How to assemble the mold

The method minimizes punch and die contact by adjusting mold positions using sensor-coordinated adjustments, enhancing tool life and machining efficiency.

JP7591736B2Active Publication Date: 2024-11-29PANASONIC INTELLECTUAL PROPERTY MANAGEMENT CO LTD
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
JP2021048978
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2021-03-23
Publication Date
2024-11-29
Estimated Expiration
2041-03-23

AI Technical Summary

Technical Problem

Existing mold assembly methods result in increased contact between the punch and die due to mold distortion when mounted on a press machine, leading to tool wear and reduced tool life.

Method used

A method that adjusts the relative positions of the upper and lower molds offline and online, using sensors to measure and compare coordinates, and applies correction values to minimize contact and extend tool life.

Benefits of technology

Significantly reduces punch and die contact, extending tool life and improving machining efficiency by reducing wear and downtime.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007591736000001
    Figure 0007591736000001
  • Figure 0007591736000002
    Figure 0007591736000002
  • Figure 0007591736000003
    Figure 0007591736000003
Patent Text Reader

Abstract

To provide an assembly method of mold which can elongate tool life by greatly reducing the occasion where a punch comes into contact with a die.SOLUTION: An assembly method of mold includes steps of: acquiring coordinates of an upper die and coordinates of a lower die off line by measuring the position of a mold, while changing the relative position of the upper die 22 and the lower die 23 of the mold 21 without installing the mold 21 including the upper die and the lower die on a processing device; acquiring the coordinates of the upper die and the coordinates of the lower die on line by measuring the position of the mold, while changing the relative position of the upper die 22 and the lower die 23 of the mold 21 by installing the mold on the processing device; calculating a difference 32 between the coordinates of the upper die and the coordinates of the lower die by comparing the coordinates of the upper die and the coordinates of the lower die off line with the coordinates of the upper die and the coordinates of the lower die on line; and calculating the value of the difference of the coordinates of the upper die and the value of a difference 31 of the coordinates of the lower die as a correction value 33 of the coordinates of the upper die off line.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a method for assembling with high precision a die used for cutting, bending, drawing, etc., and equipped with sensors capable of measuring load, position, vibration, etc. [Background technology]

[0002] In cutting, the workpiece is generally placed on a tool called a die and pressed down by a member called a stripper, while the workpiece is pushed into the die by a tool called a punch to punch out the workpiece into a desired shape. Cutting is commonly used in a wide range of manufacturing fields, such as the manufacture of home appliances, precision instruments, and automobile parts.

[0003] In cutting processes using punches and dies, it is common for on-site workers to adjust the positions of the dies and tools by trial and error depending on the individual dies. However, there are cases where such trial and error adjustments are not sufficient, and in such cases, processed products of the specified quality cannot be obtained. For this reason, a method of adjusting the positions of the dies and tools has been proposed, such as the adjustment method disclosed in Patent Document 1, in which a coating of a specified thickness is formed on the punch to make it easier to center the punch and die, and the clearance (gap) between the punch and die is ensured within a specified dimensional range. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2009-166053 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, the adjustment method of Patent Document 1 shown as a conventional example is characterized by minimizing damage such as scratches and scrapes that occur to the tool even when the punch comes into contact with the die, and is an adjustment method based on the premise that the punch and die will inevitably come into contact within a certain range of adjustment variation.

[0006] This assumption is common knowledge in the field of die adjustment, and the main reasons for this are explained below. Generally, when performing maintenance on a die for reasons such as replacing a tool or re-grinding, the desired maintenance is performed with the die removed from the press device (offline state), and then the tool is reattached to the die and assembly adjustments are performed mainly by visual inspection so that the gap between the punch and the die falls within a certain range (blade fitting adjustment).

[0007] However, when the die is loaded back onto the press machine, the die becomes distorted and bent significantly due to the die's own weight, the characteristics of each press machine (mainly distortion), and the difference in the center of gravity balance between the press machine and the die. In other words, the way the die bends differs between when it is loaded onto the press machine (online state) and when it is offline, and the assembly adjustments made by the worker in the offline state do not apply in the online state. For this reason, the punch and die often have no choice but to come into contact with each other.

[0008] The adjustment method of Patent Document 1 shown as a conventional example is based on the above-mentioned premise, and therefore has the problem that repeated punching operations by the press device increase the opportunities for contact between the punch and die, increasing the amount of load accumulated on the punch and die, and shortening the tool life.

[0009] In view of the above-mentioned conventional problems, the present disclosure aims to provide a method for assembling a die that can reduce the opportunity for contact between the punch and die and extend the tool life. [Means for solving the problem]

[0010] The mold assembly method of the present disclosure includes the steps of measuring the position of a mold having an upper mold and a lower mold while changing the relative position of the upper mold and the lower mold of the mold without installing the mold on a processing apparatus, and obtaining offline coordinates of the upper mold and coordinates of the lower mold; installing the mold on the processing apparatus and measuring the position of the mold while changing the relative position of the upper mold and the lower mold of the mold, and obtaining online coordinates of the upper mold and coordinates of the lower mold; comparing the offline coordinates of the upper mold and the lower mold coordinates with the online coordinates of the upper mold and the lower mold coordinates, and calculating the difference in coordinates of the upper mold and the difference in coordinates of the lower mold; and calculating the difference in coordinates of the upper mold and the difference in coordinates of the lower mold as a correction value for the offline coordinates of the upper mold. Effect of the Invention

[0011] According to the die assembly method of the present disclosure, the opportunity for contact between the punch and die can be significantly reduced, thereby extending the tool life. [Brief description of the drawings]

[0012] [Figure 1] 1 is a block diagram showing an overview of a die evaluation system in the processing device according to the first embodiment. [Diagram 2] 1 is a schematic cross-sectional view showing a cross-sectional configuration of a mold in which a sensor is installed in embodiment 1. FIG. [Diagram 3] 2 is a block diagram showing the configuration of an input processing unit, a learning processing unit, and an output processing unit in the first embodiment. FIG. [Figure 4] 4 shows a workflow of a die assembling method using the evaluation system in the first embodiment. [Figure 5(a)] 5 is a schematic diagram showing the results of offline measurement of coordinates at the top dead center and bottom dead center of an upper die in the first embodiment. FIG. [Figure 5(b)] 5 is a schematic diagram showing the results of online measurement of coordinates at the top dead center and bottom dead center of an upper die in the first embodiment. FIG. [Figure 5(c)] 10 is a schematic diagram showing a difference 31 between the online coordinate measurement result at the bottom dead center of the upper die and the target value. FIG. [Figure 5(d)] 10 is a schematic diagram showing a difference 32 between the online measurement result of the coordinates at the bottom dead center of the upper die and the offline measurement result of the coordinates at the bottom dead center of the upper die. FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] The method for assembling a mold according to the first aspect includes the steps of: measuring the position of the mold while changing the relative positions of the upper and lower dies of the mold without installing the mold having an upper and lower dies in a processing device, and obtaining the coordinates of the upper and lower dies offline; installing the mold in the processing device, and measuring the position of the mold while changing the relative positions of the upper mold and the lower mold to obtain online coordinates of the upper mold and the lower mold; A step of comparing the offline coordinates of the upper mold and the lower mold with the online coordinates of the upper mold and the lower mold to calculate a difference in the coordinates of the upper mold and a difference in the coordinates of the lower mold; A step of calculating a difference value between the coordinates of the upper mold and the coordinates of the lower mold as a correction value of the coordinates of the upper mold offline; Includes.

[0014] The mold assembly method according to the second aspect is the above-mentioned first aspect, further comprising a step of repeating machine learning so that the online coordinates of the upper mold and the coordinates of the lower mold approach the coordinates of the upper mold and the coordinates of the lower mold that are set as targets, thereby obtaining a difference value between the coordinates of the upper mold and the coordinates of the lower mold offline after machine learning, or a value close to the difference; In the step of calculating a correction value for the upper mold coordinates offline, the correction value for the upper mold coordinates offline may be a value of the difference between the upper mold coordinates offline after machine learning and the lower mold coordinates, or a value close to the difference.

[0015] A mold assembling method according to a third aspect is the above-mentioned first or second aspect, in which, in the step of calculating a correction value for the coordinates of the upper mold, the correction value for the coordinates of the upper mold may be calculated as the coordinate of the bottom dead center of the upper mold.

[0016] A fourth aspect of the present invention relates to a method for assembling a die in any one of the first to third aspects, comprising the steps of: measuring a load applied to an upper punch and a lower die to obtain an offline measured waveform of the load applied to the punch and the die; measuring the load applied to the upper punch and the lower die to obtain an online measured waveform of the load applied to the punch and the die; It may further include.

[0017] A die assembly method according to a fifth aspect may, in the above-mentioned fourth aspect, determine that a measured waveform of the load applied to the punch and die offline or a measured waveform of the load applied to the punch and die online is below a preset threshold value.

[0018] A mold evaluation system according to a sixth aspect includes sensors attached to a mold having an upper mold and a lower mold, the sensors measuring the position of the mold, an input processing unit which receives the coordinates of the upper and lower dies measured by the sensors as physical quantities and converts them into state variables; a learning processing unit which compares the offline coordinates of the upper mold and the lower mold with the online coordinates of the upper mold and the lower mold to calculate a difference between the coordinates of the upper mold and the coordinates of the lower mold, and which repeats machine learning so that the online coordinates of the upper mold and the lower mold approach the coordinates of the upper mold and the lower mold set as targets, thereby calculating a correction value for the offline coordinates of the upper mold after machine learning, which is a difference between the coordinates of the upper mold offline and the coordinates of the lower mold or a value close to the difference; an output processing unit that outputs a correction value of the coordinates of the upper mold offline; Equipped with.

[0019] Hereinafter, a mold assembly method and a mold evaluation system used therein according to the present disclosure will be described with reference to the accompanying drawings. Note that substantially the same members are given the same reference numerals.

[0020] (Embodiment 1) The die assembly method according to the first embodiment includes an evaluation system 1 that measures the behavior of the die and collects and analyzes data, and a method of assembling and adjusting the die using the evaluation system 1.

[0021] (Rating System) Fig. 1 is a block diagram showing an overview of a mold evaluation system 1. First, the evaluation system 1 will be described with reference to the block diagram of Fig. 1 showing the overview thereof. The evaluation system 1 of the present disclosure includes sensors 2 attached to a mold, an input processing unit 3, a learning processing unit 4, and an output processing unit 5.

[0022] This evaluation system 1 has the following functions: an input processing unit 3 that executes an input step of converting a physical quantity 6 being processed into a state variable 7 based on measurement signals from sensors 2; a learning processing unit 4 that executes a learning step such as calculating and determining 9 the state variable based on the state variable 7 and a target numerical value 8; and an output processing unit 5 that executes an output step such as a calculation result 10 and an alarm 11. The evaluation system 1 may have a typical computer configuration as a physical configuration capable of executing the functions of the input processing unit 3, the learning processing unit 4, and the output processing unit 5. For example, the evaluation system 1 may include a CPU, a memory, a storage unit, an input unit, an output unit, and the like (not shown). Furthermore, each function of the input processing unit 3, the learning processing unit 4, and the output processing unit 5 may be realized by a computer program. The computer program can be stored in a storage unit, for example, and read out as necessary and executed by a CPU to realize each function of the input processing unit 3, the learning processing unit 4, and the output processing unit 5.

[0023] (Sensor configuration) Next, the configuration of the sensors 2 attached to the mold in the evaluation system 1 will be described with reference to FIG. 2. FIG. 2 is a schematic cross-sectional view showing the cross-sectional configuration of the mold 21 in which the sensors 2 are installed. For convenience, the vertical upward direction is the Z direction, the horizontal plane perpendicular to this is the XY plane, the right direction of the paper is the X direction, and the direction from the front to the back of the paper is the Y direction. The mold 21 is equipped with at least a load sensor 29 and position sensors 30a, 30b, 30c, and 30d. The load sensor 29 can measure the load applied to the punch 24 or the die 25 when the upper mold 22 and the lower mold 23 of the mold are moved up and down. The position sensors 30a, 30b, 30c, and 30d can measure the position and inclination of the upper mold 22 (die set 26, punch plate 27, or stripper plate 28) by attaching, for example, a proximity sensor to the lower mold 23.

[0024] In FIG. 2, as an example of the arrangement of the load sensor 29, the load sensor 29 is fixed to the tip of the punch 24 with a bolt or the like, so that the load acting on the punch 24 can be measured with high sensitivity. As an example of measuring the position sensors 30a, 30b, 30c, and 30d (however, 30c and 30d are not shown), the position sensors are configured to be able to measure the positions in the X, Y, and Z directions at each of the four corners of the die set 26 of the upper mold 22.

[0025] Next, a block diagram showing an outline of the configuration of the input processing unit 3, the learning processing unit 4, and the output processing unit 5 of the evaluation system 1 will be described with reference to FIG.

[0026] (Input processing section) A physical quantity 6 is input to the input processing unit 3 from the sensors 2. The physical quantity 6 input from the sensors 2 is the coordinates of the die. The physical quantity 6 may include the load applied to the punch or die in addition to the coordinates of the die. Furthermore, the physical quantity 6 may include the sound generated during the cutting process, the vibration generated during the cutting process, the shear rate, the clearance between the die and the punch, and the temperature of the workpiece generated during the cutting process.

[0027] The physical quantity 6 measured in real time from the start to the end of one cutting process is converted as a state variable 12 and input to the learning processing unit 4. Note that, since a sufficient number of samples is required in the cutting evaluation to capture the process characteristics, for example, the tendency of the local value of the physical quantity 6 such as the curvature of the load profile curve, the sampling period (measurement period) of the physical quantity 6 is preferably 1 / 100 or less of the time required for cutting.

[0028] (Learning processing unit) In the cutting process of the present disclosure, an upper die 7 is moved up and down relative to a fixed lower die 8 to cut an object (workpiece). For this reason, one up and down movement is counted as one stroke, and the physical quantity 6 during processing is measured by a sensor while the operation is performed multiple times in succession. When the physical quantity 6 is input as a state variable 7, it is input to the learning processing unit 4 through at least two systems. The first is state variable 7a of physical quantity 6a measured by performing one or more up-down movements or cutting processes while the die 21 to which the sensor is attached is mounted on a processing device that performs cutting processes (hereinafter, online physical quantity 6a, state variable 7a). The second is state variable 7b of physical quantity 6b measured by performing one or more up-down movements while the die to which the sensor is attached is not mounted on a processing device that performs cutting processes, but is placed in a maintenance area, and is manually moved up and down using a jig or the like (hereinafter, offline physical quantity 6b, state variable 7b).

[0029] An overview of the processing in the learning processing unit 4 is shown below. First, of the state variables input to the learning processing unit 4, the difference between the online state variable 7a and a preset target numerical value 8 is calculated and stored as a difference value 31 with respect to the target numerical value 8. Furthermore, the difference between the online state variable 7a and the offline state variable 7b is also calculated and stored as a difference value 32 between online and offline.

[0030] Next, machine learning is repeated to set the difference value 32 to an optimal value (zero is not necessarily optimal) so that the difference value 31 approaches zero, thereby obtaining a correction value 33 for correcting the offline physical quantity 6b.

[0031] (Output processing section) The correction value 33 is output by the output processing unit 5 .

[0032] This learning process makes it possible to numerically calculate the offline physical quantity 6b that can optimize the online physical quantity 6a, and to clarify the target value for manual work using a jig or the like offline.

[0033] (Assembly and adjustment methods and examples) Next, a method for assembling and adjusting a die using the evaluation system 1 in the cutting process of the present disclosure will be described with reference to the workflow of FIG. Generally, at cutting processing sites, maintenance work frequently occurs in order to regrind punches or dies when tools become worn or damaged, or to switch product models, and the die must be removed from the cutting machine, disassembled, and then reassembled after the required treatment.

[0034] (1a) In the method of assembling and adjusting a die disclosed herein, first, when the die is assembled in an offline state such as during maintenance work, the upper die 22 is manually moved up and down relative to the lower die 23 at least once using a hanging jig or the like in order to align the blades between the punch 24 and the die 25 (S01).

[0035] At this point, if the combination of mold 21 and tool used here does not have correction value 33, the positions of upper mold 22 and lower mold 23 are adjusted so that when upper mold 22 and lower mold 23 are in the closest position (near bottom dead center), the load value is below a preset threshold value (ideally zero) and the X, Y and Z position coordinates at each position where the position sensor is installed have minimal deviations from the respective preset position coordinates.

[0036] (2a) In this state, signals are taken in from the load sensor 29 and the position sensors 30a, 30b, 30c, and 30d installed in the mold 21, and the physical quantity 6b is measured offline, converted into a state variable 7b, and stored (S02).

[0037] (1b) On the other hand, if the combination of the die 21 and the tool used at this time has a correction value 33 that can be obtained by a method (S05) described later, an adjustment is made offline to reflect the correction value 33. That is, the positions of the upper die 22 and the lower die 23 are adjusted (S06) so that when the upper die 22 and the lower die 23 are closest to each other (near the bottom dead center), the load value is equal to or less than a preset threshold value (ideally zero) and the X, Y and Z position coordinates at each position where the position sensor is installed have a minimum deviation from the position coordinates obtained by adding the correction value 33 to each preset predetermined position coordinate. The correction value 33 can be reflected by using, for example, an auxiliary plate or the like.

[0038] (2b) In this state, signals are taken in from the load sensor 29 and the position sensors 30a, 30b, 30c, and 30d installed in the mold 21, and the physical quantity 6b is measured offline, converted into a state variable 7b, and stored (S02).

[0039] (3) Next, at the stage where the die 21 is placed in the cutting machine, the upper die 22 is moved up and down one or more times in the cutting machine without feeding the workpiece (S03). (4) In this state, signals are received from the load sensor 29 and the position sensors 30a, 30b, 30c, and 30d installed in the mold 21, and the physical quantity 6a is measured online, converted into a state variable 7a, and stored (S04). If the load value exceeds a preset threshold value at this time, it is determined that an excessive load is being applied to the punch and die, and an alarm is issued via the output processing unit 5 to indicate an abnormal load.

[0040] It should be noted that valid data can be obtained by moving the upper die 22 up and down at least once for one stroke. It is more preferable to move the upper die 22 multiple times, since this allows data to be obtained that takes into account variations in the movement. The magnitude of the preset load threshold is preferably set to a load that generates a level that does not cause damage to the punch and die, such as chipping, etc. Ideally, it is most preferable to set the load applied to the punch and die to 0 (zero).

[0041] It is preferable that the predetermined position coordinate of the position sensor, which is set in advance, is the Z predetermined position coordinate of the position where the upper die is in surface contact with the lower die with the lower die as the reference near the bottom dead center, and that the X and Y predetermined position coordinates are the position where the outer shapes of the upper die and the lower die overlap in that state.

[0042] (5) Next, the physical quantity 6b and state variable 7b acquired offline and the physical quantity 6a and state variable 7a acquired online, which are input to the input processing unit of the evaluation system 1, are used to calculate a difference value 31 from the target numerical value 8, and a difference value 32 between the online and offline values, and are output by the evaluation system 1. A specific example of this will be described with reference to Figures 5(a) to 5(d). Here, the case where the physical quantity 6a is the coordinates of the upper mold will be described.

[0043] 5(a) shows the coordinates (X1, Y1, Z1), (X2, Y2, Z2), (X3, Y3, Z3), and (X4, Y4, Z4) of the four position sensors 30a, 30b, 30c, and 30d at the top and bottom dead centers of the upper die, which were acquired offline in S02. Points A, B, C, and D indicate predetermined position coordinates, and solid lines a, b, c, and d indicate lines connecting the predetermined position coordinates.

[0044] Fig. 5(b) shows the coordinates ('X1, 'Y1, 'Z1), ('X2, 'Y2, 'Z2), ('X3, 'Y3, 'Z3) and ('X4, 'Y4, 'Z4) of the four position sensors '30a, '30b, '30c and '30d at the top and bottom dead points of the upper die, acquired online. Dashed lines a', b', c' and d' indicate lines connecting the specified position coordinates.

[0045] First, the difference value 31 between the coordinates of the four points at the bottom dead center of the upper die and the target numerical value 8 is calculated, as shown in Fig. 5(c). The target numerical value 8, which is set manually in advance, is set to points E, F, G, and H, and the difference 31 between the position coordinates of the four points at the bottom dead center of the upper die, acquired online, and the target numerical value 8 is calculated. In this case, it is determined whether the difference 31 for each coordinate is equal to or less than a predetermined value (preferably as close to 0 as possible).

[0046] Next, Fig. 5(d) shows a diagram in which the position coordinates of four points of the upper mold acquired offline and the position coordinates of four points of the upper mold acquired online are superimposed. In this way, by calculating the difference 32 between the position coordinates of each of the upper mold acquired offline and online, the correction value 33 of each of the position coordinates of the upper mold is calculated as shown in the output processing unit of Fig. 3.

[0047] In this way, for the combination of mold 21 and tool used here, machine learning was continued until the difference between each coordinate of one of the four points of the upper mold and the target numerical value 8 in the difference value 31 became 0 (zero), that is, until the position coordinates of the four points of the upper mold obtained online approached the target numerical value 8, and the optimal value of the correction value 33 of the position coordinates of the four points of the upper mold obtained offline was continued to be calculated. (S05)

[0048] As a result, compared to conventional die assembly methods that rely on the skills of operators, the die assembly method disclosed herein can reduce the amount of wear on the punch and die for the same number of shots, thereby extending the tool life (number of shots that can be processed). [Industrial Applicability]

[0049] According to the present disclosure, mainly in the field of cutting processing, it is possible to realize an offline die assembly method that optimizes the state in which the die is placed on the processing device and operated. This significantly reduces the chances of contact between the punch and die during processing operations, extends the tool life, and is expected to improve productivity by reducing downtime of the processing device. [Explanation of symbols]

[0050] 1. Rating System 2 Sensors 3. Input Processing Section 4. Learning Processing Section 5 Output Processing Section 6 Physical quantities 7 State Variables 8 Target Values 9 Calculation and judgment of measured waveform 10 Calculation results 11 alarm 21 Mold 22 Upper mold 23 Lower mold 24 Punch 25 Die 26 Die Set 27 Punch Plate 28 Stripper plate 6a Online Physical Quantities 7a Online state variables 6b Offline physical quantities 7b Offline state variables 29 Load Sensor 30 Position sensor (30a, 30b, 30c, 30d) 31 Difference from target 32 Difference between online and offline 33 Correction value

Claims

1. a step of measuring a position of a mold having an upper mold and a lower mold while changing a relative position between the upper mold and the lower mold of the mold without installing the mold in a processing device, and obtaining coordinates of the upper mold or coordinates of the lower mold offline; a step of installing the mold in a processing device, measuring a position of the mold while changing a relative position between the upper mold and the lower mold of the mold, and obtaining online coordinates of the upper mold and the lower mold; a step of comparing the offline coordinates of the upper mold and the lower mold with the online coordinates of the upper mold and the lower mold to calculate a difference in the coordinates of the upper mold and a difference in the coordinates of the lower mold; calculating a difference value between the coordinates of the upper mold and the coordinates of the lower mold as a correction value of the coordinates of the upper mold offline; A method for assembling a mold, comprising:

2. The method further includes a step of repeating machine learning so that the online coordinates of the upper mold and the coordinates of the lower mold approach the coordinates of the upper mold and the coordinates of the lower mold that are set as targets, thereby obtaining a difference value between the coordinates of the upper mold and the coordinates of the lower mold offline after machine learning, or a value close to the difference value; A mold assembling method as described in claim 1, wherein in the step of calculating a correction value for the coordinates of the upper mold offline, a value of the difference between the coordinates of the upper mold offline after machine learning and the coordinates of the lower mold, or a value close to the difference, is set as the correction value for the coordinates of the upper mold offline.

3. 3. The die assembly method according to claim 1, wherein in the step of calculating a correction value for the coordinates of the upper die, the correction value for the coordinates of the upper die is calculated as a coordinate of a bottom dead center of the upper die.

4. measuring a load applied to the punch of the upper die and the die of the lower die to obtain a measured waveform of the load applied to the punch and the die offline; measuring a load applied to the punch of the upper die and the die of the lower die to obtain a measured waveform of the load applied to the punch and the die online; The method for assembling the mold according to claim 1 , further comprising:

5. The die assembly method according to claim 4, further comprising determining whether a waveform measured of the load applied to the punch and the die offline or a waveform measured of the load applied to the punch and the die online is equal to or less than a preset threshold value.

6. Sensors attached to a mold having an upper mold and a lower mold and for measuring the position of the mold; an input processing unit that receives input of coordinates of the upper mold and the lower mold measured by the sensors as physical quantities and converts them into state variables; a learning processing unit that compares the offline coordinates of the upper mold and the lower mold with the online coordinates of the upper mold and the lower mold to calculate a difference between the coordinates of the upper mold and the coordinates of the lower mold, and calculates a correction value for the offline coordinates of the upper mold, which is a difference between the offline coordinates of the upper mold and the coordinates of the lower mold after machine learning, or a value close to the difference, by repeating machine learning so that the online coordinates of the upper mold and the coordinates of the lower mold approach the coordinates of the upper mold and the coordinates of the lower mold that are set as targets; an output processing unit that outputs the correction value of the coordinates of the upper mold offline; A mold evaluation system comprising:

Citation Information

Patent Citations

  • Die-set system die for bending and device for automatically changing die using its die

    JP1992158926A

  • Slide guide correction device in press machine

    JP1993039799U

  • Precision punching die

    JP2008062282A

  • Structure for centering die and centering method

    JP2009166053A

  • Setting stamping dies in a press with overhead winching crane

    US20190168477A1