Classifications

• • G—PHYSICS
  • G05—CONTROLLING; REGULATING
  • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
  • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
  • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
  • G05B13/0265—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion
• • A—HUMAN NECESSITIES
  • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
  • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
  • A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
  • A61B1/00064—Constructional details of the endoscope body
  • A61B1/0011—Manufacturing of endoscope parts
• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
  • B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
  • B29C48/00—Extrusion moulding, i.e. expressing the moulding material through a die or nozzle which imparts the desired form; Apparatus therefor
  • B29C48/03—Extrusion moulding, i.e. expressing the moulding material through a die or nozzle which imparts the desired form; Apparatus therefor characterised by the shape of the extruded material at extrusion
  • B29C48/09—Articles with cross-sections having partially or fully enclosed cavities, e.g. pipes or channels
• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
  • B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
  • B29C48/00—Extrusion moulding, i.e. expressing the moulding material through a die or nozzle which imparts the desired form; Apparatus therefor
  • B29C48/15—Extrusion moulding, i.e. expressing the moulding material through a die or nozzle which imparts the desired form; Apparatus therefor incorporating preformed parts or layers, e.g. extrusion moulding around inserts
  • B29C48/151—Coating hollow articles
• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
  • B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
  • B29C48/00—Extrusion moulding, i.e. expressing the moulding material through a die or nozzle which imparts the desired form; Apparatus therefor
  • B29C48/15—Extrusion moulding, i.e. expressing the moulding material through a die or nozzle which imparts the desired form; Apparatus therefor incorporating preformed parts or layers, e.g. extrusion moulding around inserts
  • B29C48/157—Coating linked inserts, e.g. chains
• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
  • B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
  • B29C48/00—Extrusion moulding, i.e. expressing the moulding material through a die or nozzle which imparts the desired form; Apparatus therefor
  • B29C48/25—Component parts, details or accessories; Auxiliary operations
  • B29C48/92—Measuring, controlling or regulating
• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B65—CONVEYING; PACKING; STORING; HANDLING THIN OR FILAMENTARY MATERIAL
  • B65H—HANDLING THIN OR FILAMENTARY MATERIAL, e.g. SHEETS, WEBS, CABLES
  • B65H23/00—Registering, tensioning, smoothing or guiding webs
  • B65H23/04—Registering, tensioning, smoothing or guiding webs longitudinally
  • B65H23/18—Registering, tensioning, smoothing or guiding webs longitudinally by controlling or regulating the web-advancing mechanism, e.g. mechanism acting on the running web
  • B65H23/1806—Registering, tensioning, smoothing or guiding webs longitudinally by controlling or regulating the web-advancing mechanism, e.g. mechanism acting on the running web in reel-to-reel type web winding and unwinding mechanism, e.g. mechanism acting on web-roll spindle
• • C—CHEMISTRY; METALLURGY
  • C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
  • C12M—APPARATUS FOR ENZYMOLOGY OR MICROBIOLOGY; APPARATUS FOR CULTURING MICROORGANISMS FOR PRODUCING BIOMASS, FOR GROWING CELLS OR FOR OBTAINING FERMENTATION OR METABOLIC PRODUCTS, i.e. BIOREACTORS OR FERMENTERS
  • C12M41/00—Means for regulation, monitoring, measurement or control, e.g. flow regulation
  • C12M41/48—Automatic or computerized control
• • G—PHYSICS
  • G05—CONTROLLING; REGULATING
  • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
  • G05B19/00—Program-control systems
  • G05B19/02—Program-control systems electric
  • G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
  • G05B19/4184—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by fault tolerance, reliability of production system
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
  • G06N20/00—Machine learning
• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
  • B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
  • B29C2948/00—Indexing scheme relating to extrusion moulding
  • B29C2948/92—Measuring, controlling or regulating
  • B29C2948/92504—Controlled parameter
  • B29C2948/9258—Velocity
  • B29C2948/926—Flow or feed rate
• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
  • B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
  • B29C2948/00—Indexing scheme relating to extrusion moulding
  • B29C2948/92—Measuring, controlling or regulating
  • B29C2948/92819—Location or phase of control
  • B29C2948/92828—Raw material handling or dosing, e.g. active hopper or feeding device
• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
  • B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
  • B29C48/00—Extrusion moulding, i.e. expressing the moulding material through a die or nozzle which imparts the desired form; Apparatus therefor
  • B29C48/16—Articles comprising two or more components, e.g. co-extruded layers
  • B29C48/18—Articles comprising two or more components, e.g. co-extruded layers the components being layers
  • B29C48/21—Articles comprising two or more components, e.g. co-extruded layers the components being layers the layers being joined at their surfaces
• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
  • B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
  • B29C48/00—Extrusion moulding, i.e. expressing the moulding material through a die or nozzle which imparts the desired form; Apparatus therefor
  • B29C48/25—Component parts, details or accessories; Auxiliary operations
  • B29C48/30—Extrusion nozzles or dies
  • B29C48/32—Extrusion nozzles or dies with annular openings, e.g. for forming tubular articles
  • B29C48/34—Cross-head annular extrusion nozzles, i.e. for simultaneously receiving moulding material and the preform to be coated
• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B65—CONVEYING; PACKING; STORING; HANDLING THIN OR FILAMENTARY MATERIAL
  • B65H—HANDLING THIN OR FILAMENTARY MATERIAL, e.g. SHEETS, WEBS, CABLES
  • B65H2513/00—Dynamic entities; Timing aspects
  • B65H2513/10—Speed
  • B65H2513/11—Speed angular
• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B65—CONVEYING; PACKING; STORING; HANDLING THIN OR FILAMENTARY MATERIAL
  • B65H—HANDLING THIN OR FILAMENTARY MATERIAL, e.g. SHEETS, WEBS, CABLES
  • B65H2557/00—Means for control not provided for in groups B65H2551/00 - B65H2555/00
  • B65H2557/60—Details of processes or procedures
  • B65H2557/63—Optimisation, self-adjustment, self-learning processes or procedures, e.g. during start-up
• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B65—CONVEYING; PACKING; STORING; HANDLING THIN OR FILAMENTARY MATERIAL
  • B65H—HANDLING THIN OR FILAMENTARY MATERIAL, e.g. SHEETS, WEBS, CABLES
  • B65H2701/00—Handled material; Storage means
  • B65H2701/10—Handled articles or webs
  • B65H2701/17—Nature of material
  • B65H2701/175—Plastic
  • B65H2701/1752—Polymer film
• • C—CHEMISTRY; METALLURGY
  • C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
  • C12M—APPARATUS FOR ENZYMOLOGY OR MICROBIOLOGY; APPARATUS FOR CULTURING MICROORGANISMS FOR PRODUCING BIOMASS, FOR GROWING CELLS OR FOR OBTAINING FERMENTATION OR METABOLIC PRODUCTS, i.e. BIOREACTORS OR FERMENTERS
  • C12M1/00—Apparatus for enzymology or microbiology
• • C—CHEMISTRY; METALLURGY
  • C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
  • C12M—APPARATUS FOR ENZYMOLOGY OR MICROBIOLOGY; APPARATUS FOR CULTURING MICROORGANISMS FOR PRODUCING BIOMASS, FOR GROWING CELLS OR FOR OBTAINING FERMENTATION OR METABOLIC PRODUCTS, i.e. BIOREACTORS OR FERMENTERS
  • C12M3/00—Tissue, human, animal or plant cell, or virus culture apparatus
• • G—PHYSICS
  • G05—CONTROLLING; REGULATING
  • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
  • G05B19/00—Program-control systems
  • G05B19/02—Program-control systems electric
  • G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
  • G05B19/41875—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by quality surveillance of production
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
  • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
  • G06Q50/04—Manufacturing

Definitions

• the technology of the present disclosure relates to an information processing device, an operating method for the information processing device, and an operating program for the information processing device.
• a method can be considered that uses a machine learning model that outputs a predicted value of product quality according to the input of manufacturing conditions and solves an optimization problem to derive suitable manufacturing conditions where quality is the target value.
• One embodiment of the technology of the present disclosure is an information processing device, an operating method for the information processing device, and an operating program for the information processing device that can derive manufacturing conditions that are generally suitable regardless of the type of product. I will provide a.
• An information processing device of the present disclosure includes a processor, and the processor uses a machine learning model that outputs a predicted value of the quality of the divided portions according to input of manufacturing conditions of the divided portions obtained by dividing the entire product. Input the conditions to the machine learning model, output the predicted value of the divided part from the machine learning model, and determine if the value of the objective function that has a term including the difference between the predicted value of the divided part and the target value of the quality of the divided part is the minimum. By solving an optimization problem that determines manufacturing conditions, suitable manufacturing conditions that achieve the target quality of the entire product are derived.
• the objective function has a regularization term for fitting the manufacturing conditions of the divided parts into the constraint conditions regarding the manufacturing conditions of the entire product.
• the regularization term is preferably a term that smoothes the manufacturing conditions of the divided portions.
• the regularization term is preferably the sum of second-order differential values of the manufacturing conditions of the divided portions.
• the term including the difference is the sum of the values obtained by dividing the difference by the target value of the divided portion.
• At least one of the physical property information of the materials constituting the product and the product design information be input into the machine learning model.
• a machine learning model is prepared for each of the multiple types of predicted values.
• the product is a flexible tube for endoscopes that has a flexible tube base material and a resin layer that covers the flexible tube base material and is composed of an inner layer and an outer layer that have different thickness ratios in the axial direction.
• the divided portion is a portion in which the flexible tube for an endoscope is divided along the axial direction, and the manufacturing conditions are the amount of extrusion of the inner layer resin material per unit time by the extrusion molding machine, and the extrusion amount of the outer layer resin material by the extrusion molding machine.
• the predicted value includes the amount of extrusion of the resin material per unit time
• the predicted value includes the elasticity value of the divided portion, the thickness of the inner layer of the divided portion, and the thickness of the outer layer of the divided portion.
• the product is a sheet made by coating a liquid on a long support that is conveyed, and the divided portion is a portion where the sheet is divided along the width direction, and the manufacturing conditions are as follows:
• the predicted value includes a control amount of a tension adjustment roller disposed on the side and adjusts the tension applied to the support at the coating location, and the predicted value includes the coating thickness of the liquid at the divided portion.
• the product is cells cultured in a culture tank
• the divided part is a part where the culture tank is divided
• the manufacturing conditions include the cell culture conditions
• the predicted value represents the preference of the culture environment for the cells.
• an index value is included.
• the operating method of the information processing device of the present disclosure includes using a machine learning model that outputs a predicted value of the quality of the divided parts according to the input of the manufacturing conditions of the divided parts obtained by dividing the entire product;
• the machine learning model inputs the predicted value of the divided portion into a learning model, and outputs the predicted value of the divided portion from the machine learning model. This includes deriving suitable manufacturing conditions that achieve a target value for the quality of the entire product by solving an optimization problem that determines manufacturing conditions.
• the operating program of the information processing device of the present disclosure uses a machine learning model that outputs a predicted value of the quality of the divided parts according to the input of the manufacturing conditions of the divided parts obtained by dividing the entire product, and that the manufacturing conditions of the divided parts are
• the machine learning model inputs the predicted value of the divided portion into a learning model, and outputs the predicted value of the divided portion from the machine learning model.
• the computer is caused to perform a process that includes deriving suitable manufacturing conditions for achieving a target value for the quality of the entire product by solving an optimization problem for determining manufacturing conditions.
• an information processing device an operating method for the information processing device, and an operating program for the information processing device that can derive universally suitable manufacturing conditions regardless of the type of product. Can be done.
• FIG. 3 is a diagram showing divided portions treated as one unit for predicting quality.
• FIG. 2 is a block diagram showing a computer that constitutes an information processing device.
• FIG. 2 is a block diagram showing a processing unit of a CPU of the information processing device.
• FIG. 3 is a diagram showing input data.
• FIG. 3 is a diagram showing a target value group. It is a graph showing the formation of elasticity standard values.
• FIG. 3 is a diagram showing learning data. It is a figure which shows the process in the learning phase of an elasticity value prediction model. It is a figure which shows the process in the learning phase of an inner layer thickness prediction model. It is a figure which shows the process in the learning phase of an outer layer thickness prediction model.
• FIG. 3 is a diagram showing an objective function.
• FIG. 3 is a flowchart showing a processing procedure of the information processing device.
• FIG. 3 is a diagram showing how to derive suitable manufacturing conditions for a large diameter and long flexible tube for an endoscope.
• FIG. 3 is a diagram showing how to derive suitable manufacturing conditions for a flexible tube for an endoscope having a small diameter and short length.
• FIG. 7 is a diagram showing main parts of a sheet manufacturing apparatus according to a second embodiment. It is a figure which shows the division part of 2nd Embodiment. It is a figure which shows the input data of 2nd Embodiment. It is a figure showing a prediction model group of a 2nd embodiment.
• FIG. 7 is a diagram showing an objective function of a second embodiment. It is a figure which shows the process of the derivation
• FIG. 6 is a diagram showing how a more suitable oxygen supply amount is derived from the oxygen concentration under tentative suitable manufacturing conditions.
• the information processing device 10 outputs suitable manufacturing conditions 13 for the flexible tube 11 for an endoscope to an extrusion molding machine 12 that manufactures the flexible tube 11 for an endoscope.
• the preferred manufacturing conditions 13 are derived using a machine learning model as described later, and are manufacturing conditions in which the overall quality of the flexible tube 11 for an endoscope is a target value.
• “the quality meets the target value” includes not only cases where the quality completely matches the target value, but also cases where the quality falls within the range of the target value ⁇ ( ⁇ is the allowable error). .
• the information processing device 10 is, for example, a desktop personal computer, and has a display 14 that displays various screens, and an input device 15 such as a keyboard, a mouse, a touch panel, or a microphone for voice input.
• the information processing device 10 is operated, for example, by an operator of an extrusion molding machine 12.
• the information processing device 10 and the extrusion molding machine 12 are communicably connected through a communication network such as a LAN (Local Area Network).
• LAN Local Area Network
• the flexible endoscope tube 11 constitutes a part of the endoscope 16. More specifically, the endoscope 16 has an insertion section 17 that is inserted into a body cavity.
• the insertion portion 17 has a distal end portion 17A, a curved portion 17B, and a flexible portion 17C.
• the distal end portion 17A has a built-in image sensor for photographing inside the body cavity.
• the curved portion 17B is connected to the distal end portion 17A, and curves vertically and horizontally to change the direction of the distal end portion 17A.
• the flexible portion 17C is a soft elongated tubular portion that connects the curved portion 17B and the hand operation portion 18 provided with an operation knob for operating the curved portion 17B, and occupies most of the insertion portion 17.
• the flexible endoscope tube 11 constitutes this flexible portion 17C.
• the flexible tube 11 for an endoscope is an example of a "product" according to the technology of the present disclosure.
• the flexible tube 11 for an endoscope is formed by coating the outer peripheral surface of a flexible tube base material 20 with a resin layer 21.
• the flexible tube base material 20 has a configuration in which a spiral tube 22 is covered with a cylindrical mesh body 23.
• the spiral tube 22 is formed, for example, by spirally winding a metal strip made of stainless steel or the like.
• the cylindrical mesh body 23 is formed by, for example, braiding metal wires such as stainless steel fibers.
• a cap 24 is fitted to both ends of the flexible tube base material 20.
• the outer diameter of the flexible tube 11 for an endoscope is, for example, 10 mm to 14 mm, and the length is, for example, 50 cm to 150 cm.
• the resin layer 21 is continuously molded on the surface of the flexible tube base material 20 by the extrusion molding machine 12.
• the resin layer 21 has a two-layer structure in which an inner layer 25 covers the entire circumferential surface of the flexible tube base material 20 around the axis, and an outer layer 26 covers the entire circumferential surface of the inner layer 25 around the axis.
• Inner layer 25 is relatively soft and outer layer 26 is harder than inner layer 25.
• a resin containing polyurethane elastomer as a main component is used as the resin material for the inner layer 25 and the outer layer 26, for example.
• the combined thickness of the resin layer 21 including the inner layer 25 and the outer layer 26 is approximately the same in the axial direction AD of the flexible tube base material 20, and is, for example, 0.2 mm to 1.0 mm.
• the inner layer 25 and the outer layer 26 have different thickness ratios in the axial direction AD.
• the inner layer 25 is thicker than the outer layer 26 on the tip 27 side of the flexible tube base material 20 connected to the curved portion 17B.
• the outer layer 26 is thicker than the inner layer 25. From the distal end 27 to the proximal end 28, the thickness of the inner layer 25 gradually decreases and the thickness of the outer layer 26 gradually increases.
• the outer surface of the resin layer 21 is further coated with a top coat layer 29.
• the material for the top coat layer 29 may be any material as long as it does not affect the human body, has chemical resistance, and can withstand the high temperatures of steam sterilization; for example, fluorine paint is used.
• the thickness of the top coat layer 29 is, for example, 10 ⁇ m to 200 ⁇ m.
• a scale or the like indicating the length of the insertion portion 17 is printed on the surface of the top coat layer 29, a scale or the like indicating the length of the insertion portion 17 is printed.
• the extrusion molding machine 12 includes extrusion sections 35 and 36, a head section 37, a cooling section 38, a delivery drum 39, a winding drum 40, and a control section 41.
• the extrusion section 35 includes a hopper (not shown), a screw 42, and the like.
• the extrusion section 35 extrudes the resin material of the inner layer 25 in a molten state toward the head section 37 .
• the extrusion section 36 also includes a hopper (not shown), a screw 43, and the like.
• the extrusion section 36 extrudes the resin material of the outer layer 26 in a molten state toward the head section 37 .
• the head portion 37 is composed of a nipple 44, a die 45, and a support body 46 that fixedly supports the nipple 44 and the die 45.
• a circular shaped passage 47 is formed in the center of the nipple 44 .
• the entrance of the molding passage 47 is widened in a tapered shape.
• the flexible tube base material connection body 20C is inserted into the molding passage 47.
• the flexible tube base material connected body 20C is a structure in which a plurality of flexible tube base materials 20 are connected by a connecting member 48.
• the flexible tube base material connected body 20C is introduced into the molding passage 47 with the proximal end 28 of the flexible tube base material 20 at the front and the distal end 27 at the rear.
• the hole diameter of the molding passage 47 is slightly larger than the outer diameter of the flexible tube base material 20. Note that, contrary to the above, the flexible tube base material connected body 20C may be introduced into the molding passage 47 with the distal end 27 of the flexible tube base material 20 at the front and the base end 28 at the rear.
• a resin passage 49 is formed between the nipple 44 and the die 45 so as to surround the entire periphery of the molding passage 47.
• the resin passage 49 communicates with gates 50 and 51 formed in the support 46.
• the gates 50 and 51 are also formed to surround the entire periphery of the molding passage 47.
• the resin material of the inner layer 25 and the resin material of the outer layer 26 are extruded to the gates 50 and 51, and pass through the resin passage 49 into the molding passage with the resin material of the inner layer 25 on the bottom and the resin material of the outer layer 26 on the top. 47.
• the cooling unit 38 is, for example, a water tank in which cooling water is stored.
• the cooling unit 38 cools the resin material of the inner layer 25 and the resin material of the outer layer 26 extruded into the flexible tube base material connection body 20C in the head unit 37.
• the flexible tube base material connected body 20C before extrusion molding is wound around the delivery drum 39.
• the flexible tube base material connected body 20C after extrusion molding is wound up on the winding drum 40.
• a motor (not shown) is connected to the delivery drum 39 and the take-up drum 40, and the delivery drum 39 and the take-up drum 40 rotate according to the drive of the motor. As the delivery drum 39 and the winding drum 40 rotate, the flexible tube base material connected body 20C is sent out from the delivery drum 39 toward the winding drum 40 (head section 37 and cooling section 38).
• the conveyance speed of the flexible tube base material connected body 20C by the delivery drum 39 and the winding drum 40 is constant.
• the resin layer 21 is formed on the plurality of flexible tube base materials 20. can be continuously molded.
• the preferred manufacturing conditions 13 from the information processing device 10 are input to the control unit 41 .
• the control unit 41 controls the operation of each part of the extrusion molding machine 12 according to suitable manufacturing conditions 13.
• the control unit 41 controls the amount of extrusion of the resin material of the inner layer 25 per unit time and the amount of extrusion of the resin material of the outer layer 26 per unit time by controlling the rotation speed of the screws 42 and 43. Further, the control unit 41 controls the driving of the motors of the delivery drum 39 and the winding drum 40 to maintain a constant conveyance speed of the flexible tube base material connected body 20C.
• the control unit 41 controls the molding temperature by controlling the drive of the heater of the die 45, and controls the temperature of the cooling water in the cooling unit 38 to be kept constant.
• the control unit 41 makes the outer layer 26 thicker than the inner layer 25 at the proximal end 28, and decreases the thickness of the outer layer 26 toward the distal end 27. Control is performed such that the thickness of the inner layer 25 is gradually decreased and the thickness of the inner layer 25 is gradually increased, and the thickness of the inner layer 25 is made thicker than the outer layer 26 at the tip 27.
• the control unit 41 controls the amount of resin material of the inner layer 25 per unit time at the connection member 48 in order to prepare for extrusion molding of the next flexible tube base material 20. The extrusion amount of the resin material of the outer layer 26 and the extrusion amount of the resin material of the outer layer 26 per unit time are switched.
• the flexible tube base material connected body 20C wound up on the winding drum 40 is separated into individual flexible tube base materials 20 by removing the connecting member 48. Then, each flexible tube base material 20 is coated with a top coat layer 29, respectively. Finally, the caps 24 are attached to the distal end 27 and proximal end 28 of each flexible tube base material 20. As a result, the flexible tube 11 for an endoscope is completed. The endoscope flexible tube 11 is moved to the endoscope 16 assembly process.
• the information processing device 10 includes divided parts P1, P2, P3, P4, P5, P6, which are parts obtained by dividing one flexible tube base material 20 along the axial direction AD.
• P7 and P8 (lined up in this order from the base end 28 side) and divided parts P9 and P10 (lined up in this order from the base end 28 side), which are parts obtained by dividing the connecting member 48 along the axial direction, are Treat as one unit to predict.
• the divided portion P9 treated as one unit for predicting quality is a portion of the connecting member 48N that connects the flexible tube base material 20 and the next flexible tube base material 20.
• the divided portion P10 treated as one unit for predicting quality is a portion of the connecting member 48P that connects the flexible tube base material 20 concerned and the previous flexible tube base material 20.
• the lengths of the divided portions P1 to P10 may be the same or different.
• divided parts P when there is no particular need to distinguish between the divided parts P1 to P10, they may be referred to as divided parts P.
• the computer configuring the information processing apparatus 10 includes a storage 55, a memory 56, a CPU (Central Processing Unit) 57, and a communication unit 58. We are prepared. These are interconnected via a bus line 59.
• the storage 55 is a hard disk drive built into the computer that constitutes the information processing device 10 or connected through a cable or network.
• the storage 55 is a disk array in which a plurality of hard disk drives are connected in series.
• the storage 55 stores control programs such as an operating system, various application programs, and various data accompanying these programs. Note that a solid state drive may be used instead of the hard disk drive.
• the memory 56 is a work memory for the CPU 57 to execute processing.
• the CPU 57 loads the program stored in the storage 55 into the memory 56 and executes processing according to the program. Thereby, the CPU 57 centrally controls each part of the computer.
• the CPU 57 is an example of a "processor" according to the technology of the present disclosure. Note that the memory 56 may be built in the CPU 57.
• the communication unit 58 controls transmission of various information with external devices such as the extrusion molding machine 12.
• an operating program 65 is stored in the storage 55 of the information processing device 10.
• the operating program 65 is an application program for causing the computer to function as the information processing device 10. That is, the operation program 65 is an example of the "operation program for the information processing device" according to the technology of the present disclosure.
• the storage 55 also stores a predictive model group 66, an objective function 67, and the like.
• the CPU 57 of the computer constituting the information processing device 10 works with the memory 56 and the like to control the reception unit 70 and the read/write (hereinafter abbreviated as RW) control unit 71. , a prediction unit 72, a derivation unit 73, and a distribution control unit 74.
• RW read/write
• the reception unit 70 receives various information input by the operator via the input device 15. For example, the receiving unit 70 receives input data 75.
• the input data 75 is data input to the elasticity value prediction model 90, the inner layer thickness prediction model 91, and the outer layer thickness prediction model 92 (see FIG. 9) that constitute the prediction model group 66.
• the reception unit 70 outputs input data 75 to the RW control unit 71.
• the receiving unit 70 receives a target value group 76.
• the target value group 76 is a set of quality target values for each divided portion P.
• the reception unit 70 outputs the target value group 76 to the RW control unit 71.
• the reception unit 70 also receives instructions for deriving the preferred manufacturing conditions 13, instructions for distributing the preferred manufacturing conditions 13, and the like.
• the RW control unit 71 controls storage of various information in the storage 55 and reading of various information in the storage 55.
• the RW control unit 71 stores the input data 75 from the reception unit 70 and the target value group 76 in the storage 55. Further, the RW control unit 71 reads input data 75 from the storage 55 and outputs the read input data 75 to the prediction unit 72.
• the RW control unit 71 reads the predictive model group 66 from the storage 55 and outputs the read predictive model group 66 to the prediction unit 72. Further, the RW control unit 71 reads the objective function 67 and the target value group 76 from the storage 55 and outputs the read objective function 67 and target value group 76 to the derivation unit 73.
• the prediction unit 72 operates in response to an instruction to derive the suitable manufacturing conditions 13.
• the prediction unit 72 inputs the input data 75 to the elasticity value prediction model 90, the inner layer thickness prediction model 91, and the outer layer thickness prediction model 92 that constitute the prediction model group 66. Then, predicted values of the quality of the divided portion P are output from the elasticity value prediction model 90, the inner layer thickness prediction model 91, and the outer layer thickness prediction model 92.
• the prediction unit 72 outputs a predicted value group 77, which is a set of predicted values output from the elasticity value prediction model 90, the inner layer thickness prediction model 91, and the outer layer thickness prediction model 92, to the derivation unit 73.
• the derivation unit 73 also operates in response to an instruction to derive the suitable manufacturing conditions 13.
• the objective function 67 includes the target value of the target value group 76 and the predicted value of the predicted value group 77 as variables.
• the derivation unit 73 solves an optimization problem for finding manufacturing conditions that minimize the value of the objective function 67 by substituting the target value of the target value group 76 and the predicted value of the predicted value group 77 into the objective function 67.
• suitable manufacturing conditions 13 are derived.
• the derivation unit 73 outputs the suitable manufacturing conditions 13 to the distribution control unit 74.
• black box optimization methods such as genetic algorithms, evolutionary strategies, and Bayesian optimization can be used.
• the distribution control unit 74 controls distribution of the preferred manufacturing conditions 13 to the extrusion molding machine 12 specified by the distribution instruction of the preferred manufacturing conditions 13.
• the input data 75 includes manufacturing conditions for each of the divided parts P1 to P10, which are treated as one unit for predicting the quality shown in FIG. More specifically, the input data 75 includes manufacturing conditions 80_P1 for the divided portion P1, manufacturing conditions 80_P2 for the divided portion P2, . . . , manufacturing conditions 80_P9 for the divided portion P9, and manufacturing conditions 80_P10 for the divided portion P10.
• the manufacturing conditions 80_P1 and the like are expressed as "manufacturing conditions (P1)" and the like. The same applies to the elasticity predicted value 95 etc. that will appear later.
• the manufacturing conditions 80_P1 to 80_P10 for each of the divided portions P1 to P10 may be referred to as manufacturing conditions 80 if there is no particular need to distinguish them.
• the manufacturing conditions 80 include the amount of extrusion per unit time of the resin material of the inner layer 25 in the divided portion P (rotation speed of the screw 42), the amount of extrusion per unit time of the resin material of the outer layer 26 in the divided portion P (rotation of the screw 43). number), and the conveyance speed of the flexible tube base material connected body 20C (the number of rotations of the delivery drum 39 and the winding drum 40).
• the amount of extrusion of the resin material of the inner layer 25 per unit time and the amount of extrusion of the resin material of the outer layer 26 per unit time differ for each manufacturing condition 80 depending on the thickness of the inner layer 25 and the outer layer 26 of each divided portion P.
• the conveyance speed of the flexible tube base material connected body 20C is common to each manufacturing condition 80.
• the operator inputs manufacturing conditions 80 that are considered to bring the quality of the entire flexible tube 11 for an endoscope to a target value, based on past manufacturing results of the flexible tube 11 for an endoscope and his own experience.
• the manufacturing conditions 80 may include extrusion pressures for the resin material of the inner layer 25 and the resin material of the outer layer 26, molding temperature, etc., as appropriate.
• the extrusion pressure and molding temperature are also common to each manufacturing condition 80, as is the conveyance speed of the flexible tube base material connected body 20C.
• the input data 75 has physical property information 81.
• the physical property information 81 is information regarding the physical properties of the materials constituting the flexible tube 11 for an endoscope, in this case, the resin material of the inner layer 25 and the resin material of the outer layer 26.
• the physical property information 81 includes the hardness (for example, Shore A hardness) of the resin material of the inner layer 25, the tensile strength (for example, 100% modulus) of the resin material of the inner layer 25, the elasticity value (for example, flexural modulus) of the resin material of the outer layer 26, and the viscosity value (for example, melt viscosity) of the resin material of the outer layer 26.
• the physical property information 81 may include the tensile strength of the resin material of the inner layer 25, the elongation rate of the resin material of the inner layer 25, the viscosity ratio of the resin material of the outer layer 26, etc. as appropriate.
• the input data 75 has design information 82.
• the design information 82 is information regarding design values of the flexible tube 11 for an endoscope.
• the design information 82 includes the outer diameter of the flexible tube base material 20, the length of each divided portion P, and the like. The length of each divided portion P may be replaced by the distance between two adjacent divided portions P.
• the design information 82 may include, as appropriate, the amount of additives to the resin material of the inner layer 25 and the resin material of the outer layer 26.
• the target value group 76 includes standard values (hereinafter referred to as elastic standard values) of the elasticity values of the resin layer 21 in each of the divided portions P1 to P8. More specifically, the target value group 76 includes the elasticity standard value 85_P1 of the divided portion P1, the elasticity standard value 85_P2 of the divided portion P2, ..., the elasticity standard value 85_P7 of the divided portion P7, and the elasticity standard value 85_P8 of the divided portion P8. has.
• the elasticity standard values 85_P1 to 85_P8 of the divided portions P1 to P8 may be referred to as elasticity standard values 85 if there is no particular need to distinguish them.
• the elasticity standard value 85 is an example of a "target value" according to the technology of the present disclosure.
• the target value group 76 has a set value 86 for the total thickness of the resin layer 21 that is the sum of the thicknesses of the inner layer 25 and the outer layer 26 (hereinafter referred to as total thickness set value). Since the total thickness setting value 86 is common to each divided portion P, there is only one total thickness setting value 86. Like the elasticity standard value 85, the total thickness setting value 86 is also an example of a "target value" according to the technology of the present disclosure.
• the elasticity standard value 85 is, for example, the median of the upper and lower elasticity values of the resin layer 21 of each divided portion P.
• the upper and lower limits of the elasticity value of the resin layer 21 of each divided portion P, and furthermore, the elastic standard value 85 of the resin layer 21 of each divided portion P are from the divided portion P1 on the proximal end 28 side to the divided portion P8 on the distal end 27 side. It decreases as the This is because the distal end 27 side is relatively soft because the inner layer 25 is thicker than the outer layer 26, and the proximal end 28 side is relatively hard because the outer layer 26 is thicker than the inner layer 25.
• the predictive model group 66 includes an elasticity value predictive model 90, an inner layer thickness predictive model 91, and an outer layer thickness predictive model 92.
• the elasticity value prediction model 90, the inner layer thickness prediction model 91, and the outer layer thickness prediction model 92 are machine learning models configured by, for example, a neural network. That is, the elasticity value prediction model 90, the inner layer thickness prediction model 91, and the outer layer thickness prediction model 92 are examples of a "machine learning model" according to the technology of the present disclosure. Note that, hereinafter, the elasticity value prediction model 90, the inner layer thickness prediction model 91, and the outer layer thickness prediction model 92 may be collectively referred to as prediction models 90 to 92.
• the elasticity value prediction model 90 outputs a predicted value of the elasticity value (hereinafter referred to as an elasticity prediction value) of the resin layer 21 of the divided portion P according to input of manufacturing conditions 80, physical property information 81, and design information 82. do.
• the inner layer thickness prediction model 91 outputs a predicted value of the thickness of the inner layer 25 of the divided portion P (hereinafter referred to as inner layer thickness predicted value) according to the input of manufacturing conditions 80, physical property information 81, and design information 82. .
• the outer layer thickness prediction model 92 outputs a predicted value of the thickness of the outer layer 26 of the divided portion P (hereinafter referred to as an outer layer thickness predicted value) according to input of manufacturing conditions 80, physical property information 81, and design information 82. . In this way, there are multiple types of predicted values, and prediction models 90 to 92 are prepared for each of the multiple types of predicted values.
• the prediction unit 72 uses the manufacturing conditions 80_P1 of the divided portion P1 and the manufacturing conditions 80_P10 of each of the divided portions P10 and P2 on both sides of the divided portion P1. and 80_P2 are input to the elasticity value prediction model 90. Further, the prediction unit 72 inputs the physical property information 81 and the design information 82 to the elasticity value prediction model 90. Then, the prediction unit 72 causes the elasticity prediction model 90 to output the elasticity prediction value 95_P1 of the divided portion P1.
• the manufacturing conditions 80_P10 and 80_P2 are input into the elasticity value prediction model 90 in addition to the manufacturing condition 80_P1 because the elasticity values of the resin layers 21 of the divided parts P10 and P2 on both sides of the divided part P1 are the same as those of the resin layer 21 of the divided part P1. This is because it is thought to affect the elasticity value of.
• the prediction unit 72 uses manufacturing conditions 80A_P1 of the divided portion P1 excluding information on the outer layer 26, physical property information 81A excluding information on the outer layer 26, The design information 82 is input into the inner layer thickness prediction model 91. Then, the prediction unit 72 causes the inner layer thickness prediction model 91 to output the inner layer thickness prediction value 96_P1 of the divided portion P1.
• the manufacturing conditions 80A_P1 of the divided portion P1 excluding information on the outer layer 26 is information excluding the extrusion amount of the resin material of the outer layer 26 per unit time in the divided portion P1.
• the physical property information 81A excluding information on the outer layer 26 is information excluding the elasticity value of the resin material of the outer layer 26, the viscosity value of the resin material of the outer layer 26, etc.
• the reason for excluding the information on the outer layer 26 in this way is that the information on the outer layer 26 is considered unnecessary for predicting the inner layer thickness prediction value.
• the prediction unit 72 uses manufacturing conditions 80B_P1 of the divided portion P1 excluding information on the inner layer 25, physical property information 81B excluding information on the inner layer 25, The design information 82 is input into the outer layer thickness prediction model 92. Then, the prediction unit 72 causes the outer layer thickness prediction model 92 to output the outer layer thickness prediction value 97_P1 of the divided portion P1.
• the manufacturing conditions 80B_P1 of the divided portion P1 excluding information on the inner layer 25 is information excluding the extrusion amount of the resin material of the inner layer 25 per unit time in the divided portion P1.
• the physical property information 81B excluding information on the inner layer 25 is information excluding the hardness of the resin material of the inner layer 25, the tensile strength of the resin material of the inner layer 25, and the like.
• the reason for excluding the information on the inner layer 25 in this way is that, as in the case of the inner layer thickness prediction model 91, the information on the inner layer 25 is considered unnecessary for predicting the outer layer thickness prediction value.
• the prediction unit 72 adds the predicted inner layer thickness value 96_P1 and the predicted outer layer thickness value 97_P1 of the divided portion P1, and adds the predicted value of the total thickness of the resin layer 21 of the divided portion P1 (hereinafter referred to as 98_P1 (denoted as total thickness predicted value) is calculated.
• the predicted elasticity value 95_P1, the predicted inner layer thickness 96_P1, and the predicted outer layer thickness 97_P1 of the divided portion P1 are output from each prediction model 90 to 92, and the predicted total thickness 98_P1 of the divided portion P1 is calculated.
• the prediction unit 72 similarly calculates elasticity predicted values 95_P2 to 95_P8 (see FIG. 14), inner layer thickness predicted values 96_P2 to 96_P8 (not shown), and outer layer thickness for the remaining divided portions P2 to P8.
• the predicted values 97_P2 to 97_P8 are outputted from the respective prediction models 90 to 92, and the total thickness predicted values 98_P2 to 98_P8 (see FIG. 14) of the remaining divided portions P2 to P8 are determined.
• the predicted elasticity values 95_P1 to 95_P8 the predicted inner layer thickness values 96_P1 to 96_P8, the predicted outer layer thickness values 97_P1 to 97_P8, and the predicted total thickness values 98_P1 to 98_P8
• the predicted elasticity values 95 It may be expressed as an inner layer thickness predicted value 96, an outer layer thickness predicted value 97, and a total thickness predicted value 98.
• the elasticity predicted value 95, the inner layer thickness predicted value 96, the outer layer thickness predicted value 97, and the total thickness predicted value 98 are examples of "predicted values" according to the technology of the present disclosure.
• the predicted value group 77 has contents as shown in FIG. 14 as an example. That is, the predicted value group 77 includes an elasticity predicted value 95_P1 of the divided portion P1, an elasticity predicted value 95_P2 of the divided portion P2, . . . , an elasticity predicted value 95_P7 of the divided portion P7, and an elasticity predicted value 95_P8 of the divided portion P8. Further, the predicted value group 77 includes a total thickness predicted value 98_P1 of the divided portion P1, a total thickness predicted value 98_P2 of the divided portion P2, . . . , a total thickness predicted value 98_P7 of the divided portion P7, and a total thickness predicted value of the divided portion P8. It has the value 98_P8.
• the learning data 100 is data collected from previously manufactured flexible tubes 11 for endoscopes in order to be used for learning the prediction models 90 to 92.
• the learning data 100 includes learning input data 75L and correct answer data 101.
• the learning input data 75L is data corresponding to the input data 75.
• the learning input data 75L includes learning manufacturing conditions 80L, learning physical property information 81L, and learning design information 82L.
• the learning manufacturing conditions 80L is data corresponding to the manufacturing conditions 80
• the learning physical property information 81L is data corresponding to the physical property information 81
• the learning design information 82L is data corresponding to the design information 82.
• the learning data 100 includes the extrusion amount per unit time of the resin material of various inner layers 25, the extrusion amount per unit time of the resin material of the outer layer 26, and the flexible tube base material connection. This is a set of data regarding the divided portions P manufactured at the transport speed of the body 20C. Further, as can be seen from the learning physical property information 81L, the learning data 100 is also a collection of data regarding the divided portions P using the inner layer 25 and the outer layer 26 with various physical properties.
• the learning data 100 includes various design values such as the outer diameter of the flexible tube base material 20 of 10 mm, 5 mm, and 8 mm, and the length of the divided portion P of 100 mm and 50 mm. It is also a set of data regarding the divided portion P.
• correct answer data 101 is registered corresponding to each learning input data 75L.
• the correct answer data 101 is data for matching the answers with the outputs from each of the prediction models 90 to 92, that is, the elasticity prediction value 95, the inner layer thickness prediction value 96, and the outer layer thickness prediction value 97.
• Correct data 101 is the actual measured elasticity value (hereinafter referred to as actual elasticity value) of each divided portion P of the flexible tube 11 for endoscopes manufactured in the past, 95CA, for endoscopes manufactured in the past.
• Actual measurement value of the thickness of the inner layer 25 of each divided portion P of the flexible tube 11 hereinafter referred to as actual inner layer thickness value
• the measured value of the thickness of the outer layer 26 (hereinafter referred to as the measured outer layer thickness value) is 97CA.
• learning input data 75L is provided for learning.
• learning manufacturing conditions 80L are input to the elastic value prediction model 90.
• the learning manufacturing conditions 80L are the learning manufacturing conditions 80L of the divided portion P for predicting the elasticity prediction value 95, here the learning manufacturing conditions 80L_P1 of the divided portion P1.
• the learning manufacturing conditions 80L are the learning manufacturing conditions 80L for the divided portions P on both sides of the divided portion P for which the predicted elasticity value 95 is predicted, here, the learning manufacturing conditions 80L_P10 for the divided portion P10 and the learning manufacturing conditions for the divided portion P2.
• the manufacturing conditions are 80L_P2.
• the elasticity value prediction model 90 outputs a learning elasticity predicted value 95L, here a learning elasticity predicted value 95L_P1 of the divided portion P1, in response to the learning input data 75L.
• the learning elasticity predicted value 95L here the learning elasticity predicted value 95L_P1 of the divided portion P1
• the measured elasticity value 95CA corresponding to the learning input data 75L, here the divided portion P1.
• a loss calculation of the elasticity value prediction model 90 using a loss function is performed.
• various coefficients (coefficients of convolutional layer filters, etc.) of the elasticity value prediction model 90 are updated in accordance with the results of the loss calculation, and the elasticity value prediction model 90 is updated in accordance with the update settings.
• learning manufacturing conditions 80LA excluding information on the outer layer 26, learning physical property information 81LA excluding information on the outer layer 26, and learning design information 82L are applied to the inner layer. It is input into the thickness prediction model 91.
• the inner layer thickness prediction model 91 outputs a learning inner layer thickness predicted value 96L in response to the learning input data 75L.
• the loss of the inner layer thickness prediction model 91 using a loss function is calculated based on the learning inner layer thickness prediction value 96L and the inner layer thickness actual measurement value 96CA corresponding to the learning input data 75L. An operation is performed. Then, update settings are made for various coefficients of the inner layer thickness prediction model 91 according to the results of the loss calculation, and the inner layer thickness prediction model 91 is updated according to the update settings.
• the learning manufacturing conditions 80LB excluding information on the inner layer 25, the learning physical property information 81LB excluding information on the inner layer 25, and the learning design information 82L on the outer layer It is input to the thickness prediction model 92.
• the outer layer thickness prediction model 92 outputs a learning outer layer thickness predicted value 97L in response to the learning input data 75L.
• the loss of the outer layer thickness prediction model 92 using a loss function is calculated based on the learning outer layer thickness prediction value 97L and the outer layer thickness actual measurement value 97CA corresponding to the learning input data 75L. An operation is performed. Then, update settings are made for various coefficients of the outer layer thickness prediction model 92 according to the results of the loss calculation, and the outer layer thickness prediction model 92 is updated according to the update settings.
• each prediction model 90 to 92 input of learning input data 75L, output of learning elasticity predicted value 95L, learning inner layer thickness prediction value 96L, and learning outer layer thickness prediction value 97L, loss calculation, and update.
• the above-mentioned series of settings and updating processes are repeatedly performed while the learning input data 75L and the correct answer data 101 are exchanged.
• the above series of processes is repeated when the prediction accuracy of the learning elasticity prediction value 95L, the learning inner layer thickness prediction value 96L, and the learning outer layer thickness prediction value 97L reaches a predetermined setting level. .
• the prediction models 90 to 92 whose prediction accuracy has reached the set level in this way are stored in the storage 55 of the information processing device 10.
• learning may be terminated when the above series of processes is repeated a set number of times. . Furthermore, learning may be continued even after being stored in the storage 55.
• the objective function 67 has a first term 105, a second term 106, a third term 107, and a fourth term 108, which are respectively added.
• the first term 105 includes the difference between the predicted elasticity value 95 of each divided portion P and the elasticity standard value 85 of each divided portion P. More specifically, the first term 105 includes the sum of the squares of the difference between the predicted elasticity value 95 and the elasticity standard value 85 divided by the elasticity standard value 85.
• the first term 105 is a term obtained by multiplying the sum by the first weighting coefficient W1.
• the first term 105 is an example of a "term including a difference" according to the technology of the present disclosure.
• the second term 106 includes the difference between the predicted total thickness 98 and the set total thickness 86 of each divided portion P. More specifically, the second term 106 includes the sum of the squares of the difference between the total thickness predicted value 98 and the total thickness set value 86 divided by the total thickness set value 86.
• the second term 106 is a term obtained by multiplying the sum by the second weighting coefficient W2. Like the first term 105, the second term 106 is also an example of a "term including a difference" according to the technology of the present disclosure.
• the third term 107 and the fourth term 108 are regularization terms for fitting the manufacturing conditions of each of the divided portions P into the constraint conditions regarding the manufacturing conditions of the entire flexible tube 11 for an endoscope.
• the constraints regarding the manufacturing conditions of the entire flexible tube 11 for an endoscope are as follows: In order to smoothly change the elasticity value, inner layer thickness, and outer layer thickness throughout the flexible tube 11 for an endoscope, each of the divided portions P is The aim is to smoothly connect the manufacturing conditions of Therefore, the third term 107 is a term that smoothes the extrusion amount R of the resin material of each inner layer 25 of the divided portion P per unit time. Further, the fourth term 108 is a term for smoothing the extrusion amount S of the resin material of each outer layer 26 of the divided portion P per unit time.
• the third term 107 includes the sum of second-order differential values of the extrusion amount R of the resin material of the inner layer 25 of each divided portion P per unit time.
• the third term 107 is a term obtained by multiplying the sum by the third weighting coefficient W3.
• the fourth term 108 includes the sum of second-order differential values of the extrusion amount S of the resin material of the outer layer 26 of each divided portion P per unit time.
• the fourth term 108 is a term obtained by multiplying the sum by the fourth weighting coefficient W4.
• in the third term 107 is an approximate expression of the second derivative of the extrusion amount R per unit time of the resin material of each inner layer 25 of the divided portion P.
• in the fourth term 108 is an approximate expression of the second derivative of the extrusion amount S of the resin material of each outer layer 26 of the divided portion P per unit time.
• N in the first term 105 to fourth term 108 is the number of divided parts P.
• the first to fourth weighting coefficients W1 to W4 do not necessarily have to have the same value. For example, if you want to emphasize the third term 107, that is, if you want to smooth out the extrusion amount R per unit time of the resin material of the inner layer 25 of each divided portion P, set the third weighting coefficient W3 to a higher value than the others. May be set.
• the preferred manufacturing conditions 13 derived by the deriving unit 73 by solving the optimization problem for determining the manufacturing conditions that minimize the value of the objective function 67 are for each of the divided parts P1 to P8.
• a suitable extrusion amount per unit time of the resin material of the inner layer 25 in a suitable extrusion amount per unit time of the resin material of the outer layer 26 in each of the divided portions P1 to P8, and flexibility in each of the divided portions P1 to P8.
• solving the optimization problem to find the manufacturing conditions that minimize the value of the objective function 67 means that the difference between the predicted elasticity value 95 of the first term 105 and the elastic standard value 85, and the This means searching for suitable manufacturing conditions 13 in which the difference between the predicted total thickness 98 and the set total thickness 86 is minimized.
• solving the optimization problem to find the manufacturing conditions that minimize the value of the objective function 67 means that the extrusion amount R of the resin material of the inner layer 25 of each divided portion P per unit time and the This means searching for suitable manufacturing conditions 13 in which the amount of extrusion per unit time of the resin material of each outer layer 26 is smoothed.
• an input screen (not shown) for input data 75 and target value group 76 is displayed on display 14 of information processing device 10.
• the operator inputs desired input data 75 and target value group 76 on the input screen.
• the input data 75 and the target value group 76 are accepted by the reception unit 70 (step ST100).
• Input data 75 and target value group 76 are output from reception section 70 to RW control section 71 .
• the input data 75 includes manufacturing conditions 80, physical property information 81, and design information 82 for each divided portion P.
• the target value group 76 includes an elasticity standard value 85 and a total thickness setting value 86 for each of the divided portions P. Note that the input screens for the input data 75 and the target value group 76 may be separate, and the input timings for the input data 75 and the target value group 76 may be different.
• the reception unit 70 When the reception unit 70 receives an instruction to derive the suitable manufacturing conditions 13, the input data 75 is read from the storage 55 by the RW control unit 71 and output to the prediction unit 72. Further, the target value group 76 is read out from the storage 55 by the RW control unit 71 and output to the derivation unit 73.
• the input data 75 is input to the elasticity value prediction model 90, and the elasticity prediction value 95 of each divided portion P is output from the elasticity value prediction model 90.
• the input data 75 is input to the inner layer thickness prediction model 91, and the inner layer thickness prediction value 96 of each divided portion P is output from the inner layer thickness prediction model 91.
• the input data 75 is input to the outer layer thickness prediction model 92, and the outer layer thickness prediction value 97 of each divided portion P is output from the outer layer thickness prediction model 92 ( Step ST110). Then, as shown in FIG.
• the prediction unit 72 adds the predicted inner layer thickness 96 and the predicted outer layer thickness 97 to obtain the predicted total thickness 98 of each divided portion P.
• the predicted elasticity value 95 and total thickness predicted value 98 thus obtained are output from the prediction unit 72 to the derivation unit 73 as a predicted value group 77, as shown in FIG.
• Step ST120 the optimization problem for finding the manufacturing conditions that minimize the value of the objective function 67 shown in FIG. (Step ST120).
• the preferred manufacturing conditions 13 are output from the derivation unit 73 to the distribution control unit 74.
• the distribution control unit 74 distributes the preferred manufacturing conditions 13 to the extrusion molding machine 12 specified in the distribution instruction (step ST130).
• the preferred manufacturing conditions 13 from the information processing device 10 are input to the control unit 41. Then, under the control of the control unit 41, the operation of each part of the extrusion molding machine 12 is controlled under suitable manufacturing conditions 13, and the flexible tube 11 for an endoscope is manufactured.
• the CPU 57 of the information processing device 10 functions as the prediction unit 72 and the derivation unit 73.
• the prediction unit 72 calculates an elasticity prediction value 95, an inner layer thickness prediction value 96, and an outer layer thickness prediction value 97 of the divided portion P according to input of manufacturing conditions 80 of the divided portion P obtained by dividing the entire flexible tube 11 for an endoscope.
• An elasticity value prediction model 90, an inner layer thickness prediction model 91, and an outer layer thickness prediction model 92 are used.
• the prediction unit 72 inputs the manufacturing conditions 80 of each of the divided portions P into the elasticity value prediction model 90, the inner layer thickness prediction model 91, and the outer layer thickness prediction model 92, and calculates the elasticity predicted value 95 and inner layer thickness of each of the divided portions P.
• a predicted value 96 and an outer layer thickness predicted value 97 are output from the elasticity value prediction model 90, the inner layer thickness prediction model 91, and the outer layer thickness prediction model 92.
• the derivation unit 73 generates a first term 105 including the difference between the predicted elasticity value 95 of each divided portion P and the elastic standard value 85 of each divided portion P, and the total thickness predicted value 98 of each divided portion P.
• elasticity prediction values 95_P1 to 95_P10 of each divided portion P1 to P10 are shown. etc. can be outputted from the elasticity value prediction model 90 or the like, and suitable manufacturing conditions 13 can be derived. Further, as an example, as shown in FIG. 23, for example, the elasticity predicted values 95_P1 to 95_P6 of each divided portion P1 to P6 are calculated for the narrow diameter and short flexible tube 11SS of a nasal endoscope for bronchus.
• the suitable manufacturing conditions 13 can also be derived by outputting the elasticity value prediction model 90 or the like.
• the learning data 100 of each of the prediction models 90 to 92 is not the data of the entire flexible tube 11 for an endoscope, but the data of each divided portion P. Therefore, a large amount of data can be easily collected.
• the larger the number of learning data 100 the more progress will be made in the learning of each prediction model 90 to 92, and the prediction accuracy of each prediction model 90 to 92 can also be improved. Therefore, the preferred manufacturing conditions 13 can be derived with higher accuracy than the case where the quality of the entire flexible tube 11 for an endoscope is predicted and the preferred manufacturing conditions 13 are derived based on the prediction result. .
• the objective function 67 has a third term 107 and a fourth term 108, which are regularization terms for fitting the manufacturing conditions of each divided portion P into the constraint conditions regarding the manufacturing conditions of the entire flexible tube 11 for an endoscope. . Therefore, the preferable manufacturing conditions 13 can be included in the constraint conditions regarding the manufacturing conditions of the entire flexible tube 11 for an endoscope.
• the third term 107 and the fourth term 108 which are regularization terms, are the extrusion amount R per unit time of the resin material of the inner layer 25 of each of the divided portions P, and the unit of the resin material of the outer layer 26 of each of the divided portions P.
• This is a term that smoothes the extrusion amount S per hour.
• the third term 107 is the sum of second-order differential values of the extrusion amount R per unit time of the resin material of the inner layer 25 of each divided portion P
• the fourth term 108 is the sum of second-order differential values of the extrusion amount R of the inner layer 25 of each divided portion P. is the sum of second-order differential values of the extrusion amount S of the resin material of the outer layer 26 per unit time. Therefore, the preferred manufacturing conditions 13 can be set as if they were set by a skilled operator.
• the first term 105 and the second term 106 which are terms including the difference, are the sum of the values obtained by dividing the difference by the elastic standard value 85 and the total thickness setting value 86, which are the target values of the divided portion P. If the term including the difference is the sum of the differences without dividing the difference by the target value of the divided portion P, the target value is 100, the predicted value is 95 and the difference is 5, and the target value is 10 and the predicted value is 5 and the difference is 5 are treated in the same way.
• the term including the difference as the sum of the values obtained by dividing the difference by the target value of the divided portion P, if the target value is 100, the predicted value is 95, and the difference is 5 (the difference is divided by the target value)
• suitable manufacturing conditions 13 can be derived with high accuracy.
• Physical property information 81 of the material constituting the flexible endoscope tube 11 and design information 82 of the flexible endoscope tube 11 are also input to each of the prediction models 90 to 92. Therefore, the prediction accuracy of the elasticity prediction value 95, the inner layer thickness prediction value 96, and the outer layer thickness prediction value 97 can be improved compared to the case where the physical property information 81 and the design information 82 are not input to each of the prediction models 90 to 92.
• the manufacturing conditions 80 of the divided portion P that is the target of prediction are input to the elasticity value prediction model 90, so that the prediction accuracy of the elasticity predicted value 95 is further improved. can be increased.
• manufacturing conditions 80A_P1, etc. excluding information on the outer layer 26 are input to the inner layer thickness prediction model 91
• manufacturing conditions 80B_P1, etc. excluding information on the inner layer 25 are input to the outer layer thickness prediction model 92, which is unnecessary for prediction. Since information that is considered to be excluded is removed, the prediction accuracy of the inner layer thickness prediction value 96 and the outer layer thickness prediction value 97 can also be further improved.
• the physical property information 81 and the design information 82 are input to the elasticity value prediction model 90 etc. or one.
• the machine learning models include an elasticity value prediction model 90, an inner layer thickness prediction model 91, and an outer layer thickness prediction model 92. , are prepared for each of multiple types of predicted values. Therefore, the elasticity predicted value 95, the inner layer thickness predicted value 96, and the outer layer thickness predicted value 97 can all be predicted with high accuracy.
• the product includes a flexible tube base material 20 and a resin layer 21 that covers the flexible tube base material 20 and is composed of an inner layer 25 and an outer layer 26 having different thickness ratios in the axial direction AD.
• This is a flexible tube 11 for an endoscope.
• the divided portion P is a portion obtained by dividing the flexible tube 11 for an endoscope along the axial direction AD.
• the manufacturing conditions 80 include the amount of the resin material of the inner layer 25 extruded per unit time and the amount of the resin material of the outer layer 26 extruded per unit time by the extrusion molding machine 12.
• the predicted values include the elasticity value of the divided portion P (elasticity predicted value 95), the thickness of the inner layer of the divided portion P (inner layer thickness predicted value 96), and the thickness of the outer layer of the divided portion P (outer layer thickness predicted value 97).
• the flexible tube 11 for an endoscope is a large diameter and long flexible tube 11LL for an oral endoscope for the upper gastrointestinal tract shown in FIG. 22, or a nasal tube for the bronchus shown in FIG.
• a flexible tube 11SS for an endoscope with a small diameter and short length.
• specifications are frequently changed, such as changing the resin materials of the inner layer 25 and the outer layer 26 and changing the outer diameter of the flexible tube base material 20. Therefore, it is possible to fully demonstrate the effect that manufacturing conditions 13 suitable for general use can be derived regardless of the type of flexible tube 11 for an endoscope.
• the control unit 41 of the extrusion molding machine 12 may play the role of the CPU 57 of the information processing device 10. That is, the control unit 41 of the extrusion molding machine 12 may output each of the predicted values 95 to 98 and derive the suitable manufacturing conditions 13. In this case, the extrusion molding machine 12 itself becomes an example of an "information processing device" according to the technology of the present disclosure.
• the flexible tube 11 for an endoscope is exemplified as a product, and an example of deriving suitable manufacturing conditions 13 for the extrusion molding machine 12 has been described, but the technology of the present disclosure is not limited to this.
• suitable manufacturing conditions 251 see FIG. 33
• the sheet manufacturing apparatus 200 see FIG. 24
• a sheet manufacturing apparatus 200 is an apparatus that manufactures a sheet 202 by applying a liquid to a conveyed elongated support 201.
• the sheet 202 is an example of a "product" according to the technology of the present disclosure.
• the support 201 is, for example, a base material for a magnetic tape.
• the liquid is the magnetic layer material and the sheet 202 is the magnetic tape.
• the support 201 is, for example, a base material of an optical film.
• the liquid is a photosensitive layer material and the sheet 202 is an optical film.
• the sheet manufacturing apparatus 200 includes a coating section 203, conveyance rollers 204 and 205, and a tension adjustment roller 206.
• the application unit 203 is a Giesser or the like for applying a liquid to the support 201.
• Conveyance rollers 204 and 205 are disposed at symmetrical positions on the upstream and downstream sides with the application section 203 in between, and convey the support body 201 and the sheet 202 by being rotated by a motor.
• Tension adjustment roller 206 is arranged upstream of application section 203 and conveyance roller 204.
• the tension adjustment roller 206 is tilted with respect to the width direction WD of the support 201 and the sheet 202 with the center C as the center of rotation, for example, so that the tension adjustment roller 206 applies the liquid to the support 201 at the location where the liquid is applied in the application section 203. Adjust the applied tension.
• the application location is a slit 207 having a narrow interval parallel to the width direction WD.
• the sheet manufacturing apparatus 200 also includes a delivery drum that sends out the roll-shaped support 201 toward the coating section 203, a winding drum that winds up the sheet 202 coated with the liquid from the coating section 203, and the like.
• the sheet manufacturing apparatus 200 includes a driven roller that is in contact with the support 201 or the sheet 202 and rotates following the conveyance of the support 201 or the sheet 202, and a driven roller that swings according to changes in the conveyance speed of the support 201. It also includes dancer rollers and the like that suppress fluctuations in the tension applied to the support body 201.
• the quality of each of divided portions P1, P2, . . . , P7, and P8 divided along the width direction WD of the sheet 202 is predicted.
• the input data 210 of the second embodiment includes first manufacturing conditions 211 and second manufacturing conditions 212.
• the first manufacturing condition 211 includes the control amount of the tension adjustment roller 206.
• the control amount of the tension adjustment roller 206 is the pushing amount of the tension adjustment roller 206 in the width direction WD.
• the second manufacturing conditions 212 include the liquid flow rate from the application section 203, the lip clearance, and the taper angle of the pocket section.
• the lip clearance is the gap between the slit 207, which is the application location, and the support body 201.
• the pocket portion is provided within the application portion 203 and communicates with the slit 207.
• the pocket portion is a portion for spreading the liquid in the width direction WD. Note that, as the first manufacturing condition 211, the amount of swinging of the dancer roller, etc. may be added as appropriate.
• the input data 210 includes first physical property information 213_P1 of the divided portion P1, first physical property information 213_P2, . . . of the divided portion P2, and first physical property information 213_P8 of the divided portion P8.
• the input data 75 has second physical property information 214.
• the first physical property information 213_P1 to 213_P8 of each of the divided portions P1 to P8 may be referred to as first physical property information 213 if there is no need to particularly distinguish between them.
• the first physical property information 213 and the second physical property information 214 are information regarding the physical properties of the material constituting the sheet 202, here the support 201 and the liquid.
• the first physical property information 213 includes the thickness, slack amount, and waviness amount of the support body 201 in the divided portion P.
• the thickness, amount of slack, and amount of waviness of the support body 201 are values measured for each divided portion P by a sensor (not shown) installed between the coating section 203 and the conveyance roller 204. Therefore, the first physical property information 213 changes moment by moment as the support 201 is transported. Therefore, in the second embodiment, the input data 210 is input each time the support 201 is conveyed a preset length.
• the preset length is, for example, 5 cm to 10 cm.
• the thickness, slack amount, and waviness amount of the support body 201 described above are representative values, for example, average values, in a preset length.
• the second physical property information 214 includes liquid viscosity (for example, elongational viscosity). Note that the first manufacturing conditions 211, second manufacturing conditions 212, first physical property information 213, and second physical property information 214 are not limited to the exemplified contents, and other elements may be added as appropriate.
• the predictive model group 220 of the second embodiment includes a tension predictive model 221, a hydraulic pressure predictive model 222, and a coating thickness predictive model 223.
• the tension prediction model 221, the hydraulic pressure prediction model 222, and the coating thickness prediction model 223 are machine learning models configured by, for example, a neural network. That is, the tension prediction model 221, the hydraulic pressure prediction model 222, and the coating thickness prediction model 223 are examples of a "machine learning model" according to the technology of the present disclosure.
• the tension prediction model 221 calculates a predicted value of tension (hereinafter referred to as a tension predicted value) applied to the divided portion P at the slit 207, which is the coating location, in accordance with the input of the first manufacturing conditions 211 and the first physical property information 213. ) is output.
• the liquid pressure prediction model 222 calculates a predicted value of the pressure of the liquid applied to the divided portion P in the slit 207 (hereinafter referred to as Outputs the predicted hydraulic pressure value).
• the coating thickness prediction model 223 outputs a predicted value of the liquid coating thickness of the divided portion P (hereinafter referred to as a coating thickness predicted value) in response to input of the tension predicted value and the hydraulic pressure predicted value. In this way, in the second embodiment as well, there are multiple types of predicted values, and a prediction model is prepared for each of the multiple types of predicted values.
• the prediction unit 72 when predicting the quality of the divided portion P1, the prediction unit 72 inputs the first manufacturing conditions 211 and the first physical property information 213_P1 of the divided portion P1 to the tension prediction model 221. Then, the prediction unit 72 causes the tension prediction model 221 to output the tension prediction value 230_P1 of the divided portion P1.
• the prediction unit 72 when predicting the quality of the divided portion P1, calculates the second manufacturing condition 212, the first physical property information 213_P1 of the divided portion P1, and the second physical property information 214. Input to pressure prediction model 222. Then, the prediction unit 72 causes the hydraulic pressure prediction model 222 to output the hydraulic pressure prediction value 231_P1 of the divided portion P1.
• the prediction unit 72 when predicting the quality of the divided portion P1, uses the tension prediction value 230_P1 output from the tension prediction model 221 and the hydraulic pressure prediction output from the hydraulic pressure prediction model 222.
• the value 231_P1 is input to the coating thickness prediction model 223.
• the prediction unit 72 causes the coating thickness prediction model 223 to output the coating thickness prediction value 235_P1 of the divided portion P1.
• FIGS. 28 to 30 an example is shown in which the predicted tension value 230_P1, the predicted hydraulic value 231_P1, and the predicted coating thickness value 235_P1 of the divided portion P1 are output from each of the prediction models 221 to 223, but the prediction unit 72
• the predicted tension values 230_P2 to 230_P8 (not shown), predicted hydraulic values 231_P2 to 231_P8 (not shown), and predicted coating thickness values 235_P2 to 235_P8 (see FIG. 31) are similarly predicted.
• predicted coating thickness values 235_P1 to 235_P8 of the divided portions P1 to P8 they may be referred to as predicted coating thickness values 235.
• the predicted value group 240 of the second embodiment has contents as shown in FIG. 31 as an example. That is, the predicted value group 240 includes a predicted coating thickness value 235_P1 of the divided portion P1, a predicted coating thickness value 235_P2 of the divided portion P2, . . . , a predicted coating thickness value 235_P7 of the divided portion P7, and a predicted coating thickness value of the divided portion P8. 235_P8.
• the learning data for each of the predictive models 221 to 223 is data collected from the divided portions P of the sheet 202 manufactured in the past.
• the learning data for each prediction model 221 to 223 has the following contents. That is, as input data for learning, first manufacturing conditions for learning corresponding to the first manufacturing conditions 211, second manufacturing conditions for learning corresponding to the second manufacturing conditions 212, and first manufacturing conditions for learning corresponding to the first physical property information 213 are used. It includes physical property information and learning second physical property information corresponding to the second physical property information 214.
• the measured value of the tension applied to the divided part P in the slit 207, the measured value of the pressure of the liquid applied to the divided part P in the slit 207, and the measured value of the thickness of the liquid applied to the divided part P. including.
• each prediction model 221 to 223 is based on the support 201 having a relatively large width and the number of divided portions P of 10 or more; This is data collected from various types of sheets 202 manufactured in the past, such as the case of the support body 201 of .
• the objective function 245 of the second embodiment has a first term 246 and a second term 247.
• the first term 246 includes the difference between the coating thickness predicted value 235 and the coating thickness setting value 250 (see FIG. 33) for each divided portion P. More specifically, the first term 246 includes the sum of the squares of the difference between the coating thickness predicted value 235 and the coating thickness setting value 250 divided by the coating thickness setting value 250.
• the first term 246 is a term obtained by multiplying the sum by the first weighting coefficient W1. Since the coating thickness setting value 250 is common to each divided portion P, there is only one coating thickness setting value 250.
• the first term 246 is an example of a "term including a difference" according to the technology of the present disclosure.
• the coating thickness setting value 250 is an example of a "target value" according to the technology of the present disclosure.
• the second term 247 is a regularization term for fitting the manufacturing conditions of each divided portion P into the constraint conditions regarding the manufacturing conditions of the entire sheet 202.
• the constraint regarding the manufacturing conditions for the entire sheet 202 is that the control amount of the tension adjustment roller 206 (the amount of pushing of the tension adjustment roller 206) of each divided portion P is linearly connected. Therefore, the second term 247 is a term that linearly connects the control amount (pushing amount) of each tension adjustment roller 206 of the divided portion P.
• the second term 247 is, for example, the coefficient of determination obtained by linear regression analysis of the pushing amount of each tension adjustment roller 206 in the divided portion P, subtracted from 1.
• the second term 247 has a second weighting factor W2.
• the first weighting coefficient W1 and the second weighting coefficient W2 do not necessarily have to be the same value. For example, if you want to emphasize the first term 246, that is, if you want to bring the predicted coating thickness 235 of each divided portion P closer to the coating thickness setting value 250, set the first weighting coefficient W1 to a higher value than the second weighting coefficient W2. May be set.
• the preferred manufacturing conditions 251 of the second embodiment derived by the derivation unit 73 by solving an optimization problem for determining the manufacturing conditions that minimize the value of the objective function 245 are tension adjustment. It includes the control amount of the roller 206. As the control amount of the tension adjustment roller 206, one value common to each divided portion P is derived by the function of the second term 247 of the objective function 245.
• the input data 210 is input every time the support body 201 is conveyed a preset length. Therefore, the prediction unit 72 predicts the tension prediction values 230_P1 to 230_P8, the hydraulic pressure prediction values 231_P1 to 231_P8, and the coating thickness prediction value 235, and the derivation unit 73 calculates the control amount of the tension adjustment roller 206 as the preferable manufacturing condition 251. The derivation is also performed each time the support body 201 is conveyed a preset length and the input data 210 is input.
• the tension adjustment roller 206 can always apply a suitable tension to the support 201 with respect to the support 201 that is continuously conveyed.
• the liquid can be applied to the support 201 in the applied state.
• the product is a sheet 202 formed by applying a liquid to a long support 201 that is being conveyed.
• the divided portion P is a portion obtained by dividing the sheet 202 along the width direction WD.
• the first manufacturing condition 211 includes the amount of control of the tension adjustment roller 206, which is arranged upstream of the slit 207 where the liquid is applied, and adjusts the tension applied to the support 201 in the slit 207.
• the predicted value includes the coating thickness of the liquid in the divided portion P (coating thickness predicted value 235).
• the support 201 having a relatively large width dimension and the number of divided parts P being 10 or more
• the predicted value of the coating thickness of the liquid on the divided portion P is derived.
• the predicted coating thickness value may be derived without deriving the predicted tension value and the predicted hydraulic pressure value.
• a culture solution 301 is stored in a culture tank 300.
• the capacity of the culture tank 300 is, for example, 50 L or more and 5000 L or less.
• Cells 302 are seeded in a culture tank 300, and the cells 302 are cultured in a culture solution 301.
• Cell 302 is an example of a "product" according to the technology of the present disclosure.
• the cell 302 is an antibody-producing cell established by integrating an antibody gene into a cell such as a Chinese hamster ovary cell, for example.
• the cells 302 produce immunoglobulin, ie, antibodies, as a product during the culture process. Therefore, not only cells 302 but also antibodies are present in the culture solution 301.
• the antibody is, for example, a monoclonal antibody, and is an active ingredient of biopharmaceuticals.
• the culture tank 300 is provided with a culture medium supply path 303, a gas supply path 304, an exhaust path 305, a culture solution delivery/recovery path 306, a sparger 307, a gas supply path 308, a stirring blade 309, and the like.
• the medium supply path 303 is a flow path for continuously supplying fresh medium into the culture tank 300. That is, in the culture tank 300, perfusion culture is performed.
• the gas supply path 304 is a flow path for supplying gas containing air and carbon dioxide from above.
• the exhaust path 305 is a flow path for exhausting the gas supplied from the gas supply path 304 to the outside of the culture tank 300.
• An exhaust filter 310 is provided in the exhaust path 305.
• the culture solution delivery/recovery path 306 is connected to the entrance/exit 312 of the cell removal filter 311.
• the culture solution delivery/recovery path 306 is a channel for sending the culture solution 301 in the culture tank 300 to the cell removal filter 311. Further, the culture solution delivery/recovery channel 306 is a flow channel for returning the culture solution 301 (concentrated solution) from the cell removal filter 311 into the culture tank 300.
• the sparger 307 is placed at the bottom of the culture tank 300.
• the sparger 307 releases the oxygen-containing gas supplied from the gas supply path 308 into the culture tank 300 .
• the oxygen released from the sparger 307 is dissolved in the culture medium 301 and assists the antibody production activity of the cells 302.
• the stirring blade 309 is rotated by a motor or the like at a predetermined number of rotations, and stirs the culture solution 301 in the culture tank 300. This maintains the homogeneity of the culture solution 301 in the culture tank 300.
• the culture tank 300 is also provided with a flow path for cell bleed processing in which a portion of the culture solution 301 is intentionally drawn out.
• the stirring blade 309 may have a plurality of blades as shown in the figure, or may have a single disc-shaped blade, and its shape is not particularly limited. Furthermore, two or more stirring blades 309 may be arranged within the culture tank 300.
• the cell removal filter 311 connected to the culture solution delivery/recovery path 306 has a filter membrane 313 inside.
• Filter membrane 313 captures cells 302 and is permeable to antibodies.
• the cell removal filter 311 obtains a culture supernatant by removing cells 302 from the culture solution 301 using a filter membrane 313, for example, using a tangential flow filtration (TFF) method.
• the cell removal filter 311 includes a diaphragm pump 315 having an elastic membrane 314 inside, and a degassing/air supply path 316.
• the culture solution 301 in the culture tank 300 is degassed from the air below the elastic membrane 314 through the deaeration/air supply path 316 and elastically deformed so as to stick to the lower end of the diaphragm pump 315. flows into the cell removal filter 311 via the culture solution delivery/recovery path 306.
• the elastic membrane 314 is elastically deformed so as to stick to the upper side of the diaphragm pump 315, so that air cannot pass through the filter membrane 313.
• the culture solution 301 (concentrated solution) is returned into the culture tank 300 via a culture solution delivery/recovery path 306.
• the culture supernatant fluid flows out from the outlet 317 of the cell removal filter 311.
• the culture supernatant mainly contains antibodies.
• the culture supernatant is sent to a downstream purification section (not shown), where it undergoes various chromatography treatments, virus inactivation treatments, etc., and is ultimately used as an active pharmaceutical ingredient for biopharmaceuticals.
• one divided portion P corresponds to a culture vessel 320, such as a petri dish, which is smaller than the culture tank 300, and has a capacity of, for example, several mL to several liters.
• the input data 325 of the third embodiment includes manufacturing conditions 326_P1 for the divided portion P1, manufacturing conditions 326_P2 for the divided portion P2, and so on.
• the manufacturing conditions 326_P1 for the divided portion P1, the manufacturing conditions 326_P2 for the divided portion P2, etc. may be referred to as manufacturing conditions 326 if there is no particular need to distinguish between them.
• the manufacturing conditions 326 include culture conditions for the cells 302.
• the manufacturing conditions 326 include the density of the cells 302 in the divided portion P (cell density), the hydrogen ion index in the divided portion P, oxygen concentration, temperature, medium concentration, and shear energy due to stirring by the stirring blade 309. .
• These include the capacity of the culture tank 300, the position of the medium supply path 303, the position of the gas supply path 304, the position of the sparger 307, the position of the stirring blade 309, the amount of medium supplied, the amount of gas supplied, and the stirring blade. This is a value predicted by a simulation using various parameters such as the rotation speed of the 309.
• carbon dioxide concentration, etc. may be added as appropriate.
• physical property information of the cells 302 such as the size of the cells 302 and/or design information such as the capacity of the culture tank 300 may be added to the input data 325.
• a culture environment prediction model 330 is used in the third embodiment.
• the culture environment prediction model 330 is a machine learning model configured by, for example, a neural network.
• the culture environment prediction model 330 is an example of a "machine learning model" according to the technology of the present disclosure.
• the prediction unit 72 When predicting the quality of the divided portion P1, the prediction unit 72 inputs the manufacturing conditions 326_P1 to the culture environment prediction model 330. Then, the prediction unit 72 causes the culture environment prediction model 330 to output the culture environment prediction value 331_P1 of the divided portion P1.
• FIG. 37 shows an example in which the culture environment prediction value 331_P1 of the divided portion P1 is output from the culture environment prediction model 330
• the prediction unit 72 similarly predicts the culture environment for the remaining divided portions P2, .
• Values 331_P2 (see FIG. 38), . . . are output from the culture environment prediction model 330.
• the predicted culture environment value 331 is a value representing the quality of the culture environment for the cells 302 based on the proliferation rate, survival rate, etc. of the cells 302, and takes a value between 0 and 100, for example. If the value is close to 100, the culture environment for the cells 302 is good.
• the culture environment predicted value 331 is an example of an "index value" according to the technology of the present disclosure.
• the predicted value group 335 of the third embodiment has contents as shown in FIG. 38 as an example. That is, the predicted value group 335 includes a culture environment predicted value 331_P1 for the divided portion P1, a culture environment predicted value 331_P2 for the divided portion P2, and so on.
• the learning data of the culture environment prediction model 330 is data collected from cells 302 produced in the past using the small-scale culture container 320. Although not shown, the learning data of the culture environment prediction model 330 has the following contents. That is, the learning input data includes the learning manufacturing conditions corresponding to the manufacturing conditions 326. The correct data includes actual measured values of the quality of the culture environment for the cells 302 in the small-scale culture container 320.
• the learning data of the culture environment prediction model 330 is for the case of cells producing antibody A, the case of cells producing antibody B, or the case of culturing in a culture container 320 with a capacity of 50 mL, and the case of a culture vessel with a capacity of 1 L.
• This is data collected from various types of cells 302 that have been produced in the past using a small-scale culture vessel 320, such as when cultured in a small-scale culture vessel 320.
• the objective function 340 of the third embodiment has a first term 341 and a second term 342.
• the first term 341 includes the difference between the culture environment predicted value 331 and the culture environment setting value 350 (see FIG. 40) of each divided portion P. More specifically, the first term 341 includes the sum of the squares of the difference between the culture environment predicted value 331 and the culture environment setting value 350 divided by the culture environment setting value 350.
• the first term 341 is a term obtained by multiplying the sum by the first weighting coefficient W1. Since the culture environment setting value 350 is common to each divided portion P, there is only one culture environment setting value 350.
• the first term 341 is an example of a “term including a difference” according to the technology of the present disclosure. Further, the culture environment setting value 350 is an example of a "target value" according to the technology of the present disclosure.
• the second term 342 is a regularization term for fitting the manufacturing conditions of each divided portion P into the constraint conditions regarding the manufacturing conditions of the entire cells 302 in the culture tank 300.
• a constraint regarding the manufacturing conditions for the entire cells 302 in the culture tank 300 is that the manufacturing conditions for each of the divided portions P be made equal. Therefore, the second term 342 is a term that makes the manufacturing conditions of each of the divided portions P equal.
• the second term 342 is, for example, the variance or standard deviation of the manufacturing conditions of each divided portion P. In this case, by minimizing the variance or standard deviation, the manufacturing conditions for each of the divided portions P can be made equal.
• the second term 342 has a second weighting factor W2.
• the first weighting coefficient W1 and the second weighting coefficient W2 do not necessarily have to be the same value. For example, if you want to emphasize the first term 341, that is, if you want to bring the predicted culture environment value 331 of each divided portion P closer to the culture environment setting value 350, set the first weighting coefficient W1 to a higher value than the second weighting coefficient W2. May be set.
• the derivation unit 73 solves an optimization problem for finding manufacturing conditions that minimize the value of the objective function 340, thereby determining a temporary suitable manufacturing condition for the divided portion P1.
• Temporary suitable manufacturing conditions 351 for each of the divided portions P such as conditions 351_P1, . . . , are derived.
• the temporary suitable manufacturing conditions 351 include the hydrogen ion index, oxygen concentration, temperature, and medium concentration in the divided portion P.
• the hydrogen ion index, oxygen concentration, temperature, and medium concentration are almost uniform in each divided portion P due to stirring by the stirring blades 309.
• the shear energy due to stirring by the stirring blades 309 is different for each divided portion P.
• the derivation unit 73 derives more suitable manufacturing conditions from the provisional suitable manufacturing conditions 351 of each divided portion P.
• a suitable oxygen supply amount 356 to be supplied into the culture tank 300 via the sparger 307 is derived from the oxygen concentration of the provisional suitable manufacturing conditions 351 for each of the divided portions P. It shows how to do it.
• the derivation unit 73 derives a suitable oxygen supply amount 356 by predicting, through simulation, an oxygen supply amount such that the oxygen concentration of each divided portion P becomes the oxygen concentration of the tentative suitable manufacturing conditions 351.
• the suitable oxygen supply amount 356 is an example of "suitable manufacturing conditions" according to the technology of the present disclosure.
• a suitable medium supply amount from the medium supply path 303 a suitable gas supply amount from the gas supply path 304, a suitable rotation speed of the stirring blade 309, etc. may be derived as suitable manufacturing conditions.
• the product is the cells 302 cultured in the culture tank 300.
• the divided portion P is a portion into which the culture tank 300 is divided.
• the manufacturing conditions 326 include culture conditions for the cells 302, and the predicted values include an index value (culture environment predicted value 331) representing the desirability of the culture environment for the cells 302. Therefore, it is possible to deal with various aspects, such as in the case of cells producing antibody A and cells producing antibody B. In other words, it is possible to derive manufacturing conditions that are suitable for general use regardless of the type of cells 302.
• the cells 302 are not limited to the illustrated antibody-producing cells. Pluripotent stem cells such as induced pluripotent stem cells (iPS) may also be used. As the machine learning model, multiple types of models such as a model that predicts the proliferation rate of the cells 302 in each divided portion P and a model that predicts the survival rate of the cells 302 in each divided portion P may be prepared.
• iPS induced pluripotent stem cells
• the information processing device 10 may be installed in the same facility as the extrusion molding machine 12 etc., or may be installed in a data center independent from the facility where the extrusion molding machine 12 etc. are installed.
• various processors include the CPU 57, which is a general-purpose processor that executes software (operating program 65) and functions as various processing units, as well as FPGA (Field Programmable Gate Array), etc.
• Dedicated processors are processors with circuit configurations specifically designed to execute specific processes, such as programmable logic devices (PLDs), which are processors whose circuit configurations can be changed, and ASICs (Application Specific Integrated Circuits). Includes electrical circuits, etc.
• PLDs programmable logic devices
• ASICs Application Specific Integrated Circuits
• One processing unit may be composed of one of these various processors, or a combination of two or more processors of the same type or different types (for example, a combination of multiple FPGAs and/or a CPU and (in combination with FPGA). Further, the plurality of processing units may be configured with one processor.
• one processor is configured with a combination of one or more CPUs and software, as typified by computers such as clients and servers, and this There is a form in which a processor functions as multiple processing units.
• SoC system-on-chip
• various processing units are configured using one or more of the various processors described above as a hardware structure.
• an electric circuit that is a combination of circuit elements such as semiconductor elements can be used.
• the processor includes: Using a machine learning model that outputs a predicted value of the quality of the divided parts according to the input of the manufacturing conditions of the divided parts into which the whole product is divided, inputting the manufacturing conditions of the divided portion into the machine learning model and outputting the predicted value of the divided portion from the machine learning model;
• the quality of the entire product can be improved by solving an optimization problem for finding manufacturing conditions that minimize the value of an objective function that has a term that includes the difference between the predicted value of the divided portion and the target quality value of the divided portion. Deriving suitable manufacturing conditions that serve as target values, Information processing device.
• the objective function includes a regularization term for fitting the manufacturing conditions of the divided portions into the constraint conditions regarding the manufacturing conditions of the entire product.
• the regularization term is a term for smoothing the manufacturing conditions of the divided portions.
• the regularization term is a sum of second-order differential values of the manufacturing conditions of the divided portions.
• the term including the difference is a sum of values obtained by dividing the difference by the target value of the divided portion.
• the divided portion is a portion obtained by dividing the flexible tube for an endoscope along the axial direction
• the manufacturing conditions include an extrusion amount of the inner layer resin material per unit time and an extrusion amount of the outer layer resin material per unit time by an extrusion molding machine,
• the information processing device according to any one of Supplementary Notes 1 to 7, wherein the predicted value includes an elasticity value of the divided portion, a thickness of the inner layer of the divided portion, and a thickness of the outer layer of the divided portion.
• the product is a sheet formed by applying a liquid to a long support that is transported,
• the divided portion is a portion where the sheet is divided along the width direction,
• the manufacturing conditions include a control amount of a tension adjustment roller that is arranged upstream of the liquid application location and adjusts the tension applied to the support at the application location,
• the information processing device according to any one of Supplementary Notes 1, 2, and 5 to 7, wherein the predicted value includes a coating thickness of the liquid in the divided portion.
• the product is a cell cultured in a culture tank,
• the divided portion is a divided portion of the culture tank,
• the manufacturing conditions include culture conditions for the cells,
• the information processing device according to any one of Supplementary Notes 1, 2, and 5 to 7, wherein the predicted value includes an index value representing the desirability of a culture environment for the cells.
• a and/or B has the same meaning as “at least one of A and B.” That is, “A and/or B” means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, even when three or more items are expressed in conjunction with “and/or”, the same concept as “A and/or B" is applied.

Landscapes

• Engineering & Computer Science (AREA)
• Health & Medical Sciences (AREA)
• Physics & Mathematics (AREA)
• Life Sciences & Earth Sciences (AREA)
• General Engineering & Computer Science (AREA)
• General Physics & Mathematics (AREA)
• Mechanical Engineering (AREA)
• Chemical & Material Sciences (AREA)
• Automation & Control Theory (AREA)
• Manufacturing & Machinery (AREA)
• Medical Informatics (AREA)
• Artificial Intelligence (AREA)
• Software Systems (AREA)
• Quality & Reliability (AREA)
• Wood Science & Technology (AREA)
• Surgery (AREA)
• Theoretical Computer Science (AREA)
• Organic Chemistry (AREA)
• Analytical Chemistry (AREA)
• Zoology (AREA)
• Bioinformatics & Cheminformatics (AREA)
• Biomedical Technology (AREA)
• General Health & Medical Sciences (AREA)
• Evolutionary Computation (AREA)
• Computer Vision & Pattern Recognition (AREA)
• Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
• Pathology (AREA)
• Heart & Thoracic Surgery (AREA)
• Microbiology (AREA)
• Molecular Biology (AREA)
• Animal Behavior & Ethology (AREA)
• Public Health (AREA)
• Veterinary Medicine (AREA)
• Biotechnology (AREA)
• Radiology & Medical Imaging (AREA)
• Optics & Photonics (AREA)
• Biophysics (AREA)
• Sustainable Development (AREA)
• Computer Hardware Design (AREA)
• Biochemistry (AREA)

Priority Applications (3)

Applications Claiming Priority (2)

Related Child Applications (1)

Publications (1)

Family

ID=90099606

Family Applications (1)

Country Status (4)

Families Citing this family (1)

Citations (2)

Family Cites Families (4)

• 2023
  • 2023-08-23 WO PCT/JP2023/030284 patent/WO2024048388A1/ja not_active Ceased
  • 2023-08-23 EP EP23860151.2A patent/EP4583020A4/en active Pending
  • 2023-08-23 JP JP2024544164A patent/JPWO2024048388A1/ja active Pending
• 2025
  • 2025-02-21 US US19/059,281 patent/US20250189955A1/en active Pending

Patent Citations (2)

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events