Die buffer control device, die buffer control method and die buffer control program

Abstract

A die cushion control device (200A) for controlling a die cushion mechanism (3) comprises: a pressure command generating unit (50) which outputs a first pressure command (P1a) of a pressure or force generated between the die cushion mechanism and a slider (1); a deviation predicting unit (70) which obtains the generated pressure or force as a detection pressure (P3a), and predicts a pressure deviation which is a difference between the pressure or force generated by the first pressure command when the die cushion mechanism is controlled according to the first pressure command and the detection pressure based on the translation acceleration of the slider, the control parameter of the die cushion mechanism and the die cushion movement amount per one rotation of the servo motor, and outputs the predicted pressure deviation as a correction pressure command (71); a pressure command correcting unit (80A) which corrects the first pressure command by the correction pressure command, thereby calculating a second pressure command (P2a); and a pressure control unit (30) which calculates a speed command for causing the detection pressure to follow the second pressure command and outputs the predicted pressure deviation to a speed control unit (23).

Description

Technical Field

[0001] The invention relates to a die buffer control device for controlling a die buffer mechanism, a die buffer control method and a die buffer control program. Background Art

[0002] As one of the machines for performing stamping processes such as bending, drawing, and punching, there is a stamping machine having a die buffer mechanism. The die buffer mechanism applies additional pressure from a support member that supports the movable side of one die, i.e., a slider, from a support member that supports the other die, i.e., a buffer pad. Thus, the die buffer mechanism can prevent defects such as wrinkles from occurring in stamped products.

[0003] The die cushion mechanism called servo die cushion uses a servo motor as a driving source, and can arbitrarily change the applied pressure during one molding process. By using this servo die cushion, the press machine can improve the moldability, quality stability and yield rate.

[0004] In the servo die buffer, the pressure during the stamping operation is detected, and the servo motor is controlled so that the pressure follows the predetermined pressure command value. In the servo die buffer, even if the pressure is controlled, the actual pressure may drop relative to the desired pressure in the final stage of the pressurization operation. In this case, the pressure drop and the insufficient additional pressure become the cause of wrinkles in the stamped product.

[0005] In order to eliminate the pressure drop phenomenon described above, the control device of Patent Document 1 obtains the acceleration of the slide, and corrects the speed command value and the current command value to the die cushion mechanism based on a signal obtained by multiplying the acceleration by a constant.

[0006] Patent Document 1: Japanese Patent Application Laid-Open No. 2007-905 Summary of the invention

[0007] However, in the technology of the above-mentioned Patent Document 1, when the constant multiplied by the acceleration is greater than an appropriate value, overcompensation occurs, that is, the pressure becomes larger than the target value of the pressure command value. In addition, when the constant is less than an appropriate value, the pressure does not reach the target value of the pressure command value, and the pressure drop cannot be fully compensated. Therefore, in the technology of the above-mentioned Patent Document 1, in order to compensate the pressure to the level of the pressure command value, it is necessary to determine the constant by trial and error, and there is a problem that a lot of work is required to compensate for the pressure drop.

[0008] The present invention has been made in view of the above-mentioned circumstances, and an object of the present invention is to obtain a die cushion control device that can easily compensate for the pressure drop.

[0009] In order to solve the above-mentioned problems and achieve the purpose, the present invention uses a servo motor as a driving source to control a die cushion mechanism that generates pressure or force on a slide of a stamping machine, and the die cushion control device has a pressure command generating unit that outputs a command of pressure or force generated between the die cushion mechanism and the slide, that is, a first pressure command. In addition, the die cushion control device has a deviation predicting unit that obtains information on the pressure or force generated between the die cushion mechanism and the slide as a detection pressure, and predicts the difference between the pressure or force generated by the first pressure command and the detection pressure when the die cushion mechanism is controlled according to the first pressure command, that is, a pressure deviation, based on the translation acceleration of the slide, the control parameters used when controlling the pressure or force of the die cushion mechanism, and the die cushion movement amount of one rotation of the servo motor, and outputs it as a correction pressure command. In addition, the mold buffer control device has: a pressure command correction unit, which corrects the first pressure command by correcting the pressure command, thereby calculating the second pressure command; and a pressure control unit, which calculates a speed command for causing the detected pressure to follow the second pressure command, and outputs the speed command to the speed control unit, which outputs a driving current corresponding to the speed command to the servo motor.

[0010] Effects of the Invention

[0011] The die cushion control device according to the present invention has an effect of being able to easily compensate for a pressure drop. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is a diagram showing the configuration of a processing system including the die cushion control device according to the first embodiment.

[0013] Figure 2 It is a diagram showing a configuration of a pressure control unit included in the die cushion control device according to the first embodiment.

[0014] Figure 3 This is a flowchart showing a control process procedure of the die cushion mechanism by the die cushion control device according to the first embodiment.

[0015] Figure 4 It is a diagram for explaining a pressure waveform when the die cushion mechanism is controlled by the die cushion control device of the comparative example.

[0016] Figure 5 This is a diagram for explaining a pressure waveform when the die cushion mechanism is controlled by the die cushion control device according to the first embodiment.

[0017] Figure 6This is a diagram for explaining the transfer characteristics during pressure control by the die cushion control device according to the first embodiment.

[0018] Figure 7 It is a diagram showing another configuration example of the die cushion mechanism included in the machining system according to the first embodiment.

[0019] Figure 8 It is a diagram showing the configuration of a processing system including a die cushion control device according to the second embodiment.

[0020] Fig. 9 This is a flowchart showing a control processing procedure of the die cushion mechanism performed by the die cushion control device according to the second embodiment.

[0021] Fig.10 This is a diagram for explaining a pressure waveform when the die cushion mechanism is controlled by the die cushion control device according to the second embodiment.

[0022] Fig.11 This is a diagram showing the configuration of a learning device according to the third embodiment.

[0023] Fig.12 This is a flowchart showing the processing procedure of the learning process performed by the learning device involved in the third embodiment.

[0024] Fig.13 This is a diagram showing the structure of a neural network used by a learning device according to a third embodiment.

[0025] Fig.14 This is a diagram showing the structure of an inference device according to the third embodiment.

[0026] Fig.15 This is a flowchart showing the processing procedure of the estimation process performed by the estimation device involved in the third embodiment.

[0027] Fig.16 This is a diagram showing a hardware configuration example for realizing the die cushion control device according to the first embodiment. DETAILED DESCRIPTION

[0028] Hereinafter, a die cushion control device, a die cushion control method, and a die cushion control program according to embodiments of the present invention will be described in detail based on the drawings.

[0029] Implementation method 1.

[0030] Figure 1101A is a diagram showing the structure of a processing system having a die cushion control device according to Embodiment 1. The processing system 101A is a system that performs press processing on a workpiece while changing the force applied in the middle of one molding process by using a servo die cushion. In addition, the following description will be made on the case where the applied force is pressure.

[0031] The processing system 101A includes: a die cushion mechanism 3; a die cushion control device 200A that controls the die cushion mechanism 3; a slide 1; a slide control unit 2; a servo motor 10; a speed control unit 23; and a mechanical mechanism 90. The mechanical mechanism 90, the servo motor 10, the die cushion mechanism 3, and the slide 1 in the processing system 101A are structural elements of a stamping machine.

[0032] In addition, the processing system 101A has two molds (not shown). The slider 1 is a pair of molds (in Figure 1 A slider driving motor (not shown) is mounted on the slider 1, and the rotational motion of the slider driving motor is converted into up and down motion via a mechanical mechanism 90 such as a crank mechanism.

[0033] The die buffer mechanism 3 uses a servo motor 10 as a driving source, and generates force on the slider 1 of the stamping machine through the buffer pad 4 and the workpiece (not shown). The die buffer mechanism 3 has a buffer pad 4, a hydraulic cylinder 5, a pipe 6, a hydraulic pump 7 and a pressure detector 8 as a pressure detection unit. The buffer pad 4 is a supporting member that supports the other of the two dies. In the stamping machine, the slider 1 is pressed by the workpiece from the upper side of the workpiece, and the buffer pad 4 applies additional pressure to the workpiece from the lower side of the workpiece. In addition, in the stamping machine, the slider 1 can also be pressed by the workpiece from the lower side of the workpiece. In this case, the buffer pad 4 applies additional pressure to the workpiece from the upper side of the workpiece. The workpiece is also called a stamping workpiece, a stamping workpiece, etc., a processed object, that is, a stamping object. The workpiece is processed by the stamping machine, thereby becoming a stamped product.

[0034] The cushion pad 4 moves in accordance with the movement of the slider 1. The cushion pad 4 is controlled so that a specific pressure is generated on the workpiece when the slider 1 descends and contacts the cushion pad 4 via the workpiece. The cushion pad 4 is controlled based on the pressure value detected by the pressure detector 8 (hereinafter referred to as the detection pressure P3a).

[0035] The hydraulic cylinder 5 drives the cushion pad 4 in the up-down direction. The hydraulic pump 7 is a double-rotation type rotary pump. The hydraulic pump 7 is connected to the hydraulic cylinder 5 via two pipes 6. The hydraulic pump 7 supplies working oil to the hydraulic cylinder 5 via the pipe 6. The pressure detector 8 is provided in the pipe 6 to detect the pressure in the pipe 6. The pressure detector 8 sends the detected pressure value, i.e., the detected pressure P3a, to the die cushion control device 200A.

[0036] In the first embodiment, the case where the pressure detector 8 is provided in the pipe 6 is described, but the pressure detector 8 may be any device as long as it can detect the pressure generated between the slider 1 and the die cushion mechanism 3 .

[0037] The servo motor 10 drives the hydraulic pump 7. The servo motor 10 supplies torque for driving the hydraulic pump 7 to the hydraulic pump 7. The servo motor 10 is controlled by the die cushion control device 200A. The mechanical mechanism 90 is, for example, a link mechanism that converts rotational motion into linear motion. An example of the link mechanism is a crank mechanism.

[0038] The slider control unit 2 controls the slider 1 by controlling the slider driving motor. The slider control unit 2 controls the movement amount of the slider 1, the speed of the slider 1, that is, the slider speed, etc. The slider control unit 2 performs a stamping process by moving the slider 1 up and down. The slider control unit 2 sends state information indicating the state of the slider 1 to the slider acceleration calculation unit 60.

[0039] The die cushion control device 200A controls the cushion pad 4 by controlling the servo motor 10 based on the detection pressure P3a detected by the pressure detector 8. The die cushion control device 200A includes a pressure command generation unit 50, a slide acceleration calculation unit 60, a deviation prediction unit 70, a pressure command correction unit 80A, and a pressure control unit 30.

[0040] The slider acceleration calculation unit 60 calculates the information of the translation acceleration of the slider 1, that is, the slider acceleration, based on the state information from the slider control unit 2. The state information is, for example, the rotation position of the slider driving motor during the stamping process. In this case, the slider acceleration calculation unit 60 obtains the rotation position of the slider driving motor during the stamping process from the slider control unit 2 as the state information. The slider acceleration calculation unit 60 calculates the translation position of the slider 1 using the rotation position, the information of the link mechanism, the mechanical specifications, etc., and calculates the slider acceleration represented by the translation acceleration signal by performing a second-order differential on the translation position.

[0041] In addition, another example of state information is a translation position command generated by the slider control unit 2 when the slider 1 moves. In this case, the slider acceleration calculation unit 60 obtains the translation position command from the slider control unit 2 as state information. The slider acceleration calculation unit 60 calculates the slider acceleration represented by the translation acceleration signal by performing a second-order differential on the translation position corresponding to the translation position command. The slider acceleration calculation unit 60 sends the calculated slider acceleration to the deviation prediction unit 70.

[0042] The deviation prediction unit 70 calculates the pressure deviation, which is the steady-state pressure drop amount generated when the pressure to the cushion pad 4 is controlled, based on the slide acceleration, the control parameters used by the pressure control unit 30, and the die cushion movement amount per one rotation of the servo motor 10. The pressure deviation is the pressure difference between the pressure specified by the first pressure command P1a and the pressure specified by the detection pressure P3a when the pressure control unit 30 controls the pressure of the die cushion mechanism 3 using the first pressure command P1a described later.

[0043] When the pressure control unit 30 performs PI (Proportional Integral) control, the deviation prediction unit 70 uses the control parameters of the PI control to predict the pressure deviation according to the following equation (1). PD in equation (1) is the pressure deviation, and A sl is the slider acceleration. In addition, K p is the proportional gain of pressure control (control parameter of pressure control), K i is the integral gain of the pressure control (control parameter of the pressure control). Also, C is the translational movement amount of the die cushion mechanism 3 per one rotation of the servo motor 10 (hereinafter referred to as the die cushion movement amount).

[0044] [Formula 1]

[0045] PD=(1 / K p / K i / C)·A sl …(1)

[0046] As shown in equation (1), the deviation prediction unit 70 converts the slider acceleration A sl Divide by the proportional gain K p , integral gain K i , and the mold cushion movement amount C per rotation of the servo motor 10, thereby predicting the pressure deviation.

[0047] In addition, if Figure 1As shown, when the die cushion mechanism 3 includes the hydraulic cylinder 5 and the hydraulic pump 7, it can be expressed as C = the ejection volume of the working oil per one rotation of the hydraulic pump 7 / the pressure receiving cross-sectional area of ​​the hydraulic cylinder 5. In order to reduce the calculated pressure deviation, the deviation prediction unit 70 sends a correction pressure instruction 71, which is an instruction for correcting the first pressure instruction P1a, to the pressure instruction correction unit 80A. The correction pressure instruction 71 is an instruction including information on the correction pressure corresponding to the pressure deviation.

[0048] The pressure command generation unit 50 generates the desired pressure distribution to be generated by the die cushion mechanism 3 during the stamping process. The pressure distribution is information indicating the magnitude and time of the pressure applied by the cushion pad 4 to the workpiece. In the stamping process, for each workpiece, it is determined what size of pressure is applied to the workpiece and for how long. Therefore, the user of the stamping machine determines the pressure distribution desired by the user for each workpiece by setting the stamping pressure and the stamping time. The pressure command generation unit 50 generates a pressure command corresponding to the pressure distribution (hereinafter referred to as the first pressure command P1a) and sends it to the pressure command correction unit 80A.

[0049] The pressure command correction unit 80A includes an adder 85, which generates a pressure command (hereinafter referred to as a second pressure command P2a) by summing the first pressure command P1a and the correction pressure (pressure deviation) included in the correction pressure command 71. The pressure command correction unit 80A applies the second pressure command P2a to the pressure command to the pressure control unit 30, and sends it to the pressure control unit 30. As the pressure command correction unit 80A, for example, after the first pressure command P1a rises and the first pressure command P1a becomes a constant value, the second pressure command P2a is applied.

[0050] The pressure control unit 30 calculates a command (hereinafter referred to as a motor speed command 24) used when controlling the speed of the servo motor 10 based on the second pressure command P2a and the detected pressure P3a. The pressure control unit 30 calculates the motor speed command 24 corresponding to the speed generated by the servo motor 10 in such a manner that the detected pressure P3a follows the second pressure command P2a. Here, a specific configuration example of the pressure control unit 30 is described.

[0051] Figure 2 1 is a diagram showing a configuration of a pressure control unit included in the die cushion control device according to Embodiment 1. The pressure control unit 30 includes a multiplier 41 that multiplies a proportional gain K p ; Multiplier 42, which multiplies the integral gain K i ; integrator 43; adder 44; and subtractor 45.

[0052] The pressure control unit 30 receives the second pressure command P2a from the pressure command correction unit 80A, and receives the detection pressure P3a from the pressure detector 8. The pressure control unit 30 calculates the deviation between the pressure indicated by the second pressure command P2a and the pressure indicated by the detection pressure P3a by subtracting the detection pressure P3a from the second pressure command P2a, and performs proportional control processing and integral control processing on the deviation, thereby calculating the motor speed command 24.

[0053] Specifically, the subtractor 45 subtracts the pressure represented by the detection pressure P3a from the pressure represented by the second pressure command P2a, thereby calculating the deviation between the pressure represented by the second pressure command P2a and the pressure represented by the detection pressure P3a. The integrator 43 integrates the deviation calculated by the subtractor 45. The "s" shown as the integrator 43 represents a Laplace operator, which means that the integration process is performed by 1 / s.

[0054] The multiplier 42 multiplies the integrated deviation by an integral gain K as a control parameter. i The adder 44 adds the multiplication result obtained by the multiplier 42 and the subtraction result obtained by the subtractor 45. The multiplier 41 multiplies the addition result obtained by the adder 44 by the proportional gain K as a control parameter. p , and output the multiplication result as the motor speed command 24.

[0055] The speed control unit 23 supplies the drive current 25 corresponding to the motor speed command 24 to the servo motor 10 . That is, the speed control unit 23 supplies the drive current 25 to the servo motor 10 so that the speed of the servo motor 10 follows the speed indicated by the motor speed command 24 .

[0056] In addition, Figure 1 Although not shown in the figure, an encoder for detecting the rotation speed of the servo motor 10 is mounted on the servo motor 10. The speed control unit 23 can calculate the drive current 25 by performing feedback control so that the rotation speed detected by the encoder follows the motor speed command 24.

[0057] Next, a control processing procedure of the die cushion mechanism 3 by the die cushion control device 200A will be described. Figure 3 This is a flowchart showing a control process procedure of the die cushion mechanism by the die cushion control device according to the first embodiment.

[0058] In the die cushion control device 200A, the pressure command generation unit 50 generates the first pressure command P1a corresponding to the pressure distribution generated in the die cushion mechanism 3 during the press working (step S10), and sends it to the pressure command correction unit 80A.

[0059] The slider acceleration calculation unit 60 obtains the slider acceleration (step S20 ). Specifically, the slider acceleration calculation unit 60 calculates the slider acceleration based on the state information of the slider 1 sent from the slider control unit 2 .

[0060] The deviation prediction unit 70 calculates the pressure deviation, i.e., PD, based on the equation (1) (step S30). Specifically, the deviation prediction unit 70 calculates the steady-state pressure drop, i.e., the pressure deviation, generated during the pressure control of the cushion pad 4 based on the slider acceleration, the control parameters used by the pressure control unit 30, and the mold cushion movement amount per one rotation of the servo motor 10. The deviation prediction unit 70 calculates the correction pressure for reducing the calculated pressure deviation, and sends the correction pressure command 71 including the calculated correction pressure to the pressure command correction unit 80A.

[0061] The pressure command correction unit 80A corrects the first pressure command P1a using the correction pressure included in the correction pressure command 71, and calculates the second pressure command P2a (step S40). The corrected first pressure command P1a is the second pressure command P2a. The pressure command correction unit 80A sends the second pressure command P2a to the pressure control unit 30.

[0062] The pressure control unit 30 obtains the detection pressure P3a from the pressure detector 8. In addition, the process of the pressure control unit 30 obtaining the detection pressure P3a from the pressure detector 8 is always executed. The pressure control unit 30 performs pressure control in a manner that the detection pressure P3a follows the second pressure instruction P2a (step S50). Specifically, the pressure control unit 30 calculates the motor speed instruction 24 based on the second pressure instruction P2a and the detection pressure P3a. The speed control unit 23 supplies the drive current 25 to the servo motor 10 in a manner that the speed of the servo motor 10 follows the motor speed instruction 24. Thus, the pressure control unit 30 and the speed control unit 23 control the servo motor 10 connected to the mold cushion mechanism 3 using the second pressure instruction P2a and the detection pressure P3a.

[0063] The die cushion control device 200A determines whether the pressure control is completed (step S60). If the pressure control is not completed (step S60, No), the die cushion control device 200A returns to the process of step S20 and repeats the processes of steps S20 to S60. During the period when the pressure command generation unit 50 generates the first pressure command P1a and sends it to the pressure command correction unit 80A, the pressure control is not completed. If the pressure control is completed (step S60, Yes), the pressure command generation unit 50 ends the generation of the first pressure command P1a. As a result, the die cushion control device 200A ends the operation related to the pressure control.

[0064] Here, the effect obtained by the die cushion control device 200A controlling the die cushion mechanism 3 is described using Figure 4 and Figure 5 Provide explanation. Figure 4 It is a diagram for explaining a pressure waveform when the die cushion mechanism is controlled by the die cushion control device of the comparative example. Figure 4 The horizontal axis of the graph shown is time, and the vertical axis is pressure.

[0065] The die cushion control device of the comparative example is a device that directly applies the first pressure command P1a as a pressure command for pressure control to the pressure control unit. Figure 4 2 shows a waveform of a first pressure command P1a indicated by a solid line and a waveform of a detected pressure P3a indicated by a dotted line.

[0066] It is desired that the pressure distribution generated in the die cushion mechanism 3 is directly applied as the first pressure command P1a. If it is applied to the pressure control of the die cushion mechanism 3, the detected pressure P3a does not follow the first pressure command P1a. Figure 4 As shown in FIG. 1 , a waveform that decreases with respect to the first pressure command P1a is generated.

[0067] Figure 5 This is a diagram for explaining a pressure waveform when the die cushion mechanism is controlled by the die cushion control device according to the first embodiment. Figure 5 The horizontal axis of the graph shown is time, and the vertical axis is pressure. Figure 5 2 shows a waveform of a first pressure command P1a indicated by a solid line, a waveform of a detected pressure P3a indicated by a dotted line, and a waveform of a second pressure command P2a indicated by a dashed line.

[0068] The die cushion control device 200A quantitatively predicts the pressure drop, estimates the pressure drop amount, and corrects the first pressure command P1a by the correction pressure command 71 corresponding to the pressure drop amount (pressure drop prediction value), thereby generating the second pressure command P2a used for pressure control. Thus, the die cushion control device 200A can obtain a waveform in which the detection pressure P3a does not drop relative to the first pressure command P1a, which is the pressure distribution desired to be generated in the die cushion mechanism 3.

[0069] In addition, the die cushion control device 200A sets the pressure distribution desired to be applied during the stamping process as the first pressure command P1a. Thus, the die cushion control device 200A can prevent overcompensation in which the detection pressure P3a becomes larger than the first pressure command P1a, or undercompensation in which the detection pressure P3a does not reach the first pressure command P1a. Therefore, the die cushion control device 200A can control the pressure of the die cushion mechanism 3 in a stable and automatic manner as the desired pressure distribution.

[0070] Furthermore, when the die cushion control device 200A generates the second pressure command P2a, the user can obtain the correction pressure command 71 as a correction signal having an appropriate level that is neither too high nor too low, without performing operations such as adjusting a certain coefficient.

[0071] Furthermore, when the die cushion control device 200A calculates the correction pressure command 71 including the correction pressure for correcting the first pressure command P1a using the equation (1) as a prediction equation (correction equation), the equation (1) does not depend on the specification values ​​related to the workpiece and the die. Therefore, every time the workpiece or the die is changed, it is not necessary to input these specification values ​​to the die cushion control device 200A and change the equation (1), which can reduce the workload of the user.

[0072] Here, the reason why the detection pressure P3 a decreases with respect to the first pressure command P1 a when the slider 1 contacts the cushion pad 4 via the workpiece will be described.

[0073] Figure 6 1 is a diagram for explaining the transfer characteristics of the pressure control performed by the die cushion control device according to the first embodiment. Figure 6 Among the structural elements of Figure 2 Structural elements having the same functions as those shown in the drawings are denoted by the same reference numerals, and repeated descriptions are omitted.

[0074] The pressure control unit 30 performs PI control operation on the difference between the second pressure command P2a and the detected pressure P3a so that the detected pressure P3a follows the second pressure command P2a, thereby calculating and outputting the motor speed command 24.

[0075] Figure 6 The transfer characteristic 51 shown in FIG. 2 shows the transfer characteristic from the motor speed command 24 to the speed of the servo motor 10, that is, the motor speed 52. The transfer characteristic 51 is equivalent to the transfer characteristic that depends on Figure 1 The transfer characteristic is determined by the characteristics of the speed control unit 23 and the servo motor 10. Here, the control band of the pressure control is sufficiently smaller than the control band of the speed control, and the transfer characteristic of the speed control is approximately regarded as 1 to explain the pressure drop.

[0076] In the die cushion control, the motor speed 52 of the servo motor 10 is not determined solely by the speed command generated by the pressure control unit 30. After the slider 1 contacts the die cushion mechanism 3 through the workpiece, the servo motor 10 is also forcibly rotated by the external force of the slider 1. If this action is viewed from the perspective of the pressure control that controls the servo motor 10 driving the die cushion mechanism 3, it can be considered that the slider 1 is subject to speed disturbance according to the action. The situation that expresses this action from the perspective of transfer characteristics is Figure 6 Therefore, the motor speed 54 of the servo motor 10 is considered to be determined by the sum of the motor speed 52 based on the motor speed command 24 generated by the pressure control unit 30 and the disturbance speed 53 caused by the external force of the slider 1 .

[0077] Figure 6 The transfer characteristic 55 shown is a transfer characteristic from the motor speed 54 to the motor position 56. The transfer characteristic 55 can be expressed by 1 / s as an integral characteristic. Figure 6 The transfer characteristic 57 shown is a transfer characteristic from the motor position 56 to the detection pressure P3a. These transfer characteristics 55 and 57 are equivalent to those that depend on Figure 1 The transfer characteristics are determined by the characteristics of the servo motor 10 and the die cushion mechanism 3.

[0078] In the die cushion mechanism 3 , pressure is generated in proportion to the motor position 56 . K of the transmission characteristic 57 represents an elastic constant as a proportional constant. K of the transmission characteristic 57 is a constant that depends on the compressibility of the die, the workpiece, or the hydraulic oil used in the hydraulic cylinder 5 .

[0079] The detection pressure P3a becomes a signal that depends on the transfer characteristics 51, 55, 57, the motor speed command 24, and the disturbance speed 53. Then, the detected detection pressure P3a is sent to the pressure control unit 30.

[0080] When the press forming operation is performed by the slider 1 and the die cushion mechanism 3, the die cushion mechanism 3 operates in a manner of applying pressure to the slider 1. Therefore, after the slider 1 descends and the die cushion mechanism 3 contacts the slider 1 through the workpiece, the translation speed of the slider 1 and the translation speed of the die cushion mechanism 3 are substantially the same.

[0081] In this case, the translation speed of the die cushion mechanism 3 and the motor rotation speed of the servo motor 10 driving the die cushion mechanism 3 satisfy the relationship of translation speed of the die cushion mechanism 3 = C × (rotation speed of the servo motor 10 driving the die cushion mechanism 3).

[0082] Here, C is a constant, which is the amount of mold buffer movement per one rotation of the servo motor 10, as described above. If the specifications of the mold buffer mechanism 3 are determined, the constant C is the only constant that is determined. If the relationship between the translation speed of the mold buffer mechanism 3 and the rotation speed of the servo motor 10 is used, after the mold buffer mechanism 3 contacts the slider 1 via the workpiece, the relationship between the motor speed 54 of the servo motor 10 driving the mold buffer mechanism 3 and the translation speed of the slider 1 is established as follows: motor speed = (1 / C) × slider translation speed. Therefore, if the translation speed of the slider 1 is set to the translation speed V sl , then the interference speed V d It can be expressed by the following formula (2).

[0083] [Formula 2]

[0084] V d =1 / C×V sl …(2)

[0085] If the first pressure command P1a is set to P cmd , then according to Figure 6 As shown in the block diagram, if the V d The transmission characteristics of the disturbance speed 53 represented by φ and the second pressure command P2a to the detection pressure P3a represented by P give the following relationship of equation (3).

[0086] [Formula 3]

[0087]

[0088] Furthermore, if the above equation (2) is applied to equation (3), the following relationship of equation (4) is obtained.

[0089] [Formula 4]

[0090]

[0091] In formula (4), s·V slConsidered as the translation speed V of slider 1 sl The signal after the differentiation operation is thus the acceleration of the slider 1. That is, by using the acceleration of the slider 1, that is, the slider acceleration A sl , thus according to formula (4) we get the following relationship of formula (5).

[0092] [Formula 5]

[0093]

[0094] In order to consider the cause of the pressure drop, consider the pressure command P in the transfer characteristic expressed by equation (5): cmd The behavior of the steady-state time response after a period of time has passed since the value became constant. First, the steady-state time response of the transfer function that appears in the first term of equation (5) is obtained. The steady-state value that appears in the first term can be calculated by substituting s=0 into the first term of equation (5), and becomes P cmd Similarly, the steady-state time response of the transfer function that appears in the second term of equation (5) is obtained. Substituting s=0 into the second term of equation (5) yields steady-state gain=1 / (C·K p ·K i )·A sl Therefore, in a steady state, the detection pressure P can be expressed by the following equation (6).

[0095] [Formula 6]

[0096] P=P cmd +1 / (C·K p ·K i )·A sl …(6)

[0097] As shown in formula (6), the detected pressure P is not equal to the pressure command P cmd Consistent, resulting in the slider acceleration A sl Multiply by 1 / (C·K p ·K i ) and the deviation corresponding to the value obtained. In addition, when the slider 1 performs the punching action, the slider 1 decelerates and stops at the bottom dead point. Therefore, the slider acceleration A sl decreases, as the slider acceleration A sl Therefore, in equation (6), the detected pressure P becomes the relative pressure command P cmd The smaller the value, the detected pressure P is relative to the pressure command P cmd decline.

[0098] The deviation prediction unit 70 calculates the correction pressure command 71, which is a correction amount corresponding to the pressure drop, based on the pressure control characteristics of the die cushion mechanism 3 described above. Furthermore, the pressure command correction unit 80A estimates the pressure drop during pressure control, and generates the second pressure command P2a by correcting the first pressure command P1a, which is the desired pressure command, using the correction pressure command 71. The pressure command correction unit 80A applies the generated second pressure command P2a to the pressure command for pressure control. As a result, the die cushion control device 200A can obtain the pressure as the desired pressure command (first pressure command P1a) in a stable state.

[0099] In the first embodiment, an example is described in which the servo motor 10 rotates the hydraulic pump 7, which is a two rotary pump, thereby operating the hydraulic cylinder 5 to control the die cushion mechanism 3, but the processing system 101A is not limited to the case where the die cushion mechanism 3 is controlled by the hydraulic cylinder 5. For example, when the die cushion mechanism 3 is driven by a ball screw, and the ball screw is connected to the rotational motion of the servo motor 10 via various speed reducers, pulleys, timing belts, etc., and the rotational motion of the servo motor 10 is converted into the translational motion of the die cushion mechanism 3, the die cushion control device 200A can also be applied.

[0100] Figure 7 This is a diagram showing another configuration example of the die cushion mechanism included in the machining system according to Embodiment 1. The die cushion mechanism 3A of another configuration example includes pulleys 11A and 11B, a timing belt 12 , a speed reducer 13 , a ball screw 14 , a cushion pad 4 , and a pressure detector 8 .

[0101] For example, in the die cushion mechanism 3A, the cushion pad 4 is driven by a ball screw 14, and the ball screw 14 is connected to the servo motor 10 via a speed reducer 13, a pulley 11B, a timing belt 12, a pulley 11A, etc. Thus, the rotational motion of the servo motor 10 is converted into the translational motion of the die cushion mechanism 3A. The die cushion control device 200A can also be applied to the die cushion mechanism 3A as described above.

[0102] In this case, C, which represents the amount of movement of the mold buffer per rotation of the servo motor 10, is calculated based on the pitch of the ball screw (the amount of movement per rotation of the ball screw), the reduction ratio of the reducer, the pulley ratio of the timing belt, etc., and applied to formula (1). Specifically, when a ball screw, a reducer and a timing belt are used, C = pitch of the ball screw / reduction ratio / pulley ratio. In addition, when a ball screw and a reducer are used, C = pitch of the ball screw / reduction ratio. In addition, when a ball screw and a timing belt are used, C = pitch of the ball screw / pulley ratio. As a result, the mold buffer control device 200A can obtain the same effect as when the mold buffer mechanism 3 is controlled by the hydraulic cylinder 5.

[0103] In the first embodiment, the pressure control unit 30 is described as being composed of components that perform PI control, and the proportional gain K is used as a control parameter. p And integral gain K i Although the steady-state deviation of the pressure control is calculated as in equation (1), the control performed by the pressure control unit 30 is not limited to PI control. The pressure control unit 30 may be composed of a component that performs PID (Proportional Integral Differential) control, or a component that performs a control that combines phase delay compensation, phase advance compensation, etc. In these cases, if the deviation prediction unit 70 predicts the pressure deviation using the control parameter of the pressure control, it can also be applied to the first embodiment.

[0104] As described above, in the first embodiment, the deviation prediction unit 70 calculates the steady-state pressure drop amount generated when the pressure on the cushion pad 4 is controlled based on the control parameters used by the pressure control unit 30 and the mold cushion movement amount per one rotation of the servo motor 10, that is, the pressure deviation. In addition, the pressure command correction unit 80A generates the second pressure command P2a by summing the first pressure command P1a and the correction pressure (pressure deviation) included in the correction pressure command 71. Moreover, the pressure control unit 30 calculates the motor speed command 24 used when controlling the speed of the servo motor 10 based on the second pressure command P2a and the detected pressure P3a, and causes the speed control unit 23 to control the mold cushion mechanism 3. As a result, the mold cushion control device 200A can generate the second pressure command P2a corresponding to the pressure drop in the mold cushion mechanism 3 during the stamping process, so that the pressure drop during the stamping process can be easily compensated.

[0105] Furthermore, according to the first embodiment, the die cushion control device 200A does not rely on Figure 6The correction pressure command 71 can be calculated by using K, that is, the elastic constant, of the transfer characteristic 57 described above. Therefore, even when the die, the workpiece, or the hydraulic oil used in the hydraulic cylinder 5 is changed, the die cushion control device 200A can be applied without reflecting these specifications.

[0106] In the first embodiment and the second and third embodiments described later, an example of detecting pressure and performing control based on the detected pressure value is described, but force may be detected instead of pressure and control may be performed based on force. That is, the die cushion control device 200A and the die cushion control device 200B described later may detect the force generated between the die cushion mechanism 3, 3A and the slider 1 instead of pressure and control the die cushion mechanism 3, 3A based on the detected force. In this case, the die cushion control devices 200A and 200B can be applied to the control of the die cushion mechanism 3 in substantially the same manner. As described above, the pressure in the first to third embodiments refers to pressure or force.

[0107] Implementation method 2.

[0108] Next, use Figures 8 to 10 Embodiment 2 is described. In the above-mentioned embodiment 1, the die cushion control device 200A quantitatively predicts the steady-state pressure deviation of the pressure control generated by the contact between the slider 1 and the die cushion mechanism 3 via the workpiece, and corrects the first pressure command P1a by the predicted pressure deviation amount, that is, the estimated pressure drop amount. As a result, the die cushion control device 200A eliminates the pressure drop in a steady state without overcompensation or undercompensation.

[0109] When the mold cushion control device 200A performs pressure control on the mold cushion mechanism 3, by performing molding conditions such as a high slider speed, a pressure overshoot of the detection pressure P3a sometimes occurs when the first pressure instruction P1a rises. If the pressure overshoot is of an appropriate size, it will not affect the stamping process, but if an excessive pressure overshoot occurs, it may sometimes cause adverse effects such as damage to the stamped product. For situations where the pressure overshoot cannot be ignored, as described in the first embodiment, the pressure drop is estimated, and if the correction is performed in a manner that makes the first pressure instruction P1a greater than the pressure distribution, the pressure overshoot will further increase. In the second embodiment, the pressure overshoot as described above is prevented from increasing, and the pressure control of the mold cushion mechanism 3 is performed.

[0110] Figure 8 FIG. 2 is a diagram showing a structure of a processing system having a die cushion control device according to Embodiment 2. Figure 8 Among the structural elements of Figure 1In the die cushion control device 200A of the illustrated embodiment 1, components having the same functions are denoted by the same reference numerals, and duplicate descriptions thereof are omitted.

[0111] The machining system 101B of the second embodiment includes a die cushion control device 200B instead of the die cushion control device 200A, as compared with the machining system 101A. In the die cushion control device 200B of the second embodiment, the pressure command generated by the pressure command generation unit 50 is set as the first pressure command P1b, and the pressure command generated by the pressure command correction unit 80B is set as the second pressure command P2b. In addition, the pressure value detected by the pressure detection meter 8 is set as the detection pressure P3b.

[0112] Compared with the die cushion control device 200A, the die cushion control device 200B includes a pressure command correction unit 80B instead of the pressure command correction unit 80A. Compared with the pressure command correction unit 80A, the pressure command correction unit 80B includes a switch unit 81 and a timing determination unit 83.

[0113] The switch unit 81 selects and switches the correction value used for pressure correction. The switch unit 81 selects "0" or the correction pressure command 71 sent from the deviation prediction unit 70 as the correction pressure command used for pressure correction.

[0114] The switch unit 81 has a selection switch. When the selection switch selects “0”, the adder 85 is connected to the A side, and when the connection with the deviation prediction unit 70 is selected, the adder 85 is connected to the B side. The switch unit 81 inputs the selected correction pressure command to the adder 85 .

[0115] In the switch unit 81, when the selection switch is on the A side, the switch unit 81 sets the correction pressure command to "0", and when the selection switch is on the B side, the correction pressure command 71 sent from the deviation prediction unit 70 is set to the correction pressure command. That is, when the selection switch is on the A side, the first pressure command P1b becomes the second pressure command P2b directly applied to the pressure control unit 30.

[0116] As described above, the switch unit 81 switches between the first process of directly outputting the first pressure command P1b as the second pressure command P2b to the pressure control unit 30 and the second process of outputting the second pressure command P2b after correcting the first pressure command P1b by the pressure deviation to the pressure control unit 30.

[0117] The timing determination unit 83 causes the switch unit 81 to switch the selection switch. The timing determination unit 83 causes the switch unit 81 to switch the selection switch from the A side to the B side, or from the B side to the A side.

[0118] In the switch unit 81, the selection switch is at the A side at the time when the pressure control starts. The timing determination unit 83 causes the switch unit 81 to switch the selection switch to the B side if a specific time has passed or a specific condition is satisfied after the first pressure command P1b rises. A first example of the specific condition is the time when the detection pressure P3b overshoots the first pressure command P1b as the target pressure value, exceeds the first pressure command P1b, and then decreases to reach the first pressure command P1b. At this time, the timing determination unit 83 causes the switch unit 81 to switch the selection switch from the A side to the B side.

[0119] In addition, the second example of the specific condition is the moment when the detection pressure P3b overshooting the first pressure command P1b decreases and a specific time passes after reaching the first pressure command P1b. At this moment, the timing determination unit 83 causes the switch unit 81 to switch the selection switch from the A side to the B side. As described above, the specific condition may be a combination of the detection pressure P3b reaching a specific value (the first pressure command P1b) and the passage of a specific time.

[0120] Next, a control processing procedure of the die cushion mechanism 3 by the die cushion control device 200B will be described. Fig. 9 1 is a flowchart showing the control processing sequence of the die cushion mechanism by the die cushion control device involved in the second embodiment. Fig. 9 The processing shown is Figure 3 The same processing as described in is omitted.

[0121] In the die cushion control device 200B, the pressure command generation unit 50 generates the first pressure command P1b corresponding to the pressure distribution generated in the die cushion mechanism 3 during the press working (step S10), and sends it to the pressure command correction unit 80B.

[0122] The pressure command correction unit 80B determines whether a specific condition is satisfied (step S11). The specific condition is, for example, the time when the detected pressure P3b exceeds the first pressure command P1b and then decreases to reach the first pressure command P1b.

[0123] When the specific condition is not satisfied (step S11 , No), the pressure control unit 30 performs pressure control so that the detected pressure P3 b follows the first pressure command P1 b (step S12 ).

[0124] The die cushion control device 200B determines whether the pressure control is completed (step S13). If the pressure control is completed (step S13, Yes), the pressure command generation unit 50 ends the generation of the first pressure command P1b. Thus, the die cushion control device 200B ends the operation related to the pressure control.

[0125] If the pressure control is not completed (step S13, No), the die cushion control device 200B returns to the process of step S11. In this case, the pressure command correction unit 80B determines whether the specific condition is satisfied (step S11). If the specific condition is not satisfied (step S11, No), the die cushion control device 200B executes the processes of steps S12 and S13.

[0126] When the specific condition is satisfied (step S11, Yes), the die cushion control device 200B performs the processing of steps S20 to S60. In step S60, the die cushion control device 200B determines whether the pressure control is completed. When the pressure control is not completed (step S60, No), the die cushion control device 200B returns to the processing of step S20 and repeats the processing of steps S20 to S60.

[0127] If the pressure control is completed (step S60, Yes), the pressure command generating unit 50 ends the generation of the first pressure command P1b. Thus, the die cushion control device 200B ends the operation related to the pressure control.

[0128] Here, use Fig.10 Effects obtained by controlling the die cushion mechanism 3 by the die cushion control device 200B will be described. Fig.10 This is a diagram for explaining a pressure waveform when the die cushion mechanism is controlled by the die cushion control device according to the second embodiment.

[0129] exist Fig.10 In FIG. 1 , the upper graph shows a waveform related to pressure, and the lower graph shows the state of the selection switch of the switch section 81 (A side or B side). Fig.10 The horizontal axis of the graph shown in the upper part of is time, and the vertical axis is pressure. Fig.10 The upper part of shows the waveform of the first pressure command P1b indicated by the solid line, the waveform of the detected pressure P3b indicated by the dotted line, and the waveform of the second pressure command P2b indicated by the one-dot chain line. Fig.10 The horizontal axis of the graph shown in the lower part of is time, and the vertical axis is the switching timing of the selection switch.

[0130] exist Fig.10 , the waveform is shown when the detected pressure P3b overshoots the first pressure command P1b as the target pressure value and exceeds the first pressure command P1b at timing T1. Fig.10 , the case where the detected pressure P3b exceeds the first pressure command P1b and then decreases and reaches the first pressure command P1b at timing T2 is shown. In this case, the switch unit 81 switches the selection switch from the A side to the B side at timing T2.

[0131] While the selection switch is on the A side, the first pressure command P1b is not corrected, and the first pressure command P1b and the second pressure command P2b are consistent. At timing T2 when the selection switch is switched from the A side to the B side, the pressure command correction unit 80B starts correcting the first pressure command P1b using the correction pressure command 71 predicted by the deviation prediction unit 70, so the second pressure command P2b becomes larger than the first pressure command P1b.

[0132] The die cushion control device 200B can make the steady-state detection pressure P3b after the selection switch is switched to the B side coincide with the desired pressure command, i.e., the first pressure command P1b, by switching the selection switch as described above. In addition, the die cushion control device 200B does not correct the first pressure command P1b immediately after the timing at which the first pressure command P1b rises, thereby having the effect of preventing the pressure overshoot generated when the first pressure command P1b rises from becoming larger.

[0133] As described above, according to the second embodiment, in addition to the effects obtained by the first embodiment, the die cushion control device 200B can suppress further increase in pressure overshoot even when pressure overshoot occurs transiently when the first pressure command P1b increases.

[0134] Moreover, similarly to the die cushion control device 200A, the die cushion control device 200B can correct the drop in the detection pressure P3 b that occurs in a steady state with respect to the first pressure command P1 b to an appropriate level without overcompensating or undercompensating.

[0135] Implementation method 3.

[0136] Next, use Figures 11 to 15 Embodiment 3 will be described. The die cushion control device 200B of the above-mentioned embodiment 2 shifts the timing of applying the correction pressure command 71 generated by predicting the pressure deviation by the deviation prediction unit 70 from the time when the first pressure command P1b rises by a specific timing. Thus, the die cushion control device 200B does not increase the pressure overshoot and corrects the steady-state pressure drop.

[0137] The deviation prediction unit 70 can predict the steady-state pressure drop behavior using equation (1), but the detected pressure P3b shows transient behavior in the transient response time (e.g., a period of several tens of ms) immediately after the switching timing of switching the switch unit 81. Therefore, the mold cushion control device 200B can correct the first pressure instruction P1b by the correction pressure instruction 71 generated by the deviation prediction unit 70, and sometimes cannot follow the first pressure instruction P1b in the transient response time. For example, Fig.10As shown in FIG. 1 , a slight deviation occurs between the detected pressure P3b and the first pressure command P1b corresponding to the desired pressure distribution at a time immediately after the selection switch is switched from the A side to the B side. In the above case, the die cushion control device 200B finely adjusts the switching timing of switching the selection switch before and after the timing T2, thereby reducing the deviation between the first pressure command P1b and the detected pressure P3b at a transient response time immediately after the selection switch is switched.

[0138] The die cushion control device 200B of the third embodiment has a function of finely adjusting the switching timing of the switching selection switch. Thus, the die cushion control device 200B of the third embodiment reduces the deviation between the first pressure command P1b and the detected pressure P3b.

[0139] As described above, immediately after the selection switch is switched, the detection pressure P3b does not show steady-state behavior, but shows transient response behavior. The behavior of the detection pressure P3b during the transient response time is affected by the slider speed, the control parameters of the pressure control system, the characteristics of the workpiece components, the type of mold, the temperature of the working oil used by the hydraulic cylinder 5 that drives the mold buffer mechanism 3, the temperature of the working oil used by the hydraulic pump 7, and other conditions. Under certain conditions, in order to sufficiently reduce the deviation of the transient response time by fine-tuning the switching timing of the switching selection switch, trial and error adjustment is required. Even if this adjustment can be made, it is necessary to adjust the optimal switching timing again when the type of the workpiece or the type of the mold is changed, or when the temperature of the working oil changes.

[0140] In the third embodiment, a training model is generated in which the learning device 110 described later learns the maximum value of the deviation in the transient response after the selection switch is switched (after the specific condition is satisfied) based on the control conditions (the estimated input data 115A and the switching timing 116A described later) used when controlling the die cushion mechanism 3. Then, the estimation device 120 described later applies the conditions during machining to the trained model to estimate the maximum value of the deviation. And, based on the maximum value of the deviation, the estimation device 120 calculates the switching timing 116C that is most suitable for the maximum value of the deviation.

[0141] In addition, the learning device 110 may be a component of the die cushion control device 200B, or may be another structure different from the die cushion control device 200B. In addition, the estimation device 120 may be a component of the die cushion control device 200B, or may be another structure different from the die cushion control device 200B. In addition, the switching timing (switching timing 116C described later) most suitable for the maximum value of the deviation estimated by the estimation device 120 may be calculated by a device other than the estimation device 120.

[0142] When the training model is generated by the learning device 110, the user determines one estimated input data 115A and one switching timing 116A. The estimated input data 115A and the switching timing 116A are data used for learning the maximum deviation value. The estimated input data 115A is data that affects the transient response of the detection pressure P3b.

[0143] Examples of the inferred input data 115A are the slider speed, the control parameters used by the pressure control unit 30, the characteristics of the workpiece components, the type of mold, the temperature of the working oil used by the hydraulic cylinder 5, the temperature of the working oil used by the hydraulic pump 7, the first pressure instruction P1b, etc.

[0144] The die cushion control device 200B measures the maximum value of the pressure deviation (hereinafter referred to as the deviation maximum value) in the transient response of about several tens of milliseconds from the switching timing 116A when the die cushion mechanism 3 is actuated using the estimated input data 115A and the switching timing 116A. The learning device 110 acquires a data set consisting of the estimated input data 115A, the switching timing 116A, and the deviation maximum value in this case. The die cushion control device 200B actuates the die cushion mechanism 3 while making various changes to the estimated input data 115A and the switching timing 116A. Thus, the learning device 110 acquires a plurality of data sets. In addition, the deviation maximum value may be measured by a device other than the die cushion control device 200B.

[0145] When the trained model is represented by a neural network, the learning device 110 calculates the optimal weight parameters by applying the error back propagation method, etc. to the obtained multiple data groups, thereby calculating the trained model. In addition, the learning device 110 can also calculate the trained model through batch learning, online learning, etc. Here, an example of using a neural network as a specific example of a trained model is described, but the model used by the learning device 110 is not limited to a neural network, and models such as decision trees, random forests, and support vector machines can also be used. The details of the neural network are described later.

[0146] Fig.11This is a diagram showing the configuration of a learning device according to Embodiment 3. The learning device 110 includes a data acquisition unit 111 , a model generation unit 112 , and a trained model storage unit 113 .

[0147] The data acquisition unit 111 has a function of a state observation unit that acquires the estimated input data 115A, the switching timing 116A, and the maximum deviation 117A as learning data. Here, the learning data is data in which the estimated input data 115A, the switching timing 116A, and the maximum deviation 117A are associated with each other. The data acquisition unit 111 acquires the estimated input data 115A, the switching timing 116A, and the maximum deviation 117A from the die cushion control device 200B. The data acquisition unit 111 generates learning data by associating the estimated input data 115A, the switching timing 116A, and the maximum deviation 117A. The data acquisition unit 111 sends the generated learning data to the model generation unit 112.

[0148] The model generation unit 112 learns the appropriate maximum deviation value 117A corresponding to the estimated input data 115A and the switching timing 116A based on the learning data generated from the combination of the estimated input data 115A, the switching timing 116A, and the maximum deviation value 117A sent from the data acquisition unit 111. That is, the model generation unit 112 generates a trained model 114 that estimates the maximum deviation value 117A corresponding to the estimated input data 115A and the switching timing 116A based on the estimated input data 115A and the switching timing 116A. The trained model storage unit 113 stores the trained model 114 generated by the model generation unit 112. The trained model 114 stored in the trained model storage unit 113 is read by the inference device 120.

[0149] Next, use Fig.12 The processing procedure of the learning process performed by the learning device 110 will be described. Fig.12 This is a flowchart showing the processing procedure of the learning process performed by the learning device involved in the third embodiment.

[0150] The data acquisition unit 111 acquires the learning data (step S110 ). Specifically, the data acquisition unit 111 acquires the estimated input data 115A, the switching timing 116A, and the maximum deviation value 117A as the learning data.

[0151] The model generation unit 112 performs learning processing (step S120) according to the learning data, which is a combination of the estimated input data 115A, the switching timing 116A, and the maximum deviation value 117A acquired by the data acquisition unit 111. The model generation unit 112 generates the trained model 114 by so-called supervised learning, for example, according to the learning data.

[0152] The trained model storage unit 113 stores the trained model 114 generated by the model generation unit 112 (step S130 ).

[0153] The model generation unit 112 can use a known learning algorithm such as taught learning. Here, a case where the model generation unit 112 performs taught learning using a neural network will be described.

[0154] The model generation unit 112 learns the switching timing 116A of the die cushion control device 200B by so-called tutored learning, for example, according to a neural network model. Here, tutored learning refers to a method in which a set of input and result (label) data is given to a learning device, and features existing in these learning data are learned, thereby inferring the result from the input.

[0155] A neural network is composed of an input layer composed of multiple neurons, an intermediate layer (hidden layer) composed of multiple neurons, and an output layer composed of multiple neurons. The intermediate layer can be one layer or more than two layers.

[0156] Fig.13 is a diagram showing the structure of a neural network used by a learning device according to Embodiment 3. Here, for example, the trained model 114 is Fig.13 The case of a three-layer neural network shown in FIG. 1 is described below. In a three-layer neural network, if multiple data are input to the input layer, the value of each data is multiplied by a parameter called a weight and input to the middle layer, and the result is further multiplied by each weight and output from the output layer. The output result is changed by the weight when input to the middle layer and the value of the weight when input to the output layer. In the case where the trained model 114 is a neural network, the learning device 110 continuously learns the weights.

[0157] The neural network of implementation mode 3 learns the maximum deviation 117A inferred for the target product through so-called supervised learning according to the learning data (data set) generated based on the combination of the inferred input data 115A, the switching timing 116A and the maximum deviation 117A obtained by the data acquisition unit 111.

[0158] That is, the neural network adjusts the weights so that the estimated input data 115A and the switching timing 116A are input to the input layer and the result output from the output layer approaches the maximum deviation value 117A. Thus, the neural network learns the maximum deviation value 117A corresponding to the estimated input data 115A and the switching timing 116A. Examples of the estimated input data 115A input to the input layer include the slider speed, the control parameters used by the pressure control unit 30, and the like.

[0159] The model generation unit 112 generates and outputs the trained model 114 by executing the above learning. The trained model storage unit 113 stores the trained model 114 output from the model generation unit 112.

[0160] As described above, the learning device 110 receives the estimated input data 115A and the switching timing 116A through the input layer, performs calculations through the intermediate layer and the output layer, and finally outputs the maximum deviation 117A between the first pressure command P1b and the detected pressure P3b during the transient response.

[0161] Next, the inference device 120 will be described. The inference device 120 uses the trained model 114 to infer the maximum deviation value 117B described later. The maximum deviation value 117B calculated by the inference device 120 is the maximum deviation value 117B corresponding to the inference input data 115B and the switching timing 116B received by the inference device 120. The inference device 120 calculates the switching timing 116C described later based on the maximum deviation value 117B. The timing determination unit 83 of the die cushion control device 200B uses the switching timing 116C calculated by the inference device 120.

[0162] Fig.14 This is a diagram showing the configuration of an estimation device according to Embodiment 3. The estimation device 120 includes a data acquisition unit 121 , an estimation unit 122 , a trained model storage unit 123 , and a calculation unit 124 .

[0163] The inference device 120 reads the trained model 114 from the trained model storage unit 113 of the learning device 110 and stores the trained model 114 in the trained model storage unit 123. The data acquisition unit 121 of the inference device 120 is a first data acquisition unit, and the data acquisition unit 111 of the learning device 110 is a second data acquisition unit.

[0164] The data acquisition unit 121 acquires the estimated input data 115B and the switching timing 116B, which are data for estimating the maximum deviation value 117B, from the die cushion control device 200B. The estimated input data 115B is the same data as the estimated input data 115A, and the switching timing 116B is the same data as the switching timing 116A. The estimated input data 115A and the switching timing 116A are learning data used for learning, whereas the estimated input data 115B and the switching timing 116B are estimation data used when estimating the maximum deviation value 117B. The data acquisition unit 121 sends the acquired estimated input data 115B and the switching timing 116B to the estimation unit 122.

[0165] The inference unit 122 reads the trained model 114 from the trained model storage unit 123. The inference unit 122 inputs the inference input data 115B and the switching timing 116B to the trained model 114. Thus, the trained model 114 infers the maximum deviation value 117B corresponding to the inference input data 115B and the switching timing 116B. That is, the inference unit 122 inputs the inference input data 115B and the switching timing 116B used to infer the maximum deviation value 117B acquired by the data acquisition unit 121 to the trained model 114 for inferring the maximum deviation value 117B. Thus, the inference unit 122 can calculate the maximum deviation value 117B inferred from the inference input data 115B and the switching timing 116B.

[0166] In the trained model 114, the relationship between the estimated input data 115A, the switching timing 116A, and the maximum deviation value 117A is learned. Therefore, the estimation unit 122 gives a plurality of switching timings 116B and specific estimated input data 115B as inputs to the trained model 114, and calculates a plurality of maximum deviation values ​​117B corresponding to these conditions. Furthermore, the calculation unit 124 calculates the switching timing 116C that makes the maximum deviation value 117B as small as possible. The calculation unit 124 outputs the calculated switching timing 116C to the timing determination unit 83 of the pressure instruction correction unit 80B included in the die cushion control device 200B.

[0167] As described above, the inference device 120 calculates the optimal switching timing 116C for the switch unit 81 to switch the selection switch from the A side to the B side based on the inference input data 115B, the switching timing 116B, and the trained model 114, and outputs it to the timing determination unit 83. Thus, the timing determination unit 83 stores the switching timing 116C. In addition, at least one of the learning device 110 and the inference device 120 may also exist on a cloud server.

[0168] Next, use Fig.15 , describing the processing sequence of the inference processing performed by the inference device 120. Fig.15 This is a flowchart showing the processing procedure of the estimation process performed by the estimation device involved in the third embodiment.

[0169] The data acquisition unit 121 acquires the estimated input data 115B and the switching timing 116B as input information from the die cushion control device 200B (step S140). The estimation unit 122 reads the trained model 114 from the trained model storage unit 123. The estimation unit 122 inputs the input information to the trained model 114 (step S150). Thus, the trained model 114 estimates the maximum deviation value 117B using the input information, and outputs the maximum deviation value 117B as the estimation result to the calculation unit 124 (step S160).

[0170] The calculation unit 124 calculates the switching timing 116C corresponding to the maximum deviation value 117B based on the maximum deviation value 117B calculated by the estimation unit 122. The calculation unit 124 outputs the calculated switching timing 116C to the timing determination unit 83 of the pressure command correction unit 80B. The timing determination unit 83 causes the switch unit 81 to switch the selection switch according to the switching timing 116C. In this case, the estimated input data 115A used by the die cushion control device 200B is the same data as the estimated input data 115B. That is, the die cushion control device 200B uses the estimated input data 115B used for the calculation of the switching timing 116C as new estimated input data 115A, and controls the die cushion mechanism 3 using the switching timing 116C corresponding to the estimated input data 115B.

[0171] The data acquisition unit 111 of the learning device 110 can acquire the estimated input data 115A, the switching timing 116A, and the maximum deviation value 117A used as learning data when the die cushion control device 200B switches the selection switch according to the switching timing 116C acquired from the estimation device 120. In this case, the learning device 110 relearns the maximum deviation value 117A corresponding to the acquired estimated input data 115A and the switching timing 116A, and updates the trained model 114.

[0172] As described above, the learning device 110 learns the maximum deviation value 117A, the estimation device 120 estimates the maximum deviation value 117B, and calculates the switching timing 116C. Then, the die cushion control device 200B controls the die cushion mechanism 3 using the switching timing 116C. Thus, even in the transient response time immediately after the timing of switching the switch portion 81, the die cushion control device 200B can control the die cushion mechanism 3 while following the first pressure instruction P1a.

[0173] In Embodiment 3, the case where the supervised learning is applied to the learning algorithm used by the model generation unit 112 is described, but the learning algorithm is not limited to the supervised learning. Regarding the learning algorithm, in addition to the supervised learning, reinforcement learning, unsupervised learning, or semi-supervised learning can also be applied.

[0174] In addition, the model generation unit 112 may learn the maximum deviation value 117B according to the learning data generated for the plurality of die cushion control devices 200B. In addition, the model generation unit 112 may acquire the learning data from the plurality of die cushion control devices 200B used in the same area and perform the learning process, or may perform the learning process using the learning data collected from the plurality of die cushion control devices 200B independently operating in different areas. In addition, the die cushion control device 200B for which the learning data is collected may be added to the diagnosis object in the middle, or may be removed from the diagnosis object. Furthermore, the switching timing 116C calculated using the trained model 114 learned from a certain die cushion control device 200B may also be applied to other die cushion control devices 200B.

[0175] In addition, deep learning that learns to extract feature quantities themselves can also be used as a learning algorithm used in the model generation unit 112. The model generation unit 112 can perform machine learning according to other well-known methods such as genetic programming, functional logic programming, and support vector machines.

[0176] As described above, in the third embodiment, similarly to the first and second embodiments, the die cushion control device 200B predicts a steady-state pressure drop and corrects the drop amount, thereby being able to steadily make the detected pressure P3b follow the first pressure command P1b corresponding to the desired pressure distribution.

[0177] Furthermore, the die cushion control device 200B determines the switching timing 116C of the switch portion 81 based on the trained model 114 so that the pressure deviation in the transient response is reduced, thereby being able to reduce the pressure deviation in the transient response immediately after the selection switch is switched.

[0178] Furthermore, even if conditions such as the slide speed, the components of the workpiece, the type of the mold, the temperature of the hydraulic oil, etc. are changed, the learning device 110 learns the behavior of the pressure deviation corresponding to the condition change through the trained model 114. In addition, the estimation device 120 determines the switching timing 116C of the switch unit 81 based on the trained model 114. Thus, the mold cushion control device 200B can reduce the pressure deviation during transient response.

[0179] Here, the hardware configuration of the die cushion control devices 200A and 200B will be described. The die cushion control devices 200A and 200B have the same hardware configuration, so here, the hardware configuration of the die cushion control device 200A according to the first embodiment will be described.

[0180] Fig.16 This is a diagram showing a hardware configuration example for realizing the die cushion control device according to the first embodiment.

[0181] The die cushion control device 200A can be implemented by an input device 300, a processor 210, a memory 220, and an output device 400. An example of the processor 210 is a CPU (also called a Central Processing Unit, a central processing unit, a processing device, an arithmetic device, a microprocessor, a microcomputer, a DSP (Digital Signal Processor)) or a system LSI (Large Scale Integration). An example of the memory 220 is a RAM (Random Access Memory) or a ROM (Read Only Memory).

[0182] The die cushion control device 200A is implemented by the processor 210 reading and executing a die cushion control program executable by a computer for executing the operation of the die cushion control device 200A stored in the memory 220. The die cushion control program, which is a program for executing the operation of the die cushion control device 200A, can be said to be a procedure or method for causing a computer to execute the die cushion control device 200A.

[0183] The die cushion control program executed by the die cushion control device 200A has a module structure including a pressure control unit 30, a pressure command generation unit 50, a deviation prediction unit 70 and a pressure command correction unit 80A, which are downloaded to and generated on the main storage device.

[0184] The input device 300 receives the slide acceleration and the detection pressure P3a and sends them to the processor 210. The memory 220 is used as a temporary memory when the processor 210 executes various processes. In addition, the memory 220 stores pressure distribution, control parameters used by the pressure control unit 30, and the mold cushion movement amount per one rotation of the servo motor 10. The output device 400 outputs the motor speed command 24 to the speed control unit 23.

[0185] The die cushion control program can be provided as a computer program product by storing it in a computer-readable storage medium in an installable or executable form. In addition, the die cushion control program can also be provided to the die cushion control device 200A via a network such as the Internet. In addition, regarding the functions of the die cushion control device 200A, part of it can be realized by dedicated hardware such as a dedicated circuit, and part of it can be realized by software or firmware. In addition, the learning device 110 and the inference device 120 can also be realized by the same hardware structure as the die cushion control device 200A.

[0186] The configuration described in the above embodiment is merely an example, and may be combined with other known technologies, the embodiments may be combined with each other, and part of the configuration may be omitted or changed without departing from the gist.

[0187] Description of the label

[0188] 1 slider, 2 slider control unit, 3, 3A mold buffer mechanism, 4 buffer pad, 5 hydraulic cylinder, 6 piping, 7 hydraulic pump, 8 pressure detector, 10 servo motor, 11A, 11B pulley, 12 timing belt, 13 reducer, 14 ball screw, 23 speed control unit, 24 motor speed command, 25 driving current, 30 pressure control unit, 41, 42 multiplier, 43 integrator, 44, 85 adder, 45 subtractor, 50 pressure command generation unit, 51, 55, 57 transfer characteristics, 52, 54 motor speed, 53 interference speed, 56 motor position, 60 slider acceleration calculation unit, 70 deviation prediction unit, 71 correction pressure command, 80A, 80B pressure command correction unit, 81 switch unit, 83 timing determination unit, 90 mechanical mechanism, 101A processing system, 110 learning device, 111, 121 data acquisition unit, 112 model generation unit, 113, 123 trained model storage unit, 114 trained model, 115A, 115B inferred input data, 116A~116C switching timing, 117A, 117B maximum deviation value, 120 inference device, 122 inference unit, 124 calculation unit, 200A, 200B mold buffer control device, 210 processor, 220 memory, 300 input device, 400 output device, P1a, P1b first pressure instruction, P2a, P2b second pressure instruction, P3a, P3b detection pressure.

Claims

1. A die cushion control device, which uses a servo motor as a driving source to control a die cushion mechanism that generates pressure or force on a slide of a stamping machine, The mold cushion control device is characterized by having: a pressure command generating unit that outputs a command of the pressure or force generated between the die cushion mechanism and the slider, that is, a first pressure command; a deviation prediction unit that obtains information on the pressure or force generated between the die cushion mechanism and the slider as a detection pressure, and predicts a pressure deviation, which is a difference between the pressure or force under the first pressure command and the detection pressure, generated when the die cushion mechanism is controlled according to the first pressure command, based on a translation acceleration of the slider, a control parameter used when controlling the pressure or force of the die cushion mechanism, and a die cushion movement amount per one rotation of the servo motor, and outputs the pressure deviation as a correction pressure command; a pressure command correction unit that calculates a second pressure command by correcting the first pressure command using the correction pressure command; and The pressure control unit calculates a speed command for causing the detection pressure to follow the second pressure command, and outputs the speed command to the speed control unit, and the speed control unit outputs a drive current corresponding to the speed command to the servomotor.

2. The mold cushion control device according to claim 1, characterized in that: The pressure control unit performs proportional-integral control using a proportional gain and an integral gain. The deviation prediction unit predicts the pressure deviation by dividing the translation acceleration of the slide by the proportional gain, the integral gain, and the amount of movement of the die cushion per one rotation of the servo motor.

3. The mold buffer control device according to claim 1, characterized in that: The pressure instruction correction unit includes a switch unit that switches between a first process of outputting the first pressure instruction as the second pressure instruction to the pressure control unit and a second process of outputting the second pressure instruction after correcting the first pressure instruction by the pressure deviation to the pressure control unit. The switch unit executes the first process until a specific condition is satisfied, and switches from the first process to the second process if the specific condition is satisfied. The specific condition is a combination of the detection pressure reaching the first pressure command and the passage of a specific time.

4. The mold buffer control device according to claim 2, characterized in that: The pressure instruction correction unit includes a switch unit that switches between a first process of outputting the first pressure instruction as the second pressure instruction to the pressure control unit and a second process of outputting the second pressure instruction after correcting the first pressure instruction by the pressure deviation to the pressure control unit. The switch unit executes the first process until a specific condition is satisfied, and switches from the first process to the second process if the specific condition is satisfied. The specific condition is a combination of the detection pressure reaching the first pressure command and the passage of a specific time.

5. The mold cushion control device according to claim 3, characterized in that: The apparatus further comprises an inference device for inferring a maximum value of the pressure deviation in a transient response of the detected pressure after the specific condition is satisfied, that is, a maximum deviation value. The inference device comprises: a first data acquisition unit that acquires a control condition that is a condition used when controlling the die cushion mechanism and the maximum value of the deviation when the die cushion mechanism is controlled using the control condition; and an inference unit that uses a trained model for inferring the maximum deviation based on the control condition, infers the maximum deviation based on the control condition acquired by the first data acquisition unit, and calculates a switching timing from the first process to the second process that reduces the maximum deviation in the transient response based on the estimated maximum deviation. The switch unit switches from the first process to the second process at the switching timing.

6. The mold cushion control device according to claim 4, characterized in that: The apparatus further comprises an inference device for inferring a maximum value of the pressure deviation in a transient response of the detected pressure after the specific condition is satisfied, that is, a maximum deviation value. The inference device comprises: a first data acquisition unit that acquires a control condition that is a condition used when controlling the die cushion mechanism and the maximum value of the deviation when the die cushion mechanism is controlled using the control condition; and an inference unit that uses a trained model for inferring the maximum deviation based on the control condition, infers the maximum deviation based on the control condition acquired by the first data acquisition unit, and calculates a switching timing from the first process to the second process that reduces the maximum deviation in the transient response based on the estimated maximum deviation. The switch unit switches from the first process to the second process at the switching timing.

7. The mold cushion control device according to claim 5, characterized in that: Also having a learning device for generating the trained model, The learning device comprises: a second data acquisition unit that acquires learning data including the control condition and the maximum deviation value; and A model generation unit generates the trained model using the learning data.

8. The mold cushion control device according to claim 6, characterized in that: Also having a learning device for generating the trained model, The learning device comprises: a second data acquisition unit that acquires learning data including the control condition and the maximum deviation value; and A model generation unit generates the trained model using the learning data.

9. The mold cushion control device according to any one of claims 1 to 8, characterized in that: The die cushion mechanism is driven by the servo motor, hydraulic cylinder and rotary pump. The die cushion movement amount per one rotation of the servo motor is determined based on a value obtained by dividing a discharge volume of the hydraulic oil per one rotation of the rotary pump by a pressure-receiving cross-sectional area of ​​the hydraulic cylinder.

10. The die cushion control device according to any one of claims 1 to 8, characterized in that: The mold buffer mechanism is driven by the servo motor, ball screw, synchronous belt and reducer. The mold cushion movement amount per one rotation of the servo motor is determined by dividing the ball screw pitch, which is the movement amount per one rotation of the ball screw, by the pulley ratio of the timing belt and the reduction ratio of the reduction gear.

11. A die cushion control method, which uses a servo motor as a driving source to control a die cushion mechanism that generates pressure or force on a slide of a stamping machine, The die buffer control method is characterized by comprising: a pressure command generating step, wherein the die cushion control device outputs a command of the pressure or force generated between the die cushion mechanism and the slider, namely, a first pressure command; a pressure detection step, wherein the mold cushion control device detects information of the pressure or force generated between the mold cushion mechanism and the slider as a detection pressure; a deviation prediction step in which the die cushion control device predicts a pressure deviation, which is a difference between the pressure or the force under the first pressure command and the detection pressure, generated when the die cushion mechanism is controlled according to the first pressure command, based on the translation acceleration of the slider, a control parameter used when controlling the pressure or the force of the die cushion mechanism, and a die cushion movement amount per one rotation of the servo motor, and outputs the pressure deviation as a correction pressure command; a pressure command correction step, wherein the die cushion control device corrects the first pressure command by using the correction pressure command, thereby calculating the second pressure command; and In the pressure control step, the die cushion control device calculates a speed command for causing the detection pressure to follow the second pressure command, and outputs the speed command to a speed control unit, which outputs a drive current corresponding to the speed command to the servomotor.

12. A die cushion control program, using a servo motor as a driving source, controls a die cushion mechanism that generates pressure or force on a slide of a stamping machine, The die cushion control program is characterized in that it enables the computer to execute: A pressure command generating step of outputting a command of the pressure or force generated between the die cushion mechanism and the slider, that is, a first pressure command; A pressure detection step of detecting information of the pressure or force generated between the mold buffer mechanism and the slider as a detection pressure; a deviation prediction step of predicting a pressure deviation, which is a difference between the pressure or the force under the first pressure command and the detection pressure, generated when the die cushion mechanism is controlled according to the first pressure command, based on the translation acceleration of the slider, a control parameter used when controlling the pressure or the force of the die cushion mechanism, and a die cushion movement amount per one rotation of the servo motor, and outputting the pressure deviation as a correction pressure command; a pressure command correction step of correcting the first pressure command by using the correction pressure command, thereby calculating a second pressure command; and The pressure control step calculates a speed command for causing the detection pressure to follow the second pressure command, and outputs the speed command to a speed control unit, which outputs a drive current corresponding to the speed command to the servomotor.

Images