WO2022038751A1 - 熱間圧延ラインの制御装置 - Google Patents

熱間圧延ラインの制御装置 Download PDF

Info

Publication number
WO2022038751A1
WO2022038751A1 PCT/JP2020/031494 JP2020031494W WO2022038751A1 WO 2022038751 A1 WO2022038751 A1 WO 2022038751A1 JP 2020031494 W JP2020031494 W JP 2020031494W WO 2022038751 A1 WO2022038751 A1 WO 2022038751A1
Authority
WO
WIPO (PCT)
Prior art keywords
rolled material
value
hot rolling
coiler
control device
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/JP2020/031494
Other languages
English (en)
French (fr)
Japanese (ja)
Inventor
敦 鈴木
光彦 佐野
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Toshiba Mitsubishi Electric Industrial Systems Corp
Original Assignee
Toshiba Mitsubishi Electric Industrial Systems Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Toshiba Mitsubishi Electric Industrial Systems Corp filed Critical Toshiba Mitsubishi Electric Industrial Systems Corp
Priority to PCT/JP2020/031494 priority Critical patent/WO2022038751A1/ja
Priority to JP2021534898A priority patent/JP7156538B2/ja
Priority to CN202080065771.2A priority patent/CN114423537B/zh
Publication of WO2022038751A1 publication Critical patent/WO2022038751A1/ja
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
    • B21B—ROLLING OF METAL
    • B21B45/00—Devices for surface or other treatment of work, specially combined with or arranged in, or specially adapted for use in connection with, metal-rolling mills
    • B21B45/02—Devices for surface or other treatment of work, specially combined with or arranged in, or specially adapted for use in connection with, metal-rolling mills for lubricating, cooling, or cleaning
    • B21B45/0203—Cooling
    • B21B45/0209—Cooling devices, e.g. using gaseous coolants
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
    • B21B—ROLLING OF METAL
    • B21B37/00—Control devices or methods specially adapted for metal-rolling mills or the work produced thereby
    • B21B37/46—Roll speed or drive motor control
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
    • B21B—ROLLING OF METAL
    • B21B37/00—Control devices or methods specially adapted for metal-rolling mills or the work produced thereby
    • B21B37/74—Temperature control, e.g. by cooling or heating the rolls or the product
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
    • B21B—ROLLING OF METAL
    • B21B37/00—Control devices or methods specially adapted for metal-rolling mills or the work produced thereby
    • B21B37/74—Temperature control, e.g. by cooling or heating the rolls or the product
    • B21B37/76—Cooling control on the run-out table
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
    • B21B—ROLLING OF METAL
    • B21B38/00—Methods or devices for measuring, detecting or monitoring specially adapted for metal-rolling mills, e.g. position detection, inspection of the product
    • B21B38/006—Methods or devices for measuring, detecting or monitoring specially adapted for metal-rolling mills, e.g. position detection, inspection of the product for measuring temperature

Definitions

  • This disclosure relates to a control device for a hot rolling line.
  • Patent Document 1 discloses a control device for a hot rolling line.
  • the control device feedforwardly controls the amount of cooling of the rolled material by the cooling device based on the learning result of the take-up temperature.
  • An object of the present disclosure is to provide a hot rolling line control device capable of improving the accuracy in feedforward control of the cooling amount of a rolled material by a cooling device.
  • the control device for the hot rolling line predicts the temperature of the rolled material in the hot rolling line in which the rolled material rolled by the finish rolling mill is injected with water by the cooling device, cooled, and then wound by the take-up coiler.
  • a storage unit that stores information on the prediction error of the passing speed of the rolled material in the cooling device used for the temperature model, and the rolled material rolled by the finishing rolling mill are injected with water by the cooling device and cooled, and then the winding is performed.
  • the learning value of the prediction error of the passing speed of the rolled material is calculated based on the actual value of the passing speed of the rolled material, and the storage is based on the learning value of the prediction error of the passing speed of the rolled material. It is equipped with a learning unit that updates information on the prediction error of the passing speed of the rolled material stored in the unit.
  • the control device for the hot rolling line predicts the temperature of the rolled material in the hot rolling line in which the rolled material rolled by the finish rolling mill is injected with water by the cooling device, cooled, and then wound by the take-up coiler.
  • a storage unit that stores information on the prediction error of the coiler deceleration start timing used in the temperature model, and the rolled material rolled by the finishing rolling mill 1 are injected with water by the cooling device to be cooled and then wound by the winding coiler.
  • the learning value of the prediction error of the coiler deceleration start timing was calculated based on the actual value of the line speed in the coiler deceleration section, and stored in the storage unit based on the learning value of the prediction error of the coiler deceleration start timing. It is equipped with a learning unit that updates information on the prediction error of the roller deceleration start timing.
  • the control device for the hot rolling line is a hot rolling line in which the rolled material rolled by the finish rolling mill is injected with water by a cooling device, cooled, and then wound by a take-up coiler.
  • the period from passing through the take-up thermometer provided between the cooling device and the take-up coiler to reaching the take-up coiler at the tail end of the rolled material is divided into a first half part and a second half part.
  • a storage unit that stores information on actual values of line speeds at a plurality of points in the first half and information on actual values of line speeds at a plurality of points in the second half, and information stored in the storage unit.
  • the actual value of multiple line speeds in the first half is used as the input layer, and the actual value of multiple line speeds in the second half is used as the output layer. It is provided with a prediction unit that calculates predicted values of line speeds at a plurality of points in the latter half by calculating an output layer using a neural network trained by the training unit using actual values as an input layer.
  • the control device for the hot rolling line is a hot rolling line in which the rolled material rolled by the finish rolling mill is injected with water by a cooling device, cooled, and then wound by a take-up coiler.
  • Actual values of line speeds at multiple points from passing through the take-up thermometer provided between the cooling device and the take-up coiler to the time when the tail end of the rolled material reaches the take-up coiler.
  • the storage unit that stores the information of It is equipped with a learning unit that performs filtering learning of the prediction error of the line speed.
  • FIG. 1 It is a block diagram of the main part of the hot rolling line to which the control device of the hot rolling line in Embodiment 1 is applied. It is a perspective view of the cut plate to which the control device of the hot rolling line in Embodiment 1 is applied. It is a figure which shows the predicted value of the temperature of the cut plate calculated by the control device of the hot rolling line in Embodiment 1. FIG. It is a figure which shows the predicted value and the actual value of the passing speed of the cutting plate in the ROT cooling device to which the control device of the hot rolling line in Embodiment 1 is applied. It is a figure which shows the learning table of the control device of the hot rolling line in Embodiment 1. FIG.
  • FIG. It is a figure for demonstrating the method which uses the learning value of the prediction error of the passing speed of a cutting plate in the control apparatus of the hot rolling line in Embodiment 1.
  • FIG. It is a figure which shows the predicted value of the temperature drop of a cutting plate by the control device of a hot rolling line in Embodiment 1.
  • FIG. It is a figure which shows the modification of the learning table of the control device of the hot rolling line in Embodiment 1.
  • FIG. It is a hardware block diagram of the control device of the hot rolling line in Embodiment 1.
  • FIG. It is a figure which shows the speed of the hot rolling line by the control of the control device of the hot rolling line in Embodiment 2.
  • FIG. 2 It is a figure which shows the predicted value and the actual value of the speed of a hot rolling line by the control device of a hot rolling line in Embodiment 2.
  • FIG. It is a figure which shows the start prediction timing and the start actual start timing of the coiler deceleration of the hot rolling line to which the control device of the hot rolling line in Embodiment 2 is applied.
  • FIG. It is a figure for demonstrating the method of using the learning value of the prediction error of the coiler deceleration start timing in the control device of the hot rolling line in Embodiment 2.
  • FIG. 2 It is a figure which shows the predicted value of the temperature drop of a cutting plate by the control device of a hot rolling line in Embodiment 2.
  • FIG. It is a figure which shows the actual value of the speed of the hot rolling line to which the control device 7 of the hot rolling line in Embodiment 2 is applied, and the actual value of the coiler deceleration rate.
  • FIG. 3 It is a figure which shows the example of the calculation of the training of the neural network by the control device of the hot rolling line in Embodiment 3.
  • FIG. It is a figure which shows the learning table of the control device of the hot rolling line in Embodiment 3.
  • FIG. It is a figure which shows the speed of the hot rolling line to which the control device of the hot rolling line in Embodiment 4 is applied. It is a figure which shows the learning table of the control device of the hot rolling line in Embodiment 4.
  • FIG. 1 is a block diagram of a main part of a hot rolling line to which a control device for a hot rolling line according to the first embodiment is applied.
  • the finish rolling mill 1 is provided on the downstream side of the rough rolling mill (not shown).
  • the ROT cooling device 2 is provided on the downstream side of the finish rolling mill 1.
  • the pinch roll 3 is provided on the downstream side of the ROT cooling device 2.
  • the take-up coiler 4 is provided on the downstream side of the pinch roll 3.
  • the ROT cooling device 2 includes a water injection device.
  • the water injection device is divided into a plurality of banks in the cooling water supply system. Multiple banks are lined up in the length direction of the hot rolling line. Each of the plurality of banks is provided with a plurality of water injection valves. Multiple water injection valves are lined up in the length direction of the hot rolling line. A plurality of nozzles are provided for each of the plurality of water injection valves. The plurality of nozzles are arranged in the width direction of the hot rolling line.
  • the finish rolling mill exit side thermometer 5 is provided between the finish rolling mill 1 and the ROT cooling device 2.
  • the take-up thermometer 6 is provided between the ROT cooling device 2 and the pinch roll 3.
  • the finish rolling mill 1 finish-rolls the rolled material.
  • the finish rolling mill exit side thermometer 5 measures the initial temperature of the total length of the rolled material as an FDT actual value before cooling.
  • the ROT cooling device 2 cools the rolled material by injecting water at a constant pressure.
  • the take-up thermometer 6 measures the initial temperature of the entire length of the rolled material as a CT actual value.
  • the take-up coiler 4 winds up the rolled material.
  • the control device 7 includes a storage unit 7a, a learning unit 7b, and a control unit 7c.
  • the control device 7 controls the speed of the rolling material by the final stand of the finishing rolling mill 1 by the control unit 7c. At this time, the control device 7 controls the torque of the mandrel which is the winding center of the winding coiler 4 in order to keep the winding shape of the rolled material by the winding coiler 4 good by the control unit 7c.
  • the torque reference at this time is the tension between the final stand of the finishing rolling mill 1 and the mandrel of the take-up coiler 4 in consideration of the calculated value of the radius of the coil at that time and the torque required for bending the rolled material.
  • the value is set so as to be a preset tension reference value.
  • the radius of the coil is calculated from the number of rotations of the mandrel of the take-up coiler 4, the plate thickness of the rolled material, and the space factor of the rolled material.
  • the control device 7 controls the control unit 7c to sandwich the rolled material with the pinch roll 3 at a preset pressure in order to maintain a good winding shape of the rolled material in the take-up coiler 4.
  • tension is generated between the pinch roll 3 and the mandrel of the take-up coiler 4.
  • the control device 7 controls the torque of the mandrel of the take-up coiler 4 by the control unit 7c.
  • the torque reference at this time is set so that the tension between the pinch roll 3 and the mandrel of the take-up coiler 4 becomes a preset tension reference value.
  • the control device 7 feed-forward controls the ROT cooling device 2 by the control unit 7c, starting from the FDT actual value of each cutting plate measured by the finish rolling mill exit side thermometer 5. Specifically, the control device 7 starts at the FDT actual value of each cutting plate measured by the finishing rolling mill exit side thermometer 5 by the control unit 7c, and predicts the final winding temperature of the rolled material. Calculates the predicted temperature of each cutting plate on the entry side and the exit side of each bank of the ROT cooling device 2 so that the temperature matches the target value of the winding temperature. At this time, the control device 7 calculates the temperature drop of each cutting plate in each bank by using the temperature model by the control unit 7c. The control device 7 adds the learning value stored in the learning table of the storage unit 7a to the predicted value of the passing speed of each cutting plate used in the temperature model by the control unit 7c.
  • the control device 7 updates the amount of cooling water in each bank by the control unit 7c, and cuts each in each bank. Recalculate the plate temperature drop.
  • the control device 7 controls the opening and closing of all valves in the ROT cooling device 2 in order of priority so as to satisfy the cooling water amount by the control unit 7c.
  • the control device 7 uses the learning unit 7b to obtain the actual value of the rotation angular velocity of the work roll of the final stand of the finishing rolling mill 1 ⁇ F7 res , the roll diameter RF7 , the actual value of the rotation angular velocity of the take-up coiler 4, ⁇ DC res , and the coil.
  • the actual value of the passing speed of each cutting plate passing through the ROT cooling device 2 is calculated from the radius R DC of.
  • the control device 7 uses the actual value and radius of the rotational angular velocity of the lower pinch roll 3 instead of the actual value ⁇ DC res of the rotational angular velocity of the take-up coiler 4 and the radius R DC of the coil by the learning unit 7b.
  • control device 7 Each time the entire rolled material is wound by the take-up coiler 4, the control device 7 is subjected to the learning unit 7b of each cut plate based on the actual value of the passing speed of each cut plate passing through the ROT cooling device 2. Calculate the learning value of the passing speed.
  • the control device 7 updates the information of the prediction error of the passing speed of the rolled material stored in the storage unit 7a based on the learning value of the prediction error of the passing speed of the rolled material by the learning unit 7b.
  • FIG. 2 is a perspective view of a cutting plate to which the control device for the hot rolling line according to the first embodiment is applied.
  • the heat input / output is calculated after the rolled material is virtually divided into cutting plates of a certain length.
  • the constant length is set between 3 m and 5 m.
  • h w is a water-cooled heat transfer coefficient (W / mm 2 / ° C.).
  • a w is the area (mm 2 ) of the upper and lower surfaces of the cutting plate that comes into contact with the cooling water.
  • a w changes depending on the number of water injection valves opened in each bank.
  • T surf is the surface temperature (° C.) of the cutting plate.
  • T w is the temperature (° C.) of the cooling water.
  • ⁇ T i is the drop temperature (° C.) of the cutting plate k in the bank i.
  • i is the bank number.
  • t is the time (s).
  • l k is the length (mm) of the cutting plate k in the traveling direction.
  • H k is the plate thickness (mm) of the cutting plate k.
  • B k is the width (mm) of the cutting plate k.
  • ⁇ is the density of the cutting plate k (kg / mm 3 ).
  • CP is the specific heat (J / kg / ° C.) of the cutting plate k.
  • Li is the length of the bank i .
  • v k i (m / s) is the passing speed of the cutting plate k in the bank i.
  • FIG. 3 is a diagram showing predicted values of the temperature of the cutting plate calculated by the control device for the hot rolling line in the first embodiment.
  • FIG. 3 shows a predicted value of the temperature of the cutting plate when the passing speed of the cutting plate is assumed to be constant in one bank.
  • the control device 7 calculates the predicted value of the temperature of each cutting plate on the entry side and the exit side of each bank using the equations (1) and (2). According to the right side of the equation (2), the rate of decrease of the predicted value of the temperature of the cutting plate changes depending on the passing speed of the cutting plate.
  • FIG. 4 is a diagram showing predicted values and actual values of the passing speed of the cutting plate in the ROT cooling device to which the control device for the hot rolling line according to the first embodiment is applied.
  • the cooling phenomenon of the cutting plate differs between when the cutting plate moves at high speed and when the cutting plate is stationary. Therefore, in the cooling phenomenon, the influence of the passing speed of the cutting plate is taken into consideration.
  • the water-cooled heat transfer coefficient hwi in the bank i is expressed by the following equation (3).
  • f wi W / mm 2 / ° C.
  • V 0 m / s
  • b w ( ⁇ ) is an adjustment coefficient.
  • the rate of decrease of the predicted value of the temperature of the cutting plate changes greatly depending on the passing speed of the cutting plate.
  • the control device 7 learns the prediction error of the passing speed of the cut plate after the rolled material has passed through the final stand of the finishing rolling mill 1.
  • the cutting plate k passes through the fourth bank from the upstream side.
  • the control device 7 stores the information of the predicted value of the passing speed of the cutting plate k and the information of the actual value in association with the information of the position in the longitudinal direction of the ROT cooling device 2.
  • the control device 7 determines the average value of the difference between the predicted value vd ki and the actual value v k i of the passing speed of the cutting plate k in the fourth bank to the nth bank from the upstream side Verror CUR (k). To calculate. Specifically, the control device 7 calculates the Verror CUR (k) using the following (4).
  • control device 7 performs filtering update on the average value Verror CUR ( k ) of the difference between the predicted value vd ki of the passing speed of the cutting plate k and the actual value vki . Specifically, the control device 7 performs filtering update using the following equation (5).
  • Error new (k) is a learning value of the passing speed of the cutting plate k after the filtering update.
  • Verror old (k) is a learning value of the passing speed of the cutting plate k before the filtering update.
  • ⁇ is the learning gain.
  • ⁇ is a value of 0 or more and 1 or less.
  • FIG. 5 is a diagram showing a learning table of the control device for the hot rolling line according to the first embodiment.
  • the information of "learning value (m / s)" includes the information of "steel grade", the information of "target plate thickness (mm)", and the information of "cut plate number”. Is associated with.
  • the information of "learning value (m / s)” is information indicating the learning value of the passing speed of the cutting plate before the filtering update.
  • the information of "steel type” is information indicating the material of the rolled material.
  • the information of "target plate thickness (mm)” is information indicating the target plate thickness of the rolled product.
  • the information of the "cut plate number” is the information indicating the number for identifying the cut plate.
  • FIG. 6 is a diagram for explaining a method of using the learning value of the prediction error of the passing speed of the cutting plate in the control device of the hot rolling line in the first embodiment.
  • control device 7 uses a learning value Verrorold (k ) as a predicted value of the passing speed of the cut plate k in each bank that the rolled material passes after passing through the final stand of the finishing rolling mill 1. ) Is added.
  • FIG. 7 is a diagram showing a predicted value of the temperature drop of the cutting plate by the control device of the hot rolling line in the first embodiment.
  • the predicted value of the temperature drop of the cutting plate is also corrected on the entry side and the exit side of each bank through which the rolled material passes after passing through the final stand of the finishing rolling mill 1.
  • the control device 7 updates the information of the prediction error of the passing speed of the rolled material. Therefore, it is possible to improve the accuracy in feedforward control of the cooling amount of the rolled material by the cooling device.
  • control device 7 learns the prediction error of the passing speed for each of the plurality of cutting plates. Therefore, it is possible to more reliably improve the accuracy when feedforward controlling the cooling amount of the rolled material by the cooling device.
  • control device 7 switches from the final stand of the finishing rolling mill 1 in which the speed reference of the hot rolling line passes through the final stand of the finishing rolling mill 1 to the winding coiler 4 in which the tension is controlled. After the timing, the prediction error of the passing speed of the cutting plate passing through the cooling device is learned. Therefore, it is possible to efficiently learn only the points where the prediction error of the passing speed of the cutting plate becomes large.
  • control device 7 stores the information of the learning value in association with the information of the steel grade of the cutting plate and the information of the target plate thickness. Therefore, it is possible to more reliably improve the accuracy when feedforward controlling the cooling amount of the rolled material by the cooling device.
  • control device 7 learns the average value of the prediction errors of the passing speeds of the plurality of banks of the cooling device after the rolled material passes through the final stand of the finishing rolling mill 1 for each of the number of cutting plates. Therefore, it is possible to efficiently learn only the points where the prediction error of the passing speed of the cutting plate becomes large.
  • FIG. 8 is a diagram showing a modified example of the learning table of the control device for the hot rolling line according to the first embodiment.
  • the information of "learning value (m / s)" includes the information of "steel grade", the information of "target plate thickness (mm)", the information of "Part”, and “corresponding”. It is associated with the information of "cutting plate number to be used”.
  • the "Part” information is information indicating a portion of the rolled material in the longitudinal direction. Specifically, the information of "Head” is information indicating the tip end portion in the longitudinal direction of the rolled material. The information of "Middle” is information indicating an intermediate portion in the longitudinal direction of the rolled material. The information of "Tail” is information indicating the tail end portion in the longitudinal direction of the rolled material.
  • the information of the "corresponding cut plate number” is the information indicating the corresponding "cut plate number" among the “cut plate numbers” in FIG. Specifically, “m to int ⁇ (nm) / 3 ⁇ ” corresponds to “h ⁇ 3" and “Head”. “H ⁇ 3" and “Middle” correspond to “int ⁇ (nm) / 3 ⁇ + 1 to int ⁇ 2 (nm) / 3 ⁇ ". “H ⁇ 3” and “Tail” correspond to “int ⁇ 2 (nm) / 3 ⁇ + 1 to n”. “3 ⁇ h ⁇ 5" and “Head” correspond to "m' ⁇ int ⁇ (n'-m') / 3 ⁇ ".
  • the "learning value (m / s)" information is information indicating the average value of the learning values of the passing speed of the cutting plate before the corresponding plurality of filtering updates.
  • the control device 7 divides the new cutting plates into three equal parts corresponding to "Head”, “Middle”, and “Tail”. Divide into the parts to which the cutting board belongs.
  • the control device 7 is a learning value of "Part” to which each cutting plate belongs to the predicted value of the passing speed of each cutting plate in each bank that the rolled material passes through after passing through the final stand of the finishing rolling mill 1. k) is added.
  • the control device 7 divides the new cutting plate into three equal parts corresponding to "Head”, “Middle”, and “Tail”, and divides the new cutting plate into parts to which each cutting plate belongs. ..
  • the control device 7 is a learning value of "Part” to which each cutting plate belongs to the predicted value of the passing speed of each cutting plate in each bank that the rolled material passes through after passing through the final stand of the finishing rolling mill 1. k) is added. Therefore, even if the setting of the length of the cutting plate is changed or the total number of cutting plates is changed, the accuracy of feed-forward control of the cooling amount of the rolled material by the cooling device can be improved more reliably. can.
  • FIG. 9 is a hardware configuration diagram of the control device for the hot rolling line according to the first embodiment.
  • Each function of the control device 7 can be realized by a processing circuit.
  • the processing circuit comprises at least one processor 100a and at least one memory 100b.
  • the processing circuit comprises at least one dedicated hardware 200.
  • each function of the control device 7 is realized by software, firmware, or a combination of software and firmware. At least one of the software and firmware is written as a program. At least one of the software and firmware is stored in at least one memory 100b. At least one processor 100a realizes each function of the control device 7 by reading and executing a program stored in at least one memory 100b. At least one processor 100a is also referred to as a central processing unit, a processing unit, an arithmetic unit, a microprocessor, a microcomputer, and a DSP.
  • At least one memory 100b is a non-volatile or volatile semiconductor memory such as RAM, ROM, flash memory, EPROM, EEPROM, a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, a DVD, or the like.
  • the processing circuit comprises at least one dedicated hardware 200
  • the processing circuit may be implemented, for example, as a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, or a combination thereof.
  • each function of the control device 7 is realized by a processing circuit.
  • each function of the control device 7 is collectively realized by a processing circuit.
  • a part may be realized by the dedicated hardware 200, and the other part may be realized by software or firmware.
  • the function of the control unit 7c is realized by a processing circuit as dedicated hardware 200, and for the functions other than the function of the control unit 7c, at least one processor 100a reads a program stored in at least one memory 100b. It may be realized by executing the above.
  • the processing circuit realizes each function of the control device 7 by hardware 200, software, firmware, or a combination thereof.
  • FIG. 10 is a diagram showing the speed of the hot rolling line under the control of the control device for the hot rolling line according to the second embodiment.
  • the same or corresponding parts as those of the first embodiment are designated by the same reference numerals. The explanation of this part is omitted.
  • the control device 7 feed-forward controls the ROT cooling device 2 with the FDT actual value of each cutting plate measured by the finish rolling mill exit side thermometer 5 as the starting point, the coiler deceleration start timing used for the temperature model is set.
  • the learning value stored in the learning table is added to the predicted value.
  • the control device 7 uses the learning value of the deceleration rate stored in the learning table.
  • the control device 7 uses the learning value information of the coiler deceleration start timing and the learning value information of the deceleration rate in the learning table based on the actual value of the line speed.
  • FIG. 11 is a diagram showing predicted values and actual values of the speed of the hot rolling line by the control device for the hot rolling line in the second embodiment.
  • the control device 7 uses the following equation (6) to mean the average value of the line speed prediction error va line error (m / s). To calculate.
  • td DS (s) is the predicted time for starting the deceleration of the coiler.
  • v line res (t) (m / s) is the actual value of the line speed at time t.
  • vd line (t) (m / s) is a predicted value of the line speed at time t.
  • the control device 7 sets the actual value of the peripheral speed of the final stand of the finishing rolling mill 1 as the actual value of the line speed. After the tail end of the rolled material passes through the final stand of the finishing rolling mill 1, the control device 7 sets the actual value of the peripheral speed of the lower pinch roll 3 as the actual value of the line speed.
  • FIG. 12 is a diagram showing a start prediction timing and a start actual timing of the coiler deceleration of the hot rolling line to which the control device of the hot rolling line according to the second embodiment is applied.
  • the control device 7 calculates the learning value Z DS CUR (s) of the prediction error of the coiler deceleration start timing using the following equation (7).
  • ad CD (m / s 2 ) is the predicted rate of coiler deceleration.
  • the ad CD is expressed by the following equation (8).
  • the control device 7 performs filtering update for the learning value Z DS CUR (s) of the prediction error of the coiler deceleration start timing. Specifically, the control device 7 performs filtering update using the following equation (9).
  • Z DS new (s) is a learning value of the prediction error of the coiler deceleration start timing after the filtering update.
  • Z DS old (s) is a learning value of the prediction error of the coiler deceleration start timing before the filtering update.
  • ⁇ is the learning gain.
  • ⁇ is a value of 0 or more and 1 or less.
  • FIG. 13 is a diagram showing a learning table of the control device for the hot rolling line according to the first embodiment.
  • the learning value is stored in the cell corresponding to the information of "steel grade” and the information of "plate thickness".
  • FIG. 14 is a diagram for explaining a method of using the learning value of the prediction error of the coiler deceleration start timing in the control device of the hot rolling line according to the second embodiment.
  • control device 7 corrects the prediction of the timing of the start of coiler deceleration by the learning value Z DS old (s) in the prediction of the line speed on the time axis.
  • FIG. 15 is a diagram showing a predicted value of a temperature drop of a cutting plate by a control device for a hot rolling line in the second embodiment.
  • the predicted value of the temperature drop of the cutting plate is also corrected on the entry side and the exit side of each bank through which the rolled material passes after passing through the final stand of the finishing rolling mill 1.
  • FIG. 16 is a diagram showing an actual value of the speed of the hot rolling line to which the control device 7 of the hot rolling line according to the second embodiment is applied and an actual value of the coiler deceleration rate.
  • the control device 7 performs filtering update with respect to the predicted value of the coiler deceleration rate.
  • the control device 7 linearly approximates the actual value of the line speed by the least squares method after the winding of the rolled material is completed, and then the slope of the straight line is the actual result of the coiler deceleration rate.
  • a CD CUR (m / s 2 ).
  • the control device 7 calculates the a CD CUR using the following equation (10) for the a CD CUR .
  • control device 7 After that, the control device 7 performs filtering update for the predicted value of the coiler deceleration rate. Specifically, the control device 7 performs filtering update using the following equation (11).
  • ad CD new is a learning value of the coiler deceleration rate after the filtering update.
  • ad CD old is a learning value of the coiler deceleration rate before the filtering update.
  • ⁇ CD is the learning gain.
  • ⁇ CD is a value of 0 or more and 1 or less.
  • the learning value is stored in the cell in the information of "steel grade” and the information of "plate thickness".
  • the control device 7 updates the information of the prediction error of the coiler deceleration start timing. Therefore, even when the coiler deceleration is performed, it is possible to improve the accuracy in feedforward control of the cooling amount of the rolled material by the cooling device.
  • control device 7 calculates the learning value of the coiler deceleration start timing so that the predicted value of the line speed changes by the average value of the line speed prediction error in the coiler deceleration section at the intermediate point of the coiler deceleration section. .. Therefore, the prediction error of the coiler deceleration start timing can be easily calculated.
  • control device 7 stores the information of the learning value in association with the information of the steel grade of the cutting plate and the information of the target plate thickness. Therefore, it is possible to more reliably improve the accuracy when feedforward controlling the cooling amount of the rolled material by the cooling device.
  • control device 7 learns by filtering the predicted value of the coiler deceleration rate based on the actual value of the coiler deceleration rate. Therefore, it is possible to more reliably improve the accuracy when feedforward controlling the cooling amount of the rolled material by the cooling device.
  • FIG. 17 is a diagram showing the speed of the hot rolling line to which the control device for the hot rolling line according to the third embodiment is applied.
  • the same or corresponding parts as those of the first embodiment are designated by the same reference numerals. The explanation of this part is omitted.
  • control device 7 learns the change in the line speed due to the coiler deceleration after the tail end of the rolled material passes through the final stand of the finishing rolling mill 1 by the neural network.
  • the control device 7 trains the neural network by the training unit 7d.
  • the control device 7 calculates the predicted value of the line speed by the prediction unit 7e using the neural network.
  • the "first half” is a section from the tip of the rolled material passing through the take-up thermometer 6 to the middle part of the rolled material passing through the take-up thermometer 6.
  • the "second half” is a section from the middle portion of the rolled material passing through the take-up thermometer 6 to the tail end portion of the rolled material reaching the take-up coiler 4.
  • control device 7 collects 10 actual line speed values at equal intervals in the "first half" and "second half".
  • the control device 7 uses the actual value of the "first half” as the "input group”.
  • the control device 7 uses the actual value of the "second half” as the "output group”.
  • the boundary between the "first half” and the “second half” is set based on the "start of deceleration", but it may be about half of the total length of the rolled material (40% to 60%).
  • the number of actual values collected is not limited. For example, with the boundary between the "first half” and the “second half” as the position of 40% of the total length of the rolled material, 10 actual values were collected in the "first half” and 20 points were collected in the "second half”. You may collect the value.
  • the control device 7 predicts 10 points of the "second half” line speed pattern using a trained neural network, and corrects the "second half” line speed pattern. Specifically, when the intermediate portion of the rolled material passes through the finish rolling mill exit side thermometer 5 and 10 points of the actual value of the line speed of the "first half portion" are obtained, the control device 7 is concerned. The 10 points, the plate thickness, and the number of the steel grade classification to which the One-hot encoding is applied are input to the neural network. The control device 7 linearly connects 10 points of the predicted values of the line speed pattern of the "second half” obtained as the output of the neural network to obtain the predicted value of the new line speed pattern of the "second half". For the cutting plate that passes through the finish rolling mill exit side thermometer 5 after the predicted value of the new line speed pattern is obtained, the control device 7 uses the temperature model based on the predicted value of the new line speed pattern. The ROT cooling device 2 is feed-forward controlled.
  • the control device 7 stores the data required for the neural network in the database. When the data for a certain number of rolled materials are accumulated, the control device 7 adds the data to the training data and executes the training of the neuron.
  • FIG. 18 is a diagram showing an input layer, an output layer, and an intermediate layer of the neural network used by the control device for the hot rolling line in the third embodiment.
  • the neural network includes an input layer, an intermediate layer, and an output layer.
  • the input layer, the intermediate layer, and the output layer are connected in order.
  • the middle layer is at least one layer. It has an input layer, an intermediate layer, an output layer, and at least one neuron.
  • each neuron is associated with all neurons in the anterior and posterior layers.
  • each neuron is associated with some neurons in the anterior and posterior layers.
  • the control device 7 trains the neural network.
  • the control device 7 reads the number of the steel type classification of a large number of rolled materials, the plate thickness, and the actual data of the line speed from the past rolling actual data.
  • the steel type classification is mainly based on the chemical composition.
  • the actual line speed data is data of several hundred points collected from the tip to the tail of the rolled material. For example, these performance data are used for thousands of products.
  • One-hot encoding is applied to the steel type classification number. For example, if the steel grade classification number ranges from 1 to 15, 15 variables, which are the total number, are prepared. When the steel grade classification number of the rolled material is 3, the third variable is 1. Variables other than the third are set to 0.
  • the classification number of the steel grade by One-Hot encoding, the plate thickness, and the actual value of the line speed of 10 points in the "first half" are input. For example, if the steel class number ranges from 1 to 15, the number of neurons in the input layer will be 26.
  • the actual value of the line speed of 10 points in the "second half" is input to each neuron in the output layer.
  • the number of neurons in the output layer is 10.
  • the number of neurons in the middle layer is sufficiently larger than the number of neurons in the input layer and the output layer.
  • the number of neurons in each layer is 128.
  • FIG. 19 is a diagram showing an example of calculation of training of a neural network by a control device for a hot rolling line in the third embodiment.
  • the input V 1 F and the input V 2 F are neurons in the input layer.
  • N 1 F is a neuron in the middle layer.
  • the input V 1 F and the input V 2 F are input to the neuron N 1 F.
  • the bias b is required.
  • the value a 1 F is obtained.
  • the value a 1 F is converted by the activation function h 1 F.
  • each predicted value of the output layer is calculated.
  • the mean square error mes of each predicted value of the output layer is calculated by the following equation (12).
  • the control device 7 adjusts the weight w and the bias b of each neuron so that the mes is below a certain reference value.
  • the control device 7 stores the weight w and the bias b of each neuron in the learning table.
  • FIG. 20 is a diagram showing a learning table of the control device for the hot rolling line according to the third embodiment.
  • the information of "weight w" and the information of "bias b" are associated with each other.
  • the information of "weight w” is information indicating the weight w of each neuron.
  • the information of "bias b” is information indicating the bias b of each neuron.
  • control device 7 calculates the output layer using the neural network with the actual values of the line speeds of the plurality of points in the first half as the input layer, so that the plurality of points in the second half are calculated. Calculate the predicted line speed. Therefore, it is possible to improve the accuracy in feedforward control of the cooling amount of the rolled material by the cooling device.
  • control device 7 includes the plate thickness of the rolled material and the classification number of the steel grade in the input layer, and calculates the predicted value of the line speed at a plurality of points in the latter half using the neural network. Therefore, it is possible to more reliably improve the accuracy when feedforward controlling the cooling amount of the rolled material by the cooling device.
  • control device 7 sets the weight of each neuron of the neural network so that the error between the predicted value and the actual value of the line speeds of a plurality of points in the latter half or the value of the evaluation function using the error is equal to or less than the reference value. Adjust with bias. Therefore, it is possible to more reliably improve the accuracy when feedforward controlling the cooling amount of the rolled material by the cooling device.
  • the computational load in neuron training is high. Therefore, if there is no margin in the computing power of the control device 7 during rolling, the neurons are trained at the timing when the computing power of the control device 7 has a margin, such as during roll replacement of the finishing rolling mill 1 or the like, or during the repair period. You may.
  • FIG. 21 is a diagram showing the speed of the hot rolling line to which the control device for the hot rolling line according to the fourth embodiment is applied.
  • the same or corresponding parts as those of the third embodiment are designated by the same reference numerals. The explanation of this part is omitted.
  • control device 7 feeds forward the ROT cooling device 2 with the FDT actual value of each cutting plate measured by the finish rolling mill exit side thermometer 5 as a starting point, and determines the predicted value of the line speed. Correct based on the learning value stored in the learning table.
  • the control device 7 Each time the entire rolled material is wound by the take-up coiler 4, the control device 7 has a tail end portion after the tip portion of the rolled material has passed the winding temperature with respect to the prediction pattern of the line speed on the time axis. Collect the actual value of the line speed of multiple points in the range until reaching the coiler. For example, the control device 7 collects actual values of line speeds at 10 points.
  • the control device 7 calculates the prediction error of the line speed at each time using the following equation (13).
  • the control device 7 performs filtering update for the prediction error at each time using the following equation (14).
  • Error new ( tl ) is a learning value of the line speed after the filtering update.
  • Verror old (t l ) is a learning value of the line speed before the filtering update.
  • ⁇ (l) is the learning gain.
  • ⁇ (l) is a value of 0 or more and 1 or less.
  • the control device 7 stores the information of the learning value of the prediction error of the line speed in the learning table.
  • FIG. 22 is a diagram showing a learning table of the control device for the hot rolling line according to the fourth embodiment.
  • the information of the "learning value (m / s)" is the information of the "steel grade", the information of the "target plate thickness (mm)", and the information of the "sampling number". Can be associated.
  • the information of "learning value (m / s)” is information indicating the learning value of the prediction error of the line speed before the filtering update.
  • the information of "steel type” is information indicating the material of the rolled material.
  • the information of "target plate thickness (mm)” is information indicating the target plate thickness of the rolled product.
  • the information of the “sampling number” is the information indicating the number corresponding to the time when the actual value of the line speed is sampled.
  • the control device 7 performs filtering learning of the prediction error of the line speed based on the actual value of the line speed at a plurality of points. Therefore, it is possible to improve the accuracy in feedforward control of the cooling amount of the rolled material by the cooling device.
  • control device 7 stores the information of the learning value in association with the information of the steel grade of the rolled material and the information of the target plate thickness. Therefore, it is possible to more reliably improve the accuracy when feedforward controlling the cooling amount of the rolled material by the cooling device.
  • the hot rolling line control device of the present disclosure can be used for the hot rolling line.

Landscapes

  • Engineering & Computer Science (AREA)
  • Mechanical Engineering (AREA)
  • Control Of Metal Rolling (AREA)
PCT/JP2020/031494 2020-08-20 2020-08-20 熱間圧延ラインの制御装置 Ceased WO2022038751A1 (ja)

Priority Applications (3)

Application Number Priority Date Filing Date Title
PCT/JP2020/031494 WO2022038751A1 (ja) 2020-08-20 2020-08-20 熱間圧延ラインの制御装置
JP2021534898A JP7156538B2 (ja) 2020-08-20 2020-08-20 熱間圧延ラインの制御装置
CN202080065771.2A CN114423537B (zh) 2020-08-20 2020-08-20 热轧生产线的控制装置

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/JP2020/031494 WO2022038751A1 (ja) 2020-08-20 2020-08-20 熱間圧延ラインの制御装置

Publications (1)

Publication Number Publication Date
WO2022038751A1 true WO2022038751A1 (ja) 2022-02-24

Family

ID=80322768

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2020/031494 Ceased WO2022038751A1 (ja) 2020-08-20 2020-08-20 熱間圧延ラインの制御装置

Country Status (3)

Country Link
JP (1) JP7156538B2 (enExample)
CN (1) CN114423537B (enExample)
WO (1) WO2022038751A1 (enExample)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114493059A (zh) * 2022-04-19 2022-05-13 西安石油大学 一种基于机器学习的人员管控方法及系统
CN115608793A (zh) * 2022-12-20 2023-01-17 太原科技大学 一种机理融合数据的精轧温度调控方法
CN118988999A (zh) * 2024-10-24 2024-11-22 承德建龙特殊钢有限公司 一种基于数字孪生的热轧无缝钢管轧制力预测方法及装置
TWI866328B (zh) * 2022-12-28 2024-12-11 日商Ｔｍｅｉｃ股份有限公司 鋼板的溫度控制裝置
CN120105327A (zh) * 2025-01-17 2025-06-06 北京科技大学 云边端协同下基于多层级特征融合的卷取温度预测方法

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118773565B (zh) * 2024-09-09 2024-11-22 嘉兴市豪能科技股份有限公司 基于物联网的镀铝膜生产线监测系统及方法
CN119426360B (zh) * 2024-11-29 2025-10-03 鞍钢集团北京研究院有限公司 一种基于人工标签的智能热轧工艺优化方法

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS5588913A (en) * 1978-12-27 1980-07-05 Toshiba Corp Tension controller for universal rolling mill
JPH0732024A (ja) * 1993-06-28 1995-02-03 Mitsubishi Electric Corp 熱延鋼材の温度制御方法
JP2005254289A (ja) * 2004-03-12 2005-09-22 Jfe Steel Kk 圧延機の速度制御方法

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2795791B2 (ja) * 1993-06-18 1998-09-10 三菱電機株式会社 鋼板の冷却制御方法
JPH08252625A (ja) * 1995-03-16 1996-10-01 Kawasaki Steel Corp 熱間圧延巻取温度制御方法
JP3450108B2 (ja) * 1995-12-25 2003-09-22 三菱電機株式会社 熱延板材の冷却制御装置
JP5565200B2 (ja) * 2010-08-20 2014-08-06 Ｊｆｅスチール株式会社 熱間圧延における仕上温度制御装置
JP5928377B2 (ja) * 2013-03-18 2016-06-01 Ｊｆｅスチール株式会社 熱間圧延材の冷却制御方法
JP5542228B2 (ja) * 2013-03-21 2014-07-09 株式会社日立製作所 巻取り温度制御装置およびその制御方法
CN103611734B (zh) * 2013-11-25 2015-11-18 华中科技大学 一种层流冷却温度控制方法及系统
JP6187603B2 (ja) * 2014-01-24 2017-08-30 東芝三菱電機産業システム株式会社 圧延ラインのエネルギー消費量予測装置
JP6642289B2 (ja) * 2016-06-14 2020-02-05 東芝三菱電機産業システム株式会社 圧延ラインの数学モデル算出装置および圧延材の温度制御装置
CN107971351B (zh) * 2016-10-25 2020-03-27 宝山钢铁股份有限公司 用于钢管在线冷却的生产装置和生产方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS5588913A (en) * 1978-12-27 1980-07-05 Toshiba Corp Tension controller for universal rolling mill
JPH0732024A (ja) * 1993-06-28 1995-02-03 Mitsubishi Electric Corp 熱延鋼材の温度制御方法
JP2005254289A (ja) * 2004-03-12 2005-09-22 Jfe Steel Kk 圧延機の速度制御方法

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114493059A (zh) * 2022-04-19 2022-05-13 西安石油大学 一种基于机器学习的人员管控方法及系统
CN114493059B (zh) * 2022-04-19 2022-07-05 西安石油大学 一种基于机器学习的人员管控方法及系统
CN115608793A (zh) * 2022-12-20 2023-01-17 太原科技大学 一种机理融合数据的精轧温度调控方法
CN115608793B (zh) * 2022-12-20 2023-04-07 太原科技大学 一种机理融合数据的精轧温度调控方法
TWI866328B (zh) * 2022-12-28 2024-12-11 日商Ｔｍｅｉｃ股份有限公司 鋼板的溫度控制裝置
CN118988999A (zh) * 2024-10-24 2024-11-22 承德建龙特殊钢有限公司 一种基于数字孪生的热轧无缝钢管轧制力预测方法及装置
CN118988999B (zh) * 2024-10-24 2025-01-03 承德建龙特殊钢有限公司 一种基于数字孪生的热轧无缝钢管轧制力预测方法及装置
CN120105327A (zh) * 2025-01-17 2025-06-06 北京科技大学 云边端协同下基于多层级特征融合的卷取温度预测方法

Also Published As

Publication number Publication date
JP7156538B2 (ja) 2022-10-19
CN114423537B (zh) 2024-09-10
CN114423537A (zh) 2022-04-29
JPWO2022038751A1 (enExample) 2022-02-24

Similar Documents

Publication Publication Date Title
WO2022038751A1 (ja) 熱間圧延ラインの制御装置
CN101791633B (zh) 卷绕温度控制装置及控制方法
JP2000167615A (ja) 巻取温度制御方法及び制御装置
CN101535906B (zh) 用于控制和/或调节工业过程的方法
CN103108706B (zh) 用于在精轧机列中的金属热轧带材的温度和几何形状的实时测定方法
JP5679914B2 (ja) 鋼板の温度予測方法
US20060225474A1 (en) Method for regulating the temperature of a metal strip, especially in a cooling path
CN106493180A (zh) 卷绕温度控制装置以及卷绕温度控制方法
JP4208505B2 (ja) 巻取温度制御装置
JP7135962B2 (ja) 鋼板の仕上出側温度制御方法、鋼板の仕上出側温度制御装置、及び鋼板の製造方法
JPWO2020261444A1 (ja) 熱間圧延ラインの温度制御装置
JP3450108B2 (ja) 熱延板材の冷却制御装置
JP4208509B2 (ja) 圧延プロセスのモデル学習装置
KR102045652B1 (ko) 인공지능 기반 열연 런아웃 테이블 열유속계수 결정 장치
JP2744399B2 (ja) 圧延材の冷却制御装置
JP2009233716A (ja) 圧延材の冷却方法
JP5928377B2 (ja) 熱間圧延材の冷却制御方法
CN112334242B (zh) 轧制节奏控制系统
WO2024142347A1 (ja) 鋼板の温度制御装置
KR102045651B1 (ko) 인공지능 기반 열연 런아웃 테이블 열유속계수 추정 장치
JPH06238312A (ja) 熱延鋼板の冷却制御方法
US20250355426A1 (en) Material properties prediction device for rolled products
JP2007301603A (ja) 圧延材の巻き取り温度制御方法及び圧延装置
WO2022157826A1 (ja) 熱間圧延ラインの制御装置
Bulut et al. Co-ordinated control of profile and shape in hot strip finishing mills with nonlinear dynamics

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2021534898

Country of ref document: JP

Kind code of ref document: A

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20950316

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20950316

Country of ref document: EP

Kind code of ref document: A1