JP5519472B2 - Rolled material tension control device, rolled material tension control method, and hot tandem rolling mill - Google Patents

Description

  The present invention relates to a rolled material tension control device that controls the tension of the rolled material, a rolled material tension control method, and a hot tandem rolling mill equipped with the rolled material tension control device.

  In a hot tandem rolling mill that rolls rolled materials such as steel sheets, keeping the tension applied to the rolled material constant during rolling stabilizes the operation of the rolling mill and also improves the product quality of the rolled material. Is extremely important in maintaining Therefore, in a hot tandem rolling mill, a device called a looper is installed at an intermediate position between two rolling roll installation positions in order to keep the tension applied to the material to be rolled constant. The looper has a roller for supporting the material to be rolled at an intermediate position between the two rolling roll installation positions, and adjusts the tension applied to the material to be rolled by moving the roller up and down.

  In order to control the tension applied to the material to be rolled, it is necessary to detect the tension, but there is no detector that directly detects the tension. Therefore, the tension applied to the material to be rolled is obtained by calculating a detection value obtained from a load detector for detecting the load received by the looper from the material to be rolled, a pressure detector for detecting the pressure of the looper hydraulic drive device, or the like. It has been.

  In general, a detection value obtained from a load detector attached to the looper has good accuracy as a detection value for calculating the tension applied to the material to be rolled. However, the load detector is installed in an environment where water is frequently used, such as roll cooling and cooling of the material to be rolled, in the vicinity of a high temperature material to be rolled at about 1000 ° C. Often occurred.

  On the other hand, the pressure detector of the looper hydraulic drive device and the current detector of the looper electric drive device are installed at a position away from the material to be rolled at a temperature higher than that of the load detector, so there is little occurrence of failure or abnormal operation. . However, in the case of these detectors, in addition to the tension applied to the material to be rolled, the operating pressure generated when the looper hydraulic drive device or the looper electric drive device operates is also detected. Therefore, it is difficult to accurately detect only the tension applied to the material to be rolled.

  Therefore, conventionally, in order to acquire the tension applied to the material to be rolled, a load detector that is likely to cause a failure or an abnormal operation is used, or a pressure detector or a looper electric drive device of a looper hydraulic drive device that cannot obtain sufficient accuracy. There was no choice but to use a current detector.

  Patent Document 1 discloses a solution to this problem. According to Patent Document 1, the rolling material tension detection device uses a first tension calculator that calculates a tension applied to a material to be rolled using a current value of a looper electric drive machine, and a detection value of a load detector of the looper. A second tension calculator that calculates the tension applied to the material to be rolled using, and a signal switch that is connected to the output from the first tension calculator and the output from the second tension calculator. The signal switching unit appropriately switches between the tension signal from the first tension calculator and the tension signal from the second tension calculator and transmits it to the tension controller.

  Therefore, the rolling material tension detection device described in Patent Document 1 is, for example, a case where the load detector of the looper fails and the detection value of the load detector used for the calculation of the second tension calculator cannot be obtained. In addition, the calculation result of the first tension calculator can be transmitted to the tension controller. As a result, the tension control device can continue to control the tension applied to the material to be rolled.

Japanese Patent Laid-Open No. 1-237017

  However, in the rolled material tension detection device described in Patent Document 1, the accuracy and performance of the detectors used in the two tension calculators are different, so the tensions calculated by the two tension calculators are not always the same. Absent. Therefore, when there is a difference in tension calculated by the two tension calculators, for example, a failure or abnormality occurs in the load detector or the like, and a signal transmitted from the signal switch to the tension controller is transmitted to the second tension calculator. If the calculation result is switched from the calculation result to the calculation result of the first tension calculator, the control input signal in the tension control device changes suddenly, and tension control may become unstable.

  Accordingly, the present invention provides a rolled material tension control apparatus, a rolled material tension control method, and a hot material capable of stably performing tension control even when a control input signal used for tension control of the rolled material is switched. An object is to provide a tandem rolling mill.

  A rolled material tension control device according to the present invention is provided between a plurality of rolling rollers for rolling a rolled material and two rolling rollers, supports the rolled material hydraulically, and controls the tension applied to the rolled material. Is used in a hot tandem rolling mill equipped with a looper, and calculates and outputs a pressure command value of a hydraulic control device that drives the looper. A pressure detection unit that acquires the hydraulic pressure in the hydraulic cylinder, a load detection unit that acquires a load that the looper receives from the material to be rolled, a pressure command calculation unit that calculates a pressure command value to be output to the hydraulic control device, and the load A first pressure command correction unit that calculates a correction amount of a pressure command value using a load detection value obtained by the detection unit, a pressure detection value obtained by the pressure detection unit, and the first pressure command correction unit. Calculated correction amount A pressure command correction amount learning unit that learns the correlation between the pressure command value, and a second pressure command correction unit that calculates a correction amount of the pressure command value using the correlation learned by the pressure detection value and the pressure command correction amount learning unit; A correction amount switching unit that switches between the correction amount calculated by the first pressure command correction unit and the correction amount calculated by the second pressure command correction unit and transmits the correction amount to the pressure command calculation unit. Features.

  The rolled material tension control device according to the present invention includes a pressure command correction amount learning unit, and the pressure command correction amount learning unit calculates the pressure detection value acquired by the pressure detection unit and the first pressure command correction unit. The coefficient of the relational expression representing the correlation with the correction amount of the pressure command value based on the detected load value is learned. Therefore, even when the normal load detection value cannot be acquired by the load detection unit, the rolled material tension control device uses the learned correlation instead of the correction amount calculated by the first pressure command correction unit. The correction of the pressure command value can be continuously performed using the correction amount of the pressure command value based on it. Since the correction amount after the switching is correlated with the correction amount calculated by the original first pressure command correction unit, the amount of change in the correction amount at the time of switching is small. Therefore, when the correction amount is switched, the control output amount for the looper does not change greatly, so that the tension control of the material to be rolled can be stably performed.

  According to the present invention, a rolled material tension control device, a rolled material tension control method, and a rolled material tension control method capable of continuously performing stable tension control even when a control input signal used for tension control of the rolled material is switched. A hot tandem mill is provided.

The figure which showed the example of the structure of the hot tandem rolling mill provided with the to-be-rolled material tension | tensile_strength control apparatus which concerns on embodiment of this invention, and the to-be-rolled material tension | tensile_strength control apparatus. The figure which showed the example of the processing flow of the arithmetic processing in the 1st pressure command correction | amendment part of the to-be-rolled material tension | tensile_strength control apparatus which concerns on embodiment of this invention. The figure which showed the example of the processing flow of the 1st pressure command value correction amount calculation process in the 2nd pressure command correction part of the to-be-rolled material tension | tensile_strength control apparatus which concerns on embodiment of this invention. The figure which showed the example of the processing flow of the tension set value learning process in the tension set value learning part of the to-be-rolled material tension | tensile_strength control apparatus which concerns on embodiment of this invention. Example of processing flow of processing for learning coefficient of relational expression representing correlation between pressure detection value and first pressure command value correction amount in pressure command correction amount learning unit of rolled material tension control device according to embodiment of present invention FIG. The figure which showed the example of the processing flow of the switching process of the pressure command value correction amount in the correction amount switch part of the to-be-rolled material tension | tensile_strength control apparatus which concerns on embodiment of this invention. The figure which showed the example of the pressure command value calculation process which calculates the pressure command value to the hydraulic-servo-drive apparatus in the pressure command calculating part of the workpiece tension control apparatus which concerns on embodiment of this invention. The figure which showed the example of the processing flow of the learning process which learns the coefficient of the correlation formula in the tension set value learning part and pressure command correction amount learning part of the to-be-rolled material tension | tensile_strength control apparatus which concerns on embodiment of this invention. The figure which showed the example of the processing flow of the learning calculation process in the learning process which learns the coefficient of the relational expression of FIG. The figure which showed the example of the detailed process flow of the process which determines whether the load detection value in FIG. 6 is normal. The figure which showed the example of the detailed process flow of the process which determines whether the rolling performance value in FIG. 4 is normal.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a diagram showing an example of the configuration of a rolled material tension control device according to an embodiment of the present invention and a hot tandem rolling mill equipped with the rolled material tension control device.
The hot tandem rolling mill 100 is a device that continuously rolls a high-temperature rolling material 43 by rolling rollers 42 provided on a plurality of stands (not shown), and the hot tandem rolling mill 100 includes two rolling rollers 42. A looper 40 and a hydraulic servo drive device 30 for controlling the tension applied to the material to be rolled 43 are provided in the intermediate portion. Further, the hot tandem rolling mill 100 has a rolled material tension control device 1 for controlling the looper 40 and the hydraulic servo drive device 30.

  In FIG. 1, a rolling roller 42 rolls a material to be rolled 43 with the load while rotating at a predetermined speed. As the rolling roller 42 rotates, the material to be rolled 43 moves from left to right, for example. In this specification, the moving speed of the material 43 to be rolled is hereinafter referred to as a rolling speed.

  The rolling roller driving device 31 is constituted by an electric motor or the like, and rotates the rolling roller 42 at the commanded speed based on the speed command from the material to be rolled tension control device 1. Here, if there is a difference between the moving speed of the material to be rolled 43 by the entrance-side rolling roller 42a and the moving speed of the material to be rolled 43 by the exit-side rolling roller 42b, the material to be rolled 43 is slackened. , The tension applied to the material to be rolled 43 varies.

  Therefore, the material to be rolled tension control apparatus 1 detects the tension applied to the material to be rolled 43 through the looper 40 (actually, a physical quantity that can correspond to the tension), and cancels the fluctuation of the tension. That is, the looper 40 and the rolling roller driving device 31 are controlled so that the tension applied to the material to be rolled 43 is constant.

  The looper 40 rotates with the movement of the material to be rolled 43, and supports a roller 40a that supports the material to be rolled 43, and a load detector 40b that detects a load that the support roller 40a receives from the material to be rolled 43. The hydraulic cylinder 40d having a piston 40e and a piston rod 40f and reciprocatingly moving the piston 40e by hydraulic pressure, one end connected to the tip of the piston rod 40f of the hydraulic cylinder 40d, and the other end of the support roller 40a A link mechanism 40c connected to the rotary shaft and converting the reciprocating movement of the piston rod 40f into a vertical movement of the support roller 40a; and a piston position detector 40g for detecting the position of the piston 40e in the hydraulic cylinder 40d. Consists of.

  The hydraulic cylinder 40d is partitioned into two spaces by a piston 40e, and the pressure in the two spaces is given by the hydraulic servo valve 41 and the hydraulic servo drive device 30. The position of the piston 40e is determined by the pressure difference between the two spaces in the hydraulic cylinder 40d and the load received by the support roller 40a transmitted to the piston 40e via the piston rod 40f and the link mechanism 40c. This is because the position of the piston 40e, that is, the support roller 40a is controlled by controlling the valve opening of the hydraulic servo valve 41 that the hydraulic servo drive device 30 communicates with the two spaces partitioned by the piston 40e in the hydraulic cylinder 40d. This means that the vertical position (hereinafter, the vertical position is simply referred to as a looper position) can be controlled.

  By the way, the tension applied to the material 43 to be rolled increases when the looper position is moved upward, and decreases when it is moved downward. Accordingly, when the rolled material tension control device 1 detects a decrease or increase in the tension applied to the rolled material 43, the rolled material tension control device 1 moves the looper position upward or downward relative to the hydraulic servo drive device 30 in order to increase or decrease the tension. What is necessary is just to instruct | indicate to move below. That is, the control input performed by the material to be rolled tension control apparatus 1 is the tension applied to the material to be rolled 43, and the output is the looper position. Therefore, in order to perform the control, it is necessary to detect the tension and the looper position.

  In the present embodiment, the material to be rolled tension control apparatus 1 acquires the tension applied to the material to be rolled 43 by two methods. Actually, since it is difficult to detect the tension itself, two types of physical quantities corresponding to the tension are acquired. One of them is a detected value by the load detector 40b of the looper 40, and the other one is a detected value by a pressure detector (not shown) provided in the hydraulic cylinder 40d. Note that the pressure detector here detects the pressure difference between the two spaces partitioned by the piston 40e in the hydraulic cylinder 40d.

  The value detected by the load detector 40b is a physical quantity that faithfully corresponds to the tension applied to the material to be rolled 43 because the load that the support roller 40a receives from the material to be rolled 43 is detected. The hydraulic pressure generated by the pressure difference between the two spaces in the hydraulic cylinder 40d is a pressure detection because the load received by the support roller 40a from the material to be rolled 43 is balanced with the force transmitted through the link mechanism 40c. The value detected by the vessel can also be said to be a physical quantity corresponding to the tension applied to the material 43 to be rolled.

  Moreover, the to-be-rolled material tension | tensile_strength control apparatus 1 acquires the detected value of the piston position detector 40g as a physical quantity corresponding to a looper position. The piston position detector 40g measures the distance from one bottom of the hydraulic cylinder 40d to the piston 40e and detects it as the piston position. The piston position is linked to the position of the support roller 40a, that is, the looper position via the link mechanism 40c and the like. Therefore, if the piston position is detected, the looper position is the detected value and physical quantities representing the mechanical shape and arrangement of the hydraulic cylinder 40d, piston rod 40f, link mechanism 40c, etc. (these values are given in advance). Therefore, it can be obtained by calculation with a setting value based on the machine configuration of the looper 40).

  Furthermore, the structure of the detailed functional block of the to-be-rolled material tension | tensile_strength control apparatus 1 is demonstrated, referring FIG.

  The material to be rolled tension control apparatus 1 includes a tension setting value learning unit 2, a pressure command correction amount learning unit 3, a looper position setting unit 11, a rolling speed setting unit 12, a pressure command calculation unit 20, a pressure control unit 21, and a looper position control. Functional block such as unit 22, rolling speed control unit 23, first pressure command correction unit 24, second pressure command correction unit 25, correction amount switching unit 26, pressure detection unit 27, load detection unit 28, looper position detection unit 29 It is comprised including. Among these, the tension setting value learning unit 2 includes a tension value setting unit 10, a rolling record value collection unit 13, a rolling record value extraction unit 14, a rolling record value accumulation unit 15, and the like. The command correction amount learning unit 3 includes a correlation extraction unit 16, a correlation accumulation unit 17, a correlation calculation unit 18, and the like.

  The pressure detection unit 27 acquires a pressure detection value by a pressure detector (not shown) provided in the hydraulic cylinder 40 d in the looper 40, and the load detection unit 28 uses the load detection value by the load detector 40 b in the looper 40. Further, the looper position detector 29 calculates the detection value by the piston position detector 40g to acquire the looper position. The tension applied to the material to be rolled 43 includes the pressure detection value or load detection value acquired by the pressure detection unit 27 or the load detection unit 28, the looper position acquired by the looper position detection unit 29, and the dimensions of the material to be rolled 43. And a set value based on the machine configuration of the looper 40 can be obtained by calculation.

  The first pressure command correction unit 24 adjusts the looper position detection value acquired by the looper position detection unit 29, the tension setting value set by the tension value setting unit 10, the dimensions of the material 43 to be rolled, and the mechanical configuration of the looper 40. A theoretical value of the load corresponding to the detected value of the load acquired by the load detecting unit 28 is calculated using the set value based on the calculated value, and the calculated theoretical value of the load and the load acquired by the load detecting unit 28 are calculated. The difference from the detected value is calculated. Further, the first pressure command correction unit 24 uses the difference between the theoretical value of the load and the load detection value, the looper position detection value, and the set value based on the machine configuration of the looper 40 to set the first pressure command value correction amount. The calculated first pressure command value correction amount is transmitted to the correction amount switching unit 26 and the correlation extracting unit 16.

  As described above, the pressure command correction amount learning unit 3 includes the correlation extraction unit 16, the correlation accumulation unit 17, the correlation calculation unit 18, and the like. When the material to be rolled 43 is rolled, the pressure detection unit 27 is included. The correlation between the pressure detection value acquired in step 1 and the first pressure command value correction amount output from the first pressure command correction unit 24 is extracted, the coefficient of the relational expression representing the extracted correlation is acquired, and the relationship The coefficient of the relational expression is learned using the coefficient of the expression.

  That is, in the pressure command correction amount learning unit 3, the correlation extraction unit 16 detects the pressure detection value from the pressure detection unit 27 and the first pressure from the first pressure command correction unit 24 when the material to be rolled 43 is rolled. The command value correction amount is received at any time, and the received pressure detection value and first pressure command value correction amount are accumulated. When the correction amount is switched by the correction amount switching unit 26 during rolling, a switching signal for notifying the switching is transmitted from the correction amount switching unit 26 to the correlation extracting unit 16. Therefore, when the rolling is completed, the correlation extraction unit 16 determines whether or not the switching has been performed based on the presence or absence of the switching signal.

And when it determines with the determination not having switched at the time of rolling, the correlation extraction part 16 is the correlation (namely, the 1st pressure command value correction amount (that is, the accumulated pressure detection value). (Correlation coefficient). If the correlation coefficient obtained by extracting the correlation is larger than a predetermined threshold value, a relational expression representing the correlation is calculated. Here, if the relational expression is expressed by a linear expression, the pressure detection value is X, and the first pressure command value correction amount is Y, the relational expression is expressed as Expression (1).
Y = A · X + B (1)

  The correlation extraction unit 16 transmits the coefficient of the relational expression (expression (1)) between the pressure detection value thus extracted and the first pressure command value correction amount to the correlation accumulation unit 17. The correlation accumulating unit 17 receives and accumulates the coefficient of the transmitted relational expression. Therefore, each time the material 43 is rolled, the correlation accumulation unit 17 accumulates a coefficient of a relational expression between the pressure detection value and the first pressure command value correction amount.

  The correlation calculating unit 18 learns a coefficient of a relational expression representing a correlation between the pressure detection value accumulated in the correlation accumulating part 17 and the first pressure command value correction amount, and uses the learned relational expression coefficient as the second pressure. Transmit to the command correction unit 25. However, when switching is performed during rolling, or when the obtained correlation coefficient is smaller than a predetermined threshold, the correlation extraction unit 16 extracts the correlation, calculates the relational expression based on the correlation, The correlation coefficient is not transmitted to the correlation storage unit 17. Therefore, the subsequent accumulation of the coefficient of the relational expression in the correlation accumulation unit 17 and the learning of the coefficient of the relational expression in the correlation calculation unit 18 are not performed.

  Further, the correlation calculation unit 18 can refer to the rolling record value accumulated in the rolling record value accumulation unit 15 and learns the coefficient of the relational expression for deriving the correction amount of the pressure detection value from the pressure detection value. When doing, learning is also performed with reference to the rolling record value.

  The learning of the coefficient of the relational expression here means the coefficient of the relational expression representing the correlation between the detected pressure value obtained when the rolled material 43 is normally rolled and the first pressure command value correction amount. Is incorporated in the coefficient of the relational expression representing the correlation between the pressure detection value acquired and accumulated in the past rolling and the first pressure command value correction amount, details of which are separately shown in FIG. Will be described with reference to FIG.

  The second pressure command correction unit 25 learns the correlation calculation unit 18, the coefficient of the relational expression representing the correlation between the pressure detection value and the first pressure command value correction amount, and the pressure detection acquired by the pressure detection unit 27. The second pressure command value correction amount is calculated using the value, and the calculation result is transmitted to the correction amount switching unit 26.

  The correction amount switching unit 26 determines whether or not the load detection value acquired by the load detection unit 28 is normal. If the load detection value is normal, the correction amount switching unit 26 reads from the first pressure command correction unit 24. The transmitted first pressure command value correction amount is transmitted to the pressure command calculation unit 20. Further, if not normal, the correction amount switching unit 26 transmits the second pressure command value correction amount transmitted from the second pressure command correction unit 25 to the pressure command calculation unit 20. The details of the switching process will be described separately with reference to FIG. 6 and the process of determining whether or not the load detection value is normal will be described in detail with reference to FIG.

  The pressure command calculation unit 20 includes the tension set value transmitted from the tension value setting unit 10, the set value based on the dimensions of the material to be rolled 43 and the machine configuration of the looper 40, the looper position acquired by the looper position detection unit 29, and correction A pressure command value to the hydraulic servo drive device 30 is calculated using the pressure command value correction amount transmitted from the amount switching unit 26, and the calculated pressure command value is transmitted to the pressure control unit 21.

  The pressure control unit 21 calculates the difference between the pressure command value transmitted from the pressure command calculation unit 20 and the pressure detection value acquired by the pressure detection unit 27, and further hydraulic pressure so that the difference becomes zero. The opening degree of the servo valve 41 is calculated, and the calculation result is transmitted to the hydraulic servo drive device 30. The hydraulic servo drive device 30 applies a voltage to the hydraulic servo valve 41 so that the valve opening calculated by the pressure control unit 21 is reached, and drives the hydraulic servo valve 41 to thereby apply the tension applied to the material 43 to be rolled. Control.

  The looper position control unit 22 inputs the looper position acquired by the looper position detection unit 29 and the set values such as the initial position and target position of the looper 40 set by the looper position setting unit 11, and the looper position is the target. The correction amount of the rolling speed is calculated so as to be in the position, and the calculation result is transmitted to the rolling speed control unit 23.

  The rolling speed control unit 23 inputs a rolling speed correction value transmitted from the looper position control unit 22 and a setting value such as a rolling speed target value set by the rolling speed setting unit 12, and a rolling roller driving device A current for driving the rolling roller 42 is output to 31 so that the corrected rolling speed is obtained.

  As described above, the tension setting value learning unit 2 includes the tension value setting unit 10, the rolling record value collection unit 13, the rolling record value extraction unit 14, the rolling record value accumulation unit 15, and the like. When rolling is rolled, a rolling record value is collected, and a coefficient of a relational expression representing a correlation between the collected rolling record value and a tension setting value is learned. Here, the rolling actual value is the pressure value in addition to the actual thickness value, the actual width value, and the actual temperature value of the material 43 to be rolled, which are measured by the inlet side measuring device 44, the intermediate measuring device 45 and the outlet side measuring device 46. The actual tension value calculated from the pressure detection value acquired by the detection unit 27, the actual tension value calculated from the load detection value acquired by the load detection unit, and the actual looper position value acquired by the looper position detection unit 29 Shall be included.

  In the tension set value learning unit 2, the rolling record value collecting unit 13 receives the rolling record value of the material 43 being rolled at any time and accumulates the received record value. Then, after the rolling of the material to be rolled 43 is completed, the accumulated rolling record value is transmitted to the rolling record value extracting unit 14.

  The rolling record value extraction unit 14 analyzes the rolling record value collected at that time based on the past rolling record value, and determines whether or not the collected rolling record value is normal. If it is determined that the result is normal, the rolling record value extracting unit 14 transmits the collected rolling record value to the rolling record value accumulating unit 15. The rolling performance value accumulation unit 15 receives and accumulates the transmitted rolling performance values.

  The tension value setting unit 10 learns the coefficient of the relational expression representing the correlation between the rolling record value and the tension setting value using the rolling record value accumulated in the rolling record value accumulating unit 15, and uses the learned coefficient as the learned coefficient. The optimum tension setting value is obtained using the relational expression based on the relationship, and the obtained tension setting value is transmitted to the pressure command calculation unit 20. If the rolling performance value extraction unit 14 determines that the rolling performance value is not normal, the rolling performance value is transmitted to the rolling performance value accumulation unit 15 and the rolling performance value stored in the rolling performance value accumulation unit 15 is updated. Learning of the coefficient of the relational expression representing the correlation in the accumulation and tension value setting unit 10 is not performed.

  Subsequently, with reference to FIG. 1 and FIG. 2 to FIG. 11, detailed operations of main functional blocks in the material to be rolled tension control apparatus 1 will be described.

FIG. 2 is a diagram illustrating an example of a processing flow of a first pressure command value correction amount calculation process in the first pressure command correction unit 24 of the material to be rolled tension control apparatus 1.
The first pressure command correction unit 24 includes the looper position detection value acquired by the looper position detection unit 29, the tension set value set by the tension value setting unit 10, the dimensions of the material to be rolled 43, and the machine of the looper 40. Using a set value based on the configuration, a theoretical load value corresponding to the load detection value acquired by the load detection unit 28 is calculated (step S21). Next, the first pressure command correction unit 24 calculates the difference between the calculated load theoretical value and the load detection value acquired by the load detection unit 28 (step S22). Further, the first pressure command correction unit 24 uses the difference between the calculated theoretical load value and the detected load value, the detected looper position value, and the machine set value based on the machine configuration of the looper 40 to use the pressure command. A value correction amount is calculated (step S23).

FIG. 3 is a diagram illustrating an example of a processing flow of second pressure command value correction amount calculation processing in the second pressure command correction unit 25.
The second pressure command correction unit 25 is a coefficient of a relational expression representing the correlation between the pressure detection value learned by the correlation calculation unit 18 and the first pressure command value correction amount before the rolling of the material 43 to be rolled, that is, Then, the coefficient of the relational expression for deriving the first pressure command value correction amount from the pressure detection value is received (step S31). Next, the second pressure command correction unit 25 receives the pressure detection value from the pressure detection unit 27 as needed during rolling of the material to be rolled 43 (step S32), and learns from the pressure detection value learned by the correlation calculation unit 18. The second pressure command value correction amount is calculated using the coefficient of the relational expression for deriving the second pressure command value correction amount and the detected pressure value (step S33).

FIG. 4 is a diagram illustrating an example of a processing flow of the tension setting value learning process in the tension setting value learning unit 2.
The rolling performance value collection unit 13 receives and accumulates rolling performance values measured by the entry-side measuring device 44, the intermediate measuring device 45, the delivery-side measuring device 46, and the like at the time of rolling the material 43 (step S41). ). The actual rolling value includes the actual tension value calculated from the pressure detection value, the actual tension value calculated from the load detection value, the looper position actual value, and the like. When the rolling of the material to be rolled 43 is completed, the rolling performance value collection unit 13 transmits the received and accumulated rolling performance values to the rolling performance value extraction unit 14 (step S42).

  The rolling record value extraction unit 14 analyzes the rolling record value collected at that time based on the past rolling record value, determines whether or not the collected rolling record value is normal (step S43), and the rolling record value. When the value is normal (Yes in Step S44), the rolling record value extracting unit 14 transmits the collected rolling record value to the rolling record value accumulating unit 15 (Step S45). The rolling performance value accumulation unit 15 receives and accumulates the transmitted rolling performance value (step S46).

  Next, the tension value setting unit 10 learns the coefficient of the relational expression representing the correlation using the rolling result value and the tension result value accumulated in the rolling result value accumulation unit 15, and uses the relational expression. The optimum tension setting value is set by using (Step S47). On the other hand, when the collected rolling record value is not normal (No in step S44), the rolling record value accumulation unit 15 does not accumulate the rolling record value, and therefore the tension value setting unit 10 does not store the relationship. Do not learn formula coefficients. In addition, the detail of the process which determines whether the collected rolling performance value in step S43 is normal is demonstrated with reference to FIG. 11 separately. The details of the learning of the tension setting value in step S47 will be described separately with reference to FIG.

FIG. 5 is a diagram illustrating an example of a processing flow of processing for learning a coefficient of a relational expression representing a correlation between the pressure detection value and the first pressure command value correction amount in the pressure command correction amount learning unit 3.
In the pressure command correction amount learning unit 3, the correlation extraction unit 16 detects the pressure detection value from the pressure detection unit 27 and the first pressure command value from the first pressure command correction unit 24 when the material to be rolled 43 is being rolled. The correction amount and the switching signal from the correction amount switching unit 26 are received and stored as needed (step S51).

  The correlation extraction unit 16 determines whether or not a switching signal has been received after the end of rolling (step S52), and if the switching signal has not been received during rolling (No in step S53), the accumulated pressure detection value and A correlation (that is, a correlation coefficient) with the first pressure command value correction amount is extracted (step S54). If the correlation coefficient obtained by extracting the correlation is larger than a predetermined threshold (Yes in step S55), the correlation extraction unit 16 uses the relational expression (that is, the expression) representing the correlation. (1)) is calculated and transmitted to the correlation accumulating unit 17 (step S56).

  The correlation accumulating unit 17 receives the relational expression representing the transmitted correlation, and accumulates the coefficient of the relational expression (step S57). Subsequently, the correlation calculation unit 18 learns a coefficient of a relational expression representing the correlation between the pressure detection value accumulated in the correlation accumulation unit 17 and the first pressure command value correction amount (step S58). Details of learning of coefficients of relational expressions representing the correlation in step S58 will be described separately with reference to FIG.

  On the other hand, if the correlation extraction unit 16 has received a switching signal during rolling (Yes in step S53), or if the correlation coefficient extracted in step S54 is equal to or less than a predetermined threshold (No in step S55). ), The processing of step 56 to step S58 is not executed. That is, learning of the coefficient of the correlation equation in step S58 is not performed.

FIG. 6 is a diagram illustrating an example of a processing flow of the pressure command value correction amount switching processing in the correction amount switching unit 26.
The correction amount switching unit 26 includes a signal indicating that the load detection unit 28 transmitted from the load detection unit 28 is normal (hereinafter simply referred to as a normal signal), a load detection value, and a pressure transmitted from the pressure detection unit 27. The detection value, the looper position detection value transmitted from the looper position detection unit 29, the set value based on the dimensions of the material to be rolled 43 and the mechanical configuration of the looper 40 transmitted from the tension value setting unit 10, and the first pressure command correction unit 24 The first pressure command value correction amount transmitted from the second pressure command value and the second pressure command value correction amount transmitted from the second pressure command correction unit 25 are received (step S61).

  Then, the correction amount switching unit 26 uses the received load detection value, pressure detection value, looper position detection value, set size based on the dimensions of the material to be rolled 43 and the mechanical configuration of the looper 40, based on the load detection value. The tension and the tension based on the detected pressure value are respectively calculated (step S62).

  Next, the correction amount switching unit 26 includes a normal signal from the load detection unit 28, a first pressure command value correction amount, a second pressure command value correction amount, a tension based on the calculated load detection value, and a calculated pressure detection value. Whether or not the load detection value is normal is determined using the tension based on (Step S63). The details of the process of determining whether or not the load detection value in step S63 is normal will be described separately with reference to FIG.

  If it is determined in step S63 that the load detection value is normal (Yes in step S64), the correction amount switching unit 26 transmits the first pressure command value correction amount to the pressure command calculation unit 20 (step S65). If the detected load value is not normal (No in step S64), the second pressure command value correction amount is transmitted to the pressure command calculation unit 20 (step S66).

FIG. 7 is a diagram showing an example of pressure command value calculation processing for calculating a pressure command value for the hydraulic servo drive device 30 in the pressure command calculation unit 20.
The pressure command calculation unit 20 includes a looper position detection value transmitted from the looper position detection unit 29, a tension set value transmitted from the tension value setting unit 10, and a set value based on the dimensions of the material 43 to be rolled and the mechanical configuration of the looper 40. The pressure command value correction amount transmitted from the correction amount switching unit 26 is received (step S71). Next, the pressure command calculation unit 20 calculates the pressure setting value using the received looper position detection value, tension setting value, setting value based on the dimensions of the material to be rolled 43 and the mechanical configuration of the looper 40 (step S72). The pressure command value is calculated using the pressure setting value obtained by the calculation and the received pressure command value correction amount (step S73).

FIG. 8 is a diagram illustrating an example of a processing flow of a learning process for learning the coefficient of the correlation equation in the tension set value learning unit 2 and the pressure command correction amount learning unit 3.
Here, the learning process in the tension setting value learning unit 2 is performed in the tension value setting unit 10 and represents the correlation between the rolling actual value and the actual tension value for setting the optimum tension for the actual rolling value. This is a process of learning the coefficient of the correlation equation (see step S47 in FIG. 4). Further, the learning process in the pressure command correction amount learning unit 3 is performed in the correlation calculation unit 18, and learning for learning the coefficient of the correlation equation for deriving the second pressure command value correction amount from the tension setting value and the pressure detection value. Processing (see step S58 in FIG. 5). Since both processes are similar processes, both processes will be described in a form summarized in FIG. Further, in the following description of the present specification, the correlation expressions used in both are collectively referred to simply as “relational expressions”, and the tension value setting unit 10 and the correlation calculation unit 18 are collectively referred to simply as This is referred to as a “learning unit”.

  As shown in FIG. 8 (see also FIG. 1), the learning unit, when the rolling record value R is normal, the rolling record value R, the tension setting value T and the transmission result value transmitted from the rolling record value storage unit 15 The coefficient P of the relational expression is received (step S81).

The learning unit classifies the rolling performance value R, the tension setting value T, and the coefficient P of the relational expression into a plurality of layers according to the steel type G, the plate thickness target value X, and the plate width target value Y of the current material (step) S82). Here, the rolling actual value R classified into the layer i is R (i), the tension setting value T is T (i), the relational coefficient P is P (i), and the pressure actual value R is the layer i. If the function for classifying the tension set value T into the layer i is fti, and the function for classifying the coefficient P of the relational expression into the layer i is fpi, then R (i), T (i), P (I) is represented by the following equations (2) to (4), respectively.
R (i) = fri (R, G, X, Y) (2)
T (i) = fti (T, G, X, Y) (3)
P (i) = fpi (P, G, X, Y) (4)

  The learning unit stores the rolling record value R (i), the tension set value T (i), and the coefficient P (i) of the relational expression classified into the layer i. A learning table is provided, and the learning unit sets the rolling record value R (i), the tension setting value T (i), and the coefficient P (i) of the relational expression, respectively, for a long-term learning table. And it stores in the table for short-term learning (step S83). At this time, in each table, the actual rolling value 43, the tension setting value, and the coefficient of the relational expression of the rolled material 43 (hereinafter referred to as the previous material) that has been rolled up to the previous time for each layer i are already stored. Has been.

  Here, in order to make the explanation easy to understand, the actual rolling value of the current material stored in the table for long-term learning in the layer i is Rl (i), the current material stored in the table for short-term learning. The rolling actual value is Rs (i), the tension setting value of the current material stored in the long-term learning table is Tl (i), and the tension setting value of the current material stored in the short-term learning table is Ts ( i) The coefficient of the relational expression of the current material stored in the long-term learning table is represented as Pl (i), and the coefficient of the relational expression of the current material stored in the short-term learning table is represented as Ps (i). Furthermore, the long-term learned actual rolling value up to the previous material is Rlo (i), the short-term learned actual rolling value up to the previous material is Rso (i), and the long-term learned tension setting up to the previous material is set. The value is Tlo (i), and the tension setting value learned for a short time until the previous material so (i), the long-term learned coefficients of equation up to the previous material Plo (i), the learned coefficients of the relational expression of short-term up to the previous material expressed as Pso (i).

That is, Rl (i), Rs (i), Tl (i), Ts (i), Pl (i), Ps () stored in the long-term learning table and the short-term learning table, respectively. i) is represented by Formula (5)-Formula (10).
Rl (i) = R (i) (5)
Rs (i) = R (i) (6)
Tl (i) = T (i) (7)
Ts (i) = T (i) (8)
Pl (i) = P (i) (9)
Ps (i) = P (i) (10)

When the learning unit performs short-term learning, the tension setting value of the next material (the material 43 to be rolled next) and the coefficient of the relational expression are the actual rolling value Rs (i) of the current material and the tension setting value. Ts (i) and the coefficient Ps (i) of the relational expression, as well as the actual rolling actual value Rso (i), the tension setting value Tso (i), and the coefficient Pso (i) of the relational expression learned in the short term up to the previous material. Is used to calculate. Therefore, the learning unit learns for a short period of time, assuming that the function for calculating the next material tension setting value Tsn (i) in layer i is fts, and the function for calculating the coefficient Psn (i) of the next material relational expression is fps. Next, the tension setting value Tsn (i) of the next material in the layer i and the coefficient Psn (i) of the relational expression of the next material are calculated by the equations (11) and (12), respectively (step S84). .
Tsn (i) = fts (Ts (i), Tso (i)) (11)
Psn (i) = fps (Ps (i), Pso (i)) (12)

Further, when the learning unit performs long-term learning, the tension setting value of the next material and the coefficient of the relational expression are the rolling actual value Rl (i), the tension setting value Tl (i) of the current material, and the coefficient Pl of the relational expression. It is calculated using (i) and the actual rolling actual value Rlo (i), tension set value Tlo (i), and coefficient Plo (i) of the relational expression learned for a long period up to the previous material. Therefore, the learning unit assumes that a function for calculating the next material tension setting value Tln (i) in layer i is ftl, and a function for calculating the coefficient Pln (i) of the next material relational expression is fpl. Next, the tension setting value Tln (i) of the next material and the coefficient Pln (i) of the relational expression of the next material in the layer i are calculated by the equations (13) and (14), respectively (step S85). ).
Tln (i) = ftl (Tl (i), Tlo (i)) (13)
Pln (i) = fpl (Pl (i), Plo (i)) (14)

The learning unit obtains the tension setting values Tsn (i) and Tln (i) of the next material and the coefficients Psn (i) and Pln (i) of the relational expressions by the short-term learning and the long-term learning as described above, respectively. After the calculation, it is determined whether or not the next material layer matches the current material layer (step S86). Here, when the layer of the next material is expressed as in, the determination in step S86 can be expressed by the determination of Expression (15).
i = in (15)

In this determination, if the next material layer “in” matches the current material layer “i”, that is, if i = in (Yes in step S86), the learning unit obtains the short-term learning. Next, the tension setting value Tsn (i) of the next material and the coefficient Psn (i) of the next material relation are respectively set to the tension setting value Tn (i) of the next material and the coefficient Pn (i) of the next material relational expression. (Step S87). Therefore, the tension setting value Tn (i) of the next material and the coefficient Pn (i) of the relational expression of the next material can be expressed by the equations (16) and (17), respectively.
Tn (i) = Tsn (i) (16)
Pn (i) = Psn (i) (17)

If it is determined in step S86 that the next material layer in does not match the current material layer i, i.e., if i ≠ in (No in step S86), the learning unit may Next, the tension setting value Tln (in) of the next material obtained by learning and the coefficient Pln (in) of the relational expression of the next material are respectively set to the tension setting value Tsn (in) of the next material layer in the short-term learning. And it substitutes for coefficient Psn (in) of the relational expression of the layer in the next time material (step S88). Therefore, Tsn (in) and Psn (in) can be expressed by Expression (18) and Expression (19), respectively.
Tsn (in) = Tln (in) (18)
Psn (in) = Pln (in) (19)

The learning unit obtains the Tsn (in) and Psn (in) obtained by the substitution, that is, obtained by the calculation of the equations (18) and (19), the tension setting value Tn (in) of the next material and the relationship. It outputs as a coefficient Pn (in) of the equation (step S89). Therefore, when i ≠ in, Tn (in) and Pn (in) can be expressed by equations (20) and (21), respectively.
Tn (in) = Tsn (in) (20)
Pn (in) = Psn (in) (21)

FIG. 9 is a diagram illustrating an example of a processing flow of learning calculation processing (step S84 and step S85) in the learning processing for learning the coefficients of the relational expression in FIG.
Here, in order to simplify the description, in the following description, the plate thickness of the rolled material 43 detected during rolling is set as the actual plate thickness, and the set thickness of the rolled material 43 is set as the set plate. Thickness, the detected sheet width of the rolled material 43 is the actual sheet width, the set sheet width of the rolled material 43 is the set sheet width, and the detected length of the rolled material 43 is the plate length. The total plate length of the material to be rolled 43 is the total length of the plate, the tension applied to the material to be rolled 43 using the detected load detection value is the actual tension, and the tension applied to the set material to be rolled 43 is detected. The position of the looper 40 is called the actual looper position, and the detected flatness of the rolled material 43 is called the flatness of the plate.

  In the short-term learning and long-term learning, the learning unit determines the current tension setting values Ts (i) and Tl (i) based on the actual rolling values Rs (i) and Rl (i) of the current material. Further, learning degrees αts, αtl, αps, and αpl used for learning the coefficients Ps (i) and Pl (i) of the relational expressions are calculated (step S91). Note that the learning levels αts, αtl, αps, and αpl are parameters that are 0 or more and 1 or less, and are collectively referred to as the learning level α when it is not necessary to distinguish between them.

  In the calculation of the learning degree α, as the rolling actual values Rs (i) and Rl (i), the ratio of the plate length in which the deviation between the actual plate thickness and the set plate thickness is within the threshold value occupies the total plate length, the actual plate The ratio of the plate length to the total plate length where the deviation between the width and the set plate width is within the threshold, the average of the actual tension and the deviation of the set tension, the standard deviation of the actual tension, the standard deviation of the actual looper position, and the load detection value The average of the absolute value of the deviation between the tension used and the tension using the pressure detection value, the flatness of the plate, and the like are used.

Next, the learning degree α for learning the tension set values Ts (i) and Tl (i) and the coefficients Ps (i) and Pl (i) of the relational expressions is expressed as αts, αtl, αps, and αpl, respectively. The learning unit uses the calculated learning levels αts, αtl, αps, αpl and the following formulas (22) to (25) to set the tension setting values Tsn (i) and Tln (i) of the next material and the next material. The coefficients Psn (i) and Pln (i) of the relational expression are calculated (step S92).
Tsn (i) = αts · Ts (i) + (1−αts) · Tso (i)
... Formula (22)
Tln (i) = αtl · Tl (i) + (1−αtl) · Tlo (i)
... Formula (23)
Psn (i) = αps · Ps (i) + (1−αps) · Pso (i)
... Formula (24)
Pln (i) = αpl · Pl (i) + (1−αpl) · Plo (i)
... Formula (25)

In addition, since these Formula (22)-Formula (25) are equivalent to said Formula (11)-Formula (14), Formula (11)-Formula (14) are respectively Formula (26)- It is expressed as equation (29).
fts (Ts (i), Tso (i))
= Αts · Ts (i) + (1−αts) · Tso (i) (26)
ftl (Tl (i), Tlo (i))
= Αtl · Tl (i) + (1−αtl) · Tlo (i) (27)
fps (Ps (i), Pso (i))
= Αps · Ps (i) + (1−αps) · Pso (i) (28)
fpl (Pl (i), Plo (i))
= Αpl · Pl (i) + (1−αpl) · Plo (i) (29)

FIG. 10 is a diagram showing an example of a detailed processing flow of processing (steps S63 and S64) for determining whether or not the load detection value in FIG. 6 is normal.
The correction amount switching unit 26 determines whether or not a normal signal indicating that the load detection unit 28 is normal is received (step S101), and when the normal signal is not received (No in step S102), It is determined that the load detection value is abnormal (step S109). If a normal signal has been received (Yes in step S102), the correction amount switching unit 26 further determines whether or not the load detection value is normal by the following two processes. .

First, the first determination process will be described. The correction amount switching unit 26 uses a set value based on a load detection value, a pressure detection value, a looper position detection value, a dimension of the material 43 to be rolled, and a mechanical configuration of the looper 40, and a tension T _LC based on the load detection value, pressure the tension _{T PR} based on the detection values respectively calculated (step S103).

Next, the correction amount switch unit 26 calculates the difference computed tension _{T LC} and _{T PR,} whether the difference is smaller than the predetermined threshold value _ΔT LPβ1 · γ, i.e., following equation (30) Whether or not is satisfied is determined (step S104).
| T _LC −T _PR | <ΔT _LPβ1 · γ Expression (30)
Here, γ is a parameter that is greater than or equal to 0 and less than or equal to 1 and is updated as needed depending on the number of times of learning and any setting by the user (in the following description of the present specification, the meaning of γ is the same) ).

  Then, if the result of determination in step S104 is that Expression (30) is not satisfied (No in step S105), the correction amount switching unit 26 determines that the load detection value is abnormal (step S109). When Expression (30) is satisfied (Yes in Step S105), it is determined that the load detection value is normal (Step S108). Here, the expression (30) does not hold means that the expression (30) does not hold continuously for a predetermined time or more, and the expression (30) means that the negative. And

Next, the second determination process will be described. As described with reference to FIG. 1, the correction amount switching unit 26 receives the first pressure command value correction amount and the second pressure command value correction amount. Therefore, the correction amount switch unit 26, the difference between the first pressure command value correction amount P and the second pressure command value correction amount _{P s} | _{P-P s} | a calculated, the difference | _{P-P s} | is given It is determined based on the formula (31) whether or not it is smaller than the set threshold value ΔP _β1 · γ (step S106).
| P−P _s | <ΔP _β1 · γ Expression (31)

  Then, if the result of determination in step S106 is that Expression (31) is not satisfied (No in step S107), the correction amount switching unit 26 determines that the load detection value is abnormal (step S109). When Expression (31) is established (Yes in Step S107), it is determined that the load detection value is normal (Step S108). Here, the expression (31) does not hold means that the expression (31) does not hold continuously for a predetermined time or more, and the expression (30) means that the negative. And

FIG. 11 is a diagram showing an example of a detailed processing flow of processing (steps S43 and S44) for determining whether or not the rolling record value in FIG. 4 is normal.
The actual rolling value extraction unit 14 calculates the plate length L _{hactβ of the} portion where the difference between the actual plate thickness h _act of the material to be rolled 43 and the set plate thickness h _ref is smaller than a predetermined threshold value Δh _actβ , and the calculated plate length L _{The ratio} of _{hactβ to} the total _plate length L _{all is} compared with a predetermined threshold value ΔL _hβ · γ (step S111), and it is determined whether or not the comparison expression (32) is satisfied (step S112).
L _hactβ / L _all <ΔL _hβ · γ Formula (32)

  When the formula (32) is not satisfied (No in step S112), the rolling record value extracting unit 14 determines that the rolling record value is abnormal (step S126), and the formula (32) is satisfied. (Yes in step S112), it is further determined whether or not other rolling performance values are normal.

Therefore, rolling actual value extraction unit 14, the difference between the set strip width _{w ref} between the actual strip width _{w act} is calculated the plate length _{L Wactbeta} predetermined threshold [Delta] w _Actbeta smaller portions, the total of the calculated plate length _{L Wactbeta} The ratio to the plate length L _{all is} compared with a predetermined threshold value Δw _actβ · γ (step S113), and it is determined whether or not the comparison expression (33) is satisfied (step S114).
_{L wactβ / L all <ΔL wβ} · γ ··· formula (33)

  When the formula (33) is not established (No in step S114), the rolling record value extraction unit 14 determines that the rolling record value is abnormal (step S126), and the formula (33) is established. (Yes in step S114), it is further determined whether or not the other rolling performance values are normal.

Next, the actual rolling value extraction unit 14 calculates the average T _actAV of the actual tension, and compares the difference between the average T _{actAV of the} actual tension and the set tension T _ref with a predetermined threshold value ΔT _β · γ (step S115). ), It is determined whether or not the comparison expression (34) is satisfied (step S116).
| T _actAV −T _ref | <ΔT _β · γ Expression (34)

  When the formula (34) is not established (No in step S116), the rolling record value extracting unit 14 determines that the rolling record value is abnormal (step S126), and the formula (33) is established. (Yes in step S116), it is further determined whether or not other rolling performance values are normal.

Next, the actual rolling value extracting unit 14 _compares the standard deviation S _Tact of the actual tension with a predetermined threshold value ΔS _Tβ · γ (step S117), and determines whether or not the comparison expression (35) is satisfied. (Step S116).
S _Tact <ΔS _Tβ · γ Expression (35)

  When the formula (35) is not established (No in step S118), the rolling record value extracting unit 14 determines that the rolling record value is abnormal (step S126), and the formula (35) is established. (Yes in step S118), it is further determined whether or not other rolling performance values are normal.

Next, the actual rolling value extracting unit 14 _compares the standard deviation S _LAact of the actual looper position with a predetermined threshold value ΔS _LAβ · γ (step S119), and determines whether the comparison expression (36) is satisfied. (Step S120).
S _LAact <ΔS _LAβ · γ Expression (36)

  When the formula (36) is not established (No in step S120), the rolling record value extracting unit 14 determines that the rolling record value is abnormal (step S126), and the formula (36) is established. (Yes in step S120), it is further determined whether or not other rolling performance values are normal.

Next, the rolling actual value extractor 14, a first pressure instruction value correction amount P the difference between the second pressure command value correction amount P _S is compared with a predetermined threshold value _ΔP β2 · γ (step S121), the comparison It is determined whether or not the equation (37) is established (step S122).
| P−P _S | <ΔP _β2 · γ Expression (37)

  When the formula (37) is not satisfied (No in step S122), the rolling record value extracting unit 14 determines that the rolling record value is abnormal (step S126), and the formula (36) is satisfied. (Yes in step S122), it is further determined whether or not other rolling performance values are normal.

Next, the rolling actual value extractor 14 compares the difference between the tension _{T PR} based on the tension _{T LC} and pressure detection value based on the load detection value with a predetermined threshold value _ΔT LPβ2 · γ (step S123), the comparison It is determined whether or not Expression (38) is satisfied (step S124).
| T _LC −T _PR | <ΔP _LPβ2 · γ Equation (38)

  When the formula (38) is not established (No in step S124), the rolling record value extracting unit 14 determines that the rolling record value is abnormal (step S126), and the formula (36) is established. (Yes in step S124), it is determined that the rolling record value is normal (step S125).

  In the example of the processing flow shown in FIG. 11, it is determined whether or not the rolling record value is normal by comparing a plurality of rolling record values with a predetermined threshold value. It is not limited to this example. Any of the comparison items shown in the examples may be omitted, and other comparison items may be added. Further, the order of comparison determination is not limited to the example of FIG.

  As mentioned above, the to-be-rolled material tension | tensile_strength control apparatus 1 which concerns on embodiment of this invention has the pressure command correction amount learning part 3, and the pressure command correction amount learning part 3 is the load detection value obtained by the load detection part 28. When there is no abnormality, the pressure command value correction amount (first pressure command value correction amount) based on the pressure detection value acquired by the pressure detection unit 27 and the load detection value calculated by the first pressure command correction unit 24, The coefficient of the relational expression representing the correlation of Therefore, when the normal load detection value cannot be acquired by the load detection unit 28, the material to be rolled tension control apparatus 1 sets the pressure command value correction amount to the correction amount calculated by the first pressure command correction unit 24. From this, it is possible to switch to the second correction amount (second pressure command value correction amount) calculated by the second pressure command correction unit 25 based on the correlation learned by the pressure command correction amount learning unit 3.

  At this time, the second correction amount (second pressure command value correction amount) is obtained by correlating the original pressure command value correction amount (first pressure command value correction amount). small. Therefore, even if the correction amount is switched, the control output amount (pressure command value) for the looper does not change greatly, so that the tension control of the material to be rolled does not become unstable. Therefore, the tension control of the material to be rolled in the hot tandem rolling mill 100 is stabilized. As a result, the product quality of the material to be rolled is improved, and the operating rate of the hot tandem rolling mill is improved. Can be planned.

DESCRIPTION OF SYMBOLS 1 Rolled material tension | tensile_strength control apparatus 2 Tension setting value learning part 3 Pressure command correction amount learning part 10 Tension value setting part 11 Looper position setting part 12 Rolling speed setting part 13 Rolling result value collection part 14 Rolling result value extraction part 15 Rolling result Value accumulation unit 16 Correlation extraction unit 17 Correlation accumulation unit 18 Correlation calculation unit 20 Pressure command calculation unit 21 Pressure control unit 22 Looper position control unit 23 Rolling speed control unit 24 First pressure command correction unit 25 Second pressure command correction unit 26 Correction Amount switching unit 27 Pressure detection unit 28 Load detection unit 29 Looper position detection unit 30 Hydraulic servo drive unit 31 Rolling roller drive unit 40 Looper 40a Support roller 40b Load detector 40c Link mechanism 40d Hydraulic cylinder 40e Piston 40f Piston rod 40g Piston position detection 41 Hydraulic servo valve 42 Rolling roller 43 Rolled material 44 Incoming side measuring device 45 Intermediate measuring device 46 Outlet side measuring device 100 Hot tandem rolling mill

Claims (6)

A hot tandem rolling mill comprising a plurality of rolling rollers for rolling the material to be rolled, and a looper that is intermediate between the two rolling rollers and supports the material to be rolled hydraulically and controls the tension applied to the material to be rolled. In the work material tension control device for calculating and outputting the pressure command value of the hydraulic control device for driving the looper,
A pressure detector for acquiring the hydraulic pressure in the hydraulic cylinder of the looper;
A load detector for acquiring a load that the looper receives from the material to be rolled;
A pressure command calculation unit that calculates a pressure command value to be output to the hydraulic control device;
A first pressure command correction unit that calculates a correction amount of the pressure command value using a load detection value obtained by the load detection unit;
A pressure command correction amount learning unit for learning a correlation between a pressure detection value obtained by the pressure detection unit and a correction amount calculated by the first pressure command correction unit;
A second pressure command correction unit that calculates a correction amount of the pressure command value using the correlation learned by the pressure detection value and the pressure command correction amount learning unit;
A correction amount switching unit that switches between the correction amount calculated by the first pressure command correction unit and the correction amount calculated by the second pressure command correction unit and transmits the correction amount to the pressure command calculation unit;
A to-be-rolled material tension control device comprising: The pressure command correction amount learning unit
A correlation extraction unit that extracts a coefficient of a relational expression representing a correlation between the detected pressure value and the correction amount of the pressure command value calculated by the first pressure command correction unit;
A correlation accumulation unit for accumulating the coefficients of the extracted relational expression;
A correlation calculation unit that learns a coefficient of the accumulated relational expression and calculates a relational expression for deriving a correction amount of the pressure command value from the pressure detection value;
The to-be-rolled material tension | tensile_strength control apparatus of Claim 1 characterized by the above-mentioned. The correction amount switching unit
The load obtained by the load detection unit based on the difference between the correction amount of the pressure command value calculated by the first pressure command correction unit and the correction amount of the pressure command value calculated by the second pressure command correction unit. Determine whether the detection value is normal,
When the load detection value is normal, the correction amount of the pressure command value calculated by the first pressure command correction unit is transmitted to the pressure command calculation unit,
The amount of correction of the pressure command value calculated by the second pressure command correction unit is transmitted to the pressure command calculation unit when the load detection value is not normal. Rolled material tension control device. The correction amount switching unit
Whether the load detection value obtained by the load detection unit is normal based on the difference between the tension value calculated based on the pressure detection value and the tension value calculated based on the load detection value. Judgment,
When the load detection value is normal, the correction amount of the pressure command value calculated by the first pressure command correction unit is transmitted to the pressure command calculation unit,
The amount of correction of the pressure command value calculated by the second pressure command correction unit is transmitted to the pressure command calculation unit when the load detection value is not normal. Rolled material tension control device. A plurality of rolling rollers for rolling the material to be rolled, a looper that is intermediate between the two rolling rollers, supports the material to be rolled by hydraulic pressure, and controls the tension applied to the material to be rolled, and hydraulic control that drives the looper A rolled material tension control device that calculates and outputs a pressure command value of the apparatus, and a rolled material tension control method in a hot tandem rolling mill,
The material to be rolled tension control device,
Pressure detection processing for obtaining the hydraulic pressure in the hydraulic cylinder of the looper;
A load detection process for acquiring a load that the looper receives from the material to be rolled;
Pressure command calculation processing for calculating a pressure command value to be output to the hydraulic control device;
A first pressure command correction process for calculating a correction amount of the pressure command value using the load detection value obtained in the load detection process;
A pressure command correction amount learning process for learning a correlation between a pressure detection value obtained by the pressure detection unit and a correction amount calculated by the first pressure command correction unit;
A second pressure command correction process for calculating a correction amount of the pressure command value using the correlation learned by the pressure detection value and the pressure command correction amount learning unit;
A correction amount switching process for switching between the correction amount calculated by the first pressure command correction unit and the correction amount calculated by the second pressure command correction unit and transmitting the correction amount to the pressure command calculation unit;
A method for controlling the tension of the material to be rolled, characterized in that A plurality of rolling rollers for rolling the material to be rolled, a looper that is intermediate between the two rolling rollers, supports the material to be rolled by hydraulic pressure, and controls the tension applied to the material to be rolled, and hydraulic control that drives the looper A hot tandem rolling mill comprising a rolled material tension control device that calculates and outputs a pressure command value of the device,
The material to be rolled tension control device,
A pressure detector for acquiring the hydraulic pressure in the hydraulic cylinder of the looper;
A load detector for acquiring a load that the looper receives from the material to be rolled;
A pressure command calculation unit that calculates a pressure command value to be output to the hydraulic control device;
A first pressure command correction unit that calculates a correction amount of the pressure command value using a load detection value obtained by the load detection unit;
A pressure command correction amount learning unit for learning a correlation between a pressure detection value obtained by the pressure detection unit and a correction amount calculated by the first pressure command correction unit;
A second pressure command correction unit that calculates a correction amount of the pressure command value using the correlation learned by the pressure detection value and the pressure command correction amount learning unit;
A correction amount switching unit that switches between the correction amount calculated by the first pressure command correction unit and the correction amount calculated by the second pressure command correction unit and transmits the correction amount to the pressure command calculation unit;
A hot tandem rolling mill characterized by comprising:

Images