Injection molding condition generation device, injection molding condition generation method, and program

Abstract

The present invention can achieve a reduction in prototyping and other loss costs, an increase in the number of usable material candidates, and an improvement in yield by enabling non-destructive prediction of quality variations in materials prior to injection molding and optimization of process conditions.　Provided is an injection molding condition generation device comprising one or more processors and one or more memory resources. The memory resources store material information in which correspondence relationships between fluorescence fingerprint data and actual measurement values of material characteristics are registered, the fluorescence fingerprint data including an excitation wavelength of each lot of a resin material, a combination of fluorescence wavelengths, and corresponding fluorescence intensities. The processor calculates a characteristic value of a material characteristic by using a first prediction model, generated using the material information, and the fluorescence fingerprint data of a target lot of the resin material, and calculates an injection molding condition of the target lot on the basis of the calculated characteristic value.

Description

Injection molding condition generating device, injection molding condition generating method and program

[0001] The present invention relates to an injection molding condition generation device, an injection molding condition generation method, and a program. This invention claims priority to Japanese Patent Application No. 2023-184540, filed on October 27, 2023, and the contents of that application are incorporated by reference into this application in designated states where incorporation by reference of documents is permitted.

[0002] Against the backdrop of ESG (Environmental Social Governance) investments and demands for resource circulation, the proportion of recycled materials used in plastic and resin products is on the rise. However, recycled materials are prone to variations in material properties between material lots, which can lead to variations in molded product quality. Therefore, while the use of recycled materials with little variation in material properties is expanding, there is a challenge in that it is difficult to expand the use of recycled materials with relatively large variations.

[0003] Patent Document 1 discloses an analytical method for obtaining a spectrum of a plastic material or the like using near-infrared spectroscopy and measuring desired characteristic values. Specifically, Patent Document 1 describes the following: "A method for analyzing desired characteristics of plastic or rubber using near-infrared light, comprising: (i) a step of measuring absorbance in the near-infrared region of the plastic or rubber; (ii) a step of obtaining corrected spectral data by subjecting near-infrared spectral data obtained by the absorbance measurement to second-order differentiation; and (iii) a step of multiplying the second-order derivative values ​​of absorbance at each wavelength of the corrected spectral data by regression coefficients obtained by PLS regression analysis and calculating the sum of the multiplied values; the obtained value indicates an estimated value of the desired property of the plastic or rubber."

[0004] Japanese Patent Application Laid-Open No. 2022-7236

[0005] The technology in Patent Document 1 uses near-infrared spectroscopy to obtain spectra of plastic materials and estimate their characteristic values. However, this technology is not intended for injection molding using recycled materials, and does not consider how to deal with cases where the characteristic values ​​do not meet the target characteristics. Furthermore, many recycled material pellets are colored black, which absorbs light in near-infrared spectroscopy, making highly accurate measurement difficult. Therefore, it is believed that the technology in Patent Document 1 is unlikely to solve the above-mentioned problems associated with injection molding using recycled materials.

[0006] The present invention has been made in consideration of the above-mentioned problems, and aims to reduce loss costs such as prototyping, increase the number of usable material candidates, and improve yields by making it possible to non-destructively predict quality fluctuations in materials before injection molding and to optimize process conditions.

[0007] The present application includes a plurality of means for solving at least part of the above-mentioned problems, examples of which are as follows: An injection molding condition generation device according to one aspect of the present invention that solves the above-mentioned problems is an injection molding condition generation device having one or more processors and one or more memory resources, wherein the memory resources store material information in which correspondences between fluorescence fingerprint data, including combinations of excitation wavelengths and fluorescence wavelengths and corresponding fluorescence intensities for each lot of resin material, and actual measured values ​​of material properties are registered, and the processor calculates characteristic values ​​of material properties using a first prediction model generated using the material information and the fluorescence fingerprint data for a target lot of the resin material, and calculates injection molding conditions for the target lot based on the calculated characteristic values.

[0008] According to the present invention, it is possible to non-destructively predict quality fluctuations in materials before injection molding and to optimize process conditions, thereby reducing loss costs such as prototyping, increasing the number of usable material candidates, and improving yields.

[0009] Problems, configurations, and effects other than those described above will become clear from the following description of the embodiments.

[0010] FIG. 1 is a diagram showing an example of the overall configuration of an injection molding condition generation system. FIG. 2 is a diagram showing an example of fluorescence fingerprint spectrum data. FIG. 3 is a diagram showing an example of material information. FIG. 4 is a diagram showing an example of process information. FIG. 5 is a diagram showing an example of process information. FIG. 6 is a flow diagram showing an example of a generation process of fluorescence fingerprint data. An image diagram related to the generation process of fluorescence fingerprint data. FIG. 7 is a flow diagram showing an example of a generation process of a first prediction model. FIG. 8 is a flow diagram showing an example of a generation process of a second prediction model. FIG. 9 is a flow diagram showing an example of a generation process of a second prediction model. FIG. 10 is a flow diagram showing an example of a process condition calculation process. FIG. 11 is a diagram showing the relationship between the second prediction model and viscosity and set temperature. FIG. 12 is a diagram showing the relationship between the second prediction model and tensile strength and the amount of additive added. FIG. 13 is a diagram showing an example of output screen information. FIG. 14 is a diagram showing an example of the hardware configuration of an injection molding condition generation device.

[0011] Each embodiment of the present invention will be described below with reference to the drawings. The embodiments are merely examples for explaining the present invention, and appropriate omissions and simplifications have been made for clarity of explanation. The present invention can be implemented in various other forms. Unless otherwise specified, each component may be singular or plural. The position, size, shape, range, etc. of each component shown in the drawings may not represent the actual position, size, shape, range, etc., in order to facilitate understanding of the invention. Therefore, the present invention is not necessarily limited to the position, size, shape, range, etc. disclosed in the drawings. While various types of information may be described using terms such as "table," "list," and "queue," various types of information may also be expressed using other data structures. For example, various types of information such as "XX table," "XX list," and "XX queue" may also be referred to as "XX information." When describing identification information, terms such as "identification information," "identifier," "name," "ID," and "number" are used, but these terms are interchangeable. In all drawings used to explain the embodiments, identical components are generally designated by the same reference numerals, and repeated description thereof will be omitted. Furthermore, in the following embodiments, the components (including element steps, etc.) are not necessarily essential unless otherwise specified or considered to be clearly essential in principle. Furthermore, when it is said that "consists of A," "is made of A," "has A," or "includes A," other elements are not excluded unless otherwise specified to refer to only that element. Similarly, in the following embodiments, when referring to the shape, positional relationship, etc. of components, etc., it includes those that are substantially similar or similar to that shape, etc., unless otherwise specified or considered to be clearly not essential in principle.

[0012] First Embodiment The injection molding condition generation device according to this embodiment is a device that uses a prediction model generated based on spectroscopic analysis of fluorescence fingerprint spectral data to non-destructively estimate the characteristic values ​​of material properties of materials used in injection molding (for example, recycled or reclaimed resin materials, etc.), and calculates appropriate process conditions (injection molding conditions) when the characteristic values ​​do not satisfy the quality standards.

[0013] In particular, the injection molding condition generation device assumes that there may be relatively large differences in characteristic values ​​between material lots, which are delivery units from suppliers, even for the same type of material, and compares the characteristic values ​​of the material lot to be processed (hereinafter sometimes referred to as the "target lot") calculated based on the predictive model with the quality standard values, and calculates appropriate process conditions for target lots that do not meet the standard values.

[0014] Specifically, the injection molding condition generation device calculates characteristic values ​​of material properties (particularly viscosity and tensile strength) in the target lot using fluorescence fingerprint spectral data obtained by spectroscopic measurement (particularly, measurement of spectrofluorometric intensity) of the pellets before injection molding and a first prediction model.

[0015] The injection molding condition generation device generates a first prediction model for calculating (estimating) the characteristic values ​​of the target lot based on material information that registers the relationship between the wavelength combinations that affect the characteristic values, etc., extracted from the fluorescence fingerprint spectrum data, and the actual measured values ​​of the material characteristics, and a predetermined learning algorithm (e.g., a multiple regression model, etc.).

[0016] Furthermore, if the characteristic values ​​of the target lot calculated based on the first prediction model are outside the range of the specification, i.e., if the characteristic values ​​do not satisfy the quality standard, the injection molding condition generation device uses the process information and the second prediction model to calculate appropriate values ​​for the set temperature (e.g., the melting temperature of the resin material or the temperature of the nozzle part of the injection molding machine), which is a parameter for controlling viscosity and is a process condition, and the amount of additive to be added, which is a parameter for controlling tensile strength and is a process condition.

[0017] The injection molding condition generation device generates a second prediction model to be used for calculating appropriate process conditions for the target material based on process information in which actual measured values ​​of material properties corresponding to each parameter value are registered in advance.

[0018] This injection molding condition generation device makes it possible to non-destructively predict quality fluctuations for a target lot of a given material before injection molding and optimize process conditions, thereby reducing loss costs such as prototyping, increasing the number of usable material candidates, and improving yields.

[0019] In addition to the set temperature and the amount of additive added, there are various other process conditions, such as the injection speed of the molten resin, mold clamping pressure, dwell pressure, speed-pressure control switching position, and mold temperature. In this embodiment, however, attention is focused on material properties such as viscosity and tensile strength, and the following explanation focuses on the set temperature and the amount of additive added, which affect the material properties, as process conditions.

[0020] <Overall Configuration of Injection Molding Condition Generation System 1000> Fig. 1 is a diagram showing an example of the overall configuration of an injection molding condition generation system 1000 including an injection molding condition generation device 100 according to this embodiment. As shown in the figure, the injection molding condition generation system 1000 includes the injection molding condition generation device 100, injection molding equipment 200, characteristic value measurement equipment 300, and an external server 400. These devices are connected to each other so as to be able to communicate with each other via a predetermined communication network N. The communication network N is, for example, the Internet, a LAN (Local Area Network), or a WAN (Wide Area Network).

[0021] <Injection Molding Equipment 200> The injection molding equipment 200 includes an injection molding machine 210 and a spectrophotometer 220. The injection molding machine 210 is a device that performs injection molding using, for example, virgin or recycled materials. The spectrophotometer 220 is a device that measures the spectrofluorometric intensity of pellets of recycled materials, etc., before injection molding. Specifically, the spectrophotometer 220 is a spectrophotometer that includes two spectroscopes, one for excitation and one for fluorescence, and measures the fluorescence intensity of a material at a given excitation wavelength and fluorescence wavelength. The spectrophotometer 220 outputs fluorescence fingerprint spectral data, which is three-dimensional data indicating the excitation wavelength, fluorescence wavelength, and fluorescence intensity, as measurement results. The spectrophotometer 220 also transmits the fluorescence fingerprint spectral data, along with the measured material type and lot identification information, to the injection molding condition generation device 100.

[0022] The timing of measurement is not limited, and may be, for example, performed when an execution instruction is received from a user of the injection molding equipment 200, or when an execution instruction is received from the injection molding condition generation device 100.

[0023] FIG. 2 shows an example of fluorescence fingerprint spectral data. As shown in the figure, the vertical axis of the fluorescence fingerprint spectral data represents the excitation wavelength, the horizontal axis represents the fluorescence wavelength, and the color represents the fluorescence intensity. This type of fluorescence fingerprint spectral data makes it possible to detect even the slightest differences in material components based on the combinations of discretized excitation wavelengths and fluorescence wavelengths and the fluorescence intensity corresponding to each combination of wavelengths. Therefore, by comparing fluorescence fingerprint spectral data with each other or analyzing the fluorescence fingerprints, it is possible to detect not only differences in material types, but also the presence of regulated materials and differences in lots.

[0024] <Characteristic Value Measuring Equipment 300> The characteristic value measuring equipment 300 is equipment having devices for measuring characteristic values ​​of material properties. Specifically, the characteristic value measuring equipment 300 has at least a viscosity measuring instrument for measuring the viscosity of the material and a tensile strength measuring instrument for measuring the tensile strength of the material. The characteristic value measuring equipment 300 performs measurements for material information used in generating the first prediction model and measurements for process information used in generating the second prediction model.

[0025] In the material information measurement, the viscosity measuring instrument and tensile strength measuring instrument measure the viscosity and tensile strength of the material type and lot for which the spectrofluorometric measurement was performed. Here, the viscosity measuring instrument measures the viscosity of the material at a constant temperature (e.g., 200°C), and the tensile strength measuring instrument measures the tensile strength of the material without the addition of additives. The additives are adjusters used to adjust the tensile strength characteristic value, which is a material property of the resin material. Furthermore, the viscosity measuring instrument and tensile strength measuring instrument transmit information (hereinafter sometimes referred to as "material information actual measured values"), which are the measurement results, and identification information of the measured material type and lot, to the injection molding condition generation device 100.

[0026] In the process information measurement, the viscosity measuring instrument and the tensile strength measuring instrument measure the viscosity at each temperature and the tensile strength at each amount of additive added. The viscosity measuring instrument and the tensile strength measuring instrument transmit information including the actual measurement values ​​as the measurement results and the identification information of the type and lot of the measured material (hereinafter, sometimes referred to as "actual measurement values ​​for process information") to the injection molding condition generation device 100.

[0027] The timing of measurement is not limited, and may be, for example, performed when an execution instruction is received from a user of the characteristic value measuring equipment 300, or when an execution instruction is received from the injection molding condition generating device 100.

[0028] <External Server 400> The external server 400 is a cloud server that provides various services. Specifically, the external server 400 accumulates various types of information related to material properties acquired from an external system (not shown) via the network N in a database. For example, the external server 400 acquires fluorescence fingerprint spectrum data and information related to material properties (e.g., information corresponding to actual measured values ​​for material information and actual measured values ​​for process information) from service users such as factories and companies, and stores these in the database.

[0029] When the external server 400 acquires new information from an external system and updates the database, it notifies the injection molding condition generating device 100 of this and synchronizes the databases between the external server 400 and the injection molding condition generating device 100.

[0030] In addition, in response to a reference request from an external system, the external server 400 discloses (provides), for example, information about the specified material properties and their property values ​​to the company or the like that has made the reference request.

[0031] <Details of Injection Molding Condition Generation Apparatus 100> The injection molding condition generation apparatus 100 is an apparatus for generating injection molding conditions. Specifically, the injection molding condition generation apparatus 100 has a processing unit 110, a storage unit 120, a communication unit 130, an input unit 140, and a display unit 150.

[0032] The processing unit 110 is a functional unit that performs processing executed by the injection molding condition generation device 100. As shown in the figure, the processing unit 110 has an information acquisition unit 111, a spectroscopic analysis unit 112, a DB update unit 113, a prediction model generation unit 114, and a process condition calculation unit 115.

[0033] The information acquisition unit 111 is a functional unit that acquires various types of information from the injection molding equipment 200, the characteristic value measuring equipment 300, and the external server 400. Specifically, the information acquisition unit 111 acquires fluorescence fingerprint spectrum data, actual measured values ​​for material information, and actual measured values ​​for process information from the injection molding equipment 200, the characteristic value measuring equipment 300, or the external server 400, and stores the acquired information in the storage unit 120.

[0034] The spectroscopic analysis unit 112 is a functional unit that performs spectroscopic analysis of the fluorescence fingerprint spectral data. Specifically, the spectroscopic analysis unit 112 uses multivariate analysis to extract, from the fluorescence fingerprint spectral data, discretized combinations of excitation wavelengths and fluorescence wavelengths that affect material properties and the fluorescence intensities corresponding to each combination. More specifically, the spectroscopic analysis unit 112 converts the wavelength range from the fluorescence fingerprint spectral data, removing unnecessary regions, into a one-dimensional data. The spectroscopic analysis unit 112 also extracts wavelength combinations that affect material properties and the corresponding fluorescence intensities from the one-dimensional data, and generates fluorescence fingerprint data that registers this data. Here, the spectroscopic analysis unit 112 extracts, as an effective wavelength range, at least one or more fluorescence wavelengths and their fluorescence intensities for a certain excitation wavelength, or at least one or more excitation wavelengths and corresponding fluorescence intensities for a certain fluorescence wavelength. The fluorescence fingerprint data generation process will be described in detail below.

[0035] The DB (database) update unit 113 is a functional unit that updates the database of the storage unit 120. Specifically, the DB update unit 113 updates the database by storing, in the material information DB 121, material information that associates the fluorescence fingerprint data generated by the spectroscopic analysis unit 112 with the actual measured values ​​for material information of the material type and lot that correspond to the fluorescence fingerprint data acquired from the characteristic value measurement equipment 300 or the external server 400.

[0036] 3 is a diagram showing an example of material information stored in the material information DB 121. As shown in the figure, the material information has a record in which fluorescence fingerprint data X and corresponding actual measured values ​​Y of characteristic values ​​are associated with each other for each material type and lot. The fluorescence fingerprint data X is information generated by the spectroscopic analysis unit 112 based on the fluorescence fingerprint spectrum data. The actual measured values ​​Y of characteristic values ​​are actual measured values ​​for material information acquired from the characteristic value measurement equipment 300, etc., and include viscosity Y. １ indicates the viscosity of the material measured at a constant temperature (e.g., 200°C), and the tensile strength Y ２ indicates the tensile strength measured in the absence of any additives.

[0037] Such material information is used as learning data to generate and update the first prediction model.

[0038] In addition, when the DB update unit 113 obtains actual measured values ​​for process information, which indicate the relationship between the temperature and viscosity of the material, or the relationship between the amount of additive added and the tensile strength, from the characteristic value measurement equipment 300, etc., it stores these values ​​in the process information DB 122, thereby updating the database.

[0039] 4 is a diagram showing an example of process information (relationship between temperature and viscosity) stored in the process information DB 122. As shown in the figure, the process information has records in which process conditions (temperatures) are associated with viscosity characteristic values ​​corresponding to the parameter values ​​of each process condition for each material type and lot.

[0040] 5 is a diagram showing an example of process information (relationship between the amount of additive added and tensile strength) stored in the process information DB 122. As shown in the figure, the process information has records for each material type and lot in which process conditions (additive amount %) are associated with characteristic values ​​of tensile strength corresponding to the parameter values ​​of each process condition.

[0041] Such process information is used to generate and update the second predictive model.

[0042] Returning to FIG. 1 , the prediction model generation unit 114 is a functional unit that generates prediction models. Specifically, the prediction model generation unit 114 generates a first prediction model used to calculate characteristic values ​​of a target lot and a second prediction model used to calculate process conditions. Details of the prediction model generation process will be described later.

[0043] The process condition calculation unit 115 is a functional unit that calculates the process conditions during injection molding. Specifically, the process condition calculation unit 115 calculates the appropriate set temperature and the amount of additive added, which are parameters for controlling the characteristic values ​​of the material properties, to ensure the desired quality of the molded product.

[0044] More specifically, the process condition calculation unit 115 calculates the characteristic values ​​of the target lot using the first prediction model and determines whether the characteristic values ​​meet the specifications. If the characteristic values ​​meet the specifications, the process condition calculation unit 115 identifies the parameter values ​​corresponding to the characteristic values ​​as process conditions. On the other hand, if the characteristic values ​​do not meet the specifications, the process condition calculation unit 115 calculates appropriate values ​​for the set temperature and the amount of additive added using the process information and the second prediction model. Details of the process condition calculation process will be described later.

[0045] Next, the storage unit 120 will be described. The storage unit 120 is a functional unit that stores various types of information. Specifically, the storage unit 120 stores a material information DB 121, a process information DB 122, quality standard information 123, process window information 124, and a learning algorithm 125.

[0046] As described above, the material information DB 121 and the process information DB 122 are databases that store material information and process information, respectively. Note that the material information and the process information have already been described, so detailed description thereof will be omitted.

[0047] The quality standard information 123 is information in which target values ​​of characteristic values ​​necessary to achieve the quality target of a molded product and the allowable fluctuation ranges based on the target values ​​(hereinafter, sometimes referred to as standard ranges) are registered for each type of material. Specifically, the quality standard information 123 registers target values ​​and standard ranges of characteristic values ​​such as viscosity and tensile strength in association with each type of material.

[0048] The process window information 124 is information in which a process window indicating a range in which process conditions can be changed is registered for each type of material. The range of changeable process conditions indicated by the process window is set in advance in consideration of preventing variations in the composition of the material due to, for example, temperature or the amount of additive added.

[0049] The learning algorithm 125 is an algorithm used when generating the first prediction model using material information. Specifically, the learning algorithm 125 includes, for example, multiple regression, partial least squares (PLS) regression, ridge regression, least absolute shrinkage and selection operator (LASSO) regression, gradient boosting regression, and random forest regression.

[0050] Next, the communication unit 130 will be described. The communication unit 130 is a functional unit that communicates information with external devices. Specifically, the communication unit 130 communicates information with the injection molding equipment 200, the characteristic value measurement equipment 300, and the external server 400, and transmits and receives various types of information.

[0051] The input unit 140 is a functional unit that accepts various inputs. Specifically, the input unit 140 accepts input of instructions and information from a user of the injection molding condition generation apparatus 100. The display unit 150 is a functional unit that generates output screen information and displays it on a display device provided in the injection molding condition generation apparatus 100 or the like. Specifically, the display unit 150 generates output screen information including information generated by the injection molding condition generation apparatus 100. The display unit 150 also displays the generated output screen information on a display device provided in the injection molding condition generation apparatus 100 or an external device connected via the communication unit 130 (e.g., the injection molding machine 210, the spectroscopic measurement device 220, the characteristic value measurement equipment 300, the external server 400, and other devices).

[0052] The injection molding condition generating device 100 has been described above.

[0053] <Details of Processing> Next, details of the processing executed by the injection molding condition generating device 100 will be described.

[0054] <<Fluorescence Fingerprint Data Generation Process>> Fig. 6 is a flow diagram showing an example of the fluorescence fingerprint data generation process. Fig. 7 is an image diagram related to the fluorescence fingerprint data generation process. This process is started, for example, when the spectroscopic analysis unit 112 acquires the fluorescence fingerprint spectrum data via the information acquisition unit 111. Note that this process may also be started based on an instruction from the user of the injection molding condition generation apparatus 100.

[0055] When the process starts, the spectroscopic analysis unit 112 performs data preprocessing to remove unnecessary regions (step S10). Specifically, the spectroscopic analysis unit 112 removes regions representing non-fluorescent components (e.g., regions where the excitation wavelength is greater than the fluorescence wavelength) and regions representing scattered light from the fluorescence spectrum data.

[0056] Next, the spectroscopic analysis unit 112 converts the fluorescence fingerprint spectral data after the data preprocessing into one-dimensional data (step S11). Note that the conversion of multidimensional data (in this example, three-dimensional data) into one-dimensional data can be performed using a known technique.

[0057] Next, the spectroscopic analysis unit 112 extracts from the one-dimensional full-wavelength data combinations of excitation wavelengths and fluorescence wavelengths that affect the characteristic values ​​of the material properties, and the fluorescence intensities corresponding to each combination (step S12), and then ends the processing flow. Note that the combinations of excitation wavelengths and fluorescence wavelengths that affect the characteristic values ​​may be identified, for example, based on regions in the fluorescence fingerprint spectral data where peaks in fluorescence intensity are found. Through this series of processes, the spectroscopic analysis unit 112 generates fluorescence fingerprint data that associates combinations of excitation wavelengths and fluorescence wavelengths with the fluorescence intensities corresponding to each combination.

[0058] The DB update unit 113 compares the fluorescence fingerprint data (X in FIG. 3) generated by the spectroscopic analysis unit 112 with the corresponding measured values ​​of viscosity and tensile strength (Y in FIG. 3) of the material type and lot. １ , Y ２ ) and the material information associated with each other are stored in the material information DB 121.

[0059] <<Process for Generating First Prediction Model>> Figure 8 is a flow diagram showing an example of the process for generating the first prediction model. Because the first prediction model is a learning model, it is updated by re-learning every time the material information DB 121, in which learning data (material information) is stored, is updated. Therefore, this process is executed, for example, when the material information DB 121 is updated. Note that this process may also be started based on an instruction from a user of the injection molding condition generation device 100.

[0060] When the process starts, the prediction model generator 114 acquires material information for each material type from the material information DB 121 (step S20). For example, in the example of Fig. 3, material information corresponding to all lots of material type K1 is acquired.

[0061] Next, the prediction model generation unit 114 generates a first prediction model using the acquired material information and a predetermined learning algorithm 125 (step S21). Specifically, the prediction model generation unit 114 generates the first prediction model through machine learning using values ​​of each lot and a predetermined regression equation.

[0062] When multiple regression, which is one of the learning algorithms 125, is used, the first prediction model is expressed by the following formulas (1) and (2). Formula (1) is used to estimate the viscosity of the target lot, and formula (2) is used to estimate the tensile strength of the target lot. ｎ indicates the multiple regression coefficient. ｎ corresponds to the fluorescence fingerprint data obtained for each lot. １ = a １ X １ +a ２ X ２ +...a １００ X １００ ... (1) Y ２ = a １ X １ +a ２ X ２ +...a １００ X １００ ... (2)

[0063] The generated first prediction model is stored in the storage unit 120. For materials for which a first prediction model has already been generated, the corresponding first prediction model is updated by the processes in steps S20 and S21.

[0064] Next, the prediction model generation unit 114 determines whether the first prediction models have been generated or updated for all material types (step S22), and if it determines that they have been updated (Yes in step S22), ends this flow. On the other hand, if it determines that they have not been generated or updated for all material types (No in step S22), the prediction model generation unit 114 performs the processes of steps S20 to S21 for the material types that have not yet been generated or updated.

[0065] <<Generation Process of Second Prediction Model (Set Temperature)>> Fig. 9 is a flow diagram showing an example of generation process of a second prediction model used when determining the set temperature, which is a process condition. This process is started, for example, when the process information DB 122 is updated. Note that this process may also be started based on an instruction from a user of the injection molding condition generation device 100.

[0066] When the process starts, the prediction model generator 114 acquires process information indicating the relationship between temperature and viscosity for each material type from the process information DB 122 (step S30). For example, in the example of Fig. 4, process information corresponding to all lots of material type K1 is acquired.

[0067] Next, the prediction model generator 114 generates a second prediction model using the acquired process information and Andrade's equation (step S31). Specifically, the prediction model generator 114 generates a second prediction model that indicates the relationship between viscosity and temperature by substituting the values ​​of each lot into Andrade's equation. Note that the second prediction model is expressed, for example, as a curve with a slope that indicates the relationship between viscosity and temperature.

[0068] Here, Andrade's equation is expressed by the following equation (3): where μ represents the viscosity of the molten resin, T represents the temperature, and a and b represent coefficients specific to the material. μ = a·exp(b / T) (3)

[0069] The generated second prediction model is stored in the storage unit 120. For materials for which a second prediction model has already been generated, the corresponding second prediction model is updated by the processes in steps S30 and S31.

[0070] Next, the prediction model generation unit 114 determines whether the second prediction models have been generated or updated for all material types (step S32), and if it determines that the second prediction models have been generated or updated (Yes in step S32), ends this flow. On the other hand, if it determines that the second prediction models have not been generated or updated for all material types (No in step S32), the prediction model generation unit 114 performs the processes of steps S30 to S31 for the material types that have not yet been generated or updated.

[0071] <<Generation Process of Second Prediction Model (Additive Addition Amount)>> Fig. 10 is a flow diagram showing an example of generation process of a second prediction model used when determining the additive addition amount, which is a process condition. This process is started, for example, when the process information DB 122 is updated. Note that this process may also be started based on an instruction from a user of the injection molding condition generation device 100.

[0072] When the process starts, the predictive model generation unit 114 obtains process information indicating the relationship between the amount of additive added and tensile strength for each material type from the process information DB 122, similar to the aforementioned step S30 (step S40).

[0073] Next, the prediction model generator 114 generates a second prediction model showing the relationship between the amount of additive added and the tensile strength using the acquired process information (step S41). The second prediction model is expressed, for example, as a linear function (response equation) in which the tensile strength increases as the amount of additive added increases.

[0074] The generated second prediction model is stored in the storage unit 120. For materials for which a second prediction model has already been generated, the corresponding second prediction model is updated by the processes in steps S40 and S41.

[0075] Next, the prediction model generation unit 114 determines whether the second prediction models have been generated or updated for all material types (step S42), and if it determines that the second prediction models have been generated or updated (Yes in step S42), ends this flow. On the other hand, if it determines that the second prediction models have not been generated or updated for all material types (No in step S42), the prediction model generation unit 114 performs the processes of steps S40 to S41 for the material types that have not yet been generated or updated.

[0076] <<Process Condition Calculation Processing>> Figure 11 is a flow diagram showing an example of the process condition calculation processing. This processing is started, for example, before injection molding is performed on a target lot of a predetermined material, that is, when the material lot to be injection molded is changed from the previous lot to the target lot. Note that this processing may also be started based on an instruction from a user of the injection molding condition generation device 100.

[0077] When the process starts, the process condition calculation unit 115 calculates the characteristic values ​​of the target lot of the material to be processed using the first prediction model (step S100). Specifically, the process condition calculation unit 115 acquires the fluorescence fingerprint spectral data of the target lot from the storage unit 120, and generates the fluorescence fingerprint data of the target lot based on processing such as data preprocessing and one-dimensionalization.

[0078] In addition, the process condition calculation unit 115 calculates the viscosity and tensile strength of the target lot by inputting the fluorescent fingerprint data into the first prediction model (the above-mentioned formulas (1) and (2)) for each of the viscosity and tensile strength.

[0079] Next, the process condition calculation unit 115 determines whether the calculated characteristic values ​​satisfy the specification values ​​(step S110). Specifically, the process condition calculation unit 115 determines that the calculated characteristic values ​​of viscosity and tensile strength satisfy the specification values ​​if they are within the specification range and the process window range. The specification range and the process window range are determined based on the registered information of the quality specification information 123 and the process window information 124, respectively.

[0080] If it is determined that the specifications are met (Yes in step S110), the process condition calculation unit 115 determines the parameter values ​​corresponding to the calculated viscosity and tensile strength characteristics as the set temperature and the amount of additive added as process conditions (step S120). Here, these parameter values ​​correspond to the viscosity and tensile strength measurement conditions registered in the material information used to generate the first prediction model, such as a set temperature of 200°C and an amount of additive added of 0 (a state in which no additive is added). After determining the parameter values ​​that will become process conditions, the process condition calculation unit 115 proceeds to step S130.

[0081] If the process condition calculation unit 115 determines in step S110 that the quality standard is not satisfied (No in step S110), the process condition calculation unit 115 proceeds to step S140.

[0082] In step S140, the process condition calculation unit 115 calculates process conditions that satisfy the specification values ​​using the calculated characteristic values ​​and the second prediction model. Specifically, the process condition calculation unit 115 uses the second prediction model to calculate the set temperature when the characteristic value of viscosity calculated by the first prediction model (in this example, the characteristic value of viscosity at a set temperature of 200°C) falls within the specification range. Similarly, the process condition calculation unit 115 uses the second prediction model to calculate the amount of additive to be added when the characteristic value of tensile strength calculated by the first prediction model (in this example, the characteristic value of tensile strength when the amount of additive added = 0) falls within the specification range.

[0083] 12 is a diagram showing the relationship between the second prediction model and the viscosity and the set temperature. As shown in the figure, the viscosity characteristic value P calculated by the first prediction model is outside the range of the specification width. In this case, the process condition calculation unit 115 moves the second prediction model (in the illustrated example, moves it parallel upward) so that the viscosity characteristic value P is positioned on the line of the second prediction model, thereby identifying (calculating) the set temperature (470 K in the illustrated example) when the viscosity characteristic value P is within the range of the specification width.

[0084] 13 is a diagram showing the relationship between the second prediction model and the tensile strength and the amount of additive added. As shown in the figure, the characteristic value Q of tensile strength calculated by the first prediction model is outside the range of the specification width. In this case, the process condition calculation unit 115 sets the second prediction model so that the characteristic value Q of tensile strength is located on the line of the second prediction model, and specifies (calculates) the amount of additive added when the characteristic value Q of tensile strength is within the range of the specification width (in the example shown in the figure, Δt%).

[0085] Next, the process condition calculation unit 115 determines whether the calculated process conditions (set temperature and additive amount) are within the process window (step S150). If it is determined that they are within the process window (Yes in step S150), the process condition calculation unit 115 determines the set temperature and additive amount calculated in step S140 as the process conditions (step S160) and proceeds to step S130. On the other hand, if it is determined that they are not within the process window (No in step S150), the process condition calculation unit 115 determines whether to output a warning message (e.g., a message urging the user to readjust the process conditions or reconsider using the target lot) (step S170), and proceeds to step S130.

[0086] Next, in step S130, the display unit 150 generates output screen information including the process conditions and the like, and displays it on the display device 620. Specifically, the display unit 150 generates output screen information including at least the set temperature and the amount of additive added, which are the process conditions for the target lot determined in step S120 or step S160, and displays it on the display device 620. Note that the display unit 150 may also generate output screen information including, in addition to the process conditions, for example, characteristic values ​​for each material property of the target lot, or, if step S170 has been executed, a warning message.

[0087] After performing the process of step S130, the display unit 150 ends this flow.

[0088] The information content included in the output screen information is not limited, and may include, for example, the content shown in FIG.

[0089] 14 is a diagram showing an example of output screen information. As shown in the figure, output screen information 500 has an input area 510 for the target lot, a characteristic value chart display area 520, a characteristic value result display area 530, a process condition generation instruction area 540, a process condition generation result display area 550, and a fluorescence data display instruction area 560.

[0090] When the display unit 150 receives an operation by the user in the input area 510 (for example, an operation on the pull-down menu 511) and a press of the lot update button via the input unit 140, it identifies the new lot (target lot) number of the resin material (recycled material) corresponding to the selected molded article name, the control number of the fluorescence fingerprint spectrum data, the characteristic of interest (material characteristic), the target value, and the specification range from the information stored in the memory unit 120, and displays them in the respective display fields. The molded article name and new lot number may also be input directly by the user via the input device.

[0091] Furthermore, when the display unit 150 receives an operation on the lot update button via the input unit 140, it displays a chart of the characteristic values ​​calculated using the first prediction model for the target lot identified by the new lot number in the chart display area 520. Note that the example shown in the figure shows a state in which the characteristic value for the tensile strength of the target lot: HIP001-004 has been added to the chart by pressing the lot update button.

[0092] Furthermore, the display unit 150 displays the result of the characteristic value relative to the standard value in the result display area 530. Specifically, if the calculated characteristic value does not satisfy the standard value, the display unit 150 displays a predetermined message based on the relationship between the characteristic value and the standard range. In the example shown, the display unit 150 displays a message such as "The predicted strength value is below the lower limit of the standard range. We recommend process adjustment by adding an additive" in the characteristic value result display area 530. On the other hand, if the calculated characteristic value satisfies the standard value, the display unit 150 displays a message indicating that the characteristic value is within the standard range and that production is feasible.

[0093] When the type of additive to be used is input in the process condition generation instruction area 540 based on the message indicating that the characteristic value does not satisfy the specification value, the display unit 150 displays the range of the process window of the additive in the display field based on the target lot number and the process window information 124. The range of the process window may also be input directly by the user.

[0094] Furthermore, when the process condition generation button is pressed with the type of additive and the range of the process window input, the display unit 150 displays the process conditions generated based on the second prediction model in the generation result display area 550. In the illustrated example, the amount of additive added, which is the calculated process condition, is shown. In step S170 described above, if the calculated process condition is outside the range of the process window, the display unit 150 displays a message in the generation result display area 550 prompting the user to readjust the process conditions.

[0095] Furthermore, when the display unit 150 receives a user operation on the fluorescence data display instruction area 560 (operation of the pull-down menu 561) and a press of the fluorescence data display button via the input unit 140, the display unit 150 retrieves the fluorescence fingerprint spectral data corresponding to the selected lot number from the memory unit 120 and opens and displays a separate window showing the spectral data.

[0096] The process condition calculation process has been described above.

[0097] This injection molding condition generation device makes it possible to non-destructively predict quality fluctuations for a target lot of a given material before injection molding and optimize process conditions, thereby reducing loss costs such as prototyping, increasing the number of usable material candidates, and improving yields.

[0098] In particular, the injection molding condition generation device generates a prediction model using the fluorescence fingerprint data extracted from the fluorescence fingerprint spectrum data and calculates the characteristic values ​​of the target lot. Therefore, it is possible to estimate the characteristic values ​​with high accuracy even for recycled resin materials, which are often colored black.

[0099] Furthermore, if the characteristic values ​​of the target lot do not satisfy the standard values, the injection molding condition generating device can generate process conditions that will make the characteristic values ​​satisfy the standard values ​​and present them to the user.

[0100] Furthermore, the injection molding condition generation device generates a prediction model using fluorescence fingerprint data that extracts only wavelength combinations that affect characteristic values, rather than using all wavelengths of the fluorescence fingerprint spectrum data, allowing for the construction of a highly accurate prediction model. Incidentally, if an unnecessarily large number of explanatory variables are used, such as when all wavelength data is used, the computational load required to construct the regression equation (prediction model) increases, leading to overfitting to the training data and a deterioration in prediction accuracy. The injection molding condition generation device removes unnecessary data and uses fluorescence fingerprint data composed of appropriate explanatory variables, allowing for the rapid generation of highly accurate prediction models.

[0101] Furthermore, the injection molding condition generation device can calculate temperature conditions with high accuracy for controlling viscosity by using an equation that follows a physical model, such as the Andrade equation, as the explanatory variable of the second prediction model.

[0102] In addition, the injection molding condition generation device displays output screen information including the generated process conditions, the calculated material property values, and warning messages, allowing the user to check the details of the necessary information regarding the target lot in advance before injection molding.

[0103] Second Embodiment An injection molding condition generating apparatus 100 according to a second embodiment detects the inclusion of unexpected foreign matter or restricted substances by analyzing a fluorescence fingerprint using fluorescence fingerprint spectrum data, and notifies the user of the injection molding condition generating apparatus 100. Specifically, the injection molding condition generating apparatus 100 detects the inclusion of restricted substances or foreign matter based on a comparison of fluorescence fingerprint data between lots of each material, and outputs a warning message (for example, a message notifying the user that an inspection is required).

[0104] More specifically, the spectroscopic analysis unit 112 of the injection molding condition generation device 100 generates each piece of fluorescence fingerprint data from the fluorescence fingerprint spectrum data of each lot of the same type of material. Furthermore, for each combination of excitation wavelength and fluorescence wavelength corresponding to each piece of fluorescence fingerprint data, the spectroscopic analysis unit 112 compares the similarity between the lots in terms of fluorescence intensity and the distribution position / range of the excitation wavelength and fluorescence wavelength at which peaks are found. The method for determining the similarity is not limited, and may be, for example, based on a comparison with predetermined information (not shown) that defines the similarity range of the fluorescence intensity and distribution position / range.

[0105] Furthermore, if the spectroscopic analysis unit 112 detects a material lot that exhibits a value outside a predetermined similarity range as a result of the comparison, it outputs a warning message to the user of the injection molding condition generation device 100 via the display unit 150. For example, fluorescence fingerprint data of a restricted substance may be stored in advance in the storage unit 120, and the spectroscopic analysis unit 112 may detect the presence of a restricted substance or the presence of a restricted substance at a level equal to or greater than a reference value based on a comparison with the fluorescence fingerprint data of the restricted substance.

[0106] This injection molding condition generation system can detect the presence of restricted substances or foreign matter using fluorescent fingerprint data. This allows users of the injection molding condition generation system to eliminate materials containing large amounts of foreign matter and select materials that take into account components restricted in the location (country) where the molded product will be used. This can contribute to improving yield and reducing waste costs.

[0107] <Modifications> In the above-described embodiment, the process conditions were determined when temperature (the temperature of the molten resin or the temperature of the nozzle of the injection molding machine 210) was used as a parameter for controlling the viscosity of the resin material, but the present invention is not limited to this, and the amount of an additive (flow improver) that adjusts the viscosity may be calculated as a process condition. Note that in this case, as in the above-described case, the injection molding condition generation device 100 can calculate the amount of flow improver to be added as a process condition by, for example, acquiring process information in which the amount of additive to be added for each viscosity is registered and using this to generate a second prediction model.

[0108] <Hardware Configuration of Injection Molding Condition Generation Device 100>

[0109] 15 is a diagram showing an example of the hardware configuration of the injection molding condition generating apparatus 100. As shown in the figure, the injection molding condition generating apparatus 100 has an input device 610, a display device 620, a processing device 630, a main storage device 640, an auxiliary storage device 650, a communication device 660, and a bus 670 that electrically interconnects these devices.

[0110] The input device 610 is, for example, a touch panel, a keyboard, a mouse, etc. The display device 620 is a display device such as a liquid crystal display or an organic display.

[0111] The processing device 630 is a processor such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit). The main storage device 640 is a memory device (memory resource) such as a RAM (Random Access Memory) or a ROM (Read Only Memory). The injection molding condition generation device 100 has at least one processor and memory resource.

[0112] The auxiliary storage device 650 is a non-volatile storage device capable of storing digital information, such as a so-called hard disk drive, a solid state drive (SSD), or a flash memory.

[0113] The communication device 660 is a wired communication device that performs wired communication via a network cable, or a wireless communication device that performs wireless communication via an antenna.

[0114] An example of the hardware configuration of the injection molding condition generating device 100 has been described above.

[0115] The processing unit 110, input unit 140, and display unit 150 of the injection molding condition generating device 100 are realized by a program that causes the processing device 630 to perform processing. This program is stored in the main storage device 640 or the auxiliary storage device 650, and is loaded onto the main storage device 640 and executed by the processing device 630 when the program is executed.

[0116] The storage unit 120 is realized by a main storage device 640, an auxiliary storage device 650, or a combination thereof. The communication unit 130 is realized by a communication device 660.

[0117] Furthermore, the above-described configurations, functions, processing unit 110, processing means, etc. of the injection molding condition generation device 100 may be partially or entirely realized in hardware, for example, by designing them as integrated circuits. The above-described configurations and functions may also be realized in software, with a processor interpreting and executing programs that realize the respective functions. Information such as programs, tables, and files that realize the respective functions can be stored in storage devices such as memory, hard disks, and SSDs, or in recording media such as IC cards, SD cards, and DVDs.

[0118] Furthermore, the present invention is not limited to the above-described embodiments and modifications, and includes various modifications within the scope of the same technical concept. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is possible to add, delete, or replace part of the configuration of each embodiment with other configurations.

[0119] In addition, in the above explanation, the control lines and information lines are those that are considered necessary for the explanation, and do not necessarily show all the control lines and information lines in the product. In reality, it can be considered that almost all components are interconnected.

[0120] DESCRIPTION OF SYMBOLS 1000... injection molding condition generation system, 100... injection molding condition generation device, 110... processing unit, 111... information acquisition unit, 112... spectroscopic analysis unit, 113... DB update unit, 114... prediction model generation unit, 115... process condition calculation unit, 120... memory unit, 121... material information DB, 122... process information DB, 123... quality standard information, 124... process window information, 125... learning algorithm, 130... communication unit, 140... input unit, 150... display unit, 200... injection molding equipment, 210... injection molding machine, 220... spectroscopic measurement device, 300... characteristic value measurement equipment, 400... external server, N... communication network

Claims

1. An injection molding condition generating device having one or more processors and one or more memory resources, wherein the memory resource stores material information in which a correspondence between fluorescence fingerprint data, including a combination of excitation wavelengths, fluorescence wavelengths and corresponding fluorescence intensities for each lot of resin material, and actual measured values ​​of material properties is registered, and the processor calculates characteristic values ​​of the material properties using a first prediction model generated using the material information and the fluorescence fingerprint data of a target lot of the resin material, and calculates injection molding conditions for the target lot based on the calculated characteristic values.

2. An injection molding condition generating device as described in claim 1, wherein the processor generates the fluorescence fingerprint data by extracting the excitation wavelengths, the fluorescence wavelengths and the corresponding fluorescence intensities that affect the characteristic values ​​from fluorescence fingerprint spectrum data that includes a plurality of the excitation wavelengths and a spectroscopic spectrum indicating the fluorescence wavelengths and the fluorescence intensities for the excitation wavelengths and is discretized at a predetermined wavelength width, and generates the first prediction model of a regression formula by machine learning using the correspondence between the fluorescence fingerprint data and the actual measured values ​​of the material characteristics and a predetermined learning algorithm.

3. An injection molding condition generating device as described in claim 1, characterized in that: the memory resource stores process information in which a correspondence between parameters for controlling the material characteristics and actual measured values ​​of the material characteristics corresponding to each value of the parameters is registered; and the processor, when the characteristic value of the target lot calculated based on the first prediction model does not satisfy a predetermined quality standard, calculates the injection molding conditions for the target lot using the characteristic value and a second prediction model generated using the process information.

4. An injection molding condition generating device as described in claim 3, characterized in that the processor, when the material property relates to the viscosity of the resin material, generates the second prediction model using the process information and Andrade's equation, and calculates the melting temperature of the resin material or the temperature of the nozzle of the injection molding machine as the injection molding condition.

5. An injection molding condition generating device according to claim 1, characterized in that the injection molding conditions include at least one of the melting temperature of the resin material or the temperature of the nozzle of the injection molding machine, and the amount of additive added.

6. An injection molding condition generating device according to claim 1, characterized in that the processor extracts, from the fluorescence fingerprint spectrum data, at least one or more fluorescence wavelengths and their fluorescence intensities for a certain excitation wavelength, or at least one or more fluorescence intensities corresponding to a certain fluorescence wavelength as effective wavelength ranges, and generates the fluorescence fingerprint data.

7. An injection molding condition generating device according to claim 3, characterized in that the processor updates the material information or the process information using information acquired from an external device.

8. An injection molding condition generating device as described in claim 1, characterized in that the characteristic values ​​of the material properties calculated by the first prediction model are characteristic values ​​related to the viscosity and tensile strength of the resin material.

9. An injection molding condition generating device according to claim 5, characterized in that the additive is an agent for improving the tensile strength or viscosity of the resin material.

10. An injection molding condition generating device according to claim 1, characterized in that the resin material is a regenerated material or a recycled material.

11. An injection molding condition generating device as described in claim 3, characterized in that the processor calculates the characteristic values ​​using the first prediction model or calculates the injection molding conditions using the second prediction model for the resin material, which is a regenerated or recycled material, for each lot, which is a delivery unit.

12. An injection molding condition generating device as described in claim 3, characterized in that the processor generates output screen information including the injection molding conditions calculated using the second predictive model, the characteristic values ​​of the material properties calculated using the first predictive model, and a predetermined warning message, and displays the output screen information on a predetermined device.

13. An injection molding condition generating device as described in claim 1, characterized in that the processor compares the fluorescent fingerprint data of each lot of the resin material, and detects and reports the presence of restricted substances and foreign matter based on the similarity of the fluorescent fingerprint data.

14. An injection molding condition generating method executed by an injection molding condition generating device having one or more processors and one or more memory resources, wherein the processor executes the following steps: storing in the memory resource material information in which a correspondence between fluorescence fingerprint data, including a combination of excitation wavelengths, fluorescence wavelengths and corresponding fluorescence intensities for each lot of resin material, and actual measured values ​​of material properties is registered; calculating characteristic values ​​of material properties using a first prediction model generated using the material information and the fluorescence fingerprint data of a target lot of the resin material; and calculating injection molding conditions for the target lot based on the calculated characteristic values.

15. A program that causes a computer to function as an injection molding condition generation device, the injection molding condition generation device having one or more processors and one or more memory resources, the memory resource storing material information in which correspondences between fluorescence fingerprint data including a combination of excitation wavelengths, fluorescence wavelengths and corresponding fluorescence intensities for each lot of resin material and actual measured values ​​of material properties are registered, the processor calculating characteristic values ​​of material properties using a first prediction model generated using the material information and the fluorescence fingerprint data of a target lot of the resin material, and calculating injection molding conditions for the target lot based on the calculated characteristic values.

Images