Information processing system, program, and method for manufacturing injection molding machine

Abstract

Provided is information for deriving, as a characteristic to be controlled, a mechanical error in the relationship between physical quantities of a device, implementing control in accordance with the characteristic of each device, and suppressing a difference in operation of the device caused by the mechanical error. Waveform data obtained by controlling the operation of a device to be controlled is acquired, and actual values of an input variable of a first variable, which is at least some of the variables, and a second variable, which is at least the other of the variables, are extracted on the basis of the relationship between the variables included in the acquired waveform data. Then, an output value that is a variable corresponding to the second variable is derived from an input variable that is the first variable, characteristics of the device to be controlled are obtained from the derived output value and an actual value that is the second variable, and information for controlling the device to be controlled is output from the characteristics of the device to be controlled. And sending to the control object device.

Description

Technical Field

[0001] This application claims priority based on Japanese Patent Application No. 2023-131974 filed on August 14, 2023. The entire contents of the Japanese Patent Application are incorporated herein by reference.

[0002] The present invention relates to an information processing system, a program and a method for manufacturing an injection molding machine. Background Art

[0003] In the control of equipment, control gains such as PI (Proportional-Integral) control gain are used. The response of the controlled object is determined by the control gain and the characteristics of the controlled object. Therefore, if the characteristics of the equipment of the controlled object and the response to be achieved are determined, the control gain can be calculated backwards.

[0004] Patent document 1 discloses a control system comprising: a state observation unit, which observes information related to a machine as state data; a judgment data acquisition unit, which acquires information related to processing as judgment data; a reward calculation unit, which calculates a reward based on the judgment data and a reward condition; a learning unit, which performs machine learning on adjustment of the servo gain of the machine; a decision unit, which determines an adjustment action for the servo gain of the machine based on the machine learning result of the adjustment of the servo gain of the machine and the state data; and a gain change unit, which changes the servo gain of the machine based on the determined servo gain adjustment action.

[0005] Patent Document 1: Japanese Patent No. 6457472

[0006] The physical quantities detected in the devices are related to each other, but there are physical quantities that cause deviations (mechanical errors) in the relationship for each device. If there is such a mechanical error, the response to the control will differ for each device even if control is performed with the same control gain. Summary of the invention

[0007] An object of the present invention is to provide information for obtaining the characteristics of a control object in the relationship between physical quantities of a device, realizing control corresponding to the characteristics of each device, and suppressing the difference in the operation of the device caused by mechanical errors.

[0008] One embodiment of the present invention is an information processing system, characterized in that it comprises: an acquisition unit, which acquires waveform data obtained by the action of a controlled object device; an analysis unit, which extracts a first variable as at least a part of the variable and a second variable as at least another part of the variable based on the relationship of the variables contained in the acquired waveform data, derives a variable equivalent to the second variable based on the first variable, and calculates the characteristics of the controlled object device based on the derived variables and the second variable; and an output unit, which outputs information used for controlling the controlled object device based on the characteristics of the controlled object device.

[0009] Effects of the Invention

[0010] According to one embodiment of the present invention, it is possible to provide information for obtaining the characteristics of a control object in the relationship between physical quantities of a device, implementing control corresponding to the characteristics of each device, and suppressing differences in the operation of the device caused by mechanical errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is a diagram showing the structure of an injection molding machine to which the present embodiment is applied.

[0012] Figure 2 It is a diagram showing the structure of a control device.

[0013] Figure 3 It is a diagram showing the structure of a data processing device.

[0014] Figure 4 It is a diagram showing the structure of an information processing device.

[0015] Figure 5 This is a diagram showing an example of the hardware configuration of a computer that realizes the information processing device.

[0016] Figure 6 is a diagram schematically showing the adjustment of mechanical errors. Figure 6 (A) is a diagram showing an example of mechanical error. Figure 6 (B) is a diagram showing a state where the mechanical error has been adjusted.

[0017] Figure 7 is a diagram showing the structure of a linear multiple regression model.

[0018] Figure 8 It is a diagram showing an example of analysis processing by the analysis unit.

[0019] Fig. 9 This is a flowchart showing the procedure for adjusting control information according to the characteristics of the equipment in an injection molding machine.

[0020] In the figure: 10-injection molding machine, 20-injection device, 30-mold clamping device, 100-control device, 200-data processing device, 400-information processing device, 410-communication unit, 420-processing unit, 421-analysis unit, 430-storage unit, 431-machine learning model. DETAILED DESCRIPTION

[0021] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[0022] <Device Structure>

[0023] Figure 1 1 is a diagram showing the structure of an injection molding machine to which the present embodiment is applied. The injection molding machine 10 includes an injection device 20, a mold clamping device 30, a control device 100, and a data processing device 200. The injection molding machine 10 is an example of a controlled device. In addition, an information processing device 400 is connected to the control device 100 and the data processing device 200.

[0024] The injection device 20 is configured to include a cylinder for heating the molding material, a screw that can rotate in the cylinder and can move forward and backward along the axial direction, a rotary motor that drives the screw in the rotational direction, and a motor that drives the screw in the axial direction. The molding material is, for example, a resin. The injection device 20 injects the molding material that is heated in the cylinder and becomes liquid by extending the screw in the direction (front) from the injection device 20 toward the clamping device 30 while rotating, and fills it into the mold of the clamping device 30 arranged in front of the injection device 20. The injection device 20 performs, for example, a metering process, a filling process, a pressure holding process, etc. in the manufacturing process of the molded product. The filling process and the pressure holding process are also collectively referred to as an injection process.

[0025] The mold clamping device 30 is configured to include a mold, a fastening mechanism for fastening the mold, a motor for driving the fastening mechanism, and the like. The mold clamping device 30 closes the mold and receives the molding material injected from the injection device 20 into the mold. At this time, the mold clamping device 30 fastens the mold (mold clamping) using the fastening mechanism so that the mold will not open due to the filling of the molding material. The molding material filled into the mold is solidified to generate a molded product. Then, the mold clamping device 30 can open the mold and take out the generated molded product. The mold clamping device 30 performs, for example, a mold closing process, a pressure increasing process, a mold clamping process, a decompression process, a mold opening process, and the like in the manufacturing process of the molded product.

[0026] The control device 100 is a device for controlling the operation of the injection device 20 and the mold clamping device 30. The data processing device 200 is a device for processing data obtained as the injection device 20 and the mold clamping device 30 operate.

[0027] The information processing device 400 acquires waveform data obtained by the operation of the injection molding machine 10 from the control device 100 and the data processing device 200, and obtains the characteristics of the injection molding machine 10 based on the acquired waveform data. The characteristics include the elements that cause mechanical errors in each injection molding machine 10. Based on the information on the obtained characteristics of the injection molding machine 10, the information processing device 400 outputs information for performing control such as suppressing the mechanical errors of each injection molding machine 10, and provides it to the control device 100. The details of the information processed and output by the information processing device 400 will be described later.

[0028] <Configuration of Control Device 100 >

[0029] Figure 2 1 is a diagram showing the structure of the control device 100. The control device 100 controls the actions of the injection device 20 and the mold clamping device 30. The control device 100 is implemented by a computer, for example. The control device 100 includes a control unit 110 and a storage unit 120. The control device 100 controls the injection device 20 and the mold clamping device 30 to repeat the processes related to the manufacture of the molded product, thereby repeatedly manufacturing the molded product. The processes related to the manufacture of the molded product include a metering process, a mold closing process, a pressure increasing process, a mold clamping process, a filling process, a pressure holding process, a cooling process, a decompression process, a mold opening process, and an ejection process. Hereinafter, these processes related to manufacture are sometimes collectively referred to as "manufacturing processes". In addition, a series of actions for obtaining a molded product, such as the actions from the start of a metering process to the start of the next metering process in the above-mentioned manufacturing process, are referred to as "injection", "molding cycle", etc. In addition, the above-mentioned processes for manufacturing a molded product are only examples. For example, as a process performed by one injection, other processes not included in the above may be included.

[0030] The control unit 110 controls the injection device 20 and the clamping device 30 according to the control information. The control information is a condition set by the user, for example, generated according to information input by the user using an input device not shown. The control information includes, for example, molding conditions such as resin temperature, mold temperature (cylinder temperature), injection holding time, metering value, VP switching position, holding pressure, injection speed (filling speed), screw speed, screw back pressure, and clamping force. Regarding these molding conditions, multiple combinations are determined according to the molded product or mold. Hereinafter, the combination data of the molding conditions will also be referred to as a molding condition data set. In addition, the control information includes control data such as voltage, current, pressure, speed, acceleration, etc. for a driving unit such as a mechanism or a motor. The molding condition data set is prepared according to the type of molded product or mold, and is stored in the storage unit 120.

[0031] The control unit 110 uses the above-mentioned molding condition data set to control the injection device 20 and the mold clamping device 30 to implement the processes related to the manufacture (injection) of the molded product, including the above-mentioned processes. When starting to manufacture the molded product, the control unit 110 reads the molding condition data set corresponding to the molded product to be manufactured from the storage unit 120. Then, the control unit 110 controls the operation of the injection device 20 and the mold clamping device 30 based on the control information including the read molding condition data set. Specifically, the control unit 110 controls the injection device 20 and the mold clamping device 30 so that the data obtained from the injection device 20 and the mold clamping device 30 in the manufacturing process are consistent with the set value of the molding condition data set.

[0032] The storage unit 120 stores control information used by the control unit 110 to control the injection device 20 and the mold clamping device 30. The molding condition data set included in the control information is prepared by establishing a corresponding association with the molded product or mold of the manufacturing object. The storage unit 120 stores the molding condition data set for each molded product or each mold of the manufacturing object. In addition, although not shown in the figure, the storage unit 120 stores a program for the control unit 110 to control the injection device 20 and the mold clamping device 30. The details will be described later, but the functions of the control unit 110 can be realized by the processor in the control device 100 reading and executing the program stored in the storage unit 120.

[0033] <Configuration of Data Processing Device 200 >

[0034] Figure 3 2 is a diagram showing the structure of a data processing device 200. The data processing device 200 acquires and processes data obtained as the injection device 20 and the mold clamping device 30 perform the operations in the above-mentioned process related to the manufacture of the molded product. The data processing device 200 is implemented by a computer, for example. The data processing device 200 includes a data acquisition unit 210, a processing unit 220, and a storage unit 230.

[0035] The data acquisition unit 210 acquires data of a processing object from the injection device 20 and the mold clamping device 30. Various sensors, detectors, etc. are installed on the injection device 20 and the mold clamping device 30. In addition, various measuring devices are sometimes connected to the injection device 20 or the mold clamping device 30. The data acquired using these sensors, detectors, measuring devices, etc. (hereinafter referred to as "acquired data") is information indicating the molding results based on the injection device 20 and the mold clamping device 30, and can be used for quality management of molded products. The data acquisition unit 210 receives the acquired data sent from the sensors, detectors, measuring devices, etc., and stores it in the storage unit 230.

[0036] The processing unit 220 processes the acquired data stored in the storage unit 230. Specifically, the processing unit 220 extracts representative values ​​of the acquired data in each process, generates time series data by time-series the acquired data in each process, and the like. In extracting the representative value, the processing unit 220 performs statistical processing such as calculating the average value, determining the range of values, and determining the maximum or minimum value on the acquired data.

[0037] The storage unit 230 stores the acquired data acquired by the data acquisition unit 210. As the data format of the acquired data stored in the storage unit 230, for example, binary, text, CSV (Comma Separated Values), INI, YAML (YAML Ain't Markup Language), JSON (JavaScript Object Notation), etc. can be used. By using data files based on these general data formats, it is possible to exchange data between the data files stored in the storage unit 230 and other information processing devices, or edit data files acquired from external devices. In addition, although not shown in the figure, the storage unit 230 stores programs for the processing unit 220 to perform data processing. The details will be described later, but the functions of the processing unit 220 can be realized by the processor in the data processing device 200 reading and executing the program stored in the storage unit 230.

[0038] <Configuration of Information Processing Device 400 >

[0039] Figure 4 4 is a diagram showing the structure of an information processing device 400. The information processing device 400 obtains waveform data obtained by the operation of the injection molding machine 10 to obtain the characteristics of the injection molding machine 10, and outputs information for controlling the injection molding machine 10 based on the information of the obtained characteristics. The information processing device 400 is implemented by, for example, a computer. The information processing device 400 includes a communication unit 410, a processing unit 420, and a storage unit 430. The processing unit 420 includes an analysis unit 421 that analyzes the characteristics of the injection molding machine 10 using a machine learning model. The storage unit 430 stores the machine learning model 431 used in the analysis based on the analysis unit 421.

[0040] The communication unit 410 receives waveform data obtained by the operation of the injection molding machine 10 from the control device 100 and the data processing device 200. The data to be received can be various data obtained as waveform data when the injection molding machine 10 is in operation. Specifically, for example, data such as voltage and current supplied to various mechanisms, speed, acceleration, and pressure in the operation of various mechanisms can be cited. In addition, the communication unit 410 sends information for controlling the injection molding machine 10 obtained by the processing of the processing unit 420 to the control device 100. The communication unit is an example of an acquisition unit.

[0041] The processing unit 420 performs the following processing: using the waveform data received through the communication unit 410, the characteristics of the injection molding machine 10 are obtained. The processing unit 420 extracts the input variables as inputs to the machine learning model and the actual values ​​used for verification from the acquired waveform data, and inputs them to the analysis unit 421. The input variable is an example of the first variable, and the actual value is an example of the second variable. Here, the input variable and the actual value are variables that have a relationship with each other. That is, if the value of the input variable changes, the value of the actual value also changes accordingly. In addition, the processing unit 420 outputs information for controlling the injection molding machine 10 based on the analysis results based on the analysis unit 421, and causes the communication unit 410 to send it to the control device 100. The processing unit 420 is an example of an output unit.

[0042] The analysis unit 421 inputs the input variables and the input variables in the actual value into the machine learning model 431, and derives the variables equivalent to the actual value (hereinafter referred to as "output value"). The analysis unit 421 is an example of a machine learning device. Next, the analysis unit 421 compares the actual value with the output value and calculates the error. Then, the analysis unit 421 adjusts the machine learning model 431 so that the error of the output value relative to the actual value becomes smaller. The details of the processing performed by the analysis unit 421 and the machine learning model 431 will be described later.

[0043] The storage unit 430 stores waveform data received by the communication unit 410, information obtained by processing by the processing unit 420, and a machine learning model 431 used by the analysis unit 421 of the processing unit 420. As a data format of the data file stored in the storage unit 430, for example, CSV, XML (Extensible Markup Language), JSON, etc. can be used.

[0044] <Hardware Configuration of Information Processing Device 400>

[0045] Figure 5 It is a diagram showing an example of the hardware configuration of a computer that realizes the information processing device 400 . Figure 5The computer shown has a processor 401 as a computing unit, a main storage device (main memory) 402 as a storage unit, and an auxiliary storage device 403. As the processor 401, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or other various computing circuits can be used. The processor 401 reads the program stored in the auxiliary storage device 403 into the main storage device 402 for execution. As the main storage device 402, for example, a RAM (Random Access Memory) is used. As the auxiliary storage device 403, for example, a magnetic disk device or an SSD (Solid State Drive) is used.

[0046] Furthermore, the computer may be configured as follows: the computer may include a display device 404 for displaying images and an input device 405 as an input mechanism for the computer to be operated by the user. As the input device 405, for example, a keyboard, a mouse, a touch panel, etc. are used. When a touch panel integrally formed with the display device 404 is used as the input device 405, the user performs input operations by touching the operation screen displayed on the display device 404 with a finger or a pen-shaped device. Figure 5 The structure of the computer shown is only an example, and the computer used in this embodiment is not limited to Figure 5 For example, a nonvolatile memory such as a flash memory or a ROM (Read Only Memory) may be provided as a storage device.

[0047] In the information processing device 400, Figure 5 In the case of computer implementation shown, the communication unit 410 is implemented by, for example, the processor 401 that reads and executes a program and a communication interface (not shown). The function of the processing unit 420 is implemented by, for example, the processor 401 reading and executing a program. The storage unit 430 is implemented by, for example, the auxiliary storage device 403.

[0048] <Mechanical error of the controlled equipment>

[0049] Figure 6 is a diagram schematically showing the adjustment of mechanical errors. Figure 6 (A) is a diagram showing an example of mechanical error, Figure 6 (B) is a diagram showing a state where the mechanical error has been adjusted. Figure 6 (A) Figure 6 (B) is a waveform graph in which the vertical axis represents speed and the horizontal axis represents time when the speed of the control target device changes over time. Figure 6 (A) Figure 6 In (B), the dotted line is a waveform diagram representing control information, and the solid line is a waveform diagram representing actual data.

[0050] Injection molding machines 10 and other equipment generally have some mechanical errors even if they are the same model, so there are differences in operation. For example, consider changing the current value of the motor, such as Figure 6 In this case, even if the motor current value can be controlled to reach the desired state, if there is a mechanical error, Figure 6 As shown in (A), the speeds obtained in each device are deviated. Even if the speed or position is detected by a sensor and feedback is given, and the current value is controlled to finally converge to the desired speed, the time lag or overshoot generated before reaching the desired speed will vary depending on the mechanical error. Figure 6 In the example shown in (A), the speed change (solid line) of three devices controlled by one control information (dashed line) is shown. Figure 6 In the example shown in (A), three devices are controlled by the same control information but show different speed changes.

[0051] In this embodiment, such mechanical error is treated as a characteristic of each device, and the motor current value is controlled to suppress the mechanical error, thereby achieving speed control that suppresses the mechanical error. Figure 6 In the example shown in (B), according to Figure 6 The same control information (dashed line) as shown in the example (A) shows the speed change status of the three devices. Figure 6 In the example shown in (B), the mechanical errors of the three devices are suppressed, and the same speed changes are shown by control based on the same control information.

[0052] Here, the relationship between the current value and the speed is used as an example for explanation, but the application object of this embodiment is not limited to the above example. There are physical quantities that are related to each other among the physical quantities detected by sensors in the device. This embodiment can be applied to physical quantities that have such a relationship and such as the relationship that deviates depending on the device. For example, the relationship between current and voltage or the relationship between speed and acceleration can be cited. In the relationship between current and voltage, there is a mechanical error in the resistance, so even if the same amount of current flows, the voltage value obtained sometimes varies depending on the device. In addition, in the relationship between speed and acceleration, there is a mechanical error in the inertia of the action object part in the device, so even if the action is performed at the same speed, the acceleration sometimes varies depending on the device. By applying this embodiment to various physical quantities in these control object devices, it is possible to suppress the mechanical error of each device and control it to perform the same action.

[0053] The information processing device 400 analyzes as shown in reference Figure 6 (A) Figure 6 The characteristics of each injection molding machine 10 as the control target device are determined by using waveform data of physical quantities having a relationship with each other as described in (B). Then, information for performing control corresponding to the characteristics of each injection molding machine 10 is generated and provided to the control device 100. As a waveform data analysis unit, the processing unit 420 of the information processing device 400 uses a machine learning model 431. As the machine learning model 431, for example, a linear multiple regression model can be used.

[0054] In addition, as an analysis mechanism for waveform data, in addition to the linear multiple regression model, a linear regression model or other mathematical models can also be used. However, by using a linear multiple regression model, it is easier to construct a learning model than other methods. This is because the waveform data has multiple variables, and it is possible to collect correct answer data related to input and output, and the data that can become the multiple input variables of the waveform data are data that can be represented by four arithmetic operations (also including data expressed by approximate calculation formulas). Therefore, it can be said that the analysis of waveform data is well compatible with the characteristics of supervised machine learning for linear multiple regression models. Therefore, compared with the situation of manually adjusting various parameters such as voltage, current, and speed, it is possible to use a test device to make the machine move to obtain training data, build a machine learning model by learning it, and determine the parameters through calculations using the constructed machine learning model in a very short time.

[0055] Figure 74 is a diagram showing the structure of a linear multiple regression model. The linear multiple regression model used in the machine learning model 431 receives multiple inputs, performs analysis taking into account the correlation between the inputs through multiple nodes representing multiple processing units, and obtains a single output. Figure 7 In the model shown, x 1 ~x 3 is the input variable, y 1 is the output. The w recorded in each node of the middle layer 1 ~w 4 represents the weight value set at each node. Figure 7 In the model shown, if the input variable x is provided 1 ~x 3 , then by using the weight value w in each node 1 ~w 4 The output y is obtained by processing 1 Then, by changing the weight value w of each node 1 ~w 4 , with respect to the same input variable x 1 ~x 3 Different outputs y can be obtained 1 .in addition, Figure 7 The numbers of input variables and nodes shown are examples only and are not limited to the numbers shown in the figure.

[0056] In the Figure 7 The linear multiple regression model shown is applied to the case of the relationship between current and voltage, such as the actual current is input to the input variable x 1 ~x 3 , for example, the voltage command can be obtained as output y 1 Then, the weight value w of each node 1 ~w 4 For example, set the physical constant. Weight value w 1 ~w 4 The type of physical constants set in is selected according to the physical quantity of the analysis object. For example, when the actual current of the motor is used as input and the voltage command is used as output, the physical constants such as the motor resistance value (R), d-axis inductance (Ld), q-axis inductance (Lq), back electromotive force constant (Ke), torque constant (Kt), viscous friction (Dm), inertia (J) can be assigned to each node and the weight value can be set.

[0057] Figure 8 4 is a diagram showing an example of analysis processing by the analysis unit 421. The analysis unit 421 uses the reference Figure 7The supervised learning of the machine learning model 431 (linear multiple regression model) is described, and the waveform data obtained from the injection molding machine 10 as the control target device is analyzed to determine the characteristics of the device. The form of the waveform data is not particularly limited. For example, it can be data that is tabulated based on values ​​such as CSV, or it can be a waveform graph. Here, as an example, the machine learning model 431 is a model that obtains voltage and acceleration based on the actual current and actual speed of the motor. In addition, Figure 8 In the example, the intermediate layers of the machine learning model 431 are aggregated into one node for recording.

[0058] The analysis unit 421 extracts the d-axis current, q-axis current, and speed from the waveform data acquired from the injection molding machine 10 as input variables X, and inputs them to the machine learning model 431. In addition, the analysis unit 421 extracts the actual values ​​(Y (actual results)) of the d-axis voltage, q-axis voltage, and acceleration from the input waveform data. These actual values ​​correspond to the output contents calculated by the machine learning model 431 based on the d-axis current, q-axis current, and speed as input variables.

[0059] On the other hand, the machine learning model 431 derives the d-axis voltage, q-axis voltage, and acceleration as outputs (Y) based on the input d-axis current, q-axis current, and speed. Here, the d-axis voltage, q-axis voltage, and acceleration are obtained from the d-axis current, q-axis current, and speed, respectively.

[0060] Next, the analysis unit 421 compares the actual values ​​(Y(actual results)) of the d-axis voltage, q-axis voltage, and acceleration extracted from the waveform data with the output (Y) of the d-axis voltage, q-axis voltage, and acceleration obtained using the machine learning model 431, and calculates the error. Then, the weight value W of each node in the middle layer is corrected so as to minimize the error, in other words, so as to make the output (Y) of the d-axis voltage, q-axis voltage, and acceleration consistent with the actual values ​​(Y(actual results)) of the d-axis voltage, q-axis voltage, and acceleration. Then, if the weight value W that minimizes the error between the output (Y) of the d-axis voltage, q-axis voltage, and acceleration and the actual values ​​(Y(actual results)) of the d-axis voltage, q-axis voltage, and acceleration is obtained, the weight value W represents the characteristics related to the motor of the injection molding machine 10 which is the acquisition source of the input waveform data. The difference in this characteristic in the injection molding machine 10 as the control target device is an example of a mechanical error of each injection molding machine 10.

[0061] In the above description, the analysis unit 421 corrects the weight value W so that the error between the output (Y) of the d-axis voltage, the q-axis voltage, and the acceleration and the actual value (Y(actual)) of the d-axis voltage, the q-axis voltage, and the acceleration is minimized, but this is only an example. For example, instead of the above processing, the analysis unit 421 may also correct the weight value W so that the error is less than a preset threshold. At this time, by using the analysis of the machine learning model 431 with the corrected weight value W, the output (Y) of the d-axis voltage, the q-axis voltage, and the acceleration that is close to the actual value (Y(actual)) of the d-axis voltage, the q-axis voltage, and the acceleration can be obtained.

[0062] In addition, Figure 8 In the example of the analysis process shown, an example of comparing the actual values ​​(Y(actual results)) of the d-axis voltage, q-axis voltage, and acceleration extracted from the waveform data with the outputs (Y) of the d-axis voltage, q-axis voltage, and acceleration obtained using the machine learning model 431 and calculating the error is described, but these variables are not limited to the above example, and other variables with a causal relationship can also be used to compare and calculate the error. For example, the actual value and output of the above acceleration can be replaced by a value obtained using the time derivative of the speed or an approximate calculation formula. In addition, the value of the duty cycle (on-off ratio) of the inverter can also be used to replace the actual value and output of the voltage. In addition, as the speed in the input, the value of the speed calculated based on the difference of the encoder position data can also be used instead of the speed data directly obtained by the sensor, etc. In addition, as the input, the voltage value output from the current detector can also be used instead of the actual current.

[0063] <Adjustment of characteristics of controlled equipment>

[0064] The processing unit 420 of the information processing device 400 generates information (hereinafter referred to as "characteristic adjustment information") for performing control such as suppressing mechanical errors in the injection molding machine 10 with other injection molding machines based on the information of the characteristics of the injection molding machine 10 obtained as described above. As an example of the characteristic adjustment information, the processing unit 420 can generate a gain for the control device 100 to correct the control information used for controlling the injection device 20 and the clamping device 30. The gain is a coefficient multiplied by the value used as the control information in order to compensate for the deviation based on the characteristics of the injection molding machine 10. The control device 100 uses the gain obtained from the information processing device 400 to correct the control information and control the injection device 20 and the clamping device 30. In this way, it is not necessary to design and change the control information itself possessed by the control unit 110 of the control device 100 according to the characteristics (weight value W) of the injection molding machine 10 analyzed by the analysis unit 421. Since it is only necessary to correct the control information already possessed by the control unit 110, it is possible to suppress the increase in the scale of the design change.

[0065] In contrast, the processing unit 420 may generate control information used by the control device 100 for controlling the injection device 20 and the mold clamping device 30 based on the characteristics (weight value W) of the injection molding machine 10 analyzed by the analysis unit 421 as characteristic adjustment information. The control device 100 replaces the control information possessed by the control unit 110 with the control information acquired from the information processing device 400, and controls the injection device 20 and the mold clamping device 30. Furthermore, the processing unit 420 may directly send the information of the characteristics (weight value W) of the injection molding machine 10 analyzed by the analysis unit 421 to the control device 100 as characteristic adjustment information. At this time, the control device 100 generates a gain to correct the control information based on the information of the characteristics of the injection molding machine 10 acquired from the information processing device 400, or generates control information reflecting the characteristics.

[0066] Fig. 9 1 is a flowchart showing the procedure of adjusting control information according to the characteristics of the equipment in the injection molding machine 10. First, in the injection molding machine 10, the control device 100 operates the injection device 20 and the mold clamping device 30 (S101), and the data processing device 200 obtains sensor information detected along with the operation of the injection device 20 and the mold clamping device 30 (S102). Then, the control device 100 and the data processing device 200 send the waveform data obtained by the operation of the injection device 20 and the mold clamping device 30 to the information processing device 400 (S103).

[0067] The information processing device 400 analyzes the waveform data received from the control device 100 and the data processing device 200 to obtain the characteristics of the injection molding machine 10 (S104), and transmits the characteristic adjustment information based on the obtained characteristics to the control device 100 of the injection molding machine 10 (S105). When the control device 100 receives the characteristic adjustment information from the information processing device 400 (S106), the control information for controlling the injection device 20 and the mold clamping device 30 is adjusted according to the received characteristic adjustment information (S107). Thereafter, the control device 100 of the injection molding machine 10 controls the injection device 20 and the mold clamping device 30 according to the control information adjusted to reflect the inherent characteristics of each device. Then, by changing the control performed by the control device 100, the control corresponding to the characteristics of the device is performed, and the mechanical error with other devices in the control of the device can be suppressed.

[0068] The embodiments of the present invention have been described above, but the technical scope of the present invention is not limited to the above embodiments. For example, in the above embodiments, the waveform data is analyzed in the information processing device 400 connected to the injection molding machine 10, and the characteristic adjustment information is generated. In contrast, a configuration may be provided in which the waveform data is analyzed in the data processing device 200 of the injection molding machine 10, and the characteristic adjustment information is generated.

[0069] Furthermore, in the above-mentioned embodiment, the mechanical error of each device is suppressed, but in the same device, even when the characteristics change due to differences in the operating environment, the present embodiment can be applied. For example, the characteristics of the injection molding machine 10 change due to changes in the type of resin or mold used as the molding material, changes in the external air temperature, etc. Therefore, it is also possible to set a structure in which waveform data is acquired according to changes or variations in these environments, the characteristics of the device under each environment are determined, and control corresponding to the characteristics is performed.

[0070] Furthermore, in the present embodiment, when the gain in the device adjustment information generation control is used, it can be widely applied to the proportional gain and integral gain in the PI control of current, speed, pressure, etc., the coefficient of the noise filter, the gain of the feedforward, etc., for which the design method is established in the control engineering. In addition, various changes or structural substitutions that do not depart from the scope of the technical idea of ​​the present invention are also included in the present invention.

Claims

1. An information processing system, characterized in that: have: An acquisition unit that acquires waveform data obtained by controlling the operation of the target device; an analyzing unit that extracts a first variable as at least a part of the variables and a second variable as at least another part of the variables based on the relationship of the variables included in the acquired waveform data, derives a variable corresponding to the second variable based on the first variable, and obtains the characteristic of the controlled object device based on the derived variable and the second variable; and The output unit outputs information for controlling the control target device according to the characteristics of the control target device.

2. The information processing system according to claim 1, characterized in that The analysis unit uses a linear multiple regression model that receives a plurality of the first variables as input and obtains one output through a plurality of processes in an intermediate layer, and derives the variable corresponding to the second variable as the output.

3. The information processing system according to claim 2, characterized in that: The analyzing unit corrects weight values ​​set for a plurality of processes in the intermediate layer so as to reduce an error between the derived variable and the second variable, and uses the corrected weight values ​​as information indicating characteristics of the controlled device.

4. The information processing system according to claim 3, characterized in that: The information output by the output unit is information on a gain in controlling the controlled device, generated based on the corrected weight value.

5. A program, characterized in that Make the computer function as the following: Acquiring waveform data obtained by controlling the operation of the target device; Extracting a first variable as at least a part of the variables and a second variable as at least another part of the variables based on the relationship of the variables included in the acquired waveform data, deriving a variable corresponding to the second variable based on the first variable, and obtaining the characteristics of the controlled object device based on the derived variables and the second variable; and Information for controlling the control target device is outputted according to the characteristics of the control target device.

6. A method for manufacturing an injection molding machine, characterized in that: The steps include: The injection molding machine is operated to obtain waveform data; Extracting a first variable as at least a part of the variables and a second variable as at least another part of the variables based on the relationship of the variables included in the acquired waveform data, deriving a variable corresponding to the second variable based on the first variable, and obtaining the characteristics of the injection molding machine based on the derived variables and the second variable; and The control information of the injection molding machine is adjusted according to the characteristics of the injection molding machine.

Images