Characteristic evaluation system, characteristic evaluation method, and program

Abstract

According to the present invention, a characteristic related to the condition of a resin, and particularly related to the start of an oxidation reaction of the resin, is easily evaluated.　A characteristic evaluation system: stores a condition ascertainment model that receives a feature quantity regarding the constitutional structure of a resin as input and outputs a characteristic related to the start of an oxidation reaction of the resin; acquires the feature quantity regarding the constitutional structure of a target object including a resin; acquires the characteristic related to the start of an oxidation reaction of the target object by inputting the acquired feature quantity into the condition ascertainment model; and outputs the acquired characteristic.

Description

Characterization system, characterization method, and program

[0001] This disclosure relates to a characterization system, a characterization method, and a program. This invention claims priority to Japanese Patent Application No. 2025-018248, filed on 6 February 2025, and in designated countries where reference by reference is permitted, the contents described in that application are incorporated into this application by reference.

[0002] The recycled materials market is being revitalized by global plastic regulations aimed at carbon neutrality and the rise of ethical consumption. While material recycling is economically advantageous, appropriate utilization according to the condition of the waste materials and recycled materials is desirable. Patent Document 1 states, "The present invention provides a method for producing a resin waste molded article containing a thermoplastic resin and an inorganic substance powder from inorganic substance powder-containing resin waste, characterized by comprising a sorting step of sorting the inorganic substance powder-containing resin waste according to the particle size of the inorganic substance powder, a crushing step of crushing the sorted inorganic substance powder-containing resin waste, and a kneading step of kneading with an extruder. In the sorting step, it is preferable to measure the particle size of the inorganic substance powder by small-angle X-ray scattering." "In the sorting step, the inorganic substance powder-containing resin In addition to sorting the waste material by the particle size of the inorganic substance powder, the process may also include sorting the inorganic substance powder and / or thermoplastic resin that constitute the inorganic substance powder-containing resin waste material. For example, it is preferable to sort the resin and / or inorganic substance powder according to their composition by performing one or more types of analysis, such as mid-infrared spectroscopy, near-infrared spectroscopy, infrared spectroscopy, Raman spectroscopy, X-ray fluorescence analysis, X-ray diffraction analysis, and even pyrolysis GC / MS analysis and TG / DTA analysis, in addition to analysis by small-angle X-ray scattering.

[0003] Japanese Patent Publication No. 2021-102286

[0004] Patent Document 1 discloses several methods for thermal analysis, but these analyses require a tremendous amount of time, making the analysis cumbersome.

[0005] The problem that this disclosure aims to solve is to easily evaluate the state of a resin, particularly its properties related to the initiation of oxidation reactions.

[0006] This application includes several means to solve at least some of the above problems, and some examples are as follows.

[0007] One aspect of the present disclosure is a characterization system comprising a processor and a memory device, wherein the memory device stores a state recognition model that accepts feature quantities relating to the microstructure of a resin as input and outputs characteristics relating to the initiation of an oxidation reaction of the resin, and the processor acquires feature quantities relating to the microstructure of an object containing a resin, inputs the acquired feature quantities into the state recognition model to acquire characteristics relating to the initiation of an oxidation reaction of the object, and outputs the acquired characteristics.

[0008] According to this disclosure, the state of the resin, particularly its properties related to the initiation of the oxidation reaction, can be easily evaluated.

[0009] Other issues, configurations, and effects not mentioned above will be clarified by the following description of the embodiments.

[0010] This is a block diagram of the characterization system of the first embodiment. This diagram illustrates the method for evaluating oxidation induction time (OIT). This diagram illustrates an example of an exothermic curve in the measurement of oxidation induction time. This diagram illustrates the content included in the production results DB of the first embodiment. This flowchart shows the characterization method performed by the characterization system of the first embodiment. This is a block diagram of the characterization apparatus of the first embodiment. This is a block diagram of the characterization system of the second embodiment. This diagram illustrates the content included in the production results DB of the second embodiment. This flowchart shows the characterization method performed by the characterization system of the second embodiment. This diagram shows an example of the output results output by the characterization system of the second embodiment. This diagram shows an example of the output results output by the characterization system of the second embodiment. This diagram shows an example of the output results output by the characterization system of the second embodiment. This is a block diagram of the characterization system of the third embodiment. This diagram shows an example of the distribution of oxidation induction time (OIT) between virgin resin, waste resin, and recycled resin. This diagram illustrates the content included in the production results DB of the third embodiment. This flowchart shows the characterization method performed by the characterization system of the third embodiment. This flowchart shows the characterization method performed by the characterization system of the fourth embodiment. This flowchart shows the characterization method performed by the characterization system of the fifth embodiment. This diagram shows an example of the output results output by the characterization system of the fifth embodiment. This is a block diagram of the characterization system of the sixth embodiment. This is a flowchart showing a characterization method performed by the characterization system of the sixth embodiment. This is a diagram showing an example of the output results output by the characterization system of the sixth embodiment. This is a block diagram showing the characterization system of the seventh embodiment. This is a flowchart showing a characterization method performed by the characterization system of the seventh embodiment. This is a diagram showing an example of the output results output by the characterization system of the seventh embodiment. This is a diagram illustrating an example of measurement data for the fluorescence fingerprint of polypropylene.

[0011] Hereinafter, embodiments (referred to as "models") for implementing this disclosure will be described with reference to the drawings. Within the description of one embodiment below, other embodiments applicable to that embodiment will also be described as appropriate. This disclosure is not limited to the following embodiment, and different embodiments can be combined or modified as appropriate without significantly impairing the effects of this disclosure. In addition, the same reference numerals will be used for the same components, and redundant descriptions will be omitted. Furthermore, components having the same function will be given the same name. The illustrations are for illustrative purposes only, and for illustrative purposes, the actual configuration may be changed or some components may be omitted or modified between drawings without significantly impairing the effects of this disclosure. Also, the same embodiment does not necessarily need to have all the components.

[0012] Figure 1 is a block diagram showing the property evaluation system 10 of the first embodiment. The property evaluation system 10 is a system for evaluating the properties of resin (e.g., waste resin) in objects recovered from, for example, product factories, markets, and other factories, and the properties of recycled resin made from the recovered resin. In particular, resin in objects recovered from markets, factories, etc., originates from various usage environments, usage loads, products, parts, etc., resulting in differences in degradation, such as various degradation states and changes in properties during reuse. It is difficult to properly reuse resins whose degradation properties are not well understood. For this reason, the property evaluation system 10 can evaluate the degradation properties of such resins whose state is not well understood. In this disclosure, evaluating properties may include meanings such as understanding, inferring, predicting, analyzing, determining, and acquiring properties, and can be rephrased as appropriate.

[0013] Traditionally, evaluating the degradation characteristics of resins has involved, for example, processing the resin into a predetermined test specimen shape and then measuring the degradation of the specimen through degradation tests under predetermined environmental conditions. This process is costly, requiring significant expense for specimen processing, degradation testing, and evaluation, making the acquisition of degradation data burdensome. Alternatively, there are methods for evaluating degradation-related characteristics through actual measurements. Oxidation is a major cause of resin degradation. The oxidation reaction is initiated by factors such as the easily oxidizable chemical structure of the resin (electron-donating groups, double bonds, etc.), the ease of changes in crystallinity and conformation, foreign substances like metals and additives that contribute to radical generation, and environmental factors like heat. Once the oxidation reaction begins, the generated radicals and peroxides (such as peroxides) trigger a chain reaction of further resin oxidation. Therefore, the time and temperature required for the initiation of the resin oxidation reaction are evaluated. For example, oxidation induction time (OIT) and oxidation onset temperature (IOT) are evaluated and used to understand degradation-related characteristics. In OIT and IOT measurements, a small amount of resin sample is processed from a molded body such as a test piece and measured primarily using thermal analysis such as differential thermal analysis (DTA) or differential scanning thermal analysis (DSC). OIT and IOT are destructive analyses performed by heating in an oxygen-containing atmosphere. In OIT, the sample is heated to near its melting point in an inert gas atmosphere such as nitrogen, and while maintaining the temperature, the atmosphere is switched to oxygen-containing air. The time until the oxidation reaction begins is then measured. Figure 2 shows an example of OIT evaluation. In IOT, the onset temperature of the oxidation reaction is measured at a constant heating rate in an air atmosphere. Compared to evaluation by degradation testing, this reduces the burden of evaluation by eliminating the need for degradation testing time and environmental testing equipment. However, OIT and IOT are destructive analyses performed by heating during sample processing and measurement, and the entire process from sample preparation to measurement and sample recovery after cooling takes approximately 1-3 hours, depending on the type of resin. Also, because OIT and IOT differ in heating conditions and atmospheres during measurement, they cannot usually be measured simultaneously.

[0014] However, in the characterization system 10, the ultraviolet to visible and infrared spectroscopic measurements, X-ray diffraction (XRD), color measurement, hardness measurement, and specific gravity measurement described later are all measurement methods that are less constrained by the shape of the resin, thus reducing the cost of data acquisition.

[0015] The state of a resin includes, for example, the structural structure of the resin that affects the oxidation reaction of the resin (e.g., the chemical structure, such as the way atoms are bonded together, molecular weight, types of functional groups such as carbonyl groups, carboxyl groups, hydroxyl groups, halogen groups, peroxy groups, ester groups, and double bonds, the bonding position of functional groups, crystallinity and conformation, multidimensional structure such as mixtures of different resins, and the presence or contact of metals, peroxides, antioxidants, and radical scavengers). The measurement methods described above are effective in the simple extraction of features directly or indirectly related to such states, and although details will be described later, the time and temperature related to the initiation of the oxidation reaction of the resin can be evaluated as the state of the resin based on, for example, the peak width at a specific position.

[0016] In particular, recycled resins are often more susceptible to oxidation reactions than virgin materials. Therefore, by understanding the OIT and IOT of waste resins and recycled resins, it becomes possible to implement measures to restore stability, such as shortening the oxidation reaction time or lowering the temperature compared to virgin materials (e.g., adjusting the proportion of virgin materials used, blending additives, etc.), and to determine and classify their appropriate use for various applications (grading).

[0017] As for the resin, a thermoplastic resin is preferred, for example. Among these, the resin is olefin-based, and ethylene (-CH ２ -CH ２ -), propylene (-CH ２ -CH(CH ３Resins containing structural units such as )-) are preferred, and polypropylene is more preferred. Resins containing polypropylene structural units include, for example, copolymers of polypropylene structural units and polyethylene structural units, and furthermore, resins that contain hydroxyl groups, carboxyl groups, carboxylic anhydrides such as maleic acid and itaconic acid, styrene, dienes, esters, etc. in their terminals, side chains, or basic structure, or modified resins by polycondensation. It also includes mixtures of polypropylene and polyethylene that have a sea-island structure in their material structure. Hereinafter, resins containing polypropylene structural units as the main unit (most abundant unit) (excluding polypropylene) and polypropylene are collectively referred to as polypropylene, etc.

[0018] Polypropylene and similar materials are widely traded in the market. They are also used relatively frequently in products such as home appliances and industrial products. Therefore, a large amount of polypropylene and similar materials are recovered from the market. By using the characterization system 10, the condition of large quantities of polypropylene and similar materials can be quickly assessed.

[0019] The characteristic evaluation system 10 comprises an output unit 2, a production record DB (database) 3, an acquisition unit 4, an extraction unit 5, and a grasping unit 6. The production record DB 3 is stored, for example, on a server (not shown) located in a remote location. The output unit 2, the production record DB 3, the acquisition unit 4, the extraction unit 5, and the grasping unit 6 are connected to each other via a network 1 so as to be able to communicate with each other.

[0020] The output unit 2 is an output device that outputs at least one of the following to a user, a resin manufacturing device, etc.: the grasping results from the grasping unit 6, or information obtained using those grasping results. The output unit 2 outputs information obtained in the example shown below (characteristics related to the initiation of oxidation reactions, contribution, degree of degradation, manufacturing conditions, etc.). The output device is an input / output device such as a personal computer, a portable information terminal, or a mobile communication terminal. However, the output unit 2 may also be part of the functional unit that constitutes the characteristic evaluation device 20 (Figure 6).

[0021] The acquisition unit 4 is, for example, an acquisition device that acquires data (data related to feature quantities) about the characteristic state of the resin by analyzing an object containing the resin. However, the acquisition unit 4 may also be part of the functional unit that constitutes the characteristic evaluation device 20.

[0022] From the viewpoint of improving the accuracy of the characteristic evaluation system 10, the object to be evaluated for characterization is preferably made only of resin, or a mixture containing any other material but with a small amount of that material. The content is 25% by mass or less, preferably 10% by mass or less. Any other material is, for example, an additive used to improve the functionality of the coexisting resin (reinforcing materials such as fillers, elastomers, different resins, phase solvents, colorants, flame retardants, stabilizers, nucleating agents, etc.), or a contaminant during the collection, sorting, or manufacture of recycled resin.

[0023] Furthermore, it is preferable that the resin contained in the object is a single type of resin (for example, only polypropylene), or a mixture containing multiple types of resins, but with a small amount of resins other than the resin to be assessed by the evaluation of the oxidation reaction. Examples of such objects include waste resins that have been pre-classified by resin type in markets, factories, etc., and recycled resins made from waste resins. Since such waste resins are classified (separated) as, for example, polypropylene, the proportion of polypropylene, etc., in the total resin contained in the waste resin is, for example, 80% by mass or more, preferably 90% by mass or more. In the example of this disclosure, the time and temperature conditions related to the start of the oxidation reaction of the entire object containing the resin are grasped, but for convenience, expressions such as "grasping the start of the oxidation reaction of the resin" will also be used as appropriate below.

[0024] The acquisition unit 4 is specifically, for example, an analysis unit, an analytical device, etc. More specifically, for example, the acquisition unit 4 is at least one of the following measuring devices: a spectrophotometer, an X-ray diffractometer, a colorimeter, a hardness tester, a hydrometer, etc. The spectrophotometer has at least one measurement function of ultraviolet spectroscopy, visible light spectroscopy, infrared spectroscopy, and fluorescence spectroscopy; the colorimeter has a function to measure color with and without ultraviolet irradiation for at least one of total light reflectance (SCI), diffuse light reflectance (SCE), and specular reflectance (SC); the X-ray diffractometer has a measurement function of wide-angle X-ray diffraction (WAXS); the hardness tester has a function to measure hardness based on the indentation depth of the indenter; and the hydrometer has a function to measure density or specific gravity. The acquisition unit 4 can be determined according to the type of resin to be measured. Using these, for example, acquisition devices, data (measured data) relating to the characteristic quantities described below can be obtained. Furthermore, the data related to features can be the features themselves, or it can be data from which features can be obtained by fitting actual data to an arbitrary model function, for example. In other words, the data related to features can be any data associated with the features.

[0025] For example, a UV-Vis-Near-Infrared Spectrophotometer, such as the V-770DS manufactured by JASCO Corporation, can be used as a spectroscopic measuring device for ultraviolet and visible light. By using an integrating sphere unit, the optical system can be set to total internal reflection (SCI) or diffuse internal reflection (SCE) to acquire the reflection spectrum from wavelengths of 200 to 2700 nm. Furthermore, the color (L*a*b* color system) in total internal reflection (SCI) and diffuse internal reflection (SCE) can be obtained from the acquired reflection spectrum using color matching functions according to JIS Z 8701:1999. Alternatively, a color measuring device (spectrophotometer) or spectrofluorometer described later, which has the function of measuring the reflection spectrum of ultraviolet and visible light, may be used as a spectroscopic measuring device.

[0026] For wide-angle X-ray diffraction measurement, Rigaku Corporation's SmartLab can be used. The measurement method should be the Bragg-Brentano concentrated beam method at room temperature (e.g., 25°C). The applied voltage to the copper rotor should be 45kV, and the current 200mA. βThe filter should be made of Ni, the solar slits should be set to 2.5° on both the incident and receiving sides, the diverging slit to 0.15°, the receiving slit to 0.15 mm, and the long slit to 10 mm, with the scattering slit left unused. The scanning axis should be moved in a 2θ / θ linked motion with an interval of 0.01° and a speed of 5° / min, and continuous scanning should be performed in 1D at room temperature within the range of 5° ≤ 2θ ≤ 100°.

[0027] For color measurement, a spectrophotometer such as the CM-2600d manufactured by Konica Minolta, Inc. can be used. The measurement method involves measuring the lightness L* and chromaticity a*b* in the L*a*b* color space (CIELAB) in the SCE, SCI, or specular reflection (SC) obtained from the reflectance spectra of SCE and SCI, with and without irradiation of ultraviolet light with a wavelength of less than 400 nm (removing the effect on spectral intensity and color such as fluorescence emission above 400 nm due to excitation of ultraviolet light; hereinafter also referred to as UV cut). A white high-shielding, high-reflectance sheet (for example, Lumirror E20 manufactured by Toray Industries, Inc.) can be placed on a horizontal laboratory bench, and the sample can be placed on top of it. In the L*a*b* color system, the L value (L*) represents the brightness (lightness) of the color. The L value ranges from 0 to 100, with 0 being black and 100 being white, and a larger number represents a brighter color. The a-value (a*) and b-value (b*) represent the intensity of the color. A positive (+) a* indicates a reddish tint, while a negative (-) a* indicates a greenish tint. A positive (+) b* indicates a yellowish tint, while a negative (-) b* indicates a bluish-purple tint.

[0028] Furthermore, to eliminate the influence of fluorescence emission on spectral intensity and color, a spectrofluorometer may be used to measure the specular reflection spectrum and L*a*b* in specular reflection. For example, the FP-8350 manufactured by JASCO Corporation can be used as a spectrofluorometer. As a measurement method, the photodetector is positioned in the specular reflection direction relative to the direction of incidence of excitation light to the sample (resin), and the spectral intensity is measured when the excitation light wavelength is 200 to 600 nm (5 nm intervals) and the reflection and fluorescence emission wavelengths are 210 to 800 nm. The measurement results can also be represented as a fluorescence fingerprint (EEM), which is a three-dimensional spectral intensity with the excitation light wavelength on the vertical axis and the fluorescence emission wavelength on the horizontal axis. Figure 26 shows an example of fluorescence fingerprint measurement data. By extracting the spectral intensity of fluorescence wavelengths near the excitation light wavelength from the fluorescence fingerprint, spectral components related to non-emission such as reflection and scattering can be separated from the fluorescence spectrum, and the spectral integrated luminescence of fluorescence emission at each fluorescence wavelength due to the excitation light in the specific excitation light wavelength range described above, as well as characteristic quantities related to the color of the fluorescence emission, can be obtained.

[0029] As an infrared spectrometer, for example, the Spectrum 100 Fourier transform infrared spectrometer manufactured by Perkin Elmer, Inc. can be used. The measurement method is the ATR method, and the measurement range is 4500 cm⁻¹. -１ From 600cm -１ The wavenumber range should be integrated 20 times. Additionally, the sample pressure should be set to between 50 and 100 according to the device's display.

[0030] For measuring hardness, for example, the HH-300 series manufactured by Mitutoyo Corporation can be used. The measurement method involves using a Type D hardness tester (HH-338-01 manufactured by Mitutoyo Corporation) fixed to a measurement / calibration stand (CTS-102 manufactured by Mitutoyo Corporation), and measuring the durometer hardness based on the depth of indentation of the indenter due to spring force with a 4580g weight (JIS K 7215).

[0031] For measuring density or specific gravity, for example, the SD-200L manufactured by Alpha Mirage Co., Ltd. can be used. As a measurement method, using Archimedes' principle, the mass in air and the mass in the immersion liquid can be measured, and the density and specific gravity of the sample (resin) can be calculated from each mass, the density of the immersion liquid at the temperature at which the measurement was taken, and the density of the air (JIS K 7112). To minimize the influence of dissolved gases and impurities, the immersion liquid should be pure water, and for samples that float because their specific gravity is lower than that of water, such as polypropylene, it is preferable to use a sample dish that allows the entire sample to be submerged in the immersion liquid. In addition to measurement by the water displacement method, the dimensions of the sample can be measured using calipers, micrometers, or optical shape measuring devices, and the mass of the sample can be measured using an electronic balance, and the density and specific gravity can be calculated from the volume and mass of the sample.

[0032] The characteristic quantities related to the state of the resin are numerical values ​​that quantitatively represent the characteristics of the data obtained from the analysis of the object (measured data; target data). These characteristic quantities are used as clues for understanding the state in the grasping unit 6 described later.

[0033] In the examples of this disclosure, the resin is an olefin-based polypropylene, and the characteristic quantities are those that correlate with (are linked to) the time and temperature at which the oxidation reaction of polypropylene, etc., begins. This makes it possible to understand the characteristics of waste resins, recycled resins, etc., regarding the time and temperature at which the oxidation reaction of polypropylene, etc., begins. The characteristics of the time and temperature at which the oxidation reaction begins are, as described above, for example, the oxidation induction time (OIT) and the oxidation onset temperature (IOT).

[0034] As the OIT, it is measured by heating to a temperature above the melting point and maintaining the temperature in an inert gas atmosphere, and then switching the atmosphere to air containing oxygen or the like and measuring the time until the oxidation reaction starts. The holding temperature is not less than the melting point and not more than the temperature at which thermal decomposition does not occur, and in the case of polypropylene, 160 to 230 °C is preferable. In the OIT of the present disclosure, nitrogen is used as the inert gas, and the holding temperature is set to 200 °C near the melting temperature of polypropylene and evaluated in an air atmosphere. For example, it corresponds to a temperature effective for grasping the easiness of the oxidation reaction in molding processes such as injection molding, extrusion molding, and kneading by heating or melting of polypropylene products and for making judgments on process management such as the heating time during melting.

[0035] The start of the oxidation reaction is evaluated by the rise of the heat generation peak due to oxidation in FIG. 2, and any one of the intersection time obtained by extrapolating the tangent line of the peak and the baseline, the time when the heat generation reaches the peak (maximum), the time when the heat generation reaches 50% of the heat generation peak, and the time when the time derivative of the heat generation peak curve becomes maximum is extracted. As described above, virgin materials, waste resins, and recycled resins contain various additives and foreign substances, etc. Therefore, a stable baseline cannot be obtained due to these thermal reactions, and there is also a possibility that the heat of reaction other than oxidation is included in the heat generation peak. Such a measurement example is shown in FIG. 3. Therefore, in the present disclosure, the time when the time derivative of the heat generation peak curve becomes maximum is extracted to evaluate the OIT.

[0036] As the IOT, the heat generation peak due to the oxidation reaction under the heating condition of a constant heating rate in an air atmosphere is evaluated, and any one of the intersection temperature obtained by extrapolating the tangent line of the peak and the baseline, the temperature when the heat generation reaches the peak (maximum), and the temperature when the heat generation reaches 50% of the heat generation peak is extracted. Therefore, in the present disclosure, the temperature when the heat generation reaches the peak (maximum) under the heating condition with a heating rate of 10 °C / min is evaluated as the IOT.

[0037] The feature amount acquired by the acquisition unit 4 includes at least one feature amount among, for example, the first feature amount, the second feature amount, the third feature amount, the fourth feature amount, the fifth feature amount, and the sixth feature amount.

[0038] The first feature quantity is a feature quantity related to the spectrophotometry based on the ultraviolet light, visible light reflection, and absorption spectra of the resin obtained by spectroscopic measurement. The second feature quantity is a feature quantity related to the wide-angle X-ray diffraction obtained by wide-angle X-ray diffraction measurement. The third feature quantity is a feature quantity related to the color obtained by color measurement of the resin. The fourth feature quantity is a feature quantity related to the spectrophotometry obtained by Fourier transform infrared spectroscopy (FT-IR) of infrared spectroscopic measurement. The fifth feature quantity is a feature quantity related to the durometer hardness (the standard classification is Type D) obtained by hardness measurement. The sixth feature quantity is a feature quantity related to the density or specific gravity obtained by specific gravity measurement. By acquiring at least one of these feature quantities, the state of the resin can be grasped.

[0039] In one embodiment, the state of the resin is the OIT of the resin, and the feature quantity related to the data acquired by the acquisition unit 4 (the feature quantity extracted by the extraction unit 5) includes a feature quantity having a correlation with the OIT (a feature quantity specific to the OIT). Thereby, based on the feature quantity, the OIT of the resin can be grasped.

[0040] In the case of OIT (when the resin contains polypropylene), the first feature quantity includes at least one of the intensity of reflection and absorption at 360 - 380 nm in SCI and the intensity of reflection and absorption at 360 - 370 nm in SCE.

[0041] Also, in this case, the second feature quantity includes at least one of the peak intensity at diffraction angles of 17.8 - 19.0 degrees, the peak intensity and width at 27.9 - 二十八.8 degrees ((220) of the α-phase of polypropylene), the peak intensity and width at 18.4 - 19.6 degrees ((130) of the α-phase of polypropylene), and the intensity difference when the peak shape at 58.4 - 59.2 degrees is assumed to be a Gaussian function.

[0042] Further, in this case, the third feature quantity is L* of the lightness in SCI in the L*·a*·b* color space.

[0043] Further, in this case, the fourth feature quantity is 2955 cm －1 (CH3 antisymmetric stretching) peak position and intensity, 2920 cm －1 (CH2 antisymmetric stretching) peak position, including at least one of them.

[0044] Furthermore, in this case, the fifth feature includes the durometer hardness (Type D) at 1–300 seconds after loading at a measurement environment temperature of 10–30°C.

[0045] Furthermore, in this case, the sixth feature includes density or specific gravity at a measurement temperature of 15-30°C.

[0046] These features show a strong correlation with the OIT of polypropylene. Therefore, by using these features, the OIT of resins containing polypropylene can be determined.

[0047] In another embodiment, the IoT is of the resin, and the features include features that correlate with the IoT (features specific to the IoT). By using such features, the IoT of the resin can be understood.

[0048] In the case of IoT (including polypropylene resin), the first feature quantity includes at least one of the following: the reflectance and absorption intensity at 360-430 nm in SCI, and the reflectance and absorption intensity at 360 nm in SCE.

[0049] The second feature includes the peak intensity at diffraction angles of 15.0–16.9 degrees ((040) of the α-phase of polypropylene).

[0050] The third feature includes at least one of the following: a* and L* of SCI, and b* of SCI and b* of SCE in the UV-cut state where ultraviolet light below 400 nm is filtered out. In other words, SCI that does not cut out ultraviolet light is a mixture of fluorescent emission colors from resins excited by ultraviolet light, phenolic antioxidants that have conjugation and affect the initiation of oxidation reactions, and different resins mixed in from waste resins of polymer alloys and recycled resins, such as polycarbonate, polystyrene, and polyethylene terephthalate, which have aromatic rings, and object colors from the reflection of visible light.

[0051] The fourth feature is 2920 cm －1 This includes the error in the Gaussian / Lorentz component ratio of the peak (CH2 antisymmetric stretching).

[0052] The fifth feature includes the durometer hardness (Type D) at 1–300 seconds after loading at a measurement temperature of 10–30°C.

[0053] The sixth feature includes density or specific gravity at a measurement temperature of 15–30°C.

[0054] Table 1 below is a table that extracts and summarizes some of the features obtained in each of the above embodiments.

[0055]

[0056] Thus, the resin contains polypropylene, and the feature quantities are those that correlate with the OIT and IOT of polypropylene. This allows the IOT and OIT of objects containing polypropylene to be determined using the characterization system 10.

[0057] Returning to the embodiment shown in Figure 1, the extraction unit 5 is an extraction device that extracts feature quantities from the data acquired by the acquisition unit 4 (measured data and experimental data related to feature quantities). However, the extraction unit 5 may also be part of the functional unit that constitutes the characteristic evaluation device 20.

[0058] The specific method of extraction is not limited. For example, the data obtained by the acquisition unit 4 (e.g., raw data) can be extracted by fitting it to an arbitrary model function.

[0059] To explain each feature in detail, the first feature can be extracted by fitting multiple pseudo-Voigt functions to data obtained from an ultraviolet-visible spectrophotometer, the second feature to data obtained from a wide-angle X-ray diffraction analyzer, and the fourth feature to data obtained from FT-IR measurements of an infrared spectrophotometer. The fifth feature can be extracted by fitting the data obtained from a hardness tester using a power-law or logarithmic approximation of the hardness change with respect to loading time.

[0060] However, the extraction is not limited to fitting to a model function. That is, the data obtained from the spectrophotometer, colorimeter, hardness tester, and hydrometer themselves can be used as the first, third, fifth, and sixth features.

[0061] The grasping unit 6 is a grasping device that grasps the state (characteristics) of the resin from the characteristic quantities extracted by the extraction unit 5 and a state grasping model that grasps the time and temperature at which the oxidation reaction of the resin started. However, the grasping unit 6 may also be part of the functional unit that constitutes the characteristic evaluation device 20.

[0062] The state-understanding model is a pre-generated model that, when given features as input, outputs characteristics related to the initiation of the oxidation reaction of the resin (e.g., OIT). The state-understanding model is a machine learning model that can be generated, for example, through supervised learning. The state-understanding model can be generated by any method. For example, it can be generated by performing machine learning with each of the above features as explanatory variables and characteristics related to the initiation of the oxidation reaction of the resin (e.g., OIT) as the target variable.

[0063] The Manufacturing History DB3 is a database that records whether or not the characteristics related to the initiation of oxidation reactions have been known for each resin in the past. The Manufacturing History DB3 records the previously known state for each resin characteristic. These characteristics (characteristics of the target object) include, for example, the name of the resin, the location where the recovered object was collected, the time of collection, and the type of structure in which the resin contained in the object was used (e.g., home appliances, vehicles, containers and packaging). By having the Manufacturing History DB3, it is possible to check whether or not the resin whose state is to be understood by the characteristic evaluation system 10 of this disclosure has been known in the past. Furthermore, for resins whose state has already been understood, the state can be understood without extracting characteristic quantities. This reduces the time required for characteristic evaluation.

[0064] Figure 4 is a diagram illustrating the content included in the manufacturing results DB3 of the first embodiment. Although only information on polypropylene is shown as an example in Figure 4, as described above, the following information is also recorded for resins other than polypropylene.

[0065] The Manufacturing Record DB3 records catalog information of polypropylene available on the market. Specifically, it records information such as the manufacturer and model number of the polypropylene, information on the properties of the polypropylene, information on additives to the polypropylene, and recommended processing conditions for the polypropylene. Furthermore, it records literature information on polypropylene. Specifically, it records information such as the properties of polypropylene before and after compounding (how properties such as the properties related to the initiation of oxidation reactions change when virgin polypropylene is compounded with waste polypropylene), and information on additives to the polypropylene. Furthermore, it records experimental information on polypropylene. Specifically, it records information such as analytical information on polypropylene, information on the properties of polypropylene before and after compounding, information on additives to the polypropylene, and information on the degradation of polypropylene.

[0066] Figure 5 is a flowchart showing a characteristic evaluation method (which may also be called a state assessment method) performed by the characteristic evaluation system 10 of the first embodiment. The characteristic evaluation method of this disclosure includes steps S11 to S17. In the following, OIT is given as an example of a characteristic related to the initiation of an oxidation reaction (also called a characteristic of the initiation of an oxidation reaction), but the following description can be similarly applied to other characteristics (e.g., IOT).

[0067] First, a user who wants to understand the OIT of an object containing resin obtains the object (step S11). The user inputs the characteristics of the object (for example, the name of the resin) into the output unit 2. The characteristic evaluation system 10 (for example, the acquisition unit 4) then checks whether the same characteristics exist in the manufacturing record DB 3 (step S12). If they do (Yes), the characteristic evaluation system 10 (for example, the acquisition unit 4) acquires the OIT recorded in the manufacturing record DB 3 (step S13). The output unit 2 outputs the acquired OIT (the result of the acquisition by the acquisition unit 6), and for example, the acquired OIT is displayed on the display device of the output unit 2 (for example, a monitor).

[0068] On the other hand, if the material is not listed in the production history DB3 (No.), the OIT of the material is unknown. Therefore, the acquisition unit 4 acquires data on characteristic quantities related to the characteristics of the initiation of the oxidation reaction of the resin by analyzing the material containing the resin (acquisition step, step S14). Specifically, the acquisition unit 4 measures the material using at least one of the following measurement methods: spectrophotometer (ultraviolet, visible light), WAXS (wide-angle X-ray diffraction), colorimeter, FT-IR, hardness tester, and hydrometer (step S14).

[0069] Next, the extraction unit 5 extracts features from the data acquired in step S14 (extraction step, step S15). Specifically, the extraction unit 5 extracts features (for example, the first to sixth features) using an extraction method corresponding to the measurement method (step S15). More specifically, the extraction is performed according to the extraction method by the extraction unit 5 described above with reference to Figure 1.

[0070] Finally, the grasping unit 6 grasps the state of the resin from the feature quantities extracted in step S15 and a state grasping model that grasps the state (grasping step, steps S16, S17). Specifically, the grasping unit 6 inputs the extracted feature quantities into a pre-generated state grasping model (step S16). The state grasping model outputs the characteristics of the oxidation reaction initiation from the input feature quantities, thereby acquiring the characteristics of the oxidation reaction initiation (step S17, output step). Specifically, the output unit 2 outputs the acquired OIT (grasping result from the grasping unit 6), and for example, the acquired OIT is displayed on the display device of the output unit 2 (e.g., a monitor).

[0071] Figure 6 is a block diagram showing the characteristic evaluation device 20 of the first embodiment. The characteristic evaluation device 20 realizes the above-mentioned manufacturing record DB 3, acquisition unit 4, extraction unit 5, and gripping unit 6 in a single device. Therefore, the characteristic evaluation device 20 comprises the manufacturing record DB 3, acquisition unit 4, extraction unit 5, and gripping unit 6.

[0072] The characteristic evaluation device 20 is, for example, an information processing device such as a PC or a server computer, and its specific hardware configuration includes one or more processors 41, one or more memories 42, one or more storage devices 43, and one or more communication devices 44.

[0073] The processor 41 is an arithmetic unit that reads various programs stored in memory resources and executes processing corresponding to each program. The processor is, for example, a microprocessor, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an FPGA (Field Programmable Gate Array), a quantum processor, or other arithmetic semiconductor device.

[0074] The memory 42 and storage device 43 are memory resources, such as non-volatile memory and / or volatile memory. Volatile memory is, for example, RAM (Random Access Memory) or ROM (Read Only Memory). Non-volatile memory may be, for example, a rewritable storage medium such as flash memory, a hard disk, or an SSD (Solid State Drive), or it may be a USB (Universal Serial Bus) memory, memory card, or hard disk. In addition, RAM such as MRAM (Magnetoresistive RAM), PRAM (Phase Change RAM), and ReRAM (Resistive RAM) may be considered as non-volatile memory.

[0075] The communication device 44 is a communication device that communicates information with an external device. The communication device 44 communicates information with an external device via a predetermined communication network, such as the Internet or a LAN.

[0076] Of course, the characterization device 20 may be implemented using one physical or logical computer, or it may be implemented using two or more physical or logical computers. The two or more physical or logical computers may each be distributed and located on a network.

[0077] The hardware configuration described above can also be applied to other devices, such as the extraction device as the extraction unit 5 and the grasping device as the grasping unit 6.

[0078] Figure 7 is a block diagram showing the characterization system 10 of the second embodiment. The second embodiment is suitable, for example, for the development of new materials. Specifically, for a material synthesized by the user, the user is presented with the identified OIT and the features that effectively contributed to its identification (effective features). The user can then use the presented information to infer the physicochemical phenomena that cause the OIT, which can be used to make decisions regarding the formulation and utilization of antioxidants related to quality control and stability improvement.

[0079] The characteristic evaluation system 10 of the second embodiment further includes an effective feature determination unit 7. The effective feature determination unit 7 is an effective feature determination device that determines which of a plurality of features related to the state of the resin are effective, that is, which features affect the state of the resin. However, the effective feature determination unit 7 may also be part of the functional unit that constitutes the characteristic evaluation device 20. That is, the characteristic evaluation device 20 may further include an effective feature determination unit 7.

[0080] There are no particular restrictions on the method for calculating effective features (how to determine effective features). For example, a decision tree-based algorithm that can be attached to the state recognition model described above can be used. In such a state recognition model, the input is the features, and the output is the state of the resin and the effective features. Effective features are indicators that quantify the degree of effectiveness, and are hereafter referred to as, for example, the contribution rate. The higher the contribution rate, the higher the degree of effectiveness and the greater the correlation (relevance) to the state. Therefore, for example, if you want to control the characteristics related to the initiation of an oxidation reaction, it is advantageous to control features with a relatively high contribution rate.

[0081] Furthermore, decision tree-based algorithms such as gradient boosting decision trees, random forests, and decision trees can be used. Variants of these may also be used. Since the features are very spars, decision tree-based algorithms that can classify based on the presence or absence of numerical values ​​are preferable.

[0082] In the examples of this disclosure, for example, Gini impurity G(k) (indicating poor classification at node k), represented by the following formulas (1) and (2), is used.

[0083]

[0084] In equation (1), G(k) is the impurity at a node k, n is the number of target labels, and p(i) is the frequency of target label i at a node k.

[0085]

[0086] In equation (2), I(j) is the importance of a feature j. F(j) is the set of nodes to be partitioned for a feature j. Nparent(i) is the number of samples in a node i. Nleft_child(i) is the number of samples in the leftmost child node of a node i. Nright_child(i) is the number of samples in the rightmost child node of a node i. Gparent(i) is the Gini impurity in a node i. Gleft_child(i) is the Gini impurity in the leftmost child node of a node i. Gright_child(i) is the Gini impurity in the rightmost child node of a node i.

[0087] Additionally, SHAP values ​​can be used to determine effective features. SHAP values ​​are approximate values ​​calculated using a simplified method, and they indicate how much each feature influences the predicted value.

[0088] In the example of this disclosure, the effective feature determination unit 7 calculates the degree to which multiple features influence the state of the resin, i.e., their contribution. This makes it easy to determine which features are effective. A higher contribution indicates a greater influence on the state of the resin, while a lower contribution indicates a smaller influence. Therefore, by understanding the contribution, it becomes easier to control the state of the resin.

[0089] Figure 8 is a diagram illustrating the production record DB3 of the second embodiment. In the second embodiment, in addition to the contents of the production record DB3 of the first embodiment (Figure 4), features that effectively contribute to the state understanding model (effective features) are further recorded.

[0090] Figure 9 is a flowchart showing the characterization method performed by the characterization system 10 of the second embodiment. The user is, for example, a person who wants to develop a new material using characteristics related to the initiation of an oxidation reaction as an indicator. The user synthesizes a new sample (resin) (step S21). It is checked whether the newly synthesized sample exists in the production record DB3 (step S22). The confirmation method is the same as in the first embodiment (and will be checked similarly in the third embodiment and subsequent embodiments). If it exists (Yes), the user formulates other target values ​​or specifications (step S23). In step S23, if another user other than the user who synthesized the new sample had synthesized the same type of resin (similar resin) and it was recorded in the production record DB3, the user formulates other characteristics related to the initiation of an oxidation reaction or other specifications.

[0091] On the other hand, if the result is not present in step S22 (No), steps S24 to S27 are performed in the same manner as steps S14 to S16 above (step S24: acquisition step, step S25: extraction step, steps S26, S27: understanding step, output step). However, in the second embodiment, the state understanding model takes multiple feature quantities as input and outputs the state of the resin and its contribution. Therefore, the effective feature quantity determination unit 7 uses the state understanding model to determine the contribution, which is an effective feature quantity. As a result, in step S28 (output step), the feature quantities that contributed to the understanding (effective feature quantities, contribution) are also acquired. Specifically, the output unit 2 outputs the acquired characteristics and feature quantities related to the initiation of the oxidation reaction (both of which are the results of the understanding unit 6), and for example, the acquired characteristics and feature quantities related to the initiation of the oxidation reaction are displayed on the display device (e.g., monitor) of the output unit 2.

[0092] Figure 10 shows an example of the output results produced in the second embodiment. Figure 10 is a graph showing a y-y plot with the measured value of OIT, a characteristic related to the initiation of the oxidation reaction, on the horizontal axis and the predicted value of OIT on the vertical axis. Figure 10 is one result extracted from 100 random cross-validation runs performed to examine the accuracy of the state recognition model. For example, the graph shown in Figure 10 is output to an output unit 2 provided in, for example, the characteristic evaluation system 10. The output unit 2 is equipped with, for example, a display device, and for example, the graph is displayed on the display device.

[0093] The graph shown in Figure 10 plots the training data used to generate the state comprehension model (R ２ A white circle (with a value) is shown, along with an ideal straight line on which the training data is plotted. Additionally, a plot of test data (R) is shown, which captures the characteristics related to the initiation of the oxidation reaction using a state-understanding model. ２ A black circle (shown with a value) is illustrated. Users referring to Figure 10 can recognize the high accuracy of understanding the characteristics related to the initiation of oxidation reactions using the state understanding model.

[0094] Figure 11 shows an example of the output results produced in the second embodiment. The graph (box plot) shown in Figure 11 was obtained by performing random cross-validation of the y-y plot 50 times, similar to Figure 10, and the prediction accuracy R during the execution was measured. ２ This figure statistically illustrates the relationship. The horizontal axis represents OIT and IOT as characteristics related to the initiation of oxidation reactions. The vertical axis represents R ２ This is the value. Users who refer to Figure 11 can recognize the high accuracy of the state awareness model used for both OIT and IOT.

[0095] Figure 12 shows an example of the output results produced in the second embodiment. Figure 12 is a graph (box plot) showing the contribution (contribution rate, effectiveness rate, effectiveness) obtained using the state understanding model. Figure 12 is a graph with the horizontal axis representing the explanatory variable (type of feature) and the vertical axis representing the contribution. The contribution is shown for each explanatory variable (type of feature). The graph shown in Figure 12 is a graph that statistically shows the contribution during execution by performing random cross-validation of the y-y plot 100 times, similar to Figure 10 above.

[0096] The longer the width of the contribution (length along the vertical axis), the greater the contribution. Therefore, users referring to Figure 12 can intuitively grasp the level of contribution for each feature shown on the horizontal axis. In the example shown, it can be recognized that the feature corresponding to the leftmost explanatory variable among the multiple explanatory variables contributes the most (has the highest contribution).

[0097] In the second embodiment described above, the user can learn about characteristic quantities that are likely to affect the characteristics related to the initiation of oxidation reactions based on their contribution. This allows the user to infer physicochemical phenomena from the identified characteristics related to the initiation of oxidation reactions and the contributions of the characteristic quantities, and to devise measures such as experiments, prototyping, and manufacturing to suppress oxidation reactions, as well as selecting and exploring appropriate applications and utilization methods.

[0098] Figure 13 is a block diagram showing the property evaluation system 10 of the third embodiment. In the third embodiment, for waste resin and recycled resin with unknown properties, the properties related to the initiation of the oxidation reaction and the degree of resin degradation are determined using a state understanding model. Based on the determined properties related to the initiation of the oxidation reaction and the degree of resin degradation, manufacturing conditions such as the amount of virgin material to be newly added, the amount of waste resin used, and the amount of additives to be blended into the recycled resin are proposed according to the time and temperature at which the oxidation reaction of the target recycled resin is initiated. Note that the recycled resin is a resin that includes waste resin and, for example, virgin material, and is newly used by the user.

[0099] The third embodiment of the property evaluation system 10 further comprises a degradation degree detection unit 8 and a proposal unit 9. The degradation degree detection unit 8 is a device for determining the degree of degradation of waste resin whose properties are unknown. However, the degradation degree detection unit 8 may also be part of the functional unit that constitutes the property evaluation device 20. That is, the property evaluation device 20 may further include the degradation degree detection unit 8.

[0100] The degree of degradation of waste resin and recycled resin refers to the degree of degradation relative to virgin resin. Normally, resins are mixed with additives such as antioxidants, radical scavengers, light stabilizers, and light absorbers to counteract oxidation reactions. However, due to the aging of the resin, for example, the properties of the additives change or are consumed through decomposition, and as mentioned above, the chemical structure of the resin itself is also modified or decomposed by oxidation reactions, resulting in lower molecular weight. Figure 14 shows the distribution of OIT at a holding temperature of 200°C for randomly collected virgin resin, waste resin, and recycled resin for polypropylene. The central and average values ​​of OIT for waste resin and recycled resin are shorter than those for virgin resin, indicating that waste resin and recycled resin generally undergo oxidation reactions more quickly than virgin resin. Therefore, in the examples of this disclosure, by understanding the changes and decomposition of additives, the degree of degradation of waste resin and recycled resin can be determined by comparing and classifying them with the functions of virgin resin. The degree of degradation may be classified or graded based on representative values ​​such as the median, mean, interquartile range, maximum, and minimum values ​​of the oxidation induction time distribution for virgin material, waste resin, and recycled resin, as shown in Figure 14. Alternatively, the representative value of virgin resin may be stored in a memory device beforehand, and the difference between the representative value of virgin resin and the representative value of the waste resin and recycled resin being evaluated may be calculated to determine the degree of degradation of the waste resin and recycled resin.

[0101] The degree of degradation can be determined, for example, using a degradation degree determination model stored in the degradation degree determination unit 8. The degradation degree determination model is included, for example, in the state determination model described above. The degradation degree determination model can be generated, for example, by machine learning, with the physical properties obtained from degradation tests (e.g., accelerated tests) using virgin material as explanatory variables and the physical properties of the resin as the objective variable. As for specific machine learning methods, for example, it can be generated by methods similar to those used for the state determination model described above.

[0102] The proposal unit 9 is a device that proposes conditions (manufacturing conditions, recipe) that enable the production of recycled resin having the functions desired by the user. However, the proposal unit 9 may also be part of the functional unit that constitutes the character evaluation device 20. That is, the character evaluation device 20 may further include the proposal unit 9.

[0103] For example, if the degree of degradation of the waste resin is relatively high, the functions of the resin (e.g., OIT, IOT) may be lower than those of the virgin material (e.g., OIT, IOT). In this case, by using a relatively large amount of virgin material in combination with the waste resin, the function of the recycled resin composed of the waste resin and virgin material can be restored, or the function of the recycled resin compound (composition) can be restored by adding an additive that suppresses oxidation reactions to the recycled resin.

[0104] The proposed manufacturing and compounding conditions can be determined, for example, using a proposed model stored in the proposal unit 9. The proposed model can be generated, for example, by machine learning, with the performance (e.g., OIT, IOT) and degree of degradation of the waste resin as explanatory variables, and the performance (e.g., OIT, IOT) of the recycled resin made from the waste resin and virgin material as the objective variable. As for specific machine learning methods, for example, it can be generated by methods similar to those used for the state assessment model described above. For generating the proposed model, data described in catalogs, literature, etc., can be used, for example.

[0105] Figure 15 is a diagram illustrating the production results DB3 of the third embodiment. In the third embodiment, in addition to the contents of the production results DB3 of the second embodiment (Figure 8), descriptors that further contribute effectively to the proposed model (recipe design model) are recorded. Descriptors include, for example, pre-mixing characteristics, information on the mixture, and processing conditions.

[0106] Figure 16 is a flowchart showing the characteristic evaluation method performed by the characteristic evaluation system 10 of the third embodiment. The user is, for example, a person who wants to improve the OIT and IOT of recycled resin using waste resin, or recycled resin compound with appropriate additives added to recycled resin (to achieve the desired OIT and IOT). The user obtains waste resin or recycled resin (step S31). It is checked whether the obtained waste resin or recycled resin exists in the production history DB3 (step S32). If it exists (Yes), recycled resin and recycled resin compound are synthesized according to the production conditions in the production history DB3 (step S33).

[0107] On the other hand, if the data does not exist in the production record DB3 (No), steps S34 to S37 are executed in the same manner as steps S14 to S16 (Figure 5) above (step S34: acquisition step, step S35: extraction step, steps S36, S37: understanding step, output step). The degradation degree understanding unit 8 also inputs the feature quantities extracted in step S35 (same as step S15 above) to the degradation degree understanding model (step S39, understanding step). The degradation degree understanding model then outputs the degradation degree, thereby acquiring (understanding) the degradation degree (step S40, output step). Specifically, the output unit 2 outputs the acquired OIT, IOT, and degradation degree (all of which are the understanding results from the understanding unit 6), and for example, the acquired OIT, IOT, and degradation degree are displayed on the display device of the output unit 2 (e.g., a monitor).

[0108] The proposal unit 9 inputs the OIT, IOT, and degradation level obtained in steps S37 and S40 into the proposed model (step S38). This obtains the proposed manufacturing conditions (step S41, output step). Specifically, the output unit 2 outputs the manufacturing conditions, which are information obtained using the acquired OIT, IOT, and degradation level, and the obtained manufacturing conditions are displayed on the display device (e.g., monitor) of the output unit 2.

[0109] Figure 17 is a flowchart showing the characteristic evaluation method performed by the characteristic evaluation system 10 of the fourth embodiment. For example, the characteristic evaluation system 10 shown in Figure 7 can be used as the characteristic evaluation system 10 of the fourth embodiment. Furthermore, the fourth embodiment can be performed in combination with, for example, at least one embodiment of the first to third embodiments. In the fourth embodiment, an existing state understanding model is refined. This improves the accuracy of state understanding.

[0110] First, a user who wants to improve their state-understanding model obtains an object containing the resin whose state they want to understand (Step S51). Furthermore, the user designs the compounding conditions and / or process conditions (Step S51). Then, Step S52 is performed in the same manner as Step S12 (Figure 5), and if one exists (Yes), the user considers using another object or examining the compounding conditions and / or process conditions recorded in the manufacturing record DB3 (Step S53).

[0111] On the other hand, if the item does not exist (No), data is acquired and feature quantities are extracted in the same manner as in steps S14 and S15 above (steps S54 and S55). Data acquisition is performed by the acquisition unit 4, and feature quantity extraction is performed by the extraction unit 5. In addition, along with steps S54 and S55, the object (e.g., pellet) is processed into the shape of a test piece for evaluating OIT and IOT (step S57). By evaluating OIT and IOT using the test piece, OIT and IOT are obtained (i.e., measured) (step S58). The feature quantities and OIT and IOT obtained in steps S55 and S58 are linked to the OIT and IOT and stored in the production record DB3 (step S56).

[0112] In the same manner as in step S28 (Figure 9) above, features that contributed to the understanding (effective features) are acquired (step S59). The acquisition of effective features is performed by the effective feature determination unit 7. In addition, in the fourth embodiment, the effective feature determination unit 7 further limits the features (explanatory variables) that constitute the state understanding model to effective features (step S59). By doing so, it is possible to relatively increase the number of features that are likely to influence OIT and IOT (feedback). As a result, the state understanding model is composed of many features with relatively high contributions, so the understanding accuracy can be improved and the understanding calculation can be sped up. For example, features with a contribution of a predetermined value or more, or the top five or a predetermined number of features when the features are arranged in descending order of contribution, can be treated as effective features.

[0113] Then, the data stored in step S56 (including data with features limited to effective features) is added to the production record DB3, and the state understanding model is generated again (step S60). The generation of the state understanding model is performed by the understanding unit 6.

[0114] Figure 18 is a flowchart showing the characteristic evaluation method performed by the characteristic evaluation system 10 of the fifth embodiment. For example, the characteristic evaluation system 10 shown in Figure 13 can be used as the characteristic evaluation system 10 of the fifth embodiment. Furthermore, the fifth embodiment can be performed in combination with at least one of the first to third embodiments. Moreover, the fifth embodiment can also be performed in combination with the fourth embodiment. In the fifth embodiment, the existing degradation degree assessment model is refined. This improves the accuracy of assessing the degree of degradation.

[0115] A user who wants to improve the degradation assessment model obtains an object containing degraded resin (waste resin, products / parts made from recycled resin, etc.) (Step S71). The user considers other raw materials and extracts characteristic quantities in the same manner as in Steps S52 to S55 (all in Figure 17) (Steps S72 to S75). In Steps S74 and S75, the user measures the object by measuring degradation indicators (for example, the concentration of additives such as antioxidants, radical scavengers, light stabilizers, and light absorbers in the object) (Step S76). In Step S76, measurements are taken on the object from a different perspective than in Step S74.

[0116] As will be explained in detail in step S78, the features and the degradation index are linked. This creates a correlation between the features and the degradation index. The degradation degree assessment model uses the degradation index as the dependent variable and other analysis results (e.g., features based on a spectrophotometer, etc.) as the independent variables. Therefore, the degradation degree assessment model can be generated in step S80, which will be explained later.

[0117] The measurement of the object in step S76 may be performed using the same measurement method as used in step S74, for example. Also, step S74 may also serve as step S76. In step S76, a degradation index is obtained (step S77).

[0118] Features obtained from analyses different from the degradation index are linked to the degradation index and stored in the production results DB3 (step S78). Then, the explanatory variables are limited in the same way as in step S59 (Figure 17) (step S79). Finally, the data newly stored in the production results DB3 (including data with features limited to effective features) is added, and the degradation degree assessment model is regenerated (step S80).

[0119] Figure 19 shows an example of a screen displaying information regarding the degree of deterioration (grade) and the stability of the oxidation reaction (OIT, IOT) of the object. The object is assigned an ID code, and the screen may also display supplementary information (product name, product management ID, images related to appearance and evaluation area, manufacturer, manufacturing date, manufacturing lot, resin type, color tone, dimensions and shape, use of recycled materials and biomass-derived materials, contents, compliance with legal regulations such as environmental and chemical substance management, history of use, manager and usage record, etc.), information on recommended recycling after use (reuse, repair, regeneration, etc.) based on the degree of deterioration and the stability of the oxidation reaction, and a two-dimensional or three-dimensional code that encodes the address of the database where this information is stored, allowing it to be read by the camera of a portable device such as a smartphone or PDA to access and display the database. The ID code assigned to the object can be converted into a character code such as a product management ID, or a 2D code or 3D code including the address of the above-mentioned database, and then directly assigned by marking using an inkjet printer, thermal printer, laser marking machine, or engraving machine, or by attaching a label with the ID code printed on it. Furthermore, if the object has a unique appearance such as color tone, shape, or pattern (dots, marble, stripes, fibers, stone pattern, matte finish, etc.), instead of assigning an ID code to the object by marking or labeling, an ID code can be generated from information obtainable from the appearance (e.g., chromaticity, pattern coordinates and arrangement, dimensions, etc.), and this ID code can be assigned to the object as a product management ID or used for information management in the database. As a result, the degree of deterioration of the object, the stability of the oxidation reaction, and related information can be displayed on a screen, or the ID code of the object can be read with a mobile terminal, and the object can be confirmed on-site by comparing it with the product management ID or by querying the database to check detailed information.

[0120] Figure 20 is a block diagram showing the characterization system 10 of the sixth embodiment. The characterization system 10 of the sixth embodiment can be used in combination with, for example, the characterization systems 10 shown in Figures 7 and 13. Furthermore, the sixth embodiment can be implemented in combination with, for example, at least one of the first to third embodiments. Moreover, the sixth embodiment can be implemented in combination with at least one of the fourth and fifth embodiments.

[0121] In the sixth embodiment, the production record DB3 is connected to the network 11, and the production record DB3 is updated via the network 11 using publicly available information. The proposed model recorded in the proposal unit 9 is then regenerated and refined using the updated production record DB3.

[0122] Figure 21 is a flowchart showing the characteristic evaluation method performed by the characteristic evaluation system 10 of the sixth embodiment. In the sixth embodiment, a recipe design model is generated that can design (determine, propose) a manufacturing method (recipe) such as the blending and addition method of additives when manufacturing a target recycled resin from waste resin, and further a target recycled resin compound from the recycled resin.

[0123] First, resin compounding information is obtained (step S91). This is done via the network 11, for example by the acquisition unit 4. Resin compounding information includes, for example, the properties of the resin before and after compounding, the type, quantity, and size of the compound, OIT, and processing conditions for obtaining IOT. Resin compounding information is publicly available information, such as information based on experiments, information described in literature such as papers and catalogs. Next, the obtained compounding information is organized (step S92). This organization is done, for example, by the acquisition unit 4. The information unit, descriptor format, etc. differ depending on the information source. Therefore, for example, unit standardization and name matching are performed. The organized information is stored in the production results DB 3 (step S93).

[0124] From the information stored in the production record DB3, it is determined whether the explanatory variables (features) that are effective in predicting the target variables (OIT, IOT) are effective and publicly known (e.g., obvious) features (step S94). This determination is performed, for example, by the effective feature determination unit 7. For example, it is known that if an object containing resin also contains an antioxidant, the OIT and IOT of the object will improve. Therefore, if it is publicly known and effective (Yes), the effective feature determination unit 7 limits the explanatory variables to publicly known and effective features (step S95). Then, the flow from step S96 onwards is carried out.

[0125] On the other hand, if the information is not publicly known (No), the proposal unit 9 generates a recipe design model in which the OIT and IOT after blending are the target variables and the OIT, IOT and blending information before blending are the explanatory variables (step S96). The generation is performed by arbitrary machine learning. The proposal unit 9 derives the values ​​of the explanatory variables by performing Bayesian optimization on the generated recipe design model with the user's target specifications (target OIT and IOT) (step S97). In step S96 above, a recipe design model is generated in the "forward direction" in which the value of the target variable is derived from the explanatory variables, but in step S97, the values ​​of the explanatory variables are derived from the target variable using a method such as Bayesian optimization.

[0126] The derived explanatory variable values ​​are the compound information described in step S91 above. Therefore, by using the compound information derived here as recipe information (design conditions, manufacturing conditions), it is possible to design recycled resins and recycled resin compounds having the desired OIT and IOT. Figure 22 shows an example of the screen displayed when the proposed information is output. Figure 22 also shows an order button for the user to order each compound in the recipe information.

[0127] Figure 23 is a block diagram showing the characterization system 10 of the seventh embodiment. In the characterization system 10 of the seventh embodiment, for example, a product DB 30 containing information related to the manufacturing and sales of resins is connected to the characterization system 10 shown in Figure 20 via networks 1 and 11, and the gripping unit 6 has the function of evaluating mechanical properties and properties related to foreign matter. Furthermore, the seventh embodiment can be implemented in combination with, for example, at least one of the first to third embodiments. Moreover, the seventh embodiment can be implemented in combination with at least one of the fourth to sixth embodiments.

[0128] Figure 24 is a flowchart showing the characteristic evaluation method performed by the characteristic evaluation system 10 of the seventh embodiment. In the seventh embodiment, in addition to the existing degradation degree recognition model, mechanical properties and foreign matter index (foreign matter information) are obtained using a mechanical property recognition model and a foreign matter index recognition model, and information regarding grades that support quality classification and appropriate use of the resin of the target object can be provided.

[0129] First, the supplier provides an object containing degraded resin (waste resin, recycled resin, etc.) (step S98). Then, following the supplier's instructions, at least one measurement of ultraviolet-visible-infrared spectroscopy, WaxS, colorimetric measurement, hardness, and specific gravity is performed on the object, similar to steps S34 to S40 (Figure 16) above (step S99), characteristic quantities are extracted (step S100), and OIT, IOT, and degradation indices are obtained (steps S101 to S104).

[0130] Furthermore, following step S100, the characteristic evaluation system 10 acquires the mechanical properties of the object in accordance with the supplier's instructions (steps S105 to S107). Specifically, the characteristic evaluation system 10 (for example, the acquisition unit 4) checks whether the mechanical properties of the object have already been evaluated (step S105). If they have been evaluated (Yes), the grasping unit 6 acquires the mechanical properties (step S107). If they have not been evaluated (No), the grasping unit 6 inputs the extracted feature quantities into a pre-generated mechanical property acquisition model (step S106), thereby acquiring the mechanical properties (step S107).

[0131] Examples of mechanical properties include tensile strength, tensile modulus, tensile fracture strain, flexural strength, flexural modulus, Charpy impact, and Izod impact. In a mechanical property model that links features to mechanical properties, the dependent variable is the mechanical property, and the independent variables are generated as other analysis results (e.g., features based on a spectrophotometer).

[0132] Furthermore, following step S98, the characteristic evaluation system 10 acquires foreign matter indexes for the object in accordance with the supplier's instructions (steps S108 to S111). Specifically, the characteristic evaluation system 10 (e.g., acquisition unit 4) checks whether the foreign matter information of the object has already been analyzed (step S108). If it has been analyzed (Yes), the grasping unit 6 acquires the foreign matter information (step S111). If it has not been analyzed (No), the characteristic evaluation system 10 (e.g., acquisition unit 4) measures the object using a method for measuring foreign matter indexes (step S109). The grasping unit 6 then inputs the feature quantities extracted from the measurement data by the extraction unit 5 into a pre-generated foreign matter index grasping model (step S110), thereby acquiring the foreign matter index (step S111).

[0133] Foreign matter indices include copper, iron, aluminum, chromium, cadmium, mercury, lead, zinc, antimony, barium, titanium, calcium, magnesium, silicon, germanium, phosphorus, sulfur and its compounds, halides such as chlorine, bromine, and fluorine, hydroxides, glass fibers, carbon fibers, plant fibers, different resins, flame retardants, antioxidants, plasticizers such as phthalates, dyes and paints, and the presence and concentration of inclusions such as sand, glass powder, wood (wood chips, sawdust), rubber, paper, and moisture. Furthermore, in a foreign matter index recognition model that links feature quantities with foreign matter indices, the dependent variable is the foreign matter index, and the explanatory variables are generated as other analysis results (e.g., feature quantities based on spectrophotometers, etc.). Furthermore, foreign matter indicators may be obtained by actual measurements through analysis or observation using methods such as X-ray fluorescence, inductively coupled plasma emission spectrometry (ICP-AES), inductively coupled plasma mass spectrometry (ICP-MS), direct ionization mass spectrometry (DART-MS), gas chromatography-mass spectrometry (GC-MS), liquid chromatography-mass spectrometry (LC-MS), moisture meters (dry type, electric type, infrared type, high-frequency type, Karl Fischer type, etc.), microscopes, optical microscopes, fluorescence microscopes, and scanning electron microscopes (SEM).

[0134] Next, in accordance with the supplier's instructions, the characteristic evaluation system 10 links at least one of the OIT, IOT, degree of deterioration, mechanical properties, and foreign matter index obtained for the object with supplementary information regarding the object's specifications, history, performance, characteristics, etc., which can be obtained from the manufacturing record DB 3 or other sources, and inputs and stores it in the product DB 30 (step S112). Steps S99 to S112 may be performed by the user who uses or handles the object as a commercial product, or by a contractor commissioned by the supplier or user.

[0135] Next, in accordance with the user's instructions, the characteristic evaluation system 10 obtains the grade of the object determined by the user by referring to the information stored in the product DB 30 (step S113), stores the grade in the product DB 30, and updates the information of the object as necessary (step S114). Also, in accordance with the instructions of the user or customer, the characteristic evaluation system 10 outputs and displays the grade of the object and related information stored in the product DB to the customer (step S115). The user may also be a supplier.

[0136] Figure 25 shows an example of a screen displaying the object's grade and associated information as suggested information. Similar to Figure 22, an order button for the user to order the object may also be included. The output, displayed object grade and associated information, along with a color chart and a photograph of the molded object, can support customers in making appropriate decisions regarding the object's use, and support users in conducting appropriate transactions with suppliers and customers based on the object's grade.

[0137] 1 Network, 2 Output unit, 3 Production record DB, 4 Acquisition unit, 5 Extraction unit, 6 Grasping unit, 7 Effective feature determination unit, 8 Degradation degree assessment unit, 9 Proposal unit, 10 Characteristic evaluation system, 11 Network, 20 Characteristic evaluation device, 41 Processor, 42 Memory, 43 Storage device, 44 Communication device

Claims

1. A characterization system comprising a processor and a memory device, wherein the memory device stores a state recognition model that accepts feature quantities relating to the microstructure of a resin as input and outputs characteristics relating to the initiation of an oxidation reaction of the resin, and the processor acquires feature quantities relating to the microstructure of an object containing a resin, inputs the acquired feature quantities into the state recognition model to acquire characteristics relating to the initiation of an oxidation reaction of the object, and outputs the acquired characteristics.

2. A property evaluation system according to claim 1, wherein the property relating to the initiation of the oxidation reaction is the time or temperature at which the oxidation reaction of the resin is initiated.

3. A characterization system according to claim 2, characterized in that the time for the start of the oxidation reaction is the oxidation induction time (OIT), and the temperature for the start of the oxidation reaction is the oxidation onset temperature (IOT).

4. A characterization system according to claim 2, characterized in that the time of the start of the oxidation reaction is the time at which the time derivative of the exothermic peak curve due to the oxidation reaction of the object is maximized, and the temperature at which the oxidation reaction starts is the temperature at which the exothermic reaction of the object is maximized.

5. A characteristic evaluation system according to claim 3, characterized in that it calculates a representative value of the characteristics of the object from a plurality of characteristics obtained by inputting the characteristic quantities of the object into the state understanding model, calculates the degree of deterioration of the object based on the difference between a predetermined reference value and the representative value, and outputs the degree of deterioration of the object.

6. A characterization system according to claim 1, wherein the resin is an olefin-based resin, and the olefin-based resin includes either polypropylene or polyethylene.

7. A characterization system according to claim 1, wherein the feature quantity includes at least one of the following: a first feature quantity relating to spectral luminosity obtained by spectral measurement of ultraviolet and visible light; a second feature quantity relating to wide-angle X-ray diffraction obtained by wide-angle X-ray diffraction measurement; a third feature quantity relating to color obtained by color measurement; a fourth feature quantity relating to spectral luminosity obtained by Fourier transform infrared spectroscopy measurement; a fifth feature quantity relating to durometer hardness obtained by hardness measurement; and a sixth feature quantity relating to density or specific gravity obtained by specific gravity measurement.

8. A characterization system according to claim 7, wherein the first feature quantity is a feature quantity based on a reflection spectrum measurement in the range of 360-430 nm.

9. A characterization system according to claim 7, wherein the third feature quantity is a feature quantity based on color in the Lab color space under conditions in which the irradiated light includes ultraviolet light and under conditions in which ultraviolet light is not included, and the irradiated light under the ultraviolet light-free condition is characterized in that ultraviolet light less than 400 nm is cut off.

10. A characterization system according to claim 7, wherein the second feature quantity is based on a diffraction peak at at least one of the diffraction angles 15.0–19.6 degrees, 27.9–28.8 degrees, and 58.4–59.2 degrees.

11. The characteristic evaluation system according to claim 7, wherein the fourth characteristic quantity is 2920 cm due to CH2 antisymmetric expansion and contraction. ―１ And, due to the CH3 inverse symmetrical expansion and contraction, 2955 cm －１ A characterization system characterized by having a feature quantity based on at least one absorption peak.

12. A characteristic evaluation system according to claim 7, wherein the fifth characteristic quantity is a characteristic quantity based on the hardness at 1 to 300 seconds after loading, with the durometer hardness being classified as Type D according to the standards.

13. The characteristic evaluation system of claim 7, wherein the processor generates the state understanding model using at least one of the first feature, the second feature, the third feature, the fourth feature, the fifth feature, and the sixth feature.

14. A characteristic evaluation method comprising a processor and a memory device, wherein a state recognition model is stored in the memory device, which accepts feature quantities relating to the microstructure of a resin as input and outputs characteristics relating to the initiation of the oxidation reaction of the resin; and the processor acquires feature quantities relating to the microstructure of an object containing a resin, inputs the acquired feature quantities into the state recognition model to acquire characteristics relating to the initiation of the oxidation reaction of the object, and outputs the acquired characteristics.

15. A program for causing a characterization system comprising a processor and a memory device to execute a characterization method, wherein the memory device stores a state recognition model that accepts feature quantities relating to the microstructure of a resin as input and outputs characteristics relating to the initiation of an oxidation reaction of the resin, and the program causes the processor to execute the following steps: acquire feature quantities relating to the microstructure of an object containing a resin, input the acquired feature quantities into the state recognition model to acquire characteristics relating to the initiation of an oxidation reaction of the object, and output the acquired characteristics.

Images