Information processing method, information processing apparatus, and program

Abstract

To accurately predict a shape of a powder compact.SOLUTION: In an information processing method, an information processing apparatus receives a shape of a first molded object after sintering (S201), predicts the shape of the first molded object before sintering based on the shape of the first molded object after sintering and prediction conditions stored in a storage part (S203), and outputs information related to the predicted shape of the first molded object before sintering. The prediction conditions are information based on a temperature threshold at which the viscosity of the first molded object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity.SELECTED DRAWING: Figure 14

Description

[Technical Field]

[0001] The present invention relates to an information processing method, an information processing device, and a program. [Background technology]

[0002] Conventionally, techniques for predicting the shape of a powder compact have been known. For example, Patent Document 1 discloses a technique for creating a model by machine learning a training data set and predicting the shape of a three-dimensional powder compact using the model during molding. Summary of the Invention [Problem to be solved by the invention]

[0003] However, with conventional techniques, it has been difficult to predict the shape of a powder compact with high accuracy.

[0004] In view of the above problems, the present invention provides a technique for predicting the shape of a powder compact with high accuracy. [Means for solving the problem]

[0005] In view of the above-mentioned problems, the present invention provides an information processing method in which an information processing device receives a shape of a first object after sintering, predicts a shape of the first object before sintering based on the shape of the first object after sintering and prediction conditions stored in a memory unit, and outputs information related to the predicted shape of the first object before sintering, wherein the prediction conditions are information based on a temperature threshold value that is a temperature at which the viscosity of the first object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity.

[0006] The present invention also provides an information processing method in which an information processing device receives a shape of a first object before sintering, predicts a shape of the first object after sintering based on the shape of the first object before sintering and prediction conditions stored in a memory unit, and outputs information related to the predicted shape of the first object after sintering, wherein the prediction conditions are information based on a temperature threshold value that is a temperature at which the viscosity of the first object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity. [Effects of the Invention]

[0007] The present invention can predict the shape of a powder compact with high accuracy. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a system configuration diagram of a modeling apparatus utilization system according to an embodiment of the present invention; [Figure 2] 1 is a system configuration diagram of a molding system according to an embodiment of the present invention. [Figure 3] 1 is a configuration diagram of a molding apparatus according to an embodiment of the present invention. [Figure 4] FIG. 1 is a flowchart illustrating a modeling method according to an embodiment of the present invention. [Figure 5] FIG. 2 is a hardware configuration diagram of an information processing device and a cloud server according to an embodiment of the present invention. [Figure 6] 1 is a functional block diagram of an information processing apparatus according to an embodiment of the present invention; [Figure 7] FIG. 2 is a diagram showing a temperature / viscosity curve according to one embodiment of the present invention. [Figure 8] FIG. 2 is a diagram showing a temperature / viscosity curve according to one embodiment of the present invention. [Figure 9] 10 is a flowchart showing processing performed by an information processing device 2 according to an embodiment of the present invention. [Figure 10] FIG. 2 is a diagram showing a model shape of a powder compact according to an embodiment of the present invention. [Figure 11]FIG. 10 is a diagram showing an example of correspondence data (first viscosity, second viscosity, temperature threshold, and shape error of evaluation point) recorded by a shape error evaluation unit. [Figure 12] FIG. 2 is a functional block diagram of a regression model construction unit using machine learning according to an embodiment of the present invention. [Figure 13] FIG. 10 is a cross-sectional view showing the results of a sintering simulation according to one embodiment of the present invention. [Figure 14] FIG. 10 is a flowchart illustrating an example of the flow of a simulation using the optimal parameters (temperature threshold, first viscosity, second viscosity) determined by the process of FIG. 9. [Figure 15] FIG. 10 is a flowchart illustrating an example of the flow of a simulation by inverse analysis using the optimal parameters (temperature threshold, first viscosity, second viscosity) determined by the process of FIG. 9. DETAILED DESCRIPTION OF THE INVENTION

[0009] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An information processing device and an information processing method performed by the information processing device will be described below as an example of an embodiment of the present invention with reference to the accompanying drawings. <Finite element method> The finite element method (FEM) is a method for solving the overall equation by dividing an object to be analyzed into analytical elements made up of minute elements, solving the equations for each analytical element, and connecting the elements. FEM uses governing equations as equations. In this embodiment, a viscous constitutive equation is incorporated into the governing equation. The viscous constitutive equation represents the relationship between the stress and strain rate of a material. In this embodiment, an information processing device predicts the shape of a powder compact or a final compact based on the finite element method.

[0010] The shape of the powder compact is expressed by the dimensions and angles of predetermined portions, etc. The shape of the powder compact may be expressed by the difference in dimensions and angles from the model shape.

[0011] <About binder jetting> Binder jetting is a molding technique that combines powder with a binder as a molding liquid. When metal powder is used as the powder, it is called Metal Binder Jetting (MBJ).

[0012] In MBJ, a molding machine first creates a powder compact (green body) from metal powder and molding liquid. A sintering machine then sinters the powder compact to create the final compact. The sintered powder compact is called the final compact. MBJ is known for its powder compacts, which can deform during sintering. Therefore, post-processing such as cutting or polishing is often required. Furthermore, designers often must identify the deformation trends for each material or final compact shape and repeatedly create prototypes. For this reason, designers must consider corrective shapes that take deformation into account, and consider the layout within the sintering machine and setters (jigs to prevent deformation during sintering). These considerations increase costs. Furthermore, the process requires expertise, creating a situation where the process must rely on the experience of dedicated engineers.

[0013] For this reason, simulation of sintering in sintering equipment using the finite element method (hereinafter referred to as sintering simulation) is becoming increasingly important at MBJ. Sintering simulation makes it possible to predict shape deformation that occurs during sintering from the shape at sintering and model shape. Sintering simulation also makes it possible to determine the corrected shape that minimizes deformation. The finite element method minimizes trial and error, leading to cost reduction.

[0014] Sintering methods for powder compacts that use metal powders can be divided into "solid phase sintering" and "liquid phase sintering." Solid phase sintering occurs when metal powders with high melting points are used, and sintering proceeds while the material remains solid. Solid phase sintering results in little deformation, and a theoretical model that can mathematically express the phenomenon has been established. For this reason, sintering simulations can be performed using the finite element method with a viscous constitutive equation.

[0015] Liquid phase sintering is applied to metal powders with low melting points, such as pure aluminum and aluminum alloys. Liquid phase sintering occurs when a metal powder is mixed with a material that has a lower melting point than the sintering temperature of the metal powder, and some of the material forms a liquid phase during sintering. However, although liquid phase sintering causes large deformation, the detailed mechanism is not understood. Furthermore, because the physical model used in liquid phase sintering has not been established, it is not possible to predict the shape of the final compact with high accuracy.

[0016] In the case of solid-phase sintering, the viscosity in the viscous constitutive equation is theoretically derived from the Arrhenius equation. However, if the viscosity derived from the Arrhenius equation is directly applied to liquid-phase sintering, the accuracy of shape prediction is poor. Therefore, in this embodiment, the viscosity introduced into the governing equation is derived based on a temperature / viscosity curve in which the viscosity decreases from a first viscosity to a second viscosity when a predetermined temperature threshold A is reached.

[0017] Methods for optimizing parameters include techniques using sensitivity analysis and experimental design. Sensitivity analysis involves setting a central value for each parameter based on some criterion, setting multiple levels above and below that central value, evaluating the sensitivity to those values, and selecting the best level value for each. Experimental design refers to a systematic method used when conducting experiments to clarify the cause-and-effect relationship between input and output. In this embodiment, experimental design is preferably used. Using experimental design makes it possible to efficiently optimize parameters even when there are a large number of parameters.

[0018] Furthermore, in this embodiment, machine learning may also be used. In this embodiment, the shape of the final molded body can be predicted with high accuracy by combining the temperature / viscosity curve, experimental design, and machine learning.

[0019] <Terminology> The term "modeled object" refers to an object for which prediction is performed in this embodiment, and is, for example, an object that is modeled by MBJ. In this embodiment, the shape of a modeled object after sintering (e.g., a final compact) is predicted from the shape of a modeled object before sintering (e.g., a powder compact). Alternatively, the shape of a modeled object before sintering (e.g., a powder compact) is predicted from the shape of a modeled object after sintering (e.g., a final compact).

[0020] <System configuration example> 1 is a system configuration diagram of a modeling-apparatus-using system 9 according to one embodiment of the present invention. The modeling-apparatus-using system 9 includes a modeling system 1, an information processing device 2, and a cloud server 3. Note that the information processing device 2 does not need to be connected to the modeling system 1 at all times, and may not be included in the modeling-apparatus-using system 9. Furthermore, the cloud server 3 may not be essential.

[0021] The information processing device 2 and the modeling system 1 are communicatively connected via a network N1. The network N1 is, for example, a LAN (wired LAN or wireless LAN) or a WAN installed in a building where the information processing device 2 and the modeling system 1 are located. The network N1 may be a dedicated line such as a USB cable or RS485, or an ad hoc network such as Bluetooth (registered trademark).

[0022] The information processing device 2 and the modeling system 1 can communicate with the cloud server 3 via networks N2 and N3, respectively. The networks N2 and N3 include a LAN installed in a building and may include the Internet in part. The networks N2 and N3 may also be mobile phone networks such as 4G, LTE, and 5G. The networks N1 to N3 may also be VLANs (Virtual Local Area Networks).

[0023] The molding system 1 is a series of devices for molding a powder compact. The molding system 1 can use a binder jetting method, a high-speed sintering method, etc. Details of the molding system 1 will be explained with reference to FIG. 3.

[0024] The information processing device 2 is a general-purpose computer such as a PC. It is assumed that the information processing device 2 will be operated by a designer, but it may be operated by any user with any role. The information processing device 2 displays the powder compact using modeling data (e.g., STL data) and displays the predicted shape of the final compact. The information processing device 2 also transmits modeling data to the modeling system 1 and instructs the start of modeling in response to the designer's operation. During modeling, the information processing device 2 also displays the status of the modeling system in real time. The information processing device 2 that communicates with the modeling system 1 during modeling and the information processing device 2 that performs the sintering simulation may be provided separately.

[0025] The cloud server 3, for example, stores data for modeling, executes a web application on the information processing device 2, and supports the designer in creating a powder compact model. The cloud server 3 may also be a device that performs the sintering simulation or machine learning described in this embodiment. The cloud server 3 may also be a server located on-premise.

[0026] <Modeling system> Next, with reference to FIG. 2, a molding system using a binder jetting method will be described as the molding system 1 of FIG. 1. FIG. 2 is a system configuration diagram of a molding system according to one embodiment of the present invention. The molding system 1 includes a molding apparatus 100, a drying apparatus 200, an excess powder removing apparatus 300, and a sintering apparatus 400. The molding apparatus 100 molds a powder molded body. The drying apparatus 200 dries the powder molded body molded by the molding apparatus 100. The excess powder removing apparatus 300 removes excess powder adhering to the powder molded body dried by the drying apparatus 200. The sintering apparatus 400 sinters the powder molded body from which the excess powder has been removed. The sintering apparatus 400 may degrease the powder molded body before sintering it.

[0027] The molding system 1 may be configured with four separate devices: the molding device 100, the drying device 200, the excess powder removing device 300, and the sintering device 400, or the four devices may be integrated into one device, or some of the functions may be installed in the molding device 100 or the sintering device 400. Furthermore, at least some of the functions of the drying device 200 or the excess powder removing device 300 may be omitted as appropriate.

[0028] Next, with reference to FIG. 3, the modeling apparatus 100 included in the modeling system 1 will be described. FIG. 3 is a configuration diagram of the modeling apparatus 100 according to one embodiment of the present invention. The modeling apparatus 100 includes a modeling unit 10 and an application unit 20. The modeling unit 10 models a powder layer 111 containing powder 11. The application unit 20 applies a modeling liquid 21 to the powder layer 111 to form a modeling layer 112. The modeling apparatus 100 models a powder molded body in which multiple modeling layers 112 are stacked.

[0029] The modeling section 10 includes a powder tank 12 and a lamination unit 13. The powder tank 12 includes a supply tank 121, a modeling tank 122, a supply stage 123, a modeling stage 124, and an excess powder tank 125.

[0030] The powder tank 12 has a box-like shape. Three top surfaces are open: the supply tank 121, the modeling tank 122, and the surplus powder tank 125. The lamination unit 13 has a flat portion 131 and a powder removal portion 132.

[0031] The supply tank 121 is a tank that supplies the powder 11 to the modeling tank 122. The supply tank 121 also holds the powder 11 to be supplied to the modeling tank 122. A supply stage 123 is provided at the bottom of the supply tank 121. The supply stage 123 moves up and down in the vertical direction (Z direction). The side of the supply stage 123 is arranged so as to contact the inner surface of the supply tank 121.

[0032] The powder 11 required for modeling is supplied to the modeling tank 122 from the supply tank 121. A powder layer 111 and a modeling layer 112 are formed in the modeling tank 122. Furthermore, a plurality of modeling layers 112 are stacked in the modeling tank 122 to form a powder molded body.

[0033] A supply stage 123 and a modeling stage 124 are provided at the bottom of the supply tank 121 and the modeling tank 122, respectively, and move up and down in the vertical direction (Z direction). The side of the modeling stage 124 is arranged so as to contact the inner surface of the modeling tank 122. The top surfaces of the supply stage 123 and the modeling stage 124 are kept horizontal.

[0034] The surplus powder tank 125 is a tank that holds surplus powder 11 among the powder 11 flattened by the flat portion 131 when forming the powder layer 111. A means for sucking the powder 11 may be provided at the bottom of the surplus powder tank 125, or a means for removing the surplus powder tank 125 may be provided. The surplus powder tank 125 is disposed next to the modeling tank 122. The surplus powder 11 held in the surplus powder tank 125 may be returned to the supply tank 121 or may be returned to the supply tank 121 via a powder supply device. The powder supply device may be disposed above the supply tank 121 and supplies powder 11 to the supply tank 121 before modeling begins or when the amount of powder 11 in the supply tank 121 decreases. Note that the powder tank 12 is described as having two tanks, the supply tank 121 and the modeling tank 122, but it may also be configured as having only the modeling tank 122, and powder may be supplied to the modeling tank 122 from the powder supply device.

[0035] Methods for transporting the powder 11 from the powder supply device to the supply tank 121 include a screw conveyor system using a screw, a pneumatic transport system using air, and the like.

[0036] The flattening section 131 flattens the modeling layer 112 and the powder layer 111. The flattening section 131 flattens the modeling layer 112 by rotating the recoater as a rotating body. When the flattening section 131 is driven to rotate, the powder 11 from the supply stage 123 of the supply tank 121 is supplied to the modeling tank 122, forming the powder layer 111. The flattening section 131 moves back and forth in the Y direction along the stage surface (the surface on which the powder 11 is loaded) of the modeling stage 124. More specifically, the flattening section 131 moves horizontally from the outside of the supply tank 121, passing above the supply tank 121 and the modeling tank 122. As a result, the powder 11 is transferred and supplied onto the modeling tank 122, and the flattening section 131 passes over the modeling tank 122, thereby forming the powder layer 111. The flattening section 131 is a member longer than the inner dimensions of the modeling tank 122 and the supply tank 121. The flat portion 131 may be a blade or a bar as a plate-like member.

[0037] Powder removal part 132 removes powder adhering to flat part 131. Powder removal part 132 moves together with flat part 131 while being in contact with the circumferential surface of flat part 131.

[0038] The dispensing unit 20 includes a carriage 211 and a head 212. The head 212 dispenses the modeling liquid 21 onto the powder layer 111. The head 212 is, for example, an inkjet head, and is provided with a nozzle row in which a plurality of nozzles are arranged. The dispensing unit 20 may form colored powder molded bodies by dispensing a cyan modeling liquid, a magenta modeling liquid, a yellow modeling liquid, and a black modeling liquid, or may dispense a single color modeling liquid from multiple nozzles. The modeling liquid may be dispensed using an inkjet method or a dispenser method.

[0039] At least one head 212 is mounted on a carriage 211, and is reciprocated in the X (main scanning), Y (sub-scanning), and Z directions by a motor, a guide member, and the like.

[0040] <Modeling flow using the modeling system> 4 is a flowchart illustrating a molding method according to one embodiment of the present invention. The molding method is performed using a molding system 1. First, the molding process S0 includes a forming process S1 in which a powder layer containing powder is formed, an applying process S2 in which a molding liquid is applied to the powder layer, and a stacking process S3 in which the forming process S1 and the applying process S2 are repeated. In this way, a powder molded body is formed.

[0041] After the molding step S0, a drying step S4 is performed to dry the powder compact. In the drying step S4, the powder compact is sintered by the drying device 200, and liquid components such as the solvent remaining in the powder compact are vaporized and removed. Thereafter, an excess powder removing step S5 is performed by the excess powder removing device 300 to remove excess powder adhering to the powder compact.

[0042] After the excess powder removal step S5, a debinding step S6 is performed using a sintering apparatus 400. In the debinding step S6, the powder compact is sintered in an atmosphere containing an inert gas, thereby removing the resin in the powder compact. This results in a debound body. Finally, step S7, which is a sintering step using the sintering apparatus 400, is performed to obtain a powder compact in which the debound body is sintered.

[0043] When performing the debinding step S6 and the sintering step S7 using the sintering device 400, the powder compact model may be placed on a setter. By placing the powder compact model on the setter, the stress applied to the powder compact due to gravity during the debinding step S6 and the sintering step S7 can be reduced, thereby preventing cracking and collapse. While ceramic or other materials can be used as the setter, a powder compact setter created simultaneously with the powder compact model can also be used. By using a powder compact setter, the setter deforms and shrinks in tandem with the powder compact model during the debinding step S6 and the sintering step S7, reducing the stress applied in the vertical and horizontal directions and preventing cracking and collapse.

[0044] <Hardware configuration example> <<Example of hardware configuration of information processing device and cloud server>> Fig. 5 is a hardware configuration diagram of an information processing device 2 and a cloud server 3 according to an embodiment of the present invention. As shown in Fig. 5, the information processing device 2 and the cloud server 3 are constructed by a computer 500. The computer 500 includes a CPU 501, a ROM 502, a RAM 503, an HD 504, an HDD (Hard Disk Drive) controller 505, a display 506, an external device connection I / F (Interface) 508, a network I / F 509, a bus line 510, a keyboard 511, a pointing device 512, an optical drive 514, and a media I / F 516.

[0045] Of these, the CPU 501 controls the overall operation of the computer 500. The ROM 502 stores programs, such as an IPL, used to drive the CPU 501. The RAM 503 is used as a work area for the CPU 501. The HD 504 stores various data, such as programs. The HDD controller 505 controls the reading and writing of various data from and to the HD 504 under the control of the CPU 501. The display 506 displays various information, such as a cursor, menus, windows, characters, or images. The external device connection I / F 508 is an interface for connecting various external devices. In this case, external devices include, for example, USB (Universal Serial Bus) memories and printers. The network I / F 509 is an interface for data communication using the networks N2 and N3. The bus line 510 is an address bus, a data bus, or the like, for electrically connecting the components, such as the CPU 501, shown in FIG. 5.

[0046] The keyboard 511 is a type of input means having multiple keys used to input characters, numbers, various instructions, etc. The pointing device 512 is a type of input means for selecting and executing various instructions, selecting a processing target, moving a cursor, etc. The optical drive 514 controls reading and writing of various data from an optical storage medium 513, which is an example of a removable storage medium. The optical storage medium may be a CD, DVD, Blu-Ray (registered trademark), etc. The media I / F 516 controls reading and writing (storing) of data from a storage medium 515, such as a flash memory.

[0047] <About the function> 6 is a functional block diagram of an information processing device 2 according to one embodiment of the present invention. The information processing device 2 includes a parameter value determination unit 31, a parameter value selection unit 32, a shape prediction unit 33, a shape error evaluation unit 34, a regression model construction unit 35, a search unit 36, a reception unit 37, an output unit 38, and a judgment unit 39. These functions of the information processing device 2 are functions or means realized by the CPU 501 of the information processing device 2 shown in FIG. 5 executing a program stored in the HD 504 or the like and controlling the hardware shown in FIG. 5.

[0048] The parameter value determination unit 31 creates an explanatory variable database P by combining values ​​of multiple parameter items. This explanatory variable database P is stored in the third storage unit 41. A parameter item is, for example, a factor that affects deformation. Hereinafter, the value of a parameter item is referred to as a parameter value. The parameter value may be set to a value within a range set by the designer as the explanatory variable X, or may be set to a fixed value. The parameter value determination unit 31 determines the parameter value using a random number from the range of values ​​set by the designer. The parameter value determination unit 31 creates the explanatory variable database P by combining multiple parameter values.

[0049] The parameter value selection unit 32 selects a combination of multiple parameter values. The parameter value selection unit 32 selects a combination of multiple parameter values ​​from the explanatory variable database P using, for example, experimental design, so that there is no overlap and no bias. Here, the selected combination of multiple parameter values ​​is referred to as a training explanatory variable set Q. The number of combinations is, for example, 30. The second memory unit 42 stores this training explanatory variable set Q. The second memory unit 42 also stores measured shape errors associated with the training explanatory variable set Q. Data associating the training explanatory variable set Q with the shape errors is referred to as "association data." The second memory unit 42 and the third memory unit 41 may be a single memory unit.

[0050] The shape prediction unit 33 predicts the shape of the sintered powder compact, i.e., the final compact, using the combination of parameter values ​​selected by the parameter value selection unit 32. More specifically, the shape prediction unit 33 uses the temperature threshold value A, the first viscosity, and the second viscosity as parameter values ​​and predicts the shape of the final compact by the finite element method using a temperature / viscosity curve obtained by combining these.

[0051] The shape error evaluation unit 34 calculates the shape error between the shape of the final molded body predicted by the shape prediction unit 33 and the shape of the actual final molded body. Here, the shape dimensions of the actual final molded body are obtained using a 3D scanner or the like. "Evaluation" may also be referred to as "calculation."

[0052] The regression model construction unit 35 constructs a regression model through machine learning using a plurality of learning data sets each of which combines a combination of parameter values ​​and a shape error. The regression model outputs a shape error in response to an input of a combination of parameter values.

[0053] The search unit 36 ​​uses the regression model created by the regression model construction unit 35 to search for a combination of several parameter values ​​that results in a shape error output that is closest to or close to 0. Note that the learning explanatory variable set Q used for learning in the explanatory variable database P may be excluded from the search targets. The search unit 36 ​​searches for a combination of multiple parameter values ​​that results in a shape error that is closest to or close to 0 from among the parameter values ​​in the explanatory variable database P excluding the learning explanatory variable set Q. The shape prediction unit 33 determines a new combination of parameter values ​​based on the regression model and predicts the shape of the final molded body.

[0054] The receiving unit 37 receives various operations on the information processing device 2 by the designer. For example, in this embodiment, in order to express a temperature / viscosity curve, at least three parameters, namely, a temperature threshold A, a first viscosity, and a second viscosity, are received as one of the parameter items. In addition, in this embodiment, input of an object to be subjected to a sintering simulation and an evaluation point of its shape is received. Examples of objects to be subjected to a sintering simulation include a powder compact and a final compact. The evaluation point is a position at which the shape error of the powder compact and the final compact is calculated. In this embodiment, the receiving unit 37 may further receive a model shape and an actually sintered model shape.

[0055] The output unit 38 displays the predicted shape of the final molded body, the shape error between the predicted shape and the actual shape of the final molded body, and the like on the display 506. Although the example of displaying on the display 506 has been shown as an example of the output unit 38, information on the predicted shape of the final molded body may also be transmitted to the molding apparatus 100 or the sintering apparatus 400.

[0056] The determination unit 39 determines whether the shape error obtained by the shape error evaluation unit 34 achieves the target. If the shape error achieves the target, the determination unit 39 stores the temperature threshold, the first viscosity, and the second viscosity as prediction conditions in the first storage unit 43. If the shape error does not achieve the target, the determination unit 39 may store correspondence data in the second storage unit 42 that associates the temperature threshold, the first viscosity, and the second viscosity with the shape error.

[0057] In addition, when the cloud server 3 performs the sintering simulation, the cloud server 3 may have a parameter value determination unit 31, a parameter value selection unit 32, a shape prediction unit 33, a shape error evaluation unit 34, a regression model construction unit 35, and a search unit 36.

[0058] <Sintering simulation> Next, a description will be given of the sintering simulation performed by the information processing device 2 of this embodiment. In the finite element method for solid-phase sintering, a viscous constitutive equation is incorporated into the governing equation.

[0059] The viscous constitutive equation shows the relationship between stress and strain rate of a material, and is particularly effective for materials that exhibit both elastic and viscous behavior, which is common in sintering. In this embodiment, the term "viscous" is used to mean both viscoelasticity and viscoplasticity.

[0060] The governing equations, including those used in static structural problems, are as follows: Equations (1) to (8) in Mathematical formula 1 all constitute viscous constitutive equations.

[0061]

number

[0062]

number

[0063]

number

[0064] The viscosity used in conventional viscous constitutive equations is calculated using the Arrhenius equation shown in equation (8). The Arrhenius equation has poor accuracy even when the physical properties of metal powders that exhibit liquid phase sintering are substituted. In particular, it has very poor reproducibility of excessive distortion due to gravity, such as slumping, which is seen in liquid phase sintering.

[0065] The viscosity calculated from Equation (8) is suitable for simulating the gradual viscosity change observed in liquid phase sintering, but is not suitable for representing the rapid viscosity change. Therefore, in this embodiment, a temperature / viscosity curve in which the viscosity decreases at a predetermined temperature threshold is used, as shown in Figure 7.

[0066] FIG. 7 is a diagram showing a temperature / viscosity curve according to one embodiment of the present invention. The horizontal axis represents temperature [°C], and the vertical axis represents viscosity [Pa·s]. In the temperature / viscosity curve of FIG. 7, when the temperature during sintering or simulation reaches a predetermined temperature threshold A, the viscosity decreases from a first viscosity to a second viscosity. That is, the powder exhibits liquid phase sintering when the temperature exceeds the temperature threshold A. Also, in FIG. 7, the viscosity decreases by more than two orders of magnitude at the temperature threshold A. Note that the amount of viscosity decrease, the temperature threshold A, the first viscosity, and the second viscosity vary depending on the type of metal powder, the powder compact, etc. The temperatures and viscosities shown in FIG. 7 are merely examples.

[0067] In the finite element method of this embodiment, the information processing device 2 receives a temperature / viscosity curve such as that shown in Figure 7 and then executes processing. However, there are multiple types of temperature / viscosity curves, and the accuracy depends on which temperature / viscosity curve is selected. Therefore, an optimal combination of parameter values ​​may be found by combining experimental design and machine learning.

[0068] The designer sets a range of parameter values ​​that allows for values ​​that cannot be set using standard physical property values, and combines experimental design with machine learning. For a temperature / viscosity curve like the one shown in Figure 7, the designer sets temperature threshold A, the first viscosity, and the second viscosity as parameter values. The first viscosity is higher than the second viscosity. This allows the designer to combine temperature threshold A, the first viscosity, and the second viscosity in experimental design to set an appropriate temperature / viscosity curve.

[0069] Furthermore, the information processing device 2 can determine highly accurate combinations of parameter values ​​by applying machine learning to the results of using the temperature / viscosity curve. This saves the designer the trouble of manually setting parameters. Furthermore, because it is possible to select parameter values ​​far removed from the viscosity derived by equation (8) based on the physical properties of the material (activation energy, frequency factor, etc.), it is possible to predict the shape of the final molded body with high accuracy.

[0070] For example, the first viscosity exhibits a constant value from 300°C to the temperature threshold A, and the second viscosity exhibits a constant value above the temperature threshold A. The temperature / viscosity curve is preferably a step function that transitions from the first viscosity to the second viscosity when the temperature threshold A is reached. This allows the information processing device 2 to simulate a sudden temperature change.

[0071] However, there may be a transition region where the viscosity transitions from the first viscosity to the second viscosity at temperatures before and after the temperature threshold A. The transition region is a temperature range where the viscosity changes from the first viscosity to the second viscosity, rather than decreasing at a certain temperature threshold A like a step function. FIG. 8 is a diagram showing a temperature / viscosity curve according to one embodiment of the present invention. In FIG. 8, the viscosity changes sigmoidally or exponentially within the transition region. Therefore, the transition region is preferably within 100°C or 50°C before and after the temperature threshold A. Furthermore, it is preferable to set the viscosity difference between the first viscosity and the second viscosity to be at least two times.

[0072] Furthermore, the temperature / viscosity curve is not limited to two stages and may be changed in three stages. This is because the temperature at which the powder changes from solid to liquid varies depending on the powder mixture. Therefore, multiple temperature / viscosity curves may be used, each of which has a set of temperature threshold A, first viscosity, and second viscosity.

[0073] <Information processing method> Next, an information processing method according to one embodiment of the present invention will be described with reference to Fig. 9 to Fig. 13. Fig. 9 is a flowchart showing the processing performed by the information processing device 2 according to one embodiment of the present invention.

[0074] S91: First, the designer inputs the shape of a powder compact (an example of a second object before sintering) for which a sintering simulation is to be performed. That is, the receiving unit 37 receives the shape of the powder compact for which a sintering simulation is to be performed. The shape of the powder compact is three-dimensional data. In addition to the shape of the powder compact, the shape of an actual final compact (an example of a second object after sintering) and evaluation points for the shape may also be received. The shape of the actual final compact is obtained by actually molding and sintering a final compact, and the shape is acquired using a 3D scanner or the like. The shape of the actual final compact may be received in a separate process from S91. In this embodiment, an example is shown in which the receiving unit 37 receives the model shape shown in FIG. 10 as the powder compact.

[0075] Here, computer-generated 3D model data is used as the model shape of the powder compact. FIG. 10 is a diagram showing the model shape of a powder compact according to one embodiment of the present invention. The model shape makes it easy to observe slumping during liquid phase sintering, making it suitable for shape prediction. As shown in FIG. 10, for example, the designer sets nine evaluation points, W1, W2, H1, H2, H3, L1, L2, θ1, and θ2, on the input shape of the powder compact. Note that, although this embodiment illustrates an example in which the designer sets the evaluation points, the parameter value determination unit 31 may also determine the evaluation points using image processing or the like.

[0076] S92: Next, the designer sets the parameter items. That is, the reception unit 37 receives the setting of the parameter items. The parameter items are values ​​that are factors that affect the deformation. The designer sets the parameter items as explanatory variables X. The designer can set the setting items used in the finite element method as the parameter items. In this embodiment, in order to express the temperature / viscosity curve, at least three of the temperature threshold value A, the first viscosity, and the second viscosity are set as one of the parameter items. The temperature threshold value A, the first viscosity, and the second viscosity correspond to the prediction conditions for predicting the shape of the final molded body.

[0077] The designer also determines the range of values ​​that each parameter value can take. That is, the receiving unit 37 receives the ranges of the temperature threshold, the first viscosity, and the second viscosity. The range of values ​​is called a window. The window is, for example, an upper limit and a lower limit. For some parameter items, a fixed value may be set instead of a window. For example, the receiving unit 37 may further receive input of a temperature transition region. The temperature transition region is the temperature range in which the viscosity of the powder compact decreases from the first viscosity to the second viscosity through sintering, and the temperature threshold is a value within the temperature range of the temperature transition region.

[0078] S92 may be performed simultaneously with S91 or before S91.

[0079] S93: Next, the parameter value determination unit 31 determines parameter values ​​using random numbers within the range of values ​​set in S92. Then, the parameter value determination unit 31 creates an explanatory variable database P by combining the parameter values ​​of the parameter items set for the explanatory variable X. The explanatory variable database P is stored in the third storage unit 41. For example, the parameter value determination unit 31 generates 10,000 combinations of parameter values ​​using random numbers, and creates the explanatory variable database P.

[0080] S94: Next, the parameter value selection unit 32 selects approximately several tens of parameter value combinations. The parameter value selection unit 32 selects approximately several tens of parameter value combinations from the explanatory variable database P, for example, by experimental design. Known experimental design methods include an orthogonal array, a D-optimal criteria method, or a combination thereof. The parameter value selection unit 32 selects approximately several tens of parameter value combinations from the explanatory variable database P, for example, by the D-optimal criteria method, as a training data set for initial training in machine learning. For example, 30 parameter value combinations selected by the parameter value selection unit 32 are set as a training explanatory variable set Q. The training explanatory variable set Q is stored in the second storage unit 42.

[0081] S95: Next, the shape prediction unit 33 performs a sintering simulation on the shape of the powder compact inputted in S91 using the learning explanatory variable set Q selected in S94, and predicts the shape of the sintered powder compact (an example of a second shaped object after sintering), i.e., the final compact. For example, if there are 30 different learning explanatory variable sets Q, the shape of the final compact is predicted 30 times. In this embodiment, the shape prediction unit 33 extracts the temperature threshold A, the first viscosity, and the second viscosity from the learning explanatory variable set Q to create a temperature / viscosity curve, and predicts the shape of the final compact based on the created temperature / viscosity curve. The shape of the final compact is predicted using a different combination of the temperature threshold, the first viscosity, and the second viscosity for each prediction. The shape prediction unit 33 may be programmed by the designer himself or may use commercially available software such as Simufact Additive (registered trademark) from Hexagon or ANSYS (registered trademark) from ANSYS.

[0082] S96: Next, the shape error evaluation unit 34 calculates a first shape error between the shape of the final compact predicted in S95 and the shape of the actual final compact. The shape error evaluation unit 34 stores the first shape error in the second memory unit 42 in association with the learning explanatory variable set Q (first process). The shape error evaluation unit 34 executes the first process a predetermined number of times. This serves as the correspondence data. The shape of the final compact predicted in S95 is a shape predicted from the model shape after deformation due to sintering. Therefore, for example, the first shape error is a value expressed as a percentage of the difference between the shape of the final compact predicted in S95 and the actual measured value of the final compact. The shape error between the predicted final compact and the actual final compact at the evaluation point is set as the objective variable Y. Here, the evaluation points for the predicted final compact and the actual final compact may be automatically set by the shape error evaluation unit 34 based on the evaluation points for the powder compact set in S91, or the receiving unit 37 may receive input of the evaluation points for each. The closer the objective variable Y is to 0, the higher the accuracy.

[0083] FIG. 11 shows an example of correspondence data (first viscosity, second viscosity, temperature threshold, first shape error at evaluation point) recorded by the shape error evaluation unit 34. The first viscosity, second viscosity, and temperature threshold are the learning explanatory variable set Q. The positions of the evaluation points indicated by H, L, W, and θ are shown in FIG. 10. Although FIG. 11 shows four records (four rows) of correspondence data, if there are 30 different learning explanatory variable sets Q, there will be 30 records. The first shape error is shown as an error rate, but it can also be an error value. The explanatory variables X are the first viscosity, second viscosity, and temperature threshold, and the response variables Y are H, L, W, and θ.

[0084] S97: Next, the regression model construction unit 35 constructs a regression model through machine learning using a combination of multiple explanatory variables X and response variables Y as a training dataset R. Therefore, the training dataset R is the same as the training explanatory variable set Q stored in the second storage unit 42. The initial number of explanatory variables X and response variables Y is 30, the same as the number of training explanatory variables Q. When additional explanatory variables and response variables are added in S101, the number exceeds 30. Algorithms that can be used for machine learning include, for example, decision trees, neural networks, linear multiple regression analysis, nonlinear multiple regression analysis, support vector regression, Gaussian process regression, and Bayesian optimization.

[0085] S98: The search unit 36 ​​uses the regression model constructed in S97 to search the explanatory variable database P for a combination of parameter values ​​that results in a shape error output that is closest to or is closest to zero. Note that the search target may exclude the learning explanatory variable set Q used for learning from the explanatory variable database P. In other words, in S98, a combination of a temperature threshold, a first viscosity, and a second viscosity different from those in S95 is set. For example, the search unit 36 ​​searches for a combination of parameter values ​​from 10,000-30=9,970 sets. That is, the search unit 36 ​​inputs the combinations of parameter values ​​from the explanatory variable database P, excluding the learning explanatory variable set Q, into the regression model and repeats the process of outputting the shape error of the final molded body. The search unit 36 ​​searches for a combination of multiple parameter values ​​that results in a shape error that is closest to or is closest to zero from the combinations of parameter values ​​in the explanatory variable database P, excluding the learning explanatory variable set Q.

[0086] S99: The shape prediction unit 33 performs a sintering simulation using the combination of parameter values ​​searched for in S98, and predicts the shape of the final compact. By using a combination of multiple parameter values ​​whose shape error is closest to 0 or is close to 0, the shape of the final compact can be predicted with high accuracy. The shape error evaluation unit 34 calculates a second shape error between the shape of the final compact predicted in S99 and the actual shape of the final compact (second process).

[0087] S100: The determination unit 39 determines whether the second shape error obtained in S99 reaches the target. If the determination unit 39 determines that the target is reached, the search unit 36 ​​determines that the combination of parameter values ​​is the optimal parameter, and determines the temperature threshold, first viscosity, and second viscosity (corresponding to the prediction conditions). The process of FIG. 9 ends. The search unit 36 ​​stores the determined temperature threshold, first viscosity, and second viscosity in the first memory unit 43. The output unit 38 may display the shape of the final compact obtained in S99 or may display the second shape error as a numerical value, etc. Furthermore, the output unit 38 may display the combination of parameter values ​​(corresponding to the prediction conditions) used by the shape prediction unit 33 when the determination unit 39 determines that the shape error reaches the target. The displayed combination of parameter values ​​may be the optimal combination of parameter values ​​in the sintering simulation of this embodiment. Furthermore, the output unit 38 may transmit the prediction conditions to the sintering apparatus 400. The sintering device 400 can determine the sintering temperature based on the predicted conditions, and can control the rate of increase and time until the temperature threshold is reached.

[0088] Here, the term "goal" refers to satisfying a predetermined condition. The goal may be a value for the degree of coincidence between the shape output by the designer in S99 and the shape of the actual powder compact, or a tolerance for a second shape error may be set. While the model shape has been described, the shape of the final compact to be actually manufactured may also be predicted.

[0089] S101: If the judgment unit 39 determines that the evaluation result obtained in S99 does not achieve the target, the designer sets the second shape error between the shape of the final molded body output in S99 and the shape of the actual final molded body as the objective variable Y, and the combination of multiple parameter values ​​for which the shape error found in S98 is closest to 0 or closest to 0 as the explanatory variable X, and adds the objective variable Y and the explanatory variable X to the learning dataset R (adding them as corresponding data to the learning explanatory variable set Q in the second memory unit 42). The regression model construction unit 35 constructs a regression model by machine learning. The process returns to S97, and S97 to S100 are executed again. S97 to S100 are repeatedly executed until the evaluation result in S100 achieves the target.

[0090] 9, the shape of the final compact is predicted by a sintering simulation, but the information processing device 2 may also predict the amount of deformation from the model shape. For example, it may predict the degree to which each evaluation point of the model shape shown in FIG. 10, for which the sintering simulation is performed, will deform due to sintering.

[0091] <Additional information on machine learning> 12 is a functional block diagram of a regression model construction unit 35 according to one embodiment of the present invention. The regression model construction unit 35 can include a data acquisition unit 221, a training data storage unit 222, a machine learning unit 223, a trained model storage unit 224, and an inference unit 225. Each of these units will be described below.

[0092] <<Learning Phase>> The data acquisition unit 221 acquires explanatory variables X and objective variables Y as learning data and stores them in the learning data storage unit 222. A plurality of learning data sets are stored in the learning data storage unit 222. For example, 30 learning data sets R are stored.

[0093] The machine learning unit 223 acquires one set of training data at a time from the training data storage unit 222 and generates a trained model as a regression model. Specifically, the machine learning unit 223 performs machine learning using an explanatory variable X included in the training data as input data and a target variable Y associated with the explanatory variable X as training data, thereby generating a trained model. The machine learning unit 223 also stores the generated trained model in the trained model storage unit 224. The trained model storage unit 224 stores the trained model generated by the machine learning unit 223.

[0094] <<Inference Phase>> The data acquisition unit 221 acquires a combination of multiple parameter values ​​as the explanatory variable X and passes it to the inference unit 225. The inference unit 225 infers a shape error from the parameter values. Specifically, the inference unit 225 inputs the parameter values ​​into a trained model stored in the trained model storage unit 224 and outputs a shape error.

[0095] FIG. 13 is a cross-sectional view showing the results of a sintering simulation according to one embodiment of the present invention. FIG. 13 illustrates the sintering of a powder compact model shown in FIG. 13(a) using an AlSi (aluminum-silicon) alloy as powder. The powder compact shown in FIG. 13(a) is sintered using a metal binder jetting device to produce the final compact shown in FIG. 13(b). Therefore, it is desirable to reproduce a shape similar to that shown in FIG. 13(b) when performing a sintering simulation on the shape data shown in FIG. 13(a). FIG. 13(c) shows the shape of a sintering simulation result obtained by optimizing parameter values ​​through sensitivity analysis using Equation (8) without using a temperature / viscosity curve, as a comparative example. FIG. 13(d) shows the shape of a sintering simulation result obtained by optimizing parameter values ​​using experimental design and machine learning with a temperature / viscosity curve according to one embodiment of the present invention. The shapes in FIGS. 13(c) and 13(d) are all three-dimensional data.

[0096] In Figure 13(c), although the shrinkage of the external shape was reproduced to some extent, slumping was not reproduced, and there was a large shape error compared to the final molded product shown in Figure 13(b).On the other hand, Figure 13(d), which was optimized using the temperature / viscosity curve with experimental design and machine learning, reproduced slumping well, and the shape error compared to Figure 13(b) was small, demonstrating good shape prediction by simulation.

[0097] Therefore, by using the temperature / viscosity curve and optimizing the parameter values ​​through experimental design and machine learning, the shape of the final molded body can be predicted with high accuracy.

[0098] <Simulation after calculating the temperature / viscosity curve> FIG. 14 is a flowchart illustrating the flow of a simulation using the prediction conditions (temperature threshold, first viscosity, second viscosity) determined in the process of FIG.

[0099] S201: First, the receiving unit 37 receives an input of the shape of a powder compact (an example of a first object before sintering).

[0100] S202: The shape prediction unit 33 selects the prediction conditions (temperature threshold, first viscosity, second viscosity) determined in the process of Fig. 9 from the first storage unit 43. This "selection" means, for example, that when there are different parameters in the powder composition information (e.g., a temperature / viscosity curve for aluminum, a temperature / viscosity curve for titanium, etc.), the shape prediction unit 33 selects parameters according to the composition of the input powder compact.

[0101] S203: The shape prediction unit 33 predicts the shape of the sintered powder compact, i.e., the final compact, based on the selected combination of parameters.

[0102] S204: The output unit 38 outputs the shape of the final molded body predicted in S203.

[0103] <Simulation by inverse analysis after calculating the temperature / viscosity curve> Fig. 15 is a flowchart illustrating the flow of a simulation by inverse analysis using the optimal parameters (temperature threshold, first viscosity, second viscosity) determined by the process of Fig. 9. In Fig. 15, the shape of the powder compact is predicted based on the shape of the final compact using the temperature / viscosity curve obtained in Fig. 9. Performing a simulation by switching the input and output is called inverse analysis.

[0104] S301: First, the receiving unit 37 receives input of the shape of a final molded body (an example of a first shaped object after sintering).

[0105] S302: The shape prediction unit 33 selects the optimal parameters (temperature threshold, first viscosity, second viscosity) determined in the process of Fig. 9 from the first storage unit 43. This "selection" means, for example, that when there are different parameters in the powder composition information (e.g., a temperature / viscosity curve for aluminum, a temperature / viscosity curve for titanium, etc.), the parameters are selected according to the composition of the input powder compact.

[0106] S303: The shape prediction unit 33 predicts, by inverse analysis, the shape of the powder compact that will become the shape of the final compact received in S301 after sintering. The inverse analysis method involves iteratively correcting the predicted shape of the powder compact so that it converges to the target final compact. Specifically, the shape of the powder compact is initially set based on the input shape of the final compact. A sintering simulation is performed using the initially set powder compact to obtain a predicted result of the final compact. The difference between the predicted shape of the final compact obtained by this sintering simulation and the input shape of the final compact is quantified using a "cost function." The initially set shape of the powder compact is changed so that the output (difference) of the cost function is minimized. This process is then repeated until the difference becomes equal to or less than a threshold value.

[0107] One method of inverse analysis is to use a large amount of simulation data to construct a model using, for example, a neural network that predicts the shape of the powder compact from the shape of the final compact.

[0108] S304: The output unit 38 outputs the shape of the powder compact predicted in S303.

[0109] <Other application examples> The present invention is not limited to the specifically disclosed embodiments above, but various modifications and variations are possible without departing from the scope of the claims.

[0110] For example, in this embodiment, the information processing device 2 performs the sintering simulation, but the cloud server 3 may also perform the sintering simulation. In this case, the designer connects the information processing device 2 to the cloud server 3 and runs, for example, a web application. The designer sets the web application in settings related to the sintering simulation, and the cloud server 3 performs the sintering simulation in accordance with the settings.

[0111] Furthermore, although the number of parameter items is set to a plurality, the number of parameter items may be set to one.

[0112] Furthermore, a reverse analysis may be performed using the combination of parameter values ​​obtained in this embodiment. That is, the shape of a final green body to be manufactured may be input, and the shape of a powder green body that is optimal for obtaining the final green body to be manufactured may be predicted.

[0113] Furthermore, in this embodiment, the sintering simulation after molding by MBJ has been described, but molding may also be performed by high speed sintering (HSS).

[0114] In addition, although the present embodiment has been described using pure aluminum or aluminum alloy powder as an example, other metal powders may be used for manufacturing. The powder does not have to be metal, and any powder that exhibits a liquid phase upon sintering can be suitably used.

[0115] Furthermore, in this embodiment, the model shape alone has been described as an example, but the present invention may also be applied to a model placed on a ceramic setter or a powder compact setter.

[0116] In addition, the configuration examples in Fig. 6 and the like are divided according to main functions to facilitate understanding of the processing by the information processing device 2. The present invention is not limited by the way in which the processing units are divided or the names of the processing units. The processing by the information processing device 2 can also be divided into more processing units depending on the processing content. Furthermore, it can also be divided so that one processing unit includes more processes.

[0117] Each function of the above-described embodiments can be realized by one or more processing circuits. Here, the term "processing circuit" in this specification includes a processor programmed to perform each function by software, such as a processor implemented by an electronic circuit, as well as devices such as an ASIC (Application Specific Integrated Circuit), a DSP (Digital Signal Processor), an FPGA (Field Programmable Gate Array), or a conventional circuit module designed to perform each function described above.

[0118] Digital twins, which connect the real world and virtual space, are a technology that enables predictions and optimization of the real world by recreating real-world objects in virtual space and simulating their behavior. Simulation is at the heart of digital twins and is an important tool for conducting various experiments and analyses in virtual space based on real-world data. Simulation data quantifies and models complex real-world phenomena, bringing benefits such as improved prediction accuracy, optimization, risk reduction, and cost reduction. Digital twins are expected to become increasingly sophisticated through integration with AI and IoT. For example, it is thought that machine learning will be used to improve the accuracy of simulation models, and edge computing will enable real-time simulations.

[0119] An example of a method for applying the simulation data in the above-described embodiment to a digital twin is as follows.

[0120] (1) Preprocessing of simulation data We perform "data cleaning," which processes noise and missing values ​​contained in the collected simulation data to improve the quality of the data, and then we perform "data conversion," which converts the cleaned simulation data into a format that can be handled by the simulation model.

[0121] (2) Building a simulation model We carry out "physical model construction," which involves constructing a model that describes the physical characteristics of the object to be simulated; then we carry out "mathematical model construction," which involves expressing the constructed physical model in mathematical equations; and finally we carry out "numerical analysis," which involves developing an algorithm to numerically solve the constructed mathematical model.

[0122] (3) Integration into Digital Twin We will carry out "simulation result visualization," which displays the simulation results in a visually easy-to-understand format such as 3D models and graphs, and then we will carry out "model accuracy verification," which compares the visualized simulation results with real-world measurement data, sensor data, etc., to verify the accuracy of the model. We will then improve the accuracy of the verified model by adjusting real-world control parameters based on the simulation results.

[0123] By applying the simulation data in the above-described embodiment to a digital twin, it becomes possible to simulate objects made of different materials using a 3D printer, and it also becomes possible to apply the principles of a 3D printer to simulate the layering of various materials to create three-dimensional objects such as food, cosmetics, electronic circuits, and paint.

[0124] <Additional Notes> [Appendix 1] The information processing device Accepting the shape of the first sintered object; predicting a shape of the first object before sintering based on a shape of the first object after sintering and prediction conditions stored in a storage unit; an information processing method for outputting information about a predicted shape of the first object before sintering, An information processing method characterized in that the prediction conditions are information based on a temperature threshold, which is the temperature at which the viscosity of the first object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity. [Appendix 2] The information processing device Accepting the shape of the first object before sintering; predicting a shape of the first object after sintering based on a shape of the first object before sintering and prediction conditions stored in a storage unit; an information processing method for outputting information about a predicted shape of the first object after sintering, An information processing method characterized in that the prediction conditions are information based on a temperature threshold, which is the temperature at which the viscosity of the first object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity. [Appendix 3] Before receiving the shape of the first object, receiving a shape of a second object before sintering and a shape of the second object after actual sintering; accepting the temperature threshold, the first viscosity, and the second viscosity range; a first process of predicting a shape of the second object after sintering based on the temperature threshold, the first viscosity, and the second viscosity set within the ranges, calculating a first shape error between the predicted shape of the second object after sintering and an actual shape of the second object after sintering, and storing correspondence data that associates the temperature threshold, the first viscosity, and the second viscosity set within the ranges with the calculated first shape error; 3. The information processing method according to claim 1, wherein the prediction condition is determined based on a plurality of pieces of correspondence data stored by executing the first process a predetermined number of times. [Appendix 4] The information processing method according to claim 3, wherein a different combination of the temperature threshold, the first viscosity, and the second viscosity is set for each of the first processes. [Appendix 5] After the first process is executed a predetermined number of times, executing a second process of predicting a shape of the second object after sintering using the plurality of pieces of correspondence data, and calculating a second shape error between the predicted shape of the second object after sintering and the actual shape of the second object after sintering; 5. The information processing method according to claim 3, wherein the prediction condition is determined based on the second shape error calculated in the second process. [Appendix 6] The information processing method according to Appendix 5, wherein in the second process, the temperature threshold, the first viscosity, and the second viscosity are set within the ranges in a combination different from that in the first process, and the shape of the second object after sintering is predicted based on the temperature threshold, the first viscosity, and the second viscosity within the set ranges and the plurality of pieces of correspondence data. [Appendix 7] determining the prediction condition when the second shape error calculated in the second process satisfies a predetermined condition; 7. The information processing method according to claim 5, wherein if the second shape error does not satisfy a predetermined condition, the second process is executed again. [Appendix 8] An information processing method according to claim 7, wherein, if the second shape error calculated in the second process does not satisfy a predetermined condition, correspondence data is stored that associates the temperature threshold, the first viscosity, and the second viscosity values ​​within the range set in the second process with the calculated second shape error. [Appendix 9] Additionally, the range of the temperature transition region is accepted, the temperature transition region is a temperature range in which the viscosity of the second shaped object decreases from a first viscosity to a second viscosity through sintering; 9. The information processing method according to any one of claims 3 to 8, wherein the temperature threshold is a value within the temperature transition region. [Appendix 10] The information processing device further receiving an evaluation point which is a position at which the first shape error of the second object is calculated; An information processing method according to any one of claims 3 to 9, wherein the first shape error is an error between a predicted shape of the second object after sintering at the evaluation point and an actual shape of the second object after sintering. [Appendix 11] 11. The information processing method according to any one of claims 1 to 10, further comprising receiving composition information of the first object. [Appendix 12] 12. The information processing method according to any one of claims 1 to 11, wherein the first object contains aluminum. [Appendix 13] The information processing device receiving a shape of a second object before sintering and a shape of the second object after actual sintering; receiving a temperature threshold value, which is a temperature at which the viscosity of the second object decreases from a first viscosity to a second viscosity due to sintering, and a range between the first viscosity and the second viscosity; predicting a shape of the first object after sintering based on the temperature threshold, the first viscosity, and the second viscosity set within the ranges; calculating a first shape error between the predicted shape of the first object after sintering and an actual shape of the second object after sintering; and storing correspondence data that associates the temperature threshold, the first viscosity, and the second viscosity set within the ranges with the calculated first shape error. determining a prediction condition based on the plurality of correspondence data stored by executing the process a predetermined number of times; outputting the determined prediction conditions; An information processing method characterized in that the prediction conditions are information based on a temperature threshold, which is the temperature at which the viscosity of the first object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity. [Appendix 14] a receiving unit that receives the shape of the first object after sintering; a prediction unit that predicts a shape of the first object before sintering, based on a shape of the first object after sintering and prediction conditions stored in a storage unit; an output unit that outputs information about the predicted shape of the first object before sintering, The information processing device is characterized in that the prediction conditions are information based on a temperature threshold value, which is a temperature at which the viscosity of the first object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity. [Appendix 15] For information processing devices Accepting the shape of the first shaped object after sintering; predicting a shape of the first object before sintering based on a shape of the first object after sintering and prediction conditions stored in a storage unit; executing a process of outputting information about the predicted shape of the first object before sintering; The prediction conditions are information based on a temperature threshold, which is the temperature at which the viscosity of the first object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity. [Explanation of symbols]

[0125] 1 Modeling System 2. Information processing equipment 3. Cloud Server [Prior art documents] [Patent documents]

[0126] [Patent Document 1] Japanese Patent Application Laid-Open No. 2022-021956

Claims

1. The information processing device Accepting the shape of the first sintered object; predicting a shape of the first object before sintering based on a shape of the first object after sintering and prediction conditions stored in a storage unit; an information processing method for outputting information about a predicted shape of the first object before sintering, An information processing method characterized in that the prediction conditions are information based on a temperature threshold, which is the temperature at which the viscosity of the first object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity.

2. The information processing device Accepting the shape of the first object before sintering; predicting a shape of the first object after sintering based on a shape of the first object before sintering and prediction conditions stored in a storage unit; an information processing method for outputting information about a predicted shape of the first object after sintering, An information processing method characterized in that the prediction conditions are information based on a temperature threshold, which is the temperature at which the viscosity of the first object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity.

3. Before receiving the shape of the first object, receiving a shape of a second object before sintering and a shape of the second object after actual sintering; accepting the temperature threshold, the first viscosity, and the second viscosity range; a first process of predicting a shape of the second object after sintering based on the temperature threshold, the first viscosity, and the second viscosity set within the ranges, calculating a first shape error between the predicted shape of the second object after sintering and an actual shape of the second object after sintering, and storing correspondence data that associates the temperature threshold, the first viscosity, and the second viscosity set within the ranges with the calculated first shape error; 3. The information processing method according to claim 1, wherein the prediction condition is determined based on a plurality of pieces of correspondence data stored by executing the first process a predetermined number of times.

4. The information processing method according to claim 3 , wherein a different combination of the temperature threshold, the first viscosity, and the second viscosity is set for each of the first processes.

5. After the first process is executed a predetermined number of times, executing a second process of predicting a shape of the second object after sintering using the plurality of pieces of correspondence data, and calculating a second shape error between the predicted shape of the second object after sintering and the actual shape of the second object after sintering; The information processing method according to claim 3 , wherein the prediction condition is determined based on the second shape error calculated in the second process.

6. 6. The information processing method according to claim 5, wherein in the second process, the temperature threshold, the first viscosity, and the second viscosity within the range are set in a combination different from that in the first process, and the shape of the second object after sintering is predicted based on the temperature threshold, the first viscosity, and the second viscosity within the set range and a plurality of pieces of the corresponding data.

7. determining the prediction condition when the second shape error calculated in the second process satisfies a predetermined condition; 7. The information processing method according to claim 6, wherein if the second shape error does not satisfy a predetermined condition, the second process is executed again.

8. 8. The information processing method according to claim 7, wherein, if the second shape error calculated in the second process does not satisfy a predetermined condition, correspondence data is stored that associates the calculated second shape error with the temperature threshold, the first viscosity, and the second viscosity values ​​within the range set in the second process.

9. Additionally, the range of the temperature transition region is accepted, the temperature transition region is a temperature range in which the viscosity of the second shaped object decreases from a first viscosity to a second viscosity through sintering; The information processing method according to claim 3 , wherein the temperature threshold value is a value within the temperature transition region.

10. The information processing device further receiving an evaluation point which is a position at which the first shape error of the second object is calculated; The information processing method according to claim 3 , wherein the first shape error is an error between a predicted shape of the second object after sintering at the evaluation point and the actual shape of the second object after sintering.

11. The information processing method according to claim 1 or 2, further comprising receiving composition information of the first object.

12. The information processing method according to claim 1 or 2, wherein the first object contains aluminum.

13. The information processing device receiving a shape of a second object before sintering and a shape of the second object after actual sintering; receiving a temperature threshold value, which is a temperature at which the viscosity of the second object decreases from a first viscosity to a second viscosity due to sintering, and a range between the first viscosity and the second viscosity; predicting a shape of the first object after sintering based on the temperature threshold, the first viscosity, and the second viscosity set within the ranges; calculating a first shape error between the predicted shape of the first object after sintering and an actual shape of the second object after sintering; and storing correspondence data that associates the temperature threshold, the first viscosity, and the second viscosity set within the ranges with the calculated first shape error. determining a prediction condition based on the plurality of correspondence data stored by executing the process a predetermined number of times; outputting the determined prediction conditions; An information processing method characterized in that the prediction conditions are information based on a temperature threshold, which is the temperature at which the viscosity of the first object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity.

14. a receiving unit that receives the shape of the first object after sintering; a prediction unit that predicts a shape of the first object before sintering, based on a shape of the first object after sintering and prediction conditions stored in a storage unit; an output unit that outputs information about the predicted shape of the first object before sintering, The information processing device is characterized in that the prediction conditions are information based on a temperature threshold value, which is a temperature at which the viscosity of the first object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity.

15. For information processing devices Accepting the shape of the first shaped object after sintering; predicting a shape of the first object before sintering based on a shape of the first object after sintering and prediction conditions stored in a storage unit; executing a process of outputting information about the predicted shape of the first object before sintering; The prediction conditions are information based on a temperature threshold, which is the temperature at which the viscosity of the first object decreases from a first viscosity to a second viscosity due to sintering, the first viscosity, and the second viscosity.

Images