JP7750184B2 - Shrinkage rate prediction device and method, and method for manufacturing resin molded body - Google Patents

Description

The present invention relates to a shrinkage rate prediction device and method for predicting the shrinkage rate of a resin molded body formed by extrusion molding of resin, and a method for manufacturing a resin molded body.

Patent Document 1 discloses a technology for predicting changes in shrinkage rate by taking into account crystallization during the molding process of crystalline resins.

Japanese Patent Application Publication No. 9-262887

Cable coating materials and resin tubes are formed by extrusion molding. It is known that when heat is applied to the resin molded product obtained by extrusion molding, post-shrinkage (shrink back) occurs, in which the resin molded product shrinks. For example, if post-shrinkage occurs in the longitudinal direction of the cable, it may cause problems such as fluctuations in the cable's outer diameter or unintentional exposure of the conductor (core) at the end of the cable, so it is desirable to reduce post-shrinkage.

It has been known that post-molding shrinkage can be reduced by appropriately controlling the extrusion conditions (resin temperature, linear speed, rotation speed, etc.) during extrusion molding. However, no quantitative theory has been established for predicting the shrinkage rate, and workers have had to control the extrusion conditions empirically.

The present invention therefore aims to provide a shrinkage rate prediction device and method that can accurately predict shrinkage rates, as well as a method for manufacturing resin molded products.

The present invention aims to solve the above-mentioned problems by providing an apparatus for predicting the shrinkage rate of a resin molded body formed by extrusion molding of resin. The apparatus includes a regression model creation processing unit that uses the shrinkage rate of the resin molded body as a dependent variable and creates a regression model that represents the correlation between a predetermined explanatory variable and the dependent variable, and a shrinkage rate prediction processing unit that predicts the shrinkage rate of the target using the regression model, wherein the explanatory variable includes both the drawing stress applied when the resin is extruded and the elastic modulus of the resin molded body.

Furthermore, in order to solve the above-mentioned problems, the present invention provides a method for predicting the shrinkage rate of a resin molded body formed by extrusion molding of resin, comprising: a regression model creation process that uses the shrinkage rate of the resin molded body as a dependent variable and creates a regression model that represents the correlation between predetermined explanatory variables and the dependent variable; and a shrinkage rate prediction process that predicts the shrinkage rate of the target to be predicted using the regression model, wherein the explanatory variables include both the drawdown stress applied when the resin is extruded and the elastic modulus of the resin molded body.

Furthermore, in order to solve the above-mentioned problems, the present invention provides a method for manufacturing a resin molded body by extruding a resin to form a resin molded body, in which the shrinkage rate prediction method is used to derive extrusion conditions under which the shrinkage rate of the resin molded body will be a desired value, and the resin molded body is manufactured under the derived extrusion conditions.

The present invention provides a shrinkage rate prediction device and method that can accurately predict shrinkage rates, as well as a method for manufacturing resin molded products.

1 is a schematic configuration diagram of a shrinkage rate prediction device according to an embodiment of the present invention. 1A is a schematic diagram of a resin molding manufacturing apparatus, and FIG. 1B is a schematic diagram showing how a conductor wire is covered with resin extruded from a die. 1A is a diagram illustrating a learning data extraction process, FIG. 1B is a diagram illustrating a regression model creation process, and FIG. 1C is a diagram illustrating a shrinkage rate prediction process. FIG. 1 is a flow diagram of a shrinkage rate prediction method according to an embodiment of the present invention. 1A is a flow diagram of a data acquisition process, and FIG. 1B is a flow diagram of a learning data extraction process. 10A is a flow diagram of a regression model creation process, and FIG. 10B is a flow diagram of a shrinkage rate prediction process. 1 is a flow diagram of a method for manufacturing a resin molded article according to an embodiment of the present invention. 1A is a graph showing the experimentally obtained relationship between shrinkage and pull-down stress, and FIG. 1B is a graph showing the experimentally obtained relationship between shrinkage and strain. 1A is a diagram illustrating a process for creating a regression model for predicting elastic modulus, and FIG. 1B is a diagram illustrating a process for predicting elastic modulus.

[Embodiment]
Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings.

Figure 1 is a schematic diagram of a shrinkage rate prediction device 1 according to this embodiment. The shrinkage rate prediction device 1 is a device that predicts the shrinkage rate of a resin molded body formed by extrusion molding a resin. In this embodiment, the shrinkage rate prediction device 1 is configured as a personal computer.

(About the prediction target)
In this embodiment, the shrinkage rate is predicted for a resin molded article formed by extrusion molding. In the following example, a case where the resin molded article is a coating material for an electric wire will be described, but the resin molded article for which the shrinkage rate is predicted is not limited to a coating material for an electric wire and may be, for example, a resin tube. There are no particular restrictions on the resin used, but from the viewpoint that prediction accuracy can be particularly improved by taking into account the elastic modulus, which will be described later, the present invention is particularly suitable for crystalline resins such as TPU (thermoplastic polyurethane resin).

(Schematic configuration of shrinkage rate prediction device 1)
As shown in FIG. 1, the shrinkage rate prediction device 1 includes a control unit 2, a storage unit 3, a display unit 4, and an input device 5.

The control unit 2 is realized by appropriately combining a CPU or other computing element, memory, an interface, software, a storage device, etc. In this embodiment, the control unit 2 has a setting processing unit 21, a data acquisition processing unit 22, a drawdown stress calculation processing unit 23, a strain calculation processing unit 24, a learning data extraction processing unit 25, a regression model creation processing unit 26, a shrinkage rate prediction processing unit 27, and a prediction data presentation processing unit 28. Details of each unit will be described later.

The storage unit 3 is realized by a predetermined storage area of a memory or storage device. The display 4 is, for example, an LCD display, and the input device 5 is, for example, a keyboard or a mouse. The display 4 may be configured as a touch panel and also serve as the input device 5. The display 4 and input device 5 may also be configured separately from the shrinkage rate prediction device 1 and configured to be able to communicate with the shrinkage rate prediction device 1 via wireless communication or the like. In this case, the display 4 or input device 5 may be configured as a mobile terminal such as a tablet or smartphone.

(Resin Molded Body Manufacturing Apparatus 60)
Before describing the details of the shrinkage rate prediction device 1, a description will be given of a resin molded body manufacturing apparatus 60. Fig. 2(a) is a schematic diagram of the resin molded body manufacturing apparatus 60, and Fig. 2(b) is a schematic diagram showing how a conductor 65 is coated with molten resin 67 extruded from a die 62a.

As shown in Figure 2(a), raw material pellets, which are the raw material for the resin molding (here, the coating material), are fed into an extruder 61 and kneaded. Molten resin is then extruded from a die 62a via a crosshead 62. The extruded resin coats the surface of a conductor 65 traveling along a line. The resin coating the surface of the conductor 65 is then air-cooled immediately after being extruded from the die 62a, and then water-cooled in a water bath 63. In this way, the resin extruded from the die 62a solidifies during the cooling process using air and water. The outer diameter of the electric wire 66 that passes through the water bath 63 is measured using an outer diameter measuring device 64, and the wire 65 is then wound onto a drum or the like. The conductor 65 is composed of, for example, a stranded conductor formed by twisting together multiple wires made of copper, copper alloy, or the like. The resin molding manufacturing apparatus 60 includes a control device 69, which is configured to perform various controls, including control of the extrusion conditions.

(Consideration of parameters used to predict shrinkage rate)
For example, it is known that shrinkage of a resin molded body can be reduced by appropriately controlling the extrusion conditions (such as the temperature of the resin, the wire speed, and the rotation speed of the screw of the extruder 61) in the extrusion molding process. For this reason, conventional shrinkage estimation techniques have estimated the shrinkage rate based on input of the extrusion conditions. However, the shrinkage rate of a resin is affected not only by the extrusion conditions but also by the material properties of the resin (such as viscosity and crystallinity). Therefore, the present inventors focused on a physical quantity known as "drawdown stress."

In Figure 2(b), molten resin 67 extruded from the outlet of die 62a coats the surface of conductor 65, which is moving at a constant linear velocity. The molten resin 67 coated on the surface of conductor 65 is then cooled and crystallized in water tank 63. At this time, as shown in Figure 2, molten resin 67 extruded from the outlet of die 62a is pulled by conductor 65, which is moving along the travel line. As a result, the stress applied diagonally to molten resin 67 from the time it is extruded from the outlet of die 62a until it coats conductor 65 is the "drawdown stress σ." This "drawdown stress σ" is a physical quantity that can be formulated taking into account both the extrusion conditions and material properties. According to the inventor's research, the drawdown stress σ can be expressed by equation (1) shown in [Mathematical Expression 1].

The draw-down distance L in equation (1) is the distance from the exit of the die 62a to the water tank 63. The resin temperature T in equation (1) is the temperature of the molten resin extruded from the die 62a. The linear speed _vf in equation (1) is the speed of the conductor 65 moving along the travel line. The draw-down ratio D in equation (1) is expressed by equation (2) shown in [Equation 2]. Furthermore, the viscosity constant η0, the temperature coefficient of viscosity α, and the viscosity index n in equation (1) have the relationship shown in equation (3) shown in [Equation 3].

The strain rate dγ/dt included in equation (3) is expressed by equation (4) shown in [Mathematical Expression 4]. That is, the "strain rate" is defined as a physical quantity obtained by dividing the difference between the average flow velocity _vd of the molten resin at the outlet of the die 62a (hereinafter simply referred to as the average flow velocity) and the linear velocity _vf by the draw-down distance L, which is the distance from the outlet of the die 62a to the water tank 63. For example, when the average flow velocity _vd and the linear velocity _vf are equal, the strain rate becomes zero, and it can be considered that no strain is qualitatively applied to the molten resin 67.

The pull-down stress σ shown in equation (1) above is a relationship that takes into account both the material property of the resin, namely its viscosity, and the extrusion conditions. Therefore, if you have data on the resin's viscosity measured with a capillary rheometer, you can calculate the pull-down stress σ from the extrusion conditions.

The inventors' investigations revealed that when predicting shrinkage using machine learning with only the pull-down stress σ as an explanatory variable, accuracy could be insufficient depending on the type of resin used and the production line. More specifically, for rubber-based resins such as PVC (polyvinyl chloride resin), shrinkage could be predicted with high accuracy using only the pull-down stress σ as an explanatory variable, but accuracy was found to be slightly lower when predicting the shrinkage of crystalline resins such as TPU. This is thought to be due to errors in the calculation results of the pull-down stress σ depending on the type of resin and manufacturing conditions.

Therefore, in order to further improve the accuracy of predicting the shrinkage percentage, the present inventors have considered predicting the shrinkage percentage by taking into account the elastic modulus of the resin molded body measured by a viscoelasticity test in addition to the pull-down stress σ obtained by the above formula (1). More specifically, when the elastic modulus (storage modulus) of the resin molded body is E, the following formula (5) can be used:
σ=Eε (5)
The strain ε obtained by the above equation was used as an explanatory variable. It was confirmed that this method enables more accurate prediction of the shrinkage rate of crystalline resins such as TPU. The actual experimental results will be described later.

(Setting processing unit 21)
Returning to Fig. 1, each part of the control unit 2 will be described in detail. The setting processing unit 21 performs setting processing for making various settings of the shrinkage rate prediction device 1. The setting processing unit 21 can set information related to various controls, such as the method of data acquisition by the data acquisition processing unit 22 and the setting of the data acquisition date and time. The setting processing unit 21 can also register, update, delete, etc., various pieces of information stored in the storage unit 3. The input device 5 or the like can be used to input various pieces of information.

(Data acquisition processing unit 22)
The data acquisition processing unit 22 performs a data acquisition process (see FIG. 5A ) to acquire various data and store them in the storage unit 3. The data acquisition processing unit 22 associates the acquired various data with each sample and registers them in the database 31. In this embodiment, the data acquisition processing unit 22 acquires extrusion condition data 71, which is data on extrusion conditions such as the draw-down distance L and the draw-down ratio D; viscosity data 72, which is data on the viscosity of the resin such as the viscosity constant η ₀ ; elastic modulus data 73, which is data on the elastic modulus E of the resin molded body; and shrinkage rate data 76, which is data on the shrinkage rate of the resin molded body. These various data may be input via the input device 5 or may be input from an external device via a network or the like. The data acquisition processing unit 22 may be configured to actively acquire data, for example, by transmitting a signal requesting data to an external device such as the control device 69 of the resin molded body manufacturing apparatus 60. The data acquisition processing unit 22 may also have a function to present missing data for any sample, for example, by displaying the missing data on the display 4.

(Drawdown stress calculation processing unit 23)
The pull-down stress calculation processing unit 23 performs pull-down stress calculation processing to calculate the pull-down stress σ using the above formula (1) based on the information on the viscosity of the resin (viscosity data 72) and the information on the extrusion conditions of the resin (extrusion condition data 71) acquired by the data acquisition processing unit 22. The pull-down stress σ data obtained by the pull-down stress calculation processing is registered in the database 31 as pull-down stress data 74 and stored in the memory unit 3.

(Strain calculation processing unit 24)
The strain calculation processing unit 24 performs strain calculation processing to determine the strain ε using the above formula (5) based on the elastic modulus E (elastic modulus data 73) and the pull-down stress σ (pull-down stress data 74). The strain ε determined by the strain calculation processing is registered in the database 31 as strain data 75 and stored in the storage unit 3.

(Learning Data Extraction Processing Unit 25)
The learning data extraction processing unit 25 performs a learning data extraction process (see FIG. 5( b)) that extracts data necessary for machine learning from the database 31. As shown in FIG. 3( a), in this embodiment, the learning data extraction processing unit 25 extracts strain data 75 from the database 31 and sets it as explanatory variable data 81. As shown in the above formula (5), the strain ε included in the strain data 75 is a value derived using the drawdown stress σ applied when extruding the resin and the elastic modulus E of the resin molded body. Therefore, it can be said that the explanatory variable data 81 includes both the drawdown stress σ applied when extruding the resin and the elastic modulus E of the resin molded body.

In this embodiment, only strain ε is used as an explanatory variable, but items other than strain ε may also be added as explanatory variables. For example, if crosslinking is performed on a resin molded product, parameters related to the crosslinking may be added as explanatory variables. Furthermore, similar effects can be achieved by using both the pull-down stress σ and the elastic modulus E as explanatory variables instead of strain ε.

The learning data extraction processing unit 25 also extracts contraction rate data 76 from the database 31 and sets it as objective variable data 82. The extracted explanatory variable data 81 and objective variable data 82 are stored in the memory unit 3 as learning data 32. Note that in this embodiment, the learning data extraction processing unit 25 is configured to generate and store learning data 32, but this is not limiting. The explanatory variable data 81 and objective variable data 82 may be extracted and the extraction results may be used directly in the regression model creation process, or the function of generating learning data 32 may be omitted. The learning data 32 may also be temporarily stored and deleted after the regression model creation process.

(Regression model creation processing unit 26)
The regression model creation processing unit 26 performs a regression model creation process (see FIG. 6A ) in which the shrinkage rate of the resin molded body is used as a response variable to create a regression model 33 that represents the correlation between a predetermined explanatory variable and the response variable. That is, the regression model creation processing unit 26 performs machine learning on the relationship between the strain data 75 extracted as the explanatory variable data 81 and the shrinkage rate data 76 extracted as the response variable data 82, and creates a regression model 33 that represents the correlation between the explanatory variable data 81 and the response variable data 82. In this embodiment, the regression model creation processing unit 26 uses the training data 32 generated by the training data extraction processing unit 25 to create the regression model 33 that represents the correlation between the explanatory variable data 81 and the response variable data 82.

The regression model creation processing unit 26 includes software such as a learning algorithm that uses machine learning to learn the correlation between explanatory variables and response variables. There are no particular limitations on the learning algorithm, and any known learning algorithm can be used, such as so-called deep learning that uses a neural network with three or more layers. What the regression model creation processing unit 26 learns corresponds to a model structure that represents the correlation between the data used as explanatory variable data 81 (strain data 75 in this case) and the data used as response variable data 82 (shrinkage rate data 76 in this case).

As shown in Figure 3(b), the regression model creation process involves machine learning the relationship between explanatory variable data 81 and dependent variable data 82 based on training data 32, and creating a regression model 33 that represents the correlation between them. More specifically, the regression model creation processing unit 26 iteratively executes learning based on a data set including explanatory variables and dependent variables, using training data 32, and automatically interprets the correlation between the two. Note that while the correlation is unknown at the start of learning, as learning progresses the correlation between the dependent variable and the explanatory variables is gradually interpreted, and by using the resulting trained model, regression model 33, it becomes possible to interpret the correlation between the dependent variable and the explanatory variables.

The regression model creation processing unit 26 stores the created regression model 33 in the memory unit 3. In this embodiment, the regression model creation processing unit 26 updates the regression model 33 each time the database 31 is updated. However, this is not limited to this, and for example, when performing the shrinkage rate prediction process described below, the data updates may be learned collectively and the regression model 33 may be updated.

(Shrinkage rate prediction processing unit 27)
The shrinkage rate prediction processing unit 27 performs a shrinkage rate prediction process (see FIG. 6B ) to predict the shrinkage rate of the prediction target using the regression model 33. The shrinkage rate prediction processing unit 27 stores the predicted shrinkage rate in the storage unit 3 as prediction data 35.

As shown in FIG. 3(c), in the shrinkage rate prediction process, the regression model 33 and prediction source data 34 (strain data 75 that is the source of prediction, hereinafter referred to as prediction source strain data 75a) input via the input device 5 or the like are input to the shrinkage rate prediction processing unit 27. The shrinkage rate prediction processing unit 27 uses the regression model 33 to determine the shrinkage rate corresponding to the prediction source data 34, and stores the obtained shrinkage rate data 76, which is predicted shrinkage rate data 76a, in the memory unit 3 as prediction data 35.

(Prediction data presentation processing unit 28)
The predicted data presentation processing unit 28 performs a predicted data presentation process to present the predicted data 35. In the predicted data presentation process, for example, the predicted data 35 is displayed on the display 4. Note that the predicted data presentation process may be configured to also present data other than the predicted data 35, such as the predicted source strain data 75a. Furthermore, the predicted data presentation processing unit 28 may have a function to transmit the predicted data 35 to an external device such as a management computing device, and the specific means for presenting the predicted data 35 is not particularly limited.

(Shrinkage rate prediction method)
4 is a flow diagram of the shrinkage rate prediction method according to the present embodiment. As shown in FIG. 4, first, in step S1, a setting process is performed. In the setting process, for example, setting data is input from the input device 5 or the like, and a setting processing unit 21 performs various settings according to the input setting data, data update processes associated with the various settings, and the like.

After the setting process in step S1, the control unit 2 determines in step S2 whether new data has been input. If the determination in step S2 is No (N), the process proceeds to step S8. If the determination in step S2 is Yes (Y), the control unit 2 performs data acquisition processing in step S3.

In the data acquisition process of step S3, as shown in FIG. 5(a), in step S31, the data acquisition processing unit 22 receives various data, including extrusion condition data 71, viscosity data 72, elastic modulus data 73, and shrinkage rate data 76. Then, in step S32, the data acquisition processing unit 22 associates the received data, registers it in the database 31, and stores it in the memory unit 3. Then, the process returns.

After the data acquisition process in step S3, the pull-down stress calculation process is performed in step S4. In the pull-down stress calculation process, the pull-down stress calculation processing unit 23 calculates the pull-down stress using the above formula (1) based on the various data input in step S3. The calculated pull-down stress is registered in the database 31 as pull-down stress data 74 and stored in the memory unit 3.

Then, in step S5, strain calculation processing is performed. In the strain calculation processing, the strain calculation processing unit 24 calculates strain based on the pull-down stress (pull-down stress data 74) calculated in step S4 and the elastic modulus data 73. The calculated strain is registered in the database 31 as strain data 75 and stored in the memory unit 3.

Then, in step S6, a learning data extraction process is performed. In the learning data extraction process, as shown in FIG. 5(b), in step S61, the learning data extraction processing unit 25 extracts strain data 75 as explanatory variable data 81 from the database 31, and extracts shrinkage rate data 76 as response variable data 82. Then, in step S6, the learning data extraction processing unit 25 stores the extracted explanatory variable data 81 and response variable data 82 in the memory unit 3 as learning data 32. Then, the process returns.

After the learning data extraction process in step S6, a regression model creation process is performed in step S7. In the regression model creation process, as shown in FIG. 6(a), in step S71, the regression model creation processing unit 26 uses the unlearned learning data 32 (explanatory variable data 81 and target variable data 82) for machine learning to update the regression model 33. Note that in step S71, if a regression model 33 has not been created, a new regression model 33 is created. Thereafter, in step S72, the regression model creation processing unit 26 stores the updated (or created) regression model 33 in the memory unit 3 and returns. Note that the regression model creation process in step S7 corresponds to the regression model creation processing step of the present invention.

When predicting the shrinkage rate, prediction source data 34 (here, prediction source strain data 75) is input via the input device 5 or the like (step S11). Note that the data that will become the prediction source data 34 may be input into the shrinkage rate prediction device 1 in advance, and the input device 5 may be configured to select the data to be used as the prediction source data 34.

In step S8, the control unit 2 determines whether prediction source data 34 has been input. If the determination in step S8 is No (N), the process returns (returns to step S1). If the determination in step S8 is Yes (Y), the process proceeds to step S9.

In step S9, a shrinkage rate prediction process is performed. In the shrinkage rate prediction process, as shown in FIG. 6(b), first, in step S91, the shrinkage rate prediction processing unit 27 uses the regression model 33 to predict shrinkage rate data 76 (shrinkage rate to be predicted) corresponding to the prediction source data 34, and sets this as predicted data 35. Then, in step S92, the obtained predicted data 35 is stored in the memory unit 3. Note that the shrinkage rate prediction process in step S9 corresponds to the shrinkage rate prediction processing step of the present invention. Then, the process returns.

After the shrinkage rate prediction process in step S9, the predicted data presentation process is performed in step S10. In the predicted data presentation process, the predicted data presentation processing unit 28 presents the predicted data 35, for example by displaying the predicted data 35 on the display 4. Then, the process returns (returns to step S1).

(Method for manufacturing resin molded body)
7 is a flow diagram of a method for manufacturing a resin molded body according to this embodiment. As shown in FIG. 7, first, in step S101, initial values of the extrusion conditions, viscosity, and elastic modulus are set. Then, in step S102, the set values of the extrusion conditions, viscosity, and elastic modulus are output to the shrinkage factor prediction device 1. The set value of the elastic modulus may be appropriately determined based on the results of, for example, a prototype. Furthermore, as will be described in detail later, machine learning may be used to predict the elastic modulus from the extrusion conditions.

Then, in step S103, the shrinkage rate prediction device 1 predicts the shrinkage rate based on the input setting values for the extrusion conditions, viscosity, and elastic modulus. Then, in step S104, it is determined whether the predicted shrinkage rate is equal to or less than a predetermined threshold. If step S104 returns No (N), in step S105, the setting values for the extrusion conditions, etc. are changed, and the process returns to step S102, where the changed setting values are output to the shrinkage rate prediction device 1 and the shrinkage rate is predicted again. If step S104 returns Yes (Y), in step S106, a resin molded body is manufactured using the set extrusion conditions, etc. At this time, a test production is performed using the set extrusion conditions, etc., the shrinkage rate of the resin molded body manufactured in the test production is measured, and the extrusion conditions, etc. may be modified (corrected).

In this way, the method for manufacturing a resin molded body according to this embodiment uses the shrinkage prediction method of the present invention to derive extrusion conditions that will result in a desired shrinkage rate for the resin molded body (here, a value below a threshold value), and the resin molded body is manufactured under the derived extrusion conditions. Conventionally, extrusion conditions have been determined empirically by workers, but this embodiment makes it possible to set appropriate extrusion conditions prior to manufacturing. As a result, it is possible to significantly reduce the number of prototypes of resin molded bodies, and significantly reduce material waste and costs.

(Regarding the experimental results)
First, resin molded bodies were produced under the conditions shown in Table 1 while varying the extrusion conditions (resin temperature T, linear velocity _vf , and drawdown ratio D). The resin molded bodies were tubular, and TPU (thermoplastic polyurethane) was used as the resin. The extruder 61 had a screw diameter of 25 mm, a die diameter of 3.70 mm, a nipple diameter of 2.97 mm, and a vacuum pressure of 12.4 MPa.

After producing the seven samples in Table 1, each sample was cut to 400 mm (length _L1 ) and annealed at 90°C for 4 hours to measure the shrinkage. The length _L2 after annealing was measured, and the shrinkage from the original length _L1 (( _L1 - _L2 )/L1 x 100) was calculated. Meanwhile, the pull-down stress of each sample was calculated using the extrusion conditions and the viscosity data of the resin measured in advance with a capillary rheometer. The viscosity data of the resin used in the calculation is shown in Table ₂ .

The relationship between shrinkage and pull-down stress obtained as described above is shown in Figure 8(a). As shown in Figure 8(a), the correlation coefficient R between shrinkage and pull-down stress was 0.6702, a slightly low value. This is thought to be due to a large error in the calculation of pull-down stress.

Next, dynamic viscoelasticity tests were conducted on each sample to measure its modulus. Measurements were conducted under conditions of varying the temperature from 40°C to 100°C at a constant frequency. From the measurement results for each sample, the storage modulus at 90°C was used as a representative value, and the strain was calculated along with the calculated pull-down stress for each sample. The correlation between the calculated strain and the shrinkage rate for each sample was then confirmed. The results are shown in Figure 8(b). Note that while the "storage modulus at 90°C" was used as the representative value here, similar results would be obtained if the storage modulus at another temperature were used as the representative value.

As shown in Figure 8 (b), the correlation coefficient R between shrinkage and strain was 0.8114, indicating a very strong correlation. Generally, a correlation coefficient R of 0.70 or higher is considered to be a strong correlation. From the above, it was found that the accuracy of predicting shrinkage is improved when strain, which takes into account both pull-down stress and elastic modulus, is used as an explanatory variable, rather than when pull-down stress alone is used as an explanatory variable.

(Modification)
In the above embodiment, the shrinkage percentage prediction device 1 is configured as a personal computer. However, the present invention is not limited to this. For example, the shrinkage percentage prediction device 1 may be configured as a network device such as a server. In this case, the shrinkage percentage prediction device 1 may be configured to be able to communicate with a predetermined terminal device, such as a data management terminal device, and to receive various data from the terminal device. Furthermore, the shrinkage percentage prediction device 1 may be configured to transmit predicted data 35 to the terminal device, and the predicted data 35 may be displayed on the terminal device.

Furthermore, in the above embodiment, a case has been described in which the shrinkage rate prediction device 1 is configured as a single personal computer, etc., but this is not limiting. For example, some of the functions of the shrinkage rate prediction device 1 may be installed in another personal computer, etc. In other words, the shrinkage rate prediction device 1 does not have to be configured as a single piece of hardware, but may be configured as multiple pieces of hardware.

In addition, in the above embodiment, the elastic modulus was measured using a viscoelasticity test, but the elastic modulus may be obtained using other methods. Furthermore, the shrinkage rate prediction device 1 may be equipped with a function for predicting the elastic modulus using machine learning.

More specifically, as shown in Figures 9(a) and (b), the shrinkage rate prediction device 1 may further include an elastic modulus prediction regression model creation processing unit 101 that performs machine learning on the relationship between the resin extrusion conditions and the elastic modulus and creates an elastic modulus prediction regression model 102 that represents the correlation between the resin extrusion conditions and the elastic modulus, and an elastic modulus prediction processing unit 103 that predicts the elastic modulus using the elastic modulus prediction regression model 102.

As shown in FIG. 9(a), extrusion condition data 71 and elastic modulus data 73 are input to the elastic modulus prediction regression model creation processing unit 101 as learning data. These data may be extracted from database 31. The elastic modulus prediction regression model creation processing unit 101 uses the input extrusion condition data 71 as an explanatory variable and the elastic modulus data 73 as a target variable to create an elastic modulus prediction regression model 102 that shows the correlation between them.

As shown in FIG. 9(b), the elastic modulus prediction processing unit 103 receives as input the elastic modulus prediction regression model 102 and prediction source extrusion condition data 71a, which is the prediction source extrusion condition data 71. The elastic modulus prediction processing unit 103 uses the elastic modulus prediction regression model 102 to derive predicted elastic modulus data 73a, which is the elastic modulus data 73 corresponding to the prediction source extrusion condition data 71a. This makes it possible to predict the elastic modulus without producing prototypes of resin molded bodies, thereby further reducing the number of prototypes and further saving on materials and costs.

(Functions and Effects of the Embodiments)
As described above, the shrinkage rate prediction device 1 according to this embodiment includes a regression model creation processing unit 26 that uses the shrinkage rate of a resin molded body as a dependent variable and creates a regression model 33 that represents the correlation between a preset explanatory variable and the dependent variable, and a shrinkage rate prediction processing unit 27 that predicts the shrinkage rate of the prediction target using the regression model 33, and the explanatory variable includes both the pull-down stress applied when the resin is extruded and the elastic modulus of the resin molded body.

By using the drawdown stress as an explanatory variable, it is possible to easily estimate the shrinkage rate by taking into account both the material properties and the extrusion conditions, and to predict the shrinkage rate with high accuracy. However, if only the drawdown stress is used as an explanatory variable, the shrinkage rate prediction accuracy may be insufficient depending on the resin used and the production line. In contrast, in this embodiment, by using strain (strain data 75) that takes into account the elastic modulus in addition to the drawdown stress as an explanatory variable, it is possible to further improve prediction accuracy. For example, if the prediction error is large, it is possible that the extrusion conditions will be determined incorrectly, resulting in a decrease in productivity, but this embodiment makes it possible to prevent such problems.

By accurately predicting the shrinkage rate, it becomes possible to set appropriate extrusion conditions, reducing waste due to shrinkage after production, and thus improving yield. Furthermore, by setting the shrinkage rate to a sufficiently small value, it may be possible to eliminate the annealing process required to intentionally shrink the material after production.

(Summary of the embodiment)
Next, the technical ideas grasped from the above-described embodiments will be described by using the reference numerals and the like in the embodiments. However, the reference numerals and the like in the following description do not limit the components in the claims to the members and the like specifically shown in the embodiments.

[1] A device for predicting the shrinkage rate of a resin molded body formed by extrusion molding of a resin, comprising: a regression model creation processing unit (26) that uses the shrinkage rate of the resin molded body as a dependent variable and creates a regression model (33) that represents the correlation between a predetermined explanatory variable and the dependent variable; and a shrinkage rate prediction processing unit (27) that predicts the shrinkage rate of the prediction target using the regression model (33), wherein the explanatory variable includes both the drawing stress applied when the resin is extruded and the elastic modulus of the resin molded body.

[2] The shrinkage prediction device (1) described in [1] uses strain derived using the pull-down stress and the elastic modulus as the explanatory variable.

[3] The shrinkage prediction device (1) described in [1], in which both the pull-down stress and the elastic modulus are used as the explanatory variables.

[4] The shrinkage prediction device (1) described in [1], which is equipped with a pull-down stress calculation processing unit (23) that calculates the pull-down stress based on information regarding the viscosity of the resin and information regarding the extrusion conditions of the resin.

[5] The shrinkage factor prediction device (1) described in [1] further includes: an elastic modulus prediction regression model creation processing unit (101) that performs machine learning on the relationship between the resin extrusion conditions and the elastic modulus, and creates an elastic modulus prediction regression model (102) that represents the correlation between the resin extrusion conditions and the elastic modulus; and an elastic modulus prediction processing unit (103) that predicts the elastic modulus using the elastic modulus prediction regression model (102).

[6] A method for predicting the shrinkage rate of a resin molded body formed by extrusion molding of a resin, comprising: a regression model creation step of creating a regression model (33) that uses the shrinkage rate of the resin molded body as a dependent variable and expresses the correlation between a predetermined explanatory variable and the dependent variable; and a shrinkage rate prediction step of predicting the shrinkage rate of the prediction target using the regression model (33), wherein the explanatory variable includes both the drawing stress applied when the resin is extruded and the elastic modulus of the resin molded body.

[7] A method for manufacturing a resin molded body by extruding a resin to form a resin molded body, the method using the shrinkage rate prediction method described in [6] to derive extrusion conditions under which the shrinkage rate of the resin molded body will be a desired value, and the resin molded body is manufactured under the derived extrusion conditions.

(Additional Note)
Although the embodiments of the present invention have been described above, the invention according to the claims is not limited to the above-described embodiments. It should be noted that not all of the combinations of features described in the embodiments are necessarily essential to the means for solving the problems of the invention. Furthermore, the present invention can be appropriately modified and implemented within the scope of its spirit.

DESCRIPTION OF SYMBOLS 1...shrinkage rate prediction device 2...control unit 21...setting processing unit 22...data acquisition processing unit 23...drawdown stress calculation processing unit 24...strain calculation processing unit 25...learning data extraction processing unit 26...regression model creation processing unit 27...shrinkage rate prediction processing unit 28...prediction data presentation processing unit 3...storage unit 31...database 32...learning data 33...regression model 34...prediction source data 35...prediction data 71...extrusion condition data 72...viscosity data 73...elastic modulus data 74...drawdown stress data 75...strain data 76...shrinkage rate data 81...explanatory variable data 82...objective variable data 101...elastic modulus prediction regression model creation processing unit 102...elastic modulus prediction regression model 103...elastic modulus prediction processing unit

Claims (7)

An apparatus for predicting the shrinkage rate of a resin molded body formed by extrusion molding a resin, comprising:
a regression model creation processing unit that uses the shrinkage rate of the resin molded body as a response variable and creates a regression model that represents a correlation between a preset explanatory variable and the response variable;
a shrinkage rate prediction processing unit that predicts a shrinkage rate of a prediction target using the regression model,
The explanatory variables include both elements of the pull-down stress applied when the resin is extruded and the elastic modulus of the resin molded body.
Shrinkage prediction device. A strain derived using the pull-down stress and the elastic modulus is used as the explanatory variable.
The shrinkage rate prediction device according to claim 1 . Using both the pull-down stress and the elastic modulus as the explanatory variables;
The shrinkage rate prediction device according to claim 1 . a drawing stress calculation processing unit that calculates the drawing stress based on information about the viscosity of the resin and information about the extrusion conditions of the resin;
The shrinkage rate prediction device according to claim 1 . a processing unit for creating a regression model for predicting elastic modulus, which performs machine learning on the relationship between the extrusion conditions of the resin and the elastic modulus, and creates a regression model for predicting elastic modulus, which represents the correlation between the extrusion conditions of the resin and the elastic modulus;
Further provided is an elastic modulus prediction processing unit that predicts the elastic modulus using the elastic modulus prediction regression model,
The shrinkage rate prediction device according to claim 1 . A method for predicting the shrinkage rate of a resin molded body formed by extrusion molding a resin, comprising:
a regression model creation step of creating a regression model that uses the shrinkage rate of the resin molded body as a response variable and that represents a correlation between a predetermined explanatory variable and the response variable;
a shrinkage rate prediction step of predicting a shrinkage rate of a prediction target using the regression model,
The explanatory variables include both elements of the pull-down stress applied when the resin is extruded and the elastic modulus of the resin molded body.
Shrinkage prediction method. A method for manufacturing a resin molded body by extruding a resin to form a resin molded body,
deriving extrusion conditions under which the shrinkage rate of the resin molded body becomes a desired value using the shrinkage rate prediction method according to claim 6, and manufacturing the resin molded body under the derived extrusion conditions;
A method for manufacturing a resin molded product.