Method for calculating the time steel material remains in a continuous heating furnace and method for manufacturing thick steel plates.

Abstract

To provide a method for calculating the time steel material remains in the furnace in a continuous heating furnace, which can accurately calculate the time steel material remains in the furnace. [Solution] The method for calculating the time a steel material is in a continuous heating furnace according to the present invention includes: a determination process in which, at the stage of charging the steel material to be used for the calculation of the time in the furnace, the extraction explanatory variable of the target material is input into a determination formula for all target materials which are steel materials charged into the continuous heating furnace, and the rolling pattern of the target material is determined to be either single rolling or double rolling based on the value of the objective variable; and a calculation process in which the rolling time of all target materials is calculated based on the result of the determination process, and the sum of the rolling times of all target materials is calculated as the time a steel material is in the furnace when it is charged into the continuous heating furnace.

Description

Technical Field

[0001] The present invention relates to a method for calculating the residence time of a steel material in a continuous heating furnace and a method for manufacturing a thick steel plate.

Background Art

[0002] In the manufacturing process of thick steel plates produced by rolling with a hot rolling mill, steel materials such as slabs are heated to a target temperature in a heating furnace before rolling. This heating furnace is roughly classified into two types: a continuous heating furnace capable of heating steel materials on the line and a batch heating furnace that performs heating offline and then transfers the steel materials to the line. The former has higher energy efficiency.

[0003] Each steel material charged into the continuous heating furnace is given a different target temperature in order to obtain predetermined material properties. Therefore, it is necessary to calculate in advance the time from when the steel material is charged into the continuous heating furnace until it is extracted as the residence time, and calculate in advance how many degrees Celsius the steel material needs to be heated in each zone in the continuous heating furnace so that the temperature of the steel material at the time of extraction becomes the target temperature based on the residence time. Hereinafter, the heating pattern of the steel material in each zone in the continuous heating furnace is called a heat pattern.

[0004] Steel materials with different heat patterns are charged into the continuous heating furnace, and multiple steel materials are heated in one continuous heating furnace. Therefore, not only the maximization of the heating efficiency but also the reduction of the fuel unit consumption are required. For this reason, accurately calculating the residence time of the steel materials to be charged into the continuous heating furnace in the future greatly contributes to the achievement of the requirements for the continuous heating furnace. Against such a background, Patent Document 1 proposes a method for calculating the residence time of a steel material (this material) charged into a continuous heating furnace as the sum of the times required for the rolling process of the steel materials charged into the continuous heating furnace before this material (hereinafter, rolling time).

Prior Art Documents

Patent Documents

[0005] [Patent Document 1] Japanese Patent Publication No. 2006-274402 [Overview of the project] [Problems that the invention aims to solve]

[0006] In lines with only one hot rolling mill, steel materials are generally rolled using either single rolling or double rolling patterns. Single rolling refers to a rolling pattern in which no other steel materials are rolled between the start of rolling a particular steel material (the start of the first rolling pass on the heated steel material) and the end of rolling (the end of the final rolling pass in which the steel material is rolled to the product thickness). Double rolling, on the other hand, refers to a rolling pattern in which other steel materials are rolled between the start and end of rolling a particular steel material.

[0007] Generally, controlled rolling, which involves rolling in the non-recrystallized temperature range, is known as a method for manufacturing thick steel plates with high strength and toughness. This method involves, for example, rolling a steel material heated to over 1000°C to a medium thickness, then performing a temperature adjustment treatment (water cooling treatment) to lower the temperature to near the non-recrystallized temperature range, and finally rolling the steel material again once it has reached that range. In this case, the medium thickness is the thickness at which controlled rolling begins. Since it takes time for the temperature of the steel material rolled to a medium thickness to drop to near the non-recrystallized temperature range, the double rolling method described above is a rolling pattern that increases the efficiency of the hot rolling mill by rolling other steel materials during the temperature adjustment treatment.

[0008] As mentioned above, in the case of double rolling, the rolling of one steel material takes place between the start and end of the rolling of another steel material. Therefore, when the rolling pattern is double rolling, if the rolling time of the steel material is calculated as the total time required from the start to the end of the rolling of the steel material (hereinafter referred to as the total rolling time), the rolling time of the steel material cannot be calculated accurately. Specifically, as shown in Figure 5(a), the total rolling time of the target steel material A is calculated as the time including the rolling time of the other steel material B which is carried out during the temperature adjustment time (water cooling time), so the rolling time of steel material A cannot be calculated accurately. Accordingly, when the rolling pattern is double rolling, as shown in Figure 5(b), it is necessary to calculate the time when steel material A actually occupies the hot rolling mill (rolling mill occupancy time) between the start to the end of the rolling of steel material A, which does not include the rolling time of the other steel material B.

[0009] However, the method described in Patent Document 1 above calculates the rolling time of the steel material without considering whether the rolling pattern of the steel material is single rolling or double rolling. Therefore, if the method described in Patent Document 1 is applied to a line that performs double rolling, it may not be possible to accurately calculate the rolling time of the steel material, and consequently, the time the steel material stays in the furnace. If the steel material is extracted in a shorter time than the calculated time in the furnace, problems such as insufficient heating of the steel material may occur, resulting in waiting time for the hot rolling mill or additional combustion gas being added to the continuous heating furnace, leading to a decrease in the efficiency of the rolling mill and a loss of fuel per unit. Similarly, if the steel material is extracted in a longer time than the calculated time in the furnace, a loss of fuel per unit will occur due to overheating of the steel material.

[0010] The present invention was made to solve the above problems, and its objective is to provide a method for calculating the time a steel material is in a continuous heating furnace that can accurately calculate the time the steel material is in the furnace. Another objective of the present invention is to provide a method for manufacturing thick steel plates that can produce thick steel plates in an energy-efficient manner. [Means for solving the problem]

[0011] The present invention relates to a method for calculating the time a steel material remains in a furnace in a continuous heating furnace, which is used when manufacturing thick steel plates by heating and rolling a steel material in a continuous heating furnace. The method involves performing machine learning on a predetermined target variable set for each case: when the rolling pattern of the steel material is single rolling, in which no other steel material is rolled between the start and end of rolling of the steel material, and when the rolling pattern of the steel material is double rolling, in which other steel material is rolled between the start and end of rolling of the steel material, the explanatory variables are the rolling parameters of the steel material that has been heated in the continuous heating furnace in the past, and the rolling parameters of the steel material that was loaded into the continuous heating furnace immediately before the steel material. The machine learning process involves performing machine learning on a predetermined target variable set for each case, and extracting the explanatory variables that have a high influence on the target variable as extracted explanatory variables. The method includes: a second learning process in which the actual data of extracted explanatory variables are used as explanatory variables, and a regression calculation is performed on a predetermined target variable set for each case in which the rolling pattern of the steel material corresponding to the explanatory variable is single rolling and double rolling, thereby deriving a regression equation showing the relationship between the extracted explanatory variables and the target variable extracted in the first learning process as a determination equation; a determination process in which, for all target materials which are steel materials loaded into the continuous heating furnace at the stage in which the steel material to be used for the calculation of in-furnace time is loaded, the extracted explanatory variables of the target material are input into the determination equation, and a determination process in which the rolling pattern of the target material is single rolling or double rolling is determined based on the value of the target variable; and a calculation process in which the rolling time of all the target materials is calculated based on the result of the determination process, and the sum of the rolling times of all the target materials is calculated as the in-furnace time of the steel material loaded into the continuous heating furnace.

[0012] The extracted explanatory variables are preferably one or more selected from the following: the thickness of the previous material, the length of the previous material, the thickness of the current material, the target value for the end of large-plate controlled rolling of the current material, the target value for the start of controlled cooling of the previous material, and the target value for the end of large-plate controlled rolling of the previous material.

[0013] The calculation process may include the following steps: if the rolling pattern of the target material is single rolling, calculate the rolling time of the target material using the rolling time calculation formula for water-cooled material single rolling; if the rolling pattern of the target material is double rolling, calculate the rolling time of the target material using the rolling time calculation formula for water-cooled material double rolling; and if the target material is not water-cooled, calculate the rolling time of the target material using the rolling time calculation formula for air-cooled material.

[0014] The method for manufacturing thick steel plates according to the present invention includes the step of controlling the operation of a continuous heating furnace according to the furnace time calculated by the method for calculating the furnace time of steel material in a continuous heating furnace according to the present invention. [Effects of the Invention]

[0015] The method for calculating the furnace time of steel material in a continuous heating furnace according to the present invention allows for accurate calculation of the furnace time of steel material. Furthermore, the method for manufacturing thick steel plates according to the present invention allows for energy-efficient production of thick steel plates. [Brief explanation of the drawing]

[0016] [Figure 1] Figure 1 is a block diagram showing the configuration of a heating furnace control system, which is one embodiment of the present invention. [Figure 2] Figure 2 is a flowchart showing the flow of the determination formula derivation process, which is one embodiment of the present invention. [Figure 3] Figure 3 is a flowchart showing the flow of the rolling time calculation process according to one embodiment of the present invention. [Figure 4] Figure 4 shows the difference between the actual rolling time and the predicted rolling time in the inventive example and the conventional example. [Figure 5] Figure 5 shows the total rolling time and rolling mill occupancy time in double rolling. [Modes for carrying out the invention]

[0017] Hereinafter, a reheating furnace control system according to an embodiment of the present invention will be described with reference to the drawings.

[0018] 〔Configuration〕 FIG. 1 is a block diagram showing the configuration of a reheating furnace control system according to an embodiment of the present invention. As shown in FIG. 1, a reheating furnace control system 1 according to an embodiment of the present invention is a system that controls the operation of a continuous reheating furnace A that heats a steel material such as a slab to be rolled in a hot rolling mill to a target temperature, and includes a host computer 2 and a reheating furnace control device 3.

[0019] The host computer 2 is composed of an information processing device such as a business computer or a process computer, and is connected to the reheating furnace control device 3 via a telecommunication line such as the Internet or a LAN. The host computer 2 stores various information (material information, heating condition information, rolling condition information, product information, etc., which will be described later) regarding the steel material to be rolled in the hot rolling mill and the steel material for which rolling in the hot rolling mill has been completed.

[0020] The reheating furnace control device 3 is composed of an information processing device such as a personal computer, and is connected to the host computer 2 and the continuous reheating furnace A via a telecommunication line. The reheating furnace control device 3 controls the heating process of the steel material in the continuous reheating furnace A using various information regarding the steel material acquired from the host computer 2. The reheating furnace control device 3 functions as a determination formula derivation unit 31, a rolling time calculation unit 32, a furnace residence time calculation unit 33, and a reheating furnace control unit 34 when an arithmetic processing device such as a CPU in the information processing device executes a computer program. The functions of each of these units will be described later.

[0021] Note that part or all of the functions of the determination formula derivation unit 31, the rolling time calculation unit 32, and the furnace residence time calculation unit 33 may be executed by the host computer 2. In other words, part of the reheating furnace control device 3 may be constituted by the host computer 2. Further, the functions of the reheating furnace control device 3 may be distributed among a plurality of information processing devices connected to each other via a telecommunication line.

[0022] In the heating furnace control system 1 having such a configuration, the heating furnace control device 3 accurately calculates the residence time, which is the time from when a steel material is charged into the continuous heating furnace A until it is extracted, by executing the determination formula derivation process and the rolling time calculation process shown below. Hereinafter, referring to the flowcharts shown in FIGS. 2 and 3, the operation of the heating furnace control device 3 when executing the determination formula derivation process and the rolling time calculation process will be described.

[0023] 〔Determination formula derivation process〕 FIG. 2 is a flowchart showing the flow of the determination formula derivation process according to an embodiment of the present invention. The flowchart shown in FIG. 2 starts at the timing when an execution command for the determination formula derivation process is input to the heating furnace control device 3, and the determination formula derivation process proceeds to the process of step S1.

[0024] In the process of step S1, first, the determination formula derivation unit 31 acquires from the host computer 2 the performance data of the rolling parameters of the current material, which is a steel material that has been rolled in the hot rolling mill in the past, and the performance data of the rolling parameters of the previous material, which is the steel material charged into the continuous heating furnace A immediately before the current material. Here, examples of the rolling parameters include the material information (thickness, width, length, steel type, etc.) of the steel material and the rolling condition information (temperature (start, end), reduction rate per pass, thickness after each pass, rolling restart time after rolling interruption during controlled rolling, etc.).

[0025] Next, the determination formula derivation unit 31 uses the acquired actual values ​​of the rolling parameters of the current material and the previous material as explanatory variables and performs machine learning on predetermined target variables set for both the case where the rolling pattern of the current material is single rolling and the case where the rolling pattern is double rolling, thereby extracting a predetermined number of explanatory variables that have a high influence on the target variable (first learning process). Among the explanatory variables, there are some, such as slab thickness, which take almost the same value regardless of the target material, but even a small change in it can have a big effect on the result (= target variable), so the purpose of performing the first learning process is to exclude such variables. An example of a target variable is one whose value is 0 when the rolling pattern of the current material is single rolling and whose value is 1 when the rolling pattern is double rolling. With this, the processing of step S1 is completed, and the determination formula derivation process proceeds to the processing of step S2.

[0026] Here, the explanatory variables that have a high influence on the dependent variable are preferably one or more selected from the following: the thickness of the previous material (product thickness), the length of the previous material (product length), the thickness of the current material, the target value for the end of controlled rolling of the current material, the target value for the start of controlled cooling of the previous material, and the target value for the end of controlled rolling of the previous material. Thick steel plates are usually manufactured by hot-rolling a steel material such as a slab to a predetermined product thickness and then cutting or shearing the steel plate to predetermined dimensions. In this specification, a hot-rolled steel plate in the rolling stage (including the state during rolling) or the cooling stage following rolling, and in which it has not been cut or sheared to product size, etc., may be referred to as a "large plate".

[0027] In step S2, the determination formula derivation unit 31 uses the actual data of the explanatory variables extracted in the first learning process as explanatory variables, and performs regression calculations on predetermined target variables set for both the case where the rolling pattern of the steel material corresponding to the explanatory variables is single rolling and the case where the rolling pattern is double rolling. By doing so, it derives a regression equation showing the relationship between the explanatory variables extracted in the first learning process and the target variable as a determination formula (second learning process). An example of a target variable is one whose value is 0 when the rolling pattern of the material is single rolling and whose value is 1 when the rolling pattern is double rolling. With this, the process of step S2 is completed, and the series of determination formula derivation processes is finished.

[0028] [Rolling time calculation process] Figure 3 is a flowchart showing the flow of the rolling time calculation process according to one embodiment of the present invention. The flowchart shown in Figure 3 starts when an execution command for the rolling time calculation process is input to the heating furnace control device 3, and the rolling time calculation process proceeds to step S11. The rolling time calculation process is performed for all steel materials (hereinafter referred to as "target materials") that are charged into the continuous heating furnace A at the stage when the steel materials to be used for the calculation of in-furnace time are charged.

[0029] In step S11, the rolling time calculation unit 32 obtains rolling condition information for the target material from the upper-level computer 2. Based on the obtained rolling condition information, the rolling time calculation unit 32 determines whether or not each target material is water-cooled. If the determination shows that the target material is not water-cooled but air-cooled (steel material cooled on the line or in a cooling yard) (step S11: No), the rolling time calculation unit 32 proceeds to step S12 for the rolling time calculation process. On the other hand, if the target material is water-cooled (step S11: Yes), the rolling time calculation unit 32 proceeds to step S13 for the rolling time calculation process.

[0030] In step S12, the rolling time calculation unit 32 calculates the rolling time of the target material using a pre-prepared rolling time calculation formula for air-cooled materials. This completes step S12, and the series of rolling time calculation processes is finished.

[0031] In step S13, the rolling time calculation unit 32 calculates the value Y of the target variable of the target material by inputting the explanatory variables of the target material into the determination formula derived by the determination formula derivation process. The rolling time calculation unit 32 then determines whether the calculated value Y is greater than or equal to a predetermined threshold Yth. If the value Y is greater than or equal to the threshold Yth, the rolling time calculation unit 32 determines that the rolling pattern of the target material is single rolling and proceeds to step S14. On the other hand, if the value Y is less than the threshold Yth, the rolling time calculation unit 32 determines that the rolling pattern of the target material is double rolling and proceeds to step S15.

[0032] In step S14, the rolling time calculation unit 32 calculates the rolling time of the target material using a pre-prepared formula for calculating the rolling time of a single water-cooled material. This completes step S14, and the series of rolling time calculation processes is finished.

[0033] In step S15, the rolling time calculation unit 32 calculates the rolling time of the target material using a pre-prepared formula for calculating the rolling time of water-cooled double rolling. This completes step S15, and the series of rolling time calculation processes is finished.

[0034] Subsequently, the furnace time calculation unit 33 calculates the total rolling time of all target materials as the furnace time for the steel material to be loaded into the continuous heating furnace A. Then, the heating furnace control unit 34 controls the heating operation of the continuous heating furnace A by controlling its operation based on the calculated furnace time.

[0035] As is clear from the above explanation, in the heating furnace control system 1, which is one embodiment of the present invention, the heating furnace control device 3 inputs the explanatory variables of the target material into a determination formula for all target materials, which are steel materials charged into the continuous heating furnace A at the stage when the steel material to be used for calculating the time spent in the furnace is charged, determines whether the individual rolling pattern of the target material is single rolling or double rolling based on the value of the objective variable, calculates the rolling time of all target materials based on the determination result, and calculates the total value of the rolling times of all target materials as the time spent in the furnace of the steel material charged into the continuous heating furnace A. With this configuration, the rolling time of the target material is calculated considering whether the rolling pattern of the target material is single rolling or double rolling, so the time spent in the furnace of the steel material can be calculated with high accuracy. Furthermore, as a result, thick steel plates can be manufactured with energy efficiency.

[0036] [Examples] In the example invention, first, a first learning process was performed. In the first learning process, the rolling performance parameters of the current material, which is a steel material that has been rolled in the past, and the rolling performance parameters of the previous material, which is a steel material that was loaded into the continuous heating furnace one step before the current material, were obtained from the host computer. As rolling performance parameters, material information of the steel material (thickness, width, length, steel type, etc.) and rolling condition information (temperature (start, end), reduction ratio for each pass, thickness after each pass, time to restart rolling after interruption during controlled rolling, etc.) were used. Next, machine learning was performed on a target variable, with the rolling performance parameters of the current material and the previous material as explanatory variables, and the value set to 0 if the rolling pattern of the current material is single rolling, and the value set to 1 if the rolling pattern is double rolling. As a result, the following explanatory variables were identified as having a high influence on the objective variable: the thickness a of the previous material, the length b of the previous material, the thickness c of the current material, the target value d for the end of controlled rolling of the current material, the target value e for the start of controlled cooling of the previous material, and the target value f for the end of controlled rolling of the previous material.

[0037] Next, a second learning process was performed. In the second learning process, the explanatory variables were the thickness a of the previous material, the length b of the previous material, the thickness c of the current material, the target value d for the end of large-plate controlled rolling of the current material, the target value e for the start of controlled cooling of the previous material, and the target value f for the end of large-plate controlled rolling of the previous material. In the second learning process, the value of the objective variable was set to 0 when the rolling pattern of the current material was single rolling, where no other steel material was rolled between the start and end of rolling of the current material, and to 1 when the rolling pattern was double rolling, where other steel material was rolled between the start and end of rolling of the current material. By performing the regression calculation in this manner, the determination formula Y shown in the following equation (1) was obtained. Furthermore, if the value obtained by inputting the explanatory variables of the target material for which the rolling time is to be calculated is greater than or equal to the threshold Yth (=0.55), it was determined that the rolling pattern of the target material is double rolling; if it is less than the threshold Yth, it was determined that the rolling pattern of the target material is single rolling. Then, the rolling time of the target material was calculated based on the judgment result. In conventional examples, the rolling time was calculated using a regression equation obtained based on the correlation between various parameters such as slab size, product size, and various temperature conditions, and past actual rolling times. Furthermore, no distinction was made between single rolling and double rolling when determining the regression equation. Therefore, in conventional examples, the rolling time was calculated using a regression equation that did not distinguish between single rolling and double rolling.

[0038]

number

[0039] Figures 4(a) and 4(b) show the results of calculating the difference between the actual rolling time and the calculated rolling time for the inventive example and the conventional example when the rolling pattern is single rolling. For the inventive example, a frequency distribution diagram was created for the value obtained by subtracting the calculated rolling time from the actual rolling time. The white bars in Figure 4(a) represent the case where the rolling pattern is single rolling, and the white bars in Figure 4(b) represent the case where the rolling pattern is double rolling. For the conventional example, the calculated rolling time (predicted rolling time) was obtained using a common regression equation regardless of whether the rolling pattern was single or double rolling, and the frequency distribution of the value obtained by subtracting the calculated rolling time from the actual rolling time was obtained. Then, after confirming whether the rolling pattern corresponding to each individual performance value was single rolling or double rolling, all performance values ​​in the conventional example were identified as single rolling or double rolling. The frequency distribution for single rolling is shown as a shaded bar in Figure 4(a), and the frequency distribution for double rolling is shown as a shaded bar in Figure 4(b).

[0040] As shown in Figures 4(a) and 4(b), the inventive example shows a smaller variation in the difference between the actual rolling time and the predicted rolling time compared to the conventional example. This confirms that the inventive example allows for accurate calculation of the rolling time. Furthermore, comparing the combined accuracy of single rolling and double rolling with the prediction accuracy of the conventional example, the prediction accuracy improved by 30 s / roll in absolute value and the deviation by 20 s / roll.

[0041] Although embodiments applying the invention made by the present inventors have been described above, the present invention is not limited by the descriptions and drawings that constitute part of the disclosure of the present invention in this embodiment. That is, all other embodiments, examples, and operational techniques made by those skilled in the art based on this embodiment are included in the scope of the present invention. [Explanation of Symbols]

[0042] 1. Heating furnace control system 2. Higher-level computers 3. Heating furnace control device 31 Judgment formula derivation part 32 Rolling time calculation unit 33 Furnace time calculation section 34 Heating Furnace Control Unit A Continuous heating furnace

Claims

1. A method for calculating the time a steel material is in a continuous heating furnace when manufacturing thick steel plates by heating and rolling the steel material in a continuous heating furnace, For the present material, which is a steel material that has been heated in a continuous heating furnace in the past, the rolling parameters of the present material and the rolling parameters of the previous material, which is a steel material that was loaded into the continuous heating furnace immediately before the present material, are used as explanatory variables. Machine learning is performed on predetermined target variables set for each case in which the rolling pattern of the present material is single rolling, in which no other steel material is rolled between the start and end of rolling of the present material, and double rolling, in which other steel material is rolled between the start and end of rolling of the present material. The first learning process extracts the explanatory variables that have a high influence on the target variable as extracted explanatory variables. A second learning process derives a regression equation as a decision equation, which uses the actual data of the extracted explanatory variables extracted in the first learning process as explanatory variables, and performs regression calculations on predetermined target variables set for each case where the rolling pattern of the steel material corresponding to the explanatory variable is single rolling and where the rolling pattern is double rolling, thereby deriving a regression equation that shows the relationship between the extracted explanatory variables extracted in the first learning process and the target variable. At the stage of charging the steel material to be used for calculating the time spent in the furnace, for all target materials that are steel materials charged into the continuous heating furnace, the extracted explanatory variables for the target material are input into the determination formula, and a determination process is performed to determine whether the rolling pattern of the target material is single rolling or double rolling based on the value of the objective variable. A calculation process is performed to calculate the rolling time of all the target materials based on the result of the determination process, and to calculate the total value of the rolling times of all the target materials as the furnace time of the steel material to be charged into the continuous heating furnace. A method for calculating the time steel material remains in a furnace in a continuous heating furnace, including the following.

2. A method for calculating the time a steel material is in a furnace in a continuous heating furnace according to claim 1, wherein the extracted explanatory variable is one or more selected from the plate thickness of the previous material, the plate length of the previous material, the plate thickness of the current material, the target value for the end of large-plate controlled rolling of the current material, the target value for the start of controlled cooling of the previous material, and the target value for the end of large-plate controlled rolling of the previous material.

3. A method for calculating the time a steel material is in a furnace in a continuous heating furnace according to claim 1, wherein the calculation process includes the following steps: if the rolling pattern of the material is single rolling, the rolling time of the material is calculated using a rolling time calculation formula for water-cooled single rolling; if the rolling pattern of the material is double rolling, the rolling time of the material is calculated using a rolling time calculation formula for water-cooled double rolling; and if the material is not water-cooled, the rolling time of the material is calculated using a rolling time calculation formula for air-cooled material.

4. A method for manufacturing a thick steel plate, comprising the step of controlling the operation of a continuous heating furnace according to the in-furnace time calculated by the method for calculating the in-furnace time of a steel material in a continuous heating furnace according to any one of claims 1 to 3.

Images