JP7494867B2 - Method for generating a temperature prediction model for hot-rolled sheet and a transformation enthalpy prediction model for hot-rolled sheet, a coiling temperature prediction method for hot-rolled sheet, a temperature control method, and a manufacturing method - Google Patents

Description

The present invention relates to a method for generating a temperature prediction model for hot-rolled sheet that predicts and controls the coiling temperature of the hot-rolled sheet in the run-out table of a hot rolling line, a method for generating a transformation enthalpy prediction model for the hot-rolled sheet, a method for predicting the coiling temperature of the hot-rolled sheet, a method for controlling the temperature of the hot-rolled sheet, and a method for manufacturing the hot-rolled sheet.

In the manufacturing process of hot-rolled steel plate (hot-rolled plate), for example as shown in FIG. 1, a slab is heated to a predetermined temperature in a heating furnace 1, and the heated slab is rolled in a rough rolling mill 2 to produce a rough bar. Next, this rough bar is rolled to a predetermined thickness in a finishing mill (continuous hot finishing mill) 3 consisting of multiple rolling stands to produce a hot-rolled steel plate (hereinafter simply referred to as "steel plate") S.

Then, a cooling device 5 installed on the run-out table 4 sprays a refrigerant onto the upper and lower surfaces of the steel sheet S to cool it, and the steel sheet S is then wound up by a winder 6. The temperature (sheet temperature) of the steel sheet S before it is wound up by the winder 6 is called the "winding temperature." This winding temperature has a significant effect on the material properties of the steel sheet S, and is therefore an important indicator for quality control.

The winding temperature is controlled by the flow rate of the refrigerant sprayed by the cooling device 5 (hereinafter simply referred to as "flow rate") according to the transport speed of the steel sheet S on the run-out table 4. However, control of the winding temperature is extremely difficult for certain materials due to changes in cooling capacity accompanying changes in the surface temperature of the steel sheet S, changes in heat quantity accompanying transformation of the steel sheet S, etc.

In general, the cooling capacity is given by some function or a correction coefficient is given to control the coiling temperature so that the coiling temperature is accurate according to the cooling conditions of the steel sheet S. For example, Patent Document 1 discloses a technology for controlling the coiling temperature by creating an equation for cooling capacity that is a function of the refrigerant temperature, the steel sheet temperature, and the refrigerant water volume density, and further multiplying it by a correction coefficient.

JP 2015-167976 A

However, in the method of Patent Document 1, in actual operation, the accuracy of the coiling temperature depends on the correction coefficient, so how to determine the correction coefficient is important. Also, in the method of Patent Document 1, the cooling capacity is given by some kind of function, but if the function is incorrect or due to the influence of changes in heat quantity associated with the transformation of the steel sheet, the accuracy of the coiling temperature may actually decrease.

In particular, the transformation temperature of high carbon steel and high tensile steel containing a large amount of alloying components is lower than that of general low carbon steel. Therefore, the correction coefficient for hitting the coiling temperature is significantly different from that of general low carbon steel. In addition, the conveying speed of the steel plate on the run-out table generally changes depending on the position in the rolling direction, so the cooling time of the steel plate on the run-out table changes even for the same rolled material. Therefore, even for the same rolled material, the correction coefficient for hitting the coiling temperature may change, which may lead to a decrease in the accuracy of the coiling temperature. In addition, the scale condition on the steel plate surface may change due to the influence of the components contained in the steel plate, such as Si, and the descaler spraying conditions before finish rolling, which may change the cooling capacity.

The present invention has been made in consideration of the above, and aims to provide a method for generating a temperature prediction model for hot-rolled sheet, a method for generating a transformation enthalpy prediction model for hot-rolled sheet, a method for predicting the coiling temperature of hot-rolled sheet, a method for controlling the temperature of hot-rolled sheet, and a method for manufacturing hot-rolled sheet, that can predict and control the temperature (coiler temperature) of hot-rolled sheet with high accuracy during cooling in a hot rolling line.

In order to solve the above-mentioned problems and achieve the objective, the method for generating a temperature prediction model for hot-rolled sheet according to the present invention generates a temperature prediction model for predicting the coiling temperature of a hot-rolled sheet using multiple training data in a process in which a hot-rolled sheet after finish rolling is cooled and coiled on a run-out table in a hot rolling line. The input data are the components of the sheet, the operational parameters of the finishing rolling mill and the run-out table, and the actual data of the temperature of the hot-rolled sheet measured on the run-out table from the exit side of the finishing rolling mill, and the output data is the actual data of the coiling temperature of the hot-rolled sheet.

In addition, the method for generating a temperature prediction model for hot-rolled sheet according to the present invention includes, in the above invention, operational parameters of the finishing mill and the run-out table, the sheet thickness after finish rolling, the conveying speed of the run-out table, the flow rate of the coolant at the run-out table, and descaling injection performance data before finish rolling, and performance data of the temperature of the hot-rolled sheet, the temperature of the hot-rolled sheet after finish rolling.

In addition, the method for generating a temperature prediction model for hot-rolled sheet according to the present invention generates the temperature prediction model using one or more machine learning models selected from neural networks, decision tree learning, random forests, and support vector regression.

In order to solve the above-mentioned problems and achieve the objective, the method for generating a transformation enthalpy prediction model for hot-rolled sheet according to the present invention generates a transformation enthalpy prediction model for predicting the transformation enthalpy of a hot-rolled sheet using multiple teacher data in a process in which a hot-rolled sheet after finish rolling is cooled and wound on a run-out table in a hot rolling line, the input data being the components contained in the sheet, the operational parameters of the finishing rolling mill and the run-out table, and the actual data of the temperature of the hot-rolled sheet measured on the run-out table from the exit side of the finishing rolling mill, and the output data being the transformation enthalpy of the hot-rolled sheet calculated based on the actual data of the temperature of the hot-rolled sheet.

In addition, the method for generating a transformation enthalpy prediction model for hot-rolled sheet according to the present invention includes, in the above invention, operational parameters of the finishing mill and the run-out table, including the sheet thickness after finish rolling, the conveying speed of the run-out table, the flow rate of the coolant at the run-out table, and actual descaling injection data before finish rolling, and actual data of the temperature of the hot-rolled sheet, including the temperature of the hot-rolled sheet after finish rolling and the coiling temperature of the hot-rolled sheet.

In addition, the method for generating a transformation enthalpy prediction model for hot-rolled sheet according to the present invention generates the transformation enthalpy prediction model using one or more machine learning models selected from neural networks, decision tree learning, random forests, and support vector regression.

In order to solve the above-mentioned problems and achieve the objectives, the method for predicting the coiling temperature of a hot-rolled sheet according to the present invention predicts the transformation enthalpy of the hot-rolled sheet using a transformation enthalpy prediction model generated by the above-mentioned method for generating a transformation enthalpy prediction model of a hot-rolled sheet, and calculates the coiling temperature of the hot-rolled sheet based on the predicted transformation enthalpy of the hot-rolled sheet.

In order to solve the above-mentioned problems and achieve the objectives, the method for controlling the temperature of a hot-rolled sheet according to the present invention predicts the coiling temperature of the hot-rolled sheet using a temperature prediction model generated by the above-mentioned method for generating a temperature prediction model for a hot-rolled sheet, and resets the operating parameters of the run-out table based on the predicted coiling temperature of the hot-rolled sheet.

In order to solve the above-mentioned problems and achieve the objectives, the method for controlling the temperature of a hot-rolled sheet according to the present invention predicts the transformation enthalpy of the hot-rolled sheet using a transformation enthalpy prediction model generated by the above-mentioned method for generating a transformation enthalpy prediction model of a hot-rolled sheet, and resets the operating parameters of the run-out table based on the coiling temperature of the hot-rolled sheet calculated from the predicted transformation enthalpy of the hot-rolled sheet.

In order to solve the above-mentioned problems and achieve the objectives, the method for manufacturing hot-rolled sheet according to the present invention produces hot-rolled sheet by resetting the operating parameters of the run-out table using the above-mentioned method for controlling the temperature of hot-rolled sheet, and cooling the hot-rolled sheet on the run-out table based on the reset operating parameters.

The method for generating a temperature prediction model for hot-rolled sheet, the method for generating a transformation enthalpy prediction model for hot-rolled sheet, the method for predicting the coiling temperature of hot-rolled sheet, the method for controlling the temperature of hot-rolled sheet, and the method for manufacturing hot-rolled sheet according to the present invention make it possible to predict and control with high accuracy the temperature (coiling temperature) of the hot-rolled sheet during cooling in a hot rolling line, and to manufacture a hot-rolled sheet with a precisely controlled coiling temperature.

FIG. 1 is a diagram showing an example of an equipment configuration of a hot rolling line to which a method for generating a temperature prediction model for a hot-rolled sheet, a method for generating a transformation enthalpy prediction model for a hot-rolled sheet, a method for predicting a coiling temperature of a hot-rolled sheet, and a method for controlling a temperature of a hot-rolled sheet according to embodiments of the present invention are applied. FIG. 2 is a diagram showing an example of the configuration of the runout table of FIG. FIG. 3 is a graph for explaining the change in transformation enthalpy of a steel sheet with temperature. FIG. 4 is a diagram showing an example of a change in specific heat with respect to the temperature of a steel sheet. FIG. 5 is a diagram showing an example of the relationship between the measured temperature and enthalpy of an austenitic stainless steel sheet. FIG. 6 is a diagram showing an example of the relationship between the measured temperature and enthalpy of low carbon steel. FIG. 7 is a diagram showing an example of the configuration of a temperature control device for a hot-rolled strip according to an embodiment of the present invention.

A method for generating a temperature prediction model for hot-rolled sheet, a method for generating a transformation enthalpy prediction model for hot-rolled sheet, a method for predicting the coiling temperature of hot-rolled sheet, a method for controlling the temperature of hot-rolled sheet, and a method for manufacturing hot-rolled sheet according to embodiments of the present invention will be described with reference to the drawings. Note that the components in the following embodiments include those that are easily replaceable by a person skilled in the art, or those that are substantially the same.

[Hot rolling line]
The equipment configuration of a hot rolling line to which the method for generating a temperature prediction model for a hot rolled sheet, the method for generating a transformation enthalpy prediction model for a hot rolled sheet, the method for predicting the coiling temperature of a hot rolled sheet, the method for controlling the temperature of a hot rolled sheet, and the method for manufacturing a hot rolled sheet according to the embodiments of the present invention are applied will be described with reference to Fig. 1 and Fig. 2. The hot rolling line includes a heating furnace 1, a roughing mill 2, a finishing mill 3, a run-out table 4, and a winder 6.

The heating furnace 1 heats the slab to a predetermined temperature. The roughing mill 2, which is made up of multiple rolling stands, roughly rolls the heated slab to produce a rough bar. The finishing mill 3, which is made up of multiple rolling stands, rolls the rough bar to a predetermined thickness to produce steel plate S.

A cooling device 5 is installed on the run-out table 4, and the cooling device 5 cools the steel sheet S being rolled in the hot rolling line. A winder 6 winds up the steel sheet S cooled on the run-out table 4. The hot rolling line may have a number of auxiliary equipment such as width reduction equipment for adjusting the width, a welding machine, and a descaler for removing scale, as required. By passing the slab through such a hot rolling line under the desired conditions, the slab becomes a hot-rolled coil with the desired thickness and structure.

A descaler 14 may be installed before the finishing rolling mill 3. The descaler 14 is a device for removing scale from the surface of the steel sheet S by spraying high-pressure water (coolant) of 10 MPa or more, and is equipped with spray nozzles arranged to face the upper and lower surfaces of the steel sheet S. The descaler 14 may be configured with multiple devices arranged in the longitudinal direction that can spray at spray pressures of, for example, 20 MPa, 30 MPa, and 50 MPa, and different devices may be used depending on the material. The cooling capacity during water cooling changes depending on the type and thickness of the scale, which also affects the transformation state of the steel sheet, so the components contained in the steel sheet S and the descaler spray conditions are important for predicting the coiling temperature.

As shown in FIG. 2, the runout table 4 is divided into multiple cooling zones (nine in the figure), and in each cooling zone, it is possible to individually control the injection of refrigerant and the flow rate of the refrigerant. In addition, water, for example, can be used as the refrigerant, and the upper and lower surfaces of the steel plate S are cooled by injecting pressurized refrigerant from cooling devices 5 installed on the upper and lower sides of the steel plate S.

The upper cooling device 5 may be, for example, a pipe laminar cooling device with a simple pipe nozzle. The lower cooling device 5 may be, for example, a spray cooling device with a fan-shaped spray pattern spray nozzle. Both the upper and lower cooling devices 5 are equipped with multiple nozzles so that they have uniform cooling capacity in the width direction of the steel plate S. By installing multiple rows of such nozzles in the longitudinal direction of the steel plate S, each cooling zone is formed, and the flow rate of the refrigerant is controlled on a cooling zone basis.

On the runout table 4, multiple thermometers are installed for each cooling zone to measure the surface temperature of the steel sheet S. That is, a finishing mill exit thermometer 7 is installed on the exit side of the finishing rolling mill 3. A first intermediate thermometer 8 is installed at a position approximately 1/3 of the length of the runout table 4. A second intermediate thermometer 9 is installed at a position approximately 2/3 of the length of the runout table 4. A winding thermometer 10 is installed at the rear end position of the runout table 4. Temperature measurement information measured by these thermometers is input into a database 13 described below.

Each thermometer can be a type that measures a spot at a specific location in the width direction of the steel plate S and determines the representative temperature. However, a scanning thermometer that can measure the temperature distribution in the width direction of the steel plate S can also be used, or multiple thermometers can be installed in the width direction of the steel plate S to measure the temperature, or two-dimensional thermal imaging data can be acquired and analyzed for temperature, or the temperature or average temperature can be measured at a specific location in the width direction of the steel plate S and the representative temperature can be determined from that.

In addition to these thermometers, a transformation rate measuring device such as that proposed in the reference document (JP Patent Publication 56-24017) may be installed to acquire data related to the transformation rate. Furthermore, the temperature used as input data may be either the plate surface temperature or the calculated temperature at the center of the plate thickness, but the acquired data and predicted data must be the same in order to eliminate errors.

[Method of generating temperature prediction model]
A method for generating a temperature prediction model for a hot-rolled sheet according to an embodiment of the present invention will be described. In the method for generating a temperature prediction model, a temperature prediction model is generated for predicting a coiling temperature of a steel sheet S after finish rolling in a hot rolling line, in a process in which the steel sheet S is cooled on a run-out table 4 and then coiled by a coiler 6.

In the method for generating a temperature prediction model for hot-rolled sheet according to this embodiment, specifically, the components of the steel sheet S, the operating parameters of the finishing rolling mill 3 and the run-out table 4, and the actual data on the temperature of the steel sheet S measured on the run-out table 4 from the exit side of the finishing rolling mill 3 are used as input data, and the actual data on the coiling temperature of the steel sheet S is used as output data to generate a temperature prediction model that predicts the coiling temperature of the steel sheet S.

Examples of operational parameters of the finishing rolling mill 3 and the run-out table 4 include the thickness of the steel sheet S after finish rolling, the conveying speed of the steel sheet S on the run-out table 4, the flow rate of the refrigerant in the cooling device 5 installed on the run-out table 4, and the temperature of the steel sheet S measured on the run-out table 4. Examples of actual data on the temperature of the steel sheet S include the temperature of the steel sheet S after finish rolling. Details of the input data and output data used to generate the temperature prediction model are described below.

<Input data>
(1) Chemical Compositions of Steel Sheet Data on the chemical compositions of the steel sheet S is essential information for predicting the transformation of the steel sheet S. That is, the chemical compositions of the steel sheet S are information that determines the transformation temperature of the steel sheet S and the phase fractions before and after the transformation.
(2) Operation parameters As input data, it is desirable to use operation parameters of the cooling device 5 installed on the run-out table 4, but it is also possible to use set values of the operation parameters determined by a cooling condition calculation unit 121 described later. Specifically, necessary information includes the thickness of the steel sheet S after finish rolling, i.e., on the run-out table 4, the conveying speed of the steel sheet S, the flow rate of the coolant in the cooling device 5, and measured values of multiple temperatures of the steel sheet S.

It is also desirable to use descaling injection data before finish rolling as an operational parameter. Since the cooling capacity during water cooling varies depending on the type and thickness of scale, it is desirable to use parameters that can predict the type and thickness of scale as input data in order to accurately predict the coiling temperature. Specifically, in addition to the components contained in the steel sheet S and operational parameters, it is necessary to use descaling injection data.

Here, the descaling injection performance data refers to information related to the descaling injection performance (water volume, injection pressure, injection angle, etc.) of the descaler 14 installed before the finishing rolling mill 3, the conveying speed when passing through the descaler 14, the steel sheet surface temperature before descaling injection, etc. Among these, the injection pressure in particular is an important parameter since it has a significant effect on scale removal. The steel sheet surface temperature before descaling injection is temperature data measured by a radiation thermometer installed before the descaler 14 of the finishing rolling mill 3.

As the measured temperature of the steel sheet S, it is preferable to use the measured temperature of any one of the finishing rolling mill outlet thermometer 7, the first intermediate thermometer 8, and the second intermediate thermometer 9 shown in FIG. 2. These measured temperatures are used to grasp at which cooling position the transformation behavior of the steel sheet S occurs, so it is preferable to obtain the temperature of the steel sheet S by dividing the area between the finishing rolling mill 3 and the winder 6 as finely as possible. In practice, it is preferable to obtain the temperature of the steel sheet S between the cooling zones where the measurement error due to the influence of the refrigerant is small. Specifically, it is preferable to obtain the temperature by dividing the area between the finishing rolling mill 3 and the winder 6 into distances of about 30 m or less.

The temperature of the steel sheet S is preferably measured when the coil tip reaches the winding machine 6. This is because, when no tension is applied to the steel sheet S, the shape of the steel sheet S is easily disturbed and the accuracy of the temperature measurement is easily reduced, and because importance is placed on the prediction accuracy of the tip, where the temperature needs to be predicted only by feedforward control.

Furthermore, the input data may include various other information stored in the database 13. This is because the accuracy of predicting the coiling temperature may be improved by using operational performance data that may directly or indirectly affect the coiling temperature. Examples of the various information include the width of the steel sheet S, environmental temperatures such as the outside air temperature, and hot rolling information such as heating conditions.

<Output data>
(1) Coiling Temperature As output data, actual data on the coiling temperature of the steel sheet S is used. The actual data on the coiling temperature of the steel sheet S is, for example, a measurement value of the coiling thermometer 10 shown in FIG.

In the temperature prediction model generation method, these input data and output data are used to generate a temperature prediction model that predicts the coiling temperature of the steel sheet S. In the temperature prediction model generation method, by using the components contained in the steel sheet S, the temperature of the steel sheet S at the exit side of the finishing rolling mill 3, and the flow rate of the refrigerant in the cooling device 5 as input data, a temperature prediction model that predicts the coiling temperature taking into account the transformation behavior of the steel sheet S can be generated.

In addition, the input data when generating a temperature prediction model may be not only actual values, but also set values and calculated values calculated from rolling actual information. In addition, the acquisition position and number of temperatures of the steel sheet S used as input data may be changed appropriately depending on the transportation status of the steel sheet S, the contained components of the steel sheet S, and the required characteristics. Furthermore, it is also possible to generate a temperature prediction model that predicts the temperature of the steel sheet S at any position on the run-out table 4, not limited to the coiling temperature. This makes it possible to grasp the amount of temperature change due to the transformation of the steel sheet S on the run-out table 4, making it possible to reset the necessary cooling conditions for each cooling zone.

[Method of generating transformation enthalpy prediction model]
A method for generating a transformation enthalpy prediction model for a hot-rolled sheet according to an embodiment of the present invention will be described. In the transformation enthalpy prediction model, a transformation enthalpy prediction model is generated for predicting the transformation enthalpy of the steel sheet S on the run-out table 4 in a process in which the steel sheet S after finish rolling is cooled on the run-out table 4 and then wound by a winder 6 in a hot rolling line.

In the method for generating a transformation enthalpy prediction model for hot-rolled sheets according to this embodiment, specifically, the input data are the components of the steel sheet S, the operating parameters of the finishing rolling mill 3 and the run-out table 4, and the actual data on the temperature of the steel sheet S measured on the run-out table 4 from the exit side of the finishing rolling mill 3, and the output data is the transformation enthalpy of the steel sheet S calculated based on the actual data on the temperature of the steel sheet S. A transformation enthalpy prediction model that predicts the transformation enthalpy of the steel sheet S is generated using multiple teacher data.

Examples of operational parameters of the finishing rolling mill 3 and the run-out table 4 include the thickness of the steel sheet S after finish rolling, the conveying speed of the steel sheet S on the run-out table 4, the flow rate of the coolant in the cooling device 5 installed on the run-out table 4, and the temperature of the steel sheet S measured on the run-out table 4. Examples of actual data on the temperature of the steel sheet S include the temperature of the steel sheet S after finish rolling and the coiling temperature of the steel sheet S. Details of the input data and output data used to generate the transformation enthalpy prediction model are described below.

<Input data>
(1) Chemical Compositions of Steel Sheet Data on the chemical compositions of the steel sheet S is essential information for predicting the transformation of the steel sheet S. That is, the chemical compositions of the steel sheet S are information that determines the transformation temperature of the steel sheet S and the phase fractions before and after the transformation.
(2) Operation Parameters As input data, it is desirable to use operation parameters of the cooling device 5 installed on the run-out table 4, but it is also possible to use set values of the operation parameters determined by a cooling condition calculation unit 121 described later. Specific information required is the plate thickness of the steel sheet S, the conveying speed of the steel sheet S, the flow rate of the coolant in the cooling device 5, and measured values of multiple temperatures of the steel sheet S.

Furthermore, in order to predict the cooling capacity on the run-out table and accurately predict the transformation of the steel sheet S, it is preferable to use descaling injection performance data. Here, descaling injection performance data refers to information on the descaling injection performance (water volume, injection pressure, injection angle, etc.) of the descaler 14 installed before the finishing rolling mill 3, the conveying speed when passing through the descaler 14, the steel sheet surface temperature before descaling injection, etc. Among these, the injection pressure is an especially important parameter because it has a large effect on scale removal. The steel sheet surface temperature before descaling injection is temperature data measured by a radiation thermometer installed before the descaler 14 of the finishing rolling mill 3.

As the temperature measurement value of the steel sheet S, it is desirable to use the measurement value of any one of the finishing rolling mill outlet thermometer 7, the first intermediate thermometer 8, the second intermediate thermometer 9, and the winding thermometer 10 shown in FIG. 2. These temperature measurements are intended to grasp at which cooling position the transformation behavior of the steel sheet S occurs, so it is desirable to obtain the temperature of the steel sheet S by dividing the area between the finishing rolling mill 3 and the winding machine 6 as finely as possible. Therefore, in addition to the finishing rolling mill outlet thermometer 7 mentioned above, it is desirable to obtain the temperature at two or more points on the run-out table 4. In practical terms, it is desirable to obtain the temperature of the steel sheet S between the cooling zones, where the measurement error due to the influence of the refrigerant is small.

The temperature of the steel sheet S is preferably measured when the coil tip reaches the winding machine 6. This is because, when no tension is applied to the steel sheet S, the shape of the steel sheet S is easily disturbed and the accuracy of the temperature measurement is easily reduced, and because importance is placed on the prediction accuracy of the tip, where the temperature needs to be predicted only by feedforward control.

Furthermore, the input data may include various other information stored in the database 13. This is because the accuracy of predicting the coiling temperature may be improved by using operational performance data that may directly or indirectly affect the coiling temperature. Examples of the various information include the width of the steel sheet S, environmental temperatures such as the outside air temperature, and hot rolling information such as heating conditions.

<Output data>
(1) Transformation enthalpy As output data, a calculated value of transformation enthalpy (hereinafter also referred to as "enthalpy") at each position of the steel sheet S is used. In steel, specific heat change and latent heat occur due to transformation. In controlling the coiling temperature of the steel sheet S, the energy required from the exit temperature of the finishing rolling mill 3, which is the start of cooling, to the coiling temperature, which is the end of cooling, that is, the difference in enthalpy, is important. If the temperature change in enthalpy is known, the difference in enthalpy between the exit side of the finishing rolling mill 3 and the entry side of the coiler 6, that is, the amount of cooling required at the run-out table 4, can be known.

Figure 3 shows examples of enthalpy for general low carbon steel and high carbon steel. Enthalpy is the integration of specific heat with respect to temperature, and is expressed as a function of temperature. Compared to low carbon steel, high carbon steel has a lower transformation temperature, so enthalpy increases rapidly at low temperatures. In addition, with general low carbon steel, transformation is already complete by the time the coiling thermometer 10 is reached, so it is less susceptible to enthalpy changes due to transformation. However, with high carbon steel and high tensile steel, as the subject of this invention, there are cases where transformation is not complete by the time the coiling thermometer 10 is reached, and it is more susceptible to enthalpy changes due to transformation.

Enthalpy may be found using a calculation model from temperature information, or multiple temperature dependence patterns of enthalpy may be created from publicly available data or independently acquired data based on the components and temperature history of the steel sheet S within the range expected for operation, and it may be determined which of these patterns applies. Also, even if not all data is obtained experimentally, a formula for calculating enthalpy from one or more parameters may be created, and the enthalpy may be found in a format that identifies the parameters.

The specific method for calculating enthalpy using a calculation model involves applying a modeled cooling capacity equation to the heat conduction equation as a boundary condition, but the density, specific heat, and thermal conductivity must be provided as the necessary physical property information for the steel plate S. For density and thermal conductivity, publicly available data that is close to the content of the steel plate S can be used.

An example of a formula for calculating enthalpy is explained below. Since specific heat is the derivative of enthalpy with respect to temperature, enthalpy can be calculated if there is a formula that expresses the temperature dependence of specific heat.

4 shows an example of the change in specific heat with respect to the temperature of the steel sheet S. A to E in the figure represent the following.
A: Specific heat of the austenite phase (reference specific heat)
B: Specific heat of ferrite phase before phase transformation C: Specific heat of magnetic transformation D: Heat of transformation of ferrite phase E: Heat of transformation of pearlite layer

After finish rolling, typical hot-rolled steel sheets undergo a phase transformation from austenite to ferrite, pearlite, etc. Furthermore, after transformation to the ferrite phase, magnetic transformation occurs. The specific heat differs between the austenite phase before transformation and the ferrite phase after transformation, and also differs between the ferrite phase before magnetic transformation and the ferrite phase after magnetic transformation. Measurements of specific heat show that before phase transformations such as the ferrite phase, the austenite phase and the ferrite phase before magnetic transformation have roughly the same specific heat.

Heat is also generated during phase transformations, such as from the austenite phase to the ferrite phase, and from the austenite phase to the pearlite phase. For simplicity's sake, as shown in Figure 4, the specific heat B of the ferrite phase before the phase transformation is assumed to be an extension of the specific heat A of the reference austenite phase, and the difference between A and B is included in the transformation heat D of the ferrite phase.

Based on the results of the specific heat measurements, the approximate formulas for B to E in Figure 4 are shown as the following formulas (1) to (4).

The specific heat can be expressed by applying the transformation temperature and phase fraction to the above formulas (1) to (4), and enthalpy can be calculated by integrating this with respect to temperature.

An example of the relationship between the measured temperature and enthalpy of an austenitic stainless steel sheet is shown in Figure 5. In the figure, F to I represent the following:
F: Enthalpy calculated from the temperature measured by the finishing rolling mill delivery thermometer 7 and a physical model G: Enthalpy calculated from the temperature measured by the first intermediate thermometer 8 and a physical model H: Enthalpy calculated from the temperature measured by the second intermediate thermometer 9 and a physical model I: Enthalpy calculated from the temperature measured by the winding thermometer 10 and a physical model J: Enthalpy (reference enthalpy) diagram of the austenite phase

Curve J in Figure 5 is the enthalpy diagram for the austenite phase, and on this line are plots of the temperatures measured by each thermometer and the enthalpies F, G, H, and I calculated using a physical model.

An example of the relationship between the measured temperature and enthalpy of low carbon steel is shown in Fig. 6. In the figure, K and L represent the following:
K: Enthalpy diagram at the completion of transformation L: Enthalpy curve

In Figure 6, cooling was performed at the same conveying speed and cooling pattern as in Figure 5. Curve K in the figure is an enthalpy diagram when the transformation to the ferrite phase and pearlite phase and the magnetic transformation are complete. In this figure, unlike Figure 5 described above, the temperature drop is smaller relative to the drop in enthalpy at the positions of the second intermediate thermometer 9 and the winding thermometer 10. This is the effect of the change in heat quantity due to transformation, and it is possible to obtain the enthalpy curve L for this steel type by applying the transformation temperatures to the above formulas (1) to (4).

Furthermore, a method for obtaining teacher data for enthalpy based on the above calculations and input data will be described. If the target is an austenitic stainless steel sheet, neither phase transformation nor magnetic transformation occurs, and the specific heat can be expressed only by the above formula (1). Using this, a physical model is created that expresses the temperature history in the cooling device 5 of the run-out table 4. Then, using this physical model, the relationship between the measured temperature of the steel sheet S after finish rolling and the enthalpy can be obtained by inputting data on the thickness of the steel sheet S after finish rolling, the exit temperature of the finish rolling mill 3, the intermediate temperature, and the coiling temperature, the conveying speed of the steel sheet S, the descaling injection performance before finish rolling, and the flow rate of the refrigerant.

In the method for generating a transformation enthalpy prediction model, these input data and output data are used to generate a transformation enthalpy prediction model that predicts the transformation enthalpy of steel sheet S. Note that, although the case where a prediction model is generated using transformation enthalpy as training data has been described here, any indicator that can unambiguously show the transformation of steel sheet S with respect to temperature change may be used as training data other than transformation enthalpy.

In the method for generating a transformation enthalpy prediction model, a transformation enthalpy prediction model for predicting transformation enthalpy can be generated by using as input data the components contained in the steel sheet S, the temperature and thickness of the steel sheet S at the exit side of the finishing rolling mill 3, the descaling injection performance before finishing rolling, and the flow rate of the coolant in the cooling device 5. Furthermore, it is possible to predict the coiling temperature based on the predicted transformation enthalpy by heat transfer calculation using the coolant conditions, and to output the predicted coiling temperature. This makes it possible to reset the cooling conditions required for each cooling zone based on the predicted coiling temperature calculated based on the predicted transformation behavior.

[Temperature Control Device]
Hereinafter, the configuration of a temperature control device 100 for a hot-rolled sheet according to an embodiment of the present invention will be described with reference to Fig. 3. The temperature control device 100 is configured by a general-purpose information processing device such as a personal computer or a workstation. The temperature control device 100 includes a host computer 11, a control device 12, and a database 13.

The control device 12 acquires from the host computer 11 size information such as the conveying speed of the steel sheet S, the thickness of the steel sheet S, and the width of the steel sheet S, as well as the target winding temperature required to give the steel sheet S predetermined characteristics. The control device 12 then calculates the operating conditions for achieving the above conditions and determines the operating parameters of the cooling device 5.

The control device 12 receives input from the host computer 11 the finish rolling conditions, including the descaler injection conditions before finish rolling, the temperature of the steel plate S before and after finish rolling, the cooling conditions, including the cooling pattern of the cooling device 5, and other operating conditions, including the components of the steel plate S, the conveying speed of the steel plate S, the thickness of the steel plate S, the width of the steel plate S, etc.

The control device 12 includes a cooling condition calculation unit 121, a prediction model 122, and a cooling condition output unit 123.

The cooling condition calculation unit 121 performs heat transfer calculations and determines cooling conditions (operation parameters) such as the flow rate of the refrigerant in the cooling device 5 and the conveying speed of the steel sheet S on the run-out table 4 so as to satisfy the range of the target winding temperature. The cooling condition output unit 123 outputs the cooling conditions determined by the cooling condition calculation unit 121 to the cooling device 5.

Here, the operating parameters of the cooling device 5 include whether or not refrigerant is sprayed in each cooling zone, and the flow rate of the refrigerant. Even if the total flow rate of the refrigerant is the same, if the cooling zone from which the refrigerant is sprayed and the flow rate of the refrigerant at that time are different, the winding temperature may change. There are many combinations of cooling zones from which the refrigerant is sprayed and the flow rates of the refrigerant, but it is desirable to create patterns to some extent depending on the target winding temperature and the components contained in the steel sheet S. For example, under conditions where the winding temperature is high, the flow rate of the refrigerant can be set to a constant flow rate and sprayed preferentially from the previous zone.

The database 13 stores temperature measurement information measured by each thermometer (finishing rolling mill exit thermometer 7, first intermediate thermometer 8, second intermediate thermometer 9, and winding thermometer 10). In addition to the temperature measurement information, the database 13 also stores information set by the host computer 11, such as the thickness, width, and contained components of the steel sheet S after cooling, and the target winding temperature.

In addition, the database 13 also stores the set values of the operating parameters of the cooling device 5 input from the cooling condition output unit 123 (or the cooling condition calculation unit 121) together with the temperature measurement information. Furthermore, it is preferable that the database 13 also stores the flow rate of the refrigerant in the cooling device 5 and the actual values of the conveying speed of the steel sheet S, etc.

The prediction model 122 is a temperature prediction model or a transformation enthalpy prediction model generated by a machine learning model. The prediction model 122 may be generated using any machine learning model as long as it is capable of predicting temperature or transformation enthalpy sufficient for practical use. The prediction model 122 may be generated by one or more machine learning models selected from, for example, commonly used neural networks, decision tree learning, random forests, and support vector regression. The prediction model 122 may also be generated by an ensemble model that combines multiple machine learning models.

[Temperature prediction method]
When cooling the steel sheet S, the temperature control device 100 predicts the coiling temperature of the steel sheet S by using the prediction model 122. In this case, as shown in Fig. 7, first, information on the specifications of the steel sheet S and operation conditions is input from the host computer 11 to the control device 12.

Then, the cooling condition calculation unit 121 of the control device 12 determines cooling conditions (operation parameters) such as the flow rate of the refrigerant in the cooling device 5 and the conveying speed of the steel sheet S on the run-out table 4 so as to satisfy the range of the preset target winding temperature. Next, the cooling condition calculation unit 121 calculates a predicted value of the winding temperature using the prediction model 122, using these cooling conditions, necessary specifications, operation conditions, etc. as input data. Note that when the prediction model 122 is a transformation enthalpy prediction model, the transformation enthalpy is first predicted using the transformation enthalpy prediction model, and then the predicted value of the winding temperature is calculated from the predicted value of the transformation enthalpy by heat transfer calculation.

The initial operating conditions can be set using table values categorized by manufacturing conditions such as the thickness, width, and steel type of the steel sheet S. In addition, relationships obtained from operating experience can be organized using approximate formulas.

[Temperature control method]
The cooling condition calculation unit 121 resets the operation parameters of the cooling device 5 so that the predicted winding temperature approaches a preset target winding temperature. That is, the cooling condition calculation unit 121 compares the predicted value of the winding temperature obtained as described above with a preset allowable range of temperature deviation of the winding temperature, and determines that the predicted value of the winding temperature is acceptable if it is within the allowable range of temperature deviation.

On the other hand, if the predicted winding temperature value is outside the allowable temperature deviation range, the cooling condition calculation unit 121 changes the initial operating conditions so that the temperature deviation falls within the allowable temperature deviation range. Furthermore, a check may be performed to see if any equipment malfunctions have occurred. The allowable temperature deviation range is preferably set to, for example, "within ±25°C of the target winding temperature." This is because if a temperature deviation occurs beyond this range, the possibility of poor quality increases.

The cooling condition output unit 123 then outputs the reset operating parameters of the cooling device 5 to the cooling device 5. The reset operating parameter is preferably, for example, the flow rate of the refrigerant in the cooling device 5. This makes it possible to realize appropriate operating conditions for the cooling device 5 according to the attributes and manufacturing conditions of each steel plate S.

According to the above-described embodiments of the method for generating a temperature prediction model for hot-rolled sheet, the method for generating a transformation enthalpy prediction model for hot-rolled sheet, the method for predicting the coiling temperature of hot-rolled sheet, the method for controlling the temperature of hot-rolled sheet, and the method for manufacturing hot-rolled sheet, the temperature (coiling temperature) of the hot-rolled sheet during cooling can be predicted and controlled with high accuracy in a hot rolling line, and a hot-rolled sheet with a precisely controlled coiling temperature can be manufactured.

In addition, according to the method for generating a temperature prediction model for hot-rolled sheet, the method for generating a transformation enthalpy prediction model for hot-rolled sheet, the method for predicting the coiling temperature of hot-rolled sheet, the method for controlling the temperature of hot-rolled sheet, and the method for manufacturing hot-rolled sheet, which are related to the embodiment, it is possible to predict the position where transformation occurs on the run-out table 4 and the amount of heat required for transformation by focusing on the transformation behavior of the steel sheet S on the run-out table 4. By reflecting this in the cooling conditions, it is possible to control the coiling temperature with greater accuracy. In addition, by controlling the cooling conditions of the run-out table 4 based on the prediction model related to the embodiment, it is possible to manufacture a hot-rolled sheet that is precisely controlled at a coiling temperature that is closer to the target temperature.

〔Example〕
Examples will be described below to demonstrate the effects of the method for generating a temperature prediction model for a hot-rolled sheet, the method for generating a transformation enthalpy prediction model for a hot-rolled sheet, the coiling temperature prediction method for a hot-rolled sheet, the temperature control method for a hot-rolled sheet, and the manufacturing method for a hot-rolled sheet according to the present invention. In these examples, high tensile steel was manufactured in a hot rolling line. Note that the numerical values and the like described in these examples are presented for further understanding of the present invention, and the present invention is not limited by these examples.

The runout table in this embodiment has the same configuration as that shown in Figure 2. The length of the runout table is 130 m, and nine cooling zones, each 10 m long, are arranged. The three cooling zones from the finishing mill are called the "front stage", the next three cooling zones are called the "middle stage", and the last three cooling zones are called the "rear stage".

As in FIG. 2, a finishing mill exit thermometer is installed on the exit side of the finishing mill, a first intermediate thermometer is installed at a position about 1/3 of the length of the run-out table, a second intermediate thermometer is installed at a position about 2/3 of the length of the run-out table, and a winding thermometer 10 is installed at the rear end of the run-out table. Multiple descalers are installed in the conveying direction before the finishing mill, with descalers having injection pressures of 18 MPa, 30 MPa, and 50 MPa.

The following data was used as input when generating the prediction model: steel plate thickness: 1.8-4.5 mm; steel plate width: 800-1200 mm; steel plate components: C: 0.01-0.1%, Si: 0.1-1.0%, Mn: 0.5-2.0%; steel plate conveying speed: 8-12 m/s at the tip of the material; and descaler spray pressure before finish rolling: 18 MPa, 30 MPa, 50 MPa. Water was used as the coolant, the temperature of the coolant sprayed from the cooling device and the air temperature were 30°C, the target temperature at the exit of the finish rolling mill was 900°C, and the target coiling temperature was 550-650°C.

The input data used were the temperatures measured by the finishing rolling mill exit thermometer, the first intermediate thermometer, and the second intermediate thermometer. For the temperature of the steel plate, the temperature of the center of the steel plate in the width direction was measured using a spot-type radiation thermometer, and the results of measuring the temperature of the top surface of the steel plate were used.

In this way, the steel plate's composition, cooling device operating parameters, and temperature measurement information were used as input data, and the enthalpy information obtained by the above-mentioned method was used as output data to generate a transformation enthalpy prediction model using a neural network. Note that since the enthalpy information is subject to compositional variation, an approximation equation for the transformation temperature for each composition was created to take into account the compositional influence.

The representative components of the steel plate used to predict the coiling temperature using the transformation enthalpy prediction model were the same as those of the steel plate used to generate the transformation enthalpy prediction model, with a plate thickness of 3.2 mm and a plate width of 1,000 mm. The conveying speed of the steel plate was 10 m/s at the tip of the material, and the target coiling temperature was 600°C.

The transformation enthalpy of the steel sheet was predicted using a transformation enthalpy prediction model, and the coiling temperature was calculated from these predicted values by heat transfer calculations. The cooling pattern, particularly the amount of cooling water, was adjusted for each cooling zone so that the coiling temperature calculated in this way would match the target coiling temperature. As a result, the variation in coiling temperature within one coil was reduced, and it was possible to reduce it to within ±25°C of the target value.

The above describes in detail the method for generating a temperature prediction model for hot-rolled sheet, the method for generating a transformation enthalpy prediction model for hot-rolled sheet, the coiling temperature prediction method for hot-rolled sheet, the temperature control method for hot-rolled sheet, and the manufacturing method for hot-rolled sheet according to the present invention, using the form and examples for implementing the invention. However, the gist of the present invention is not limited to these descriptions, and must be interpreted broadly based on the claims. Needless to say, various changes and modifications based on these descriptions are also included in the gist of the present invention.

For example, in the method for generating a transformation enthalpy prediction model according to the embodiment, the transformation enthalpy of the steel sheet S is used as output data for learning, but when grasping the transformation behavior of the steel sheet S during cooling, data other than the transformation enthalpy may be used as long as the information changes due to the transformation of the steel sheet S. In addition, the method can be applied to materials other than the steel sheet S as long as the material undergoes transformation during the cooling process.

REFERENCE SIGNS LIST 1 heating furnace 2 roughing mill 3 finishing mill 4 run-out table 5 cooling device 6 winding machine 7 finishing mill outlet thermometer 8 first intermediate thermometer 9 second intermediate thermometer 10 winding thermometer 11 host computer 12 control device 13 database 100 temperature control device 121 cooling condition calculation unit 122 prediction model 123 cooling condition output unit 14 descaler S steel plate

Claims (8)

In the hot rolling line, the hot-rolled sheet after finish rolling is cooled on the run-out table and coiled up.
A temperature prediction model is generated for predicting the coiling temperature of the hot-rolled sheet using a plurality of teacher data, the teacher data being the components of the sheet, the operation parameters of the finishing rolling mill and the run-out table, and the actual data of the temperature of the hot-rolled sheet measured on the run-out table from the exit side of the finishing rolling mill, and the actual data of the coiling temperature of the hot-rolled sheet being output as data ;
The operation parameters of the finishing rolling mill and the run-out table include the plate thickness after the finish rolling, the conveying speed of the run-out table, the flow rate of the coolant of the run-out table, and the descaling injection performance data before the finish rolling,
The actual data of the temperature of the hot-rolled sheet includes the temperature of the hot-rolled sheet after finish rolling, measured at the time when the tip of the hot-rolled sheet reaches the winder;
A method for generating a temperature prediction model for hot-rolled strip. generating the temperature prediction model by one or more machine learning models selected from a neural network, a decision tree learning, a random forest, and a support vector regression;
The method for generating a temperature prediction model for a hot-rolled strip according to claim 1 . In the hot rolling line, the hot-rolled sheet after finish rolling is cooled on the run-out table and coiled up.
The components contained in the plate, the operation parameters of the finishing rolling mill and the run-out table, and the actual data of the temperature of the hot-rolled plate measured on the run-out table from the exit side of the finishing rolling mill are used as input data, and the transformation enthalpy of the hot-rolled plate calculated based on the actual data of the temperature of the hot-rolled plate is used as output data.
A transformation enthalpy prediction model is generated to predict the transformation enthalpy of the hot-rolled sheet ;
The operation parameters of the finishing rolling mill and the run-out table include the plate thickness after the finish rolling, the conveying speed of the run-out table, the flow rate of the coolant of the run-out table, and the descaling injection performance data before the finish rolling,
The actual data of the temperature of the hot-rolled sheet includes the temperature of the hot-rolled sheet after finish rolling and the coiling temperature of the hot-rolled sheet measured at the time when the tip of the hot-rolled sheet reaches the coiler.
A method for generating a transformation enthalpy prediction model for hot-rolled sheet. generating the transformation enthalpy prediction model by one or more machine learning models selected from a neural network, a decision tree learning, a random forest, and a support vector regression;
The method for generating a transformation enthalpy prediction model of a hot-rolled sheet according to claim 3 . The transformation enthalpy of the hot-rolled sheet is predicted using a transformation enthalpy prediction model generated by the method for generating a transformation enthalpy prediction model of the hot-rolled sheet according to claim 3 or 4 , and a coiling temperature of the hot-rolled sheet is calculated based on the predicted transformation enthalpy of the hot-rolled sheet.
A method for predicting the coiling temperature of hot-rolled strip. A coiling temperature of the hot-rolled sheet is predicted using a temperature prediction model generated by the method for generating a temperature prediction model of the hot-rolled sheet according to claim 1 or 2 , and an operation parameter of a run-out table is reset based on the predicted coiling temperature of the hot-rolled sheet.
A method for controlling the temperature of hot-rolled strip. The transformation enthalpy of the hot-rolled sheet is predicted using a transformation enthalpy prediction model generated by the method for generating a transformation enthalpy prediction model of the hot-rolled sheet according to claim 3 or 4 , and an operation parameter of a run-out table is reset based on a coiling temperature of the hot-rolled sheet calculated from the predicted transformation enthalpy of the hot-rolled sheet.
A method for controlling the temperature of hot-rolled strip. A method for manufacturing a hot-rolled sheet, comprising the steps of: resetting operational parameters of a run-out table by the method for controlling the temperature of a hot-rolled sheet according to claim 6 or claim 7 ; and cooling the hot-rolled sheet on the run-out table based on the reset operational parameters.

Images