Surface defect prediction method, steel strip manufacturing method, and surface defect prediction device

Abstract

To provide a surface defect prediction method capable of predicting the presence or absence of surface defects of a steel strip that become apparent after cold rolling, using operation parameters from processes prior to a pickling process.SOLUTION: A surface defect prediction method for a steel strip manufactured by making molten steel refined in a secondary refining device a slab with a continuous casting device, hot rolling the slab with a hot rolling device, annealing with an annealing device, pickling with a pickling device, and cold rolling with a cold rolling device, includes: a parameter acquisition step of acquiring one or more parameters in operation parameters from the secondary refining device, one or more parameters in operation parameters from the continuous casting device, one or more parameters in operation parameters from the hot rolling device, and one or more parameters in operation parameters from the pickling device; and a surface defect prediction step of inputting input data including the parameters acquired in the parameter acquisition step into a surface defect prediction model, and outputting the presence or absence of surface defects in the steel strip after cold rolling to predict the presence or absence of occurrence of surface defects.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a surface defect prediction method capable of predicting the occurrence of surface defects in a steel strip, a surface defect prediction device, and a steel strip manufacturing method using the surface defect prediction method. [Background technology]

[0002] In the production of stainless steel cold-rolled annealed steel strips, an annealing process is carried out following the hot rolling process, and then an oxidized scale on the steel strip surface is removed by a pickling process. In the pickling process, pickling conditions are controlled to increase the amount of dissolution, thereby removing surface defects of the steel strip that occurred during the steelmaking and hot rolling processes. However, increasing the amount of dissolution in the pickling process can reduce the yield of the steel strip and increase the amount of sludge produced. For this reason, it is not preferable to increase the amount of dissolution in the pickling process, and it is preferable to minimize the amount of dissolution in the pickling process within the range in which surface defects are removed.

[0003] Since the degree of surface defects varies from product to product, the optimum conditions for the pickling equipment also vary from product to product. As such, it is difficult to find the optimum conditions for the pickling equipment for each steel strip product, for which the degree of surface defects varies from product to product. To address this issue, Patent Document 1 discloses a technology for predicting the properties of a final product based on the properties of an intermediate product, and calculating a new manufacturing route if it is predicted that the deviation from the desired properties will exceed a threshold value. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Special Publication No. 2022-515094 Summary of the Invention [Problem to be solved by the invention]

[0005] Even if the technology disclosed in Patent Document 1 is applied, there is a problem in that it is not possible to predict the presence or absence of surface defects in a steel strip that will become apparent after cold rolling based on the characteristics of the intermediate product. An object of the present invention is to provide a surface defect prediction method and surface defect prediction device that can predict the presence or absence of surface defects in a steel strip that will become apparent after cold rolling using operational parameters before the pickling process, and a steel strip manufacturing method that uses the surface defect prediction method. [Means for solving the problem]

[0006] The means for solving the above problems are as follows. [1] A method for predicting surface defects in a steel strip produced by forming molten steel refined in a secondary refining device into a slab in a continuous casting device, hot rolling the slab in a hot rolling device, annealing the slab in an annealing device, pickling in a pickling device, and cold rolling the slab in a cold rolling device, the method comprising: a parameter acquisition step of acquiring one or more parameters among the operation parameters of the secondary refining device, one or more parameters among the operation parameters of the continuous casting device, one or more parameters among the operation parameters of the hot rolling device, and one or more parameters among the operation parameters of the pickling device; and a surface defect prediction step of inputting input data including the parameters acquired in the parameter acquisition step into a surface defect prediction model, outputting the presence or absence of surface defects in the steel strip after the cold rolling, and predicting the presence or absence of the surface defects. [2] The surface defect prediction method according to [1], wherein the parameter acquisition step acquires operational parameters of a secondary refining device including a converter reduction time, operational parameters of a continuous casting device including a mold current value, operational parameters of a hot rolling mill including an angular temperature of the slab extracted from the heating furnace and an average load of a roughing mill, and operational parameters of a pickling device including a pickling speed, an acid temperature, an immersion rate, and an acid concentration. [3] The method for predicting surface defects according to [2], wherein a combined parameter of a pickling speed, an acid temperature, an immersion rate, and an acid concentration is used as the operational parameter of the pickling apparatus. [4] A method for producing a steel strip by forming molten steel refined in a secondary refining device into a slab in a continuous casting device, hot rolling the slab in a hot rolling device, annealing it in an annealing device, pickling it in a pickling device, and cold rolling it in a cold rolling device to produce the steel strip, the method comprising: a parameter identification step for identifying operating parameters of a pickling device that are predicted not to produce surface defects in the surface defect prediction step of the surface defect prediction method described in any one of [1] to [3]; and a steel strip production step for producing the steel strip under production conditions that include the operating parameters of the pickling device identified in the parameter identification step. [5] A method for manufacturing a steel strip according to [4], wherein the operational parameters of the pickling apparatus that are predicted to cause no surface defects in the surface defect prediction step are identified so that the pickling speed is the fastest and the acid temperature, acid concentration, and immersion rate are the lowest. [6] A method for manufacturing a steel strip as described in [4], wherein the surface defect prediction step predicts whether or not a surface defect will occur for each of a plurality of steel strips of different steel types manufactured in the same manufacturing equipment, and the parameter identification step identifies the operational parameters of the pickling equipment for which the surface defect prediction step predicted that there would be no surface defects for each of a plurality of steel strips of different steel types, and identifies, from the identified operational parameters of the pickling equipment, the operational parameters of the pickling equipment that are common to the plurality of steel strips. [7] The method for manufacturing a steel strip according to [6], wherein the parameter identifying step identifies the operational parameters of the common pickling equipment so that the pickling speed is the fastest and the acid temperature, immersion rate, and acid concentration are the lowest. [8] A surface defect prediction device for a steel strip produced by forming molten steel refined in a secondary refining device into a slab in a continuous casting device, hot rolling the slab in a hot rolling device, annealing the slab in an annealing device, pickling in a pickling device, and cold rolling the slab in a cold rolling device, the surface defect prediction device comprising: a parameter acquisition unit that acquires one or more parameters of the operation parameters of the secondary refining device, one or more parameters of the operation parameters of the continuous casting device, one or more parameters of the operation parameters of the hot rolling device, and one or more parameters of the operation parameters of the pickling device; and a surface defect prediction unit that inputs input data including the parameters acquired by the parameter acquisition unit into a surface defect prediction model, outputs the presence or absence of surface defects, and predicts the presence or absence of the surface defects. [9] The surface defect prediction device according to [8], wherein the parameter acquisition unit acquires operational parameters of a secondary refining device including a converter reduction time, operational parameters of a continuous casting device including a mold current value, operational parameters of a hot rolling mill including an angular temperature of a slab extracted from a heating furnace and an average load of a roughing mill, and operational parameters of a pickling device including a pickling speed, an acid temperature, an immersion rate, and an acid concentration.

[10] The surface defect prediction device according to [9], wherein the parameter acquisition unit uses a combined parameter of a pickling speed, an acid temperature, an immersion rate, and an acid concentration as an operational parameter of the pickling device.

[11] The surface defect prediction device according to [9] or

[10] , further comprising a parameter specifying unit that specifies operational parameters of the pickling device for which the surface defect prediction unit predicts that no surface defects will occur. [Effects of the Invention]

[0007] By carrying out the surface defect prediction method according to the present invention, it becomes possible to predict the presence or absence of surface defects in a steel strip that will become apparent after cold rolling, based on the operational parameters before the pickling process. Furthermore, by identifying the operational parameters of the pickling apparatus that are predicted by the surface defect prediction method to prevent the occurrence of surface defects, and by producing a steel strip under operational conditions that include the identified operational parameters, it becomes possible to produce a steel strip while suppressing the occurrence of surface defects. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a schematic diagram showing an example of a manufacturing facility for cold-rolled annealed stainless steel strip including a surface defect prediction device capable of implementing the surface defect prediction method according to this embodiment. [Figure 2] FIG. 2 is a schematic diagram showing an example of the configuration of a surface defect prediction device. [Figure 3] FIG. 3 is a schematic cross-sectional view showing an example of a continuous casting apparatus and a perspective view of a mold. [Figure 4] FIG. 4 is a flow chart showing an example of a surface defect prediction method and a steel strip manufacturing method according to this embodiment. [Figure 5] FIG. 5 is a diagram showing an example of combination parameters of a pickling apparatus. DETAILED DESCRIPTION OF THE INVENTION

[0009] The present invention will be described below through embodiments of the present invention. Note that the embodiments shown below are merely examples of devices and methods for realizing the technical concept of the present invention, and the present invention is not limited to these embodiments.

[0010] 1 is a schematic diagram showing an example of a manufacturing facility 100 for a stainless steel cold-rolled and annealed steel strip (hereinafter, the stainless steel cold-rolled and annealed steel strip may be referred to as a "steel strip") including a surface defect prediction device 24 capable of implementing the surface defect prediction method according to this embodiment. The steel strip is manufactured through a refining process, a continuous casting process, a hot rolling process, an annealing process, a pickling process, and a cold rolling process. Therefore, the steel strip manufacturing facility 100 includes a secondary refining device 10, a continuous casting device 12, a hot rolling device 14, an annealing device 16, a pickling device 18, a cold rolling device 20, an annealing and pickling device 21, a process computer 22, and a surface defect prediction device 24.

[0011] Examples of the secondary refining equipment 10 include VOD (Vacuum Oxygen Decarburization) and AOD (Argon Oxygen Decarburization). VOD is a refining equipment in which chromium-containing molten steel (hereinafter, chromium-containing molten steel may be referred to as "molten steel") is decarburized while suppressing chromium oxidation by injecting argon gas from the bottom of a ladle in a vacuum furnace, stirring the molten steel, and then injecting oxygen gas onto the surface. AOD is a refining equipment in which oxygen gas diluted with argon gas is injected from the bottom of the furnace, and decarburized while suppressing chromium oxidation. The continuous casting equipment 12 is an equipment in which molten steel contained in a tundish is continuously cooled in a mold and a secondary cooling equipment to form a slab.

[0012] The hot rolling mill 14 heats slabs in a heating furnace and hot rolls the heated slabs into steel strips of a predetermined width and thickness using roughing mills, intermediate mills, and finishing mills. The annealing mill 16 anneals the hot-rolled steel strips in a predetermined atmosphere. The pickling mill 18 immerses the annealed steel strips in sulfuric acid and a mixture of nitric acid and hydrofluoric acid, in that order, to remove scale and other deposits on the steel strip surface. The cold rolling mill 20 cold-rolls the pickled steel strips to a desired product thickness. The annealing and pickling mill 21 anneals and pickles the cold-rolled steel strips to produce the product (stainless steel cold-rolled annealed steel strip).

[0013] The process computer 22 is, for example, a general-purpose computer such as a workstation or a personal computer. The process computer 22 is connected to each device of the steel strip manufacturing facility 100 by wire or wirelessly and controls the steel strip manufacturing process. The process computer 22 stores manufacturing conditions for operating the secondary refining equipment 10, the continuous casting equipment 12, the hot rolling equipment 14, the annealing equipment 16, the pickling equipment 18, the cold rolling equipment 20, and the annealing and pickling equipment 21. The process computer also collects and stores measurement values ​​measured during operation of each device. In this embodiment, the operational parameters include not only the manufacturing conditions of each device but also the measurement values ​​measured during operation of each device.

[0014] In a steel strip manufactured using such a steel strip manufacturing facility 100, linear surface defects may occur in the steel strip after cold rolling. Linear surface defects are surface defects that originate when a slab containing inclusions is hot rolled, and become apparent when the origin is not removed in the pickling process and the steel strip is cold rolled. The surface defect prediction device 24 according to this embodiment is used to predict whether or not such linear surface defects will occur.

[0015] The surface defect prediction device 24 is a device capable of implementing the surface defect prediction method. The surface defect prediction device 24 acquires the operation parameters of the secondary refining device 10, the operation parameters of the continuous casting device 12, the operation parameters of the hot rolling device 14, and the operation parameters of the pickling device 18 from the process computer 22. The surface defect prediction device 24 inputs these acquired operation parameters as input data into a surface defect prediction model, outputs the presence or absence of surface defects, and predicts the presence or absence of surface defects in the steel strip after cold rolling.

[0016] Furthermore, the surface defect prediction device 24 preferably identifies operational parameters of the pickling device 18 that are predicted to prevent the occurrence of surface defects in the steel strip after cold rolling. The surface defect prediction device 24 outputs the identified operational parameters of the pickling device to the process computer 22 so that the identified operational parameters of the pickling device 18 are set as the production conditions of the pickling device 18.

[0017] Next, we will explain the surface defect prediction device 24 that predicts whether or not surface defects will occur in a steel strip after cold rolling. Figure 2 is a schematic diagram showing an example configuration of the surface defect prediction device 24. The surface defect prediction device 24 is, for example, a general-purpose computer such as a workstation or a personal computer. The surface defect prediction device 24 has a control unit 30, an input unit 32, an output unit 34, and a storage unit 36. The control unit 30 is, for example, a CPU, and functions as a parameter acquisition unit 38, a surface defect prediction unit 40, and a parameter identification unit 42 by executing a program stored in the storage unit 36.

[0018] The input unit 32 is, for example, a keyboard, a touch panel integrated with a display, or the like. The output unit 34 is, for example, an LCD or CRT display, or the like. The storage unit 36 ​​is, for example, an updatable flash memory, a built-in hard disk or a hard disk connected via a data communication terminal, an information recording medium such as a memory card, and a read / write device therefor. The storage unit 36 ​​stores programs and data for realizing each function of the surface defect prediction device 24. The storage unit 36 ​​also stores a database 44 and a surface defect prediction model 46. The database 44 stores 200 or more, and more preferably 1000 or more, data sets, each set consisting of the actual values ​​of one or more operational parameters of the secondary refining equipment 10 for steel strips previously produced using the same steel strip manufacturing equipment 100, the actual values ​​of one or more operational parameters of the continuous casting equipment 12, the actual values ​​of one or more operational parameters of the hot rolling equipment 14, the actual values ​​of one or more operational parameters of the pickling equipment 18, and binary data indicating the presence or absence of surface defects in the steel strip after cold rolling.

[0019] The surface defect prediction model 46 is a trained machine learning model that has been trained using the dataset stored in the database 44 as training data. This trained machine learning model receives as input one or more operational parameters of the secondary refining apparatus 10, one or more operational parameters of the continuous casting apparatus 12, one or more operational parameters of the hot rolling mill 14, and one or more operational parameters of the pickling apparatus 18, and outputs binary data indicating the presence or absence of surface defects in the steel strip after cold rolling. The machine learning model may be any of commonly used neural networks, gradient boosting decision trees (XGBoost), random forests, and support vector regression. The machine learning model and the dataset stored in the database 44 may be acquired from the process computer 22 or may be input to the storage unit 36 ​​by an operator via the input unit 32.

[0020] Next, the processing executed by the parameter acquisition unit 38 and the surface defect prediction unit 40 will be described. The parameter acquisition unit 38 acquires operational parameters of the steel strip manufacturing equipment 100 from the process computer 22 as input data. The processing by this parameter acquisition unit 38 is the parameter acquisition step. The parameter acquisition unit 38 acquires, for example, operational parameters 1 shown below from the process computer 22.

[0021] <Operation parameter 1> Secondary refining equipment 10: Converter reduction time Continuous casting device 12: Mold current value Hot rolling mill 14: angular temperature of slab extracted from heating furnace, average load of roughing mill Pickling equipment 18: pickling speed, acid temperature, acid concentration and immersion rate

[0022] The parameter acquisition unit 38 outputs the operational parameters acquired from the process computer 22 as input data to the surface defect prediction unit 40. When the surface defect prediction unit 40 acquires the input data from the parameter acquisition unit 38, it reads out a surface defect prediction model 46 from the storage unit 36, inputs the input data into the surface defect prediction model 46, and causes it to output binary data indicating the presence or absence of surface defects. As a result, the surface defect prediction unit 40 predicts the presence or absence of surface defects in the steel strip after cold rolling. This processing by the surface defect prediction unit 40 constitutes a surface defect prediction step. The surface defect prediction unit 40 may output the binary data indicating the presence or absence of surface defects to the output unit 34, and display it on the output unit 34. As a result, an operator can confirm the presence or absence of surface defects in the steel strip after cold rolling by visually checking the output unit 34.

[0023] Next, the operational parameters of each device used as input data will be described. As the converter reduction time of the secondary refining equipment 10 increases, the time required for inclusions to float increases, resulting in fewer inclusions that cause linear surface defects. In this way, the converter reduction time affects the presence or absence of surface defects in the steel strip, which is the output of the surface defect prediction model 46. In this way, the prediction accuracy of the surface defect prediction model 46 is improved by including the converter reduction time in the input data for the surface defect prediction model 46. The converter reduction time is the time from when ferrosilica is added during vacuum killed treatment to when the pressure is restored and the steel strip is opened to the atmosphere.

[0024] FIG. 3 is a schematic cross-sectional view showing an example of a continuous casting apparatus 12 and a perspective view of a mold 50. The mold current value is the actual value of the current applied to electromagnetic stirring coils 52 provided on both long side surfaces of the mold 50. The mold current value applied to the electromagnetic stirring coils 52 affects the stirring of the molten steel poured into the mold 50, and this stirring affects the floating and removal of inclusions that cause linear surface defects. In this way, the mold current value affects the presence or absence of surface defects in the steel strip, which is the output of the surface defect prediction model 46. Therefore, the prediction accuracy of the surface defect prediction model 46 is improved by including the mold current value in the input data for the surface defect prediction model 46.

[0025] The angular temperature of the slab removed from the heating furnace is the temperature difference between the center and end portions of the slab in the width direction after removal from the heating furnace. Here, the temperatures of the center and end portions of the slab in the width direction are the surface temperatures of the center and end portions of the slab in the width direction measured by a radiation thermometer or the like. The average load of the roughing mill is the average load of the roughing rolls.

[0026] If the slab temperature is low and the rolling load is large when rolling in the hot rolling mill 14, surface cracks are more likely to occur, and the origin of linear surface defects is more likely to occur. In this way, the corner temperature of the slab and the average load of the roughing mill affect the presence or absence of surface defects in the steel strip, which is the output of the surface defect prediction model 46. For this reason, the prediction accuracy of the surface defect prediction model 46 is improved by including the corner temperature of the slab and the average load of the roughing rolling mill in the input data of the surface defect prediction model 46.

[0027] By pickling the steel strip under appropriate pickling conditions in the pickling device 18, it is possible to remove the origins of linear surface defects that occur during the hot rolling process. Therefore, the pickling speed, acid temperature, acid concentration, and immersion rate, which are operational parameters of the pickling device 18, affect the presence or absence of surface defects that occur in the steel strip after cold rolling. In this way, the pickling speed, acid temperature, acid concentration, and immersion rate of the pickling device 18 affect the presence or absence of surface defects in the steel strip, which is the output of the surface defect prediction model, and therefore, by including these as input data for the surface defect prediction model 46, the prediction accuracy of the surface defect prediction model 46 is improved.

[0028] The parameter acquisition unit 38 is not limited to the above-mentioned operation parameters, and may further acquire, for example, operation parameter 2 shown below, and include the operation parameter 2 in the input data. In addition to operation parameter 1 and operation parameter 2, other operation parameters for the secondary refining equipment 10, the continuous casting equipment 12, the hot rolling equipment 14, the annealing equipment 16, and the pickling equipment 18 may further be acquired, and the operation parameters may be included in the input data.

[0029] <Operation parameter 2> Secondary refining equipment 10: Molten steel temperature at the end of refining Continuous casting equipment 12: Temperature of molten steel in the tundish Hot rolling mill 14: Finishing mill entry temperature Pickling equipment 18: Iron concentration in sulfuric acid, iron concentration in nitric hydrofluoric acid

[0030] Next, the processing of the parameter specifying unit 42 will be described. The parameter specifying unit 42 specifies operational parameters predicted by the surface defect prediction unit 40 as those that will not cause surface defects on the steel strip. The parameter specifying unit 42 outputs the operational parameters to the process computer 22, thereby setting them as the manufacturing conditions of the steel strip manufacturing equipment 100. Below, the processing by the parameter specifying unit 42 will be described using an example in which the parameter specifying unit 42 specifies operational parameters of the pickling device 18.

[0031] 4 is a flow diagram showing an example of the surface defect prediction method and steel strip manufacturing method according to this embodiment. The flow shown in FIG. 4 is started, for example, by receiving an input from an operator via the input unit 32 to start the processing.

[0032] First, the parameter acquisition unit 38 acquires operational parameters of the steel strip to be manufactured from the process computer 22 (step S101). This process is the acquisition step in the surface defect prediction method and steel strip manufacturing method. The parameter acquisition unit 38 outputs the acquired operational parameters to the surface defect prediction unit 40.

[0033] The surface defect prediction unit 40 reads out the surface defect prediction model 46 from the storage unit 36, inputs the operational parameters into the surface defect prediction model, and outputs binary data indicating the presence or absence of a surface defect, thereby predicting the presence or absence of a surface defect in the steel strip (step S102). This processing is the surface defect prediction step in the surface defect prediction method and steel strip manufacturing method. The surface defect prediction unit 40 outputs the prediction result to the parameter identification unit 42.

[0034] The parameter specifying unit 42 determines whether or not surface defects will occur in the steel strip to be produced (step S103). If the prediction result is that no surface defects will occur (step S103: Yes), the parameter specifying unit 42 specifies that the operational parameters of the pickling apparatus 18 used in the input data are operational parameters that will not cause surface defects in the steel strip (step S104). The processing of steps S103 and S104 constitutes the parameter specifying step.

[0035] The parameter specifying unit 42 outputs the specified operational parameters of the pickling apparatus 18 to the process computer 22 (step S105). As a result, the specified operational parameters of the pickling apparatus 18 are reflected in the production conditions of the steel strip production equipment 100. Then, a steel strip is produced under production conditions in which the operational parameters of the pickling apparatus 18 are reflected (step S106), and this flow ends. The processing of this step S106 is the steel strip production step. By producing a steel strip in this manner, it is possible to produce a steel strip under production conditions that include operational parameters that are predicted to prevent the occurrence of surface defects in the steel strip, and therefore it is possible to produce a steel strip while suppressing the occurrence of surface defects.

[0036] On the other hand, if the prediction result indicates that surface defects will occur (step S103: No), the parameter specifying unit 42 changes the operation parameters of the pickling apparatus 18 used in the input data in a direction that makes the pickling conditions stricter (step S107). The direction that makes the pickling conditions stricter means that the acid concentration, acid temperature, and immersion rate all increase, and the pickling speed decreases. The order in which the pickling conditions are changed and to what extent are determined in advance, and the operation parameters of the pickling apparatus 18 are changed according to the determination.

[0037] The surface defect prediction unit 40 and the parameter identification unit 42 repeatedly execute the processes of steps S102 and S103 using the changed operational parameters. As a result, the parameter identification unit 42 can identify operational parameters of the pickling apparatus 18 that are predicted to prevent linear surface defects from occurring on the steel strip. If the surface defect prediction unit 40 predicts that surface defects will occur under all of the predetermined pickling conditions, the parameter identification unit 42 causes the output unit 34 to display a message indicating that the operational parameters of the pickling apparatus 18 cannot be identified or an error message, and the flow ends.

[0038] As described above, by using the surface defect prediction device 24 and the surface defect prediction method according to this embodiment, it is possible to predict the occurrence of linear surface defects that will become apparent on a steel strip after cold rolling. In addition, by identifying operational parameters that are predicted to prevent the occurrence of surface defects using the surface defect prediction method, and producing a steel strip under operational conditions that include these operational parameters, it becomes possible to produce a steel strip while suppressing the occurrence of linear surface defects.

[0039] The present invention is not limited to the above-described embodiment, and various modifications may be made. In the description of steps S101 to S104 in FIG. 4, an example has been shown in which operational parameters of the pickling apparatus 18 are acquired, the presence or absence of surface defects for the operational parameters is predicted, and operational parameters of the pickling apparatus 18 that will not cause surface defects are identified. However, the present invention is not limited to this. In the processing of step S102, the surface defect prediction unit 40 may predict the presence or absence of surface defects for all combinations of operational parameters of the pickling apparatus 18. Then, the parameter identification unit 42 may identify operational parameters that have the fastest pickling speed and the lowest acid temperature, acid concentration, and immersion rate among all combinations of operational parameters of the pickling apparatus 18 that are predicted not to cause surface defects.

[0040] The slower the pickling speed and the higher the acid temperature, acid concentration, and immersion rate, the stronger the pickling conditions, and the greater the amount of steel strip dissolved by the pickling. An increase in the amount of steel strip dissolved not only reduces the yield of the steel strip, but also increases the amount of sludge produced, which is waste material generated during the pickling of the steel strip. Furthermore, a slower pickling speed may become the rate-limiting factor for the production speed of the steel strip manufacturing facility 100, resulting in a decrease in the production volume of the steel strip manufacturing facility 100.

[0041] In contrast, as described above, by identifying the operational parameters that provide the fastest pickling speed and the lowest acid temperature, acid concentration, and immersion rate, and by performing pickling under conditions that include these operational parameters, the amount of steel strip dissolved can be reduced, thereby suppressing a decrease in steel strip yield and an increase in sludge generation. Furthermore, it is also possible to suppress a decrease in the production volume of the steel strip manufacturing facility 100.

[0042] 4, an example has been described in which the presence or absence of surface defects is predicted for one steel strip produced by the steel strip production facility 100 and the operational parameters of the pickling device 18 are identified, but the present invention is not limited to this. The surface defect prediction unit 40 may predict the presence or absence of surface defects for all combinations of operational parameters of the pickling device 18 for multiple steel strips of different steel types that are scheduled to be produced by the steel strip production facility 100 over a predetermined period. Then, the parameter identification unit 42 may identify operational parameters common to the multiple steel strips from among the operational parameters of the pickling device 18 that are predicted to result in no surface defects.

[0043] If different operational parameters of the pickling apparatus 18 are set for a plurality of different steel strips of different steel types that are to be produced in the steel strip production facility 100 during a given period, it may take a long time to change the pickling conditions and may incur costs for changing the pickling conditions. Furthermore, depending on the scale of the change in the operational parameters of the pickling apparatus 18, it may be necessary to prepare a plurality of pickling apparatuses 18.

[0044] In contrast, as described above, by identifying common operational parameters for multiple steel strips of different steel grades and performing pickling under pickling conditions that include these operational parameters, the number of times the pickling conditions need to be changed can be reduced, thereby reducing the time and cost required to change the pickling conditions. It is preferable to identify common operational parameters that provide the fastest pickling speed and the lowest acid temperature, acid concentration, and immersion rate. This reduces the amount of steel strip dissolution, thereby preventing a decrease in steel strip yield and an increase in sludge generation. Furthermore, it is also possible to prevent a decrease in the production volume of the steel strip manufacturing equipment 100.

[0045] In addition, in the description of this embodiment, the operational parameters of the pickling apparatus 18 include the pickling speed, acid temperature, acid concentration, and immersion ratio. However, this is not limiting. Combination parameters of the pickling speed, acid temperature, acid concentration, and immersion ratio may also be used as the operational parameters of the pickling apparatus 18. FIG. 5 shows an example of the combination parameters of the pickling apparatus 18. As shown in FIG. 5, weak pickling conditions, medium pickling conditions, and strong pickling conditions may be set by combining the three operational parameters of acid temperature, acid concentration, and immersion ratio, and combination parameters (P1 to P12) may be used that combine each condition with the pickling speed. The immersion ratio refers to the percentage of the length of the steel strip immersed in the acid solution in the pickling apparatus 18. An immersion ratio of 100% is the pickling condition under which the immersion length of the steel strip in the acid solution in the pickling apparatus 18 is maximized, and an immersion ratio of 50% is the pickling condition under which the immersion length of the steel strip is half of the maximum.

[0046] By using the combined parameters in this way, it is possible to reduce the number of input data and the calculation load on the surface defect prediction device 24. Furthermore, in the process of specifying the operation parameters by the parameter specifying unit 42, the use of the combined parameters also reduces the number of times the operation parameters of the pickling device 18 are changed. This makes it possible to quickly specify the operation parameters of the pickling device 18 that are predicted not to cause surface defects.

[0047] 5, the combination parameters P1 to P12 are set in order from lowest to highest in acid temperature, acid concentration, and immersion rate, and from highest to lowest in pickling rate. When specifying the operational parameters of the pickling apparatus 18, the surface defect prediction unit 40 may use the combination parameters P1 to P12 in order to predict surface defects. This makes it possible to specify the operational parameters of the pickling apparatus that have the fastest pickling rate and the lowest acid temperature, acid concentration, and immersion rate within a range in which surface defects do not occur on the steel strip.

[0048] Furthermore, when identifying the operational parameters of the pickling apparatus 18, the surface defect prediction unit 40 may predict the presence or absence of surface defects for all combinations of parameters P1 to P12. Then, the parameter identification unit 42 may identify the combination of parameters with the smallest number among the combinations of parameters of the pickling apparatus 18 that are predicted to have no surface defects. This makes it possible to identify the operational parameters of the pickling apparatus that have the fastest pickling speed and the lowest acid temperature, acid concentration, and immersion rate within a range in which surface defects do not occur on the steel strip.

[0049] Furthermore, the surface defect prediction unit 40 may use the above-mentioned combination parameters to predict the presence or absence of surface defects for a plurality of steel strips of different steel types that are scheduled to be produced in the steel strip production facility 100 during a predetermined period.The parameter identification unit 42 may then identify a combination parameter that is common to the plurality of steel strips from among the combination parameters of the pickling apparatus 18 that are predicted to produce no surface defects.Also, from among the combination parameters common to the plurality of steel strips, the combination parameter with the smallest number may be identified.

[0050] In the description of this embodiment, an example has been shown in which the steel strip manufacturing equipment 100 has the process computer 22 and the surface defect prediction device 24, but this is not limiting. For example, the process computer 22 may have the function of the surface defect prediction device 24, and these may be configured as a single device. Furthermore, the surface defect prediction device 24 may be directly connected to the secondary refining equipment 10, the continuous casting equipment 12, the hot rolling equipment 14, the annealing equipment 16, and the pickling equipment 18 by wire or wirelessly, and the parameter acquisition unit 38 may directly acquire operation parameters from these equipment.

[0051] In the description of this embodiment, an example has been given in which the control unit 30 of the surface defect prediction device 24 shown in Fig. 3 includes the parameter acquisition unit 38, the surface defect prediction unit 40, and the parameter identification unit 42, but this is not limiting. If the surface defect prediction device 24 predicts the occurrence of linear surface defects that become apparent after cold rolling, the control unit 30 does not need to include the parameter identification unit 42. [Example]

[0052] Next, an example will be described in which a surface defect prediction model was actually generated to predict the occurrence of surface defects in steel strips produced by a steel strip production facility 100. In this example, martensitic stainless steel cold-rolled and annealed steel strips and ferritic stainless steel cold-rolled and annealed steel strips were inspected after cold rolling to confirm the presence or absence of linear surface defects for each product. Here, the martensitic stainless steel cold-rolled and annealed steel strips and ferritic stainless steel cold-rolled and annealed steel strips are martensitic stainless steels and ferritic stainless steels specified in JIS G 4305:2021. For the martensitic stainless steel cold-rolled and annealed steel strips and ferritic stainless steel cold-rolled and annealed steel strips for which the presence or absence of surface defects was confirmed, actual data from the secondary refining equipment 10, continuous casting equipment 12, hot rolling equipment 14, and pickling equipment 18 was collected to prepare a total of 678 sets of training data.

[0053] A surface defect prediction model was generated by training a machine learning model using the training data. The machine learning model used was XGBoost, a machine learning model that combines ensemble learning and decision trees.

[0054] In Example 1, the following operational parameters 1 were used as input data for the surface defect prediction model. The input data for the pickling apparatus 18 was a combination of parameters that combine the pickling rate, acid concentration, acid temperature, and immersion rate shown in Fig. 5. The output data was binary data that indicated the presence or absence of linear surface defects occurring on the martensitic stainless cold-rolled and annealed steel strip and the ferritic stainless cold-rolled and annealed steel strip after cold rolling.

[0055] <Operation parameter 1> Secondary refining equipment 10: Converter reduction time Continuous casting device 12: Mold current value Hot rolling mill 14: Corner temperature of slab extracted from heating furnace, rough rolling average load Pickling equipment 18: pickling speed, acid temperature, acid concentration and immersion rate

[0056] In Example 2, the above operational parameters and the following operational parameter 2 were used as input data for the surface defect prediction model. The output data, like Example 1, is binary data indicating the presence or absence of linear surface defects occurring on the martensitic stainless cold-rolled annealed steel strip and the ferritic stainless cold-rolled annealed steel strip after cold rolling.

[0057] <Operation parameter 2> Secondary refining equipment 10: Molten steel temperature at the end of refining Continuous casting equipment 12: Temperature of molten steel in the tundish Hot rolling mill 14: Finishing mill entry temperature Pickling equipment 18: Iron concentration in sulfuric acid, iron concentration in nitric hydrofluoric acid

[0058] The surface defect prediction model thus generated was used to predict the presence or absence of linear surface defects occurring in a martensitic stainless cold-rolled and annealed steel strip and a ferritic stainless cold-rolled and annealed steel strip after cold rolling, and 200 rolls of martensitic stainless cold-rolled and annealed steel strip and a ferritic stainless cold-rolled and annealed steel strip were manufactured under the same conditions. The presence or absence of linear surface defects was confirmed for the manufactured 200 rolls of martensitic stainless cold-rolled and annealed steel strip and the ferritic stainless cold-rolled and annealed steel strip, and the answers were judged as correct or incorrect as follows:

[0059] <Correct answer> 1. When the occurrence of surface defects is predicted in the surface defect prediction step and surface defects are found in the manufactured martensitic stainless cold-rolled annealed steel strip and ferritic stainless cold-rolled annealed steel strip. 2. When it is predicted that no surface defects will occur in the surface defect prediction step, and the produced martensitic stainless cold-rolled annealed steel strip and ferritic stainless cold-rolled annealed steel strip have no surface defects.

[0060] <Wrong answer> 1. When the occurrence of surface defects is predicted in the surface defect prediction step, and the produced martensitic stainless cold-rolled annealed steel strip and ferritic stainless cold-rolled annealed steel strip have no surface defects. 2. When it is predicted that no surface defects will occur in the surface defect prediction step, but surface defects are found in the manufactured martensitic stainless cold-rolled annealed steel strip and ferritic stainless cold-rolled annealed steel strip.

[0061] The number of correct answers and the number of incorrect answers were confirmed for Invention Examples 1 and 2, and the correct answer rate (number of correct answers / (number of correct answers+number of incorrect answers)) was calculated. The correct answer rates for Invention Examples 1 and 2 are shown in Table 1 below.

[0062] [Table 1]

[0063] As shown in Table 1, it was confirmed that the occurrence of surface defects in martensitic stainless cold-rolled annealed steel strip and ferritic stainless cold-rolled annealed steel strip can be predicted with high accuracy by using a surface defect prediction model (Invention Example 1) using operation parameter 1 as input data. Furthermore, it was confirmed that the occurrence of surface defects in martensitic stainless cold-rolled annealed steel strip and ferritic stainless cold-rolled annealed steel strip can be predicted with even higher accuracy by using a surface defect prediction model (Invention Example 2) using operation parameter 2 as input data in addition to operation parameter 1. [Explanation of symbols]

[0064] 10 Secondary refining equipment 12 Continuous casting equipment 14 Hot rolling equipment 16 Annealing equipment 18 Pickling equipment 20 Cold rolling equipment 21 Annealing and pickling equipment 22 Process Computer 24 Surface defect prediction device 30 Control Unit 32 Input section 34 Output section 36 Storage area 38 Parameter Acquisition Unit 40 Surface Defect Prediction Department 42 Parameter Identification Section 44 databases 46 Surface defect prediction model 50 Mold 52 Electromagnetic stirring coil 100 Steel strip manufacturing equipment

Claims

1. A method for predicting surface defects of a steel strip produced by forming molten steel refined in a secondary refining device into a slab in a continuous casting device, hot rolling the slab in a hot rolling device, annealing the slab in an annealing device, pickling the slab in a pickling device, and cold rolling the slab in a cold rolling device, comprising: a parameter acquisition step of acquiring one or more parameters among the operation parameters of the secondary refining apparatus, one or more parameters among the operation parameters of the continuous casting apparatus, one or more parameters among the operation parameters of the hot rolling apparatus, and one or more parameters among the operation parameters of the pickling apparatus; a surface defect prediction step of inputting input data including the parameters acquired in the parameter acquisition step into a surface defect prediction model, outputting the presence or absence of surface defects in the steel strip after the cold rolling, and predicting the presence or absence of the surface defects; A surface defect prediction method comprising:

2. 2. The surface defect prediction method according to claim 1, wherein the parameter acquisition step acquires operation parameters of a secondary refining device including a converter reduction time, operation parameters of a continuous casting device including a mold current value, operation parameters of a hot rolling device including an angular temperature of the slab extracted from the heating furnace and an average load of a roughing mill, and operation parameters of a pickling device including a pickling speed, an acid temperature, an immersion rate, and an acid concentration.

3. 3. The method for predicting surface defects according to claim 2, wherein a combined parameter of a pickling speed, an acid temperature, an immersion rate, and an acid concentration is used as the operational parameter of the pickling apparatus.

4. A method for producing a steel strip, comprising the steps of: forming molten steel refined in a secondary refining device into a slab in a continuous casting device; hot rolling the slab in a hot rolling device; annealing the slab in an annealing device; pickling the slab in a pickling device; and cold rolling the slab in a cold rolling device to produce a steel strip; a parameter specifying step of specifying an operational parameter of a pickling apparatus for which it is predicted that no surface defects will occur in the surface defect prediction step in the surface defect prediction method according to any one of claims 1 to 3; a steel strip manufacturing step of manufacturing the steel strip under manufacturing conditions including the operational parameters of the pickling apparatus identified in the parameter identifying step; A method for manufacturing a steel strip, comprising:

5. 5. The method for producing a steel strip according to claim 4, wherein the operational parameters of the pickling apparatus that are predicted to cause no surface defects in the surface defect prediction step are specified so that the pickling speed is the fastest and the acid temperature, acid concentration, and immersion rate are the lowest.

6. In the surface defect prediction step, the presence or absence of surface defects is predicted for each of a plurality of steel strips of different steel types produced in the same production facility, In the parameter identification step, operational parameters of the pickling apparatus predicted to have no surface defects in the surface defect prediction step are identified for each of a plurality of steel strips of different steel types; The method for manufacturing a steel strip according to claim 4, wherein, from among the identified operational parameters of the pickling apparatus, an operational parameter of the pickling apparatus that is common to the plurality of steel strips is identified.

7. 7. The method for producing a steel strip according to claim 6, wherein the parameter specifying step specifies operational parameters of the common pickling apparatus such that the pickling speed is the fastest and the acid temperature, immersion rate, and acid concentration are the lowest.

8. A surface defect prediction device for a steel strip produced by forming molten steel refined in a secondary refining device into a slab in a continuous casting device, hot rolling the slab in a hot rolling device, annealing the slab in an annealing device, pickling the slab in a pickling device, and cold rolling the slab in a cold rolling device, a parameter acquiring unit that acquires one or more parameters among the operation parameters of the secondary refining apparatus, one or more parameters among the operation parameters of the continuous casting apparatus, one or more parameters among the operation parameters of the hot rolling apparatus, and one or more parameters among the operation parameters of the pickling apparatus; a surface defect prediction unit that inputs input data including the parameters acquired by the parameter acquisition unit into a surface defect prediction model, outputs the presence or absence of a surface defect, and predicts the presence or absence of the surface defect; A surface defect prediction device comprising:

9. 9. The surface defect prediction device according to claim 8, wherein the parameter acquisition unit acquires operation parameters of a secondary refining device including a converter reduction time, operation parameters of a continuous casting device including a mold current value, operation parameters of a hot rolling device including an angular temperature of the slab extracted from the heating furnace and an average load of a roughing mill, and operation parameters of a pickling device including a pickling speed, an acid temperature, an immersion rate, and an acid concentration.

10. 10. The surface defect prediction device according to claim 9, wherein the parameter acquisition unit uses a combined parameter of a pickling speed, an acid temperature, an immersion rate, and an acid concentration as the operation parameter of the pickling device.

11. 11. The surface defect prediction device according to claim 9, further comprising a parameter specifying unit that specifies an operation parameter of the pickling device for which the surface defect prediction unit has predicted that no surface defects will occur.

Images