JP2024101309A - Information processing device, method for generating inference model, inference method, and inference program - Google Patents

Abstract

To manage molds used for molding resins with low decomposition temperatures properly.SOLUTION: An information processing device (2) has a data acquisition section (201) to acquire data for inference including the difference between the decomposition temperature at which molten resin thermally decomposes in injection molding and the temperature of the flow path leading the molten resin to the mold, and an inference section 202 that uses an inference model (211) generated by learning the relationship between the data for inference and predetermined matters related to the management of the mold for inference.SELECTED DRAWING: Figure 1

Description

The present invention relates to an information processing device that can be used to support injection molding.

Injection molding has been a widely used technology for a long time, and the development of technologies to support injection molding has also been progressing. For example, the following Patent Document 1 discloses a mold management system that estimates the degree of contamination of the mold used in molding based on product images taken of the injection molded product.

JP 2022-107954 A

The above-mentioned conventional technologies have room for improvement in terms of proper management of molds used to mold resins with low decomposition temperatures. For example, biodegradable plastics, which are decomposed by living organisms, have been attracting attention in recent years, but they have a lower decomposition temperature than general resins. For this reason, when biodegradable plastics are injection molded, contamination of the flow paths and molds due to thermal decomposition products and clogging of air vents that exhaust gases inside the mold are likely to occur, which can cause molding defects such as insufficient filling (short circuits) and burrs due to overfilling.

However, because minute changes such as clogged air vents are unlikely to be reflected in the appearance of an injection-molded product, the mold management system of Patent Document 1 cannot be said to be sufficient for managing molds used to mold resins with low decomposition temperatures. One aspect of the present invention aims to realize an information processing device or the like that enables appropriate management of molds used to mold resins with low decomposition temperatures.

In order to solve the above problems, an information processing device according to one aspect of the present invention includes a data acquisition unit that acquires inference data including at least one of the difference between the decomposition temperature at which the molten resin thermally decomposes during injection molding and the temperature of the flow path that guides the molten resin to the mold, and the time that the molten resin remains in the flow path, and an inference unit that infers the matters from the inference data acquired by the data acquisition unit using an inference model generated by learning the relationship between the inference data and predetermined matters related to the management of the mold.

In order to solve the above problems, a method for generating an inference model according to one aspect of the present invention is a method for generating an inference model executed by one or more information processing devices, and includes a data acquisition step of acquiring training data in which correct answer data for a predetermined matter related to mold management is associated with inference data including at least one of the difference between the decomposition temperature at which molten resin thermally decomposes in injection molding and the temperature of the flow path that guides the molten resin to the mold, and the time that the molten resin remains in the flow path, and a learning step of generating an inference model for inferring the matter from the inference data by machine learning using the training data.

In order to solve the above problems, an inference method according to one aspect of the present invention is an inference method executed by one or more information processing devices, and includes a data acquisition step of acquiring inference data including at least one of the difference between the decomposition temperature at which the molten resin thermally decomposes in injection molding and the temperature of the flow path that guides the molten resin to the mold, and the time that the molten resin remains in the flow path, and an inference step of inferring the items from the inference data acquired in the data acquisition step using an inference model generated by learning the relationship between the inference data and predetermined items related to the management of the mold.

According to one aspect of the present invention, it becomes possible to properly manage molds used for molding resins with low decomposition temperatures.

1 is a block diagram showing an example of a configuration of a main part of an information processing device according to a first embodiment of the present invention; 1 is a diagram showing an overview of an information processing system according to a first embodiment of the present invention. FIG. 2 is a diagram showing an example of the configuration of an injection molding machine and its flow paths. FIG. 11 is a diagram showing the relationship between the screw position and the injection pressure of the injection molding machine in the normal state and when the air vent is blocked. FIG. 1 is a diagram showing the relationship between the number of shots and the defective rate in injection molding. A flowchart showing an example of a method for generating an inference model relating to embodiment 1 of the present invention. 3 is a flowchart showing an example of an inference method according to the first embodiment of the present invention. FIG. 11 is a block diagram showing an example of a configuration of a main part of an information processing device according to a second embodiment of the 10 is a flowchart illustrating an example of processing executed by an information processing device according to a second embodiment of the present invention.

[Embodiment 1]
(System Overview)
An overview of the information processing system 5 according to this embodiment will be described with reference to Fig. 2. Fig. 2 is a diagram showing an overview of the information processing system 5. The information processing system 5 is a system that can be used to support injection molding, and includes an information processing device 1 and an information processing device 2. Below, the injection molding to be supported will be described, and then the information processing devices 1 and 2 will be described.

For injection molding, an injection molding machine including an injection unit, a mold, and a mold clamping unit is used. In injection molding, resin pellets are first fed from a hopper into a cylinder, and the resin in the cylinder is heated by a heater to become molten resin. Next, the screw is advanced by a motor, and the molten resin is injected into the mold. The injected molten resin is cooled and solidified inside the mold. After the molten resin has solidified, the crosshead is moved to open the mold along the tie bars, and the molded product is removed from the mold by an ejector mechanism.

The molded product is removed and checked for quality issues before being used as a finished product. Figure 2 shows an example in which the molded product is a spoon, and its quality is judged by an appearance inspection machine. The information processing system 5 can support any injection molding, and the configuration of the injection molding machine used for the injection molding and the molded product to be injection molded are not particularly limited, but the information processing system 5 is particularly suitable for molding resins with low decomposition temperatures. For example, the information processing system 5 can be suitably used for molding biodegradable plastics and the like. To give a specific example, the resin may be a polyhydroxyalkanoate resin (hereinafter referred to as a PHA resin) such as a poly(3-hydroxyalkanoate)-based resin or poly(3-hydroxybutyrate-co-3-hydroxyhexanoate). A PHA resin is an aliphatic polyester that has biodegradability.

The information processing device 1 generates an inference model through machine learning using training data, and the information processing device 2 uses the inference model to infer specific inference items related to the mold from the inference data.

As will be described in detail later, the training data is inference data including at least one of the difference between the decomposition temperature at which molten resin thermally decomposes during injection molding and the temperature of the flow path that guides the molten resin to the mold, and the time that the molten resin remains in the flow path, and is associated with correct answer data for specific matters related to mold management. This makes it possible to properly manage molds used to mold resins with low decomposition temperatures.

(Configuration of information processing device)
Fig. 1 is a block diagram showing an example of a configuration of a main part of information processing devices 1 and 2. Below, each component of the information processing devices 1 and 2 will be described in order with reference to Fig. 1.

As shown in the figure, the information processing device 1 includes a control unit 10 that controls each unit of the information processing device 1, and a storage unit 11 that stores various data used by the information processing device 1. The information processing device 1 also includes a communication unit 12 that enables the information processing device 1 to communicate with other devices, an input unit 13 that accepts input of various data to the information processing device 1, and an output unit 14 that enables the information processing device 1 to output various data. The control unit 10 also includes a training data acquisition unit 101 and a learning unit 102. And the storage unit 11 stores training data 111.

The training data acquisition unit 101 acquires training data 111 in which correct answer data for specific matters related to mold management is associated with inference data including at least one of the difference between the decomposition temperature at which the molten resin thermally decomposes during injection molding and the temperature of the flow path that guides the molten resin to the mold, and the time that the molten resin remains in the flow path. The inference data can be rephrased as explanatory variables in an inference model, and the above matters and the correct answer data can be rephrased as objective variables in the inference model.

The learning unit 102 generates an inference model for inferring the above items from the inference data through machine learning using the above training data. The machine learning algorithm is not particularly limited. For example, the learning unit 102 may generate the inference model through regression analysis or multiple regression analysis, or may generate an inference model such as a neural network model or a tree model.

The specified matters related to mold management may be matters related to mold management that are related to the inference data described above. For example, the inference model may predict the quality of a molded product molded by injection molding using a mold. In this case, the learning unit 102 may generate the inference model by learning using training data 111 that indicates the relationship between the inference data and the quality of a molded product molded by injection molding using a mold. The quality of a molded product can be expressed, for example, by the product defect rate or an evaluation value that evaluates the quality of the product. The inference model may also infer, for example, the degree of dirtiness of the mold, the quality of the manufacturing conditions, the need to clean the mold, the number of injections remaining until cleaning is required, or the next time to clean the mold.

On the other hand, the information processing device 2 includes a control unit 20 that controls each unit of the information processing device 2, and a storage unit 21 that stores various data used by the information processing device 2. The information processing device 2 also includes a communication unit 22 that allows the information processing device 2 to communicate with other devices, an input unit 23 that accepts input of various data to the information processing device 2, and an output unit 24 that allows the information processing device 2 to output various data. The control unit 20 also includes a data acquisition unit 201, an inference unit 202, and an output control unit 203. And the storage unit 11 stores an inference model 211.

The data acquisition unit 201 acquires inference data including at least one of the difference between the decomposition temperature at which the molten resin thermally decomposes during injection molding and the temperature of the flow path that guides the molten resin to the mold, and the time that the molten resin remains in the flow path.

The inference unit 202 infers the above items from the inference data acquired by the data acquisition unit 201 using an inference model 211 generated by learning the relationship between the inference data and specific items related to mold management.

The output control unit 203 causes the inference result of the inference unit 202 to be output to the output unit 24. The form of output is arbitrary, and for example, the output control unit 203 may cause the inference result to be output in at least one of the following forms: display output, audio output, and print output. In addition, the device that outputs the inference result may be a device external to the information processing device 2.

The output control unit 203 may also cause the inference unit 202 to output information according to the inference result. For example, the output control unit 203 may cause the inference unit 202 to output a notification encouraging cleaning of the mold when the inference result of the inference unit 202 indicates signs of an increase in the defect rate or when the degree of mold contamination indicated in the inference result exceeds a threshold value. When outputting such a notification, the output control unit 203 may also cause the predicted defect rate and the degree of mold contamination to be output together.

In addition, when the inference data includes condition data indicating the injection molding conditions (e.g., heating temperature, type of resin, physical properties, molding cycle, number of products molded simultaneously, etc.), the output control unit 203 may display information indicating the progress of the condition data and the inference results. For example, the output control unit 203 may display the progress of the condition data and the inference results as a graph. In this case, the output control unit 203 may also display a graph indicating the progress of ideal conditions under which the mold is less likely to be contaminated (e.g., conditions under which stable production was possible for a long period of time in the past). This allows the operator of the injection molding machine to use these graphs as indicators and perform appropriate operations.

The operator can also perform a simulation of injection molding by inputting various injection molding conditions and having the inference unit 202 perform inference. In this way, the information processing device 2 can also be used as an auxiliary tool for determining conditions that are less likely to cause contamination of the mold.

As described above, the information processing device 2 includes a data acquisition unit 201 that acquires inference data including at least one of the difference between the decomposition temperature at which the molten resin thermally decomposes during injection molding and the temperature of the flow path that guides the molten resin to the mold, and the time that the molten resin remains in the flow path, and an inference unit 202 that infers the above items from the inference data acquired by the data acquisition unit 201 using an inference model 211 generated by learning the relationship between the inference data and specific items related to mold management.

The smaller the difference between the decomposition temperature at which the molten resin thermally decomposes and the temperature of the flow path that leads the molten resin to the mold, the higher the possibility of the molten resin thermally decomposing. Thermal decomposition of the molten resin can cause contamination of the flow path and the mold with thermal decomposition products of the molten resin, and can cause clogging of air vents that exhaust gas from within the mold, resulting in molding defects such as insufficient filling (short circuits) and burrs due to overfilling.

In this regard, the above-mentioned configuration, which uses the difference between the decomposition temperature and the flow path temperature, allows for inferences that take into account the ease with which the resin is thermally decomposed, making it possible to appropriately manage molds used to mold resins with low decomposition temperatures.

In addition, the longer the molten resin remains in the flow path, the more likely it is that thermal decomposition will progress. For example, there are cases where the injection interval is extended due to the disposal of defective products, and in such cases the retention time becomes longer than usual, posing a risk of thermal decomposition progressing.

In this regard, the above-mentioned configuration, which performs inference using inference data including the time that the molten resin was retained in the flow path, allows inference to be performed taking into account the time that the molten resin was retained in the flow path, making it possible to appropriately manage the mold used to mold resins with low decomposition temperatures.

As described above, the inference model 211 may be generated by learning the relationship between the inference data and the quality of a molded product formed by injection molding using a mold. In this case, the inference unit 202 can predict the quality of a molded product formed by injection molding using the above-mentioned mold. This makes it possible to take measures such as adjusting the injection molding conditions or cleaning the mold at the appropriate time depending on the prediction result, making it possible to form a molded product of the desired quality.

(Regarding inference data)
As described above, the inference data includes at least one of the difference between the decomposition temperature at which the molten resin thermally decomposes during injection molding and the temperature of the flow path that guides the molten resin to the mold, and the time that the molten resin stays in the flow path. The inference data may also include data other than these data. Here, data that can be included in the inference data will be described with reference to Fig. 3. Fig. 3 is a diagram showing an example of the configuration of an injection molding machine and its flow paths.

A1 in Figure 3 shows an injection molding machine similar to that in Figure 2. As shown in the figure, the cylinder is provided with heaters h1 to h3 that heat the resin inside it. In addition, a heater h4 is also provided in the flow path that guides the molten resin injected from the injection port e at the end of the cylinder to the cavity c in the mold. This flow path can also be called a hot runner. The rear stage of the cylinder (near the injection port e) can also be considered part of this flow path.

The temperature of the flow path can be measured by installing a temperature sensor near heater h4. Therefore, by using the measurement value of the temperature sensor, it is possible to calculate the difference between the decomposition temperature at which the molten resin thermally decomposes and the temperature of the flow path that leads the molten resin to the mold. In addition, since the measurement value of the temperature sensor changes significantly depending on whether or not molten resin is flowing in the flow path, the measurement value of the temperature sensor can also be used to determine the time that the molten resin remains in the flow path.

The above-mentioned flow paths may be branched. A2 in FIG. 3 shows an example of a branched flow path configuration. The flow paths include a first-stage flow path r1, a second-stage flow path r2, and a third-stage flow path r3. The molten resin injected from the cylinder first flows into the first-stage flow path r1. The flow path r1 branches into two at its end, and each branch leads to the second-stage flow path r2. Therefore, there are two flow paths r2. Each of the two flow paths r2 branches into four, and each branch leads to the third-stage flow path r3. Therefore, there are eight flow paths r3. These eight flow paths r3 are connected to each of the cavities c1 to c8. By using a mold having such branched flow paths, eight molded products can be manufactured simultaneously in one injection molding. In a hot runner, the temperature of each of the branched flow paths is adjusted by a heater. In addition, in order to eliminate bias between cavities, it is common to design each flow path to be equal in length.

When the mold contains multiple cavities in this way, inference data including parameters related to multiple flow paths connected to each of the multiple cavities may be used. In this case, the inference unit 202 may infer a predetermined inference item for each of the multiple cavities. This makes it possible to manage the mold on a cavity-by-cavity basis.

For example, by providing a temperature sensor in each of the flow paths connected to each cavity, it is possible to measure the temperature of the molten resin flowing through each flow path. This makes it possible to calculate for each cavity the difference between the decomposition temperature of the molten resin and the temperature of the flow path that guides the molten resin to the mold, as well as the time that the molten resin remains in the flow path. This also makes it possible, for example, to predict the quality of the molded product for each cavity.

The inference data may also include parameters that indicate the flow state of the molten resin calculated using the output value of a heater that is provided along the flow path that guides the molten resin to the mold and that maintains the molten resin in the flow path at a predetermined temperature. This makes it possible to obtain reasonable inference results that reflect changes in the flow state that are a sign of molding defects caused by mold contamination, using an easily obtainable value, the heater output value.

For example, in the example of A1 in FIG. 3, the output value of heater h4 can be used to calculate a parameter indicating the flow state of the molten resin. For example, when controlling the output of heater h4 so that the temperature of the molten resin flowing through the flow path falls within a predetermined range, the parameter may be the ratio or difference of the output of heater h4 relative to the normal state. When the flow state of the molten resin deteriorates and the flow rate decreases, the amount of heat required per given time to keep the temperature of the molten resin within the predetermined range decreases, and the output of heater h4 drops significantly compared to normal.

Similarly, in the example of A2 in Figure 3, the above parameters can be calculated from the output values of each heater provided in each flow path. For example, when predicting the quality of a molded product molded in cavity c1, the above parameters can be calculated from the output values of each heater in flow paths r1 to r3. Note that each of the parameters calculated for each of the multiple flow paths corresponding to one cavity may be used as inference data, or one parameter may be calculated for multiple flow paths corresponding to one cavity. In the latter case, the above parameters may be, for example, a representative value (e.g., average value or maximum value) of the ratio or difference of the heater output in each flow path relative to the normal state.

In addition, when a mold has multiple cavities, the ratio or difference of the heater output value of each heater to the cavity average may be used as a parameter indicating the flow state. For example, in the example of A2 in FIG. 3, the parameter may be the ratio or difference of the heater output value in each of the 11 flow paths to the average heater output value in the 11 flow paths. Also, as described above, such parameters may be calculated for each cavity.

The inference data may include the temperature, pressure, and heater output values measured during injection molding, or may include parameters calculated using these values. For example, in the example of A1 in FIG. 3, the inference data may include the output values of heaters h1 to h4, the temperature near these heaters, and the injection pressure at the injection port e. The inference data may also include the cumulative temperature value and the rate of change of pressure. In addition, the inference data may include, for example, the position of the screw. In the example of A2 in FIG. 3, the inference data may include the output values of the heaters provided in each flow path and the temperatures near these heaters. In this way, the temperature of each part of the flow path up to the cavity and the heater output value can be used as inference data.

(Other examples of data for inference)
4 is a diagram showing the relationship between the screw position of the injection molding machine and the injection pressure in both normal and closed air vent conditions. As described above, during injection molding, the molten resin is extruded and injected into the mold by the screw, and at this time, as shown in the figure, the injection pressure increases as the screw position moves toward the mold. This increase in injection pressure accompanying the transition of the screw position is a phenomenon common to both normal and closed air vent conditions, but when the air vent is closed, gas is less likely to escape from the mold, so the injection pressure increases at an earlier timing than normal.

In this way, the manner in which the injection pressure changes reflects whether or not the air vent is blocked. For this reason, the inference data may include a parameter that indicates the manner in which the injection pressure changes. This makes it possible to perform highly accurate inference that reflects minute changes such as whether or not the air vent is blocked.

The above parameter may be, for example, the ratio of the change in injection pressure to the change in screw position (the slope of the graph shown in FIG. 4). The above parameter may also be the rate of pressure increase per unit time during the injection period (which can also be said as the period during which the screw moves toward the injection port). Each of the above parameters may be an average value or a maximum value during the injection period. The above parameter may also be the value of the injection pressure at a specified timing (for example, the timing when the screw moves to a specified position).

(Further examples of inference data)
Figure 5 is a diagram showing the relationship between the number of shots and the defect rate in injection molding. In addition, in this figure, the timing of cleaning the mold is indicated by an arrow. As shown in the figure, cleaning of the mold is performed when the defect rate becomes high, and the defect rate is kept low for a while after cleaning. Although mold dirt is not the only cause of molded product defects, mold dirt has a large influence especially on insufficient filling (short), so cleaning the mold is often an effective solution when insufficient filling occurs frequently.

As a measure to avoid insufficient filling, it is common to increase the fluidity of the molten resin by raising the molding temperature. However, when this measure is taken, the high temperature of the molten resin can easily cause thermal decomposition, resulting in blockage of the air vents due to pyrolysis products, which can lead to a vicious cycle of insufficient filling. For this reason, it is preferable to clean the mold at an appropriate time rather than taking measures to raise the molding temperature.

As shown in Figure 5, the defect rate tends to increase as the cumulative number of injection moldings after the mold used for injection molding is last cleaned increases. For this reason, the inference data may include the cumulative number of injection moldings after the mold used for injection molding is last cleaned. This makes it possible to obtain a reasonable inference result that reflects the cumulative number of injections, which is a sign of molding defects caused by mold contamination. For example, it becomes possible to take measures such as changing other molding conditions without increasing the molding temperature to make it difficult for the mold to become contaminated, or to predict the appropriate timing for cleaning.

(How to generate an inference model)
A method for generating the inference model 211 executed by the information processing device 1 will be described with reference to Fig. 6. Fig. 6 is a flowchart showing an example of the method for generating the inference model 211.

In S11, the training data acquisition unit 101 acquires training data 111 generated based on the difference between the decomposition temperature at which the molten resin thermally decomposes during injection molding and the temperature of the flow path that guides the molten resin to the mold. More specifically, the training data 111 is data for inference that includes the difference between the decomposition temperature at which the molten resin thermally decomposes during injection molding and the temperature of the flow path that guides the molten resin to the mold, and corresponds correct answer data for a specific matter related to mold management.

The training data 111 may include, as an explanatory variable, the time that the molten resin remains in the flow path, instead of or in addition to the difference between the decomposition temperature and the flow path temperature. The explanatory variable may also include various data such as the cumulative number of injection moldings after the mold was last cleaned.

In S12, the learning unit 102 generates an inference model 211 for inferring a predetermined matter from the inference data by machine learning using the training data 111 acquired in S11, and the process in FIG. 6 ends. The learning unit 102 may store the generated inference model 211 in the storage unit 11 or the like, or may transmit it to the information processing device 2.

As described above, the method for generating the inference model shown in FIG. 6 includes a data acquisition step (S11) for acquiring training data 111 in which correct answer data for a predetermined matter related to mold management is associated with inference data including at least one of the difference between the decomposition temperature at which the molten resin thermally decomposes in injection molding and the temperature of the flow path that guides the molten resin to the mold, and the time the molten resin remains in the flow path, and a learning step (S12) for generating an inference model 211 for inferring the predetermined matter from the inference data by machine learning using the training data 111. This makes it possible to generate an inference model 211 that contributes to the appropriate management of a mold used to mold a resin with a low decomposition temperature. This makes it possible to appropriately manage a mold used to mold a resin with a low decomposition temperature.

(Inference Method)
The inference method executed by the information processing device 2 will be described with reference to Fig. 7. Fig. 7 is a flowchart showing an example of the inference method.

In S21, the data acquisition unit 201 acquires inference data including the difference between the decomposition temperature at which the molten resin thermally decomposes during injection molding and the temperature of the flow path that guides the molten resin to the mold. Note that the inference data may include, as an explanatory variable, the time that the molten resin remains in the flow path instead of or in addition to the difference between the decomposition temperature and the flow path temperature, or may include other data.

In S22, the inference unit 202 infers the above items from the inference data acquired in S21 using an inference model 211 generated by learning the relationship between the inference data and specific items related to mold management.

In S23, the output control unit 203 causes the inference result of S22 to be output to the output unit 24. This ends the processing in FIG. 7. Note that the inference result of S22 may be output to a unit other than the output unit 24. Also, the processing in S23 may be omitted, and the inference result may be stored in the storage unit 21 or the like to end the inference.

As described above, the inference method shown in FIG. 7 includes a data acquisition step (S21) for acquiring inference data including at least one of the difference between the decomposition temperature at which the molten resin thermally decomposes in injection molding and the temperature of the flow path that guides the molten resin to the mold, and the time that the molten resin remains in the flow path, and an inference step (S22) for inferring the above items from the inference data acquired in the data acquisition step using an inference model 211 generated by learning the relationship between the inference data and specific items related to mold management. This makes it possible to properly manage molds used to mold resins with low decomposition temperatures.

[Embodiment 2]
(Configuration of information processing device)
8 is a block diagram showing an example of a main configuration of the information processing device 3 according to this embodiment. As with the information processing device 2 of embodiment 1, the information processing device 3 infers a predetermined inference item related to a mold from inference data using an inference model generated by machine learning. The information processing device 3 also has a function of automatically searching for suitable conditions for injection molding using the inference result.

As shown in the figure, the information processing device 3 includes a control unit 30 that controls each unit of the information processing device 3, and a storage unit 31 that stores various data used by the information processing device 3. The information processing device 3 also includes a communication unit 32 that allows the information processing device 3 to communicate with other devices, an input unit 33 that accepts input of various data to the information processing device 3, and an output unit 34 that allows the information processing device 3 to output various data. The control unit 30 also includes a data acquisition unit 301, an inference unit 302, a condition search unit 303, a data generation unit 304, and an output control unit 305. The storage unit 31 stores an inference model 311.

The data acquisition unit 301 acquires inference data including at least one of the difference between the decomposition temperature at which the molten resin thermally decomposes during injection molding and the temperature of the flow path that guides the molten resin to the mold, and the time that the molten resin remains in the flow path. The inference data also includes condition data that indicates the conditions of the injection molding. In addition, the inference data may include various explanatory variables, such as the cumulative number of injection moldings since the mold was last cleaned.

The inference unit 302 uses the inference model 311 to infer certain matters related to mold management from the inference data acquired by the data acquisition unit 301. Specifically, the inference unit 302 predicts the quality of the product.

The inference model 311 is generated by learning the relationship between the inference data and specific items related to mold management. Specifically, the specific items are the quality of the product. The inference model 311 can be generated, for example, by the information processing device 1 of embodiment 1. The flow of the generation process is as shown in Figure 6, and the only difference from the generation of the inference model 211 is the training data.

The training data used to generate the inference model 311 is obtained by associating correct answer data indicating the quality of a product generated under the conditions with inference data including condition data indicating the conditions of injection molding, and at least one of the difference between the decomposition temperature at which the molten resin thermally decomposes during injection molding and the temperature of the flow path that guides the molten resin to the mold, and the time that the molten resin remains in the flow path. The quality of the product may be, for example, the defect rate. Furthermore, when a mold including multiple cavities is used, the inference model 311 may be used to predict the quality of the product for each cavity.

The explanatory variables of the inference model 311 may include data measured during injection molding or parameters calculated using the data. In this case, a prediction model that predicts the data measured during injection molding or the values of parameters calculated using the data may be generated in advance. Then, by inputting the predicted values calculated using the prediction model to the inference model 311, it is possible to output prediction results that reflect the data.

The condition search unit 303 searches for suitable conditions for injection molding. More specifically, the condition search unit 303 searches for suitable conditions for injection molding by repeatedly executing a prediction by the inference unit 302 using the condition data generated by the data generation unit 304 and the generation of condition data by the data generation unit 304 according to the prediction result.

The data generation unit 304 generates condition data indicating other conditions different from the conditions in question based on the results of the prediction of the quality of the product by the inference unit 302 using the inference data including condition data indicating the injection molding conditions.

The output control unit 305 causes the output unit 34 to output the conditions detected by the condition search unit 303. The output mode is arbitrary, and for example, the output control unit 305 may cause the conditions to be output in at least one of the following modes: display output, audio output, and print output. The device that outputs the conditions may be a device external to the information processing device 3. The output control unit 305 may also output the conditions detected by the condition search unit 303 to a device that controls injection molding, and apply the conditions to the injection molding.

As described above, the information processing device 3, like the information processing device 2 of embodiment 1, is equipped with a data acquisition unit 301 that acquires inference data including at least one of the difference between the decomposition temperature at which the molten resin thermally decomposes in injection molding and the temperature of the flow path that leads the molten resin to the mold, and the time that the molten resin remains in the flow path, and an inference unit 302 that infers the above items from the inference data acquired by the data acquisition unit 301 using an inference model 311 generated by learning the relationship between the inference data and specific items related to mold management. This makes it possible to appropriately manage molds used for molding resins with low decomposition temperatures.

The data for inference in this embodiment also includes condition data indicating the conditions for injection molding. The information processing device 3 is equipped with a data generation unit 304 that generates condition data indicating conditions different from the above conditions based on the prediction result of the inference unit 302 as to whether the product is good or bad, and a condition search unit 303 that repeatedly executes a prediction by the inference unit 302 using the condition data generated by the data generation unit 304 and the generation of condition data by the data generation unit 304 according to the prediction result. This makes it possible to automatically search for suitable conditions for injection molding.

(Processing flow)
The flow of the process executed by the information processing device 3 will be described with reference to Fig. 9. Fig. 9 is a flowchart showing an example of the process executed by the information processing device 3.

In S31, the data acquisition unit 301 acquires inference data including the difference between the decomposition temperature at which the molten resin thermally decomposes during injection molding and the temperature of the flow path that guides the molten resin to the mold, and the initial value of the condition data. Note that the inference data may include the time that the molten resin remains in the flow path instead of or in addition to the difference between the decomposition temperature and the flow path temperature, and may also include parameters other than these.

The difference between the decomposition temperature and the flow path temperature included in the inference data may be the difference between the decomposition temperature and the flow path temperature in the most recent injection molding. For example, the data acquisition unit 301 may use a prediction model that predicts the flow path temperature from the condition data included in the inference data acquired in S31 to predict the flow path temperature from the condition data, and calculate the difference between the decomposition temperature and the flow path temperature from the predicted flow path temperature. The same applies to the time that the molten resin is retained in the flow path, and the data acquisition unit 301 may include the retention time predicted from the condition data in the inference data, or may include the retention time in the most recent injection molding in the inference data.

The initial value of the condition data may be set by the user, or may be set by the data acquisition unit 301, for example, by using random numbers for each explanatory variable from within the range of the condition data learned by the inference model 311. In addition, if there is a combination of explanatory variables that is known to obtain a good value of the objective variable indicating the quality of the product, the data acquisition unit 301 may set the nearby value as the initial value of the condition data.

In S32, the inference unit 302 predicts the quality of the product from the inference data acquired in S31 using the inference model 311. Note that multiple pieces of inference data may be acquired in S31, and when multiple pieces of inference data are acquired, the inference unit 302 makes a similar prediction for each piece of inference data.

In S33, the condition search unit 303 determines whether or not the end condition of the condition search is satisfied. The end condition of the condition search may be set arbitrarily. For example, if the defect rate is predicted in S32, the end condition may be that the predicted value of the defect rate is within a predetermined tolerance range. If the result of S33 is NO, the process proceeds to S34, and if the result of S33 is YES, the process proceeds to S35.

In S34, the data generation unit 304 generates new condition data that is at least partially different from the condition data used for the prediction based on the prediction result in S32. The method of generating the condition data may be determined in advance, and for example, the data generation unit 304 may generate new condition data by increasing or decreasing a predetermined explanatory variable included in the condition data by a predetermined amount. After S34, the process returns to S32, and a prediction is made using the condition data generated in S34.

In S35, the condition search unit 303 determines the injection molding conditions to be the conditions indicated in the condition data used in the most recent prediction, and the process in FIG. 9 ends. The output control unit 305 may output the determined conditions to the output unit 34, etc.

(Application of genetic algorithm)
Various methods can be applied to search for suitable condition data. For example, it is also possible to search for condition data using a genetic algorithm. In this case, in S31, the condition search unit 303 generates a plurality of pieces of inference data including candidates for optimal condition data. Each explanatory variable included in the condition data is a "gene," one piece of inference data including those genes is one "individual," and a group of generated inference data is one "generation."

Next, in S32, the inference unit 302 uses the inference model 311 to predict the quality of the product using each of the inference data generated in S31. This prediction result becomes the "fitness" of each individual. Then, in S33, the condition search unit 303 checks whether there is an individual whose fitness satisfies a predetermined condition, and if there is such an individual, the process proceeds to S35, where the conditions indicated in the condition data of the individual are determined as the conditions for injection molding. On the other hand, if there is no individual whose fitness satisfies the predetermined condition, the condition search unit 303 proceeds to S34, where the next generation of individuals is generated. When generating the next generation of individuals, selection of individuals with high fitness, crossover, mutation, and other processes may be performed. Note that the method of searching for suitable condition data is arbitrary, and other methods such as mathematical optimization may be used.

[Modifications]
The execution subject of each process described in each of the above-mentioned embodiments is arbitrary and is not limited to the above-mentioned examples. In other words, the same functions as the information processing devices 1 to 3 can be realized by a plurality of information processing devices capable of communicating with each other. For example, each process shown in Fig. 6, Fig. 7, and Fig. 9 may be shared and executed by a plurality of information devices. In addition, an information processing device having the functions of the information processing devices 1 and 2, or an information processing device having the functions of the information processing devices 1 and 3, is also included in the scope of the present invention.

[Software implementation example]
The functions of information processing devices 1 to 3 (hereinafter referred to as "devices") can be realized by a program for causing a computer to function as the device, and a program (inference program/learning program) for causing a computer to function as each control block of the device (particularly each part included in controls 10, 20, 30).

In this case, the device includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the program. The functions described in each of the above embodiments are realized by executing the program using the control device and storage device.

The program may be recorded on one or more computer-readable recording media, not on a temporary basis. The recording media may or may not be included in the device. In the latter case, the program may be supplied to the device via any wired or wireless transmission medium.

In addition, some or all of the functions of each of the above control blocks can be realized by a logic circuit. For example, an integrated circuit in which a logic circuit that functions as each of the above control blocks is formed is also included in the scope of the present invention. In addition, it is also possible to realize the functions of each of the above control blocks by, for example, a quantum computer.

The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. The technical scope of the present invention also includes embodiments obtained by appropriately combining the technical means disclosed in different embodiments.

2 Information processing device 201 Data acquisition unit 202 Inference unit 211 Inference model 3 Information processing device 301 Data acquisition unit 302 Inference unit 303 Condition search unit 304 Data generation unit 311 Inference model

Claims (10)

a data acquisition unit that acquires data for inference including at least one of a difference between a decomposition temperature at which a molten resin is thermally decomposed in injection molding and a temperature of a flow path through which the molten resin is guided to a mold, and a time during which the molten resin is retained in the flow path;
An information processing device comprising: an inference unit that infers the matters from the inference data acquired by the data acquisition unit using an inference model generated by learning the relationship between the inference data and specified matters related to the management of the mold. the inference model is generated by learning a relationship between the inference data and the quality of a molded product molded by injection molding using the mold,
The information processing device according to claim 1 , wherein the inference unit predicts whether a molded product produced by injection molding using the metal mold is good or bad. The information processing device according to claim 1 or 2, wherein the inference data includes the cumulative number of injection moldings after the mold used in the injection molding was last cleaned. The information processing device according to claim 1 or 2, wherein the inference data includes a parameter indicating the flow state of the molten resin calculated using the output value of a heater provided along a flow path that guides the molten resin to the mold, for maintaining the molten resin in the flow path at a predetermined temperature. The information processing device according to claim 1 or 2, wherein the inference data includes a parameter indicating the manner in which the injection pressure changes. the inference data includes parameters related to a plurality of flow paths connected to each of a plurality of cavities included in the mold,
The information processing device according to claim 1 , wherein the inference unit infers the item for each of the plurality of cavities. The inference data includes condition data indicating conditions of the injection molding,
a data generating unit that generates condition data indicating other conditions different from the above-mentioned conditions based on a result of the prediction of the quality of the product by the inference unit;
The information processing device according to claim 2 , further comprising a condition searching unit that repeatedly executes a prediction by the inference unit using the condition data generated by the data generating unit and generation of the condition data by the data generating unit in accordance with the prediction result. A method for generating an inference model executed by one or more information processing devices, comprising:
a data acquisition step of acquiring training data in which correct answer data for a predetermined matter related to mold management is associated with inference data including at least one of a difference between a decomposition temperature at which a molten resin thermally decomposes during injection molding and a temperature of a flow path through which the molten resin is introduced into a mold, and a residence time of the molten resin in the flow path;
A method for generating an inference model, comprising: a learning step of generating an inference model for inferring the matter from the inference data by machine learning using the training data. 1. An inference method executed by one or more information processing devices, comprising:
a data acquisition step of acquiring data for inference including at least one of a difference between a decomposition temperature at which a molten resin is thermally decomposed in injection molding and a temperature of a flow path through which the molten resin is guided to a mold, and a residence time of the molten resin in the flow path;
an inference step of inferring the matters from the inference data acquired in the data acquisition step using an inference model generated by learning the relationship between the inference data and specified matters related to the management of the mold. An inference program for causing a computer to function as the information processing device according to claim 1, the inference program causing a computer to function as the data acquisition unit and the inference unit.

Images