JP7411518B2 - Plant control device, rolling control device, plant control method, and plant control program - Google Patents

Description

The present invention relates to a plant control device, a rolling control device, a plant control method, and a plant control program.

BACKGROUND ART In a rolling mill, which is a plant that efficiently produces thin metal materials by rolling metal plates, defective plate thickness may occur due to uneven hardness of the metal plate that is the material to be rolled. Uneven hardness refers to a state in which the hardness of a rolled material is not uniform throughout the rolled material. The hardness of the rolled material acts as deformation resistance during rolling, so if there is uneven hardness in the rolling direction, which is the direction in which the rolled material is conveyed during rolling, the hardness of the rolled material may vary depending on the position. The way they collapse is different, and the thickness of the plate after rolling varies.

Rolling is generally performed by passing the material to be rolled through a rolling mill multiple times from the original thickness of the metal plate to the product thickness. When hardness unevenness exists, the hardness of the material to be rolled differs depending on the position, resulting in variation in plate thickness, and a new deviation in plate thickness occurs each time in multiple rolling operations. In order to improve the thickness accuracy of products, plate thickness control is carried out in rolling mills, but it is difficult to eliminate plate thickness fluctuations that occur every time rolling due to uneven hardness using conventional plate thickness control. there were.

For example, it is possible to detect plate thickness variations due to hardness unevenness that occurred during a certain rolling process using an entry-side plate thickness gage during the next rolling process, and to suppress plate thickness variations through feedforward plate thickness control. However, although the plate thickness control suppresses the previous plate thickness variations, new plate thickness variations occur due to hardness unevenness. In such a case, a control gain larger than the normal control gain is required to suppress new plate thickness fluctuations. Therefore, in the plate thickness control method disclosed in Patent Document 1, the presence or absence of hardness unevenness is determined by frequency analysis, and the control gain of feedforward plate thickness control is changed.

Furthermore, in feedforward control, in order to expect a sufficient control effect, the amount of phase shift of the control output is important as well as the control gain. Therefore, the plate thickness control device disclosed in Patent Document 2 attempts to bring out the maximum control effect by adjusting the control gain and phase shift amount based on the phase relationship between a plurality of control state quantities.

In addition, a patent document describes a method to bring out the maximum control effect by adjusting the control gain and phase shift amount using fast Fourier transform in order to accurately determine the phase relationship and attenuation amount relationship between multiple control state quantities. It is disclosed in 3.

Japanese Patent Application Publication No. 2000-33409 Special Publication No. 6404195 Japanese Patent Application Publication No. 2018-05544

In the technology disclosed in Patent Document 1, in order to eliminate variations in deformation resistance in the conveyance direction of the rolled material due to hardness unevenness, the plate thickness variation that occurred during the previous rolling is replaced with the entrance side plate thickness variation during the next rolling. This is removed by feedforward control. At that time, the control gain of the feedforward control is changed depending on the presence or absence of hardness unevenness.

Feedforward control is proportional control, and it is possible to maximize the control effect by providing a control output whose phase and amplitude match the deviation of the target control state quantity. Here, assuming a sine wave as the deviation of the controlled state quantity of the controlled object, we will examine how the control gain amplitude changes with the deviation of the controlled state quantity.

For example, a sine wave with a control gain G and a phase shift amount Δ is created as a control output for a sine wave sin(ωt) representing a deviation of the control state quantity, and the control result of the feedforward control is set as y. At this time, y is expressed as in equation (1).

Here, the amplitude X and phase difference δ of y in equation (1) are expressed by equations (2-1) and (2-2), respectively.

FIG. 25 is a diagram showing the relationship between the phase shift amount Δ of the control output in feedforward control, the phase difference δ and the amplitude X of the control state quantities before and after the control, and (a) shows the relationship between the phase shift amount Δ and the A diagram showing the relationship between the phase difference δ and (b) is a diagram showing the relationship between the phase shift amount Δ and the amplitude X of the control state quantity after control. As shown in FIG. 25(b), as the controlled phase shift amount Δ increases, the amplitude also increases, and depending on the control gain G, if the phase shift amount Δ exceeds plus or minus 60 degrees, no control effect can be obtained. It turns out that it has the opposite effect. That is, it can be seen that when the phase shift amount Δ is included in the control output, the phase of the obtained control result y deviates from the original sine wave sin(ωt).

In other words, even if the control gain G of feedforward control, which is proportional control, is increased, if the phase of the control output is out of phase with the controlled state quantity of the controlled object, that is, there is a phase shift amount Δ (not zero). ), the control effect not only decreases but may even worsen.

Here, when plate thickness variation occurs due to hardness unevenness, not only plate thickness control but also tension control is performed in the rolling control. Therefore, the phase relationship between plate thickness variation and hardness unevenness is shifted. This phase relationship indicates by what angle the peak position of each waveform deviates from one cycle of 360 degrees. Therefore, even if feedforward control is performed using the entrance side plate thickness deviation of the rolled material, a sufficient control effect cannot be obtained because the phase relationship is shifted from the original hardness unevenness.

In addition, hardness unevenness may occur due to multiple causes, and in that case, plate thickness fluctuations also occur at multiple frequencies accordingly. In addition to hardness unevenness, sheet thickness fluctuations also occur due to factors such as mechanical fluctuations during rolling, differences in the surface condition of the rolled material, and changes in the entrance thickness of the base material.The phase relationship between the factors and the sheet thickness fluctuations is Generally different. Therefore, it is necessary to adjust the feedforward control according to the frequency of the plate thickness fluctuation, but the entrance plate thickness deviation that can be measured with the entrance plate thickness gauge is a composite waveform of these, and it is necessary to adjust the feedforward control for each frequency component. It's impossible.

Note that such a situation is not limited to uneven hardness of a rolled material during rolling of metal materials, but may also occur in general plant control. In particular, in cases where control results are obtained by controlling a controlled object that includes pre-control variation factors that are generated based on a reference variation factor, the phase between the reference variation factor and the pre-control variation factor may be If there is a deviation, a sufficient control effect cannot be obtained as described above.

In particular, when multiple fluctuation factors exist, the resulting state quantity has multiple frequencies, and for each of them, the reference fluctuation factor and the fluctuation factor before control are out of phase, and the detected Even if feedforward control is implemented using state quantities, sufficient control effects cannot be obtained.

Patent Document 2 discloses that in feedforward control of a plant such as a rolling mill in which the controlled state quantity of a controlled object includes a plurality of fluctuation factors with different phases, a phase shift amount Δ of a control output is suitably adjusted. Techniques for increasing control effectiveness are disclosed. According to this technology, first, when processing such as rolling is performed, the variation in the control state quantity before control (pre-control state quantity) and the variation in the control state quantity after control (post-control state quantity) are A phase difference δ is acquired by a phase difference acquisition section. Then, based on the phase difference δ, the feedforward adjustment section determines the phase shift amount Δ when the measurement result of the pre-control state quantity is reflected in the feedforward control. Therefore, it becomes possible to appropriately determine the control gain G and the phase shift amount Δ used in the control output of the feedforward control, and the control effect can be improved.

However, in the invention disclosed in Patent Document 2, the phase difference acquisition unit creates a time-series table of pre-control state quantities and post-control state quantities, and calculates the phase difference δ while comparing both tables. decide. Therefore, if the pre-control state quantity and post-control state quantity contain many frequency components and their waveforms become complex, it is necessary to identify the frequency of the plate thickness disturbance (hardness unevenness) that is the target of control, and to determine the phase difference. It becomes difficult to determine δ. As a result, it has been found that there are problems such as difficulty in accurately determining the phase shift amount Δ of the control output.

As a countermeasure, based on the pre-control state quantity which is the control state quantity before control when processing a workpiece as disclosed in Patent Document 3, the post-control state quantity which is the control state quantity after the control is calculated. A plant control device that performs feedforward control, wherein the position of the post-control state quantity relative to the pre-control state quantity is determined based on results of fast Fourier transform of time-series data of the pre-control state quantity and the post-control state quantity. A frequency response measuring means for acquiring a phase difference and an attenuation amount, and a control output timing shift amount, which is a delay time until the pre-control state quantity is reflected in the feedforward control, based on the acquired phase difference and attenuation amount. A means for adjusting feedforward control parameters has been devised. However, in the case where the detection results of the pre-control state quantity and the post-control state quantity include a plurality of frequency components, even if the feedforward control parameter adjustment means is performed for one frequency component, other frequency components cannot be adjusted. This is not possible, and the situation where the control effect is insufficient cannot be improved.

In view of the problems of the prior art as described above, an object of the present invention is to provide a larger feed even when the pre-control state quantity and the post-control state quantity have complex waveforms composed of multiple frequency components. To provide a plant control device, a rolling control device, a plant control method, and a plant control program that can efficiently determine a control timing shift amount (phase shift amount Δ) and control gain of a control output that can realize the effect of forward control. It is in.

In order to achieve the object of the invention, the plant control device according to the present invention calculates the control state amount after the control based on the pre-control state amount which is the control state amount before the control when processing the workpiece. A plant control device that performs feedforward control of a post-control state quantity, wherein a plurality of frequency components included in the pre-control state quantity are extracted based on a result of fast Fourier transform of time-series data of the pre-control state quantity. and a pre-control state quantity synthesis means (e.g., control disturbance provisional value creation device 320, control disturbance composite value creation device 325) that creates a pre-control state amount composite waveform by synthesizing the extracted frequency components; and frequency response measuring means (for example, frequency response a measuring device 201), and a control output timing shift amount, which is a delay time until the pre-control state quantity is reflected in the feedforward control, based on the acquired phase difference and attenuation amount for each of the plurality of frequency components. Feedforward control parameter adjustment means (for example, feedforward control adjustment device 101) that determines the control gain of the feedforward control, and control output timing shift amount and feedforward control determined by the feedforward control parameter adjustment means. a pre-control state quantity composite value correction means (for example, control disturbance composite correction value creation device 330) that corrects the pre-control state quantity composite waveform using a gain and determines a pre-control state quantity composite value correction value; It is characterized by having a feedforward control means (for example, feedforward control 307) that performs feedforward control using the state quantity composite value correction value. Other aspects of the present invention will be explained in the embodiments described below.

According to the present invention, even if the pre-control state quantity and the post-control state quantity have complex waveforms composed of multiple frequency components, the control timing of the control output can realize a greater feedforward control effect. A plant control device, a rolling control device, a plant control method, and a plant control program that can efficiently determine a shift amount (phase shift amount Δ) and a control gain are provided.

1 is a diagram showing an example of the overall configuration of a rolling mill and a rolling control device according to an embodiment of the present invention. The figure which showed the rolling phenomenon of the material to be rolled by a rolling mill, and the example of the parameter related to rolling control. The figure which showed the example of the control model of a rolling phenomenon. The figure which showed the example of the basic control structure of plate thickness control in a plate thickness control apparatus. The figure which showed the example of the basic control structure of tension control in a tension control device. The figure which showed the example of the simulation result when neither plate thickness control nor tension control is performed. The figure which showed the example of the simulation result when tension control of an entry side and an exit side was implemented by proportional-integral control, and only feedback control of plate thickness control of an exit side was implemented. The figure which showed the example of the simulation result when the feedback control of the board thickness control of the exit side of the stand rolling mill of the front stage is carried out in addition to the conditions in the case of FIG. 7. The figure which is the simulation result at the time of a multi-frequency component, and shows the state before adjustment. This is a simulation result for multi-frequency components, and is a diagram showing a state in which adjustment is performed using the frequency component with the maximum entrance side plate thickness deviation. This is a simulation result for multi-frequency components, and is a diagram showing a state in which adjustment is performed using the frequency component with the maximum exit side plate thickness deviation. FIG. 4 is a diagram showing simulation results for multi-frequency components, and shows a state adjusted using the method of the present embodiment. The figure which showed the control method of the plant control system of a comparative example. The figure which showed the control method of the plant control system of this embodiment. The figure which showed the example of the multiple frequency extended control structure of the board thickness control apparatus, feedforward control adjustment apparatus, and multiple frequency control adjustment apparatus which concern on this embodiment. The figure which showed the structure of the entrance side plate thickness deviation provisional value creation device. The figure which showed the structure of the entrance side board thickness deviation composite value creation device. FIG. 2 is a diagram showing an example of a detailed configuration of a control gain/timing shift amount setting device. FIG. 3 is a diagram for explaining an overview of the frequency response method, in which (a) is a diagram showing a response model in time space, and (b) is a diagram showing a response model in frequency space. FIGS. 2A and 2B are diagrams showing examples of frequency response simulation results by FFT, in which (a) is an example when the data collection time is 10.24 seconds, and (b) is an example when the data collection time is 5.12 seconds. Figures showing examples of frequency response simulation results by FFT, where (c) is an example when the data collection time is 2.56 seconds, and (b) is an example when the input signal is a single frequency and the data collection time is 2. .Example for 56 seconds. The figure which showed the example of the sampling period/data number search table. The figure which showed the example of a structure of a board thickness disturbance measuring device. The figure which showed the example of the frequency dependence characteristic of the entrance side plate thickness deviation amplitude and the exit side plate thickness deviation amplitude. The figure which showed the example of a structure of a frequency response estimation device. FIG. 3 is a diagram showing the configuration of an entrance side plate thickness deviation composite correction value creation device according to the present embodiment. FIG. 1 is a diagram showing an example of the hardware configuration of an information processing device that constitutes a rolling control device according to an embodiment of the present invention. It is a diagram showing the relationship between the phase shift amount of the control output in feedforward control and the phase difference and amplitude of the control state quantities before and after the control, and (a) is a diagram showing the relationship between the phase shift amount and the phase difference. , (b) is a diagram showing the relationship between the phase shift amount and the amplitude of the control state amount after control.

Embodiments of the present invention will be described in detail below with reference to the drawings. In addition, in each drawing, common components are given the same reference numerals, and redundant explanations will be omitted. Further, in this specification, a rolling control device for a rolling mill that rolls a material to be rolled such as metal will be described as a specific example of a plant control device.

≪1. Basic control configuration≫
FIG. 1 is a diagram showing an example of the overall configuration of a rolling mill 1 and a rolling control device 2 according to an embodiment of the present invention. Here, it is assumed that the rolling mill 1 is a tandem rolling mill with a 4-stand configuration, and the rolling control device 2 mainly performs control to minimize plate thickness fluctuations caused by uneven hardness when rolling the material 3 to be rolled. I do.

As shown in FIG. 1, the rolling mill 1 (tandem rolling mill) according to the present embodiment is configured by four stand rolling mills 11 to 14 arranged in series, and the rolled material 3 is It is continuously rolled by stand rolling mills 11 to 14. At this time, the material to be rolled 3 moves from the left side to the right side in FIG. 1 while being rolled.

Each of the stand rolling mills 11 to 14 is composed of six upper and lower rolls, and the six upper and lower rolls are called a work roll, an intermediate roll, and a backup roll from the inside with the material 3 to be rolled in between. In addition, plate thickness gauges 41 to 44 and tension gauges 51 to 54 are provided on the exit side of the stand rolling mills 11 to 14 in order to obtain control state quantities necessary for control by the rolling control device 2. .

Further, the rolling control device 2 includes motor speed control devices 21 to 25, roll gap control devices 31 to 34, plate thickness control devices 61 to 64, tension control devices 71 to 74, and the like. In this embodiment, the plate thickness control devices 61 to 64 and the tension control devices 71 to 74 play important roles, and the details thereof will be explained in sequence below.

First, before explaining the details of plate thickness control, the rolling phenomenon of the rolled material 3 will be explained.
FIG. 2 is a diagram showing an example of parameters related to the rolling phenomenon of the material 3 to be rolled by the rolling mill 1 and the rolling control. As shown in FIG. 2, rolling is performed by crushing the material 3 to be rolled between upper and lower work rolls of the rolling mill 1. At this time, the rolled material 3 is pulled by the entry tension T _b and the exit tension T _f and crushed by the rolling load P, so that the entry side plate thickness H becomes the exit side plate thickness h. Such a rolling phenomenon causes a forward rate f and a reverse rate b, and when the work roll speed is _VR , the entry side speed V _e and the exit side speed V _o are respectively calculated using the forward rate f and the backward rate b. It is expressed by the formula shown in FIG.

FIG. 3 is a diagram showing an example of a control model for rolling phenomena. In the case of a tandem rolling mill, the entry tension T b and the exit tension are determined by the entry speed V _e and exit speed V _o of the own stand rolling mill, the entry speed of the rear stand rolling mill, _and the exit speed of the front stand rolling mill. T _f changes. When these tensions change, the rolling load P, the outlet side plate thickness h, the inlet side speed V _e , and the outlet side speed V _o change.

As shown in FIG. 3, the rolling load P, the advancing rate f, and the backward rate b are all determined by the entrance side plate thickness H, the exit side plate thickness h, the entrance side tension T _b , the exit side tension T _f , the deformation resistance k, and the friction It is expressed as a function depending on the coefficient μ. Furthermore, the parameter L included in the equation shown in the lower right corner of FIG. 3 represents the distance between adjacent stands of the stand rolling mills 11 to 14. Further, input V _-1 is the exit speed from the adjacent front stand rolling mill, and V ₊₁ is the input speed to the adjacent rear stand rolling mill.

As described above, the rolling phenomenon is a phenomenon whose inputs are the input plate thickness H, work roll speed V _R , and roll gap S, and whose outputs are the input side tension T _b , the output side tension T _f , and the output side plate thickness h. However, this is a complex phenomenon that is also related to the rolling phenomenon in the stand rolling mills in the front and rear stages through tension.

Referring to FIG. 1, motor speed control devices 21 to 24 that control the work roll speed _VR and roll gaps S that are the intervals between the work rolls are set so as to correspond to each of the four stand rolling mills 11 to 14. Roll gap control devices 31-34 are provided for operation. In rolling processing, the thickness of the rolled material 3 that becomes the product is particularly important for the quality of the product. Thickness gauges 41 to 44 are installed. In addition, since the tension applied to the rolled material 3 is important for the stability of the rolling operation and also relates to the accuracy of the plate thickness, tension gauges 51 to 54 are installed on the exit side of the stand rolling mills 11 to 14. ing. Further, on the exit side of the #4 stand rolling mill 14, there is an exit bridle roll 15 for controlling the tension on the exit side, and a motor speed control device 25 for controlling the speed of the motor for driving the exit bridle roll 15. is installed.

In the rolling mill 1 and rolling control device 2 configured as described above, the plate thickness control device 61 of the #1 stand rolling mill 11 controls the roll gap S of the #1 stand rolling mill 11 via the roll gap control device 31. control. In addition, the plate thickness control devices 62 to 64 of the #2 to #4 stand rolling mills 12 to 14 control the work roll speed V _R of the previous stage, that is, the #1 to #3 stand rolling mills 11 to 13, to the motor speed control devices 21 to 14. 23.

At this time, the plate thickness control devices 62 to 64 after the #2 stand rolling mill 12 perform feedforward control using the detection results of the plate thickness gauges 41 to 43 on the inlet side, and Feedback control using the detection results 42 to 44 is performed. For example, the plate thickness control device 62 performs feedforward control using the detection results of the inlet side plate thickness gauge 41, and further performs feedback control using the detection results of the outlet side plate thickness gauge 42. .

In addition, the tension control devices 71 to 73 of the #1 to #3 stand rolling mills 11 to 13 control the rolls of the next stage stand rolling mills 12 to 14 based on the tension detected by the tensiometers 51 to 55 on the exit side. Find the gap S. The roll gap control devices 32 to 34 operate the positions of the work rolls according to the determined roll gap S. For example, the tension control device 71 determines the roll gap S of the #2 stand rolling mill 12 based on the tension detected by the tension meter 51 on the exit side of the #1 stand rolling mill 11, and the roll gap control device 32 determines the roll gap S of the #2 stand rolling mill 12. The position of the work roll of the #2 stand rolling mill 12 is operated based on the following.

Further, the tension control device 73 of the #4 stand rolling mill 14 controls the tension on the exit side of the #4 stand rolling mill 14 by controlling the speed of the exit bridle roll 15 via the motor speed control device 25.

FIG. 4 is a diagram showing an example of a basic control configuration for plate thickness control in the plate thickness control device 64. As shown in FIG. 4 (see also FIG. 2), the plate thickness control device 64 calculates the entry side plate thickness deviation ΔH measured by the plate thickness gauge 43 on the exit side of the #3 stand rolling mill 13, A transfer process is performed to delay the time _TFF until the measurement position of the material 3 reaches directly below the #4 stand rolling mill 14. Here, the measurement result of the entrance side plate thickness deviation ΔH is a control state quantity before rolling, and can be called a so-called pre-control state quantity.

Next, the plate thickness control device 64 multiplies the transfer processing result by a control gain _GFF to obtain a feedforward control amount. In addition, the plate thickness control device 64 multiplies the outlet side plate thickness deviation Δh measured by the plate thickness meter 44 on the outlet side of the #4 stand rolling mill 14 by a control gain G _FB and performs an integral process to obtain a feedback control amount. . The plate thickness control device 64 outputs the amount obtained by adding the feedforward control amount and the feedback control amount thus obtained to the motor speed control device 23 of the #3 stand rolling mill 13. Here, the measurement result of the exit plate thickness deviation Δh is a controlled state quantity after rolling, and can be called a so-called post-control state quantity.

It should be noted that plate thickness variation due to hardness unevenness cannot be detected directly below the #4 stand rolling mill 14 where it occurs, but is detected by a plate thickness meter 44 installed at a position away from the #4 stand rolling mill 14. Therefore, since there is a dead time from the occurrence of plate thickness variation to its detection, the integral control amount is included in the calculation of the feedback control amount.

The configurations of the plate thickness control devices 62 and 63 are similar to that of the plate thickness control device 64, so the description thereof will be omitted below. On the other hand, since the plate thickness control device 61 is for controlling the roll gap S of the #1 stand rolling mill 11, its configuration and control method are different from the plate thickness control device 64. However, in this embodiment, description of the configuration and control method of the plate thickness control device 61 will be omitted.

FIG. 5 is a diagram showing an example of a basic control configuration for tension control in the tension control device 73. As shown in FIG. 5 (see also FIG. 2), the tension control device 73 is configured to control the tension measured by the tension meter 53 installed between the #3 stand rolling mill 13 and the #4 stand rolling mill 14. The configuration is such that proportional-integral control is performed using the deviation ΔT ₃₄ between the value T _34FB and the tension command value T _34ref . In this integral control, since the control output is out of phase with the control state quantity by 90 degrees, the resulting sheet thickness h on the exit side of the #4 stand rolling mill 14 is different from the original hardness uneven position. The phase of the deviation Δh shifts.

≪2. Simulation based on basic control configuration≫
Next, simulation results of rolling phenomena in a tandem rolling mill having a four-stand configuration as shown in FIG. 1 will be explained using FIGS. 6 to 8. In the simulation, it was calculated how the plate thickness variation, tension variation, and load variation of the #4 stand rolling mill 14 would vary over time due to variation in deformation resistance, which is hardness unevenness.

FIG. 6 is a diagram showing an example of simulation results when neither plate thickness control nor tension control is performed. In addition, FIG. 7 shows that the tension control on the entry side and the exit side of the #4 stand rolling mill 14 is carried out by proportional-integral control, and only the feedback control of the plate thickness control on the exit side of the #4 stand rolling mill 14 is carried out. It is a figure showing an example of a simulation result in the case of doing so. In addition, FIG. 8 shows an example of simulation results when, in addition to the conditions shown in FIG. 7, feedback control is applied to the plate thickness control on the exit side of the #3 stand rolling mill 13 preceding the #4 stand rolling mill 14. This is a diagram.

In addition, in FIGS. 6 to 8, "plate thickness variation" indicates the variation in the inlet side plate thickness H (inlet side plate thickness deviation ΔH) as a solid line, and the variation in the outlet side plate thickness h (outlet side plate thickness deviation Δh) as a broken line. has been done. Similarly, for "tension fluctuations", fluctuations in inlet tension are shown by solid lines and fluctuations in exit tension are shown by broken lines, and for "load fluctuations", fluctuations in rolling load are shown by solid lines, and fluctuations in deformation resistance are shown by broken lines. has been done.
Also, time flows from the left side to the right side of the diagram, with the left end showing the current state and the right end showing the most past state.

In the simulation in the case of FIG. 6, the hardness unevenness directly appears as plate thickness variation. Therefore, the waveform peak positions of the fluctuations in deformation resistance, the fluctuations in the entrance side plate thickness H and the fluctuations in the exit side plate thickness h in the #4 stand rolling mill 14 coincide, and there is no deviation in their phase relationship (for example, vertical (see solid line position).

On the other hand, in the simulation in the case of FIG. 7, a phase advance occurs in which the phase of the variation in the plate thickness h on the exit side of the #4 stand rolling mill 14 is faster than the variation in the plate thickness on the input side. This is because the plate thickness control device 64 of the #4 stand rolling mill 14 performs integral control, resulting in a control output with a phase delay of 90 degrees, as shown in equations (1) to (3) and FIG. 25. This is because the phase shift amount Δ is negative due to the following relationship. As a result, the phase shift δ of the variation in the outlet side plate thickness h of the #4 stand, which is the result of plate thickness control, becomes positive.

In addition, in the simulation in the case of FIG. 8, since feedback control is also implemented in the plate thickness control of the #3 stand rolling mill 13 in the preceding stage of the #4 stand rolling mill 14, the change in the entrance side plate thickness H of the #4 stand rolling mill 14 is in a leading phase than the deformation resistance.

As described above, by performing predetermined control on fluctuation factors that the controlled object originally has, such as hardness unevenness, other fluctuation factors with different phases occur, and the difference between the controlled state quantity of the controlled object is generated. The phase relationship may change. This variation in phase relationship varies depending on the generation frequency of the control state quantity because the predetermined control response differs. In addition, in rolling mills, not only hardness unevenness but also changes in mechanical conditions due to changes in the thickness of the base material, eccentricity of the rolls used in the rolling mill, etc., and changes in the surface condition of the rolled material can cause changes in thickness. Occur. In addition, even hardness unevenness occurs at various length cycles due to multiple factors such as cooling unevenness and annealing unevenness in the paper process of the rolling mill, and multiple frequency components exist as state quantities before control of the rolling mill. However, the attenuation amount and phase difference of the inlet plate thickness deviation, which is a pre-control state quantity, and the outlet plate thickness deviation, which is a post-control state quantity, are different.

Normally, in a tandem rolling mill, plate thickness control is performed in each of the stand rolling mills 12 to 14, including the #1 stand rolling mill 11, so fluctuations in deformation resistance and the resultant change in exit side plate thickness h. This means that the phase is out of phase with the fluctuation (outgoing plate thickness deviation Δh). The phase shift differs depending on the frequency of plate thickness fluctuation. Therefore, when performing feedforward control using the entrance side plate thickness deviation ΔH of the stand rolling mill, a sufficient control effect cannot be obtained due to the influence of the phase shift between the deformation resistance fluctuation and the entrance side plate thickness deviation ΔH.

Conventionally, as a method for adjusting control parameters for feedforward control, the control output timing shift amount ΔT _FF for feedforward control in FIG. 4 is set in consideration of the dead time and response from the control output to the control operation end, and The control gain G has been changed based on the exit side plate thickness deviation Δh, which is the control result. However, when this method is used, a sufficient control effect is often not obtained because there is a phase difference between the entry side plate thickness deviation ΔH, which is the target control state quantity, and the deformation resistance fluctuation, which is the hardness unevenness. .

Furthermore, even though the entrance plate thickness deviation ΔH has a complex waveform that includes various frequency components, only one control output timing shift amount ΔT _FF and one control gain G can be set each, and the effect depends on the frequency components. The results showed different degrees of

As shown in equations (1), (2-1), (2-2) and FIG. 25, in feedforward control, the control output timing shift amount corresponding to the control gain G and the phase shift amount Δ It is necessary to appropriately set _ΔTFF . This setting needs to be determined in consideration of the rolling speed and other controls being implemented, resulting in a complicated adjustment. In the case of rolling speed, the frequency of plate thickness variation changes, so the response from the control output to the control operation end changes. Furthermore, in the case of a tandem rolling mill, the response differs depending on which rolling mill stand performs the plate thickness control and tension control.

Further, since the above differs depending on the type of control disturbance, it is not sufficient to change only one control gain G and one phase shift amount Δ.

A specific frequency component is selected from a plurality of inlet thickness fluctuation frequency components based on some standard, and the control gain G of feedforward control is determined using the inlet thickness deviation, the attenuation amount of the outlet thickness deviation, and the phase relationship of that frequency component. A method of adjusting the control output timing shift amount ΔTFF is shown in Patent Document 3. Therefore, it is possible to adjust the feedforward control for each frequency component, but the problem is how to adjust them when there are multiple inlet thickness deviation frequency components.

9A, 9B, 9C, and 9D show the results of a simulation in which the entrance side plate thickness deviation has three types of frequency components. The upper diagram in each figure shows the waveform in the time domain, and the lower diagram shows the waveform in the frequency domain.

FIG. 9A shows a case where the control gain and control output timing shift amount set by feedforward control using the detected value of entrance plate thickness deviation in a normal rolling mill are set in a state before feedforward AGC adjustment. A sine wave with three frequency components, frequency (A), frequency (B), and frequency (C), is used as the entrance plate thickness deviation (the time series waveform is displayed at the top of the figure). The FFT results of the entrance side plate thickness deviation and the exit side plate thickness deviation are shown at the bottom of the figure, and the plate thickness deviation amplitude is detected at frequency (A), frequency (B), and frequency (C).

FIG. 9B shows the results when the control gain and control output timing shift amount are adjusted using the single frequency shown in Patent Document 3 for the frequency component with the largest amplitude of the entrance side plate thickness deviation, and is shown in the lower part of the figure. It can be seen that the adjustment for the frequency component (C) was insufficient and variations in the thickness of the exit side remained. The upper part of the figure shows the time series data of plate thickness deviation, and it can be seen that a single frequency component remains in the exit plate thickness deviation.

FIG. 9C shows the result when the control gain and the control output timing shift amount are adjusted using the single frequency shown in Patent Document 3 for the frequency component with the largest amplitude of the exit side plate thickness deviation which is the control result. As shown in the lower part of FIG. 9C, it can be seen that the adjustment of the frequency component (A) and the frequency component (C) is insufficient, and variations in the outlet side plate thickness remain.

In the upper part of FIG. 9C, time-series data of plate thickness deviation is displayed, and it can be seen that multiple frequency components remain in the exit side plate thickness deviation. When the frequency component (A) of the outlet side plate thickness deviation decreases due to adjustment of the feedforward control, the control system is adjusted for the frequency component (C) of the outlet side plate thickness deviation, and as a result, when the frequency component (C) decreases, the frequency component Adjustment for frequency component (A) and frequency component (C) was repeated in such a way that (A) became the maximum frequency component of the exit side plate thickness deviation, and the control system was adjusted for frequency component (A). As a result, almost the same amount of exit side plate thickness deviation remains for the frequency component (A) and the frequency component (C).

FIG. 9D shows the results when the control gain and control output timing shift amount are adjusted for a plurality of frequency components in this embodiment. In the upper part of FIG. 9D, time-series data of plate thickness deviation is displayed, and it can be seen that the exit side plate thickness deviation is almost suppressed. Furthermore, looking at the frequency response waveform at the bottom of FIG. 9D, it can be seen that the frequency components (A), (B), and (C) have been sufficiently adjusted and the variation in the outlet side plate thickness is suppressed.

As described above, when performing feedforward control using the detected value of the inlet thickness deviation that includes multiple frequency components, only one control gain and one control output timing shift amount can be changed by adjusting the control system. , it is impossible to perform adjustment for all frequency components.

In addition to hardness unevenness, actual factors contributing to the thickness deviation on the entrance side include plate thickness variations due to mechanical vibrations (roll eccentricity, etc.) in the upstream process of the rolling mill, and plate thickness variations due to surface treatment unevenness. Depending on these factors, plate thickness fluctuations have characteristics. For example, hardness unevenness is caused by the hardness of the rolled material periodically fluctuating in the longitudinal direction of the rolled material, and requires a control gain larger than a normal feedforward control gain. Unlike hardness unevenness, plate thickness fluctuations due to machine fluctuations in the upstream process can be suppressed using normal feedforward control gain. Thickness fluctuations caused by surface treatment unevenness occur during the first rolling, but after that, they can basically be suppressed using the feedforward control gain for normal sheet thickness fluctuations, just like the thickness fluctuations caused by mechanical vibrations. It is.

From the above, when there are frequency components of the entrance plate thickness deviation caused by multiple factors, the optimal control gain and control output timing shift amount are different, so one control gain and control output timing shift amount can cover all frequency components. It is impossible to adjust the control gain and control timing shift amount for each frequency component.

For this purpose, we utilize the fact that the entrance side plate thickness deviation detection value is a combination of plate thickness deviations of multiple frequency components, separate the multiple plate thickness deviation frequency components, and control gain for each frequency component. , the control timing shift amount is set to correct the plate thickness deviation waveform of each frequency component, and finally they are combined to create an estimated waveform of the entrance side plate thickness deviation to be used in feedforward control. By doing so, it becomes possible to separately adjust the control gain of feedforward control and the control output timing shift amount for each frequency component.

The concept of a plant control system for realizing the above will be explained using FIGS. 10 and 11. FIG. 10 is a diagram showing a control method of a comparative example of a plant control system. FIG. 11 is a diagram showing a control method of the plant control system of this embodiment.

Here, a case is shown in which feedback control 306 and feedforward control 307 are performed on controlled object 300. A controlled disturbance d(t) (301) is input to the controlled object 300 via a disturbance transfer 305, which is the time during which the influence is transmitted from the disturbance occurrence position to the controlled object. Furthermore, a control operation amount 309 that is an output of various controls is also input to the controlled object 300, and as a result of these changes, the controlled object state amount 302 changes. The controlled object state quantity 302 is detected by the detector as an observed quantity 303 after a detection dead time 304, and feedback control 306 is performed on it. Regarding the control disturbance d(t) (301), if it is observable, the control disturbance transfer value d'(t) (310) with the same delay time as the disturbance transfer 305 is used in the control disturbance transfer 308. Then, the feedforward control 307 performs control to suppress the influence on the controlled state quantity 302.

Although the feedforward control 307 has a large disturbance suppression effect, it is impossible to remove the offset error from the control target value that occurs in the controlled object state quantity 302, so the feedback control 306 removes the offset error.

In the plant control system of the comparative example shown in FIG. 10, the control disturbance 301 is transferred by the control disturbance transfer 308, and the feedforward control 307 performs control at the same timing as the control disturbance 301 affects the controlled object 300. Therefore, when the control disturbance 301 includes a plurality of frequency components, it is impossible to adjust the control gain and control output timing separately for them. Incidentally, in FIG. 10, the control gain corresponds to _GFF in the feedforward control 307, and the control output timing corresponds to _ΔFF in the control disturbance transfer 308.

FIG. 11 is a diagram showing a control method of the plant control system of this embodiment. In FIG. 11, the difference from the plant control system of the comparative example in FIG. Rather, the disturbance frequency components included in the control disturbance d(t) (301) are corrected for each frequency, and the feedforward control 307 is performed using the control disturbance composite correction value _dFFEST (t) (350), which is a combination of the disturbance frequency components. This point is being implemented. Therefore, it is possible to implement the control gain/timing shift amount adjustment method for any frequency component, and the control accuracy of the feedforward control 307 can be improved.

From the control disturbance d(t) (301), frequency analysis is performed using FFT (Fourier transform), frequency components with large amplitudes are extracted, and a control disturbance provisional value creation device 320 generates a composite waveform of them. A tentative control disturbance value d _DUMMY (t) (321) is created. Here, it is assumed that the control disturbance d(t) (301) has n frequency components ωi, and the amplitude of the frequency component ωi is A(ωi). The control disturbance provisional value d _DUMMY (t) (321) does not include the phase difference of each frequency component. By performing FFT on the control disturbance d(t) (301), it is also possible to obtain the phase difference between each frequency component. However, since FFT is actually implemented using DFT (digital Fourier transform), the frequency component ωi includes a quantization error, so the control disturbance provisional value d _DUMMY (t) (321) is the control disturbance d(t) (301 ) will result in a slightly different waveform.

Therefore, in order to get as close as possible to the control disturbance transfer value d'(t) (310), the control disturbance composite value creation device 325 generates D'( ω) and D _DUMMY (ω) which is the FFT result of the control disturbance provisional value d _DUMMY (t) (321). (including δB(ω)), and corrects the control disturbance provisional value d _DUMMY (t) (321) with the response B(ω) to obtain the control disturbance composite value d' _DUMMY (t) (326).

By doing the above, the control disturbance composite value d' _DUMMY (t) (326) has a waveform that approximates the control disturbance transfer value d'(t) (310). The degree of approximation changes depending on the number of frequency components n of the control disturbance d(t) (301), but the value of n is, for example, the control disturbance composite value d' _DUMMY (t) (326) It is possible to determine this by, for example, using the correlation coefficient of the transfer value d'(t) (310).

In the feedforward control adjustment device 101, a control gain/timing shift amount adjustment method is determined based on the control disturbance d(t) (301) and the observed quantity 303 of the controlled object state quantity x(t) (302) of the controlled object 301. Implement.

As adjustment results, the control gain G _FF (ωi) and control output timing shift amount δ _FF (ωi) for each frequency component ωi are determined. The control disturbance composite correction value creation device 330 uses these values to correct the control disturbance composite value d' _DUMMY (t) (326) to obtain the control disturbance composite correction value d _FFEST (t) (350). create. The control disturbance synthesis correction value d _FFEST (t) (350) is the control gain G _FF (ωi) so that the feedforward control effect is maximized for a typical frequency component ωi of the control disturbance d(t) (301). , the control output timing shift amount δ _FF (ωi) is corrected. By implementing the feedforward control 307 using this, it becomes possible to maximize the control effect of the feedforward control.

Hereinafter, a case will be described in which the control method of this embodiment is applied to feedforward control of a tandem rolling mill described in the example of Patent Document 3.

<Control gain/timing shift amount adjustment>
In feedforward control, it is important to appropriately set the control output timing shift amount ΔT _FF (phase shift amount Δ) and control gain G; They are connected by the relationship explained using 2). For example, when the control gain G is changed, the phase difference δ between the control state quantities before and after the control also changes. Conversely, if the control output timing shift amount _ΔTFF is changed, the amplitude X of the control state amount will also vary. Therefore, it is actually difficult to adjust the settings so that both are set appropriately.

As shown in equation (2-2) above, the phase difference δ between the control state quantities before and after the control is an arctangent function, so it varies from -90 degrees to +90 degrees with respect to -∞ to +∞. Let degree be the domain. Also, as is clear from equation (2-2), if it exceeds +∞ and becomes -, it becomes larger than 90 degrees, so the phase difference δ is conveniently set to exceed 90 degrees as shown in Fig. 25. It is said that Furthermore, according to equation (2-2), if the control gain G is not greater than 1, the phase difference δ between the control state quantities does not exceed 90 degrees. Therefore, if the phase difference δ between the control state quantities exceeds 90 degrees, it can be predicted that the control gain G is too large.

In addition, since the phase shift amount Δ and the phase difference δ between the control state quantities before and after the control are in opposite directions, if the phase difference δ between the control state quantities is known, the phase shift amount Δ, in other words, the control output timing It is possible to predict how the shift amount _ΔTFF will be changed. For example, when the phase difference δ between the control state quantities is in the positive direction, the phase shift amount Δ may be changed in the increasing direction, that is, in the direction from the negative side to the positive side. In the opposite case, the phase shift amount Δ may be changed in the decreasing direction, that is, from the plus side to the minus side.

As described above, in the case of feedforward control in plate thickness control, the phase relationship between the inlet plate thickness deviation ΔH detected by the inlet plate thickness gauge 43 and the outlet plate thickness deviation Δh detected by the outlet plate thickness gauge 44 is , can be regarded as the phase difference δ between the control state quantities. Similarly, the control output timing shift amount _ΔTFF from the entrance side plate thickness deviation ΔH to the control output can be regarded as the phase shift amount Δ. Therefore, using these control state quantities, the control output timing shift amount ΔT _FF and control gain G _FF in feedforward control can be adjusted.

Therefore, a configuration in which a function of adjusting the control output timing shift amount ΔT _FF and control gain G _FF is added to the basic control configuration of the plate thickness control device 64 shown in FIG. It is called composition. Furthermore, a configuration in which a function of adjusting the control output timing shift amount ΔT _FF and control gain G _FF for multiple frequency components, which is a feature of this embodiment, is added will be hereinafter referred to as multiple frequency extended control of the plate thickness control device 64. It is called composition.

≪3. Multiple frequency expansion control configuration≫
<3.1 Feedforward control adjustment device>
FIG. 12 is a diagram showing an example of a multi-frequency extended control configuration of the plate thickness control device 64, feedforward control adjustment device 101, and multi-frequency control adjustment device 400 according to the embodiment of the present invention. Here, the feedforward control adjustment device 101 is a device that calculates the control output timing shift amount ΔT _FF and control gain G _FF for the feedforward control performed by the multiple frequency control adjustment device 400. That is, the multi-frequency control adjustment device 400 is a device that realizes a multi-frequency extended control configuration of the plate thickness control device 64, and is a major feature of this embodiment.

As shown in FIG. 12, in the feedforward control adjustment device 101, in response to the entry side plate thickness deviation ΔH detected by the plate thickness gauge 43 on the entry side of the #4 stand rolling mill 14, the Transfer processing is performed up to _the timing _of ) (410) to perform feedforward control. Here, the entry side plate thickness deviation composite correction value ΔH _4FFEST (0) is used.

The multi-frequency control adjustment device 400 detects the detected value ΔH ₄ of the thickness deviation on the entrance side, the value ΔH _4MILL of the thickness deviation on the #4 stand that has been transferred to the #4 stand directly below the rolling mill 14, and the feedforward control adjustment device 101. An entry side plate thickness deviation composite correction value ΔH _4FFEST (tcal) (410) is created from the obtained control output timing shift amount ΔT _FF and control gain G _FF , and is output to the plate thickness control device 64.

The multi-frequency control adjustment device 400 is an entrance side plate thickness deviation provisional value creation device 420 that has the same functions as the control disturbance provisional value creation device 320, the control disturbance composite value creation device 325, and the control disturbance composite correction value creation device 330 in FIG. , an inlet thickness deviation composite value creation device 425, and an inlet thickness deviation composite correction value creation device 430.

FIG. 13 is a diagram showing the configuration of the entry side plate thickness deviation provisional value creation device 420. Using the input side plate thickness deviation detection value ΔH ₄ as input, sampling is performed at the sampling period Δt of the computer, and an input side plate thickness deviation table 4201 is created. It is assumed that the entry side thickness deviation table 4201 stores m pieces of data, and the m pieces of data are set at appropriate times according to the frequency component of the expected entry side plate thickness deviation.

The entry side plate thickness deviation FFT device 4202 performs FFT processing using the data of the entry side plate thickness deviation table 4201 to calculate the entry side plate thickness deviation amplitude H _4G (f) in the frequency space. The inlet thickness frequency selection device 4203 uses the inlet thickness deviation amplitude H _4G (f) to select the frequency component fi (selected inlet thickness frequency fi(i=1,n) (4206 )). The selection method is arbitrary, but for example, it is possible to select n pieces (any predetermined value) in descending order of the entry side plate thickness deviation amplitude H _4G (f), or to set a threshold value in advance and select the entry side plate thickness from the threshold value. Possible methods include selecting one with a large deviation amplitude H _4G (f), performing FFT on the exit side plate thickness deviation, and selecting n pieces in order of decreasing attenuation rate. According to the selected inlet thickness frequency fi (i=1,n) (4206), the temporary time series data creation device 4204 performs calculations at each sampling period Δt based on the calculation formula of ΔH _4DUMMY (t), and m The calculation results are set in the entry side plate thickness deviation tentative value table ΔH _4DUMMY (j) (4205). Here, t=0 is stored at the beginning of the table, and thereafter, the value of (m-1)×Δt is stored at the end of the table.

FIG. 14 is a diagram showing the configuration of the entrance side plate thickness deviation composite value creation device 425. A sampling process is performed on the entry side plate thickness deviation mill value ΔH _4MILL , and an entry side plate thickness deviation table ΔH _4MILL (j) (4251) having m storage areas is created. By storing the current sampling value at the beginning of the table and sequentially shifting the stored data, the data from (m-1)×Δt seconds ago is stored in the last table. As a result, the side plate thickness deviation tentative value table ΔH _4DUMMY (j) (4205) (see Fig. 13) and the entry side plate thickness deviation table ΔH _4MILL (j) (4251) just below the mill have a time series starting point directly below the mill. The complete data is stored.

Using these two tables, the respective frequency space values X(ω) and Y(ω) are determined by the entry side plate thickness deviation provisional value FFT device 4252 and the entry side plate thickness deviation right below the mill FFT device 4253, and the entry side deviation provisional value The correction device 4254 obtains those responses, obtains the attenuation rate H _B (fi) and the phase difference Δ _B (fi) at the selected inlet plate thickness frequency fi (4206), and uses them to calculate the provisional time series data ΔH _4DUMMY (t) is corrected, and the entrance side plate thickness deviation composite value ΔH' _DUMMY (t) is obtained. Through the above processing, the inlet side plate thickness deviation composite value ΔH' _DUMMY (t) becomes a reproduction of the inlet side plate thickness deviation value ΔH _4MILL (t) just below the mill within the selected frequency component range.

The FFT used here is DFT (Digital Fourier Transform), and the frequency of the calculation result is also discretized, so strictly speaking, it is the frequency component included in the input side plate thickness deviation value ΔH _4MILL (t) directly below the mill. is different. Therefore, the processing of the input side plate thickness deviation provisional value creation device 420 (see FIG. 13) and the input side plate thickness deviation composite value creation device 425 is always executed, and the input side plate thickness deviation composite value ΔH' _DUMMY (t) is generated in the vicinity directly below the mill. It is necessary to match the inlet side plate thickness deviation value ΔH _4MILL (t) just below the mill within a range that does not cause problems in feedforward control.

Note that in FIG. 12, the feedforward control adjustment device 101 and the multiple frequency control adjustment device 400 are depicted as separate devices provided outside the plate thickness control device 64; It may be a device included.

<3.2 Control gain/timing shift amount setting device>
FIG. 15 is a diagram showing an example of a detailed configuration of the control gain/timing shift amount setting device 102. As shown in FIG. 15, the control gain/timing shift amount setting device 102 includes a frequency response measuring device 201, three membership functions 105, 106, 107, a fuzzy inference device 108, a parameter changing device 109, etc. Ru.

As described above, the control gain/timing shift amount setting device 102 inputs the inlet thickness deviation ΔH _TRK , the outlet thickness deviation Δh, the rolling load _PTRK , and the outlet thickness deviation Δh, and sets the selected inlet thickness frequency as described above. A control gain G _FF (fi) for feedforward control and a control output timing shift amount ΔT _FF (fi) in fi (4206) are calculated. The calculated control gain G _FF (fi) and control output timing shift amount ΔT _FF (fi) are output to the entrance side plate thickness deviation composite correction value creation device 430 (see FIG. 12).

The entrance side plate thickness deviation composite correction value creation device 430 (see FIG. 12) uses _the control gain G _FF (fi) to generate an input side plate thickness deviation composite correction value ΔH _4FFEST ( 0) (410). The plate thickness control device 64 (#4 stand plate thickness control) performs feedforward control by multiplying the entrance side plate thickness deviation composite correction value ΔH _4FFEST (0) (410) by the control gain G _FF . That is, the control gain/timing shift amount setting device 102 plays a role of setting and adjusting control parameters during feedforward control in the plate thickness control device 64. This is one of the major features of this embodiment that is not found in the prior art.

The purpose of the feedforward control in the plate thickness control device 64 is to make the outlet side plate thickness deviation Δh smaller than the inlet side plate thickness deviation ΔH. Therefore, when the feedforward control works properly, the outlet side plate thickness deviation Δh becomes small. However, when the outlet side plate thickness deviation Δh becomes smaller, it becomes difficult to judge the phase relationship between the inlet side plate thickness deviation ΔH and the outlet side plate thickness deviation Δh. In that case, it may be difficult to obtain the control gain G _FF (fi) for feedforward control and the control output timing shift amount ΔT _FF (fi). Therefore, in the control gain/timing shift amount setting device 102 according to the present embodiment, the control gain G for feedforward control is also used by using the phase relationship between the rolling load P _TRK , which is affected by hardness unevenness, and the entrance plate thickness deviation ΔH _TRK . _FF (fi) and control output timing shift amount ΔT _FF (fi) are calculated. This is also one of the major features of this embodiment.

Therefore, the control gain/timing shift amount setting device 102 determines the attenuation amount and phase relationship between time-series signals such as the inlet side plate thickness deviation ΔH _TRK and the outlet side plate thickness deviation Δh, the inlet side plate thickness deviation ΔH _TRK and the rolling load P _TRK . It is necessary to find out.

In the invention disclosed in Patent Document 2, a "squared error for one period" is calculated while shifting the phases of two time series signals, and the phase that minimizes the squared error is calculated as the phase difference between the two time series signals. There is. In addition to the need to recognize one period of the reference signal, this method is applicable when the amplitudes of the reference signal and comparison signal differ significantly due to control effects, or when multiple frequency components overlap. Sometimes it became difficult. Therefore, in this embodiment, a frequency response method is used that allows the signal attenuation amount and phase relationship between two time-series signals to be determined relatively easily.

(Reference 1: Regarding frequency response method)
FIG. 16 is a diagram for explaining the outline of the frequency response method, in which (a) is a diagram showing an example of a time response model, and (b) is a diagram showing an example of a frequency response model. In the rolling control, the material to be rolled 3 enters, for example, from the entry side of the #4 stand rolling mill 14 and comes out from the exit side of the #4 stand rolling mill 14 after its thickness is reduced by a rolling phenomenon. That is, the entrance side plate thickness deviation ΔH of the rolled material 3 changes to the exit side plate thickness deviation Δh due to the rolling phenomenon.

Here, as shown in FIG. 16(a), if the time change of the entrance side plate thickness deviation ΔH is represented by x(t), and the time change of the exit side plate thickness deviation Δh is represented by y(t), the rolling phenomenon is as follows. It can be expressed as a time response function g(t) that satisfies y(t)=g(t)·x(t). In other words, the entrance thickness deviation x(t), which is a time-space signal (time-series signal), is converted into the exit thickness deviation y(t), which is a time-space signal, by the time response function g(t) of the rolling phenomenon. be done.

The rolling phenomenon represented by such a time response function g(t) can be represented using a frequency response function G(ω) shown in FIG. 16(b). In other words, if the inlet side plate thickness deviation ΔH and the outlet side plate thickness deviation Δh are expressed as the inlet side plate thickness deviation X(ω) and the outlet side plate thickness deviation Y(ω), which are signals (frequency component values) in the frequency space, The relationship can be expressed as Y(ω)=G(ω)·X(ω). That is, the inlet plate thickness deviation X(ω), which is a signal in the frequency space, is converted into the outlet plate thickness deviation Y(ω), which is a signal in the frequency space, by the frequency response function G(ω) of the rolling phenomenon.

The entrance thickness deviation x(t) and the exit thickness deviation y(t) in time and space are detected, for example, by the entrance thickness gauge 43 and the exit thickness gauge 44 of the #4 stand rolling mill 14. It can be obtained as a time series signal. On the other hand, the inlet plate thickness deviation X(ω) and the outlet plate thickness deviation Y(ω) in the frequency space are obtained by Fourier transforming x(t) and y(t) obtained in the time space, respectively.

The advantage of expressing the rolling phenomenon using the input signal X(ω), output signal Y(ω), and frequency response function G(ω) in the frequency space is that the relationship between the amplitude and phase of the input signal and output signal can be expressed for each frequency. The purpose is to make it easier to compare. That is, in the frequency space, the attenuation amount and phase difference of the plate thickness deviation signal due to the rolling phenomenon can be easily determined.

That is, in this embodiment, the entrance side plate thickness deviation x(t) and the exit side plate thickness deviation y(t) are obtained as values detected by the plate thickness meters 43 and 44. In addition, the inlet thickness deviation X(ω) and the outlet thickness deviation Y(ω) in the frequency space are obtained by Fourier transforming the inlet thickness deviation x(t) and the outlet thickness deviation y(t), respectively. . Then, the frequency response function G(ω) can be determined by the following equation (3).

Furthermore, from this frequency response function G(ω), the attenuation amount gain and the phase difference phase at the frequency ω can be determined by the following equations (4-1) and (4-2).

(Reference 2: About discrete Fourier transform and FFT)
Here, the Fourier transform (discrete Fourier transform) of discrete data used to obtain the inlet side plate thickness deviation X(ω) and the outlet side plate thickness deviation Y(ω) in the frequency space will be explained. Generally, when a time series signal f(t) whose one cycle consists of N pieces of sampling data is expressed using a sine wave signal with N independent frequencies of k, it is expressed as the following equation (5). Ru.

Here, if the number n = 0, 1, ..., N representing the order of sampling data for one cycle is associated with the time t representing the phase from 0 to 2π, it is expressed as t = 2π・n/N. be able to. Therefore, equation (5) can be expressed as the following equation (6).

Then, by performing discrete Fourier transform on equation (6), the following equation (7) is obtained.

Here, the coefficient _cm is a complex number. Furthermore, in equation (7), 2πm/N corresponds to frequency. That is, the coefficient _cm represents the frequency component of the time series signal f(t) expressed by equation (5) when the frequency is 2πm/N. Therefore, the absolute value and argument of the coefficient _cm represent the amplitude and phase of the frequency component of the time series signal f(t) when the frequency is 2πm/N, respectively.

Furthermore, when the discrete Fourier transform is processed by a computer, a fast Fourier transform (hereinafter abbreviated as FFT) is usually used. FFT requires that the number of data to be transformed be a power of 2 as a condition for application, but has the great advantage that the amount of calculation is significantly smaller than ordinary discrete Fourier transform.

In general, when performing Fourier transform on N pieces of data, ordinary discrete Fourier transform requires a calculation amount proportional to N ² , but it is known that with FFT, the calculation amount is proportional to N log ₂ N. It is being For example, when Fourier transform is performed on 1024 pieces of data, the amount of calculation for FFT is compared to the amount of calculation for normal discrete Fourier transform.
log ₂ 1024/1024=10/1024
becomes. That is, the amount of calculation required for FFT is about 1/100 of that of ordinary discrete Fourier transform.

<3.3 FFT frequency resolution and data collection time>
As described above, since FFT requires a power of 2 data count, there are also restrictions on the time interval of data sampling (sampling interval). Here, if the sampling frequency which is the reciprocal of the sampling interval is f _s , and the number of samplings (number of data) is N, the frequency resolution Δf can be calculated by Δf=f _s /N, and the data collection time MT is , MT=N/f _s =1/Δf.

Here, the data collection time MT refers to the time from the start to the end of sampling of data to be input to the FFT, and the frequency resolution Δf refers to the resolution in the frequency axis direction when performing the FFT. Further, the theoretical maximum frequency f _r that can resolve two frequency components in data sampled at the sampling frequency f _s is given by f _r =f _s /2. That is, two frequency components cannot be separated unless they are separated by at least twice the frequency resolution Δf.

It is better for both the frequency resolution Δf and the data collection time MT to be small. However, since there is a relationship of MT=1/Δf as described above, it is not possible to reduce both at the same time. Therefore, when using FFT, it is important to set the frequency resolution Δf and the data acquisition time MT to practically appropriate values.

By the way, the purpose of the feedforward control of plate thickness control according to the embodiment of the present invention is to adjust the control gain G and phase shift amount Δ to increase the control effect. To this end, it is necessary to perform calculations at as short time intervals as possible, and it is also necessary to make the data collection time MT as short as possible.

On the other hand, if the disturbance regarding the entrance side plate thickness, that is, the entrance side plate thickness deviation ΔH, includes multiple frequency components, if the frequency of each disturbance cannot be separated, the attenuation amount gain and phase difference in the frequency components of each disturbance Unable to calculate phase. Therefore, it is necessary to select a data collection time that satisfies these conditions.

FIGS. 17 and 18 are diagrams showing examples of frequency response simulation results using FFT. The simulation results in FIG. 17A are for a case where the data collection time MT is 10.24 seconds and the frequency resolution Δf is about 0.1 Hz. In this simulation, sine waves of 0.5 Hz, 1.0 Hz, 2.0 Hz, and 3.0 Hz were mixed and used as an input signal as the entrance side plate thickness deviation ΔH, which means disturbance to the plate thickness. The sine wave input at this time is set so that the attenuation amount gain and phase difference phase of the output side plate thickness deviation Δh, which is the output signal, are as follows at each of the above-mentioned frequencies.
(Frequency) (Attenuation gain) (Phase difference phase)
0.5Hz -6.0dB 60 degrees 1.0Hz -4.4dB -45 degrees 2.0Hz -3.1dB -30 degrees 3.0Hz -1.9dB 30 degrees

In FIG. 17(a), the upper graph shows the time change of the inlet thickness deviation ΔH and the outlet thickness deviation Δh in the time space, and the lower graph shows the inlet thickness deviation in the frequency space after performing FFT. It is a graph showing the frequency characteristics of ΔH and outlet side plate thickness deviation Δh (output signal). The lower graph also shows the attenuation amount gain and the phase difference phase.

As can be seen from the frequency space graph in the lower part of Fig. 17(a), the four frequency components are clearly separated even in the exit plate thickness deviation Δh, and the attenuation amount gain and phase difference phase are also accurately determined. . However, judging from the fact that the minimum value of the frequency of the disturbance, that is, the entrance side plate thickness deviation ΔH, is 0.5 Hz, it must be said that the data collection time of 10 seconds is long. That is, in this case, it takes more than five cycles of the plate thickness variation cycle (2 seconds) to adjust the feedforward AGC (Automatic Gain Control).

The simulation result in FIG. 17(b) is a case where the data collection time MT is 5.12 seconds and the frequency resolution Δf is about 0.2 Hz. The entrance plate thickness deviation ΔH input in the simulation in this case is the same as in the case of FIG. 17(a), and the display format of the graph representing the simulation result is also based on FIG. 17(a).

As can be seen from the frequency space graph in the lower part of Fig. 17(b), the four frequency components included in the input signal are almost clearly separated even with the exit plate thickness deviation Δh, and both the attenuation amount gain and the phase difference phase are generally accurate. is required. Note that this figure shows that if the minimum separation width (0.5 Hz) of any two frequencies included in the input signal is more than twice the frequency resolution Δf (0.2 Hz), the attenuation gain and phase difference phase This is an example showing that it is possible to obtain almost exactly.

The simulation result in FIG. 18(c) is a case where the data collection time MT is 2.56 seconds and the frequency resolution Δf is about 0.4 Hz. The entry side plate thickness deviation ΔH input in the simulation in this case is the same as in the case of FIG. 17(a), and the display format of the graph representing the simulation result is based on FIG. 17(a).

As can be seen from the frequency space graph in the lower part of FIG. Both ΔH and exit side plate thickness deviation Δh are not sufficiently separated. Therefore, both the obtained attenuation amount gain and the phase difference phase are inaccurate.

The simulation results in Fig. 18(d) are for a case where the data collection time MT is 2.56 seconds and the frequency resolution Δf is about 0.2 Hz. A 0.5Hz sine wave is input. In this case, as shown in the lower graph of Fig. 18(d), the frequency of 0.5 Hz is correctly separated even with the exit plate thickness deviation Δh, and both the attenuation amount gain and the phase difference phase are almost accurately calculated. It is being

In addition, the data collection time MT in this case is 2.56 seconds, which is the time of 2 seconds + α in which the frequency 0.5 Hz of the outlet side plate thickness deviation Δh can be reproduced, and the feedforward AGC adjustment can be performed in almost the minimum time. It turns out that it is possible.

By the way, in FFT, the calculation time is significantly reduced by limiting the number of data to be processed to a power of two. Therefore, the number of input data to the FFT cannot be set to an arbitrary number. Therefore, the data collection time MT varies significantly depending on the combination of the data sampling period and the number of data.

For example, consider a case where the frequency resolution Δf is 0.1 Hz (period is 10 seconds). In this case, if the sampling period is 10 ms and the number of data is 1024, the data collection time MT will be 10.24 seconds. This data collection time MT is approximately equal to the period of 10 seconds obtained from the frequency resolution Δf=0.1 Hz. On the other hand, if the sampling period is 8 ms and the number of data is 2048, the data collection time MT will be 16.384 seconds, which is significantly longer than the period of 10 seconds.

Next, consider a case where the frequency resolution Δf is 0.5 Hz (period is 2 seconds). In this case, if the sampling period is 10 ms and the number of data is 256, the data collection time MT will be 2.56 seconds, which is longer than the period of 2 seconds. Therefore, if the sampling period is 8 ms and the number of data is 256, the data collection time MT will be 2.048 seconds, which is almost the same as the period of 2 seconds.

FIG. 19 is a diagram showing an example of a sampling period/data number search table. As shown in FIG. 19, the sampling period/data number search table is a table storing the sampling period and data number that can obtain the actual data collection time most suitable for the minimum data collection time according to the frequency resolution Δf. It is. Here, "most suitable" means greater than the minimum data collection time and closest to the minimum data collection time.

In order to adjust the control parameters of feedforward control using the frequency response method using FFT, it is necessary to adjust the attenuation amount gain and phase difference phase of the outlet side plate thickness deviation Δh with respect to the inlet side plate thickness deviation ΔH at the corresponding frequency. It is necessary to obtain it as fast as possible (in a short period of time). For this purpose, it is important to minimize the data collection time MT used for FFT. The data collection time MT is determined based on the minimum resolution required according to the actually occurring plate thickness deviation (inlet side plate thickness deviation ΔH), and the number of samplings and data required for this are set. be done.

<3.4 Frequency response measurement device>
Next, the frequency response method is used to obtain the control gain G _FF (fi) and the control output timing shift amount ΔT _FF (fi) at the selected inlet thickness frequency fi (4206) in feedforward control of thickness control. I will explain about it. According to the frequency response method, the amplitude at the data collection time MT can be determined by setting the data collection time MT and performing FFT processing on the entrance side plate thickness deviation ΔH and the exit side plate thickness deviation Δh. As mentioned above, hardness unevenness is a change in hardness in the longitudinal direction of the rolled material 3 and occurs every time it is rolled, so it is caused by a change in thickness that occurred in the previous rolling due to hardness unevenness. A certain inlet plate thickness deviation ΔH and an outlet plate thickness deviation Δh, which is a variation in plate thickness after rolling, have almost the same frequency. Further, unlike the normal entrance side plate thickness deviation ΔH, the exit side plate thickness deviation Δh due to hardness unevenness is expected to have a small attenuation amount.

Therefore, by determining the control gain G _FF (fi) and the control output timing shift amount ΔT _FF (fi) in the feedforward control using the following procedure, it is possible to efficiently adjust the control parameters of the feedforward control. can. In the following description, the entry side plate thickness deviation ΔH often means the entry side plate thickness deviation ΔH _TRK after the transfer process, but even in that case, it will simply be written as the entry side plate thickness deviation ΔH.

(Step 1) Perform FFT processing on the inlet side plate thickness deviation _ΔHTRK and the outlet side plate thickness deviation Δh. Note that this FFT processing is performed at a cycle according to the frequency resolution Δf necessary for detecting a predetermined plate thickness disturbance.
The frequency at which the deviation of the selected inlet plate thickness frequency fi 4206 is the smallest is determined and set as the disturbance identification frequency resolution Δf _c .
(Procedure 2) A minimum data collection time is determined based on the disturbance identification frequency resolution Δf _c , and the number of samplings and sampling period are set in consideration of FFT.
(Step 3) Perform FFT with the set sampling number and sampling period to determine the attenuation amount and phase relationship at the selected entrance thickness frequency fi 4206 for the entrance thickness deviation _ΔHTRK and the exit thickness deviation Δh.
(Step 4) Find the control gain G _FF (fi) for feedforward control and the control output timing shift amount ΔT _FF (fi) based on the phase relationship, and output these to the inlet side plate thickness deviation composite correction value creation device 430. .

The above steps 1 to 4 are performed by the control gain/timing shift amount setting device 102 shown in FIG. 15. That is, the plate thickness disturbance measuring device 202 configuring the frequency response measuring device 201 implements step 1, the plate thickness disturbance estimating device 203 implements step 2, and the frequency response estimating device 204 performs step 3 and step 4. Implement. Further, step 4 is performed by the membership functions 105 to 107, the fuzzy inference device 108, and the parameter change device 109. Then, by repeating the above steps 3 and 4, _the control parameter (i.e., control gain G _FF (fi) The feed forward control of the plate thickness control device 64 (#4 stand plate thickness control) is adjusted by adjusting the transfer time ΔT _FF (fi)) of the entrance plate thickness deviation ΔH.

The detailed configuration and control contents of the frequency response measuring device 201 that constitutes the control gain/timing shift amount setting device 102 will be described below. As shown in FIG. 15, the frequency response measurement device 201 includes a thickness disturbance measurement device 202, a thickness disturbance estimation device 203, and a frequency response estimation device 204.

FIG. 20 is a diagram showing an example of the configuration of the plate thickness disturbance measuring device 202. As shown in FIG. 20, the plate thickness disturbance measurement device 202 includes an inlet thickness deviation table 2021, an outlet thickness deviation table 2022, an inlet thickness deviation FFT device 2023, an outlet thickness deviation FFT device 2024, and the like. Ru.

Generally, the frequency of a disturbance that causes plate thickness deviation (hereinafter referred to as plate thickness disturbance) differs depending on not only the rolling speed but also the type of plate thickness disturbance. Here, the frequency resolution Δf is set to 0.1 Hz considering that, for example, a plate thickness disturbance having a frequency of 0.5 Hz or more is removed from the plate thickness deviation. Note that these values can be appropriately set and changed by the user depending on the actual plate thickness disturbance situation and operating conditions.

When the frequency resolution Δf is 0.1 Hz, the minimum data collection time is 10 seconds. Therefore, when referring to the sampling period/data number search table shown in FIG. 19, sampling period=0.01 seconds and data number=1024 are obtained. Subsequent processing such as FFT in the plate thickness disturbance measuring device 202 is performed using these numerical values.

A storage device (not shown) of the plate thickness disturbance measuring device 202 is provided with an inlet plate thickness deviation table 2021 and an outlet plate thickness deviation table 2022, each of which can store 1024 pieces of data. Then, ΔH _TKR , which is the inlet side plate thickness deviation ΔH after the transfer process, and the outlet side plate thickness deviation Δh are inputted to the plate thickness disturbance measuring device 202 at every sampling period of 0.01 seconds, and each number is 1023 from address 0 of the table. The addresses are written in order.

When the writing of data to the inlet thickness deviation table 2021 and the outlet thickness deviation table 2022 is completed, the inlet thickness deviation FFT device 2023 executes FFT processing using the data written in the inlet thickness deviation table 2021 as input data. do. Similarly, the outlet thickness deviation FFT device 2024 executes FFT processing using the data written in the outlet thickness deviation table 2022 as input data. As a result of these FFT processes, an inlet thickness deviation frequency component H(f) and an outlet thickness deviation frequency component h(f) are obtained.

Here, the values of the inlet side plate thickness deviation frequency component H(f) and the outlet side plate thickness deviation frequency component h(f) when the frequency f=m·Δf (Δf is the frequency resolution) are calculated by the formula ( It is obtained by calculating _cm defined in 7). In this case, the data of the time series signal f(n) included in equation (7) is given by the inlet side plate thickness deviation table 2021 and the outlet side plate thickness deviation table 2022, respectively.

Therefore, when the frequency f=m・Δf, the attenuation amount and phase difference of the amplitude of the time series signal f(n), that is, the inlet side plate thickness deviation ΔH and the outlet side plate thickness deviation Δh, are calculated by the above equation (4-1). and (4-2), it can be expressed as in formulas (8-1) and (8-2).

As a result of the above processing, in the plate thickness disturbance measuring device 202, the inlet plate thickness deviation amplitude Hg(m) and the inlet plate thickness deviation phase Hp(m) are output from the inlet plate thickness deviation FFT device 2023. Similarly, the exit side plate thickness deviation amplitude hg(m) and the exit side plate thickness deviation phase hp(m) are outputted from the exit side plate thickness deviation FFT device 2024.

FIG. 21 is a diagram showing an example of the frequency dependence characteristics of the inlet thickness deviation amplitude Hg(m) and the outlet thickness deviation amplitude hg(m). In other words, FIG. 21 is a diagram in which the horizontal axis is the frequency, and the vertical axis is the values of the inlet side plate thickness deviation amplitude Hg(m) and the outlet side plate thickness deviation amplitude hg(m) for each frequency, indicated by broken lines and solid lines. This is an example. Hereinafter, with reference to FIG. 21, the contents of the processing executed by the plate thickness disturbance estimation device 203 will be described.

In the example of FIG. 21, the broken line graph representing the entry side plate thickness deviation amplitude Hg(m) corresponds to the frequency positions of (A), (B), and (C), that is, the frequencies are m _A・Δf, m _B・Δf It becomes a large value at the position where and m _C ·Δf. This indicates that the entrance side plate thickness deviation ΔH fluctuates due to plate thickness disturbances having these frequencies.

Normally, the actually measured inlet thickness deviation ΔH and outlet thickness deviation Δh include a noise component (actual noise or a portion that can be considered as noise). The deviation amplitude hg(m) also includes a noise component. Therefore, here, a noise level L _n is determined in advance for the inlet side plate thickness deviation amplitude Hg(m) and the outlet side plate thickness deviation amplitude hg(m). Then, it is determined that the control parameters of the feedforward control need to be adjusted for the frequency at which the outlet plate thickness deviation amplitude hg(m) exceeds the noise level _Ln .

In this embodiment, the selected entrance side plate thickness frequency fi is the disturbance frequency fc _i .
In the example of FIG. 21, m _A ·Δf, m _B ·Δf, and m _C ·Δf are determined as the disturbance frequency _fci . Here, i=1, 2. ..., which is an identification number when there are a plurality of disturbance frequencies fc _i .

As can be seen from the above description, in this embodiment, the adjustment target frequency fc is the selected entrance side plate thickness frequency fi. Therefore, it is necessary to appropriately adjust the control gain G _FF and the control output timing shift amount ΔT _FF , which are control parameters of the feedforward control, for the adjustment target frequency fc determined in this way. Note that the selection-side plate thickness frequency fi is the frequency to be adjusted fc, but also includes frequencies of plate thickness fluctuations caused by causes other than hardness unevenness.

Further, the plate thickness disturbance estimating device 203 determines the minimum value of the difference between the adjustment target frequency fci and the adjustment target frequency fcj excluding the adjustment target frequency fci. Then, a value obtained by multiplying the minimum value by 1/2 is obtained as the disturbance identification frequency resolution Δf _c necessary for adjusting the control parameters of the feedforward control. That is, the plate thickness disturbance estimating device 203
Δf _c = (1/2)・min {|fc _i −fcj|: fc _i ≠fcj}
Calculate.

The plate thickness disturbance estimating device 203 estimates the adjustment target frequency fci obtained through the above processing to be the frequency of the plate thickness deviation caused by the hardness unevenness disturbance. Then, the obtained adjustment target frequency fci and disturbance identification frequency resolution Δf _c are output to the frequency response estimation device 204.

FIG. 22 is a diagram showing an example of the configuration of the frequency response estimation device 204. The frequency response estimation device 204 performs FFT based on the disturbance identification frequency resolution Δf _c obtained by the plate thickness disturbance estimation device 203. Therefore, the frequency response estimation device 204 first refers to the sampling period/data number search table shown in FIG. 19 to determine the sampling period (1/Δf _s ) and the data number N _c based on the disturbance identification frequency resolution Δf _c . Determine.

As shown in FIG. 22, the frequency response estimating device 204 is provided with an inlet thickness deviation table 2041, an outlet thickness deviation table 2042, and a rolling load table 2043, each of which can store _Nc pieces of data. Then, the inlet plate thickness deviation ΔH _TRK , the outlet plate thickness deviation Δh, and the rolling load P _TRK , which are input to the frequency response estimating device 204 at every sampling period Δf _s , are stored from address 0 to address N _c -1 of the table corresponding to each. written in order.

When the writing of data up to address N _c -1 is completed, the inlet thickness deviation FFT device 2044 executes FFT processing on the data written in the inlet thickness deviation table 2041. Similarly, the exit side plate thickness deviation FFT device 2045 performs FFT processing on the data written in the exit side plate thickness deviation table 2042, and the rolling load FFT device 2046 performs FFT processing on the data written in the rolling load table 2043. Execute.

Note that the number Nc of data used in the FFT processing of the frequency response estimation device 204 is usually a sufficiently smaller value than the number N of data used in the FFT processing of the plate thickness disturbance estimation device 203, for example, 1. /10. Therefore, the FFT processing of the plate thickness disturbance estimating device 203 can be completed in a short time.

As a result of these FFT processes, the inlet thickness deviation FFT device 2044, the outlet thickness deviation FFT device 2045, and the rolling load FFT device 2046 calculate the inlet thickness deviation frequency component Hc(f) and the outlet thickness deviation frequency component hc, respectively. (f) and rolling load frequency component Pc(f) are obtained.

The inlet plate thickness to outlet plate thickness response measuring device 2047 calculates the inlet plate thickness to outlet plate thickness based on the inlet plate thickness deviation frequency component Hc(f) and the outlet plate thickness deviation frequency component hc(f) obtained as described above. Calculate the side plate thickness response Gh(f). Similarly, the inlet plate thickness to rolling load response measuring device 2048 calculates the inlet plate thickness to outlet plate thickness response GP(f) based on the inlet plate thickness deviation frequency component Hc(f) and the rolling load frequency component Pc(f). calculate.

Here, the entry side plate thickness to exit side plate thickness response Gh(f) and rolling load frequency component Pc(f) are calculated according to the following equations (9-1) and (9-2), respectively.

Next, the inlet side plate thickness to outlet side plate thickness response measuring device 2047 substitutes the adjustment target frequency fci obtained by the plate thickness disturbance estimating device 203 into the frequency f of equation (9-1), and calculates the deviation angle. Accordingly, the phase difference ΔT _ED between the inlet side plate thickness and the outlet side plate thickness is calculated. Similarly, the entry side plate thickness-rolling load response measurement device 2048 calculates the phase difference between the entry side plate thickness and rolling load by substituting the frequency fci to be adjusted into the frequency f of equation (9-2) and finding its deviation angle. Calculate ΔT _EP . In addition, the outlet thickness deviation FFT device 2045 calculates the outlet thickness deviation PP value Δh _PP by substituting the adjustment target frequency fci into the frequency f of the outlet thickness deviation frequency component hc(f) and finding its absolute value. calculate.

In other words, the phase difference ΔT _ED between the inlet plate thickness and the outlet plate thickness, and the phase difference ΔT _E between the inlet plate thickness and the rolling load.
_P and exit side plate thickness deviation PP value Δh _PP are calculated by the following equations (10-1) to (10-3).

In the above explanation, it is assumed that the plate thickness disturbance measurement device 202 and the frequency response estimation device 204 perform FFT after acquiring, for example, N pieces of performance data for one period. It is also possible to perform FFT. To do this, in tables that store performance data such as the entry side plate thickness deviation tables 2021 and 2041, when writing new performance data to address 0, the data at addresses 0 to N-1 must be shifted to addresses 1 to N. Write it down. By doing so, the latest performance data is always written in tables such as the entry side plate thickness deviation tables 2021 and 2041. Therefore, the entry side plate thickness deviation FFT devices 2023, 2044, etc. can perform FFT at the actual data acquisition cycle at the shortest.

<3.5 Membership function and fuzzy inference device>
By the way, in the feedforward control, the purpose is to use the inlet side plate thickness deviation ΔH to reduce the outlet side plate thickness deviation Δh. Therefore, what is to be controlled is the outlet side plate thickness deviation Δh. The influence of variation in deformation resistance, which is hardness unevenness, appears on the entry side of the #4 stand as the entry side plate thickness deviation ΔH, so the plate thickness control device 64 of the #4 stand performs feedforward control using the entry side plate thickness deviation ΔH. Implement. Then, the feedforward control adjustment device 101 adjusts the feedforward control based on the phase relationship between the inlet side plate thickness deviation ΔH and the outlet side plate thickness deviation Δh.

As a result, when the effect of the feedforward control appears suitably, the detected value of the exit side plate thickness deviation Δh becomes small, and ideally becomes 0. In this case, it is difficult to determine the phase relationship between the inlet side plate thickness deviation ΔH and the outlet side plate thickness deviation Δh. On the other hand, since the rolling load P fluctuates greatly as a result of removing the exit side plate thickness deviation Δh due to hardness unevenness, this can be used as a substitute for the exit side plate thickness deviation Δh. Therefore, in this embodiment, the feedforward control adjustment device 101 has a function of adjusting the control output timing shift amount ΔT _FF of the feedforward control based on the phase relationship between the entrance side plate thickness deviation ΔH and the rolling load P.

As shown in FIG. 15, the control gain/timing shift amount setting device 102 includes membership functions 105, 106, 107 and a fuzzy inference device 108, thereby realizing the above adjustment function.

First, the membership function 105 receives the exit side plate thickness deviation PP value Δh _PP as input, and calculates the values of SHS and SHB. Here, SHS is a value indicating the degree when the outlet side plate thickness deviation Δh is small, and SHB is a value indicating the degree when the outlet side plate thickness deviation Δh is large.

Similarly, the membership function 106 takes as input the phase difference ΔT _ED between the inlet side plate thickness and the outlet side plate thickness, and calculates the values of TEDB, TEDM, TEDZ, TEDP, and TEDT.
Here, TEDB is a value indicating the degree when the negative value of the phase difference ΔT _ED between the inlet side plate thickness and the outlet side plate thickness is large, and TEDM is a value indicating the degree when the phase difference ΔT _ED between the inlet side plate thickness and the outlet side plate thickness is on the negative side. This is a value indicating the degree when . Further, TEDZ is a value indicating the degree to which the phase difference ΔT _ED between the inlet side plate thickness and the outlet side plate thickness is zero. Further, TEDP is a value indicating the degree when the phase difference ΔT _ED between the inlet side plate thickness and the outlet side plate thickness is on the positive side, and TEDT is a value indicating the degree when the phase difference ΔT _ED between the inlet side plate thickness and the outlet side plate thickness is positive. This is a value indicating the degree when the value is large.

Further, the membership function 107 takes as input the phase difference ΔT _EP between the entrance plate thickness and the rolling load, and calculates the values of TEPM, TEPZ, and TEPP.
Here, TEPM is a value indicating the degree when the phase difference ΔT _EP between the entrance plate thickness and the rolling load is on the negative side. TEPZ is a value indicating the degree to which the phase difference _ΔTEP between the entrance plate thickness and the rolling load is zero. TEPP is a value indicating the degree to which the phase difference ΔT _EP between the entrance plate thickness and the rolling load is on the positive side.

Note that in FIG. 15, predetermined threshold values are used for each of the threshold values provided on the horizontal axis in the membership functions 105, 106, and 107. SB in the membership function 105 is a threshold value used to determine whether adjustment of the feedforward control using the outlet side plate thickness deviation Δh can be performed. For example, if the outlet side plate thickness deviation Δh is 1 μm or less and the outlet side plate thickness deviation Δh is not used when adjusting the feedforward control, SB=1 μm. As described above, the control gain/timing shift amount setting device 102 according to the present embodiment adjusts the fluctuation range of the rolling load P instead of the exit side plate thickness deviation Δh when the variation width of the exit side plate thickness deviation Δh is within a predetermined range. Refer to phase.

DB and DT in the membership function 106 are threshold values used to determine whether the control gain is too high. For example, if the phase difference ΔT _ED between the inlet side plate thickness and the outlet side plate thickness exceeds 90 degrees, it is determined that the control gain is high. In this case, DB=−90 degrees and DT=90 degrees are set, and control to lower the control gain is performed.

DM and DP in the membership function 106 and PM and PP in the membership function 107 are threshold values used to determine that adjustment of the output timing shift amount is unnecessary. For example, if the phase difference ΔT _ED between the inlet side plate thickness and the outlet side plate thickness is within ±20 degrees, it is determined that adjustment of the output timing shift amount ΔT _FF is unnecessary. In this case, DM=-20 degrees, DP=20 degrees, and similarly PM=-20 degrees and PP=20 degrees. Note that these values are just examples, and can be changed as appropriate depending on rolling conditions and equipment characteristics.

In addition, DZ and PZ are the phase difference ΔT _ED between the inlet side plate thickness and the outlet side plate thickness, and the position between the inlet side plate thickness and the rolling load when the outlet side plate thickness deviation Δh is minimized and the effect of feedforward control is maximized. A phase difference ΔT _EP is set. It is assumed that the values of these phase differences are determined in advance based on, for example, rolling simulation or actual rolling data during manual adjustment in actual rolling. The control gain/timing shift amount setting device 102 determines the output timing shift amount ΔT _FF by comparing the input phase difference, such as the phase difference ΔT _ED between the input side plate thickness and the output side plate thickness, with a predetermined value. do.

The fuzzy inference device 108 uses SHS, SHB, TEDB, TEDM, TEDZ, TEDP, TEDT, TEPM, TEPZ, and TEPP determined by the membership functions 105, 106, and 107 to determine TFFP, TFFM, GFFP, and GFFM. . Here, TFFP and TFFM are values representing the degree to which the control output timing shift amount _ΔTFF for feedforward control is changed to the increasing side and the degree to which it is changed to the decreasing side, respectively. Further, GFFP and GFFM are values representing the degree to which the control gain _GFF is changed to the increasing side and the degree to which the control gain GFF is changed to the decreasing side, respectively.

Generally, there are various inference rules, but the fuzzy inference device 108 according to this embodiment performs processing expressed by the following conditions.
If IF (A and B) then C: C=min(A, B)
If IF (A or B) then C: C=max(A,B)

When the exit side plate thickness deviation Δh is large and the phase difference ΔT _ED between the input side plate thickness and the output side plate thickness is zero, it is considered that the control gain G _FF of the feedforward control is small, so the following inference rule is applied.
IF (SHB and TEDZ) then GFFP

Further, if the exit side plate thickness deviation Δh is large and there is a phase difference ΔT _ED between the input side plate thickness and the output side plate thickness, it is determined that the control output timing shift amount ΔT _FF is shifted. Therefore, it is expected that by eliminating this deviation, the exit plate thickness deviation Δh will be reduced, and therefore the following inference rule is applied.
IF (SHB and TEDP) then TFFP
IF (SHB and TEDM) then TFFM

Further, when the exit side plate thickness deviation Δh is large and the phase difference between the input side plate thickness and the output side plate thickness is large and exceeds 90 degrees, it is determined that the control gain G _FF of the feedforward control is too large. In this case, it is considered better to first lower the control gain G _FF and adjust the control output timing shift amount ΔT _FF after the control gain becomes an appropriate control gain. Therefore, in this case, the following inference rules apply:
IF (SHB and TEDT) then GFFM
IF (SHB and TEDB) then GFFM

In addition, when the exit side plate thickness deviation Δh is small and the phase difference between the input side plate thickness and rolling load ΔT _EP is large, it is expected that the exit side plate thickness deviation Δh can be further reduced by adjusting the control control output timing shift amount ΔT _FF . be done. Therefore, in this case, the following inference rules apply:
IF (SHS and TEPP) then TFFM
IF (SHS and TEPM) then TFFP

According to the simulation of the rolling phenomenon, when the phase difference ΔT _EP between the entrance plate thickness and the rolling load is on the negative side and the control output timing shift amount ΔT _FF is changed to the increasing side, the phase difference between the entrance plate thickness and the rolling load ΔT _EP becomes smaller. Further, when the phase difference ΔT _EP between the entrance plate thickness and the rolling load is on the positive side, if the control output timing shift amount ΔT _FF is changed to the decreasing side, the phase difference ΔT _EP between the entrance plate thickness and the rolling load becomes smaller. The inference rules shown above were determined based on the results of this simulation.

The relationship in FIG. 25 shows how control state quantities such as the inlet side plate thickness deviation ΔH and the outlet side plate thickness deviation Δh change before and after the control. Since the rolling load P is determined by the plate thickness variation on the entry and exit sides and the tension on the entry and exit sides, the phase difference between the entry side plate thickness and the rolling load ΔT _EP and the control output timing shift amount ΔT _FF are The relationship is different from that in FIG. However, if the change tendency of the phase difference ΔT _EP between the entrance plate thickness and the rolling load when the control output timing shift amount ΔT _FF is changed is known, it is possible to change the phase difference ΔT _EP between the entrance plate thickness and the rolling load as in this embodiment. can be used to adjust the control output timing shift amount _ΔTFF .

By using the above inference rules, it is possible to obtain TFFP, which is the degree to which the control output timing shift amount ΔT _FF for feedforward control is changed to the increasing side, and FFM, which is the degree to which the control output timing shift amount ΔT FF for feedforward control is changed to the decreasing side. Furthermore, GFFP, which is the degree to which the control gain _GFF for feedforward control is changed to the increasing side, and GFFM, which is the degree to which the control gain GFF for feedforward control is changed to the decreasing side, can be determined.

Note that the inference rule explained above is an example, and the inference rule is not limited to this inference rule. For example, any inference rule that reduces the exit side plate thickness deviation Δh by changing the control state quantity in feedforward control, the control gain G _FF for feedforward control, the control output timing shift amount ΔT _FF , etc. But that's fine. Further, the inference rule may be determined based on actual data manually adjusted in an actual rolling operation, rather than based on a simulation of rolling phenomena. This is often more consistent with actual rolling phenomena.

The parameter change device 109 uses the change degrees TFFP, TFFM, GFFP, and GFFM obtained above to change the control gain G _FF and control output for feedforward control according to the following equations (11-1) and (11-2). Change the timing shift amount _ΔTFF .

Here, C _TFFP , C _TFFM , C _GFFP , and C _GFFM are adjustment parameters. C _TFFP is a value indicating the amount of change on the increasing side per one time of the control output timing shift amount ΔT _FF , and C _TFFM is a value indicating the amount of change on the decreasing side. Moreover, _{C_GFFP} is a value indicating the amount of change on the increasing side of the control gain G per one time, and _{C_GFFM} is a value indicating the amount of change on the decreasing side.
By performing the above processing for each selected inlet thickness frequency fi (4206), the control gain G _FF (fi) and the control output timing shift amount ΔT _FF (fi) can be determined.

As described above, the feedforward control adjustment device 101 always adjusts the control gain G _FF (fi) and the control output timing shift amount ΔT _FF (fi) in the inlet thickness deviation composite correction value creation device 430 to the optimal state. becomes possible. As a result, the control effect of feedforward control is significantly improved.

The control gain G _FF (fi) and control output timing shift amount ΔT _FF (fi) for each selected inlet thickness frequency fi 4206 obtained by the feedforward control adjustment device 101 are input to the inlet thickness deviation composite correction value creation device 430. be done.

FIG. 23 is a diagram showing an outline of the entrance side plate thickness deviation composite correction value creation device 430. The entry side plate thickness deviation composite correction value calculation formula creation device 4301 converts the entry side plate thickness deviation composite value 4255 created by the entry side plate thickness deviation composite value creation device 425 into the selected input calculated by the feedforward control adjustment device 101. By correcting with the control gain G _FF (fi) for each side plate thickness frequency fi (4206) and the control output timing shift amount ΔT _FF (fi), the input side plate thickness deviation composite correction value calculation formula ΔH _4FFEST (t) can be obtained. . Here, t is the time when t=0, which is the time when the input side plate thickness deviation value immediately below the mill is written in the first table of the input side plate thickness deviation mill table 4251 in the input side plate thickness deviation composite value creation device 425. The plate thickness control device 64 (#4 stand plate thickness control) performs control processing at a certain control period Δt, but since FFT etc. require calculation time, the inlet side plate thickness deviation composite value creation device 425 and This is because the feedforward control adjustment device 101 is executed asynchronously with Δt.

Since the plate thickness control device 64 (#4 stand plate thickness control) is executed at the control period Δt, the inlet side plate thickness deviation composite correction value _ΔH4FFEST to be used also needs to be determined at the same control period Δt. Therefore, the entry side plate thickness deviation composite correction value calculation device 4302 periodically executes the operation at the control period Δt, and writes the entry side plate thickness deviation value immediately below the mill into the head table of the entry side plate thickness deviation mill table 4251 at the time t= The elapsed time from 0 is calculated as the plate thickness deviation calculation time tcal, and the entry side Itatsu deviation composite correction value ΔH _4FFEST (tcal) (410) is calculated and output to the plate thickness control device 64 (#4 stand plate thickness control). . Here, the plate thickness deviation calculation time tcal is set to tcal=0 when the entry side plate thickness deviation value directly below the mill is written in the first table of the input side plate thickness deviation table 4251.
With the above configuration, it is possible to adjust feedforward control for multiple frequencies.

<3.5 Simulation results>
As described above, FIG. 9D shows the results when the control gain and control output timing shift amount are adjusted for a plurality of frequency components using the method of this embodiment. It can be seen that all of the frequency components (A), frequency components (B), and frequency components (C) have been adjusted, and there is almost no variation in the exit side plate thickness. In the upper part of FIG. 9D, time-series data of plate thickness deviation is displayed, and most of the plate thickness deviation on the outlet side has been removed.
As described above, by using the adjustment method of this patent, it was possible to adjust feedforward control for multiple frequency components.

As described above, according to the present embodiment, the control output timing shift amount ΔT _FF and the control gain G _FF in the feedforward control can be adjusted while taking in the rolling performance data during the rolling operation even when there is an inlet plate thickness deviation of multiple frequency components. The effect of feedforward control can be improved by making modifications. Furthermore, in the present embodiment, the control output timing shift amount ΔT _FF and the control gain G _FF are basically determined based on the results of FFT processing of the inlet side plate thickness deviation ΔH and the outlet side plate thickness deviation Δh. Therefore, even if the inlet thickness deviation ΔH and the outlet thickness deviation Δh contain many frequency components, it is possible to find the frequency of plate thickness fluctuation due to hardness unevenness among them, and to find the inlet thickness deviation that should be controlled. It is facilitated to obtain the phase difference δ between ΔH and the exit side plate thickness deviation Δh. As a result, the control output timing shift amount ΔT _FF and the control gain G _FF can be determined more appropriately, so that the effect of feedforward control can be greatly improved. That is, it can be said that this embodiment makes it possible to efficiently adjust the control output timing for feedforward control in a short time based on the frequency characteristics of fluctuations in the pre-control state quantity and the post-control state quantity.

≪4. Modifications of the embodiment≫
<Modification 1>
In the embodiment described above, the control output timing shift amount ΔT _FF and control gain G _FF for feedforward control are determined using the phase difference ΔT _ED between the input side plate thickness and the output side plate thickness and the phase difference ΔT _EP between the input side plate thickness and the rolling load. This will be adjusted accordingly. However, the method of adjusting the control output timing shift amount ΔT _FF and control gain G _FF for feedforward control is not limited to this method.

As shown in FIG. 22, the inlet plate thickness to rolling load response measuring device 2048 and the inlet plate thickness to outlet plate thickness response measuring device 2047 measure the inlet plate thickness to rolling load response GP(f) and the inlet plate thickness to outlet plate thickness. The thickness response GPh(f) is being calculated. Therefore, the attenuation factors |GP(fc)| and |Gh(fc)| at the adjustment target frequency fc can be determined. Therefore, we will increase the number of control rules in the fuzzy inference device 108 using data on these attenuation rates. For example, in the control rule IF (SHB and TEDP) then TFFM, if |Gh(fc)| is large (attenuation amount is small), GFFP is also changed to be performed at the same time.

In this way, the control output timing shift amount ΔT _FF and control gain G _FF for feedforward control can be adjusted. In this case, effects such as a reduction in the response time required for adjustment can be expected.

<Modification 2>
In modification example 2, the actual values of the control output timing shift amount ΔT _FF and control gain G _FF for feedforward control when the exit side plate thickness deviation Δh that satisfies the predetermined manufacturing quality is obtained in the actual rolling process are stored. Consider an embodiment in which a database is further provided. This database includes the control output timing shift amount ΔT based on the rolling conditions such as the steel type, rolling speed, and target thickness of the rolled material 3 when the exit side plate thickness deviation Δh that satisfies the predetermined manufacturing quality is obtained in the rolling process. _{The FF} and the control gain _GFF are stored in association with each other.

In this case, search the database at the start of rolling, and if data for similar rolling conditions is stored, retrieve and use the control output timing shift amount ΔT _FF and control gain G _FF for the similar rolling conditions. Can be done. Therefore, in this modification, it is possible to use control parameters for feedforward control that have been proven in past rolling processes and to further modify them. As a result, the control effect in feedforward control can be increased.

<Modification 3>
The basic concept of adjusting the control gain and phase of the feedforward control in the embodiment described above can also be applied to roll eccentric plate thickness control in a single stand rolling mill. In this case, based on the inlet thickness deviation detected by the inlet thickness gauge of the single-stand rolling mill, the outlet thickness deviation is controlled using, for example, the roll gap (distance between upper and lower work rolls) as an operating end. Such rolling control is often called gauge meter type, and the basic formula for the rolling phenomenon is expressed as the following formula (12).

Here, when considering only the inlet side plate thickness deviation ΔH and the outlet side plate thickness deviation Δh as the rolling load deviation ΔP, the rolling load deviation ΔP can be expressed by the following equation (13).

In equation (13), in order to set the exit side plate thickness deviation Δh = 0, when considering equation (12), the following equation (14) is required between the roll gap deviation ΔS and the input side plate thickness deviation ΔH. It can be seen that the relationship holds true.

Equation (14) means that the outlet side plate thickness deviation Δh can be made zero by performing feedforward proportional control of the roll gap deviation ΔS based on the inlet side plate thickness deviation ΔH. That is, the roll gap deviation ΔS can be obtained by multiplying the entrance side plate thickness deviation ΔH by a control gain expressed by the following equation (15).

Furthermore, when considering the hardness unevenness of the rolled material 3, that is, the deformation resistance variation Δk, as the rolling load deviation ΔP, the rolling load deviation ΔP can be expressed by the following equation (16).

In equation (16), in order to set the exit side plate thickness deviation Δh = 0, when considering equation (12), the following relationship between the roll gap deviation ΔS, the input side plate thickness deviation ΔH, and the deformation resistance variation Δk is required. It can be seen that the relationship of equation (17) holds true.

Here, when the entrance plate thickness deviation ΔH and the deformation resistance variation Δk have the same frequency component, equation (17) can be expressed as the following equation (18).

Furthermore, equation (18) can be transformed as shown in equation (18) below.

The values of X and δ included in equation (19) are given by equations (2-1) and (2-2) above when G and Δ are calculated using equation (20) below.

Equation (19) indicates that feedforward control of the roll gap deviation ΔS requires adjustment of the control gain G to be multiplied by the entrance side plate thickness deviation ΔH and the phase shift of the input side plate thickness deviation ΔH. Therefore, the adjustment can be performed using a configuration similar to that described in the above embodiment.

≪5. Supplement≫
FIG. 24 is a diagram showing an example of the hardware configuration of the information processing device 500 that constitutes the rolling control device 2 according to the embodiment of the present invention. The rolling control device 2 including the plate thickness control device 64, the feedforward control adjustment device 101, and the like used in the embodiment of the present invention and its modifications is realized by a combination of software and hardware. Such an information processing device 500 has a configuration similar to a general PC (Personal Computer), workstation, or the like.

That is, the information processing device 500 includes a so-called CPU (Central Processing Unit) 501, a RAM (Random Access Memory) 502, a ROM (Read Only Memory) 503, an HDD (Hard Disk Drive) 504, an I/F (Interface Circuits) 505, etc. are connected via bus 508. Further, the I/F 505 is connected to a display unit 506 such as an LCD (Liquid Crystal Display) and an operation unit 507 such as a keyboard.

The CPU 501 is not only a program execution unit but also a calculation unit that executes various calculations. The RAM 502 is a volatile storage medium in which information can be read and written at high speed, and stores programs to be executed by the CPU 501, as well as various types of information necessary for executing the programs. The ROM 503 is a read-only nonvolatile storage medium, and stores programs such as firmware.

The HDD 504 is a nonvolatile magnetic storage medium in which information can be read and written, and stores an OS (Operating System), control programs necessary for sheet thickness control, control information, general application programs, and the like. The I/F 505 connects the devices forming the display section 506 and the operation section 507 to the bus 508, and controls the exchange of information with the devices. Furthermore, the I/F 505 is connected to various measuring instruments (for example, plate thickness gauge 41, tension gauge 51, etc.) and various equipment control devices (for example, roll gap control device 31, etc.) provided in the rolling mill 1. It is also used as an interface to exchange information between.

In the information processing device 500 configured as described above, the rolling control device 2 according to the embodiment of the present invention executes the program read from the recording medium such as the ROM 503 and the HDD 504 and developed in the RAM 502. functions are realized. In this case, the functions of the rolling control device 2 may be realized by one information processing device 500 or by a plurality of information processing devices 500.

The present invention is not limited to the embodiments and modified examples described above, and further includes various modified examples. For example, the above-described embodiments and modified examples have been described in detail to explain the present invention in an easy-to-understand manner, and are not necessarily limited to those having all the configurations described. Furthermore, it is possible to replace a part of the configuration of one embodiment or modification with the configuration of another embodiment or modification, and the configuration of one embodiment or modification may be replaced with another embodiment or modification. It is also possible to add the following configuration. Further, it is also possible to add, delete, or replace a part of the configuration of each embodiment or modified example with the configuration included in other embodiments or modified examples.

1 Rolling mill 2 Rolling control device 3 Rolled material 11, 12, 13, 14 Stand rolling mill 15 Output bridle roll 21, 22, 23, 24, 25 Motor speed control device 31, 32, 33, 34 Roll gap control device 41, 42, 43, 44 Plate thickness meter 51, 52, 53, 54 Tension meter 61, 62, 63, 64 Plate thickness control device 71, 72, 73, 74 Tension control device 101 Feedforward control adjustment device (feedforward control parameter adjustment means)
102 Control gain/timing shift amount setting device 105, 106, 107 Membership function 108 Fuzzy inference device (feedforward control parameter adjustment means)
109 Parameter changing device (feedforward control parameter adjusting means)
201 Frequency response measurement device (frequency response measurement means)
202 Plate thickness disturbance measuring device (first frequency response measuring means)
2021 Inlet thickness deviation table 2022 Outlet thickness deviation table 2023 Inlet thickness deviation FFT device 2024 Outlet thickness deviation FFT device 203 Thickness disturbance estimation device (second frequency response measurement means)
204 Frequency response estimation device (third frequency response measurement means)
2041 Inlet thickness deviation table 2042 Outlet thickness deviation table 2043 Rolling load table 2044 Inlet thickness deviation FFT device 2045 Outlet thickness deviation FFT device 2046 Rolling load FFT device 2047 Inlet thickness to outlet thickness response measurement device 2048 Inlet thickness ~Rolling load response measuring device 300 Controlled object 301 Controlled disturbance 302 Controlled object state quantity 303 Observable quantity 304 Detection dead time 305 Disturbance transfer 306 Feedback control 307 Feedforward control (feedforward control means)
308 Control disturbance transfer 309 Control operation amount 310 Control disturbance transfer value 311 Disturbance transfer value 320 Control disturbance provisional value creation device (pre-control state quantity synthesis means)
321 Control disturbance provisional value 325 Control disturbance composite value creation device (pre-control state quantity composition means)
326 Control disturbance composite value 330 Control disturbance composite correction value creation device 350 Control disturbance composite correction value 400 Multiple frequency control adjustment device 410 Inlet side plate thickness deviation composite correction value 420 Inlet side plate thickness deviation provisional value generator 4201 Inlet side plate thickness deviation table 4202 Input Side thickness deviation FFT device 4203 Entrance side thickness frequency selection device 4204 Temporary time series data creation device 4205 Entrance side thickness deviation tentative value table 425 Entrance side thickness deviation composite value creation device 4251 Entrance side thickness deviation table directly below the mill 4252 Entrance side thickness deviation provisional table Value FFT device 4253 Entrance side thickness deviation FFT device directly below the mill 4254 Entrance side deviation provisional value correction device 4255 Entrance side thickness deviation composite value 430 Entrance side thickness deviation composite correction value creation device 4301 Entrance side thickness deviation composite correction value calculation formula creation device 4302 Entrance side plate thickness deviation composite correction value calculation device 500 Information processing device (computer)
501 CPU
502 RAM
503 ROM
504 HDD
505 I/F
506 Display section 507 Operation section 508 Bus δ Phase difference δ _FF control output timing shift amount Δ Phase shift amount Δf Frequency resolution Δf _c disturbance identification frequency resolution Δf _s sampling period ΔH Inlet side thickness deviation ΔH _TRK inlet side thickness deviation Δh Outlet side plate thickness Deviation Δh _PP outlet plate thickness deviation PP value ΔT ₃₄ Tension deviation ΔT _FF control output timing shift amount ΔT Phase difference between _ED inlet plate thickness and outlet plate thickness ΔT Phase difference between _EP inlet plate thickness and rolling load G, G _BF , G _FF Control gain T _34FB tension actual value T _34ref tension command value T _FF transfer time f _s sampling frequency f _r maximum frequency (= f _s /2)
fc _i disturbance frequency fc frequency to be adjusted L _n noise level P _TRK rolling load Hc(f) entrance plate thickness deviation frequency component hc(f) exit side plate thickness deviation frequency component Pc(f) rolling load frequency component Hg(m) entrance plate Thickness deviation amplitude Hp(m) Entrance side plate thickness deviation phase hg(m) Output side plate thickness deviation amplitude hp(m) Output side plate thickness deviation phase

Claims (8)

A plant control device that performs feedforward control of a post-control state quantity that is a control state quantity after the control based on a pre-control state quantity that is a control state quantity before control when processing a workpiece,
Based on the result of fast Fourier transform of the time series data of the pre-control state quantity, a plurality of frequency components included in the pre-control state quantity are extracted, and a pre-control state quantity composite waveform is created by synthesizing the extracted frequency components. Pre-control state quantity synthesis means;
Frequency response measuring means for obtaining a phase difference and an attenuation amount of the post-control state quantity with respect to the pre-control state quantity based on results of fast Fourier transform of time-series data of the pre-control state quantity and the post-control state quantity. and,
For each of the plurality of frequency components, a control output timing shift amount, which is a delay time until the pre-control state quantity is reflected in the feedforward control, and control of the feedforward control, based on the acquired phase difference and attenuation amount. feedforward control parameter adjusting means for determining the gain;
A pre-control state in which a pre-control state quantity composite waveform is corrected using the control output timing shift amount obtained by the feedforward control parameter adjustment means and the control gain of the feedforward control, and a pre-control state quantity composite value correction value is determined. Quantity composite value correction means;
A plant control device comprising: feedforward control means for performing feedforward control using the pre-control state quantity composite value correction value. The pre-control state quantity synthesis means uses the result of fast Fourier transform of the time series data of the pre-control state quantity and the result of fast Fourier transform of the time series data of the pre-control state quantity composite waveform, and calculates the difference between each frequency component. The plant control device according to claim 1, further comprising: determining an amplitude and a phase difference. The feedforward control parameter adjusting means includes:
The plant control device according to claim 1, wherein the control gain of the feedforward control is determined based on the phase difference and the attenuation amount acquired by the frequency response measuring means for each of the plurality of frequency components. The control output timing shift used in the feedforward control during the processing is added to the processing condition data during the processing of the workpiece when the post-control state quantity is within a predetermined range. further comprising a database storing data in which quantities and control gains of the feedforward control are associated with each other;
The feedforward control parameter adjusting means includes:
When the processing starts, the database is searched, and if data with the same processing conditions as the processing is stored, the control output timing is determined for each of the plurality of frequency components based on the data stored in the database. The plant control device according to claim 3, further comprising determining a shift amount and a control gain of the feedforward control. The frequency response measuring means includes:
Based on the result of fast Fourier transform of the time series data of the pre-control state quantity, a plurality of frequency components included in the pre-control state quantity are extracted, and the minimum value of the difference value between the extracted plurality of frequency components is Obtain a value larger than 1/2 as the disturbance identification frequency resolution,
Fast Fourier transform is performed on each time series data of the pre-control state quantity and the post-control state quantity acquired according to the sampling period and the number of data that are compatible with the fast Fourier transform determined by the disturbance identification frequency resolution, and the respective fast Fourier transforms are performed. The plant according to claim 1, wherein a phase difference and an attenuation amount of the post-control state quantity with respect to the pre-control state quantity at the adjustment target frequency of the extracted plurality of frequency components are calculated based on the conversion result. Control device. A rolling control device that performs feedforward control of an exit side plate thickness deviation, which is a control state quantity after the control, based on an entry side plate thickness deviation, which is a control state quantity before control when rolling a material to be rolled,
Control that extracts a plurality of frequency components included in the pre-control state quantity based on the results of fast Fourier transform of time-series data of the pre-control state quantity, and creates a pre-control state quantity composite waveform by synthesizing the extracted frequency components. a pre-state quantity synthesis means;
Frequency response measuring means for obtaining the phase difference and attenuation amount of the exit side plate thickness deviation with respect to the input side plate thickness deviation based on the results of fast Fourier transform of the respective time series data of the input side plate thickness deviation and the exit side plate thickness deviation. and,
For each of the plurality of frequency components, based on the acquired phase difference and attenuation amount, a control output timing shift amount, which is a delay time until the input side plate thickness deviation is reflected in the feedforward control, and control of the feedforward control. feedforward control parameter adjusting means for determining the gain;
Pre-control state quantity composite value correction means for correcting a pre-control state quantity composite waveform using the control output timing shift amount and the control gain of the feedforward control to determine a pre-control state quantity composite value correction value;
A rolling control device comprising: feedforward control means for performing feedforward control using the pre-control state quantity composite value correction value. A plant control device that performs feedforward control of a post-control state quantity, which is a control state quantity after the control, based on a pre-control state quantity, which is a control state quantity before control when processing a workpiece,
Based on the result of fast Fourier transform of the time series data of the pre-control state quantity, a plurality of frequency components included in the pre-control state quantity are extracted, and a pre-control state quantity composite waveform is created by synthesizing the extracted frequency components. step and
calculating a phase difference and an attenuation amount of the post-control state quantity with respect to the pre-control state quantity based on results of fast Fourier transform of time-series data of the pre-control state quantity and the post-control state quantity;
For each of the plurality of frequency components, a control output timing shift amount, which is a delay time until the pre-control state quantity is reflected in the feedforward control, and control of the feedforward control, based on the acquired phase difference and attenuation amount. a step of determining a gain;
correcting the pre-control state quantity composite waveform using the control output timing shift amount and the control gain of the feedforward control, and determining a pre-control state quantity composite value correction value;
A plant control method, comprising: performing feedforward control using the pre-control state quantity composite value correction value. A computer constituting a plant control device that performs feedforward control of a post-control state quantity, which is a control state quantity after the control, based on a pre-control state quantity, which is a control state quantity before control when processing a workpiece. ,
Based on the result of fast Fourier transform of the time series data of the pre-control state quantity, a plurality of frequency components included in the pre-control state quantity are extracted, and a pre-control state quantity composite waveform is created by synthesizing the extracted frequency components. step and
calculating a phase difference and an attenuation amount of the post-control state quantity with respect to the pre-control state quantity based on results of fast Fourier transform of time-series data of the pre-control state quantity and the post-control state quantity;
For each of the plurality of frequency components, a control output timing shift amount, which is a delay time until the pre-control state quantity is reflected in the feedforward control, and control of the feedforward control, based on the acquired phase difference and attenuation amount. a step of determining a gain;
correcting a pre-control state quantity composite waveform using the control output timing shift amount and the control gain of the feedforward control, and determining a pre-control state quantity composite value correction value;
performing feedforward control using the pre-control state quantity composite value correction value;
A plant control program to run.

Images