Presumption device, identification device, presumption method, identification method

Abstract

To provide a value of transverse intensity from a plurality of parameters relating to the production of a casting mold.SOLUTION: A processor (11) of an estimating device (1) executes a transverse intensity estimation step (M11 or M31) for estimating the transverse intensity of a casting mold from the transverse intensity of the casting mold and a parameter set of a plurality of parameters relating to the production of the casting mold. The plurality of parameters includes a raw material parameters, a kneading parameter, and a molding parameter.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an apparatus and method for estimating the flexural strength of a mold from a number of parameters relating to the creation of the mold. [Background technology]

[0002] A mold is used to create a casting mold. A mold is completed by pouring the mixed sand, which is made by adding additives to molding sand, into a mold. The mold must have a flexural strength that corresponds to the shape and type of casting to be cast. It is known that the flexural strength of a mold varies depending on various parameters, such as the amount and quality of the raw materials for the mold, the conditions when mixing the raw materials, and the conditions when molding the mixed sand. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2010-160007 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, it was unclear how the various parameters described above affect the flexural strength. Therefore, in the past, mold making experiments were actually performed using parameters provisionally set based on the creator's experience, and conditions for making a mold with a desired flexural strength were identified. In other words, the flexural strength of a mold was unknown until it was actually made.

[0005] One aspect of the present invention has been made in consideration of the above problems, and has an object to estimate the value of flexural strength from multiple parameters related to the creation of a mold. [Means for solving the problem]

[0006] In order to solve the above problems, an estimation device according to an aspect of the present invention includes one or more processors that execute a first estimation step. Also, an estimation method according to an aspect of the present invention includes the first estimation step.

[0007] In the estimation device and the estimation method, a first estimation step is a step in which the processor estimates the flexural strength of the mold from a parameter set of a plurality of parameters related to the creation of the mold, where the plurality of parameters include (1) raw material parameters related to the raw material of the mold, (2) mixing parameters measured or set in a mixing process of the raw material, and (3) molding parameters measured or set in a molding process of the mixed sand created in the mixing process.

[0008] In order to achieve the above object, a determination device according to an aspect of the present invention includes one or more processors that execute a determination step. Also, a determination method according to an aspect of the present invention includes a determination step.

[0009] In the identification device and the identification method, the identification step is a step of identifying a relational expression or a trained model for estimating the flexural strength from a parameter set of a plurality of parameters. Here, the plurality of parameters are a plurality of parameters related to the creation of a mold. The plurality of parameters include (1) raw material parameters related to the raw material of the mold, (2) mixing parameters measured or set in a mixing process of the raw material, (3) molding parameters measured or set in a molding process of the mixed sand created in the mixing process, and (4) the flexural strength of the mold. Effect of the Invention

[0010] According to one aspect of the present invention, the value of the flexural strength can be estimated from multiple parameters related to the creation of the mold. [Brief description of the drawings]

[0011] [Figure 1]FIG. 2 is a diagram illustrating an outline of a process for producing a template in the first embodiment of the present invention. [Diagram 2] FIG. 1 is a schematic diagram of an estimation system according to the embodiment. [Diagram 3] 1 is a block diagram showing a configuration of an estimation device according to a first embodiment. [Figure 4] 4 is a flowchart showing a flow of a first estimation method. [Diagram 5] FIG. 10 is a diagram illustrating a parameter set selection method in the first estimation method. [Figure 6] 1 is a block diagram showing a configuration of a relational equation specifying device according to a first embodiment. [Figure 7] 10 is a diagram showing the types and acquisition timing of actual parameter values ​​acquired by the relational equation identifying device in the mold creation process. FIG. [Figure 8] 1 is a flowchart showing a flow of a first identification method. [Figure 9] FIG. 2 is a diagram illustrating a genetic algorithm executed in a first identification method. [Figure 10] 10 is a flowchart showing a flow of a first relational equation specifying step executed by the relational equation specifying device. [Figure 11] 10 is a flowchart showing a flow of processes included in a first relational expression specifying step. [Figure 12] FIG. 13 is a diagram illustrating an example of the contents of a crossover. [Figure 13] FIG. 1 illustrates the contents of a subtree mutation. [Figure 14] FIG. 1 illustrates the content of a hoist mutation. [Figure 15] FIG. 1 illustrates the contents of point mutations. [Figure 16] 10 is a flowchart showing the flow of a second estimation method according to the second embodiment. [Figure 17] FIG. 11 is a diagram illustrating a parameter set selection method in the second estimation method. [Figure 18] 20 is a flowchart showing the flow of a second identification method according to the sixth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] [Embodiment 1] A first embodiment of the present invention will be described with reference to Figs. 1 to 8. The flexural strength estimation system according to the first embodiment of the present invention is a system for estimating the flexural strength of a mold. Note that the "mold" in this embodiment means a sand mold made from molding sand. The mold according to this embodiment may be a mold made using a core, or a mold made without using a core.

[0013] <Mold making process> First, the process from molding sand to a mold will be briefly described with reference to Fig. 1. Fig. 1 is a diagram for explaining an outline of the mold making process in this embodiment. The mold is made through a sand pouring process C1, a kneading process C2, and a molding process C3.

[0014] The sand charging process C1 is a process of adding recycled sand to unused foundry sand, i.e., new sand. The sand lump after casting goes through a crushing process, a separation process, a pre-recycling cooling process, a sand regeneration process, and a post-regeneration cooling process to become reusable recycled sand. In this embodiment, as an example, the mixture of new sand and recycled sand is used as the foundry sand in the following processes.

[0015] The kneading step C2 is a step of adding additives to the molding sand obtained in the sand pouring step C1 and kneading the mixture. The additives are, for example, water glass, a hardener such as resin, and a surfactant. The molding step C3 is a step of filling a flask with the molding sand kneaded in the kneading step C2 to form a mold. In this embodiment, an upper mold corresponding to the upper part of the mold and a lower mold corresponding to the lower part of the mold are each molded.

[0016] The mold made in the molding process C3 is ready for pouring molten metal by going through the mold-stripping process and the mold-coating process in that order. The mold-stripping process is a process for separating the upper and lower molds made in the molding process C3 from the flask. The mold-coating process is a process for applying a mold wash to the product surfaces of the upper and lower molds removed in the mold-stripping process. When a mold is made using a core, the mold made in the molding process C3 may go through a frame-fitting process after the mold-stripping and mold-coating processes. The frame-fitting process is a process for assembling the mold by combining the upper and lower molds to which the mold wash has been applied in the mold-coating process.

[0017] Among the steps shown in Fig. 1, the sand-injecting step C1 is not an essential step. For example, if recycled sand is not used as the raw material for the mold, the sand-injecting step C1 may not be performed, and new sand may be injected in the mixing step C2. The molding step C3 may also be a non-heating step. For example, the molding step C3 may be a step in which the mixed sand is filled into a molding flask and then hardened without heating.

[0018] <System Overview> 2 is a schematic diagram of the estimation system S1. The estimation system S1 is a system for estimating the bending strength of a mold from a plurality of parameters related to the creation of the mold. The estimation system S1 may also estimate at least one of the cycle time for creating the mold and the cost required for creating the mold.

[0019] The parameters related to the creation of the mold are, for example, parameters that are measured or set in any one of the sand pouring step C1, the kneading step C2, and the molding step C3. The parameters include raw material parameters, kneading parameters, and molding parameters.

[0020] The raw material parameters are parameters related to the raw materials of the mold. For example, the raw material parameters may be at least one of the amount of molding sand, the amount of water, and the amount of additives added in the kneading step C2, and the amount of residual additives in the molding sand. Note that the "amount of additives added" refers to at least one of the amount of hardener added, such as the amount of water glass or the amount of resin added, and the amount of surfactant added. The raw material parameters may include one or more of these parameters for the amount of additives added.

[0021] The mixing parameters are parameters that are measured or set in the mixing process C2. The mixing parameters may be values ​​that are measured or set at the timing of the mixing process C2, and may be data that is not directly related to the mixing process C2. For example, the mixing parameters may be at least one of the mixing speed of the raw material, the mixing time, and the temperature of the mixed sand in the mixing process C2.

[0022] The molding parameters are parameters that are measured or set in the molding process C3. The molding parameters may be values ​​that are measured or set at the timing of the molding process C3, and may be data that is not directly related to the molding process C3. For example, the molding parameters may be at least one of the speed at which the mixed sand is pressed into the mold frame, the temperature of the mold frame, the amount of release agent applied to the mold frame, and the baking time of the mixed sand.

[0023] Hereinafter, the multiple parameters related to the creation of a mold will be simply referred to as "parameters." Furthermore, the multiple parameters will be collectively referred to as a "parameter set."

[0024] (Estimation device 1) The estimation system S1 includes at least an estimation device 1. The estimation device 1 estimates the flexural strength of a mold when the mold is created with a certain parameter set value. The estimation device 1 stores a flexural strength relational equation F1. The flexural strength relational equation F1 is a relational equation that indicates the relationship between the flexural strength of the mold and multiple parameters related to the mold.

[0025] By substituting the values ​​of each parameter in the parameter set into the flexural strength relational equation F1, the estimation device 1 can estimate, as a solution to the flexural strength relational equation F1, the flexural strength of the mold when it is created with the various parameter values ​​indicated by the input parameter set.

[0026] The estimation device 1 may estimate a cycle time for making a mold when the mold is made with a certain set of parameter values. The estimation device 1 may also estimate a cost for making a mold when the mold is made with a certain set of parameter values. When estimating the cycle time, the estimation device 1 stores a cycle time relational expression F2. When estimating the cost, the estimation device 1 stores a cost relational expression F3.

[0027] The cycle time relational equation F2 is a relational equation that indicates the relationship between at least some of the multiple parameters included in the parameter set and the cycle time for mold making. The cost relational equation F3 is a relational equation that indicates the relationship between at least some of the multiple parameters included in the parameter set and the cost required for mold making. In this embodiment, the cycle time relational equation F2 and the cost relational equation F3 are assumed to be created in advance and stored in the estimation device 1.

[0028] The cycle time relation F2 can be expressed, for example, by the following Equation 1.

[0029] (Formula 1) Cycle time = A parameter indicating the time required for a part of the process of mold creation, and the total value of the said time calculated from any of the parameters Here, the "parameter indicating the time required for a part of the process of mold creation" is, for example, the mixing time and the baking time of the mixed sand. Also, the "time calculated from any of the parameters" is, for example, the mixed sand charging time, which is the time required to charge the mixed sand into the molding flask. The units of each time on the right side of Equation 1 should be consistent. The mixed sand charging time may be calculated by multiplying the amount of mixed sand charged by the injection speed of the mixed sand. The amount of mixed sand charged may be the sum of the amount of foundry sand charged, the amount of water charged, and the amount of additives charged.

[0030] A parameter indicating the time required for the sand charging step C1 may be added to the right side of Equation 1. Furthermore, the molding sand charging time, the water charging time, and the additive charging time may be added to the right side of Equation 1. These three parameters indicate the time required to charge molding sand, water, and additives into the kneader, respectively. These three parameters may be calculated by multiplying the molding sand charging amount, water charging amount, and additive charging amount, which are included in the parameter set, by the molding sand charging speed, water charging speed, and additive charging speed, which are also included in the parameter set.

[0031] Moreover, the cost relational expression F3 can be expressed, for example, by the following expression 2.

[0032] (Formula 2) Cost = parameters related to the amount of mold material x price per unit of each material Here, parameters related to the amount of raw materials for the mold are, for example, the amount of casting sand, the amount of water, and the amount of additives. The units of amount are the same as the unit amounts of price. In addition, electricity charges for facility maintenance and operation of equipment in the mold making processes C1 to C3 may be added to the right side of Equation 2.

[0033] The estimation device 1 substitutes the values ​​of at least some of the parameters in the parameter set into the cycle time relational equation F2, thereby making it possible to estimate the cycle time when a mold is produced using the various parameter values ​​indicated by the input parameter set as a solution to the cycle time relational equation F2.

[0034] The estimation device 1 substitutes the values ​​of at least some of the parameters in the parameter set into the cost relational equation F3, thereby making it possible to estimate the cost of creating a mold using the various parameter values ​​indicated by the input parameter set as a solution to the cost relational equation F3.

[0035] In this embodiment, the estimation device 1 estimates the above-mentioned bending strength, cycle time, and cost. The estimation device 1 according to this embodiment can identify a parameter set that satisfies predetermined conditions for the bending strength, cycle time, and cost by comparing the estimated values ​​of these three types of values ​​with thresholds (a target value, a predetermined time, and a predetermined value, which will be described later).

[0036] The values ​​of the various parameters indicated by the parameter set may be prepared in response to an input operation by a user, or may be determined by the estimation device 1. In this embodiment, the values ​​of the various parameters are determined randomly by the estimation device 1.

[0037] When the estimation device 1 randomly determines the values ​​of each parameter, the range of possible values ​​of the random values ​​may be determined by an input operation by the user. When the estimation device 1 randomly determines the values ​​of various parameters, a huge number of parameter sets may be generated. Therefore, when the estimation device 1 randomly determines the values ​​of various parameters, the number of parameter sets to be generated may be determined in advance. The number of generated parameter sets may be freely changeable by the user.

[0038] The estimation device 1 selects, from among a specified number of parameter sets, a parameter set that will provide a flexural strength equal to or greater than the target value when a mold is created with parameters indicated by the parameter set, in accordance with a user's input operation. In the following description, the target value of the flexural strength will also be referred to simply as the target value. The parameter set that will provide a flexural strength equal to or greater than the target value will also be referred to as the "primary selection parameter set." The range that the target value can take is not particularly limited. The target value may be appropriately determined in accordance with the user's input operation.

[0039] The estimation device 1 further selects a parameter set whose cycle time is within a predetermined time from the first selected parameter set. Hereinafter, the parameter set further selected from the first selected parameter set is referred to as a "second selected parameter set." In this embodiment, the parameter set whose cycle time is within a predetermined time from among the first selected parameter sets is the second selected parameter set. Note that the specific value of the predetermined time is not particularly limited. Also, the above-mentioned predetermined time, i.e., the cycle time threshold value, may be appropriately determined according to an input operation by the user.

[0040] The estimation device 1 identifies a parameter set from the secondary selected parameter sets that results in a cost equal to or less than a predetermined value. Alternatively, the estimation device 1 may identify one parameter set from the secondary selected parameter sets that results in the lowest cost. That is, the estimation device 1 identifies a parameter set that has a flexural strength equal to or greater than a target value, a cycle time within a predetermined time, and a cost equal to or less than a predetermined value (or the lowest).

[0041] According to the above process, the estimation device 1 can identify a parameter set that has a flexural strength equal to or greater than a target value, a cycle time within a predetermined time, and a cost equal to or less than a predetermined value. Since the thresholds of the flexural strength, cycle time, and cost can be set by the user, the estimation device 1 can identify a parameter set (target parameter set) that satisfies the conditions desired by the user for the flexural strength, cycle time, and cost. The estimation device 1 may present this target parameter set to the user.

[0042] (Relationship identification device 2, data logger 3, and manufacturing device 4) The estimation system S1 may include a relational equation specifying device (specifying device) 2, a data logger 3, and a manufacturing device 4. The manufacturing device 4 is a device disposed in at least one of the sand charging process C1, the kneading process C2, and the molding process C3 shown in FIG. 1. In the estimation system S1, the manufacturing device 4 may be one or more. Each manufacturing device 4 is connected to the data logger 3, and transmits actual values ​​of various parameters at the time of creating the mold to the data logger 3. Hereinafter, the actual parameter values ​​at the time of creating the mold are also referred to as actual parameter values.

[0043] The data logger 3 is a device that collects actual parameter values ​​from the manufacturing apparatus 4. In the estimation system S1, the data logger 3 may be one or more. When there are multiple data loggers 3, each data logger 3 may collect actual parameter values ​​for each manufacturing apparatus 4 or for each mold production process. The data logger 3 is connected to the relational equation specifying device 2 and transmits the collected actual parameter values ​​to the relational equation specifying device 2. The data logger 3 may also receive the flexural strength (i.e., the flexural strength of the actual mold) from a device that measures the flexural strength of the produced mold (for example, a flexural strength meter 48 described later) and transmit the flexural strength to the relational equation specifying device 2.

[0044] The relational equation specifying device 2 is a device that creates the bending strength relational equation F1. The relational equation specifying device 2 according to this embodiment first constructs a trained model LM that has learned the relationship between a parameter set and bending strength by machine learning using the learning dataset DS as training data. Here, the training data is data consisting of a combination of a parameter set of actual parameter values ​​and the actual bending strength of a mold created with the actual parameter values.

[0045] The relational equation identifying device 2 according to the present embodiment then identifies a nonlinear relational equation that expresses the relationship between the input values ​​and the output values ​​of the constructed trained model LM using a predetermined algorithm. That is, the relational equation identifying device 2 visualizes the relationship between the input values ​​and the output values ​​of the trained model LM as a relational equation. Here, the relational equation identified by the relational equation identifying device 2 is the bending strength relational equation F1. The relational equation identifying device 2 transmits the identified bending strength relational equation F1 to the estimation device 1.

[0046] The type of the "predetermined algorithm" is not particularly limited. The "predetermined algorithm" is, for example, a nonlinear regression algorithm. Furthermore, the specific type of the nonlinear regression algorithm is not particularly limited. In this embodiment, as an example of the nonlinear regression algorithm, a genetic algorithm is used to visualize the input and output of the trained model LM into a nonlinear relational equation (flexural strength relational equation F1). However, the trained model LM may be visualized using a nonlinear regression algorithm other than the genetic algorithm, such as logistic regression.

[0047] (Main components of the estimation device 1) The configuration of the estimation device 1 will be described with reference to Fig. 3. Fig. 3 is a block diagram showing the configuration of the estimation device 1.

[0048] The estimation device 1 is realized using a general-purpose computer, and includes a processor 11, a primary memory 12, a secondary memory 13, an input / output interface (IF) 14, a communication interface 15, and a bus 16. The processor 11, the primary memory 12, the secondary memory 13, the input / output interface 14, and the communication interface 15 are connected to each other via the bus 16.

[0049] The secondary memory 13 stores a first estimation program P1, a bending strength relational expression F1, a cycle time relational expression F2, and a cost relational expression F3. The processor 11 expands the first estimation program P1, the bending strength relational expression F1, the cycle time relational expression F2, and the cost relational expression F3 stored in the secondary memory 13 onto the primary memory 12. The processor 11 then executes each step included in the first estimation method M1 according to the instructions included in the first estimation program P1 expanded onto the primary memory 12. The bending strength relational expression F1 expanded onto the primary memory 12 is used when the processor 11 executes a bending strength estimation step M11 (described later) of the first estimation method M1. The cycle time relational expression F2 expanded onto the primary memory 12 is used when the processor 11 executes a cycle time estimation step M13 (described later) of the first estimation method M1. The cost relational formula F3 expanded on the primary memory 12 is used when the processor 11 executes a cost estimation step M15 (described later) of the first estimation method M1. Note that the first estimation program P1 being stored in the secondary memory 13 means that the source code or an executable file obtained by compiling the source code is stored in the secondary memory 13.

[0050] Examples of devices that can be used as the processor 11 include a central processing unit (CPU), a graphic processing unit (GPU), a digital signal processor (DSP), a micro processing unit (MPU), a floating point number processing unit (FPU), a physics processing unit (PPU), a microcontroller, or a combination of these. The processor 11 is also called a "computing device."

[0051] Moreover, examples of devices that can be used as the primary memory 12 include, for example, a semiconductor RAM (Random Access Memory). The primary memory 12 is sometimes called a "main memory device." Examples of devices that can be used as the secondary memory 13 include, for example, a flash memory, a hard disk drive (HDD), a solid state drive (SSD), an optical disk drive (ODD), a floppy (registered trademark) disk drive (FDD), or a combination of these. The secondary memory 13 is sometimes called an "auxiliary memory device." The secondary memory 13 may be built in the estimation device 1, or may be built in another computer (for example, a computer constituting a cloud server) connected to the estimation device 1 via the input / output interface 14 or the communication interface 15. In this embodiment, the storage in the estimation device 1 is realized by two memories (the primary memory 12 and the secondary memory 13), but is not limited to this. That is, the storage in the estimation device 1 may be realized by one memory. In this case, for example, one storage area of ​​the memory may be used as the primary memory 12 and another storage area of ​​the memory may be used as the secondary memory 13.

[0052] An input device and / or an output device are connected to the input / output interface 14. Examples of the input / output interface 14 include interfaces such as Universal Serial Bus (USB), Advanced Technology Attachment (ATA), Small Computer System Interface (SCSI), and Peripheral Component Interconnect (PCI). Examples of input devices connected to the input / output interface 14 include a keyboard, a mouse, a touchpad, a microphone, or a combination of these. Data acquired from a user in the first estimation method M1 is input to the estimation device 1 via these input devices and stored in the primary memory 12. Examples of output devices connected to the input / output interface 14 include a display, a projector, a printer, a speaker, a headphone, or a combination of these. Information provided to a user in the first estimation method M1 is output from the estimation device 1 via these output devices. The estimation device 1 may incorporate a keyboard functioning as an input device and a display functioning as an output device, as in a laptop computer. Alternatively, the estimation device 1 may incorporate a touch panel functioning as both an input device and an output device, as in a tablet computer.

[0053] The communication interface 15 is connected to other computers via a network in a wired or wireless manner. Examples of the communication interface 15 include interfaces such as Ethernet (registered trademark) and Wi-Fi (registered trademark). Available networks include a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), a global area network (GAN), or an internetwork including these networks. The internetwork may be an intranet, an extranet, or the Internet. Data (e.g., bending strength relational formula F1) acquired by the estimation device 1 from another computer (e.g., the relational formula identification device 2) in the first estimation method M1 and data provided by the estimation device 1 to another computer in the first estimation method M1 are transmitted and received via these networks.

[0054] In this embodiment, a configuration in which the first estimation method M1 is executed using a single processor (processor 11) is adopted, but the present invention is not limited to this. That is, a configuration in which the first estimation method M1 is executed using multiple processors may be adopted. In this case, the multiple processors that execute the first estimation method M1 in cooperation with each other may be provided in a single computer and configured to be able to communicate with each other via a bus, or may be provided in a distributed manner in multiple computers and configured to be able to communicate with each other via a network. As an example, a mode in which a processor built in a computer that constitutes a cloud server and a processor built in a computer owned by a user of the cloud server execute the first estimation method M1 in cooperation with each other may be considered.

[0055] In the present embodiment, the bending strength relational expression F1, the cycle time relational expression F2, and the cost relational expression F3 are stored in a memory (secondary memory 13) built in the same computer as the processor (processor 11) that executes the first estimation method M1, but the present invention is not limited to this. That is, a configuration may be adopted in which the bending strength relational expression F1, the cycle time relational expression F2, and the cost relational expression F3 are stored in a memory built in a computer different from the processor that executes the first estimation method M1. In this case, the computer that has the memory that stores the bending strength relational expression F1, the cycle time relational expression F2, and the cost relational expression F3 is configured to be able to communicate with the computer that has the processor that executes the first estimation method M1 via a network. As an example, a mode in which the bending strength relational expression F1, the cycle time relational expression F2, and the cost relational expression F3 are stored in a memory built in a computer that constitutes a cloud server, and a processor built in a computer owned by a user of the cloud server executes the first estimation method M1, etc., can be considered.

[0056] In addition, in this embodiment, a configuration is adopted in which the bending strength relational expression F1 is stored in a single memory (secondary memory 13), but the present invention is not limited to this. That is, a configuration in which the bending strength relational expression F1 is distributed and stored in multiple memories may be adopted. In this case, the multiple memories that store the trained model LM may be provided in a single computer (which may or may not be a computer with a built-in processor that executes the first estimation method M1), or may be distributed and provided in multiple computers (which may or may not include a computer with a built-in processor that executes the first estimation method M1). As an example, a configuration in which the trained model LM is distributed and stored in memories built into each of the multiple computers that constitute the cloud server can be considered.

[0057] <<Flow of the first estimation method M1>> The flow of the first estimation method M1 will be described with reference to Fig. 4 and Fig. 5. Fig. 4 is a flowchart showing the flow of the first estimation method M1. Fig. 5 is a diagram showing a schematic diagram of a parameter set selection method in the first estimation method M1. The first estimation method M1 includes a parameter set creation step M10, a bending strength estimation step (first estimation step) M11, a bending strength selection step M12, a cycle time estimation step (second estimation step) M13, a cycle time selection step M14, a cost estimation step (third estimation step) M15, and a cost selection step M16.

[0058] In the parameter set creation step M10, the processor 11 creates a parameter set to be subjected to the subsequent processing. The processor 11 creates one parameter set, for example, by randomly determining each parameter within a value range specified by the user. After creating a predetermined number of parameter sets, the processor 11 executes the next step M11, which is to estimate the bending strength.

[0059] The flexural strength estimation step M11 is a step in which the processor 11 estimates the flexural strength when a mold is created with each parameter set, using the flexural strength relational expression F1. Specifically, as shown in Fig. 5, the processor 11 inputs each parameter set into the flexural strength relational expression F1 to obtain an estimated value of the flexural strength as a solution to the flexural strength relational expression F1. When the processor 11 has executed the estimation of the flexural strength for all the parameter sets created in the parameter set creation step M10, the processor 11 executes the flexural strength selection step M12.

[0060] The flexural strength selection step M12 is a step of selecting a parameter set whose flexural strength specified in the flexural strength estimation step M11 is equal to or greater than the target value. The processor 11 selects a parameter set whose flexural strength is equal to or greater than the target value from the parameter sets created in the parameter set creation step M10, and sets this as a first selected parameter set. The processor 11 executes the next process, the cycle time estimation step M13, for the first selected parameter set.

[0061] The cycle time estimation step M13 is a step in which the processor 11 uses the cycle time relational expression F2 to estimate the cycle time when a mold is made with parameters indicated by each of the first selection parameter sets. Specifically, as shown in FIG. 5, the processor 11 inputs the parameter values ​​indicated by the first selection parameter sets into the cycle time relational expression F2, respectively, to obtain an estimated value of the cycle time as a solution of the cycle time relational expression F2. Note that not all parameters of the parameter sets are necessarily included as variables in the cycle time relational expression F2. That is, the processor 11 may input at least a part of the parameters indicated by the first selection parameter set into the cycle time relational expression F2 to estimate the cycle time when a mold is made with the first selection parameter set. When the cycle time estimation is performed for all the first selection parameter sets, the processor 11 executes the cycle time selection step M14.

[0062] The cycle time selection step M14 is a step of selecting a parameter set whose cycle time specified in the cycle time estimation step M13 is within a predetermined time. The processor 11 selects a parameter set whose cycle time is within a predetermined time from the first selection parameter set, and sets this as a second selection parameter set. The processor 11 executes the next process, the cost estimation step M15, for the second selection parameter set.

[0063] The cost estimation step M15 is a step in which the processor 11 estimates the cost of creating a mold with each parameter set using the cost relational expression F3. Specifically, as shown in FIG. 5, the processor 11 inputs the parameter values ​​indicated by the secondary selection parameter set into the cost relational expression F3, respectively, to obtain an estimated value of the cost as a solution to the cost relational expression F3. Note that not all parameters of the parameter set are necessarily included as variables in the cost relational expression F3. That is, the processor 11 inputs at least a part of the parameters indicated by the secondary selection parameter set into the cost relational expression F3 to estimate the cost of creating a mold with the secondary selection parameter set. When the cost estimation is performed for all the secondary selection parameter sets, the processor 11 executes the cost selection step.

[0064] The cost selection step M16 is a step of selecting a parameter set whose cost specified in the cost estimation step M15 is equal to or less than a predetermined value. The processor 11 selects a parameter set whose cost is equal to or less than a predetermined value from the secondary selection parameter set, and determines this as a target parameter set. Note that in the cost selection step M16, the processor 11 may select only one parameter set whose cost is the lowest from the secondary selection parameter set, and set this as the target parameter set.

[0065] According to the above process, the estimation device 1 can identify a parameter set that can obtain a bending strength that satisfies a target value. Also, according to the above process, the estimation device 1 can identify a parameter set that provides a cycle time within a predetermined time. Also, according to the above process, the estimation device 1 can identify a parameter set that provides a cost equal to or less than a predetermined value.

[0066] More specifically, the estimation device 1 estimates the flexural strength, cycle time, and cost in the order of the above-mentioned first estimation method M1, and selects a parameter set after each estimation. This allows the estimation device 1 to identify a parameter set that provides a target flexural strength, a cycle time within a predetermined time, and a cost equal to or less than a predetermined value. Therefore, the user can identify the values ​​of various parameters that satisfy the desired flexural strength, cycle time, and cost conditions without conducting experiments.

[0067] The first estimation method M1 may include a first output step of outputting the flexural strength estimated in the flexural strength estimation step M11. The first estimation method M1 may also include a second output step of outputting the cycle time estimated in the cycle time estimation step M13. The first estimation method M1 may also include a third output step of outputting the cost estimated in the cost estimation step M15. In the first to third output steps, the processor 11 outputs the estimated flexural strength, cycle time, or cost to a display. This allows the user of the estimation system S1 to know the flexural strength, cycle time, or cost when a mold is created with parameters indicated by a certain parameter set. The first estimation method M1 may also include a condition setting step of setting the values ​​of various parameters indicated by the target parameter set as setting values ​​of the manufacturing device 4.

[0068] <<Configuration of main parts of the relational expression identifying device 2>> The configuration of the relational equation specifying device 2 will be described with reference to Fig. 6. Fig. 6 is a block diagram showing the configuration of the relational equation specifying device 2.

[0069] The relational equation identifying device 2 is realized by using a general-purpose computer, and includes a processor 21, a primary memory 22, a secondary memory 23, an input / output interface 24, a communication interface 25, and a bus 26. The processor 21, the primary memory 22, the secondary memory 23, the input / output interface 24, and the communication interface 25 are connected to each other via the bus 26.

[0070] The secondary memory 23 stores the first relational expression specifying program P2 and the learning data set DS. The learning data set DS is a collection of teacher data DS1, DS2, .... The processor 21 expands the first relational expression specifying program P2 stored in the secondary memory 23 onto the primary memory 22. Then, the processor 21 executes each step included in the first specifying method M2 according to the instructions included in the first relational expression specifying program P2 expanded onto the primary memory 22. The learning data set DS stored in the secondary memory 23 is constructed in a learning data set construction step M21 (described later) of the first specifying method M2, and is used in a trained model construction step M22 (described later) of the first specifying method M2. In addition, the trained model LM constructed in the trained model construction step M22 of the first specifying method M2 is also stored in the secondary memory 23. Note that the first relational expression specifying program P2 being stored in the secondary memory 23 refers to the source code, or an executable file obtained by compiling the source code, being stored in the secondary memory 23. In addition, storing the learned model LM in the secondary memory 23 means that parameters that define the learned model LM are stored in the secondary memory 23.

[0071] Examples of devices that can be used as the processor 21 include a central processing unit (CPU), a graphic processing unit (GPU), a digital signal processor (DSP), a micro processing unit (MPU), a floating point number processing unit (FPU), a physics processing unit (PPU), a microcontroller, or a combination of these. The processor 21 is also called a "computing device."

[0072] Moreover, examples of devices that can be used as the primary memory 22 include, for example, semiconductor RAM (Random Access Memory). The primary memory 22 is sometimes called a "main memory". Examples of devices that can be used as the secondary memory 23 include, for example, flash memory, HDD (Hard Disk Drive), SSD (Solid State Drive), ODD (Optical Disk Drive), FDD (Floppy (registered trademark) Disk Drive), or a combination of these. The secondary memory 23 is sometimes called an "auxiliary memory". The secondary memory 23 may be built in the relational equation identifying device 2, or may be built in another computer (for example, a computer constituting a cloud server) connected to the relational equation identifying device 2 via the input / output interface 24 or the communication interface 25. In this embodiment, the storage in the relational equation identifying device 2 is realized by two memories (the primary memory 22 and the secondary memory 23), but is not limited to this. That is, the storage in the relational equation identifying device 2 may be realized by one memory. In this case, for example, one storage area of ​​the memory may be used as the primary memory 22 and another storage area of ​​the memory may be used as the secondary memory 23.

[0073] An input device and / or an output device are connected to the input / output interface 24. Examples of the input / output interface 24 include interfaces such as Universal Serial Bus (USB), Advanced Technology Attachment (ATA), Small Computer System Interface (SCSI), and Peripheral Component Interconnect (PCI). Examples of input devices connected to the input / output interface 24 include a data logger 3. Examples of input devices connected to the input / output interface 24 include a keyboard, a mouse, a touchpad, a microphone, or a combination of these. Data acquired from a user in the first identification method M2 is input to the relational equation identifying device 2 via these input devices and stored in the primary memory 22. Examples of output devices connected to the input / output interface 24 include a display, a projector, a printer, a speaker, a headphone, or a combination of these. Information provided to a user in the first identification method M2 is output from the relational equation identifying device 2 via these output devices. The relational equation identifying device 2 may each have a built-in keyboard functioning as an input device and a built-in display functioning as an output device, like a laptop computer. Alternatively, the relational equation specifying device 2 may have a built-in touch panel that functions as both an input device and an output device, like a tablet computer.

[0074] The communication interface 25 is connected to other computers via a network in a wired or wireless manner. Examples of the communication interface 25 include interfaces such as Ethernet (registered trademark) and Wi-Fi (registered trademark). Available networks include a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), a global area network (GAN), or an internetwork including these networks. The internetwork may be an intranet, an extranet, or the Internet. Data (e.g., the trained model LM) provided by the relational equation identification device 2 to other computers (e.g., the estimation device 1) is transmitted and received via these networks.

[0075] In this embodiment, a configuration is adopted in which the first identification method M2 is executed using a single processor (processor 21), but the present invention is not limited to this. That is, a configuration may be adopted in which the first identification method M2 is executed using multiple processors. In this case, the multiple processors that execute the first identification method M2 in cooperation with each other may be provided in a single computer and configured to be able to communicate with each other via a bus, or may be provided in a distributed manner in multiple computers and configured to be able to communicate with each other via a network. As an example, a processor built in a computer that constitutes a cloud server and a processor built in a computer owned by a user of the cloud server may execute the first identification method M2 in cooperation with each other.

[0076] In addition, in the present embodiment, a configuration is adopted in which the training dataset DS is stored in a memory (secondary memory 23) built into the same computer as the processor (processor 21) that executes the first identification method M2, but the present invention is not limited to this. That is, a configuration may be adopted in which the training dataset DS is stored in a memory built into a computer different from the processor that executes the first identification method M2. In this case, the computer with the built-in memory for storing the training dataset DS is configured to be able to communicate with the computer with the built-in processor that executes the first identification method M2 via a network. As an example, a mode in which the training dataset DS is stored in a memory built into a computer constituting a cloud server, and a processor built into a computer owned by a user of the cloud server executes the first identification method M2, etc., can be considered.

[0077] In addition, in the present embodiment, a configuration in which the training dataset DS is stored in a single memory (secondary memory 23) is adopted, but the present invention is not limited to this. That is, a configuration in which the training dataset DS is distributed and stored in multiple memories may be adopted. In this case, the multiple memories that store the training dataset DS may be provided in a single computer (which may or may not be a computer with a built-in processor that executes the first identification method M2), or may be distributed and provided in multiple computers (which may or may not include a computer with a built-in processor that executes the first identification method M2). As an example, a configuration in which the training dataset DS is distributed and stored in memories built into each of the multiple computers that constitute the cloud server is conceivable.

[0078] <Example of getting actual parameter values> FIG. 7 is a diagram showing the types and acquisition timings of actual parameter values ​​acquired by the relational equation identifying device 2 from each manufacturing device 4 via the data logger 3 in the mold making process. The mold production line shown in FIG. 7 includes a molding sand property measuring device 40 and a molding machine 47 as the manufacturing devices 4. The molding machine 47 includes a molding sand injection device 41, an additive injection device 42, a water injection device 43, a kneader 44, a mixed sand property measuring device 45, and a mold 46. A flexural strength meter 48 is attached to the molding machine 47. The molding sand property measuring device 40, the molding sand injection device 41, the additive injection device 42, the water injection device 43, the kneader 44, and the mixed sand property measuring device 45 can be said to be the manufacturing devices 4 in the kneading process C2 shown in FIG. 1. The mold 46 is the manufacturing device 4 in the molding process C3 shown in FIG. 1.

[0079] The molding sand property measuring device 40 is a device for measuring the sand properties of molding sand (hereinafter also referred to as "molding sand properties"). In this embodiment, the molding sand property measuring device 40 measures the amount of residual additives in the molding sand as the molding sand properties. Here, the amount of residual additives in the molding sand refers to, for example, the weight or volume of a hardener contained in a unit weight or unit volume of molding sand. The molding sand property measuring device 40 may also measure at least one of the temperature and moisture content of the molding sand as the molding sand properties. Here, the moisture content of the molding sand refers to, for example, the weight or volume of moisture contained in a unit weight or unit volume of molding sand. The molding sand property measuring device 40 provides the measured molding sand properties to the data logger 3.

[0080] The foundry sand charging device 41 is a device for charging a preset amount of foundry sand (hereinafter also referred to as the "foundry sand charging amount") into the kneader 44. Here, the foundry sand charging amount refers to, for example, the weight or volume of foundry sand charged into the kneader 44 per unit time (in the case of continuous processing) or per processing (in the case of batch processing). The foundry sand charging device 41 provides the set foundry sand charging amount to the data logger 3.

[0081] The additive feed device 42 is a device for feeding a preset amount of additives (hereinafter also referred to as "additive feed amount") to the kneader 44. The additive feed amount refers to, for example, the weight or volume of additives fed to the kneader 44 per unit time (in the case of continuous processing) or per processing (in the case of batch processing). In the additive feed device 42 according to this embodiment, the curing agent feed amount and the surfactant feed amount are preset as the additive feed amounts. The additive feed device 42 also provides the set curing agent feed amount and surfactant feed amount to the data logger 3.

[0082] The water injection device 43 is a device for injecting a preset amount of water (hereinafter also referred to as the “water input amount”) into the kneader 44. Here, the water input amount refers to, for example, the weight or volume of water input into the kneader 44 per unit time (in the case of continuous processing) or per processing (in the case of batch processing). The water injection device 43 provides the set water input amount to the data logger 3.

[0083] The kneader 44 is a device that kneads additives and water into foundry sand according to preset kneading conditions. Examples of the kneading conditions include the rotation speed of the kneading blade (in the case of continuous processing and batch processing) and the rotation time (in the case of batch processing). The rotation speed of the kneading blade refers to, for example, the number of times the kneading blade rotates per unit time. Furthermore, the rotation time of the kneading blade refers to, for example, the time it takes for the kneading blade to rotate in one processing. In this embodiment, the rotation speed of the kneading blade is considered to be the "kneading speed". Furthermore, in this embodiment, the rotation time of the kneading blade is considered to be the "kneading time". The kneader 44 provides the set kneading conditions to the data logger 3.

[0084] The mixed sand property measuring device 45 is a device for measuring the sand properties of the mixed sand (hereinafter, also referred to as "mixed sand properties"). In this embodiment, the mixed sand property measuring device 45 measures the temperature of the mixed sand as the mixed sand property. The mixed sand property measuring device 45 may also measure at least one of the temperature, moisture content, compactability (CB value), and air permeability of the mixed sand as the mixed sand property. The moisture content of the mixed sand refers to, for example, the weight or volume of the moisture contained in the mixed sand per unit weight or unit volume. The compactability of the mixed sand is a value that indicates the moisture state of the particle surface layer of the mixed sand, and is a value used as a green sand evaluation index. The air permeability of the mixed sand refers to, for example, the ease with which gases (hydrogen, oxygen, nitrogen, carbon monoxide, carbon dioxide, hydrocarbons, etc.) that escape from inside the mold can pass through. The mixed sand property measuring device 45 provides the measured mixed sand properties to the data logger 3.

[0085] The mixed sand mixed in the mixer 44 is poured into a mold 46. An injection speed sensor is attached to the mixer 44, and the injection speed of the mixed sand into the mold 46 is measured. The injection speed sensor provides the measured injection speed to the data logger 3.

[0086] The mold 46 is a formwork for molding a casting mold. In this embodiment, the mold is the formwork for the casting mold, but in the present invention, the material of the mold work is not limited to metal. Various molding sensors are attached to the mold 46. The molding sensor is one or more sensors that measure various parameters related to the molding process. In this embodiment, the molding sensors measure the baking time of the mixed sand, the temperature of the mold 46, and the amount of release agent applied to the mold 46. The molding sensors provide the various measured parameters to the data logger 3.

[0087] In addition, a flexural strength meter 48 is attached to the mold 46. The flexural strength meter 48 is a device that measures the flexural strength of a mold completed through a molding process. The flexural strength meter 48 provides the measured flexural strength to the data logger 3.

[0088] The data logger 3 collects parameters provided from the above-mentioned manufacturing apparatus 4 and provides them to the relational equation specifying apparatus 2. That is, as shown in Fig. 7, the data logger 3 provides raw material parameters, kneading parameters, and molding parameters to the relational equation specifying apparatus 2. The data logger 3 also acquires the actual flexural strength of a mold actually made with these actual parameter values ​​from a flexural strength meter 48. The data logger 3 provides the acquired actual flexural strength to the relational equation specifying apparatus 2.

[0089] The relational expression specifying device 2 may record the parameters acquired from the data logger 3 as one parameter set each time a mold is created, or may create and record a plurality of parameter sets at a time at a predetermined timing. When kneading is repeated k times, k parameter sets are accumulated in the relational expression specifying device 2 using the parameters acquired from the first to kth times. Thereafter, the relational expression specifying device 2 performs a first specification method M2 including a first relational expression specifying step (specifying step) M24. The first relational expression specifying step M24 is a step of specifying a flexural strength relational expression F1, which is a nonlinear relational expression that expresses the relationship between various parameters included in the parameter set and the flexural strength, by referring to the accumulated parameter set. The flexural strength relational expression F1 is expressed, for example, in the form of z=f(x1, x2, ..., xn). Here, z indicates the flexural strength. Also, x1, x2, ..., xn respectively indicate the parameters included in the parameter set. The configuration of the relational expression specifying device 2 and the flow of the first specification method M2 will be described later with reference to the drawings.

[0090] <<Flow of the first identification method M2>> The flow of the first specifying method M2 will be described with reference to Fig. 8. Fig. 8 is a flowchart showing the flow of the first specifying method M2.

[0091] The first identification method M2 includes a learning dataset construction step M21, a trained model construction step (first construction step) M22, an estimated dataset creation step M23, and a first relational equation identification step M24.

[0092] The learning data set construction step M21 is a step in which the processor 21 constructs a learning data set DS which is a collection of teacher data DS1, DS2, . . .

[0093] Each training data DSi (i=1, 2, ...) is training data consisting of a combination of the actual parameter values ​​(i.e., raw material parameters, kneading parameters, and molding parameters) of the parameter set collected via the data logger 3 and the actual bending strength value. The types of parameters included in the parameter set of the training data DSi are the same as the parameters included in the parameter set input to the trained model LM. The processor 21 builds the training dataset DS by repeatedly creating training data.

[0094] The trained model construction step M22 is a step in which the processor 21 constructs the trained model LM. In the trained model construction step M22, the processor 21 constructs the trained model LM by supervised learning using the training dataset DS. Then, the processor 21 stores the constructed trained model LM in the secondary memory 23. When the construction and storage of the trained model LM are completed, the processor 21 executes the next estimated dataset creation step M23.

[0095] The estimated data set creation step M23 is a step in which the processor 32 creates an estimated data set used when identifying a relational expression using a genetic algorithm in the first relational expression identification step M24 described later. The estimated data set is a collection of multiple estimated data EDS. The processor 21 inputs random or specified parameter set values ​​to the learned model LM constructed in the learned model construction step M22, thereby obtaining the flexural strength as an output from the learned model LM. The processor 21 sets data in which the obtained flexural strength (i.e., the estimated flexural strength) is associated with the values ​​of various parameters input to the learned model LM as one piece of estimated data EDS. The processor 21 stores the created estimated data EDS in the secondary memory 23. The creation and storage of such estimated data EDS is repeated one or more times to create an estimated data set.

[0096] The processor 21 creates an estimated data set using the learned model LM for the amount used in the first relational expression identification step M24. When the estimated data set necessary for executing the first relational expression identification step M24 is prepared, the processor 21 executes the next first relational expression identification step M24.

[0097] The first relational equation specifying step M24 is a step of specifying a relational equation showing the relationship between the input value and the output value of the learned model LM using a nonlinear regression algorithm.

[0098] In the first relational equation specifying step M24 according to this specific example, the relational equation z=f(x1, x2, ..., xn) is specified using a genetic algorithm as a type of nonlinear regression algorithm. Here, the genetic algorithm refers to an algorithm in which a plurality of individuals i each expressing a solution candidate by genes are prepared, and an individual i with high fitness Di is preferentially selected, and a solution is searched for by repeating operations such as crossover and mutation. In this embodiment, the individual i is a nonlinear relational equation represented by a tree structure, and the operators and arguments included in the relational equation are represented by the nodes of the tree. The fitness Di is given by a fitness function.

[0099] The processor 21 executes the first relational equation specifying step M24, for example, by using a predetermined program module. Hereinafter, the predetermined program module is simply referred to as an A module. The A module is a program module in which processing related to a genetic algorithm is described. The A module may be a publicly known library. The processor 11 performs the following processing by executing the program described in the A module.

[0100] First, the processor 11 creates a random equation that represents the relationship between known independent variables and a dependent variable target that depends on the independent variables in order to predict new data. The processor 11 creates a plurality of such random equations to create a population of random equations. This population is the first generation population to be subjected to genetic manipulation.

[0101] Next, the processor 11 selects the most suitable individual from the population of random formulas. That is, the random formula that best represents the relationship between the independent variables and the dependent variables in the population is selected as the optimal individual. Next, the processor 11 evolves the population to generate the next generation population. At this time, the optimal individual may be added to the next generation population as it is without evolving it. By repeating the above operations, a random formula that appropriately represents the relationship between the independent variables and the dependent variables can be identified as the relational formula between the independent variables and the dependent variables, like survival of the fittest in living organisms.

[0102] In this specific example, a module that executes genetic programming is used as the A module. Genetic programming is an extension of a genetic algorithm, and uses a tree structure as an expression of a gene type. Note that the flow of the first relational expression specification step M24 shown in FIG. 10 is an example, and the method of specifying a relational expression using the genetic algorithm GA is not limited to the method shown in FIG. 10. Various other methods may be adopted as the method of specifying a relational expression using the genetic algorithm GA.

[0103] (Specific example of the first relational expression determination step M24) A specific example of the first relational equation specifying step M24 will be described with reference to Fig. 9 to Fig. 15. Fig. 9 is a diagram illustrating a genetic algorithm GA. Fig. 10 is a flowchart showing the process of the first relational equation specifying step M24 executed by the processor 21. In the example of Fig. 9, the genetic algorithm GA includes a first generation G1 to a fourth generation G4.

[0104] In step M121, the processor 21 acquires an estimated data set. In this operation example, the processor 21 acquires the estimated data set by reading the estimated data set stored in the secondary memory 23.

[0105] In step M122, the processor 21 acquires parameters (hereinafter referred to as "individual parameters") used in the genetic algorithm GA. The individual parameters include, for example, the number of individuals to be generated N, the tournament size Nt, the crossover probability Pc, the mutation probability Pms, the number of evolutionary generations Ng, the operator Oj used in the syntax tree, the maximum depth d of the syntax tree, and the event occurrence probabilities Pk1 to Pk5. The value of each individual parameter is input to the relational expression specifying device 2 by, for example, the user.

[0106] The number of generated individuals N represents the number of individuals i to be included in the set. The tournament size Nt is the number of individuals i to be randomly selected from the current generation set. The mutation probability Pms is the probability that a gene will mutate. Operators Oi used in the syntax tree are, for example, Max, Min, sqrt (root), log (natural logarithm), +, -, ×, ÷, sin (radians), cos (radians), tan (radians), abs, neg, and inv. Max is an operator that selects the maximum value. Min is an operator that selects the minimum value. neg is an operator that makes the sign negative. inv is an operator that sets arguments close to zero to 0.

[0107] Event occurrence probabilities Pk1 to Pk5 are probabilities that operations m1 to m5 are selected as operations for evolving the next-generation set. Processor 21 evolves the next-generation set by one of operations m1 to m5. The sum of event occurrence probabilities Pk1 to Pk5 is 1. As an example, the values ​​of event occurrence probabilities Pk1, Pk2, Pk3, Pk4, and Pk5 are "0.1", "0.2", "0.3", "0.4", and "0.1", respectively. Operations m1 to m5 will be described later with reference to different drawings.

[0108] In step M123, the processor 21 randomly generates N individuals i based on specified individual parameters (the operator Oi used in the syntax tree, the maximum depth d of the syntax tree, etc.), and generates a set of N individuals i that will become the initial current generation. In this embodiment, each of the N individuals i is a randomly generated relational expression z=f(x1, x2, ..., xn).

[0109] In step M124, the processor 21 calculates the fitness Di of each individual i included in the set of the current generation. The fitness Di is given by a fitness function. In this embodiment, the fitness Di may be calculated based on the degree of approximation between the parameter value or the flexural strength indicated by each estimated data and the other value when the parameter value or the flexural strength indicated by each estimated data is substituted into the random formula of the individual i, for example.

[0110] In step M125, processor 21 randomly selects individuals i of tournament size Nt from the set of the current generation, selects the individual i with the highest fitness Di among them, and adds it to the set of the next generation. The individual i selected in step M125, i.e., the individual i added to the set of the next generation, is also called the "winner tree."

[0111] Processor 21 repeats the process of step M125 until the number of individuals in the next generation becomes N, the same as the current generation, that is, while the number of individuals in the next generation has not reached N (step M126; NO). When the number of individuals in the next generation reaches N (step M126; YES), processor 21 executes the process of step M127.

[0112] In step M127, the processor 21 executes a process of evolving the next generation set. Details of step M127 will be described later with reference to another drawing.

[0113] In step M130, the processor 21 overwrites the next-generation set to the current-generation set. In step M131, the processor 21 determines whether the number of evolutionary generations Ng has been reached. If the number of evolutionary generations Ng has not been reached (step M131; NO), the processor 21 returns to the processing of step M124. On the other hand, if the number of evolutionary generations Ng has been reached (step M131; YES), the processor 21 proceeds to the processing of step M132.

[0114] In step M132, the processor 21 identifies an individual i having the highest fitness Di from among the individuals i included in the set of the current generation. Through the above processing, the processor 21 identifies a nonlinear relational expression that represents the relationship between the multiple parameters included in the estimated data EDS and the flexural strength.

[0115] FIG. 11 is a flow chart illustrating the flow of step M127 executed by the processor 21. In step M201, the processor 21 selects one of the operations m1 to m5 based on the event occurrence probabilities Pk1 to Pk5 set by the user. If the operation m1 is selected (step M201; "operation m1"), the processor 21 proceeds to the process of step M202. If the operation m2 is selected (step M201; "operation m2"), the processor 21 proceeds to the process of step M211. If the operation m3 is selected (step M201; "operation m3"), the processor 21 proceeds to the process of step M221. If the operation m4 is selected (step M201; "operation m4"), the processor 21 proceeds to the process of step M231. If the operation m5 is selected (step M201; "operation m5"), the processor 21 ends the process.

[0116] Operation m1 is crossover. Crossover is a method of mixing genetic material between individuals. In the case of crossover, in step M202, processor 21 randomly selects subtrees included in each winning tree for the winning trees included in the next generation set.

[0117] In step M203, the processor 21 generates a next-generation set for donors. The processing content of step M203 is the same as the processing content of steps M123 to M125 in FIG. 11. That is, the processor 21 first randomly generates N individuals i based on individual parameters designated by the user, and generates a set of N individuals i (hereinafter referred to as a "donor set"). Next, the processor 21 calculates the fitness Di of each of the individuals i included in the donor set. Next, the processor 21 randomly extracts individuals i in the number of tournament size Nt from the donor set, selects the individual i with the highest fitness Di among them, and adds it to the donor set of the next generation. The individual i selected by this processing is also referred to as a "donor tree". The processor 21 repeats the donor tree selection processing until the number of donor trees of the next generation reaches N.

[0118] In step M204, the processor 21 randomly selects a subtree included in the donor tree (hereinafter referred to as a "donor subtree").

[0119] In step M205, the processor 21 replaces the subtrees in the winning tree. In this embodiment, the processor 21 removes the subtree selected in step M202 from the winning tree, and transplants the donor subtree selected in step M203 into the location where the subtree was. That is, the processor 21 replaces the subtree included in the winning tree with the donor subtree. The winning tree with the replaced subtree becomes the descendant (individual) of the next generation.

[0120] Operation m2 is an operation for mutating a subtree. By mutating a subtree, extinct functions and operators can be reintroduced into the population to maintain diversity. In this case, in step M211, processor 21 randomly selects a subtree to be included in the winning tree.

[0121] In step M212, processor 21 randomly generates a subtree. In step M213, processor 21 replaces the subtrees in the winning tree. In this embodiment, processor 21 removes the subtree selected in step M202 from the winning tree, and transplants the subtree generated in step M212 into the location of the removed subtree. That is, processor 21 replaces the subtree included in the winning tree with the subtree generated in step M212. The winning tree with the replaced subtree becomes the next generation descendant (individual).

[0122] Operation m3 is hoist mutation. Hoist mutation is a mutation operation that fights tree bloat. In step M221, the processor randomly selects a subtree to be included in the winning tree. In step M222, the processor randomly selects a subtree to be included in the subtree selected in step M221.

[0123] In step M223, processor 21 rolls up the subtree selected in step M222 to the position of the original subtree (the subtree selected in step M221). The winning tree into which this subtree has been rolled up becomes the next generation descendant (individual).

[0124] Operation m4 is point mutation. Point mutation is an operation for reintroducing extinct relations and operators into a population in order to maintain diversity. In step M231, processor 21 randomly selects a node of the winning tree. In step M232, processor 21 replaces the node selected in step M231 with another node. As a result, the relational expression represented by the winning tree is replaced with another relational expression that requires the same number of arguments as the original node. The winning tree obtained by the replacement becomes the descendant (individual) of the next generation.

[0125] Operation m5 is regeneration, where the winning tree is replicated and included in the next generation without modification.

[0126] 12 to 15 are diagrams illustrating the contents of operations performed on the winning tree. Fig. 12 is a diagram illustrating the contents of operation m1 (crossover). In the example of Fig. 12, the subtree tr111 of the winning tree tr11 is replaced with the subtree tr121 of the donor tree tr12 to become the winning tree tr13. The winning tree tr13 becomes the descendant (individual) of the next generation.

[0127] Figure 13 is a diagram illustrating the contents of operation m2 (subtree mutation). In the example of Figure 13, subtree tr111 of winner tree tr11 is replaced with subtree tr22 to become winner tree tr23. Winner tree tr23 becomes a descendant (individual) of the next generation.

[0128] Figure 14 is a diagram illustrating the contents of operation m3 (hoist mutation). In the example of Figure 14, subtree tr1121 of winner tree tr11 is rolled up to the position of subtree tr112 to become winner tree tr31. Winner tree tr31 becomes a descendant (individual) of the next generation.

[0129] Fig. 15 is a diagram illustrating the contents of operation m4 (point mutation). In the example of Fig. 15, node n21 and node n34 included in the winning tree tr11 are replaced with node n421 and node n434, resulting in a winning tree tr41. The winning tree tr41 becomes a descendant (individual) of the next generation.

[0130] [Embodiment 2] The estimation system S1 of the present invention may select a parameter set that satisfies the target value of the flexural strength and has a cost below a predetermined value, and further determine, from among the parameter sets, a parameter set whose cycle time is within a predetermined time as a target parameter set.

[0131] Hereinafter, a second embodiment of the present invention will be described with reference to Fig. 16 and Fig. 17. For ease of explanation, the same reference numerals are given to members having the same functions as those described in the previous embodiment, and the explanations thereof will not be repeated. This also applies to the following embodiments.

[0132] The estimation device 1 according to this embodiment has a hardware configuration similar to that of the estimation device 1 according to embodiment 1. However, the secondary memory 13 of the estimation device 1 according to this embodiment stores a second estimation program P3 together with the first estimation program P1 or instead of the first estimation program P1.

[0133] The processor 11 expands the second estimation program P3 stored in the secondary memory 13 onto the primary memory 12. The processor 11 then executes each step included in the second estimation method M3 according to instructions included in the second estimation program P3 expanded onto the primary memory 12. The flexural strength relational formula F1 expanded onto the primary memory 12 is used when the processor 11 executes a flexural strength estimation step M31 (described later) of the second estimation method M3. The cycle time relational formula F2 expanded onto the primary memory 12 is used when the processor 11 executes a cycle time estimation step M35 (described later) of the second estimation method M3. The cost relational formula F3 expanded onto the primary memory 12 is used when the processor 11 executes a cost estimation step M33 (described later) of the second estimation method M3.

[0134] The second estimation program P3 being stored in the secondary memory 13 means that the source code or an executable file obtained by compiling the source code is stored in the secondary memory 13.

[0135] (Flow of the second estimation method M3) The flow of the second estimation method M3 will be described with reference to Fig. 16 and Fig. 17. Fig. 16 is a flowchart showing the flow of the second estimation method M3. Fig. 17 is a diagram showing a schematic diagram of a parameter set selection method in the second estimation method M3. The second estimation method M3 includes a parameter set creation step M30, a deflection strength estimation step (first estimation step) M31, a deflection strength selection step M32, a cost estimation step (fourth estimation step) M33, a cost selection step M34, a cycle time estimation step (fifth estimation step) M35, and a cycle time selection step M36.

[0136] The parameter set creation step M30, the deflection strength estimation step M31, and the deflection strength selection step M32 are similar to the parameter set creation step M10, the deflection strength estimation step M11, and the deflection strength selection step M12 described in the first embodiment, respectively, and therefore will not be described repeatedly. Processor 11 executes the next process, the cost estimation step M33, for the primary selected parameter set.

[0137] The cost estimation step M33 is a step in which the processor 11 uses the cost relational expression F3 to estimate the cost of creating a mold with each parameter set. Specifically, as shown in Fig. 17, the processor 11 inputs each of the primary selection parameter sets into the cost relational expression F3 to obtain a cost estimate as a solution to the cost relational expression F3. When cost estimation has been performed for all of the primary selection parameter sets, the processor 11 executes the cost selection step M34.

[0138] The cost selection step M34 is a step of selecting a parameter set whose cost identified in the cost estimation step M33 is equal to or less than a predetermined value. The processor 11 selects a parameter set whose cost is equal to or less than a predetermined value from the first selection parameter set. That is, in this embodiment, a parameter set whose cost is equal to or less than a predetermined value among the first selection parameter set is a second selection parameter set. The processor 11 executes the next process, the cycle time estimation step M35, for the second selection parameter set.

[0139] The cycle time estimation step M35 is a step in which the processor 11 uses the cycle time relational expression F2 to estimate the cycle time when a mold is created with each parameter set. Specifically, as shown in Fig. 17, the processor 11 inputs each of the secondary selection parameter sets into the cycle time relational expression F2 to obtain an estimated value of the cycle time as a solution to the cycle time relational expression F2. When the cycle time estimation has been performed for all of the secondary selection parameter sets, the processor 11 executes the cycle time selection step M36.

[0140] A cycle time selection step M36 is a step of selecting a parameter set whose cycle time specified in the cycle time estimation step M35 is within a predetermined time. Processor 11 selects one or more parameter sets whose cycle time is within the predetermined time from the secondary selection parameter sets, and determines them as a target parameter set.

[0141] According to the above process, the estimation device 1 can identify a parameter set that can obtain a bending strength that satisfies a target value. Also, according to the above process, the estimation device 1 can identify a parameter set that has a cost equal to or less than a predetermined value. Also, according to the above process, the estimation device 1 can identify a parameter set that has a cycle time within a predetermined time.

[0142] More specifically, the estimation device 1 estimates the flexural strength, cost, and cycle time in the order of the above-mentioned first estimation method M1, and selects a parameter set after each estimation. This allows the estimation device 1 to identify a parameter set that provides a target flexural strength, a cost that is equal to or less than a predetermined value, and a cycle time that is within a predetermined time. Therefore, the user can identify the values ​​of various parameters that satisfy the desired flexural strength, cost, and cycle time conditions without conducting experiments.

[0143] The second estimation method M3 may include a fourth output step of outputting the flexural strength estimated in the flexural strength estimation step M31. The second estimation method M3 may also include a fifth output step of outputting the cost estimated in the cost estimation step M33. The second estimation method M3 may also include a sixth output step of outputting the cycle time estimated in the cycle time estimation step M35. In the fourth to sixth output steps, the processor 11 outputs the estimated flexural strength, cost, or cycle time to a display. The second estimation method M3 may also include a condition setting step of setting the values ​​of various parameters indicated by the target parameter set as setting values ​​of the manufacturing device 4.

[0144] [Embodiment 3] The cycle time relational expression F2 and / or the cost relational expression F3 may be derived in a manner similar to the method of deriving the flexural strength relational expression F1 described in the first and second embodiments above.

[0145] When deriving cycle time relational equation F2 in a manner similar to that used to derive flexural strength relational equation F1, training data can be used that is a combination of a parameter set used to derive cycle time relational equation F2 and an actual measured value or an estimated value of the cycle time, and is a parameter related to cycle time relational equation F2 among the parameters used to derive flexural strength relational equation F1.

[0146] The same is true when deriving the cost relational formula F3 in a manner similar to that of deriving the flexural strength relational formula F1. It is sufficient to use training data that combines a parameter set for deriving the cost relational formula F3 and actual calculated values ​​or estimated values ​​of costs, which are parameters related to the cost relational formula F3 among the parameters used in deriving the flexural strength relational formula F1.

[0147] [Embodiment 4] In the above-described embodiments, a configuration has been described in which the flexural strength relational expression F1 is constructed using a nonlinear regression algorithm (specifically, a genetic algorithm). However, the method for constructing the flexural strength relational expression F1 is not limited to the nonlinear regression algorithm. For example, the flexural strength relational expression F1 may be constructed using the Monte Carlo method. In this case, for example, a plurality of function expressions are created by randomly selecting the length of the flexural strength relational expression F1 and the elements (variables, operators, etc.) of the flexural strength relational expression F1, and the flexural strength relational expression F1 is constructed by selecting the function expression with the smallest error from among these function expressions.

[0148] [Embodiment 5] The estimation device 1 according to the present invention may estimate the target parameter set using the trained model LM without using the bending strength relational equation F1. The trained model LM in each of the above-described embodiments is a model that has learned the relationship between a parameter set and the bending strength corresponding to the parameter set. In other words, in the trained model construction step (identification step) M22, the relational equation specification device 2 constructs a trained neural network model (more specifically, a deep neural network model) as the trained model LM, which, when a target value of the bending strength is input, estimates and outputs a target parameter set for creating a mold exhibiting the bending strength.

[0149] In this embodiment, the estimation device 1 is configured integrally with a relational equation specifying device (specifying device) 2, or the learned model LM is downloaded to the estimation device 1, or the estimation device 1 is configured to be able to access the learned model LM stored externally. The estimation device 1 then inputs a target value of the bending strength into the learned model LM to obtain a target parameter set. This makes it possible to derive the target parameter set directly from the learned model LM without the need to specify the bending strength relational equation F1.

[0150] [Embodiment 6] The relational equation specifying device 2 according to the present invention may specify the bending strength relational equation F1 from the actual parameter values ​​and the actual bending strength values ​​acquired by the data logger 3, without using a learned model. Hereinafter, the sixth embodiment of the present invention will be described with reference to FIG.

[0151] The relational equation specifying device 2 according to this embodiment has the same hardware configuration as the relational equation specifying device 2 according to embodiments 1 and 2. However, a second relational equation specifying program P4 is stored in the secondary memory 23 of the relational equation specifying device 2 according to this embodiment instead of the first relational equation specifying program P2.

[0152] The second relational expression specifying program P4 and the actual data set are stored in the secondary memory 23. The processor 21 expands the second relational expression specifying program P4 and the actual data set stored in the secondary memory 23 onto the primary memory 22. The processor 21 then executes each step included in the second specifying method M4 according to instructions included in the second relational expression specifying program P4 expanded onto the primary memory 22. The actual data set expanded onto the primary memory 22 is used when the processor 21 executes a first relational expression specifying step M24 (described later) of the second specifying method M4. Note that the second relational expression specifying program P4 being stored in the secondary memory 23 refers to the source code or an executable file obtained by compiling the source code being stored in the secondary memory 23.

[0153] <<Flow of the second identification method M4>> The flow of the second identification method M4 will be described with reference to Fig. 18. Fig. 18 is a flowchart showing the flow of the second identification method M4.

[0154] The second specifying method M4 includes an actual data set generating step M41 and a second relational expression specifying step M42.

[0155] The actual data set creation step M41 is a step in which the processor 21 creates an actual data set from the actual parameter values ​​acquired by the data logger 3 and the actual bending strength value. The actual data set is used in a second relational equation determination step M42, which will be described later. The processor 21 stores the created actual data set in the secondary memory 23.

[0156] The processor 21 creates an actual data set for the amount used in the second relational expression specifying step M42. When the actual data set required for executing the second relational expression specifying step M42 is prepared, the processor 21 executes the next second relational expression specifying step M42.

[0157] The second relational expression specifying step M42 is a step of specifying a bending strength relational expression F1 indicating the relationship between the parameter set and the bending strength from the actual data set using a nonlinear regression algorithm (a genetic algorithm in this embodiment). The specific processing content of the second relational expression specifying step M42 is similar to that of the first relational expression specifying step M24 described in the first embodiment. However, while the first relational expression specifying step M24 uses an estimated data set that is a combination of input data and output data to the trained model LM, the processor 21 according to this embodiment executes the same processing using an actual data set instead of an estimated data set.

[0158] [Embodiment 7] The estimation device 1 according to the present invention may be configured to derive the flexural strength from the parameter set without using the flexural strength relational expression F1 and the trained model LM as a neural network. For example, the estimation device 1 may derive the flexural strength from the parameter set by referring to a multidimensional table in which each parameter included in the parameter set is associated with the flexural strength.

[0159] [Modifications] The estimation device 1 according to the present invention may independently perform the bending strength estimation step (M11 or M31), the cycle time estimation step (M13 or M35), and the cost estimation step (M15 or M33). That is, each of the above steps in the first estimation program P1 or the second estimation program P3 may be a separate program module, and each may be a program that can be independently called and processed.

[0160] As a result, for example, from among a certain parameter set group, a parameter set that can obtain a flexural strength equal to or greater than a target value can be specified as a target parameter set.Also, for example, from among a certain parameter set group, a parameter set that has a cycle time within a predetermined value can be specified as a target parameter set.Also, for example, from among a certain parameter set group, a parameter set that has a cost equal to or less than a predetermined value or that has the lowest cost can be specified as a target parameter set.

[0161] In each embodiment of the present invention, a configuration is adopted in which the first estimation method M1, the first identification method M2, the second estimation method M3, and the second identification method M4 are executed by different processors (the processor 11 and the processor 21). However, the execution entity of the methods M1 to M4 in the present invention is not limited to the combination of the respective embodiments described above.

[0162] That is, two or more of the first estimation method M1, the first specification method M2, the second estimation method M3, and the second specification method M4 may be executed using the same processor. In this case, by executing the first specification method M2 or the second specification method M4, the bending strength relational formula F1 is stored in a memory built in the same computer as the processor. Then, when executing the first estimation method M1 or the second estimation method M3, the processor uses the bending strength relational formula F1 stored in the memory.

[0163] 〔summary〕 The estimation device of aspect 1 includes one or more processors, and the processor executes a first estimation step of estimating the flexural strength of the mold from a parameter set of multiple parameters related to the creation of the mold, the multiple parameters including (1) raw material parameters related to the raw material of the mold, (2) mixing parameters measured or set in a mixing process of the raw material, and (3) molding parameters measured or set in a molding process of the mixed sand created in the mixing process.

[0164] According to the above configuration, the estimation device can estimate the value of the bending strength from a plurality of parameters related to the creation of the mold.

[0165] The estimation device according to the second aspect has the following features in addition to the features of the estimation device according to the first aspect. That is, in the estimation device according to the second aspect, the first estimation step is a step of estimating the flexural strength from the parameter set by using a flexural strength relational equation indicating a relationship between the flexural strength of the mold and a plurality of parameters related to the creation of the mold.

[0166] According to the above configuration, the estimation device can estimate the value of the flexural strength from a plurality of parameters related to the creation of the mold, using the flexural strength relational equation.

[0167] The estimation device according to the third aspect has the following features in addition to the features of the estimation device according to the second aspect: In the estimation device according to the third aspect, the bending strength relational expression is a relational expression identified using a nonlinear regression algorithm.

[0168] According to the above configuration, the estimation device can identify the bending strength relational expression by using a nonlinear regression algorithm.

[0169] The estimation device according to aspect 4 has the following features in addition to the features of the estimation device according to aspect 2 or 3. That is, in the estimation device according to aspect 8, the flexural strength relational equation is a relational equation identified by a nonlinear regression algorithm using a parameter set that combines output data obtained by inputting specific values ​​of the parameter set into a trained model that has been machine-learned to learn the relationship between the flexural strength of the mold and a plurality of parameters related to the production of the mold, and the specific values.

[0170] According to the above configuration, the estimation device can estimate the flexural strength by using the parameter set created by the trained model and further by using the flexural strength relational expression specified by the nonlinear regression algorithm. Therefore, the estimation device can estimate the flexural strength more accurately.

[0171] The estimation device according to aspect 5 has the following features in addition to the features of the estimation device according to any one of aspects 1 to 4. That is, in the estimation device according to aspect 4, the processor executes a second estimation step of estimating the cycle time by substituting at least a portion of the parameters included in the parameter set in which the flexural strength estimated in the first estimation step is equal to or greater than a target value into a cycle time relational equation that is a relational equation showing a relationship between at least a portion of the plurality of parameters and a cycle time for producing the mold.

[0172] According to the above configuration, the estimation device can further estimate the cycle time for a parameter set capable of producing a mold having a flexural strength equal to or greater than a target value.

[0173] The estimation device according to aspect 6 has the following features in addition to the features of the estimation device according to aspect 5. That is, in the estimation device according to aspect 5, the processor executes a third estimation step of estimating the cost by substituting at least a portion of the parameters included in the parameter set for which the cycle time estimated in the second estimation step is within a predetermined time into a cost relational equation that is a relational equation showing a relationship between at least a portion of the multiple parameters and the cost of creating the mold.

[0174] According to the above configuration, the estimation device can further estimate the cost of mold creation for a parameter set that can create a mold having a flexural strength equal to or greater than a target value and for which the cycle time of mold creation is within a specified time.

[0175] The estimation device according to aspect 7 has the following features in addition to the features of the estimation device according to any one of aspects 1 to 4. That is, in the estimation device according to aspect 6, the processor executes a fourth estimation step of estimating the cost by substituting at least a portion of the parameters included in the parameter set in which the flexural strength estimated in the first estimation step is equal to or greater than a target value into a cost relational equation that is a relational equation showing a relationship between at least a portion of the plurality of parameters and the cost of creating the mold.

[0176] According to the above configuration, the estimation device can further estimate the cost for a parameter set capable of producing a mold having a flexural strength equal to or greater than a target value.

[0177] The estimation device according to aspect 8 has the following features in addition to the features of the estimation device according to aspect 7. That is, in the estimation device according to aspect 7, the processor executes a fifth estimation step of estimating the cost by substituting at least a portion of the parameters included in a parameter set for which the cost estimated in the fourth estimation step is equal to or less than a predetermined value into a cycle time relational equation, which is a relational equation showing the relationship between at least a portion of the plurality of parameters and a cycle time for creating the mold.

[0178] According to the above configuration, the estimation device can further estimate the cycle time for mold production for a parameter set that can produce a mold having a flexural strength equal to or greater than a target value and whose cost is within a predetermined value.

[0179] The estimation device according to the ninth aspect has the following features in addition to the features of the estimation device according to any one of the first to eighth aspects: In the estimation device according to the ninth aspect, the raw material parameters include at least a part of the amounts of foundry sand, water, and additives charged in the kneading step, and the amounts of additives remaining in the foundry sand.

[0180] According to the above-described configuration, the estimation device can estimate the bending strength from at least a part of the amounts of molding sand, water, and additives charged in the kneading process, and the amount of residual additives in the molding sand.

[0181] The estimation device according to the tenth aspect has the following features in addition to the features of the estimation device according to any one of the first to ninth aspects: In the estimation device according to the tenth aspect, the mixing parameters include at least a part of the mixing speed of the raw material, the mixing time, and the temperature of the mixed sand in the mold mixing step.

[0182] According to the above configuration, the estimation device can estimate the bending strength from at least a part of the mixing speed of the raw material, the mixing time, and the temperature of the mixed sand in the mixing process of the mold.

[0183] The estimation device according to aspect 11 has the following features in addition to the features of the estimation device according to any one of aspects 1 to 10. That is, in the estimation device according to aspect 11, the molding parameters include at least a part of the injection speed of the mixed sand into the formwork of the casting mold, the temperature of the formwork, the amount of release agent applied to the formwork, and the baking time of the mixed sand.

[0184] According to the above configuration, the estimation device can estimate the flexural strength from at least some of the following: the injection speed of the mixed sand into the formwork of the casting mold, the temperature of the formwork, the amount of release agent applied to the formwork, and the baking time of the mixed sand.

[0185] The relational equation identifying device of aspect 12 includes one or more processors, and the processors execute an identifying step of identifying a relational equation or a learned model for estimating flexural strength from a parameter set of multiple parameters related to the creation of a mold, including (1) raw material parameters related to the raw material of the mold, (2) mixing parameters measured or set in a mixing process of the raw material, (3) molding parameters measured or set in a molding process of the mixed sand created in the mixing process, and (4) flexural strength of the mold.

[0186] According to the above configuration, it is possible to identify a relational equation or a learned model for estimating flexural strength from multiple parameters related to mold creation.

[0187] The estimation method of aspect 13 includes a first estimation step in which one or more processors estimate the flexural strength of the mold from a parameter set of a plurality of parameters related to the creation of the mold, the plurality of parameters including (1) raw material parameters related to the raw material of the mold, (2) mixing parameters measured or set in a mixing process of the raw material, and (3) molding parameters measured or set in a molding process of the mixed sand created in the mixing process.

[0188] According to the above-described configuration, the value of the bending strength can be estimated from a plurality of parameters related to the creation of the mold.

[0189] The identification method of aspect 14 includes an identification step in which one or more processors identify a relational equation or a learned model for estimating flexural strength from a parameter set of multiple parameters related to the creation of a mold, including (1) raw material parameters related to the raw material of the mold, (2) mixing parameters measured or set in a mixing process of the raw material, (3) molding parameters measured or set in a molding process of the mixed sand created in the mixing process, and (4) flexural strength of the mold.

[0190] According to the above configuration, it is possible to identify a relational equation or a learned model for estimating flexural strength from multiple parameters related to mold creation.

[0191] [Additional Note 1] In the first to fifth embodiments described above, a trained model trained by machine learning through supervised learning is used as the trained model LM, but a trained model trained by machine learning through unsupervised learning may be used. For the trained model trained by machine learning through unsupervised learning, the processor 21 also identifies a nonlinear relational expression that expresses the relationship between the input and output of the trained model by using a nonlinear regression algorithm.

[0192] [Additional Note 2] In the above-described first to fifth embodiments, the relational equation identifying device 2 constructs the learned model LM, but the learned model LM may be constructed in advance by a device other than the relational equation identifying device 2. In this case, the relational equation identifying device 2 performs the above-described first identifying method M2 using a learned model constructed in advance by the other device.

[0193] [Additional Note 3] In the first to fifth embodiments described above, the processor 21 identifies the relational equation by nonlinear regression using a genetic algorithm in which a nonlinear relational equation representing the relationship between the input data and the output data of the trained model LM is set as a solution candidate. However, the method of identifying the relational equation representing the relationship between the input data and the output data is not limited to the method described in the first and second embodiments described above. For example, the processor 21 may identify the first relational equation and the second relational equation by nonlinear regression using a genetic algorithm in which a nonlinear first relational equation representing the relationship between the numerical value representing the input data and each output value of some or all nodes belonging to the hidden layer, and a nonlinear second relational equation representing the relationship between the output value of each node and the numerical value representing the output data are set as solution candidates. In this case, the processor 21 identifies the nonlinear relational equation representing the relationship between the numerical value representing the input data and the numerical value representing the output data by, for example, solving a simultaneous equation of the identified first relational equation and the second relational equation.

[0194] [Additional Note 4] The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention. [Explanation of symbols]

[0195] 1 Estimation device 2. Relational equation identification device 3 Data Logger 4 Manufacturing equipment 11, 21, Processor 12, 22, Primary memory 13, 23, Secondary memory 14, 24 Input / Output Interface 15, 25 Communication Interface 16, 26 Bus 40. Casting sand property measuring device 41 Casting sand charging device 42 Additive dosing device 43 Water injection device 44 Kneader 45 Sand Mixing Property Measuring Device 46 Molds 47 Molding machine 48 Deflection intensity meter M1 1st estimation method M3 2nd estimation method M10, M30 Parameter Set Creation Steps M11, M31 flexural strength estimation step M12, M32 flexural strength selection step M13, M35 Cycle time estimation steps M14, M36 Cycle Time Selection Step M15, M33 Cost Estimation Steps M16, M34 cost selection step M2 1st identification method M21 Dataset Construction Steps M22 Trained model construction steps M23 Estimation Dataset Creation Steps M24 First relational equation identification step M4 2nd identification method M41 Actual Data Set Creation Step M42 Second relational expression identification step S1 Estimation System

Claims

1. one or more processors; The processor executes a first estimation step of estimating a flexural strength of the mold from a parameter set of a plurality of parameters related to the creation of the mold; The plurality of parameters are: (1) material parameters relating to the material of the mold; (2) kneading parameters measured or set in the raw material kneading process; and (3) A molding parameter measured or set in a molding process of the mixed sand produced in the mixing process, The raw material parameters include an amount of molding sand, an amount of water, an amount of additives, and an amount of additives remaining in the molding sand in the kneading process; The mixing parameters include a mixing speed of the raw material, a mixing time, and a temperature of the mixed sand in the mold mixing process; The molding parameters include a rate at which the mixed sand is pressed into the mold frame, a temperature of the mold frame, an amount of release agent applied to the mold frame, and a baking time of the mixed sand. An estimation device comprising:

2. 2. The estimation device according to claim 1, wherein the first estimation step is a step of estimating the flexural strength from the parameter set using a flexural strength relational equation that indicates a relationship between the flexural strength of the mold and a plurality of parameters related to the creation of the mold.

3. The estimation device according to claim 2 , wherein the bending strength relational expression is a relational expression specified using a nonlinear regression algorithm.

4. 4. The estimation device according to claim 2 or 3, wherein the flexural strength relational equation is a relational equation identified by a nonlinear regression algorithm using a parameter set that combines output data obtained by inputting specific values ​​of the parameter set into a trained model that has been machine-learned to determine the relationship between the flexural strength of the mold and a plurality of parameters related to the production of the mold, and the specific values.

5. The processor, A cycle time relational expression which is a relational expression showing a relationship between at least a part of the plurality of parameters and a cycle time of creating the mold, 5. The estimation device according to claim 1, further comprising: a second estimation step of estimating the cycle time by substituting at least a portion of the parameters included in the parameter set in which the flexural strength estimated in the first estimation step is equal to or greater than a target value into the cycle time relational expression.

6. The processor, A cost relational equation which is a relational equation showing a relationship between at least some of the plurality of parameters and a cost required for creating the mold, 6. The estimation device according to claim 5, further comprising: a third estimation step of estimating the cost by substituting at least a portion of the parameters included in the parameter set for which the cycle time estimated in the second estimation step is within a predetermined time, into the cost relational equation.

7. The processor, A cost relational equation which is a relational equation showing a relationship between at least some of the plurality of parameters and a cost required for creating the mold, 5. The estimation device according to claim 1, further comprising: a fourth estimation step of estimating the cost by substituting at least a portion of the parameters included in the parameter set in which the flexural strength estimated in the first estimation step is equal to or greater than a target value into the cost relational equation.

8. The processor, A cycle time relational expression which is a relational expression showing a relationship between at least a part of the plurality of parameters and a cycle time of creating the mold, 8. The estimation device according to claim 7, further comprising: a fifth estimation step of estimating the cost by substituting at least a portion of the parameters included in the parameter set for which the cost estimated in the fourth estimation step is equal to or less than a predetermined value into the cycle time relational expression.

9. one or more processors; The processor includes: (1) material parameters relating to the material of the mold; (2) Kneading parameters measured or set in the raw material kneading process; (3) Molding parameters measured or set in a molding process of the mixed sand produced in the mixing process; and (4) The flexural strength of the mold A step of identifying a relational expression or a learned model for estimating a flexural strength from a parameter set of a plurality of parameters, comprising: The raw material parameters include an amount of molding sand, an amount of water, an amount of additives, and an amount of additives remaining in the molding sand in the kneading process; The mixing parameters include a mixing speed of the raw material, a mixing time, and a temperature of the mixed sand in the mold mixing process; The molding parameters include a rate at which the mixed sand is pressed into the mold frame, a temperature of the mold frame, an amount of release agent applied to the mold frame, and a baking time of the mixed sand. A specific device characterized by the above.

10. A first estimation step in which one or more processors estimate a flexural strength of the mold from a parameter set of a plurality of parameters related to the creation of the mold, The plurality of parameters are: (1) material parameters relating to the material of the mold; (2) kneading parameters measured or set in the raw material kneading process; and (3) A molding parameter measured or set in a molding process of the mixed sand produced in the mixing process, The raw material parameters include an amount of molding sand, an amount of water, an amount of additives, and an amount of additives remaining in the molding sand in the kneading process; The mixing parameters include a mixing speed of the raw material, a mixing time, and a temperature of the mixed sand in the mold mixing process; The molding parameters include a rate at which the mixed sand is pressed into the mold frame, a temperature of the mold frame, an amount of release agent applied to the mold frame, and a baking time of the mixed sand.

13. An estimation method comprising:

11. The one or more processors determine a plurality of parameters related to the creation of the mold, (1) material parameters relating to the material of the mold; (2) Kneading parameters measured or set in the raw material kneading process; (3) Molding parameters measured or set in a molding process of the mixed sand produced in the mixing process; and (4) The flexural strength of the mold A step of identifying a relational expression or a learned model for estimating a flexural strength from a parameter set of a plurality of parameters, comprising: The raw material parameters include an amount of molding sand, an amount of water, an amount of additives, and an amount of additives remaining in the molding sand in the kneading process; The mixing parameters include a mixing speed of the raw material, a mixing time, and a temperature of the mixed sand in the mold mixing process; The molding parameters include a speed at which the mixed sand is pressed into the mold frame, a temperature of the mold frame, an amount of release agent applied to the mold frame, and a baking time of the mixed sand. The method for identifying a target object.

Images