JPH10250888A - Tensile force control method and device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To adjust control gain imparted to a tensile force control loop automatically, at high speed and to the optimum in accordance with a situation at the time of carrying a web material by using a plural number of driving rolls. SOLUTION: This tensile force control method uses genetic algorism at the time of adjusting control gain of a feedback controller 22 to carry out tensile force control by a control gain adjusting device 25. At this time, output of the feedback controller 22 is given to an approximate model 23, an estimated value is computed in accordance with response deviation between the approximate model 23 and a normative response model 27 to show a response to be a norm, and genetic operation is carried out in accordance with this estimated value.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for controlling the tension of a web material in a facility for handling the web material such as a metal strip, roll paper, film, and the like. The present invention relates to a tension control method and device in a drive control system.

[0002]

2. Description of the Related Art In a facility for manufacturing or processing web materials such as metal strips, roll paper, and polymer films, a multi-drive control provided with a plurality of drive mechanisms for transporting these web materials is provided. The system is commonly used. Each drive mechanism is provided with a drive roll. By controlling each of the plurality of drive rolls separately, it is possible to independently control the speed and tension of the web material. FIG. 2 is a block diagram showing an example of a configuration of a control system in such a multi-drive control system. In order to transport the web material 1 in the direction of the arrow shown in the figure, a first
, And a second drive roll 3 is provided on the downstream side. Further, a tension detector 4 for detecting the tension of the web material 1 and outputting the detected tension as a tension detection signal _Tfb is provided at an intermediate position between the driving rolls 2 and 3. The driving rolls 2 and 3 are connected to electric motors (motors) 7 and 8 via reduction gears 5 and 6, respectively, and are driven by the electric motors 7 and 8. The motors 7 and 8 are connected to speed detectors (pulse generators) 9 and 10 for detecting their speeds and outputting them as speed detection signals V _fb1 and V _fb2 , respectively. The electric motors 7 and 8 are drive devices 11 and 1 respectively.
2, and control amounts are input to the drive devices 11 and 12 from the speed controllers 13 and 14, respectively.

The tension command T _ref and the speed command V _ref are set by a tension setting device 19 and a speed setting device 20, respectively. Here, in the first drive roll 2 on the upstream side,
Tension control is performed by performing control based on the speed command V _ref and the tension command T _ref, and the second drive roll 3 on the downstream side is controlled.
In, speed control is performed by performing control based only on the speed command _Vref . As for the second drive roll that performs speed control, a subtractor 16 that calculates a deviation between the speed command _Vref and the speed detection signal _Vfb2 is provided, and the output of the subtractor 16 is input to the speed controller 14. . Speed controller 14
For example, PI control or PID control is used. For the first drive roll 2, the tension command V _ref
A subtractor 17 for calculating a deviation between the tension and the tension detection signal _Tfb is provided, and an output of the subtractor 17 is input to a tension controller 18 for performing tension control. Further, a subtractor 15 for _obtaining a deviation between the sum of the output of the tension controller 18 and the speed command _Vref and the speed detection signal _Vfb1 is provided, and the output of the subtractor 15 is input to the speed controller 13. As a result, a minor loop for controlling the speed based on the _difference between the speed command _Vref and the speed detection signal Vfb1 is provided for the first drive roll for performing the tension control. On the other hand, the tension of the web material 1 is detected by the tension detector 4, and the tension controller 18 performs the tension control using the tension detection signal T _fb. Vessel 1
The modulation corresponding to the output from the control unit 8 is added. Conventionally, the speed controller 13 uses PI control or PID control,
Similarly, the tension controller 18 also uses PI control or PID control.

In this case, it is necessary for the tension controller 18 to set an appropriate control gain in order to satisfy the requirements for responsiveness and stability. However, in the above-described multi-drive control system, the control gain cannot be set as theoretically due to mutual interference with the preceding and following sections, restrictions on the control gain due to required specifications, and the like. Therefore, conventionally, an operator observes the control amount and changes the control gain by trial and error.

[0005]

As described above, PI control or PI control in a multi-drive control system is performed.
When the tension control is performed by the D control, it takes a long time to set the control gain, and it depends on the experience of the operator. SUMMARY OF THE INVENTION An object of the present invention is to provide a tension control method and device capable of automatically and quickly adjusting a control gain according to a situation.

[0006]

According to the present invention, there is provided a tension control method comprising the steps of: transporting the web material using a plurality of drive rolls arranged along a transport direction of the web material and connected to an electric motor; One of the electric motors, each controlled by a speed control loop based on a speed command based on a line speed, and connected to two adjacent drive rolls, respectively, uses the speed control loop as a minor loop and A tension control loop having a tension control loop using a detection output of the tension of the web material between the drive rolls as a feedback signal, and controlling a tension of the web material between the two adjacent drive rolls; When adjusting the control gain for the tension control loop, the control gain is converted into a symbol string and the symbol string is connected. A code string representing the control gain and a combination thereof, a first step of inputting the control gain to be adjusted and the adjustment condition thereof, and the code string serving as a solution candidate according to the adjustment condition. A second step of generating an initial group of, and a third step of decoding each of the code strings into the control gain, and obtaining an evaluation value from a response obtained by the decoded control gain, A fourth step of performing a genetic operation based on a genetic algorithm on each of the code strings in accordance with the evaluation value to generate a solution candidate code string to be a child; and a fifth step of determining convergence. The control gain is optimized by repeatedly executing the third step, the fourth step, and the fifth step.

In the tension control method of the present invention, an approximate model of a tension control system including the tension control loop and a reference response model that gives a reference response are used, and the approximate model and the reference response model are used. The evaluation value can be determined based on the response deviation. The tension control device according to the present invention includes a first drive roll and a second drive roll disposed along the transport direction of the web material, and a first drive roll connected to the first drive roll and the second drive roll, respectively. And a second electric motor, and a tension detector disposed between the first driving roll and the second driving roll to detect a tension of the web material, wherein each of the electric motors is a line. Controlled by a speed control loop based on a speed command based on speed, one of the first electric motor and the second electric motor has the speed control loop as a minor loop and a detection output from the tension detector. Has a tension control loop with a feedback signal
A tension control device that conveys the web material while controlling the tension by each of the drive rolls, a controller provided in the drive control loop to perform tension control according to a control gain, and a tension control system including the tension control loop An approximate model, a reference response model that gives a reference response, and an evaluation function comparison unit that calculates an evaluation value based on a response deviation between the approximate model and the reference response model.
A switching means for switching the output signal of the controller from the approximation model to the control target side when a response within the reference range is obtained, and giving the same command to the control target and the approximation model by a response deviation at that time. A model identification unit that corrects each parameter of the approximate model, and a control gain adjustment unit that adjusts the control gain, wherein the control gain adjustment unit converts the control gain into a symbol string and converts the symbol string into a symbol string. The connected one is a code string representing the control gain and a combination thereof, and an initial group of the code string that is a solution candidate is generated according to an adjustment condition. After the initial group is generated, a predetermined convergence condition is satisfied. Up to this point, each of the code strings is decoded into the control gain and input to the controller. Performing a genetic operation based on genetic algorithm to repeatedly generating a solution candidate code string to be a child, to optimize the control gain.

In the present invention, since the genetic algorithm is used, the control gain of the tension control system can be automatically adjusted at high speed and optimally according to the situation.

[0009]

Next, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram illustrating a configuration of a tension control device according to an embodiment of the present invention.
The tension control device shown in FIG. 1 determines the control gain of the tension controller 18 by the tension control method of the present invention in a multi-drive control system as shown in FIG. 1. In FIG. 1, a control target 21 is a collective abstraction of each component (on the web material 1 side) from the tension controller 18 in the multi-drive control system shown in FIG. Reference numeral 22 has the same configuration as the tension controller 18 in FIG. Then, a model identification unit 2 that identifies the approximate model 23 by comparing the output of the control object 21 and the output of the approximate model 23 with an approximate model 23 that is a model that approximates the control object 21.
4, a control gain adjusting device 25 for adjusting the control gain of the feedback controller 22, a preset reference response model 27 showing an ideal response, and an approximate model 23
Evaluation value calculation unit 28 for calculating an evaluation value by comparing the response by the reference response model 27 with the response by the reference response model 27, and the output (control signal) of the feedback controller 22 to the control object 21
And a switching unit 29 for switching between input to the approximate model 23 and input to the approximate model 23.

The control gain adjusting device 25 automatically adjusts the control gain of the feedback controller 22 by using a genetic algorithm based on the evaluation value calculated by the evaluation value calculator 28. In the present embodiment,
A model appropriate for performing the adjustment is set in advance as the approximate model 23, and only the unknown parameters of the approximate model 23 are identified by the model identification unit 24 by the least square method or the like. Specifically, control target 2
1 and the same command are given to the approximation model 23, and the model identification unit 24 corrects each parameter of the approximation model 23 based on the response deviation. Next, a method for adjusting the control gain in the present embodiment will be described with reference to the flowchart in FIG.

First, each condition is input as a first step (step 51), and as a second step, an initial group of code strings to be used as individuals in a genetic algorithm is created (step 52). In the present embodiment, each control gain is coded by a lower limit of each control gain, scaling based on a required or settable numerical value, and binary conversion. A combination of the bit strings representing the control gains coded as described above is defined as one code string (individual) that is a control gain combination candidate. Then, M code strings in which each bit is set at random are generated to form an initial group. Next, as a third step, a response simulation is performed using a control gain obtained by decoding each code string included in the group, and an evaluation value of each candidate is obtained (step 5).
3). That is, the switch 29 is switched to the approximation model 23 side, and each control source obtained by decoding the code string is supplied to the feedback controller 22 to obtain a response of the approximation model 23 to the tension step command. Then, by the evaluation value calculation unit 28 as the evaluation function comparison means,
The response of the approximate model 23 compared to the response of the nominal response model 27, and calculates the evaluation value V _i. In the present embodiment, the evaluation value V _i is obtained by the following evaluation function.

[0012]

(Equation 1)

Here, N is the number of state variables to be evaluated, T is the evaluation time, x _j is the value of each state variable (such as tension) in the approximate model 23, and x _jmodel is each state variable in the reference response model 27. , K _j represents the weight for each state variable. Next, as a fourth step, the selection probability P _i of each candidate is calculated based on the evaluation value V _i as in the following equation, and this selection probability P _i is calculated.
According to the above, M / 2 pairs of code sequences are formed while including duplication, and genetic operations of crossover and mutation are performed to generate a code sequence of M children as a code sequence group of the next generation. (Step 54).

[0014]

(Equation 2)

First, the crossover process will be described. FIG. 4 is a flowchart showing the crossover process, and FIG. 5 is a diagram for explaining the concept of the crossover process. Based on the selection probabilities P _i determined in the third step for each code string, it is determined that one set is composed of two code strings, and a combination of M / 2 sets including duplication is determined (step 61). One of the pairs determined in this way is shown in FIG. In the figure, "●"
And “○” indicate one and the other of the binary values “0” and “1”, respectively. Next, the crossover position is determined randomly for each pair (step 62). FIG. 5B shows an example of the crossover position set for a certain pair. If a portion of the code sequence on one side (the right side in the illustrated example) of the crossover position is called a partial code sequence, the partial code sequence is exchanged between the two code sequences forming the pair (step 63). ). FIG. 5C shows a code string after such replacement. By performing such crossover processing, finally, M code strings serving as new solution candidates are generated (step 64).

Next, the mutation process will be described.
FIG. 6 is a flowchart showing the mutation process.
Explains the concept of mutation processing. According to a predetermined mutation probability, a mutation occurrence position is declared for each code sequence at random (step 71). Fig. 7 (a)
Shows an example of the code sequence before the mutation occurrence position is declared, and FIG. 7B shows the mutation occurrence position declared in the code sequence shown in FIG. 7A. . And FIG.
As shown in (c), the bit at the position where the occurrence of the mutation is declared is inverted (step 72). When the code string group of the next generation is generated through the crossover and the mutation in this way, as a fifth step, it is determined whether or not a preset convergence condition is satisfied (step 55). If the convergence condition is satisfied, the determination of the control gain based on the genetic algorithm is terminated. If the convergence condition is not satisfied, the process returns to the third step (step 53). By repeating the processes from the third step to the fifth step, the control gain for the tension control is optimized.

Then, the control gain optimized as described above is set in the feedback controller 22, the switch 29 is switched to the control target 21 side, and the normal operation of the multi-drive control system is started.

[0018]

As described above, according to the present invention, by automatically adjusting the control gain of the tension control system using a genetic algorithm, the control gain can be adjusted at high speed and with high accuracy according to the situation. There is an effect that it becomes. When adjusting the control gain using an approximate model,
Since an actual operation system is not used, automatic adjustment can be performed safely. By providing an approximation model identification unit,
Even if the parameters of the model change due to deterioration of the material of the web or deterioration of the speed reducer, the control gain can be adjusted again and again.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a tension control device according to an embodiment of the present invention.

FIG. 2 is a block diagram illustrating an example of a configuration of a multi-drive control system.

FIG. 3 is a flowchart illustrating a procedure for adjusting a control gain.

FIG. 4 is a flowchart illustrating a crossover process.

FIG. 5 is a diagram illustrating the concept of crossover processing.

FIG. 6 is a flowchart illustrating a mutation process.

FIG. 7 is a diagram illustrating the concept of mutation processing.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Web material 2,3 Drive roll 4 Tension detector 5,6 Reduction gear 7,8 Electric motor 9,10 Speed detector 11,12 Drive device 13,14 Speed controller 15-17 Subtractor 18 Tension controller 19 Tension setting Device 20 speed setting device 21 controlled object 22 feedback controller 23 approximate model 24 model identification unit 25 control gain adjustment device 27 reference response model 28 evaluation value calculation unit 29 switching unit 51-55, 61-64, 71, 72 steps

Claims (4)

[Claims] 1. A web material is conveyed using a plurality of drive rolls arranged along a conveying direction of the web material and connected to an electric motor, and each of the electric motors is driven by a speed based on a speed command based on a line speed. One of the electric motors, each controlled by a control loop and coupled to two adjacent drive rolls, respectively, comprises a speed control loop as a minor loop and a tension of the web material between the two adjacent drive rolls. A tension control method for controlling a tension of the web material between the two adjacent driving rolls, the method including a tension control loop having a detection output as a feedback signal, wherein a control gain for the tension control loop is adjusted. The control gain is converted to a symbol string and the symbol string is connected to the control gain and a set thereof. A first step of inputting the control gain to be adjusted and the adjustment condition thereof as a code string representing the matching, and a second step of generating an initial group of the code string as a solution candidate in accordance with the adjustment condition And a third step of performing decoding to the control gain for each of the code strings and obtaining an evaluation value from a response obtained by the decoded control gain; and each of the code strings according to the evaluation value. A fourth step of performing a genetic operation based on a genetic algorithm on to generate a solution candidate code sequence as a child, and a fifth step of determining convergence, wherein the third step, the fourth step A tension control method, characterized in that the control gain is optimized by repeatedly executing the step 4 and the fifth step. 2. An evaluation model based on a response deviation between the approximate model and the reference response model, using an approximate model of a tension control system including the tension control loop and a reference response model giving a reference response. The tension control method according to claim 1, wherein the value is determined. 3. An evaluation function comparing means for calculating the evaluation value based on a response deviation between the approximation model and the reference response model, and a controller to which the control gain is given and which performs feedback control for the tension control. Switching means for switching the output signal of the controller from the approximation model to the control object side when a response within the reference range is obtained, and giving the same command to the control object and the approximation model; 3. The tension control method according to claim 2, wherein a model identification unit that corrects each parameter of the approximate model based on a deviation and an adjustment gain adjustment unit that performs the first to fifth steps are used. 4. A first driving roll and a second driving roll disposed along a transport direction of the web material, a first electric motor connected to the first driving roll and the second driving roll, respectively, and A second motor, and a tension detector disposed between the first drive roll and the second drive roll to detect a tension of the web material, wherein each of the motors is based on a line speed. Is controlled by a speed control loop based on a speed command, and one of the first electric motor and the second electric motor makes the speed control loop a minor loop and outputs a detection output from the tension detector as a feedback signal. A tension control loop that conveys the web material while controlling the tension by each of the drive rolls, wherein the drive control loop A controller that performs tension control in accordance with a control gain, an approximate model of a tension control system including the tension control loop, a reference response model that provides a reference response, and a response between the approximate model and the reference response model Evaluation function comparing means for calculating an evaluation value based on the deviation; switching means for switching an output signal of the controller from the approximation model to the control object side when a response within a reference range is obtained; and A model identification unit that corrects each parameter of the approximation model by giving the same command to the approximation model and a response deviation at that time; and a control gain adjustment unit that adjusts the control gain. The control gain is converted into a symbol string and the symbol string is connected. And to the code string, to generate an initial population of the code string to be the solution candidates in accordance with the adjustment conditions,
After the generation of the initial group, until the predetermined convergence condition is satisfied, the code strings are decoded into the control gain and input to the controller, and for each of the code strings according to the evaluation value at that time. And performing a genetic operation based on a genetic algorithm to generate a solution candidate code sequence as a child, thereby optimizing the control gain.