WO2024089851A1

WO2024089851A1 - Determination device and determination method

Info

Publication number: WO2024089851A1
Application number: PCT/JP2022/040234
Authority: WO
Inventors: 淳史堀内
Original assignee: ファナック株式会社
Priority date: 2022-10-27
Filing date: 2022-10-27
Publication date: 2024-05-02

Abstract

A determination device according to the present disclosure is provided with: a data acquisition unit for acquiring the number of molding cycles in an injection molding machine, step identification data for identifying a plurality of molding steps, and time-series data pertaining to a predetermined physical quantity as data indicating a state pertaining to the injection molding machine; a feature amount calculation unit that calculates a feature amount indicating a feature of the time-series data; a difference calculation unit that calculates the difference between the current feature amount and the feature amount before a predetermined number of molding cycles on the basis of the feature amount and the number of molding cycles; a determination reference value calculation unit that calculates a determination reference value on the basis of at least one of the most recent difference values; a determination unit that determines the molding state in the injection molding machine on the basis of the difference value in the current molding cycle and the determination reference value, and a notification unit for notifying the determination result from the determination unit.

Description

Determination device and determination method

This disclosure relates to a determination device and a determination method.

Data observed during molding (e.g., time-series data, feature quantities) are monitored to determine whether the molding state of a molded product or an injection molding machine is good or bad. For example, it is known to observe "time-series data (e.g., pressure, torque, etc.)" that indicates the molding state of an injection molding machine for each molding cycle, and compare it with the time-series data observed during a molding cycle when the product is good, to determine whether the molding state of a molded product or an injection molding machine is good or bad.

Furthermore, the molding cycle that produces one molded product is made up of multiple molding processes (e.g., injection process, weighing process, etc.). Therefore, it is also known to calculate "feature values (e.g., maximum pressure value in the injection process)" based on the time-series data included in each molding process, and compare the calculated feature values with a predetermined allowable range to determine whether the molding state of the molded product or the injection molding machine is good or bad (e.g., Patent Documents 1 to 4, etc.).

JP 2010-076177 A International Publication No. 2022/075181 International Publication No. 2022/075242 JP 2017-087588 A

Even if the feature value falls within the monitoring range (lower limit to upper limit) that determines whether the molding state is good or bad, the feature value may suddenly change and cause the molding state to become abnormal. More specifically, abnormalities may be triggered by sudden factors such as damage to the sensor or moving parts, the intrusion of foreign matter into moving parts or production materials, or an operator error (missetting of operation command values). Previously, even if the feature value suddenly changed, it was judged to be normal until the "feature value or the difference value of the feature value" exceeded the monitoring range, which caused the problem of defective products being produced during this period.

The determination device according to the present disclosure acquires the number of molding cycles, which is the number of times an injection molding machine repeats a molding cycle to manufacture a molded product, process identification data identifying the multiple molding processes that make up the molding cycle, and time series data related to predetermined physical quantities that indicate the state of the injection molding machine, and calculates a feature quantity that indicates the characteristics of the time series data included in the predetermined molding process. Next, based on the feature quantity and the number of molding cycles, it calculates a difference value (amount of change) between the current feature quantity and the feature quantity a predetermined number of molding cycles ago. Then, using the most recent difference value or the average of the most recent difference values as a determination reference value, it compares the difference value of the current molding cycle with the determination reference value to determine whether the molding state is normal or abnormal, and if an abnormality is determined, it issues a signal indicating an abnormality, thereby resolving the above-mentioned problem.

An aspect of the present disclosure is a judgment device that includes a data acquisition unit that acquires a molding cycle number, which is the number of times a molding cycle for manufacturing a molded product in an injection molding machine is repeated, process identification data that identifies a plurality of molding processes that constitute the molding cycle, and time series data related to a predetermined physical quantity as data indicating a state of the injection molding machine; a feature calculation unit that calculates a feature amount that indicates a feature of the time series data included in a predetermined molding process among the plurality of molding processes identified by the process identification data; a difference calculation unit that calculates a difference value between a current feature amount and a feature amount a predetermined number of molding cycles ago based on the feature amount and the molding cycle number; a judgment reference value calculation unit that calculates a judgment reference value based on at least one or more of the most recent difference values; a judgment unit that judges the molding state of the injection molding machine based on the difference value in the current molding cycle calculated by the difference calculation unit and the judgment reference value; and a notification unit that notifies the judgment result by the judgment unit.

1 is a schematic hardware configuration diagram of a determination device according to a first embodiment of the present disclosure. FIG. 1 is a schematic configuration diagram of an injection molding machine. 1 is a block diagram showing schematic functions of a determination device according to a first embodiment; FIG. 2 is a diagram illustrating molding steps constituting a molding cycle of an injection molding machine. 1 is a graph illustrating a transition of a feature amount of time-series data relating to a predetermined physical amount; 11 is a graph illustrating changes in a difference value, an upper limit judgment value, and a lower limit judgment value; 11 is a graph illustrating a transition of a difference value. 1 is a graph illustrating a transition of a feature amount; FIG. 11 is a block diagram showing schematic functions of a determination device according to a third embodiment.

Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.
[First embodiment]
1 is a schematic hardware configuration diagram showing a main part of a determination device according to a first embodiment of the present disclosure. The determination device 1 according to this embodiment can be implemented as a control device that controls an industrial machine based on a control program, for example. The determination device 1 according to this embodiment can be implemented on a computer such as a personal computer attached to a control device that controls an industrial machine, or a personal computer, cell computer, fog computer 6, or cloud server 7 that is connected to the control device via a wired/wireless network 5. In this embodiment, an example is shown in which the determination device 1 is implemented on a personal computer that is connected to a control device 3 that controls an injection molding machine 4 as an industrial machine via a network 5.

The CPU 11 provided in the determination device 1 according to this embodiment is a processor that controls the entire determination device 1. The CPU 11 reads out a system program stored in the ROM 12 via the bus 22, and controls the entire determination device 1 according to the system program. The RAM 13 temporarily stores temporary calculation data, display data, and various data input from outside.

The non-volatile memory 14 is composed of, for example, a memory backed up by a battery (not shown) or an SSD (Solid State Drive), and the memory state is maintained even when the power supply of the determination device 1 is turned off. The non-volatile memory 14 stores programs and data read from the external device 72 via the interface 15, programs and data input from the input device 71 via the interface 18, and programs and data acquired from the control device 3 or other devices via the network 5. The stored data may include, for example, data related to physical quantities such as the motor current, voltage, torque, position, speed, acceleration, injection cylinder temperature, resin pressure, flow rate, flow rate, mold temperature and pressure, mold temperature and pressure, mold temperature regulator temperature and pressure, molded product removal machine position and speed, and vibration and sound generated in each part of the injection molding machine 4 detected by the sensor 8 attached to the injection molding machine 4 controlled by the control device 3. The programs and data stored in the non-volatile memory 14 may be expanded to the RAM 13 when executed/used. In addition, various system programs, such as well-known analysis programs, are pre-programmed into ROM 12.

The interface 15 is an interface for connecting the CPU 11 of the determination device 1 to an external device 72 such as a USB device. For example, system programs, programs related to the operation of the injection molding machine 4, setting data, etc. are read from the external device 72. In addition, programs and setting data created and edited within the determination device 1 can be stored in an external storage means via the external device 72.

The interface 20 is an interface for connecting the CPU 11 of the determination device 1 to a wired or wireless network 5. The network 5 may communicate using technologies such as serial communication such as RS-485, Ethernet (registered trademark), optical communication, wireless LAN, Wi-Fi (registered trademark), Bluetooth (registered trademark), etc. The network 5 is connected to a control device 3 that controls at least one injection molding machine 4, a fog computer 6, a cloud server 7, etc., and exchanges data with the determination device 1.

The display device 70 displays the various data loaded into the memory, data obtained as a result of executing programs, etc., output via the interface 17. The input device 71, which is comprised of a keyboard, pointing device, etc., passes instructions and data based on operations by the operator to the CPU 11 via the interface 18.

FIG. 2 is a schematic diagram of the injection molding machine 4. The injection molding machine 4 is mainly composed of a mold clamping unit 401 and an injection unit 402. The mold clamping unit 401 is equipped with a movable platen 416 and a fixed platen 414. A movable mold 412 is attached to the movable platen 416, and a fixed mold 411 is attached to the fixed platen 414. A servo motor 50 is attached to the mold clamping unit 401. By driving the servo motor 50, a ball screw (not shown) is driven via power transmission means such as a belt 420 and a pulley 422, and the movable platen 416 can be advanced or retreated toward the fixed platen 414.

On the other hand, the injection unit 402 is composed of an injection cylinder 426, a hopper 436 that stores the resin material to be supplied to the injection cylinder 426, and a nozzle 440 provided at the tip of the injection cylinder 426. The injection unit 402 can move the injection cylinder 426 forward or backward in the direction of the fixed platen 414 by driving a servo motor (not shown).

In a molding cycle for producing one molded product, the mold clamping unit 401 closes and clamps the mold by moving the movable platen 416, and the injection unit 402 presses the nozzle 440 against the fixed mold 411 and then injects a measured amount of resin into the injection cylinder 426 into the mold. These operations are controlled by commands from the control device 3 (not shown).

In addition, sensors 8 (not shown) are attached to each part of the injection molding machine 4, and various physical quantities necessary for controlling the molding operation are detected. Examples of the detected physical quantities include the motor current, voltage, torque, position, speed, and acceleration of the drive unit, the temperature of the injection cylinder 426, the pressure of the resin in the injection cylinder 426, the flow rate of the resin, the temperature and pressure of the mold, the temperature and pressure of the mold temperature regulator, the position and speed of the molded product removal machine, and vibrations and sounds generated in each part of the injection molding machine 4. The detected physical quantities are sent and output to the determination device 1. In the determination device 1, each detected physical quantity is stored in the RAM 13, non-volatile memory 14, etc.

FIG. 3 is a schematic block diagram showing the functions of the determination device 1 according to the first embodiment of the present disclosure. Each function of the determination device 1 according to this embodiment is realized by the CPU 11 of the determination device 1 shown in FIG. 1 executing a system program and controlling the operation of each part of the determination device 1.

The judgment device 1 of this embodiment includes a data acquisition unit 110, a leveling unit 120, a feature calculation unit 130, a difference calculation unit 140, a judgment reference value calculation unit 150, a judgment unit 160, and a notification unit 170. The RAM 13 to the non-volatile memory 14 of the judgment device 1 are provided with an acquired data storage unit 210, which is an area for storing data acquired by the data acquisition unit 110 from the sensor 8 attached to the injection molding machine 4, a feature storage unit 220, which is an area for storing data indicating the feature calculated by the feature calculation unit 130, and a difference value storage unit 230, which is an area for storing the difference value calculated by the difference calculation unit 140.

When the injection molding machine 4 performs a molding cycle operation, the data acquisition unit 110 acquires from the control device 3 the number of molding cycles, which is the number of times the molding cycle for manufacturing a molded product is repeated in the injection molding machine, a plurality of process identification data for identifying the plurality of molding processes constituting the molding cycle, and time series data related to a predetermined physical quantity as data indicating the state of the operation of the molding cycle. The process identification data for identifying the molding process is data for identifying which molding process is currently being performed. As illustrated in FIG. 4, the molding processes constituting the molding cycle include a mold closing process, an injection process, a pressure holding process, a weighing process, a decompression process, a cooling process, a mold opening process, an ejection process, and a removal process. The data indicating the state of the operation of the molding cycle may include, for example, motor current, voltage, torque, position, speed, and acceleration acquired from the servo motor that operates the injection molding machine 4, and detection value data such as the temperature of the injection cylinder 426 detected by the sensor 8, the pressure of the resin in the injection cylinder 426, the vibration of the drive unit, the amount of opening and closing of the mold, the pressure in the mold cavity, and the temperature in the mold cavity, as time series data. The data acquisition unit 110 may acquire data input by the operator from the input device 71 or data input via the external device 72. The data acquisition unit 110 associates the acquired molding cycle number, process identification data, and data indicating the state related to the operation of the molding cycle, and stores them in the acquired data storage unit 210.

The smoothing unit 120 calculates smoothed data by smoothing the time series data stored in the acquired data storage unit 210. The smoothing process may be carried out using a known smoothing method such as the moving average method, weighted moving average, moving median, or Savitzky-Golay filter.

The feature amount calculation unit 130 calculates the feature amount of the time series data related to the predetermined physical quantity acquired in a predetermined molding process among the multiple molding processes based on the number of molding cycles and process identification data stored in the acquired data storage unit 210, and the time series data related to the predetermined physical quantity. The feature amount calculation unit 130 may calculate the feature amount of the smoothed data related to the predetermined physical quantity based on the number of molding cycles and process identification data stored in the acquired data storage unit 210, and the smoothed data obtained by smoothing the time series data related to the predetermined physical quantity by the smoothing unit 120. The feature amount of the time series data related to the predetermined physical quantity may be a statistical amount such as a minimum value, a maximum value, a minimum value, or a maximum value of the time series data related to the predetermined physical quantity acquired in the predetermined molding process. The feature amount of the time series data related to the predetermined physical quantity may also be the value of the time series data at a predetermined timing (for example, a time when a predetermined time has elapsed, a time when the axis of the injection molding machine 4 reaches a predetermined position, etc.) after the start of the predetermined molding process. Furthermore, the feature amount of the time series data related to the predetermined physical quantity may be the time required from the start of the predetermined molding process until the time series data related to the predetermined physical quantity reaches any one of the minimum value, maximum value, minimum value, and maximum value, the movement amount of the screw, the rotation amount of the screw, the movement amount of the movable platen, the movement amount of the ejector, the opening and closing amount of the mold, the rotation amount of the mold, etc. The feature amount calculation unit 130 stores the calculated feature amount of the time series data related to the predetermined physical quantity in each molding process in the feature amount storage unit 220 in association with the number of molding cycles and the process identification data.

FIG. 5 is a graph illustrating the transition of the feature quantity of time-series data related to a specified physical quantity. In FIG. 5, the black circles represent plots of the feature quantity V(n) (n is the number of molding cycles) related to the specified time-series data calculated by the feature quantity calculation unit 130 for each molding cycle.

Based on the feature values stored in the feature value storage unit 220 and the number of molding cycles, the difference calculation unit 140 calculates, for each feature value, a difference value between the feature value of the time series data acquired before the molding cycle in which the time series data used to calculate the feature value was acquired. In other words, the difference calculation unit 140 calculates the amount of change in the feature value over the molding cycle. For example, the difference calculation unit 140 may calculate a difference value between a specific feature value and a feature value calculated a specific number of molding cycles before (one cycle before, five cycles before, etc.) the molding cycle in which the feature value was calculated. The difference value may also be calculated using feature values for a specific number of molding cycles immediately preceding the molding cycle in which the feature value was calculated. The difference calculation unit 140 stores the calculated difference value in the difference value storage unit 230 in association with the number of molding cycles and process identification data.

The judgment reference value calculation unit 150 calculates the judgment reference value based on the most recent difference value stored in the difference value storage unit 230, or the difference values for the most recent multiple predetermined cycles. The judgment reference value calculation unit 150 may use the most recent difference value as the judgment reference value. Also, the judgment reference value may be the average value obtained by statistically processing the difference values for the most recent multiple predetermined cycles.

The judgment unit 160 judges the validity of the difference value of the current molding cycle calculated by the difference calculation unit 140 based on the judgment reference value calculated by the judgment reference value calculation unit 150. If it is judged to be inappropriate, it judges that the molding state of the injection molding machine 4 is abnormal. The judgment unit 160 may calculate, for example, an upper limit judgment value by adding a predetermined upper limit tolerance width (offset) α to the judgment reference value and a lower limit judgment value by subtracting a predetermined lower limit tolerance width (offset) β from the judgment reference value, and judge that the difference value is inappropriate when the difference value of the current molding cycle deviates from the range between the lower limit judgment value and the upper limit judgment value. In addition, the judgment unit 160 may calculate a deviation degree indicating the degree to which the difference value of the current molding cycle calculated by the difference calculation unit 140 deviates from the judgment reference value calculated by the judgment reference value calculation unit 150, and judge the molding state of the injection molding machine 4 based on whether the deviation degree is greater than at least one predetermined threshold value. If the judgment unit 160 judges that the molding state in the injection molding machine 4 is abnormal, it outputs a message to that effect to the notification unit 170.

6 is a graph showing an example of the transition of the difference value, upper limit judgment value, and lower limit judgment value calculated by the difference calculation unit 140. In FIG. 6, the black circles are plots of the difference value D(n) (n is the number of molding cycles) of the feature amount related to the predetermined time series data calculated by the difference calculation unit 140 for each molding cycle. The black triangles are plots of the upper limit judgment value ALM _U (n) determined based on the judgment reference value calculated by the judgment reference value calculation unit 150 for each molding cycle, and the black squares are plots of the lower limit judgment value ALM _L (n) determined based on the judgment reference value. In the example of FIG. 6, the upper limit judgment value ALM _U (n) at the molding cycle number n is the value obtained by adding a predetermined upper limit tolerance width α to the most recent difference value D(n-1), and the lower limit judgment value ALM _L (n) is the value obtained by subtracting a predetermined lower limit tolerance width β from the most recent difference value D(n-1). 6, the judgment unit 160 judges whether the molding state of the injection molding machine 4 is normal/abnormal depending on whether the difference value D(n) deviates from the range between a monitor line connecting the upper limit judgment value ALM _U (n) calculated in each molding cycle and a monitor line connecting the lower limit judgment value ALM _L (n). By using the difference value D(n) to judge the molding state, it becomes possible to judge the molding state based on the changing tendency (degree of increase or decrease) of the feature amount.

The notification unit 170 notifies the judgment result indicating an abnormality when the judgment unit 160 judges that the molding state of the injection molding machine 4 is abnormal. The notification unit 170 may display, for example, the judgment result of the molding state of the injection molding machine 4 input from the judgment unit 160 on the display device 70. At that time, the molding process to which the feature value judged to be abnormal belongs may be displayed. A similar notification may be transmitted to the control device 3 that controls the injection molding machine 4, the fog computer 6, the cloud server 7, etc. In addition, it may display predetermined information indicating the judgment result by the judgment unit 160, a graph plotting a set of the number of molding cycles and the difference value, etc. Furthermore, the notification unit 170 may output a command to the control device 3 to limit the operation of the injection molding machine 4 when the judgment unit 160 judges that the molding state of the injection molding machine 4 is abnormal. Examples of commands output to the control device 3 include a command to stop the operation of the injection molding machine 4 and a command to limit the drive torque of the servo motor 50 equipped in the injection molding machine 4 to a low torque.

The judgment device 1 according to this embodiment, which has the above configuration, judges the molding state of the injection molding machine 4 based on the difference value of the feature amount of the time series data acquired from the injection molding machine 4, and can pick up signs of changes in the feature amount at an early stage and detect abnormalities in the molding state at an early stage. This makes it possible to notify the operator of an abnormality at an early stage and to safely stop the operation of the machine.

In the injection molding machine 4, an unexpected factor may trigger the time series data observed by the position speed sensor and pressure sensor of the moving part to fluctuate significantly from the normal value. Examples of unexpected factors include improper assembly or damage of the sensor 8 or driving parts, contamination of the moving parts or production materials with foreign matter, and operator operation errors. More specifically, as an example of poor contact of the sensor 8, the temperature detected by the temperature sensor equipped in the nozzle 440 or the injection cylinder 426 rises toward the set temperature when the control of the heater equipped in the injection cylinder 426 is turned on, overshoots the set temperature, and then gradually approaches the set temperature while repeating a drop and rise. Here, if the assembly between the heater equipped in the injection cylinder 426 and the injection cylinder 426 becomes loose and poor contact occurs between them, the current temperature of the injection cylinder 426 drops suddenly. Here, as a feature quantity related to the temperature of the injection cylinder 426, the maximum temperature value of the injection cylinder 426 in the injection process falls within the allowable range (lower limit to upper limit) as the molding cycle is repeated, but this feature quantity may suddenly drop as the molding cycle is repeated. In addition, if the current temperature does not stabilize at the set temperature and fluctuates, physical quantities related to the injection molding machine other than temperature, such as pressure and torque, also fluctuate. Therefore, for example, values such as the maximum pressure value in the injection process and the maximum rotational torque value in the metering process also change suddenly.

In conventional technology, even if the feature value changes suddenly, as long as the feature value falls within the monitoring range (lower limit to upper limit), the sign of an abnormality is overlooked and the value is erroneously determined to be normal. However, the judgment device 1 according to this embodiment makes it possible to detect abnormalities such as poor contact of the sensor 8 and damage to the injection molding machine, its components, and the mold at an early stage. In addition, since abnormalities that can lead to molding defects (e.g. burrs, shorts, etc.) in molded products are detected at an early stage, the outflow of defective products is reduced.

As a modified example of the determination device 1 according to this embodiment, the determination device 1 may be configured to thin out the determination process. When the change in the characteristic amount is small, for example, the amount of backflow during injection, which varies according to the temperature of the mold or the progress of wear of the backflow prevention ring provided on the screw, varies slowly over a long period of time. In such a case, the molding cycle may be thinned out and a determination may be made every predetermined molding cycle period, for example, by calculating and determining the difference value D(n) every three shots (cycles). This reduces the calculation load on the CPU and the number of determinations.

As another variation of the determination device 1 according to this embodiment, the determination device 1 may acquire data related to multiple injection molding machines 4 via the network 5 and determine the molding state of each injection molding machine 4. By configuring in this manner, it becomes possible to manage the molding state of multiple injection molding machines 4 with one determination device 1. For example, when the determination device 1 is implemented on a higher-level computer such as a fog computer 6 installed in a factory, the configuration for managing the molding state of multiple injection molding machines 4 makes it possible to optimally utilize the functions of the determination device 1 according to the present disclosure.

[Second embodiment]
A determination device according to a second embodiment of the present disclosure will be described below.
The determination device 1 according to the present embodiment has a hardware configuration similar to that of the determination device 1 according to the first embodiment. Similarly to the determination device 1 according to the first embodiment, the determination device 1 according to the present embodiment also has a data acquisition unit 110, a leveling unit 120, a feature amount calculation unit 130, a difference calculation unit 140, a determination reference value calculation unit 150, a determination unit 160, and a notification unit 170.

The data acquisition unit 110, the leveling unit 120, the feature calculation unit 130, the difference calculation unit 140, the judgment unit 160, and the notification unit 170 provided in the judgment device 1 according to this embodiment have the same functions as those in the first embodiment.

The judgment reference value calculation unit 150 of this embodiment differs from the first embodiment in that it obtains a regression equation by a predetermined regression analysis based on the multiple difference values stored in the difference value storage unit 230 and the number of molding cycles, and calculates the estimated difference value using the obtained regression equation as the judgment reference value.

FIG. 7 is a graph illustrating the transition of the difference value. The judgment reference value calculation unit 150 according to this embodiment calculates the difference value Dest(i+1) estimated based on a plurality of difference values calculated in molding cycles performed prior to the specified molding cycle (i+1) as the judgment reference value in the specified molding cycle (i+1). The judgment reference value calculation unit 150 performs a regression analysis based on a specified number of molding cycles prior to the molding cycle (i+1). The regression analysis performed by the judgment reference value calculation unit 150 may be a known regression analysis method such as linear regression, ridge regression, lasso regression, exponential regression, polynomial regression, or autoregressive model. In the example of FIG. 7, the regression analysis is performed using the difference values D(i-3) to D(i) calculated in the molding cycles (i-3) to (i) prior to the molding cycle (i+1). The solid line graph in FIG. 7 is a graph of a linear regression formula calculated by linear regression analysis based on the difference values D(i-3) to D(i). The judgment reference value calculation unit 150 calculates the difference value Dest(i+1) estimated in the molding cycle (i+1) using the linear regression equation calculated by the regression analysis, and the calculated difference value Dest(i+1) is set as the judgment reference value. Based on the calculated judgment reference value, the judgment unit 160 judges the molding state of the injection molding machine 4 using the difference value D(i+1) in the molding cycle (i+1), which is similar to the judgment device 1 according to the first embodiment. The graph shown by the dashed line in FIG. 7 is a graph showing, for example, an upper limit judgment value obtained by adding a predetermined upper limit tolerance width (offset) α to the judgment reference value, and a lower limit judgment value obtained by subtracting a predetermined lower limit tolerance width (offset) β from the judgment reference value. Then, when the difference value D(i+1) in the molding cycle (i+1) deviates from the range between the lower limit judgment value and the upper limit judgment value, it is judged that the molding state is abnormal.

As another example of the present embodiment, an example in which a difference value is estimated using an autoregressive model will be described.
FIG. 8 is a graph illustrating the transition of a predetermined feature quantity Y(n). The difference calculation unit 140 calculates the feature quantity difference value D(n) based on the feature quantity. In the example of FIG. 8, the difference between the feature quantity Y(i) in the molding cycle (i) and the feature quantity Y(i-1) in the immediately previous molding cycle (i-1) is set as the molding cycle (i) difference value D(i). When calculating the difference value Dest(i+1) estimated by regression analysis using an autoregressive model, for example, the following formula 1 is used. In formula 1, N is a value obtained by subtracting 1 from the total number of difference values used in the regression analysis.

The difference value Dest(i+1) estimated in this manner is used as the judgment reference value, and the judgment unit 160 judges the molding state of the injection molding machine 4 using the difference value D(i+1) in the molding cycle (i+1), which is the same as in the judgment device 1 according to the first embodiment.

The determination device 1 according to this embodiment, which has the above configuration, is expected to be able to use regression analysis to accurately calculate a determination reference value based on the trends in past difference values.

[Third embodiment]
A determination device according to a third embodiment of the present disclosure will be described below.
The determination device 1 according to the present embodiment has the same hardware configuration as the determination device 1 according to the first embodiment.
9 is a schematic block diagram showing functions of the determination device 1 according to the present embodiment. The determination device 1 according to the present embodiment includes a data acquisition unit 110, a leveling unit 120, a feature amount calculation unit 130, a difference calculation unit 140, a determination reference value calculation unit 150, a determination unit 160, and a notification unit 170, similar to the determination device 1 according to the first embodiment.

The judgment criterion value calculation unit 150 in this embodiment learns the change trend of the difference value using machine learning technology, and estimates the difference value based on the learning result. It differs from the first embodiment in that the estimated difference value is used as the judgment criterion value. The judgment criterion value calculation unit 150 includes a learning unit 152 and an estimation unit 154.

The learning unit 152 generates a learning model for estimating the difference value of the next molding cycle using multiple consecutive difference values among the difference values stored in the difference value storage unit 230 as learning data. For example, learning is performed with the difference values of the feature amounts calculated in the molding cycles (i-k) to (i-1) as explanatory variables and the difference value of the feature amount calculated in the molding cycle (i) as the objective variable. Note that k is a natural number greater than 1. This makes it possible to generate a learning model for estimating the next difference value based on the past (k) difference values. The learning model generated by the learning unit 152 may be a known supervised learning model such as a multilayer perceptron, an RNN (Recurrent Neural Network) model, or an LSTM (Long Short Term Memory) model. The learning model generated by the learning unit 152 is stored in the learning model storage unit 158 provided on, for example, the RAM 13 or the non-volatile memory 14. After a learning model that has sufficiently learned the transitions in the feature quantities is generated and stored in the learning model storage unit 158, the learning unit 152 is no longer necessarily required as part of the configuration of the determination device 1.

The estimation unit 154 uses the learning model generated by the learning unit 152 and stored in the learning model storage unit 158 to estimate the difference value of the next molding cycle based on the most recent consecutive multiple difference values stored in the difference value storage unit 230. The difference value estimated by the estimation unit 154 is used as a judgment reference value.

The determination device 1 according to this embodiment having the above configuration is expected to be able to accurately calculate a determination reference value based on the trends in past difference values by using machine learning technology. In addition, the learning model can be stored in a compressed state in memory and decompressed for use when making an estimation. This allows the determination device 1 to be realized with a small memory capacity, which has the advantage of reducing costs. Furthermore, it is preferable from the standpoint of security and information confidentiality to encrypt and store the learning model and then decrypt and use it in estimation mode. It is also possible to update an existing learning model by additional learning in accordance with the operating conditions of the injection molding machine 4.

The determination device 1 according to each embodiment of the present disclosure described so far determines the molding state of the injection molding machine 4 based on the difference values of the feature quantities of the time-series data acquired from the injection molding machine 4, thereby capturing early signs of changes in the feature quantities and detecting abnormalities in the molding state at an early stage. This makes it possible to notify the operator of an abnormality at an early stage and to safely stop the operation of the machine.

Although the embodiments of the present disclosure have been described in detail above, the present disclosure is not limited to the individual embodiments described above. Various additions, substitutions, modifications, partial deletions, etc. are possible in these embodiments, without departing from the gist of the invention, or without departing from the idea and intent of the present disclosure derived from the contents described in the claims and their equivalents. For example, in the above-mentioned embodiments, the order of each operation and the order of each process are shown as examples, and are not limited to these. The same applies when numerical values or formulas are used in the explanation of the above-mentioned embodiments.

Below, notes relating to the embodiments of the present disclosure are provided.
(Appendix 1)
A determination device (1) according to one aspect of the present disclosure includes a data acquisition unit (110) that acquires a molding cycle number, which is the number of times a molding cycle for manufacturing a molded product has been repeated in an injection molding machine (4), process identification data that identifies a plurality of molding processes that constitute the molding cycle, and time series data related to a predetermined physical quantity as data indicating a state of the injection molding machine, a feature calculation unit (130) that calculates a feature amount that indicates a feature of the time series data included in a predetermined molding process among the plurality of molding processes identified by the process identification data, and The system includes a difference calculation unit (140) that calculates a difference value between a current feature value and a feature value a predetermined number of molding cycles ago based on the feature value and the number of molding cycles, a judgment reference value calculation unit (150) that calculates a judgment reference value based on at least one of the most recent difference values, a judgment unit (160) that judges a molding state in the injection molding machine (4) based on the difference value in the current molding cycle calculated by the difference calculation unit (140) and the judgment reference value, and an alarm unit (170) that alarms the result of the judgment by the judgment unit (160).
(Appendix 2)
The judgment reference value calculation unit (150) provided in the judgment device (1) according to another aspect of the present disclosure sets the difference value calculated in the most recent molding cycle or the average value of the difference values calculated in the most recent multiple molding cycles as the judgment reference value.
(Appendix 3)
The judgment reference value calculation unit (150) provided in the judgment device (1) according to another aspect of the present disclosure performs regression analysis based on the difference values calculated in the most recent multiple molding cycles, and sets the difference value of the current molding cycle estimated using the regression equation calculated by the regression analysis as the judgment reference value.
(Appendix 4)
The judgment reference value calculation unit (150) provided in the judgment device (1) according to another aspect of the present disclosure estimates a difference value for the current molding cycle based on difference values calculated in the most recent multiple molding cycles using a learning model for estimating the next difference value of the difference value based on multiple difference values, and sets the estimated difference value as the judgment reference value.
(Appendix 5)
The determination device (1) according to another aspect of the present disclosure further includes a smoothing unit (120) that calculates smoothed data by smoothing the time-series data.
(Appendix 6)
The notification unit (170) provided in the judgment device (1) according to another aspect of the present disclosure outputs a command to limit the operation of the injection molding machine (4) when it judges that the molding state of the injection molding machine (4) is abnormal.
(Appendix 7)
The determination device (1) according to another aspect of the present disclosure is further implemented on a management device (6, 7) that manages the injection molding machine (4) and is connected to the injection molding machine (4) via a wired or wireless network.
(Appendix 8)
A judgment method according to one aspect of the present disclosure includes the steps of: acquiring a molding cycle number, which is the number of times a molding cycle for manufacturing a molded product in an injection molding machine (4) has been repeated, process identification data identifying a plurality of molding processes that constitute the molding cycle, and time series data related to a predetermined physical quantity as data indicating a state of the injection molding machine; calculating a feature amount indicating a feature of the time series data included in a predetermined molding process among the plurality of molding processes identified by the process identification data; calculating a difference value between a current feature amount and a feature amount a predetermined number of molding cycles ago based on the feature amount and the molding cycle number; calculating a judgment reference value based on at least one of the most recent difference values; judging a molding state of the injection molding machine (4) based on the difference value in the current molding cycle calculated in the difference value calculation step and the judgment reference value; and notifying a result of the judgment.

REFERENCE SIGNS LIST 1 Determination device 3 Control device 4 Injection molding machine 5 Network 6 Fog computer 7 Cloud server 8 Sensor 11 CPU
12 ROM
13 RAM
14

Non-volatile memory

15, 17, 18, 20 Interface 22 Bus 70 Display device 71 Input device 72 External device 110 Data acquisition unit 120 Leveling unit 130 Feature amount calculation unit 140 Difference calculation unit 150 Judgment reference value calculation unit 152 Learning unit 154 Estimation unit 158 Learning model storage unit 160 Judgment unit 170 Notification unit 210 Acquired data storage unit 220 Feature amount storage unit 230 Difference value storage unit

Claims

a data acquisition unit that acquires a molding cycle count, which is the number of times a molding cycle for manufacturing a molded product in an injection molding machine has been repeated, process identification data that identifies a plurality of molding processes that constitute the molding cycle, and time-series data relating to a predetermined physical quantity as data indicating a state of the injection molding machine;
a feature amount calculation unit that calculates a feature amount indicating a feature of the time-series data included in a predetermined molding process among the plurality of molding processes identified by the process identification data;
a difference calculation unit that calculates a difference between a current feature value and a feature value a predetermined number of molding cycles ago based on the feature value and the number of molding cycles;
a judgment reference value calculation unit that calculates a judgment reference value based on at least one of the most recent difference values;
a determination unit that determines a molding state of the injection molding machine based on the difference value in the current molding cycle calculated by the difference calculation unit and the determination reference value;
a notification unit that notifies a result of the determination by the determination unit;
A determination device comprising:
The judgment reference value calculation unit sets the difference value calculated in the most recent molding cycle or the average value of the difference values calculated in the most recent multiple molding cycles as the judgment reference value.
The determination device according to claim 1 .
the judgment reference value calculation unit performs a regression analysis based on the difference values calculated in the most recent multiple molding cycles, and sets the difference value of the current molding cycle estimated using the regression equation calculated by the regression analysis as the judgment reference value.
The determination device according to claim 1 .
the judgment reference value calculation unit estimates a difference value of a current molding cycle based on difference values calculated in a plurality of most recent molding cycles using a learning model for estimating a next difference value of the difference value based on the plurality of difference values, and sets the estimated difference value as a judgment reference value.
The determination device according to claim 1 .
A smoothing unit that smoothes the time series data and calculates smoothed data.
The determination device according to claim 1 .
the notification unit outputs a command to limit operation of the injection molding machine when it determines that the molding state of the injection molding machine is abnormal.
The determination device according to claim 1 .
The system is implemented on a management device that manages the injection molding machine and is connected to the injection molding machine via a wired or wireless network.
The determination device according to claim 1 .
acquiring a molding cycle count, which is the number of times a molding cycle for manufacturing a molded product in an injection molding machine has been repeated, process identification data for identifying a plurality of molding processes constituting the molding cycle, and time-series data relating to a predetermined physical quantity as data indicating a state of the injection molding machine;
calculating a feature quantity indicating a feature of the time-series data included in a predetermined molding process among the plurality of molding processes identified by the process identification data;
calculating a difference value between a current feature value and a feature value a predetermined number of molding cycles ago based on the feature value and the number of molding cycles;
calculating a judgment reference value based on at least one of the most recent difference values;
a step of determining a molding state of the injection molding machine based on the difference value in the current molding cycle calculated in the step of calculating the difference value and the determination reference value;
notifying the result of the determination;
A determination method implemented by a computer.