Molding condition parameter estimation method, computer program, estimation device, and estimation system

Abstract

Provided are a molding condition parameter estimation method and the like capable of appropriately adjusting a molding condition parameter of a film molding machine. Detection data relating to the state of a film forming machine detected by a detector is acquired, and forming condition parameters in the film forming machine are estimated by inputting the acquired detection data into an estimation model. The estimation model estimates molding condition parameters in the film forming machine when detection data relating to the state of the film forming machine is input.

Description

TECHNICAL FIELD

[0001] The present application relates to a molding condition parameter inference method, a computer program, an inference device, and an inference system. BACKGROUND

[0002] In the past, a molding machine for molding a resin material by a molding method such as injection molding or extrusion molding has been widely used. In the molding machine, it is necessary to adjust the set values of various molding condition items in the molding machine so that the molded body satisfies the required specifications. The adjustment of these molding conditions is made on the basis of the experience of an operator, and in order to obtain appropriate molding conditions, trial and error is required. Therefore, a technology for supporting the adjustment work of the molding condition parameters performed by the operator has been proposed.

[0003] For example, Patent Literature 1 discloses an injection molding machine system capable of appropriately adjusting molding condition parameters of an injection molding machine using a machine learning device that learns by reinforcement learning.

[0004] PRIOR ART DOCUMENTS

[0005] PATENT LITERATURE

[0006] Patent Literature 1: JP-A No. 2017-30152 SUMMARY

[0007] However, the technology described in Patent Literature 1 is a technology related to an injection molding machine, and is not a technology aimed at inferring molding condition parameters of a film molding machine that molds a film by extrusion molding. Since the machine configuration of the injection molding machine is different from that of the film molding machine, it is difficult to directly apply the technology related to the injection molding machine to the film molding machine. A technology capable of inferring molding condition parameters of the film molding machine is desired.

[0008] An object of the present disclosure is to provide a molding condition parameter inference method and the like capable of appropriately inferring molding condition parameters of a film molding machine.

[0009] In the molding condition parameter inference method of one aspect of the present disclosure, detection data related to the state of a film molding machine detected by a detector is acquired, and molding condition parameters in the film molding machine are inferred by inputting the acquired detection data to an inference model that infers the molding condition parameters in the film molding machine when detection data related to the state of the film molding machine is input.

[0010] The computer program of one aspect of the present disclosure is a computer program that causes a computer to execute processing of acquiring detection data detected by a detector regarding a state of a film forming machine, and inferring a forming condition parameter in the film forming machine by inputting the acquired detection data to an inference model that infers the forming condition parameter in the film forming machine when detection data regarding a state of a film forming machine is input.

[0011] The inference device of one aspect of the present disclosure includes an acquisition unit that acquires detection data detected by a detector regarding a state of a film forming machine, and an inference unit that infers a forming condition parameter in the film forming machine by inputting the acquired detection data to an inference model that infers the forming condition parameter in the film forming machine when detection data regarding a state of a film forming machine is input.

[0012] The inference system of one aspect of the present disclosure includes a film forming machine, a detector, and an inference device that includes an acquisition unit that acquires detection data detected by the detector regarding a state of the film forming machine, and an inference unit that infers a forming condition parameter in the film forming machine by inputting the acquired detection data to an inference model that infers the forming condition parameter in the film forming machine when detection data regarding a state of a film forming machine is input.

[0013] Effects of Invention

[0014] According to the present disclosure, it is possible to appropriately infer a forming condition parameter of a film forming machine. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 is a schematic diagram of an inference system of the first embodiment.

[0016] Figure 2 is a block diagram showing a configuration example of a data collection device.

[0017] Figure 3 is a block diagram showing a configuration example of an inference device.

[0018] Figure 4 is a diagram showing a content example of information stored in a forming information DB.

[0019] Figure 5 is an explanatory diagram showing an outline of an inference model.

[0020] Figure 6 is a flowchart showing an example of a processing sequence performed by an inference device.

[0021] Figure 7 is a flowchart showing an example of a processing sequence performed by an inference device of the second embodiment. DETAILED DESCRIPTION

[0022] The present disclosure will be specifically described with reference to the drawings showing embodiments of the present disclosure.

[0023] (First Embodiment)

[0024] Figure 1 is a schematic diagram of an inference system 100 of the first embodiment. The inference system 100 has, as main devices, a film forming machine 1, a plurality of detectors 2, and an inference device 4. The inference device 4 is communicably connected with a data collection device 3 and a terminal device 5 via a network N1 such as the Internet.

[0025] Figure 1 One film forming machine 1 and the data collection device 3 are shown, but the inference device 4 is connected with a plurality of data collection devices 3 not shown via the network N1. One or a plurality of film forming machines 1 are connected with each data collection device 3. Each data collection device 3 and one or a plurality of film forming machines 1 connected with the data collection device 3 are provided, for example, in the same factory.

[0026] The inference device 4 is a device capable of various information processing, information transmission and reception, and is, for example, a server computer, a personal computer, a quantum computer, or the like. The inference device 4 collects molding information of each of the plurality of film forming machines 1. The inference device 4 infers molding condition parameters in the film forming machine 1 in such a manner that the quality of the film molded in each film forming machine 1 satisfies a target quality based on the collected molding information. The inference device 4 of the present embodiment 1 adjusts a plurality of molding condition parameters set in the film forming machine 1 in such a manner that the thickness (film thickness) of the film becomes within a prescribed range. The inference device 4 transmits the inference result of the molding condition parameters to the terminal device 5.

[0027] The terminal device 5 is an information processing terminal having a communication function and a display function, and is, for example, a personal computer, a tablet terminal, a smartphone, or the like. The terminal device 5 can also be provided in the same factory as the film forming machine 1. The terminal device 5 can be portable. The terminal device 5 is a terminal used by a user of the film forming machine 1. The user of the film forming machine 1 includes an organization such as a corporation that owns the film forming machine 1, an operator belonging to the organization, and the like. The user of the film forming machine 1 can also be a service provider such as a sales person, a maintenance and management person, or the like, associated with the film forming machine 1. The number of terminal devices 5 can be two or more.

[0028] < Film Forming Machine 1 >

[0029] The film forming machine 1 forms a resin film (film) by extrusion molding. The film forming machine 1 has an extruder 11, a casting device 12, an MD stretching device 13, a TD stretching device 14, a winding machine 15, and a control device 16.

[0030] The extruder 11 is, for example, a single-screw extruder or a twin-screw extruder, and has a raw material supply part 111 into which a raw material of a resin is fed, a cylinder 112, a screw 113, and a die 114. The raw material supply part 111 includes a hopper that stores the raw material of the resin and a feeder that feeds the raw material of the resin into the hopper, and discharges the raw material of the resin fed into the hopper into the cylinder 112 at a constant flow rate. The screw 113 is rotatably inserted into a hole of the cylinder 112, conveys the raw material of the resin discharged from the raw material supply part 111 to an extrusion direction (right direction in the drawing), and melts and kneads the raw material. The screw 113 is rotated at a predetermined rotational speed by a driving force output from a driving device 115 provided with, for example, a motor and a speed reducer. The extruder 11 extrudes the melted raw material of the resin from a front end narrow gap (die lip) of the die 114 in a film shape. The die 114 is, for example, a T-die. Figure 1

[0031] The casting device 12 is provided with a plurality of casting rollers 121 that cool and shape the high-temperature melt extruded from the die 114. The casting rollers 121 are rotated at a predetermined rotational speed by a driving force output from a driving device 122 provided with, for example, a motor and a speed reducer. The plurality of casting rollers 121 include a first roller 121a and a second roller 121b.

[0032] The first roller 121a is, for example, a metal roller having a temperature adjustment part (not shown) that cools the melt, and is pivotally supported below the die 114. The film-shaped melt extruded from the die 114 is sandwiched between the first roller 121a and the second roller 121b, and is cooled in a short time and shaped into a film (sheet) shape together with the second roller 121b. The casting device 12 controls the film thickness within a predetermined range. Thus, an unstretched film is obtained. The temperature adjustment method of the first roller 121a is not particularly limited, but for example, a method using a heat medium based on air, water, oil, or the like, or a method using an electric heater, dielectric heating, or the like is exemplified. In the example shown in the drawing, the plurality of casting rollers 121 further include a roller for cooling or conveying the melt. Figure 1

[0033] The MD stretching device 13 is provided with a plurality of tension rollers 131, and stretches the unstretched film conveyed from the casting device 12 while the film is interposed between the tension rollers 131 in the MD direction. The MD direction is a direction along the conveying direction of the film, and is also referred to as the longitudinal direction. The tension rollers 131 are rotated at a predetermined rotational speed by a driving force output from a driving device 132 provided with, for example, a motor and a speed reducer.

[0034] ​​The plurality of tension rollers 131 include a heating roller having a temperature adjustment section that heats the film, and a cooling roller having a temperature adjustment section that cools the film. The same method as that of the casting roller 121 described above is cited as a temperature adjustment method of the tension rollers 131. The film is heated to a prescribed temperature range in which the film can be stretched while being in contact with the heating roller, and thereafter, is stretched in the MD direction by the difference in rotational speed of each cooling roller. The stretching ratio in the MD direction can be adjusted by the speed ratio of each tension roller 131. Further, Figure 1 The number of tension rollers 131 is not limited to Figure 1 the example shown.

[0035] The TD stretching device 14 stretches the film that has been stretched in the MD direction by the MD stretching device 13 in the TD direction. The TD direction is a direction that intersects the conveyance direction of the film (the width direction of the film), and is also called the transverse direction.

[0036] The TD stretching device 14 is, for example, a tenter stretching device, and has a heating device such as a hot air blowing device, and heats the film to a prescribed temperature range in which the film can be stretched and stretches the film in the transverse direction.

[0037] The TD stretching device 14 has an oven 141 having a heating device such as a hot air blowing device, and a traveling mechanism including a pair of rails 142 and a plurality of links 143 arranged on the rails 142 is provided in the oven 141. The rails 142 are arranged so as to expand in the TD direction toward the downstream direction in the MD direction. The plurality of links 143 form a looped chain. The looped chain formed by the plurality of links 143 travels on the rails 142 at a prescribed speed by rotational drive force output from a link drive device 145 arranged on the outside of the oven 141.

[0038] A gripper 144 that is a film holding member is attached to each of the plurality of links 143. The number of links 143 and the number of grippers 144 arranged on each of the left and right rails 142 are the same. That is, the TD stretching device 14 has a pair of left and right links 143 and a pair of left and right grippers 144. The pair of links 143 and the pair of grippers 144 are arranged on the rails 142 so as to be located at both ends of the film, that is, so as to be arranged in a row in the TD direction. By synchronizing the pre-drive devices, the pair of left and right grippers 144 are adjusted so as not to have a phase difference, that is, so as to have the same conveyance cycle. The TD stretching device 14 can also be a bearing clip type or a sliding clip type.

[0039] At the inlet of the TD stretching device 14, the ends of both sides of the film are held by the grippers 144. The grippers 144, with the film held, are conveyed together with the link 143 in the film conveying direction, and the film is conveyed in the film conveying direction by traveling on the rail 142. The two surfaces of the film in the oven are heated by air blown from the hot air blowers provided on the upper side and the lower side of the traveling mechanism, and the film is stretched in the width direction in conjunction with the movement of the grippers 144 holding the ends of the film. The stretching ratio in the TD direction can be adjusted by the amplitude of the rail 142 in the TD direction. The oven 141 is, for example, divided into a plurality of zones in the conveying direction, and the set temperature or the air volume can be adjusted for each zone. The above-mentioned zones include a preheating zone, a stretching zone, and a heat fixing zone.

[0040] The grippers 144 release the film at the outlet of the TD stretching device 14. The film stretched in the width direction is wound by the winder 15 via a puller not shown. Furthermore, the TD stretching device 14 is not limited to a gripper tenter, and can be, for example, a pin tenter having pins as film holders. The above-mentioned explanation of the configuration of each part of the TD stretching device 14 is an example, and the configuration of the temperature adjustment, the air volume adjustment, the link 143 of the rail 142, and the like is not particularly limited.

[0041] The film forming machine 1 is continuously operated without stopping at ordinary times, except for a specific timing such as occurrence of a device abnormality, a preset stop period.

[0042] The control device 16 is a computer that performs operation control of the film forming machine 1. The control device 16 has a control section such as a CPU (Central Processing Unit) not shown, a storage section such as a ROM (Read Only Memory), a RAM (Random Access Memory), a hard disk, a communication section that transmits and receives information with external devices through a wire or wirelessly, a display section that displays various images on a liquid crystal display or an organic EL (Electro Luminescence) display of a user, an operation section that receives operations from the user, and the like.

[0043] The control device 16 receives a setting of a molding condition parameter with respect to the film molding machine 1 through an operation section or a communication section. In addition, the control device 16 receives the detection data detected by the detector 2 from the data collection device 3 through the communication section. The control device 16 controls the operation of the film molding machine 1 based on the received molding condition parameter and the detection data. In addition, the control device 16 transmits operation data indicating the operation state of the film molding machine 1 to the data collection device 3 through the communication section. The data collection device 3 transmits the received operation data to the inference device 4. Further, the operation data can be transmitted directly to the inference device 4 from devices such as motors, inverters, and the like in the film molding machine 1 without passing through the collection device 3.

[0044] The film molding machine 1 is provided with a molding condition parameter of a prescribed molding condition, and operates in accordance with the molding condition parameter. As the molding condition parameter, for example, the feeder supply amount in the extruder 11, the rotational speed or rotation rate of the screw 113, the rotational speed or rotation rate of the motor of the driving device 115, and the die lip opening of the die 114, the rotational speed or rotation rate of the casting roll 121 in the casting device 12, the rotational speed or rotation rate of the motor of the driving device 122, and the discharge amount of the film, the rotational speed or rotation rate of the tension roll 131 in the MD stretching device 13, the rotational speed or rotation rate of the motor of the driving device 132, the stretching ratio, and the discharge amount of the film, the line speed of the film conveying line in the TD stretching device 14, the temperature in the oven 141 (particularly for each zone), the pressure in the oven 141 (particularly for each zone), the air volume or air speed in the inlet and outlet (particularly the air blowing nozzles of the air blowing device near the inlet and outlet) of the oven 141, the stretching angle, the stretching ratio, and the discharge amount of the film, and the like are exemplified. Further, instead of the discharge amount of the film, the film speed, the film width, and the film thickness can be set. The optimum molding condition parameter changes at any time during the operation depending on the environment of the film molding machine 1 and the state of the molded product.

[0045] As the operation data of the film molding machine 1, for example, the feeder supply amount in the extruder 11, the temperature of the cylinder 112, the rotational speed of the screw 113, the rotational speed of the motor, the torque of the motor, and the motor current, the rotational speed of the casting roll 121 in the casting device 12, the rotational speed of the motor, the torque of the motor, and the motor current, the rotational speed of the tension roll 131 in the MD stretching device 13, the rotational speed of the motor, the torque of the motor, and the motor current, and the line speed and the line tension in the TD stretching device 14, and the like are exemplified. The operation data corresponds to the set value of each item, that is, the control value based on the control device 16.

[0046] <Detector 2>

[0047] The detector 2 is a sensor that detects a physical quantity indicating a state of the film forming machine 1 in a time series. The state of the film forming machine 1 detected by the detector 2 includes, on the basis of a state of the film forming machine 1 itself, also a state of a film (resin) formed by the film forming machine 1. The detector 2 is connected to the data collection device 3 by wire or wirelessly, and outputs detection data obtained by detection directly or indirectly to the data collection device 3.

[0048] The inference system 100 is provided with a plurality of detectors 2 that detect various detection data. Some of the plurality of detectors 2 are connected to the data collection device 3, and the data collection device 3 acquires detection data from the detectors 2. Some of the plurality of detectors 2 are connected to the control device 16, and the data collection device 3 acquires detection data from the detectors 2 via the control device 16. Furthermore, in Figure 1

[0049] The detector 2 includes a detector provided for the purpose of detecting a sensor value necessary for operation control of the film forming machine 1, and a detector provided for the purpose of detecting a sensor value necessary for implementation of inference of a molding condition parameter of the film forming machine 1 based on the present inference method. The detector 2 can be a configuration separate from the film forming machine 1, or can be a configuration embedded in the film forming machine 1.

[0050] As the detector 2, for example, an acceleration sensor, a speed sensor, a displacement sensor, an AE (Acoustic Emission) sensor, a viscometer, a gel counter, a deformation sensor, a crystal orientation sensor, an anemometer, an air flow meter, a thermometer, a surface potential sensor, a laser sensor, an infrared sensor, an X-ray sensor, a position sensor, a tension sensor, a load cell, a torque sensor, an encoder, a current sensor, a tachometer, an angle sensor, a manometer, a gravimeter, a length measuring sensor, a photoelectric sensor (camera), a flowmeter, a voltmeter, a timer, a power meter, and the like are exemplified.

[0051] As the detection data detected by the detector 2, for example, acceleration, speed, displacement, viscosity, gel amount, deformation, degree of crystal orientation, wind speed, air volume, temperature, amount of electrification, position, phase difference, tension, torque, rotational speed, angle, pressure, weight, length, thickness, image, flow rate, current, voltage, time, force, power consumption, and the like are exemplified. In the present specification, the detection data detected by the detector 2 is not limited to a sensor value directly detected by the detector 2, and the meaning also includes a calculated value calculated indirectly from the sensor value.

[0052] ​As the detection data of the extruder 11 and the detector 2 of the detection data, for example, an acceleration sensor that detects acceleration (vibration) of a decelerator, a tachometer that detects a rotational speed of the screw 113 or a motor, a rotational speed meter that detects a rotational speed of the screw 113 or a motor, a torque sensor that detects a torque of a motor, a current sensor that detects a current of a motor, a viscometer that detects a viscosity of a resin (particularly, a viscosity of a resin immediately after discharge of the die 114), a contact type thermometer or a non-contact type thermal imaging camera that detects a temperature in the cylinder 112 or a resin (particularly, a temperature of a resin immediately after discharge of the die 114), a laser displacement sensor that detects a die lip opening of the die 114, and a weight meter that detects a feeding amount of a feeder, and the like are exemplified.

[0053] As the detection data of the extruder 11 and the detector 2 of the detection data, for example, an acceleration sensor that detects acceleration (vibration) of a decelerator, a tachometer that detects a rotational speed of the screw 113 or a motor, a rotational speed meter that detects a rotational speed of the screw 113 or a motor, a torque sensor that detects a torque of a motor, a current sensor that detects a current of a motor, a viscometer that detects a viscosity of a resin (particularly, a viscosity of a resin immediately after discharge of the die 114), a contact type thermometer or a non-contact type thermal imaging camera that detects a temperature in the cylinder 112 or a resin (particularly, a temperature of a resin immediately after discharge of the die 114), a laser displacement sensor that detects a die lip opening of the die 114, and a weight meter that detects a feeding amount of a feeder, and the like are exemplified.

[0054] As the detection data of the extruder 11 and the detector 2 of the detection data, for example, an acceleration sensor that detects acceleration (vibration) of a decelerator, a tachometer that detects a rotational speed of the screw 113 or a motor, a rotational speed meter that detects a rotational speed of the screw 113 or a motor, a torque sensor that detects a torque of a motor, a current sensor that detects a current of a motor, a viscometer that detects a viscosity of a resin (particularly, a viscosity of a resin immediately after discharge of the die 114), a contact type thermometer or a non-contact type thermal imaging camera that detects a temperature in the cylinder 112 or a resin (particularly, a temperature of a resin immediately after discharge of the die 114), a laser displacement sensor that detects a die lip opening of the die 114, and a weight meter that detects a feeding amount of a feeder, and the like are exemplified.

[0055] As the detection data of the TD stretching device 14 and the detector 2 of the detection data, for example, an acceleration sensor that detects acceleration (vibration) of the TD stretching device 14 (particularly, the rails 142 near the start and end portions of the stretching zones in the oven 141), a velocity sensor that detects velocity (vibration), a displacement sensor that detects displacement (vibration), an AE sensor that detects elastic waves, a deformation sensor that detects deformation of the jaws 144, a laser sensor that detects a phase difference (deviation of the passing time of the reference position) of the left and right pair of jaws 144, a photoelectric sensor or a photoelectric sensor that detects the temperature in the oven 141 (particularly, for each zone), a thermometer or a thermal imaging camera that detects the pressure in the oven 141 (particularly, for each zone), a pressure gauge that detects the air volume in the inlets and outlets of the oven 141 (particularly, the air blowing nozzles of the hot air blowing devices near the inlets and outlets), an air volume gauge that detects the air speed in the inlets and outlets of the oven 141 (particularly, the air blowing nozzles of the hot air blowing devices near the inlets and outlets), a pressure gauge that detects the pressure in the inlets and outlets of the oven 141 (particularly, the air blowing nozzles of the hot air blowing devices near the inlets and outlets), a tachometer that detects the fan rotation speed in the hot air blowing device, a velocity sensor that detects the line speed of the film conveying line, a tension sensor that detects the tension of the film conveying line or the jaws 144, a crystal orientation sensor that detects the degree of crystalline orientation of the film (particularly, the film at the most downstream of the TD stretching device 14), a gel counter that detects the number of gels of the film (particularly, the film at the most downstream of the TD stretching device 14), a laser sensor that detects the width of the film, and a laser sensor that detects the thickness of the film, and the like are exemplified.

[0056] As the detection data of the winding machine 15 and the detector 2 of the detection data, for example, a surface potential sensor that detects the amount of electrification of the film is exemplified.

[0057] The detector 2 is provided at an appropriate place in the film forming machine 1 or the periphery of the film forming machine 1 in a manner to detect the state of the film and the components at the desired position in the film forming machine 1. The detector 2 preferably has heat resistance, water resistance, oil resistance, and the like according to the setting position. Furthermore, the detector 2 and the content of the detection data are not limited to the above-described examples. In addition, the above-described detection data is not limited to the data detected by the exemplified detector 2 in combination, and the detector 2 is not limited to the detector used to detect the exemplified detection data in combination.

[0058] The above-described forming condition parameters, operation data, and detection data are examples of parameters that have an influence on the forming of the film based on the film forming machine 1, and the forming condition parameters correspond to parameters that can be controlled by device settings.

[0059] <DATA COLLECTION DEVICE 3>

[0060] Figure 2 is a block diagram showing a configuration example of the data collection device 3. The data collection device 3 is a computer, and includes a control section 31, a storage section 32, a communication section 33, and a data input section 34. The storage section 32, the communication section 33, and the data input section 34 are connected to the control section 31. The data collection device 3 is, for example, a PLC (Programmable Logic Controller).

[0061] The control section 31 includes a CPU (Central Processing Unit), a multi-core CPU, an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), an internal storage device such as a ROM (Read Only Memory) and a RAM (Random Access Memory), and an I / O terminal. The control section 31 collects molding information including detection data and operation data by executing a control program stored in the storage section 32 as described later, and performs processing of transmission to the inference device 4. Note that each functional section of the data collection device 3 can be implemented in software, in hardware, or by a combination of these.

[0062] The storage section 32 includes a hard disk, an EEPROM (Electrically Erasable Programmable ROM), a flash memory, and a nonvolatile memory. The storage section 32 stores a control program for causing the computer to perform collection processing of the molding information.

[0063] The communication section 33 includes a communication device that enables communication via the networks N1 and N2. The communication section 33 is connected to the control device 16 via a network N2 such as a LAN (Local Area Network). The control section 31 can transmit and receive various information to and from the control device 16 via the communication section 33. In addition, the communication section 33 is connected to the inference device 4 on the cloud as the network N1 via a relay device not shown. The relay device is, for example, a router, a gateway, or the like. The control section 31 can transmit and receive various information to and from the inference device 4 via the relay device and the communication section 33.

[0064] The data input section 34 is an input interface for inputting a signal output from the detector 2. The detector 2 is connected to the data input section 34. The control section 31 acquires detection data output from the detector 2 at all times via the data input section 34.

[0065] The data collection device 3 and the control device 16 are not limited to separate devices, and can be constituted by a common device.

[0066] <Inference Device 4>

[0067] Figure 3 is a block diagram showing a configuration example of the inference device 4. The inference device 4 is provided with a control section 41, a storage section 42, and a communication section 43. The storage section 42 and the communication section 43 are connected to the control section 41. The inference device 4 can be constituted by a plurality of computers and perform distributed processing, can be realized by a plurality of virtual machines provided in one server, or can be realized using a cloud server.

[0068] The control section 41 has an arithmetic processing circuit such as a CPU, a multi-core CPU, an ASIC, an FPGA, an internal storage device such as a ROM, a RAM, and the like, and an I / O terminal. Each functional section of the inference device 4 can be realized in software, in hardware, or by a combination of these.

[0069] The storage section 42 is provided with a nonvolatile memory such as a hard disk, a flash memory, an SSD (Solid State Drive), and the like, for example. The storage section 42 can also be an external storage device connected to the inference device 4. The storage section 42 stores various computer programs and data referred to by the control section 41. The storage section 42 of the present embodiment stores a program 4P for causing a computer to execute processing related to inference of a molding condition parameter of the film molding machine 1, a molding information DB (Data Base) 421 and an inference model 422 required for execution of the program 4P. The molding information DB 421 is a database that holds molding information received from each data collection device 3. The inference model 422 is a learning model generated by machine learning. It is assumed that the inference model 422 utilizes a program module that is part of an artificial intelligence software.

[0070] The computer program (computer program product) including the program 4P can also be provided by a non-transitory recording medium 4A in which the computer program is recorded in a readable manner. The storage section 42 stores the computer program read from the recording medium 4A by a not-illustrated reading device. The recording medium 4A is, for example, a magnetic disk, an optical disk, a semiconductor memory, or the like. In addition, the computer program can also be downloaded from an external server connected via a communication network and stored in the storage section 42. The program 4P can also be a single computer program, or can be configured using a plurality of computer programs, and can be executed on a single computer or on a plurality of computers connected to each other via a communication network.

[0071] The communication section 43 has a communication device that enables communication via the network N1. The control section 41 can transmit and receive various information between the data collection device 3 and the terminal device 5 via the communication section 43.

[0072] <Shaping information DB 421>

[0073] Figure 4 is a diagram showing the content example of the information stored in the shaping information DB 421. In the shaping information DB 421, records in which information such as the machine ID, the date and time, the operation data, the detection data, and the like are associated with each other using the data ID as a key are stored. The machine ID indicates the machine identifier of the film shaping machine 1. The date and time indicates the year, month, day, and time at which the operation data or the detection data stored as a record is obtained. The operation data includes the various operation data described above. The detection data includes the various detection data described above. The inference device 4 stores the operation data and the detection data in the shaping information DB 421 in chronological order each time it receives them from the data collection device 3. The content of the shaping information DB 421 is updated in real time.

[0074] <Inference model 422>

[0075] Figure 5 is an explanatory diagram showing an outline of the inference model 422. The inference model 422 is a learning model that outputs the shaping condition parameter in the film shaping machine 1 using the shaping information including the detection data and the operation data of the film shaping machine 1 as input.

[0076] The inference model 422 is, for example, a model that learns the shaping condition parameter corresponding to the shaping information by deep reinforcement learning, and is a DQN (Deep Q Network). The inference model 422 is a neural network model that outputs the action value of each of a plurality of actions a when a state s is input.

[0077] The inference model 422 has an input layer into which the state s is input, an output layer that outputs the action value Q (s, a) of the action a in the input state s, and an intermediate layer that extracts the features of the state s.

[0078] The state s input to the inference model 422 includes detection data and operation data obtained from the film forming machine 1. The detection data and the operation data include the above-described various data. The detection data and the operation data to be input to the inference model 422 can also be a combination of appropriate detection data and operation data selected from among the above-described example detection data and operation data, respectively. The detection data preferably includes both detection data detected by the detector 2 for state inference and detection data detected by the detector 2 for operation control. For example, the detection data includes acceleration, pressure, air volume, cycle difference, deformation, rotation speed, rotational speed, temperature, pressure, torque, thickness of the film, and the like, and the operation data includes rotation speed, rotational speed, temperature, pressure, torque, and the like. The forming information can also include a forming condition parameter set to the film forming machine 1. The detection data and the operation data can also be time series data for a period from the start of the operation to a data acquisition time point or a predetermined number of detection times. The detection data and the operation data can also be input as image data representing the time series data.

[0079] As shown in FIG. 2, the state s can also include at least one of a target quality (for example, a target film thickness) required for the manufactured film and actual quality data (for example, a measured value of the film thickness) in the formed film. The target quality can be included in the operation data, and the target quality can be input as a part of the operation data. The quality data can be included in the detection data, and the quality data can be input as a part of the detection data. Figure 5

[0080] Further, the input information input to the inference model 422 can be the detection data and the operation data themselves, but can also be data subjected to a prescribed preprocessing based on the detection data and the operation data, or the like.

[0081] The output layer of the inference model 422 has nodes corresponding to a plurality of actions a, respectively, and outputs an action value Q(s, a) with respect to the corresponding action a from each node. The action a corresponds to a change value of the forming condition parameter. The change value can be a value of the changed forming condition parameter, or can be a change amount of the forming condition parameter. Hereinafter, the change value is set to the value of the forming condition parameter. An action a with a high value indicates an appropriate forming condition parameter to be set to the film forming machine 1. The output layer of the inference model 422 can also output an action value Q with respect to a plurality of forming condition parameters.

[0082] The inference device 4 selects a change value for which the action value is the largest, based on the action values output from the inference model 422, and thereby can derive a change value of the forming condition parameter corresponding to the forming information.

[0083] ​The inference device 4 generates the inference model 422 by using deep reinforcement learning using prescribed data. In the reinforcement learning, the molding information is set as "state", the change value is set as "action", the desired "reward" is calculated, and the value of the Q value or Q function (action value function) is learned.

[0084] The update formula of the action value function is expressed by the following formula (1).

[0085] Q(s, a)←Q(s, a) + a(R + γmaxQ(s_next, a_next) - Q(s, a))...(1)

[0086] Herein,

[0087] s: state

[0088] a: action selected in the state s

[0089] a: learning coefficient

[0090] R: reward obtained by the result of the action

[0091] γ: discount rate

[0092] maxQ(s_next, a_next): maximum value of the action value Q with respect to the action that can be taken in the next state

[0093] The inference device 4 optimizes the parameters of the inference model 422, for example, by the error backpropagation method, in such a manner that the TD error, that is, the second term on the right side of the above formula (1) approaches zero, whereby the inference model 422 is caused to perform machine learning. The parameters of the inference model 422 are updated in such a manner that the expected value of the reward approaches 0 with respect to the error of the current action evaluation.

[0094] The inference device 4 performs processing of learning the molding condition parameters using the above-described reinforcement learning. Specifically, the inference device 4 acquires the target quality of the film, the actual quality data, and the molding information based on the information stored in the molding information DB 421. The target quality is, for example, a value set in advance by an operator. The target quality can also be acquired as part of the operation data. The quality data is obtained by measuring the film thickness of the film molded in the operation state indicated by the molding information. The quality data can also be acquired as part of the detection data. The quality data is not limited to a measured value, and can be, for example, a predicted value obtained using a machine learning model or a prescribed simulation model that predicts the quality data corresponding to the molding information.

[0095] The inference device 4 sets the molding condition parameter before the change as the state s, sets the change value of the molding condition parameter as the action a, and performs reinforcement learning based on a reward (score) calculated from the target quality and the quality data. The reward is, for example, automatically calculated by the inference device 4 based on the target quality included in the operation data stored in the molding information DB 421 and the quality data included in the detection data. The inference device 4 calculates the reward in such a manner that the smaller the deviation between the film thickness as the target quality and the film thickness as the quality data, the larger the value. In the case where the deviation between the target quality and the quality data is large, the reward is assigned to zero or a negative reward. The calculation formula of the reward is not particularly limited.

[0096] Further, the target quality is not limited to the film thickness, and can be, for example, the degree of crystallization in the outlet of the TD stretching device 14, the film thickness distribution, the air permeability, the degree of crystallization in the casting device 12, the transparency, and the like. The target quality can include a plurality of properties.

[0097] The inference device 4 causes the inference model 422 to learn based on the molding information input to the inference model 422, the action value Q output when each state variable is input, and the calculated reward. The inference device 4 can also cause the inference model 422 to learn using the molding information acquired from a plurality of film molding machines 1. When the learning ends, the inference device 4 stores the definition information related to the learned inference model 422 to the storage section 42 as the learned inference model 422.

[0098] The inference model 422 is not limited to being generated and learned by the inference device 4. The inference model 422 can be a model learned by an external server and transmitted to the inference device 4, and stored in the storage section 42. The inference model 422 can also be generated by an external server and learned by the inference device 4.

[0099] The configuration of the inference model 422 is not limited, as long as it can infer the change value of the molding condition parameter with respect to the detection data and the operation data. The inference model 422 can also be configured to adjust the molding condition parameter using a graph in which the state quantity and the action value are associated. The inference model 422 can also be a model that uses training data in which the change value of the molding condition parameter that becomes a correct value is associated with the molding information, and learns by supervised learning. The inference model 422 is, for example, a model based on other learning algorithms such as RNN (Recurrent Neural Network), CNN (Convolution Neural Network), GNN (Graph Neural Network), Transformer, support vector machine, decision tree, logistic regression, random forest, and the like.

[0100] The inference model 422 can accurately infer the molding condition parameters by considering various characteristics by setting a variety of detection data and operation data as input elements. For example, it is considered that the deterioration state of the device components can be considered by inputting the mechanical vibration represented by acceleration. It is considered that the property change of the resin, particularly the feedback that contributes to the improvement of the feeder supply amount, the mixing condition in the extruder 11, can be considered by inputting the resin viscosity. It is considered that the quality abnormality caused by the presence or absence of unmixed parts of the resin raw material, the presence or absence of foreign matter, can be considered by inputting the gel amount. By inputting the clamp deformation, the deterioration state of the device can be considered, and it is considered that the feedback of the parameter that particularly contributes to the reduction of product defects can be improved. It is considered that the quality abnormality can be considered by inputting the crystalline orientation degree. By inputting the air speed or air volume in the oven 141, the environment inside the oven 141 can be considered, and it is considered that the feedback of the conditions inside the oven 141 can be improved. It is considered that various temperature environments can be considered by inputting the temperature of each detection site and the detection object. It is considered that the film winding failure, the quality abnormality around the winding machine 15 can be considered by inputting the film charge amount. It is considered that the device load, the sliding load due to deterioration, the quality change can be considered by inputting the synchronization error shown by the phase difference of the left and right clamps. It is considered that the rail wear state can be considered by inputting the wire tension. It is considered that the deterioration state of the drive section can be considered by inputting the motor torque, the rotation speed, the rotation speed of each detection site and the detection object.

[0101] Figure 6 is a flowchart showing an example of the processing order performed by the inference device 4. The control section 41 of the inference device 4 performs the following processing according to the program 4P stored in the storage section 42. The control section 41 starts the following processing at a predetermined or appropriate interval.

[0102] The control section 41 of the inference device 4 acquires molding information including detection data and operation data in continuous operation by the data collection device 3 as a function of the acquisition section, and stores the acquired molding information in the molding information DB 421 (step S11). The machine ID of the film molding machine 1 is associated with the molding information. The detection data can also include quality information, and the operation data can also include target quality. The control section 41 can also acquire a plurality of detection data and operation data together.

[0103] The control unit 41 also functions as an inference unit, thereby inputting molding information, including acquired detection data and operating data, into the inference model 422 (step S12). The input to the inference model 422 may also include target quality and quality data. The control unit 41 acquires molding condition parameters output from the inference model 422 (step S13). In detail, the control unit 41 selects the optimal action based on the output data from the inference model 422 and acquires the change values ​​of one or more molding condition parameters.

[0104] The control unit 41 sends information indicating the changed values ​​of the acquired molding condition parameters to the user's terminal device 5 corresponding to the machine ID of the film forming machine 1 (step S14). The control unit 41 then completes a series of processing steps.

[0105] The user confirms the changed values ​​of the molding condition parameters through the terminal device 5 and operates the control device 16 as needed to change the setting of the molding condition parameters. Upon receiving the changed molding condition parameter settings, the control device 16 outputs a control signal to the film forming machine 1 to control its operation according to the accepted changed molding condition parameters. The film forming machine 1 operates according to the control signal based on the changed molding condition parameters.

[0106] In step S14, the control unit 41 may also automatically send the change value of the obtained molding condition parameter or the adjustment instruction corresponding to the change value of the molding condition parameter to the control device 16 via the data collection device 3, etc. The control device 16 outputs the control signal corresponding to the change value of the molding condition parameter or the adjustment instruction obtained from the inference device 4 to the film forming machine 1.

[0107] The control unit 41 can also determine whether the changed value of the acquired molding condition parameter exceeds the pre-allowed range. If it is determined that the changed value of the molding condition parameter exceeds the pre-allowed range, the control unit 41 can also output an alarm to the terminal device 5 or the control device 16. If the changed value is large, it is possible that the molding condition parameter is inappropriate or that the film forming machine 1 is in an abnormal state. Based on the above configuration, the possibility of such an abnormality can be detected early.

[0108] The control unit 41 repeatedly performs the above-described processing during the continuous operation of the film forming machine 1. Based on the detection data and operating data detected during continuous operation, the forming condition parameters are inferred while the continuous operation continues, thereby enabling feedback control corresponding to the inferred results. Preferably, the control unit 41 performs the above-described processing in real time whenever new detection data and operating data are detected. By performing the processing at relatively short intervals, such as every 1 minute or every 5 minutes, more appropriate operation control can be achieved.

[0109] The above illustrates an example of the inference device 4 performing a series of processes, but the processing entity for each process is not limited. Each process in the flowchart can also be performed by, for example, the control device 16, the data collection device 3, or the terminal device 5. The inference model 422 can also be configured on the control device 16 or the data collection device 3, where the inference processing of molding condition parameters is performed.

[0110] The inference device 4 can also send detailed information related to molding condition parameters, inspection data, and operating data to the terminal device 5. It can also output detailed information based on inferences of changes in molding condition parameters, or upon request from the terminal device 5. For example, based on information stored as detailed information in the molding information DB421, the inference device 4 generates a trend curve representing at least one trend of molding condition parameters, inspection data, and operating data during a specified period.

[0111] The inference device 4 can also relearn the inference model 422 based on molding information obtained during continuous operation. The inference device 4 receives molding information, target quality, and quality data obtained from the data collection device 3 within a specified period. Based on the received molding information, target quality, and quality data, the inference device 4 relearns and updates the inference model 422. The data used for relearning may also include molding information obtained from multiple film forming machines 1.

[0112] The inference device 4 can also evaluate the accuracy of the inference model 422 using methods such as cross-validation, and update the inference model 422 if the evaluation indicators meet the specified benchmarks. The inference device 4 can also receive feedback from the service provider regarding whether the inference model 422 needs to be updated. Therefore, the inference accuracy of the molding condition parameters can be improved by using this system.

[0113] According to this embodiment, based on the diverse detection data obtained by the detector 2, which detects the state of the film forming machine 1, the forming condition parameters of the film forming machine 1 can be inferred with high accuracy. In existing film forming machines, the detection data from the detector 2 is mainly used for operation control, and the state of the film forming machine 1 during operation is not considered when adjusting the forming condition parameters. In this embodiment, by setting up the detector 2 for the purpose of maintaining the film forming machine 1, more detection data not previously available for the film forming machine 1 can be collected and applied to state analysis.

[0114] During the continuous operation of the film forming machine 1, detection data is collected, and forming condition parameters are inferred based on the collected data, thereby enabling appropriate feedback.

[0115] By using inference model 422, the forming condition parameters of film forming machine 1 can be easily and accurately inferred. By using multiple unrelated and independent detection data representing the state of film forming machine 1 as input information for inference model 422, the inference accuracy of the forming condition parameters can be improved.

[0116] Regarding the film forming machine 1, a single production line comprises numerous devices and components, all operating in interconnected manner. Therefore, even slight deterioration or minor defects in a portion of a device or component upstream and downstream of the film forming machine 1 can lead to a reduction in the quality of the molded product. If operation continues while a portion of the device remains deteriorated, the overall load on the device increases, accelerating overall deterioration. Furthermore, many parameters within the film forming machine 1 interact with each other. Changing one parameter to approach the target quality can alter other parameters, potentially worsening the overall quality. By using the inference model 422, considering diverse molding information based on the state of each device and component within the film forming machine 1, multiple parameters are simultaneously optimized. By appropriately adjusting parameters in response to the deterioration of the film forming machine 1, the operating load on the film forming machine 1 can be reduced.

[0117] The inference device 4 accumulates a large amount of data in the cloud and can provide the inference results of molding condition parameters based on the accumulated large amount of data to the terminal device 5. The user can clearly identify the molding condition parameters.

[0118] The inference device 4 can output the inference results of the molding condition parameters to the film forming machine 1 via the control device 16. It can automatically perform the process from inference to adjustment of the molding condition parameters, thus improving convenience.

[0119] (Second Implementation)

[0120] In the second embodiment, the configuration using the inference model 422 or feature quantity corresponding to the film forming machine 1 will be described. The following embodiments will primarily describe the differences from the first embodiment; the same reference numerals will be used for components common to the first embodiment, and detailed descriptions will be omitted.

[0121] The inference device 4 of the second embodiment stores multiple inference models 422 generated according to different algorithms in the storage unit 42. Before inferring the molding condition parameters corresponding to the molding information, the inference device 4 determines the inference model 422 used for inferring the molding condition parameters based on the type of film forming machine 1 used by the user and the molding information collected from that film forming machine 1. The determination of the inference model 422 can be based on the inference accuracy of each inference model 422. For example, the accuracy of each inference model 422 is evaluated by evaluation methods such as cross-validation based on the inference results of each inference model 422 for the molding information in the film forming machine 1. The inference model 422 is determined by selecting the inference model 422 with the highest inference accuracy from all the prepared inference models 422.

[0122] The inference device 4 can generate an inference model 422 that focuses on inferring the molding condition parameters of the film forming machine 1 used by the user by fine-tuning the pre-generated learned inference model 422.

[0123] The inference device 4 can also select, from a set of pre-set detection and operation data, the detection and operation data used for input information into the inference model 422, depending on the type of film forming machine 1 used by the user. The detection and operation data used for input information can be determined, for example, based on the contribution of the detection and operation data. For instance, using a set of pre-set detection and operation data, inference based on the inference model 422 can be performed, and the contribution of the input data corresponding to each detection and operation data can be calculated using SHAP (SHapley Additive exPlanation), LIME (Local Interpretable Model-Agnostic Explanations), etc. The inference device 4 can determine the detection and operation data used for input information, for example, by selecting detection and operation data with a calculated contribution value greater than or equal to a predetermined value, or a predetermined number of detection and operation data with contributions ranging from large to small.

[0124] The inference device 4 stores the machine ID of the determined inference model 422, the detection data used for input information and the type of operating data, the inference model 422, and the machine ID of the film forming machine 1 that applies the detection data and operating data in a corresponding association in the storage unit 42.

[0125] Figure 7 This is a flowchart illustrating an example of the processing sequence performed by the inference device 4 of the second embodiment.

[0126] The control unit 41 of the inference device 4 obtains molding information, which includes detection data and operation data during continuous operation, in association with the machine ID of the film forming machine 1, and stores the obtained molding information in the molding information DB421 (step S21).

[0127] The control unit 41 selects the inference model 422 associated with the acquired machine ID of the film forming machine 1 (step S22). In step S22, the control unit 41 determines the inference model 422 corresponding to the acquired machine ID based on the pre-stored correspondence between the film forming machine 1 and the applied inference model 422.

[0128] Subsequently, the control unit 41 performs the same processing as in the first embodiment, and infers the molding condition parameters using the selected inference model 422. When molding condition parameters are set for each type of film forming machine 1, the control unit 41 determines the molding condition parameters associated with the acquired machine ID of the film forming machine 1 based on the pre-stored correspondence between the film forming machine 1 and the applied molding condition parameters.

[0129] According to this embodiment, by using an inference model 422, detection data, and operating data corresponding to the type of film forming machine 1 owned by the user, the accuracy of inferring the forming condition parameters can be improved. The configuration of film forming machines 1 varies greatly depending on the specific device. By setting the type of inference model 422, the detection data used, and the operating data according to the type of film forming machine 1, the forming condition parameters can be inferred with high accuracy based on the state of the film forming machine 1.

[0130] It should be understood that the embodiments disclosed herein are illustrative in all respects and not limiting. The technical features described in the various embodiments can be combined with each other, and the scope of the invention includes all modifications within the scope of protection and scopes equivalent to the scope of protection. The timing shown in the various embodiments is not limited; the order of each process can be changed to execute it without contradiction, and multiple processes can also be executed in parallel. The processing entity of each process is not limited, and the processes of each device can also be performed by other devices without contradiction.

[0131] The items described in each embodiment can be combined with each other. Furthermore, the independent items and dependent items described in the scope of protection can be combined with each other in all combinations, regardless of the form of reference. Moreover, the scope of protection uses a form that refers to two or more other items (multiple forms), but is not limited to this. It may also use a form that describes multiple dependent items that refer to at least one other item (multiple references to multiple items).

[0132] Regarding the above implementation methods, the following notes are also disclosed.

[0133] (Note 1)

[0134] A method for inferring molding condition parameters, wherein,

[0135] Acquire detection data related to the status of the film forming machine, as detected by the testing instrument.

[0136] By inputting the acquired detection data into the inference model, the molding condition parameters in the membrane forming machine are inferred. The inference model infers the molding condition parameters in the membrane forming machine when detection data related to the state of the membrane forming machine is input.

[0137] (Note 2)

[0138] The method for inferring molding condition parameters described in Appendix 1, wherein,

[0139] The detection data is acquired during the continuous operation of the film forming machine.

[0140] While continuing continuous operation, the molding condition parameters are inferred based on the detection data.

[0141] (Note 3)

[0142] The method for inferring the molding condition parameters described in Appendix 1 or Appendix 2, wherein,

[0143] Acquire quality data, which represents the quality of the film formed by the film forming machine.

[0144] By inputting the obtained detection data and quality data into the inference model, the molding condition parameters are inferred. The inference model infers the molding condition parameters in the film forming machine when the detection data and quality data of the film forming machine are input.

[0145] (Note 4)

[0146] The method for deducing the molding condition parameters as described in any one of Appendices 1 to 3, wherein,

[0147] The inferred molding condition parameters are sent to the film forming machine.

[0148] (Note 5)

[0149] The method for deducing the molding condition parameters as described in any one of Appendices 1 to 4, wherein,

[0150] The information representing the inferred molding condition parameters is sent to the user of the film forming machine.

[0151] (Note 6)

[0152] The method for deducing the molding condition parameters as described in any one of Appendices 1 to 5, wherein,

[0153] The film forming machine includes an extruder, a casting device, an MD stretching device, a TD stretching device, and a winding machine.

[0154] The molding condition parameters include at least one of the following: feed rate of the extruder, rotational speed of the extruder, rotational speed of the extruder, die lip opening of the extruder, rotational speed of the casting device, rotational speed of the casting device, rotational speed of the MD stretching device, rotational speed of the MD stretching device, linear speed of the TD stretching device, temperature of the TD stretching device, pressure of the TD stretching device, air volume of the TD stretching device, and air velocity of the TD stretching device.

[0155] (Note 7)

[0156] The method for deducing the molding condition parameters as described in any one of Appendices 1 to 6, wherein,

[0157] The film forming machine includes an extruder, a casting device, an MD stretching device, a TD stretching device, and a winding machine.

[0158] The detection data includes physical quantities related to the extruder, physical quantities related to the casting device, physical quantities related to the MD stretching device, physical quantities related to the TD stretching device, or physical quantities related to the winding machine.

[0159] (Note 8)

[0160] The method for deducing the molding condition parameters as described in any one of Appendices 1 to 7, wherein,

[0161] The film forming machine includes an extruder, a casting device, an MD stretching device, a TD stretching device, and a winding machine.

[0162] The detection data includes at least one of the following: acceleration in the extruder, resin viscosity in the extruder, acceleration in the casting device, amount of gel in the film in the casting device, acceleration in the MD stretching device, acceleration in the TD stretching device, wind speed in the TD stretching device, air volume in the TD stretching device, deformation of the film holder in the TD stretching device, phase difference between the left and right pairs of film holders in the TD stretching device, crystal orientation in the TD stretching device, amount of gel in the film in the TD stretching device, and charge on the film in the winding machine.

[0163] (Note 9)

[0164] The method for deriving the molding condition parameters as described in any one of Appendices 1 to 8, wherein,

[0165] The detection data selected from a variety of pre-set detection data will be input into the inference model.

[0166] (Postscript 10)

[0167] The method for deducing the molding condition parameters as described in any one of Appendices 1 to 9, wherein,

[0168] Prepare multiple inference models generated based on different algorithms.

[0169] The inference model used is selected based on the type of membrane forming machine.

[0170] Explanation of reference numerals in the attached figures

[0171] 100 Inference System

[0172] 1. Membrane forming machine

[0173] 11 Extruder

[0174] 12 Casting Equipment

[0175] 13MD stretching device

[0176] 14TD tensioning device

[0177] 15 winding machine

[0178] 16 control devices

[0179] 2 Detectors

[0180] 3 Data Collection Device

[0181] 31 Control Department

[0182] 32 Storage Unit

[0183] 33 Ministry of Communications

[0184] 34 Data Input Section

[0185] 4. Inference device

[0186] 41 Control Department

[0187] 42 Storage Unit

[0188] 43 Ministry of Communications

[0189] 4A recording media

[0190] 4P program

[0191] 421 Molding Information DB

[0192] 422 Inference Model

[0193] 5. Terminal devices.

Claims

1. An inference method of a molding condition parameter, wherein detection data relating to a state of a film molding machine is acquired by a detector, a molding condition parameter in the film molding machine is inferred by inputting the acquired detection data to an inference model that infers a molding condition parameter in a film molding machine when detection data relating to a state of a film molding machine is input.

2. The inference method of a molding condition parameter according to claim 1, wherein the detection data in a continuous operation of the film molding machine is acquired, the molding condition parameter is inferred based on the detection data in a state where the continuous operation is continued.

3. The inference method of a molding condition parameter according to claim 1 or 2, wherein quality data indicating a quality of a film molded by the film molding machine is acquired, the molding condition parameter is inferred by inputting the acquired detection data and the quality data to the inference model that infers a molding condition parameter in a film molding machine when detection data and quality data of a film molding machine are input.

4. The inference method of a molding condition parameter according to any one of claims 1 to 3, wherein the inferred molding condition parameter is transmitted to the film molding machine.

5. The inference method of a molding condition parameter according to any one of claims 1 to 4, wherein information indicating the inferred molding condition parameter is transmitted to a user of the film molding machine.

6. The inference method of a molding condition parameter according to any one of claims 1 to 5, wherein the film molding machine is provided with an extruder, a casting device, an MD stretching device, a TD stretching device, and a winding machine, the molding condition parameter includes at least one of a feeder supply amount in the extruder, a rotational speed in the extruder, a rotational velocity in the extruder, a die lip opening of a die in the extruder, a rotational speed in the casting device, a rotational velocity in the casting device, a rotational speed in the MD stretching device, a rotational velocity in the MD stretching device, a linear speed in the TD stretching device, a temperature in the TD stretching device, a pressure in the TD stretching device, an air volume in the TD stretching device, and an air speed in the TD stretching device.

7. The inference method of a molding condition parameter according to any one of claims 1 to 6, wherein the film molding machine is provided with an extruder, a casting device, an MD stretching device, a TD stretching device, and a winding machine, the detection data includes a physical quantity relating to the extruder, a physical quantity relating to the casting device, a physical quantity relating to the MD stretching device, a physical quantity relating to the TD stretching device, or a physical quantity relating to the winding machine.

8. The inference method of a molding condition parameter according to any one of claims 1 to 7, wherein the film molding machine is provided with an extruder, a casting device, an MD stretching device, a TD stretching device, and a winding machine, The detection data includes at least one of acceleration in the extruder, resin viscosity in the extruder, acceleration in the casting device, gel amount of a film in the casting device, acceleration in the MD stretching device, acceleration in the TD stretching device, air speed in the TD stretching device, air volume in the TD stretching device, deformation of a film holder in the TD stretching device, phase difference of a left and right pair of film holders in the TD stretching device, crystallization degree in the TD stretching device, gel amount of a film in the TD stretching device, and charge amount of a film in the winding machine.

9. The inference method of a molding condition parameter according to any one of claims 1 to 8, wherein The detection data selected from the plurality of detection data set in advance is input to the inference model.

10. The inference method of a molding condition parameter according to any one of claims 1 to 9, wherein A plurality of inference models generated according to different algorithms are prepared, The inference model used is selected according to the kind of the film molding machine.

11. A computer program for causing a computer to execute processing of: acquiring detection data related to a state of a film molding machine detected by a detector, inference of a molding condition parameter in the film molding machine by inputting the acquired detection data to an inference model that infers a molding condition parameter in a film molding machine when detection data related to a state of a film molding machine is input.

12. An inference device comprising: an acquisition unit that acquires detection data related to a state of a film molding machine detected by a detector; and an inference unit that infers a molding condition parameter in the film molding machine by inputting the acquired detection data to an inference model that infers a molding condition parameter in a film molding machine when detection data related to a state of a film molding machine is input.

13. An inference system having a film molding machine, a detector, an inference device, the inference device comprising: an acquisition unit that acquires detection data related to a state of a film molding machine detected by the detector; and an inference unit that infers a molding condition parameter in the film molding machine by inputting the acquired detection data to an inference model that infers a molding condition parameter in a film molding machine when detection data related to a state of a film molding machine is input.

Images