Anomaly detection system, anomaly detection device, anomaly detection method, program, and learned model

Abstract

To provide an abnormality detection system, an abnormality detection device, an abnormality detection method, a program, and a learnt model for securely detecting a clogging of a gate of a molding device.SOLUTION: An abnormality detection system includes a molding device and an abnormality detection device for detecting an abnormality of the molding device. The molding device includes a mold part provided with a plurality of gates and a mold cavity connecting with the plurality of gates, formed inside, an injection part for filling a molding material in a molten state to the mold cavity via the plurality of gates, a first sensor closest to a first gate among the plurality of gates for detecting a pressure or a temperature of the molding material in the mold cavity, and a second sensor closest to a second gate for detecting a pressure or a temperature of the molding material in the mold cavity. The abnormality detection device includes a data acquisition unit for acquiring a first evaluation value of the pressure or the temperature detected by the first sensor and a second evaluation value of the pressure or the temperature detected by the second sensor and an abnormality detection unit for detecting a clogging in the first gate or the second gate based on the first evaluation value and the second evaluation value.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an abnormality detection system, an abnormality detection device, an abnormality detection method, a program, and a learned model.

Background Art

[0002] There is known an injection molding apparatus that forms a molded product by supplying a molten liquid of a molding material to a cavity formed between a plurality of molds. For example, Patent Document 1 discloses a technique in which an injection speed is detected by an encoder attached to an injection servo motor, and when the injection speed has not reached a monitoring speed during a monitoring time of an injection process, the injection operation is stopped.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In a molding apparatus, when a gate becomes clogged, the molding material may not flow into the cavity as designed, and an abnormality may occur in the molded product. In the technique of Patent Document 1, when the gate becomes clogged, the load pressure during injection becomes higher than the load pressure during normal product molding, and the injection speed becomes slower than the injection speed during normal product molding, and this is used to detect clogging of the gate.

[0005] However, the inventors have found that even when the gate becomes clogged, the load pressure during injection may hardly change from that during normal product molding. In such a case, the technique of Patent Document 1 cannot detect clogging of the gate.

[0006] Therefore, an object of the present invention is to provide an abnormality detection system, an abnormality detection device, an abnormality detection method, a program, and a learned model that can more reliably detect gate clogging.

Means for Solving the Problems

[0007] (1) The abnormality detection system of the present disclosure is an abnormality detection system including a molding device for molding a molded product and an abnormality detection device for detecting an abnormality of the molding device. The molding device includes a mold part having a plurality of gates and a cavity connected to the plurality of gates formed therein, an injection part for filling the cavity with a molten molding material via the plurality of gates, a first sensor provided in a region closest to a first gate among the plurality of gates for detecting the pressure or temperature of the molding material in the cavity, and a second sensor provided in a region closest to a second gate different from the first gate among the plurality of gates for detecting the pressure or temperature of the molding material in the cavity. The abnormality detection device includes a data acquisition part for acquiring a first evaluation value calculated based on first time-series data of the pressure or temperature detected by the first sensor and a second evaluation value calculated based on second time-series data of the pressure or temperature detected by the second sensor, and an abnormality detection part for detecting clogging of the first gate or the second gate based on the first evaluation value and the second evaluation value.

[0008] When the first gate or the second gate is clogged, a difference occurs between the first time-series data and the second time-series data. To make it easier to compare this difference, the first time-series data is quantified as the first evaluation value, the second time-series data is quantified as the second evaluation value, and clogging of the first gate or the second gate is detected based on these first evaluation value and second evaluation value. As a result, clogging of some gates that could not be detected by conventional detection can be detected. As a result, gate clogging can be detected more reliably.

[0009] (2) Preferably, a first distance through which the molding material passes from the first gate to the first sensor is equal to a second distance through which the molding material passes from the second gate to the second sensor.

[0010] By configuring in this way, when the gate is not clogged, the time points at which the molding material reaches the first sensor and the second sensor can be made to coincide. As a result, a difference can be made to occur only between the first time-series data and the second time-series data when some of the gates are clogged. Consequently, the accuracy of clogging detection can be improved.

[0011] (3) Preferably, the first evaluation value includes a first rising time point of the first time-series data or a first peak time point at which the pressure or temperature of the first time-series data reaches a maximum value, the second evaluation value includes a second rising time point of the second time-series data or a second peak time point at which the pressure or temperature of the second time-series data reaches a maximum value, and the abnormality detection unit detects clogging of the first gate when the first rising time point is later than the second rising time point or when the first peak time point is later than the second peak time point.

[0012] (4) Preferably, the first sensor and the second sensor detect the pressure of the molding material, the first evaluation value includes a first peak value that is the maximum value of the pressure of the first time-series data, the second evaluation value includes a second peak value that is the maximum value of the pressure of the second time-series data, and the abnormality detection unit detects clogging of the first gate when the first peak value is lower than the second peak value.

[0013] (5) Preferably, the first sensor and the second sensor detect the pressure of the molding material, the first evaluation value includes a first integration value that is the time integral from the first rising point of the pressure of the first time-series data to the first peak point where the pressure of the first time-series data reaches the maximum value, the second evaluation value includes a second integration value that is the time integral from the second rising point of the pressure of the second time-series data to the second peak point where the pressure of the second time-series data reaches the maximum value, and the abnormality detection unit detects clogging of the first gate when the first integration value is smaller than the second integration value.

[0014] According to the configurations of (3) to (5) above, the first evaluation value includes at least one of the first rising point, the first peak point, the first peak value, and the first integration value, and the second evaluation value includes at least one of the second rising point, the second peak point, the second peak value, and the second integration value corresponding to the first evaluation value. Then, by comparing the first rising point and the second rising point, the first peak point and the second peak point, the first peak value and the second peak value, or the first integration value and the second integration value, it is possible to identify that the first gate among the first gate and the second gate is clogged. That is, according to the abnormality detection system, not only can clogging of some gates that could not be detected by conventional detection be detected, but it is also possible to identify which gate is clogged.

[0015] (6) Preferably, the abnormality detection unit inputs the first evaluation value and the second evaluation value into a learned model obtained by machine learning the correlation between the first evaluation value and the second evaluation value and clogging of the plurality of gates, thereby detecting clogging of at least one of the first gate and the second gate.

[0016] With such a configuration, after a learned model is once generated, it becomes possible to detect clogging of some gates from the first evaluation value and the second evaluation value. By using the learned model, even in a state where there are variations in the molding conditions, it is possible to more accurately detect clogging of some gates.

[0017] (7) The abnormality detection device of the present disclosure is an abnormality detection device that detects an abnormality in a molding device that molds a molded product. The molding device includes a mold part having a plurality of gates and a cavity connected to the plurality of gates formed therein, an injection part that fills the cavity with a molten molding material via the plurality of gates, a first sensor provided in a region closest to a first gate among the plurality of gates and detecting the pressure or temperature of the molding material in the cavity, and a second sensor provided in a region closest to a second gate different from the first gate among the plurality of gates and detecting the pressure or temperature of the molding material in the cavity. The abnormality detection device has a data acquisition part that acquires a first evaluation value calculated based on first time-series data of the pressure or temperature detected by the first sensor and a second evaluation value calculated based on second time-series data of the pressure or temperature detected by the second sensor, and an abnormality detection part that detects clogging of at least one of the first gate and the second gate based on the first evaluation value and the second evaluation value.

[0018] When the first gate or the second gate is clogged, a difference occurs between the first time-series data and the second time-series data. To make it easier to compare this difference, the first time-series data is quantified as a first evaluation value, the second time-series data is quantified as a second evaluation value, and clogging of the first gate or the second gate is detected based on these first and second evaluation values. As a result, clogging of some gates that could not be detected conventionally can be detected. Consequently, gate clogging can be detected more reliably.

[0019] (8) The abnormal detection method of the present disclosure is an abnormal detection method for detecting an abnormality in a molding apparatus including a mold part in which a plurality of gates and a cavity connected to the plurality of gates are formed inside, and an injection part for filling a molten molding material into the cavity via the plurality of gates. The method includes: a data acquisition step of acquiring a first evaluation value calculated based on first time-series data of the pressure or temperature of the molding material in the cavity detected by a first sensor provided in a region closest to a first gate among the plurality of gates, and a second evaluation value calculated based on second time-series data of the pressure or temperature of the molding material in the cavity detected by a second sensor provided in a region closest to a second gate different from the first gate among the plurality of gates; and an abnormal detection step of detecting clogging of at least one of the first gate and the second gate based on the first evaluation value and the second evaluation value.

[0020] When the first gate or the second gate is clogged, a difference occurs between the first time-series data and the second time-series data. To facilitate comparison of this difference, the first time-series data is quantified as a first evaluation value, the second time-series data is quantified as a second evaluation value, and clogging of the first gate or the second gate is detected based on these first and second evaluation values. Thereby, clogging of some gates that could not be detected by conventional detection can be detected. As a result, gate clogging can be detected more reliably.

[0021] (9) The program of the present disclosure is a program for detecting an abnormality in a molding apparatus including a mold part in which a plurality of gates and a cavity connected to the plurality of gates are formed inside, and an injection part for filling the cavity with a molten molding material via the plurality of gates. Based on first time-series data of the pressure or temperature of the molding material in the cavity detected by a first sensor provided in a region closest to a first gate among the plurality of gates, a first evaluation value is calculated. And a second evaluation value calculated based on second time-series data of the pressure or temperature of the molding material in the cavity detected by a second sensor provided in a region closest to a second gate different from the first gate among the plurality of gates. A data acquisition step of acquiring the first evaluation value and the second evaluation value, and an abnormality detection step of detecting clogging of at least one of the first gate and the second gate based on the first evaluation value and the second evaluation value. It is a program that causes a computer device to execute.

[0022] When the first gate or the second gate is clogged, a difference occurs between the first time-series data and the second time-series data. To make it easier to compare this difference, the first time-series data is quantified as a first evaluation value, the second time-series data is quantified as a second evaluation value, and clogging of the first gate or the second gate is detected based on these first evaluation value and second evaluation value. As a result, clogging of some gates that could not be detected conventionally can be detected. As a result, gate clogging can be detected more reliably.

[0023] (10) The learned model of the present disclosure is a learned model for detecting an abnormality of a molding apparatus including a mold part in which a plurality of gates and a cavity connected to the plurality of gates are formed inside, and an injection part for filling the cavity with a molten molding material via the plurality of gates. The explanatory variables include a first evaluation value calculated based on first time-series data of the pressure or temperature of the molding material in the cavity detected by a first sensor provided in a region closest to a first gate among the plurality of gates, and a second evaluation value calculated based on second time-series data of the pressure or temperature of the molding material in the cavity detected by a second sensor provided in a region closest to a second gate different from the first gate among the plurality of gates. The objective variable includes a value related to a clogged gate among the plurality of gates.

[0024] When the first gate or the second gate is clogged, a difference occurs between the first time-series data and the second time-series data. To make it easier to compare this difference, the first time-series data is quantified as a first evaluation value, the second time-series data is quantified as a second evaluation value, and clogging of the first gate or the second gate is detected based on these first and second evaluation values. Thereby, clogging of some gates that could not be detected conventionally can be detected. As a result, gate clogging can be detected more reliably.

Advantages of the Invention

[0025] According to the present invention, gate clogging can be detected more reliably.

Brief Description of the Drawings

[0026]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Mode for Carrying Out the Invention

[0027] <First Embodiment> Hereinafter, a first embodiment of the present invention will be described with reference to the drawings.

[0028] <Overall Configuration of Abnormality Detection System> FIG. 1 is a block diagram schematically showing an abnormality detection system 10 according to the first embodiment. The abnormality detection system 10 includes a plurality of molding devices 20, a learning device 30, an abnormality detection device 40, an input unit 50, and a display unit 60.

[0029] The molding device 20, the learning device 30, the abnormality detection device 40, the input unit 50, and the display unit 60 are each provided so as to be communicable wirelessly or by wire. The learning device 30 and the abnormality detection device 40 are configured by an information processing device (computer device) having an arithmetic unit (for example, CPU, GPU, etc.) and a storage unit (for example, HDD, SSD, etc.). The learning device 30 and the abnormality detection device 40 may be configured by the same information processing device or by separate information processing devices.

[0030] In this embodiment, a plurality of molding devices 20 are connected to one learning device 30 and one abnormality detection device 40, and the learning device 30 and the abnormality detection device 40 perform learning and abnormality detection based on various data transmitted from the plurality of molding devices 20. Note that the molding device 20 may correspond one-to-one with the learning device 30 and the abnormality detection device 40. That is, in the abnormality detection system 10, there may be one molding device 20, or there may be a plurality of learning devices 30 and abnormality detection devices 40.

[0031] The input unit 50 is, for example, a keyboard or a mouse, and receives various inputs from an operator. The display unit 60 is, for example, a display or a speaker, and displays various information in the abnormality detection system 10. The input unit 50 and the display unit 60 may be integrated, for example, like a touch panel. Further, the input unit 50 and the display unit 60 may be provided as a portable terminal device so as to be movable to a location away from the molding device 20, the learning device 30, and the abnormality detection device 40.

[0032] <Schematic Configuration of Molding Device> FIGS. 2 and 3 are explanatory views conceptually showing the molding device 20. In FIGS. 2 and 3, hatching is applied to the portions shown as cross-sections. The molding device 20 includes a bed 21, an injection unit 22, a clamping unit 23, a mold unit 24, a plurality of sensors 25, a temperature sensor 26, and a control panel 27. FIG. 2 shows the molding device 20 in a state where the mold unit 24 is open, and FIG. 3 shows the molding device 20 in a state where the mold unit 24 is combined. The molding device 20 is a device that performs clamping type injection molding.

[0033] The control panel 27 includes a control unit 271 and a communication unit 272. The control unit 271 is electrically connected to each drive unit (such as the motor 237) of the molding device 20 and outputs an operation command to each of the drive units. Further, the control unit 271 is electrically connected to each sensor (such as the sensor 25) of the molding device 20, and a signal detected by each of the sensors is input to the control unit 271. The control unit 271 is configured by an information processing device having an arithmetic unit (for example, a CPU, a GPU, etc.) and a storage unit (for example, an HDD, an SSD, etc.).

[0034] The communication unit 272 communicates with other parts (such as the learning device 30) of the abnormality detection system 10. The communication unit 272 transmits, for example, a signal detected by each of the sensors to the learning device 30 or the abnormality detection device 40. Further, the communication unit 272 receives detection information described later from the abnormality detection device 40.

[0035] The clamping unit 23 includes a fixed platen 231, a movable platen 232, a tie bar 233, a ball screw 234, a support platen 235, a clamping force sensor 236, and a motor 237. The fixed platen 231 and the support platen 235 are fixed to the bed 21. The support platen 235 supports the ball screw 234. The ball screw 234 is connected to the motor 237. When the motor 237 is rotated by an operation command of the control unit 271, the ball screw 234 moves. A movable platen 232 is fixed to an end of the ball screw 234 on the side opposite to the end connected to the motor 237.

[0036] Here, in the molding device 20, the direction in which the ball screw 234 moves is referred to as the "axial direction". The side on which the motor 237 is located with respect to the ball screw 234 is referred to as the "one side" of the axial direction, and the side on which the movable platen 232 is located with respect to the ball screw 234 is referred to as the "other side" of the axial direction.

[0037] The movable plate 232 moves axially as the ball screw 234 moves. A through hole 232a penetrating axially is formed in the movable plate 232. The tie bar 233 has one end fixed to the support plate 235 on one side in the axial direction and the other end fixed to the fixed plate 231 on the other side in the axial direction. The tie bar 233 is inserted into the through hole 232a of the movable plate 232. Thereby, the tie bar 233 guides the axial movement of the movable plate 232.

[0038] The clamping force sensor 236 detects the pressure (the reaction force of the clamping force) applied from the ball screw 234 to the support plate 235. The clamping force sensor 236 outputs a detection signal related to the pressure to the control unit 271. Note that the clamping force sensor 236 may be installed at other positions as long as it can detect the clamping force in the mold part 24 described later. A through hole 231a with a widened diameter is formed in the fixed plate 231 on the other side in the axial direction. The cylinder 222 described later is inserted into the through hole 231a.

[0039] The mold part 24 has a plurality of molds 241 and 242. The mold 241 is fixed to the movable plate 232. When the ball screw 234 moves axially, the mold 241 also moves axially together with the movable plate 232. That is, the mold 241 is a movable mold. The mold 242 is fixed to the fixed plate 231. That is, the mold 242 is a fixed mold.

[0040] Referring to FIG. 3, when the mold 241 moves axially to the other side by the ball screw 234 and the mold 241 contacts the mold 242 (that is, when the plurality of molds 241 and 242 are combined), a cavity C1 is formed between the molds 241 and 242. The cavity C1 of the present embodiment is an annular space filled with a molding material.

[0041] Figure 4 is a cross-sectional view showing an enlarged view of the mold part 24 shown in Figure 3. In the mold 242, a sprue 243, a plurality (four in this embodiment) of runners 244 branching from the sprue 243, and a plurality (four in this embodiment) of gates 245 respectively connected to the plurality of runners 244 are formed inside. The cavity C1 is connected to the plurality of gates 245.

[0042] The sprue 243 is a passage through which the molding material injected from the injection part 22 into the mold part 24 first flows. The end 243a of the sprue 243 on the injection part 22 side is open so as to communicate with a nozzle 224 described later. The sprue 243 has the smallest inner diameter at the end 243a, and the inner diameter expands from the end 243a to the runner 244.

[0043] The plurality of runners 244 branch slightly on the end 243a side rather than on the end 243b side of the cavity C1 side (that is, the opposite side of the end 243a) of the sprue 243. The space on the end 243b side of the sprue 243 from the position where the runner 244 branches is a resin reservoir for storing cold slugs, and is also referred to as a "cold slug well".

[0044] Figure 5 is a cross-sectional view showing the mold part 24 cut along the cutting line shown by the arrow V in Figure 4. The plurality of gates 245 correspond one-to-one with the plurality of runners 244. Also, the inner diameter of the plurality of gates 245 decreases as it approaches the cavity C1. Hereinafter, when particularly distinguishing the plurality of gates 245, a predetermined one gate 245 (for example, the gate 245 located above in Figure 5) is referred to as gate GT1, and starting from this gate GT1 in the clockwise direction, the remaining three gates 245 are respectively referred to as gates GT2, GT3, and GT4.

[0045] When the position of the gate GT1 in the mold 24 is set to 0 degrees, the gate GT2 is at 90 degrees, the gate GT3 is at 180 degrees, and the gate GT4 is at 270 degrees. The gate GT3 is located on the opposite side of the gate GT1, and the gate GT4 is located on the opposite side of the gate GT2.

[0046] The plurality of sensors 25 are provided at positions facing the cavity C1 of the molds 241 and 242. The plurality of sensors 25 each detect the pressure of the molding material in the cavity C1. Note that the sensor 25 may be a sensor that detects temperature. The sensor 25 outputs a detection signal regarding pressure (or temperature) to the control unit 271. In the present embodiment, as an example, the plurality of sensors 25 are provided in the mold 241, but the sensors 25 may be provided in the mold 242.

[0047] The plurality of sensors 25 are provided at positions corresponding to the plurality of gates 245, respectively. In the present embodiment, four sensors 25 correspond one-to-one to the four gates 245. Hereinafter, when particularly distinguishing, the sensor 25 corresponding to the gate GT1 among the plurality of sensors 25 is referred to as the sensor SN1. Similarly, the sensors 25 corresponding to the gates GT2, GT3, and GT4 are referred to as the sensors SN2, SN3, and SN4, respectively.

[0048] The sensor SN1 (corresponding to the "first sensor" in the present disclosure) is provided in the region closest to the gate GT1 (corresponding to the "first gate" in the present disclosure) among the plurality of gates 245. That is, the distance from the sensor SN1 to the gate GT1 is shorter than the distances from the sensor SN1 to the other gates GT2, GT3, and GT4, respectively. In other words, the sensor 25 closest to the gate GT1 is the sensor SN1.

[0049] Similarly, the sensors SN2, SN3, and SN4 (all corresponding to the "second sensor" in the present disclosure) are provided in the regions closest to the gates GT2, GT3, and GT4 (all corresponding to the "second gate" in the present disclosure) among the plurality of gates 245.

[0050] The first distance d1 through which the molding material passes from gate GT1 to sensor SN1 is equal to the second distance d2 through which the molding material passes from gate GT2 to sensor SN2 (d1 = d2). Also, the first distance d1 is equal to both the third distance d3 through which the molding material passes from gate GT3 to sensor SN3 and the fourth distance d4 through which the molding material passes from gate GT4 to sensor SN4 (d1 = d3 = d4). That is, the distances from the plurality of sensors 25 to the nearest gate 245 are equal to each other.

[0051] Refer to FIG. 4. The temperature sensor 26 is built into the mold 242 and detects the temperature of the mold 242. The temperature sensor 26 outputs a detection signal regarding temperature to the control unit 271. Note that the temperature sensor 26 may be installed in the mold 241. Also, the temperature sensor 26 may be installed in the cylinder 222. That is, the temperature sensor 26 only needs to be able to directly or indirectly detect the temperature of the molding material supplied into the cavity C1.

[0052] Refer to FIG. 2. The injection unit 22 includes a hopper 221, a cylinder 222, a screw 223, a ball screw 225, a motor 226, a pressure sensor 227, a movement amount sensor 228, and a heater 229. The hopper 221 is connected to the cylinder 222 and supplies the molding material into the cylinder 222. The cylinder 222 is a member having a hollow cylindrical shape extending in the axial direction. One end on the axial direction side of the cylinder 222 has a diameter that narrows as it approaches the outermost end on one side in the radial direction, and a nozzle 224 is provided at the outermost end on one side in the axial direction of the cylinder 222. The nozzle 224 is connected to the sprue 243 of the mold 242.

[0053] The screw 223 is inserted into the cylinder 222 from the other end on the axial direction side of the cylinder 222. A ball screw 225 is connected to the other end on the axial direction side of the screw 223, and a motor 226 is connected to the other end on the axial direction side of the ball screw 225. When the motor 226 rotates according to an operation command from the control unit 271, the ball screw 225 moves in the axial direction. Along with this, the screw 223 also moves in the axial direction. At this time, the screw 223 rotates in the circumferential direction with the axial direction as the central axis.

[0054] The pressure application sensor 227 detects the pressure applied from the ball screw 225 to the motor 226 (the reaction force of the pushing force of the screw 223). That is, the pressure application sensor 227 detects the pressure that the screw 223 receives from the molding material. The pressure application sensor 227 outputs a detection signal related to the pressure to the control unit 271. Note that the pressure application sensor 227 may be installed at other positions as long as it can detect the pushing force of the screw 223.

[0055] The displacement sensor 228 detects the displacement in the axial direction of the ball screw 225. The displacement sensor 228 outputs a detection signal related to the displacement to the control unit 271. Note that the displacement sensor 228 may be installed at other positions as long as it can detect the displacement in the axial direction of the ball screw 225.

[0056] The heater 229 is, for example, a resistance heating heater in which a resistance wire is wound in a coil shape. When an electric current is passed through the resistance wire according to an operation command of the control unit 271, the heater 229 heats the inside of the cylinder 222 by resistance heat.

[0057] <Manufacturing method by the molding apparatus> With appropriate reference to FIGS. 2 and 3, a manufacturing method of a molded product by the molding apparatus 20 will be described. The manufacturing method of the molded product by the molding apparatus 20 includes a pre-process ST1, a mold clamping process ST2, a filling process ST3, a pressure holding process ST4, a pressure holding release process ST5, and a mold release process ST6, which are executed in this order. In the present embodiment, the molded product is a resin-made cage used for a rolling bearing. However, this is an example of the molded product, and the molded product molded by the molding apparatus according to the present invention may be a molded product having other shapes and uses.

[0058] Refer to Fig. 2. First, the previous process ST1 is executed. In the previous process ST1, while the screw 223 is rotated by the motor 226 and the inside of the cylinder 222 is heated by the heater 229, pellets of the molding material are supplied from the hopper 221 into the cylinder 222. The pellets of the molding material are melted in the cylinder 222 by the frictional heat generated by the rotation of the screw 223 and the heating by the heater 229, and become a molten molding material (melting process).

[0059] Next, while the screw 223 rotates and moves axially to the other side, a predetermined amount of the molding material is stored in the cylinder 222 on the one side in the axial direction of the screw 223 (measurement process). Thus, the previous process ST1 ends.

[0060] Next, when the mold clamping process ST2 is started, in the molding apparatus 20 in the state of Fig. 2, the ball screw 234 moves axially to the other side according to the operation command of the control unit 271, and the mold 241 is brought into contact with the mold 242 as shown in Fig. 3. In the state where the mold 241 and the mold 242 are combined in this way, the ball screw 234 further presses the mold 241 against the mold 242 with a predetermined mold clamping force axially to the other side. That is, a plurality of molds 241 and 242 are clamped. As a result, an annular cavity C1 is formed between the plurality of molds 241 and 242. Thus, the mold clamping process ST2 ends.

[0061] Here, the mold clamping force is one of the molding conditions and is determined according to other molding conditions such as the shapes of the molds 241 and 242. The mold clamping force is detected by the mold clamping force sensor 236.

[0062] Subsequently, when the filling process ST3 is started, while maintaining the above-mentioned mold clamping force, the ball screw 225 moves axially to the one side. As a result, the screw 223 pushes the molding material axially to the one side, and the molten molding material is injected into the cavity C1 from the nozzle 224 of the cylinder 222 via the sprue 243, the plurality of runners 244, and the plurality of gates 245 (filling operation).

[0063] When the cavity C1 is filled with the molding material, the filling process ST3 ends. In the filling process ST3, the molten molding material is supplied to the cavity C1 while gradually solidifying from the vicinity of the surfaces of the molds 241 and 242.

[0064] Subsequently, when the pressure holding process ST4 is started, the screw 223 further pushes the molding material axially to one side, and the molding material is further injected from the nozzle 224 of the cylinder 222 into the cavity C1. As a result, a predetermined pressure (for example, several tens to several hundreds of MPa) is applied to the molding material filled in the cavity C1. Then, the screw 223 holds this state for a predetermined time, thereby continuously applying a predetermined pressure to the molding material for a predetermined time (for example, several seconds) (pressure holding operation). The pressure (applied pressure) at which the screw 223 extrudes the molding material into the cavity C1 is detected by the applied pressure sensor 227.

[0065] Subsequently, when the pressure release process ST5 is started, the screw 223 moves axially to the other side to release the holding of the pressure of the molding material (pressure release operation). After the pressure release operation, when a predetermined time has elapsed and the pressure of the molding material in the cavity C1 becomes equal to or less than a predetermined value, the pressure release process ST5 ends. Thereafter, when the mold release process ST6 is started, the mold part 24 is cooled, so that the molding material in the cavity C1 solidifies and a molded product is formed. Then, the ball screw 234 moves axially to one side, and the mold 241 separates from the mold 242, so that the molded product is taken out. Note that the cooling of the mold part 24 may be started simultaneously with the pressure release process ST5.

[0066] <Regarding clogging of the gate> FIG. 6 is a diagram for explaining the clogging of the gate 245 detected in the present embodiment. In FIG. 6, as an example of clogging, a state in which the gate GT1 is clogged by the foreign matter Fm1 is shown. FIG. 6(a) is a diagram showing the state during the filling process ST3, and FIG. 6(b) is a diagram showing the state during the pressure holding process ST4. FIGS. 6(a) and 6(b) both show the same cross section as FIG. 5.

[0067] As shown in FIG. 6(a), when the filling process ST3 is executed, the molding material L1 flows into the cavity C1 via the spool 243, the runner 244, and the gate 245. Here, since the gate GT1 is clogged by the foreign matter Fm1, the molding material L1 does not flow into the cavity C1 from the gate GT1, and the molding material L1 flows into the cavity C1 from the other gates GT2, GT3, and GT4.

[0068] Conventionally, no sensor 25 was provided in the mold part 24, and the pressure of the molding material L1 was monitored by the pressure boosting sensor 227 provided in the injection part 22. For example, when all the gates 245 are clogged, the pressure increase of the pressure boosting sensor 227 in the filling process ST3 becomes clearly earlier than that in the normal state (when there is no clogging), so it is possible to detect the clogging of the gate 245 even with the conventional configuration.

[0069] However, when some of the gates 245 (particularly, only less than half of the plurality of gates 245) are clogged, as shown in FIG. 6(a), since the molding material L1 flows into the cavity C1 from the other unclogged gates GT2, GT3, and GT4, the pressure detected by the pressure boosting sensor 227 hardly differs from the normal pressure. For this reason, it is impossible to detect the clogging of some of the gates 245 based on the pressure of the pressure boosting sensor 227.

[0070] Also, as shown in FIG. 6(b), finally, the molding material L1 flowing in from the other gates GT2, GT3, and GT4 flows into the cavity C1 near the gate GT1, so the cavity C1 is filled with the molding material L1. As a result, even if an abnormality occurs in which some of the gates 245 are clogged, a molded product can be produced for the time being.

[0071] However, since the molding material L1 does not flow into the cavity C1 near the clogged gate GT1 as usual, the dimensions (e.g., roundness), weight, and quality (e.g., strength) of the molded product may deviate from the design range. For example, a weld area may be formed in an unintended area (an area other than the designed area), which may reduce the strength of the molded product. A molded product that deviates from the design range becomes a defective product. In addition, since the density of the molding material L1 is low near the gate GT1, there is also a risk of abnormalities such as sink marks and voids occurring in the molded product.

[0072] Such abnormalities in the molded product have conventionally been detected by measuring the dimensions, appearance, weight, etc. of the molded product. In the molding apparatus 20, since the molding of the molded product is performed continuously, in the conventional detection method, abnormal molded products continue to be molded uselessly until after the molding of the molded product and until the abnormality is detected by measuring the molded product outside the molding apparatus 20. For this reason, a technique that can sense clogging of some of the gates 245 in the molding apparatus 20 and determine the abnormality immediately after molding without measuring the molded product is important.

[0073] Therefore, the inventors conceived of providing a plurality of sensors 25 at positions corresponding to the plurality of gates 245 in the cavity C1, and detecting clogging of some of the gates 245 based on the difference in the time-series data of the pressure detected by these plurality of sensors 25. Hereinafter, the difference in the time-series data caused by clogging of some of the gates 245 will be described.

[0074] <Difference in time-series data due to clogging of some gates> FIG. 7 is an example of a graph showing the time-series data of the pressure detected by the plurality of sensors 25. In FIG. 7, the vertical axis represents pressure and the horizontal axis represents time. In the present embodiment, the origin of time (t = 0) is the point in time when the control unit 271 issues an operation command for the motor 226 to move to the other axial side of the screw 223. That is, it is the point in time when the injection unit 22 starts supplying the molding material to the mold unit 24 in the filling process ST3. Note that the origin of time (t = 0) is not limited to this, and it may be the point in time when the pressure sensor 227 detects a predetermined pressure, or the point in time when the movement amount sensor 228 detects a predetermined movement amount.

[0075] The graph line F11 in FIG. 7 shows the time-series data of the pressure detected by the sensor SN1 (corresponding to the "first time-series data" of the present disclosure). Similarly, the graph line F12 shows the time-series data of the pressure detected by the sensor SN2, the graph line F13 shows the time-series data of the pressure detected by the sensor SN3, and the graph line F14 shows the time-series data of the pressure detected by the sensor SN4 (all corresponding to the "second time-series data" of the present disclosure). Also, the graph line AvF1 is the time-series data of the average value of the pressures of the four graph lines F11 to F14. The graph lines F11 to F14 may be such that the pressures detected by the sensors SN1 to SN4 at each predetermined time are plotted as they are, or the moving average (for example, a 5-point moving average) of the pressures detected at each predetermined time may be plotted.

[0076] Although the rising positions of the graph lines and the peak values etc. are different respectively, the overall trends of the graph lines F11 to F14 are the same. When the filling process ST3 is started, the pressures of the graph lines F11 to F14 rise and then increase monotonically. This is because the molding material L1 is continuously supplied from the injection unit 22 and the molding material L1 gradually presses the sensors SN1 to SN4 more strongly.

[0077] The monotonic increase of the graph lines F11 to F14 continues until midway through the pressure holding step ST4. The pressures of the graph lines F11 to F14 reach their maximum values midway through the pressure holding step ST4, and then monotonically decrease until the end of the pressure release step ST5. The monotonic increase until midway through the pressure holding step ST4 is due to additional pressure being applied from the injection part 22 to the molding material L1 filled in the cavity C1. Also, the monotonic decrease in pressure starting midway through the pressure holding step ST4 is due to the gate 245 being sealed by the solidified molding material, the application of pressure from the injection part 22 to the cavity C1 stopping, and the molding material L1 in the cavity C1 cooling and shrinking.

[0078] Next, focus on the individual trends of the graph lines F11 to F14. First, the graph line F11 of the sensor SN1 corresponding to the clogged gate GT1 rises later than the other graph lines F12 to F14. This is because, as shown in FIG. 6, since the molding material L1 wraps around from the gates GT2 to GT4 to the position of the sensor SN1, the time when the molding material L1 reaches the sensor SN1 is later than the time when it reaches the other sensors SN2 to SN4.

[0079] Also, for the same reason, the peak time Xt1 when the pressure of the graph line F11 reaches the maximum value Pt1 is later than the peak times Xt2 to Xt4 of the other graph lines F12 to F14 (FIG. 7 typically shows the peak time Xt3 when the pressure of the graph line F13 reaches the maximum value Pt3).

[0080] Here, the time when the pressure rises can be defined as appropriate, but in this embodiment, the time when a predetermined pressure Ps1 is reached is defined as the rising time. The predetermined pressure Ps1 is, for example, a value determined based on the average value AvPt of the maximum values of the pressures of the graph lines F11 to F14, and more specifically, a value that is a predetermined percentage (for example, 10 to 20%) of the average value AvPt. Note that in each of the graph lines F11 to F14, the time when the slope of the pressure reaches a predetermined value may be used as the rising time for each.

[0081] As shown in Fig. 7, the rising point Xs1 of the graph line F11 is later than the rising points Xs2 to Xs4 of the other graph lines F12 to F14 (in Fig. 7, the rising point Xs3 of the graph line F13 is typically shown). Also, the rising point Xs1 is later than the rising point AXs of the graph line AvF1.

[0082] The maximum pressure value Pt1 of the graph line F11 becomes lower than the maximum pressure values Pt2 to Pt4 of the other graph lines F12 to F14 (in Fig. 7, the maximum pressure value Pt3 of the graph line F13 is typically shown) and the maximum value AvPt of the graph line AvF1. This is due to the fact that the molding material L1 flowing into the cavity C1 contracts as it is cooled by the mold part 24. Since the gate GT1 is clogged, the time it takes for the molding material L1 to reach the sensor SN1 becomes longer than the time it takes for the molding material L1 to reach the other sensors SN2 to SN4, and the contraction of the molding material L1 progresses for that amount of time. For this amount of contraction, the force with which the molding material L1 presses the sensor SN1 becomes weaker than the force pressing the other sensors SN2 to SN4, so the maximum pressure value Pt1 of the graph line F11 becomes lower than the maximum values of the other graph lines F12 to F14.

[0083] Based on the above, when any of the following tendencies (1) to (3) appear, there is a high possibility that the gate GT1 is clogged. (1) The rising point Xs1 of the graph line F11 is delayed compared to the rising points Xs2 to Xs4 of the other graph lines F12 to F14 (2) The peak point Xt1 of the graph line F11 is delayed compared to the peak points Xt2 to Xt4 of the other graph lines F12 to F14 (3) The maximum pressure value Pt1 of the graph line F11 is lower than the maximum pressure values Pt2 to Pt4 of the other graph lines F12 to F14

[0084] Thus, by focusing on the differences in the time-series data detected by the plurality of sensors 25 corresponding to the plurality of gates 245 respectively, it is possible to detect that some of the gates 245 are clogged. In FIGS. 6 and 7, an example where the gate GT1 is clogged is described, but the same tendency appears when any of the other gates GT2 to GT4 is clogged. For example, when only the gate GT2 is clogged, the rising point Xs2 and the peak point Xt2 of the graph line F12 become relatively late, and the maximum value Pt2 of the pressure of the graph line F12 becomes relatively low.

[0085] Here, due to noise or the like, there may be a case where the pressure of the graph line F11 suddenly exceeds a predetermined pressure Ps1 at a time point earlier than the rising points Xs2 to Xs4. In this case, the rising point Xs1 of the graph line F11 becomes earlier than the rising points Xs2 to Xs4. Similarly, due to the influence of noise or the like, the maximum value Pt1 may suddenly exceed the other maximum values Pt2 to Pt4. In this case, if the comparison is made based only on the rising point and the maximum value, the clogging of some of the gates 245 cannot be correctly detected.

[0086] Therefore, in order to reduce the influence of noise or the like, instead of (or in addition to) comparing the rising point and the maximum value, the time integral values of the graph lines F11 to F14 may be compared. As the time integral value, for example, an integral value IV1 that is the time integral from the rising point Xs1 of the pressure of the graph line F11 to the peak point Xt1 is calculated. Similarly, for the graph lines F12 to F14, integral values IV2 to IV4 that are the time integrals from the rising points Xs2 to Xs4 to the peak points Xt2 to Xt4 are calculated. When the gate GT1 is clogged, the integral value IV1 is smaller than the integral values IV2 to IV4. Also, the integral value IV1 is smaller than the average value AIV of the integral values IV1 to IV4.

[0087] <Differences in time-series data when the sensor detects temperature> FIG. 8 is an example of a graph showing the time-series data of the temperature detected by the plurality of sensors 25. In FIG. 8, the vertical axis represents the temperature and the horizontal axis represents the time.

[0088] Here, FIG. 7 explains the difference in time-series data when a plurality of sensors 25 detect pressure. However, the plurality of sensors 25 may each detect temperature. When detecting pressure, the surface of the sensor 25 (or an indirect member that presses the surface of the sensor 25) needs to face the cavity C1 and contact the molding material L1. Therefore, if there is no space for exposing the sensor 25 in the cavity C1, it is difficult to provide the sensor 25 for detecting pressure.

[0089] On the other hand, if the sensor 25 detects temperature, even if the sensor 25 is not exposed in the cavity C1 (that is, even if it is buried in the mold 24 like the temperature sensor 26), the temperature of the molding material L1 can be detected by heat conduction. And even when the sensor 25 detects temperature, clogging of some gates 245 can be detected based on the difference in time-series data.

[0090] The graph line F21 in FIG. 8 shows the time-series data of the temperature detected by the sensor SN1 (corresponding to the "first time-series data" in the present disclosure). Similarly, the graph line F22 shows the time-series data of the temperature detected by the sensor SN2, the graph line F23 shows the time-series data of the temperature detected by the sensor SN3, and the graph line F24 shows the time-series data of the temperature detected by the sensor SN4 (all corresponding to the "second time-series data" in the present disclosure). Also, the graph line AvF2 is the time-series data of the average value of the temperatures of the four graph lines F21 to F24. The graph lines F21 to F24 may be the temperatures detected by the sensors SN1 to SN4 plotted as they are at each predetermined time, or may be the moving average (for example, 5-point moving average) of the temperatures detected at each predetermined time plotted.

[0091] Although the rising positions of the graph lines and peak values etc. are different respectively, the overall trends of the graph lines F21 to F24 are consistent. First, the graph lines F21 to F24 rise in temperature and then increase monotonically. This is due to the molding material L1 supplied to the cavity C1 heating the sensors SN1 to SN4. After the graph lines F21 to F24 reach their respective maximum values, they decrease monotonically. This is due to the supply of the molding material L1 from the injection part 22 to the cavity C1 stopping and the molding material L1 in the cavity C1 cooling down.

[0092] Next, focus on the individual trends of the graph lines F21 to F24. First, the rising point Xs1 (the point when reaching a predetermined temperature Ts1) and the peak point Xt1 of the graph line F21 are later than the rising points Xs2 to Xs4 and the peak points Xt2 to Xt4 of the other graph lines F22 to F24. Also, the rising point Xs1 is later than the rising point AXs of the graph line AvF2. This is for the same reason as the delay in the rise of the graph line F11 in FIG. 7, because the gate GT1 is clogged, and the time when the molding material L1 reaches the sensor SN1 is later than the time when it reaches the other sensors SN2 to SN4.

[0093] Based on the above, when considering the case based on the difference in the time-series data of temperature, when the following tendency of (1) or (2) appears, there is a high possibility that the gate GT1 is clogged. (1) The rising point Xs1 of the graph line F21 is delayed compared to the rising points Xs2 to Xs4 of the other graph lines F22 to F24 (2) The peak point Xt1 of the graph line F21 is delayed compared to the peak points Xt2 to Xt4 of the other graph lines F22 to F24

[0094] Thus, even when the sensor 25 detects temperature, by paying attention to the differences in the plurality of time-series data detected by the plurality of sensors 25 respectively corresponding to the plurality of gates 245, it is possible to detect that some of the gates 245 are clogged.

[0095] As described above, as a result of intensive research, the inventors found that when pressure or temperature is detected by a plurality of sensors 25 respectively corresponding to a plurality of gates 245, if clogging occurs in some of the gates 245, the time-series data detected by the sensor 25 (for example, sensor SN1) corresponding to the clogged gate 245 (for example, gate GT1) is different from the time-series data detected by other sensors 25 (for example, sensors SN2 to SN4).

[0096] Further, the inventors found that the difference in the time-series data appears particularly significantly in the rising points Xs1 to Xs4 of the pressure or temperature, the peak points Xt1 to Xt4 at which the pressure or temperature reaches the maximum value, the maximum values Pt1 to Pt4 of the pressure, and the integrated values IV1 to IV4 obtained by integrating the time from the rising point of the pressure to the peak point. Therefore, the inventors conceived an invention of quantifying these values as evaluation values and detecting clogging of some of the gates 245 based on the evaluation values.

[0097] The evaluation values are obtained, for example, for each of the plurality of sensors 25. In the present embodiment, four evaluation values R1 to R4 are calculated based on the time-series data detected by the four sensors 25 respectively. The evaluation value R1 (corresponding to the "first evaluation value" in the present disclosure) is a value calculated based on the time-series data (graph line F11 or F21) of the pressure or temperature detected by the sensor SN1, and is, for example, at least one value among the rising point Xs1, the peak point Xt1, the maximum value Pt1, and the integrated value IV1.

[0098] The evaluation values R2 to R4 (all corresponding to the "second evaluation value" in the present disclosure) are values calculated based on the time-series data of the pressure or temperature detected by the sensors SN2 to SN4. The evaluation values R1 to R4 are corresponding values respectively. For example, if the evaluation value R1 includes the rising point, the evaluation values R2 to R4 also include the rising point. If the evaluation value R1 includes the peak point, the evaluation values R2 to R4 also include the peak point.

[0099] In the abnormality detection system 10 according to the present embodiment, a learned model Tm1 in which the learning device 30 learns the correlation between a plurality of evaluation values R1 to R4 and the clogging of the gate 245 is generated, and the abnormality detection device 40 acquires detection information for detecting the clogging of the gate 245 based on the learned model Tm1. Hereinafter, the learning device 30 and the abnormality detection device 40 will be described.

[0100] <Description of the learning device> FIG. 9 is a block diagram showing the functional configuration of the learning device 30 according to the present embodiment. The learning device 30 includes a training data acquisition unit 31, a learning calculation unit 32, a shaping information storage unit 33, and a learned model storage unit 34. Each of these units is realized by a computer device having a calculation unit such as a CPU and a storage unit such as an HDD.

[0101] Various types of shaping information are stored in the shaping information storage unit 33. The shaping information is, for example, information in a table format associating various types of first information and second information. For example, when the first information is the type of mold, the second information includes various dimensions of the mold and the volume of the cavity C1. When the first information is the type of molding material or the lot number, the second information includes the physical properties (viscosity, moisture content, etc.) of the molding material.

[0102] The training data acquisition unit 31 acquires information related to training data from each unit of the abnormality detection system 10. The training data includes gate information, shaping information corresponding to the gate information, and a plurality of evaluation values R1 to R4. The training data is acquired, for example, based on data detected by each unit (for example, the sensor 25) of the abnormality detection system 10 when molding a molded product to be learned.

[0103] The gate information is a value related to the clogging of gate 245. More specifically, it is a value obtained by quantifying the way gate 245 is clogged. For example, the state where there is no clogging in all gates 245 is "0", the state where only gate GT1 is clogged is "1", the state where only gate GT2 is clogged is "2", the state where only gate GT3 is clogged is "3", and the state where only gate GT4 is clogged is "4", which are quantified as such.

[0104] The training data acquisition unit 31 acquires gate information based on the input of the operator. For example, for the training of the learning device 30, the operator artificially clogs a predetermined gate 245 and inputs to the input unit 50 about the way gate 245 is clogged. Thereby, the training data acquisition unit 31 acquires the gate information.

[0105] Also, the training data acquisition unit 31 acquires molding information based on the information input by the operator to the input unit 50 and the molding information storage unit 33. For example, when molding a molded product to be learned, the operator inputs the lot number of the molding material related to the molded product. The training data acquisition unit 31 acquires molding information regarding the physical properties (e.g., viscosity) of the molding material corresponding to the lot number from the molding information storage unit 33.

[0106] The molding device 20 uses a gate 245 without clogging and a gate 245 artificially clogged in a predetermined manner to mold a molded product a plurality of times respectively. Then, the training data acquisition unit 31 acquires a plurality of evaluation values R1 to R4 to which gate information is given as teacher data based on the time-series data of the pressure or temperature detected by the sensor 25. For example, "0" is given as gate information to the plurality of evaluation values R1 to R4 obtained when molding a molded product using a gate 245 without clogging.

[0107] Further, the training data acquisition unit 31 further acquires a plurality of environmental values as training data. The environmental value is, for example, a value related to the temperature detected by the temperature sensor 26 when molding the molded product to be learned. The environmental value may further include values related to the environment around the molding apparatus 20 and inside the molding apparatus 20 detected by the pressure application sensor 227, the clamping force sensor 236, and other sensors (e.g., humidity sensor) (not shown) when molding the molded product to be learned.

[0108] For example, every time one molded product to be learned is molded, the training data acquisition unit 31 acquires one set of training data (a set of gate information, molding information, a plurality of evaluation values R1 to R4, and environmental values acquired when molding the molded product). The training data acquisition unit 31 acquires a plurality of sets of training data for a predetermined number of times by molding the molded product to be learned a predetermined number of times.

[0109] Based on a plurality of sets of training data, the learning calculation unit 32 performs a calculation for performing supervised machine learning to generate a learned model Tm1 that models the correlation between the molding information, the plurality of evaluation values R1 to R4, the environmental values, and the gate information. In the present embodiment, a support vector machine (SVM) is used as the machine learning model, but other models may be used. For example, a convolutional neural network (CCN) may be used, or a regression tree model, which is a model related to data grouping, may be used.

[0110] Specifically, when generating the learned model Tm1, the correlation between the input information and the gate information is modeled by using the molding information, the plurality of evaluation values R1 to R4, and the environmental values (these information are collectively referred to as "input information") as explanatory variables and the gate information as the target variable. That is, the learning calculation unit 32 generates a learned model Tm1 that constitutes an intermediate layer for outputting the gate information corresponding to the input information from the output layer when the input information is input.

[0111] Here, the tendency of the learned model Tm1 will be described. Consider a case where a plurality of evaluation values R1 to R4 are the rising points Xs1 to Xs4 of pressure or temperature respectively. In this case, when the value of the evaluation value R1 input to the input layer is later than the other evaluation values R2 to R4 (the evaluation value R1 is a larger value than the other evaluation values R2 to R4), in the output layer, the accuracy of the gate information indicating that the gate GT1 is clogged (in the above example, "1") becomes high. Also, for example, when the value of the evaluation value R4 input to the input layer is later than the other evaluation values R1 to R3, in the output layer, the accuracy of the gate information indicating that the gate GT4 is clogged (in the above example, "4") becomes high.

[0112] Also, in the case where a plurality of evaluation values R1 to R4 are respectively the maximum values of pressure, when the value of the evaluation value R1 input to the input layer is smaller than the other evaluation values R2 to R4, in the output layer, the accuracy of the gate information indicating that the gate GT1 is clogged (in the above example, "1") becomes high.

[0113] Note that when generating the learned model Tm1, it is only necessary that the input information includes a plurality of evaluation values R1 to R4, and it is not necessary to include the molding information and the environmental value. Here, the more the contained moisture (molding information) of the molding material L1 is, or the higher the temperature (environmental value) detected by the temperature sensor 26 is, the lower the viscosity of the molding material L1 becomes, and the easier it is for the molding material L1 to flow. And when the molding material L1 becomes easier to flow, since the molding material L1 is filled into the entire cavity C1 earlier, the difference between the rising point Xs1 (evaluation value R1) and the rising point Xs3 (evaluation value R3) shown in FIG. 7 becomes smaller.

[0114] Thus, the plurality of evaluation values R1 to R4 also change depending on the molding information and the environmental value. Therefore, when the variations in the environmental value and the molding information are large, or when the influence of such variations on the plurality of evaluation values R1 to R4 is large, in order to more accurately predict the gate information corresponding to the input information, it is preferable to include the environmental value and the molding information in the input information when generating the learned model Tm1.

[0115] The learned model Tm1 generated by the learning calculation unit 32 is stored in the learned model storage unit 34. When new information is input to the learning device 30 and new training data is acquired by the training data acquisition unit 31, the learned model Tm1 stored in the learned model storage unit 34 is appropriately updated according to the content of the training data. Further, the learned model Tm1 is transmitted from the learning device 30 to the anomaly detection device 40 described later and is also stored in the learned model storage unit 45 of the anomaly detection device 40.

[0116] <Method for generating learned model> Next, a method for generating the learned model Tm1 by the learning device 30 will be described. The method for generating the learned model Tm1 includes a training data acquisition step and a learning calculation step. These steps are realized by a computer device constituting the learning device 30 executing a predetermined program.

[0117] First, when the training data acquisition step is started, the training data acquisition unit 31 acquires a plurality of sets of training data. For example, an operator artificially clogs a predetermined gate 245 and inputs to the input unit 50 how the gate 245 is clogged, whereby the training data acquisition unit 31 acquires gate information. Then, a molded product is molded using the gate 245, and based on the time-series data of the pressure or temperature detected by the sensor 25 at that time, a plurality of evaluation values R1 to R4 (for example, rising points Xs1 to Xs4) corresponding to the gate information are acquired. Thus, the training data acquisition step ends.

[0118] Next, when the learning calculation step is started, the learning calculation unit 32 learns the correspondence between the input information and the gate information based on a plurality of sets of training data and generates the learned model Tm1. The learned model Tm1 is stored in the learned model storage units 34 and 45. Thus, the learning calculation step ends.

[0119] <Explanation of anomaly detection device> FIG. 10 is a block diagram showing the functional configuration of the abnormality detection device 40 according to the present embodiment. The abnormality detection device 40 includes a data acquisition unit 41, an abnormality detection unit 42, an output unit 43, a molding information storage unit 44, and a learned model storage unit 45. Each of these units is realized by a computer device having an arithmetic unit such as a CPU and a storage unit such as an HDD. The arithmetic unit executes data acquisition processing and abnormality detection processing described below based on a program stored in the storage unit.

[0120] In the molding information storage unit 44, in the same manner as the molding information storage unit 33, molding information in a table format in which various pieces of first information and second information are associated with each other is stored. In the learned model storage unit 45, a learned model Tm1 generated by the learning device 30 is stored.

[0121] The molding information storage unit 44 and the learned model storage unit 45 may be realized by the same storage area as the molding information storage unit 33 and the learned model storage unit 34 of the learning device 30 in the computer device, or may be realized by another storage area. That is, the learning device 30 and the abnormality detection device 40 may be configured to share the same molding information storage unit 33 and learned model storage unit 34, or the learning device 30 and the abnormality detection device 40 may be configured to have independent molding information storage units 33, 44 and learned model storage units 34, 45, respectively.

[0122] The data acquisition unit 41 executes data acquisition processing for acquiring information for performing abnormality detection from each unit of the abnormality detection system 10. The information for performing abnormality detection is, for example, a set of environmental values, molding information, and a plurality of evaluation values R1 to R4 (hereinafter, these pieces of information are collectively referred to as "detection data set") acquired by the molding device 20 when molding a molded product to be detected.

[0123] The abnormality detection unit 42 inputs the detection data set acquired by the data acquisition unit 41 to the learned model Tm1. Based on these inputs, the learned model Tm1 outputs detection information D1 for detecting clogging of the gate 245.

[0124] The detection information D1 is information including the probability of each gate information. Specifically, the detection information D1 includes the first to fourth probabilities which are the probabilities of jams occurring in the gates GT1 to GT4 respectively, and the fifth probability which is the probability that no jam has occurred in all the gates 245. As an example, when a detection dataset is input to the learned model Tm1, the learned model Tm1 outputs detection information D1 with the first probability being 60%, the second probability being 10%, the third probability being 10%, the fourth probability being 10%, and the fifth probability being 10%.

[0125] Based on the detection information D1 acquired by the abnormality detection unit 42, the output unit 43 determines the way of jamming of the gates 245 (that is, among the plurality of gates 245, the gates 245 with a high possibility of jamming). For example, the output unit 43 obtains, as the determination result of the way of jamming of the gates 245, the way of jamming with the highest probability among the detection information D1. For example, when the first probability is the highest among the above first to fifth probabilities, the output unit obtains "the gate GT1 is jammed" as the determination result.

[0126] The output unit 43 outputs the determination result to the display unit 60 and the control unit 271. The determination result is displayed on the display unit 60. When it is determined that any of the plurality of gates 245 is jammed, the determination result may be displayed in an emphasized color such as red on the display of the display unit 60, and an alert may be issued by the speaker.

[0127] Also, in this case, the molding apparatus 20 including the gate 245 in which jamming is determined may be configured to stop with the mold part 24 in an open state by an operation command of the control unit 271. In this case, the operator inspects the mold part 24 based on an alert or the like by the display unit 60, and cleans or replaces the mold part 24 as necessary.

[0128] Note that in this embodiment, it may be configured such that the detection information D1 obtained by the abnormality detection unit 42 is directly displayed on the display unit 60 without providing the output unit 43. In this case, based on the detection information D1 displayed on the display unit 60, an operator may determine whether there is a blockage in the gate 245.

[0129] <Abnormality Detection Method by Abnormality Detection Device> Next, an abnormality detection method by the abnormality detection device 40 will be described. The abnormality detection method includes a data acquisition step and an abnormality detection step. These steps are realized by a computer device constituting the abnormality detection device 40 executing a predetermined program.

[0130] When the data acquisition step is started, the data acquisition unit 41 acquires a detection data set acquired when molding a molded product to be detected. Thus, the data acquisition step ends. Next, when the abnormality detection step is started, the abnormality detection unit 42 inputs the detection data set to the learned model Tm1 to acquire detection information D1. Next, the output unit 43 acquires a determination result based on the detection information D1. Finally, the output unit 43 outputs the determination result to the display unit 60 and the control unit 271. Thus, the abnormality detection step ends.

[0131] <Operation and Effect of Abnormality Detection System> The abnormality detection system 10 according to this embodiment acquires a plurality of evaluation values R1 to R4 based on the time-series data of the pressure detected by the sensor 25, and inputs the plurality of evaluation values R1 to R4 to the learned model Tm1 to acquire detection information D1 for detecting a blockage in some of the gates 245.

[0132] More specifically, the abnormality detection system 10 includes a molding apparatus 20 and an abnormality detection apparatus 40. The molding apparatus 20 includes a mold part 24 having a plurality of gates 245 and a cavity C1 connected to the plurality of gates 245 formed therein, an injection part 22 for filling the molten molding material L1 into the cavity C1 via the plurality of gates 245, a first sensor (e.g., sensor SN1) provided in a region closest to the first gate (e.g., gate GT1) among the plurality of gates 245 for detecting the pressure or temperature of the molding material L1 in the cavity C1, and a second sensor (e.g., at least one of sensors SN2 to SN4) provided in a region closest to a second gate (e.g., at least one of gates GT2 to GT4) different from the first gate among the plurality of gates 245 for detecting the pressure or temperature of the molding material L1 in the cavity C1.

[0133] The abnormality detection apparatus 40 includes a data acquisition unit 41 that acquires a first evaluation value (e.g., evaluation value R1) calculated based on first time-series data (e.g., graph line F11 or F21) of the pressure or temperature detected by the first sensor and a second evaluation value (e.g., at least one of evaluation values R2 to R4) calculated based on second time-series data (e.g., at least one of graph lines F12 to F14 or F22 to F24) of the pressure or temperature detected by the second sensor, and an abnormality detection unit 42 that detects clogging of the first gate or the second gate based on the first evaluation value and the second evaluation value.

[0134] When the first gate or the second gate is clogged, a difference occurs between the first time-series data and the second time-series data. To facilitate comparison of this difference, the first time-series data is quantified as the first evaluation value, the second time-series data is quantified as the second evaluation value, and clogging of the first gate or the second gate is detected based on these first evaluation value and second evaluation value. Thereby, clogging of some of the gates 245 that could not be detected by conventional detection can be detected.

[0135] In this embodiment, the first distance (for example, the first distance d1) through which the molding material L1 passes from the first gate to the first sensor is equal to the second distance (for example, at least one of the second to fourth distances d2 to d4) through which the molding material L1 passes from the second gate to the second sensor. By configuring in this way, when the gate 245 is not clogged, the time points (i.e., the rising time point and the peak time point) at which the molding material L1 reaches the first sensor and the second sensor can be aligned. As a result, it is possible to cause a delay in the rising time point and a delay in the peak time point only when some of the gates 245 are clogged. Consequently, the accuracy of clogging detection can be improved.

[0136] In this embodiment, the first evaluation value includes the first rising time point (for example, the time point Xs1) of the first time-series data, or the first peak time point (for example, the time point Xt1) at which the pressure or temperature of the first time-series data reaches the maximum value. The second evaluation value includes the second rising time point (for example, at least one of the time points Xs2 to Xs4) of the second time-series data, or the second peak time point (for example, at least one of the time points Xt2 to Xt4) at which the pressure or temperature of the second time-series data reaches the maximum value. The abnormality detection unit 42 detects clogging of the first gate when the first rising time point is later than the second rising time point, or when the first peak time point is later than the second peak time point.

[0137] Further, in this embodiment, the first sensor and the second sensor detect the pressure of the molding material L1. The first evaluation value includes the first peak value (for example, the maximum value Pt1) which is the maximum value of the pressure in the first time-series data. The second evaluation value includes the second peak value (for example, at least one of the maximum values Pt2 to Pt4) which is the maximum value of the pressure in the second time-series data. The abnormality detection unit 42 detects clogging of the first gate when the first peak value is lower than the second peak value.

[0138] In addition, in the present embodiment, the first sensor and the second sensor detect the pressure of the molding material L1, the first evaluation value includes a first integration value (for example, integration value IV1) that is the time integration from the first rising point of the pressure of the first time series data to the first peak point at which the pressure of the first time series data reaches the maximum value, the second evaluation value includes a second integration value (for example, at least one of integration values IV2 to IV4) that is the time integration from the second rising point of the pressure of the second time series data to the second peak point at which the pressure of the second time series data reaches the maximum value, and when the first integration value is smaller than the second integration value, the abnormality detection unit 42 detects clogging of the first gate.

[0139] As described above, in the present embodiment, the first evaluation value includes at least one of the first rising point, the first peak point, the first peak value, and the first integration value, and the second evaluation value includes at least one of the second rising point, the second peak point, the second peak value, and the second integration value corresponding to the first evaluation value. Then, by comparing the first rising point and the second rising point, the first peak point and the second peak point, the first peak value and the second peak value, or the first integration value and the second integration value, it is possible to identify that the first gate among the first gate and the second gate is clogged. That is, according to the abnormality detection system 10, not only can clogging of some gates 245 that could not be detected by conventional detection be detected, but also which gate 245 is clogged can be identified.

[0140] Further, the abnormality detection unit 42 inputs the first evaluation value and the second evaluation value to a learned model Tm1 in which the correlation between the first evaluation value and the second evaluation value and clogging of a plurality of gates 245 is machine-learned, thereby detecting clogging of at least one of the first gate and the second gate.

[0141] With such a configuration, after the learned model Tm1 is once generated, it becomes possible to detect clogging of some gates 245 from the first evaluation value and the second evaluation value. By using the learned model Tm1, even in a state where there are variations in the molding conditions, clogging of some gates 245 can be detected more accurately.

[0142] <Second Embodiment> The abnormal detection system 10 according to the first embodiment has been described above. However, the implementation of the present invention is not limited to this, and various modifications can be made. Hereinafter, the abnormal detection system 11 according to the second embodiment of the present invention will be described. In the following description, parts that are not changed from the first embodiment are denoted by the same reference numerals, and the description thereof will be omitted.

[0143] FIG. 11 is a block diagram schematically showing the abnormal detection system 11 according to the second embodiment. The abnormal detection system 11 includes a plurality of molding devices 20, an abnormal detection device 40a, an input unit 50, and a display unit 60.

[0144] In the present embodiment, the abnormal detection device 40a of the abnormal detection system 11 detects that the gate GT1 is clogged when the absolute value of the difference between the average value AvR of the plurality of evaluation values R1 to R4 and the evaluation value R1 (that is, the absolute value DIV1 of the deviation of the evaluation value R1) exceeds a predetermined reference value REF1. That is, the abnormal detection system 11 is different from the abnormal detection system 10 according to the first embodiment in that it detects an abnormality by comparing the absolute value of the deviation DIV1 with the reference value REF1 without using the learned model Tm1.

[0145] FIG. 12 is a block diagram showing the functional configuration of the abnormal detection device 40a according to the present embodiment. The abnormal detection device 40a includes a data acquisition unit 41, an abnormal detection unit 42a, an output unit 43a, a molding information storage unit 44, and a reference value storage unit 46. Each of these units is realized by a computer device having a computing unit such as a CPU and a storage unit such as an HDD. The data acquisition unit 41 acquires a plurality of evaluation values R1 to R4 in the same manner as the data acquisition unit 41 according to the first embodiment.

[0146] The reference value REF1 is stored in the reference value storage unit 46. The reference value REF1 is, for example, a value obtained by adding a predetermined margin to the standard deviation of the plurality of evaluation values R1 to R4 acquired when none of the plurality of gates 245 is clogged.

[0147] The abnormality detection unit 42a calculates the absolute value of the deviation DIV1 of the evaluation value R1 based on a plurality of evaluation values R1 to R4 acquired by the data acquisition unit 41. Then, by comparing the absolute value of the deviation DIV1 with the reference value REF1 stored in the reference value storage unit 46, detection information D2 for detecting clogging of the gate GT1 is acquired. The detection information D2 is, for example, the difference (DIV1 - REF1) between the absolute value of the deviation DIV1 and the reference value REF1.

[0148] The output unit 43a determines whether or not the gate GT1 is clogged based on the detection information D2 acquired by the abnormality detection unit 42a. For example, when the detection information D2 is the above difference (DIV1 - REF1), the output unit 43a determines that there is clogging in the gate GT1 when the detection information D2 is a positive value (that is, when the absolute value of the deviation DIV1 exceeds the reference value REF1). The output unit 43a outputs the determination result to the display unit 60 and the control unit 271.

[0149] Also, when determining whether or not the gate GT2 is clogged, the abnormality detection unit 42a calculates the absolute value of the deviation DIV2 of the evaluation value R2 based on a plurality of evaluation values R1 to R4 acquired by the data acquisition unit 41, and compares the absolute value of the deviation DIV2 with the reference value REF1 to acquire detection information D3 for detecting clogging of the gate GT2. In this way, by comparing the absolute values of the deviations of the evaluation values R1 to R4 corresponding to the gates GT1 to GT4 for which clogging is to be detected with the reference value REF1, detection information for detecting clogging of each of the gates GT1 to GT4 can be acquired.

[0150] Note that in this embodiment, it may be configured such that the detection information obtained by the abnormality detection unit 42a is directly displayed on the display unit 60 without providing the output unit 43a. In this case, based on the detection information displayed on the display unit 60, an operator may determine the clogging of the gate 245.

[0151] According to the abnormality detection system 11 according to this embodiment, clogging of some gates 245 can be easily detected by comparing a plurality of evaluation values R1 to R4 with a reference value REF1.

[0152] <Others> The embodiments disclosed as above are illustrative in all respects and not restrictive. That is, the abnormality detection system of the present invention is not limited to the illustrated form and may be in other forms within the scope of the present invention.

Explanation of Signs

[0153] 10 Abnormality detection system 11 Abnormality detection system 20 Molding device 21 Bed 22 Injection part 221 Hopper 222 Cylinder 223 Screw 224 Nozzle 225 Ball screw 226 Motor 227 Pressure sensor 228 Displacement sensor 229 Heater 23 Mold clamping part 231 Fixed platen 231a Through hole 232 Movable platen 232a Through hole 233 Tie bar 234 Ball screw 235 Support platen 236 Mold clamping force sensor 237 Motor 24 Mold part 241 Mold 242 Mold 243 Spool 243a End 243b End 244 Runner 245 Gate 25 Sensor 26 Temperature sensor 27 Control panel 271 Control unit 272 Communication part 30 Learning device 31 Training data acquisition part 32 Learning calculation part 33 Molding information storage part 34 Learned model storage part 40 Abnormality detection device 40a Abnormality detection device 41 Data acquisition part 42 Abnormality detection part 42a Abnormality detection part 43 Output part 43a Output part 44 Molding information storage part 45 Learned model storage part 46 Reference value storage part 50 Input part 60 Display part C1 cavity, L1 molding material, Fm1 foreign matter GT1 - GT4 gates, SN1 - SN4 sensors, d1 first distance d2 second distance, d3 third distance, d4 fourth distance R1 - R4 evaluation values, Tm1 learned model, D1 detection information REF1 reference value, DIV1 absolute value of deviation

Claims

1. An abnormality detection system comprising a molding apparatus for molding a molded product and an abnormality detection apparatus for detecting an abnormality in the molding apparatus, wherein the molding apparatus includes a mold part formed therein with a plurality of gates and a cavity connected to the plurality of gates, an injection part for filling the cavity with a molten molding material via the plurality of gates, a first sensor provided in a region closest to a first gate among the plurality of gates for detecting the pressure or temperature of the molding material in the cavity, a second sensor provided in a region closest to a second gate different from the first gate among the plurality of gates for detecting the pressure or temperature of the molding material in the cavity, and has the abnormality detection apparatus includes a data acquisition part for acquiring a first evaluation value calculated based on first time series data of the pressure or temperature detected by the first sensor and a second evaluation value calculated based on second time series data of the pressure or temperature detected by the second sensor, an abnormality detection part for detecting clogging of the first gate or the second gate based on the first evaluation value and the second evaluation value, and has the abnormality detection part inputs the first evaluation value and the second evaluation value to a learned model obtained by machine learning the correlation between the first evaluation value and the second evaluation value and clogging of the plurality of gates, thereby detecting clogging of at least one of the first gate and the second gate. An abnormality detection system.

2. The abnormality detection system according to claim 1, wherein a first distance through which the molding material passes from the first gate to the first sensor is equal to a second distance through which the molding material passes from the second gate to the second sensor.

3. The first evaluation value is a first rising point of the first time series data, or includes a first peak point at which the pressure or temperature of the first time series data reaches a maximum value, the second evaluation value is a second rising point of the second time series data, or includes a second peak point at which the pressure or temperature of the second time series data reaches a maximum value, the abnormality detection part detects clogging of the first gate when the first rising point is delayed compared to the second rising point, or when the first peak point is delayed compared to the second peak point. The abnormality detection system according to claim 2.

4. The first sensor and the second sensor detect the pressure of the molding material, The first evaluation value includes a first peak value which is the maximum value of the pressure of the first time-series data, The second evaluation value includes a second peak value which is the maximum value of the pressure of the second time-series data, The abnormality detection unit detects clogging of the first gate when the first peak value is lower than the second peak value. The abnormality detection system according to any one of claims 1 to 3.

5. The first sensor and the second sensor detect the pressure of the molding material, The first evaluation value includes a first integral value which is the time integral from the first rising point of the pressure of the first time-series data to the first peak point at which the pressure of the first time-series data becomes the maximum value, The second evaluation value includes a second integral value which is the time integral from the second rising point of the pressure of the second time-series data to the second peak point at which the pressure of the second time-series data becomes the maximum value, The abnormality detection unit detects clogging of the first gate when the first integral value is smaller than the second integral value. The abnormality detection system according to any one of claims 1 to 4.

6. An abnormality detection device for detecting an abnormality of a molding device for molding a molded product, The molding device includes A mold part having a plurality of gates and a cavity connected to the plurality of gates formed therein, An injection part for filling the cavity with a molten molding material through the plurality of gates, A first sensor provided in a region closest to the first gate among the plurality of gates, for detecting the pressure or temperature of the molding material in the cavity, A second sensor provided in a region closest to a second gate different from the first gate among the plurality of gates, for detecting the pressure or temperature of the molding material in the cavity, And has The abnormality detection device includes A data acquisition unit for acquiring a first evaluation value calculated based on first time-series data of the pressure or temperature detected by the first sensor and a second evaluation value calculated based on second time-series data of the pressure or temperature detected by the second sensor, An abnormality detection unit for detecting clogging of at least one of the first gate and the second gate based on the first evaluation value and the second evaluation value, And has The abnormality detection unit inputs the first evaluation value and the second evaluation value into a learned model obtained by machine learning the correlation between the first evaluation value, the second evaluation value, and the clogging of the plurality of gates, and thereby detects clogging of at least one of the first gate and the second gate. An abnormality detection device.

7. An abnormality detection method for detecting an abnormality in a molding apparatus including a mold part in which a plurality of gates and a cavity connected to the plurality of gates are formed inside, and an injection part for filling the cavity with a molten molding material via the plurality of gates, A first evaluation value calculated based on first time-series data of the pressure or temperature of the molding material in the cavity detected by a first sensor provided in a region closest to a first gate among the plurality of gates, and a second evaluation value calculated based on second time-series data of the pressure or temperature of the molding material in the cavity detected by a second sensor provided in a region closest to a second gate different from the first gate among the plurality of gates, and a data acquisition step of obtaining the second evaluation value; An abnormality detection step of inputting the first evaluation value and the second evaluation value into a learned model obtained by machine learning the correlation between the first evaluation value, the second evaluation value, and the clogging of the plurality of gates, and thereby detecting clogging of at least one of the first gate and the second gate. An abnormality detection method comprising:

8. A program for detecting an abnormality in a molding apparatus including a mold part in which a plurality of gates and a cavity connected to the plurality of gates are formed inside, and an injection part for filling the cavity with a molten molding material via the plurality of gates, A first evaluation value calculated based on first time-series data of the pressure or temperature of the molding material in the cavity detected by a first sensor provided in a region closest to a first gate among the plurality of gates, and a second evaluation value calculated based on second time-series data of the pressure or temperature of the molding material in the cavity detected by a second sensor provided in a region closest to a second gate different from the first gate among the plurality of gates, and a data acquisition step of obtaining the second evaluation value; An abnormality detection step of detecting clogging of at least one of the first gate and the second gate by inputting the first evaluation value and the second evaluation value into a learned model obtained by machine learning the correlation between the first evaluation value, the second evaluation value, and the clogging of the plurality of gates; A program for causing a computer device to execute the above. **Claim 9** A learned model for detecting an abnormality in a molding apparatus including a mold part in which a plurality of gates and a cavity connected to the plurality of gates are formed inside, and an injection part for filling the cavity with a molten molding material via the plurality of gates, A first evaluation value calculated based on first time-series data of the pressure or temperature of the molding material in the cavity detected by a first sensor provided in a region closest to a first gate among the plurality of gates; A second evaluation value calculated based on second time-series data of the pressure or temperature of the molding material in the cavity detected by a second sensor provided in a region closest to a second gate different from the first gate among the plurality of gates; Explanatory variables including the above are input, A learned model for causing a computer to function so as to output an objective variable including a value related to a clogged gate among the plurality of gates.

Images