Mold manufacturing assistance system, mold manufacturing assistance method, and program

Abstract

A mold manufacturing assistance system 100 includes: an assistance device 1 having an estimation unit 12 for estimating, by a trained model, target mold shape data, as output data, that represents the target shape of a mold from input data related to at least one of a mold and a molded article; and a GUI terminal 2 for performing input / output in connection with the assistance device 1. The GUI terminal 2 has a display unit 22 for displaying a data input screen for selecting input data and a data output screen for displaying output data, with an estimation button 54 for executing estimation by an estimation unit 12 being displayed on the data input screen.

Description

Mold manufacturing support system, mold manufacturing support method, and program

[0001] The present disclosure relates to a mold manufacturing support system, a mold manufacturing support method, and a program.

[0002] Patent Document 1 discloses a technique for obtaining expected shape data of a mold by so-called CAE (Computer Aided Engineering).

[0003] JP 2012-119010 A

[0004] In the conventional techniques described above, the number of times the die shape needs to be revised can increase due to insufficient forecast accuracy, etc. In particular, in recent years, in addition to problems of cracking and wrinkles due to the increased strength of materials, it has become difficult to satisfy the dimensional accuracy of the molded product due to springback after molding, and the number of times the die shape needs to be revised tends to increase. In addition, there has been no detailed consideration of the GUI (Graphical User Interface), leaving room for improvement.

[0005] An object of the present disclosure is to reduce the number of times a mold shape is modified and to provide a suitable GUI in a mold production support system, a mold production support method, and a program.

[0006] A first aspect of the present disclosure provides a mold production support system comprising: an assistance device having an estimation unit that estimates mold target shape data representing a target shape of the mold as output data from input data related to at least one of the mold and a molded product using a trained model; and a GUI terminal that performs input and output in relation to the assistance device, wherein the GUI terminal has a display unit that displays a data input screen for selecting the input data and a data output screen for displaying the output data, and the data input screen displays an estimation button for executing estimation by the estimation unit.

[0007] According to this configuration, the mold target shape data can be estimated with high accuracy using the trained model of the support device. Furthermore, the data input screen, data output screen, and estimation button of the GUI terminal enable intuitive input operations and confirmation of output results. Therefore, it is possible to reduce the number of times the mold shape needs to be corrected and provide a suitable user interface. The support device and the GUI terminal may be separate or integrated.

[0008] The support device may have an acquisition unit that acquires mold actual shape data that represents the actual shape of the mold, molded product actual shape data that represents the actual shape of a molded product molded by the mold, and molded product target shape data that represents the target shape of the molded product, and molded product difference data that represents the difference between these data and molded product target shape data that represents the target shape of the molded product, and the trained model may have trained by machine learning the relationship between the mold actual shape data and molded product difference data as the input data and the mold target shape data as the output data.

[0009] According to this configuration, by using actual mold shape data and molded product difference data as input data, it is possible to specifically estimate target mold shape data with high accuracy.

[0010] The support device may have an acquisition unit that acquires molded product actual shape data that represents the actual shape of a molded product molded using the actual shape of the mold, molded product transformed shape data that represents the transformed shape of the molded product converted from the mold actual shape data that represents the actual shape of the mold by mold-molded product conversion using simulation, and molded product target shape data that represents the target shape of the molded product, and the trained model may have trained by machine learning the relationship between the molded product target shape data as the input data plus the difference between the molded product transformed shape data and the molded product actual shape data, and the mold target shape data as the output data.

[0011] According to this configuration, by using the molded product target shape data plus errors in the molded product transformed shape data and the molded product actual shape data as input data, the mold target shape data can be estimated specifically and with high accuracy. In particular, the input data can be limited to data related to the molded product. In other words, data related to the mold can be omitted from the input data.

[0012] The GUI terminal may have a difference reflection level setting unit that accepts the setting of a difference reflection level that changes the magnitude of the difference, the difference reflection level setting unit may display a difference reflection level setting field for setting the difference reflection level on the data input screen, and the estimation unit may correct the input data based on the difference reflection level to estimate the mold target shape data.

[0013] According to this configuration, by changing the difference reflection level, it is possible to select a mold shape by trial and error while comparing the mold shapes.

[0014] The GUI terminal may have an AI reflection level setting unit that accepts the setting of an AI reflection level that adjusts the amount of correction of the mold target shape data from the mold actual shape data, and a correction execution unit that corrects and displays the mold target shape data at the AI ​​reflection level, the AI ​​reflection level setting unit may display an AI reflection level setting field for setting the AI ​​reflection level on the data output screen, and a mold correction button for executing correction by the correction execution unit may be displayed on the data output screen.

[0015] According to this configuration, by changing the AI ​​reflection level, it is possible to select a mold shape by trial and error while comparing them.

[0016] The GUI terminal may have a smoothing level setting unit that accepts the setting of a smoothing level for the mold target shape data, and a smoothing execution unit that smooths the mold target shape data at the smoothing level and displays it, the smoothing level setting unit may display a smoothing level setting field for setting the smoothing level on the data output screen, and a smoothing button for executing smoothing by the smoothing execution unit may be displayed on the data output screen.

[0017] According to this configuration, the smoothness of the die target shape data as output data can be adjusted.

[0018] The GUI terminal may have an output target setting unit that accepts settings for an output target, an output data format setting unit that accepts settings for an output data format for the output target, and a save execution unit that saves the output target in the output data format, wherein the output target setting unit may display an output target setting field on the data output screen for setting the output target, and the output data format setting unit may display an output data format setting field on the data output screen for accepting settings for an output data format for the output target, and a save button that executes file saving by the save execution unit may be displayed on the data output screen.

[0019] This configuration allows necessary files to be saved in a desired format.

[0020] The estimation unit may estimate new die target shape data using the die target shape data as new die actual shape data.

[0021] This configuration makes it possible to repeatedly estimate the die target shape data.

[0022] The die manufacturing support system may include a CAM unit that converts the die target shape data into NC data, and a machine tool that corrects the actual shape of the die based on the NC data.

[0023] According to this configuration, the actual shape of the mold can be corrected based on the mold target shape data estimated by the trained model.

[0024] The die production support system may include a shape measuring instrument that measures the actual shape of a molded product formed by the die corrected by the machine tool.

[0025] According to this configuration, it is possible to measure the actual shape of the molded product molded by the corrected mold.

[0026] The support device may re-train the trained model using the mold target shape data estimated by the trained model and molded product actual shape data representing the actual shape of the molded product measured by the shape measuring instrument.

[0027] This configuration makes it possible to improve the estimation accuracy of the trained model.

[0028] A second aspect of the present disclosure provides a mold production support method, which includes: estimating output data including mold target shape data representing a target shape of a mold from input data related to at least one of the mold and a molded product using a trained model; displaying a data input screen for selecting the input data and a data output screen for displaying the output data; and displaying an estimation button on the data input screen for performing the estimation.

[0029] A third aspect of the present disclosure provides a program for a computer to execute the following steps: estimate output data including mold target shape data representing a target shape of a mold from input data related to at least one of the mold and a molded product using a trained model; display a data input screen for selecting the input data and a data output screen for displaying the output data; and display an estimation button on the data input screen for performing the estimation.

[0030] According to the present disclosure, in a mold manufacturing support system, a mold manufacturing support method, and a program, it is possible to reduce the number of times mold shape modifications are made and to provide a suitable GUI.

[0031] A system diagram showing an example of a mold production support system. A screen diagram showing an example of a data input screen. A screen diagram showing another example of the data input screen. A screen diagram showing an example of a data output screen. A screen diagram showing another example of the data output screen. A conceptual diagram explaining an example of input / output data in a first algorithm. A block diagram showing an example of a support device and a GUI terminal. A flowchart showing an example of a mold production support method in a first algorithm. A conceptual diagram explaining an example of input / output data in a second algorithm. A flowchart showing an example of a mold production support method in a second algorithm.

[0032] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings.

[0033] The functions of the elements disclosed herein can be performed using circuits or processing circuits, including general-purpose processors, special-purpose processors, integrated circuits, application-specific integrated circuits (ASICs), conventional circuits, and / or combinations thereof, configured or programmed to perform the disclosed functions. A processor is considered a processing circuit or circuit because it includes transistors and other circuitry. In this disclosure, a circuit, unit, or means may be hardware that performs the recited functions or hardware that is programmed to perform the recited functions. The hardware may be hardware disclosed herein or other known hardware that is programmed or configured to perform the recited functions. Where hardware is a processor, which is considered a type of circuit, the circuit, means, or unit is a combination of hardware and software, and the software is used to configure the hardware and / or processor.

[0034] 1 is a diagram showing an example of the configuration of a die production support system 100 according to this embodiment. The die production support system 100 is a system for supporting the production of dies used in press molding.

[0035] The mold production support system 100 of this embodiment has a support device 1, a GUI (Graphical User Interface) terminal 2, a CAM / machine tool 3, a press 4, a shape measuring instrument 5, and a molding analysis DB (Data Base) 6.

[0036] The support device 1 has a computer including a CPU, RAM, ROM, non-volatile memory, an input / output interface, etc. The GUI terminal 2 is a terminal that provides a GUI to the user U and performs input / output related to the support device 1, and has a computer like the support device 1.

[0037] The CPU executes information processing in accordance with a program loaded from the ROM or non-volatile memory into the RAM. The program may be supplied via an information storage medium or a communication network.

[0038] In this embodiment, the support device 1 and the GUI terminal 2 are in a server-client relationship. However, the support device 1 and the GUI terminal 2 may be integrated. In other words, the functions of the support device 1 and the GUI terminal 2 may be configured as a single device.

[0039] The CAM / machine tool 3 has a CAM section that generates NC data and a machine tool that manufactures a mold based on the NC data.

[0040] The press 4 produces a molded product by press molding using a mold.

[0041] The shape measuring instrument 5 measures the shape of the produced molded product. The shape measuring instrument 5 is, for example, a 3D scanner or a contact or non-contact displacement sensor.

[0042] The molding analysis DB 6 is a database that stores data on molds and molded products that have been manufactured in the past. The mold data and molded product data include not only actual shape data obtained by testing and measurement, but also data obtained by CAE.

[0043] The operation and display of the GUI terminal 2 will now be described.

[0044] 2 and 3 are screen diagrams showing two examples of data input screens on the GUI terminal 2. FIG.

[0045] On the data input screen, the topmost tab 40 is labeled "Input." Below the tab 40, there are a molded product target shape data input field 41, a molded product actual shape data input field 42, a molded product transformed shape data input field 43, and a mold actual shape data input field 44, each of which accepts input of corresponding data from the user. The data corresponding to these input fields 41-44 (molded product target shape data, molded product actual shape data, molded product transformed shape data, and mold actual shape data) will be described in detail below. Browse buttons 45-48 are displayed next to these input fields 41-44, and files can be selected by pressing these buttons 45-48. Alternatively, data may be input by dragging and dropping files corresponding to these input fields 41-44.

[0046] A display area 49 for displaying data is provided below the input fields 41 to 44. In the illustrated example, the display area 49 displays a molded product P that is stepped and curved rather than linear in the longitudinal direction.

[0047] The display area 49 may be divided into left and right halves as shown in Fig. 2, or may be undivided as shown in Fig. 3. The display area 49 is provided with check boxes 50 for selecting the data to be displayed. Depending on the selection of the check boxes 50, one or two of the molded product target shape data, molded product actual shape data, molded product transformed shape data, mold actual shape data, and comparison contour data are displayed. The comparison contour data, as shown in Fig. 3, is a contour display of the difference between two pieces of data. In the illustrated example, the magnitude of the difference between the molded product target shape data and the molded product actual shape data is displayed as a contour.

[0048] A difference reflection level setting field 51 for setting the difference reflection level on the data input screen is displayed in the upper right corner of the display area 49. The difference reflection level setting field 51 includes a slide bar 52 and a window 53 arranged side by side above and below. By moving the slide bar 52, the difference reflection level can be changed in increments of 0.1, for example, from 0.1 to 2.0. The numerical value of the difference reflection level changed by the slide bar 52 is displayed in the window 53. The difference reflection level will be described in detail later.

[0049] An estimation button 54 for executing estimation by the estimation unit 12 (described later) on the data input screen is provided below the difference reflection level setting field 51. By pressing the estimation button 54, one of two algorithms is executed for estimation (described later), and the display can be switched from the data input screen to the data output screen.

[0050] 4 and 5 are screen diagrams showing two examples of data output screens on the GUI terminal 2. FIG.

[0051] On the data output screen, the top tab 40 is displayed as "Output." Below the tab 40, a save folder input field 55, a file name input field 56, and a mold actual shape input field 57 are provided, each of which accepts input of corresponding data (save destination folder, file name, mold actual shape data) from the user. Browse buttons 58, 59 are displayed next to these input fields 55, 57, and files can be selected by pressing the browse buttons 58, 59. Alternatively, data can be input by dragging and dropping files into these input fields 58, 59, respectively. A file name can be input into the file name input field 56 via the reception unit 21, which will be described later.

[0052] A display area 60 for visually checking data is provided below the input fields 55 to 57. In the illustrated example, a stepped mold M that is curved rather than linear in the longitudinal direction is displayed in the display area 60. The mold M and molded product P (FIGS. 2 and 3) have complementary shapes.

[0053] The display area 60 may be divided as shown in FIG. 4 or may be undivided as shown in FIG. 5. The display area 60 is provided with check boxes 61 for selecting the data to be displayed. Depending on the selection of the check boxes 61, one or two of the molded product target shape data, molded product actual shape data, mold actual shape data, mold target shape data, corrected mold target shape data, mold target shape data after smoothing, and comparison contour data are displayed. As shown in FIG. 4, the comparison contour data is a contour display of the difference between two pieces of data. In the illustrated example, the magnitude of the difference between the mold target shape data and the actual mold shape data is displayed as a contour. Furthermore, in FIG. 5, the mold target shape data and the mold target shape data after smoothing are displayed overlapping each other.

[0054] At the top right of the display area 60, there are displayed an AI reflection level setting field 62 for setting the AI ​​reflection level on the data output screen, and a mold correction button 63 for re-executing estimation by the estimation unit 12 (described later) at the AI ​​reflection level set in the AI ​​reflection level setting field 62. The AI ​​reflection level setting field 62 includes a slide bar 64 and a window 65 arranged vertically. By moving the slide bar 64, the AI ​​reflection level can be changed in increments of 0.1, for example, from 0.1 to 1.0. The numerical value of the AI ​​reflection level changed by the slide bar 64 is displayed in the window 65. The AI ​​reflection level will be described in detail later.

[0055] Below the AI ​​reflection level setting field 62, there is provided a mold correction button 63 for correcting the mold target shape data at the AI ​​reflection level set in the AI ​​reflection level setting field 62 on the data output screen.

[0056] Also displayed in the upper right corner of the display area 60 are a smoothing level setting field 66 for setting the smoothing level on the data output screen, and a smoothing button 67 for executing smoothing at the smoothing level set in the smoothing level setting field 66. The smoothing level setting field 66 includes a slide bar 68 and a window 69 arranged vertically. By moving the slide bar 68, the smoothing level can be changed in increments of 1, for example, from 1 to 10. The numerical value of the smoothing level changed by the slide bar 68 is displayed in the window 69. The smoothing level will be described in detail later.

[0057] Also displayed in the upper right corner of the display area 60 are an output target setting field 70 for setting the output target on the data output screen, an output data format setting field 71 for accepting the setting of the output data format for the output target, and a save button 72 for saving the file in the output target set in the output target setting field 70 and the output data format set in the output data format setting field 71. The output target setting field 70 allows the user to select one of molded product target shape data, molded product actual shape data, mold actual shape data, mold target shape data, corrected mold target shape data, mold target shape data after smoothing, and comparison contour data. The output data format setting field 71 allows the user to select one of a predetermined extension list. The extension list is a list of extensions related to shape representation data such as IGS files and STL files.

[0058] In this embodiment, two algorithms can be used for the estimation: the first algorithm treats both the mold and molded product data as input data for the trained model, and the second algorithm treats only the molded product data as input data for the trained model.

[0059] First, steps (a) to (j) shown in FIG. 1 will be described for the first algorithm.

[0060] (a) A user U inputs mold actual shape data representing the actual shape of a mold, molded product actual shape data representing the actual shape of a molded product molded by the mold, and molded product target shape data representing the target shape of the molded product into the GUI terminal 2. These data are, for example, IGS files or STL files.

[0061] (b) Based on the above data, the GUI terminal 2 generates data to be input to the trained model and outputs it to the support device 1. Specifically, the GUI terminal 2 calculates molded product difference data that represents the difference between the actual shape and target shape of the molded product. The actual mold shape data and the molded product difference data are then used as input data.

[0062] (c) The support device 1 uses the trained model to estimate mold target shape data representing the target shape of the mold from the input data (mold actual shape data and molded product difference data) acquired from the GUI terminal 2 (see FIG. 6 ), and outputs the data to the GUI terminal 2. Note that, as will be described later, the input data may be corrected before being used for estimation.

[0063] The trained model has been trained by machine learning to understand the relationship between the actual mold shape data and molded product difference data as input data and the target mold shape data as output data. The trained model is, for example, a regression model such as a neural network or Gaussian process.

[0064] (d) The GUI terminal 2 displays the die target shape data output from the support device 1. The user U checks the die target shape data displayed on the GUI terminal 2 and determines whether to accept it. (e) When the GUI terminal 2 receives the die target shape determination by the user, (f) it outputs the die target shape data.

[0065] (g) A user U inputs die target shape data into the CAM / machine tool 3. (h) The CAM / machine tool 3 converts the die target shape data into NC data and modifies the actual shape of the die based on the converted NC data. (i) The press 4 forms a molded product using the modified die.

[0066] (j) The shape measuring instrument 5 measures the actual shape of the molded product molded using the modified mold. Thereafter, the mold target shape data is used as new actual mold shape data, and modified difference data representing the difference between the measured actual shape of the molded product and the target shape is used as new molded product difference data, and steps (a) to (j) are performed again. Note that data obtained by measuring the actual mold may be used as the actual mold shape data. Furthermore, once the shape measuring instrument 5 measures the actual shape of the molded product, the actual mold shape data and molded product shape data may be registered in the molding analysis DB 6, and the trained model may be retrained. This makes it possible to improve estimation accuracy.

[0067] The above-described steps (a) to (j) are repeated until the actual shape and the target shape of the molded product match.

[0068] 7 is a block diagram showing an example of the configuration of the support device 1 and the GUI terminal 2. The support device 1 has an acquisition unit 11 and an estimation unit 12. These functional units 11 and 12 are realized by the CPU of the support device 1 executing information processing in accordance with a program loaded from the ROM or non-volatile memory to the RAM.

[0069] The GUI terminal 2 has a reception unit 21 and a display unit 22. The reception unit 21 is, for example, a keyboard, a mouse, or a touch panel, and receives operations from the user U. The display unit 22 is, for example, a liquid crystal display, an organic EL display, or a plasma display.

[0070] The GUI terminal 2 also has a subtraction unit 23, a conversion calculation unit 24, a difference reflection level setting unit 25, an AI reflection level setting unit 26, a correction execution unit 27, a smoothing level setting unit 28, a smoothing execution unit 29, an output target setting unit 30, an output data format setting unit 31, and a storage execution unit 32. These functional units 23 to 32 are realized by the CPU of the GUI terminal 2 executing information processing in accordance with a program loaded from the ROM or non-volatile memory to the RAM.

[0071] The subtraction unit 23 calculates the difference between the actual shape data of the molded product and the target shape data of the molded product as molded product difference data (see FIG. 6). The subtraction unit 23 is used only in the first algorithm. In contrast, the conversion calculation unit 24 is used only in the second algorithm, which will be described later. The other functional units 25 to 32 are used in common by both the first and second algorithms.

[0072] The difference reflection level setting unit 25 accepts the setting of a difference reflection level that changes the magnitude of the difference. The difference reflection level setting unit 25 displays a difference reflection level setting field 51 for setting the difference reflection level on the data input screen (see FIGS. 2 and 3). The estimation unit 12 then corrects the input data based on the difference reflection level as follows. Specifically, the difference between the molded product target shape data and the molded product actual shape data is multiplied by the difference reflection level as a coefficient. In the example of FIGS. 2 and 3, the difference reflection level is, for example, a value between 0.1 and 2.0. The estimation unit 12 then estimates the mold target shape data.

[0073] The AI ​​reflection level setting unit 26 accepts the setting of an AI reflection level that adjusts the amount of correction of the mold target shape data from the mold actual shape data. The AI ​​reflection level setting unit 26 displays an AI reflection level setting field 62 for setting the AI ​​reflection level on the data output screen (see FIGS. 4 and 5). In the examples of FIGS. 4 and 5, the AI ​​reflection level is a value between 0.1 and 1.0.

[0074] The correction execution unit 27 accepts re-execution of estimation by the estimation unit 12 at the AI ​​reflection level set by the AI ​​reflection level setting unit 26. A mold correction button 63 for executing correction by the correction execution unit 27 is displayed on the data output screen (see FIGS. 4 and 5).

[0075] The smoothing level setting unit 28 accepts the setting of the smoothing level of the mold target shape data. The smoothing level setting unit 28 displays a smoothing level setting field 66 for setting the smoothing level on the data output screen (see FIGS. 4 and 5). In the example of FIGS. 4 and 5, the smoothing level is a value from 1 to 10.

[0076] The smoothing execution unit 29 smooths and displays the mold target shape data at the smoothing level set in the smoothing level setting field 66. A smoothing button 67 for executing smoothing by the smoothing execution unit 29 is displayed on the data output screen (see FIGS. 4 and 5).

[0077] The output target setting section 30 accepts the setting of the output target and displays an output target setting field 70 for setting the output target on the data output screen (see FIGS. 4 and 5).

[0078] The output data format setting unit 31 accepts the setting of the output data format of the output target. The output data format setting unit 31 displays an output data format setting field 71 on the data output screen for accepting the setting of the output data format of the output target (see FIGS. 4 and 5).

[0079] The save execution unit 32 saves the output target in the output data format. On the data output screen, a save button 72 is displayed to save the file in the output target set in the output target setting field 70 and the output data format set in the output data format setting field 71 (see Figures 4 and 5).

[0080] 8 is a flowchart showing an example of the procedure of the mold production support method of the first algorithm realized in the mold production support system 100. Each of the support device 1 and the GUI terminal 2 executes the information processing shown in the figure according to a program.

[0081] First, the GUI terminal 2 receives input of mold actual shape data, molded product actual shape data, and molded product target shape data from the user (step S8-1, corresponding to the above-mentioned step (a)). These data are represented as point cloud data such as an STL file. Note that, when there are a large number of measurement points, dimension reduction by feature extraction may be performed.

[0082] Next, the GUI terminal 2 calculates molded product difference data that indicates the difference between the actual shape and the target shape of the molded product (step S8-2, processing as the subtraction unit 23, corresponding to the above step (b)).

[0083] Next, the system accepts the user's input for setting the difference reflection level and pressing the "Estimate" button (step S8-3). The difference reflection level is a value used to adjust the magnitude of the difference, and in the examples shown in FIGS. 2 and 3, it is set in increments of 0.1 from 0.1 to 2.0. After setting the difference reflection level, the user presses the "Estimate" button. This causes the support device 1 to acquire the mold actual shape data and molded product difference data as input data from the GUI terminal 2 (step S8-4, processing as the acquisition unit 11). At this time, the difference reflection level is also sent to the support device 1 and is used to correct the input data as follows:

[0084] Next, the support device 1 corrects the input data by reflecting the difference reflection level (step S8-5). The difference reflection level is a coefficient multiplied by the difference between the actual shape and the target shape of the molded product. For example, if the difference reflection level is 1.0, the molded product difference data representing the difference between the actual shape and the target shape of the molded product is used as is. However, if the difference reflection level is less than 1.0, the difference is reduced by a corresponding ratio before use. If the difference reflection level is greater than 1.0, the difference is increased by a corresponding ratio before use.

[0085] The support device 1 uses the trained model to estimate output data (mold target shape data) from input data (mold actual shape data and molded product difference data) (step S8-6, processing as the estimation unit 12, corresponding to the above step (c)).

[0086] The GUI terminal 2 displays the mold target shape data estimated by the trained model in the support device 1 on the display unit 22 (step S8-7, corresponding to the above process (d)).

[0087] Next, the GUI terminal 2 receives from the user the setting of the AI ​​reflection level and the pressing of the mold correction button (step S8-8). In the examples of FIGS. 4 and 5 described above, the AI ​​reflection level is a value ranging from 0.1 to 1.0, set in increments of 0.1. After setting the AI ​​reflection level, the user presses the mold correction button. This causes the GUI terminal 2 to correct the mold target shape data based on the AI ​​reflection level. Specifically, the AI ​​reflection level is a coefficient for adjusting the amount of correction from the mold actual shape data to the mold target shape data estimated by the estimation unit 12. For example, when the AI ​​reflection level is 1.0, the mold target shape data estimated by the estimation unit 12 is used as is without correction. When the AI ​​reflection level is less than 1.0, the amount of correction is reduced by that ratio (approaching the mold actual shape data).

[0088] Next, the GUI terminal 2 accepts the setting of the smoothing level and the pressing of the smoothing button from the user (step S8-9). In the examples of FIGS. 4 and 5 described above, the smoothing level is a numerical value from 1 to 10, and is set in increments of 1. After setting the smoothing level, the user presses the smoothing button. This causes the GUI terminal 2 to smooth the mold target shape data based on the smoothing level. Specifically, the smoothing level is a coefficient for adjusting the smoothness of the mold target shape data estimated by the estimation unit 12. For example, when the smoothing level is 1, the mold target shape data is jagged, and the higher the smoothing level, the smoother it becomes.

[0089] Next, the GUI terminal 2 accepts the user's decision on the die target shape at the accepting unit 21 (step S8-10). The user decides whether further correction is necessary by looking at the die target shape data finally displayed on the display unit 22. When the decision is accepted (determined that no correction is necessary) (step S8-10: YES), the GUI terminal 2 accepts the selection of a save file and the pressing of a save button (step S8-11). Note that the die target shape data obtained in this way may be used as new die actual shape data to estimate new die target shape data.

[0090] When selecting a file to save, the output target and output data format are input. In the examples of FIGS. 4 and 5 described above, the output target and output data format can be selected from a list prepared in advance. Then, by pressing the Save button, the output target (mold target shape data, etc.) is output in the selected output data format (S8-11, corresponding to steps (e) and (f) above). If the decision is not accepted (S8-10: NO), the process returns to step S8-3, step S8-8, or step S8-9. This return process can be selected by the user at will.

[0091] As described above, the mold target shape data is used to modify the actual shape of the mold using the CAM / machine tool 3, and the shape of the molded product formed using the modified mold is measured by the shape measuring instrument 5 (corresponding to steps (g) to (j) above).

[0092] The GUI terminal 2 calculates corrected difference data that indicates the difference between the actual shape of the molded product measured by the shape measuring instrument 5 and the target shape (processing as the subtraction unit 23).

[0093] The support device 1 estimates new mold target shape data from the GUI terminal 2 by using the mold target shape data as new mold actual shape data and the corrected difference data as new molded product difference data (processing as the estimation unit 12).

[0094] Next, the second algorithm will be described with reference again to FIG.

[0095] Unlike the first algorithm, the second algorithm limits input data to data related to molded products. Accordingly, the above (b) and (c) differ from the first algorithm. In other words, (a) and (d) through (j) are the same as the first algorithm, and therefore will not be described here.

[0096] (b) The GUI terminal 2 converts the actual mold shape data into transformed molded product shape data that represents the transformed shape of the molded product through a mold-to-molded product conversion simulation. For example, the molded product shape converted from the mold shape is obtained through a simulation using FEM (finite element method), a type of CAE. Then, the input data is the product target shape data plus the difference between the transformed molded product shape data and the actual molded product shape data.

[0097] (c) Using the trained model, the support device 1 estimates mold target shape data representing the target shape of the mold from input data acquired from the GUI terminal 2 (molded product target shape data plus the difference between the molded product transformed shape data and the molded product actual shape data) (see FIG. 9 ), and outputs the data to the GUI terminal 2. Note that, as will be described later, the input data may be corrected before being used for estimation.

[0098] The trained model has been trained by machine learning to understand the relationship between the input data (the target shape data of the molded product plus the difference between the transformed shape data of the molded product and the actual shape data of the molded product) and the output data (the target shape data of the mold).The trained model is, for example, a regression model such as a neural network or a Gaussian process.

[0099] 10 is a flowchart showing an example of the procedure of the mold production support method of the second algorithm realized in the mold production support system 100. Each of the support device 1 and the GUI terminal 2 executes the information processing shown in the figure according to a program.

[0100] First, the GUI terminal 2 accepts input of mold actual shape data, molded product actual shape data, and molded product target shape data from the user (step S10-1, corresponding to step (a) above). These data are represented as point cloud data such as an STL file. If there are many measurement points, dimensionality reduction may be performed by feature extraction. Note that if mold-to-molded product conversion has been performed in advance, the user may directly input molded product conversion shape data rather than mold actual shape data. In this case, the data file may be dragged and dropped into the molded product conversion shape data input field 43 on the data input screen.

[0101] Next, the GUI terminal 2 converts the mold actual shape data into molded product converted shape data representing the converted shape of the molded product through mold-to-molded product conversion by simulation, and calculates, as input data, the molded product target shape data plus the difference between the molded product converted shape data and the molded product actual shape data (step S10-2, processing as the conversion calculation unit 24, corresponding to the above process (b)).

[0102] Next, the system accepts the user's input for setting the difference reflection level and pressing the "Estimate" button (step S10-3). The difference reflection level is a value used to adjust the magnitude of the difference, and in the examples shown in FIGS. 2 and 3, it is set in increments of 0.1 from 0.1 to 2.0. After the user sets the difference reflection level, the system presses the "Estimate" button. This causes the support device 1 to acquire, as input data, the molded product target shape data plus the difference between the molded product transformed shape data and the molded product actual shape data from the GUI terminal 2 (step S10-4, processing as the acquisition unit 11). At this time, the difference reflection level is also sent to the support device 1 and is used to correct the input data as follows:

[0103] Next, the support device 1 corrects the input data by reflecting the difference reflection level (step S8-5). The difference reflection level is a coefficient multiplied by the difference between the molded product transformed shape data and the molded product actual shape data. For example, if the difference reflection level is 1.0, the difference between the molded product transformed shape data and the molded product actual shape data is used as is. However, if the difference reflection level is less than 1.0, the difference is reduced by a corresponding ratio before use. If the difference reflection level is greater than 1.0, the difference is increased by a corresponding ratio before use.

[0104] The support device 1 uses the trained model to estimate output data (mold target shape data) from input data (molded product target shape data plus the difference between the molded product converted shape data and the molded product actual shape data) (step S10-6, processing as the estimation unit 12, corresponding to the above step (c)).

[0105] The subsequent processing from step S10-7 to step S10-11 is substantially the same as the processing from step S8-7 to step S8-11 in the first algorithm.

[0106] According to this embodiment, the following advantageous effects are achieved.

[0107] The mold target shape data can be estimated with high accuracy using the trained model of the support device 1. In addition, the data input screen, data output screen, and estimation button 54 of the GUI terminal 2 enable intuitive input operations and confirmation of output results. Therefore, it is possible to reduce the number of times the mold shape needs to be corrected and provide a suitable user interface.

[0108] In the first algorithm, the mold target shape data can be estimated specifically and with high accuracy by using the mold actual shape data and molded product difference data as input data.

[0109] The second algorithm uses input data obtained by adding errors from the transformed shape data and the actual shape data of the molded product to the target shape data of the molded product, thereby enabling specific and highly accurate estimation of the target shape data of the mold. In particular, the input data can be limited to data related to the molded product. In other words, data related to the mold can be omitted from the input data.

[0110] By changing the difference reflection level, it is possible to select a mold shape by trial and error while comparing them.

[0111] By changing the AI ​​reflection level, it is possible to select mold shapes by trial and error, comparing them.

[0112] By setting the smoothing level, the smoothness of the die target shape data as output data can be adjusted. Furthermore, the smoothing may be applied to data other than the die target shape data.

[0113] File saving settings allow you to save the files you need in the format you want.

[0114] Since new die target shape data is estimated using the die target shape data as new die actual shape data, it is possible to repeat the estimation of die target shape data.

[0115] The CAM / machine tool 3 can correct the actual shape of the mold based on the mold target shape data estimated by the learned model.

[0116] The shape measuring instrument 5 can measure the actual shape of the molded product formed by the corrected mold.

[0117] While specific embodiments of the present disclosure have been described above, the present disclosure is not limited to the above embodiments and can be modified and implemented within the scope of the present invention. For example, only the first algorithm may be implemented, or only the second algorithm may be implemented. Furthermore, the functional units 23 to 32 (see FIG. 7) and corresponding processes may be selectively omitted as necessary.

[0118] The present disclosure may include the following aspects. (Aspect 1) A mold production support system comprising: an assistance device having an estimation unit that estimates mold target shape data representing a target shape of the mold as output data from input data related to at least one of the mold and a molded product using a trained model; and a GUI terminal that performs input and output in connection with the assistance device, wherein the GUI terminal has a display unit that displays a data input screen for selecting the input data and a data output screen that displays the output data, and an estimation button for executing estimation by the estimation unit is displayed on the data input screen. (Aspect 2) The mold production support system according to Aspect 1, wherein the assistance device has an acquisition unit that acquires mold actual shape data representing an actual shape of the mold, molded product actual shape data representing an actual shape of the molded product molded by the mold, and molded product target shape data representing the difference between the molded product actual shape data and the molded product difference data representing the target shape of the molded product, and the trained model has learned through machine learning the relationship between the mold actual shape data and the molded product difference data as input data, and the mold target shape data as output data. (Aspect 3) The mold production support system according to Aspect 1, wherein the support device includes an acquisition unit that acquires molded product actual shape data representing the actual shape of the molded product molded using the actual shape of the mold, molded product transformed shape data representing the transformed shape of the molded product converted from the mold actual shape data representing the actual shape of the mold by mold-to-molded product conversion using simulation, and molded product target shape data representing the target shape of the molded product, and the trained model has trained by machine learning a relationship between the molded product target shape data as the input data plus a difference between the molded product transformed shape data and the molded product actual shape data, and the mold target shape data as the output data. (Aspect 4) The mold production support system according to Aspect 2 or 3, wherein the GUI terminal includes a difference reflection level setting unit that accepts setting of a difference reflection level that changes the magnitude of the difference, and the difference reflection level setting unit displays a difference reflection level setting field on the data input screen for setting the difference reflection level, and the estimation unit corrects the input data based on the difference reflection level to estimate the mold target shape data.(Aspect 5) The mold production support system according to any of Aspects 2 to 4, wherein the GUI terminal has an AI reflection level setting unit that accepts a setting of an AI reflection level that adjusts the amount of correction of the mold target shape data from the mold actual shape data, and a correction execution unit that corrects and displays the mold target shape data at the AI ​​reflection level, the AI ​​reflection level setting unit displays an AI reflection level setting field for setting the AI ​​reflection level on the data output screen, and the data output screen displays a mold correction button for executing correction by the correction execution unit. (Aspect 6) The mold production support system according to any of Aspects 2 to 5, wherein the GUI terminal has a smoothing level setting unit that accepts a setting of a smoothing level of the mold target shape data, and a smoothing execution unit that smooths and displays the mold target shape data at the smoothing level, the smoothing level setting unit displays a smoothing level setting field for setting the smoothing level on the data output screen, and the data output screen displays a smoothing button for executing smoothing by the smoothing execution unit. (Aspect 7) The mold production support system according to any one of Aspects 2 to 6, wherein the GUI terminal has an output target setting unit that accepts setting of an output target, an output data format setting unit that accepts setting of an output data format of the output target, and a save execution unit that saves the output target in the output data format, wherein the output target setting unit displays an output target setting field for setting the output target on the data output screen, wherein the output data format setting unit displays an output data format setting field on the data output screen that accepts setting of the output data format of the output target, and wherein a save button for saving a file by the save execution unit is displayed on the data output screen. (Aspect 8) The mold production support system according to any one of Aspects 2 to 7, wherein the estimation unit estimates new mold target shape data from the mold target shape data as new mold actual shape data.(Aspect 9) A mold production support system according to any one of Aspects 1 to 8, comprising a CAM unit that converts the mold target shape data into NC data, and a machine tool that modifies the actual shape of the mold based on the NC data. (Aspect 10) A mold production support system according to Aspect 9, comprising a shape measuring instrument that measures the actual shape of a molded product molded by the mold modified by the machine tool. (Aspect 11) A mold production support system according to Aspect 10, wherein the support device re-learns the trained model using the mold target shape data estimated by the trained model and molded product actual shape data that represents the actual shape of the molded product measured by the shape measuring instrument. (Aspect 12) A mold production support method, comprising: estimating output data including mold target shape data that represents the target shape of the mold from input data related to at least one of the mold and the molded product, using a trained model; displaying a data input screen for selecting the input data and a data output screen that displays the output data; and displaying an estimation button on the data input screen for executing the estimation. (Aspect 13) A program for executing on a computer the following: estimating output data including mold target shape data representing a target shape of a mold from input data related to at least one of the mold and a molded product using a trained model; displaying a data input screen for selecting the input data and a data output screen for displaying the output data; and displaying an estimation button on the data input screen for executing the estimation.

[0119] This application claims priority from Japanese Patent Application No. 2024-113516, filed July 16, 2024. Japanese Patent Application No. 2024-113516 is incorporated herein by reference.

[0120] REFERENCE SIGNS LIST 1 Support device 2 GUI terminal 3 CAM / machine tool 4 Press machine 5 Shape measuring instrument 6 Molding DB 11 Acquisition unit 12 Estimation unit 21 Reception unit 22 Display unit 23 Subtraction unit 24 Conversion calculation unit 25 Difference reflection level setting unit 26 AI reflection level setting unit 27 Correction execution unit 28 Smoothing level setting unit 29 Smoothing execution unit 30 Output target setting unit 31 Output data format setting unit 32 Save execution unit 100 Mold production support system 40 Tab 41 Molded product target shape data input field 42 Molded actual shape data input field 43 Molded product conversion shape data input field 44 Mold actual shape data input field 45-48 Reference buttons 49 Display area 50 Check box 51 Difference reflection level setting field 52 Slide bar 53 Window 54 Estimation button 55 Save folder input field 56 File name input field 57 Mold actual shape input field 58, 59 Browse button 60 Display area 61 Check box 62 AI reflection level setting field 63 Mold correction button 64 Slide bar 65 Window 66 Smoothing level setting field 67 Smoothing button 68 Slide bar 69 Window 70 Output target setting field 71 Output data format setting field 72 Save button 100 Mold production support system P Molded product M Mold

Claims

1. A mold production support system comprising: a support device having an estimation unit that estimates mold target shape data representing the target shape of the mold as output data from input data related to at least one of the mold and molded product using a trained model; and a GUI terminal that performs input and output in relation to the support device, wherein the GUI terminal has a display unit that displays a data input screen for selecting the input data and a data output screen for displaying the output data, and the data input screen displays an estimation button for executing estimation by the estimation unit.

2. The mold production support system described in claim 1, wherein the support device has an acquisition unit that acquires mold actual shape data representing the actual shape of the mold, molded product actual shape data representing the actual shape of the molded product molded by the mold, and molded product target shape data representing the target shape of the molded product, and the trained model has trained the relationship between the mold actual shape data and molded product difference data as the input data and the mold target shape data as the output data through machine learning.

3. The mold production support system of claim 1, wherein the support device has an acquisition unit that acquires molded product actual shape data representing the actual shape of the molded product molded using the actual shape of the mold, molded product transformed shape data representing the transformed shape of the molded product converted from the mold actual shape data representing the actual shape of the mold by mold-to-molded product conversion using simulation, and molded product target shape data representing the target shape of the molded product, and the trained model has trained by machine learning the relationship between the molded product target shape data as the input data plus the difference between the molded product transformed shape data and the molded product actual shape data, and the mold target shape data as the output data.

4. A mold production support system as described in claim 2 or 3, wherein the GUI terminal has a difference reflection level setting unit that accepts the setting of a difference reflection level that changes the magnitude of the difference, the difference reflection level setting unit displays a difference reflection level setting field on the data input screen for setting the difference reflection level, and the estimation unit corrects the input data based on the difference reflection level and estimates the mold target shape data.

5. The mold production support system according to claim 2 or 3, wherein the GUI terminal has an AI reflection level setting unit that receives the setting of an AI reflection level that adjusts the amount of correction of the mold target shape data from the mold actual shape data, and a correction execution unit that corrects and displays the mold target shape data at the AI ​​reflection level, the AI ​​reflection level setting unit displays an AI reflection level setting field for setting the AI ​​reflection level on the data output screen, and the data output screen displays a mold correction button for executing correction by the correction execution unit.

6. The mold production support system described in claim 2 or 3, wherein the GUI terminal has a smoothing level setting unit that accepts the setting of a smoothing level for the mold target shape data, and a smoothing execution unit that smooths and displays the mold target shape data at the smoothing level, the smoothing level setting unit displays a smoothing level setting field for setting the smoothing level on the data output screen, and the data output screen displays a smoothing button for executing smoothing by the smoothing execution unit.

7. The mold production support system described in claim 2 or 3, wherein the GUI terminal has an output target setting unit that accepts settings for an output target, an output data format setting unit that accepts settings for an output data format for the output target, and a save execution unit that saves the output target in the output data format, wherein the output target setting unit displays an output target setting field on the data output screen for setting the output target, wherein the output data format setting unit displays an output data format setting field on the data output screen for accepting settings for the output data format for the output target, and wherein a save button is displayed on the data output screen for executing file saving by the save execution unit.

8. A mold production support system as described in claim 2 or 3, wherein the estimation unit estimates new mold target shape data as new mold actual shape data.

9. A mold production support system according to claim 1, comprising a CAM unit that converts the mold target shape data into NC data, and a machine tool that corrects the actual shape of the mold based on the NC data.

10. A mold production support system according to claim 9, further comprising a shape measuring instrument for measuring the actual shape of a molded product formed by the mold corrected by the machine tool.

11. A mold production support system as described in claim 10, wherein the support device re-learns the trained model using the mold target shape data estimated by the trained model and molded product actual shape data representing the actual shape of the molded product measured by the shape measuring instrument.

12. A mold production support method comprising: estimating output data including mold target shape data representing the target shape of the mold from input data related to at least one of the mold and molded product using a trained model; displaying a data input screen for selecting the input data and a data output screen for displaying the output data; and displaying an estimation button on the data input screen for executing the estimation.

13. A program for executing on a computer the following steps: estimating output data including mold target shape data representing the target shape of a mold from input data related to at least one of a mold and a molded product using a trained model; displaying a data input screen for selecting the input data and a data output screen for displaying the output data; and displaying an estimation button on the data input screen for executing the estimation.

Images