Injection molding machine, control system for injection molding machine, and control method for injection molding machine

The injection molding machine control system optimizes rotation and stopping of rotating parts with attached molds using machine learning, addressing the challenge of adapting to mold weight changes and ensuring accurate positioning.

JP7731808B2Active Publication Date: 2025-09-01THE JAPAN STEEL WORKS LTD
View PDF 9 Cites 0 Cited by

Patent Information

Application Number
JP2022003777
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2022-01-13
Publication Date
2025-09-01
Estimated Expiration
2042-01-13

AI Technical Summary

Technical Problem

Existing injection molding machines struggle to rotate a rotating part with attached molds quickly and accurately stop at the next position, failing to adapt to changes in mold weight and other conditions.

Method used

An injection molding machine control system equipped with a machine learning device that verifies and determines parameters for rotating and stopping the rotating part using a servo motor, optimizing rotation control through machine learning algorithms.

Benefits of technology

Enables precise and efficient rotation and stopping of the rotating part, adapting to changes in mold weight and conditions, thereby improving productivity and accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007731808000001
    Figure 0007731808000001
  • Figure 0007731808000002
    Figure 0007731808000002
  • Figure 0007731808000003
    Figure 0007731808000003
Patent Text Reader

Abstract

To provide an injection molding machine, an injection molding machine control system, and an injection molding machine control method that can perform appropriate rotation control of a rotating part in the injection molding machine in which a rotating part to which a mold is attached is rotatable by a servomotor.SOLUTION: In a control method for an injection molding machine 11 in which a rotary part 14 to which molds 13a and 13b are attached is rotatable by a servomotor 15, a target stop position P1 and P2 during rotation of the rotary part 14 is determined by using a position of the rotary part 14 in a state in which molds 13a and 13b are attached to the rotary part 14 and stopped the rotary part 14 after rotation by the servomotor 15, molds 13a and 13b are closed against the other mold 23 and at least part of the molds are in contact with each other.SELECTED DRAWING: Figure 2
Need to check novelty before this filing date? Find Prior Art

Description

[Technical Field]

[0001] The present invention relates to an injection molding machine in which a rotating part to which a mold is attached is rotatable by a servo motor, a control system for the injection molding machine, and a control method for the injection molding machine. [Background technology]

[0002] In an injection molding machine in which a rotating part to which a mold is attached is rotatable by a servo motor, it is required to rotate the rotating part as quickly as possible and stop it accurately at the next stopping position. A known device for rotating a rotary table, which is a rotating part, in a short time is described in Patent Document 1. In Patent Document 1, the rotation of the servo motor is controlled to accelerate the rotary table by approximately half of a predetermined rotation angle and decelerate the rotary table by approximately the remaining half to rotate the rotary table.

[0003] Furthermore, a known control device that can appropriately and quantitatively set the control gain or time constant of a driver using information about the acceleration of a driver such as an injection molding machine, and an adjustment device that has a function for adjusting the control gain or time constant, is described in Patent Document 2. Patent Document 2 aims to perform appropriate control that suppresses vibration of a servo motor by adjusting the control gain and time constant. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 10-646 [Patent Document 2] Japanese Patent Application Laid-Open No. 2007-141189 Summary of the Invention [Problem to be solved by the invention]

[0005] However, in Patent Document 1, although the rotation angle is decelerated by approximately half, there are cases where the optimal deceleration start position for the rotary table rotation is not reached. It is also unable to respond to changes in conditions such as the weight of the mold attached to the rotary table, which is the rotating part. As a result, it is often impossible to rotate the rotary table as quickly as possible and stop it accurately at the next stopping position. Furthermore, Patent Document 2 provides suggestions regarding the control gain and time constant of the servo motor, but, like Patent Document 1, it is unable to respond to changes in conditions such as the weight of the mold attached to the rotating part of the injection molding machine.

[0006] Therefore, an object of the present invention is to provide an injection molding machine, a control system for an injection molding machine, and a control method for an injection molding machine that can appropriately control the rotation of a rotating part to which a mold is attached, which is rotatable by a servo motor.

[0007] Other objects and novel features will become apparent from the description of this specification and the accompanying drawings. [Means for solving the problem]

[0008] The injection molding machine described in claim 1 of the present invention is an injection molding machine control system in which a rotating part to which a mold is attached is rotatable by a servo motor, and is equipped with a control device in which a machine learning device verifies parameters for rotating and stopping the rotating part by the servo motor with a mold attached to the rotating part, and determines the parameters. [Effects of the Invention]

[0009] The injection molding machine of the present invention is an injection molding machine control system in which a rotating part to which a mold is attached is rotatable by a servo motor, and is equipped with a control device in which a machine learning device verifies parameters for rotating and stopping the rotating part by the servo motor with a mold attached to the rotating part, and determines the parameters, thereby enabling appropriate rotation control of the rotating part. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a front view showing an injection molding machine according to an embodiment of the present invention. [Figure 2] FIG. 1 is a plan view showing an injection molding machine according to an embodiment of the present invention. [Figure 3] FIG. 2 is a block diagram showing a machine learning device of the control device of the injection molding machine of the present embodiment. [Figure 4] FIG. 2 is a diagram showing a neural network of a machine learning device of the laminate molding apparatus of the present embodiment. [Figure 5] FIG. 10 is a flowchart illustrating the verification of parameters for rotating and stopping the rotary table using the machine learning device of the injection molding machine of this embodiment. [Figure 6] FIG. 10 is a flowchart for correcting a target stop position of the injection molding machine according to the present embodiment. [Figure 7] 10A and 10B are explanatory diagrams illustrating the operation content when correcting the target stop position of the injection molding machine according to the embodiment. [Figure 8] FIG. 4 is a front view showing an injection molding machine according to a second embodiment. [Figure 9] FIG. 10 is a front view showing an injection molding machine according to a third embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0011] <Configuration of injection molding machine> The configuration of a vertical rotary injection molding machine 11, which is a type of injection molding machine, will be described with reference to Figures 1 and 2. The mold clamping device 12 of the vertical rotary injection molding machine 11 has a rotary table 14, which is a rotating part to which lower molds 13a and 13b, which are the first molds among the molds, are attached, and which is rotatable by a servo motor 15, which is a rotation mechanism. In this embodiment, the lower molds 13a and 13b are attached to mold attachment parts 14a and 14b of the rotary table 14, respectively, but the number of molds that can be attached to the rotary table 14 is not limited.

[0012] The fixed platen 16 of the mold clamping unit 12 is fixed to and supported by a plurality of support columns 18 on a base 17. Three tie bars 21a, 21b, and 21c are vertically disposed between a pressure platen 19 disposed below the fixed platen 16 and a movable platen 20 disposed above the fixed platen 16. One of the tie bars, 21a, also serves as the central axis of rotation of the rotary table 14 and is disposed at the center of the rotary table 14. The rotary table 14 is rotatably supported by a bearing (not shown) disposed around the tie bar 21a. Known sliding members, rollers, and rollers are disposed between the rotary table 14 and the fixed platen 16 to reduce resistance during rotation of the rotary table 14. A mold opening / closing / clamping mechanism, such as a toggle mechanism 22, is disposed between the pressure platen 19 and the fixed platen 16, and is operated by a servo motor 37 for mold clamping / opening / closing. The structure of the mold opening / closing / clamping mechanism is not limited.

[0013] An upper mold 23, which is a second mold and a movable mold, is attached to the mold attachment portion 20a on the underside of the movable platen 20. In this embodiment, there is one upper mold 23, but the number of upper molds 23 is not limited. The molds consisting of the lower mold 13 and the upper mold 23 have one convex mold and the other concave mold, and when the molds are closed and fitted, they abut and form a cavity inside. To ensure a good fit of the convex mold with the concave mold, guide pins 24 are provided near the four corners of the cavity surface of one of the molds, the lower mold 13 or the upper mold 23. The other mold has guide holes 25, into which the guide pins 24 are inserted, at positions corresponding to the guide pins 24 around the cavity surface. The guide pins 24 and guide holes 25 are well known, but even if the molds are attached at slightly different positions, the guide holes 25 guide the guide pins 24, allowing one mold to be closed onto the other mold without damaging the main parts of the molds.

[0014] A servo motor 15 for rotating the rotary table is fixed to the side of the stationary platen 16 with its drive shaft 26 facing upward. The servo motor 15 is composed of a motor section 15a and a rotary encoder 15b that detects the rotation angle. A timing belt 39 is stretched between a pulley 27 fixed to the drive shaft 26 and the outer periphery 14c of the rotary table 14. Driven by the servo motor 15, the rotary table 14 and the lower dies 13a and 13b attached to the rotary table 14 are rotated horizontally around a tie bar 21a, which serves as the center of rotation. The rotational position (rotation angle) of the rotary table 14 can be detected by a rotary encoder 15b attached to the motor section 15a of the servo motor 15. In this embodiment, the rotary encoder 15b is preferably an absolute rotary encoder that can set an origin position and detect position information as an absolute value from the origin position. An ejector mechanism 40 is provided on the rear side of the ejection position 35 of the stationary platen 16. The servo motor 15 including the rotary encoder 15b is connected to a control device 29 via a servo amplifier .

[0015] The rotation mechanism of the turntable 14 may be one in which a gear attached to the drive shaft 26 of the servomotor 15 for rotating the turntable rotates a gear wheel formed on the outer periphery of the turntable 14. Furthermore, the servomotor 15 for rotating the turntable may rotate the central axis of the turntable 14 provided around the tie bar 21a, and the rotation mechanism is not limited.

[0016] A proximity sensor 30 (proximity switch) serving as an interlock mechanism is attached to one side of the fixed platen 16, corresponding to the stop position of the rotary table 14. A cam member 31 serving as a detection target that is read by the proximity sensor 30 is also provided on the upper surface of the rotary table 14. However, these interlock mechanisms may be limit switches that directly contact the detection target, or other detection means. In this embodiment, no positioning mechanism is provided to mechanically position the rotary table 14, which is the rotating part, or no mechanical stopper is provided to stop the rotary table 14, which is the rotating part. This has the advantages of simplifying the structure of the injection molding machine 11 and preventing problems such as wear on the positioning mechanism or the positioning mechanism becoming stuck. Another advantage is that it does not require adjustment of the stopper position.

[0017] However, the present invention does not completely exclude the use of a positioning mechanism. Even when a positioning mechanism is used, it may be used only once for multiple table rotations during continuous molding, rather than every time. The positioning mechanism may also be operated once for multiple table rotations during continuous molding, at least once when the molding mold is replaced, when the molding conditions are changed, or when the molding conditions are generated. When a positioning mechanism is used, it is assumed that a positioning member (pin or wedge) is inserted into a positioning portion (positioning hole or positioning recess) on the rotary table 14. However, a simple mechanism in which a protrusion on the rotary table 14 abuts against a stopper portion on the fixed platen 16 may also be used.

[0018] In this embodiment, an injection unit 32 is provided on one side of the mold clamping unit 12 of the injection molding machine 11. The injection unit 32 is arranged so that a nozzle 34 fixed to a heating cylinder 33 is directed horizontally toward the mold. However, the injection unit may be arranged vertically relative to the mold clamping unit 12. The number of injection units 32 is also not limited.

[0019] <Configuration of injection molding machine control device and machine learning device> The control device 29 of this embodiment will be described with reference to the block diagram of the machine learning device of the control device in FIG. 3. The control device 29 includes an input unit 41, an operation processing unit 42 that processes operations such as sequence control of the injection molding machine, a memory unit 43, and an output unit 44. The operation processing unit 42 is connected to the input unit 41, the memory unit 43, and the output unit 44, respectively. The control device 29 also includes a machine learning device 38. In this embodiment, the memory unit 43 of the machine learning device 38 is shown in a block diagram as being shared with the memory unit 43 of the control device 29. The memory unit 43 of the machine learning device 38 includes a pass / fail memory unit 45 and a condition memory unit 46. The machine learning device 38 also includes a judgment unit 47 connected to the input unit 41, a condition correction unit 48 connected to the judgment unit 47, and a pass / fail determination unit 49 connected to the condition correction unit 48. The pass / fail determination unit 49 is connected to the output unit 44.

[0020] While the control device 29 has been described above as being divided into functional blocks, the operation processing unit 42 and the judgment unit 47, condition correction unit 48, and pass / fail determination unit 49 of the machine learning device 38 correspond to arithmetic devices such as a CPU. The judgment unit 47, condition correction unit 48, and pass / fail determination unit 49 of the machine learning device 38 are equipped with neural networks, and generate parameters related to the rotation control of the turntable 14 and perform pass / fail determination when the rotation of the turntable 14 is controlled using the parameters. As shown in FIG. 4, the specific network configuration of the machine learning device 38 is that the machine learning device 38 includes an input layer 72, an intermediate layer 73, and an output layer 74. Note that the machine learning device 38 may be configured to perform deep learning by providing multiple intermediate layers 73.

[0021] The pass / fail memory unit 45 and condition memory unit 46 of the control device 29 correspond to storage devices such as volatile memory and nonvolatile memory. The input unit 41 is connected to the current detection unit 28a of the servo amplifier 28 as well as the rotary encoder 15b of the servo motor 15. It is also connected to other detection devices of the injection molding machine 11. The input unit 41 is also connected to a touch-panel setting input device 50 that displays set values ​​and detected values ​​on a screen. The input unit 41 is also connected to peripheral devices other than the injection molding machine 11, such as a take-out machine. The input unit 41 of the control device 29 can also be connected to control devices (not shown) other than the injection molding machine 11. Specifically, it can be connected via wire or wireless to other injection molding machines or central control devices in the same or another factory. It may also be connected to control devices including servers of the same company, another company such as an injection molding machine manufacturer, or a cloud service company.

[0022] The output unit 44 is also connected to the command unit 28b of the servo amplifier 28. Furthermore, like the input unit 41, the output unit 44 is also connected to another control device, other actuators, and devices that control their operation. Furthermore, like the input unit 41, the output unit 44 is also connected to a setting input device 50 of the injection molding machine 11 and peripheral devices other than the injection molding machine 11, such as a take-out machine.

[0023] <Servo amplifier configuration> The servo amplifier 28 is connected to the control device 29 and the servo motor 15. As is well known, the servo amplifier is equipped with a power transistor, and supplies power of a predetermined current value to the servo motor 15 based on a command signal from the control device 29. The servo amplifier 28 is also connected to the rotary encoder 15b of the servo motor 15, and forms a speed closed loop or a position closed loop with the servo motor 15. The servo amplifier 28 further includes a current detection unit 28a and a command unit 28b, and is capable of detecting the value of the current supplied to the servo motor 15. The command unit 28b transmits and receives signals to and from the control device 29, and performs feedback control with the servo motor 15.

[0024] <Configuration of setting input device> Next, the setting input device 50 of the injection molding machine 11 will be described with reference to Figure 3. The setting input device 50 is composed of a touch panel or the like, and has a setting input section 51 for inputting various setting values ​​for the injection molding machine 11. The setting input device 50 also has a display section 52 for displaying the status detected by each sensor of the injection molding machine 11.

[0025] The setting input unit 51 includes a molding condition input unit 53, a pass / fail input unit 54, a control gain input unit 55, and a time constant input unit 56. The control gain input unit 55 is used to input the gain of the servo motor 15 and other components installed in the injection molding machine. It includes a speed proportional gain input unit 57 (speed P gain input unit) for inputting the value of the speed proportional gain, a speed integral gain input unit 58 (speed I gain input unit) for inputting the value of the speed integral gain, and a position loop gain input unit 59 for inputting the value of the position loop gain. It also includes a time constant input unit 56 for inputting the value of the current command filter. Correcting the speed proportional gain is essential and changes the responsiveness across all frequency bands. Increasing the gain value increases responsiveness, but increases the likelihood of overshoot and vibration. Correcting the speed integral gain is also essential and changes the responsiveness at low frequencies. Increasing the value of the speed integral gain increases the servo lock force, improving responsiveness, but increases the likelihood of overshoot and vibration. The position loop gain optimizes the responsiveness, positioning speed, and accuracy of feedback control.

[0026] <How to set parameters for rotating and stopping the rotary table> Next, a control method for the injection molding machine 11, particularly a method for setting parameters for rotating and stopping the rotating table 14 (rotating unit) using the servo motor 15, will be described with reference to the flowchart in FIG. 5. In the present invention, a machine learning device 38 verifies parameters for rotating and stopping the rotating unit (rotating unit) using the servo motor 15 with molds (lower molds 13a, 13b) attached to the rotating unit (rotating unit) (rotating table 14) and determines the parameters. One possible stopping method involves determining the target stopping positions P1 and P2 of the rotating table 14 by rotating the rotating unit (rotating unit) using the servo motor 15 with the lower molds 13a, 13b attached to the rotating table 14, and then stopping the rotating unit. Then, the lower molds 13a, 13b are closed against the upper mold 23, the other mold, so that at least portions of the molds are in contact with each other.

[0027] First, injection molding machine information is input to the machine learning device 38 (s1). The injection molding machine information is an essential input item for the machine learning device 38 to determine parameters for controlling the rotation of the turntable 14. The injection molding machine information includes at least two of the following: the type or capacity of the servo motor 15; the reduction ratio of the reducer or belt; the friction coefficient and power loss rate of the bearings of the turntable 14 and the reducer; the standard mold weight attached to the turntable 14; and the weight of the turntable 14. In particular, the type or capacity of the servo motor 15 constitutes essential injection molding machine information. It is desirable that this injection molding machine information is already stored in the control device 29 before shipping from the injection molding machine manufacturer's factory. The injection molding machine information may also include the number of years of use of the servo motor 15 of the injection molding machine 11 or the rotation mechanism of the rotating part rotated by the servo motor 15.

[0028] Next, the standard control gain (a type of parameter) of the servo motor 15 that rotates the rotary table is set and input through the control gain input unit 55 of the setting input device 50 (s2). As described above, the control gains are typically a speed proportional gain, a speed integral gain, and a position loop gain, but other gains may also be provided. A time constant is also set as necessary. It is desirable that standard values ​​for these gains, etc., are stored in the control device 29 at the time of shipment from the injection molding machine manufacturer's factory.

[0029] Next, the weight of the mold attached for molding is input from molding condition input unit 53 (s3). Generally, the weight of the mold attached to turntable 14 varies depending on the molded product to be molded. Furthermore, the weight of the mold attached to turntable 14 is an important factor in rotating turntable 14 with servo motor 15 and stopping it at the correct target stopping position in the shortest possible time. However, when replacing and attaching a mold of exactly the same weight or a mold of approximately the same weight, the mold weight conditions input initially or previously may remain as they are, and it may even be possible to omit inputting a new mold weight.

[0030] Further, molding conditions are input from the molding condition input unit 53 (s4). The molding conditions correspond to parameters for controlling the rotation of the turntable 14. While the mold opening / closing speed and mold clamping pressure are input as molding conditions for the mold clamping unit 12, the speed setting of the servo motor 15 is important for shortening the cycle time and stopping the turntable 14 at an accurate position. Among these, when the speed of the turntable 14 increases from the start of rotation, a value that allows the servo motor 15 to fully utilize its capabilities within a range that can withstand continuous load is selected. Whether the speed is maintained for a while after reaching the maximum speed or the maximum speed is immediately decelerated, the most important parameters are the deceleration start position (or deceleration start time) and the acceleration during deceleration. These input values ​​are initially determined and input based on the molding conditions for the turntable rotation when the same or similar molds are used in a different injection molding machine, or the molding conditions for the turntable rotation when a similar mold is used in the same injection molding machine 11.

[0031] Next, the injection molding machine 11 is actually operated using parameters including the molding conditions initially set and input by manual molding or semi-automatic molding, and a verification is started using the machine learning device 38 to determine whether good stop control of the turntable 14 has been achieved. That is, a molding condition optimization loop and a control gain optimization loop are started (S5). During the verification, the servo motor 15 is first driven to perform rotation control and stop control of the turntable 14 (s6). With a mold attached to the rotating part (turntable 14) during this stop control, the load on the servo motor 15, reducer, etc., and the behavior of the turntable 14 during rotation are detected when the servo motor rotates and stops the rotating part, and these results are sent to the control device 29.

[0032] The determination of whether a good stop (stop within tolerance) has been achieved in the above is made by the determination unit 47 of the machine learning device 38. Indicators for this include: (1) whether the servo motor 15 has been accurately stopped at a commanded position within a predetermined time and within a predetermined error range without detecting any overshoot or vibration greater than the tolerance in the detection value of the rotary encoder 15b; (2) whether the current value (especially the peak torque) measured by the current detection unit 28a of the servo amplifier 28 is within a predetermined tolerance range for the servo motor 15; and (3) whether the acceleration during deceleration calculated using the detection value of the rotary encoder 15b is within the tolerance range of the drive mechanism including the deceleration mechanism.

[0033] If the rotation of the turntable 14 is performed using the combination of molding conditions and control gains and a satisfactory stop is achieved, there is no need to change the molding conditions or control gains, and the molding condition optimization loop and control gain optimization loop are terminated (S10). However, since the stop control of the turntable 14 is usually not optimal from the first rotation control, the condition correction unit 48 corrects the molding conditions (s8) and the control gains (s9). The correction of the molding conditions includes the rotation speed and deceleration start position of the turntable 14. The correction of the control gain also includes correction of the speed proportional gain, speed integral gain, and position loop gain. It also includes correction of the time constant, which is performed depending on the situation.

[0034] Then, the servo motor rotates the rotary table again (s6), and the results are verified again. At this time, the machine learning device 38 of the injection molding machine 11 performs rotation using many combinations of the molding conditions and control gains, and the judgment unit 47 judges whether or not the combination matches the conditions for a good stop (stop within the tolerance) stored in the pass / fail memory unit 45. If a good stop is next achieved, the combination of molding conditions and control gains that will result in a good stop (stop within the tolerance) in the shortest time is corrected.

[0035] These parameters are corrected in a predetermined order. In this embodiment, the molding conditions are corrected (s8) before the control gains are corrected (s9). The molding conditions, such as the rotation speed and deceleration start position of the turntable 14, are adjusted and the results are verified. The control gains are corrected when the satisfactory stopping control stored in the pass / fail memory unit 45 is no longer possible. This is because the rotation of the turntable 14 does not affect the quality of the molded product, and shortening the molding cycle time is prioritized above all else in terms of productivity. However, the order of parameter correction is not limited to this. For example, if the rotation time of the turntable 14 (depending on the combination of the maximum rotation speed and deceleration start position) is too short, resulting in unsatisfactory rotation and stopping control, and the control gains are corrected, but the rotation and stopping control cannot be improved, the rotation time of the turntable 14 is extended and the control gains are readjusted.

[0036] In this case, the control gains are adjusted by first adjusting the velocity proportional gain, followed by adjusting the velocity integral gain. Adjustment of the position loop gain is not essential, but is performed when problems such as poor response, overshoot, or vibration cannot be resolved by adjusting the velocity proportional gain and velocity integral gain alone. Adjustment of the time constant is also performed when problems cannot be resolved by adjusting the gains alone. For these adjustments, a conventional function for auto-tuning the control gain of a servo motor may be used or may be used in combination.

[0037] As described above, the control parameters vary widely, including the rotation speed of the rotary table driven by the servo motor (maximum rotation speed, deceleration start position, acceleration during deceleration) and control gain, in response to injection molding machine information including the mold weight. Therefore, rather than determining the optimal value for any one parameter, the optimal parameters are determined based on the correlation between multiple parameters. Reinforcement learning is used to determine these optimal parameters. For example, if increasing a certain parameter causes a problem such as overshoot or vibration, a negative reward is given for increasing the parameter. Conversely, if the problem such as overshoot or vibration is resolved, a positive reward is given. Furthermore, if increasing a certain parameter shortens the time required for the rotary table 14 to stop, a positive reward is given for increasing the parameter. Conversely, if the time required for the rotary table 14 to stop increases, a negative reward is given for increasing the parameter.

[0038] Naturally, the magnitude of the reward is also set, such that if the injection molding machine 11 exhibits behavior that results in a load abnormality, a large negative reward is given to the parameter that caused the abnormality. This is performed by the condition modification unit 48, and the parameters of the optimal solution are determined based on the results of these reward calculations. Deep learning may also be incorporated into reinforcement learning. Furthermore, analysis using mathematical programming may also be added in the process of determining the parameters of the optimal solution. Furthermore, when an optimal solution cannot be reached through adjustments, the weighting of the reward in reinforcement learning is generally modified, but the machine learning function formula may also be modified using a technique such as backpropagation.

[0039] The pass / fail determining unit 49 then determines parameters that will allow for satisfactory stopping of the rotary table. Once the parameters have been determined, the molding condition optimization loop and the control gain optimization loop are terminated (s10). The parameters obtained at this time are stored in the condition storage unit 46 of the machine learning device 38 and used as molding conditions for continuous operation. The parameter data that allows for satisfactory stopping control and that are stored in the condition storage unit 46 are stored as correct answer data (labeled data) for supervised learning and are also used as reference data when determining parameters the next time the mold is replaced. Similarly, data when satisfactory stopping cannot be achieved during parameter setting is also stored in the condition storage unit 46 as incorrect answer data and is also used as reference data when determining parameters the next time the mold is replaced.

[0040] The adjustment of the rotation control of the rotary table 14 by changing the parameters such as the control gain and molding conditions is preferably performed automatically by a machine learning device using reinforcement learning. However, some of the changes to the control gain and molding conditions, as well as confirmation of the results of rotating the rotary table 14 (stopping accuracy at the target stop position, load, time, etc.), may be performed by an operator. Furthermore, the adjustment by changing the parameters may be performed continuously during continuous operation, not just before continuous operation. For example, when continuous operation begins, the temperature of the mold and various parts of the injection molding machine 11 rises due to operation, and the parameters may be fine-tuned to accommodate this thermal expansion.

[0041] The supervised learning data (labeled data) stored in the condition storage unit 46 can also be used when a different mold is used. Alternatively, the data can be sent to an injection molding machine other than the injection molding machine 11 and used. In these cases, by inputting the variable portion of the injection molding machine information into the machine learning device 38, the machine learning device 38 can calculate a regression equation and narrow down the combinations of molding conditions and control gains to be tried from the beginning. In other words, by taking into account the regularity of the relationship between the mold weight and other factors from the accumulated data, combinations of molding conditions and control gains that result in inappropriate rotation and stop control of the rotary table 14 can be excluded from the beginning. Then, by starting a molding trial from a combination of molding conditions and control gains that the machine learning device determines to be optimal, the optimal combination of molding conditions and control gains can be quickly arrived at, even if trials are required.

[0042] <Rotary table stop position and rotary table positioning method> In the method for positioning the turntable of this embodiment, a mold is attached to the turntable, and the turntable is rotated and then stopped by a servo motor, and a target stop position for the turntable during rotation is determined using the position of the turntable when one mold is closed against another mold and at least a portion of the molds are in contact with each other, as shown in the flowchart of Fig. 6. The method of not using a stopper to control the stopping of the turntable 14 after rotation, as shown in the flowchart of Fig. 6, is particularly effective when using the machine learning device 38 to adjust parameters as described above.

[0043] Regarding the order of parameter adjustment by the machine learning device 38 described in the flowchart of Fig. 5 and determination of the target stop positions P1, P2 according to the flowchart of Fig. 6, it is desirable to determine the target stop positions P1, P2 first. After the target stop positions P1, P2 are determined, it is desirable to adjust the parameters during rotation control from the current stop position to the target stop positions P1, P2 after rotation. Furthermore, although the target stop positions P1, P2 differ depending on the mold, the difference is slight. Therefore, a method may be used in which parameters during rotation control are first determined based on rotation angles such as 90° rotation, 120° rotation, and 180° rotation, and then the rotation control is performed using the parameters to determine the target stop positions P1, P2.

[0044] Next, the control for stopping the rotary table at the target stop position will be specifically described with reference to the flowchart in FIG. 6 and the explanatory diagram of the operation in FIG. 7. When a new mold (a set of lower molds 13a, 13b and upper mold 23) is attached to the mold clamping device 12 of the injection molding machine 11, the absolute rotary encoder 15b is set to the origin so that the position where the cam member 31 is detected by the proximity sensor 30 is set to the origin (S1). Alternatively, the absolute rotary encoder 15b is set to the origin so that the position where the rotary table 14 is positioned by a positioning member, for example, by abutting a protrusion on the rotary table 14 against a stopper portion on the fixed platen 16, is set to the origin. Then, the target stop positions P1 and P2 of the rotary table 14 are set manually, while other molding details are confirmed. Note that the target stop positions P1 and P2 of the rotary table 14 may be set exclusively for this purpose, or may be set together with adjustments to other molding conditions during automatic continuous molding.

[0045] Next, the servo motor 15 for rotating the rotary table, which was servo-locked while the lower mold 13a was opening, is released, and the servo motor 15 begins to rotate (S2). At the same time, the rotary table 14 is rotated, and the lower mold 13a, which was at the removal position 35 and into which an insert has been inserted, is moved from the removal position 35 to the molding position 36 facing the upper mold 23. At the same time, the lower mold 13b, which holds the molded product, is moved from the molding position 36 to the removal position 35. The rotation command value sent from the control device 29 to the servo motor 15 when the rotary table 14 moves to the target stop position P0 is a pulse command value required to rotate the rotary table 14 180°, only for the first time after the mold change. The first time includes both the movement of the lower mold 13a to the molding position 36 and the movement of the lower mold 13b to the molding position 36. From the second time onwards, the target stop positions P1 and P2 are the positions corrected by the previous mold closing of the rotary table 14.

[0046] Next, when the rotary encoder 15b detects that the servo motor 15 has reached the target stop position P1 or P2 (S3=Y), the servo motor 15 is stopped and servo-locked (S4). At this time, the servo motor 15 performs position control using an absolute value and stops at a position that is controlled relative to the origin. Simultaneously, or around the same time, the proximity sensor 30 detects whether the turntable 14 is stopped at a position that allows mold closing (S5). The range that the proximity sensor 30 can detect for the cam member 31 is sufficient if the guide pin 24 of one mold can be inserted into the guide hole 25 during mold closing and the upper mold 23 can be closed relative to the lower mold 13a while the turntable 14 is being rotated with correction. In other words, even if the turntable 14 and the lower mold 13a are not stopped at the optimal molding position 36 designed for the mold clamping device 12 relative to the upper mold 23, the proximity sensor 30 or the like can detect that they are stopped at a position that allows mold closing.

[0047] If the proximity sensor 30 detects the cam member 31 (S5=Y), the servo motor 37 for mold opening / closing / mold clamping is activated to move the upper mold 23, which is a movable mold attached to the movable platen 20, toward the lower mold 13a attached to the rotary table 14, to start mold closing (S7). Alternatively, if the proximity sensor 30 does not detect anything even after a predetermined time has elapsed (S6=Y), an abnormality is determined to have occurred, and the injection molding machine 11 is stopped without transitioning to mold closing control. When mold closing is performed, the servo motor 15 that rotates the rotary table 14 remains in a servo-locked state. However, if there is a manufacturing error in the mold, or if the lower mold 13a or the like is not accurately attached to the mold attachment portion 14a or the like of the rotary table 14, a force other than the mold closing direction will be applied to the rotary table 14 when the guide pin 24 of the lower mold 13a is inserted into the guide hole 25 of the upper mold 23 during mold closing.

[0048] In this embodiment, the servo-locked servo motor 15 detects an increase in torque as it attempts to maintain its current position (S8). If the torque increases (S8=Y), the servo-lock of the servo motor 15 is switched to zero-speed control to correct the rotational position of the turntable 14 (S9). Zero-speed control is a control in which the servo motor 15 is in a pre-excited state where the speed is set to zero and the motor stops at its current position. However, if the turntable 14 receives an external force, the turntable 14 is moved from its current position to another position, and the servo motor 15 does not move to return the turntable 14 to its original position. Therefore, even if the turntable 14 is moved by an external force, the load on the servo motor 15 does not increase. Furthermore, in the present invention, the servo-locked state may be changed to a servo-free state when the turntable 14 receives an external force during mold closing. The servo-free state refers to a state in which the servo motor 15 is not excited, and the turntable 14, which is the controlled object, can move freely without being affected by the servo motor 15.

[0049] Furthermore, the servo motor 15 may be changed from a servo-locked state to zero speed control or servo-free by reading a change in the value (absolute value) of the rotary encoder 15b of the servo motor 15 that actually rotates the turntable 14. Furthermore, the servo-locked state may be released when the upper mold 23, which is the movable mold, reaches a predetermined mold closing position. In either case, since the upper mold 23, which is the movable mold, is not closed to the mold closing completion position while the servo motor 15 is in the servo-locked state, galling of the guide pin 24 and other defects are prevented or suppressed when the molds are closed together.

[0050] Next, regardless of whether the torque of the servo motor 15 has increased or not, when the upper mold 23, which is the movable mold relative to the lower mold 13a, reaches the mold closing completion position (S10=Y), the position (absolute value) of the turntable 14 relative to the control origin position of the servo motor 15 is detected by the rotary encoder 15b of the servo motor 15. In the present invention, the position of the turntable 14 at the time of mold closing completion is regarded as the target stop position P1 to be used for subsequent rotation control of the turntable 14, and the previous target stop position P0 is corrected and stored in the control device 29 (S11).

[0051] The target stop position P1 can be corrected and determined by detecting the position of the rotary table 14, which is the rotating part, when at least a portion of the molds are in contact with each other. In other words, the position of the rotary table 14, which is the rotating part, after it has been moved without control of the servo motor 15 during mold closing is detected, and the target stop positions P1 and P2 during rotation of the rotary part are determined. Specifically, the target stop position P1 may be corrected when the guide pin 24 and the guide hole 25 come into contact and the rotary table 14 moves, even if mold closing is not yet complete. Alternatively, the target stop position P1 may be corrected when the convex and concave portions of the mold come into contact and the rotary table 14 moves. Furthermore, if the rotary table 14 does not move during mold closing, the position of the rotary table 14 at the mold closing completion position becomes the target stop position P1. Furthermore, if the rotary table 14 moves multiple times during mold closing, it is preferable to set the position after the first movement as the target stop position P1. Furthermore, the actual correction of the target stop position P1 (or P2) in the control device 29 may be performed not only when the turntable 14 moves, but also until the next time the turntable 14 starts to rotate. Furthermore, the target stop positions P1, P2 during rotation of the rotating part may be determined using the position of the turntable 14, which is the rotating part, in a state where at least parts of the molds are in contact with each other after multiple mold closings.

[0052] 6, at the same time as or around the time when the target stop position P1 of the servo motor 15 is corrected at the mold closing completion position, if the servo motor 15 is under zero speed control, the mold clamping process is carried out in a servo-free state (S12). When the mold clamping process is completed, the servo motor 15 is locked again, and the mold opening / closing / mold clamping servo motor 37 is operated to open the movable platen 20 and the upper mold 23. When the upper mold 23 reaches the mold opening completion position, the servo lock is released.

[0053] Next, the lower mold 13a is moved from the molding position 36 toward the removal position 35. In the flowchart of FIG. 6, this series of control is indicated as (S12). When the lower mold 13a is stopped at the removal position 35, the servo motor 15 is servo-locked, and the molded product is removed and the insert is inserted. In parallel with the movement of the lower mold 13a from the molding position 36 toward the removal position 35, the lower mold 13b is moved from the removal position 35 to a target stop position P2 at the molding position 36. In this embodiment, the turntable 14 does not rotate continuously, and the movement of the lower mold 13a from the molding position 36 to the removal position 35 is reversed (inverted) from the movement from the removal position 35 to the molding position 36. In other words, the direction of rotation when the lower mold 13a is moved from the removal position 35 to the molding position 36 is different from the direction of rotation when the lower mold 13b is moved from the removal position 35 to the molding position 36. In the case of continuous molding, the molding cycle continues until the planned number of molded pieces is completed (S13).

[0054] In the present invention, the target stop positions P1 and P2 of the turntable 14 are determined according to the actual states of the lower molds 13a and 13b and the upper mold 23. Therefore, the target stop position P1 (including the rotation angle of the turntable 14) when the lower mold 13a is moved from the removal position 35 to the molding position 36 is often different from the target stop position P2 (including the rotation angle of the turntable 14) when the lower mold 13b is moved from the removal position 35 to the molding position 36.

[0055] Therefore, the stopping position of lower die 13b or lower die 13a at removal position 35 will differ depending on the target stopping positions P1, P2 of lower die 13a or lower die 13b at molding position 36. However, slight differences in the stopping positions of the take-out machine that holds the molded product and the insert machine that inserts the insert can be accommodated without changing the control. Alternatively, in the case of high-precision molding where different stopping positions of lower die 13a and lower die 13b during removal will have an impact, the robot can be programmed so that the take-out machine and insert machine are operated under different controls for lower die 13a and lower die 13b.

[0056] In this embodiment, the rotation of the turntable 14 and the mold closing are performed without using a positioning member, so the positioning member can be omitted. Furthermore, the molding cycle time required for inserting and removing the positioning member can be shortened. Furthermore, when the mold is closed with the positioning member still inserted, if a strong force is generated in the direction of rotation of the turntable 14, there is a risk that the positioning member will become stuck in the positioning hole, but this problem does not occur.

[0057] The above description of this embodiment follows a molding cycle. While the next target stop position may be determined each time based on the mold closing position at which the molds at least abutted against each other the previous time, in practice, the target stop position is often determined as follows: Backlash exists in the transmission mechanism that connects the driving force of the servo motor 15 for rotating the turntable 14 to the rotation of the turntable 14. Due to various disturbances, even if the same command pulse is given to the servo motor 15 from the origin, there is a slight error in the actual stop position of the turntable 14. Errors due to backlash and other factors also vary depending on the weight of the lower molds 13a and 13b attached to the turntable 14. Furthermore, if the amount of movement of the turntable 14 during mold closing is extremely small (below a certain amount), the presence of backlash may or may not cause the movement of the turntable 14 to be transmitted to the rotation shaft of the servo motor 15.

[0058] Therefore, in the present invention, the position of the turntable 14 when it contacts the mold during mold closing after rotating the turntable through multiple molding cycles is detected by the rotary encoder 15b, and the data is transmitted to and stored in the control device 29. The control device 29 then calculates the center value or optimal value of the target stop positions P1 and P2 of the turntable 14 from the data. More specifically, the position (absolute position from the origin) after the turntable 14 is rotated and corrected when the lower mold 13a is moved from the removal position 35 to the molding position 36 and closed with the upper mold 23 is detected as the position after the rotation correction through multiple molding cycles. The center value or optimal value is then calculated from the detected value and used as the target stop position P1 of the lower mold 13a for subsequent control of the servo motor 15. Separately, the position (absolute position from the origin) after the turntable 14 is rotated and corrected when the lower mold 13b is moved from the removal position 35 to the molding position 36 and closed with the upper mold 23 is also used as the target stop position P2 of the lower mold 13b for subsequent control of the servo motor 15.

[0059] During continuous molding, control is performed to stop the lower dies 13a, 13b at the target stop positions P1, P2 at the molding position 36 calculated by the above calculation. Note that even during continuous molding, in response to the above-mentioned backlash problem, thermal expansion of the dies, and other disturbances, the target stop positions P1, P2 of the lower dies 13a and 13b at the molding position 36 may be detected at least once during multiple molding cycles, and the target stop positions P1, P2 may be corrected based on the detected values.

[0060] Furthermore, in the present invention, a machine learning device 38 that performs machine learning to learn the position of the rotating part when at least a portion of the molds are in contact with each other as an index for determining target stop positions P1, P2 during rotation of the rotary table 14, which is the rotating part, may be provided in the control device 29 of the injection molding machine 11 or in a control device (not shown) other than the injection molding machine 11, and the control device 29 of the injection molding machine 11 and the control device other than the injection molding machine 11 may be connected. The machine learning device 38 is equipped with a neural network, including one that performs deep learning, and performs supervised learning that can determine whether a detected value is good or bad. However, reinforcement learning or unsupervised learning based on the learning results may also be performed.

[0061] In a device equipped with a machine learning device 38, such as the control device 29, the amount of movement of the turntable 14 after the lower molds 13a, 13b stop at the target stop positions P1, P2 is verified. A move less than a predetermined distance is judged as good, whereas a move greater than the predetermined distance is judged as bad. If the number of bad judgments exceeds a certain number during continuous molding, the target stop position is reviewed. Alternatively, the maximum rotation speed, deceleration start position, and deceleration rate of the turntable 14 driven by the servo motor 15 are reviewed, and a pass / fail judgment is made again to determine whether the amount of movement of the turntable 14 after stopping at the target stop positions P1, P2 is less than the predetermined distance.

[0062] As mentioned above, the present invention does not exclude the use of a positioning member, but if a positioning member is present, it may be operated at least once during multiple molding cycles to correct the origin position, etc. This makes it possible to eliminate or suppress any discrepancy that occurs between the actual position of the turntable 14 and the position detected by the rotary encoder 15b during continuous molding due to backlash, etc. When a positioning member is used, it is possible to consider a method in which the turntable 14 is positioned by the positioning member before the mold is closed and then the positioning member is removed, or a method in which the mold is closed with the positioning member inserted.

[0063] In the present invention, two to eight molds may be attached to the turntable 14, and the turntable 14 may always rotate in the same direction. Although not shown, as an example, in the case of four lower molds, 13a, 13b, 13c, and 13d, the amount of movement of the lower mold 13a when it moves from the previous stage to the position at the molding position after the molds are closed (the position after the turntable is corrected), the amount of movement of the lower mold 13b when it moves from the previous stage to the position at the molding position after the molds are closed, the amount of movement of the lower mold 13c when it moves from the previous stage to the position at the molding position after the molds are closed, and the amount of movement of the lower mold 13d when it moves from the previous stage to the position at the molding position after the molds are closed are detected by the rotary encoder 15b, stored in the control device 29, and similarly used for controlling the rotation of the turntable 14 thereafter. Generally, the amount of movement to the target stop position differs for each of the lower molds 13a, 13b, 13c, and 13d.

[0064] In such a case where the mold is rotated forward in one direction, the sum of the movement amounts (pulses) of each mold when it is moved from the previous stage to the post-mold-closed position at the molding position is, in principle, the movement amount (pulses) required for the turntable to rotate 360°. Also, in a case where the mold is rotated forward in one direction, the number of pulses corresponding to the rotation angle of the turntable 14 from the target stop position to the post-mold-closed position (the position after the turntable correction) is detected, and this can be used as a command value for the servo motor 15 from the next time onwards. Therefore, an incremental type rotary encoder 15b can be used. However, even in the case of forward rotation, as the table rotates repeatedly during continuous molding, a discrepancy may occur between the position detected by the rotary encoder 15b of the servo motor 15 and the actual position of the turntable 14. In such a case, it is desirable to correct the target stop position again during the continuous molding.

[0065] Next, a vertical injection molding machine 81 of another embodiment shown in Figure 8 will be described. In the mold clamping device 82 of the injection molding machine 81, each tie bar 84 is fixed vertically to a fixed platen 83, which is the lower platen, and the top of each tie bar 84 is fixed to a pressure platen 85, which is the upper platen. Also, guided by the tie bars 84, a movable platen 86 is provided so as to be able to move up and down by two mold opening and closing mechanisms 98 (only one is shown in Figure 8, with the mechanisms omitted). Also, a mold clamping mechanism 97 such as a mold clamping cylinder is provided on the pressure platen 85, and a ram 88 of the mold clamping cylinder is fixed to the back surface of the movable platen 86.

[0066] The injection molding machine 81 of this embodiment is an injection molding machine 81 for multicolor molded products, and a rotary table 87, which is a rotating part, is provided on the underside of a movable platen 86 so as to be rotatable relative to the movable platen 86. The rotary table 87 is rotated via a belt 99 by a servo motor 89 for table rotation, which is fixed to the side of the movable platen 86. Two movable molds 90 are attached to the underside of the rotary table 87. Furthermore, two fixed molds 91 are attached to the upper surface of the fixed platen 83.

[0067] The rotation angle of the rotary table 87 can be detected by an encoder of the servo motor 89. Also, a device such as a proximity switch (not shown) is provided to confirm the stopped state of the rotary table 87 when rotation has stopped. The rotary table 87 is provided with positioning holes 92 for a positioning mechanism. Meanwhile, the movable platen 86 is provided with an actuator such as a hydraulic cylinder 94 for inserting positioning pins 93 of a positioning member into the positioning holes 92. Also, two injection units 95 and 96 are arranged on the side of the mold clamping unit 82.

[0068] The control method and machine learning method of the injection molding machine 81 shown in FIG. 8 are common to the injection molding machine 11 of the embodiment shown in FIGS. 1 to 7 in terms of the basic technical concept of the invention, but the following description will focus on the differences. In the injection molding machine 81 shown in FIG. 8, the origin of the servo motor 89 is set with the positioning pin 93 inserted into the positioning hole 92. The positioning pin 93 may also be inserted once every several moldings to perform origin correction. In the injection molding machine 81, two movable molds 90 are simultaneously closed against two fixed molds 91. Therefore, the behavior of the rotary table 87 becomes complex when there is an error in the mold itself or in the mold mounting position. Therefore, the stopping position of the rotary table 87 from the next mold closing onward may be the position after the greatest movement of the rotary table 87 during mold closing (between mold contact and mold closing completion) rather than the position at which mold closing is completed.

[0069] Next, a horizontal injection molding machine 101 according to yet another embodiment shown in Figure 9 will be described. In the injection molding machine 101, a mold clamping unit 103 is disposed on the upper surface of a bed 102. In the mold clamping unit 103, an intermediate platen 109 is provided between a fixed mold 105, which is a first mold, attached to a fixed platen 104, which is a first platen, and a movable mold 107, which is a second mold, attached to a movable platen 106, which is a second platen, and the intermediate platen 109 is rotatable about an axis member 108 perpendicular to the mold opening / closing direction. Intermediate molds 110, 110 that are mated with both the fixed mold 105 and the movable mold 107 are disposed on both sides of the intermediate platen 109.

[0070] Clamping cylinders 111 of the mold clamping means are provided near the four corners of the fixed platen 104, and the rods of the clamping cylinders 111 form tie bars 112. Each tie bar 112 is inserted through both a movable platen 106 that is movable in the mold opening / closing direction relative to the fixed platen 104, and an intermediate member 113 that is located between the fixed platen 104 and the movable platen 106 and is movable in the mold opening / closing direction. The movable platen 106 is opened and closed by a mold opening / closing mechanism 114. The intermediate member 113 and the intermediate platen 109 are opened and closed by a separate mold opening / closing mechanism 115. Note that the mold opening / closing mechanisms 114 and 115 are shown schematically in Figure 9, and the locations and number of mechanisms to be attached are not limited.

[0071] An intermediate platen 109 is fixed to a shaft member 108 that is rotatably attached to the intermediate member 113. A servo motor 116 is provided on one of the upper and lower intermediate members 113, and the drive shaft of the servo motor 116 is connected to the shaft member 108 by a belt or gear (including a reducer). Therefore, the intermediate platen 109, which is a rotating part, is rotated by the drive of the servo motor 116.

[0072] Furthermore, a positioning mechanism 117 is provided on at least one of the upper and lower intermediate members 113. The positioning mechanism 117 has a positioning member 119 that moves forward and backward by an actuator 118. Furthermore, a positioning hole 120 into which the positioning member 119 is inserted is provided on the upper or lower surface of the intermediate platen 109. When the positioning member 119 is inserted into the positioning hole 120, the intermediate platen 109 and the intermediate mold 110 are positioned (the rotation angle is adjusted) so that they directly face the fixed platen 104 and the movable platen 106. Note that the positioning mechanism 117 is not essential.

[0073] In the injection molding machine 101, half nuts 121 are attached to the movable platen 106, and a groove into which the half nuts 121 are engaged is formed in the tie bar 112. A first injection unit 122 is disposed on the outside of the fixed platen 104 (the opposite side to the mold mounting surface), and a second injection unit 123 is disposed on the outside of the movable platen 106 (the opposite side to the mold mounting surface).

[0074] The control method for the injection molding machine 101 shown in Fig. 9 is almost the same as that for the injection molding machines 11 and 81 of the embodiments shown in Figs. 1 to 8, but differences will be mainly described below. In the injection molding machine 101 shown in Fig. 9, the intermediate platen 109, which is a rotating part, is rotated by the servo motor 116 toward a target stop position, and then the servo motor 116 is locked when it is stopped at the target stop position. Then, the mold opening and closing mechanisms 114 and 115 close the fixed mold 105 and the intermediate mold 110, and the intermediate mold 110 and the movable mold 107. At this time, the servo motor 116 remains locked at least until the molds come into contact during the mold closing process.

[0075] Then, when the molds come into contact with each other, the rotation angle of the intermediate platen 109 is changed based on the relationship between the guide pins and guide holes, and the position (angle) of the intermediate platen 109 is corrected so that it becomes the target stop position for the next and subsequent stops. In the case of the injection molding machine 101 shown in Figure 9, even if the guide pins or mold members come into strong contact between one of the intermediate molds 110 and the fixed mold 105 or the movable mold 107, causing the intermediate platen 109 to rotate, once the molds are completely closed after that, the intermediate platen 109 and the intermediate mold 110 are often sandwiched between the fixed mold 105 and the movable mold 107, and the intermediate platen 109 is returned to a position directly opposite the fixed platen 104. Therefore, it is desirable to set the position to which the intermediate platen 109 is moved during mold closing (from the start of mold abutment to the completion of mold closing) as the target stop position for the next and subsequent stops.

[0076] Although not listed individually, the present invention is not limited to the above-described embodiments, and it goes without saying that the present invention can be applied to modifications made by a person skilled in the art based on the spirit of the present invention, or to individual combinations of parts of the above-described embodiments. The injection molding machine 11 of the first embodiment and the injection molding machine 81 of the second embodiment may be of a type that opens and closes the mold horizontally.

[0077] The method for controlling an injection molding machine of the present invention, in which a machine learning device verifies parameters for rotating and stopping a rotating part using a servo motor with a mold attached to the rotating part, and determines the parameters, may be a method in which the rotating part, such as a rotary table, is always rotated by a fixed angle. For example, the rotary table may be rotated by a predetermined angle such as 90°, 120°, or 180°, and then moved depending on the state of the mold.

[0078] Furthermore, the control method of the injection molding machine of the present invention, in which the machine learning device 38 verifies and determines parameters for rotating and stopping the rotating part by the servo motor 15 with a mold attached to the rotating part, may be a method in which the rotating part, i.e., the rotary table, is rotated and then stopped at a stopping position by a stopper or the like. In this case, there is no need to pursue precision in stopping the rotary table at a stopping position or to worry about overshooting, but there are issues such as shortening the molding cycle time, the load on the servo motor, and control to stop the stopper at a predetermined speed or below, and therefore it is still important to use the machine learning device to generate molding conditions and control gain values ​​for optimally rotating the rotary table, etc.

[0079] In addition, in the above embodiment, the method of finding an optimal solution for the rotation control and stop control of the rotary table 14, which is a rotating part, by the machine learning device 38 can also be applied to the movement control and stop control of parts other than the rotating part. Specifically, with regard to the mold opening and closing movement of the movable platen, it is desirable to change the optimal molding conditions and servo motor control gains by changing the mold weight of the movable mold or the mold opening and closing speed, and these can be determined by a method using a machine learning device similar to that of this embodiment. [Explanation of symbols]

[0080] 11 Injection molding machine 12 Mold clamping device 13a, 13b Lower mold (mold) 14 Rotating table (rotating part) 15 Servo motor 23 Upper mold (mold) 29 Control Device 38 Machine Learning Device 47 Judgment Department 48 Condition correction section 49 Acceptance / Rejection Decision Section 50 Setting input device P1,P2 Target stop position

Claims

1. In an injection molding machine in which a rotating part to which a mold is attached is rotatable by a servo motor, a control device that verifies, with a machine learning device, parameters for rotating and stopping the rotating part by the servo motor with a mold attached to the rotating part during a plurality of molding cycles, and a result of detecting, with a rotary encoder of the servo motor, a position to which the rotating part is again moved in the rotation direction due to an external force after the rotation of the rotating part has stopped, and determines parameters including parameters for subsequently stopping the rotating part by the servo motor at a rotation stop position; Injection molding machine.

2. 2. The injection molding machine according to claim 1, wherein no mechanical stopper is provided for stopping the rotating portion.

3. In an injection molding machine in which a rotating part to which a mold is attached is rotatable by a servo motor, a machine learning device verifies parameters when the servo motor rotates and stops the rotating part with a mold attached to the rotating part during a plurality of molding cycles, and the load amount, rotation speed, and rotation stop position of the servo motor detected using a servo amplifier of the servo motor and a rotary encoder of the servo motor when the servo motor rotates and stops the rotating part; a control device that corrects parameters including a rotation speed of the rotating part caused by the servo motor and a parameter for stopping the rotation at a rotation stop position; Injection molding machine.

4. In a control system for an injection molding machine in which a rotating part to which a mold is attached is rotatable by a servo motor, a control device that verifies, with a machine learning device, parameters for rotating and stopping the rotating part by the servo motor with a mold attached to the rotating part during a plurality of molding cycles and a result of detecting, with a rotary encoder of the servo motor, a position to which the rotating part has been moved again in the rotation direction without being controlled by the servo motor after the rotation of the rotating part has stopped, and determines parameters including parameters for subsequently stopping the rotating part by the servo motor at a rotation stop position; Injection molding machine control system.

5. 5. The control system for an injection molding machine according to claim 4, wherein the parameters include at least a control gain of the servo motor.

6. A control method for an injection molding machine in which a rotating part to which a mold is attached is rotatable by a servo motor, comprising: a machine learning device verifies parameters when the servo motor rotates and stops the rotating part with a mold attached to the rotating part during a plurality of molding cycles, and the load amount, rotation speed, and rotation stop position of the servo motor detected using a servo amplifier of the servo motor and a rotary encoder of the servo motor when the servo motor rotates and stops the rotating part; correcting parameters including a rotation speed of the rotating part caused by the servo motor and a parameter for stopping the rotation at a rotation stop position; How to control an injection molding machine.

Citation Information

Patent Citations

  • Upright rotary injection-molding machine

    JP1998000646A

  • Controller using acceleration sensor and adjusting device

    JP2007141189A

  • Injection molding machine

    JP2008055625A

  • Method for control of molding machine, and molding machine

    JP2011235523A

  • Die rotating type injection molding machine, method for rotating rotary table for the same

    JP2012206398A