JP2019210535A - Coating weight control method, manufacturing method of hot dip steel strip, coating weight control device, and coating weight control program - Google Patents

Abstract

To provide a coating weight control method capable of improving preset control accuracy for calculating an operation amount for realizing a target coating weight of a steel strip and simplifying a logic of coating weight control, a manufacturing method of a hot dip steel strip, a coating weight control device, and a coating weight control program.SOLUTION: A coating weight control method controls a coating weight of a steel strip 32 by spraying a gas from a gas wiping nozzle 36 to the steel strip 32 taken continuously out from a hot dip bath. A coating weight is categorized into multiple groups on the basis of an operation condition by using a statistical method in which the operation condition is an explanatory variable and the coating weight corresponding thereto is an objective function. The coating weight of the steel strip 32 is controlled by estimating the coating weight in actual operation time on the basis of information of the categorization.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a plating adhesion amount control method, a hot-dip plated steel strip manufacturing method, a plating adhesion amount control device, and a plating adhesion amount control program.

  In the continuous plating line of the steel process, the amount of plating adhering to the steel strip (hereinafter referred to as “plating adhesion amount”) is sprayed from the tip of the gas wiping nozzle to the steel strip (nozzle gap) and from the gas wiping nozzle. It is generally determined by the gas pressure (nozzle pressure).

  In many plants that automatically control the amount of plating deposited, the nozzle gap and nozzle pressure that achieve the target value of the amount of plated coating sent from the host computer (hereinafter referred to as the “target plating amount of deposition”) are predicted. Automatic control is performed by calculating using a model. In this automatic control of the plating adhesion amount, preset control is performed when the target plating adhesion amount is changed (when the welding point of the steel strip is passed). This is solved by feedback control using the meter detection value.

  Here, the plating adhesion meter described above measures the amount of plating actually attached to the steel strip, but it is attached to the back (downstream side) from several tens to several hundred meters from the gas wiping nozzle. Therefore, it takes, for example, 1 to 2 minutes before the plating adhesion amount corresponding to the current nozzle gap and nozzle pressure can be measured. Therefore, when it is going to control plating adhesion amount using feedback control, the precision of plating adhesion amount will fall in the wide range in the longitudinal direction of a steel strip. Accordingly, improvement in control of preset control has been conventionally demanded. For example, in Patent Document 1, high-precision preset control is achieved by appropriately selecting a plurality of preset portions and a compensation method for a plating adhesion amount prediction model. A method for performing the above has been proposed.

JP 2008-50680 A

  However, in the method proposed in Patent Document 1, the difference between the plating adhesion amount prediction model and the actual plating plant behavior is stored separately in a universal model error and a temporarily generated model error. Therefore, there is a problem that the logic for controlling the amount of plating is complicated.

  The present invention has been made in view of the above problems, and improves the accuracy of preset control for calculating the operation amount for realizing the target plating adhesion amount of the steel strip, and simplifies the logic of the plating adhesion amount control. It is to provide a plating adhesion amount control method, a method for manufacturing a hot dipped steel strip, a plating adhesion amount control device, and a plating adhesion amount control program.

  In order to solve the above-described problems and achieve the object, the plating adhesion amount control method according to the present invention is characterized in that the steel is obtained by blowing gas from a nozzle to a steel strip continuously taken out from a hot dipping bath. In the plating adhesion amount control method for controlling the plating adhesion amount of the belt, a plurality of the plating adhesion amounts are determined depending on the operation conditions by using a statistical method with the operation condition as an explanatory variable and the corresponding plating adhesion amount as an objective variable. It is characterized in that the plating adhesion amount of the steel strip is controlled by estimating the plating adhesion amount during actual operation based on the classification information.

  Further, the plating adhesion amount control method according to the present invention uses a plurality of data sets including past operation conditions and corresponding plating adhesion amounts in the above invention, and a plurality of plating adhesion amounts are determined according to the past operation conditions. A learning step for generating a decision tree model by classifying into the following groups, an adhesion amount estimation step for estimating a current plating adhesion amount based on the decision tree model and actual operating conditions, and an estimated plating adhesion amount And a deposit amount control step for controlling the plating amount of the steel strip to be a target value by feeding back.

  Further, in the plating adhesion amount control method according to the present invention, in the above invention, the plurality of data sets are, as the past operation conditions, a steel type that is character data corresponding to the composition of the steel strip, and the steel strip It includes a plate width, a plate thickness of the steel strip, a distance from the tip of the nozzle to the steel strip, and a pressure of gas blown from the nozzle.

  In the plating adhesion amount control method according to the present invention, in the above invention, the plurality of data sets may further include an indentation amount with respect to the steel strip by a roll installed in the hot dipping bath as the past operation condition. It is characterized by including.

  Further, the plating adhesion amount control method according to the present invention is characterized in that, in the above-mentioned invention, in the learning step, the steel type in the past operation conditions is set as a first branch condition of the decision tree model. And

  Further, the plating adhesion amount control method according to the present invention, in the above invention, in the learning step, as a second branch condition of the decision tree model, the steel strip plate width of the past operation conditions and the Either one of the plate thickness of the steel strip is set, and as the third branching condition of the decision tree model, either the plate width of the steel strip or the plate thickness of the steel strip of the past operation conditions Is set.

  Moreover, the plating adhesion amount control method according to the present invention, in the above invention, when a new steel type that has not been operated in the past is added to the operating conditions, the learning step includes the first branch of the decision tree model in the learning step. A partial tree having a new steel type as a vertex is added, and the decision tree model is updated.

  Further, the plating adhesion amount control method according to the present invention is characterized in that, in the above-mentioned invention, in the learning step, the plating adhesion amount is classified into a plurality of groups according to the past operation conditions using an F test. To do.

  In order to solve the above-described problems and achieve the object, the plating adhesion amount control method according to the present invention is characterized in that the steel is obtained by blowing gas from a nozzle to a steel strip continuously taken out from a hot dipping bath. In the plating adhesion amount control method that controls the plating adhesion amount of the belt, plating is performed during actual operation using a decision tree model with operation conditions including character data as explanatory variables and the corresponding plating adhesion amount as an objective variable. The plating adhesion amount of the steel strip is controlled by estimating the adhesion amount.

  In the plating adhesion amount control method according to the present invention, in the above invention, the decision tree model is such that a first branch condition is a steel type, and a second branch condition is a steel strip width and a steel strip thickness. The third branch condition is either the steel strip width or the steel strip thickness, and the fourth and subsequent branch conditions are determined based on the F test. And

  In order to solve the above-described problems and achieve the object, the manufacturing method of the hot-dip galvanized steel strip according to the present invention manufactures the hot-dip galvanized steel strip whose plating adhesion amount is controlled by the above-mentioned plating adhesion amount control method. It is characterized by.

  In order to solve the above-described problems and achieve the object, the plating adhesion amount control device according to the present invention is characterized in that the steel is obtained by blowing a gas from a nozzle to a steel strip continuously taken out from a hot dipping bath. In the plating adhesion amount control device for controlling the plating adhesion amount of the belt, a plurality of data sets including past operation conditions and corresponding plating adhesion amounts are used, and the plating adhesion amount is divided into a plurality of groups according to the past operation conditions. By classifying them, the learning unit that generates the decision tree model, the adhesion amount estimation unit that estimates the current plating adhesion amount based on the decision tree model and the actual operation conditions, and the estimated plating adhesion amount are fed back And a preset control unit that controls the plating adhesion amount of the steel strip to be a target value.

  Moreover, the plating adhesion amount control apparatus according to the present invention is the above invention, wherein the plurality of data sets are, as the past operation conditions, a steel type that is character data corresponding to the composition of the steel strip, and It includes a plate width, a plate thickness of the steel strip, a distance from the tip of the nozzle to the steel strip, and a pressure of gas blown from the nozzle.

  In order to solve the above-mentioned problems and achieve the object, the plating adhesion amount control program according to the present invention is characterized in that the steel is obtained by blowing a gas from a nozzle to a steel strip continuously taken out from a hot dipping bath. In the plating adhesion amount control program for controlling the plating adhesion amount of the belt, the computer uses a plurality of data sets including past operation conditions and corresponding plating adhesion amounts, and the plating adhesion amount is determined according to the past operation conditions. By classifying into a plurality of groups, a learning means for generating a decision tree model, an adhesion amount estimation means for estimating a current plating adhesion amount based on the decision tree model and actual operating conditions, an estimated plating adhesion amount By feeding back, it functions as a preset control means for controlling the amount of plating on the steel strip to a target value. And wherein the Rukoto.

  According to the present invention, using a statistical method, the plating adhesion amount is classified into a plurality of groups according to the operation conditions, and the plating adhesion amount at the time of actual operation is estimated based on the classification information. It is possible to improve the accuracy of the preset control for calculating the operation amount for realizing the plating adhesion amount and simplify the logic of the plating adhesion amount control.

FIG. 1 is a schematic diagram showing a configuration of a plating adhesion amount control device and a plating plant according to an embodiment of the present invention. FIG. 2 is a flowchart showing the flow of the plating adhesion amount control method by the plating adhesion amount control apparatus according to the embodiment of the present invention. FIG. 3 is a diagram for explaining a decision tree model generation method in the learning step of the plating adhesion amount control method according to the embodiment of the present invention. FIG. 4 is a learning step of the plating adhesion amount control method according to the embodiment of the present invention. It is a figure which shows an example of a case. FIG. 5 is a learning step of the plating adhesion amount control method according to the embodiment of the present invention, and updates the decision tree model by adding a steel type as character data corresponding to the composition of the steel strip as a new operation condition. It is a figure which shows another example of a case. FIG. 6 is a diagram illustrating an example of the data structure of the decision tree model generated in the learning step of the plating adhesion amount control method according to the embodiment of the present invention.

  Hereinafter, a plating adhesion amount control method, a hot-dip plated steel strip manufacturing method, a plating adhesion amount control device, and a plating adhesion amount control program according to an embodiment of the present invention will be described with reference to the drawings. Note that the present invention is not limited to the embodiments described below.

[Plating plant]
First, the structure of the plating plant 3 used by this embodiment is demonstrated. As shown in FIG. 1, the plating plant 3 includes a pot 31, a sink roll 33, a collect roll 34, a guide roll 35, a gas wiping nozzle (hereinafter referred to as “nozzle”) 36, and a plating adhesion meter 37. And.

  The pot 31 stores molten metal. In the pot 31, a steel strip (steel plate) 32 is continuously conveyed and immersed in molten metal. Further, a sink roll 33 and a collect roll 34 are installed in the pot 31.

  The collect roll 34 is for suppressing the curvature of the steel strip 32 at the position where the nozzle 36 is installed. One roll (lower roll in FIG. 1) and the other roll (upper roll in FIG. 1) of the pair of collect rolls 34 are arranged with their height positions shifted from each other, and one fixed roll On the other hand, by pushing the other roll, the steel strip 32 is sandwiched and its curvature is suppressed. In addition, in this embodiment, the pushing amount of the other roll with respect to one roll at that time is defined as “roll pushing amount C”.

  The steel strip 32 is supported by a sink roll 33 and a collect roll 34 inside the pot 31, and is supported by a guide roll 35 outside the pot 31. The steel strip 32 is once immersed in the molten metal in the pot 31, and then high pressure gas is blown from the nozzle 36 to the lifting edge. In this way, by removing unnecessary molten metal adhering to the steel strip 32, the plating adhesion amount is controlled to a desired value (target plating adhesion amount). The amount of plating on the steel strip 32 includes the distance from the tip of the nozzle 36 to the surface of the steel strip 32 (hereinafter referred to as “nozzle gap D”) and the pressure of the gas blown from the nozzle 36 (hereinafter referred to as “nozzle pressure P”). It is generally determined by.

  Specifically, the steel strip 32 is configured by connecting a plurality of steel strips by welding points PW. This welding point PW corresponds to a switching portion of the target plating adhesion amount in the plating adhesion amount control described later.

  The plating adhesion meter 37 measures the actual plating adhesion amount of the steel strip 32. The plating adhesion meter 37 is attached at a position separated from the position of the nozzle 36 by several tens to several hundreds m behind (downstream side). Moreover, since the plating adhesion meter 37 measures the steel strip 32 cyclically (periodically) in the width direction, until the plating adhesion amount corresponding to the current nozzle gap D and nozzle pressure P is measured, for example, Requires 1-2 minutes.

[Plating adhesion amount control device]
Hereinafter, the structure of the plating adhesion amount control apparatus 1 according to the present embodiment will be described. The plating adhesion amount control device 1 controls the above-described plating plant 3 to adhere a plating having a desired thickness to the steel strip 32. That is, the plating adhesion amount control device 1 controls the plating adhesion amount of the steel strip 32 by blowing gas from the nozzle 36 to the steel strip 32 continuously taken out from the hot dipping bath (pot 31). .

  In terms of hardware, the plating adhesion amount control device 1 includes an arithmetic processing device such as a CPU (Central Processing Unit) and a storage device such as a hard disk device. The arithmetic processing unit described above functions as a preset control unit (preset control unit) 11, an adhesion amount estimation unit (adhesion amount estimation unit) 12, and a learning unit (learning unit) 13 by executing a computer program. The storage device described above functions as the learning result storage unit 14. Although not shown in FIG. 1, the plating adhesion amount control device 1 includes an input unit for capturing various information from the host computer 2 via a network, and an output unit (for example, a display) for outputting various processing information. Apparatus, printing apparatus, etc.).

  The preset control unit 11 uses the above-mentioned input unit to send the target plating adhesion amount of each steel strip 32 connected at the welding point PW and the manufacturing information such as the plate width and thickness of each steel strip 32 to the upper computer. 2 and the control command value at the operation end is calculated so that a new target plating adhesion amount is realized at the transition of the steel strip 32. Here, the above-described “control command value” is specifically a control command value related to the nozzle gap D and the nozzle pressure P of the nozzle 36.

  The preset control unit 11 controls the plating adhesion amount of the steel strip 32 to be the target plating adhesion amount by feeding back the current plating adhesion amount estimated by the adhesion amount estimation unit 12 described later. Specifically, the preset control unit 11 calculates a control command value related to the nozzle gap D and the nozzle pressure P of the nozzle 36 based on the estimated current plating adhesion amount and the target plating adhesion amount.

  The preset control unit 11 is input from the host computer 2 with the steel type, plate width, plate thickness, and target plating adhesion amount of the steel strip 32, and from the adhesion amount estimation unit 12, an estimated value of the plating adhesion amount (hereinafter referred to as “estimation”). "Plating adhesion amount") is input. In response to this, the preset control unit 11 calculates the control command value and outputs the control command value to the nozzle 36.

  The adhesion amount estimation unit 12 takes in the actual value of the nozzle gap D of the nozzle 36 and the actual value of the nozzle pressure P of the gas sprayed from the nozzle 36 against the steel strip 32 in the plating plant 3, and adheres to the plating. Based on the quantity prediction model, the current plating adhesion amount is estimated. Here, the aforementioned “plating adhesion amount prediction model” specifically indicates a decision tree model described later. The adhesion amount estimation unit 12 estimates the current plating adhesion amount based on the decision tree model and actual operation conditions.

  The adhesion amount estimation unit 12 receives the nozzle gap D, the nozzle pressure P, the roll push-in amount C, and the plating adhesion amount measured by the plating adhesion meter 37 as actual values from the plating plant 3. In response to this, the adhesion amount estimation unit 12 estimates the plating adhesion amount described above, and outputs the estimated plating adhesion amount to the preset control unit 11.

  The learning unit 13 receives information from the host computer 2 and the plating plant 3 via the adhesion amount estimation unit 12, and mechanically learns the correlation between the plating adhesion amount and the operation conditions using the information. . Specifically, the learning unit 13 uses a plurality of data sets including past operation conditions and corresponding plating adhesion amounts, and determines the plating adhesion amounts by classifying them into a plurality of groups according to the past operation conditions. Generate a tree model.

  Further, when a new operation condition is added, the learning unit 13 updates the decision tree model by adding the new operation condition as a branch condition (explanatory variable) to the generated decision tree model. . Here, as an example of the above-mentioned “operation conditions”, the steel type that is character data corresponding to the composition of the steel strip 32, the plate width of the steel strip 32, the plate thickness of the steel strip 32, the nozzle gap D, Nozzle pressure P and roll pushing amount C are mentioned.

  The learning unit 13 receives from the adhesion amount estimation unit 12 the steel type of the steel strip 32, the plate width, the plate thickness, the nozzle gap D, the nozzle pressure P, the roll pressing amount C, and the plating adhesion amount measured by the plating adhesion meter 37. Is entered. In response to this, the learning unit 13 generates the above-described decision tree model and outputs this decision tree model to the learning result storage unit 14. A method for generating a decision tree model by the learning unit 13 and the structure of the decision tree model will be described later.

[Plating adhesion amount control method]
Hereinafter, the plating adhesion amount control method by the plating adhesion amount control apparatus 1 according to the present embodiment will be described with reference to FIGS. The plating adhesion amount control method described below can be widely applied to, for example, plating adhesion amount control in a steel process line.

  As will be described later, the plating adhesion amount control method according to the present embodiment uses a statistical method in which the operation condition is an explanatory variable and the corresponding plating adhesion amount is a target variable, and the plating adhesion amount is plural depending on the operation condition. The plating adhesion amount of the steel strip 32 is controlled by estimating the plating adhesion amount during actual operation based on the classification information.

  Further, as described later, the plating adhesion amount control method according to the present embodiment uses a decision tree model in which operation conditions including character data such as steel types are used as explanatory variables and the corresponding plating adhesion amount is used as an objective variable. Thus, the plating adhesion amount of the steel strip 32 is controlled by estimating the plating adhesion amount during actual operation.

  In the plating adhesion amount control method according to the present embodiment, specifically, as shown in FIG. 2, a learning step S1, an adhesion amount estimation step S2, and an adhesion amount control step S3 are performed in this order.

<Learning step>
In learning step S1, the learning unit 13 uses a plurality of data sets including past operation conditions and corresponding plating adhesion amounts, and classifies the plating adhesion amounts into a plurality of groups according to the past operation conditions. Generate a decision tree model.

  In the learning step S1, the amount of plating adhesion is classified into a plurality of groups according to past operating conditions using the F test. That is, in learning step S1, as shown in FIG. 3, the F test is performed for each explanatory variable (operation condition) on the entire population (parent data) of the plating adhesion amount measured by the plating adhesion meter 37. Then, the branching condition is determined by searching for a dividing point of data that can minimize the p value. The child data classified from the parent data of the plating adhesion amount is represented by an average value of the plating adhesion amount obtained by averaging the classified plating adhesion amounts as shown in FIG.

  In the learning step S1, the F test is performed for all explanatory variables (operation conditions), and among the explanatory variables, the explanatory variable that can make the parent data the most homogeneous (equally distributed) is the first (first ) Branch condition. In the learning step S1, such a branch condition determination process is performed on each child data, and an explanatory variable that can make the branched child data most uniform is determined as the next branch condition. A decision tree model as shown in FIG. 3 is generated. In the learning step S1, the branching of the child data is terminated when it is determined by the F-test that the child data is not equally distributed (not homogeneous).

  In the decision tree model, as the operation condition constituting the branch condition, as described above, in addition to numerical information such as the plate width, plate thickness, nozzle gap D, nozzle pressure P, and roll push-in amount C of the steel strip 32, the steel type Can be used without converting branch conditions other than numerical information into numerical information.

(Data structure of decision tree model)
In the learning step S1, for example, a decision tree model as shown in FIGS. 4 and 5 can be generated. In the decision tree model shown in FIG. 4, the steel type (steel grade A or B) is set as the first branch condition, and the plate width (whether the plate width is larger or smaller than Xmm) is set as the second branch condition. Has been. Here, in the case where the decision tree model shown on the “existing” side of the figure has already been generated, for example, when adding a new steel type C having no operation record in the past to the operation condition, As shown on the “new” side of the figure, the decision tree model is updated by adding a subtree having the new steel type C as the apex to the first branch of the decision tree model. As shown in the figure, when the steel type is set as the first branching condition, there is an advantage that it is possible to distinguish between existing and new installations.

  On the other hand, in the decision tree model shown in FIG. 5, the plate width (whether the plate width is larger or smaller than Xmm) is set as the first branch condition, and the steel type (steel type A or steel type B) is set as the second branch condition. Is set. Here, in the case where the decision tree model shown on the “existing” side of the figure has already been generated, for example, when adding a new steel type C having no operation record in the past to the operation condition, As shown on the “new” side of the figure, the decision tree model is updated by adding a subtree having the new steel grade C as the vertex to the second branch of the decision tree model. As shown in the figure, when the steel type is set as the second branch condition, it is necessary to remodel each branch of the first branch (plate width).

  Here, in the learning step S1, as shown in FIG. 6, the branch condition is manually determined without using the F test for the upper branch close to the parent data, and the F test is performed for the lower branch far from the parent data. It is preferable to generate a decision tree model by automatically determining a branch condition using.

  In the decision tree model shown in FIG. 6, the steel type is set as the first branch condition, the plate width of the steel strip 32 is set as the second branch condition, and the plate thickness of the steel strip 32 is set as the third branch condition. Is set, roll push amount C and nozzle pressure (table) are set as the fourth branch condition, nozzle pressure (back) and nozzle gap (table) are set as the fifth branch condition, and the sixth branch condition is Nozzle pressure (table) is set. Among these, the first to third branch conditions are determined manually without using the F test, and the fourth and subsequent (fourth to sixth) branch conditions are based on the F test. It is determined automatically.

  The above-mentioned “nozzle pressure (table)” indicates the nozzle pressure P of the nozzle (left nozzle in FIG. 1) arranged on the front side of the steel strip 32 of the pair of nozzles 36. The “nozzle pressure (back)” indicates the nozzle pressure P of the nozzle (the right nozzle in the figure) arranged on the back side of the steel strip 32 of the pair of nozzles 36. “Table)” indicates the nozzle gap D between the steel strip 32 and the nozzle (left nozzle in the figure) arranged on the front side of the steel strip 32 of the pair of nozzles 36.

  In this way, in the decision tree model generated in the learning step S1, by setting the steel type above the branching condition, a new steel type is added as the operation condition as shown in FIG. 4 described above. Even in such a case, the decision tree model can be extended without modifying the existing decision tree model. Therefore, the reusability of the existing decision tree model is improved. Hereinafter, the description after the adhesion amount estimation step S2 will be continued.

<Adhesion amount estimation step>
In the adhesion amount estimation step S2, the adhesion amount estimation unit 12 estimates the current plating adhesion amount based on the decision tree model and the actual operation conditions.

<Adhesion amount control step>
In the adhesion amount control step S3, the preset control unit 11 feeds back the estimated current plating adhesion amount so as to control the plating adhesion amount of the steel strip 32 to be the target plating adhesion amount.

  As described above, in the plating adhesion amount control method according to the present embodiment, the actual data of the steel type, plate width, plate thickness, nozzle gap D, nozzle pressure P, and roll push-in amount C of the steel strip 32, and the nozzle 36 In addition, the correlation with the actual amount data of the plating adhesion measured by the latter plating adhesion meter 37 is periodically and mechanically learned (learning step S1). Then, the current plating adhesion amount is estimated based on the decision tree model that is the learning result (adhesion amount estimation step S2), and the actual plating adhesion amount is controlled based on the estimated plating adhesion amount (adhesion amount control step S3). ).

[Plating adhesion amount control program]
The plating adhesion amount control program according to this embodiment is for controlling the plating adhesion amount of the steel strip 32 by blowing gas from the nozzle 36 to the steel strip 32 continuously taken out from the hot dipping bath. The computer is caused to function as a learning unit (learning unit 13), an adhesion amount estimation unit (attachment amount estimation unit 12), and a preset control unit (preset control unit 11).

  The plating adhesion amount control program according to the present embodiment may be distributed by being stored in a computer-readable recording medium such as a hard disk, a flexible disk, or a CD-ROM, or distributed through a network. Also good.

[Method of manufacturing hot-dip galvanized steel strip]
The manufacturing method of the hot dipped steel strip according to this embodiment manufactures a hot dipped steel strip whose plating adhesion amount is controlled by the plating adhesion amount control method according to this embodiment described above.

  According to the plating adhesion amount control method, the hot-dip plated steel strip manufacturing method, the plating adhesion amount control device 1 and the plating adhesion amount control program according to the present embodiment as described above, the plating adhesion amount is obtained using a statistical method. Are classified into a plurality of groups according to the operating conditions, and the amount of operation for realizing the target plating adhesion amount of the steel strip 32 by estimating the plating adhesion amount during actual operation based on the classification information (decision tree model) It is possible to improve the accuracy of the preset control for calculating the plating and simplify the logic of the plating adhesion amount control.

  Moreover, in the plating adhesion amount prediction model used in the prior art (Patent Document 1), the plating adhesion amount is predicted only by three items of nozzle pressure, nozzle gap, and plate speed. Therefore, when other operating conditions (such as steel strip grade, plate width, plate thickness, etc.) fluctuate, it is necessary to correct the model itself. There is a need to.

  Further, in the prior art, when other operation conditions are newly added and the difference between the plating adhesion amount prediction model and the actual result becomes large, it is necessary to newly create a plating adhesion amount prediction model. In this case, in the prior art, for example, it is assumed that the model accuracy is improved by incorporating a large number of additional elements such as composition data of steel types into the plating adhesion amount prediction model and performing multiple regression. However, in this case, it is necessary to recreate a plating adhesion amount prediction model every time additional elements increase. In addition, since only numerical data can be used for multiple regression, the steel grade (steel grade symbol) itself, which is character data corresponding to the composition of the steel strip, cannot be reflected in the plating adhesion prediction model, and the steel grade adheres to the plating. In order to reflect in the quantity prediction model, it is necessary to create a plating adhesion quantity prediction model for each steel type.

  In addition, since the composition of the steel strip changes greatly when the steel type changes, when trying to incorporate the composition data into a mathematical model such as a regression equation, for example, C, Si, Mn, S, P, Al, Cr, Ni, It is necessary to incorporate all the concentrations of many additive alloy elements such as Cu, W, Mo, V, Co, Ti, Nb, B, Sb, and In, and the calculation amount of numerical analysis becomes enormous and the number of man-hours increases.

  On the other hand, in the plating adhesion amount control method, the hot-dip plated steel strip manufacturing method, the plating adhesion amount control device 1 and the plating adhesion amount control program according to the present embodiment, plating is performed on the basis of the result of regular learning by the learning unit 13. Since the adhesion amount prediction model (decision tree model) is updated, the stability determination logic or the like as in the conventional technique is not required, and the logic is simplified. Furthermore, in this embodiment, since it is possible to use the steel type symbol itself as a branching condition instead of the steel type composition, the decision tree model generated by the learning unit 13 is handled as one model including the steel type symbol. Is possible.

  Further, in the plating adhesion amount control method, the hot-dip plated steel strip manufacturing method, the plating adhesion amount control device 1 and the plating adhesion amount control program according to the present embodiment, character data such as a steel type is obtained by using a decision tree model. Can be taken as an explanatory variable, and the steel type that has an influence on the plating adhesion amount can also be used for creating a model, and the estimation accuracy of the plating adhesion amount that is the objective variable can be improved.

  Moreover, in the plating adhesion amount control method, the hot dipped steel strip manufacturing method, the plating adhesion amount control device 1 and the plating adhesion amount control program according to the present embodiment, the steel type is determined as shown in FIGS. 4 and 6 described above. By setting the first branch of the tree model, even if a steel grade is newly added, the decision tree model can be updated by adding a new subtree without destroying the existing decision tree model. Therefore, the existing decision tree model can be reused. Further, by using a decision tree model as a plating adhesion amount prediction model, visibility is improved, and for example, program verification and lateral development to other facilities are facilitated.

  Hereinafter, the present invention will be described more specifically with reference to examples. In this example, as shown in Table 1 below, when learning was performed including the steel type that is character data as the operation condition (branch condition) (learning step S1), learning was performed without including the steel type. For the case, the estimated accuracy of the amount of plating was confirmed.

  As shown in Table 1, when a decision tree model is generated by learning including the steel type in the operation condition, and the plating adhesion amount is estimated based on the decision tree model, the estimated value of the plating adhesion amount is “488. 89 ”, the actual difference indicating the difference between the estimated value and the actual value of the plating adhesion amount was“ 0.0644730 ”, and the difference ratio was“ 1.0188% ”. On the other hand, when a decision tree model is generated by learning without including the steel type in the operation condition, and the plating adhesion amount is estimated based on the decision tree model, the estimated value of the plating adhesion amount is “488.88” The difference was “0.0462449” and the difference ratio was “1.1345%”. Therefore, it was confirmed that the estimation accuracy of the amount of plating adhesion was improved by learning including the steel type in the operation conditions.

  As mentioned above, although the plating adhesion amount control method, the manufacturing method of the hot dip plated steel strip, the plating adhesion amount control device, and the plating adhesion amount control program according to the present invention have been specifically described by the embodiments and examples for carrying out the invention. The gist of the present invention is not limited to these descriptions, but should be broadly interpreted based on the description of the scope of claims. Needless to say, various changes and modifications based on these descriptions are also included in the spirit of the present invention.

  For example, in the plating adhesion amount control device 1 shown in FIG. 1 described above, the decision tree model generated by the learning unit 13 is temporarily stored in the learning result storage unit 14, but the learning unit 13 generates the generated decision tree. The model may be output as it is to the adhesion amount estimation unit 12 without being stored in the learning result storage unit 14.

  In the decision tree model shown in FIG. 6, the steel type is set as the first branch condition, the plate width of the steel strip 32 is set as the second branch condition, and the steel strip 32 is set as the third branch condition. However, the plate thickness of the steel strip 32 may be set as the second branch condition, and the plate width of the steel strip 32 may be set as the third branch condition.

  Moreover, in the above-mentioned plating adhesion amount control method, as a character data that can be included in the operation condition (branching condition), a steel type is given as an example. Other character data such as “GA” representing an alloyed hot-dip galvanized steel sheet may be used as the operation condition.

  In the plating adhesion amount control method described above, the case where a new steel type that has not been operated in the past is added to the operating conditions as an example of the timing at which the decision tree model is updated has been exemplified (see FIGS. 4 and 5). ) In addition to the case where the operation conditions such as the steel type and the new steel plate size are added, the decision tree model may be updated, for example, at a set predetermined time interval or operation stop timing such as periodic inspection.

1 Plating adhesion amount control device 11 Preset control unit (preset control means)
12 Adhesion amount estimation unit (adhesion amount estimation means)
13 Learning part (learning means)
14 Learning result storage unit 2 Host computer 3 Plating plant 31 Pot 32 Steel strip (steel plate)
33 Sink roll 34 Collect roll 35 Guide roll 36 Nozzle (gas wiping nozzle)
37 Plating adhesion meter PW welding point

Claims (14)

In the plating adhesion amount control method for controlling the plating adhesion amount of the steel strip by blowing gas from the nozzle to the steel strip continuously taken out from the hot dipping bath,
Using statistical methods with the operating conditions as explanatory variables and the corresponding plating deposits as objective variables, the plating deposits are classified into a plurality of groups according to the operating conditions, and during actual operation based on the classification information The plating adhesion amount control method characterized by controlling the plating adhesion amount of the said steel strip by estimating the plating adhesion amount in. A learning step for generating a decision tree model by classifying the plating adhesion amount into a plurality of groups according to the past operation conditions, using a plurality of data sets including past operation conditions and corresponding plating adhesion amounts; ,
Based on the decision tree model and actual operating conditions, an adhesion amount estimation step for estimating a current plating adhesion amount;
An adhesion amount control step for controlling the plating adhesion amount of the steel strip to be a target value by feeding back the estimated plating adhesion amount;
The plating adhesion amount control method according to claim 1, comprising:   The plurality of data sets are, as the past operation conditions, a steel type that is character data corresponding to the composition of the steel strip, a plate width of the steel strip, a plate thickness of the steel strip, and a tip of the nozzle The plating adhesion amount control method according to claim 2, comprising a distance to the steel strip and a pressure of a gas blown from the nozzle.   The plating amount control method according to claim 3, wherein the plurality of data sets further include, as the past operation conditions, an indentation amount with respect to the steel strip by a roll installed in the hot dipping bath. .   5. The plating adhesion amount control method according to claim 3, wherein, in the learning step, the steel type of the past operation condition is set as a first branch condition of the decision tree model. .   In the learning step, as the second branch condition of the decision tree model, one of the plate width of the steel strip and the plate thickness of the steel strip among the past operation conditions is set, and the decision tree model 6. The plating adhesion amount according to claim 5, wherein, as the third branching condition, any one of the plate width of the steel strip and the plate thickness of the steel strip among the past operation conditions is set. Control method.   In the case of adding a new steel type having no operation results in the past to the operation condition, in the learning step, a subtree having the new steel type as a vertex is added to a first branch of the decision tree model, and the decision tree model The plating adhesion amount control method according to claim 5 or 6, wherein the plating amount is updated.   8. The plating according to claim 2, wherein, in the learning step, the plating adhesion amount is classified into a plurality of groups according to the past operation conditions using an F test. Amount control method. In the plating adhesion amount control method for controlling the plating adhesion amount of the steel strip by blowing gas from the nozzle to the steel strip continuously taken out from the hot dipping bath,
By using a decision tree model with operating conditions including character data as explanatory variables and the corresponding plating adhesion amount as a target variable, the plating adhesion amount in the actual operation is estimated by estimating the plating adhesion amount in actual operation. The plating adhesion amount control method characterized by controlling.   In the decision tree model, the first branching condition is a steel type, the second branching condition is one of the steel strip width and the steel strip thickness, and the third branching condition is a steel strip. The plating adhesion amount control method according to claim 9, wherein the other one of the width and the thickness of the steel strip and the fourth and subsequent branching conditions are determined based on an F test.   A method for producing a hot-dip galvanized steel strip, comprising producing a hot-dip steel strip whose plating adhesion amount is controlled by the plating adhesion amount control method according to any one of claims 1 to 10. In the plating adhesion amount control device for controlling the plating adhesion amount of the steel strip by blowing gas from the nozzle to the steel strip continuously taken out from the hot dipping bath,
A learning unit that generates a decision tree model by classifying the plating adhesion amounts into a plurality of groups according to the past operation conditions, using a plurality of data sets including past operation conditions and corresponding plating adhesion amounts; ,
Based on the decision tree model and actual operation conditions, an adhesion amount estimation unit that estimates the current plating adhesion amount,
By feeding back the estimated plating adhesion amount, a preset control unit that controls the plating adhesion amount of the steel strip to be a target value;
A plating adhesion amount control device comprising:   The plurality of data sets are, as the past operation conditions, a steel type that is character data corresponding to the composition of the steel strip, a plate width of the steel strip, a plate thickness of the steel strip, and a tip of the nozzle The plating adhesion amount control device according to claim 12, comprising a distance to the steel strip and a pressure of a gas blown from the nozzle. In a plating adhesion amount control program for controlling the plating adhesion amount of the steel strip by blowing gas from a nozzle to the steel strip continuously taken out from the hot dipping bath,
A learning means for generating a decision tree model by classifying the plating adhesion amount into a plurality of groups according to the past operation conditions, using a plurality of data sets including past operation conditions and corresponding plating adhesion amounts,
Based on the decision tree model and actual operating conditions, the adhesion amount estimation means for estimating the current plating adhesion amount,
By feeding back the estimated plating adhesion amount, preset control means for controlling the plating adhesion amount of the steel strip to a target value;
Adhesion amount control program to function as

Images