CONDITION DETERMINATION DEVICE AND CONDITION DETERMINATION METHOD

Abstract

State determination device (1) for determining an operating state of an injection molding machine (2), wherein the state determination device (1) comprises: a data acquisition unit (30) configured to obtain a single time series data item from the injection molding machine (2), wherein the single time series data item is obtained within a single-cycle molding operation; an extraction condition storage unit (52) configured to store an extraction condition for extracting data used for processing related to machine learning from the data obtained by the data acquisition unit (30); a learning data extraction unit (32) configured to extract the data used for processing related to machine learning from the data obtained by the data acquisition unit (30) according to the extraction condition stored by the extraction condition unit (52); and a machine learning device (100) configured to perform the processing related to machine learning using the data extracted by the learning data extraction unit (32);wherein the extraction condition is at least one of the following: exclusion of acquired data during an alarm from the learning data, exclusion of acquired data for a predetermined number of cycles since the start of molding from the learning data, exclusion of acquired data for a predetermined number of cycles since the mold change from the learning data, exclusion of acquired data after completion of production from the learning data, exclusion of acquired data for a predetermined number of cycles since the change of injection conditions from the learning data, exclusion of acquired data for a predetermined number of cycles since the change of dispensing conditions from the learning data, extraction of only the acquired data in the mold closing process as learning data, and extraction of only the acquired data in the injection and compression process as learning data.;

Description

BACKGROUND OF THE INVENTIONField of the invention

[0001] The present invention relates to a condition determining device and a condition determining method, and more particularly relates to a condition determining device and a condition determining method for assisting maintenance work for injection molding machines. Description of related technology

[0002] Maintenance of an industrial machine, such as an injection molding machine, is performed regularly or when abnormalities occur. During maintenance of an industrial machine, maintenance personnel determine the abnormality in the operating condition of the industrial machine using physical quantities that indicate the machine's operating condition and were recorded during the machine's operation, and perform maintenance work, such as replacing abnormal components.

[0003] For example, for maintenance work on a check valve of an injection cylinder of an injection molding machine, a type of industrial machine, there is a well-known method in which a screw is periodically removed from the injection cylinder so that the dimensions of the check valve can be directly measured. However, this method requires production to be stopped for the measurement work, which inevitably limits productivity.

[0004] To solve this problem, there is a well-known abnormality diagnostic method. In this method, an abnormality is diagnosed by indirectly detecting the wear level of the injection cylinder's check valve, without stopping production for removing the screw from the injection cylinder, etc. Furthermore, in this diagnostic method, the abnormality is diagnosed by detecting the torque on the screw or the occurrence of resin backflow relative to the screw.

[0005] For example, JP H01-168421 A discloses a method in which a torque influence on a screw driving process is measured and an abnormality is determined when the measured value exceeds a tolerance range. Furthermore, JP 2017-030221 A and JP 2017-202632 A disclose methods in which an abnormality is diagnosed through supervised learning of a drive part load, resin pressure, and the like. Furthermore, JP 2018-097616 A and JP 2017-188030 A disclose a method in which machine learning is performed using time series data.

[0006] However, in an injection molding machine whose drive part has components of different specifications, there is a problem that the deviation between the measured values ​​obtained by the machine and the numerical values ​​of the learning data input during machine learning is so large that diagnosis cannot be performed correctly by machine learning. Another problem arises because if the type of resin used as the raw material of molded articles produced by the injection molding machine, or the type of injection mold, mold temperature controller, resin dryer, and the like as auxiliary equipment of the injection molding machine differ from those used in machine learning, diagnosis cannot be performed correctly by machine learning.

[0007] To improve the diagnostic accuracy of machine learning by solving these problems, it is necessary to prepare a wide variety of learning conditions for machine learning when building the machine learning learning model. However, machine learning based on the selection of a wide variety of injection molding machines, resins, and auxiliary equipment requires high costs. In addition, machine operation also requires the preparation of raw materials, such as resins and workpieces, and the cost of the raw materials required to obtain the learning data is also high. In addition, the work of obtaining the learning data is time-consuming. This poses the problem that the learning data cannot be collected efficiently.

[0008] In some cases, time series data obtained from the injection molding machine may be unsuitable for machine learning immediately after replacing a resin as a raw material for molded articles produced by the injection molding machine, replacing molds of an auxiliary device of the injection molding machine, or starting operation of a peripheral device such as a mold temperature controller or resin dryer or the injection molding machine, or when operating conditions such as injection conditions or compaction conditions related to the operation of the injection molding machine have changed, or when the injection molding machine is in an alarm state where it does not operate normally.However, according to the prior art, a problem exists in that the operating state of the machine cannot be correctly diagnosed because even unsuitable learning data is used to perform machine learning to introduce a learning model, or the unsuitable learning data is diagnosed. Approaches to this end are known from the prior art, which can be found, for example, in documents WO 2020 / 136836 A1, JP 2017 102826 A, WO 2018 / 229881 A1, and US Pat. No. 5,121,467 A. SUMMARY OF THE INVENTION

[0009] Thus, there is a need for a condition determination device and a condition determination method capable of easily eliminating unsuitable learning data to perform accurate machine learning and support the maintenance of various industrial machines using the learning results. This object is achieved by the invention specified in the independent claims. Advantageous further developments can be found in the subclaims.

[0010] Accordingly, a state determination device and method according to the present invention solve the above problems by introducing a highly accurate learning model by performing machine learning with time series data including changes in the operating or operable state of the injection molding machine or an unstable injection molding state, such as time series data during alarming, time series data immediately after the start of machine operation or replacement of a mold, or changes in mold condition target values ​​including injection conditions and compaction conditions related to the machine operation, excluding learning data, in conjunction with the learning data to be input to the machine learning.

[0011] A state determination device according to one aspect of the present invention is designed to determine an operating state of an injection molding machine and comprises a data acquisition unit designed to obtain a single time-series data item from the injection molding machine, the single time-series data item being obtained within a single-cycle molding process, an extraction condition storage unit designed to store an extraction condition for extracting data used for processing related to machine learning from the data obtained by the data acquisition unit, a learning data extraction unit designed to extract the data used for processing related to machine learning from the data obtained by the data acquisition unit according to the extraction conditions stored by the extraction condition unit,and a machine learning device configured to perform the processing related to machine learning using the data extracted by the learning data extraction unit; wherein the extraction condition is at least one of the following: excluding acquired data during an alarm from the learning data, excluding acquired data for a predetermined number of cycles since the start of molding from the learning data, excluding acquired data for a predetermined number of cycles since the mold change from the learning data, excluding acquired data after completion of production from the learning data, excluding acquired data for a predetermined number of cycles since the change of injection conditions from the learning data, excluding acquired data for a predetermined number of cycles since the change of metering conditions from the learning data,Extraction of only the data acquired in the mold closing process as learning data and extraction of only the data acquired in the injection and compression process as learning data.

[0012] The machine learning device may include a learning unit configured to perform machine learning using the data extracted by the learning data extraction unit, thereby creating a learning model. Furthermore, the learning unit may perform at least one of machine learning modes, including supervised learning, unsupervised learning, and reinforcement learning.

[0013] The machine learning device may include a learning model storage unit configured to store the learning model created by machine learning using the data extracted by the learning data extraction unit, and an estimation unit configured to perform estimation of the state of the industrial machine using the learning model based on the data extracted by the learning data extraction unit.

[0014] The estimation unit can estimate a degree of abnormality regarding the operating state of the injection molding machine, and the state determination device can display a warning message on a display device when the degree of abnormality estimated by the estimation unit exceeds a predetermined threshold. The estimation unit can estimate a degree of abnormality regarding the operating state of the injection molding machine, and the state determination device can display a warning symbol on a display device when the degree of abnormality estimated by the estimation unit exceeds a predetermined threshold.

[0015] The estimation unit may estimate a degree of abnormality regarding the operating state of the injection molding machine, and the state determining device may issue a command to stop the operation, decelerate, and / or limit the torque of a motor to the injection molding machine.

[0016] The data obtained by the learning data extraction unit may include at least one of the pieces of information comprising information for identifying an in-operation state, a stop state, a temperature increase state, a temperature increase completion state, a mold replacement state, a mold replacement completion state, an alarm state, or a manufacturing completion state indicating a machine state of the injection molding machine; information for identifying the occurrence of a change in an injection condition, a compression condition, a measurement condition, a mold opening / closing condition, an ejection condition, or a temperature condition representing an operation state of the injection molding machine; and information for identifying a mold closing operation, a mold clamping operation, an injection operation, a compression operation, a measurement operation, a mold opening operation, an ejection operation, or a waiting operation as a molding operation of the injection molding machine.

[0017] The data obtained by the data acquisition unit may include at least one of the data obtained from a plurality of injection molding machines connected via a wired / wireless network.

[0018] A machine learning method in a state determination device for obtaining data on an industrial machine according to another aspect of the present invention includes a data acquisition step for obtaining a single time-series data item from the injection molding machine, the single time-series data item being obtained within a one-cycle molding process; a learning data extraction step for extracting data used for machine learning processing from the data obtained from the injection molding machine from the data obtained in the data acquisition step according to an extraction condition for extracting the data used for machine learning processing; and a step of executing machine learning processing using the data extracted in the learning data extraction step.wherein the extraction condition is at least one of the following: excluding acquired data during an alarm from the data for learning, excluding acquired data for a predetermined number of cycles since the start of molding from the learning data, excluding acquired data for a predetermined number of cycles since the mold change from the learning data, excluding acquired data after completion of production from the learning data, excluding acquired data for a predetermined number of cycles since the change of injection conditions from the learning data, excluding acquired data for a predetermined number of cycles since the change of dispensing conditions from the learning data, extracting only the acquired data in the mold closing process as learning data, and extracting only the acquired data in the injection and compression process as learning data.;

[0019] The present invention having the configuration described above can perform machine learning with data obtained in the case of changes in the operating or operable states of an industrial machine and excluding data obtained in an unstable injection molding state, so that an improvement in the determination accuracy of the machine learning can be expected. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic hardware configuration diagram of a state determination device according to an embodiment; Fig. 2 is a schematic functional block diagram of a state determination device according to a first embodiment; Fig. 3 is a diagram showing examples of extraction conditions; Fig. 4 is a diagram showing an example of extraction of data for learning by a learning data extraction unit; Fig. 5 is a diagram showing another example of extraction of data for learning by the learning data extraction unit; Fig. 6 is a diagram showing another example of extraction of data for learning by the learning data extraction unit; Fig. 7 is a schematic functional block diagram of a state determination device according to a second embodiment; and Fig. 8 is a diagram showing a display example of an abnormal state. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] Fig. 1 is a schematic hardware configuration diagram showing main parts of a state determination device including a machine learning device according to an embodiment.

[0021] A state determination device 1 of the present invention can be installed, for example, on a controller for controlling industrial machines. Alternatively, it can be implemented as a personal computer associated with the controller for controlling the industrial machines, a management device 3 connected to the controller via a wired / wireless network, or a computer such as an edge computer, fog computer, or cloud server. In the following description, the state determination device 1 of the present invention will be exemplified as the computer connected via the network to the controller for controlling injection molding machines as the industrial machines.Although an injection molding machine is described as an industrial machine in each of the embodiments described below, the industrial machines as possible objects of state determination by the state determination device 1 of the present invention include an injection molding machine, a machine tool, a robot, a mining machine, and the like.

[0022] A CPU 11 included in the state determination device 1 according to the present embodiment is a processor for overall control of the state determination device 1. The CPU 11 reads system programs stored in a ROM 12 via a bus 20 and controls the entire state determination device 1 according to these system programs. A RAM 13 is temporarily loaded with temporary calculation data, various data input by an operator via an input device 71, and the like.

[0023] A non-volatile memory 14 consists of, for example, a memory backed up by a battery (not shown) or an SSD (semiconductor drive), and its storage state can be maintained even when the state determination device 1 is turned off. The non-volatile memory 14 stores a setting area filled with setting information of the operation of the state determination device 1, data input from the input device 71, and statistical data (machine type, mass and material of a mold, resin type, etc.) obtained via a network 7 of injection molding machines 2, time-series data on physical quantities (the temperature of a nozzle, the position, speed, acceleration, current, voltage, and torque of a motor for driving the nozzle, the temperature of the mold, the flow rate, flow velocity, and pressure of the resin, etc.).) detected during molding operations of the injection molding machines 2, time-series data of information (information for identifying a mold closing operation, mold clamping operation, injection operation, compaction operation, measurement operation, mold opening operation, ejection operation, cycle start and cycle end as molding operations of the injection molding machine 2, information indicating the state of alarm occurrence, etc.), data read from other computers via external storage devices (not shown) or the network 7, or the like. The programs and various data stored in the non-volatile memory 14 can be developed in the RAM 13 during execution and use.In addition, the system programs including a conventional analysis program for analyzing the various data, a program for controlling exchanges with a machine learning device 100 (described later), and the like are written in advance in the ROM 12.

[0024] The condition determination device 1 is connected to the wired / wireless network 7 via an interface 16. The network 7 is connected to the injection molding machines 2, the management device 3 for managing the production work by the injection molding machine 2, and the like, and exchanges data with the condition determination device 1.

[0025] Each injection molding machine 2 is a machine designed to produce molded articles from a resin, such as plastic. The injection molding machine 2 melts the resin as a material and fills (injects) it into the mold to perform molding. The injection molding machine 2 has various equipment parts, including the nozzle, the motor, a transmission mechanism, a reduction gear, and the moving part. The states of various parts are detected by sensors or the like, and the operation of the various parts is controlled by the controller. For example, an electric motor, an oil-hydraulic cylinder, an oil-hydraulic motor, or an air motor can be used as the motor for the injection molding machine 2. In addition, a ball screw, gears, pulleys, a belt, and the like can be used for the transmission mechanism for the injection molding machine 2.

[0026] Data read into the memory, data obtained as a result of executing the programs, and the like, data output from the machine learning device 100 (described later), and the like are output via an interface 17 and displayed on a display device 70. Furthermore, the input device 71, which consists of a keyboard, pointing device, or the like, supplies commands, data, and the like based on the operator's operation to the CPU 11 via an interface 18.

[0027] An interface 21 connects the state determination device 1 and the machine learning device 100. The machine learning device 100 includes a processor 101, a ROM 102, a RAM 103, and a non-volatile memory 104. The processor 101 controls the entire machine learning device 100. The ROM 102 stores system programs and the like. The RAM 103 is used for temporary storage at each step of machine learning processing. The non-volatile memory 104 is used to store learning models and the like.The machine learning device 100 can observe various pieces of information (for example, various data such as the type of the injection molding machine 2, the mass and material of the mold, and the type of resin, and time series data of various physical quantities such as the temperature of the nozzle, the position, speed, acceleration, current, voltage, and torque of the motor for driving the nozzle, the temperature of the mold, and the flow rate, flow velocity, and pressure of the resin) that can be obtained by the state determination device 1 via the interface 21. Furthermore, the state determination device 1 receives the processing result output from the machine learning device 100, stores the obtained result, displays it, and transmits it to other devices via the network 7 or the like.

[0028] Fig. 2 is a schematic functional block diagram of the state determination device 1 and the machine learning device 100 according to the first embodiment.

[0029] The state determination device 1 of the present embodiment has a configuration required when the machine learning device 100 performs learning (learning mode). Each functional block included in Fig. 2 is implemented as the CPU 11 of the state determination device 1 and the processor 101 of the machine learning device 100 shown in Fig. 1, executes the respective system programs and controls the operation of the individual parts of the state determination device 1 and the machine learning device 100.

[0030] The state determination device 1 of the present embodiment includes a data acquisition unit 30, a learning data extraction unit 32, a preprocessing unit 34, and the machine learning device 100. The machine learning device 100 includes a learning unit 110. Furthermore, an obtained data storage unit 50 and an extraction condition storage unit 52 are provided on a nonvolatile memory 14 of the state determination device 1. The obtained data storage unit 50 stores data obtained from external machines or the like. The extraction condition storage unit 52 stores conditions for extracting data for learning from the obtained data. A learning model storage unit 130 is provided on the nonvolatile memory 104 of the machine learning device 100. The learning model storage unit 130 stores learning models created by machine learning by the learning unit 110.

[0031] The data acquisition unit 30 receives various data input from the injection molding machine 2, the input device 71, and the like. The data acquisition unit 30 receives, for example, various pieces of information including statistical data such as the type of the injection molding machine 2, the mass and material of the mold, and the type of resin; time series data on various physical quantities such as the temperature of the nozzle, the position, speed, acceleration, current, voltage, and torque of the motor for driving the nozzle; the temperature of the mold related to the injection molding operation of the injection molding machine 2; and the flow rate, flow velocity, and pressure of the resin; information indicating machine states of the injection molding machine 2, such as a running state, a stop state, a temperature increase state, and a mold replacement state;a completion of mold replacement, an alarm state, a manufacturing completion state, and the like; information for identifying the occurrence of a change in injection conditions, compression conditions, measurement conditions, mold opening / closing conditions, and ejection conditions representing operating states of the injection molding machine 2; information for identifying a mold closing operation, mold clamping operation, injection operation, compression operation, measurement operation, mold opening operation, ejection operation, waiting operation, cycle start, and cycle end as molding operations of the injection molding machine 2; information indicating the state of the occurrence of an alarm; information on maintenance work for the injection molding machine input by an operator, and the like; and stores these data in the acquired data storage unit 50. Upon obtaining the time-series data, the data acquisition unit 30 considers the time-series data,obtained within a predetermined time range (e.g., a single-cycle molding operation range) as a single time-series data and then stores it in the obtained data storage unit 50 based on changes in signal data obtained from the injection molding machine 2 and other time-series data. The data acquisition unit 30 may be configured to obtain the data from the management device 3 or from other computers via the external storage devices (not shown) or the wired / wireless network 7.

[0032] In the machine learning phase of the learning unit 110, the learning data extraction unit 32 extracts obtained data to be used for machine learning from the obtained data obtained by the data acquisition unit 30 (and stored in the obtained data storage unit 50) based on the extraction conditions stored in the extraction condition storage unit 52. In other words, the learning data extraction unit 32 excludes unsuitable obtained data for machine learning from the obtained data obtained by the data acquisition unit 30 based on the extraction conditions stored in the extraction condition storage unit 52.

[0033] Fig. 3 is a diagram illustrating the extraction conditions stored in the extraction condition storage unit 52.

[0034] The extraction condition storage unit 52 stores at least one of the extraction conditions, which are organized and managed, for example, based on condition categories or the like. The extraction conditions stored by the extraction condition storage unit 52 may be conditions for specifying obtained data to be used for machine learning or conditions for specifying those obtained data that are not used for machine learning (or are excluded from it). The extraction conditions stored by the extraction condition storage unit 52 include, at least, conditions for categorizing the obtained data based on predetermined data values ​​included in the obtained data and specifying whether or not the obtained data categorized by those conditions as data for learning should be used.

[0035] Fig. 4 is a diagram illustrating an example of extraction of the obtained data by the learning data extraction unit 32 based on the extraction conditions regarding the machine states stored in the extraction condition storage unit 52.

[0036] Suppose in a case where the learning data extraction unit 32 extracts the waveform data of a current value for each cycle as data for learning, when the obtained data as shown in Fig. 4, are stored in the acquired data storage unit 50. In this case, if an extraction condition “exclude acquired learning data during alarming of data for learning” is set in the extraction condition storage unit 52, the learning data extraction unit 32 operates such that no current value data obtained in the cycle of the molding process is extracted as data for learning when an alarm is issued during the cycle. Specifically, in the case of the example shown in Fig. 4, does not extract data on the current values ​​obtained in the (i + 2)th and (1 + 3)th cycles in which the occurrence of an alarm is detected as data for learning, but extracts data on current values ​​obtained in and before the (i + 1)th cycle and in and after the (i + 4)th cycle as data for learning.

[0037] Fig. 5 is a diagram illustrating an example of extraction of the obtained data by the learning data extraction unit 32 based on the extraction conditions related to the operation states stored in the extraction condition storage unit 52.

[0038] Suppose in a case where the learning data extraction unit 32 extracts the waveform data of a voltage value for each cycle as data for learning, when the obtained data as shown in Fig. 5, are stored in the acquired data storage unit 50. In this case, if an extraction condition “exclude acquired data for 10 cycles since the change of the injection conditions from the data for learning” is set in the extraction condition storage unit 52, the learning data extraction unit 32 operates such that no voltage value data obtained during 10 cycles following a cycle in which the injection conditions are changed is extracted as data for learning when the injection condition change is performed during molding process cycles (or when an injection condition change signal is turned ON). Specifically, in the case of the example shown in Fig. 5, does not extract data on the voltage values ​​obtained during 10 cycles (up to the (i + 10)th cycle) since the (i + 1)th cycle in which the injection conditions are changed as data for learning, but extracts data on voltage values ​​obtained in and before an i-th cycle and in and after the (i + 11)th cycle as data for learning.

[0039] Fig. 6 is a diagram illustrating an example of extraction of the obtained data by the learning data extraction unit 32 based on the extraction conditions regarding the molding operations stored in the extraction condition storage unit 52.

[0040] Suppose in a case where the learning data extraction unit 32 extracts the waveform data of a current value for each cycle as data for learning, when the obtained data as shown in Fig. 6, are stored in the obtained data storage unit 50. In this case, if an extraction condition “extract only data obtained during injection and compression processes as data for learning” is set in the extraction condition storage unit 52, then the learning data extraction unit 32 operates such that current value data obtained during the injection and compression processes from the individual molding processes are extracted as data for learning. Specifically, in the case of the example, the learning data extraction unit 32 specifies Fig. 6, time periods for the injection and compression processes from each molding process based on start and end signals in each process, and extracts current value data obtained during these time periods as data for learning.

[0041] A plurality of extraction conditions may be set in the extraction condition storage unit 52. In this case, conflicts may arise between specifications of use and non-use of two or more extraction conditions as learning data. Then, the learning data extraction unit 32 may prioritize the non-use specification as the learning data. Alternatively, the priority order among the extraction conditions is stored in advance along with the extraction conditions in the extraction condition storage unit 52 so that the learning data extraction unit 32 can resolve the conflicts by specifying use or non-use as learning data based on the stored priority order.

[0042] In the machine learning phase of the machine learning device 100, the preprocessing unit 34 creates learning data to be used for learning by the machine learning device 100 based on the data for learning extracted by the learning data extraction unit 32. The preprocessing unit 34 creates learning data obtained by converting (or quantifying or sampling) data input from the learning data extraction unit 32 into a uniform format to be used in the machine learning device 100. For example, when the machine learning device 100 performs unsupervised learning, the preprocessing unit 34 creates state data S having a predetermined format during learning as the learning data.If the machine learning device 100 performs supervised learning, the preprocessing unit 34 creates a set of state data S and label data L having a predetermined format during learning as the learning data. If the machine learning device 100 performs reinforcement learning, the preprocessing unit 34 creates a set of state data S and determination data D having a predetermined format during learning as the learning data.

[0043] The learning unit 110 of the machine learning device 100 performs machine learning using the learning data created by the preprocessing unit 34 based on the data for learning extracted by the learning data extraction unit 32. The learning unit 110 creates a learning model by performing machine learning using the data obtained from the injection molding machine 2 based on a conventional machine learning method such as unsupervised learning, supervised learning, or reinforcement learning, and stores the created learning model in the learning model storage unit 130.The unsupervised learning method performed by the learning unit 110 can be represented, for example, by the autoencoder method or the K-means method, while the supervised learning method can be represented, for example, by the multilayer perceptron method, the recurrent neural network method, the long-short-term memory method, or the convolutional neural network method. The reinforcement learning method can be represented, for example, by the Q-learning method.

[0044] The learning unit 110 may perform unsupervised learning based on, for example, learning data obtained by processing the data obtained from the injection molding machine 2 in a normal operation state by the learning data extraction unit 32 and the pre-processing unit 34, and create the distribution of data obtained in a normal state as a learning model.

[0045] Furthermore, the learning unit 110 may, for example, perform supervised learning using learning data obtained by processing the obtained data by the learning data extraction unit 32 and the pre-processing unit 34, in such a manner that a normal label is assigned to the obtained data obtained from the normally operated injection molding machine and an abnormal label is assigned to the obtained data obtained from the injection molding machine 2 before and after the occurrence of an abnormality, thereby establishing discrimination boundaries between the normal and abnormal data as learning models.

[0046] In the state determination device 1 according to the first embodiment having the configuration described above, the learning data extraction unit 32 extracts data for learning from the obtained data contained in the obtained data storage unit 50 from the injection molding machine 2 according to the extraction conditions stored in the extraction condition storage unit 52. The operator can set the extraction conditions in the extraction condition storage unit 52 so that appropriate data can be extracted as data for learning according to the purpose of machine learning at that time.In this way, in conjunction with the data for learning extracted by the learning data extraction unit 32, it is possible to exclude from learning data time series data including changes in the operating or operable state of the injection molding machine or an unstable injection molding state, such as time series data during alarming, time series data immediately after the start of machine operation or replacement of a mold, or changes in mold condition setpoints, including injection conditions and compaction conditions, related to the machine operation, so that only those time series data belonging to predetermined operations necessary for determining the operating state can be used.When determining the state on the injection molding machine 2 using the learning data created in this way, it can be expected that the accuracy of determining the operating state of the injection molding machine 2 is improved compared with the case where a learning model created by a conventional method is used.

[0047] Fig. 7 is a schematic functional block diagram of a state determination device 1 and a machine learning device 100 according to a second embodiment.

[0048] The state determination device 1 of the present embodiment has a configuration required when the machine learning device 300 performs estimation (estimation mode). Each of the functional blocks shown in Fig. 7 is implemented as the CPU 11 of the state determination device 1 and the processor 101 of the machine learning device 100 as shown in Fig. 1, executes the respective system programs and controls the operation of the individual parts of the state determination device 1 and the machine learning device 100.

[0049] The state determination device 1 of the present embodiment, like the first embodiment, includes a data acquisition unit 30, a learning data extraction unit 32, a preprocessing unit 34, and the machine learning device 100. The machine learning device 100 includes an estimation unit 120. Furthermore, an obtained data storage unit 50 and an extraction condition storage unit 52 are provided on a nonvolatile memory 14 of the state determination device 1. The obtained data storage unit 50 stores data obtained from external machines or the like. The extraction condition storage unit 52 stores conditions for extracting data for learning from the obtained data. A learning model storage unit 130 is provided on the nonvolatile memory 104 of this machine learning device 100.The learning model storage unit 130 stores learning models created by machine learning by the learning unit 110.

[0050] The control unit 30 according to the present embodiment has the same function as the data acquisition unit 30 in the first embodiment.

[0051] Although the basic operation of the learning data extraction unit 32 according to the present embodiment is the same as that of the learning data extraction unit 32 of the first embodiment, the second embodiment differs from the first embodiment in that the data extracted by the learning data extraction unit 32 is data for estimation used for the machine learning device 100 to estimate the state of the injection molding machine 2.

[0052] In the phase of estimating the state of the injection molding machine 2 by the machine learning device 100 using the learning data, the preprocessing unit 34 creates state data S having a predetermined format to be used for estimation by the machine learning device 100 by converting (quantifying or sampling) data for estimation estimated by the learning data extraction unit 32 into a unit format to be used in the machine learning device 100.

[0053] Based on the state data S created by the preprocessing unit 34, the estimation unit 120 estimates the state of the injection molding machine using the learning models stored in the learning model storage unit 130. In the case where the learning model stored in the learning model storage unit 130 is a learning model created by unsupervised learning (or for which parameters have been set), the estimation unit 120 of this embodiment inputs the state data S obtained by the preprocessing unit 34 into the learning model and then estimates the degree of deviation of the state data S from the state data obtained during normal operation, thereby calculating an abnormality degree as the result of the estimation.On the other hand, in the case where the learning model stored in the learning model storage unit 130 is a learning model created by supervised learning, the estimation unit 120 of this embodiment inputs the state data obtained by the preprocessing unit 34 into the learning model, thereby estimating and calculating a category (a degree of normality or abnormality) to which the operating state of the injection molding machine belongs. The result of the estimation by the estimation unit 120 (the degree of abnormality regarding the state of the injection molding machine, the category to which the operating state of the injection molding machine belongs, etc.) can be used by outputting it for display on the display device 70 or outputting it for transmission via a wired / wireless network (not shown) to a host computer, cloud computer, or the like.Furthermore, if the result of the estimation by the estimation unit 120 turns out to be a predetermined state (for example, if a predetermined threshold is exceeded by the degree of abnormality estimated by the estimation unit 120, or if the category to which the operating state of the injection molding machine estimated by the estimation unit 120 belongs turns out to be “abnormal”), a warning message and icon may be selected for display on the display device 70, as shown in FIG. Fig.8, or a command to stop the operation, decelerate, or limit the torque of the motor for driving the injection molding machine may be issued to the injection molding machine. In the state determination device 1 according to the second embodiment having the configuration described above, the learning data extraction unit 32 extracts data for estimation from obtained data included in the obtained data storage unit 50 of the injection molding machine 2 according to the extraction conditions stored in the extraction condition storage unit 52. The operator can set the extraction conditions in the extraction condition storage unit 52 so that appropriate data can be extracted as data for estimation according to the purpose of state determination on the injection molding machine 2 at that time.Thus, the data for estimation extracted by the learning data extraction unit 32 is data that does not include time series data that indicate changes in the operating or operable state of the injection molding machine or an unstable molding state, such as time series data during alarms, time series data immediately after the start of machine operation or mold replacement, or changes in mold condition target values, including injection conditions and compaction conditions, related to the machine operation. Therefore, only suitable time series data for determining the operating state of the injection molding machine 2 can be used for state determination. Accordingly, the accuracy of determining the operating state of the injection molding machine 2 can be expected to improve through machine learning.Although the state determining devices 1 according to the first and second embodiments described above are applicable to the case of determining states related to industrial machines such as robots and machine tools, they can be suitably used for industrial machines that behave unstably within expectations, for example, at the start of production or at the beginning of an operation to resume production. In particular, an injection molding machine often operates unstably within expectations at the start of its operation or immediately after the injection conditions have changed, even if production is performed under the same injection conditions. Since the operation approaches a stable normal behavior if it continues without change, this operation state is not considered abnormal and falls out of the scope of maintenance and inspection.Therefore, the condition determining device of the present invention is particularly useful for the injection molding machine having such characteristics.

[0054] Although embodiments of the present invention have been described above, the invention is not limited to the above-described embodiments and can be appropriately modified and implemented in various forms.

[0055] For example, although the state determination device 1 and the machine learning device 100 are described as devices having different CPUs (processors) in the above embodiments, the machine learning device 100 may alternatively be implemented by the CPU 11 of the state determination device 1 and the system programs stored in the ROM 12. Furthermore, if a plurality of injection molding machines 2 are connected to each other via the network, their respective operating states may be determined by a single state determination device 1, or the state determination device 1 may be installed on the controller of the injection molding machine.

Claims

[1] State determining device (1) for determining an operating state of an injection molding machine (2), wherein the state determining device (1) comprises: a data acquisition unit (30) configured to obtain a single time series data item from the injection molding machine (2), wherein the single time series data item is obtained within a single-cycle molding operation; an extraction condition storage unit (52) configured to store an extraction condition for extracting data used for processing related to machine learning from the data obtained by the data acquisition unit (30); a learning data extraction unit (32) configured to extract the data used for processing related to machine learning from the data obtained by the data acquisition unit (30) according to the extraction condition stored by the extraction condition unit (52); and a machine learning device (100) configured to perform the processing related to machine learning using the data extracted by the learning data extraction unit (32);wherein the extraction condition is at least one of the following: exclusion of acquired data during an alarm from the learning data, exclusion of acquired data for a predetermined number of cycles since the start of molding from the learning data, exclusion of acquired data for a predetermined number of cycles since the mold change from the learning data, exclusion of acquired data after completion of production from the learning data, exclusion of acquired data for a predetermined number of cycles since the change of injection conditions from the learning data, exclusion of acquired data for a predetermined number of cycles since the change of dispensing conditions from the learning data, extraction of only the acquired data in the mold closing process as learning data, and extraction of only the acquired data in the injection and compression process as learning data.; [2] The state determination device (1) according to claim 1, wherein the machine learning device (100) comprises a learning unit (110) configured to perform machine learning using the data extracted by the learning data extraction unit (32), thereby creating a learning model. [3] The state determining device (1) according to claim 2, wherein the learning unit performs at least one type of machine learning including supervised learning, unsupervised learning and reinforcement learning. [4] The state determination device (1) according to claim 1, wherein the machine learning device (100) comprises a learning model storage unit (130) configured to store the learning model created by the machine learning using the data extracted by the learning data extraction unit (32); and an estimation unit (120) configured to perform estimation of the state of the industrial machine using the learning model based on the data extracted by the learning data extraction unit (32). [5] The condition determining device (1) according to claim 4, wherein the estimation unit (120) estimates a degree of abnormality with respect to the operating state of the injection molding machine (2), and the condition determining device (1) displays a warning message on a display device (70) if the degree of abnormality estimated by the estimation unit (120) exceeds a predetermined threshold. [6] The condition determining device (1) according to claim 4, wherein the estimation unit (120) estimates a degree of abnormality with respect to the operating state of the injection molding machine (2), and the condition determining device displays a warning symbol on a display device (70) if the degree of abnormality estimated by the estimation unit (120) exceeds a predetermined threshold. [7] The state determining device (1) according to claim 4, wherein the estimation unit (120) estimates a degree of abnormality with respect to the operating state of the injection molding machine (2), and the state determining device (1) outputs a command to stop the operation, decelerate, and / or restrict the torque of a motor to the injection molding machine (2). [8] The state determination device according to claim 1, wherein, among the data obtained by the learning data extraction unit, at least one of the pieces of information comprising information for identifying a state in operation, a stop state, a temperature increase state, a completion of temperature increase, a mold replacement state, a completion of mold replacement, an alarm state, or a production completion state indicating a machine state of the injection molding machine, information for identifying the occurrence of a change in an injection condition, a compression condition, a measurement condition, a mold opening / closing condition, an ejection condition, or a temperature condition representing an operation state of the injection molding machine, and information for identifying a mold closing operation, a mold clamping operation, an injection operation, a compression operation, a measurement operation, a mold opening operation,Ejection process or waiting process as a molding process of the injection molding machine., [9] The condition determining device (1) according to claim 1, wherein the data obtained by the data acquisition unit includes at least one of the data obtained from a plurality of industrial machines connected via a wired / wireless network. [10] Method relating to machine learning in a state determination device (1) for determining an operating state of an injection molding machine (2), the state determination method comprising: a data acquisition step for obtaining a single time series data from the injection molding machine (2), wherein the single time series data is obtained within a one-cycle molding process; a learning data extraction step for extracting data used for processing related to machine learning from the data obtained from the injection molding machine (2) from the data obtained in the data acquisition step, according to an extraction condition for extracting the data used for processing related to machine learning; and a step of executing processing related to machine learning using the data extracted in the learning data extraction step;wherein the extraction condition is at least one of the following: excluding acquired data during an alarm from the data for learning, excluding acquired data for a predetermined number of cycles since the start of molding from the learning data, excluding acquired data for a predetermined number of cycles since the mold change from the learning data, excluding acquired data after completion of production from the learning data, excluding acquired data for a predetermined number of cycles since the change of injection conditions from the learning data, excluding acquired data for a predetermined number of cycles since the change of dispensing conditions from the learning data, extracting only the acquired data in the mold closing process as learning data, and extracting only the acquired data in the injection and compression process as learning data.;

Images