Plating adhesion amount prediction method, hot-dip galvanized steel strip manufacturing method, plating adhesion amount prediction model generation method, and plating adhesion amount prediction device

Abstract

The present invention provides a plating adhesion amount prediction method which makes it possible to predict the plating adhesion amount of a hot-dip galvanized steel strip with high accuracy.　Provided is a plating adhesion amount prediction method in which the plating adhesion amount of a hot-dip galvanized steel strip is predicted in hot-dip galvanized steel strip manufacturing equipment having a plating treatment unit that immerses, in a zinc plating bath, a steel strip annealed by an annealing unit, said method comprising: a parameter acquisition step for acquiring at least one operation parameter among operation parameters of the plating treatment unit and at least one operation parameter among operation parameters of the annealing unit; and a plating adhesion amount prediction step for inputting, into a plating adhesion amount prediction model, input data including the operation parameter acquired in the parameter acquisition step, and causing the plating adhesion amount prediction model to output the plating adhesion amount, thereby predicting the plating adhesion amount.

Description

Coating weight prediction method, hot-dip galvanized steel strip manufacturing method, coating weight prediction model generation method, and coating weight prediction device

[0001] The present invention relates to a coating weight prediction method, a manufacturing method of a hot-dip galvanized steel strip, a coating weight prediction model generation method, and a coating weight prediction device.

[0002] A hot-dip galvanized steel sheet manufacturing facility for producing hot-dip galvanized steel sheets generally includes an annealing device for annealing a steel strip, a coating treatment device for coating the steel strip, and a reheating device for reheating the steel strip. The annealing device has a preheating zone, a heating zone, a soaking zone, and a cooling zone. The coating treatment device has a zinc bath in which the steel strip is immersed and a device for adjusting the amount of zinc applied. The reheating device has an alloying zone, a holding zone, and a cooling zone.

[0003] Various methods have been proposed to predict and control the amount of zinc coating applied to steel sheets in such hot-dip galvanized steel sheet manufacturing equipment. Patent Document 1 discloses a method for controlling the amount of molten zinc coating that uses a database in which various operational data from the plating processing unit are input items and the distance between the wiping nozzle and the steel sheet or the wiping nozzle gas pressure is output items. According to Patent Document 1, the zinc coating amount, sheet thickness, sheet width, steel sheet conveying speed, etc. of the steel sheet to be plated are used as input items to determine the distance between the wiping nozzle and the steel sheet or the wiping nozzle gas pressure, thereby enabling the amount of molten zinc coating to be controlled.

[0004] Japanese Patent Application Laid-Open No. 2007-262503

[0005] However, the method disclosed in Patent Document 1 did not provide sufficient accuracy in predicting the coating weight of zinc on a steel sheet. The present invention has been made in view of the above-described conventional technology, and an object of the present invention is to provide a coating weight prediction method and coating weight prediction device that can predict the coating weight of a hot-dip galvanized steel strip with high accuracy. Furthermore, another object of the present invention is to provide a manufacturing method of a hot-dip galvanized steel strip using the coating weight prediction method and a coating weight prediction model.

[0006] The above-mentioned problems are solved by the following means. [1] A coating weight prediction method for predicting a coating weight of a hot-dip galvanized steel strip in a manufacturing facility for hot-dip galvanized steel strip having a coating treatment section in which a steel strip annealed in an annealing section is immersed in a zinc coating bath, the method comprising: a parameter acquisition step of acquiring one or more operational parameters of the coating treatment section and one or more operational parameters of the annealing section; and a coating weight prediction step of inputting input data including the operational parameters acquired in the parameter acquisition step into a coating weight prediction model and outputting the coating weight to predict the coating weight. [2] The coating weight prediction method according to [1], wherein the parameter acquisition step further acquires one or more quality characteristic parameters of quality characteristic parameters of the steel strip, and the coating weight prediction step inputs input data including the operational parameters and the quality characteristic parameters acquired in the parameter acquisition step into a coating weight prediction model and outputs the coating weight to predict the coating weight. [3] The coating weight prediction method according to [1] or [2], wherein the coating weight prediction model is a trained machine learning model obtained by machine learning using a plurality of data sets, each set consisting of a pair of actual values ​​of the input data and an actual value of the coating weight. [4] The coating weight prediction method according to [3], wherein the machine learning model is any one of a neural network, decision tree learning, random forest, and support vector regression. [5] A method for producing a hot-dip galvanized steel strip using hot-dip galvanized steel strip production equipment having a coating processing section that immerses a steel strip annealed in an annealing section in a zinc coating bath, the method comprising: an operational parameter specifying step of specifying operational parameters of the coating processing section that make the coating weight predicted in the coating weight prediction step in the coating weight prediction method according to any one of [1] to [4], within a predetermined target value range; and a hot-dip galvanized steel strip production step of producing the hot-dip galvanized steel strip under production conditions that include the operational parameters of the coating processing section specified in the operational parameter specifying step.[6] A method for generating a coating weight prediction model for predicting a coating weight of a hot-dip galvanized steel strip in a manufacturing facility for hot-dip galvanized steel strip having a coating treatment section in which a steel strip annealed in an annealing section is immersed in a zinc coating bath, the method comprising: a dataset acquisition step for acquiring a plurality of datasets, each set consisting of actual values ​​of input data including one or more operational parameters of the coating treatment section and one or more operational parameters of the annealing section, and the actual values ​​of the coating weight; and a coating weight prediction model generation step for training a machine learning model using the plurality of datasets as training data to generate the coating weight prediction model. [7] The method for generating a coating weight prediction model according to [6], wherein the dataset acquisition step acquires a plurality of datasets, each set consisting of actual values ​​of input data including one or more operational parameters of the coating treatment section, one or more operational parameters of the annealing section, and one or more quality characteristic parameters of the steel strip, and the actual values ​​of the coating weight. [8] The method for generating a coating weight prediction model according to [6] or [7], wherein the machine learning model is any one of a neural network, decision tree learning, random forest, and support vector regression. [9] A coating weight prediction device for predicting a coating weight of a hot-dip galvanized steel strip in a manufacturing facility for hot-dip galvanized steel strip having a coating treatment section that immerses a steel strip annealed in an annealing section in a zinc coating bath, the coating weight prediction device having: a parameter acquisition section that acquires one or more operation parameters of the coating treatment section and one or more operation parameters of the annealing section; and a coating weight prediction section that inputs input data including the operation parameters acquired by the parameter acquisition section into a coating weight prediction model and outputs the coating weight.

[10] The coating weight prediction device described in [9], wherein the parameter acquisition unit further acquires one or more quality characteristic parameters from among the quality characteristic parameters of the steel strip, and the coating weight prediction unit inputs input data including the operational parameters and quality characteristic parameters acquired by the parameter acquisition unit into a coating weight prediction model and outputs the coating weight.

[0007] According to the present invention, the coating weight of a hot-dip galvanized steel strip is predicted using the operational parameters of the annealing zone, thereby making it possible to predict the coating weight with higher accuracy than in the past. Furthermore, by controlling the production conditions of the hot-dip galvanized steel strip using the predicted coating weight, it becomes possible to stably produce a hot-dip galvanized steel strip whose coating weight falls within a target value range.

[0008] FIG. 1 is a schematic diagram showing an example of a hot-dip galvanized steel strip manufacturing facility including a coating weight prediction device capable of implementing the coating weight prediction method according to the present embodiment. FIG. 2 is a cross-sectional schematic diagram of a coating treatment section. FIG. 3 is a graph showing the thermal history of a steel strip in an annealing section and a reheating section of a hot-dip galvanized steel strip manufacturing apparatus for manufacturing hot-dip galvanized steel strip. FIG. 4 is a schematic diagram showing an example of the configuration of a coating weight prediction device. FIG. 5 is a flow chart showing an example of a process for identifying operational parameters of a coating treatment section by an operational parameter identification section. FIG. 6 is a schematic diagram showing a method for generating a coating weight prediction model.

[0009] The present inventors have confirmed that the coating weight of a hot-dip galvanized steel strip is also affected by operational parameters other than those in the galvanizing section. Specifically, they have confirmed that a change in the temperature of the steel strip at the cooling zone exit in the annealing section, which is located upstream of the galvanizing section, changes the viscosity of the zinc coating deposited on the steel strip surface in the galvanizing bath, resulting in a change in the coating weight even when other wiping conditions are the same. Furthermore, they have confirmed that a change in the temperature of the steel strip changes the mechanical properties of the steel strip, resulting in a difference in the shape in the strip width direction during transport by rolls, which in turn changes the distance between the gas wiping device and the steel strip, resulting in a change in the coating weight even when other wiping conditions are the same. Because operational parameters in the annealing section also affect the coating weight of a hot-dip galvanized steel strip, the present inventors have conceived the idea of ​​using operational parameters in the annealing section that affect the coating weight to predict the coating weight of a hot-dip galvanized steel strip. The present invention will be specifically described below through embodiments of the present invention. The following embodiments illustrate preferred examples 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 hot-dip galvanized steel strip manufacturing facility 100 including a coating weight prediction device capable of implementing the coating weight prediction method according to the present embodiment. The hot-dip galvanized steel strip manufacturing facility 100 is a facility that continuously anneals a steel strip 10 that has been reduced to a predetermined thickness through a hot-rolling process, a pickling process, and a cold-rolling process, and then immerses the steel strip in a galvanizing bath for coating treatment, thereby continuously manufacturing a hot-dip galvanized steel strip. Arrow A in FIG. 1 indicates the conveying direction of the steel strip 10.

[0011] The hot-dip galvanized steel strip manufacturing facility 100 includes a hot-dip galvanized steel strip manufacturing device 12, a process computer 54, and a coating weight prediction device 56. The hot-dip galvanized steel strip manufacturing device 12 includes an entry section 13, an annealing section 14, a coating treatment section 15, a reheating section 16, and an exit section 17.

[0012] The entry-side section 13 has a payoff reel 20, a welding machine 22, an electrolytic cleaning device 23, and an entry-side looper 24. An annealing section 14 is provided downstream of the entry-side section 13. The annealing section 14 has a heating zone 28, a soaking zone 30, and a cooling zone 32. A preheating zone 26 may be provided upstream of the heating zone 28.

[0013] In the annealing section 14, the steel strip 10 is heated from near room temperature, maintained at a predetermined temperature, and then cooled to a temperature suitable for galvanizing, thereby annealing the steel strip 10. The heating zone 28 is equipment for heating the steel strip 10 to a temperature predetermined depending on the steel type. Direct flame or radiant combustion burners are used in the heating zone 28, but the heating method is not particularly limited. The soaking zone 30 is equipment for maintaining the steel strip 10 at a predetermined temperature, and is equipment with a heating capacity sufficient to compensate for the heat dissipated from the furnace body, etc.

[0014] The cooling zone 32 is a facility for cooling the steel strip 10 to a temperature suitable for galvanizing. Gas jet cooling is generally used as a cooling means, but the cooling method is not particularly limited as long as it can cool the steel strip to a predetermined cooling temperature. The cooling zone 32 may be divided into multiple zones, such as a first cooling zone 32a and a second cooling zone 32b, and the thermal history of the steel strip during cooling may be controlled by changing the cooling conditions of the cooling means.

[0015] A plating treatment section 15 is provided downstream of the annealing section 14. Figure 2 is a cross-sectional schematic diagram of the plating treatment section 15. The plating treatment section 15 has a snout 34 connected to the outlet of the cooling zone 32, a zinc plating bath 36, and a wiping device 38. The snout 34 is a member with a rectangular cross section that defines the space through which the steel strip 10 passes. A mixed gas containing hydrogen, nitrogen, and water vapor is supplied inside the snout 34, and the atmospheric gas is adjusted until the steel strip 10 is immersed in the zinc plating bath 36.

[0016] The galvanizing bath 36 is a bottomed, rectangular cylindrical container containing molten zinc and including a sink roll 58 and a support roll 60. The sink roll 58 immerses the steel strip 10, which has passed through the snout 34, facing downward into the galvanizing bath 36, and pulls the steel strip 10, with molten zinc adhering to its surface, above the plating bath. The support roll 60 is provided above the sink roll 58 and corrects any warpage or other shape of the steel strip 10, stabilizing its path. The wiping device 38 sprays wiping gas from a pair of nozzles 62, one on each side of the steel strip 10, to scrape off excess molten zinc adhering to the surface of the steel strip 10 and adjust the amount of molten zinc applied.

[0017] Here, we will explain the offset amount and gap amount of the wiping device 38. The offset amount means the amount of deviation in the vertical direction (longitudinal direction of the steel strip) of the position where the wiping gas sprayed from the pair of left and right nozzles 62 collides with the steel strip 10. The gap amount is the width of the opening of the nozzle 62 from which the wiping gas is sprayed.

[0018] Referring again to FIG. 1 , a reheating zone 16 is provided downstream of the plating treatment section 15. The reheating zone 16 includes an alloying zone 40, a holding zone 42, and a final cooling zone 44. An induction heating device is disposed in the alloying zone 40. In the alloying zone 40, the temperature of the steel strip 10, which was lowered in the plating treatment section 15, is raised again to promote the Zn-Fe alloying reaction. In the holding zone 42, the temperature of the steel strip 10 is maintained to promote the alloying reaction. In the final cooling zone 44, the steel strip 10 is cooled to room temperature. Similar to the cooling zone 32, the final cooling zone 44 may be divided into multiple zones, such as a first final cooling zone 44a and a second final cooling zone 44b, to control the thermal history of the steel strip 10 during cooling.

[0019] An outlet section 17 is provided downstream of the reheating section. The outlet section 17 has a temper rolled strip 46, an outlet looper 48, an inspection zone 50, and a tension reel 52. The outlet looper 48 is a device that temporarily stores the steel strip 10 in order to adjust the conveying speed of the steel strip 10 and the processing speed in the outlet section 17. The inspection zone 50 is a device that inspects the dimensional accuracy and surface quality of the steel strip 10. The tension reel 52 is a device that winds the steel strip 10 into a coil. The steel strip 10 that has been wound into a coil by the tension reel 52 and that has passed the quality inspection in the inspection zone 50 is shipped as a product coil. On the other hand, the steel strip 10 that has failed the quality inspection in the inspection zone 50 is sent to a recoil line, where the steel strip 10 undergoes adjustment of its size and weight, sampling for quality confirmation, shape and size inspection, recoiling, etc.

[0020] It is preferable that thermometers for measuring the surface temperature of the steel strip 10 are installed in the annealing section 14, the plating treatment section 15, and the reheating section 16. Among these, it is particularly preferable to install thermometers from the annealing section 14 to just before the plating treatment section 15 to measure the temperature of the steel strip 10. As the thermometer for measuring the temperature of the steel strip 10, it is preferable to use a radiation thermometer that continuously measures the surface temperature at the center of the steel strip 10's width, but a profile radiation thermometer that measures the temperature distribution in the width direction of the steel strip 10 may also be used. In addition to the surface temperature of the steel strip 10, furnace thermometers that measure the atmospheric temperature inside the furnace in each zone of the annealing section 14 and the reheating section 16 may also be installed. The measured surface temperature and atmospheric temperature of the steel strip 10 are output to the process computer 54.

[0021] 3 is a graph showing the temperature of the steel strip 10 from the annealing section 14 to the reheating section 16 of a manufacturing apparatus for producing a hot-dip galvanized steel strip. In FIG. 3, the horizontal axis represents time, and the vertical axis represents the temperature of the steel strip 10. The temperature of the steel strip 10 is the surface temperature of the steel strip 10. The annealing process is carried out in the preheating zone 26, heating zone 28, soaking zone 30, and cooling zone 32, and then the steel strip passes through the coating treatment section 15 and is reheated in the alloying zone 40, holding zone 42, and final cooling zone 44.

[0022] In order to suppress variations in coating weight depending on the position in the longitudinal direction of the steel strip 10, the conveying speed of the steel strip 10 in the annealing section 14 is kept constant. However, when steel strips 10 of different thicknesses, widths, steel types, etc. are welded, the line speed may change before and after the weld. For this reason, the thermal history of the steel strip 10 may vary depending on the position at which the temperature of the steel strip 10 is measured.

[0023] The process computer 54 is, for example, a general-purpose computer such as a workstation or a personal computer. The process computer 54 is connected to the hot-dip galvanized steel strip manufacturing apparatus 12 by wire or wirelessly and controls the hot-dip galvanized steel strip manufacturing process. The process computer 54 also acquires quality characteristic parameters of the steel strip 10 from a higher-level computer. The quality characteristic parameters include, for example, the thickness, width, yield stress, tensile strength, and surface roughness of the steel strip 10. The surface roughness of the steel strip 10 may be measured non-contact and in real time using a laser irradiation type surface roughness meter provided in front of the zinc plating bath 36.

[0024] The process computer 54 acquires and stores operational parameters for the entry section 13, the annealing section 14, the plating treatment section 15, the reheating section 16, and the exit section 17 in the production of a hot-dip galvanized steel strip. The process computer 54 acquires, as operational parameters for the annealing section 14, for example, the line speed and tension of the steel strip 10 in the soaking zone 30, and the surface temperature of the steel strip 10 when immersed in the galvanizing bath 36. It is preferable that a radiation thermometer be provided in front of the galvanizing bath 36 and that the surface temperature of the steel strip 10 be measured non-contact and in real time using the radiation thermometer.

[0025] Furthermore, the process computer 54 acquires operating parameters of the plating processing unit 15, such as the plating temperature of the zinc plating bath 36, the amount of pressure the support roll 60 places on the steel strip 10, the height of the nozzle 62 from the bath surface, the wiping gas spray angle from the nozzle 62, the gap amount at the opening of the nozzle 62, the distance between the nozzle 62 and the steel strip 10, the offset amount of the nozzle 62, and the wiping gas spray pressure from the nozzle 62.

[0026] The coating weight prediction device 56 acquires the operation parameters of the galvanizing section 15, the operation parameters of the annealing section 14, and the quality characteristic parameters of the steel strip 10 from the process computer 54. The coating weight prediction device 56 inputs input data including the acquired parameters into a coating weight prediction model, which outputs the coating weight. As a result, the coating weight prediction device 56 predicts the coating weight of the hot-dip galvanized steel strip manufactured by the hot-dip galvanized steel strip manufacturing apparatus 12. The coating weight prediction device 56 identifies the operation parameters of the coating section that will bring the predicted coating weight within a predetermined target value range. The coating weight prediction device 56 outputs the identified operation parameters of the coating section to the process computer 54 to be set as the operation parameters of the coating section.

[0027] Next, a coating weight prediction device 56 that predicts the coating weight of a hot-dip galvanized steel strip will be described. FIG. 4 is a schematic diagram showing an example configuration of the coating weight prediction device 56. The coating weight prediction device 56 is, for example, a general-purpose computer such as a workstation or a personal computer. The coating weight prediction device 56 has a control unit 70, an input unit 72, an output unit 74, and a storage unit 76. The control unit 70 is, for example, a CPU, and functions as a parameter acquisition unit 78, a coating weight prediction unit 80, an operation parameter identification unit 82, and a coating weight prediction model generation unit 84 by executing a program stored in the storage unit 76.

[0028] The input unit 72 may be, for example, a keyboard, a touch panel integrated with a display, or the like. The output unit 74 may be, for example, an LCD or CRT display. The storage unit 76 may be, for example, an updatable flash memory, a built-in hard disk or a memory card connected via a data communication terminal, or an information recording medium and a read / write device for the same. The storage unit 76 stores programs and data for implementing each function of the coating weight prediction device 56. The storage unit 76 also stores a database 86 and a coating weight prediction model 88. The database 86 stores 200 or more, preferably 1,000 or more, data sets consisting of actual values ​​of hot-dip galvanized steel strips previously manufactured by the same hot-dip galvanized steel strip manufacturing apparatus 12. The above data sets are sets of actual values ​​of input data, including operational parameters of the coating treatment unit 15, operational parameters of the annealing unit 14, and quality characteristic parameters of the steel strip 10, and actual values ​​of the coating weight.

[0029] The coating weight prediction model 88 is a trained machine learning model that has been trained using a data set stored in the database 86 as training data. The coating weight prediction model 88 according to this embodiment is a trained machine learning model that receives input data including operation parameters of the plating processing unit 15, operation parameters of the annealing unit 14, and quality characteristic parameters of the steel strip 10 as input, and outputs the coating weight. In this embodiment, an example will be described in which the operation parameters of the plating processing unit 15, operation parameters of the annealing unit 14, and quality characteristic parameters of the steel strip 10 are used as input data, but this is not limiting. The input data may include parameters other than those described above, and may not include the quality characteristic parameters of the steel strip 10.

[0030] Next, a description will be given of the processing executed by the parameter acquisition unit 78 and the coating weight prediction unit 80. The parameter acquisition unit 78 acquires, as input data, operation parameters of the plating processing unit 15 and operation parameters of the annealing unit 14 from the process computer 54. This processing by the parameter acquisition unit 78 corresponds to a parameter acquisition step.

[0031] The parameter acquisition unit 78 acquires, as input data, one or more of the following operation parameters of the plating processing unit 15 from the process computer 54:

[0032] <Operation parameters of the plating treatment section 15> Plating temperature of the zinc plating bath 36 Push-in amount of the support roll 60 to the steel strip 10 Height of the nozzle 62 from the bath surface Wiping gas ejection angle from the nozzle 62 Gap amount of the opening of the nozzle 62 Distance between the nozzle 62 and the steel strip 10 Offset amount of the nozzle 62 Injection pressure of the wiping gas from the nozzle 62

[0033] The smaller the amount of pressure the support roll 60 exerts on the steel strip 10, the lower the tension on the steel strip 10, making the steel strip 10 more likely to bend. As the steel strip 10 bends, the distance from the nozzle 62 changes on the front and back sides when wiping gas is sprayed. This change in distance causes a change in the wiping gas pressure that scrapes off the plating, and therefore a change in the plating deposition weight.

[0034] The higher the height of the nozzle 62 from the bath surface, the greater the amount of coating that falls under its own weight before wiping. Therefore, even if other wiping conditions are the same, the higher the height of the nozzle 62 from the bath surface, the less coating adhesion there is. The greater the distance between the nozzle 62 and the steel strip 10, the lower the pressure of the gas that scrapes off the coating on the steel strip surface. Therefore, even if other wiping conditions are the same, the greater the distance between the nozzle 62 and the steel strip 10, the less coating adhesion there is.

[0035] As the wiping gas jet pressure from the nozzle 62 decreases, the pressure of the gas scraping the plating off the steel strip surface decreases, resulting in an increase in the coating weight even if other wiping conditions are the same. As the coating temperature in the zinc plating bath 36 increases, the viscosity of the zinc on the steel sheet surface at the time the wiping gas is sprayed onto the steel strip 10 decreases, resulting in a decrease in the coating weight even if other wiping conditions are the same. As the wiping gas jet angle from the nozzle 62 (the smaller of the angle between the wiping gas jet direction and the steel strip) decreases, the pressure of the gas scraping the plating off the steel strip surface decreases, resulting in an increase in the coating weight even if other wiping conditions are the same.

[0036] When the gap amount at the opening of the nozzle 62 increases, the spray pressure of the wiping gas from the nozzle 62 decreases, which reduces the pressure of the gas scraping off the coating on the steel strip surface, resulting in an increase in the coating weight even if other wiping conditions remain the same. When the offset amount of the nozzle 62 decreases, the wiping gas on both sides collides with the edge of the steel strip 10, causing edge splashing and reducing the coating weight. In this way, these operational parameters affect the coating weight. Therefore, by acquiring one or more of the operational parameters of the plating processing unit 15 and including those parameters in the input data for the coating weight prediction model, the accuracy of the coating weight prediction can be improved.

[0037] Furthermore, the parameter acquisition unit 78 acquires, as input data, one or more of the following operational parameters of the annealing unit 14 from the process computer 54:

[0038] <Operation parameters of the annealing section 14> Dew point Temperature Time in furnace Line speed of the steel strip 10 Line tension of the steel strip 10 Surface temperature of the steel strip 10 when immersed in the galvanizing bath 36

[0039] When the dew point of the annealing portion 14 increases, oxides are formed on the surface of the steel strip 10, reducing the coating adhesion. As a result, the coating weight decreases even if other wiping conditions remain the same. When the temperature of the annealing portion 14 increases or the furnace time increases, the surface hardness of the steel strip 10 decreases, reducing the amount of plastic strain in the surface. The plastic strain activates the surface of the steel strip 10 and increases the growth rate of the alloy layer during zinc plating. As a result, when the temperature of the annealing portion 14 increases or the furnace time increases, the coating weight decreases even if other wiping conditions remain the same.

[0040] When the line speed of the steel strip 10 in the annealing section 14 increases, the amount of coating lifted from the steel strip surface per unit time increases. Therefore, even if other wiping conditions are the same, the coating weight increases. When the tension of the steel strip 10 decreases, the steel strip 10 becomes more likely to bend. When the steel strip 10 bends, the distance from the nozzle 62 changes between the front and back surfaces when the wiping gas is sprayed, which causes fluctuations in the wiping gas pressure that scrapes off the coating, and thus fluctuations in the coating weight.

[0041] If the surface temperature of the steel strip 10 increases when immersed in the zinc plating bath 36, the viscosity of zinc on the steel strip surface at the time when the wiping gas is sprayed onto the steel strip 10 decreases. As a result, even if other wiping conditions are the same, the coating weight decreases. In this way, these operational parameters affect the coating weight. Therefore, by acquiring one or more of the operational parameters of the annealed portion 14 and including those parameters in the input data of the coating weight prediction model, the accuracy of the coating weight prediction can be improved.

[0042] Furthermore, it is preferable that the parameter acquisition unit 78 acquires, as input data, one or more quality characteristic parameters of the steel strip 10 from the process computer 54, for example, the quality characteristic parameters listed below.

[0043] <Quality characteristic parameters of steel strip 10> Plate thickness Plate width Yield stress Tensile strength Surface roughness

[0044] The thinner the thickness and wider the width of the steel strip 10, the lower the rigidity of the steel strip 10, making it more likely to bend during threading. When the steel strip 10 bends, the distance from the nozzle 62 changes on the front and back sides when wiping gas is sprayed. This change in distance causes fluctuations in the wiping gas pressure that scrapes off the coating, resulting in fluctuations in the coating weight. The lower the yield stress and tensile strength of the steel strip 10, the lower the rigidity even for the same steel strip dimensions, making it more likely to bend during threading. When the steel strip 10 bends in this way, the coating weight fluctuates for the same reasons as when the thickness is thinner and the width is wider.

[0045] The greater the surface roughness of the steel strip 10, the greater the coefficient of friction between the steel strip 10 and the coating, and therefore the coating weight increases even if other wiping conditions are the same. In this way, these quality characteristic parameters affect the coating weight. Therefore, by acquiring one or more of the quality characteristic parameters of the steel strip 10 and including those parameters as input data for the coating weight prediction model, the prediction accuracy of the coating weight is improved. The parameter acquisition unit 78 outputs the acquired input data to the coating weight prediction unit 80.

[0046] When the coating weight prediction unit 80 acquires input data from the parameter acquisition unit 78, it reads out a coating weight prediction model 88 from the storage unit 76, inputs the input data into the coating weight prediction model 88, and outputs the coating weight to predict the coating weight. This processing by the coating weight prediction unit 80 corresponds to a coating weight prediction step. The coating weight prediction unit 80 may output the output coating weight to the output unit 74, and may display the coating weight of the hot-dip galvanized steel strip on the output unit 74. This allows the operator to visually check the output unit 74 and confirm the predicted value of the coating weight of the hot-dip galvanized steel strip.

[0047] Next, a description will be given of the processing of the operation parameter specifying unit 82 and the coating weight prediction model generating unit 84. The operation parameter specifying unit 82 specifies operation parameters of the coating treatment unit that will cause the coating weight of the hot-dip galvanized steel strip predicted by the coating weight predicting unit 80 to fall within a range of preset target values. The operation parameter specifying unit 82 outputs the specified operation parameters to the process computer 54, thereby setting them as operation parameters of the coating treatment unit.

[0048] Fig. 5 is a flow diagram showing an example of a process for specifying operational parameters of a plating processing unit by the operational parameter specifying unit 82. The flow shown in Fig. 5 is started, for example, by receiving an input from an operator to start the process.

[0049] First, the parameter acquisition unit 78 acquires, from the process computer 54, operational parameters of the plating processing unit 15 of the hot-dip galvanized steel strip to be manufactured (step S101). The parameter acquisition unit 78 acquires operational parameters of the annealing unit 14 from the process computer 54 (step S102). Furthermore, the parameter acquisition unit 78 acquires quality characteristic parameters of the steel strip 10 from the process computer 54 (step S103). The parameter acquisition unit 78 outputs the acquired operational parameters of the plating processing unit 15, operational parameters of the annealing unit 14, and quality characteristic parameters to the coating weight prediction unit 80.

[0050] The coating weight prediction unit 80 reads out the coating weight prediction model 88 from the storage unit 76. The coating weight prediction unit 80 inputs the acquired operation parameters of the plating processing unit 15, operation parameters of the annealing unit 14, and quality characteristic parameters into the coating weight prediction model 88, thereby outputting the coating weight. In this manner, the coating weight prediction unit 80 predicts the coating weight of the hot-dip galvanized steel sheet to be manufactured (step S104). The coating weight prediction unit 80 outputs the output predicted value of the coating weight to the operation parameter identification unit 82.

[0051] The range of target coating weights for the hot-dip galvanized steel sheet may be predetermined and stored in the storage unit 76, or may be input by the operator via the input unit 72. After acquiring the predicted coating weight and the target coating weight range for the hot-dip galvanized steel sheet, the operational parameter specifying unit 82 determines whether the predicted coating weight falls within the target coating weight range for the hot-dip galvanized steel sheet (step S105). The operational parameter specifying unit 82 compares the predicted coating weight with the target coating weight range. If the operational parameter specifying unit 82 determines that the predicted coating weight falls outside the target coating weight range for the hot-dip galvanized steel sheet (step S105: No), it changes the operational parameters of the coating unit 15 in accordance with predetermined conditions (step S106) and returns the process to step S104. The coating weight prediction unit 80 inputs the changed operational parameters of the coating unit 15, the previously acquired operational parameters of the annealing unit 14, and the quality characteristic parameters of the steel strip 10 into the coating weight prediction model 88, and predicts the coating weight again. The processes of steps S104 to S106 are repeated until it is determined in step S105 that the predicted coating weight is within the target value range.

[0052] On the other hand, if it is determined in step S105 that the predicted coating weight is within the target range (step S105: Yes), the operational parameter specifying unit 82 specifies that the operational parameters of the plating processing unit 15 used to predict the coating weight are the operational parameters of the plating processing unit 15 that can bring the coating weight within the target range (step S107). The processes of steps S105 to S107 constitute operational parameter specifying steps in the method for producing a hot-dip galvanized steel strip.

[0053] When the operational parameters of the plating processing unit 15 are identified by the operational parameter identification unit 82, the flow of the operational parameter identification process shown in Fig. 5 is terminated. The operational parameter identification unit 82 outputs the identified operational parameters of the plating processing unit 15 to the process computer 54. The process computer 54 sets the identified operational parameters of the plating processing unit 15 as the operational parameters of the plating processing unit 15, and produces a hot-dip galvanized steel strip under production conditions including the operational parameters. This is the hot-dip galvanized steel strip production step in the hot-dip galvanized steel strip production method.

[0054] In this way, by specifying the operation parameters of the plating processing section 15, it is possible to specify the operation parameters of the plating processing section 15 that can control the coating weight of the hot-dip galvanized steel strip to be within the range of the target value. Then, by manufacturing the hot-dip galvanized steel strip under manufacturing conditions that include the operation parameters, it becomes possible to stably manufacture the hot-dip galvanized steel strip whose coating weight falls within the range of the target value.

[0055] Next, a method for generating a coating weight prediction model used to predict the coating weight will be described. FIG. 6 is a schematic diagram showing the method for generating a coating weight prediction model. The parameter acquisition unit 78 acquires, from the process computer 54, actual values ​​of the operational parameters of the coating treatment unit 15 of previously produced hot-dip galvanized steel strips, actual values ​​of the operational parameters of the annealing unit 14, actual values ​​of the quality characteristic parameters of the steel strip 10, and actual values ​​of the coating weight of the hot-dip galvanized steel strips, and stores these as a set of data in the database 86 of the storage unit 76. The number of data sets stored in the database 86 is preferably at least 200 or more, and more preferably 1000 or more. The processing by the parameter acquisition unit 78 corresponds to the data set acquisition step in the method for generating a coating weight prediction model.

[0056] The coating weight prediction model generation unit 84 reads out a pre-stored machine learning model from the storage unit 76, and performs machine learning on the machine learning model using a plurality of data sets stored in the database 86 as training data to generate a trained machine learning model. This trained machine learning model becomes a coating weight prediction model 88. The processing by the coating weight prediction model generation unit 84 corresponds to the coating weight prediction model generation step in the coating weight prediction model generation method. The machine learning model used in the coating weight prediction device 56 according to this embodiment may be any of commonly used neural networks, decision tree learning, random forests, and support vector regression.

[0057] It is preferable to update the coating weight prediction model to a new shape prediction model by re-training it through machine learning, for example, once a month or once a year. The parameter acquisition unit 78 acquires actual values ​​each time a hot-dip galvanized steel strip is manufactured and stores them in the database 86. The database 86 stores data sets of newly manufactured hot-dip galvanized steel strips, so the number of data sets stored increases. The greater the number of data sets, the higher the accuracy with which the coating weight can be predicted. Therefore, by periodically performing machine learning using the data sets stored in the database 86 and updating the shape prediction model, the coating weight of a hot-dip galvanized steel strip can be predicted with even higher accuracy.

[0058] As described above, the coating weight prediction model used in the coating weight prediction device 56 according to this embodiment includes operational parameters of the annealing section 14 as input data. As described above, the operational parameters of the annealing section 14 affect the coating weight of the hot-dip galvanized steel strip. Therefore, by including the operational parameters of the annealing section 14 as input data for the coating weight prediction model, the coating weight prediction model becomes a coating weight prediction model that takes the annealing section 14 into account, thereby increasing the accuracy of predicting the coating weight of the hot-dip galvanized steel strip. In this way, if the coating weight of the hot-dip galvanized steel strip can be predicted with high accuracy, it is possible to prevent the production of hot-dip galvanized steel strips whose coating weights exceed the target value range, thereby preventing increases in the manufacturing costs of the hot-dip galvanized steel strips and decreases in productivity.

[0059] Although the hot-dip galvanized steel strip manufacturing facility 100 shown in FIG. 1 includes the process computer 54 and the coating weight prediction device 56, this is not limiting. For example, the coating weight prediction device 56 may have the functions of the process computer 54, and these may be configured as a single device. In the coating weight prediction device 56 shown in FIG. 4, the control unit 70 includes the operation parameter specifying unit 82 and the coating weight prediction model generating unit 84, but this is not limiting. If the coating weight prediction device 56 predicts the coating weight of the hot-dip galvanized steel strip, the control unit 70 does not need to include the operation parameter specifying unit 82. Furthermore, if the coating weight prediction model 88 is generated externally and stored in the storage unit 76 via the parameter acquiring unit 78, the coating weight prediction device 56 does not need to include the coating weight prediction model generating unit 84.

[0060] Next, an example will be described. 200 coils of galvanized steel strip were produced using the hot-dip galvanized steel sheet production equipment 100 shown in Figure 1. From this production record, actual values ​​of the operation parameters of the plating processing unit 15, the operation parameters of the annealing unit, the quality characteristic parameters of the steel strip 10, and the coating weight were obtained. As the operation parameters of the plating processing unit 15, actual values ​​of the pushing amount of the support roll 60 into the steel strip 10, the height of the nozzle 62 from the bath surface, the distance between the nozzle 62 and the steel strip 10, and the spray pressure of the wiping gas from the nozzle 62 were obtained.

[0061] As operational parameters of the annealing section 14, actual values ​​were obtained for the line speed of the steel strip 10 in the annealing section 14 and the surface temperature of the steel strip 10 when immersed in the zinc plating bath 36. A radiation thermometer was installed upstream of the zinc plating bath 36, and the value measured using this radiation thermometer was used as the surface temperature of the steel strip 10.

[0062] As quality characteristic parameters of the steel strip 10, actual values ​​of the thickness, width, yield stress, and surface roughness of the steel strip 10 were obtained. The surface roughness of the steel strip 10 was measured using a surface roughness meter that uses laser irradiation and was installed upstream of the zinc plating bath 36. The actual coating weight of the hot-dip galvanized steel strip was measured using a fluorescent X-ray film thickness meter.

[0063] For each actual value obtained in this way, different actual values ​​of each parameter used as training data were changed to generate different coating weight prediction models 1 to 3. The conditions for generating coating weight prediction models 1 to 3 are as follows:

[0064] Coating weight prediction model 1: A coating weight prediction model generated by machine learning using 10,000 data sets as training data, each set consisting of a set of actual values ​​of operational parameters of the plating processing unit 15 and an actual value of the coating weight. Coating weight prediction model 2: A coating weight prediction model generated by machine learning using 10,000 data sets as training data, each set consisting of actual values ​​of operational parameters of the plating processing unit 15 and the annealing unit 14 and an actual value of the coating weight. Coating weight prediction model 3: A coating weight prediction model generated by machine learning using 10,000 data sets as training data, each set consisting of actual values ​​of operational parameters of the plating processing unit 15 and the annealing unit 14 and an actual value of the quality characteristic parameters of the steel strip 10 and an actual value of the coating weight.

[0065] Using the coating weight prediction models 1 to 3 thus generated, the predicted coating weight of the hot-dip galvanized steel strip was set to the target value of 48 g / m according to the flow chart shown in FIG. ２ 50g / m or more ２ The operation parameters of the plating section were specified to fall within the following ranges: Then, 100 coils of hot-dip galvanized steel strip were produced under production conditions including the specified operation parameters of the plating section.

[0066] In coating weight prediction model 1 (comparative example), the coating weight of the hot-dip galvanized steel strip fell within the target range for 85% of the 100 coils. In contrast, in coating weight prediction model 2 (inventive example), which includes operational parameters of the annealing section as input data, the coating weight of the hot-dip galvanized steel strip fell within the target range for 90% of the 100 coils. Furthermore, in coating weight prediction model 3 (inventive example), which includes operational parameters of the annealing section and quality characteristic parameters of the steel strip 10 as input data, the coating weight of the hot-dip galvanized steel strip fell within the target range for 95% of the 100 coils.

[0067] Thus, more coils of hot-dip galvanized steel strips had coating weights within the target range when coating weight prediction models 2 and 3 were used than when coating weight prediction model 1 was used. From these results, it was confirmed that the use of coating weight prediction models 2 and 3 makes it possible to predict the coating weight with higher accuracy than when coating weight prediction model 1 is used, and that it is possible to identify operating parameters of the coating processing section that will bring the coating weight within the target range with a high probability.

[0068] 10 Steel strip 12 Manufacturing apparatus for hot-dip galvanized steel strip 13 Entry section 14 Annealing section 15 Coating processing section 16 Reheating section 17 Exit section 20 Payoff reel 22 Welding machine 23 Electrolytic cleaning device 24 Entry looper 26 Preparing zone 28 Heating zone 30 Soaking zone 32 Cooling zone 32a First cooling zone 32b Second cooling zone 34 Snout 36 Galvanizing bath 38 Wiping device 40 Alloying zone 42 Holding zone 44 Final cooling zone 44a First final cooling zone 44b Second final cooling zone 46 Temper rolling zone 48 Exit looper 50 Inspection zone 52 Tension reel 54 Process computer 56 Coating weight prediction device 58 Sink roll 60 Support roll 62 Nozzle 70 Control unit 72 Input unit 74 Output unit 76 Storage unit 78 Parameter acquisition unit 80 Coating weight prediction unit 82 Operation parameter identification unit 84 Coating weight prediction model generation unit 86 Database 88 Coating weight prediction model 100 Manufacturing facility for hot-dip galvanized steel strip

Claims

1. A coating weight prediction method for predicting a coating weight of a hot-dip galvanized steel strip in a manufacturing facility for hot-dip galvanized steel strip having a coating processing section in which a steel strip annealed in an annealing section is immersed in a zinc coating bath, the method comprising: a parameter acquisition step for acquiring one or more operational parameters of the coating processing section and one or more operational parameters of the annealing section; and a coating weight prediction step for inputting input data including the operational parameters acquired in the parameter acquisition step into a coating weight prediction model, outputting the coating weight, and predicting the coating weight.

2. A coating weight prediction method as described in claim 1, wherein in the parameter acquisition step, one or more quality characteristic parameters of the steel strip are further acquired, and in the coating weight prediction step, input data including the operational parameters and quality characteristic parameters acquired in the parameter acquisition step are input into a coating weight prediction model, and the coating weight is output to predict the coating weight.

3. The plating adhesion weight prediction method according to claim 1 or 2, wherein the plating adhesion weight prediction model is a trained machine learning model that is trained using multiple data sets, each set being a pair of the actual value of the input data and the actual value of the plating adhesion weight, as training data.

4. The plating coating weight prediction method according to claim 3, wherein the machine learning model is one of a neural network, a decision tree learning, a random forest, and a support vector regression.

5. A method for manufacturing a hot-dip galvanized steel strip using a hot-dip galvanized steel strip manufacturing facility having a coating processing section in which a steel strip annealed in an annealing section is immersed in a zinc coating bath, the method comprising: an operational parameter identification step of identifying operational parameters of the coating processing section such that the coating adhesion amount predicted in the coating adhesion weight prediction step in the coating adhesion weight prediction method described in claim 1 or claim 2 falls within a predetermined target value range; and a hot-dip galvanized steel strip manufacturing step of manufacturing the hot-dip galvanized steel strip under manufacturing conditions including the operational parameters of the plating processing section identified in the operational parameter identification step.

6. A method for producing a hot-dip galvanized steel strip using a hot-dip galvanized steel strip manufacturing facility having a coating processing section in which a steel strip annealed in an annealing section is immersed in a zinc coating bath, the method comprising: an operational parameter identification step of identifying operational parameters of the coating processing section that will result in a coating adhesion weight predicted in the coating adhesion weight prediction step in the method for predicting coating weight described in claim 3 falling within a range of a preset target value; and a hot-dip galvanized steel strip manufacturing step of manufacturing the hot-dip galvanized steel strip under manufacturing conditions including the operational parameters of the plating processing section identified in the operational parameter identification step.

7. A method for producing a hot-dip galvanized steel strip using a hot-dip galvanized steel strip manufacturing facility having a coating processing section in which a steel strip annealed in an annealing section is immersed in a zinc coating bath, the method comprising: an operational parameter identification step of identifying operational parameters of the coating processing section that will bring the coating adhesion weight predicted in the coating adhesion weight prediction step in the coating adhesion weight prediction method described in claim 4 within a range of a preset target value; and a hot-dip galvanized steel strip manufacturing step of manufacturing the hot-dip galvanized steel strip under manufacturing conditions including the operational parameters of the coating processing section identified in the operational parameter identification step.

8. A method for generating a coating weight prediction model for predicting a coating weight of a hot-dip galvanized steel strip in a manufacturing facility for hot-dip galvanized steel strip having a coating processing section in which a steel strip annealed in an annealing section is immersed in a zinc coating bath, the method comprising: a dataset acquisition step for acquiring a plurality of datasets each consisting of a set of actual values ​​of input data including one or more operational parameters of the coating processing section and one or more operational parameters of the annealing section, and an actual value of the coating weight; and a coating weight prediction model generation step for training a machine learning model using the plurality of datasets as training data to generate the coating weight prediction model.

9. A method for generating a coating weight prediction model as described in claim 8, wherein in the dataset acquisition step, a plurality of datasets are acquired, each of which is a set of actual values ​​of input data including one or more operational parameters of the plating processing section, one or more operational parameters of the annealing section, and one or more quality characteristic parameters of the quality characteristic parameters of the steel strip, and an actual value of the coating weight.

10. A method for generating a plating adhesion weight prediction model as described in claim 8 or claim 9, wherein the machine learning model is any one of a neural network, a decision tree learning, a random forest, and a support vector regression.

11. A coating weight prediction device for predicting a coating weight of a hot-dip galvanized steel strip in a manufacturing facility for hot-dip galvanized steel strip having a coating processing section in which a steel strip annealed in an annealing section is immersed in a zinc coating bath, the coating weight prediction device having: a parameter acquisition section for acquiring one or more operational parameters of the coating processing section and one or more operational parameters of the annealing section; and a coating weight prediction section for inputting input data including the operational parameters acquired by the parameter acquisition section into a coating weight prediction model and outputting the coating weight.

12. A coating weight prediction device as described in claim 11, wherein the parameter acquisition unit further acquires one or more quality characteristic parameters of the steel strip, and the coating weight prediction unit inputs input data including the operational parameters and quality characteristic parameters acquired by the parameter acquisition unit into a coating weight prediction model and outputs the coating weight.

Images