Method for determining manufacturing conditions of resin compositions

Abstract

To determine a control parameter which is hardly affected by destroy and deterioration and the like accompanying secondary processing, and makes impact resistance of a resin composition suitable.SOLUTION: A method for determining a production condition of a resin composition executes a machine learning algorithm, using a data set including production condition data including at least a blending component of a resin composition, a mixing condition, and a kneaded product temperature in melting and kneading as production condition items, first physical property measurement data including at least impact resistance of the resin composition produced by the production condition indicated by the production condition data as a characteristic value item, and second physical property measurement data obtained by measuring the resin composition produced under the production condition indicated by the production condition data in a non-destructive manner online, and thereby determines an item where importance for a change of the characteristic value of a characteristic improvement object item is high and the second physical property measurement data is within a predetermined range, when the characteristic value of the characteristic improvement object item of the resin composition is used as an objective variable.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a method for determining the manufacturing conditions of a resin composition. [Background technology]

[0002] Conventionally, there has been a technology for producing resin compositions by a melt-kneading process. In such a technology, two or more types of resin pellets are mixed, the mixed pellets are heated while being stirred by rotating them with a screw, and the kneaded resin is extruded to perform polymer blending (see, for example, Patent Document 1). [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2017-149002 [Overview of the project] [Problems that the invention aims to solve]

[0004] The melt-mixing process described above involves a very large number of parameters to control as manufacturing conditions, and these parameters interact with each other in a complex manner. Therefore, it has been difficult to find suitable manufacturing conditions to improve the impact resistance of the resin composition. Furthermore, resin compositions manufactured under conventional conditions may not be able to be properly evaluated for their properties due to the effects of fracture and degradation associated with secondary processing, making it difficult to determine suitable manufacturing conditions. In other words, with conventional technology, it has been difficult to optimize properties such as impact resistance due to the effects of fracture and degradation associated with secondary processing of the resin composition. Especially when performing online measurements, the time spent on secondary processing makes it even more difficult to determine suitable manufacturing conditions.

[0005] Therefore, the present invention aims to provide a method for determining manufacturing conditions for a resin composition that can determine control parameters that make the resin composition less susceptible to damage and deterioration associated with secondary processing and that provide suitable impact resistance. [Means for solving the problem]

[0006] As a result of diligent research to solve the above problems, the inventors of this invention have found that it is possible to determine suitable control parameters by using a dataset obtained from the acquired data described later and performing analysis using a machine learning algorithm, thereby completing the present invention.

[0007] In other words, the present invention has the following aspects.

[0008] (1) One aspect of the present invention is a method for determining the manufacturing conditions of a resin composition, in which, when the characteristic value of the item to be improved in the resin composition is used as the objective variable, a machine learning algorithm is executed using a dataset that includes manufacturing condition data which includes at least manufacturing condition items such as the blending components of the resin composition, mixing conditions and the temperature of the kneaded material during melt kneading; first physical property measurement data which includes at least the impact resistance of the resin composition manufactured under the manufacturing conditions indicated by the manufacturing condition data as a characteristic value item; and second physical property measurement data which is obtained by measuring the resin composition manufactured under the manufacturing conditions indicated by the manufacturing condition data online and non-destructively.

[0009] (2) One aspect of the present invention is a method for determining the manufacturing conditions of the resin composition described in (1) above, wherein the resin composition includes one or more resins selected from the group consisting of polyarylene sulfide resin, polyamide resin, polyester resin, polyphenylene oxide resin, polyetheretherketone resin, polycarbonate resin, and polystyrene resin.

[0010] (3) In one aspect of the present invention, in a method for determining the manufacturing conditions of the resin composition described in (1) or (2) above, the second physical property measurement data includes at least the scattering intensity obtained by Raman scattering measurement.

[0011] (4) One aspect of the present invention is a method for determining the manufacturing conditions of a resin composition described in any of (1) to (3) above, wherein the Raman scattering measurement includes at least the value of baseline intensity (INTENSITY) as an explanatory variable.

[0012] (5) One aspect of the present invention is a method for determining the manufacturing conditions of a resin composition described in any of (1) to (4) above, wherein the second physical property measurement data includes at least the diffuse reflectance or Kuberkamunck function obtained by near-infrared spectral measurement.

[0013] (6) One aspect of the present invention is a method for determining the manufacturing conditions of a resin composition described in any of (1) to (5) above, wherein the second physical property measurement data includes at least the diffuse reflectance or Kuberkamunck function obtained by near-infrared diffuse reflectance measurement.

[0014] (7) One aspect of the present invention is a method for determining the manufacturing conditions of a resin composition described in any of (1) to (6) above, wherein in the near-infrared spectral measurement, the measurement method is diffuse reflectance, and the scattering coefficient value based on the baseline position is included as at least an explanatory variable.

[0015] (8) One aspect of the present invention is a method for determining the manufacturing conditions of a resin composition as described in any of (1) to (7) above, wherein the manufacturing condition data includes, as manufacturing condition items, first manufacturing condition data which is controlled by the resin composition manufacturing apparatus and second manufacturing condition data which is not controlled by the manufacturing apparatus.

[0016] (9) One aspect of the present invention is a method for determining the manufacturing conditions of a resin composition according to any one of (1) to (8) above, wherein the second manufacturing condition data includes the internal temperature of each of the multiple kneading sections in the manufacturing apparatus where the resin is kneaded.

Advantages of the Invention

[0017] According to the present invention, it is possible to provide a method for determining the manufacturing conditions of a resin composition, which is hardly affected by destruction, deterioration, etc. associated with secondary processing and can determine control parameters that make the impact resistance of the resin composition suitable.

Brief Description of the Drawings

[0018] [Figure 1] It is a diagram for explaining the functional configuration of the twin-screw extruder according to the present embodiment. [Figure 2] It is a diagram for explaining the machine learning algorithm according to the present embodiment. [Figure 3] It is a diagram for explaining the dataset according to the present embodiment. [Figure 4] It is a diagram for explaining an example of the importance calculated by the machine learning algorithm according to the present embodiment. [Figure 5] It is a diagram showing the relationship between the measurement result by the spectroscopic sensor according to the present embodiment and the Charpy impact value. [Figure 6] It is a flowchart for explaining a series of processes of the method for determining the manufacturing conditions of the resin composition according to the present embodiment.

Modes for Carrying Out the Invention

[0019] Hereinafter, a method for determining the manufacturing conditions of a resin composition according to an embodiment of the present invention will be described with reference to the drawings. In the following description, as an example of the resin composition, an example in the case where the resin composition is a polyarylene sulfide resin composition will be described. However, the resin composition according to the present embodiment is not limited to this example and widely includes various resin compositions. As an example of various resin compositions, those containing resins such as polyamide resin, polyester resin, polyphenylene oxide resin, polyether ether ketone resin, polycarbonate resin, and polystyrene resin may be included. Details of the resin composition in the embodiment will be described later.

[0020] [Overview of a twin-screw extruder] Figure 1 is a diagram illustrating the functional configuration of the twin-screw extruder 10 according to this embodiment. The functional configuration of the twin-screw extruder 10 according to this embodiment will be explained with reference to this figure. The twin-screw extruder 10 comprises a drive unit 11, a feeder 12, a cylinder 13, a screw 14, and an infrared temperature sensor IR.

[0021] The feeder 12 is an inlet for introducing the raw materials of the polyarylene sulfide resin composition according to the embodiment. In this embodiment, the term "polyarylene sulfide resin composition" broadly encompasses resin compositions containing polyarylene sulfide. The raw materials of the polyarylene sulfide resin composition may include polyarylene sulfide. Polyphenylene sulfide is an example of polyarylene sulfide. The raw materials of the polyarylene sulfide resin composition include the compounding components that constitute the polyarylene sulfide resin composition, and precursors of the compounding components.

[0022] The raw materials for the polyarylene sulfide resin composition may include any additional components that are mixed with polyarylene sulfide. When mixing polyarylene sulfide with an elastomer, the elastomer is also fed in from the feeder 12. The number of feeders 12 may be one, or there may be two or more to individually feed in the raw materials for the polyarylene sulfide resin composition.

[0023] The cylinder 13 has a cylindrical shape. One end of the cylinder 13 is connected to the feeder 12, and the other end is connected to the die 19. In the following description, the feeder 12 side of the cylinder 13 may be referred to as the upstream side, and the die 19 side as the downstream side. The cylinder 13 houses a screw 14 inside. The raw material introduced into the feeder 12 is heated by a heater (not shown), and at least a portion of the raw material (e.g., polyarylene sulfide) melts inside the cylinder 13, while the entire introduced raw material is kneaded by the screw 14. In the following description, the cylinder 13 will also be referred to as the kneading section. In the kneading section, the raw materials for the polyarylene sulfide resin composition are kneaded. Hereinafter, the contents inside the cylinder during melt kneading will be referred to as the kneaded material.

[0024] In this embodiment, multiple heating units (not shown) may be installed at different positions along the x-axis. Multiple heating units installed at different positions along the x-axis heat the cylinder 13 at different temperatures. The twin-screw extruder 10 has multiple heating units and heats the cylinder 13 at different temperatures, thereby heating the kneaded material at different temperatures on the upstream and downstream sides of the cylinder 13. An example of a heating unit is a barrel that covers the cylinder 13.

[0025] The infrared temperature sensor IR measures the temperature of the cylinder 13. Specifically, the infrared temperature sensor IR measures the temperature of the molded material inside the cylinder 13 during the melting and kneading process. The twin-screw extruder 10 may be equipped with multiple infrared temperature sensors IR. In this embodiment, the twin-screw extruder 10 is equipped with the following infrared temperature sensors IR, in order from the upstream side of the cylinder 13: a first infrared temperature sensor IR1, a second infrared temperature sensor IR2, a third infrared temperature sensor IR3, and a fourth infrared temperature sensor IR4.

[0026] The screw 14 is rotationally driven by the drive unit 11. By rotating, the screw 14 guides the kneaded material inside the cylinder from the upstream side to the downstream side. The raw material fed into the feeder 12 passes through the cylinder 13 and is extruded from the die 19 as a polyarylene sulfide resin composition obtained by melt kneading. The drive unit 11 includes a motor and gearbox (not shown). The drive unit 11 controls the motor's rotational speed and torque, etc., to rotate the screw 14 so that the screw rotational speed (rpm) is predetermined.

[0027] [Machine learning algorithms] Figure 2 is a diagram illustrating the machine learning algorithm 20 according to this embodiment. The machine learning algorithm 20 will be explained with reference to this figure. The machine learning algorithm 20 is trained using supervised learning with the dataset DS stored in the memory device 30 as training data. The machine learning algorithm 20 also performs inference based on the information stored in the memory device 30 and outputs the inference results to the memory device 30. The memory device 30 stores manufacturing condition data CD, physical property measurement data MD, and high-priority items HC.

[0028] The manufacturing conditions data CD contains data related to the manufacturing conditions of the resin composition. The manufacturing conditions data CD includes at least the following manufacturing conditions: the compositional components of the resin composition, the mixing conditions, and the temperature of the mixture during melt kneading. The manufacturing conditions data CD includes a first manufacturing conditions data CD1 and a second manufacturing conditions data CD2. The first manufacturing condition data CD1 includes controllable variables among the data relating to the manufacturing conditions of the resin composition. Examples of the first manufacturing condition data CD1 include the amount of elastomer modification, the amount of elastomer blended, and the screw rotation speed. The second manufacturing condition data CD2 includes uncontrollable (in other words, variable) variables related to the manufacturing conditions of the resin composition. Examples of the second manufacturing condition data CD2 include the current value required to rotate the screw, the pressure of the mixed material, and the internal temperature of the cylinder.

[0029] The physical property measurement data MD includes data relating to the results of measuring the physical properties of a resin composition manufactured under the manufacturing conditions included in the manufacturing conditions data CD. The physical property measurement data MD includes the first physical property measurement data MD1 and the second physical property measurement data MD2. The first physical property measurement data MD1 includes, as a characteristic value item, at least the impact resistance (e.g., Charpy impact value) of the resin composition manufactured under the manufacturing conditions indicated in the manufacturing conditions data CD, from the physical property measurement data MD. The second physical property measurement data MD2 includes at least measurement data obtained online and non-destructively from the physical property measurement data MD, which is obtained from the resin composition manufactured under the manufacturing conditions indicated in the manufacturing conditions data CD. Examples of online and non-destructive measurement data include measurements using a spectroscopic sensor and the degree of damping when vibration is applied to the sample. Examples of spectroscopic sensors include Raman sensors and near-infrared (NIR) spectroscopic sensors.

[0030] As an example, machine learning algorithm 20 calculates the importance of each item to the Charpy impact value, using the Charpy impact value (Charpy impact intensity) as the dependent variable and items other than the Charpy impact value as independent variables. In the following explanation, the dependent variable set as the target for improving the characteristics will also be referred to as the item to be improved. High-importance items HC are items whose importance is high in terms of changes in characteristic values ​​when the characteristic values ​​of the items targeted for improvement in the properties of the resin composition are used as the objective variable. In the following explanation, the impact resistance of the resin composition will be described as an example of an item targeted for improvement in properties. However, this embodiment is not limited to this example, and characteristic values ​​other than impact resistance may be used as the objective variable. For example, the measurement results measured by a Raman sensor, near-infrared spectrometer, etc., included in the second physical property measurement data MD2 may be used as the objective variable.

[0031] The machine learning algorithm 20 receives the dataset DS from the storage device 30 as input. Based on the input dataset DS, the machine learning algorithm 20 calculates high-importance items HC. The calculated high-importance items HC are stored in the storage device 30. The learning and inference processes using the machine learning algorithm 20 are described in detail below, for each stage.

[0032] First, the learning phase of the machine learning algorithm 20 will be explained. The machine learning algorithm 20 is learned by supervised learning. The dataset DS used for learning is created in advance by measuring the physical properties of a resin composition manufactured under predetermined manufacturing conditions. The dataset DS used for learning includes manufacturing condition data CD and physical property measurement data MD, which includes measurement results of the physical properties of a resin composition manufactured under the manufacturing conditions specified by the manufacturing condition data CD.

[0033] Next, the inference stage of the machine learning algorithm 20 will be described. In the method for determining the manufacturing conditions of the resin composition in this embodiment, the machine learning algorithm is executed using the dataset DS to determine high-importance items HC, which are items of high importance for improving properties. In the method for determining the manufacturing conditions of the resin composition in this embodiment, when the characteristic value of the item to be improved in the resin composition is used as the objective variable, the method determines items among a plurality of items included in the manufacturing condition data CD and physical property measurement data MD that are of high importance for the change in the characteristic value of the item to be improved and for which the second physical property measurement data MD2 is within a predetermined range. If the second physical property measurement data MD2 includes measurement results measured by a Raman sensor, near-infrared spectrometer, etc., the method for determining the manufacturing conditions of the resin composition in this embodiment may determine items for which either the measurement result measured by the Raman sensor or the measurement result measured by the near-infrared spectrometer is within a predetermined range. Alternatively, the method for determining the manufacturing conditions of the resin composition in this embodiment may determine items for which both the measurement result measured by the Raman sensor and the measurement result measured by the near-infrared spectrometer are within a predetermined range.

[0034] The machine learning algorithm 20 used in the method for determining the manufacturing conditions of the resin composition in this embodiment will be described in more detail. First, the machine learning algorithm 20 calculates the importance (%) for each item included in the acquired dataset DS. Specifically, the machine learning algorithm 20 calculates the importance using the random forest method. The machine learning algorithm 20 analyzes all control variables and measured variables included in the manufacturing condition data CD and all physical property variables included in the physical property measurement data MD from the acquired dataset DS, and calculates their importance (%). Here, the hyperparameters used to calculate importance can be adjusted as needed to maximize the coefficient of determination (Score). Alternatively, optimized values ​​obtained using grid search or Bayesian optimization methods may be used.

[0035] [Example of a dataset] Figure 3 is a diagram illustrating the dataset DS according to this embodiment. An example of the dataset DS according to this embodiment will be described with reference to this figure. The dataset DS includes manufacturing condition data CD and physical property measurement data MD.

[0036] The manufacturing conditions data CD includes the components of the resin composition, mixing conditions, and the temperature of the mixture during melt kneading. Specifically, the manufacturing conditions data CD includes two manufacturing condition items: first manufacturing conditions data CD1, which is controlled by the resin composition manufacturing equipment (i.e., a control variable), and second manufacturing conditions data CD2, which is not controlled by the resin composition manufacturing equipment (i.e., an eventual variable).

[0037] Examples of the first manufacturing condition data CD1 include the amount of elastomer modification, the amount of elastomer blended, and the screw rotation speed. The amount of elastomer modification corresponds to, for example, the amount of functional groups that thermoplastic elastomer (B) may have. The amount of elastomer blended can be the percentage (mass%) of elastomer content relative to the total mass of the polyarylene sulfide resin composition. The screw rotation speed is the screw rotation speed (rpm) of the twin-screw extruder 10 described above, and is determined by the drive device 11 controlling the motor speed, torque, etc.

[0038] Examples of the second manufacturing condition data CD2 include current, kneading pressure, and cylinder internal temperature. Current refers to the current required to rotate the screw 14 when extruding the kneading material from the cylinder 13. As a variation, the extrusion torque may be used as the measured variable instead of current. Kneading pressure is a value measured by a pressure sensor (not shown) located in the die section downstream of the cylinder 13.

[0039] The cylinder internal temperature is the value measured by the infrared temperature sensor IR described above. There may be multiple cylinder internal temperatures depending on the position of the cylinder. For example, the cylinder internal temperature may include multiple temperature information measured by the first infrared temperature sensor IR1 to the fourth infrared temperature sensor IR4. That is, the second manufacturing condition data CD2 may include the temperature at multiple locations in the kneading section (cylinder 13) where the resin composition is kneaded. Note that the temperature at multiple locations in the kneading section includes the internal temperature of the kneading section, specifically the temperature of the kneaded material or the resin temperature.

[0040] In this embodiment, items from the second manufacturing condition data CD2, which are measured variables, that are determined to be important may be used as the first manufacturing condition data CD1, which are control variables, or items from the first manufacturing condition data CD1, which are control variables, that are determined to be unimportant may be used as the second manufacturing condition data CD2, which are measured variables.

[0041] The physical property measurement data MD includes characteristic values ​​of the resin composition manufactured under the manufacturing conditions indicated in the manufacturing conditions data CD. Specifically, the physical property measurement data MD includes first physical property measurement data MD1, which is data obtained by offline and destructive testing, and second physical property measurement data MD2, which is data obtained by online and non-destructive testing.

[0042] Examples of the first physical property measurement data MD1 include melt viscosity, elastomer mean dispersion diameter, and Charpy impact value. The first physical property measurement data MD1 may also include other values ​​that can serve as targets for improving properties, such as gas generation amount, high-temperature heat resistance, and high-temperature modulus.

[0043] Melt viscosity is measured by applying a load to the obtained resin composition. For example, melt viscosity may be measured by placing pellets of the resin composition into a flow tester with a cylinder temperature of 300°C, an orifice length of 10 mm, and an orifice diameter of 1 mm, preheating for 6 minutes, and then applying a load of 50 kg.

[0044] The average dispersion diameter of the elastomer is a value measured from the obtained resin composition. Specifically, the average dispersion diameter of the elastomer may be obtained by using the obtained resin composition as a molding material, molding a multipurpose test piece with an injection molding machine, cutting the molded multipurpose test piece at the center in the longitudinal direction, polishing the cut surface, immersing it in xylene, performing ultrasonic treatment at a temperature of 50°C, removing the elastomer dispersion in the cross-section by xylene extraction, then drying it at 130°C for 2 hours, observing the cross-section with an SEM, and measuring the image. In this case, the areas where the elastomer has been removed become void phases and are displayed as dark black circles. Alternatively, the average dispersion diameter of the elastomer may be obtained by measuring the diameter equivalent to the area of ​​a circle (the value obtained by calculating the diameter of a true circle equivalent to the area of ​​a circle) of all such black circular objects observed in the image field of view using image analysis software and dividing by the number of circular objects by the average value obtained by dividing by the number of circular objects.

[0045] The Charpy impact value is a value measured from the obtained resin composition. The Charpy impact value may be measured, for example, by using the obtained resin composition as a molding material, molding it in an injection molding machine under the conditions of a cylinder setting temperature of 300°C and a mold setting temperature of 130°C to obtain a test piece measuring 80 mm in length, 10.0 mm in width, and 4.0 mm in thickness, then cutting a notch into the test piece according to ISO 2818, and conducting a test at 23°C according to ISO 179-1.

[0046] Examples of the second physical property measurement data MD2 include NIR diffuse reflectance and Raman scattering. NIR diffuse reflectance is a value measured by an NIR sensor, and is the diffuse reflectance or Kuberkamunck function obtained by near-infrared spectral measurement. In other words, the second physical property measurement data MD2 includes at least the diffuse reflectance or Kuberkamunck function obtained by near-infrared spectral measurement. Furthermore, NIR diffuse reflectance is a value that represents the diffuse reflectance or Kuberkamunck function obtained by near-infrared diffuse reflectance measurement. In other words, the second physical property measurement data MD2 includes at least the diffuse reflectance or Kuberkamunck function obtained by near-infrared diffuse reflectance measurement.

[0047] Furthermore, the measurement method used in the near-infrared spectral measurement may be the diffuse reflectance method. In addition, the near-infrared spectral measurement may include at least the diffuse reflectance or Kuberkamunck function based on the baseline position as an explanatory variable.

[0048] Raman scattering is a value measured by a Raman sensor, representing the scattering intensity determined by Raman scattering measurement. That is, the second physical property measurement data MD2 includes at least the scattering intensity determined by Raman scattering measurement. Furthermore, the baseline intensity (INTENSITY) value may be included as at least an explanatory variable in the said Raman scattering measurement.

[0049] In the example shown in Figure 3, dataset DS1 and dataset DS2 are shown as dataset DS. Dataset DS1 and dataset DS2 show the measurement results for resin compositions manufactured under different manufacturing conditions. Dataset DS may include measurement results for resin compositions manufactured under multiple manufacturing conditions, and it is preferable to have a large number of measurement results. Furthermore, the manufacturing condition data CD and physical property measurement data MD may use the average value based on the results of multiple measurements.

[0050] Dataset DS1 has the following values: elastomer modification amount is "CD11_1", elastomer blending amount is "CD12_1", screw rotation speed is "CD13_1", current is "CD21_1", kneading pressure is "CD22_1", cylinder internal temperature is "CD23_1", melt viscosity is "MD11_1", elastomer mean dispersion diameter is "MD12_1", Charpy impact value is "MD13_1", NIR diffuse reflectance is "MD21_1", and Raman scattering is "MD22_1". Dataset DS2 has the following values: elastomer modification amount is "CD11_2", elastomer blending amount is "CD12_2", screw rotation speed is "CD13_2", current is "CD21_2", kneading pressure is "CD22_2", cylinder internal temperature is "CD23_2", melt viscosity is "MD11_2", elastomer mean dispersion diameter is "MD12_2", Charpy impact value is "MD13_2", NIR diffuse reflectance is "MD21_2", and Raman scattering is "MD22_2".

[0051] [Example of importance level] Figure 4 is a diagram illustrating an example of importance calculated by the machine learning algorithm 20 according to this embodiment. An example of importance calculated by the machine learning algorithm 20 will be explained with reference to this figure. The machine learning algorithm 20 calculates importance for each item corresponding to the dataset DS. In the example shown in the figure, the Charpy impact value is labeled "target" because it is the result of calculating the importance of each item in relation to the Charpy impact value. Specifically, the importance of each item in relation to the Charpy impact value is "I1" for elastomer modification amount, "I2" for elastomer blending amount, "I3" for screw rotation speed, "I4" for current, "I5" for compound pressure, "I6" for cylinder internal temperature, "I7" for melt viscosity, "I8" for elastomer dispersion diameter, "I9" for NIR diffuse reflectance, and "I10" for Raman scattering. Note that these importance values ​​may also be calculated by multiplying each calculated value by 100 so that the sum is 100.

[0052] Next, the machine learning algorithm 20 determines the high-importance items HC using a predetermined method. Specifically, the machine learning algorithm 20 calculates the importance of each of the multiple items included in the manufacturing condition data CD and the physical property measurement data MD, thereby determining the items that are of high importance (i.e., high-importance items HC) regarding the change in the characteristic value of the item to be improved (in this example, the Charpy impact value).

[0053] In the example shown in Figure 4, the high-importance items HC are the elastomer content and the cylinder internal temperature. In other words, to improve the Charpy impact value characteristics, it is preferable to appropriately control the elastomer content and the cylinder internal temperature. Here, the items determined to be high-importance items HC by the machine learning algorithm 20 are those that have a high importance in terms of changes in the characteristic values ​​of the items to be improved, and whose second physical property measurement data MD2 is within a predetermined range. The predetermined range used to determine the second physical property measurement data MD2 may be predetermined or may be learned by the machine learning algorithm 20.

[0054] Furthermore, if the internal cylinder temperature is divided into multiple locations within the cylinder, the location with the highest importance may be designated as the high-importance item HC. For example, if the location on the upstream side of the cylinder is deemed to have high importance, it indicates that among the temperatures of the mixed material at multiple locations in the mixing section, the upstream side where the raw materials for the resin composition are introduced into the mixing section is of higher importance than the downstream side where the mixed resin composition is extruded.

[0055] A predetermined method for determining high-importance items HC may be configured such that, among items whose second physical property measurement data MD2 falls within a predetermined range, the items with the highest calculated importance (e.g., one or more items) are designated as high-importance items HC. As a variation, the machine learning algorithm 20 may be configured to determine whether an item is a high-importance item (HC) by comparing its importance with a predetermined threshold. For example, the predetermined threshold may be set to 10, and items whose importance is equal to or greater than the threshold may be designated as high-importance items (HC).

[0056] Next, the machine learning algorithm 20 uses the calculated high-importance items as a new target variable and determines which items are highly important in relation to the change in the characteristic value of the new target variable. Specifically, in the example shown in Figure 4, since elastomer content and cylinder internal temperature are high-importance items HC, the elastomer content, which is one of the high-importance items HC, is used as the target variable, and the items other than elastomer content are used as explanatory variables, and the importance of each item in relation to elastomer content is calculated using the random forest algorithm.

[0057] As described above, according to the machine learning algorithm 20, a predetermined item in the dataset DS is used as the target variable, and the items other than the target variable are used as explanatory variables. The importance of each item relative to the target variable is calculated using the random forest algorithm. The machine learning algorithm 20 then determines the items with high importance by further calculating the importance of a specific item among the calculated high-importance items HC, again using the random forest algorithm, with that item as the target variable.

[0058] Next, the machine learning algorithm 20 performs support vector regression on the items determined to be high-importance items (HC). Specifically, the machine learning algorithm 20 uses the calculated high-importance items (high-importance items HC) as the analysis axis and performs a regression operation using the dataset DS to estimate the correspondence between the change in the characteristic value of the high-importance items (high-importance items HC) and the change in the characteristic value of the target variable. Here, the items determined to be high-importance items (HC) may be determined comprehensively from the items determined in the first operation and the items determined in the second and subsequent operations. The hyperparameters used when performing support vector regression may be changed as appropriate to maximize the coefficient of determination (Score). Alternatively, optimized values ​​may be used using grid search or Bayesian optimization programs. Furthermore, by performing a regression calculation using the dataset DS with the high-importance item HC and the second physical property measurement item MD2 as the analysis axes, it is possible to estimate the correspondence between the changes in the characteristic values ​​of the high-importance items (high-importance item HC), the changes in the characteristic values ​​of the second physical property measurement item MD2, and the changes in the characteristic values ​​of the objective variable.

[0059] [Relationship between Charpy impact value and measurement values ​​from a spectroscopic sensor] Figure 5 shows the relationship between the measurement results from the spectroscopic sensor according to this embodiment and the Charpy impact value. Figures 5(A) and 5(B) are both three-dimensional graphs created from the same measurement results. Figures 5(A) and 5(B) are graphs viewed from different angles. The x-axis shows the measurement results by Raman, the y-axis shows the measurement results by NIR, and the z-axis shows the Charpy impact value. Referring to the figure, the relationship between the Charpy impact value and the measured values ​​(NIR and Raman) from the spectroscopic sensor will be explained. The figure shows the measurement results for a molded product containing 20% ​​elastomer.

[0060] The measured values ​​of Raman, NIR, and Charpy impact values ​​are shown as the measurement results (meas). The results of support vector regression analysis based on these measurement results are also shown as the fit. As shown in Figure 5, a smaller Raman sensor measurement and a smaller NIR sensor measurement result indicate a higher Charpy impact value, which represents impact resistance. If the NIR range for a good molded product is defined as the threshold TH_NIR, then TH_NIR may be, for example, 40. Similarly, if the Raman range for a good molded product is defined as the threshold TH_RAMAN, then TH_RAMAN may be, for example, 300. Since the NIR sensor measurement indicates the dispersion diameter, a smaller value is better, and values ​​below the threshold are considered good. Similarly, since the Raman sensor measurement indicates thermal degradation, a smaller value is better, and values ​​below the threshold are considered good. From the results shown in the figure, it can be seen that when both the Raman and NIR values ​​are small, the Charpy impact value is good (high Charpy impact value).

[0061] Next, based on the measurement results, we will explain how the Charpy impact value and the measurement results from the spectroscopic sensor change in accordance with the screw rotation speed (not shown). The Charpy impact value is 46.1 kJ / m when the screw rotation speed is 300 rpm. 2 ] and when the screw rotation speed is 600 [rpm], it is 54.0 [kJ / m 2 ] and when the screw rotation speed is 1000 [rpm], it is 58.6 [kJ / m 2 ] and when the screw rotation speed is 1500 [rpm], it is 55.4 [kJ / m 2 ] and when the screw rotation speed is 2000 [rpm], it is 45.8 [kJ / m 2 ] and when the screw rotation speed is 2500 [rpm], it is 35.0 [kJ / m 2 ] and when the screw rotation speed is 3000 [rpm], it is 26.1 [kJ / m 2 In other words, the Charpy impact value peaks at a screw rotation speed of 1000 rpm and increases in the range of 500 rpm to 1500 rpm.

[0062] The NIR is 57.9 when the screw rotation speed is 300 rpm, 53.0 when the screw rotation speed is 600 rpm, 39.8 when the screw rotation speed is 1000 rpm, 35.2 when the screw rotation speed is 1500 rpm, 30.7 μm when the screw rotation speed is 2000 rpm, 29.7 when the screw rotation speed is 2500 rpm, and 38.2 when the screw rotation speed is 3000 rpm. In other words, the NIR is greater than the threshold of 40 when the screw rotation speed is 1000 rpm or less.

[0063] Raman is obtained from the Raman spectrum obtained from a laser with an excitation wavelength of 785 nm, with a Raman shift of 300 [cm²]. -1This indicates the intensity (intensity (au)) of scattered light at the position ]. Specifically, it was 27924 when the screw rotation speed was 300 [rpm], 27254 when the screw rotation speed was 600 [rpm], 28525 when the screw rotation speed was 1000 [rpm], 30601 when the screw rotation speed was 1500 [rpm], 36725 when the screw rotation speed was 2000 [rpm], 41432 when the screw rotation speed was 2500 [rpm], and 40666 when the screw rotation speed was 3000 [rpm]. In other words, the Raman value is below the threshold when the screw rotation speed is around 1400 [rpm] or less.

[0064] Therefore, it can be seen that the range in which the Charpy impact value is large and the NIR and Raman values ​​are below the threshold is the range in which the screw rotation speed is between 1000 [rpm] and 1400 [rpm].

[0065] [A series of steps for determining the manufacturing conditions of a resin composition] Figure 6 is a flowchart illustrating the sequence of steps for determining the manufacturing conditions of the resin composition according to this embodiment. The sequence of steps for determining the manufacturing conditions of the resin composition according to this embodiment will be explained with reference to this figure. (Step S110) The machine learning algorithm 20 acquires the dataset DS from the storage device 30 using a predetermined communication method. (Step S120) The machine learning algorithm 20 calculates importance using a random forest based on the acquired dataset DS. In this process, the machine learning algorithm 20 calculates importance for all control variables, observed variables, and physical properties included in the dataset DS. (Step S130) The machine learning algorithm 20 selects items from among the calculated importance levels that have high importance and yield a favorable range from the second physical property measurement data MD2, according to predetermined conditions.

[0066] (Step S140) The machine learning algorithm 20 determines whether or not to set an additional target variable based on predetermined conditions. If the machine learning algorithm 20 decides to set an additional target variable and calculate its importance (i.e., Step S140; YES), it proceeds to Step S120. If the machine learning algorithm 20 decides not to set an additional target variable and calculate its importance (i.e., Step S140; NO), it proceeds to Step S150. The predetermined conditions for the machine learning algorithm 20 to determine whether or not to further set a target variable may be obtained by presenting the calculated high-importance items HC to the user and receiving a response from the user. Furthermore, the predetermined conditions for the machine learning algorithm 20 to determine whether or not to set an additional target variable may be automatically determined based on the difference between the item ranked 1st in importance and the item ranked 2nd in importance.

[0067] (Step S150) The machine learning algorithm 20 obtains suitable manufacturing conditions by performing support vector regression analysis with the desired characteristic as the objective variable and items including the calculated high-importance item HC and the second physical property measurement data MD2 as explanatory variables. (Step S160) The machine learning algorithm 20 determines the improvement conditions. Furthermore, the machine learning algorithm 20 may be configured to determine improvement conditions by obtaining a response from the user in response to presenting the results of the support vector regression analysis performed in step S150 to the user.

[0068] [Summary of Embodiments] According to the embodiments described above, the method for determining the manufacturing conditions of the resin composition involves executing a machine learning algorithm 20 using a dataset DS to determine items for which the change in characteristic values ​​of the items to be improved is of high importance and for which the second physical property measurement data MD2 is within a predetermined range. The dataset DS includes manufacturing condition data CD and physical property measurement data MD. The manufacturing condition data CD includes at least the compounding components of the resin composition, mixing conditions, and the temperature of the kneaded material during melt kneading. The physical property measurement data MD includes first physical property measurement data MD1 and second physical property measurement data MD2. The first physical property measurement data MD1 includes at least the impact resistance of the resin composition manufactured under the manufacturing conditions indicated in the manufacturing condition data CD as a characteristic value item. The second physical property measurement data MD2 includes data obtained by measuring the resin composition manufactured under the manufacturing conditions indicated in the manufacturing condition data CD online and non-destructively.

[0069] Conventionally, the number of parameters controlled as manufacturing conditions was very large, and each parameter interacted in a complex manner, making it difficult to find suitable manufacturing conditions to impart higher impact resistance to resin compositions. Furthermore, resin compositions manufactured under conventional manufacturing conditions sometimes could not be properly evaluated due to the effects of fracture and degradation associated with secondary processing. In other words, with conventional technology, it was difficult to optimize properties such as impact resistance of resin compositions due to the effects of fracture and degradation associated with secondary processing. However, the method for determining the manufacturing conditions of the resin composition according to this embodiment determines items for which the change in characteristic values ​​of the items to be improved is of high importance and for which the second physical property measurement data MD2 is within a predetermined range. Therefore, it is possible to determine control parameters that make the properties of the resin composition suitable without being affected by the impact resistance of the resin composition or by fracture or deterioration due to secondary processing.

[0070] Furthermore, according to the embodiments described above, the resin composition may contain one or more resins selected from the group consisting of polyarylene sulfide resin, polyamide resin, polyester resin, polyphenylene oxide resin, polyetheretherketone resin, polycarbonate resin, and polystyrene resin. Therefore, the method for determining the manufacturing conditions of the resin composition according to this embodiment can determine the manufacturing conditions of various resin compositions as described above.

[0071] Furthermore, according to the embodiments described above, the second physical property measurement data MD2 includes at least the scattering intensity obtained by Raman scattering measurement. Raman scattering measurement enables non-destructive evaluation of the material. According to this embodiment, the machine learning algorithm 20 determines items for which the change in characteristic values ​​of the item to be improved is of high importance and for which the second physical property measurement data is within a predetermined range. Therefore, according to this embodiment, it is possible to determine manufacturing conditions for which the change in characteristic values ​​of the item to be improved is of high importance and which are resistant to thermal degradation.

[0072] Furthermore, according to the embodiments described above, in Raman scattering measurements, the baseline intensity (INTENSITY) value may be included as at least one explanatory variable, and the value of the light intensity of the baseline intensity (INTENSITY) may also be included as at least one explanatory variable. In particular, aromatic polymers such as polyarylene sulfide may produce fluorescent substances as decomposition products due to thermal reactions, and by adopting the light intensity value, it is possible to non-destructively evaluate the thermal degradation of the material. In addition, the light intensity value correlates with the IR1 temperature, making it possible to non-destructively predict the deterioration of properties such as impact resistance due to the thermal degradation of the material. Furthermore, in the embodiment, the full width at half maximum (FWHM) of the peaks originating from the main chain structure of the resin, such as the polyarylene sulfide resin, obtained by Raman scattering measurement, may be included as at least one explanatory variable. When the degree of crystallinity of the crystalline structure contained in the resin decreases, an increase in the FWHM value is observed, and by adopting the FWHM value, non-destructive evaluation of the crystallinity of the material is possible.

[0073] Furthermore, according to the embodiments described above, the second physical property measurement data MD2 includes at least the diffuse reflectance or Kuberkamunck function obtained by near-infrared spectral measurement. A value including the diffuse reflectance or Kuberkamunck function (for example, the reciprocal of the Kuberkamunck function, 1 / f) shows a high correlation with the particle size of the dispersed phase of the component contained in the resin composition as a dispersed phase. Therefore, it can serve as a non-destructive substitute evaluation that determines that the better the particle size of the dispersed phase, the better the impact resistance.

[0074] Furthermore, according to the embodiments described above, the second physical property measurement data MD2 includes at least the diffuse reflectance or Kuberkamunck function obtained by near-infrared diffuse reflectance measurement. Near-infrared diffuse reflectance measurement enables non-destructive evaluation of the dispersion state of components contained in the resin composition. According to this embodiment, the machine learning algorithm 20 determines items for which the change in characteristic values ​​of the item to be improved is of high importance and for which the second physical property measurement data is within a predetermined range. Therefore, according to this embodiment, it is possible to determine manufacturing conditions for which the change in characteristic values ​​of the item to be improved is of high importance and which are highly uniform (for example, the dispersion diameter of the resin that becomes the dispersed phase becomes smaller).

[0075] Furthermore, according to the embodiments described above, in near-infrared spectral measurement, the measurement method is diffuse reflectance, and the scattering coefficient value based on the baseline position is included as at least one explanatory variable. As a result, even in opaque materials, it is possible to non-destructively determine that better particle size of the dispersed phase leads to better impact resistance, and it can be used as an explanatory variable instead of the particle size evaluation of the dispersed phase, which is time-consuming to measure.

[0076] Furthermore, according to the embodiments described above, the manufacturing condition data CD includes control variables that are controlled by the resin composition manufacturing apparatus and measured variables that are not controlled by the manufacturing apparatus. In other words, according to this embodiment, even when parameters that change due to circumstances and are not controlled as control parameters are important, those parameters that change due to circumstances can be determined as high-importance items HC.

[0077] Furthermore, according to the embodiment described above, the measured variables include the internal cylinder temperature at multiple locations within the cylinder 13. That is, the machine learning algorithm 20 uses the internal cylinder temperature at multiple locations within the cylinder 13 (temperature of the mixed material during melting and kneading) as measured variables to calculate the items with high importance regarding the change in characteristic values ​​of the items targeted for improvement. Therefore, according to this embodiment, the machine learning algorithm 20 can calculate the high-importance item HC with high accuracy.

[0078] Furthermore, according to the embodiment described above, among the internal cylinder temperatures (temperature of the mixed material during melting and kneading) at multiple locations within the cylinder 13 included in the measured variables, the upstream side of the cylinder 13 is of higher importance than the downstream side. In other words, according to this embodiment, important parameters that previously could only be identified by relying on the experience and intuition of skilled technicians can be determined as control parameters.

[0079] Furthermore, the entirety or a part of the functions of each part of the machine learning algorithm 20 in the above-described embodiment may be realized by recording a program for realizing these functions on a computer-readable recording medium, having a computer system read the program recorded on this recording medium, and executing it. The term "computer system" here includes hardware such as an operating system and peripheral devices.

[0080] Furthermore, "computer-readable recording media" refers to portable media such as magneto-optical disks, ROMs, and CD-ROMs, as well as storage units such as hard disks built into computer systems. In addition, "computer-readable recording media" may also include those that dynamically hold programs for a short period of time, such as communication lines used when transmitting programs over a network such as the Internet, and those that hold programs for a certain period of time, such as volatile memory inside computer systems that act as servers or clients in such cases. Moreover, the above-mentioned program may be for the purpose of realizing some of the functions described above, and may also be able to realize the above-mentioned functions in combination with programs already recorded in the computer system.

[0081] [Resin composition] Examples of resin compositions according to this embodiment include those containing resins such as polyamide resin, polyester resin, polyphenylene oxide resin, polyetheretherketone resin, polycarbonate resin, and polystyrene resin.

[0082] The resin composition according to the embodiment is preferably a resin composition containing one or more resins selected from the group consisting of polyamide resins, polyester resins, polyphenylene oxide resins, polyetheretherketone resins, polycarbonate resins, and polystyrene resins, and a thermoplastic elastomer (B).

[0083] In particular, the resin composition according to the embodiment is more preferably a polyarylene sulfide resin composition containing a polyarylene sulfide resin (A) and a thermoplastic elastomer (B).

[0084] <Polyarylene sulfide resin (A)> The polyarylene sulfide resin (A) according to the embodiment has a resin structure having a structure in which an aromatic ring and a sulfur atom are bonded as a repeating unit. Specifically, examples thereof include a resin having a structural site represented by the following structural formula (1) as a repeating unit.

[0085]

Chemical formula

[0086] (In the formula, R 1 and R 2 each independently represent a hydrogen atom, an alkyl group having 1 to 4 carbon atoms, a nitro group, an amino group, a phenyl group, a methoxy group, or an ethoxy group.)

[0087] Here, the structural site represented by the structural formula (1) is such that R 1 and R 2 in the formula are preferably hydrogen atoms from the viewpoint of the mechanical strength of the polyarylene sulfide resin (A). In that case, examples include those bonded at the para position represented by the following structural formula (2) and those bonded at the meta position represented by the following structural formula (3).

[0088]

Chemical formula

[0089] Among these, particularly, the bond of the sulfur atom to the aromatic ring in the repeating unit is preferably a structure bonded at the para position represented by the structural formula (2) from the viewpoints of the heat resistance and crystallinity of the polyarylene sulfide resin (A).

[0090] Further, the polyarylene sulfide resin (A) may contain not only the structural site represented by the structural formula (1) but also one or more structural sites represented by the following structural formulas (4) to (7) in an amount of 30 mol% or less based on the total amount with the structural site represented by the structural formula (1) being 100 mol%.

[0091] [ka]

[0092] In particular, it is preferable that the polyarylene sulfide resin (A) contains one or more structural parts represented by the above structural formulas (4) to (7) in an amount of 10 mol% or less, from the viewpoint of heat resistance and mechanical strength. When the polyarylene sulfide resin (A) contains one or more structural parts represented by the above structural formulas (4) to (7), the bonding mode between them may be either a random copolymer or a block copolymer.

[0093] The aforementioned polyarylene sulfide resin (A) includes crosslinked polyarylene sulfide resins and so-called linear polyarylene sulfide resins having a substantially linear structure. Such polyarylene sulfide resins (A) are easy to control and have excellent industrial productivity, and can be produced, for example, by reacting sodium sulfide with p-dichlorobenzene in an amide solvent such as N-methylpyrrolidone or dimethylacetamide, or a sulfone solvent such as sulfolane.

[0094] The raw material polyarylene sulfide resin (A) is preferably polyarylene sulfide resin particles (a) having a volume-average particle diameter in the range of 1.0 mm to 3.0 mm. When the volume-average particle diameter of the polyarylene sulfide resin particles (a) is 1.0 mm or more, the polyarylene sulfide resin particles (a) are less likely to re-aggregate, are easy to handle, and mix uniformly with the thermoplastic elastomer particles (b). Furthermore, when the volume-average particle diameter of the polyarylene sulfide resin particles (a) is 3.0 mm or less, they mix uniformly with the thermoplastic elastomer particles (b), thus improving the strength-improving effect on the polyarylene sulfide resin composition. Among these, polyarylene sulfide resin particles (a) with a volume-average particle diameter in the range of 1.5 mm to 2.5 mm are more preferable.

[0095] The polyarylene sulfide resin particles (a) having a volume-average particle diameter in the range of 1.0 mm to 3.0 mm can be manufactured, for example, by the following method (a) or (b). (i) A method for obtaining a plate-shaped solid by compressing and fixing the polyarylene sulfide resin (A) particles obtained after the polymerization of the polyarylene sulfide resin (A) has been completed, cooling the reaction solution of the polyarylene sulfide resin (A), washing it several times with water or warm water, and then drying it, using a press such as a belt press, and then grinding it to obtain polyarylene sulfide resin particles (a) having a volume average particle diameter in the range of 1.0 mm to 3.0 mm. (b) A method for obtaining polyarylene sulfide resin particles (a) having a volume average particle diameter in the range of 1.0 mm to 3.0 mm, by adding water to the polyarylene sulfide resin (A) while it is dissolved in the reaction solvent, after the polymerization of the polyarylene sulfide resin (A) is completed and before the reaction solution of the polyarylene sulfide resin (A) is cooled.

[0096] It is preferable that the polyarylene sulfide resin (A) has a carboxyl group as a functional group having an active hydrogen atom in its molecular structure, as this can increase the reactivity of the polyarylene sulfide resin (A). Specifically, it is preferable that the content of the carboxyl group in the polyarylene sulfide resin (A), as measured by neutralization titration, is in the range of 10 μmol / g to 200 μmol / g, and more preferably in the range of 10 μmol / g to 100 μmol / g. When the content of the carboxyl group in the polyarylene sulfide resin (A), as measured by the neutralization titration, is 10 μmol / g or more, the reactivity of the polyarylene sulfide resin (A) can be increased, while when it is 200 μmol / g or less, the reactivity of the polyarylene sulfide resin (A) can be easily controlled.

[0097] A method for producing the polyarylene sulfide resin (A) having a carboxyl group as a functional group with an active hydrogen atom in its molecular structure includes, after polymerization of the polyarylene sulfide resin (A), cooling to room temperature and washing with water, filtering the polyarylene sulfide resin (A), treating it with acid, and then washing with water. The acids that can be used in this process are preferably acetic acid, hydrochloric acid, sulfuric acid, phosphoric acid, silicic acid, carbonic acid, oxalic acid, or propionic acid, as these can efficiently reduce the amount of residual metal ions without decomposing the polyarylene sulfide resin (A), and among these, acetic acid or hydrochloric acid is more preferred.

[0098] Furthermore, the polyarylene sulfide resin (A) according to the embodiment is preferably one whose melt viscosity, measured at 300°C, is in the range of 60 Pa·s to 240 Pa·s. When the melt viscosity of the polyarylene sulfide resin (A) is 60 Pa·s or higher, the toughness of the polyarylene sulfide resin composition is improved, while when the melt viscosity of the polyarylene sulfide resin (A) is 240 Pa·s or lower, it becomes easier to suppress the heat generation of the polyarylene sulfide resin composition under high shear. Among these, from the viewpoint of balancing the strength improvement effect due to the addition of the thermoplastic elastomer (B) and the fluidity of the polyarylene sulfide resin composition, the melt viscosity of the polyarylene sulfide resin (A) is particularly preferably in the range of 80 Pa·s to 180 Pa·s.

[0099] Here, the melt viscosity of the polyarylene sulfide resin (A) measured at 300°C refers to the melt viscosity (Pa·s) of the polyarylene sulfide resin (A) measured using a high-flow type flow tester with an orifice of 10 mm in length and 1 mm in diameter, after holding it for 6 minutes under the conditions of 300°C and a test load of 50 kg.

[0100] In the polyarylene sulfide resin composition according to the embodiment, the content of polyarylene sulfide resin (A) is preferably 50 to 95% by mass, more preferably 70 to 90% by mass, and even more preferably 75 to 85% by mass, based on the total mass (100% by mass) of the polyarylene sulfide resin composition. The upper and lower limits of the numerical range for the content of polyarylene sulfide resin (A) exemplified above can be freely combined.

[0101] <Thermoplastic elastomer (B)> The thermoplastic elastomer (B) is preferably, for example, one with a melting point of 300°C or less and rubber elasticity at room temperature. Such a thermoplastic elastomer (B) is more preferable because it provides an excellent effect in improving the impact resistance of the polyarylene sulfide resin composition. Furthermore, from the standpoint of excellent heat resistance, the thermoplastic elastomer (B) is preferably a polyolefin-based thermoplastic elastomer or a nitrile-based thermoplastic elastomer, and more preferably a polyolefin-based thermoplastic elastomer.

[0102] The polyolefin-based thermoplastic elastomer mentioned above can be exemplified by a thermoplastic elastomer containing constituent units derived from an olefin. α-olefin is preferred as the olefin. The polyolefin-based thermoplastic elastomer is preferably one that has one or more functional groups or structures selected from the group consisting of hydroxyl groups, carboxyl groups, amino groups, mercapto groups, epoxy groups, acid anhydride structures, ester structures, and isocyanate groups in its molecular structure, as this provides excellent reactivity and compatibility with the polyarylene sulfide resin (A). Among these, the polyolefin-based thermoplastic elastomer having a carboxyl group, epoxy group, acid anhydride structure, or ester structure in its molecular structure is more preferable because it exhibits superior reactivity with the polyarylene sulfide resin (A), resulting in improved compatibility and a more uniformly mixed polyarylene sulfide resin composition.

[0103] The polyolefin-based thermoplastic elastomers having carboxyl groups, epoxy groups, acid anhydride structures, or ester structures in their molecular structure can be obtained, for example, by copolymerization of an α-olefin with a vinyl polymerizable compound having carboxyl groups, epoxy groups, acid anhydride structures, or ester structures in its molecular structure. Examples of the α-olefins include 2-8 carbon olefins such as ethylene, propylene, and butene.

[0104] The glycidyl-modified polyolefin-based thermoplastic elastomer having the aforementioned carboxyl group in its molecular structure can be obtained, for example, by copolymerization of an α,β-unsaturated carboxylic acid such as acrylic acid or methacrylic acid, or a carbon-4 to carbon-10 unsaturated dicarboxylic acid such as maleic acid, fumaric acid, itaconic acid, or other carbon-4 to carbon-10 unsaturated carboxylic acid with the aforementioned α-olefin.

[0105] The polyolefin-based thermoplastic elastomer having the aforementioned epoxy group in its molecular structure can be obtained by copolymerization of glycidyl acrylate, glycidyl methacrylate, or the like with the α-olefin.

[0106] Polyolefin-based thermoplastic elastomers having the aforementioned acid anhydride structure in their molecular structure can be obtained by copolymerization of an α,β-unsaturated dicarboxylic acid anhydride, such as maleic acid, fumaric acid, itaconic acid, or other unsaturated dicarboxylic acids having 4 to 10 carbon atoms, with the aforementioned α-olefin.

[0107] Polyolefin-based thermoplastic elastomers having the aforementioned ester structure in their molecular structure can be obtained by copolymerization of alkyl esters of α,β-unsaturated carboxylic acids such as acrylic acid esters and methacrylic acid esters, mono and diesters of maleic acid, fumaric acid, itaconic acid, and other unsaturated dicarboxylic acids having 4 to 10 carbon atoms with the aforementioned α-olefin.

[0108] Furthermore, copolymers containing multiple of these two or more functional groups or structures simultaneously can be used. Preferred examples of these include terpolymers of α-olefins, acrylic acid esters, and glycidyl methacrylate.

[0109] The thermoplastic elastomer (B) preferably contains 40 to 95% by mass of α-olefin-derived constituent units, more preferably 50 to 90% by mass, and even more preferably 60 to 80% by mass, based on the total mass (100% by mass) of the constituent units that make up the thermoplastic elastomer (B).

[0110] The thermoplastic elastomer (B) preferably contains 0.1 to 30% by mass of constituent units derived from glycidyl (meth)acrylate, more preferably 0.5 to 15% by mass, and even more preferably 1 to 7% by mass, based on the total mass (100% by mass) of constituent units constituting the thermoplastic elastomer (B).

[0111] The thermoplastic elastomer (B) preferably contains 0.1 to 50% by mass of constituent units derived from methyl acrylate, more preferably 3 to 40% by mass, and even more preferably 10 to 35% by mass, based on the total mass (100% by mass) of constituent units that make up the thermoplastic elastomer (B).

[0112] As an example of the above copolymer, with respect to the total mass (100% by mass) of the constituent units that make up thermoplastic elastomer (B), It contains 0.1 to 30% by mass of constituent units derived from glycidyl (meth)acrylate. Examples include those containing 0.1 to 50% by mass of constituent units derived from methyl acrylate.

[0113] As an example of the above copolymer, with respect to the total mass (100% by mass) of the constituent units that make up thermoplastic elastomer (B), It contains 40-95% by mass of constituent units derived from α-olefins. It contains 0.1 to 30% by mass of constituent units derived from glycidyl (meth)acrylate. Examples include those containing 0.1 to 50% by mass of constituent units derived from methyl acrylate.

[0114] Next, as the nitrile-based thermoplastic elastomer, copolymers of unsaturated nitrile and conjugated diene are mentioned. Examples of the unsaturated nitrile include acrylonitrile or methacrylonitrile, and examples of the conjugated diene include 1,3-butadiene, 2-methyl-1,3-butadiene, 2,3-dimethyl-1,3-butadiene, and 1,3-pentadiene. Among these, acrylonitrile-butadiene copolymers are preferred, and hydrogenated nitrile-based thermoplastic elastomers are even more preferred, in which some or all of the double bonds of the conjugated diene are hydrogenated to improve heat resistance while maintaining the triple bonds of the nitrile group.

[0115] Furthermore, the hydrogenated nitrile thermoplastic elastomer is preferably one or more functional groups selected from the group consisting of vinyl groups, hydroxyl groups, carboxyl groups, acid anhydride structures, glycidyl groups, amino groups, isocyanate groups, mercapto groups, oxazoline groups, isocyanurate groups, and maleimide groups in its molecular structure, as this is preferable in terms of excellent reactivity and compatibility with the polyarylene sulfide resin (A). Among these, the hydrogenated nitrile thermoplastic elastomer having a carboxyl group is particularly preferred in terms of excellent heat resistance and reactivity.

[0116] The raw material thermoplastic elastomer (B) is preferably thermoplastic elastomer particles (b) having a volume-average particle diameter in the range of 0.1 mm to 3.0 mm. When the volume-average particle diameter of the thermoplastic elastomer particles (b) is 0.1 mm or more, the specific surface area of ​​the thermoplastic elastomer particles (b) becomes smaller, making re-aggregation of the thermoplastic elastomer particles (b) less likely to occur, making them easier to handle, and facilitating the blending of a predetermined amount of the thermoplastic elastomer particles (b). On the other hand, when the volume-average particle diameter of the thermoplastic elastomer particles (b) is 3.0 mm or less, it becomes easier to uniformly mix them with the polyarylene sulfide resin (A), and the strength improvement effect on the polyarylene sulfide resin composition is well expressed. Among the volume-average particle diameters mentioned above, it is preferable that the volume-average particle diameter of the thermoplastic elastomer particles (b) be in the range of 0.3 mm to 2.0 mm, considering the balance of the effects of ease of handling during work, ease of uniform mixing, and improvement of impact resistance and flexural strength of the thermoplastic elastomer particles (b).

[0117] Methods for producing the thermoplastic elastomer particles (b) having a volume-average particle diameter in the range of 0.1 mm to 3.0 mm include a method of finely cutting thermoplastic elastomer particles having a volume-average particle diameter greater than 3.0 mm using a cutting machine, or a method of freeze-pulverizing the thermoplastic elastomer particles having a volume-average particle diameter greater than 3.0 mm. Freeze-pulverization methods include freezing with dry ice or liquid nitrogen, and then pulverizing using a conventional hammer-type pulverizer, cutter-type pulverizer, or millstone-type pulverizer. Among the above methods, the method of freeze-pulverizing to produce the thermoplastic elastomer particles (b) is preferred because it allows for easy production of the thermoplastic elastomer particles (b).

[0118] In the resin composition according to the embodiment, the content of thermoplastic elastomer (B) is preferably 5 to 30% by mass, more preferably 10 to 28% by mass, and even more preferably 15 to 25% by mass, based on the total mass (100% by mass) of the polyarylene sulfide resin composition. The upper and lower limits of the numerical range for the blending ratio of thermoplastic elastomer (B) exemplified above can be freely combined.

[0119] From another perspective, in the polyarylene sulfide resin composition, the blending ratio of the thermoplastic elastomer (B) to the total content (100 parts by mass) of the polyarylene sulfide resin (A) and thermoplastic elastomer (B) is preferably 5 to 30 parts by mass, more preferably 10 to 28 parts by mass, and even more preferably 15 to 25 parts by mass.

[0120] When the content of the thermoplastic elastomer (B) is above the lower limit, the effect of improving the impact resistance of the polyarylene sulfide resin composition is well exhibited. When the content of the thermoplastic elastomer (B) is below the upper limit, the amount of gas generated during molding of the polyarylene sulfide resin composition can be effectively reduced.

[0121] The resin composition according to the embodiment may contain, in addition to the polyarylene sulfide resin (A) and the thermoplastic elastomer (B), other optional components such that their total content (mass%) does not exceed 100% by mass. For example, it may further contain other resins that do not fall under the category of polyarylene sulfide resin (A) and the thermoplastic elastomer (B). Other resins include homopolymers or copolymers of monomers such as ethylene, butylene, pentene, butadiene, isoprene, chloroprene, styrene, α-methylstyrene, vinyl acetate, vinyl chloride, acrylic acid esters, methacrylic acid esters, and (meth)acrylonitrile; polyurethane, polyester, polyesters such as polybutylene terephthalate and polyethylene terephthalate; polyacetal, polycarbonate, polysulfone, polyallyl sulfone, polyethersulfone, polyphenylene ether, polyether ketone, polyether ether ketone, polyimide, polyamide imide, polyetherimide; silicone resin, epoxy resin, phenoxy resin, liquid crystal polymer, and polyaryl ether homopolymers, random copolymers or block copolymers, and graft copolymers.

[0122] The resin composition according to the embodiment may further contain an epoxysilane coupling agent (C) in addition to the above components. This is preferable because, due to the excellent reactivity between the polyarylene sulfide resin (A) and the thermoplastic elastomer (B) and the epoxysilane coupling agent, the uniform dispersion of the thermoplastic elastomer (B) is improved, and the adhesion at the interface between the polyarylene sulfide resin (A) and the thermoplastic elastomer (B) is improved, resulting in an even more pronounced strength improvement effect of the polyarylene sulfide resin composition.

[0123] The epoxysilane coupling agent (C) is preferably a silane compound having a structure in which an epoxy structure-containing group, such as a glycidoxyalkyl group or a 3,4-epoxycyclohexylalkyl group, which has a linear alkyl group with 1 to 4 carbon atoms as the alkyl group, and two or more methoxy groups and ethoxy groups are bonded to a silicon atom.

[0124] Examples of such epoxysilane coupling agents (C) include γ-glycidoxypropyltrimethoxysilane, β-(3,4-epoxycyclohexyl)ethyltrimethoxysilane, γ-glycidoxypropyltriethoxysilane, and epoxy-based silicone oils.

[0125] The epoxy-based silicone oils mentioned above include compounds having polyalkylene oxy groups composed of 2 to 6 repeating units of alkoxy groups having 2 to 6 carbon atoms.

[0126] Among the epoxysilane coupling agents (C), glycidoxyalkyltrialkoxysilane compounds, such as γ-glycidoxypropyltrimethoxysilane and γ-glycidoxypropyltriethoxysilane, are particularly preferred due to their excellent reactivity with the polyarylene sulfide resin (A) and the thermoplastic elastomer (B).

[0127] The content of the epoxysilane coupling agent (C) is preferably in the range of 0.1% to 5% by mass relative to the total mass of the polyarylene sulfide resin composition. When the content is 0.1% or more by mass, the compatibility between the polyarylene sulfide resin (A) and the thermoplastic elastomer (B) improves, and when it is 5% or less by mass, the amount of gas generated during melt molding of the polyarylene sulfide resin composition decreases. Among these, the content relative to the total amount of the polyarylene sulfide resin composition is particularly preferably in the range of 0.1% to 2% by mass, from the viewpoint of balancing the compatibility between the polyarylene sulfide resin (A) and the thermoplastic elastomer (B) and the amount of gas generated during melt molding of the polyarylene sulfide resin composition.

[0128] The resin composition according to the embodiment may contain, in addition to the above-mentioned compound, an inorganic filler as appropriate. The inorganic filler may include a fibrous inorganic filler and a non-fibrous inorganic filler.

[0129] Examples of the fibrous inorganic fillers include glass fibers, PAN-based or pitch-based carbon fibers, silica fibers, zirconia fibers, boron nitride fibers, silicon nitride fibers, boron fibers, aluminum borate fibers, potassium titanate fibers, inorganic fibrous materials such as stainless steel, aluminum, titanium, copper, and brass, and organic fibrous materials such as aramid fibers.

[0130] Furthermore, examples of non-fibrous inorganic fillers include silicates such as mica, talc, warlastenite, sericite, kaolin, clay, bentonite, asbestos, alumina silicate, zeolite, and pyrophyllite; carbonates such as calcium carbonate, magnesium carbonate, and dolomite; sulfates such as calcium sulfate and barium sulfate; metal oxides such as alumina, magnesium oxide, silica, zirconia, titania, and iron oxide; glass beads, ceramic beads, boron nitride, silicon carbide, and calcium phosphate. These fibrous inorganic fillers and non-fibrous inorganic fillers may be used individually or in combination of two or more. The timing of blending these non-fibrous inorganic fillers is not particularly limited, but it is preferable that they be blended when the polyarylene sulfide resin (A) and the thermoplastic elastomer (B) are dry-blended by the Nauta mixer.

[0131] The content ratio of the polyarylene sulfide resin (A) to the inorganic filler is preferably in the range of 30 to 100 parts by mass / 70 to 0 parts by mass in terms of the ratio of the former to the latter, from the viewpoint of the melting characteristics of the polyarylene sulfide resin composition and the mechanical properties of the molded product. Furthermore, the mixing ratio of the fibrous inorganic filler to the non-fibrous inorganic filler may be any content from the viewpoint of the mechanical properties required of the molded product, but it is preferably in the range of 20 to 100 parts by mass / 80 to 0 parts by mass in terms of the ratio of the former to the latter.

[0132] In the production of the resin composition according to the embodiment, it is particularly preferable to feed the fibrous filler into the extruder from the side feeder of the twin-screw extruder, as this ensures good dispersibility of the fibrous filler. The position of the side feeder is preferably such that the ratio of the distance from the extruder resin input section to the side feeder to the total length of the twin-screw extruder is in the range of 0.1 to 0.6, and among these, it is particularly preferable that it is in the range of 0.2 to 0.4.

[0133] Furthermore, the resin composition according to the embodiment may contain appropriate amounts of additives such as antioxidants, processing heat stabilizers, plasticizers, mold release agents, colorants, lubricants, weather-resistant stabilizers, foaming agents, rust inhibitors, and waxes.

[0134] The resin composition according to the embodiment can be provided, for example, as pellets. The target molded product can be obtained by melt-molding these pellets of the resin composition in a molding machine. The melt-molding method can be, but is not limited to, injection molding, extrusion molding, compression molding, etc.

[0135] The molded articles of the resin composition produced by the manufacturing conditions obtained by the determination method of the embodiment exhibit excellent dispersibility of the dispersed phase components and suppressed thermal decomposition of the resin, thereby exhibiting exceptionally good impact resistance. As for impact resistance, the Charpy impact value of the molded product of the resin composition according to the embodiment can be used.

[0136] The resin composition produced by the manufacturing conditions obtained by the determination method of the embodiment can be suitably used, for example, in automotive parts, or in vehicle parts or vehicle components of electrical or electronic components used as automotive parts. Examples of vehicle components include drivetrain components in the engine compartment (e.g., transmission gears, drive motor components, etc.), control components (PCU, etc.), cooling components (piping, valves, pump components, etc.), and battery components.

[0137] According to the manufacturing conditions obtained by the determination method of the embodiment, it is possible to provide a resin composition exhibiting excellent impact resistance, particularly a polyarylene sulfide resin composition. [Examples]

[0138] The present invention will now be described in more detail with reference to examples, but the present invention is not limited to the following examples.

[0139] <Raw materials> • Polyphenylene sulfide (manufactured by DIC Corporation, MA-520) • Polyethylene elastomer A (manufactured by Sumitomo Chemical Co., Ltd., Bond First BF-7L) • Polyethylene-based elastomer B (manufactured by Sumitomo Chemical Co., Ltd., BondFirst BF-E)

[0140] The proportions of each monomer used as a raw material for the polyethylene elastomer described above are shown below. The values ​​represent the amount of each monomer (mass %) relative to the total mass of monomers used as a raw material for the polyethylene elastomer.

[0141] [Table 1]

[0142] <Dataset Acquisition> [Melting and kneading using a twin-screw extruder] (Manufacturing 1) Polyphenylene sulfide resin MA-520 was blended with polyethylene elastomer A at concentrations of 10%, 15%, or 20% by mass relative to 100% by mass of the total mass of the resin composition. The mixtures were hand-blended and fed into a twin-screw extruder. The twin-screw extruder used was a co-rotating twin-screw extruder with a screw diameter of 15 mm and an L / D ratio of 90. The temperature of the 3rd to 15th barrels from the upstream side of the 15-part barrel was set to 300°C. Each mixture was melt-kneaded at 150, 300, 600, 1000, 1500, 2000, or 3000 rpm. The strands exiting the die were cooled and cut to obtain pellets of the polyphenylene sulfide resin composition.

[0143] (Manufacturing 2) In the same procedure as in Manufacturing 1 described above, polyethylene elastomer B was added to polyphenylene sulfide resin MA-520 in amounts of 10%, 15%, or 20% by mass relative to 100% by mass of the total mass of the resin composition. The mixture was then hand-blended, and the resulting mixture was melt-kneaded in a twin-screw extruder to obtain pellets of the resin composition.

[0144] [Melting and mixing conditions] During melt-mixing using a twin-screw extruder, infrared thermometers were installed at the 5th, 8th, 11th, and 14th barrels (from the raw material input side) of the 15 divided barrels. The temperature inside the cylinder (°C) as the resin passed through each barrel was measured at a total of four locations from the screw base to the tip (IR1, IR2, IR3, and IR4 temperatures). Resin pressure (MPa) and extruder current (A) were also measured. These experimental data were used as transient variables (transient features).

[0145] [Charpy impact test] The pellets of each polyphenylene sulfide resin composition obtained above were used as molding materials and molded in an injection molding machine under the conditions of a cylinder setting temperature of 300°C and a mold setting temperature of 130°C to obtain test specimens measuring 80 mm in length, 10.0 mm in width, and 4.0 mm in thickness. Next, notches (B) were made according to ISO 2818. N A piece measuring 8mm was cut, and a Charpy impact test was performed according to ISO 179-1, yielding a notched Charpy impact value (kJ / m²) at 23°C. 2 ) was obtained.

[0146] [Elastomer dispersion particle size] Using pellets of each polyphenylene sulfide resin composition obtained above as molding material, multipurpose test specimens conforming to ISO 3617 Type A were molded using an injection molding machine. Next, the cross-section of the center of the multipurpose test specimen was polished, immersed in xylene, and ultrasonically treated at a temperature of 50°C to extract and remove the elastomer dispersion in the cross-section. After drying at 130°C for 2 hours, the cross-section was observed using a scanning electron microscope (SEM) and an image was obtained. The areas where the elastomer was removed became circular void phases and appeared black. The image was binarized using image analysis software, and the diameter of all the circles where the elastomer had been dispersed as particulate matter was measured. The average value of these measurements was taken as the average particle diameter of the elastomer.

[0147] [Melting viscosity] Pellets of each polyphenylene sulfide resin composition obtained above were placed in a flow tester with a cylinder temperature of 300°C, an orifice length of 1 mm, and an orifice diameter of 1 mm. After preheating for 5 minutes, a load of 50 kg was applied, and the melt viscosity was measured.

[0148] [Measurement of Raman scattering intensity] The pellets of each polyphenylene sulfide resin composition obtained above were used as molding materials and molded in an injection molding machine to obtain test specimens measuring 80 mm in length, 10.0 mm in width, and 4.0 mm in thickness. Using a laser Raman spectrophotometer (JASCO, RMP-520), a laser with a wavelength of 785 nm was irradiated onto the test specimens, and the Raman scattering intensity was measured. The baseline was 300 cm². -1The intensity was obtained.

[0149] [Measurement of near-infrared diffuse reflectance] Pellets of each polyphenylene sulfide resin composition obtained above were used as molding materials and molded in an injection molding machine to obtain test specimens measuring 80 mm in length, 10.0 mm in width, and 4.0 mm in thickness. Using an ultraviolet-visible-near-infrared spectrophotometer (JASCO, V-770) and an integrating sphere unit (JASCO, ISN-923), the diffuse reflectance R (%) value at a wavelength of 1290 nm, which corresponds to the trough of the absorption band, was read, and the Kuberkamunck function f was calculated from the formula f = (1 - R / 100)^2 / (2R / 100). It is known that the Kuberkamunck function f is equal to the ratio of the absorption coefficient K to the scattering coefficient S (f = K / S), but at the baseline, which corresponds to the trough of the absorption band, the value of the absorption coefficient K is not considered to change significantly. Therefore, the reciprocal of the Kuberkamunck function, 1 / f, is considered to be proportional to the scattering coefficient S. Furthermore, when the wavelength of near-infrared light (1290 nm) is larger than the dispersion particle size (several hundred nm), light scattering by the particles becomes Rayleigh scattering, and the scattering coefficient S increases with the dispersion particle size. In fact, when we investigated the relationship between the reciprocal value of the Kuberkamunck function (1 / f) and the above-mentioned dispersion particle size, we found a high positive correlation between the two (R² = 0.8318), demonstrating that the dispersion particle size of elastomers can be evaluated nondestructively by using the reciprocal value of the Kuberkamunck function (1 / f).

[0150] <Evaluation of Machine Learning Algorithms> Using the same procedure as in Manufacturing 1 described above, polyethylene elastomer B was blended with polyphenylene sulfide resin MA-520 in an amount of 20% by mass relative to 100% by mass of the total mass of the resin composition. The mixture was then melt-kneaded in a twin-screw extruder at 300 rpm, 600 rpm, 1000 rpm, 1500 rpm, or 2000 rpm to obtain pellets of the polyphenylene sulfide resin composition. The Raman scattering intensity, the reciprocal of the Kuberkamunck function (1 / f), and the Charpy impact value of the resin composition were then measured.

[0151] Furthermore, regression analysis was performed using a support vector machine algorithm with the Charpy impact value as the dependent variable and the reciprocal 1 / f obtained from Raman scattering intensity and near-infrared diffuse reflectance measurements as independent variables. The regression yielded a coefficient of determination of 0.89. From the regression curve, it was found that the conditions under which the Raman scattering intensity was 32000 or less and the reciprocal 1 / f was 40 or less corresponded to a higher range for the dependent variable.

[0152] To achieve the desired result within the specified range, the screw rotation speed was set to 1200 rpm, and the resin composition was manufactured under the same conditions otherwise. The Charpy impact value was then measured and found to be 60.2 kJ / m². 2 That's what happened. The following shows the results of determining the Charpy impact value of the obtained resin composition in a series of manufacturing processes with varying screw rotation speeds. The Charpy impact value was 55 kJ / m². 2 Those exceeding the above are designated as A, 40kJ / m³. 2 More than ~55kJ / m 2 Those less than 25 kJ / m³ are designated as B. 2 More than ~40kJ / m 2 Those less than 25 kJ / m³ are designated as C. 2 Let F be the value less than F.

[0153] [Table 2]

[0154] These results indicate that the resin composition obtained under manufacturing conditions where the Raman scattering intensity was 32,000 or less and the reciprocal 1 / f was 40 or less, which were predicted to result in a high Charpy impact value through regression analysis, actually exhibited a high Charpy impact value.

[0155] Furthermore, it was found that the Raman scattering intensity showed a high correlation with the measured IR1 temperature, and the reciprocal of the above, 1 / f, showed a high correlation with the measured elastomer dispersion diameter. The correlation coefficient was approximately 0.9. From this, it was inferred that it was possible to determine the conditions for reducing the IR1 resin temperature and suppressing thermal degradation, as well as the conditions for making the elastomer dispersion diameter smaller, and as a result, it was possible to predict favorable conditions for improving impact resistance. [Explanation of symbols]

[0156] 10...Twin-screw extruder, 11...Drive unit, 12...Feeder, 13...Cylinder, 14...Screw, 19...Die, 20...Machine learning algorithm, 30...Storage device, DS...Dataset, CD...Manufacturing condition data, MD...Physical property measurement data, HC...High importance item, IR...Infrared temperature sensor, IR1...First infrared temperature sensor, IR2...Second infrared temperature sensor, IR3...Third infrared temperature sensor, IR4...Fourth infrared temperature sensor

Claims

1. By executing a machine learning algorithm using a dataset that includes manufacturing condition data, which at least includes the compounding components of the resin composition, mixing conditions, and the temperature of the kneaded material during melt kneading as manufacturing condition items; first physical property measurement data, which at least includes the impact resistance of the resin composition manufactured under the manufacturing conditions indicated by the manufacturing condition data as a characteristic value item; and second physical property measurement data, which is obtained by measuring the resin composition manufactured under the manufacturing conditions indicated by the manufacturing condition data online and non-destructively, the algorithm determines, when the characteristic value of the item to be improved in the resin composition is the objective variable, which item among a plurality of items included in the manufacturing condition data, the first physical property measurement data and the second physical property measurement data has a high importance in relation to the change in the characteristic value of the item to be improved in the item to be improved, and the second physical property measurement data is within a predetermined range. A method for determining the manufacturing conditions of a resin composition.

2. The resin composition comprises one or more resins selected from the group consisting of polyarylene sulfide resin, polyamide resin, polyester resin, polyphenylene oxide resin, polyetheretherketone resin, polycarbonate resin, and polystyrene resin. A method for determining the manufacturing conditions of the resin composition described in claim 1.

3. The second physical property measurement data includes at least the scattering intensity obtained by Raman scattering measurement. A method for determining the manufacturing conditions of the resin composition described in claim 1.

4. In the aforementioned Raman scattering measurement, the baseline intensity (INTENSITY) value is included as at least one explanatory variable. A method for determining the manufacturing conditions of the resin composition according to claim 3.

5. The second physical property measurement data includes at least the diffuse reflectance or Kuberkamunck function obtained by near-infrared spectral measurement. A method for determining the manufacturing conditions of the resin composition according to claim 1 or claim 3.

6. The second physical property measurement data includes at least the diffuse reflectance or Kuberkamunck function obtained by near-infrared diffuse reflectance measurement. A method for determining the manufacturing conditions of the resin composition described in claim 5.

7. In the aforementioned near-infrared spectral measurement, the measurement method is diffuse reflectance, and the scattering coefficient value based on the baseline position is included as at least one explanatory variable. A method for determining the manufacturing conditions of the resin composition according to claim 6.

8. The manufacturing condition data includes, as manufacturing condition items, first manufacturing condition data which is controlled by the resin composition manufacturing apparatus and second manufacturing condition data which is not controlled by the manufacturing apparatus. A method for determining the manufacturing conditions of the resin composition according to claim 1 or claim 2.

9. The second manufacturing condition data includes the internal temperature of each of the multiple kneading sections in the manufacturing apparatus where the resin is kneaded. A method for determining the manufacturing conditions of the resin composition according to claim 8.

Images