JPH0985328A - Hot rolled steel sheet coiling temperature control device and its coiling temperature control method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To determine the appropriate amount of the cooling water even when the carrying speed of a hot rolled steel plate is changed. SOLUTION: The learning coefficient Cni corresponding to the elapsed time tn or the distance Ln is retrieved from a storage means by measuring the elapsed time tn or the distance Ln from the time point when a tip part of a hot rolled steel plate 2 passes a finish outlet side thermometer 7 to the time point when a part of each control length (n) passes the finish outlet side thermometer 7, and the water cooling quantity ΔTp of each control length (n) is operated based on the retrieved learning coefficient Cni .

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a hot-rolled steel sheet wound from a final stand of a rolling mill, which cools the hot-rolled steel sheet to a predetermined target temperature before being wound on a coiler. The present invention relates to a temperature control device and a winding temperature control method thereof.

[0002]

2. Description of the Related Art FIG. 12 is a block diagram showing a conventional coiling temperature controller for hot-rolled steel sheets. In the figure, 1 is the final stand of the finishing rolling mill, and 2 is the output from the final stand 1 of the finishing rolling mill. hot-rolled steel sheet, coiler 3 is for winding the hot-rolled steel sheet 2, the thickness meter for measuring the thickness H _F of the hot-rolled steel sheet 2 is 4, the motor 5 which drives the final stand 1 of the finishing mill, the 6 An arithmetic unit for measuring the rotation speed of the motor 5 to obtain the conveying speed V and the acceleration a of the hot-rolled steel sheet 2, and 7 is the finishing of each location in the longitudinal direction of the hot-rolled steel sheet 2 output from the final stand 1 of the finishing mill Finishing outlet thermometer for measuring outlet temperature FDT,
Reference numeral 8 is a winding thermometer for measuring the winding temperature CT immediately before the hot-rolled steel sheet 2 is wound around the coiler 3, and 9 is an initial data input device for setting a predetermined initial value for controlling the winding temperature of the hot-rolled steel sheet 2. Reference numeral 10 is a water injection on-off pattern (hereinafter referred to as a water injection pattern) based on the outputs of various sensors such as the finish side thermometer 7 and the initial values set in the initial data input device 9.
11 is a valve control device for controlling a cooling device for injecting cooling water into the hot-rolled steel plate 2 according to a water injection pattern, 1U, 2U, 3U ... NU is an upper bank of the cooling device, 1L, 2L, 3L ... NL are lower banks of the cooling device.

Next, the operation will be described. The hot-rolled steel sheet 2 output from the final stand 1 of the finish rolling mill is conveyed on the illustrated hot run table and wound around the coiler 3. The strength and workability of the hot-rolled steel sheet 2 wound around the coiler 3, etc. In order to ensure the quality of the hot rolled steel sheet 2, it is necessary to cool the temperature of the hot rolled steel sheet 2 to a predetermined target winding temperature CT ^* before the hot rolled steel sheet 2 is wound around the coiler 3.

Therefore, the hot-rolled steel sheet 2 is divided into a plurality (N pieces) (for example, the length of each divided portion is equal to the length of each bank of the cooling device. (The part is called the control length), and the water injection pattern is determined for each control length to control each bank of the cooling device.

Specifically, first, when the hot-rolled steel plate 2 is output from the final stand 1 of the finish rolling mill, the finish-side thermometer 7 and the plate thickness gauge 4 cause the control lengths n of the hot-rolled steel plate 2 to be controlled. (N = 1,
The finish-out temperature FDT and the sheet thickness H _F of the hot-rolled steel sheet 2 are measured for each of 2, 3, ... N). Also, the computing unit 6
However, the conveyance speed V and the acceleration a when each control length n of the hot-rolled steel sheet 2 passes through the finishing-side thermometer 7 are obtained by measuring the rotation speed of the motor 5.

When the conveying speed V of the hot-rolled steel sheet 2 and the like are obtained, the computer 10 reaches the take-up thermometer 8 after each control length n of the hot-rolled steel sheet 2 has passed through the finish-side thermometer 7. The required time t until it is calculated is calculated as shown below. However, the distance L from the finish-side thermometer 7 to the take-up thermometer 8 is determined at the time of designing the rolling mill, and thus is input from the initial data input device 9 as an initial value.

[0007]

[Equation 1]

When the required time t is calculated, the computer 10 calculates the air cooling temperature drop amount ΔT _air as shown below based on the required time t and the finishing outlet temperature FDT for each control length n. Calculate However, since the emissivity ε, the specific heat C _P and the specific gravity γ in the formula (2) are determined at the time of designing the rolling mill, they are input from the initial data input device 9 as initial values.

[0009]

[Equation 2]

Then, when the air-cooling temperature drop amount ΔT _air is calculated, the computer 10 subtracts the target winding temperature CT ^* and the air-cooling temperature drop amount ΔT _air from the finish outlet temperature FDT for each control length n. Then, the water cooling temperature drop amount ΔT _water is calculated. ΔT _water = FDT-CT ^* -ΔT _air (3)

[0011] Then, the computer 10 has calculated the water cooling temperature drop [Delta] T _water, the water cooling temperature drop [Delta] T _water of each control length n, by multiplying a preset learning coefficient C _i, each control The water cooling amount ΔTp for the length n is calculated. ΔTp = C _i · ΔT _water (4)

Here, the learning coefficient C _i is the ratio of the actual water cooling temperature drop amount to the predicted water cooling temperature drop amount, and is a coefficient common to each control length n. In this way, the reason for multiplying the water cooling temperature drop amount ΔT _water by the learning coefficient C _i is that when there is an error between the predicted water cooling temperature drop amount and the actual water cooling temperature drop amount, they are conveyed next time. This is because the same degree of error is likely to occur in cooling the hot-rolled steel sheet 2, and it is necessary to absorb such an error. For example, if the predicted water cooling temperature drop amount is 100 ° C. and the actual water cooling temperature drop amount is 120 ° C., an error of 20 ° C. has occurred. Therefore, it is necessary to absorb the error of 20 ° C. is there. Therefore, the learning coefficient used for cooling the hot-rolled steel sheet 2 conveyed next time is C _{i + 1} = 120/100 = 1.2.
(For the sake of simplicity, here, the learning coefficient C _i used for the cooling this time is not considered), and when the water cooling temperature drop amount of the hot rolled steel sheet 2 to be conveyed next time is predicted to be 100 ° C. The water cooling amount is ΔTp = 1.2 × 100 =
It becomes 120 ° C. Therefore, even if the predicted water cooling temperature drop amount is 100 ° C., the water injection pattern for lowering the temperature of the hot-rolled steel sheet 2 by 120 ° C. is determined, and 20 ° C.
The error of can be absorbed.

When the water cooling amount ΔTp for each control length n is calculated, the computer 10 determines the water injection pattern based on the water cooling amount ΔTp for each control length n (the number of banks to be used, The number N of nozzles actually used among the nozzles mounted in the bank to be used is determined). For example, the water cooling amount ΔTp is 100 ° C
Then, assuming that the maximum cooling capacity of each bank (cooling capacity when all mounted nozzles are used) is 40 ° C, the maximum cooling capacity of two banks is 80 ° C.
The water injection amount is determined, and the half of the nozzles in one bank are used to determine the remaining water cooling amount of 20 ° C. In addition, the following equation (5)
Is a calculation formula for determining the number of nozzles to be used. However, the number of mounted nozzles N in each bank in equation (5) is N
₀ , the cooling bank length M, and the heat flux coefficient f are determined at the time of designing the rolling mill, and thus are input as initial values from the initial data input device 9.

[0014]

(Equation 3)

When the water injection pattern is determined by the computer 10, the valve controller 11 controls the valves of each bank according to the water injection pattern, and the hot rolled steel sheet 2
Cooling water is poured into. As a result, the hot-rolled steel sheet 2 is cooled to the target winding temperature CT ^* and wound around the coiler 3, but the winding temperature CT of the hot-rolled steel sheet 2 does not always match the target winding temperature CT ^*. There may be an error, and if the learning coefficient C _i used for the current cooling is also used as it is for cooling the hot-rolled steel sheet 2 that is conveyed next time, there is a high possibility that an error of the same degree will occur. Therefore, in order to absorb such an error, it is necessary to update the learning coefficient C _i based on the current cooling result.

The update of the learning coefficient C _i will be described below. First, the learning coefficient C _i is, as described above, the hot rolled steel sheet 2
Is a coefficient common to the control lengths n, the control length k is focused on one arbitrary control length when the learning coefficient C _i is updated (for convenience of description, the control length k will be focused hereinafter).
The learning coefficient C _i is updated based on the data related to. Specifically, when the hot-rolled steel sheet 2 is cooled by the cooling device as described above, the coiling thermometer 8 measures the coiling temperature CT of the control length k immediately before being coiled by the coiler 3, and the computer 10 actually measures the actual value t _res of the time required for the control length k to reach the take-up thermometer 8 after passing through the finishing outlet thermometer 7.

When the actual value t _res of the required time is measured, the computer 10 determines the actual value ΔT of the air cooling temperature drop amount of the control length k based on the actual value t _{res of the} required time.
Calculate _{air- res} .

[0018]

[Equation 4]

Then, the computer 10 calculates the actual value ΔT _air-res of the air cooling temperature drop amount of the control length k from the actual finish side temperature FDT of the control length k to the actual winding temperature CT and the air cooling temperature drop. by subtracting the amount of actual values [Delta] T _air-res calculating the actual value ΔT _{wa ter-res} water cooling temperature drop. ΔT _water-res = FDT-CT-ΔT _air-res ... (7)

The computer 10 also includes a valve controller 11
Calculates the water cooling temperature drop amount ΔT _water-ant for the control length k, which is predicted by calculation from the water injection pattern used for cooling the control length k (the number N of nozzles used).

[0021]

(Equation 5)

In this way, when the predicted value ΔT _water-ant and the actual value ΔT _water-res of the water cooling temperature drop amount of the control length k are calculated, the computer 10 calculates the predicted value ΔT _water-ant.
From the actual value ΔT _water-res , the learning coefficient C _{i + 1} used for cooling the hot rolled steel sheet 2 to be conveyed next time is calculated. C _{i + 1} = ΔT _water-res / ΔT _water-ant (9)

Since the learning coefficient C _{i + 1} obtained from the equation (9) does not consider the learning coefficient C _i used this time, the computer 10 considers the learning coefficient C _i used this time, In order to further improve the accuracy of the learning coefficient, the learning coefficient C _{i + 1} or the like is multiplied by an appropriate weighting coefficient α as shown below,
The learning coefficient C _{i + 1} is modified. C _{i + 1} ← (1-α) C _i + α · C _{i + 1} (10)

Incidentally, the algorithm for updating the learning coefficient described above is disclosed in Japanese Patent Publication No. 58-47924.

[0025]

Since the conventional hot-rolled steel sheet winding temperature control device is constructed as described above, it is relatively accurate when the hot-rolled steel sheet 2 is always conveyed at a constant speed. Although the winding temperature can be controlled well, when winding the hot-rolled steel sheet 2 in the coiler 3, the hot-rolled steel sheet 2 is slowly wound at the front end and the end portion of the hot-rolled steel sheet 2 while being wound at a high speed in the middle portion. The required time t from passing through the finishing outlet thermometer 7 to reaching the winding thermometer 8 is different for each control length n of 2. For this reason, since the air cooling temperature drop amount ΔT _air is different for each control length n, the required water cooling amount ΔTp is calculated for each control length n, but the transport speed V of the hot-rolled steel sheet 2 (winding of the coiler 3 is calculated. Since the error of the winding temperature slightly changes for each control length n with the change of the take-up speed),
Even if the common learning coefficient C _i is used for each control length n, the error cannot be completely absorbed, and there is a problem that an appropriate water cooling amount ΔTp cannot be determined for each control length n.

The present invention has been made in order to solve the above problems. A hot-rolled steel sheet capable of determining an appropriate water cooling amount for each control length even if the hot-rolled steel sheet conveyance speed changes. An object of the present invention is to obtain a winding temperature control device and a winding temperature control method thereof.

[0027]

According to a first aspect of the present invention, there is provided a coiling temperature control device for a hot rolled steel sheet, wherein each of the hot rolled steel sheet is controlled from the time when the tip of the hot rolled steel sheet passes a predetermined reference point. The elapsed time until the point passes the reference point is measured, the learning coefficient corresponding to each elapsed time is referred from the storage means, and each learning coefficient is multiplied by the water cooling temperature drop amount at each point. Water cooling amount calculation means for calculating the amount of water cooling at a location is provided.

A hot-rolled steel sheet winding temperature control device according to a second aspect of the present invention measures the distance from the tip of the hot-rolled steel sheet to each location, and stores a learning coefficient corresponding to each distance from the storage means. For reference, a water cooling amount calculation means for calculating each water cooling amount at each position by multiplying each learning coefficient by the water cooling temperature drop amount at each position is provided.

According to a third aspect of the present invention, there is provided a hot-rolled steel sheet winding temperature control device, in which a drop amount calculation means for calculating an actual water cooling temperature drop amount at each location and a water injection pattern determined by a water injection pattern determination means. A drop amount predicting unit that predicts the water cooling temperature drop amount at each location from the pattern is provided, and the learning coefficient is calculated by dividing the actual water cooling temperature drop amount by the predicted water cooling temperature drop amount for each location. It is the one.

In the coiling temperature control device for a hot rolled steel sheet according to the invention of claim 4, when the learning coefficient stored in the storage means is updated, the elapsed time after passing the reference point or the tip end portion Updating means that divides the distances into a plurality of sections, distributes the learning coefficients obtained by calculation to corresponding sections, obtains the average value of the learning coefficients of each section, and updates the average value as the learning coefficient of each section. Is provided.

In the coiling temperature control device for a hot-rolled steel sheet according to the invention of claim 5, when updating the learning coefficient stored in the storage means, the elapsed time after passing the reference point or the tip end portion The distance of is divided into a plurality of sections, and the learning coefficient obtained by calculation is distributed to the corresponding section to obtain the linear approximation formula of the learning coefficient of each section, and the change pattern of the linear approximation is used as the learning coefficient of each section. The updating means for updating is provided.

In the winding temperature control method for a hot-rolled steel sheet according to a sixth aspect of the present invention, from the time when the tip of the hot-rolled steel sheet passes a predetermined reference point on the line, each point of the hot-rolled steel sheet becomes While measuring the elapsed time until passing the reference point, the learning coefficient corresponding to each measured elapsed time is searched from among the learning coefficients set in advance according to the elapsed time, and each searched learning coefficient Is multiplied by the water cooling temperature drop amount at each location to calculate the water cooling amount at each location.

According to a seventh aspect of the present invention, there is provided a method for controlling a winding temperature of a hot-rolled steel sheet, which measures a distance from a tip portion of the hot-rolled steel sheet to each position, and a learning coefficient preset according to the distance. Among them, the learning coefficient corresponding to each of the measured distances is searched, and the calculated learning coefficient is multiplied by the water cooling temperature drop amount at each location to calculate the water cooling amount at each location. Is.

In the winding temperature control method for hot rolled steel sheet according to the present invention, the actual water cooling temperature drop amount at each location is calculated, while the water cooling temperature drop amount at each location is predicted from the water injection pattern. Then, the learning coefficient is obtained by dividing the actual water cooling temperature drop amount at each location by the predicted water cooling temperature drop amount.

In the winding temperature control method for a hot rolled steel sheet according to a ninth aspect of the present invention, when the preset learning coefficient is updated, the elapsed time after passing the reference point or the distance from the tip portion. Is divided into a plurality of sections, and the learning coefficient calculated is distributed to the corresponding section to obtain the average value of the learning coefficient of each section, and the average value is updated as the learning coefficient of each section. Is.

According to a tenth aspect of the present invention, there is provided a method for controlling a coiling temperature of a hot-rolled steel sheet, wherein when a preset learning coefficient is updated, an elapsed time after passing a reference point or a distance from a tip portion. Is divided into a plurality of sections, and the learning coefficient obtained by calculation is divided into corresponding sections to obtain a linear approximation formula of the learning coefficient of each section, and the change pattern of the linear approximation is updated as the learning coefficient of each section. It was done like this.

[0037]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below. Embodiment 1. 1 is a block diagram showing a coiling temperature control device for hot-rolled steel sheets according to Embodiment 1 of the present invention, in which 1 is a final stand of a finishing rolling mill, and 2 is an output from a final stand 1 of the finishing rolling mill. has been hot-rolled steel sheet, coiler 3 is for winding the hot-rolled steel sheet 2, 4 thickness gauge for measuring the thickness H _F of the hot-rolled steel sheet 2, 5 drives the final stand 1 of the finishing mill motor, 6 Is a calculator (measuring means) for measuring the rotation speed of the motor 5 to obtain the conveying speed V and the acceleration a of the hot-rolled steel sheet 2, and 7 is the longitudinal direction of the hot-rolled steel sheet 2 output from the final stand 1 of the finishing rolling mill. , A finish-side thermometer (measuring means) for measuring the finish-side temperature FDT at each position (control length n), and 8 is a winding for measuring the winding temperature CT immediately before the hot-rolled steel sheet 2 is wound around the coiler 3. Intake thermometer (measurement means), 9
Is an initial data input device for setting a predetermined initial value relating to the winding temperature control of the hot-rolled steel sheet 2, and 20 is an output of various sensors such as the finishing side thermometer 7 and an initial value set in the initial data input device 9. Calculator that determines water injection pattern based on 1
1 is a valve control device (control means) for controlling a cooling device for injecting cooling water into a hot-rolled steel plate 2 according to a water injection pattern, 1U, 2U, 3U ... NU is an upper bank of the cooling device, 1L, 2L, 3L ... NL is the lower bank of the cooling device.

FIG. 2 is a configuration diagram showing a detailed configuration of the computer 20 according to the first embodiment. In FIG.
1 is the finish outlet temperature F measured by the finish outlet thermometer 7.
The air cooling temperature drop amount ΔT _air for each control length n is estimated based on the transport speed V calculated by the DT and the calculator 6, and the finish side temperature FDT and the air cooling temperature drop amount ΔT _air are estimated.
Estimating means for estimating the water cooling temperature drop amount ΔT _water of each control length n from T _air and the target winding temperature CT ^* , and 22 is the end of the hot-rolled steel plate 2 at the finish side thermometer 7 (predetermined on the line). It is a storage unit that stores the learning coefficient C _ni set according to the elapsed time t _n after passing the reference point).

Reference numeral 23 indicates the elapsed time t _n from the time when the tip of the hot-rolled steel sheet 2 passes the finishing outlet side thermometer 7 until each control length n passes the finishing outlet side thermometer 7, The learning coefficient C _ni corresponding to each elapsed time t _n is referred to from the storage means 22, and each learning coefficient C _ni is the water cooling temperature drop Δ of each control length n.
T _water is multiplied by each water cooling amount ΔT for each control length n
A water cooling amount calculation means for calculating p, and 24 is a water injection pattern from the water cooling amount ΔTp of each control length n calculated by the water cooling amount calculation means 23 and the transport speed V of each control length n calculated by the calculator 6. Is a means for determining the water injection pattern.

Numeral 25 is a finish side thermometer 7 for each control length n.
The measuring means for measuring the required time t _res from the measurement of the finish outlet temperature FDT to the measurement of the winding temperature by the take-up thermometer 8, 26 is the finish outlet temperature FDT and the required time of each control length n. based on the actual values t _res well as calculating the actual value [Delta] t _air-res empty cooling temperature drop of each control length n of finishing delivery temperature FDT of each control length n, air cooling temperature drop [Delta] t _{air -res} and winding temperature CT to control length n
A drop amount calculation means for calculating the actual value ΔT _water-res of the water cooling temperature drop amount of 27, and a water cooling temperature drop amount ΔT _water-ant of each control length n from the water injection pattern determined by the water injection pattern determination means 24. Predicting amount of descent amount, 28
Is the actual value ΔT of the water cooling temperature drop for each control length n
The _air-res divided by the predicted value [Delta] T _water-ant water cooling temperature drop seek learning coefficient C _ni, an update means for updating the learning coefficient C _ni in the storage unit 22 are stored. By the way,
FIG. 3 is a flowchart showing an algorithm for controlling the cooling device by the coiling temperature controller for the hot rolled steel sheet, and FIG. 4 is a flowchart showing an algorithm for updating the learning coefficient by the coiling temperature controller for the hot rolled steel sheet.

Next, the operation will be described. The hot-rolled steel sheet 2 output from the final stand 1 of the finish rolling mill is conveyed on the illustrated hot run table and wound around the coiler 3. The strength and workability of the hot-rolled steel sheet 2 wound around the coiler 3, etc. In order to ensure the quality of the hot rolled steel sheet 2, it is necessary to cool the temperature of the hot rolled steel sheet 2 to a predetermined target winding temperature CT ^* before the hot rolled steel sheet 2 is wound around the coiler 3.

Therefore, the hot-rolled steel plate 2 is divided into a plurality (N pieces) (for example, the length of each divided portion is equal to the length of each bank of the cooling device. (The part is called a control length), and a water injection pattern is determined for each control length n to control each bank of the cooling device.

Specifically, first, when the hot-rolled steel plate 2 is output from the final stand 1 of the finish rolling mill, the finish outlet side thermometer 7 and the plate thickness gauge 4 are respectively hot-rolled for each control length n. The finish outlet temperature FDT and the plate thickness H _F of the steel plate 2 are measured (step ST1). Further, the computing unit 6 obtains the transport speed V and the acceleration a when each control length n of the hot-rolled steel sheet 2 passes through the finishing outlet thermometer 7 by measuring the rotation speed of the motor 5 (step ST2). ).

When the conveying speed V of the hot-rolled steel sheet 2 and the like are obtained, the estimation means 21 of the computer 20 causes the estimation temperature 21 of the hot-rolled steel sheet 2 to pass after the control lengths n of the hot-rolled steel sheet 2 have passed through the finishing outlet thermometer 7 and the coiling temperature. 8 in total
The time t required to reach (1) is calculated as shown below (step ST3). However, the distance L from the finish-side thermometer 7 to the take-up thermometer 8 is determined at the time of designing the rolling mill, and thus is input from the initial data input device 9 as an initial value.

[0045]

(Equation 6)

After calculating the required time t, the estimating means 21 of the computer 20 performs the following calculation for each control length n based on the required time t and the finish outlet temperature FDT to obtain the air cooling temperature. The amount of fall ΔT _air is estimated (step ST4). However, the emissivity ε and the specific heat C _P in the equation (12) are
Since the specific gravity γ and the specific gravity γ are determined at the time of designing the rolling mill, they are input as initial values from the initial data input device 9.

[0047]

(Equation 7)

Then, the estimating means 21 of the computer 20 estimates the air cooling temperature drop amount ΔT _air , and then, for each control length n, from the finish outlet temperature FDT to the target winding temperature CT ^* and the air cooling temperature drop amount ΔT. Water cooling temperature drop Δ by subtracting _air
Estimate T _water (step ST4). ΔT _water = FDT-CT ^* -ΔT _air (13)

Then, when the water cooling temperature drop amount ΔT _water is estimated, the water cooling amount calculating means 23 of the computer 20 starts each time the tip of the hot-rolled steel sheet 2 passes through the finishing outlet thermometer 7. The elapsed time t _n until the control length n passes through the finishing outlet side thermometer 7 is measured (step ST5). When measuring the elapsed time t _n of the respective control length n, as shown in FIG. 5, corresponding to the elapsed time t _n the learning coefficient C _ni is the storage means 2
2, the learning coefficient C _ni corresponding to the elapsed time t _n of each control length n is retrieved from the storage unit 22 (step ST6). Incidentally, FIG. 5 shows that when the elapsed time t _n is t ₀ , the value of C _0i is acquired as the learning coefficient C _ni .

Then, the water cooling amount calculation means 2 of the computer 20
3 is a water cooling temperature drop amount ΔT _water for each control length n.
Is multiplied by each retrieved learning coefficient C _ni to calculate the water cooling amount ΔTp for each control length n (step ST7). ΔTp = C _ni · ΔT _water (14)

Here, the learning coefficient C _ni is the ratio of the actual water cooling temperature drop amount to the predicted water cooling temperature drop amount, and is a coefficient corresponding to the elapsed time t _n . in this way,
The reason for multiplying the water cooling temperature drop amount ΔT _water by the learning coefficient C _ni is that when there is an error between the predicted water cooling temperature drop amount and the actual water cooling temperature drop amount, the hot rolled steel sheet 2 to be conveyed next time This is because there is a high possibility that the same degree of error will occur in the case of cooling, and it is necessary to absorb such an error. For example, if the predicted water cooling temperature drop is 100 ° C,
If the actual water cooling temperature drop is 120 ° C, 20
Since an error of ° C has occurred, it is necessary to absorb the error of 20 ° C. Therefore, although the details will be described later, the learning coefficient used for cooling the hot rolled steel sheet 2 to be conveyed next time is C.
_{n (i + 1)} = 120/100 = 1.2 (for simplicity of explanation, the learning coefficient C _ni used for the cooling this time is not considered), and the hot rolled steel sheet to be conveyed next time When the water cooling temperature drop amount of 2 is predicted to be 100 ° C., the water cooling amount is ΔTp = 1.2 × 100 = 120 ° C. Therefore, even if the predicted water cooling temperature drop is 100 ° C,
The water injection pattern for lowering the temperature of the hot-rolled steel sheet 2 by 120 ° C. is determined, and the error of 20 ° C. can be absorbed.

When the water cooling amount ΔTp for each control length n is calculated, the water injection pattern determining means 24 of the computer 20 determines the water injection pattern for each control length n based on the water cooling amount ΔTp. (The number of banks used,
Among the nozzles installed in the bank used,
Determine the number N of nozzles actually used). For example,
If the water cooling amount ΔTp is 100 ° C and the maximum cooling capacity of each bank (cooling capacity when all mounted nozzles are used) is 40 ° C, it is possible to use two banks with the maximum cooling capacity of 80 The water cooling amount of ℃ is secured, and by using half the nozzles of one bank, the remaining 2
A water injection pattern such as ensuring a water cooling amount of 0 ° C. is determined (step ST8). The formula (15) shown below is
It is a calculation formula for determining the number of nozzles to be used.
However, the number of mounted nozzles N ₀ in each bank in Expression (15),
Since the cooling bank length M and the heat flux coefficient f are determined at the time of designing the rolling mill, they are input as initial values from the initial data input device 9.

[0053]

(Equation 8)

When the water injection pattern determining means 24 of the computer 10 determines the water injection pattern, the valve control device 11 controls the valves of each bank in accordance with the water injection pattern to inject the cooling water into the hot rolled steel sheet 2. Yes (step ST9). As a result, the hot-rolled steel sheet 2 is cooled to the target winding temperature CT ^* and wound around the coiler 3, but the winding temperature CT of the hot-rolled steel sheet 2 does not always match the target winding temperature CT ^*. An error may occur, and if the learning coefficient C _ni used for the current cooling is also used as it is for cooling the hot-rolled steel sheet 2 that is conveyed next time, there is a high possibility that the same degree of error will occur. Therefore, in order to absorb such an error, it is necessary to update the learning coefficient C _ni based on the current cooling result.

The updating of the learning coefficient C _ni will be described below. First, when the hot-rolled steel sheet 2 is cooled by the cooling device as described above, the coiling thermometer 8 measures the coiling temperature CT of each control length n immediately before being coiled by the coiler 3, and at the same time, the computer 20 calculates the coiling temperature. The measuring means 25 actually determines each control length n.
Measures the actual value t _res of the time required from reaching the take-up thermometer 8 after passing through the finishing outlet thermometer 7 (step ST11).

When the actual value t _res of the required time is measured, the drop amount calculating means 26 of the computer 20 determines the air cooling temperature drop of each control length n based on the actual value t _{res of the} required time. The actual value ΔT _air-res of the amount is calculated (step ST12).

[0057]

[Equation 9]

Then, the descent amount calculating means 26 of the computer 20
Is the actual value ΔT of the air-cooling temperature drop amount for each control length n.
_{When air-res} is calculated, the actual winding temperature CT and the actual value ΔT of the _air cooling temperature drop amount are subtracted from the actual finish-side temperature FDT of each control length n, and the actual value ΔT of the water cooling temperature drop amount is subtracted from _air-res. Calculate _water-res (step ST12). ΔT _water-res = FDT-CT-ΔT _air-res (17)

Further, the descent amount prediction means 27 of the computer 20
Calculates the water cooling temperature drop amount ΔT _water-ant, which is computationally predicted from the water injection pattern (the number N of nozzles used) used for cooling by the valve control device 11 for each control length n (step ST13). .

[0060]

(Equation 10)

In this way, the predicted value ΔT _water-ant and the actual value ΔT _water-res of the water cooling temperature drop amount of each control length n are obtained.
Is calculated, the updating unit 28 of the computer 20 uses the predicted value ΔT _water-ant and the actual value ΔT _water-res to learn coefficient C used for cooling the hot-rolled steel sheet 2 to be conveyed next time.
Calculate _{n (i + 1)} (step ST14). Cn _{(i + 1)} = ΔT _water-res / ΔT _water-ant (19)

The updating means 28 of the computer 20 does not consider the learning coefficient C _ni used this time in the learning coefficient C _{n (i + 1)} obtained from the equation (19), so that the learning coefficient C used this time is used. _In order to further improve the accuracy of the learning coefficient in consideration of _ni , an appropriate weighting coefficient α is set to the learning coefficient C as shown below.
_The learning coefficient C _{n (i + 1)} is corrected by multiplying _{n (i + 1)} or the like. C _{n (i + 1)} ← (1-α) C _ni + α · C _{n (i + 1)} ... (20)

When the updating means 28 of the computer 20 corrects the learning coefficient C _{n (i + 1)} , the elapsed time t _n after each control length n has passed through the finishing outlet thermometer 7 is shown in FIG. As shown in FIG. 5, the learning coefficient C _{n (i + 1)} is obtained by dividing into a plurality of sections (7 sections in the case of FIG. 5) and obtained for each control length n.
Are divided into corresponding sections, and the average value of the learning coefficient in each section is obtained (step ST15). For example, in the case where the section 1 is from 0 second to 10 seconds, if the control length 1 to the control length 3 have passed through the finishing outlet thermometer 7 in the range from 0 second to 10 seconds, at least the following three Learning coefficient C _{1 (i + 1)} ,
Since C _{2 (i + 1)} and C _{3 (i + 1)} belong to the interval 1, (3 × (d) learning coefficients are obtained for each control length n). This means that d learning coefficients belong to the interval 1. The number of learning coefficients to be obtained may differ for each control length n), and at least the average of the three learning coefficients C _{n (i + 1)} . Find the value.

Then, the updating means 28 of the computer 20 obtains the average value of the learning coefficients C _{n (i + 1)} belonging to the section for each section, and updates the average value as the learning coefficient of each section. (Step ST16), a series of processing ends.

As described above, according to the first embodiment, from the time when the tip of the hot-rolled steel sheet 2 passes the finish-side thermometer 7, each control length n of the hot-rolled steel sheet 2 is changed to the finish-side temperature. The elapsed time t _n until passing the total 7 is measured, and each elapsed time t
_The learning coefficient C _ni corresponding to _n is referred from the storage means 22,
Since the water cooling amount ΔTp for each control length n is calculated from each learning coefficient C _ni , even if the transport speed V of the hot-rolled steel sheet 2 changes, an appropriate water cooling amount ΔTp for each control length n can be obtained. It can be determined, and as a result, the coiling temperature after cooling can be matched with the target coiling temperature with extremely high accuracy. That is, since the winding speed of the coiler 3 is different between the front end portion, the intermediate portion, and the end portion of the hot-rolled steel sheet 2, the error generated in each portion of the hot-rolled steel sheet 2 is different, but the change pattern of the winding speed is Since the hot-rolled steel sheet 2 does not change significantly and is substantially constant, the learning coefficient C _ni corresponding to each part of the hot-rolled steel sheet 2 is used, that is, the learning coefficient C corresponding to each elapsed time t _{n. When ni} is used, learning control applied to the change pattern of the winding speed becomes possible. Therefore, even if the transport speed V of the hot-rolled steel sheet 2 (winding speed of the coiler 3) changes, each control length n A proper water cooling amount ΔTp can be determined.

Further, according to the first embodiment, each control length n
For each time, the learning coefficient C _{n (i + 1)} is obtained by dividing the actual water cooling temperature drop ΔT _water-res by the predicted water cooling temperature drop ΔT _water-ant. The learning coefficient C _{n (i + 1)} used for cooling the hot rolled steel sheet 2 can be updated so as to correspond to the elapsed time t _n , and each control is performed also in the cooling of the hot rolled steel sheet 2 that is conveyed next time. An appropriate amount of water cooling ΔTp can be determined for each length n.

Further, according to the first embodiment, the obtained learning coefficient C _{n (i + 1)} is distributed to the corresponding section, and the average value of the learning coefficient C _{n (i + 1)} in each section is calculated, Since the average value is updated as the learning coefficient of each section, the learning coefficient corresponding to the elapsed time t _n can be maintained even if the number of learning coefficients C _{n (i + 1)} to be obtained is small.

Embodiment 2 FIG. 6 is a configuration diagram showing a detailed configuration of the computer 20 according to the second embodiment. In the figure, 31 stores the learning coefficient C _ni set according to the distance L _n from the tip of the hot-rolled steel sheet 2. The storage means 32 measures the distance L _n from the tip of the hot-rolled steel sheet 2 to each control length n, and refers to the learning coefficient C _ni corresponding to each distance L _n from the storage means 31 to learn each learning coefficient. This is a water cooling amount calculation means for multiplying C _ni by the water cooling temperature drop amount ΔT _water of each control length n to calculate the water cooling amount ΔTp of each control length n. Incidentally, FIG. 7 is a flowchart showing an algorithm for controlling the cooling device by the coiling temperature control device for the hot rolled steel sheet.

Next, the operation will be described. Storage means 31
The storage unit 31 and the water cooling amount calculating unit 3 are the same as those in the first embodiment except the water cooling amount calculating unit 32.
Only 2 will be described. When the water cooling temperature drop amount ΔT _water is estimated by the estimating means 21 of the computer 20 in the same manner as in the first embodiment, the water cooling amount calculating means 32 of the computer 20 calculates each from the tip of the hot-rolled steel sheet 2. The distance L _n to the control length n is measured (step ST21). Specifically, when determining the water injection pattern for each control length n, the length of each control length n can be recognized when the hot-rolled steel plate 2 is divided into a plurality of parts. Hot rolled steel plate 2
The distance L _n from the tip of the control length n to each control length _n is calculated.

Then, the water cooling amount calculation means 3 of the computer 20
When the distance L _n to each control length n is measured, the learning coefficient C _ni corresponding to the distance L _n is stored in the storage means 31 as shown in FIG. The learning coefficient C _ni corresponding to L _n is retrieved from the storage means 31 (step ST22). Incidentally, FIG. 8 shows that when the distance L _n is L ₀ , the value of C _0i is acquired as the learning coefficient C _ni .

The water cooling amount calculation means 3 of the computer 20
2 is a water cooling temperature drop amount ΔT _water for each control length n
Is multiplied by each retrieved learning coefficient C _ni to calculate the water cooling amount ΔTp for each control length n (step ST2).
3). ΔTp = C _ni · ΔT _water (21)

The subsequent processing is the same as that of the above-described first embodiment, and therefore its explanation is omitted. However, in the second embodiment, since the learning coefficient C _ni corresponding to the distance L _n is stored in the storage means 31, when the learning coefficient is updated, the distance L _{n is set} to a plurality of _values instead of the elapsed time t _n . The learning coefficient C _{n (i + 1)} is divided into sections (in FIG. 8, it is divided into 7 sections), and the learning coefficient C _{n (i + 1)} obtained for each control length n is distributed to the corresponding section to obtain the learning coefficient of each section. An average value is obtained (step ST15 in FIG. 4).

As described above, according to the second embodiment, the distance L _n from the tip of the hot rolled steel sheet 2 to each control length n of the hot rolled steel sheet 2 is measured and corresponds to each distance L _n . Learning coefficient C
_Since the water cooling amount ΔTp of each control length n is calculated from each learning coefficient C _ni by referring to _ni from the storage means 31, even if the transport speed V of the hot-rolled steel sheet 2 changes, each control length n. An appropriate water cooling amount ΔTp can be determined for each of the conditions, and as a result, the coiling temperature after cooling can be matched with the target coiling temperature extremely accurately. That is, since the winding speed of the coiler 3 is different between the front end portion, the intermediate portion, and the end portion of the hot-rolled steel sheet 2, the error generated in each portion of the hot-rolled steel sheet 2 is different, but the change pattern of the winding speed is Since it does not significantly change for each hot-rolled steel sheet 2 and is substantially constant, the learning coefficient C _ni corresponding to each part of the hot-rolled steel sheet 2 is used, that is, the learning coefficient C _ni corresponding to each distance L _n. By using, the learning control applied to the change pattern of the winding speed becomes possible. Therefore, even if the transport speed V of the hot-rolled steel sheet 2 (the winding speed of the coiler 3) changes, each control length n An appropriate water cooling amount ΔTp can be determined.

Embodiment 3 In the first and second embodiments, when the learning coefficient is updated, the average value of the learning coefficient is obtained for each section, and the average value is updated as the learning coefficient for each section. A linear approximation formula for the learning coefficient may be obtained, and the change pattern of the linear approximation may be updated as the learning coefficient for each section.

First, for convenience of explanation, as shown in FIG.
It is assumed that S learning coefficients have been obtained for each section.
In this case, the linear approximation formula between each point can be described as follows (In the formula below, t _j was used, but L
_j may be used). C _j = a · t _j + b (22) where C _j : learning coefficient belonging to the section t _j : elapsed time L _j : distance a, b: unknown number

Therefore, if the unknowns a and b are obtained, the above linear approximation formula can be specified. However, there are S formulas for each of the unknowns a and b, and in the simultaneous equation of two elements, the unknown a , B cannot be obtained, the unknowns a and b are determined using regression calculation. That is, the unknowns a and b are determined so that the sum of squares of the error is minimized by the least squares method of the S expressions. Specifically, a and b that minimize the following formula are obtained.

[0077]

[Equation 11]

Then, as shown below from the equation (23), when the unknowns a and b are obtained for each section, the unknowns a and b are obtained.
By substituting values into equation (22), the elapsed time t _j (or distance L _j) corresponding to the search of the learning coefficient C _j, the learning coefficient C _j for cooling the hot-rolled steel sheet 2 conveyed next The learning coefficient to be used is updated (see FIGS. 10 and 11).

[0079]

(Equation 12)

As described above, according to the third embodiment, the linear approximation formula of the learning coefficient C _j is obtained for each section, and the change pattern of the linear approximation is updated as the learning coefficient of each section. The learning coefficient can be continuously changed over the adjacent sections, and therefore, learning that is more suitable for the change pattern of the transport speed V than the case where the learning coefficient is changed stepwise as in the first and second embodiments. As a result, the winding temperature can be controlled with higher accuracy.

[0081]

As described above, according to the invention of claim 1, from the time when the tip of the hot-rolled steel sheet passes a predetermined reference point to the time when each point of the hot-rolled steel sheet passes the reference point. Of the water, the learning coefficient corresponding to each elapsed time is referred from the storage means, and each learning coefficient is multiplied by the water cooling temperature drop amount at each location to calculate the water cooling amount at each location. Since the cooling amount calculation means is provided, even if the transport speed of the hot-rolled steel sheet changes, an appropriate water cooling amount can be determined for each location, and as a result, the winding temperature after cooling can be determined. There is an effect that the target winding temperature can be matched very accurately.

According to the invention of claim 2, the distance from the tip of the hot-rolled steel sheet to each location is measured, the learning coefficient corresponding to each distance is referred from the storage means, and each learning coefficient is determined for each location. Since the water cooling amount calculation means for calculating the water cooling amount at each place by multiplying the water cooling temperature drop amount is provided, even if the hot-rolled steel sheet transfer speed changes, it is possible to obtain an appropriate value for each place. The amount of water cooling can be determined, and as a result, there is an effect that the winding temperature after cooling can be matched with the target winding temperature extremely accurately.

According to the third aspect of the present invention, the drop amount calculating means for calculating the actual drop amount of the water cooling temperature at each location, and the drop amount of the water cooling temperature at each location based on the water injection pattern determined by the water injection pattern determining means. Since it is configured so that the learning coefficient is obtained by dividing the actual water cooling temperature drop amount by the predicted water cooling temperature drop amount for each location, it is carried next time. The learning coefficient used for cooling the hot-rolled steel sheet can be updated based on the cooling result this time, and therefore, when cooling the hot-rolled steel sheet that will be conveyed next time, an appropriate water cooling amount must be determined for each location. There is an effect that can be.

According to the invention of claim 4, when the learning coefficient stored in the storage means is updated, the elapsed time after passing the reference point or the distance from the tip is divided into a plurality of sections. , The learning coefficient obtained by the calculation is distributed to the corresponding section, the average value of the learning coefficient of each section is obtained, and the updating means for updating the average value as the learning coefficient of each section is provided. Even if the number of learning coefficients calculated as described above is small, it is possible to maintain the learning coefficient suitable for the change pattern of the transport speed.

According to the invention of claim 5, when updating the learning coefficient stored in the storage means, the elapsed time after passing the reference point or the distance from the tip is divided into a plurality of sections. , The learning coefficient obtained by calculation is distributed to the corresponding section, and the linear approximation formula of the learning coefficient in each section is calculated,
Since the updating means for updating the change pattern of the linear approximation as the learning coefficient of each section is provided, the learning coefficient can be continuously changed over the adjacent section, and therefore the learning coefficient is changed stepwise. Compared with the case of performing it, learning control suitable for the change pattern of the transport speed becomes possible, and there is an effect that more accurate winding temperature control can be performed.

According to the invention of claim 6, the elapsed time from the time when the tip of the hot-rolled steel sheet passes a predetermined reference point on the line to the time when each point of the hot-rolled steel sheet passes the reference point. While measuring, the learning coefficient corresponding to each measured elapsed time is searched from among the learning coefficients set in advance according to the elapsed time, and each searched learning coefficient is used as the water cooling temperature drop amount at each location. Since it is configured to calculate the water cooling amount at each location by multiplying each, it is possible to determine an appropriate water cooling amount for each location even if the transport speed of the hot-rolled steel sheet changes, and as a result, There is an effect that the winding temperature after cooling can be matched with the target winding temperature with extremely high accuracy.

According to the invention of claim 7, the distance from the tip portion of the hot-rolled steel sheet to each position is measured, and the measured distance is corresponded to from the learning coefficient preset according to the distance. Since the learning coefficient to be searched is searched for and the water cooling temperature drop at each location is multiplied by each of the found learning coefficients to calculate the water cooling quantity at each location, the transport speed of the hot-rolled steel sheet changes. Even so, it is possible to determine an appropriate water cooling amount for each location, and as a result, there is an effect that the winding temperature after cooling can be made to match the target winding temperature with extremely high accuracy. .

According to the invention of claim 8, while calculating the actual water cooling temperature drop amount of each location, the water cooling temperature drop amount of each location is predicted from the water injection pattern, and the actual water cooling temperature of each location is calculated. Since the learning coefficient is calculated by dividing the temperature drop amount by the predicted water cooling temperature drop amount, the learning coefficient used for cooling the hot rolled steel sheet that will be conveyed next time should be updated based on the cooling result this time. Therefore, in cooling the hot-rolled steel sheet that is conveyed next time, there is an effect that an appropriate water cooling amount can be determined for each location.

According to the ninth aspect of the present invention, when the preset learning coefficient is updated, the elapsed time after passing the reference point or the distance from the tip is divided into a plurality of sections and the calculation is performed. The learning coefficient obtained in step S1 is distributed to the corresponding sections to obtain the average value of the learning coefficients in each section, and the average value is updated as the learning coefficient in each section. Even if the number is small, there is an effect that the learning coefficient suitable for the change pattern of the transport speed can be maintained.

According to the tenth aspect of the invention, when the preset learning coefficient is updated, the elapsed time after passing the reference point or the distance from the tip portion is divided into a plurality of sections and the calculation is performed. The learning coefficients obtained in this way are distributed to the corresponding sections to obtain the linear approximation formulas of the learning coefficients of each section, and the change pattern of the linear approximation is configured to be updated as the learning coefficient of each section. Since the learning coefficient can be changed continuously, learning control that is more suitable for the change pattern of the transport speed is possible than when changing the learning coefficient stepwise, and more accurate winding temperature control is possible. There is an effect that can be done.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a winding temperature control device for hot-rolled steel sheets according to Embodiment 1 of the present invention.

FIG. 2 is a configuration diagram showing a detailed configuration of a computer according to the first embodiment.

FIG. 3 is a flowchart showing an algorithm for controlling a cooling device by a coiling temperature control device for hot-rolled steel sheets.

FIG. 4 is a flowchart showing an algorithm in which a coiling temperature control device for hot-rolled steel sheets updates a learning coefficient.

FIG. 5 is a graph showing a learning coefficient corresponding to elapsed time.

FIG. 6 is a configuration diagram showing a detailed configuration of a computer according to the second embodiment.

FIG. 7 is a flowchart showing an algorithm for controlling the cooling device by the coiling temperature control device for the hot-rolled steel sheet.

FIG. 8 is a graph showing a learning coefficient corresponding to a distance.

FIG. 9 is a graph illustrating a linear approximation formula between points.

FIG. 10 is a graph showing a learning coefficient corresponding to elapsed time.

FIG. 11 is a graph showing a learning coefficient corresponding to a distance.

FIG. 12 is a configuration diagram showing a conventional winding temperature control device for hot-rolled steel sheets.

[Explanation of symbols]

1 final stand, 2 hot rolled steel plate, 6 calculator (measuring means), 7 finishing thermometer (measuring means), 8 winding thermometer (measuring means), 11 valve control device (control means),
21 estimation means, 22, 31 storage means, 23, 32
Water cooling amount calculation means, 24 Water injection pattern determination means, 25
Measuring means, 26 descent amount calculating means, 27 descent amount predicting means, 28 updating means.

Claims (10)

[Claims] 1. A measuring means for measuring a finish outlet temperature and a conveying speed at each position in a longitudinal direction of a hot rolled steel sheet output from a final stand of a rolling mill, and a finish outlet temperature measured by the measuring means and Estimating means for estimating the air-cooling temperature drop amount at each location based on the transport speed, and estimating the water-cooling temperature drop amount at each location from the finish side temperature, the air-cooling temperature drop amount, and the target winding temperature. And a storage unit that stores a learning coefficient set according to the elapsed time after the tip of the hot-rolled steel sheet passes a predetermined reference point on the line, and the tip of the hot-rolled steel sheet is the reference. The elapsed time from the point of passing the point until each of the points passes the reference point is measured, the learning coefficient corresponding to each elapsed time is referred from the storage means, and the learning coefficient of each of the points is calculated. Water cooling temperature drop it Water cooling amount calculating means for multiplying each to calculate the water cooling amount at each location, and water cooling amount at each location calculated by the water cooling quantity computing means and conveyance of each location measured by the measuring means A coiling temperature control device for a hot-rolled steel sheet, comprising: a water-pouring pattern determining means for determining a water-pouring pattern based on a speed; 2. A measuring means for measuring the finish-side temperature and the conveying speed at each position in the longitudinal direction of the hot-rolled steel sheet output from the final stand of the rolling mill, and the finish-side temperature and the finish-side temperature measured by the measuring means. Estimating means for estimating the air-cooling temperature drop amount at each location based on the transport speed, and estimating the water-cooling temperature drop amount at each location from the finish side temperature, the air-cooling temperature drop amount, and the target winding temperature. And storage means for storing a learning coefficient set according to the distance from the tip of the hot-rolled steel sheet, and measuring the distance from the tip of the hot-rolled steel sheet to each location, and corresponding to each distance A water cooling amount calculation means for calculating the water cooling amount at each location by multiplying the learning coefficient by the learning coefficient from the storage means and multiplying each learning coefficient by the water cooling temperature drop amount at each location. Each calculated by means A water cooling pattern determining means for determining the water injection pattern from the water cooling amount of the location and the transport speed of each location measured by the measuring means, and a cooling device for injecting cooling water into the hot rolled steel sheet are controlled according to the water injection pattern. A coiling temperature control device for hot-rolled steel sheet, comprising: a control means. 3. The winding temperature of each of the points after the cooling water is poured by the cooling device is measured, and the winding temperature is measured after the finish outlet temperature is measured by the measuring means at each of the points. Measuring means for actually measuring the time required to measure the temperature, and the finish-out side of each of the above-mentioned locations. A drop amount calculating means for calculating the water cooling temperature drop amount at each location from the temperature, the air cooling temperature drop amount, and the winding temperature, and the water cooling temperature drop amount at each location from the water injection pattern determined by the water injection pattern determining means. And a learning coefficient by dividing the water cooling temperature drop amount calculated by the drop amount calculation means for each location by the water cooling temperature drop amount predicted by the drop amount prediction means. The hot-rolled steel sheet winding temperature control device according to claim 1 or 2, further comprising: update means for obtaining and updating the learning coefficient stored in the storage means. 4. The updating means divides an elapsed time after passing the reference point or a distance from the tip portion into a plurality of sections when updating the learning coefficient stored in the storage means. The hot-rolled steel sheet according to claim 3, further comprising: dividing the obtained learning coefficient into corresponding sections to obtain an average value of learning coefficients in each section, and updating the average value as a learning coefficient in each section. Winding temperature control device. 5. The updating means divides the elapsed time after passing the reference point or the distance from the tip portion into a plurality of sections when updating the learning coefficient stored in the storage means. Along with this, the obtained learning coefficient is distributed to the corresponding section to obtain a linear approximation formula of the learning coefficient of each section, and the change pattern of the linear approximation is updated as the learning coefficient of each section. Winding temperature control device for hot rolled steel sheet. 6. The finish outlet temperature and the transport speed of each location in the longitudinal direction of the hot-rolled steel sheet output from the final stand of the rolling mill are measured, and based on the finish outlet temperature and the transport speed, each of the above locations is measured. While estimating the air cooling temperature drop amount, the water cooling temperature drop amount at each of the above locations is estimated from the finish outlet temperature, the air cooling temperature drop amount, and the target winding temperature, while the tip of the hot-rolled steel sheet is lined. From the time of passing the above predetermined reference point, the elapsed time until each of the above points passes the reference point is measured, and from the learning coefficient preset according to the elapsed time, the above is measured. After searching for the learning coefficient corresponding to each elapsed time, each of the searched learning coefficients is multiplied by the water cooling temperature drop amount at each location to calculate the water cooling amount at each location, and Water cooling amount and transportation A method for controlling a coiling temperature of a hot-rolled steel sheet, which comprises determining a water injection pattern from a feed rate and controlling a cooling device for injecting cooling water into the hot-rolled steel sheet according to the water injection pattern. 7. The finish exit side temperature and the transport speed of each location in the longitudinal direction of the hot rolled steel sheet output from the final stand of the rolling mill are measured, and the finish exit side temperature and the transport speed of each location are measured based on the finish exit side temperature and the transport speed. While estimating the air-cooling temperature drop, the water-cooling temperature drop at each of the above locations is estimated from the finish outlet temperature, the air-cooling temperature drop, and the target coiling temperature, while the tip of the hot-rolled steel sheet is used to The distance to each location is measured, and the learning coefficient corresponding to each of the above measured distances is searched from among the learning coefficients set in advance according to the above distance. The water cooling temperature drop amount is calculated to calculate the water cooling amount at each location, and the water injection amount is determined from the water cooling rate at each location and the transport speed, and the cooling water is injected into the hot rolled steel sheet. Cooler up A method for controlling a winding temperature of a hot rolled steel sheet, which is controlled according to a water injection pattern. 8. The winding temperature is measured at each of the points after cooling water is poured by the cooling device, and the finish side temperature is measured at each of the points, and then the winding temperature is measured. The time required to measure the air cooling temperature drop at each location is calculated based on the finish temperature and the required time at each location.
The water cooling temperature drop amount at each location is calculated from the air cooling temperature drop amount and the winding temperature, while the water cooling temperature drop amount at each location is predicted from the water injection pattern, and the water cooling temperature is calculated at each location. The hot rolling steel sheet according to claim 6 or 7, wherein the learning coefficient is obtained by dividing the temperature drop amount by the predicted water cooling temperature drop amount, and the preset learning coefficient is updated. Winding temperature control method. 9. When updating the preset learning coefficient, the elapsed time after passing the reference point or the distance from the tip portion is divided into a plurality of sections,
The winding of the hot-rolled steel sheet according to claim 8, wherein the obtained learning coefficient is distributed to corresponding sections to obtain an average value of the learning coefficient of each section, and the average value is updated as a learning coefficient of each section. Temperature control method. 10. When updating the preset learning coefficient, the elapsed time after passing the reference point or the distance from the tip is divided into a plurality of sections, and the obtained learning coefficient is calculated. 9. The hot-rolled steel sheet winding according to claim 8, wherein the linear approximation formula of the learning coefficient of each section is obtained by allocating to the corresponding section, and the change pattern of the linear approximation is updated as the learning coefficient of each section. Temperature control method.