JP7118941B2 - RESIN MOLDING ANALYSIS METHOD, PROGRAM AND RECORDING MEDIUM - Google Patents

Description

The present invention relates to a resin molding analysis method, program and recording medium.

Conventionally, a resin molding analysis method is known (see Patent Document 1, for example).

In Patent Document 1, conditions for resin molding are set, molding simulation including flow analysis is performed, and molding constraint conditions regarding welds (resin confluence portions) and air traps (mixture of air bubbles) are set for the results of the molding simulation. A molding simulation method (resin molding analysis method) is disclosed in which it is determined whether or not the conditions are satisfied, and the molding simulation is repeated by changing the resin molding conditions.

JP 2005-169766 A

However, in the molding simulation method (resin molding analysis method) of Patent Document 1, the molding simulation is repeatedly performed by changing the resin molding conditions. Molding simulation should be performed. Therefore, there is a problem that the processing time increases when analyzing the resin molding.

The present invention has been made to solve the above-described problems, and one object of the present invention is to provide a resin molding method capable of suppressing an increase in processing time when analyzing resin molding. It is to provide a molding analysis method, a program and a recording medium.

In order to achieve the above object, a resin molding analysis method according to a first aspect of the present invention relates resin molding condition information to output values based on actual measurements of characteristics of the corresponding resin molded product. Estimates the characteristics of the resin molded product based on a plurality of accumulated resin molding information from the resin molding conditions that are accumulated as molding information and outputs them as characteristic information. Based on the obtained plural pieces of resin molding information, parameters for simulating the state of the resin at the time of molding and details of the characteristics of the molded product are estimated, and the estimated parameters are used to perform the simulation.

In the method for analyzing resin molding according to the first aspect of the present invention, by configuring as described above, the characteristic information of the resin molded product can be output without performing a simulation, so the simulation is performed once or multiple times. Compared to the case, it is possible to suppress the processing time from increasing when analyzing the resin molding. In addition, since the characteristics of the resin molded product are estimated based on a plurality of resin molding information that accumulates actual measured values of the characteristics of the resin molded product, the characteristics of the resin molded product are much faster than when estimating the characteristics only by simulation. can be estimated based on actual measurements. In addition, since a plurality of stored resin molding information can be used, the characteristics of the resin molded product can be estimated with higher accuracy than when the characteristics of the resin molded product are estimated individually.

In addition, since the parameters for simulation can be adjusted based on a plurality of stored resin molding information, the simulation parameters can be easily set to accurately estimate the characteristics of the resin molded product. be able to.

In the resin molding analysis method according to the first aspect, preferably, machine learning is performed based on a plurality of pieces of resin molding information including resin molding condition information and measured values obtained by actually measuring characteristics of the corresponding resin molded product, and a plurality of Based on the resin molding information, parameters are optimized when performing a simulation of the state of the resin at the time of molding and the details of the characteristics of the molded product. With this configuration, the parameters are optimized when the simulation is performed, so that the parameters for the simulation that can accurately estimate the characteristics of the resin molded product can be more easily set.

In the configuration for performing machine learning based on the plurality of pieces of resin molding information, preferably, the resin molding conditions are classified based on the plurality of pieces of resin molding information, and based on the input resin molding conditions, the resin of the corresponding classification Using the parameters of the molding information, a simulation is performed of the state of the resin during molding and the details of the characteristics of the molded product. According to this configuration, the resin molding information to be used can be narrowed down by classifying a plurality of accumulated resin molding information, so that the parameters of the corresponding resin molding information can be easily set.

In the resin molding analysis method according to the first aspect, preferably, the resin molding conditions are classified based on a plurality of pieces of resin molding information, and from the input resin molding conditions, based on the resin molding information of the corresponding classification, The characteristics of the resin molded product are estimated and output as characteristic information. With this configuration, it is possible to more accurately and quickly estimate the characteristics of the resin molded product based on the corresponding classification of the resin molding information.

In the configuration for classifying the resin molding conditions based on the plurality of pieces of resin molding information, preferably, the classification of the resin molding conditions includes the size of the molded product, the complexity of the molded product, the shape of the molded product, the resin data, and the molding conditions. At least one of these is digitized. With this configuration, by quantifying at least one of the size of the molded product, the complexity of the molded product, the shape of the molded product, the resin data, and the molding conditions, the resin molding conditions can be easily determined for each similar condition. can be classified into

In the resin molding analysis method according to the first aspect, preferably, the characteristics of the molded product are at least warpage of the molded product, pressure during molding, temperature during molding, orientation of the molded product, and weld state of the molded product. including one. With such a configuration, by inputting the resin molding conditions, the characteristics of the molded product can be selected from among the warpage of the molded product, the pressure during molding, the temperature during molding, the orientation of the molded product, and the weld state of the molded product. At least one can be obtained.

In the resin molding analysis method according to the first aspect, preferably, the characteristics of the molded product include the filling pattern of the resin, the appearance characteristics of the molded product, the resin physical property values of the molded product, the distortion during molding, the mass of the molded product, and the molded product. Product length, part distortion, part stress, part mechanical properties, part thermal properties, part electrical properties, part chemical properties, part molding properties, molding It includes at least one of the optical properties of the product and the color tone of the molded product. With this configuration, by inputting the resin molding conditions, the characteristics of the molded product, such as the filling pattern of the resin, the appearance characteristics of the molded product, the resin physical property values of the molded product, the strain during molding, the mass of the molded product, part length, part strain, part stress, part mechanical properties, part thermal properties, part electrical properties, part chemical properties, part molding properties, At least one of the optical properties of the molded article and the color tone of the molded article can be obtained.

In the resin molding analysis method according to the second aspect of the present invention, the resin molding condition information and the output value based on the actual measurement value of the characteristics of the corresponding resin molded product are associated with each other, accumulated as resin molding information, and inputted. Based on the accumulated resin molding information from the resin molding conditions obtained, the characteristics of the resin molded product are estimated and output as characteristic information. and an interpolated value obtained by interpolating the actual measured value of the characteristic by performing machine learning.
In the resin molding analysis method according to the first aspect, preferably, the output value based on the measured value of the characteristics of the resin molded product is obtained by performing machine learning on the measured value of the characteristics and the measured value of the characteristics. and interpolated values interpolated by With this configuration, even if the number of measured values is not sufficient, the interpolated value can be obtained by machine learning, so the number of output value data can be easily increased.

In the resin molding analysis method according to the first or second aspect, preferably, physical property values of the same kind of resin material in different states are predicted based on known physical property values of the resin material. With this configuration, even if all the physical property values of the resin material required for the simulation are not known due to differences in resin material grade (indicator indicating the superiority or inferiority of properties in stages), composition, or manufacturing method, a known resin material can be used. Based on the physical property values of the material, it is possible to predict and acquire the physical property values of the resin materials of the same type and in different states used for resin molding, so that simulation can be performed based on the acquired physical property values of the resin material. . Moreover, since there is no need to actually measure the physical property values of the unknown resin material, there is no need to separately provide a measuring device for measuring the physical property values of the resin material.

In this case, preferably, the physical property values of the same kind of resin material in different states are predicted based on the known basic information of the resin material and the content of the recycled material contained in the resin material. With this configuration, even when the recycled material that has undergone molding and has received heat is mixed with the resin material and used, the physical property values of the resin material can be predicted and obtained. It is possible to accurately perform a simulation when resin molding is performed using the material.

A program according to a third aspect of the present invention causes a computer to execute the resin molding analysis method according to the first aspect.

In the program according to the third aspect of the present invention, by causing a computer to execute the resin molding analysis method according to the first aspect, it is possible to suppress an increase in the processing time when analyzing the resin molding.

A storage medium according to a fourth aspect of the present invention stores a program according to the second aspect and is computer-readable.

A storage medium according to a fourth aspect of the present invention is a computer-readable storage medium capable of suppressing an increase in processing time when analyzing resin molding by recording the program according to the second aspect. A recording medium can be provided.

According to the present invention, as described above, it is possible to suppress an increase in processing time when analyzing resin molding.

1 is a block diagram showing a configuration example for implementing a resin molding analysis method according to one embodiment; FIG. FIG. 4 is a diagram for explaining accumulation of resin molding information according to one embodiment; FIG. 4 is a diagram for explaining output of characteristic information according to one embodiment; FIG. 4 is a diagram for explaining parameter optimization when performing a simulation according to an embodiment; FIG. 4 is a diagram for explaining classification of resin molding conditions according to one embodiment; FIG. 4 is a diagram for explaining interpolation of characteristics of a resin molded product according to one embodiment; 4 is a flowchart for explaining a first example of resin molding analysis processing according to one embodiment; 9 is a flowchart for explaining a second example of resin molding analysis processing according to one embodiment; FIG. 4 is a diagram for explaining prediction processing of resin physical property values according to one embodiment; 4 is a flow chart for explaining the learning phase according to one embodiment; 4 is a flow chart for explaining the operation phase according to one embodiment;

Embodiments embodying the present invention will be described below with reference to the drawings.

A resin molding analysis method according to one embodiment will be described with reference to FIGS.

The resin molding analysis method according to the present embodiment includes at least one of the warpage of the molded product, the pressure during molding, the temperature during molding, the orientation of the molded product, and the weld state of the molded product when molding the resin. This is a method for analyzing the properties of molded products. In addition, when developing a new molded product, we propose suitable parameters based on the product category and shape feature values of the product.

(Device configuration example)
The resin molding analysis method according to this embodiment can be implemented by causing the computer 1 to execute the program 3a. The resin molding analysis method can be implemented by, for example, an apparatus configuration as shown in FIG. The computer 1 is configured to be able to execute the program 3a. A resin molding analysis apparatus 100 is configured by causing the computer 1 to execute the program 3a. Part or all of the processing performed by causing the computer 1 to execute the program 3a may be performed by hardware such as a dedicated arithmetic circuit.

In the configuration example of FIG. 1, a computer 1 includes one or more processors 2 such as a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), a storage unit including a storage device, and the like. 3. The storage device is, for example, a hard disk drive, a semiconductor storage device, or the like.

The computer 1 can perform resin molding analysis by causing the processor 2 to execute the program 3 a stored in the storage unit 3 . The program 3a may be read from the recording medium 7, or may be provided from an external server or the like via a network such as the Internet or a transmission path 8 such as a LAN (Local Area Network). The recording medium 7 is a computer-readable recording medium such as an optical disk, a magnetic disk, or a nonvolatile semiconductor memory, and records the program 3a.

In addition to the program 3a, the storage unit 3 stores various kinds of analysis data 3b used for resin molding analysis. The analysis data 3b includes resin molding information including resin molding condition information and characteristics of the resin molded product, parameters for simulating the state of the resin at the time of molding and details of the characteristics of the molded product, product category information, and shape characteristics. Quantities, groups of optimal parameters, error rates, numerical data used for analysis, data on analysis conditions, and the like are stored.

The computer 1 also includes a display unit 4 such as a liquid crystal display device, an input unit 5 comprising input devices such as a keyboard and a mouse, and a reading unit 6 for reading programs 3a and various data from a recording medium 7 . The reading unit 6 is a reader device or the like corresponding to the type of the recording medium 7 . A user can input the data of the analysis conditions using the input unit 5 . The analysis data 3b may be read from a recording medium created by the user, or may be created by the user in an external server or the like and acquired from the external server via the transmission path 8. FIG.

(Resin molding analysis method)
Next, a resin molding analysis method will be described. In the present embodiment, as shown in FIG. 2, the resin molding condition information and the output values based on the measured values of the characteristics of the corresponding resin molded product are associated and accumulated as the resin molding information. Then, as shown in FIG. 3, based on the accumulated plural pieces of resin molding information, the characteristics of the resin molded product are estimated and output as characteristic information.

The resin molding analysis method includes a learning phase and an operation phase. In the learning phase, resin molding information is accumulated as shown in FIG. In the learning phase, the accumulated resin molding information is learned by machine learning. Specifically, resin molding information is stored by inputting input condition definitions as resin molding condition information in association with measured values obtained by actually measuring characteristics of a molded product. Input condition definitions include product category information (use, field), analysis (calculation) mesh information (element type, number of elements, number of nodes, division conditions, element quality), resin data (resin manufacturer, grade name, base resin, latent heat , solidification temperature, density, specific heat, thermal conductivity, melt viscosity, PVT data, elastic modulus, Poisson's ratio, coefficient of linear expansion, molding shrinkage, mechanical strength, physical properties of reinforcing material, content of reinforcing material, viscoelastic properties (prony series, shift factor), optical properties (stress optical coefficient, photoelastic coefficient, refractive index, molecular structure, gelation reaction rate, curing reaction heat)), molding conditions (time, filling rate, upper limit of pressure, screw position, screw speed , flow rate, metering position, resin temperature, mold temperature, VP switching timing, holding pressure, holding pressure time, in-mold cooling conditions, cycle time), mold conditions (gate position, number of gates, runner layout, cooling circuit, protrusion pin arrangement), boundary conditions (nozzle flow rate/pressure, heat transfer coefficient, ambient temperature, ambient humidity), molding machine information (molding machine manufacturer name, molding machine model number, maximum injection speed, maximum injection pressure, maximum holding pressure, (screw diameter, maximum clamping force, maximum injection volume), and molded product rigidity information. The properties of the molded product include warpage of the molded product, pressure during molding, temperature during molding, orientation of the molded product, and weld condition of the molded product. For example, the characteristics of a molded product include the appearance quality of the molded product (weld lines, flow marks, sink marks, poor transferability, etc.), physical property quality (residual stress, strength, rigidity), and dimensional quality (warpage, sink marks, transferability, etc.). ), including optical properties (birefringence).

In addition, the characteristics of the molded product are the filling pattern of the resin, the appearance characteristics of the molded product, the resin physical properties of the molded product, the distortion during molding, the mass of the molded product, the length of the molded product, the distortion of the molded product, and the At least one of stress, mechanical properties of the molded article, thermal properties of the molded article, electrical properties of the molded article, chemical properties of the molded article, molding properties of the molded article, optical properties of the molded article, color tone of the molded article contains one. Further, the properties of a molded product include at least those defined by actual measurements, and include not only quantitative properties but also qualitative indices. Properties of molded products as measurable actual values (physical quantities) include weight, length, warpage (strain), stress, mechanical properties, thermal properties, electrical properties, resin physical properties, and chemical properties. , molding properties, optical properties, appearance properties, and color.

Also, in the learning phase, the shape feature amount is calculated from the three-dimensional CAD data or the analysis mesh. Also, a recommended module is selected from user information (license information). Also, analysis conditions are generated from the parameter factors and the number of levels. Then, a simulation is performed and an analysis is performed. The difference between the analysis result output by the simulation and the measured value is obtained, and the error rate is evaluated. An optimum group of parameters is obtained from the error rate between the analysis result and the actual measurement. In other words, machine learning is performed based on a plurality of pieces of resin molding information including resin molding condition information and measured values obtained by actually measuring the characteristics of the corresponding resin molded product, and based on the plurality of pieces of resin molding information, the amount of resin during molding is determined. Optimizing parameters during simulation with details of conditions and part properties. Then, product category information, shape features, optimal parameter groups, and error rates are saved. Also, the correlation between the product category information, the shape feature amount, the group of optimal parameters, and the error rate is calculated.

Also, in the learning phase, the degree of contribution of multiple influencing factors to the properties of the molded product (for example, warpage) is calculated. Influencing factors include the shape of the molded product, molding conditions, and properties of the resin. As for the degree of contribution, the ratio of contribution is calculated for each of the plurality of influencing factors. The ratio of the contribution of an influencing factor can be determined, for example, by varying one influencing factor and comparing a plurality of resin molding information in which the other influencing factors are fixed, how much the variation of the influencing factor contributes to the properties of the molded product. Calculate whether

The parameters are parameters that affect mold cooling analysis (refrigerant temperature, refrigerant flow rate), parameters that affect filling/holding pressure cooling analysis (resin data, product thickness, influence of air in the mold), and orientation distribution analysis. (coefficient of interference between fibers, fiber content, fiber length, aspect ratio), parameters that affect analysis of shrinkage distribution (timing of shrinkage start, solidification state), parameters that affect analysis of physical property distribution (elastic modulus, Poisson's ratio, viscoelastic properties), etc. Parameters are product category information (use, field), analysis mesh information (element type, number of elements, number of nodes, division conditions, element quality), resin data (resin manufacturer, grade name, base resin, latent heat, solidification temperature , density, specific heat, thermal conductivity, melt viscosity, PVT data, elastic modulus, Poisson's ratio, coefficient of linear expansion, molding shrinkage, mechanical strength, physical properties of reinforcing material, content of reinforcing material, viscoelastic properties (Prony series, shift factor), optical properties (stress optical coefficient, photoelastic coefficient, refractive index, molecular structure, gelling reaction rate, curing reaction heat)), molding conditions (time, filling rate, upper limit of pressure, screw position, screw speed, flow rate, metering position, resin temperature, mold temperature, VP switching timing, holding pressure, holding pressure time, cooling conditions in the mold, cycle time), mold conditions (gate position, number of gates, runner layout, cooling circuit, ejector pin arrangement) , boundary conditions (nozzle flow rate/pressure, heat transfer coefficient, ambient temperature, ambient humidity), molding machine information (molding machine manufacturer name, molding machine model number, maximum injection speed, maximum injection pressure, maximum holding pressure, screw diameter, (maximum mold clamping force, maximum injection capacity). The parameters include the shape to be analyzed (molded article shape, runner shape, mold shape, gate position), calculation model to be analyzed, resin physical property values, and shape feature amounts. Here, as the number of types of parameters increases, the work for adjusting the parameters becomes more complicated. In the present embodiment, since an optimal parameter group can be obtained by machine learning, it is possible to prevent the task of setting parameters from becoming complicated.

Also, in the learning phase, the accuracy of simulation using parameters optimized by machine learning from a plurality of accumulated resin molding information is confirmed. Specifically, the accuracy of the simulation is confirmed by inputting resin molding condition information having measured values and comparing the characteristics of the resin molded product output by the simulation with the characteristics of the resin molded product based on the measured values. For example, about 80% of the measured value data is used for machine learning, and about 20% of the measured value data is used to confirm the accuracy of simulation using parameters optimized by machine learning.

In the operation phase, as shown in FIG. 3, the characteristics of the resin molded product are estimated from the input resin molding conditions based on a plurality of accumulated resin molding information, and output as characteristic information. Also, from the input resin molding conditions, based on a plurality of accumulated resin molding information, parameters for simulating the state of the resin at the time of molding and details of the characteristics of the molded product are estimated, and the estimated parameters A simulation is performed using In other words, in the operation phase, simulation is performed using the optimum parameters registered for each category. Also, after selecting a category, a simulation is performed using the optimum parameters of information with similar shape feature amounts.

Moreover, in this embodiment, as shown in FIG. 5, the resin molding conditions are classified based on a plurality of pieces of resin molding information. The resin molding conditions are classified by quantifying at least one of the size of the molded product, the complexity of the molded product, the shape of the molded product, the resin data, and the molding conditions. For example, the classification of resin molding conditions is performed by quantifying feature quantities (for example, volume, surface area, thickness distribution, complexity, self-organizing map, etc.) representing the shape of the molded product. The size of the molded product is quantified based on the surface area, volume, etc. of the molded product. Also, the complexity of the molded product and the shape of the molded product are quantified based on protrusions, notches, holes, curvatures, etc. of the molded product. Resin data are quantified based on density, specific heat, thermal conductivity, Young's modulus, Poisson's ratio, linear expansion coefficient, melt viscosity, PVT data, reinforcing material physical properties, viscoelastic properties, and optical properties. Molding conditions are quantified based on injection time, resin temperature, mold temperature, VP switching timing, holding pressure conditions, cooling conditions, cycle time, molding machine conditions, and the like.

In addition, in this embodiment, based on the input resin molding conditions, the resin molding information parameters of the corresponding classification are used to simulate the state of the resin at the time of molding and the details of the characteristics of the molded product. For example, resin molding information of the same classification or a similar classification is presented based on the input resin molding conditions. By selecting resin molding information by the user, a simulation is performed with parameters set based on the selected resin molding information. In this case, the characteristics of the resin molded product are estimated by mathematical analysis.

Also, based on the resin molding information of the corresponding classification from the input resin molding conditions, the characteristics of the resin molded product are estimated and output as characteristic information. For example, resin molding information of the same classification or a similar classification is selected based on the input resin molding conditions. Then, from the resin molding information of the selected class, the characteristics of the resin molded product under the input resin molding conditions are estimated and output. In this case, the properties of the resin molded product are estimated by statistical analysis.

Further, in the present embodiment, the output value based on the actual measurement value of the characteristics of the resin molded product is obtained by combining the actual measurement value of the characteristics and the interpolated value obtained by interpolating the actual measurement values of the characteristics by performing machine learning. include. As shown in FIG. 6, based on the resin molding condition information and the actually measured values of the characteristics of the resin molded product associated with the resin molding condition information, machine learning is used to interpolate insufficient conditions.

(Measurement method of actual value)
A method for measuring actual values will be described.

The orientation of the resin of the molded product is measured by an X-ray transmission method. The birefringence of the molded product is measured with a birefringence measuring instrument. The pressure during molding is measured by a pressure sensor provided in the mold. The temperature at the time of molding is measured by a temperature sensor provided in the mold or by thermography by capturing an image of the mold. The displacement of the resin material is measured by a contact three-dimensional measuring device or a non-contact three-dimensional measuring device.

(machine learning)
An example of machine learning used in the learning phase will be described.

Machine learning is performed by methods such as regression, decision tree, neural network, Bayes, clustering, and ensemble learning, for example.

(Resin molding analysis processing)
An overview of a first example of the resin molding analysis process will be described with reference to FIG. The resin molding analysis process is executed by the computer 1 (processor 2).

As shown in FIG. 7, in step S1, input resin molding conditions (analysis mesh, molding conditions, resin data) are acquired. In step S2, based on the accumulated resin molding information, the characteristic (warp) of the resin molded product is statistically estimated and output. In this case, the contribution of the warpage influencing factor is calculated and presented.

At step S3, it is determined whether or not an instruction to perform a warpage analysis has been given. If the warp analysis is to be performed, the process proceeds to step S4, and if the warp analysis is not to be performed, the process proceeds to step S5. In step S4, simulation analysis is performed using the optimum parameters. As a result, the warpage is also analyzed by simulation separately from the statistical analysis in step S2. After that, the process proceeds to step S5.

In step S5, it is determined whether or not an instruction to take countermeasures against warpage has been issued. If warp countermeasures are to be taken, the process proceeds to step S6, and if warp countermeasures are not to be taken, the process proceeds to step S7. In step S6, the form factors are analyzed, the optimum conditions are analyzed, and a warp reduction material is adopted to take countermeasures against warpage. After that, the process proceeds to step S7.

At step S7, it is determined whether or not an instruction to end the resin molding analysis process has been given. If not terminated, the process proceeds to step S8, and if terminated, the resin molding analysis process is terminated. In step S8, the conditions are changed and the process returns to step S3.

An outline of a second example of the resin molding analysis process will be described with reference to FIG. The resin molding analysis process is executed by the computer 1 (processor 2).

As shown in FIG. 8, in step S11, input resin molding conditions (analysis mesh, molding conditions, resin data) are acquired. In step S12, based on the accumulated resin molding information, the characteristics of the resin molded product are statistically estimated and output. In this case, the contribution of the influencing factor is calculated and presented.

In step S13, it is determined whether or not a molding simulation (including mold cooling analysis, material filling analysis, holding pressure/cooling analysis, warpage analysis, etc.) is instructed. When molding simulation is performed, the process proceeds to step S14, and when molding simulation is not performed, the process proceeds to step S15. In step S14, simulation analysis is performed using the optimum parameters. As a result, the warpage is also analyzed by simulation separately from the statistical analysis in step S12. After that, the process proceeds to step S15.

In step S15, it is determined whether or not an instruction to take countermeasures has been given. If countermeasures are to be taken, the process proceeds to step S16, and if countermeasures are not to be taken, the process proceeds to step S17. In step S16, the shape of the molded product, the molding conditions, and the mold requirements (gate position, runner layout, nested structure) are changed, and countermeasures are taken. After that, the process proceeds to step S17.

At step S17, it is determined whether or not an instruction to end the resin molding analysis process has been given. If not terminated, the process proceeds to step S18, and if terminated, the resin molding analysis process is terminated. In step S18, the conditions are changed and the process returns to step S13.

Further, in this embodiment, the physical property values of the same type of resin material in different states are predicted based on the known physical property values of the resin material. Specifically, based on the known basic information of the resin material and the content of the recycled material contained in the resin material, the physical property values of the same kind of resin material in different states are predicted.

Here, it is said that there are about 100,000 resin grades (indices indicating the superiority or inferiority of properties in stages) of resin materials handled by resin manufacturers around the world. Formability, thermal properties, mechanical properties, etc. are required for forming simulation. However, some of the data necessary for simulation of resin molding (density, deflection temperature under load, linear expansion coefficient, molding shrinkage, MFR, elasticity ratio, reinforcement content, etc.). In addition, users who use molding simulation usually use the material data built into the software, but often measure separately if the necessary material data is not available. In most cases, special measuring equipment is used to measure material data. Therefore, the cost is also high.

In addition, in order to reduce costs, some recycled materials are sometimes mixed with virgin materials to use resin materials for molding. Recycled materials are known to have already undergone heat history or fractured reinforcement, and their physical properties are significantly different from those of virgin materials. The physical properties of the resin material also change depending on the content of the recycled material. However, considering the labor and cost of measurement, it is difficult to prepare material data for accurate simulation when recycled materials are mixed.

Therefore, in the process of predicting physical property values of an unknown resin material, material physical property values for molding simulation are generated when materials having a plurality of compositions are combined.

For example, as shown in FIG. 8, when the basic information of the resin material corresponding to the resin grade and the recycled material content rate are input, the resin physical properties are output. The basic information is based on the catalog information provided by the resin manufacturer. As shown in FIG. 8, the physical properties of the resin materials of grade A, grade B, and grade C are known, and the physical properties of the resin material of grade C are unknown. In this case, the physical property values of an unknown resin material used for resin molding are predicted based on the physical property values of a plurality of known resin materials. Also, by machine learning to output resin physical property values based on the basic information and the recycled material content rate, the material physical property values when an arbitrary recycled material is contained are predicted.

As shown in FIG. 9, in the learning phase, resin molding conditions (basic information, recycled material content rate) are input in step S21. In step S22, molding simulation is performed. In step S23, the difference (error) between the analysis result (simulation result) and the measured value is extracted. In step S24, machine learning is performed on the characteristics (feature amounts) and errors of the resin molded product.

As shown in FIG. 10, in the operation phase, resin molding conditions (basic information, recycled material content rate) are input in step S31. In step S32, optimum parameters (resin physical property values) are estimated from the feature values of the molded product. In step S33, molding simulation is performed based on the estimated parameters to predict characteristics after molding.

(Effect of this embodiment)
Next, the effects of this embodiment will be described.

In this embodiment, as described above, the characteristics of the resin molded product are estimated from the input resin molding conditions based on a plurality of accumulated resin molding information, and output as characteristic information. As a result, it is possible to output the characteristic information of the resin molded product without performing a simulation, so the increase in processing time when analyzing the resin molding can be avoided compared to the case where the simulation is performed once or multiple times. can be suppressed. In addition, since the characteristics of the resin molded product are estimated based on a plurality of resin molding information that accumulates actual measured values of the characteristics of the resin molded product, the characteristics of the resin molded product are much faster than when estimating the characteristics only by simulation. can be estimated based on actual measurements. In addition, since a plurality of stored resin molding information can be used, the characteristics of the resin molded product can be estimated with higher accuracy than when the characteristics of the resin molded product are estimated individually.

In addition, in this embodiment, as described above, the resin state at the time of molding and details of the characteristics of the molded product are simulated based on a plurality of stored resin molding information from the input resin molding conditions. Estimate the actual parameters, and perform a simulation using the estimated parameters. This makes it possible to adjust parameters for simulation based on a plurality of stored resin molding information, making it easy to set simulation parameters that enable accurate estimation of the characteristics of resin molded products. can do.

Further, in the present embodiment, as described above, machine learning is performed based on a plurality of pieces of resin molding information including the resin molding condition information and the measured values obtained by actually measuring the characteristics of the corresponding resin molded products. Based on the information, parameters are optimized for simulating the state of the resin during molding and details of the properties of the molded product. As a result, the parameters for the simulation are optimized, so that the parameters for the simulation that can accurately estimate the characteristics of the resin molded product can be set more easily.

Further, in this embodiment, as described above, the resin molding conditions are classified based on a plurality of pieces of resin molding information, and the parameters of the resin molding information of the corresponding classification are used based on the input resin molding conditions. Simulate the state of the resin during molding and the details of the characteristics of the molded product. Accordingly, by classifying a plurality of accumulated resin molding information, the resin molding information to be used can be narrowed down, so that the parameters of the corresponding resin molding information can be easily set.

Further, in the present embodiment, as described above, the resin molding conditions are classified based on a plurality of pieces of resin molding information, and from the input resin molding conditions, based on the resin molding information of the corresponding classification, the resin molded product characteristics are estimated and output as characteristic information. As a result, it is possible to more accurately and quickly estimate the characteristics of the resin molded product based on the resin molding information of the corresponding classification.

In addition, in the present embodiment, as described above, the classification of the resin molding conditions is performed by quantifying at least one of the size of the molded product, the complexity of the molded product, the shape of the molded product, the resin data, and the molding conditions. . By quantifying at least one of the size of the molded product, the degree of complexity of the molded product, the shape of the molded product, the resin data, and the molding conditions, it is possible to easily classify the resin molding conditions for each similar condition. can be done.

Further, in the present embodiment, as described above, the characteristics of the molded product include at least one of the warpage of the molded product, the pressure during molding, the temperature during molding, the orientation of the molded product, and the weld state of the molded product. include. By inputting resin molding conditions, at least one of molded product warpage, pressure during molding, temperature during molding, orientation of molded product, and weld state of molded product can be obtained. can be obtained.

In addition, in the present embodiment, as described above, the characteristics of the molded product are the filling pattern of the resin, the appearance characteristics of the molded product, the resin physical property values of the molded product, the strain during molding, the mass of the molded product, the length of the molded product, part distortion, part stress, part mechanical properties, part thermal properties, part electrical properties, part chemical properties, part molding properties, part optics characteristics, and at least one of the color tone of the molded product. As a result, by inputting the resin molding conditions, the characteristics of the molded product, such as the resin filling pattern, the appearance characteristics of the molded product, the resin physical properties of the molded product, the strain during molding, the mass of the molded product, the length of the molded product, etc. part distortion, part stress, part mechanical properties, part thermal properties, part electrical properties, part chemical properties, part molding properties, part optics At least one of the physical characteristics and the color tone of the molded product can be acquired.

In addition, in the present embodiment, as described above, the output value based on the actual measurement value of the characteristics of the resin molded product is interpolated by performing machine learning on the actual measurement value of the characteristics and the actual measurement value of the characteristics. and interpolated values. As a result, even if the number of measured values is not sufficient, the interpolated value can be obtained by machine learning, so the number of output value data can be easily increased.

Further, in this embodiment, as described above, the physical property values of the same type of resin material in different states are predicted based on the known physical property values of the resin material. As a result, even if all the physical property values of the resin material required for the simulation are not known due to differences in the grade, composition, and manufacturing method of the resin material, the same type and state used for resin molding can be calculated based on the known physical property values of the resin material. Since the physical property values of different resin materials can be predicted and obtained, a simulation can be performed based on the obtained physical property values of the resin material. Moreover, since there is no need to actually measure the physical property values of the unknown resin material, there is no need to separately provide a measuring device for measuring the physical property values of the resin material.

In addition, in this embodiment, as described above, the physical property values of the same type of resin material in different states are predicted based on the known basic information of the resin material and the content of the recycled material contained in the resin material. This makes it possible to predict and obtain the physical property values of the resin material even when the resin material is mixed with the recycled material that has undergone molding and has received heat. A simulation for resin molding can be performed with high accuracy.

(Modification)
It should be noted that the embodiments disclosed this time should be considered as examples and not restrictive in all respects. The scope of the present invention is indicated by the scope of the claims rather than the description of the above-described embodiments, and includes all modifications (modifications) within the scope and meaning equivalent to the scope of the claims.

For example, in the above embodiment, the properties of the molded product include at least one of warpage of the molded product, pressure during molding, temperature during molding, orientation of the molded product, and weld state of the molded product. Although shown, the invention is not so limited. In the present invention, the properties of the molded article may include properties other than those described above.

In the above embodiment, classification of the resin molding conditions is performed by quantifying at least one of the size of the molded product, the complexity of the molded product, the shape of the molded product, the resin data, and the molding conditions. However, the present invention is not limited to this. In the present invention, the classification of resin molding conditions may be performed by quantifying factors other than the above, or may be performed without quantifying the above factors and other factors.

In the present invention, the plurality of pieces of resin molding information may be stored in a computer for estimating the characteristic information of the resin molded product, or may be stored in an external computer provided separately from the computer for estimating the characteristic information of the resin molded product. It may be stored in a device such as a server.

Further, in the above-described embodiment, for convenience of explanation, a flow-driven flowchart in which processing operations of a computer are performed in order along a processing flow has been described, but the present invention is not limited to this. In the present invention, the processing operation of the computer may be performed by event-driven processing that executes processing on an event-by-event basis. In this case, it may be completely event-driven, or a combination of event-driven and flow-driven.

1 computer 3a program 7 recording medium

Claims (13)

accumulating resin molding information in association with resin molding condition information and output values based on actual measurements of characteristics of the corresponding resin molded product;
estimating characteristics of a resin molded product from the input resin molding conditions based on a plurality of the accumulated resin molding information and outputting the characteristics as characteristic information ;
estimating parameters for simulating the state of the resin at the time of molding and the details of the characteristics of the molded product from the input resin molding conditions, based on the plurality of accumulated resin molding information;
A resin molding analysis method that performs simulation using estimated parameters . performing machine learning based on a plurality of pieces of resin molding information including resin molding condition information and measured values obtained by actually measuring characteristics of corresponding resin molded products;
2. The resin molding analysis method according to claim 1 , wherein parameters are optimized when performing a simulation of the state of the resin during molding and the details of the characteristics of the molded product based on the plurality of pieces of resin molding information. Classifying resin molding conditions based on a plurality of pieces of resin molding information,
3. The resin molding according to claim 2 , wherein, based on the input resin molding conditions, the parameters of the resin molding information of the corresponding classification are used to simulate the state of the resin at the time of molding and details of the characteristics of the molded product. analysis method. Classifying resin molding conditions based on a plurality of pieces of resin molding information,
The resin molding according to any one of claims 1 to 3 , wherein characteristics of the resin molded product are estimated based on the resin molding information of the corresponding classification from the input resin molding conditions and output as characteristic information. analysis method. 5. The resin molding analysis according to claim 3 or 4 , wherein the classification of the resin molding conditions is performed by quantifying at least one of the size of the molded product, the complexity of the molded product, the shape of the molded product, the resin data, and the molding conditions. Method. The properties of the molded product include at least one of warpage of the molded product, pressure during molding, temperature during molding, orientation of the molded product, and weld state of the molded product. The resin molding analysis method described in . The characteristics of the molded product are the filling pattern of the resin, the appearance characteristics of the molded product, the resin physical properties of the molded product, the distortion during molding, the mass of the molded product, the length of the molded product, the distortion of the molded product, the stress of the molded product, At least one of the mechanical properties of the molded article, the thermal properties of the molded article, the electrical properties of the molded article, the chemical properties of the molded article, the molding properties of the molded article, the optical properties of the molded article, and the color tone of the molded article. The resin molding analysis method according to any one of claims 1 to 6 , comprising accumulating resin molding information in association with resin molding condition information and output values based on actual measurements of characteristics of corresponding resin molded products;
estimating characteristics of a resin molded product from the input resin molding conditions based on a plurality of the accumulated resin molding information and outputting the characteristics as characteristic information;
A resin molding analysis method in which the output value based on the actual measured value of the characteristics of the resin molded product includes the actual measured value of the characteristic and the interpolated value obtained by interpolating the measured value of the characteristic by performing machine learning. Claims 1 to 7 , wherein the output value based on the measured values of the characteristics of the resin molded product includes the measured values of the characteristics and the interpolated values obtained by interpolating the measured values of the characteristics by performing machine learning. The resin molding analysis method according to any one of 1. The resin molding analysis method according to any one of claims 1 to 9, wherein physical property values of the same kind of resin material in different states are predicted based on known physical property values of the resin material. 11. The resin molding analysis method according to claim 10, wherein physical property values of the same kind of resin material in different states are predicted based on known basic information of the resin material and the content of recycled material contained in the resin material. A program that causes a computer to execute the resin molding analysis method according to any one of claims 1 to 11. A computer-readable recording medium on which the program according to claim 12 is recorded.

Images