Control device, control system, and control method for wire straightening machine

Abstract

To improve the accuracy of feedforward control for a wire straightening machine.SOLUTION: A feedback control amount rb to a straightening machine 200 is determined by feedback control means 110 based on an error between a target curvature qr and a straightening curvature q. Feedforward compensation means 120 determines a feedforward compensation value rf from a measurement value do of the straightening machine 200 using a prediction model Mp. Learning means 130 performs machine learning on the prediction model Mp using teacher data Ds. The learning means 130 adds at least one combination including the measurement value do and a control amount r when the absolute value of an error eq is smaller than a reference value α to the teacher data Ds.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to a control device, a control system, and a control method for a wire straightening machine.

Background Art

[0002] Conventionally, a configuration for controlling a wire straightening machine has been known. For example, Japanese Patent Application Laid-Open No. 2016-196037 (Patent Document 1) discloses a wire straightening device that individually moves a straightening roller to a position for correcting the deformation based on information on the deformation of a wire. According to the wire straightening device, in the straightening of a wire, the burden of correction adjustment during actual operation can be eliminated and accurate correction can be automatically performed.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the wire straightening device disclosed in Patent Document 1, the position of the straightening roller is changed according to the amount of deformation (straightening curvature) of the wire after straightening (feedback control). Further, the position of the straightening roller is changed according to the winding position of the bobbin (feedforward compensation). In the feedforward compensation, the movement amount of the straightening roller is expressed as a linear expression of the winding position of the bobbin. However, the movement amount of the straightening roller is not uniform with respect to the winding position of the bobbin, and there are cases where it cannot be expressed as a linear expression of the winding position of the bobbin. Further, the winding position of the bobbin that determines the initial curvature of the wire is only one of the factors that affect the straightening curvature. In Patent Document 1, fluctuations in factors other than the winding position of the bobbin that affect the straightening accuracy are not considered.

[0005] The present disclosure has been made to solve the above problems, and an object thereof is to improve the accuracy of feedforward control for a wire straightening machine.

Means for Solving the Problems

[0006] A control device for a straightening machine according to an aspect of the present disclosure determines a feedforward compensation value corresponding to a feedback operation amount to the straightening machine so that the straightening curvature of a wire straightened by the straightening machine approaches a target curvature. The straightening machine includes a wire supply unit, a plurality of rollers, and a wire moving unit. The wire supply unit has a wound wire. The plurality of rollers are arranged alternately so as to sandwich the wire drawn from the wire supply unit and press the wire. The wire moving unit draws the wire from the wire supply unit and passes it through the plurality of rollers. The straightening curvature is the curvature of a first portion of the wire that has passed through the plurality of rollers. The feedback operation amount is determined by feedback control means based on the error between the target curvature and the straightening curvature. The sum of the feedback operation amount and the feedforward compensation value is output to the straightening machine as an operation amount. The control device includes feedforward compensation means and learning means. The feedforward compensation means determines a feedforward compensation value from a measured value obtained from the straightening machine using a prediction model. The learning means performs machine learning on the prediction model using teacher data. The learning means acquires at least one first combination including the target curvature, the operation amount corresponding to the target curvature, the measured value, and the straightening curvature corresponding to both the measured value and the operation amount. The learning means adds at least one second combination including the measured value and the operation amount when the absolute value of the error is smaller than a first reference value among the at least one first combination to the teacher data.

[0007] According to this disclosure, since teacher data is selected from the results of feedforward control, the prediction model can be adapted to the measured value and the characteristics of the straightening machine in real time in parallel with the feedforward control. As a result, the accuracy of feedforward control for the straightening machine can be improved.

[0008] In the above disclosure, the operation amount may be changed by changing at least one of the positions of a plurality of rollers and the moving speed of the wire. The control device may obtain, from the measured values, at least one of the moving speed of the wire, the position of the first part, the position of the second part of the wire before being pressed by the plurality of rollers, the dimension of the second part, the curvature of the second part, the winding position of the wire in the wire supply section, the position of each of the plurality of rollers, the radius of each of the plurality of rollers, the pitch of the plurality of rollers, and the pressure of the plurality of rollers.

[0009] According to this disclosure, the prediction model is optimized based on various factors that affect the correction curvature. As a result, the accuracy of the feedforward control for the straightening machine can be further improved.

[0010] In the above disclosure, the learning means may calculate the average value of the operation amount when the absolute value of the measured value is smaller than the second reference value in the teacher data. The learning means may approximate the relationship between the measured value represented by the prediction model and the feedforward compensation value as the difference between the operation amount and the average value, with the feedforward compensation value as the target variable and the measured value as the explanatory variable as a function. The learning means may end the machine learning when the ratio of the number of the fourth combinations in which the absolute value of the measured value is greater than the second reference value among at least one of the second combinations to the number of the third combinations in which the absolute value of the measured value is greater than the second reference value among at least one of the first combinations is greater than the third reference value.

[0011] According to this disclosure, since the machine learning is continued until the accuracy of the prediction model becomes sufficiently high, the accuracy of the feedforward control for the straightening machine can be sufficiently improved.

[0012] In the above disclosure, the learning means may calculate the average value of the operation amount when the absolute value of the measured value is smaller than the second reference value in the teacher data. The learning means may approximate the relationship between the measured value and the operation amount represented by the prediction model as a function having the operation amount as the objective variable and the measured value as the explanatory variable. The learning means may terminate the machine learning when the ratio of the number of the fourth combinations in which the absolute value of the measured value is greater than the second reference value among at least one second combination to the number of the third combinations in which the absolute value of the measured value is greater than the second reference value among at least one first combination is greater than the third reference value. The feedforward compensation means may determine, as the feedforward compensation value, a value obtained by subtracting the average value from the operation amount predicted from the measured value by the prediction model.

[0013] According to this disclosure, since the machine learning is continued until the accuracy of the prediction model becomes sufficiently high, the accuracy of the feedforward control for the correction machine can be sufficiently improved.

[0014] In the above disclosure, the learning means may resume the machine learning when the machine learning is terminated and the ratio is smaller than the third reference value.

[0015] According to this disclosure, since the prediction model is re-adapted to the characteristics of the correction machine in response to the change in the characteristics of the correction machine, it is possible to suppress a decrease in the accuracy of the feedforward control due to the change in the characteristics of the correction machine.

[0016] A control system according to another aspect of the present disclosure outputs, as an operation amount, the sum of a feedback operation amount and a feedforward compensation value to a straightening machine so that the straightening curvature of the wire straightened by the straightening machine approaches a target curvature. The straightening machine includes a wire supply unit, a plurality of rollers, and a wire moving unit. The wire supply unit has a wound wire. The plurality of rollers are arranged alternately so as to sandwich the wire drawn from the wire supply unit and press the wire. The wire moving unit draws the wire from the wire supply unit and passes it through the plurality of rollers. The straightening curvature is the curvature of the portion of the wire that has passed through the plurality of rollers. The control system includes a feedback control device and a learning device. The feedback control device determines a feedback operation amount based on the error between the target curvature and the straightening curvature. The feedforward compensation device determines a feedforward compensation value from the measured values obtained from the straightening machine using a prediction model. The learning device performs machine learning on the prediction model using teacher data. The learning device acquires at least one first combination including a target curvature, an operation amount corresponding to the target curvature, a measured value, and a straightening curvature corresponding to both the measured value and the operation amount. The learning device adds, to the teacher data, at least one second combination including a measured value and an operation amount when the absolute value of the error is smaller than a first reference value among the at least one first combination.

[0017] According to this disclosure, since teacher data is selected from the results of feedforward control, the prediction model can be adapted to the measured values and the characteristics of the straightening machine in real time in parallel with the feedforward control. As a result, the accuracy of the feedforward control for the straightening machine can be improved.

[0018] A control method according to another aspect of the present disclosure is a control method for a straightening machine that determines a feedforward compensation value corresponding to a feedback operation amount to the straightening machine so that the straightening curvature of the wire rod straightened by the straightening machine approaches a target curvature. The straightening machine includes a wire supply unit, a plurality of rollers, and a wire moving unit. The wire supply unit has a wound wire rod. The plurality of rollers are arranged alternately so as to sandwich the wire rod drawn from the wire supply unit and press the wire rod. The wire moving unit draws the wire rod from the wire supply unit and passes it through the plurality of rollers. The straightening curvature is the curvature of the portion of the wire rod that has passed through the plurality of rollers. The feedback operation amount is determined by a feedback control means based on the error between the target curvature and the straightening curvature. The sum of the feedback operation amount and the feedforward compensation value is output to the straightening machine as an operation amount. The control method includes a step of determining a feedforward compensation value from a measurement value obtained from the straightening machine using a prediction model, and a step of performing machine learning on the prediction model using teacher data. The step of performing machine learning includes a step of obtaining at least one first combination including a target curvature, an operation amount corresponding to the target curvature, a measurement value, and a straightening curvature corresponding to both the measurement value and the operation amount, and a step of adding to the teacher data at least one second combination including a measurement value and an operation amount when the absolute value of the error is smaller than a first reference value among the at least one first combination.

[0019] According to this disclosure, since teacher data is selected from the results of feedforward control, the prediction model can be adapted to the measurement value and the characteristics of the straightening machine in real time in parallel with the feedforward control. As a result, the accuracy of the feedforward control for the straightening machine can be improved.

Advantages of the Invention

[0020] According to the control device, control system, and control method for a wire rod straightening machine according to the present disclosure, the accuracy of the feedforward control for the wire rod straightening machine can be improved.

Brief Description of the Drawings

[0021]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Embodiments for Carrying Out the Invention

[0022] Hereinafter, the embodiments will be described in detail with reference to the drawings. In the drawings, the same or corresponding parts are denoted by the same reference numerals, and the description thereof will not be repeated in principle.

[0023] [Embodiment 1] [Application Example] FIG. 1 is a block diagram showing the functional configuration of the control device 100 of the wire straightening machine 200 according to Embodiment 1. As shown in FIG. 1, the control device 100 includes a feedback control means 110, a feedforward compensation means 120, a learning means 130, a storage means 140, a subtractor 150, and an adder 160.

[0024] The control device 100 outputs, as an operation amount r, the sum of the feedback operation amount rb to the straightening machine 200 and the feedforward compensation value rf to the straightening machine 200 so that the straightening curvature q of the wire straightened by the straightening machine 200 approaches the target curvature qr (for example, 0). Note that the control device 100 may output a plurality of operation amounts r to the straightening machine 200.

[0025] The storage means 140 stores a prediction model Mp and teacher data Ds. The control device 100 and the straightening machine 200 may be connected via a network (for example, the Internet or a cloud system) and may be arranged remotely from each other. Examples of the machine learning algorithm for constructing the prediction model Mp include a deep binary tree or a support vector machine.

[0026] Hereinafter, a configuration including the feedback control means 110 and the subtractor 150 is also referred to as a feedback control system, and a configuration including the feedforward compensation means 120 and the adder 160 is also referred to as a feedforward control system.

[0027] The subtractor 150 outputs an error eq (= qr - q) between the target curvature qr and the corrected curvature q to the feedback control means 110. The feedback control means 110 determines a feedback operation amount rb based on the error eq and outputs it to the adder 160. The feedforward compensation means 120 predicts a feedforward compensation value rf from a measurement value set do obtained from the corrector 200 using a prediction model Mp and outputs it to the adder 160. The measurement value set do includes at least one measurement value. The adder 160 outputs the sum of the feedback operation amount rb and the feedforward compensation value rf as an operation amount r to the corrector 200 and the learning means 130.

[0028] The learning means 130 performs machine learning on the prediction model Mp using teacher data Ds. The learning means 130 obtains at least one combination Cm1 (first combination) including the target curvature qr, the operation amount r corresponding to the target curvature qr, the measurement value set do, and the corrected curvature q corresponding to both the measurement value set do and the operation amount r. Among the at least one combination Cm1, when the absolute value of the error eq is smaller than a reference value α (first reference value) (when high-precision control is performed on the corrector 200), the learning means 130 adds at least one combination Cm2 (second combination) including the measurement value set do and the operation amount r as correct data for machine learning to the teacher data Ds. The reference value α can be appropriately determined based on, for example, actual machine experiments, simulations, product standard values, or manufacturing process control values.

[0029] The learning means 130 calculates the average value rb0 of the operation amount r when, in the teacher data Ds, the absolute value of each of at least one measurement value included in the measurement value set do is smaller than the reference value β (second reference value) corresponding to the measurement value (when the control system is in a steady state). The reference value β can be appropriately determined based on, for example, actual machine experiments, simulations, product standard values, or management values of the manufacturing process. The learning means 130 approximates the relationship between the measurement value set do and the feedforward compensation value rf as the difference (= r - rb0) between the operation amount r and the average value rb0 as a function (regression curve) with the feedforward compensation value rf as the objective variable and the measurement value set do as the explanatory variable. The prediction model Mp includes the function. When the ratio Cr of the number of combinations Cm4 (fourth combination) in which the absolute value of the measurement value is greater than the reference value β among at least one combination Cm2 to the number of combinations Cm3 (third combination) in which the absolute value of any of the measurement values included in the measurement value set do is greater than the reference value β among at least one combination Cm1 is greater than the reference value δ (third reference value), the machine learning for the prediction model Mp is terminated.

[0030] When the machine learning (initial learning or additional learning) for the prediction model Mp is completed and the ratio Cr is less than or equal to the reference value δ, assuming that the characteristics of the correction machine 200 have changed, the machine learning (additional learning) for the prediction model Mp is restarted. The reference value δ can be appropriately determined based on, for example, actual machine experiments, simulations, product standard values, or management values of a certain manufacturing process. The characteristics of the correction machine 200 include, for example, the correspondence relationship between the measurement value set do and the operation amount r and the correction curvature q.

[0031] According to the control device 100, since the teacher data Ds is selected from the results of the feedforward control, the prediction model Mp can be real-time adapted to the measured value set do and the characteristics of the corrector 200 in parallel with the feedforward control. As a result, the burden on the user required for the adjustment of the corrector 200 can be reduced, and the accuracy of the feedforward control for the corrector 200 can be improved to reduce the deviation and variation of the correction curvature. Also, since the prediction model is optimized based on various factors that affect the correction curvature q, the accuracy of the feedforward control for the corrector can be further improved. Further, since the machine learning is continued until the accuracy of the prediction model Mp becomes sufficiently high, the accuracy of the feedforward control for the corrector 200 can be sufficiently improved. Furthermore, since the prediction model Mp is re-adapted to the characteristics according to the change in the characteristics of the corrector 200, the decrease in the accuracy of the feedforward control due to the change in the characteristics of the corrector 200 can be suppressed.

[0032] FIG. 2 is a diagram showing a specific configuration of the corrector 200 in FIG. 1. In FIG. 2, the X-axis, Y-axis, and Z-axis are orthogonal to each other. The same applies to FIG. 3 described later. As shown in FIG. 2, the corrector 200 includes a bobbin Bn (wire supply unit), a plurality of rollers Ro1 to Ro7, a cutting unit Ct, a wire moving unit Mu, sensors Sd1, Sd2, Sd3, Sd4, Sd5, Sd6, Sd7, sensors Sd10, Sd11, Sd12, Sd13, Sd14, Sd15, Sd16, Sd17, sensors SL2, SL3, SL4, SL5, SL6, and servo motors Ms2, Ms4, Ms6.

[0033] In the bobbin Bn, a wire Wr is wound around a winding shaft Ws extending in the X-axis direction. The cutting unit Ct cuts the wire Wr at regular intervals to generate a plurality of cut wires Cw.

[0034] The plurality of rollers Ro1 to Ro7 are arranged alternately in the Y-axis direction so as to sandwich the wire rod Wr drawn from the bobbin Bn in the Z-axis direction, and press the wire rod Wr in the Z-axis direction. The servo motors Ms2, Ms4, and Ms6 change the positions of the rollers Ro2, Ro4, and Ro6 in the Z-axis direction according to the operation amount r from the control device 100. Each of the rollers Ro2, Ro4, and Ro6 presses the wire rod Wr in the positive direction of the Z-axis. Each of the rollers Ro3 and Ro5 presses the wire rod Wr in the negative direction of the Z-axis.

[0035] The wire rod moving unit Mu draws the wire rod Wr from the bobbin Bn and moves it along the Y-axis direction to the cutting unit Ct via the plurality of rollers Ro1 to Ro7. The wire rod moving unit Mu includes a servo motor Ms8 and an encoder Se. The servo motor Ms8 changes the moving speed of the wire rod Wr in the Y-axis direction by changing the rotation speed according to the operation amount r from the control device 100. The measured value regarding the wire rod Wr measured by the encoder Se is included in the measured value set do. The said measured value is used by the control device 100 to acquire the moving speed of the wire rod Wr in the Y-axis direction.

[0036] The plurality of measured values regarding the plurality of rollers Ro1 to Ro7 respectively measured by the sensors Sd1 to Sd7 are included in the measured value set do. The said plurality of measured values are used by the control device 100 to acquire the respective positions of the plurality of rollers Ro1 to Ro7 or the roller pitch (the interval between adjacent rollers in the Y-axis direction) of the plurality of rollers Ro1 to Ro7. Each of the sensors Sd1 to Sd7 includes, for example, a displacement sensor or a temperature sensor.

[0037] The measured value regarding the wire rod Wr measured by the sensor Sd10 is included in the measured value set do. The said measured value is used by the control device 100 to acquire the winding position of the bobbin Bn (the position of the wire rod Wr that is the outermost in the radial direction of the winding axis Ws). The sensor Sd10 includes, for example, a displacement sensor.

[0038] The measured values related to the wire rod Wr measured by the sensor Sd11 are included in the set of measured values do. The measured values are used by the control device 100 to obtain the position, dimensions, and initial curvature of the portion (second portion) of the wire rod Wr before being pressed by the plurality of rollers Ro1 to Ro7. The sensor Sd11 includes, for example, a length measuring sensor or a displacement sensor.

[0039] The plurality of measured values related to the plurality of rollers Ro2 to Ro6 respectively measured by the sensors Sd12 to Sd16 are included in the set of measured values do. The plurality of measured values are used by the control device 100 to respectively obtain the positions of the rollers Ro2 to Ro6. Each of the sensors Sd12 to Sd16 includes, for example, a displacement sensor.

[0040] The plurality of measured values measured by the sensors SL2 to SL6 are included in the set of measured values do. The plurality of measured values are used by the control device 100 to respectively obtain the plurality of pressures applied to the wire rod Wr by the rollers Ro2 to Ro6. Each of the sensors SL2 to SL6 includes, for example, a load sensor.

[0041] The measured value related to the wire rod Wr measured by the sensor Sd17 is used by the control device 100 to obtain the corrected curvature q of the portion (first portion) of the wire rod Wr after being pressed by the plurality of rollers Ro1 to Ro7 and before being cut by the cutting portion Ct. The sensor Sd17 includes, for example, a displacement sensor.

[0042] Note that the corrected curvature q may be calculated from the measured values of the cut wire rod Cw. FIG. 3 is a diagram showing another example of the specific configuration of the straightening machine 200 in FIG. 1. The configuration of the straightening machine 200 shown in FIG. 3 is a configuration in which a sensor Sd18 is added in place of the sensor Sd17 in FIG. 2. Since the other configurations are the same as those in FIG. 2, the description of the same configurations will not be repeated.

[0043] As shown in FIG. 3, the measured values regarding the cut wire Cw measured by the sensor Sd18 are used by the control device 100 to obtain the correction curvature q.

[0044] FIG. 4 is a diagram showing the correction process of the wire Wr by the plurality of rollers Ro1 to Ro7 in FIGS. 2 and 3. In FIG. 4, the vertical axis represents the bending moment applied to the wire Wr by each of the plurality of rollers Ro1 to Ro7, and the horizontal axis represents the curvature of the wire Wr. The region of the bending moments Me1 to Me2 represents the elastic region where the wire Wr elastically deforms. The bending moment Me2 is positive, and the bending moment Me1 is negative. The absolute value of the bending moment Me1 is equal to the absolute value of the bending moment Me2. When the absolute value of the bending moment exceeds the elastic region, plastic deformation due to the yield stress occurs in the wire Wr. The elastic region changes according to the yield stress of the wire Wr.

[0045] With reference to FIGS. 2 and 4 together, the concept of correction will be described. Strictly speaking, the curvature and bending moment of the wire Wr are determined by the front and rear rollers, but here it will be described in a simplified manner. The curvature and bending moment at the point Ps are κ s (<0) and 0, respectively. The curvature κ s is the initial curvature. The process from the point Ps to the point P2 is the correction process by the roller Ro2. In this process, the roller Ro2 presses the wire Wr in the positive direction of the Z-axis. The curvature and bending moment at the point P2 are κ2 (>0) and M2 (>Me2), respectively.

[0046] The process from the point P2 to the point P3 is the correction process by the roller Ro3. In this process, the roller Ro3 presses the wire Wr in the negative direction of the Z-axis. The curvature and bending moment at the point P3 are κ3 (<0) and M3 (<Me1), respectively.

[0047] The process from point P3 to point P4 is the correction process by roller Ro4. In this process, roller Ro4 presses the wire Wr in the positive direction of the Z-axis. The curvature and bending moment at point P4 are κ4(>0) and M4(>Me2), respectively. The curvature κ4 is smaller than the curvature κ2. The bending moment M4 is smaller than the bending moment M2.

[0048] The process from point P4 to point P5 is the correction process by roller Ro5. In this process, roller Ro5 presses the wire Wr in the negative direction of the Z-axis. The curvature and bending moment at point P5 are κ5(<0) and M5(<Me1), respectively. The curvature κ5 is larger than the curvature κ3. The bending moment M5 is larger than the bending moment M3.

[0049] The process from point P5 to point P6 is the correction process by roller Ro6. In this process, roller Ro6 presses the wire Wr in the positive direction of the Z-axis. The curvature and bending moment at point P6 are κ6(>0) and M6(>Me2), respectively. The curvature κ6 is smaller than the curvature κ4. The bending moment M6 is smaller than the bending moment M4.

[0050] After the pressing of the wire Wr by the plurality of rollers Ro1~Ro7 is completed, as the bending moment of the wire Wr decreases, the curvature of the wire Wr changes from the curvature κ6 to the corrected curvature κ f (<κ6).

[0051] Figure 5 is a distribution diagram of the measurement value set do and the operation amount r. In Figure 5, for the sake of easy understanding of the explanation regarding the teacher data Ds and the prediction model Mp, the distribution diagram in the simplest case where there is one measurement value included in the measurement value set do (the dimension of the measurement value set do is 1) is shown. The same applies to Figures 6, 7, 9, and 10 which will be described later.

[0052] In FIG. 5, the square points represent the combinations Cm1 among the plurality of combinations Cm1 where the absolute value of the error eq is greater than or equal to the reference value α, and the round points represent the combinations Cm2 where the absolute value of the error eq is less than the reference value α. The learning means 130 adds the plurality of combinations Cm2 represented by the round points to the teacher data Ds.

[0053] FIG. 6 is a diagram showing the relationship between the operation amount r in FIG. 5 and the target variable rf of the prediction model Mp. As shown in FIG. 6, the data obtained by subtracting the average value rb0 from each operation amount r of the plurality of combinations Cm2 corresponds to the target variable rf.

[0054] FIG. 7 is a diagram showing the relationship between the target variable rf and the explanatory variable do represented by the learned prediction model Mp. This relationship is approximated as a regression curve Rc by the learning means 130. As shown in FIG. 7, the prediction model Mp receives the measurement value set do and outputs the target variable rf corresponding to the measurement value set do on the regression curve Rc.

[0055] FIG. 8 is a Venn diagram showing the inclusion relationships of four data sets Scm1, Scm2, Scm3, and Scm4, each including four combinations Cm1 to Cm4. The data set Scm1 is a data set including all of the combination Cm1 including a target curvature qr, an operation amount r, a measurement value set do, and a correction curvature q. The data set Scm2 is a data set including all of the combination Cm2 including the measurement value set do and the operation amount r when the absolute value of the error eq is smaller than a reference value α among the data set Scm1. The data set Scm3 is a data set including all of the combination Cm3 when the absolute value of each measurement value of the measurement value set do is larger than a reference value β corresponding to the measurement value among the data set Scm1. The data set Scm4 is a data set including all of the combination Cm4 when the absolute value of each measurement value of the measurement value set do is larger than the reference value β corresponding to the measurement value among the data set Scm2. As shown in FIG. 8, the data set Scm1 includes the data sets Scm2 and Scm3. The data set Scm4 is the common part (intersection) of the data sets Scm2 and Scm3. The average value rb0 is the average value of the data set obtained by removing the data set Scm4 from the data set Scm2.

[0056] FIG. 9 is a diagram showing time charts of each of a measurement value set do, an operation amount r, and a correction curvature q when the prediction model Mp is unlearned. As shown in FIG. 9, at the time of 500 seconds, a measurement value set do that changes in the same manner as the step response of the first-order lag system is input. Pulse-like noise is superimposed on the correction curvature q in the time interval from 500 seconds to 600 seconds in response to the measurement value set do.

[0057] FIG. 10 is a diagram showing time charts of each of the measurement value set do, the manipulated variable r, and the correction curvature q when the prediction model Mp has been learned. Also in FIG. 10 as in FIG. 9, at the time of 500 seconds, the measurement value set do that changes in the same manner as the step response of the first-order lag system is input. However, almost no pulse-like noise as shown in FIG. 9 occurs in the correction curvature q. In FIG. 10, the fluctuation of the correction curvature q is suppressed by correcting the manipulated variable r by the feedforward compensation value rf predicted by the learned prediction model Mp.

[0058] FIG. 11 is a diagram showing a flowchart showing the flow of processing performed by each of the feedback control system, the feedforward control system, and the learning means 130 of FIG. 1. The routines corresponding to the flowcharts of the feedback control system and the feedforward control system are executed, for example, every sampling time. The routine corresponding to the flowchart of the learning means 130 is executed, for example, in response to the first execution of the routine corresponding to each flowchart of the feedforward control system. Hereinafter, the steps will be simply described as S.

[0059] As shown in FIG. 11, the subtractor 150 calculates the error eq between the target curvature qr and the correction curvature q in S111 and outputs it to the feedback control means 110. The feedback control means 110 determines the feedback manipulated variable rb based on the error eq in S312, outputs it to the adder 160, and ends the process.

[0060] The feedforward compensation means 120 determines the feedforward compensation value rf from the measurement value set do in S121 and outputs it to the adder 160. The adder 160 outputs the sum of the feedback manipulated variable rb and the feedforward compensation value rf as the manipulated variable r to the corrector 200 and the learning means 130 in S122 and ends the process.

[0061] The learning means 130 performs machine learning on the prediction model Mp using the teacher data Ds in S130 and ends the process. When a plurality of operation amounts r are output from the control device 100 to the correction machine 200, the process shown in FIG. 11 is performed for each of the plurality of operation amounts r.

[0062] FIG. 12 is a flowchart showing the specific process flow of the machine learning process S130 in FIG. 11. In S131, the learning means 130 acquires the target curvature qr, the operation amount r, the measurement value set do, and the correction curvature q, calculates the error eq, and advances the process to S132. In S132, the learning means 130 determines whether the absolute value of the error eq is smaller than the reference value α. If the absolute value of the error eq is greater than or equal to the reference value α (NO in S132), the learning means 130 advances the process to S137.

[0063] If the absolute value of the error eq is smaller than the reference value α (YES in S132), in S133, the learning means 130 adds the combination Cm2 including the measurement value set do and the operation amount r to the teacher data Ds and advances the process to S134. In S134, when the absolute value of each measurement value in the measurement value set do in the teacher data Ds is smaller than the reference value β corresponding to the measurement value, the learning means 130 calculates the average value rb0 of the operation amount r and advances the process to S135. In S135, the learning means 130 approximates the relationship between the measurement value set do expressed by the prediction model Mp and the feedforward compensation value rf as the difference between the operation amount r and the average value rb0 as a function (regression curve) with the feedforward compensation value rf as the objective variable and the measurement value set do as the explanatory variable, and advances the process to S136.

[0064] In S136, the learning means 130 calculates the ratio Cr (= N2 / N1) of the number N2 of combinations Cm4 in which the absolute value of the measurement value set do is greater than the reference value β among at least one combination Cm2 to the number N1 of combinations Cm3 in which the absolute value of the measurement value set do is greater than the reference value β among at least one combination Cm1, and proceeds with the process to S137. In S137, the learning means 130 determines whether the ratio Cr is greater than the reference value δ. If the ratio Cr is less than or equal to the reference value δ (NO in S137), the learning means 130 assumes that the accuracy of the feedforward control to the corrector 200 is insufficient and returns the process to S131. If the ratio Cr is greater than the reference value δ (YES in S137), assuming that the accuracy of the feedforward control to the corrector 200 has sufficiently increased, the learning means 130 ends the machine learning.

[0065] FIG. 13 is a diagram showing the flow of additional learning processing performed by the learning means 130 in FIG. 1. The processing shown in FIG. 13 is executed, for example, every sampling time after the first machine learning starts. In S131A, the learning means 130 determines whether the machine learning has ended and whether the ratio Cr is less than or equal to the reference value δ. If the machine learning has not ended or the ratio Cr is greater than the reference value δ (NO in S131A), assuming that the characteristics of the corrector 200 have not changed and the existing learned prediction model Mp is suitable for the corrector 200, the learning means 130 ends the process.

[0066] If the machine learning has ended and the ratio Cr is less than or equal to the reference value δ (YES in S131A), the learning means 130 assumes that the characteristics of the corrector 200 have changed after the previous machine learning ended, and in order to re - adapt the prediction model Mp to the characteristics of the corrector 200, resumes the machine learning for the prediction model Mp in S130 similar to FIG. 8. According to the control device 100, since additional learning for the prediction model Mp is performed in response to the change in the characteristics of the corrector 200, it is possible to suppress a decrease in the accuracy of the feedforward control due to the change in the characteristics of the corrector 200.

[0067] [Modification Example 1 of Embodiment 1] In Embodiment 1, the case where the relationship between the measurement value set do, the operation amount r, and the feedforward compensation value rf as the difference between the operation amount r and the average value rb0 is expressed by the prediction model was described. In Modification 1 of Embodiment 1, the case where the relationship expressed by the prediction model is the relationship between the measurement value set do and the operation amount r will be described.

[0068] FIG. 14 is a flowchart showing another example of the specific processing flow of the machine learning process S130 in FIG. 11. The flowchart shown in FIG. 14 is a flowchart in which S135 in FIG. 12 is replaced with S135B. As shown in FIG. 14, after performing S131 to S134 in the same manner as in Embodiment 1, the learning means 130 approximates, in S135B, the relationship between the measurement value set do and the operation amount r expressed by the prediction model Mp as a function (regression curve) having the operation amount r as the target variable and the measurement value set do as the explanatory variable, and proceeds with the processing to S136. The learning means 130 performs S136 and S137 in the same manner as in Embodiment 1 to end the processing. The feedforward compensation means 120 predicts the operation amount r from the measurement value set do using the prediction model Mp, and outputs the value obtained by subtracting the average value rb0 from the operation amount r (= r - rb0) to the adder 160 as the feedforward compensation value rf.

[0069] [Modification 2 of Embodiment 1] In Embodiment 1, the configuration including both the feedback control system and the feedforward control system was described. In Modification 2 of Embodiment 1, the configuration not including the feedback control system will be described.

[0070] FIG. 15 is a block diagram showing the functional configuration of the control device 100A of the wire straightening machine 200 according to Modification 2 of Embodiment 1. The configuration of the control device 100A is a configuration in which the subtractor 150 and the feedback control means 110 are removed from the control device 100 in FIG. 1. Since the rest is the same, the description of the same configuration will not be repeated. Note that the adder 160 may not be included in the control device 100A.

[0071] The control device 100A determines a feedforward compensation value rf such that the correction curvature q of the wire rod corrected by the straightening machine 200 approaches the target curvature qr. According to the control device 100A, while retaining the existing feedback control system, by adding a control device to the feedback control system, the existing feedback control system can be easily extended to a control system including a feedforward control system and a learning function.

[0072] As described above, according to the control device and the control method according to Embodiment 1 and Modifications 1 and 2, the accuracy of the feedforward control for the straightening machine of the wire rod can be improved.

[0073] [Embodiment 2] In Embodiment 1, the case where the feedback control system, the feedforward control system, and the configuration for performing machine learning on the prediction model are included in one control device has been described. In Embodiment 3, a configuration in which the feedback control system, the feedforward control system, and the configuration for performing machine learning on the prediction model are separated into different devices will be described.

[0074] FIG. 16 is a block diagram showing the functional configuration of the control system 2 according to Embodiment 2. In FIG. 16, the components with the same reference numerals as those in FIG. 1 have the same functions as the components specified by the corresponding reference numerals described in Embodiment 1, and thus the description of the same components will not be repeated.

[0075] As shown in FIG. 16, the control system 2 includes a feedback control device 11, a feedforward compensation device 12, and a learning device 13. The feedback control device 11 includes feedback control means 110 and a subtractor 150. The feedforward compensation device 12 includes feedforward compensation means 120 and an adder 160. The learning device 13 includes learning means 130 and storage means 140. The feedback control device 11, the feedforward compensation device 12, the learning device 13, and the straightening machine 200 may be connected to each other via a network and may be arranged remotely from each other. Note that the adder 160 may be included in the feedback control device 11 instead of the feedforward compensation device 12.

[0076] According to the control system 2, the existing control system can be easily expanded by adding a feedforward compensation device and a learning device to the feedback control device while leaving the existing feedback control device.

[0077] As described above, according to the control system and the control method according to Embodiment 2, the accuracy of the feedforward control for the wire straightening machine can be improved.

[0078] [Embodiment 3] In Embodiment 3, as an example of the control device according to Embodiment 1, a configuration in which the control device includes a PLC (Programmable Logic Controller) will be described.

[0079] <Example of Network Configuration of Control System> FIG. 17 is a schematic diagram showing an example of the network configuration of the control system 3 according to Embodiment 3. As shown in FIG. 17, the control system 3 includes a group of devices configured such that a plurality of devices can communicate with each other. Typically, the device may include a control device 300 that is a processing entity that executes a control program and a peripheral device connected to the control device 300. The control device 300 has the same functional configuration as the control device 100 shown in FIG. 1.

[0080] The control device 300 corresponds to an industrial controller that controls control targets such as various facilities or devices. The control device 300 is a kind of computer that executes control calculations and typically includes a PLC (Programmable Logic Controller). The control device 300 is connected to the straightening machine 200 via the field network 20. The control device 300 exchanges data with at least one straightening machine 200 via the field network 20.

[0081] The control calculations executed in the control device 300 include a process of collecting data collected or generated in the straightening machine 200, a process of generating data such as command values (operation amounts) for the straightening machine 200, and a process of transmitting the generated output data to the target straightening machine 200. The data collected or generated in the straightening machine 200 includes data related to measurement values obtained from the straightening machine 200 and control amounts as a result of the actual operation of the straightening machine 200 according to the command values. The command value for the straightening machine 200 is determined by adding a feedforward compensation value predicted from the measurement values by a prediction model to a temporarily calculated operation amount based on the error between the control target value (target curvature) calculated based on the control program executed by the control device 300 and the actual control amount (straightening curvature).

[0082] The field network 20 preferably employs a bus or network that performs fixed-cycle communication. Known examples of such a bus or network that performs fixed-cycle communication include EtherCAT (registered trademark), EtherNet / IP (registered trademark), DeviceNet (registered trademark), or CompoNet (registered trademark). EtherCAT (registered trademark) is preferred in terms of ensuring the arrival time of data.

[0083] The field network 20 can connect to other correction machines 200 and any field devices. The field devices include actuators that exert some physical action on robots or conveyors on the field side, and input / output devices that exchange information with the field.

[0084] The control device 300 is also connected to other devices via the upper network 32. The upper network 32 is connected to the Internet 900, which is an external network, via the gateway 700. The upper network 32 may adopt Ethernet (registered trademark), which is a common network protocol, or EtherNet / IP (registered trademark). More specifically, at least one server device 600 and at least one display device 500 may be connected to the upper network 32.

[0085] As the server device 600, a database system or a manufacturing execution system (MES) is assumed. The manufacturing execution system acquires information from manufacturing devices or facilities to be controlled, monitors and manages the entire production, and can also handle order information, quality information, or shipping information, etc. Not limited to these, devices that provide information system services may be connected to the upper network 32. As the information system service, a process of acquiring information from manufacturing devices or facilities to be controlled and performing macro or micro analysis is assumed. For example, as the information system service, data mining for extracting some characteristic trends included in the information from manufacturing devices or facilities to be controlled, or a machine learning tool for performing machine learning based on the information from the facilities or machines to be controlled is assumed.

[0086] The display device 500 receives an operation from the user, outputs a command according to the user operation to the control device 300, and graphically displays the calculation result, etc. in the control device 300.

[0087] A support device 400 can be connected to the control device 300. The support device 400 may be connected to the control device 300 via the upper-level network 32 or the Internet 900. The support device 400 is a device that supports the preparations necessary for the control device 300 to control the controlled object. Specifically, the support device 400 provides a development environment for programs executed on the control device 300 (such as program creation and editing tools, parsers, and compilers), a setting environment for setting configuration information (configuration) of the control device 300 and various devices connected to the control device 300, a function of outputting the generated program to the control device 300, and a function of online correcting and changing programs executed on the control device 300, etc.

[0088] In the control system 3, the control device 300, the support device 400, and the display device 500 are each configured as separate entities, but a configuration in which all or part of these functions are integrated into a single device may be adopted.

[0089] The control device 300 is used not only at one production site but also at other production sites. Also, it may be used in a plurality of different lines within one production site.

[0090] <Hardware configuration example of the control device> FIG. 18 is a block diagram showing a hardware configuration example of the control device 300 in FIG. 17. As shown in FIG. 18, the control device 300 includes a processor 302, a main memory 304, a storage 360, a memory card interface 312, an upper-level network controller 306, a field network controller 308, a local bus controller 316, and a USB controller 370 that provides a USB (Universal Serial Bus) interface. These components are connected via a processor bus 318.

[0091] The processor 302 corresponds to an arithmetic processing unit that executes control operations and is composed of a CPU (Central Processing Unit) and / or a GPU (Graphics Processing Unit), etc. Specifically, the processor 302 reads out the program stored in the storage 360, expands it in the main memory 304, and executes it, thereby realizing control operations for the correction machine 200.

[0092] The main memory 304 is composed of a volatile storage device such as a DRAM (Dynamic Random Access Memory) and / or an SRAM (Static Random Access Memory), etc. The storage 360 is composed of a non-volatile storage device such as an SSD (Solid State Drive) and / or an HDD (Hard Disk Drive), etc.

[0093] Stored in the storage 360 are a control program Pc, teacher data Ds, and a prediction model Mp. The storage 360 corresponds to the storage means 140 in FIG. 1. The control program Pc includes a program for integrally controlling the control device 300 and realizing each function of the control device 300. That is, the processor 302 that executes the control program Pc corresponds to the feedback control system (feedback control means 110 and subtractor 150), the feedforward control system (feedforward compensation means 120 and adder 160), and the learning means 130 in FIG. 1.

[0094] The memory card interface 312 accepts a memory card 314, which is an example of a removable storage medium. The memory card interface 312 enables arbitrary data to be read from and written to the memory card 314.

[0095] The upper network controller 306 exchanges data with any information processing device connected to the upper network 32 via the upper network 32 (for example, a local area network).

[0096] The field network controller 308 exchanges data with the straightening machine 200 via the field network 20.

[0097] The local bus controller 316 exchanges data with any functional unit 380 that constitutes the control device 300 via the local bus 122. The functional unit 380 includes, for example, an analog I / O unit responsible for input and / or output of analog signals, a digital I / O unit responsible for input and / or output of digital signals, and a counter unit that receives pulses from an encoder or the like.

[0098] The USB controller 370 exchanges data with any information processing device via a USB connection. For example, a support device 400 is connected to the USB controller 370.

[0099] As described above, according to the control device and the control method according to Embodiment 3, the accuracy of the feedforward control for the straightening machine of the wire rod can be improved.

[0100] <Appendix> The present embodiment as described above includes the following technical ideas.

[0101] [Configuration 1] A control device (100, 100A, 300) for the straightening machine (200), which determines a feedforward compensation value (rf) corresponding to a feedback operation amount (rb) to the straightening machine (200) so that the straightening curvature (q) of the wire rod (Wr) corrected by the straightening machine (200) approaches a target curvature (qr). The straightening machine (200) includes: A wire supply unit (Bn) having the wound wire rod (Wr); A plurality of rollers (Ro1 to Ro7) that are alternately arranged so as to sandwich the wire rod (Wr) drawn out from the wire supply unit (Bn) and press the wire rod (Wr). It includes a wire moving part (Mu) that pulls out the wire (Wr) from the wire supply part (Bn) and passes it through the plurality of rollers (Ro1 to Ro7). The correction curvature (q) is the curvature of the first part of the wire (Wr) that has passed through the plurality of rollers (Ro1 to Ro7). The feedback operation amount (rb) is determined by the feedback control means (110) based on the error (eq) between the target curvature (qr) and the correction curvature (q). The sum of the feedback operation amount (rb) and the feedforward compensation value (rf) is output as the operation amount (r) to the straightening machine (200). The control devices (100, 100A, 300) are as follows: A feedforward compensation means (120) that determines the feedforward compensation value (rf) from the measured value (do) obtained from the straightening machine (200) using a prediction model (Mp). A learning means (130) that performs machine learning on the prediction model (Mp) using teacher data (Ds). The learning means (130) is as follows: At least one first combination (Cm1) including the target curvature (qr), the operation amount (r) corresponding to the target curvature (qr), the measured value (do), and the correction curvature (q) corresponding to both the measured value (do) and the operation amount (r) is obtained. Among the at least one first combination (Cm1), when the absolute value of the error (eq) is smaller than the first reference value (α), at least one second combination (Cm2) including the measured value (do) and the operation amount (r) is added to the teacher data (Ds). The control devices (100, 100A, 300) of the straightening machine (200).

[0102] [Configuration 2] The operation amount (r) changes at least one of the positions of the plurality of rollers (Ro2 to Ro6) and the moving speed of the wire (Wr). The control devices (100, 100A, 300) obtain from the measured value (do): The moving speed of the wire (Wr). The position of the first part The position of the second part of the wire (Wr) before being pressed by the plurality of rollers (Ro1 to Ro7), the dimension of the second part, and the curvature of the second part The winding position of the wire (Wr) in the wire supply unit (Bn) Among the positions of each of the plurality of rollers (Ro2 to Ro6), the radius of each of the plurality of rollers (Ro1 to Ro7), the pitch of the plurality of rollers (Ro1 to Ro7), and the pressure of the plurality of rollers (Ro2 to Ro6) The control device (100, 100A, 300) of the straightening machine (200) according to Configuration 1, which acquires at least one of them

[0103] [Configuration 3] The learning means (130) is In the teacher data (Ds), when the absolute value of the measurement value (do) is smaller than the second reference value (β), the average value (rb0) of the operation amount (r) is calculated The relationship between the measurement value (do) represented by the prediction model (Mp), the operation amount (r), and the feedforward compensation value (rf) as the difference from the average value (rb0) is approximated as a function with the feedforward compensation value (rf) as the target variable and the measurement value (do) as the explanatory variable When the ratio (Cr) of the number of the fourth combinations (Cm4) in which the absolute value of the measurement value (do) is greater than the second reference value (β) among the at least one second combination (Cm2) to the number of the third combinations (Cm3) in which the absolute value of the measurement value (do) is greater than the second reference value (β) among the at least one first combination (Cm1) is greater than the third reference value (δ), the machine learning is terminated. The control device (100, 100A, 300) according to Configuration 1 or 2

[0104] [Configuration 4] The learning means (130) is In the teacher data (Ds), when the absolute value of the measurement value (do) is smaller than the second reference value (β), the average value (rb0) of the operation amount (r) is calculated The relationship between the measured value (do) and the manipulated variable (r) represented by the prediction model (Mp) is approximated as a function with the manipulated variable (r) as the target variable and the measured value (do) as the explanatory variable. When the ratio (Cr) of the number of the fourth combinations (Cm4) in which the absolute value of the measured value (do) is greater than the second reference value (β) among the at least one second combination (Cm2) to the number of the third combinations (Cm3) in which the absolute value of the measured value (do) is greater than the second reference value (β) among the at least one first combination (Cm1) is greater than the third reference value (δ), the machine learning is terminated. The feedforward compensation means (120) determines, as the feedforward compensation value (rf), a value obtained by subtracting the average value (rb0) from the manipulated variable (r) predicted from the measured value (do) by the prediction model (Mp). The control device (100, 100A, 300) according to Configuration 1 or 2.

[0105] [Configuration 5] When the machine learning is terminated and the ratio (Cr) is smaller than the third reference value (δ), the learning means (130) resumes the machine learning. The control device (100, 100A, 300) according to Configuration 3 or 4.

[0106] [Configuration 6] A control system (2) that outputs, as the manipulated variable (r), the sum of the feedback manipulated variable (rb) and the feedforward compensation value (rf) to the straightening machine (200) so that the straightening curvature (q) of the wire (Wr) straightened by the straightening machine (200) approaches the target curvature (qr). The straightening machine (200) A wire supply unit (Bn) having the wound wire (Wr), A plurality of rollers (Ro1 to Ro7) that are alternately arranged so as to sandwich the wire (Wr) drawn from the wire supply unit (Bn) and press the wire (Wr), And a wire moving unit (Mu) that draws the wire (Wr) from the wire supply unit (Bn) and passes it through the plurality of rollers (Ro1 to Ro7). The correction curvature (q) is the curvature of the portion of the wire (Wr) that has passed through the plurality of rollers (Ro1 to Ro7), a feedback control device (11) that determines the feedback operation amount (rb) based on the error (eq) between the target curvature (qr) and the correction curvature (q); a feedforward compensation device (12) that determines the feedforward compensation value (rf) from the measured value (do) obtained from the straightening machine (200) using a prediction model (Mp); a learning device (13) that performs machine learning on the prediction model (Mp) using teacher data (Ds), wherein the learning device (13) acquires at least one first combination (Cm1) including the target curvature (qr), the operation amount (r) corresponding to the target curvature (qr), the measured value (do), and the correction curvature (q) corresponding to both the measured value (do) and the operation amount (r); a control system (2) that adds at least one second combination (Cm2) including the measured value (do) and the operation amount (r) to the teacher data (Ds) when the absolute value of the error (eq) is smaller than a first reference value (α) among the at least one first combination (Cm1).

[0107] [Configuration 7] A control method for the straightening machine (200) that determines a feedforward compensation value (rf) corresponding to a feedback operation amount (rb) to the straightening machine (200) so that the correction curvature (q) of the wire (Wr) corrected by the straightening machine (200) approaches the target curvature (qr), wherein the straightening machine (200) includes a wire supply unit (Bn) having the wound wire (Wr), a plurality of rollers (Ro1 to Ro7) that are alternately arranged so as to sandwich the wire (Wr) drawn from the wire supply unit (Bn) and press the wire (Wr), and a wire moving unit (Mu) that draws the wire (Wr) from the wire supply unit (Bn) and passes it through the plurality of rollers (Ro1 to Ro7). The correction curvature (q) is the curvature of the first portion of the wire material (Wr) that has passed through the plurality of rollers (Ro1 to Ro7), The feedback operation amount (rb) is determined by the feedback control means (110) based on the error (eq) between the target curvature (qr) and the correction curvature (q), The sum of the feedback operation amount (rb) and the feedforward compensation value (rf) is output as the operation amount (r) to the straightening machine (200), The control method is as follows: Determining the feedforward compensation value (rf) from the measured value (do) obtained from the straightening machine (200) using a prediction model (Mp); Performing machine learning on the prediction model (Mp) using teacher data (Ds), The step of performing the machine learning includes: Obtaining at least one first combination (Cm1) including the target curvature (qr), the operation amount (r) corresponding to the target curvature (qr), the measured value (do), and the correction curvature (q) corresponding to both the measured value (do) and the operation amount (r); Adding to the teacher data (Ds) at least one second combination (Cm2) including the measured value (do) and the operation amount (r) when the absolute value of the error (eq) is smaller than a first reference value (α) among the at least one first combination (Cm1). A control method.

[0108] Each embodiment disclosed this time is also planned to be implemented by being appropriately combined within a non - conflicting range. It should be considered that all the embodiments disclosed this time are illustrative in all respects and not restrictive. The scope of the present invention is shown by the claims rather than the above description, and it is intended that all modifications within the meaning and scope equivalent to the claims are included.

Explanation of Reference Numerals

[0109] 2,3 control system, 11 feedback control device, 12 feedforward compensation device, 13 learning device, 20 field network, 32 upper network, 100, 100A, 300 control device, 110 feedback control means, 120 feedforward compensation means, 122 local bus, 130 learning means, 140 memory means, 150 subtractor, 160 adder, 200 corrector, 302 processor, 304 main memory, 306 upper network controller, 308 field network controller, 312 memory card interface, 314 memory card, 316 local bus controller, 318 processor bus, 360 storage, 370 controller, 380 functional unit, 400 support device, 500 display device, 600 server device, 700 gateway, 900 Internet, Bn bobbin, Cm1~Cm4 combination, Cr ratio, Ct cutting part, Cw cutting wire, Ds teacher data, M2~M6, Me1, Me2 moment, Mp prediction model, Ms2, Ms4, Ms6, Ms8 servo motor, Mu wire movement part, P2~P6, Ps point, Pc control program, Rc regression curve, Ro1~Ro7 roller, SL2~SL6, Sd1~Sd7, Sd10~Sd18 sensor, Scm1~Scm4 data set, Se encoder, Wr wire, Ws winding shaft, do measurement value set, eq error, q correction curvature, qr target curvature, r operation amount, rb feedback operation amount, rb0 average value, rf feedforward compensation value.

Claims

1. A control device for the straightening machine, which determines a feedforward compensation value corresponding to a feedback operation amount to the straightening machine so that the straightening curvature of the wire straightened by the straightening machine approaches a target curvature, wherein the straightening machine comprises a wire supply unit having the wound wire, a plurality of rollers arranged alternately so as to sandwich the wire drawn from the wire supply unit and pressing the wire, and a wire moving unit that draws the wire from the wire supply unit and passes it through the plurality of rollers, wherein the straightening curvature is the curvature of a first portion of the wire that has passed through the plurality of rollers, wherein the feedback operation amount is determined by feedback control means based on an error between the target curvature and the straightening curvature, wherein the sum of the feedback operation amount and the feedforward compensation value is output to the straightening machine as an operation amount, wherein the control device comprises a feedforward compensation means for determining the feedforward compensation value from a measured value obtained from the straightening machine using a prediction model, and a learning means for performing machine learning on the prediction model using teacher data, wherein the learning means acquires at least one first combination including the target curvature, the operation amount corresponding to the target curvature, the measured value, and the straightening curvature corresponding to both the measured value and the operation amount, and adds at least one second combination including the measured value and the operation amount when the absolute value of the error is smaller than a first reference value among the at least one first combination to the teacher data, a control device for a straightening machine.

2. wherein the operation amount changes at least one of positions of the plurality of rollers and a moving speed of the wire, wherein the control device obtains, from the measured value, the moving speed of the wire, the position of the first portion, the position of a second portion of the wire before being pressed by the plurality of rollers, the dimension of the second portion, and the curvature of the second portion, the winding position of the wire in the wire supply unit, and at least one of positions of each of the plurality of rollers, radii of each of the plurality of rollers, pitches of the plurality of rollers, and pressures of the plurality of rollers, the control device for a straightening machine according to claim 1.

3. wherein the learning means calculates an average value of the operation amount when the absolute value of the measured value is smaller than a second reference value in the teacher data, The relationship between the measured value represented by the prediction model, the feedforward compensation value as the difference between the manipulated variable and the average value, with the feedforward compensation value as the target variable and the measured value as the explanatory variable, is approximated as a function. The control device according to claim 1 or 2, wherein when the ratio of the number of the fourth combinations in which the absolute value of the measured value among the at least one second combination is greater than the second reference value to the number of the third combinations in which the absolute value of the measured value among the at least one first combination is greater than the second reference value is greater than the third reference value, the machine learning is terminated.

4. The learning means calculates the average value of the manipulated variable when the absolute value of the measured value in the teacher data is smaller than the second reference value. The relationship between the measured value and the manipulated variable represented by the prediction model is approximated as a function with the manipulated variable as the target variable and the measured value as the explanatory variable. When the ratio of the number of the fourth combinations in which the absolute value of the measured value among the at least one second combination is greater than the second reference value to the number of the third combinations in which the absolute value of the measured value among the at least one first combination is greater than the second reference value is greater than the third reference value, the machine learning is terminated. The feedforward compensation means determines, as the feedforward compensation value, a value obtained by subtracting the average value from the manipulated variable predicted from the measured value by the prediction model. The control device according to claim 1 or 2.

5. The learning means resumes the machine learning when the machine learning is terminated and the ratio is smaller than the third reference value. The control device according to claim 3 or 4.

6. A control system that outputs, as the manipulated variable, the sum of the feedback manipulated variable and the feedforward compensation value to the straightening machine so that the straightening curvature of the wire straightened by the straightening machine approaches the target curvature. The straightening machine includes a wire supply unit having the wound wire, a plurality of rollers arranged alternately so as to sandwich the wire drawn from the wire supply unit and pressing the wire, and a wire moving unit that draws the wire from the wire supply unit and passes it through the plurality of rollers. The straightening curvature is the curvature of the portion of the wire that has passed through the plurality of rollers. The control system A feedback control device that determines the feedback operation amount based on the error between the target curvature and the correction curvature, A feedforward compensation device that determines the feedforward compensation value from the measurement values obtained from the corrector using a prediction model, A learning device that performs machine learning on the prediction model using teacher data, and The learning device, obtains at least one first combination including the target curvature, the operation amount corresponding to the target curvature, the measurement value, and the correction curvature corresponding to both the measurement value and the operation amount, A control system that adds at least one second combination including the measurement value and the operation amount when the absolute value of the error is smaller than a first reference value among the at least one first combination to the teacher data. **Claim 7** A control method for the corrector that determines a feedforward compensation value corresponding to a feedback operation amount to the corrector so that the correction curvature of the wire corrected by the corrector approaches the target curvature, The corrector includes a wire supply unit having the wound wire, a plurality of rollers arranged alternately so as to sandwich the wire drawn from the wire supply unit and pressing the wire, and a wire moving unit that draws the wire from the wire supply unit and passes it through the plurality of rollers, The correction curvature is the curvature of a first portion of the wire that has passed through the plurality of rollers, The feedback operation amount is determined by a feedback control means based on the error between the target curvature and the correction curvature, The sum of the feedback operation amount and the feedforward compensation value is output to the corrector as an operation amount, The control method includes a step of determining the feedforward compensation value from the measurement values obtained from the corrector using a prediction model, and a step of performing machine learning on the prediction model using teacher data, The step of performing the machine learning includes a step of obtaining at least one first combination including the target curvature, the operation amount corresponding to the target curvature, the measurement value, and the correction curvature corresponding to both the measurement value and the operation amount, and a step of adding at least one second combination including the measurement value and the operation amount when the absolute value of the error is smaller than a first reference value among the at least one first combination to the teacher data.

Images