Operation condition determination device, operation condition determination method, and program

Abstract

To provide an operation condition determination device, etc., capable of determining an operation condition in a lower process based on order information after execution of an upper process.SOLUTION: An operation condition determination device 100 comprises a material property value prediction model which uses operation conditions in upper and lower processes of a steel material as input and a probability distribution of a material property value of the steel material as output. This device determines an operation condition in the lower process by reversely analyzing the material property value prediction model so that an acceptance probability of the material property value of the steel material calculated on the basis of the probability distribution of the material property value of the steel material output when an operation achievement in the upper process and the operation condition in the lower process of the steel material are input to the material property value prediction model and specifications of the material property value of the steel material included in order information can be equal to or higher than a prescribed threshold value.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an operating condition determination device, an operating condition determination method, and a program for causing a computer to function as an operating condition determination device, which can determine appropriate operating conditions for a downstream process in accordance with order information after the upstream process has been executed for steel material manufactured in a manufacturing process having an upstream process and a downstream process. [Background technology]

[0002] 2. Description of the Related Art Steel materials such as steel plates are produced through a manufacturing process that includes an upstream process such as a steelmaking process and a downstream process such as a continuous annealing process. Here, in order to prevent a decrease in product yield caused by the material property values, such as the tensile strength of the manufactured steel material not satisfying the order specifications (the material property values ​​of the steel material being unacceptable), when the operational results of the upstream process deviate from appropriate operating conditions (operating conditions that have a high probability of the material property values ​​of the steel material being acceptable), it is important to appropriately determine the operating conditions in the downstream process to be executed from now on, taking into account the operational results of the upstream process, so as to increase the probability that the material property values ​​of the steel material will be acceptable (in this specification, this is referred to as "material property value FF (feedforward) technology").

[0003] Conventional material property value FF technologies often use methods based on table methods or linear regression models, and because the number of control ends is limited, there is a problem that it is not possible to sufficiently increase the probability that the material property values ​​of steel will pass.

[0004] Recently, as a material characteristic value FF technology, a technology has been proposed for determining operation conditions in a downstream process so that the output of a material characteristic value prediction model constructed by machine learning such as a neural network approaches a desired value, as described in Patent Document 1, for example. Specifically, the technology described in Patent Document 1 is a technology for determining operation conditions in a downstream process so that the predicted value of the material characteristic value approaches a desired value by using the confirmed operation results (production result data in Patent Document 1) in an upstream process as a fixed value of the operation conditions (production specifications in Patent Document 1), and then performing a reverse analysis of a material characteristic value prediction model (prediction model in Patent Document 1) that predicts the material characteristic value from the operation conditions.

[0005] However, the technology described in Patent Document 1 uses a material property value prediction model that outputs only expected values ​​as predicted values ​​of material property values. Therefore, if the operating conditions in downstream processes are determined by relying on this predicted expected value, there is a high risk that the material property values ​​will fail in areas of operating conditions where there is a large variation in the material property values.

[0006] Non-Patent Documents 1 and 2 describe heteroscedastic Gaussian processes (Gaussian process regression taking heteroscedasticity into account). [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Patent Publication No. 2022-14878 [Non-patent literature]

[0008] [Non-Patent Document 1] Alan D. Saul, James Hensman, Aki Vehtari, Neil D. Lawrence, "Chained Gaussian Processes", Proceedings of the 19th International Conference on Artificial Intelligence and Statistics, PMLR 51:1431-1440, 2016 [Non-Patent Document 2] "Heteroskedastic Likelihood and Multi-Latent GP", [online], [Retrieved December 20, 2022], Internet<URL:https: / / gpflow.github.io / GPflow / develop / notebooks / advanced / heteroskedastic.html> Summary of the Invention [Problem to be solved by the invention]

[0009] The present invention has been made to solve the problems of the conventional technology described above, and has an object to provide an operating condition determination device, an operating condition determination method, and a program that can determine appropriate operating conditions for a downstream process in accordance with order information after the upstream process has been executed for steel materials manufactured in a manufacturing process having an upstream process and a downstream process. [Means for solving the problem]

[0010] In order to solve the above problem, the present invention provides an operating condition determination device that determines operating conditions for a downstream process in accordance with order information for a steel material manufactured through a manufacturing process having an upstream process and a downstream process after the upstream process has been executed, the operating condition determination device having a material property value prediction model that receives as input the operating conditions for the upstream process and the downstream process for the steel material and outputs a probability distribution of the material property values ​​of the steel material, and determines the operating conditions for the downstream process by inversely analyzing the material property value prediction model so that the probability of passing the material property values ​​of the target steel material, calculated based on the probability distribution of the material property values ​​of the target steel material output when the operational performance of the upstream process for the target steel material and the operating conditions for the downstream process are input into the material property value prediction model, and the specifications of the material property values ​​of the target steel material included in the order information, is equal to or greater than a predetermined threshold value.

[0011] In the present invention, examples of the "upstream process" include a steelmaking process, a hot rolling process, and a cold rolling process. Examples of the "downstream process" include a continuous annealing process and a continuous hot-dip galvanizing process. The "manufacturing process" refers to the process up to the process of fabricating a product by changing the thickness, surface condition, and internal metal structure of the steel material (such as a hot rolling process, a cold rolling process, a continuous annealing process, and a continuous hot-dip galvanizing process), and does not necessarily include a finishing process in which only cutting and inspection of the steel material are performed. In the present invention, the "operating conditions" refer to conditions that affect the quality of steel in the manufacturing process, and examples thereof include the product size of the steel (the size of the steel after the cold rolling process), component values ​​of carbon (C) and manganese (Mn) and the like in the steelmaking process, the coiling temperature in the hot rolling process, the width of the steel in the cold rolling process (cold rolling width), the heating temperature (heating plate temperature) and soaking temperature (soaking plate temperature) in the continuous annealing process, and the conveying speed of the steel in a continuous annealing furnace (furnace speed). In the present invention, the "operational results" are actual values ​​of the operating conditions, and are data that affect the quality of the steel material. The operational results may include measured values ​​and set values ​​obtained in the manufacturing process of the steel material, as well as data obtained by processing these values ​​using physical formulas, etc. In the present invention, the "material characteristic value" means the mechanical properties of a steel material, and examples thereof include tensile strength (TS), yield point (YP), surface hardness, bending strength, toughness value, and fatigue life. In the present invention, the term "target steel material" refers to a steel material for which operating conditions in downstream processes are to be determined. Furthermore, in the present invention, the "specifications of the material property values ​​of the target steel material included in the order information" may include, for example, the required upper limit value or the required lower limit value of the material property value, or both.

[0012] The operating condition determination device according to the present invention is equipped with a material property value prediction model that receives operating conditions in upstream and downstream processes of steel material as input and outputs a probability distribution of material property values ​​of the steel material. The material property value prediction model outputs the probability distribution of material property values ​​of steel material, in other words, the expected value and variation of material property values, as a prediction result, so that it is expected that the material property values ​​(especially the variation) can be predicted with high accuracy even for operating conditions of the input steel material that have no past operating results. In addition, it is expected that the prediction accuracy of the material property values ​​can be improved even for operating conditions of steel material where it is difficult to control the material property values ​​and variation in the material property values ​​is likely to occur. In the operation condition determination device according to the present invention, the pass probability of the material property value of the target steel is calculated based on the probability distribution of the material property value of the target steel output when the operation results of the upstream process and the operation conditions of the downstream process for the target steel are input into the material property value prediction model, and the specification of the material property value of the target steel included in the order information. For example, if the specification of the material property value of the target steel included in the order information is the required upper limit and required lower limit of the material property value, the ratio of the area within the range from the required lower limit to the required upper limit to the area of ​​the entire probability distribution of the material property value is calculated as the pass probability. Therefore, by reverse-analyzing this material property value prediction model so that the probability of passing the material property value of the target steel material is above a predetermined threshold, appropriate operating conditions for downstream processes can be determined in accordance with the order information, and a decrease in the product yield of the target steel material can be suppressed.

[0013] The material property value prediction model of the present invention is constructed by using known data, for example, in which the operation results of the manufacturing process (upstream and downstream processes) of steel materials manufactured in the past are input, and the material property values ​​measured on test pieces obtained from the steel materials manufactured in the past are output. For example, the model may be constructed by machine learning using such known data as teacher data. In addition, the material property value prediction model may be constructed by the operating condition determination device of the present invention, or a material property value prediction model constructed by another device may be stored in the material property prediction device of the present invention.

[0014] The operating condition determination device according to the present invention may be configured to determine the operating conditions in the downstream process independently for each target steel product so that the pass probability of the material characteristic value is equal to or greater than a predetermined threshold value. However, for example, when the downstream process is a continuous annealing process, since the continuous annealing process is a process in which a plurality of steel materials are welded and continuously processed, it is necessary to determine the transport order of the steel materials so that the size difference (thickness difference or width difference) between adjacent steel materials is small so that the plurality of steel materials can be welded together. Also, it is necessary to determine the operating conditions so that the difference in heating temperature, soaking temperature, etc. between adjacent steel materials is small enough to be actually specified.

[0015] In view of the above, a first preferred configuration of the operating condition determination device of the present invention is to determine operating conditions in the downstream process for multiple target steel materials, whose transport order in the downstream process is predetermined, by inversely analyzing the material characteristic value prediction model under constraints imposed by the transport order so that all of the calculated pass probabilities or the geometric mean of the pass probabilities is equal to or greater than a predetermined threshold value.

[0016] According to a first preferred configuration, when the order of transporting multiple target steel materials in the downstream process is determined in advance taking into consideration the weldability between the multiple target steel materials, the operating conditions in the downstream process are determined under the constraints imposed by the order of transporting the multiple target steel materials (for example, the difference in heating temperature or soaking temperature between adjacent target steel materials) so that all of the pass probabilities of the material property values ​​of the multiple target steel materials or the geometric mean of the pass probabilities is equal to or greater than a predetermined threshold value.Therefore, even if the downstream process is a process in which multiple target steel materials are processed continuously, appropriate operating conditions in the downstream process can be determined in accordance with the order information, and a decrease in the product yield of the multiple target steel materials can be suppressed.

[0017] In the first preferred configuration, the transport order of the multiple target steel products in the downstream process is determined in advance, but the preferred configuration is not limited to this. For example, for multiple target steel products whose transport order has not yet been determined, the operating conditions in the downstream process may be determined independently for each target steel product, and then the transport order of the multiple target steel products in the downstream process may be determined taking into account the constraints imposed by the transport order of the multiple target steel products.

[0018] In other words, in a second preferred configuration of the operating condition determination device of the present invention, the material characteristic value prediction model is reverse-analyzed so that all of the pass probabilities calculated for each of the multiple target steel materials or the geometric mean of the pass probabilities is equal to or greater than a predetermined threshold value, thereby determining the operating conditions for the multiple target steel materials in the downstream process, and then the transport order of the multiple target steel materials in the downstream process is determined by solving an optimization problem that takes into account constraints due to the transport order of the multiple target steel materials in the downstream process.

[0019] According to a second preferred configuration, the operation conditions for the downstream process for the multiple target steel products are determined first. Then, the transport order of the multiple target steel products in the downstream process is determined by solving an optimization problem that takes into account constraints due to the transport order of the multiple target steel products in the downstream process (for example, differences in size between adjacent target steel products, differences in heating temperature and soaking temperature between adjacent target steel products, etc.). Therefore, even if the downstream process is a process that continuously processes multiple target steel products, appropriate operation conditions for the downstream process according to the order information can be determined, and a decrease in the product yield of the multiple target steel products can be suppressed.

[0020] In addition, in the second preferred configuration, it is more preferable to determine the transport order of the multiple target steel materials in the downstream process, and then determine the operating conditions in the downstream process for the multiple target steel materials by re-inverse analyzing the material characteristic value prediction model under the constraints imposed by the determined transport order so that any of the multiple pass probabilities calculated for each of the multiple target steel materials, or the geometric mean of the pass probabilities, is equal to or greater than a predetermined threshold value. According to the above-mentioned more preferred configuration, after the conveying order is determined, the material property value prediction model is again reverse-analyzed under the constraints imposed by the determined conveying order to determine the operating conditions for the downstream processes for the multiple target steel products (in other words, after the conveying order is determined, the operating conditions for the downstream processes are determined as in the first preferred configuration), so that it is expected that more appropriate operating conditions will be determined.

[0021] In the first and second preferred configurations, one of the operation conditions and the transport order in the downstream process for a plurality of target steel products is first determined, and then the other is determined using the determined one, but the preferred configuration is not limited to this. For example, the operation conditions and the transport order in the downstream process may be determined at the same time.

[0022] In other words, in a third preferred configuration of the operating condition determination device of the present invention, it is considered that the operating conditions and transport order in the downstream process for the multiple target steel materials are determined at once by inversely analyzing the material characteristic value prediction model under constraints imposed by the transport order of the multiple target steel materials in the downstream process so that all of the multiple pass probabilities calculated for each of the multiple target steel materials or the geometric mean of the pass probabilities is equal to or greater than a predetermined threshold value.

[0023] In the third preferred configuration, "determining at the same time" means determining both the operating conditions and the conveying order in the downstream process at the same time, rather than first determining one of them and then using the determined one to determine the other. According to the third preferred configuration, the operating conditions and transport order in the downstream process for a plurality of target steel products are determined at once, which has the advantage that the operating conditions and transport order with a higher probability of passing can be determined.

[0024] In the operation condition determination device according to the present invention, the material characteristic value prediction model is, for example, a model using Bayesian estimation. Also, the material characteristic value prediction model is, for example, a model using a heteroscedastic Gaussian process.

[0025] When a downstream process is to be carried out, taking into account the operational performance of the upstream process, there is a high need to predict material property values ​​by giving the parameters that make up the operating conditions in the downstream process as a range rather than as a single value (or predicting material property values ​​by giving the parameters that make up the operating conditions in the downstream process as target values), or to determine the parameters that make up appropriate operating conditions as a range rather than as a single value (or to determine them as target values) so that the predicted material property values ​​satisfy the order specifications. In order to meet the above-mentioned needs, it is preferable that the operating condition determination device of the present invention further includes an operating condition prediction model that inputs the range of operating conditions or the target values ​​of the operating conditions in the downstream process for the steel and outputs the probability distribution of the operating conditions in the downstream process for the steel, and determines the range or target values ​​of the operating conditions in the downstream process by inversely analyzing the operating condition prediction model and the material characteristic value prediction model so that the pass probability of the material characteristic value of the target steel calculated based on the probability distribution of the operating conditions in the downstream process for the target steel output when the range of operating conditions or the target values ​​of the operating conditions in the downstream process for the target steel are input into the operating condition prediction model, and the probability distribution of the material characteristic value of the target steel output when the operational performance of the target steel in the upstream process for the target steel is input into the material characteristic value prediction model, is equal to or greater than a predetermined threshold value.

[0026] In the above preferred configuration, "inversely analyzing the operating condition prediction model and the material characteristic value prediction model" means inversely analyzing a composite function of a function represented by the operating condition prediction model and a function represented by the material characteristic value prediction model. The above-mentioned preferred configuration includes, in addition to the material property value prediction model, an operating condition prediction model that inputs the range of operating conditions or the target value of the operating conditions in the downstream process for the steel material and outputs the probability distribution of the operating conditions in the downstream process for the steel material. Therefore, if the range of operating conditions or the target value of the operating conditions in the downstream process for the target steel material are input to the operating condition prediction model, the probability distribution of the operating conditions in the downstream process for the target steel material is output, and if the probability distribution of the operating conditions in the downstream process for the target steel material and the operational performance in the upstream process for the target steel material are input to the material property value prediction model, the probability distribution of the material property value of the target steel material according to the range of operating conditions in the downstream process for the target steel material or the target value of the operating conditions is output. Therefore, by reverse-analyzing the operating condition prediction model and the material characteristic value prediction model so that the pass probability of the material characteristic values ​​of the target steel calculated based on the probability distribution of the material characteristic values ​​of the target steel (specifically, based on the probability distribution of the material characteristic values ​​of the target steel and the specifications of the material characteristic values ​​of the target steel included in the order information) is above a predetermined threshold value, it is possible to determine the range or target values ​​of appropriate operating conditions in downstream processes according to the order information, thereby meeting the above-mentioned needs.

[0027] In order to solve the above problem, the present invention is also provided as an operating condition determination method for a steel material manufactured through a manufacturing process having an upstream process and a downstream process, which determines operating conditions in the downstream process according to order information after the upstream process is executed, and which uses a material property value prediction model that inputs the operating conditions in the upstream process and the downstream process for the steel material and outputs a probability distribution of the material property values ​​of the steel material, and determines the operating conditions in the downstream process by inversely analyzing the material property value prediction model so that the probability of passing the material property values ​​of the target steel material, which is calculated when the operational performance in the upstream process for the target steel material and the operating conditions in the downstream process for the target steel material are input into the material property value prediction model, and the specifications of the material property values ​​of the target steel material included in the order information, is equal to or greater than a predetermined threshold value.

[0028] Furthermore, in order to achieve the above object, the present invention also provides a program for causing a computer to function as the operation condition determination device. Effect of the Invention

[0029] According to the present invention, for steel products manufactured in a manufacturing process having an upstream process and a downstream process, after the upstream process is executed, appropriate operating conditions for the downstream process can be determined according to order information, thereby making it possible to suppress a decrease in the product yield of the steel products. [Brief description of the drawings]

[0030] [Figure 1] 1 is a diagram for explaining an outline of an operation condition determination device according to an embodiment of the present invention; [Diagram 2] 1 is a block diagram showing a schematic configuration of an operation condition determination device according to a first embodiment. [Diagram 3] 3 is a flowchart showing an outline of a procedure of an operation condition determining method executed by using the operation condition determining device shown in FIG. 2. [Figure 4] 3 is a diagram showing an example of a manufacturing record acquired by a manufacturing record acquiring unit 1 shown in FIG. 2. FIG. [Diagram 5] FIG. 2 is a diagram for explaining a construction example of a material characteristic value prediction model according to the first embodiment in comparison with a construction example of a conventional machine learning model. [Figure 6] 3 is a diagram showing an example of order information acquired by an order information acquiring unit 4 shown in FIG. 2 and operation results in an upstream process acquired by an upstream process operation results acquiring unit 5. FIG. [Figure 7] 3 is a diagram showing an example of the contents output from a downstream process operation condition output unit 9 shown in FIG. 2.

[0023] FIG. [Figure 8] FIG. 6 is a block diagram showing a schematic configuration of an operation condition determination device according to a second embodiment. [Figure 9] 9 is a flowchart showing an outline of a procedure of an operation condition determining method executed by using the operation condition determining device shown in FIG. 8. [Figure 10] 9 is a diagram showing an example of order information acquired by the order information acquisition unit 4 shown in FIG. 8, a conveying order in a downstream process acquired by the conveying order acquisition unit 10, and operation results in an upstream process acquired by the upstream process operation results acquisition unit 5. [Figure 11] FIG. 9 is a diagram showing an example of the contents output from a downstream process operating condition output unit 9 shown in FIG. 8. [Figure 12] FIG. 11 is a block diagram showing a schematic configuration of an operation condition determination device according to a third embodiment. [Figure 13] 13 is a flowchart showing an outline of a procedure of an operation condition determining method executed by using the operation condition determining device shown in FIG. 12. [Figure 14] 13 is a diagram showing an example of order information acquired by an order information acquiring unit 4 and upstream operation results acquired by an upstream operation results acquiring unit 5 shown in FIG. 12. FIG. [Figure 15] 13 is a diagram showing an example of the contents output from a downstream process operating condition output unit 9 shown in FIG. 12. FIG. [Figure 16] FIG. 13 is a conceptual diagram of an optimization problem applied in the third embodiment. [Figure 17] FIG. 13 is a block diagram showing a schematic configuration of an operation condition determination device according to a fourth embodiment. [Figure 18] 18 is a flowchart showing an outline of a procedure of an operation condition determining method executed by using the operation condition determining device shown in FIG. 17. [Figure 19] 18 is a diagram showing an example of the contents output from the downstream process operation condition / conveying sequence output unit 14 shown in FIG. 17. FIG. [Figure 20] FIG. 13 is a block diagram showing a schematic configuration of an operation condition determination device according to a fifth embodiment. [Figure 21] 21 is a flowchart showing an outline of a procedure of an operation condition determining method executed by using the operation condition determining device shown in FIG. 20. [Figure 22] FIG. 21 is a diagram showing an example of the contents output from a downstream process operating condition output unit 9 shown in FIG. 20. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0031] [Outline of the operation condition determination device] Hereinafter, an operation condition determination device according to an embodiment (first to fifth embodiments) of the present invention will be described with reference to the accompanying drawings as appropriate, taking a case where the steel material is a steel plate as an example. First, an overview of the operation condition determination device according to the present embodiment will be described.

[0032] FIG. 1 is a diagram for explaining an outline of an operation condition determination device according to the present embodiment. The operation condition determination device according to the present embodiment is a device that determines operation conditions in a downstream process according to order information after the upstream process is executed for a steel material manufactured through a manufacturing process having an upstream process and a downstream process. Specifically, as shown in FIG. 1(a), the operation condition determination device according to the present embodiment is provided with a material property value prediction model that receives the operation conditions in the upstream and downstream processes of the steel material as input and outputs the probability distribution of the material property value of the steel material. By using this material property value prediction model, the ratio of the area within the range from the required lower limit value to the required upper limit value (the area of ​​the hatched area shown in FIG. 1(a)) to the area of ​​the entire probability distribution of the material property value can be calculated as the pass probability based on the probability distribution of the material property value of the steel material output from the material property value prediction model and the specifications of the material property value of the steel material included in the order information (for example, the required upper limit value and the required lower limit value of the material property value). As shown in Figure 1(b), the operating condition determination device of this embodiment determines the operating conditions for the downstream process by inversely analyzing the material property value prediction model (solving an optimization problem) so that the pass probability of the material property values ​​of the target steel material (the area of ​​the hatched area shown in Figure 1(b) relative to the area of ​​the entire probability distribution) calculated based on the probability distribution of the material property values ​​of the target steel material output when the operational performance of the target steel material in the upstream process after the upstream process has been executed and the operating conditions in the downstream process to be determined (i.e., the downstream process before execution) are input into the material property value prediction model, and the specifications of the material property values ​​of the target steel material included in the order information (e.g., the required upper and lower limits of the material property values) is greater than or equal to a predetermined threshold value.

[0033] In this embodiment, the upper process is a steelmaking process, a hot rolling process, and a cold rolling process, and the lower process is a continuous annealing process. An example will be given of a case in which the heating temperature (heating plate temperature) and soaking temperature (soaking plate temperature) of the steel material are determined as operating conditions in the lower process. Hereinafter, first to fifth embodiments, which differ from one another in the procedures for determining the operating conditions in the downstream process, will be described in order.

[0034] [First embodiment] Fig. 2 is a block diagram showing a schematic configuration of an operation condition determination device according to the first embodiment. Fig. 3 is a flow chart showing a schematic procedure of an operation condition determination method executed by using the operation condition determination device shown in Fig. 2. The operating condition determination device 100 according to the first embodiment is configured to determine the operating conditions in the downstream process so that the pass probability of the material characteristic value (also referred to as "mechanical test value" in this specification) is equal to or higher than a predetermined threshold value, independently for each target steel material. As shown in Fig. 2, the operating condition determination device 100 according to the first embodiment includes a manufacturing record acquisition unit 1, a material characteristic value prediction model construction unit 2, a material characteristic value prediction model storage unit 3, an order information acquisition unit 4, an upstream process operating record acquisition unit 5, a material characteristic value prediction model acquisition unit 6, a constraint condition setting unit 7, a downstream process operating condition determination unit 8, and a downstream process operating condition output unit 9.

[0035] The operation condition determination device 100 is composed of a computer having one or more hardware processors such as a CPU (Central Processing Unit), one or more memories such as a RAM (Random Access Memory) and a ROM (Read Only Memory), and one or more programs stored in the memory are executed by the one or more hardware processors to perform various calculations. As a result, the operation condition determination device 100 functions as a manufacturing record acquisition unit 1, a material characteristic value prediction model construction unit 2, a material characteristic value prediction model storage unit 3, an order information acquisition unit 4, an upstream process operation record acquisition unit 5, a material characteristic value prediction model acquisition unit 6, a constraint condition setting unit 7, a downstream process operation condition determination unit 8, and a downstream process operation condition output unit 9. The operation condition determination device 100 may be a PLC (Programmable Logic Controller) or may be realized by dedicated hardware such as an ASIC (Application Specific Integrated Circuit). Each unit 1 to 9 included in the operation condition determination device 100 will be described in order below.

[0036] <Manufacturing Record Acquisition Department 1> The manufacturing record acquiring unit 1 executes step ST11 shown in Fig. 3. Specifically, the manufacturing record acquiring unit 1 acquires manufacturing records of the manufacturing process (upstream process and downstream process) of a steel material manufactured in the past (herein, this is also referred to as a "reference steel material" to distinguish it from the above-mentioned "target steel material") from a predetermined database (not shown). The manufacturing record is combined data of the operation record of the manufacturing process of the reference steel material and material characteristic values ​​(mechanical test values) measured on a test piece obtained from the reference steel material. Fig. 4 is a diagram showing an example of the manufacturing results acquired by the manufacturing results acquisition unit 1. In the example shown in Fig. 4, for each reference steel, the operation results in the upstream process are acquired, such as the product thickness included in the product size of the reference steel (the size of the reference steel after the cold rolling process), the component values ​​of carbon (C) and manganese (Mn) in the steelmaking process, the coiling temperature in the hot rolling process, and the cold rolling width in the cold rolling process. In addition, the heating plate temperature, soaking plate temperature, and furnace speed in the continuous annealing process are acquired as the operation results in the downstream process. Furthermore, the tensile strength (TS) and yield point (YP) are acquired as mechanical test values.

[0037] <Material property value prediction model construction part 2> The material characteristic value prediction model construction unit 2 executes step ST12 shown in FIG. 3. Specifically, the material characteristic value prediction model construction unit 2 constructs a material characteristic value prediction model by using known data in which the operation results in the manufacturing process of the reference steel material are input and the mechanical test values ​​measured on the test piece obtained from the reference steel material are output, as shown in FIG. 4. For example, the material characteristic value prediction model construction unit 2 may construct a material characteristic value prediction model by machine learning using the known data as teacher data. In the first embodiment, in order to quantify the uncertainty of the prediction due to the lack of past operation results, a material characteristic value prediction model is constructed using Gaussian process regression, which is a type of Bayesian estimation, but the present invention is not limited to this. For example, the material characteristic value prediction model construction unit 2 may construct a material characteristic value prediction model by using a statistical method such as quantile regression using known data, or may construct a material characteristic value prediction model by other machine learning such as a Bayesian neural network. Below, we will explain an overview of the Gaussian process regression used in the first embodiment. Note that in the following equations, variables written in bold represent vectors or matrices.

[0038] The input x expressed by the following formula (1) i When the output vector f, which is expressed by the following equation (3) and corresponds to the input x, which is expressed by the following equation (2), follows a Gaussian distribution N(0,K) with a mean of 0 and a covariance matrix K, which is expressed by the following equation (4), the output f is said to follow a Gaussian process, and can be written as shown in the following equation (5).

number

number

number

[0039] Although it is possible to use normal Gaussian process regression as shown in the above formula (7) as a material characteristic value prediction model, normal Gaussian process regression cannot take into account changes in the variation that occurs in mechanical test values ​​due to differences in operational performance. For this reason, in the first embodiment, Gaussian processes gP(0,K1) and gP(0,K2) that individually estimate the expected value (average value) and variance (standard deviation) expressed by the following formulas (8) to (12) are defined, and a non-uniform variance Gaussian process (a Gaussian process considering non-uniform variance) is used in which a Gaussian distribution with the output of each Gaussian process as a parameter is the final prediction result.

number

[0040] By developing the above method, it is possible to estimate the covariance matrix Σ(x) between output variables, for example, as described below. Let D be the number of explanatory variables (input variables to the material property value prediction model), M be the number of objective variables (output variables from the material property value prediction model), and N be the number of records of actual data. For example, the past output actual data Y expressed by the following formula (13) is decorrelated based on formula (15) using the load matrix W expressed by the following formula (14) to obtain T expressed by the following formula (16).

number

number

[0041] That is, the output mean value vector μ corresponding to the new input x is expressed by the following equation (19). g (x) and the output standard deviation vector σ g Based on these prediction vectors, the output mean vector μ(x) and covariance matrix Σ(x) can be calculated by the following equations (21) and (22), respectively.

number

[0042] The material property value prediction model construction unit 2 of the first embodiment constructs a material property value prediction model (a model expressed by the above-mentioned formula (12) when correlation between output variables is not considered, and a model expressed by the above-mentioned formulas (18), (21) and (22) when correlation between output variables is considered) by the above-mentioned method. The material property value prediction model is a model that receives the operating conditions in the upstream and downstream processes of the steel material as input and outputs the probability distribution of mechanical test values ​​of the steel material.

[0043] FIG. 5 is a diagram for explaining a construction example of a material characteristic value prediction model of the first embodiment in comparison with a construction example of a conventional machine learning model. Specifically, in FIG. 5, for convenience of illustration, only the hot plate temperature of the steel material in the downstream process (continuous annealing process) is plotted among a plurality of parameters constituting the operating conditions, and the construction example of a model predicting the tensile strength (TS) as a mechanical test value is shown. FIG. 5(a) shows a construction example of a conventional machine learning model, and FIG. 5(b) shows a construction example of a material characteristic value prediction model of the first embodiment. Data plotted with "x" in FIG. 5 is known data. In the conventional machine learning model shown in Fig. 5(a), the average value of the mechanical test value (TS) is output as the prediction result for the input operating conditions (heating plate temperature). Therefore, in the area surrounded by the dashed line in Fig. 5(a) where the mechanical test value varies widely even under the same operating conditions, there is a problem that the prediction accuracy of the mechanical test value is poor. In contrast, the material property value prediction model of the first embodiment shown in FIG. 5(b) outputs a probability distribution of mechanical test values ​​as a prediction result. Specifically, the material property value prediction model of the present embodiment is a model using a non-uniform variance Gaussian process, and outputs a mean value and a standard deviation as a prediction result. The hatched area shown in FIG. 5(b) is a 95% prediction range (95% confidence interval) of a normal distribution obtained from the mean value and the standard deviation. In the material property value prediction model, as shown by the arrow A1, an area with a large variation in the mechanical test values ​​(an area corresponding to the area surrounded by a dashed line in FIG. 5(a)) is included in the probability distribution (95% prediction range), and is reflected as an output. In addition, in the material property value prediction model, as shown by the arrows A2 and A3, a probability distribution (95% prediction range) is output even in an area with no operating history (an area corresponding to the hatched area in FIG. 5(a)), so that the uncertainty of the prediction is taken into consideration.

[0044] <Material characteristic value prediction model storage unit 3> The material characteristic value prediction model storage unit 3 executes step ST13 shown in Fig. 3. Specifically, the material characteristic value prediction model storage unit 3 stores (memorizes) the material characteristic value prediction model constructed by the material characteristic value prediction model construction unit 2.

[0045] <Order information acquisition unit 4> The order information acquisition unit 4 executes step ST14 shown in Fig. 3. Specifically, the order information acquisition unit 4 acquires, as order information, the product size of the target steel material and the specifications of the mechanical test values ​​of the target steel material from a predetermined database (not shown). The order information acquisition unit 4 of the first embodiment acquires the product thickness as the product size of the target steel material, and the required upper limit value or the required lower limit value, or both, of the mechanical test values ​​as the specifications of the mechanical test values ​​of the target steel material.

[0046] <Upstream Operation Record Acquisition Department 5> The upstream process operational results acquisition unit 5 executes step ST15 shown in Fig. 3. Specifically, the upstream process operational results acquisition unit 5 acquires operational results in the upstream process of the target steel material, obtained after the upstream process of the target steel material is executed, from a predetermined database (not shown). Specifically, the upstream process operational results acquisition unit 5 in the first embodiment acquires operational results in the upstream process, such as the product thickness included in the product size of the target steel material (the size of the target steel material after the cold rolling process), component values ​​of carbon (C) and manganese (Mn) in the steelmaking process, operational results such as the coiling temperature in the hot rolling process, and operational results such as the cold rolled width in the cold rolling process, as operational results in the upstream process.

[0047] FIG. 6 is a diagram showing an example of order information acquired by the order information acquisition unit 4 and operation results in the upstream process acquired by the upstream process operation results acquisition unit 5. In the example shown in FIG. 6, the product thickness, which is the order information, is also used as the upstream process operation results as is. This is because in many cases, the thickness of the target steel material after the cold rolling process is manufactured without any error with the product thickness, which is the order information. However, this is not limited to this, and it is also possible to actually measure the thickness of the target steel material after the cold rolling process using a known thickness gauge and use the measured value as the upstream process operation results.

[0048] <Material characteristic value prediction model acquisition unit 6> The material characteristic value prediction model acquisition unit 6 executes step ST16 shown in Fig. 3. Specifically, the material characteristic value prediction model acquisition unit 6 acquires the material characteristic value prediction model stored in the material characteristic value prediction model storage unit 3.

[0049] <Constraint Condition Setting Section 7> The constraint condition setting unit 7 executes step ST17 shown in Fig. 3. Specifically, the constraint condition setting unit 7 sets constraint conditions for the inverse analysis executed by the downstream process operation condition determination unit 8 described below as necessary. The constraint condition setting unit 7 in the first embodiment sets, for each target steel material, an upper limit value, a lower limit value, or both of parameters constituting the operation conditions in the downstream process to be determined (in the first embodiment, the heating plate temperature and the soaking plate temperature in the continuous annealing process) as necessary. The constraint condition setting unit 7 may also set constraint conditions between parameters. The constraint condition setting unit 7 of the first embodiment sets, for example, 850°C as the upper limit of the heating plate temperature. In addition, for example, a constraint condition is set such that the difference between the heating plate temperature and the soaking plate temperature is 10°C or less. This is because the heating process and the soaking process constituting the continuous annealing process are performed in adjacent facilities, and therefore it is impossible to issue an operation instruction that would cause a sudden change in plate temperature between the two processes.

[0050] <Downstream operation condition determination unit 8> The downstream process operating condition determination unit 8 executes step ST18 shown in Fig. 3. Specifically, as outlined with reference to Fig. 1, the downstream process operating condition determination unit 8 determines the operating conditions (heating plate temperature and soaking plate temperature) in the downstream process by performing a reverse analysis of the material characteristic value prediction model acquired by the material characteristic value prediction model acquisition unit 6 so that the pass probability of the mechanical test value of the target steel material is equal to or higher than a predetermined threshold value (maximum in the first embodiment). The contents of the back analysis executed by the downstream process operating condition determination unit 8 will be specifically described below.

[0051] As described above, by using the material property value prediction model, when the operating condition x expressed by the following formula (23) is given, the mechanical test value y (tensile strength y ts , yield point y yp ) can output the conditional probability distribution p(y|x) of (equation (12) or (18)).

number

number

[0052] In the above formula (25), y yp upper , y yp lower are the upper and lower limits of the yield point (YP) determined by the order information. ts upper , y ts lower are the upper and lower limit values ​​of the tensile strength (TS) determined by the order information. If the order information does not include either the upper limit or the lower limit (in the example shown in FIG. 6, there is no upper limit value for TS), it is sufficient to set the upper limit to a sufficiently large value and the lower limit to a sufficiently small value. In the above equation (25), μ(x) and Σ(x) are the mean value vector and covariance matrix of the probability distribution p(y|x) of the mechanical test value y output from the material characteristic value prediction model. In the above formulas (26), (27), and (29), u ht is the heating plate temperature. In the above formula (26) and formula (29), u st is the heating plate temperature. In the above formula (30), v fs is the furnace velocity. In the above formula (31), v t is the order thickness, ν C is the carbon content, ν Mn is the manganese content, ν CT is the coiling temperature, ν wis the cold rolled width. ν shown in the above formula (28) prev As shown in the above formula (31), v capl As shown in the above formula (30), v is a group of parameters that are fixed and set in advance for each product thickness in the downstream process (continuous annealing process). prev and v capl are all treated as constants. The above equations (26) and (27) are the constraint conditions set by the constraint condition setting unit 7.

[0053] In addition, each mechanical test value (tensile strength y ts , yield point y yp ) are assumed to be mutually independent, the operating conditions x (operating conditions u) that maximize the pass probability P(u), which is expressed as the geometric mean of the pass probability of each mechanical test value, may be searched for and determined without using the covariance matrix Σ(x), as shown in the following equation (32).

number

[0054] <Downstream operation condition output section 9> The downstream process operating condition output unit 9 executes step ST19 shown in Fig. 3. Specifically, the downstream process operating condition output unit 9 outputs the operating conditions (heating plate temperature and soaking plate temperature) in the downstream process determined by the downstream process operating condition determination unit 8. It is preferable that the downstream process operating condition output unit 9 not only outputs the determined operating conditions in the downstream process, but also outputs the probability distribution (average value, standard deviation) of the mechanical test value output from the material characteristic value prediction model when the determined operating conditions (including the operational results in the upstream process) are input to the material characteristic value prediction model, and the pass probability of the mechanical test value calculated from this probability distribution and order information. Fig. 7 is a diagram showing an example of the contents output from the downstream process operational condition output unit 9. Fig. 7 shows that for the target steel material No. 001, when the target steel material is manufactured under the operational results in the upstream process shown in Fig. 6 and the conditions of a heating plate temperature of 842°C and a soaking plate temperature of 837°C determined by the downstream process operational condition determination unit 8, the pass probability of the yield point (YP) and tensile strength (TS) is as high as 0.998.

[0055] [Second embodiment] The operating condition determination device 100 according to the first embodiment described above is configured to determine the operating conditions in the downstream process for each target steel material independently so that the pass probability of the material characteristic value is equal to or greater than a predetermined threshold value. However, since the downstream continuous annealing process is a process in which a plurality of steel materials are welded and continuously processed, it is necessary to determine the transport order of the steel materials so that the difference in size (difference in plate thickness or plate width) between adjacent steel materials in the transport order is small so that the plurality of steel materials can be welded together. It is also necessary to determine the operating conditions so that the difference in the heating plate temperature, the soaking plate temperature, etc. between adjacent steel materials in the transport order is small enough to be actually instructed. Therefore, the operating condition determination device of the second embodiment (corresponding to the first preferred configuration of the present invention) is configured to determine operating conditions in downstream processes for multiple target steel materials, for which the transport order in the downstream processes is predetermined, by inversely analyzing a material property value prediction model under constraints due to the transport order, so that the calculated pass probability of the mechanical test values ​​of the multiple target steel materials is all above a predetermined threshold value.

[0056] Fig. 8 is a block diagram showing a schematic configuration of an operation condition determination device according to the second embodiment. Fig. 9 is a flow chart showing a schematic procedure of an operation condition determination method executed by using the operation condition determination device shown in Fig. 8. As shown in FIG. 8, the operation condition determination device 100A according to the second embodiment is configured, for example, by a computer, similar to the operation condition determination device 100 according to the first embodiment, and includes a conveying sequence acquisition unit 10 in addition to the respective units similar to those of the operation condition determination device 100 according to the first embodiment. Among the units 1 to 10 included in the operating condition determination device 100A, the operations of the manufacturing performance acquisition unit 1, the material characteristic value prediction model construction unit 2, the material characteristic value prediction model storage unit 3 and the material characteristic value prediction model acquisition unit 6 (steps ST21, ST22, ST23, and ST27 shown in FIG. 9) are the same as those in the first embodiment, and therefore will not be described here. The order information acquiring unit 4, the transport sequence acquiring unit 10, the upstream process operation result acquiring unit 5, the constraint condition setting unit 7, the downstream process operation condition determining unit 8, and the downstream process operation condition output unit 9 will be described below in order.

[0057] <Order information acquisition unit 4> The order information acquisition unit 4 executes step ST24 shown in Fig. 9. Specifically, the order information acquisition unit 4 acquires, as order information, the product sizes of the multiple target steel products and the specifications of the mechanical test values ​​of the multiple target steel products from a predetermined database (not shown). The order information acquisition unit 4 of the second embodiment acquires the product thickness as the product size of the multiple target steel products, and the required upper limit value or the required lower limit value, or both, of the mechanical test values ​​as the specifications of the mechanical test values ​​of the multiple target steel products.

[0058] <Transportation order acquisition unit 10> The conveyance order acquisition unit 10 executes step ST25 shown in Fig. 9. Specifically, the conveyance order of a plurality of steel materials in a downstream process that has been determined in advance is acquired from a predetermined database (not shown).

[0059] <Upstream Operation Record Acquisition Department 5> The upstream process operation result acquisition unit 5 executes step ST26 shown in Fig. 9. Specifically, the upstream process operation result acquisition unit 5 acquires operation results in the upstream process of a plurality of target steel materials obtained after the upstream process of the plurality of target steel materials is executed from a predetermined database (not shown). Specifically, the upstream process operation result acquisition unit 5 of the second embodiment acquires operation results in the upstream process such as product thickness included in the product size (size of the target steel materials after the cold rolling process) of the plurality of target steel materials, component values ​​of carbon (C) and manganese (Mn) in the steelmaking process, operation results such as coiling temperature in the hot rolling process, and operation results such as cold rolled width in the cold rolling process, as operation results in the upstream process.

[0060] Fig. 10 is a diagram showing an example of order information acquired by the order information acquisition unit 4, the conveying order in the downstream process acquired by the conveying order acquisition unit 10, and the operation results in the upstream process acquired by the upstream process operation results acquisition unit 5. In the example shown in Fig. 10, the product thickness, which is the order information, is also used as the upstream process operation results as it is.

[0061] <Constraint Condition Setting Section 7> The constraint condition setting unit 7 executes step ST28 shown in Fig. 9. Specifically, the constraint condition setting unit 7 sets constraint conditions for the back analysis executed by a downstream process operation condition determination unit 8, which will be described later, as necessary. The constraint condition setting unit 7 of the second embodiment sets a constraint condition based on the transport order of multiple target steel products, in addition to the constraint condition for each target steel product as in the constraint condition setting unit 7 of the first embodiment. The constraint condition setting unit 7 of the second embodiment further sets a constraint condition that the change in the heating plate temperature and the soaking plate temperature between the target steel products is 5°C or less, for example.

[0062] <Downstream operation condition determination unit 8> The downstream process operating condition determination unit 8 executes step ST29 shown in Fig. 9. Specifically, the downstream process operating condition determination unit 8 determines the operating conditions (heating plate temperature and soaking plate temperature) in the downstream process for the multiple target steel materials by performing a reverse analysis of the material characteristic value prediction model acquired by the material characteristic value prediction model acquisition unit 6 so that the pass probabilities of the mechanical test values ​​of the multiple target steel materials are all equal to or higher than a predetermined threshold value (in the second embodiment, so that the geometric mean of the pass probabilities of the mechanical test values ​​of the multiple target steel materials is maximized).

[0063] Specifically, when the multiple target steel products are, for example, four target steel products, the downstream process operating condition determination unit 8 determines operating conditions u in the downstream process (continuous annealing process) for the multiple target steel products by solving an optimization problem P2 formulated, for example, as in the following equations (33) to (41) (determines operating conditions u that maximize the geometric mean P(u[1], ...u[4]) of the pass probability).

number

[0064] As described in the first embodiment, if it is assumed that each mechanical test value is independent of the others, the operating conditions x (operating conditions u) that maximize the pass probability P(u), which is expressed as the geometric mean of the pass probability of each mechanical test value, may be searched for and determined without using the covariance matrix Σ(x[i]), as shown in the following equation (42).

number

[0065] <Downstream operation condition output section 9> The downstream process operating condition output unit 9 executes step ST29A shown in Fig. 9. Specifically, the downstream process operating condition output unit 9 outputs the operating conditions (heating plate temperature and soaking plate temperature) in the downstream process for the multiple target steel materials determined by the downstream process operating condition determination unit 8. It is preferable that the downstream process operating condition output unit 9 not only outputs the determined operating conditions in the downstream process, but also outputs the probability distribution (average value, standard deviation) of the mechanical test value output from the material characteristic value prediction model when the determined operating conditions (including the operational results in the upstream process) are input to the material characteristic value prediction model, and the pass probability of the mechanical test value calculated from this probability distribution and order information. Fig. 11 is a diagram showing an example of the contents output from the downstream process operating condition output unit 9. As shown in Fig. 11, it can be seen that the downstream process operating conditions with a high probability of passing can be determined while satisfying the constraint on the conveying order, that is, the change in the heating plate temperature and the soaking plate temperature between the target steel products is within 5°C.

[0066] [Third embodiment] In the operating condition determination device 100A of the second embodiment described above, the transport order of multiple steel products in the downstream process is determined in advance, but the operating condition determination device of the third embodiment (corresponding to the second preferred configuration of the present invention) is configured to first determine the operating conditions in the downstream process independently for each of multiple target steel products for which the transport order has not yet been determined, and then determine the transport order of the multiple target steel products in the downstream process taking into account the constraints imposed by the transport order of the multiple target steel products.

[0067] Fig. 12 is a block diagram showing a schematic configuration of an operation condition determination device according to the third embodiment. Fig. 13 is a flow chart showing a schematic procedure of an operation condition determination method executed by using the operation condition determination device shown in Fig. 12. As shown in FIG. 12, the operation condition determination device 100B according to the third embodiment is composed of, for example, a computer, similar to the operation condition determination device 100 according to the first embodiment, and includes a conveying order determination unit 11 and a conveying order output unit 12 in addition to the respective units similar to those of the operation condition determination device 100 according to the first embodiment. Of the units 1 to 12 included in the operating condition determination device 100B, the operations of the manufacturing record acquisition unit 1, the material characteristic value prediction model construction unit 2, the material characteristic value prediction model storage unit 3 and the material characteristic value prediction model acquisition unit 6 (steps ST31, ST32, ST33, and ST36 shown in FIG. 13) are the same as those in the first embodiment, and therefore will not be described here. Furthermore, the operations of the order information acquiring unit 4 and the upstream process operation results acquiring unit 5 (steps ST34, ST35 shown in FIG. 13) are the same as those in the second embodiment, and therefore detailed explanations will be omitted here. However, for example, the order information and operation results in the upstream process as shown in FIG. 14 are acquired. FIG. 14 is a diagram showing an example of order information acquired by the order information acquiring unit 4 and operation results in the upstream process acquired by the upstream process operation results acquiring unit 5. In the example shown in FIG. 14, the product thickness, which is the order information, is also used as the upstream process operation results as it is. Furthermore, the operations of the constraint condition setting unit 7 and downstream process operating condition determination unit 8 (steps ST37 and ST38 shown in FIG. 13) are the same as those in the first embodiment, and therefore detailed explanations will be omitted here, but the constraint condition setting unit 7 sets constraint conditions for each target steel material, as in the first embodiment. Also, the downstream process operating condition determination unit 8 determines the operating conditions (heating plate temperature and soaking plate temperature) in the downstream process by performing inverse analysis of the material characteristic value prediction model acquired by the material characteristic value prediction model acquisition unit 6, so that the pass probability of the mechanical test value of the target steel material is equal to or higher than a predetermined threshold value (maximum in the third embodiment), as in the first embodiment. The downstream process operation condition output unit 9, the conveying sequence determination unit 11, and the conveying sequence output unit 12 will be described below in order.

[0068] <Downstream operation condition output section 9> The downstream process operating condition output unit 9 executes step ST39 shown in Fig. 13. Specifically, the downstream process operating condition output unit 9 outputs the operating conditions (heating plate temperature and soaking plate temperature) in the downstream process for the multiple target steel materials determined by the downstream process operating condition determination unit 8. It is preferable that the downstream process operating condition output unit 9 not only outputs the determined operating conditions in the downstream process, but also outputs the probability distribution (average value, standard deviation) of the mechanical test value output from the material characteristic value prediction model when the determined operating conditions (including the operational results in the upstream process) are input to the material characteristic value prediction model, and the pass probability of the mechanical test value calculated from this probability distribution and order information. Fig. 15 is a diagram showing an example of the contents output from the downstream process operation condition output unit 9. As shown in Fig. 15, it can be seen that the downstream process operation conditions with a high probability of passing can be determined for each of the target steel products No. 001 to 004. However, it can be seen that if the steel products No. 001 to 004 are transported in the order of the downstream process, the constraint condition on the transport order, that is, the change in the heating plate temperature and the soaking plate temperature between the target steel products is 5°C or less, as described in the second embodiment, is not satisfied.

[0069] <Conveyance order determination unit 11> The transport order determination unit 11 executes step ST39A shown in FIG. 13. Specifically, the transport order determination unit 11 determines the transport order of multiple target steel materials in the downstream process, taking into consideration constraint conditions due to the transport order of multiple target steel materials in the downstream process. Constraint conditions due to the transport order can include the difference in size (product thickness, cold-rolled width) between adjacent target steel materials, and the difference in heating temperature and soaking temperature between adjacent target steel materials. The transport order determination unit 11, for example, searches for and determines the transport order that minimizes the weighted linear sum of these differences. That is, when the multiple target steel materials are four target steel materials, for example, the transport order determination unit 11 determines the transport order of multiple target steel materials in the downstream process by solving the optimization problem P3 formulated as shown in the following equations (43) to (49) (z being the steel material No. of the transport order k that minimizes the evaluation function J). k (Determine).

number

[0070] Examples of a method for solving the optimization problem P3 include a 2-opt optimization method (such as a traveling salesman problem solution using 2-opt or a genetic algorithm). FIG. 16 is a conceptual diagram of an optimization problem applied in the third embodiment. As shown in FIG. 16, the transportation order of the target steel materials can be determined by using a 2-opt optimization method that aims to smoothly transition the product thickness, cold-rolled width, and temperature (heating plate temperature, soaking plate temperature) between the target steel materials (steel materials with diagonal hatching in FIG. 16) for which the transportation order is to be determined, and also smoothly transition between the steel materials before and after the target steel materials (steel materials with dotted hatching in FIG. 16) as necessary. For example, if the 2-opt optimization method is applied under the condition of W={w1, w2, w3, w4}={1, 1, 10000, 10}, Z={z1, z2, z3, z4}={1, 4, 2, 3}. That is, the transportation order of the target steel materials is determined to be steel material No. 001 → steel material No. 004 → steel material No. 002 → steel material No. 003.

[0071] <Transportation order output unit 12> The transport sequence output unit 12 executes step ST39B shown in Fig. 13. Specifically, the transport sequence output unit 12 outputs the transport sequence of the multiple target steel products in the downstream process determined by the transport sequence determination unit 11.

[0072] In addition, after the transport order determination unit 11 determines the transport order of multiple target steel materials in the downstream process, as in the second embodiment, the downstream process operation condition determination unit 8 can determine the operating conditions in the downstream process for the multiple target steel materials by re-inverse analyzing the material property value prediction model under the constraints of the determined transport order so that the pass probability of the calculated material property values ​​of the multiple target steel materials for the multiple target steel materials whose transport order has been determined is all above a predetermined threshold value (maximum).

[0073] [Fourth embodiment] The operating condition determination device 100B of the third embodiment described above is configured to first determine the operating conditions for the downstream process independently for each target steel material, and then determine the transport order of multiple target steel materials in the downstream process, whereas the operating condition determination device of the fourth embodiment (corresponding to the third preferred configuration of the present invention) is configured to determine the operating conditions and transport order for multiple target steel materials in the downstream process at once.

[0074] Fig. 17 is a block diagram showing a schematic configuration of an operation condition determination device according to the fourth embodiment. Fig. 18 is a flow chart showing a schematic procedure of an operation condition determination method executed by using the operation condition determination device shown in Fig. 17. As shown in FIG. 17, the operation condition determination device 100C according to the fourth embodiment is configured, for example, by a computer like the operation condition determination device 100 according to the first embodiment, and includes a downstream process operation condition / conveying order determination unit 13 and a downstream process operation condition / conveying order output unit 14 instead of the downstream process operation condition determination unit 8 and the downstream process operation condition output unit 9 of the operation condition determination device 100 according to the first embodiment. Of the various units 1 to 14 included in the operating condition determination device 100C, the operations of the manufacturing performance acquisition unit 1, the material characteristic value prediction model construction unit 2, the material characteristic value prediction model storage unit 3 and the material characteristic value prediction model acquisition unit 6 (steps ST41, ST42, ST43, and ST46 shown in FIG. 18) are the same as those in the first embodiment, and therefore will not be described here. Moreover, the operations of the order information acquiring unit 4, the upstream process operation record acquiring unit 5, and the constraint condition setting unit 7 (steps ST44, ST45, and ST47 shown in FIG. 18) are the same as those in the second embodiment, and therefore detailed description thereof will be omitted here. The downstream process operation condition / conveyance sequence determination unit 13 and the downstream process operation condition / conveyance sequence output unit 14 will be described below in order.

[0075] <Downstream process operation condition / transport order determination unit 13> The downstream process operation condition / transport order determination unit 13 executes step ST48 shown in Fig. 18. Specifically, the downstream process operation condition / transport order determination unit 13 determines the operation conditions (heating plate temperature and soaking plate temperature) and transport order in the downstream process for the multiple target steel materials at once by performing reverse analysis of the material characteristic value prediction model acquired by the material characteristic value prediction model acquisition unit 6 under the constraint condition by the transport order of the multiple target steel materials in the downstream process so that the pass probabilities of the mechanical test values ​​of the multiple target steel materials are all equal to or higher than a predetermined threshold value (so that the geometric mean of the pass probabilities of the mechanical test values ​​of the multiple target steel materials is maximized in the fourth embodiment).

[0076] Specifically, when the number of target steel products is four, the downstream process operation condition / transportation order determination unit 13 solves an optimization problem P4 formulated as in the following equations (50) to (59), for example, to determine the operation condition u and the transport order (z, which is the steel product No. of the transport order k) in the downstream process (continuous annealing process) for the multiple target steel products. k ) is determined (the operating conditions and transportation sequence that minimize the evaluation function J are determined).

number

[0077] <Downstream operation conditions / transport order output section 14> The downstream process operation condition / transport order output unit 14 executes step ST49 shown in Fig. 18. Specifically, the downstream process operation condition / transport order output unit 14 outputs the operation conditions (heating plate temperature and soaking plate temperature) and transport order in the downstream process for the multiple target steel products determined by the downstream process operation condition / transport order determination unit 13. It is preferable that the downstream process operation condition / transport order output unit 14 not only outputs the determined operation conditions and transport order in the downstream process, but also outputs the probability distribution (average value, standard deviation) of the mechanical test value output from the material characteristic value prediction model when the determined operation conditions (including the operational results in the upstream process) are input to the material characteristic value prediction model, and the pass probability of the mechanical test value calculated from this probability distribution and order information. Fig. 19 is a diagram showing an example of the contents output from the downstream process operation condition / transportation sequence output unit 14. The example shown in Fig. 19 is a result obtained under the condition of W = {w1, w2, w3, w4, w5} = {1, 1, 10000, 10, 500000} in the above-mentioned formula (54). As shown in Fig. 19, it can be seen that downstream process operation conditions with a high probability of passing can be determined while satisfying the constraint condition due to the transport sequence, that is, the change in the heating plate temperature and the soaking plate temperature between the target steel products is within 5°C.

[0078] [Fifth embodiment] The operating condition determination device 100 according to the first embodiment described above is configured to determine the parameters constituting the operating conditions in the downstream process according to the order information with one value. However, when the downstream process is to be executed in the future, taking into account the operational performance of the upstream process, there is a high need to predict the material characteristic value by giving the parameters constituting the operating conditions in the downstream process with a range rather than with one value (or by giving the parameters constituting the operating conditions in the downstream process with a target value), or to determine the parameters constituting appropriate operating conditions with a range rather than with one value (or by determining with a target value) so that the predicted material characteristic value satisfies the order specification. Therefore, the operating condition determination device of the fifth embodiment further includes an operating condition prediction model that inputs the range of operating conditions or target values ​​of operating conditions in downstream processes for the steel material and outputs a probability distribution of operating conditions in downstream processes for the steel material, and is configured to determine the range or target value of operating conditions in downstream processes by inversely analyzing the operating condition prediction model and the material characteristic value prediction model (inversely analyzing a composite function of a function represented by the operating condition prediction model and a function represented by the material characteristic value prediction model) so that the probability of passing the material characteristic values ​​of the target steel material calculated based on the probability distribution of operating conditions in downstream processes for the target steel material output when the range of operating conditions or target values ​​of operating conditions in downstream processes for the target steel material are input into the operating condition prediction model, and the operational performance of the target steel material in upstream processes is input into the material characteristic value prediction model, is equal to or greater than a predetermined threshold value.

[0079] Fig. 20 is a block diagram showing a schematic configuration of an operation condition determination device according to the fifth embodiment. Fig. 21 is a flow chart showing a schematic procedure of an operation condition determination method executed by using the operation condition determination device shown in Fig. 20. As shown in FIG. 20, the operating condition determination device 100D according to the fifth embodiment is composed of, for example, a computer, similarly to the operating condition determination device 100 according to the first embodiment, and includes an operating condition prediction model construction unit 15, an operating condition prediction model storage unit 16 and an operating condition prediction model acquisition unit 17 in addition to the respective units similar to those of the operating condition determination device 100 according to the first embodiment. Of the various units 1 to 17 included in the operating condition determination device 100D, the operations of the material characteristic value prediction model construction unit 2, the material characteristic value prediction model storage unit 3, the order information acquisition unit 4, the upstream operation performance acquisition unit 5 and the material characteristic value prediction model acquisition unit 6 (steps ST52, ST53, ST56, ST57, and ST58 shown in Figure 21) are the same as those in the first embodiment, and therefore will not be described here. The manufacturing performance acquisition unit 1, operating condition prediction model construction unit 15, operating condition prediction model storage unit 16, operating condition prediction model acquisition unit 17, constraint condition setting unit 7, downstream process operating condition determination unit 8 and downstream process operating condition output unit 9 will be described in order below.

[0080] <Manufacturing Record Acquisition Department 1> The manufacturing record acquiring unit 1 executes step ST51 shown in FIG. 21. Specifically, the manufacturing record acquiring unit 1 acquires manufacturing records in the manufacturing process (upstream process and downstream process) of a reference steel material manufactured in the past from a predetermined database (not shown). In the fifth embodiment, in addition to the operational records in the manufacturing process of the reference steel material described in the first embodiment and the mechanical test values ​​measured on the test pieces acquired from the reference steel material, the manufacturing record acquired by the manufacturing record acquiring unit 1 includes the range (instruction range) of the operating conditions in the downstream process for the reference steel material or the target value of the operating conditions. The following Table 1 partially shows an example of the manufacturing record including the range (instruction range) of the operating conditions in the downstream process, and the following Table 2 partially shows an example of the manufacturing record including the target value of the operating conditions in the downstream process. [Table 1] [Table 2]

[0081] <Operation Condition Prediction Model Building Division 15> The operating condition prediction model construction unit 15 executes step ST54 shown in Fig. 21. Specifically, the operating condition prediction model construction unit 15 constructs an operating condition prediction model in which the range (indication range) of operating conditions in downstream processes for steel materials is input and the probability distribution of operating conditions in downstream processes for steel materials is output. Here, the parameter group that constitutes the operational performance of the upstream process is v capl The parameter group that is fixed and set in advance for each product thickness in the downstream process (continuous annealing process) is v prev The operating conditions in the downstream process (continuous annealing process) are u = [u1, u M ], the mechanical test value is y, and the lower limit of the range of the operation conditions in the downstream process is u. lo =[u1 lo ,···u M lo ], and the upper limit of the indication range is u up =[u1 up ,···u M up ], the probability distribution p(y|x lo ,x up ) is expressed by the following equation (60).

number

[0082] The operation condition prediction model construction unit 15 may, for example, calculate the lower limit value u of the indication range of the heating plate temperature and the soaking plate temperature in the continuous annealing process for the reference steel material as shown in Table 1 above. lo =[u ht lo ,u st lo ], the upper limit of the heating plate temperature and soaking plate temperature indication range u up =[u ht up ,u st up ] is input, and the operating record of the heating plate temperature and the soaking plate temperature for the reference steel is calculated. ht , u st Using the known data with the output, the operating condition prediction model p(u|u lo ,u up )=p(u ht ,u st |u ht lo ,u ht up ,u st lo ,u st up For example, such known data may be used as training data and constructed by machine learning. For example, it is possible to construct an operating condition prediction model by using known data as training data and Bayesian estimation such as Gaussian process regression.

[0083] In addition, the operating condition prediction model construction unit 15 may construct an operating condition prediction model that takes target values ​​of operating conditions in downstream processes for steel as input and outputs a probability distribution of operating conditions in downstream processes for steel. As in the previous section, the parameters that make up the operational performance of the upstream process are denoted as ν capl The parameter group that is fixed and set in advance for each product thickness in the downstream process (continuous annealing process) is v prevThe operating conditions in the downstream process (continuous annealing process) are u = [u1, u M ], the mechanical test value is y, and the target value of the operating conditions in the downstream process is u aim =[u1 aim ,···u M aim ], the probability distribution p(y|x lo ,x up ) is expressed by the following equation (61).

number

[0084] The operation condition prediction model construction unit 15 calculates the target values ​​u of the heating plate temperature and the soaking plate temperature in the continuous annealing process for the reference steel material, for example, as shown in Table 2 above. aim =[u ht aim ,u st aim ] is input, and the operating results of the heating plate temperature and the soaking plate temperature for the reference steel are calculated. ht , u st Using the known data with the output, the operating condition prediction model p(u|u aim )=p(u ht ,u st |u ht aim ,u st aim For example, such known data may be used as training data and constructed by machine learning. For example, it is possible to construct an operating condition prediction model by using known data as training data and Bayesian estimation such as Gaussian process regression.

[0085] <Operation condition prediction model storage unit 16> The operating condition prediction model storage unit 16 executes step ST55 shown in Fig. 21. Specifically, the operating condition prediction model storage unit 16 stores the operating condition prediction model p(u ht ,u st |u ht lo ,u ht up ,u st lo ,u st up ) or p(u ht ,u st |u ht aim ,u st aim ) is stored (memorized).

[0086] <Operation condition prediction model acquisition unit 17> The operating condition prediction model acquisition unit 17 executes step ST59 shown in Fig. 21. Specifically, the operating condition prediction model acquisition unit 17 acquires the operating condition prediction model p(u ht ,u st |u ht lo ,u ht up ,u st lo ,u st up ) or p(u ht ,u st |u ht aim ,u st aim ) to get the

[0087] <Constraint Condition Setting Section 7> The constraint condition setting unit 7 executes step ST60 shown in Fig. 21. Specifically, the constraint condition setting unit 7 sets constraint conditions for the back analysis executed by the downstream process operation condition determination unit 8 described below, as necessary. Examples of the constraint conditions set by the constraint condition setting unit 7 include equations (63) to (67) described in the description of the downstream process operation condition determination unit 8 below.

[0088] <Downstream operation condition determination unit 8> The downstream process operating condition determination unit 8 executes step ST61 shown in Fig. 21. Specifically, the downstream process operating condition determination unit 8 determines the range or target value of the operating conditions in the downstream process by performing a reverse analysis of the operating condition prediction model and the material characteristic value prediction model so that the pass probability of the material characteristic value of the target steel material calculated based on the probability distribution of the operating conditions in the downstream process for the target steel material output when the range of the operating conditions or the target value of the operating conditions in the downstream process for the target steel material are input to the operating condition prediction model acquired by the operating condition prediction model acquisition unit 17 and the probability distribution of the material characteristic value of the target steel material output when the operational performance in the upstream process for the target steel material is input to the material characteristic value prediction model is equal to or greater than a predetermined threshold value.

[0089] Specifically, the downstream process operation condition determination unit 8 determines the range of operation conditions u in the downstream process (continuous annealing process) (u lo ,u up ) (pass probability P(u lo ,u up ) is maximized by determining the operating condition u.

number

[0090] In the fifth embodiment, even if the material characteristic value prediction model is a deterministic model (a model that outputs only expected values, rather than a probability distribution), it is possible to search for and determine the range of appropriate operating conditions in downstream processes. In this case, ∫N(y s ;μ s (x),σ s (x))p(u|u lo ,u up )du can be rewritten as the following equation (72):

number

[0091] <Downstream operation condition output section 9> The downstream process operational condition output unit 9 executes step ST62 shown in Fig. 21. Specifically, the downstream process operational condition output unit 9 outputs the range or target value of the operational conditions (heating plate temperature and soaking plate temperature) in the downstream process for the target steel material determined by the downstream process operational condition determination unit 8. It is preferable that the downstream process operational condition output unit 9 not only outputs the determined range or target value of the operational conditions in the downstream process, but also outputs the probability distribution (average value, standard deviation) of the mechanical test value output from the material characteristic value prediction model when the determined range or target value of the operational conditions (including the operational performance in the upstream process) is input to the material characteristic value prediction model, and the pass probability of the mechanical test value calculated from this probability distribution and the order information. Fig. 22 is a diagram showing an example of the contents output from the downstream process operating condition output unit 9. As shown in Fig. 22, when the range of the operating conditions in the downstream process is determined to be 832°C as the lower limit and 852°C as the upper limit of the heating plate temperature in the continuous annealing process, and 827°C as the lower limit and 847°C as the upper limit of the soaking plate temperature, the pass probability of the tensile strength (TS) and the yield point (YP) is a high value of 0.987. However, in the fifth embodiment, unlike the first embodiment, the parameters (heating plate temperature and soaking plate temperature) constituting the operating conditions in the downstream process are determined as a range rather than as a single value, and the pass probability is calculated taking into account the variation in the operating conditions. Therefore, the pass probability is calculated to be smaller than the result shown in Fig. 6 described above.

[0092] In the embodiment described above, the function of constructing a material property value prediction model and the function of calculating the pass probability of the material property value of steel using the material property value prediction model are provided in the same operational condition determination device 100 (or operational condition determination device 100A, 100B, 100C or 100D), but these functions may also be provided in different devices. In addition, in the fifth embodiment, the function of constructing an operating condition prediction model and the function of calculating the probability distribution of operating conditions in a downstream process using the operating condition prediction model are provided in the same operating condition determination device 100D, but these functions may be provided in different devices.

[0093] Furthermore, the operation condition determination apparatus 100 (or the operation condition determination apparatus 100A, 100B, 100C, or 100D) may be implemented using a plurality of computers. For example, the operation condition determination apparatus 100 may be implemented using an apparatus such as a cloud server. Furthermore, the hardware processor and memory of the operation condition determination apparatus 100 may be distributed and implemented in a plurality of computers.

[0094] Although one embodiment of the present invention has been described in detail above with reference to the drawings, the specific configuration is not limited to this embodiment, and designs that do not deviate from the gist of the present invention are also included. [Explanation of symbols]

[0095] 1. Manufacturing performance acquisition department 2. Material property value prediction model construction section 3. Material property value prediction model storage section 4. Order information acquisition section 5. Upstream Operational Performance Acquisition Department 6. Material property value prediction model acquisition section 7. Constraint Condition Setting Section 8. Downstream operation condition determination section 9. Downstream process operation condition output section 10...Transportation order acquisition unit 11. Transport order determination section 12. Conveying order output section 13. Downstream process operation conditions and transport order determination section 14. Downstream process operation conditions and transport order output section 15. Operational condition prediction model construction section 16. Operating condition prediction model storage section 17. Operational condition prediction model acquisition section 100, 100A, 100B, 100C... Operating condition determination device

Claims

1. An operating condition determination device for determining operating conditions in a downstream process according to order information after an upstream process is executed for a steel material manufactured through a manufacturing process having an upstream process and a downstream process, comprising: A material characteristic value prediction model is provided that inputs operational conditions in the upstream process and the downstream process for the steel material and outputs a probability distribution of material characteristic values ​​of the steel material, The operating conditions in the downstream process are determined by inversely analyzing the material property value prediction model so that the pass probability of the material property value of the target steel material, calculated from the probability distribution of the material property value of the target steel material output when the operational performance of the upstream process and the operating conditions in the downstream process for the target steel material are input into the material property value prediction model and the specifications of the material property value of the target steel material included in the order information, is equal to or greater than a predetermined threshold value. Operation condition determination device.

2. The method determines operating conditions in the downstream process for the plurality of target steel materials, the transport order of which has been determined in advance, by performing a reverse analysis on the material characteristic value prediction model under constraints imposed by the transport order so that any of the calculated pass probabilities or the geometric mean of the pass probabilities is equal to or greater than a predetermined threshold value. The operating condition determining device according to claim 1 .

3. determining operating conditions for the plurality of target steel materials in the downstream process by reverse analyzing the material characteristic value prediction model so that all of the pass probabilities calculated for the plurality of target steel materials or the geometric mean of the pass probabilities is equal to or greater than a predetermined threshold value; and then determining the transport order of the plurality of target steel materials in the downstream process by solving an optimization problem that takes into account constraints imposed by the transport order of the plurality of target steel materials in the downstream process. The operating condition determining device according to claim 1 .

4. After determining the transport order of the plurality of target steel materials in the downstream process, the material characteristic value prediction model is reverse-analyzed again under the constraint conditions imposed by the determined transport order so that each of the plurality of pass probabilities calculated for the plurality of target steel materials or the geometric mean of the pass probabilities is equal to or greater than a predetermined threshold value, thereby determining the operating conditions in the downstream process for the plurality of target steel materials. The operating condition determining device according to claim 3.

5. The material characteristic value prediction model is reverse-analyzed under constraints imposed by the transport order of the plurality of target steel materials in the downstream process so that all of the plurality of pass probabilities calculated for the plurality of target steel materials or the geometric mean of the pass probabilities is equal to or greater than a predetermined threshold value, thereby determining the operating conditions and transport order in the downstream process for the plurality of target steel materials at once. The operating condition determining device according to claim 1 .

6. An operating condition prediction model that inputs a range of operating conditions or a target value of operating conditions in the downstream process for the steel material and outputs a probability distribution of operating conditions in the downstream process for the steel material, The range or target value of the operating conditions in the downstream process for the target steel material is determined by reverse analysis of the operating condition prediction model and the material characteristic value prediction model so that the pass probability of the material characteristic value of the target steel material calculated based on the probability distribution of the material characteristic value of the target steel material output when the range of operating conditions or the target value of the operating conditions in the downstream process for the target steel material are input into the operating condition prediction model and the operational performance of the target steel material in the upstream process is input into the material characteristic value prediction model is equal to or greater than a predetermined threshold value. The operating condition determining device according to claim 1 .

7. The material characteristic value prediction model is a model using Bayesian estimation.

7. An operating condition determining device according to claim 1.

8. The material characteristic value prediction model is a model using a heteroscedastic Gaussian process.

7. An operating condition determining device according to claim 1.

9. A method for determining operating conditions for a steel material manufactured through a manufacturing process having an upstream process and a downstream process, the method determining operating conditions for the downstream process according to order information after the upstream process is executed, comprising the steps of: A material characteristic value prediction model is used, which inputs the operational conditions in the upstream process and the downstream process for the steel material and outputs a probability distribution of the material characteristic value of the steel material, The operating conditions in the downstream process are determined by inversely analyzing the material property value prediction model so that the pass probability of the material property value of the target steel material, calculated from the probability distribution of the material property value of the target steel material output when the operational performance of the upstream process and the operating conditions in the downstream process for the target steel material are input into the material property value prediction model and the specifications of the material property value of the target steel material included in the order information, is equal to or greater than a predetermined threshold value. How to determine operating conditions.

10. A program for causing a computer to function as the operation condition determination device according to any one of claims 1 to 6.

Images