Molding condition derivation device, machine learning device, inference device, information processing method, machine learning method, and inference method

Abstract

To obtain a molding condition derivation device, a machine learning device, an inference device, an information processing method, a machine learning method, and an inference method, in which an appropriate blow molding condition can be easily derived.SOLUTION: There is provided a molding condition derivation device 6 that determines output data comprising information for target molding condition that is a target molding condition based on input data comprising information for a bottle shape that is information about a shape of a molded body 4. When input data is entered, a target molding condition is determined using a learning model in which correlation between input data and output data is learned by machine learning, and information about a determined target molding condition is output as output data. The target molding condition is a target value of one or more blow molding conditions set in a blow molding step. The bottle shape information includes a bottle volume.SELECTED DRAWING: Figure 10

Description

[Technical Field]

[0001] The present invention relates to a molding condition derivation device, a machine learning device, an inference device, an information processing method, a machine learning method, and an inference method. [Background technology]

[0002] In conventional container wall thickness design systems, the mechanical properties of a container are calculated based on CAD data related to the shape of the container to be blow-molded, and the ideal wall thickness of the container is then determined. The ideal wall thickness is a wall thickness that provides mechanical properties that satisfy predetermined conditions. Then, blow molding conditions are determined such that the ideal wall thickness matches or approximates the wall thickness of the container calculated using a simulation model (for example, see Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2003-62896 Summary of the Invention [Problem to be solved by the invention]

[0004] However, with conventional container wall thickness design systems, it was difficult to match the container wall thickness calculated using a simulation model based on blow molding conditions with the wall thickness of the container that was actually blow molded, and in order to achieve this matching, it was necessary to spend a lot of time improving the simulation model.

[0005] The present invention has been made to solve the above-mentioned problems, and aims to provide a molding condition derivation device, a machine learning device, an inference device, an information processing method, a machine learning method, and an inference method that can easily derive appropriate blow molding conditions. [Means for solving the problem]

[0006] The molding condition derivation device according to the present invention derives target molding conditions, which are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding. The molding condition derivation device calculates output data consisting of target molding condition information, which is information about the target molding conditions, based on input data consisting of bottle shape information, which is information about the shape of the molded body. When the input data is input, the molding condition derivation device determines target molding conditions using a learning model in which the correlation between the input data and the output data is learned by machine learning, and outputs information about the determined target molding conditions as output data, where the bottle shape information includes the capacity. [Effects of the Invention]

[0007] According to the molding condition derivation device, machine learning device, inference device, information processing method, machine learning method, and inference method of the present invention, appropriate blow molding conditions can be easily derived. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a schematic diagram illustrating an example of a container molding system according to a first embodiment. [Figure 2] 1 is a block diagram showing an example of a blow molding device according to a first embodiment. [Figure 3] FIG. 2 is a schematic configuration diagram showing an example of a molding unit according to the first embodiment. [Figure 4] 1 is a flowchart showing the flow of a blow molding process. [Figure 5] FIG. 10 is a diagram for explaining the amount of stretching. [Figure 6] FIG. 1 is a block diagram illustrating an example of a machine learning device according to a first embodiment. [Figure 7] FIG. 2 is a diagram illustrating an example of a learning model and learning data according to the first embodiment. [Figure 8] 1 is a flowchart illustrating an example of a machine learning routine executed by the machine learning device. [Figure 9] 1 is a block diagram showing an example of a molding condition deriving device according to a first embodiment. [Figure 10] FIG. 10 is a functional explanatory diagram showing an example of the function of the molding condition derivation device of FIG. [Figure 11] FIG. 2 is a hardware configuration diagram illustrating an example of a computer. [Figure 12] FIG. 10 is a functional explanatory diagram showing an example of a molding condition deriving device according to a second embodiment. [Figure 13] 13 is a functional explanatory diagram showing an example of the function of the molding condition derivation device of FIG. 12. FIG. [Figure 14] 10 is a flowchart illustrating an algorithm for inference using a mathematical optimization model. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, an embodiment for carrying out the present invention will be described with reference to the drawings. The scope necessary for the explanation to achieve the object of the present invention will be schematically shown, and the scope necessary for explaining the relevant parts of the present invention will be mainly explained, and the parts that are omitted from the explanation will be based on publicly known techniques.

[0010] (First embodiment) FIG. 1 is a schematic diagram showing an example of a container molding system according to a first embodiment. The container molding system 1 functions as a production system that uses a mold 20 to produce a hollow molded body 4 by blow molding. In this embodiment, a case will be described in which a blow-molded container (PET bottle), which is one form of the hollow molded body 4, is produced from a preform 3 by a biaxial stretch blow molding method. The preform 3 is one form of a molded body made from a synthetic resin containing polyethylene terephthalate (PET) as a raw material.

[0011] The container molding system 1 includes, as its main components, an injection molding device 1000, a preform removal device 1100, a preform transport conveyor 1200, a preform supply device 1300, a blow molding device 2, a molded body removal device 1400, and a molded body transport conveyor 1500.

[0012] In the container molding system 1, when synthetic resin, which is a raw material, is fed into the injection molding device 1000, the injection molding device 1000 molds the fed synthetic resin into preforms 3. A preform removal device 1100 removes the preforms 3 from the injection molding device 1000 and transfers them to a preform transfer conveyor 1200. A heating device 1210 is provided on the preform transfer conveyor 1200. The heating device 1210 heats the preforms 3 that pass through the heating device 1210 as they are transported by the preform transfer conveyor 1200.

[0013] The preform supply device 1300 sequentially supplies heated preforms 3 to the blow molding device 2. The blow molding device 2 molds the preforms 3 into molded articles 4 (PET bottles). Thereafter, the molded article removal device 1400 removes the molded articles 4 from the blow molding device 2 and transfers them to a molded article transfer conveyor 1500.

[0014] Fig. 2 is a block diagram showing an example of a blow molding apparatus according to the first embodiment. Fig. 3 is a schematic configuration diagram showing an example of a molding unit according to the first embodiment. The blow molding apparatus 2 includes, as its main components, a rotary unit 21, a plurality of molding units 22, and a control unit 23.

[0015] The rotary unit 21 has a rotary support and a rotation mechanism. The rotary support is formed in a disk shape and supports a plurality of molding units 22 arranged at equal intervals in the circumferential direction of the rotary support. The rotation mechanism rotates the rotary support at a predetermined rotation speed.

[0016] Each molding unit 22 includes a mold support mechanism 220 , a seal support mechanism 221 , a stretch rod 222 , a stretch rod support mechanism 223 , a blow fluid supply / discharge mechanism 224 , and a temperature adjustment mechanism 225 .

[0017] The mold support mechanism 220 supports the mold 20 so that it can be opened and closed. The seal support mechanism 221 supports the preform 3. When the mold 20 is closed by the mold support mechanism 220, it sandwiches the preform 3. This seals the preform 3 within the mold 20. The stretch rod 222 is arranged so that it can be inserted into the preform 3 through an opening in the preform 3. The stretch rod support mechanism 223 supports the stretch rod 222 so that it can move back and forth.

[0018] The blow fluid supply / discharge unit 224 supplies or discharges blow fluid to or from the preform 3. In this embodiment, air is used as the blow fluid. The temperature adjustment mechanism 225 adjusts the temperature of the mold 20. Note that, in this embodiment, the blow fluid is described as being air, but it may be any gas other than air, or may be a liquid.

[0019] The blow fluid supply / discharge unit 224 has a main pipe 2240, three branch pipes 2241A, 2241B, and 2241C, a module group, and a sensor group. The module group includes a pre-blow valve 2242, a main blow valve 2243, and an exhaust valve 2244. The sensor group includes a pressure sensor 2245, a flow rate sensor 2246, and a temperature sensor 2247.

[0020] The main pipe 2240 is connected to the stretch rod 222 via the seal support portion 221. Three branch pipes 2241A, 2241B, and 2241C branch off from the main pipe 2240. The branch pipe 2241A is connected to a pre-blow air supply source (not shown). The pre-blow air supply source is a supply source of pre-blow air as a pre-blow fluid. The branch pipe 2241B is connected to a main blow air supply source (not shown). The main blow air supply source is a supply source of main blow air as a main blow fluid. The pressure of the main blow air supply source is higher than the pressure of the pre-blow air supply source. The branch pipe 2241C is connected to an exhaust system (not shown).

[0021] The pre-blow valve 2242 is provided in the branch pipe 2241 A. The main blow valve 2243 is provided in the branch pipe 2241 B. The exhaust valve 2244 is provided in the branch pipe 2241 C.

[0022] The pressure sensor 2245 is provided in the main pipe 2240. The pressure sensor 2245 measures the pressure of the air supplied into the preform 3 at predetermined time intervals and outputs the result to the control unit 23. The flow rate sensor 2246 is provided in the main pipe 2240. The flow rate sensor 2246 measures the flow rate of the air supplied into the preform 3 at predetermined time intervals and outputs the result to the control unit 23. The temperature sensor 2247 measures the temperature of the preform 3 at predetermined time intervals and outputs the result to the control unit 23.

[0023] 3 does not show the specific configurations of the mold support mechanism 220, the seal support mechanism 221, and the stretch rod support mechanism 223. These mechanisms are configured by appropriately combining, for example, modules for generating driving force such as servo motors and cylinders, driving force transmission mechanisms such as linear guides, ball screws, gears, cams, belts, couplings, and bearings, and sensors such as linear sensors, encoder sensors, and limit sensors.

[0024] 3 does not show the specific configuration of the temperature adjustment mechanism 225. The temperature adjustment mechanism 225 is configured by appropriately combining, for example, a temperature adjustment module such as an electric heater and a sensor such as a temperature sensor. The pressure sensor 2245 and the flow rate sensor 2246 may be provided in the seal support part 221 instead of in the main pipe 2240.

[0025] The control unit 23 is electrically connected to the module group and the sensor group provided in the rotary unit 21. The control unit 23 is also electrically connected to the module group and the sensor group provided in each molding unit 22.

[0026] The control unit 23 is configured by, for example, a general-purpose or dedicated computer. The control unit 23 has, as its main components, a control unit 230, a communication unit 231, an input unit 232, an output unit 233, and a storage unit 234.

[0027] The control unit 230 is configured by, for example, an arithmetic processing device or a sequencer. The control unit 230 functions as a blow molding control unit 2300, a pressure monitoring unit 2301, a flow rate monitoring unit 2302, and a temperature monitoring unit 2303, for example, by executing a blow molding program 2340 stored in the storage unit 234.

[0028] The blow molding control unit 2300 operates the modules included in the rotary unit 21 and the modules included in each molding unit 22. The pressure monitoring unit 2301 monitors the measurement results of each pressure sensor 2245 and notifies the blow molding control unit 2300 of the monitored air pressure. The flow rate monitoring unit 2302 monitors the measurement results of each flow rate sensor 2246 and notifies the blow molding control unit 2300 of the monitored air flow rate. The temperature monitoring unit 2303 monitors the measurement results of each temperature sensor 2247 and notifies the blow molding control unit 2300 of the monitored air temperature. The temperature monitoring unit 2303 also monitors the measurement results of a temperature sensor (not shown) provided in the heating device 1210.

[0029] The communication unit 231 is connected to a communication network and functions as a communication interface for transmitting and receiving various types of data to and from, for example, a terminal device used by a user of the blow molding apparatus 2. The input unit 232 accepts various input operations by a user of the blow molding apparatus 2. The output unit 233 functions as a user interface by outputting various types of information to the user through a screen display, lighting up a signal tower, and sounding a buzzer.

[0030] The storage unit 234 stores various programs and data used in the operation of the blow molding apparatus 2. The programs include an operating system and a blow molding program 2340. The data includes apparatus setting information 2341. The apparatus setting information 2341 is information that can register various operating conditions when the blow molding apparatus 2 executes a blow molding process, and is configured to be editable by the user via a display screen, for example.

[0031] 4 is a flowchart showing the flow of the blow molding process. The blow molding process is a process for carrying out a blow molding treatment, in which the preform 3 is heated, the heated preform 3 is set in a mold 20, and the set preform 3 is blow molded to obtain a molded body 4. The blow molding process includes a heating process, a pre-blow process, and a main blow process.

[0032] The heating step is a step of heating the preform 3 (step S0). The pre-blow step is a step of introducing pre-blow air into the heated preform 3. The pre-blow step is performed in the following order: setting the preform 3 (step S1), closing the mold 20 (step S2), stretching (step S3), starting the pre-blow air supply (step S4), and maintaining the pre-blow pressure (step S5). The main blow step is a step of introducing main blow air into the preform 3 following the pre-blow step. The pre-blow pressure is the pressure of the pre-blow air. The main blow step is performed in the following order: starting the main blow air supply (step S6), maintaining the main blow pressure (step S7), starting the blow air exhaust (step S8), opening the mold 20 (step S9), and removing the molded body (step S10). The main blow pressure is the pressure of the main blow air.

[0033] More specifically, in step S0, the heating device 1210 heats the preform 3 while the preform 3 passes through the heating device 1210.

[0034] In step S1, the blow molding control unit 2300 causes the heated preform 3 to be supported by the seal support unit 221. In step S2, the blow molding control unit 2300 causes the mold support mechanism unit 220 to close the mold 20. In step S3, the blow molding control unit 2300 advances the stretch rod 222 along the central axis of the preform 3, thereby stretching the preform 3.

[0035] In step S4, the blow molding control unit 2300 opens the pre-blow valve 2242. In step S5, the blow molding control unit 2300 advances the stretch rod 222 while continuing to supply pre-blow air to maintain the pre-blow pressure.

[0036] In step S6, the blow molding control unit 2300 closes the pre-blow valve 2242 to stop the supply of pre-blow air, and opens the main blow valve 2243 to increase the main blow pressure. In step S7, the blow molding control unit 2300 continues the supply of main blow air to maintain the main blow pressure at the target pressure. In step S8, the blow molding control unit 2300 closes the main blow valve 2243 and opens the exhaust valve 2244 to discharge the blow fluid from the main pipe 2240 to the outside.

[0037] In step S9, the blow molding control unit 2300 causes the mold support mechanism unit 220 to open the mold 20 while causing the stretch rods 222 to retract from inside the preform 3. In step S10, the blow molding control unit 2300 releases the state in which the molded body is fixed by the seal support unit 221. This makes the molded body removable from the seal support unit 221.

[0038] That is, the pre-blow process is a blow molding process in which a stretch rod 222 is inserted into a heated preform 3 along the central axis of the preform 3 to stretch the preform 3 in the central axis direction and introduce air into the preform 3.

[0039] The pre-blow conditions in the pre-blow process include the pre-blow pressure, the pressure maintenance period, the flow rate of pre-blow air while the pre-blow pressure is maintained, the stretch amount of the preform 3, and the stretch speed of the preform 3. The pressure maintenance period is the period during which the pre-blow pressure is maintained, and corresponds to the period of step S5 in FIG. 4. The stretch amount of the preform 3 is the amount by which the preform 3 stretches in the central axis direction. The stretch speed of the preform 3 is the speed at which the preform 3 stretches in the central axis direction.

[0040] Figure 5 is a diagram illustrating the stretch amount. The diagram on the left side of Figure 5 shows the relative positional relationship between the preform 3 and the stretch rod 222 in step S2 of Figure 4. The diagram in the center of Figure 5 shows the state in step S3 of Figure 4, where the advancing stretch rod 222 hits the preform 3. The diagram on the right side of Figure 5 shows the state in step S5 of Figure 4, where the stretch rod 222 stops. Thus, the stretch amount Lx is the amount by which the preform 3 is stretched by inserting the stretch rod 222 into it.

[0041] 6 is a block diagram showing an example of a machine learning device according to the first embodiment. The machine learning device 5 includes a control unit 50, a communication unit 51, a learning data storage unit 52, and a trained model storage unit 53.

[0042] The control unit 50 functions as a learning data acquisition unit 500 and a machine learning unit 501. The communication unit 51 is connected to an external device via the network 7 and functions as a communication interface for transmitting and receiving various data. The external device is, for example, the blow molding machine 2 and the operator terminal device 8.

[0043] The learning data acquisition unit 500 is connected to the blow molding machine 2 and the worker terminal device 8 via the communication unit 51 and the network 7, and acquires learning data 13 from at least one of the blow molding machine 2 and the worker terminal device 8. The learning data 13 is composed of input data and output data.

[0044] The input data is composed of bottle shape information and preform shape information. The bottle shape information is information relating to the shape of the molded body 4, and includes at least one of drawing information about the molded body 4, image information about the molded body 4, and design parameter information about the molded body 4. The drawing information includes, for example, CAD data and three-view drawings. The image information includes image data.

[0045] The design parameter information includes the type, strength characteristics, measurement characteristics, material characteristics, and other characteristics of the molded body 4. The types of molded body 4 are divided into those for carbonated beverages and those for aseptic beverages, for example. The strength characteristics include, for example, compressive strength, drop strength, and environmental stress crack resistance.

[0046] The measurement characteristics include the barrel diameter, minimum barrel diameter, wall thickness, property distribution, capacity, height, and false bottom height. The material characteristics include current-voltage characteristics, transparency, thermal properties, crystallinity, and carbon dioxide permeability. Other characteristics include, for example, the inversion angle, perpendicularity, longitudinal stretch ratio, transverse stretch ratio, volume ratio, and circumference ratio. The longitudinal stretch ratio, transverse stretch ratio, volume ratio, and circumference ratio are all the ratios of the molded body 4 to the preform 3.

[0047] The preform shape information is information relating to the shape of the preform 3, and includes at least one of drawing information about the preform 3, image information about the preform 3, and design parameter information about the preform 3.

[0048] The design parameter information includes the type of preform 3, its metrological characteristics, and its material characteristics. The types of preform 3 are divided into, for example, those for carbonated beverages, those for aseptic beverages, and a plurality of sizes corresponding to those for carbonated beverages and those for aseptic beverages, respectively. The metrological characteristics include the body diameter, wall thickness, property distribution, capacity, and height. The material characteristics include current-voltage characteristics, transparency, thermal properties, crystallinity, and carbon dioxide gas permeability.

[0049] The output data is composed of molding condition information, which includes preform conditions, pre-blow conditions, and main blow conditions.

[0050] The preform conditions include at least one of the preform temperature and the preform temperature control conditions. The preform temperature is the temperature of the preforms 3 when they are heated in the preform transport conveyor 1200 and supplied to the blow molding device 2 by the preform supply device 1300. The preform temperature control conditions include, for example, the heater temperature in the heating device 1210, the length of the heating section, and the transport speed of the preform transport conveyor 1200.

[0051] The pre-blow conditions include at least one of the pre-blow pressure, pre-blow time (pre-blow start timing, pre-blow maintenance time), pre-blow air flow rate, the stretching amount of the preform 3 by the stretch rod 222, and the stretching speed by the stretch rod 222. The main blow conditions include at least one of the main blow pressure, main blow time (main blow start timing, main blow maintenance time), main blow air flow rate, and main blow maintenance time.

[0052] The learning data 13 is data used as teacher data (training data), verification data, and test data in supervised learning. The molding condition information is data used as a correct answer label in supervised learning.

[0053] The learning data storage unit 52 is a database that stores multiple sets of learning data 13 acquired by the learning data acquisition unit 500. The specific configuration of the database that constitutes the learning data storage unit 52 is designed as appropriate.

[0054] The machine learning unit 501 performs machine learning using multiple sets of learning data 13 stored in the learning data storage unit 52. That is, the machine learning unit 501 inputs multiple sets of learning data 13 to the learning model 12, and generates a trained learning model 12 by having the learning model 12 learn the correlation between the input data included in the learning data 13, i.e., bottle shape information and preform shape information, and the output data, i.e., molding condition information.

[0055] The trained model storage unit 53 is a database that stores the trained learning model 12 generated by the machine learning unit 501. Specifically, the trained learning model 12 is a group of adjusted weight parameters. The trained learning model 12 stored in the trained model storage unit 53 is provided to an actual system, for example, the blow molding apparatus 2, via the network 7 or a recording medium. Note that although the training data storage unit 52 and the trained model storage unit 53 are shown as separate storage units in FIG. 6, they may also be configured as a single storage unit.

[0056] 7 is a diagram showing an example of the learning model 12 and learning data 13 according to the first embodiment. The learning data 13 used for machine learning of the learning model 12 is composed of bottle shape information and preform shape information.

[0057] The learning data acquisition unit 500 receives bottle shape information and preform shape information from the blow molding apparatus 2 or the worker terminal device 8. As a result, the learning data acquisition unit 500 acquires learning data 13. In addition, the learning data acquisition unit 500 receives target molding condition information as a correct label from the blow molding apparatus 2 or the worker terminal device 8.

[0058] The target molding condition information is information relating to target molding conditions, which are based on the results of adjustments made by a skilled worker to, for example, preform temperature conditions, preform temperature control conditions, pre-blow conditions, and main blow conditions.

[0059] The learning model 12 employs, for example, a neural network structure. The learning model 12 includes an input layer 120, an intermediate layer 121, and an output layer 122. A plurality of synapses are established between the input layer 120, the intermediate layer 121, and the output layer 122, connecting a plurality of neurons, and each synapse is associated with a weight. A group of weight parameters consisting of the weights of each synapse is adjusted by machine learning.

[0060] The input layer 120 has neurons in a number corresponding to the bottle shape information and preform shape information as input data. Values ​​of the bottle shape information and preform shape information are input to each neuron in the input layer 120. The output layer 122 has neurons in a number corresponding to the molding condition information as output data. The output layer 122 outputs the prediction result of the target molding condition information for the molding condition information, i.e., the inference result, as output data.

[0061] 8 is a flowchart showing an example of a machine learning routine executed by the machine learning device 5. When the routine of FIG. 8 starts, in step S100, the learning data acquisition unit 500 acquires multiple pieces of learning data 13 as preparation for starting machine learning, and stores the acquired learning data 13 in the learning data storage unit 52. Here, the number of pieces of learning data acquired for preparation may be set in consideration of the inference accuracy required for the learning model 12 to be finally obtained.

[0062] Next, in step S110, the machine learning unit 501 prepares a pre-learning learning model 12 in order to start machine learning. The pre-learning learning model 12 prepared here is configured by the neural network model exemplified in Fig. 7. At this point, the weights of each synapse are set to their initial values.

[0063] Next, in step S120, the machine learning unit 501 acquires, for example, one set of training data 13 at random from the multiple sets of training data 13 stored in the training data storage unit 52.

[0064] Next, in step S130, the machine learning unit 501 inputs the bottle shape information and preform shape information (input data) included in one set of learning data 13 to the input layer 120 of the prepared learning model 12 before learning or during learning. As a result, target molding condition information (output data) is output as an inference result from the output layer 122 of the learning model 12. Since this output data is data generated by the learning model 12 before learning or during learning, the output data output as an inference result indicates information different from the target molding condition information (correct label) included in the learning data 13.

[0065] Next, in step S140, the machine learning unit 501 compares the target molding condition information (correct label) included in the set of learning data 13 acquired in step S120 with the target molding condition information (output data) output as an inference result from the output layer 122 in step S130. The machine learning unit 501 performs machine learning by performing a process of adjusting the weight of each synapse, i.e., backpropagation, based on the comparison result between the correct label and the output data. In this way, the machine learning unit 501 causes the learning model 12 to learn the correlation between the bottle shape information, preform shape information, and the target molding condition information.

[0066] Next, in step S150, the machine learning unit 501 determines whether a learning termination condition is met. For example, the machine learning unit 501 determines whether the learning termination condition is met based on at least one of the evaluation value of an error function based on the correct label and the output data and the remaining number of unlearned learning data 13 stored in the learning data storage unit 52.

[0067] If the learning termination condition is not met in step S150, the machine learning unit 501 performs the processes of steps S120 to S140 multiple times on the learning model 12 under training, using unlearned learning data 13. On the other hand, if the learning termination condition is met in step S150, the machine learning unit 501 stores the generated trained learning model 12 (adjusted weight parameter group) in the trained model storage unit 53 in step S160, and temporarily ends this routine.

[0068] In the machine learning method, step S100 corresponds to a learning data storage step, steps S110 to S150 correspond to a machine learning step, and step S160 corresponds to a trained model storage step.

[0069] 9 is a block diagram showing an example of a molding condition deriving device according to the first embodiment. The molding condition deriving device 6 derives target molding conditions. The target molding conditions are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body 4 from a preform 3 by blow molding. The molding condition deriving device 6 includes a communication unit 61, a control unit 60, and a trained model storage unit 62.

[0070] The communication unit 61 is connected to an external device via the network 7, and functions as a communication interface for transmitting and receiving various data. The external device is, for example, the blow molding device 2 and the operator terminal device 8.

[0071] The control unit 60 functions as an information acquisition unit 600 , an inference unit 601 , and an output processing unit 602 .

[0072] The information acquisition unit 600 is connected to an external device via the communication unit 61 and the network 7, and executes an information acquisition process to acquire input data from the external device. Specifically, the information acquisition unit 600 acquires bottle shape information and preform shape information as input data from the blow molding device 2 or the operator terminal device 8.

[0073] The inference unit 601 inputs the bottle shape information and preform shape information acquired by the information acquisition unit 600 as input data to the learning model 12. The inference unit 601 executes inference processing using the learning model 12 stored in the trained model storage unit 62.

[0074] The trained model storage unit 62 is a database that stores trained learning models 12 used in the inference unit 601. The number of training models 12 stored in the trained model storage unit 62 is not limited to one, and multiple trained models with different conditions, such as machine learning techniques, bottle shapes, and preform shapes, may be stored and selectively used.

[0075] The trained model storage unit 62 may be replaced by a storage unit of an external computer, in which case the inference unit 601 simply accesses the external computer. Examples of external computers include server-based computers and cloud-based computers.

[0076] The output processing unit 602 executes an output process to output to an external device output data including the target molding conditions inferred by the inference unit 601. For example, the output processing unit 602 may transmit the generated target molding condition information to the blow molding apparatus 2 or to the operator terminal device 8.

[0077] In this way, the control unit 60 obtains output data consisting of target molding condition information based on input data consisting of bottle shape information and preform shape information.

[0078] Fig. 10 is a functional explanatory diagram showing an example of the functions of the molding condition derivation device 6 of Fig. 9. In the molding condition derivation device 6, an information acquisition unit 600 acquires bottle shape information and preform shape information as input data and inputs them to an inference unit 601. The inference unit 601 inputs the bottle shape information and preform shape information to a learning model 12, and generates target molding condition information as an inference result by the learning model 12.

[0079] As input data, for example, bottle shape information and preform shape information are provided to the information acquisition unit 600. The bottle shape information is information relating to the shape of the molded body 4. The preform shape information is information relating to the shape of the preform 3. As described above, when input data is input, the molding condition derivation device 6 determines target molding conditions using a learning model in which the correlation between the input data and the output data has been learned by machine learning, and outputs information relating to the determined target molding conditions as output data.

[0080] 11 is a hardware configuration diagram showing an example of a computer. The control unit 23, the machine learning device 5, and the molding condition derivation device 6 are configured by a general-purpose or dedicated computer 900.

[0081] 11, the computer 900 includes, as its main components, a bus 910, a processor 912, a memory 914, an input device 916, an output device 917, a display device 918, a storage device 920, a communication I / F (interface) unit 922, an external device I / F unit 924, an I / O (input / output) device I / F unit 926, and a media input / output unit 928. Note that the above components may be omitted as appropriate depending on the application for which the computer 900 is used.

[0082] The processor 912 is composed of one or more arithmetic processing devices (such as a central processing unit (CPU), a micro processing unit (MPU), a digital signal processor (DSP), or a graphics processing unit (GPU)), and operates as a control unit that controls the entire computer 900. The memory 914 stores various data and programs 930, and is composed of, for example, a volatile memory (such as a DRAM or SRAM) that functions as a main memory, a non-volatile memory (ROM), a flash memory, etc.

[0083] The input device 916 is configured, for example, by a keyboard, a mouse, a numeric keypad, an electronic pen, etc., and functions as an input unit. The output device 917 is configured, for example, by a sound (audio) output device, a vibration device, etc., and functions as an output unit. The display device 918 is configured, for example, by a liquid crystal display, an organic EL display, electronic paper, a projector, etc., and functions as an output unit. The input device 916 and the display device 918 may be configured integrally, such as a touch panel display. The storage device 920 is configured, for example, by an HDD, an SSD (Solid State Drive), etc., and functions as a storage unit. The storage device 920 stores various data necessary for executing the operating system and the program 930.

[0084] The communication I / F unit 922 is connected to a network 940 (which may be the same as network 7 in FIG. 6 ) such as the Internet or an intranet via a wired or wireless connection, and functions as a communication unit that transmits and receives data to and from other computers in accordance with a predetermined communication protocol. The external device I / F unit 924 is connected to an external device 950 (such as a camera, printer, scanner, or reader / writer) via a wired or wireless connection, and functions as a communication unit that transmits and receives data to and from the external device 950 in accordance with a predetermined communication protocol. The I / O device I / F unit 926 is connected to I / O devices 960 (such as various sensors and actuators), and functions as a communication unit that transmits and receives various signals and data, such as detection signals from sensors and control signals to actuators, to and from the I / O devices 960. The media input / output unit 928 is configured by a drive device such as a DVD drive or CD drive, and reads and writes data from and to media (non-transitory storage media) 970 (such as DVDs and CDs).

[0085] In the computer 900 having the above configuration, the processor 912 loads a program 930 stored in the storage device 920 into the memory 914, executes the program, and controls each unit of the computer 900 via the bus 910. The program 930 may be stored in the memory 914 instead of the storage device 920. The program 930 may be recorded on the medium 970 in an installable file format or an executable file format and provided to the computer 900 via the media input / output unit 928. The program 930 may be provided to the computer 900 by being downloaded via the network 940 via the communication I / F unit 922. Furthermore, the computer 900 may implement various functions realized by the processor 912 executing the program 930 using hardware such as an FPGA or an ASIC.

[0086] The computer 900 is, for example, a desktop computer or a portable computer, and is an electronic device of any type. The computer 900 may be a client computer, a server computer, or a cloud computer. The computer 900 may be applied to devices other than the control unit 23, the machine learning device 5, and the molding condition derivation device 6.

[0087] In the above embodiment, the input data is composed of bottle shape information and preform shape information, but the input data may be composed of bottle shape information only. In other words, in this case, the machine learning unit 501 only needs to make the learning model 12 learn the correlation between the bottle shape information and the target molding condition information.

[0088] In this way, the molding condition derivation device 6 according to the first embodiment is a device that derives target molding conditions. The target molding conditions are target values ​​of one or more blow molding conditions that are set in the blow molding process for obtaining a molded body 4 from a preform 3 by blow molding. The molding condition derivation device 6 determines output data consisting of target molding condition information based on input data consisting of bottle shape information. The bottle shape information is information about the shape of the molded body 4. The target molding condition information is information about the target molding conditions.

[0089] This allows target molding conditions to be derived based on shape information about the molded body 4. That is, appropriate blow molding conditions can be easily derived. Furthermore, even if the type of molded body 4 is changed, for example, the target molding conditions can be determined without relying on the skills of a skilled worker.

[0090] The input data further includes preform shape information, which is information relating to the shape of the preform 3.

[0091] This makes it possible to derive target molding conditions based on the shape information about the molded body 4 and the shape information about the preform 3. In other words, more appropriate blow molding conditions can be easily derived.

[0092] The bottle shape information includes at least one of design drawing information of the molded body 4, image information of the molded body 4, and design parameter information of the molded body 4.

[0093] The design drawing information of the molded body 4, the image information of the molded body 4, and the design parameter information of the molded body 4 are all information that is easily available to the user of the molding condition deriving device 6. Therefore, this makes it possible to more easily derive appropriate blow molding conditions.

[0094] The preform shape information includes at least one of design drawing information of the preform 3, image information of the preform 3, and design parameter information of the preform 3.

[0095] The design drawing information of the preform 3, the image information of the preform 3, and the design parameter information of the preform 3 are all information that is easily available to the user of the molding condition deriving device 6. Therefore, this makes it possible to more easily derive appropriate blow molding conditions.

[0096] The blow molding process includes a heating process, a pre-blow process, and a main-blow process. The heating process is a process of heating the preform 3. The pre-blow process is a process of introducing a pre-blow fluid into the heated preform 3. The main-blow process is a process of introducing a main-blow fluid into the preform 3 following the pre-blow process.

[0097] The one or more blow molding conditions include at least one of the heating temperature conditions of the preform 3 in the heating process, the temperature control conditions of the preform 3 in the heating process, the pre-blow conditions which are the set conditions in the pre-blow process, and the main blow conditions which are the set conditions in the main blow process.

[0098] The shape of the molded body 4 is highly dependent on the heating temperature conditions of the preform 3, the temperature control conditions of the preform 3, the pre-blow conditions, and the main blow conditions. Therefore, by setting at least one of these parameters as a molding condition and adjusting at least one parameter, the shape of the molded body 4 can be made more appropriate without relying on the skills of a skilled worker.

[0099] In addition, when input data is input, the molding condition derivation device 6 determines target molding conditions using a learning model 12 in which the correlation between the input data and the output data is learned by machine learning, and outputs information about the determined target molding conditions as output data.

[0100] According to this, the trained learning model 12 can easily derive the appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0101] The machine learning device 5 also generates a learning model 12 for inferring target molding conditions. The target molding conditions are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body 4 from a preform 3 by blow molding. The machine learning device 5 includes a learning data storage unit 52, a machine learning unit 501, and a trained model storage unit 53. The learning data storage unit 52 stores multiple sets of learning data, each set including input data consisting of bottle shape information and output data associated with the input data and consisting of target molding condition information. The machine learning unit 501 receives multiple sets of learning data and causes the learning model 12 to learn the correlation between the input data and the output data. The trained model storage unit 53 stores the learning model 12 trained by the machine learning unit 501.

[0102] This makes it possible to easily generate a learning model 12 for inferring appropriate blow molding conditions.

[0103] Furthermore, in the machine learning device 5, the input data further includes preform shape information.

[0104] This makes it easier to generate the learning model 12 for inferring appropriate blow molding conditions.

[0105] The inference device also includes a memory 914 and at least one processor 912, and infers target molding conditions. The target molding conditions are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body 4 from a preform 3 by blow molding. The at least one processor 912 executes information acquisition processing, inference processing, and output processing. The information acquisition processing is processing for acquiring input data consisting of bottle shape information, which is information about the shape of the molded body 4. The inference processing is processing for inferring target molding conditions using a learning model 12 based on machine learning stored in the memory 914, once the input data has been acquired. The output processing is processing for outputting output data including the inferred target molding conditions.

[0106] This makes it possible to easily deduce the appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0107] In the inference device, the input data further includes preform shape information.

[0108] This makes it easier to determine the appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0109] The information processing method is a method for deriving target molding conditions. The target molding conditions are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body 4 from a preform 3 by blow molding. The information processing method determines output data consisting of target molding condition information, which is information related to the target molding conditions, based on input data consisting of bottle shape information.

[0110] This makes it possible to easily derive the appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0111] In the information processing method, the input data further includes preform shape information.

[0112] This makes it possible to more easily derive the appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0113] The machine learning method is a method for generating a learning model for inferring target molding conditions. The machine learning method executes a learning data storage step, a machine learning step, and a learned model storage step. The learning data storage step is a step of storing multiple sets of learning data each consisting of input data consisting of bottle shape information and output data consisting of target molding condition information. The machine learning step is a step of inputting multiple sets of learning data and having the learning model learn the correlation between the input data and the output data. The learned model storage step is a step of storing the learning model learned in the machine learning step in a learned model storage unit.

[0114] This makes it possible to easily generate a learning model 12 for inferring appropriate blow molding conditions.

[0115] In the machine learning method, the input data further includes preform shape information.

[0116] This makes it easier to generate the learning model 12 for inferring appropriate blow molding conditions.

[0117] The inference method is a method executed by an inference device to infer target molding conditions. The inference device includes a memory 914 and at least one processor 912. The target molding conditions are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body 4 from a preform 3 by blow molding. The at least one processor 912 executes an information acquisition process, an inference process, and an output process. The information acquisition process is a process of acquiring input data consisting of bottle shape information. The inference process is a process of inferring target molding conditions using a learning model based on machine learning stored in the memory 914 upon acquiring the input data. The output process is a process of outputting output data including the inferred target molding conditions.

[0118] This makes it possible to easily deduce the appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0119] In the inference method, the input data further includes preform shape information.

[0120] This makes it easier to determine the appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0121] (Second embodiment) In the first embodiment, the molding condition derivation device 6 stores the learning model 12 created using the machine learning device 5 in the learned model storage unit 62, and infers the target molding conditions of the bottle using the stored learning model 12. The second embodiment differs from the first embodiment in that, instead of the learning model 12 based on machine learning as in the first embodiment, the molding condition derivation device determines the target molding conditions using a mathematical optimization model consisting of a predetermined calculation algorithm.

[0122] In the molding condition derivation device according to the second embodiment, input data is input to a mathematical optimization model, and output data is output from the mathematical optimization model. In other words, in the molding condition derivation device 6, the mathematical optimization model is stored in the memory 914 instead of the learning model 12. The processor 912 calculates the target molding conditions using the mathematical optimization model stored in the memory 914.

[0123] 12 is a block diagram showing an example of a molding condition deriving device according to the second embodiment. The molding condition deriving device 6 includes a communication unit 61, an existing bottle information storage unit 63, and a control unit 60. The communication unit 61 is connected to an external device via a network 7 and functions as a communication interface for transmitting and receiving various data. The external device is, for example, the blow molding device 2 and an operator terminal device 8.

[0124] The existing bottle information storage unit 63 is a database that stores preform shape information about molded bodies 4 manufactured in the past as existing preform shape information, and also stores bottle shape information about molded bodies 4 manufactured in the past as existing bottle shape information.

[0125] Furthermore, the existing bottle information storage unit 63 stores existing preform shape information for one type of molded body 4 and existing bottle shape information corresponding to the existing preform shape information, linking them to each other. Note that the existing bottle information storage unit 63 stores the variable values ​​in the existing preform shape information and the variable values ​​in the existing bottle shape information in the form of vector data. Note that in this embodiment, the values ​​of each variable are stored in the form of vector data, but the values ​​of each variable may also be stored as a relational database.

[0126] The control unit 60 functions as an information acquisition unit 600 , an inference unit 601 , and an output processing unit 602 .

[0127] The information acquisition unit 600 is connected to an external device via the communication unit 61 and the network 7, and executes an information acquisition process to acquire input data from the external device. Specifically, the information acquisition unit 600 acquires bottle shape information and preform shape information as input data from the blow molding device 2 or the operator terminal device 8.

[0128] The inference unit 601 has a mathematical optimization model 14 as a functional block. The inference unit 601 inputs the bottle shape information and preform shape information acquired by the information acquisition unit 600 to the mathematical optimization model 14. The inference unit 601 also inputs the existing bottle information stored in the existing bottle information storage unit 63 to the mathematical optimization model 14. The inference unit 601 uses the mathematical optimization model 14 to calculate target molding conditions.

[0129] The output processing unit 602 executes an output process for outputting to an external device output data including the target molding conditions calculated by the inference unit 601. For example, the output processing unit 602 may transmit the generated target molding condition information to the blow molding apparatus 2 or to the operator terminal device 8.

[0130] In this way, the control unit 60 obtains output data consisting of target molding condition information based on input data consisting of bottle shape information and preform shape information. When the input data is input, the molding condition derivation device 6 determines target molding conditions using a mathematical optimization model consisting of a predetermined calculation algorithm, and outputs information related to the determined target molding conditions as output data.

[0131] Fig. 13 is a functional explanatory diagram showing an example of the functions of the molding condition derivation device 6 of Fig. 12. In the molding condition derivation device 6, an information acquisition unit 600 acquires bottle shape information and preform shape information as input data and inputs them to an inference unit 601. The inference unit 601 inputs the bottle shape information and preform shape information to a mathematical optimization model 14, and generates target molding condition information as a calculation result by the mathematical optimization model 14.

[0132] 14 is a flowchart for explaining the algorithm of calculation by the mathematical optimization model 14. The mathematical optimization model 14 uses, for example, the following molding condition derivation algorithm.

[0133] In step S101, the inference unit 601 acquires bottle shape information and preform shape information as input data from the information acquisition unit 600. The input data may be extracted based on the bottle type, for example, the type of contents such as a beverage bottle or a detergent bottle, or the volume type such as 2 liters, 1 liter, or 500 milliliters. That is, data extracted based on a rule base may also be used as the input data. In this case, a step of extracting input data based on a rule base may be added.

[0134] Next, in step S102, the inference unit 601 converts the values ​​of the input bottle variables and the input preform variables into vector data. The input bottle variables are each variable of the bottle shape information of the input data. The input preform variables are each variable of the preform shape information of the input data.

[0135] Next, in step S103, the inference unit 601 calculates the vector distance between each value of the input bottle variable and each value of the existing bottle variable corresponding to each value of the input bottle variable. The existing bottle variables are each variable of the existing bottle shape information. In this embodiment, the vector distance is calculated as the Manhattan distance. The Manhattan distance is the distance between two points, calculated as the sum of the absolute values ​​of the differences between each coordinate. Furthermore, the weight in calculating the vector distance can be adjusted appropriately based on the type of the input bottle variable and the existing bottle variable. Note that the vector distance may be calculated as the Euclidean distance instead of the Manhattan distance.

[0136] Furthermore, the inference unit 601 calculates a first vector distance from the vector distance for each variable of the bottle shape information. The first vector distance is calculated as Manhattan distance. Alternatively, the first vector distance may be calculated as Euclidean distance. For example, the first vector distance is calculated as the square root of the sum of squares of the vector distance for each variable of the bottle shape information. That is, if the input bottle variables are bottle type, strength characteristics, measurement characteristics, and material characteristics, the inference unit 601 calculates the square root of the sum of squares of the four variables as the first vector distance. Also, for example, if existing bottle shape information for 10 types of bottles is stored in the existing bottle information storage unit 63, the inference unit 601 calculates 10 different first vector distances. That is, the first vector distance is the distance between the vector data of the input bottle variable and the vector data of the existing bottle variable corresponding to the input bottle variable. Also, the first vector distance may be a combination of the vector distances for each variable. For example, when the vector data is [body diameter|stretching ratio|···], the first vector distance may be set to the vector distance of the body diameter+the vector distance of the stretching ratio+···.

[0137] Next, in step S104, the inference unit 601 calculates the vector distance between each value of the input preform variable and each value of the existing preform variable corresponding to each value of the input preform variable. The existing preform variables are each variable of the existing preform shape information.

[0138] Furthermore, the inference unit 601 calculates a second vector distance from the vector distance for each variable of the preform shape information. The second vector distance is calculated as Manhattan distance. Alternatively, the second vector distance may be calculated as Euclidean distance. For example, the second vector distance is calculated as the square root of the sum of squares of the vector distance for each variable of the preform shape information. That is, if the input preform variables are preform type, measurement characteristics, and material characteristics, the inference unit 601 calculates the square root of the sum of squares of the three variables as the second vector distance. Also, for example, if existing preform shape information for five types of preforms is stored in the existing bottle information storage unit 63, the inference unit 601 calculates five different second vector distances. That is, the second vector distance is the distance between the vector data of the input preform variable and the vector data of the existing preform variable corresponding to the input preform variable. Also, the second vector distance may be a combination of the vector distances for each variable. For example, when the vector data is [body diameter|transparency|···], the second vector distance may be set to the vector distance of the body diameter+the vector distance of the transparency+···.

[0139] Next, in step S105, the inference unit 601 extracts existing bottle shape information corresponding to the shortest first vector distance from among the ten first vector distances. The inference unit 601 also extracts existing preform shape information corresponding to the shortest second vector distance from among the five second vector distances. In this embodiment, the existing bottle shape information and existing preform shape information corresponding to the shortest first vector distance and the shortest second vector distance are extracted. However, for example, cosine similarity may be used to extract existing bottle shape information and existing preform shape information. That is, the similarity between each variable of the bottle shape information and each variable of the existing bottle shape information, and the similarity between each variable of the preform shape information and each variable of the existing preform shape information may be calculated. The closer the distance between two vectors, the higher the similarity and the larger the cosine similarity value. Therefore, existing bottle shape information corresponding to the first vector distance with the largest cosine similarity from among multiple first vector distances may be extracted. Furthermore, existing preform shape information corresponding to the second vector distance with the largest cosine similarity from among multiple second vector distances may be extracted. In this way, in step S105, the optimal existing bottle shape information is identified based on the first vector distance, and the optimal existing preform shape information is identified based on the second vector distance. Furthermore, for example, the final existing bottle shape information may be identified by taking a weighted average of the existing bottle shape information corresponding to each of the ten first vector distances. In this case, the existing bottle shape information may be identified based on a combination of the ten first vector distances. Furthermore, for example, the final existing preform shape information may be identified by taking a weighted average of the existing preform shape information corresponding to each of the five second vector distances. In this case, the existing preform shape information may be identified based on a combination of the five second vector distances. In other words, the existing bottle shape information and the existing preform shape information may be identified based on a combination of multiple vectors. In the above example, a weighted average is taken to identify the final existing bottle shape information and the existing preform shape information, but a suitable arithmetic operation may also be used.

[0140] Next, in step S106, the inference unit 601 extracts bottles corresponding to the extracted existing bottle shape information as similar bottles. The inference unit 601 also extracts preforms corresponding to the extracted existing preform shape information as similar preforms. When extracting similar bottles from existing bottle shape information extracted by combining multiple first vector distances, the bottle most similar to the existing bottle shape information is extracted as the similar bottle. When extracting similar preforms from existing preform shape information extracted by combining multiple second vector distances, the preform most similar to the existing preform shape information is extracted as the similar preform.

[0141] Next, in step S107, the inference unit 601 acquires molding condition information corresponding to the similar bottle as the first molding condition information from the existing bottle information storage unit 63. The inference unit 601 also acquires molding condition information corresponding to the similar preform as the second molding condition information from the existing bottle information storage unit 63. The first molding condition information can also be acquired by combining multiple first vector distances. In this case, existing bottle shape information corresponding to each first vector distance can be extracted, similar bottles corresponding to each extracted existing bottle shape information can be extracted, and a weighted average of the molding condition information corresponding to each extracted similar bottle can be obtained as the first molding condition information. The second molding condition information can also be acquired by combining multiple second vector distances. In this case, existing preform shape information corresponding to each second vector distance can be extracted, similar preforms corresponding to each extracted existing preform shape information can be extracted, and a weighted average of the molding condition information corresponding to each extracted similar preform can be obtained as the second molding condition information. In the above example, a weighted average is taken to obtain the final first molding condition information and second molding condition information, but any suitable arithmetic operation may be used.

[0142] Next, in step S108, the inference unit 601 evaluates each molding condition included in the first molding condition information and each molding condition included in the second molding condition information based on a specific index. Then, the inference unit 601 selects molding condition information including more appropriate molding conditions from either the first molding condition information or the second molding condition information as target molding condition information.

[0143] Next, in step S109, the inference unit 601 outputs the selected molding condition information to the output processing unit 602 as the calculation result.

[0144] In the above embodiment, the input data is composed of bottle shape information and preform shape information, but the input data may be composed of bottle shape information only. In other words, in this case, the mathematical optimization model 14 only needs to calculate the target molding conditions for the molded body 4 from the bottle shape information and existing bottle shape information.

[0145] In this way, when input data is input, the molding condition derivation device 6 of the second embodiment determines target molding conditions using a mathematical optimization model 14 consisting of a predetermined calculation algorithm, and outputs information about the determined target molding conditions as output data.

[0146] According to this, the mathematical optimization model 14 can easily derive the appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0147] Further, the predetermined calculation algorithm calculates a first vector distance. The first vector distance is the distance between the vector data of the input bottle variable and the vector data of the existing bottle variable corresponding to the input bottle variable. The input bottle variable is each variable of the bottle shape information of the input data. The existing bottle variable is each variable of the existing bottle shape information. The predetermined calculation algorithm identifies the existing bottle shape information based on the first vector distance, and extracts first molding condition information based on molding condition information linked to the existing bottle shape information. The predetermined calculation algorithm selects the first molding condition information as target molding condition information.

[0148] According to this, by inputting bottle shape information of the molded body 4 into the molding condition deriving device 6, it is possible to easily derive appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0149] The predetermined calculation algorithm also calculates a first vector distance. The predetermined calculation algorithm identifies existing bottle shape information based on the first vector distance, and extracts first molding condition information based on molding condition information linked to the existing bottle shape information.

[0150] The predetermined calculation algorithm further calculates a second vector distance. The second vector distance is the distance between the vector data of the input preform variables and the vector data of the existing preform variables corresponding to the input preform variables. The input preform variables are each variable of the preform shape information of the input data. The existing preform variables are each variable of the existing preform shape information.

[0151] The predetermined calculation algorithm identifies the existing preform shape information based on the second vector distance, and extracts the second molding condition information based on the molding condition information linked to the existing preform shape information. The predetermined calculation algorithm selects either the first molding condition information or the second molding condition information as the target molding condition information based on a specific index.

[0152] According to this, by inputting bottle shape information of the molded body 4 and preform shape information of the preform 3 into the molding condition derivation device 6, it is possible to easily derive appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0153] The inference device also includes a memory 914 and at least one processor 912, and infers target molding conditions. The target molding conditions are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body 4 from a preform 3 by blow molding. The at least one processor 912 executes information acquisition processing, inference processing, and output processing. The information acquisition processing is processing for acquiring input data consisting of bottle shape information, which is information about the shape of the molded body 4. The inference processing is processing for inferring target molding conditions using a mathematical optimization model 14 stored in memory upon acquiring the input data. The output processing is processing for outputting output data including the inferred target molding conditions.

[0154] This makes it possible to easily deduce the appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0155] In the inference device, the input data further includes preform shape information.

[0156] This makes it easier to determine the appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0157] The inference method is a method executed by an inference device to infer target molding conditions. The inference device includes a memory 914 and at least one processor 912. The target molding conditions are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body 4 from a preform 3 by blow molding. The at least one processor 912 executes an information acquisition step, an inference step, and an output step. The information acquisition step is a step of acquiring input data consisting of bottle shape information. The inference step is a step of inferring target molding conditions using a mathematical optimization model 14 stored in the memory 914 upon acquiring the input data. The output step is a step of outputting output data including the inferred target molding conditions.

[0158] This makes it possible to easily deduce the appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0159] In the inference method, the input data further includes preform shape information.

[0160] This makes it easier to determine the appropriate blow molding conditions for obtaining the molded body 4 from the preform 3.

[0161] In this embodiment, the vector distance is calculated as the Manhattan distance, but the vector distance may be calculated as any one of the Euclidean distance, the standardized Euclidean distance, the Chebyshev distance, the Minkowski distance, the Mahalanobis distance, and the Hellinger distance.

[0162] Furthermore, depending on the variables, instead of calculating the vector distance, as described above, the similarity between each variable of the bottle shape information and each variable of the existing bottle shape information, and the similarity between each variable of the preform shape information and each variable of the existing preform shape information may be calculated from the angle formed by two vectors. In addition to cosine similarity, for example, deviation pattern similarity and Pearson's correlation coefficient can also be used to calculate the similarity.

[0163] Furthermore, the inference unit 601 may evaluate each molding condition included in the first molding condition information and each molding condition included in the second molding condition information, and select the more appropriate molding condition for each molding condition.

[0164] Furthermore, the inference unit 601 may extract molding condition information using only the bottle shape information without using the preform shape information. In this case, the processes of step S104 and step S108 in Fig. 14 are unnecessary. That is, in this case, the mathematical optimization model 14 selects the first molding condition information.

[0165] (Other embodiments) The present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit and scope of the present invention, all of which are included in the technical concept of the present invention.

[0166] For example, a device for heating the stretch rod 222 may be provided outside or inside the stretch rod 222 to heat the stretch rod 222. In this case, the target molding condition information may include at least one of information on the temperature condition of the stretch rod 222 and information on the temperature control condition of the stretch rod 222.

[0167] Furthermore, since it is not essential to monitor the temperature of the preforms 3 in each molding unit 22, each molding unit 22 does not necessarily need to be provided with a temperature sensor 2247.

[0168] Furthermore, in the above embodiment, the pre-blow conditions include the stretch amount, but instead of the stretch amount, the amount by which the stretch rods 222 advance may also be included.

[0169] Furthermore, in the first embodiment, a case has been described in which a neural network is employed as the learning model 12 that realizes machine learning by the machine learning unit 501, but other machine learning models may also be employed. Examples of other machine learning models include tree types such as decision trees and regression trees, ensemble learning such as bagging and boosting, neural network types (including deep learning) such as recurrent neural networks, convolutional neural networks, and LSTM (Long Short Term Memory), clustering types such as hierarchical clustering, non-hierarchical clustering, k-nearest neighbors, and k-means, multivariate analyses such as principal component analysis, factor analysis, and logistic regression, and support vector machines.

[0170] The present invention can also be provided in the form of a program (machine learning program) that causes the computer 900 to function as each part of the machine learning device 5, or a program (machine learning program) that causes the computer 900 to execute each step of the machine learning method.

[0171] The present invention can also be provided in the form of a program (inference program) that causes a computer 900 to function as each unit included in the molding condition derivation device 6. In this case, the molding condition derivation device (inference program) includes a memory 914 and a processor 912, and the processor 912 can execute a series of processes.

[0172] This series of processes includes an information acquisition process (information acquisition step) for acquiring bottle shape information, which is information relating to one or more bottle shapes, and preform shape information, which is information relating to one or more preform shapes, and an inference process (inference step) for inferring target molding conditions using a learning model stored in memory 914 once the bottle shape information and preform shape information have been acquired by the information acquisition process.

[0173] By providing it in the form of a molding condition derivation device (inference method or inference program), it can be easily applied to various devices. It is naturally understandable to those skilled in the art that when the molding condition derivation device (inference method or inference program) infers the target molding conditions, the inference method implemented by the inference unit 601 may be applied using the machine learning device 5 and the trained learning model 12 generated by the machine learning method according to the above embodiment.

[0174] The above describes in detail preferred embodiments, but the present invention is not limited to the above-described embodiments, and various modifications and substitutions can be made to the above-described embodiments without departing from the scope of the claims.

[0175] Various aspects of the present invention will be summarized below as appendices.

[0176] (Appendix 1) A molding condition deriving device that derives target molding conditions that are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding, the molding condition derivation device determines output data consisting of target molding condition information, which is information about the target molding conditions, based on input data consisting of bottle shape information, which is information about the shape of the molded body; When the input data is input, the target molding conditions are determined using a learning model in which a correlation between the input data and the output data is learned by machine learning, and information about the determined target molding conditions is output as the output data; The bottle shape information includes the capacity Molding condition derivation device. (Appendix 2) A molding condition deriving device that derives target molding conditions that are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding, the molding condition derivation device determines output data consisting of target molding condition information, which is information about the target molding conditions, based on input data consisting of bottle shape information, which is information about the shape of the molded body; When the input data is input, the target molding conditions are determined using a learning model in which a correlation between the input data and the output data is learned by machine learning, and information about the determined target molding conditions is output as the output data; the input data further includes preform shape information which is information regarding the shape of the preform, The preform shape information includes material properties and metrology properties. Molding condition derivation device. (Appendix 3) In the blow molding step, Heating the preform with a heating device; The heated preform is fed into a blow molding apparatus; The preform is stretched by advancing the stretch rod along the central axis of the preform, and a blowing fluid is introduced into the preform to form it into a molded body. 10. The molding condition deriving device according to claim 1 or 2. (Appendix 4) The bottle shape information used as the input data further includes at least one of design drawing information of the molded body, image information of the molded body, and design parameter information of the molded body. 4. A molding condition deriving device according to any one of appendix 1 to appendix 3. (Appendix 5) the blow molding process includes a heating process of heating the preform, a pre-blow process of introducing a pre-blow fluid into the heated preform, and a main blow process of introducing a main blow fluid into the preform, The one or more blow molding conditions are a heater temperature in a heating device in the heating step, a pre-blow pressure in the pre-blow step, and a main blow pressure in the main blow step. 5. A molding condition deriving device according to any one of appendix 1 to appendix 4. (Appendix 6) the blow molding process includes a pre-blow process which is a process of introducing a pre-blow fluid into the heated preform, and a main-blow process which is a process of introducing a main-blow fluid into the preform, The one or more blow molding conditions further include a pre-blow start timing in the pre-blow process and a main-blow start timing in the main-blow process. 6. A molding condition deriving device according to any one of appendixes 1 to 5. (Appendix 7) the blow molding step includes a pre-blowing step of introducing a pre-blowing fluid into the heated preform; The one or more blow molding conditions further include a stretching speed by a stretch rod in the pre-blow process. 10. The molding condition deriving device according to claim 1, wherein the molding condition deriving device is a molding condition deriving device. (Appendix 8) The bottle shape information includes at least one of strength characteristics and measurement characteristics. 8. A molding condition deriving device according to any one of appendixes 1 to 7. (Appendix 9) A machine learning device that generates a learning model for inferring target molding conditions, which are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding, a learning data storage unit that stores a plurality of sets of learning data, each set including input data consisting of bottle shape information, which is information about the shape of the molded body, and output data, which is associated with the input data and consists of target molding condition information, which is information about the target molding conditions; a learning data acquisition unit that acquires the learning data; a machine learning unit that receives a plurality of sets of the learning data and causes a learning model to learn a correlation between the input data and the output data; a trained model storage unit that stores the trained model trained by the machine learning unit; Equipped with The bottle shape information includes the capacity Machine learning device. (Appendix 10) A machine learning device that generates a learning model for inferring target molding conditions, which are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding, a learning data storage unit that stores a plurality of sets of learning data, each set including input data consisting of bottle shape information, which is information about the shape of the molded body, and output data, which is associated with the input data and consists of target molding condition information, which is information about the target molding conditions; a learning data acquisition unit that acquires the learning data; a machine learning unit that receives a plurality of sets of the learning data and causes a learning model to learn a correlation between the input data and the output data; a trained model storage unit that stores the trained model trained by the machine learning unit; Equipped with the input data further includes preform shape information which is information regarding the shape of the preform, The preform shape information includes metrology and material properties. Machine learning device. (Appendix 11) In the blow molding step, Heating the preform with a heating device; The heated preform is fed into a blow molding apparatus; The preform is stretched by advancing the stretch rod along the central axis of the preform, and a blowing fluid is introduced into the preform to form it into a molded body. 11. The machine learning device according to claim 9 or 10. (Appendix 12) An information processing method for deriving target molding conditions, which are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding, comprising: Based on input data consisting of bottle shape information which is information about the shape of the molded body, output data consisting of target molding condition information which is information about the target molding conditions is obtained; The bottle shape information includes the capacity Information processing methods. (Appendix 13) An information processing method for deriving target molding conditions, which are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding, comprising: Based on input data consisting of bottle shape information which is information about the shape of the molded body, output data consisting of target molding condition information which is information about the target molding conditions is obtained; the input data further includes preform shape information which is information regarding the shape of the preform, The preform shape information includes metrology and material properties. Information processing methods. (Appendix 14) In the blow molding step, Heating the preform with a heating device; The heated preform is fed into a blow molding apparatus; The preform is stretched by advancing the stretch rod along the central axis of the preform, and a blowing fluid is introduced into the preform to form it into a molded body. 14. The information processing method according to claim 12 or 13. (Appendix 15) A program that causes a computer to execute the information processing method according to any one of appendixes 12 to 14. [Explanation of symbols]

[0177] 2 Blow molding device, 3 Preform, 4 Molded body, 5 Machine learning device, 6 Molding condition derivation device, 12 Learning model, 13 Learning data, 14 Mathematical optimization model, 52 Learning data storage unit, 53 Trained model storage unit, 501 Machine learning unit, 912 Processor, 914 Memory.

Claims

1. A molding condition deriving device that derives target molding conditions that are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding, the molding condition derivation device determines output data consisting of target molding condition information, which is information about the target molding conditions, based on input data consisting of bottle shape information, which is information about the shape of the molded body; When the input data is input, the target molding conditions are determined using a learning model in which a correlation between the input data and the output data is learned by machine learning, and information about the determined target molding conditions is output as the output data; The bottle shape information includes the capacity Molding condition derivation device.

2. A molding condition deriving device that derives target molding conditions that are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding, the molding condition derivation device determines output data consisting of target molding condition information, which is information about the target molding conditions, based on input data consisting of bottle shape information, which is information about the shape of the molded body; When the input data is input, the target molding conditions are determined using a learning model in which a correlation between the input data and the output data is learned by machine learning, and information about the determined target molding conditions is output as the output data; the input data further includes preform shape information which is information regarding the shape of the preform, The preform shape information includes material properties and metrology properties. Molding condition derivation device.

3. In the blow molding step, Heating the preform with a heating device; The heated preform is fed into a blow molding apparatus; The preform is stretched by advancing the stretch rod along the central axis of the preform, and a blowing fluid is introduced into the preform to form it into a molded body. The molding condition deriving device according to claim 1 or 2.

4. The bottle shape information used as the input data further includes at least one of design drawing information of the molded body, image information of the molded body, and design parameter information of the molded body. The molding condition deriving device according to claim 1 or 2.

5. the blow molding process includes a heating process of heating the preform, a pre-blow process of introducing a pre-blow fluid into the heated preform, and a main blow process of introducing a main blow fluid into the preform, The one or more blow molding conditions are a heater temperature in a heating device in the heating step, a pre-blow pressure in the pre-blow step, and a main blow pressure in the main blow step. The molding condition deriving device according to claim 1 or 2.

6. the blow molding process includes a pre-blow process which is a process of introducing a pre-blow fluid into the heated preform, and a main-blow process which is a process of introducing a main-blow fluid into the preform, The one or more blow molding conditions further include a pre-blow start timing in the pre-blow process and a main-blow start timing in the main-blow process. The molding condition deriving device according to claim 1 or 2.

7. the blow molding step includes a pre-blowing step of introducing a pre-blowing fluid into the heated preform; The one or more blow molding conditions further include a stretching speed by a stretch rod in the pre-blow process. The molding condition deriving device according to claim 1 or 2.

8. The bottle shape information includes at least one of strength characteristics and measurement characteristics. The molding condition deriving device according to claim 1 or 2.

9. A machine learning device that generates a learning model for inferring target molding conditions, which are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding, comprising: a learning data storage unit that stores a plurality of sets of learning data, each set including input data consisting of bottle shape information, which is information about the shape of the molded body, and output data, which is associated with the input data and consists of target molding condition information, which is information about the target molding conditions; a learning data acquisition unit that acquires the learning data; a machine learning unit that receives a plurality of sets of the learning data and causes a learning model to learn a correlation between the input data and the output data; a trained model storage unit that stores the trained model trained by the machine learning unit; Equipped with The bottle shape information includes the capacity Machine learning device.

10. A machine learning device that generates a learning model for inferring target molding conditions, which are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding, comprising: a learning data storage unit that stores a plurality of sets of learning data, each set including input data consisting of bottle shape information, which is information about the shape of the molded body, and output data, which is associated with the input data and consists of target molding condition information, which is information about the target molding conditions; a learning data acquisition unit that acquires the learning data; a machine learning unit that receives a plurality of sets of the learning data and causes a learning model to learn a correlation between the input data and the output data; a trained model storage unit that stores the trained model trained by the machine learning unit; Equipped with the input data further includes preform shape information which is information regarding the shape of the preform, The preform shape information includes metrology and material properties. Machine learning device.

11. In the blow molding step, Heating the preform with a heating device; The heated preform is fed into a blow molding apparatus; The preform is stretched by advancing the stretch rod along the central axis of the preform, and a blowing fluid is introduced into the preform to form it into a molded body. The machine learning device according to claim 9 or 10.

12. An information processing method for deriving target molding conditions, which are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding, comprising: Based on input data consisting of bottle shape information which is information about the shape of the molded body, output data consisting of target molding condition information which is information about the target molding conditions is obtained; The bottle shape information includes the capacity Information processing methods.

13. An information processing method for deriving target molding conditions, which are target values ​​of one or more blow molding conditions set in a blow molding process for obtaining a molded body from a preform by blow molding, comprising: Based on input data consisting of bottle shape information which is information about the shape of the molded body, output data consisting of target molding condition information which is information about the target molding conditions is obtained; the input data further includes preform shape information which is information regarding the shape of the preform, The preform shape information includes metrology and material properties. Information processing methods.

14. In the blow molding step, Heating the preform with a heating device; The heated preform is fed into a blow molding apparatus; The preform is stretched by advancing the stretch rod along the central axis of the preform, and a blowing fluid is introduced into the preform to form it into a molded body.

14. The information processing method according to claim 12 or 13.

15. A program for causing a computer to execute the information processing method according to claim 12 or 13.

Images