Injection molding device

Abstract

To provide an injection molding apparatus capable of grasping a wear state of a gate without relying on knowledge and experience of an expert.SOLUTION: An injection molding apparatus 1 includes: an injection device 20 comprising a cylinder 22, a screw 23, and a nozzle 24 provided at a front end of the cylinder 22 for discharging a molten resin as the screw 23 advances; a mold 30 having a molded product cavity C and a resin flow path P between the molded product cavity C and a portion that abuts on the nozzle 24; a channel pressure measuring device 33 for acquiring flow path pressure data in the resin flow path P; a screw pressure measuring device 27 for acquiring screw pressure data that the screw 23 receives from the molten resin in the cylinder 22; and an output unit 61 for outputting a value representing a wear state of a gate P3 of the resin flow path P based on a feature amount of the flow path pressure data and a feature amount of the screw pressure data.SELECTED DRAWING: Figure 7

Description

[Technical field]

[0001] The present invention relates to an injection molding apparatus. [Background technology]

[0002] Patent Document 1 describes a method in which sensors are installed in an injection molding device and a mold, and the quality of a molded product is estimated by machine learning using detection data from the sensors. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-49929 A Summary of the Invention [Problem to be solved by the invention]

[0004] A mold used in injection molding has a molded product cavity as a portion for molding a molded product, and a resin flow path between the molded product cavity and a portion that abuts against the nozzle of an injection device. The resin flow path includes a spool (also called a sprue), a runner, and a gate.

[0005] When molding cycles for molding a molded product are performed multiple times in succession, the resin flow path wears out. In particular, the gate of the resin flow path is more susceptible to wear than other parts because the cross-sectional area of ​​the flow path is smaller. When the gate wears out, the pressure in the molded product cavity increases, which may affect the dimensional accuracy of the molded product. Therefore, when the wear of the gate becomes too great, the gate part of the mold is replaced.

[0006] The more times a molding cycle is performed (number of shots), the greater the amount of wear on the gate. However, the amount of wear on the gate changes depending on the state of the molten resin flowing through the gate. The state of the molten resin changes depending on the type of resin and the moisture content of the resin, etc. Therefore, the timing to replace the gate part cannot be determined solely by the number of molding cycles performed, but has had to be determined based on the knowledge and experience of an expert. Therefore, in order to determine the timing to replace the gate part, there is a need to understand the wear condition of the gate without relying on the knowledge and experience of an expert.

[0007] The present invention has been made in view of the above background, and aims to provide an injection molding apparatus that makes it possible to grasp the wear state of a gate without relying on the knowledge or experience of a skilled operator. [Means for solving the problem]

[0008] One aspect of the present invention is an injection device including a cylinder, a screw, and a nozzle provided at a front end of the cylinder for ejecting molten resin as the screw advances; a mold including a molded product cavity and a resin flow path between the molded product cavity and a portion that contacts the nozzle; a flow channel pressure measuring device for acquiring flow channel pressure data in the resin flow channel; A screw pressure measuring device for acquiring screw pressure data that the screw receives from the molten resin in the cylinder; an output unit that outputs a value representing a wear state of a gate of the resin flow path based on a feature amount of the flow path pressure data and a feature amount of the screw pressure data; The present invention relates to an injection molding apparatus comprising: Effect of the Invention

[0009] The flow path pressure data is pressure data in the resin flow path of the mold, so it is affected by gate wear and also by the state of the molten resin flowing. On the other hand, the screw pressure data is far from the resin flow path of the mold, so it is less affected by gate wear, but is affected by the state of the molten resin in the cylinder. Here, the state of the molten resin in the cylinder and the state of the molten resin in the mold are similar.

[0010] The output unit uses the feature amount of the flow path pressure data and the feature amount of the screw pressure data. The feature amount of the flow path pressure data is a value influenced by gate wear and the state of the molten resin, and the feature amount of the screw pressure data is a value influenced by the state of the molten resin. If the influence of the state of the molten resin can be removed from the feature amount of the flow path pressure data, the influence of the gate wear can be extracted.

[0011] Therefore, the output unit performs processing using the feature amount of the flow path pressure data and the feature amount of the screw pressure data, thereby extracting the influence of gate wear and outputting a value representing the gate wear state. Since the worker can grasp the gate wear state based on the value representing the gate wear state, the worker can appropriately determine the timing of replacing the gate portion without relying on the knowledge and experience of an expert. [Brief description of the drawings]

[0012] [Figure 1] FIG. 2 is a diagram showing a mechanical configuration of an injection molding device. [Diagram 2] FIG. 2 is an enlarged cross-sectional view of the mold in FIG. [Diagram 3] FIG. 3 is a diagram showing only the spatial portion of the mold (the molded product cavity and the resin flow path) as viewed from the axial direction of the spool (as viewed from the right in FIG. 2). [Figure 4] 1 is a flowchart illustrating an injection molding method. [Diagram 5]1 is a graph showing the change in screw pressure data over time in an injection process and a pressure holding process, in which the solid line indicates the case without gate wear and the dashed line indicates the case with gate wear. [Figure 6] 1 is a graph showing a change in flow channel pressure data over time during an injection process and a pressure holding process, in which the solid line indicates a case where there is no gate wear, and the dashed line indicates a case where there is gate wear. [Figure 7] FIG. 1 is a functional block diagram showing an injection molding apparatus according to a first embodiment. [Figure 8] FIG. 2 is a functional block diagram showing the processing of a gate wear calculation computer device constituting the injection molding apparatus of the first embodiment. [Figure 9] FIG. 11 is a functional block diagram showing the processing of a gate wear calculation computer device constituting the injection molding apparatus of the second embodiment. [Figure 10] 1 is a graph showing the relationship between the number of molding cycles (number of shots) and the gate wear index value. [Figure 11] 13 is a flowchart showing the processing of an output unit of a gate wear calculation computer device constituting an injection molding apparatus of a third embodiment. [Figure 12] 13 is a flowchart showing the processing of an output unit of a gate wear calculation computer device constituting the injection molding apparatus of the fourth embodiment. [Figure 13] FIG. 13 is a functional block diagram showing an injection molding apparatus according to a fifth embodiment. [Figure 14] FIG. 13 is a functional block diagram showing the processing of a gate wear calculation computer device constituting the injection molding apparatus of the fifth embodiment. [Figure 15] FIG. 13 is a functional block diagram showing the processing of a gate wear calculation computer device constituting the injection molding apparatus of the sixth embodiment. [Figure 16] FIG. 13 is a functional block diagram showing the processing of a gate wear calculation computer device constituting the injection molding apparatus of the seventh embodiment. [Figure 17] FIG. 13 is a functional block diagram showing the processing of a gate wear calculation computer device constituting the injection molding apparatus of the eighth embodiment. [Figure 18]13 is a flowchart showing the processing of an output unit of a gate wear calculation computer device constituting the injection molding apparatus of the ninth embodiment. [Figure 19] 1 is a graph showing the change in screw pressure data over time in an injection process and a pressure holding process, in which the thick line indicates the case where the resin viscosity is high, and the thin line indicates the case where the resin viscosity is low. [Figure 20] 20 is an enlarged graph showing the periods from time t1 to t2a and from time t1 to t2b in FIG. 19, in which the thick line indicates the case where the resin viscosity is high, and the thin line indicates the case where the resin viscosity is low. [Figure 21] 1 is a graph showing a change in flow channel pressure data over time during an injection process and a pressure holding process, in which a thick line indicates a case where the resin viscosity is high, and a thin line indicates a case where the resin viscosity is low. [Figure 22] 22 is an enlarged graph showing the periods from time t1 to t2a and from time t1 to t2b in FIG. 21, in which the thick line indicates the case where the resin viscosity is high, and the thin line indicates the case where the resin viscosity is low. [Figure 23] 13 is a graph showing the change in nozzle pressure data over time in a purging process, showing the case where the resin viscosity is low. [Figure 24] 13 is a graph showing the change over time in nozzle pressure data during a purging process, in the case where the resin viscosity is high. [Diagram 25] 1 is a graph showing the change in screw position over time in a metering process. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] (1. Configuration of Injection Molding Apparatus 1) An injection molding apparatus 1 will be described with reference to Fig. 1. The injection molding apparatus 1 is an apparatus for molding a resin molded product using a mold 30. The injection molding apparatus 1 mainly comprises a bed 10, an injection unit 20, a mold 30, a mold clamping unit 40, a control unit 50, and a computer device 60 for calculating gate wear. The bed 10 is a member that is installed on an installation surface.

[0014] The injection device 20 is disposed on the bed 10. The injection device 20 is a device that melts resin, which is a molding material, and applies pressure to the molten resin to supply it to a molded product cavity C of a mold 30. The injection device 20 includes a hopper 21, a cylinder 22, a screw 23, a nozzle 24, a heater 25, a drive device 26, a screw pressure measuring device 27, and a nozzle pressure measuring device 28.

[0015] The hopper 21 is an inlet for resin pellets (granular molding material), which are the raw material for the molding material. The cylinder 22 stores molten resin produced by heating and melting the pellets fed into the hopper 21. The cylinder 22 is provided so as to be movable in the axial direction of the cylinder 22 relative to the bed 10. The screw 23 is disposed inside the cylinder 22 and is provided so as to be rotatable and movable in the axial direction. The nozzle 24 is an outlet provided at the front end of the cylinder 22, and discharges the molten resin inside the cylinder 22 as the screw 23 advances.

[0016] The heater 25 is provided, for example, on the outer circumferential surface of the cylinder 22 or embedded inside the cylinder 22, and heats the resin inside the cylinder 22. That is, the heater 25 melts the pellets and keeps the molten resin in a molten state. The drive device 26 moves the cylinder 22 in the axial direction (forward and backward), rotates the screw 23, and moves it in the axial direction (forward and backward).

[0017] The screw pressure measuring device 27 is provided, for example, near the base end of the screw 23, and acquires pressure data that the screw 23 receives from the molten resin in the cylinder 22 (hereinafter referred to as "screw pressure data").

[0018] The nozzle pressure measuring device 28 is provided in the nozzle 24, and acquires pressure data (hereinafter referred to as "nozzle pressure data") that the nozzle 24 receives from the molten resin when the molten resin flows through the nozzle 24. In addition to the screw pressure measuring device 27 and the nozzle pressure measuring device 28, the injection device 20 also includes sensors that acquire the position of the cylinder 22, the position of the screw 23, the moving speed of the screw 23, the temperature of the heater 25, the state of the drive device 26, and the like.

[0019] The mold 30 includes a first mold 31 which is a fixed side, and a second mold 32 which is a movable side. The mold 30 forms a molded product cavity C between the first mold 31 and the second mold 32 by clamping the first mold 31 and the second mold 32. The first mold 31 and the second mold 32 include a resin flow path P between the molded product cavity C and a portion that abuts against the nozzle 24 of the injection device 20. The resin flow path P is a flow path (spool, runner, gate) that guides the molten material supplied from the nozzle 24 of the injection device 20 to the molded product cavity C.

[0020] Furthermore, the mold 30 includes a flow path pressure measuring device 33 that acquires pressure data (hereinafter referred to as “flow path pressure data”) in the resin flow path P. The flow path pressure data is pressure data that the inner wall surface of the resin flow path P receives from the molten resin flowing through the resin flow path P.

[0021] The mold clamping device 40 is disposed opposite the injection device 20 on the bed 10. The mold clamping device 40 opens and closes the attached mold 30, and prevents the mold 30 from opening due to the pressure of the molten material injected into the molded product cavity C when the mold 30 is clamped.

[0022] The mold clamping unit 40 includes a fixed platen 41, a movable platen 42, a diver 43, a drive unit 44, and a measurement device for the mold clamping unit 45. The first mold 31 is fixed to the fixed platen 41. The second mold 32 is fixed to the movable platen 42. The movable platen 42 can move toward and away from the fixed platen 41. The diver 43 supports the movement of the movable platen 42. The drive unit 44 is, for example, configured by a cylinder device, and moves the movable platen 42. The measurement device for the mold clamping unit 45 acquires the mold clamping force, mold temperature, the state of the drive unit 44, and the like.

[0023] The control device 50 controls the drive device 26 of the injection device 20 and the drive device 44 of the mold clamping device 40. The gate wear calculation computer device 60 performs a calculation process of a value representing the wear state of the gate P3 (shown in FIG. 2) in the resin flow path P of the mold 30. The gate wear calculation computer device 60 is composed of an arithmetic device and a storage device, and processes by executing a computer program. The gate wear calculation computer device 60 calculates a value representing the wear state of the gate using, for example, control data from the control device 50, pressure data acquired by the screw pressure measuring device 27, the nozzle pressure measuring device 28, the flow path pressure measuring device 33, etc.

[0024] (2. Detailed configuration of mold 30) The detailed configuration of the mold 30 will be described with reference to Figs. 2 and 3. The mold 30 has a molded product cavity C for molding a molded product. The molded product cavity C is formed by a first die 31 and a second die 32. In this embodiment, the molded product cavity C is formed, for example, in a circular ring shape, but it can be formed in any shape such as a C-shape or a U-shape. The molded product cavity C may be formed in only one location, or may be formed in multiple locations. In Figs. 2 and 3, one molded product cavity C is illustrated for ease of explanation.

[0025] Furthermore, the mold 30 has a resin flow path P that connects between the portion that contacts the nozzle 24 (shown in FIG. 1) and the molded product cavity C. The resin flow path P includes a spool P1 (also called a sprue), a runner P2, and a gate P3. The spool P1 is a passage through which the molten material is introduced from the nozzle 24. The spool P1 is formed, for example, in a straight line from the portion that contacts the nozzle 24.

[0026] The runner P2 is a flow path formed at an angle from the spool P1. In other words, the molten resin introduced into the spool P1 flows into the runner P2. For example, as shown in Fig. 3, in this embodiment, a plurality of runners P2 are formed branching out radially from the spool P1 toward one molded product cavity C. Note that even when a plurality of molded product cavities C are formed, a plurality of runners P2 are formed branching out from the spool P1 toward each of the plurality of molded product cavities C.

[0027] The gate P3 is located at the tip of the runner P2, and is a flow path that guides the molten resin from the runner P2 to the molded product cavity C. The flow path cross-sectional area of ​​the gate P3 is formed smaller than the flow path cross-sectional area of ​​the runner P2. In this embodiment, a plurality of gates P3 are formed, each connecting the plurality of runners P2 to one molded product cavity C. Therefore, even if one of the plurality of gates P3 becomes clogged, the molten resin will flow from the other gates P3 to the molded product cavity C.

[0028] 2, the mold 30 is provided with a flow path pressure measuring device 33. The flow path pressure measuring device 33 is provided at any one of the middle of the spool P1, the end of the spool P1, the middle of the runner P2, the tip of the runner P2, etc. in the resin flow path P. In other words, the flow path pressure measuring device 33 acquires pressure data at a position different from the gate P3 in the resin flow path P. The flow path pressure measuring device 33 may be provided at one location in the mold 30, or at multiple locations.

[0029] (3. Injection molding method) The injection molding method for a molded product using the injection molding apparatus 1 will be described with reference to Fig. 4. The injection molding method is executed by a control device 50 of the injection molding apparatus 1.

[0030] First, before a series of multiple molding cycles, the control device 50 executes a purging step S1 in which the nozzle 24 is separated from the mold 30 and the molten resin in the cylinder 22 is discharged from the nozzle 24. The purging step S1 is performed, for example, for the purpose of discharging thermally deteriorated resin or replacing it with a different resin material. In the purging step S1, the screw 23 is advanced to discharge the molten resin in the cylinder 22 from the nozzle 24.

[0031] Next, the control device 50 executes a metering step S2 in which the screw 23 is retracted to a predetermined position while rotating the screw 23, thereby storing a predetermined amount of molten resin on the front side of the cylinder 22. In the metering step S2, the screw 23, which is located in the forward position, is rotated to move the molten resin to the front end side of the cylinder 22, and the screw 23 is retracted to a predetermined position by a reaction force of the forward movement of the molten resin. In this way, a predetermined amount of molten resin is stored between the tip of the screw 23 and the nozzle 24 in the cylinder 22.

[0032] Next, the control device 50 executes a nozzle touch step S3 in which the cylinder 22 is advanced to bring the nozzle 24 into contact with the mold 30. At this time, the mold 30 is assumed to be clamped. However, the mold may be clamped after the nozzle touch step S3.

[0033] Next, the control device 50 executes the following molding cycles S4 to S11. The control device 50 executes the injection step S4 in which the screw 23 is advanced by controlling the speed of the screw 23, and the molten resin is injected from the nozzle 24 into the mold 30. In the injection step S4, the molten resin flows from the nozzle 24 into the resin flow path P, and from the resin flow path P into the molded product cavity C. In the injection step S4, the molten resin is supplied to most of the molded product cavity C (e.g., 90 to 95%).

[0034] Following the injection step S4, the control device 50 executes a pressure holding step S5 in which a holding pressure is applied to the molten resin in the molded product cavity C by controlling the pressure applied to the screw 23. In the pressure holding step S5, a predetermined pressure is applied to the screw 23, thereby causing the screw 23 to move further forward and supplying the molten resin from the nozzle 24 through the resin flow path P to the molded product cavity C. In the pressure holding step S5, the molded product cavity C is completely filled with the molten resin.

[0035] Next, the control device 50 stops applying pressure to the screw 23 and stops heating the mold 30, and executes a cooling step S6 in which the mold 30 is cooled to cool the molten resin in the mold 30. In the cooling step S6, the molten resin in the mold 30 is solidified. Following the cooling step S6, the control device 50 controls the mold clamping device 40 to separate the second mold 32 from the first mold 31 and executes a mold release / molded product removal step S7 in which the molded product is removed. Next, the control device 50 controls the mold clamping device 40 to execute a mold clamping step S8 in which the second mold 32 is aligned with the first mold 31 and the molds are clamped.

[0036] Furthermore, after completing the pressure holding step S5, the control device 50 executes a nozzle separating step S9 in which the nozzle 24 is separated from the die 30 by retracting the cylinder 22. Following the nozzle separating step S9, the control device 50 executes a measuring step S10 in which the control device 50 moves the molten resin to the front end side of the cylinder 22 by rotating the screw 23 located at the front position, and then retracts the screw 23 to a predetermined position due to the reaction of the forward movement of the molten resin, thereby storing a predetermined amount of molten resin on the front side of the cylinder 22.

[0037] After the mold clamping step S8 and the measurement step S10 are completed, the control device 50 executes the nozzle touch step S11 in which the cylinder 22 is advanced to bring the nozzle 24 into contact with the mold 30. However, the mold clamping step S8 may be executed after the nozzle touch step S11. Then, after the nozzle touch step S11, the above-mentioned injection step S4 is repeated again.

[0038] (4. Behavior of screw pressure data and flow path pressure data) The behavior (change over time) of the screw pressure data and the behavior (change over time) of the flow channel pressure data in the injection process S4 and the pressure holding process S5 will be described with reference to Fig. 5 and Fig. 6. First, the state where there is no wear of the gate P3 will be described with reference to the solid line in Fig. 5 and the solid line in Fig. 6.

[0039] 5, at time t1, the injection process S4 is started, whereby the screw 23 advances and the screw pressure data increases. After that, in this embodiment, the forward movement speed of the screw 23 is initially set to a low speed and then switched to a high speed.

[0040] As shown by the solid lines in Fig. 5 and Fig. 6, when the forward speed of the screw 23 is switched to high speed, the molten resin flows into the resin flow path P of the mold 30, causing the screw pressure data and the flow path pressure data to rise sharply. Then, as the resin flows from the gate P3 of the resin flow path P into the molded product cavity C, the screw pressure data and the flow path pressure data drop slightly and then change to rise gradually. Therefore, in the injection process S4, the screw pressure data and the flow path pressure data reach maximum values ​​at the pressure peak time t2.

[0041] The time when the injection process S4 ends is t4, and the time between the pressure peak time t2 and the injection process end time t4 in the injection process S4 is t3. The time t3 is, for example, the intermediate time between the pressure peak time t2 and the injection process end time t4, but may be a time shifted from the intermediate time.

[0042] By switching from the injection process S4 to the dwelling process S5, the screw pressure data and the flow channel pressure data decrease to near the desired dwelling pressure by the pressure control in the dwelling process S5. Usually, the screw pressure data and the flow channel pressure data increase toward near the desired dwelling pressure immediately after decreasing to a pressure slightly lower than the desired dwelling pressure.

[0043] Then, the pressure holding step S5 continues, and when the molded product cavity C is completely filled with the molten resin, that is, at the time of filling completion t5, the screw pressure data and the flow path pressure data increase. The screw pressure data is pressure controlled, so it reaches the desired holding pressure immediately thereafter. On the other hand, the flow path pressure data gradually increases after the time of filling completion t5.

[0044] Next, a case where the amount of wear of the gate P3 is large will be described with reference to the dashed lines in Fig. 5 and Fig. 6. In the injection process S4, the screw pressure data and the flow channel pressure data are generally smaller than when there is no wear of the gate P3. Here, the rate of increase in the flow channel pressure data is smaller than the rate of increase in the screw pressure. The rate of increase is the ratio of the pressure increase when the gate P3 is worn to the case when there is no wear of the gate P3. Therefore, the wear of the gate P3 affects both the screw pressure data and the flow channel pressure data.

[0045] The flow path pressure data is pressure data in the resin flow path P of the mold 30, and is therefore affected by the wear of the gate P3 as well as the state of the molten resin flowing through it. On the other hand, the screw pressure data is far from the resin flow path P of the mold 30, and is therefore less affected by the wear of the gate P3, but is affected by the state of the molten resin in the cylinder 22. The state of the molten resin in the cylinder 22 and the state of the molten resin in the mold 30 are similar.

[0046] The flow path pressure data is a value influenced by gate wear and the state of the molten resin, and the screw pressure data is a value influenced by the state of the molten resin. If the influence of the state of the molten resin can be removed from the flow path pressure data, the influence of gate wear can be extracted.

[0047] (5. Overview of Each Embodiment) Regarding the calculation of the value representing the wear of the gate P3, an overview of each embodiment will be described. (A1) First embodiment: The gate wear amount is calculated by machine learning based on the flow path pressure data and the screw pressure data. (A2) Second embodiment: A gate wear index value is calculated from relationship information (such as a relational expression) using flow path pressure data and screw pressure data, and the gate wear amount is calculated by machine learning based on the gate wear index value. (A3) Third embodiment: A gate wear index value is calculated from relationship information (such as a relational expression) using flow path pressure data and screw pressure data. (A4) Fourth embodiment: A gate wear index value is calculated from first relationship information (such as a relational expression) using flow path pressure data and screw pressure data, and a gate wear amount is calculated using second relationship information (such as a relational expression) and the gate wear index value. (A5) Fifth embodiment: The gate wear amount is calculated by machine learning based on the flow path pressure data, the screw pressure data, and the estimated resin viscosity. (A6) Sixth embodiment: The gate wear amount is calculated by machine learning based on the flow path pressure data and the screw pressure data for each estimated resin viscosity. (A7) Seventh embodiment: A gate wear index value is calculated from relationship information using flow path pressure data and screw pressure data, and the gate wear amount is calculated by machine learning based on the gate wear index value and estimated resin viscosity. (A8) Eighth embodiment: A gate wear index value is calculated from relationship information using flow path pressure data and screw pressure data, and the gate wear amount is calculated by machine learning based on the gate wear index value for each estimated resin viscosity. (A9) Ninth embodiment: A gate wear index value is calculated from first relationship information using flow path pressure data and screw pressure data, and a gate wear amount is calculated using the gate wear index value, estimated resin viscosity, and second relationship information. (B1) First specific example of a method for estimating resin viscosity: Using flow path pressure data, the time to the pressure peak in the injection process S4 is used. (B2) A second specific example of a method for estimating resin viscosity: Using flow path pressure data, the time until the molding cavity is completely filled in the pressure holding process S5 is used. (B3) A third specific example of a method for estimating resin viscosity: Using screw pressure data, the time until the pressure peak in the injection process S4 is used. (B4) A fourth specific example of a method for estimating resin viscosity: Using screw pressure data, the time until the molding cavity is completely filled in the pressure holding step S5 is used. (B5) Fifth specific example of the resin viscosity estimation method: Nozzle pressure data and screw movement speed data in the purging step S1 are used. (B6) Sixth specific example of the resin viscosity estimation method: Nozzle pressure data and screw movement speed data in the injection process S4 are used. (B7) A seventh specific example of the resin viscosity estimation method: Using the time required for the measurement steps S2 and S10. (B8) Eighth specific example of the resin viscosity estimation method: A plurality of the above first to seventh specific examples are used in combination.

[0048] (6. Injection molding apparatus 1 of first embodiment) An injection molding apparatus 1 of a first embodiment will be described with reference to Figures 7 and 8. As shown in Figure 7, the injection molding apparatus 1 includes an injection device 20, a mold 30, a clamping device 40, a control device 50, and a computer device 60. In Figure 7, only some of the functional parts of the injection molding apparatus 1 are shown. In the following, the computer device 60 will be described.

[0049] The computer device 60 includes an output unit 61 and a storage unit 62. The output unit 61 is configured by the arithmetic unit constituting the computer device 60 described above, and functions by executing a computer program. The storage unit 62 is configured by the storage device constituting the computer device 60 described above.

[0050] The output unit 61 outputs a value representing the wear state of the gate P3 in the resin flow path P. As described above, the gate P3 is located between the runner P2 and the molded product cavity C, and has a flow path cross-sectional area smaller than that of the runner P2. Therefore, high pressure is applied to the gate P3 when the molten resin flows through it, causing the gate P3 to wear. In particular, when the above-mentioned molding cycle is performed multiple times in succession, the wear of the gate P3 progresses. Therefore, the output unit 61 outputs a value representing the wear state of the gate P3 as information for grasping the wear state of the gate P3 every time a molding cycle is performed.

[0051] The output unit 61 also outputs a value representing the wear state of the gate P3 based on the flow path pressure data acquired by the flow path pressure measuring device 33 and the screw pressure data acquired by the screw pressure measuring device 27. Furthermore, the output unit 61 outputs a value representing the wear state of the gate P3 based on the feature quantities of the flow path pressure data and the feature quantities of the screw pressure data, rather than the flow path pressure data and the screw pressure data themselves. Therefore, the output unit 61 extracts the feature quantities from the acquired flow path pressure data and also extracts the feature quantities from the acquired screw pressure data.

[0052] In this embodiment, the output unit 61 outputs a value representing the wear state of the gate P3 using the feature amount of the channel pressure data in the injection process S4 (times t1 to t4) and the feature amount of the screw pressure data in the injection process S4. In particular, the output unit 61 uses the feature amount of the channel pressure data and the feature amount of the screw pressure data after the pressure reaches the pressure peak in the injection process S4, that is, from time t2 to t4.

[0053] More specifically, in this embodiment, the output unit 61 uses the characteristic quantities of the flow path pressure data in a later time period (t3 to t4) of the injection process S4 closer to the pressure holding process S5, and the characteristic quantities of the screw pressure data in a later time period (t3 to t4) of the injection process S4 closer to the pressure holding process S5.

[0054] The characteristic amount may be, for example, a time integral value of the flow passage pressure data in the later time period (t3 to t4) and a time integral value of the screw pressure data in the later time period (t3 to t4). Note that the characteristic amount may be a maximum value, a minimum value, a median value, a mean value, a first quartile, a third quartile, a variance, a standard deviation, a kurtosis, a skewness, or the like, of the flow passage pressure data in the later time period (t3 to t4). Also, the output unit 61 may use a plurality of characteristic amounts.

[0055] In this embodiment, the output unit 61 applies machine learning to output a value representing the wear state of the gate P3. Therefore, the output unit 61 uses a trained model generated in advance to output a value representing the wear state of the gate P3. In particular, when applying machine learning, the output unit 61 can easily apply multiple feature amounts.

[0056] The storage unit 62 stores information used for outputting a value representing the wear state of the gate P3 by the output unit 61. In this embodiment, the storage unit 62 stores a trained model generated by performing machine learning using a training data set. The output unit 61 uses the trained model stored in the storage unit 62.

[0057] The function of the computer device 60 when machine learning is applied will be described with reference to FIG. 8. As a learning phase, first, a training data set is prepared. The training data set includes flow path pressure data, screw pressure data, and the actual wear amount of the gate P3. In this embodiment, a predetermined feature amount calculated from the flow path pressure data and a predetermined feature amount calculated from the screw pressure data are set as the training data set. In addition, the actual wear amount of the gate P3 is set as, for example, the wear amount in the opening radial direction at the opening of the gate P3 (the boundary portion with the molded product cavity C).

[0058] In the learning phase, the machine learning processing unit 63 in the computer device 60 performs machine learning using a training data set to generate one trained model. The trained model is stored in the storage unit 62. In this embodiment, the trained model uses the feature amounts of the flow path pressure data and the feature amounts of the screw pressure data as explanatory variables, and uses the wear amount of the gate P3 as a target variable, which is a value representing the wear state of the gate P3.

[0059] Next, in the estimation phase, the output unit 61 acquires the flow path pressure data and the screw pressure data as the detection data. Then, the output unit 61 calculates the feature amount of the flow path pressure data and calculates the feature amount of the screw pressure data. Next, the output unit 61 uses the trained model stored in the storage unit 62 to input the feature amount of the flow path pressure data and the feature amount of the screw pressure data, thereby outputting the wear amount of the gate P3 as a value representing the wear state of the gate P3.

[0060] Furthermore, the output unit 61 judges whether or not the gate P3 portion requires maintenance based on the amount of wear of the gate P3 as a value representing the wear state of the gate P3. If the output unit 61 judges that the gate P3 portion requires maintenance, it issues a maintenance alert for the gate P3 portion. As a method of issuing the notice, a display device (not shown) may display a maintenance alert indicating that it is time to replace the gate P3 portion, or an alarm may be issued. Furthermore, a management device (not shown) or a fixed terminal or mobile terminal of an operator may be notified of the maintenance alert.

[0061] (7. Effects of the First Embodiment) As described above, the flow path pressure data is pressure data in the resin flow path P of the mold 30, and is therefore affected by the wear of the gate P3 as well as the state of the molten resin flowing. On the other hand, the screw pressure data is far from the resin flow path P of the mold 30, and is therefore less affected by the wear of the gate P3, but is affected by the state of the molten resin in the cylinder 22. Here, the state of the molten resin in the cylinder 22 and the state of the molten resin in the mold 30 are similar.

[0062] The output unit 61 uses the feature amount of the flow path pressure data and the feature amount of the screw pressure data. The feature amount of the flow path pressure data is a value influenced by the wear of the gate P3 and the state of the molten resin, and the feature amount of the screw pressure data is a value influenced by the state of the molten resin. If the influence of the state of the molten resin can be removed from the feature amount of the flow path pressure data, the influence of the wear of the gate P3 can be extracted.

[0063] Therefore, the output unit 61 performs processing using the feature amount of the flow path pressure data and the feature amount of the screw pressure data, thereby extracting the influence of the wear of the gate P3 and outputting a value representing the wear state of the gate P3. Since the worker can grasp the wear state of the gate P3 based on the value representing the wear state of the gate P3, the worker can appropriately determine the timing of replacing the gate P3 portion without relying on the knowledge and experience of an expert.

[0064] In particular, the machine learning processing unit 63 of the computer device 60 includes the actual wear amount of the gate P3 in the training data set in the machine learning, and learns the objective variable as the wear amount of the gate P3. Therefore, the output unit 61 outputs the wear amount of the gate P3 as a value representing the wear state of the gate P3. Since the worker can check the estimated value of the wear amount of the gate P3, the worker can more easily determine the timing of replacing the gate P3.

[0065] Furthermore, the output unit 61 uses the data in the injection process S4 from among the flow path pressure data and the screw pressure data. Particularly in the injection process S4, the pressure changes depending on the presence or absence of wear of the gate P3. Therefore, the output unit 61 can output a value representing the wear state of the gate P3 with high accuracy by using the pressure data in the injection process S4.

[0066] Furthermore, the output unit 61 uses the flow path pressure data and the screw pressure data in a later time period (t3 to t4) closer to the pressure holding step S5 in the injection process S4. In the later time period (t3 to t4) in the injection process S4, the pressure data is relatively stable. Therefore, the output unit 61 can output a value representing the wear state of the gate P3 with high accuracy. In addition, the output unit 61 outputs a value representing the wear state of the gate P3 by machine learning. In particular, when multiple feature amounts are used, judgment by machine learning is useful.

[0067] (8. Second embodiment of injection molding apparatus 1) An injection molding apparatus 1 according to a second embodiment will be described with reference to Figures 9 and 10. In the following, the function of a computer device 60 that applies machine learning will be described.

[0068] In the learning phase, first, the detection data, ie, the flow path pressure data and the screw pressure data, are prepared, and then a training data set is prepared. The training data set includes a gate wear index value y and an actual wear amount of the gate P3. The actual wear amount of the gate P3, which is one of the training data sets, is, for example, the wear amount in the radial direction of the opening of the gate P3 (the boundary portion with the molded product cavity C).

[0069] The gate wear index value y as one of the training data sets is calculated by the gate wear index value calculation unit 64 of the computer device 60 using the flow path pressure data and the screw pressure data. The gate wear index value y is defined by the equation (1) as the relationship information using the feature amount of the flow path pressure data and the feature amount of the screw pressure data. In other words, the equation (1) corresponds to the relationship information that defines the relationship between the feature amount of the flow path pressure data, the feature amount of the screw pressure data, and the gate wear index value y. Here, the constant a in the equation (1) is determined so that the gate wear index value y is a constant value when the gate P3 is not worn.

[0070] y = b × x (1) y: Gate wear index value x: Feature quantity of screw pressure data a: constant b: Feature quantity of flow path pressure data

[0071] When the number of molding cycles (number of shots) and the gate wear index value y defined by formula (1) are plotted on a graph, the results are as shown by the points in Figure 10. In the graph of Figure 10, the plot points are divided into the initial, middle, and final periods, and linear approximation is performed, and an approximation line is shown. According to the approximation line, it can be seen that the gate wear index value y decreases as the number of molding cycles increases. And, the amount of wear of gate P3 increases as the number of molding cycles increases. In other words, it can be seen that the gate wear index value y is a value that changes so that it decreases as the amount of wear of gate P3 increases.

[0072] Next, in a learning phase, the machine learning processing unit 63 in the computer device 60 performs machine learning using the training data set to generate one trained model. The trained model is stored in the storage unit 62. In this embodiment, the trained model uses the gate wear index value y as an explanatory variable and the wear amount of the gate P3 as a target variable, which is a value representing the wear state of the gate P3.

[0073] Next, in the estimation phase, the gate wear index value calculation unit 61a of the output unit 61 acquires the channel pressure data and the screw pressure data as the detection data. Then, the gate wear index value calculation unit 61a calculates the feature amount of the channel pressure data and calculates the feature amount of the screw pressure data. Next, the gate wear index value calculation unit 61a calculates the gate wear index value y by the above-mentioned formula (1) using the feature amount of the channel pressure data and the feature amount of the screw pressure data.

[0074] Then, the wear amount calculation unit 61b of the output unit 61 inputs the calculated gate wear index value y using the learned model stored in the memory unit 62, and outputs the wear amount of gate P3 as a value representing the wear state of gate P3.

[0075] In this embodiment, a gate wear index value y is calculated using the feature amount of the flow path pressure data and the feature amount of the screw pressure data, and a trained model is generated by machine learning that defines the relationship between the gate wear index value y and the wear amount of the gate P3 as a value representing the wear state of the gate P3. Then, the wear amount of the gate P3 is calculated using the trained model generated by machine learning.

[0076] By performing machine learning using the gate wear index value y, which is an index highly related to the wear amount of the gate P3, as an explanatory variable, the generated trained model can easily adapt the relationship with the objective variable, the wear amount of the gate P3. As a result, the wear amount of the gate P3 output by the wear amount calculation unit 61b of the output unit 61 using the trained model can be output with high accuracy.

[0077] (9. Injection molding apparatus 1 of third embodiment) An injection molding apparatus 1 according to a third embodiment will be described with reference to Fig. 11. The process of the output unit 61 of the computer device 60 will be described below. In this embodiment, the output unit 61 does not apply machine learning.

[0078] 11, the output unit 61 executes a detection data input step S21 for inputting the detection data, ie, the channel pressure data and the screw pressure data. Then, the output unit 61 executes a feature amount calculation step S22 for calculating the feature amount of the detection data. As described above, the feature amount of the detection data is the feature amount of the channel pressure data and the feature amount of the screw pressure data.

[0079] Next, the output unit 61 executes an index value calculation step S23 in which the gate wear index value y is calculated from the calculated feature amount of the flow channel pressure data and the feature amount of the screw pressure data by the above-mentioned formula (1). Formula (1) is relationship information that defines the relationship between the feature amount of the flow channel pressure data, the feature amount of the screw pressure data, and the gate wear index value y.

[0080] In this embodiment, the output unit 61 can output the gate wear index value y. As described above, the gate wear index value y is a value that changes to become smaller as the amount of wear of the gate P3 increases. Therefore, the worker can grasp the wear state of the gate P3 based on the gate wear index value y, and can appropriately determine the timing of replacing the gate P3 without relying on the knowledge or experience of a skilled worker.

[0081] (10. Injection molding apparatus 1 of the fourth embodiment) An injection molding apparatus 1 according to the fourth embodiment will be described with reference to Fig. 12. The process of the output unit 61 of the computer device 60 will be described below. In this embodiment, the output unit 61 does not apply machine learning.

[0082] 12, the output unit 61 executes a detection data input step S21, a feature amount calculation step S22, and an index value calculation step S23, similarly to the third embodiment. Here, the process in the index value calculation step S23 uses formula (1), which is first relationship information that defines the relationship between the feature amount of the channel pressure data, the feature amount of the screw pressure data, and the gate wear index value y.

[0083] Next, the output unit 61 executes a wear amount calculation step S24 of calculating the wear amount of the gate P3 by using the second relationship information stored in the storage unit 62 and inputting the gate wear index value y.

[0084] Here, the second relationship information is defined by, for example, equation (2) and represents the relationship between the gate wear index value y and the wear amount z of the gate P3. z = F(y) (2) z: Wear amount of gate P3 F(y): Function of gate wear index value y

[0085] The function F(y) is expressed, for example, by a linear function with a negative slope, a function having a term that is the reciprocal of the gate wear index value y, or the like. The function F(y) is set taking into consideration the actual wear amount of the gate P3. Therefore, the output unit 61 can output the wear amount z of the gate P3. Therefore, the worker can grasp the wear state of the gate P3 based on the wear amount z of the gate P3, and can appropriately determine the timing of replacing the gate P3 portion without relying on the knowledge or experience of a skilled worker.

[0086] (11. Injection molding apparatus 1 of the fifth embodiment) An injection molding apparatus 1 of a fifth embodiment will be described with reference to Fig. 13. As shown in Fig. 13, a computer device 60 of the injection molding apparatus 1 includes an output unit 61, a storage unit 62, and a viscosity estimation unit 65. The output unit 61 and the viscosity estimation unit 65 are configured by the arithmetic unit constituting the above-mentioned computer device 60, and function by executing a computer program.

[0087] The viscosity estimation unit 65 estimates the viscosity of the molten resin flowing through the resin flow path P in the mold 30. The viscosity estimation unit 65 estimates the viscosity of the molten resin based on the flow path pressure data, the screw pressure data, the nozzle pressure data, the control data in the control device 50, etc. Several specific examples of the viscosity estimation unit 65 will be described below.

[0088] The output unit 61 adds the viscosity estimated by the viscosity estimation unit 65 (hereinafter referred to as "estimated resin viscosity") to the channel pressure data and the screw pressure data, and outputs a value indicating the wear state of the gate P3.

[0089] Moreover, the output unit 61 applies machine learning to output a value representing the wear state of the gate P3. Therefore, the output unit 61 outputs a value representing the wear state of the gate P3 using a trained model generated in advance. The storage unit 62 stores the trained model for the output unit 61 to output a value representing the wear state of the gate P3.

[0090] The function of the computer device 60 when machine learning is applied will be described with reference to Fig. 14. In the learning phase, a training data set is first prepared. The training data set includes the feature amount of the flow path pressure data, the feature amount of the screw pressure data, the estimated resin viscosity estimated by the viscosity estimation unit 65, and the actual wear amount of the gate P3 as a value representing the wear state of the gate P3.

[0091] In the learning phase, the machine learning processing unit 63 in the computer device 60 performs machine learning using a training data set to generate one trained model. The trained model is stored in the storage unit 62. In this embodiment, the trained model has the feature amount of the flow path pressure data, the feature amount of the screw pressure data, and the estimated resin viscosity as explanatory variables, and has the wear amount of the gate P3 as a value representing the wear state of the gate P3 as a response variable.

[0092] Next, in the estimation phase, the output unit 61 acquires the flow path pressure data and the screw pressure data as the detection data, and the estimated resin viscosity estimated by the viscosity estimation unit 65. Then, the output unit 61 calculates the feature amounts of the flow path pressure data and the feature amounts of the screw pressure data. Next, the output unit 61 uses the trained model stored in the storage unit 62 to input the feature amounts of the flow path pressure data, the feature amounts of the screw pressure data, and the estimated resin viscosity, and outputs the wear amount of the gate P3 as a value representing the wear state of the gate P3.

[0093] If the resin viscosity is different, the flow path pressure data and the screw pressure data will show different values. Therefore, the output unit 61 outputs the wear amount of the gate P3 as a value representing the wear state of the gate P3 by adding the estimated resin viscosity to the flow path pressure data and the screw pressure data. Therefore, the output unit 61 can output the wear amount of the gate P3 with higher accuracy.

[0094] (12. Injection molding apparatus 1 of the sixth embodiment) An injection molding apparatus 1 according to the sixth embodiment will be described with reference to Fig. 15. The functions of a computer device 60 to which machine learning is applied will be described below.

[0095] In the learning phase, first, the detection data, ie, the flow path pressure data and the screw pressure data, are prepared, and then a training data set is prepared. The training data set includes the feature amount of the flow path pressure data, the feature amount of the screw pressure data, and the actual wear amount of the gate P3, which are classified by the resin viscosity estimated by the viscosity estimation unit 65.

[0096] Next, in a learning phase, the machine learning processing unit 63 in the computer device 60 performs machine learning using a training data set for each estimated resin viscosity to generate multiple trained models for each estimated resin viscosity. The multiple trained models for each estimated resin viscosity are stored in the storage unit 62. In this embodiment, each trained model uses the feature amount of the flow path pressure data and the feature amount of the screw pressure data as explanatory variables, and uses the wear amount of the gate P3 as a target variable representing the wear state of the gate P3.

[0097] Next, in the estimation phase, the output unit 61 acquires the flow path pressure data and the screw pressure data as the detection data. Then, the output unit 61 calculates the feature amount of the flow path pressure data and calculates the feature amount of the screw pressure data. In addition, the output unit 61 acquires the estimated resin viscosity estimated by the viscosity estimation unit 65.

[0098] Next, the output unit 61 selects one trained model corresponding to the estimated resin viscosity from among the multiple trained models stored in the storage unit 62. Next, the output unit 61 uses the selected trained model to input the feature amount of the flow path pressure data and the feature amount of the screw pressure data, thereby outputting the wear amount of the gate P3 as a value representing the wear state of the gate P3. That is, the output unit 61 classifies the estimated resin viscosity, and performs a calculation process of the wear amount of the gate P3 according to the classified estimated resin viscosity. In this case, too, the same effect as in the fifth embodiment described above is exhibited.

[0099] (13. Seventh embodiment of injection molding apparatus 1) An injection molding apparatus 1 according to the seventh embodiment will be described with reference to Fig. 16. In the following, the functions of a computer device 60 to which machine learning is applied will be described.

[0100] In the learning phase, first, the detection data, ie, the flow path pressure data and the screw pressure data, are prepared. Then, a training data set is prepared. The training data set includes a gate wear index value y, an estimated resin viscosity estimated by the viscosity estimation unit 65, and an actual wear amount of the gate P3 as a value representing the wear state of the gate P3. The gate wear index value y is defined by the above-mentioned formula (1).

[0101] Next, in a learning phase, the machine learning processing unit 63 in the computer device 60 performs machine learning using the training data set to generate one trained model. The trained model is stored in the storage unit 62. In this embodiment, the trained model uses the gate wear index value y and the estimated resin viscosity as explanatory variables, and uses the wear amount of the gate P3 as a value representing the wear state of the gate P3 as a response variable.

[0102] Next, in the estimation phase, the gate wear index value calculation unit 61a of the output unit 61 acquires the channel pressure data and the screw pressure data as the detection data. Then, the gate wear index value calculation unit 61a calculates the feature amount of the channel pressure data and calculates the feature amount of the screw pressure data. Next, the gate wear index value calculation unit 61a calculates the gate wear index value y by the above-mentioned formula (1) using the feature amount of the channel pressure data and the feature amount of the screw pressure data.

[0103] Then, the wear amount calculation unit 61b of the output unit 61 acquires the gate wear index value y calculated by the gate wear index value calculation unit 61a and the estimated resin viscosity estimated by the viscosity estimation unit 65. Next, the wear amount calculation unit 61b of the output unit 61 uses the trained model stored in the storage unit 62 to input the gate wear index value y and the estimated resin viscosity, and outputs the wear amount of gate P3 as a value representing the wear state of gate P3.

[0104] According to this embodiment, machine learning is applied using the estimated resin viscosity in addition to the gate wear index value y, which is an index highly related to the wear amount of the gate P3. Therefore, the wear amount of the gate P3 output by the wear amount calculation unit 61b of the output unit 61 using the trained model generated by the machine learning can be output with high accuracy.

[0105] (14. Eighth embodiment of injection molding apparatus 1) An injection molding apparatus 1 according to the eighth embodiment will be described with reference to Fig. 17. The functions of a computer device 60 to which machine learning is applied will be described below.

[0106] In the learning phase, first, the detection data, ie, the flow path pressure data and the screw pressure data, are prepared, and then a training data set is prepared. The training data set includes the gate wear index value y and the actual wear amount of the gate P3 classified by the resin viscosity estimated by the viscosity estimation unit 65. The gate wear index value y is defined by the above-mentioned formula (1).

[0107] Next, in a learning phase, the machine learning processing unit 63 in the computer device 60 performs machine learning using a training data set for each estimated resin viscosity to generate multiple trained models for each estimated resin viscosity. The multiple trained models for each estimated resin viscosity are stored in the storage unit 62. In this embodiment, each trained model uses the gate wear index value y as an explanatory variable and the wear amount of gate P3 as a target variable, which is a value representing the wear state of gate P3.

[0108] Next, in the estimation phase, the gate wear index value calculation unit 61a of the output unit 61 acquires the channel pressure data and the screw pressure data as the detection data. Then, the gate wear index value calculation unit 61a calculates the feature amount of the channel pressure data and calculates the feature amount of the screw pressure data. Next, the gate wear index value calculation unit 61a calculates the gate wear index value y by the above-mentioned formula (1) using the feature amount of the channel pressure data and the feature amount of the screw pressure data.

[0109] Then, the wear amount calculation section 61b of the output section 61 acquires the gate wear index value y calculated by the gate wear index value calculation section 61a and the estimated resin viscosity estimated by the viscosity estimation section 65.

[0110] Next, the wear amount calculation unit 61b of the output unit 61 selects one trained model corresponding to the estimated resin viscosity from among the multiple trained models stored in the storage unit 62. Next, the wear amount calculation unit 61b of the output unit 61 uses the selected trained model to input the gate wear index value y, thereby outputting the wear amount of the gate P3 as a value representing the wear state of the gate P3. That is, the wear amount calculation unit 61b of the output unit 61 classifies the estimated resin viscosity, and performs a calculation process of the wear amount of the gate P3 according to the classified estimated resin viscosity. In this case, the same effect as that of the seventh embodiment described above is also exhibited.

[0111] (15. Ninth embodiment of injection molding apparatus 1) An injection molding apparatus 1 according to a ninth embodiment will be described with reference to Fig. 18. The process of the output unit 61 of the computer device 60 will be described below. In this embodiment, the output unit 61 does not apply machine learning.

[0112] 18, the output unit 61 executes a detection data input step S31 in which the detection data, ie, the channel pressure data and the screw pressure data, are input. Then, the output unit 61 executes an estimated resin viscosity input step S32 in which the resin viscosity v estimated by the viscosity estimation unit 65 is input. Then, the output unit 61 executes a feature amount calculation step S33 in which a feature amount of the detection data is calculated. As described above, the feature amount of the detection data is the feature amount of the channel pressure data and the feature amount of the screw pressure data.

[0113] Next, the output unit 61 executes an index value calculation step S34 in which the gate wear index value y is calculated from the first relationship information, which is the above-mentioned formula (1), using the calculated feature amount of the flow path pressure data and the feature amount of the screw pressure data. The formula (1), which is the first relationship information, defines the relationship between the feature amount of the flow path pressure data, the feature amount of the screw pressure data, and the gate wear index value y.

[0114] Next, the output unit 61 executes a wear amount calculation step S35 in which the output unit 61 uses the second relationship information stored in the memory unit 62 to input the gate wear index value y and the estimated resin viscosity v to calculate the wear amount of the gate P3.

[0115] The second relationship information is defined by, for example, formula (3) and represents the relationship among the gate wear index value y, the estimated resin viscosity v, and the wear amount z of the gate P3. z = F(y,v) (3) z: Wear amount of gate P3 F(y,v): Function of gate wear index value y and estimated resin viscosity v

[0116] Therefore, the output unit 61 can output the wear amount z of the gate P3. Therefore, the worker can grasp the wear state of the gate P3 based on the wear amount z of the gate P3, and can appropriately determine the timing of replacing the gate P3 without depending on the knowledge or experience of a skilled worker.

[0117] (16. Behavior of screw pressure data and flow path pressure data) The behavior (change over time) of the screw pressure data and the behavior (change over time) of the flow path pressure data in the injection process S4 and the pressure holding process S5 will be described with reference to Figures 19 to 22. In Figures 19 to 22, thick lines indicate the case where the resin viscosity in the resin flow path P is high, and thin lines indicate the case where the resin viscosity in the resin flow path P is low.

[0118] 19 and 20, the screw pressure data has different values ​​between pressure peak times t2a, t2b and injection process end time t4 in the injection process S4. That is, in that time period (t2a to t4, t2b to t4), if the resin viscosity is low, the screw pressure data shows a low value, and if the resin viscosity is high, the screw pressure data shows a high value.

[0119] 20, in the screw pressure data, the time (t1 to t2a, t1 to t2b) from the start time t1 of the injection process S4 to the pressure peak times t2a, t2b in the injection process S4 differs depending on the resin viscosity. That is, the time (t1 to t2b) when the resin viscosity is low is shorter than the time (t1 to t2a) when the resin viscosity is high.

[0120] As shown in Fig. 19, in the screw pressure data, the time (t4-t5a, t4-t5b) from the start time t4 of the pressure holding process S5 to the filling completion time t5a, t5b of the molded product cavity C in the pressure holding process S5 varies depending on the resin viscosity. In other words, the time (t4-t5b) when the resin viscosity is low is shorter than the time (t4-t5a) when the resin viscosity is high. The filling completion times t5a, t5b are the time when the pressure of the screw pressure data in the pressure holding process S5 rises, albeit slightly. In other words, the filling completion times t5a, t5b are the time when the pressure change of the screw pressure data in the pressure holding process S5 becomes larger than a predetermined value.

[0121] Here, the injection process S4 ends when a predetermined time has elapsed since the start, so the end time t4 of the injection process S4 is constant regardless of the resin viscosity. Therefore, the same applies even if the above times (t4 to t5a, t4 to t5b) are replaced with the times (t1 to t5a, t1 to t5b) from the start time t1 of the injection process S4 to the completion times t5a, t5b of filling the molded product cavity C in the pressure holding process S5.

[0122] As shown in Figures 21 and 22, the flow path pressure data behaves in the same way as the screw pressure data. The flow path pressure data has different values ​​between pressure peak times t2a, t2b and the end time t4 of the injection process S4. That is, in that time period (t2a to t4, t2b to t4), if the resin viscosity is low, the screw pressure data shows a low value, and if the resin viscosity is high, the screw pressure data shows a high value.

[0123] 22, in the flow path pressure data, the time (t1 to t2a, t1 to t2b) from the start time t1 of the injection process S4 to the pressure peak times t2a, t2b in the injection process S4 differs depending on the resin viscosity. That is, the time (t1 to t2b) when the resin viscosity is low is shorter than the time (t1 to t2a) when the resin viscosity is high.

[0124] As shown in FIG. 21, in the flow path pressure data, the time (t4-t5a, t4-t5b) from the start time t4 of the pressure holding process S5 to the filling completion time t5a, t5b of the molded product cavity C in the pressure holding process S5 varies depending on the resin viscosity. In other words, the time (t4-t5b) when the resin viscosity is low is shorter than the time (t4-t5a) when the resin viscosity is high. The filling completion times t5a, t5b are the time when the pressure of the flow path pressure data in the pressure holding process S5 rises. In other words, the filling completion times t5a, t5b are the time when the pressure change of the flow path pressure data in the pressure holding process S5 becomes larger than a predetermined value. The pressure change of the flow path pressure data is larger than that of the screw pressure data.

[0125] Here, the same applies if the above times (t4 to t5a, t4 to t5b) are replaced with the times (t1 to t5a, t1 to t5b) from the start time t1 of the injection process S4 to the completion time t5a, t5b of filling the molded product cavity C in the pressure holding process S5.

[0126] (17. First Example of Resin Viscosity Estimation) In the fifth to ninth embodiments, the viscosity estimation unit 65 estimates the viscosity of the molten resin in the resin flow path P. A first specific example of a method for estimating the resin viscosity by the viscosity estimation unit 65 will be described with reference to FIGS. 21 and 22.

[0127] 21 and 22, the time (t1-t2a, t1-t2b) from the start time t1 of the injection process S4 to the pressure peak times t2a, t2b of the flow path pressure data in the injection process S4 varies depending on the resin viscosity of the resin flow path P. Therefore, the viscosity estimation unit 65 estimates the resin viscosity using the time (t1-t2a, t1-t2b) from the start time t1 of the injection process S4 to the pressure peak times t2a, t2b of the flow path pressure data in the injection process S4 as a component representing the resin viscosity.

[0128] (18. Second Example of Resin Viscosity Estimation) A second specific example of a method for estimating resin viscosity by the viscosity estimation unit 65 will be described with reference to Figures 21 and 22. As shown in Figures 21 and 22, in the flow path pressure data, the time (t1 to t5a, t1 to t5b, or t4 to t5a, t4 to t5b) from the start time t1 of the injection process S4 or the start time t4 of the pressure holding process S5 to the completion time t5a, t5b of filling of the molded product cavity C in the pressure holding process S5 differs depending on the resin viscosity of the resin flow path P.

[0129] Therefore, the viscosity estimation unit 65 estimates the resin viscosity using the time (t1 to t5a, t1 to t5b, or t4 to t5a, t4 to t5b) from the start time t1 of the injection process S4 or the start time t4 of the pressure holding process S5 to the completion time t5a, t5b of filling of the molded product cavity C in the pressure holding process S5 in the flow path pressure data as a component representing the resin viscosity.

[0130] (19. Third Example of Resin Viscosity Estimation) A third specific example of a method for estimating the resin viscosity by the viscosity estimation unit 65 will be described with reference to Figures 19 and 20. As shown in Figures 19 and 20, the time (t1-t2a, t1-t2b) from the start time t1 of the injection process S4 to the pressure peak times t2a, t2b of the screw pressure data in the injection process S4 varies depending on the resin viscosity of the resin flow path P. Therefore, the viscosity estimation unit 65 estimates the resin viscosity using the time (t1-t2a, t1-t2b) from the start time t1 of the injection process S4 to the pressure peak times t2a, t2b of the screw pressure data in the injection process S4 as a component representing the resin viscosity.

[0131] (20. Fourth Specific Example of Resin Viscosity Estimation) A fourth specific example of a method for estimating the resin viscosity by the viscosity estimation unit 65 will be described with reference to Figures 19 and 20. As shown in Figures 19 and 20, the time (t1 to t5a, t1 to t5b, or t4 to t5a, t4 to t5b) from the start time t1 of the injection process S4 or the start time t4 of the pressure holding process S5 to the completion time t5a, t5b of filling the molded product cavity C in the pressure holding process S5 varies in the screw pressure data depending on the resin viscosity of the resin flow path P.

[0132] Therefore, the viscosity estimation unit 65 estimates the resin viscosity using the time (t1 to t5a, t1 to t5b, or t4 to t5a, t4 to t5b) from the start of the injection process S4 or the start of the pressure holding process S5, to the completion of filling of the molded product cavity C in the pressure holding process S5, as a component representing the resin viscosity, in the screw pressure data.

[0133] (21. Fifth Example of Resin Viscosity Estimation) A fifth specific example of a method for estimating a resin viscosity by the viscosity estimation unit 65 will be described with reference to Fig. 23 to Fig. 24. The viscosity estimation unit 65 estimates the resin viscosity using nozzle pressure data acquired by the nozzle pressure measuring device 28 and control data in the control device 50.

[0134] The behavior (change over time) of the nozzle pressure data in the purging process S1 is as shown in Figures 23 and 24. Here, the moving speed of the screw 23 is the same in Figures 23 and 24. When the moving speed of the screw 23 is the same, when the resin viscosity is low as shown in Figure 23, the average value of the nozzle pressure data in the purging process S1 is smaller than when the resin viscosity is high as shown in Figure 24. Therefore, the nozzle pressure data in the purging process S1 depends on the resin viscosity.

[0135] The nozzle pressure data increases as the movement speed of the screw 23 increases. Therefore, the viscosity estimation unit 65 estimates the resin viscosity using the relationship between the nozzle pressure data and the movement speed data of the screw 23 in the purging process S1 as a component representing the resin viscosity.

[0136] (22. Sixth Specific Example of Resin Viscosity Estimation) In the fifth specific example, the viscosity estimation unit 65 estimated the resin viscosity using the nozzle pressure data and the movement speed data of the screw 23 in the purging process S1. Here, the nozzle pressure data changes similarly in the injection process S4. Therefore, in the sixth specific example, the viscosity estimation unit 65 estimates the resin viscosity using the relationship between the nozzle pressure data and the movement speed data of the screw 23 in the injection process S4 as a component representing the resin viscosity.

[0137] (23. Seventh Specific Example of Resin Viscosity Estimation) A seventh specific example of a method for estimating the resin viscosity by the viscosity estimation unit 65 will be described with reference to Fig. 25. The viscosity estimation unit 65 uses control data in the control device 50 to estimate the resin viscosity.

[0138] The time required for the metering steps S2, S10 is affected by the resin viscosity. As shown in Fig. 25, when the resin viscosity is low, the time from the start of the metering steps S2, S10 to the end times t6a, t6b of the metering steps S2, S10 is longer than when the resin viscosity is high. This is because the pressure that the screw 23 receives from the molten resin decreases as the resin viscosity decreases. Therefore, the viscosity estimation unit 65 estimates the resin viscosity using the time required for the metering steps S2, S10 as a component representing the resin viscosity.

[0139] (24. Eighth Specific Example of Resin Viscosity Estimation) A description will be given of an eighth specific example of a method for estimating a resin viscosity by the viscosity estimation unit 65. In the above first to seventh specific examples, the viscosity estimation unit 65 estimates the resin viscosity by taking each element as a component representing the resin viscosity.

[0140] In addition, the viscosity estimation unit 65 estimates the resin viscosity by considering the elements described in the first to seventh specific examples as one of the components expressing the resin viscosity. That is, the viscosity estimation unit 65 estimates the resin viscosity by using a plurality of elements, each of which is considered as one of the components expressing the resin viscosity. Since the viscosity estimation unit 65 uses a plurality of elements, it is preferable to apply machine learning. Of course, the viscosity estimation unit 65 can also estimate the viscosity by using a database in which a plurality of elements are accumulated. [Explanation of symbols]

[0141] 1 Injection molding equipment 20 Injection device 22 cylinders 23 Screw 24 Nozzles 27 Screw pressure measuring device 28 Nozzle pressure measuring device 30 Mold 33 Flow path pressure measuring device 61 Output section C Molding cavity P Resin flow path P1 spool P2 Runner P3 Gate

Claims

1. an injection device including a cylinder, a screw, and a nozzle provided at a front end of the cylinder for ejecting molten resin as the screw advances; a mold including a molded product cavity and a resin flow path between the molded product cavity and a portion that contacts the nozzle; a flow channel pressure measuring device for acquiring flow channel pressure data in the resin flow channel; A screw pressure measuring device for acquiring screw pressure data that the screw receives from the molten resin in the cylinder; an output unit that outputs a value representing a wear state of a gate of the resin flow path based on a feature amount of the flow path pressure data and a feature amount of the screw pressure data; An injection molding apparatus comprising:

2. A memory unit that stores a learned model generated by performing machine learning using a training data set including a feature amount of the flow path pressure data, a feature amount of the screw pressure data, and a value representing a wear state of the gate, The injection molding apparatus according to claim 1 , wherein the output unit uses the learned model to input features of the flow path pressure data and features of the screw pressure data, and outputs a value representing the wear state of the gate.

3. the value representing the wear state of the gate included in the training data set is the actual wear amount of the gate, The injection molding apparatus according to claim 2 , wherein the output unit outputs the amount of wear of the gate as a value representing a wear state of the gate.

4. A gate wear index value is defined using the feature amount of the flow path pressure data and the feature amount of the screw pressure data according to the formula (1), A memory unit that stores a learned model generated by performing machine learning using a training data set including the gate wear index value and the actual wear amount of the gate, The output unit is Calculating the gate wear index value using the feature amount of the flow path pressure data and the feature amount of the screw pressure data; The injection molding apparatus according to claim 1 , wherein the gate wear index value calculated using the trained model is input to output the amount of wear of the gate. y=ba-a×x... (1) y: Gate wear index value x: Feature quantity of screw pressure data a: constant b: Feature quantity of flow path pressure data

5. 5. The injection molding apparatus according to claim 4, wherein the constant a in equation (1) is determined so that y is a constant value when the gate is not worn.

6. The injection molding apparatus according to claim 1 , wherein the output unit calculates a gate wear index value as a value representing a wear state of the gate by using an equation (1). y=ba-a×x... (1) y: Gate wear index value x: Feature quantity of screw pressure data a: constant b: Feature quantity of flow path pressure data

7. 7. The injection molding apparatus according to claim 6, wherein the constant a in equation (1) is determined so that y is a constant value when the gate is not worn.

8. A storage unit that stores relationship information between the gate wear index value and an actual wear amount of the gate, 8. The injection molding apparatus according to claim 6, wherein the output unit outputs the amount of wear of the gate by using the gate wear index value output by equation (1) and the relationship information stored in the memory unit.

9. A viscosity estimation unit that estimates the viscosity of the molten resin flowing through the resin flow path in the mold is further provided, The injection molding apparatus according to any one of claims 1 to 8, wherein the output unit outputs a value representing a wear state of the gate, taking into account the viscosity estimated by the viscosity estimation unit.

10. A control device for controlling the injection device is further provided. The control device includes: an injection step of injecting a molten resin from the nozzle into the mold by controlling the speed of the screw; Following the injection step, a pressure holding step is performed in which a pressure is applied to the screw to apply a holding pressure to the molten resin in the molded product cavity. Executing a molding cycle including 10. The injection molding apparatus according to claim 9, wherein the viscosity estimation unit estimates the viscosity using a time from a start of the injection process to a pressure peak of the flow path pressure data in the injection process as one of components representing the viscosity.

11. A control device for controlling the injection device is further provided. The control device includes: an injection step of injecting a molten resin from the nozzle into the mold by controlling the speed of the screw; Following the injection step, a pressure holding step is performed in which a pressure is applied to the screw to apply a holding pressure to the molten resin in the molded product cavity. Executing a molding cycle including 10. The injection molding apparatus according to claim 9, wherein the viscosity estimation unit estimates the viscosity using a time from a start of the injection process or a start of the pressure holding process to a completion of filling of the molded product cavity in the pressure holding process as one of the components representing the viscosity.

12. The injection molding apparatus according to claim 11, wherein the filling completion time is a time when the pressure of the flow passage pressure data rises in the pressure holding step.

13. 13. The injection molding apparatus according to claim 12, wherein the filling completion time is when a pressure change in the flow channel pressure data during the pressure holding step becomes larger than a predetermined value.

14. The injection molding apparatus according to claim 9 , wherein the viscosity estimation unit estimates the viscosity using the screw pressure data as one of components representing the viscosity.

15. A control device for controlling the injection device is further provided. The control device includes: an injection step of injecting a molten resin from the nozzle into the mold by controlling the speed of the screw; Following the injection step, a pressure holding step is performed in which a pressure is applied to the screw to apply a holding pressure to the molten resin in the molded product cavity. Executing a molding cycle including The injection molding apparatus according to claim 14, wherein the viscosity estimation unit estimates the viscosity using a time from a start of the injection process to a pressure peak of the screw pressure data in the injection process as one of components representing the viscosity.

16. a nozzle pressure measuring device for acquiring nozzle pressure data that the nozzle receives from the molten resin when the molten resin flows through the nozzle; A control device that controls the injection device; Further equipped with The control device includes: an injection step of injecting a molten resin from the nozzle into the mold by controlling the speed of the screw; Following the injection step, a pressure holding step is performed in which a pressure is applied to the screw to apply a holding pressure to the molten resin in the molded product cavity. Executing a molding cycle including The control device further executes a purging step of discharging the molten resin in the cylinder from the nozzle in a state in which the nozzle is spaced from the mold before a plurality of consecutive molding cycles, The injection molding apparatus according to claim 9 , wherein the viscosity estimation unit estimates the viscosity using a relationship between the nozzle pressure data in the purging process and the moving speed data of the screw in the purging process as one of the components representing the viscosity.

17. a nozzle pressure measuring device for acquiring nozzle pressure data that the nozzle receives from the molten resin when the molten resin flows through the nozzle; A control device that controls the injection device; Further equipped with The control device includes: an injection step of injecting a molten resin from the nozzle into the mold by controlling the speed of the screw; Following the injection step, a pressure holding step is performed in which a pressure is applied to the screw to apply a holding pressure to the molten resin in the molded product cavity. Executing a molding cycle including The injection molding apparatus according to claim 9 , wherein the viscosity estimation unit estimates the viscosity using a relationship between the nozzle pressure data in the injection process and the moving speed data of the screw in the injection process as one of the components representing the viscosity.

18. A control device for controlling the injection device is further provided. The control device includes: a metering process in which the screw located at a forward position is rotated to move the molten resin toward the front end side of the cylinder, and the screw is retracted to a predetermined position by a reaction force of the forward movement of the molten resin, thereby storing a predetermined amount of molten resin at the front side of the cylinder; Following the metering step, an injection step of injecting the molten resin from the nozzle into the mold by controlling the speed of the screw; Following the injection step, a pressure holding step is performed in which a pressure is applied to the screw to apply a holding pressure to the molten resin in the molded product cavity. Executing a molding cycle including The injection molding apparatus according to claim 9 , wherein the viscosity estimation unit estimates the viscosity using a time required for the metering step as one of components representing the viscosity.

19. A control device for controlling the injection device is further provided. The control device includes: an injection step of injecting a molten resin from the nozzle into the mold by controlling the speed of the screw; Following the injection step, a pressure holding step is performed in which a pressure is applied to the screw to apply a holding pressure to the molten resin in the molded product cavity. Executing a molding cycle including The output unit outputs a value representing the wear state of the gate based on a characteristic amount of the flow path pressure data in a later time period of the injection process closer to the pressure holding process, and a characteristic amount of the screw pressure data in a later time period of the injection process closer to the pressure holding process. The injection molding apparatus according to any one of claims 1 to 8.

20. The injection molding apparatus according to any one of claims 1 to 19, wherein the output section issues a maintenance alert regarding the gate portion when it determines that the gate portion has reached a state requiring maintenance based on a value representing the wear state of the gate.

Images