts the presence or absence of defects in each specified range by inputting manufacturing conditions for each specified range into a defect prediction model. The defect prediction model is generated by using the manufacturing conditions for each specified range of the metal material in each step as input variables and the presence or absence of defects in each specified range as output variables. The characteristic prediction step predicts the characteristics of each specified range by inputting manufacturing conditions for each specified range into a characteristic prediction model. The characteristic prediction model is generated by using the manufacturing conditions for each specified range in each step as input variables and the characteristics of each specified range as output variables. The manufacturing condition determination step involves extracting manufacturing conditions predicted as defect-free by the defect prediction model from among multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model, and determining the manufacturing conditions for the metal material whose characteristics are predicted to be the maximum or minimum from among the extracted manufacturing conditions. In this manufacturing condition determination step, when the same manufacturing conditions are set for the specified range covering the entire length of the metal material, the manufacturing conditions that are predicted as defective across the entire length of the metal material from among multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model have a proportion within an acceptable range are extracted. Finally, the manufacturing conditions for the metal material whose statistical value of the characteristics predicted by the characteristic prediction model is the maximum or minimum are determined as the manufacturing conditions for the metal material.

2. A method for determining manufacturing conditions to improve the quality of a metallic material manufactured through one or more steps, the method comprising: The defect prediction step predicts the presence or absence of defects in each specified range by inputting manufacturing conditions for each specified range into a defect prediction model. The defect prediction model is generated by using the manufacturing conditions for each specified range of the metal material in each step as input variables and the presence or absence of defects in each specified range as output variables. The characteristic prediction step predicts the characteristics of each specified range by inputting manufacturing conditions for each specified range into a characteristic prediction model. The characteristic prediction model is generated by using the manufacturing conditions for each specified range in each step as input variables and the characteristics of each specified range as output variables. The manufacturing condition determination step involves extracting manufacturing conditions predicted as defect-free by the defect prediction model from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model, and determining the manufacturing conditions for the metal material as those whose characteristics are predicted to be the maximum or minimum from the extracted manufacturing conditions. In this step, when the same manufacturing conditions are set for the specified range covering the entire length of the metal material, the manufacturing condition determination step extracts manufacturing conditions from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model whose proportion of the specified range predicted as defective covering the entire length of the metal material is within an acceptable range. It then extracts manufacturing conditions from the extracted manufacturing conditions that are predicted as defect-free within the specified range, and determines the manufacturing conditions for the metal material as those whose statistical value of the characteristic predicted by the characteristic prediction model is the maximum or minimum from the extracted manufacturing conditions.

3. A method for determining manufacturing conditions to improve the quality of a metallic material manufactured through one or more steps, the method comprising: The defect prediction step predicts the presence or absence of defects in each specified range by inputting manufacturing conditions for each specified range into a defect prediction model. The defect prediction model is generated by using the manufacturing conditions for each specified range of the metal material in each step as input variables and the presence or absence of defects in each specified range as output variables. The characteristic prediction step predicts the characteristics of each specified range by inputting manufacturing conditions for each specified range into a characteristic prediction model. The characteristic prediction model is generated by using the manufacturing conditions for each specified range in each step as input variables and the characteristics of each specified range as output variables. The manufacturing condition determination step involves extracting manufacturing conditions predicted as defect-free by the defect prediction model from among multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model, and determining the manufacturing conditions for the metal material whose characteristics are predicted to be the maximum or minimum from among the extracted manufacturing conditions. This manufacturing condition determination step, when setting the same manufacturing conditions for the entire length of the metal material within the specified range, extracts manufacturing conditions predicted as defect-free within the specified range from among multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model, and determines the manufacturing conditions for the metal material whose statistical value of the characteristics predicted by the characteristic prediction model is the maximum or minimum from among the extracted manufacturing conditions.

4. A method for determining manufacturing conditions to improve the quality of a metallic material manufactured through one or more steps, the method comprising: The defect prediction step predicts the presence or absence of defects in each specified range by inputting manufacturing conditions for each specified range into a defect prediction model. The defect prediction model is generated by using the manufacturing conditions for each specified range of the metal material in each step as input variables and the presence or absence of defects in each specified range as output variables. The characteristic prediction step predicts the characteristics of each specified range by inputting manufacturing conditions for each specified range into a characteristic prediction model. The characteristic prediction model is generated by using the manufacturing conditions for each specified range in each step as input variables and the characteristics of each specified range as output variables. The manufacturing condition determination step extracts manufacturing conditions predicted as defect-free by the defect prediction model from among multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model. It then determines the manufacturing conditions for the metal material based on the manufacturing conditions whose characteristics are predicted to be at their maximum or minimum from among the extracted manufacturing conditions. Furthermore, when setting the same manufacturing conditions for the entire length of the metal material, the manufacturing condition determination step determines the manufacturing conditions for the metal material based on the manufacturing conditions with the largest number of manufacturing conditions predicted as defect-free and with characteristics within the allowable range from among the multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model.

5. A method for manufacturing a metallic material, wherein the metallic material is manufactured by means of manufacturing conditions determined by the manufacturing conditions determination method as described in any one of claims 1 to 4.

6. A manufacturing condition determining apparatus for improving the quality of a metallic material manufactured through one or more steps, the manufacturing condition determining apparatus comprising: The system comprises: a defect prediction unit, which predicts the presence or absence of defects in each specified range by inputting manufacturing conditions for each range into a defect prediction model, wherein the defect prediction model is generated by using the manufacturing conditions for each specified range of the metal material in each step as input variables and the presence or absence of defects in each specified range as output variables; a characteristic prediction unit, which predicts the characteristics of each specified range by inputting manufacturing conditions for each range into a characteristic prediction model, wherein the characteristic prediction model is generated by using the manufacturing conditions for each specified range in each step as input variables and the characteristics of each specified range as output variables; and a manufacturing condition determination unit, which extracts manufacturing conditions predicted as defect-free by the defect prediction model from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model, and determines the manufacturing conditions for the metal material as the manufacturing conditions for the metal material, wherein the manufacturing condition determination unit sets the same manufacturing conditions for the entire specified range of the metal material. The proportion of manufacturing conditions predicted as defective along the entire length of the metal material from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model is taken as the allowable manufacturing conditions. The manufacturing conditions with the largest or smallest statistical value of the characteristic predicted by the characteristic prediction model among the extracted manufacturing conditions are determined as the manufacturing conditions of the metal material.

7. A manufacturing condition determining apparatus for improving the quality of a metallic material manufactured through one or more steps, the manufacturing condition determining apparatus comprising: The system comprises: a defect prediction unit, which predicts the presence or absence of defects in each specified range by inputting manufacturing conditions for each range into a defect prediction model, wherein the defect prediction model is generated by using the manufacturing conditions for each specified range of the metal material in each step as input variables and the presence or absence of defects in each specified range as output variables; a characteristic prediction unit, which predicts the characteristics of each specified range by inputting manufacturing conditions for each range into a characteristic prediction model, wherein the characteristic prediction model is generated by using the manufacturing conditions for each specified range in each step as input variables and the characteristics of each specified range as output variables; and a manufacturing condition determination unit, which extracts manufacturing conditions predicted as defect-free by the defect prediction model from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model, and determines the manufacturing conditions for the metal material as the manufacturing conditions for the metal material, wherein the manufacturing condition determination unit sets the same manufacturing conditions for the entire specified range of the metal material. The proportion of manufacturing conditions predicted as defective along the entire length of the metal material from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model is extracted as the allowable manufacturing conditions. The manufacturing conditions predicted as defect-free within the specified range from the extracted manufacturing conditions are extracted. The manufacturing condition with the largest or smallest statistical value of the characteristic predicted by the characteristic prediction model is determined as the manufacturing condition of the metal material.

8. A manufacturing condition determining apparatus for improving the quality of a metallic material manufactured through one or more steps, the manufacturing condition determining apparatus comprising: The system comprises: a defect prediction unit, which predicts the presence or absence of defects in each specified range by inputting manufacturing conditions for each range into a defect prediction model, wherein the defect prediction model is generated by using the manufacturing conditions for each specified range of the metal material in each step as input variables and the presence or absence of defects in each specified range as output variables; a characteristic prediction unit, which predicts the characteristics of each specified range by inputting manufacturing conditions for each range into a characteristic prediction model, wherein the characteristic prediction model is generated by using the manufacturing conditions for each specified range in each step as input variables and the characteristics of each specified range as output variables; and a manufacturing condition determination unit, which extracts manufacturing conditions predicted as defect-free by the defect prediction model from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model, and determines the manufacturing conditions for the metal material as the manufacturing conditions for the metal material, wherein the manufacturing condition determination unit sets the same manufacturing conditions for the entire specified range of the metal material. The manufacturing conditions within the specified range that are predicted to be defect-free from multiple manufacturing conditions input into the defect prediction model and the characteristic prediction model are extracted. The manufacturing conditions with the largest or smallest statistical value of the characteristic predicted by the characteristic prediction model among the extracted manufacturing conditions are determined as the manufacturing conditions of the metal material.

9. A manufacturing condition determining apparatus for improving the quality of a metallic material manufactured through one or more steps, the manufacturing condition determining apparatus comprising: The system comprises: a defect prediction unit, which predicts the presence or absence of defects in each specified range by inputting manufacturing conditions for each range into a defect prediction model, wherein the defect prediction model is generated by using the manufacturing conditions for each specified range of the metal material in each step as input variables and the presence or absence of defects in each specified range as output variables; a characteristic prediction unit, which predicts the characteristics of each specified range by inputting manufacturing conditions for each range into a characteristic prediction model, wherein the characteristic prediction model is generated by using the manufacturing conditions for each specified range in each step as input variables and the characteristics of each specified range as output variables; and a manufacturing condition determination unit, which extracts manufacturing conditions predicted as defect-free by the defect prediction model from multiple manufacturing conditions input to the defect prediction model and the characteristic prediction model, and determines the manufacturing conditions for the metal material as the manufacturing conditions for the metal material, wherein the manufacturing condition determination unit sets the same manufacturing conditions for the entire specified range of the metal material. The manufacturing condition that is predicted to be defect-free and have characteristics within the allowable range among the multiple manufacturing conditions input into the defect prediction model and the characteristic prediction model is determined as the manufacturing condition of the metal material.

Images