Estimation device, learned model generation device, teacher data generation device, control program, estimation method, learned model generation method, and teacher data generation method

Abstract

To facilitate estimation of a heat transfer coefficient.SOLUTION: An estimation device (30) includes a second estimation unit (35) that acquires second heat transfer coefficient data corresponding to second pressure data by inputting the second pressure data indicating a pressure value acting on an interface to a learned model that has machine-learned a correlation between first pressure data and first heat transfer coefficient data using a basic data set in which the first pressure data indicating a pressure value acting on the interface between a metal part of a mold and a molten material and the first heat transfer coefficient data indicating the first heat transfer coefficient on the interface are associated with each other as teacher data.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an estimation device used for simulation of die forming, a learned model generation device, a teacher data generation device, a control program, an estimation method, a learned model generation method, and a teacher data generation method.

Background Art

[0002] Conventionally, methods for simulating the flow state and solidification behavior of a melt during die forming have been actively studied. Since the flow state and solidification behavior of the melt are affected by the metal part of the die and the respective temperatures of the melt, in this simulation method, it is important to accurately estimate the respective temperature distributions (hereinafter abbreviated as "temperature distributions") of the metal part and the melt.

[0003] In the estimation of the temperature distribution, it is necessary to estimate the heat transfer coefficient (HTC) at the interface between the metal part and the melt as a premise. The heat transfer coefficient is a value representing the ease of heat transfer at the interface and varies depending on the contact state between the metal part and the melt. Here, since the contact state between the metal part and the melt changes over time and also varies depending on the contact location, it has been difficult to estimate the heat transfer coefficient.

[0004] To solve this problem, various studies have been conducted. For example, Non-Patent Document 1 discloses the results of investigating the influence of the decrease in the heat transfer coefficient due to the formation of an air gap on the temperature history, shrinkage cavity prediction parameters, etc. in the solidification analysis of castings.

Prior Art Documents

Non-Patent Documents

[0005]

Non-Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] The investigation method disclosed in Non-Patent Document 1 estimates the heat transfer coefficient by analyzing a model of the heat transfer coefficient that depends on the air gap amount at the time of air gap generation. Therefore, in this investigation method, it is necessary to measure the air gap amount as a premise for estimating the heat transfer coefficient. However, measuring the air gap amount is time-consuming for preparation and difficult to measure itself. Therefore, it has not always been easy to estimate the heat transfer coefficient using the investigation method disclosed in Non-Patent Document 1.

[0007] One aspect of the present invention has been made in view of the above problems, and an object thereof is to facilitate the estimation of the heat transfer coefficient.

Means for Solving the Problems

[0008] An estimation device according to one aspect of the present invention uses, as teacher data, a basic data set in which first pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface are associated with each other. A second estimation unit that inputs second pressure data indicating a pressure value acting on an interface between a metal part of the mold and a melt filled in a space surrounded by the metal part to a learned model that has learned the correlation between the first pressure data and the first heat transfer coefficient data and acquires second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to the second pressure data is provided.

[0009] The generation device of a learned model according to one aspect of the present invention uses, as teacher data, a basic data set in which first pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part is associated with first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface, machine-learns the correlation between the first pressure data and the first heat transfer coefficient data, and includes a first estimation unit that generates a learned model for outputting second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to second pressure data by inputting the second pressure data indicating a pressure value acting on an interface between the metal part of the mold and the melt filled in the space surrounded by the metal part.

[0010] The generation device of teacher data according to one aspect of the present invention includes a generation unit that generates teacher data used for machine-learning the correlation between first pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface.

[0011] An estimation method according to one aspect of the present invention is an estimation method executed by a computer, and includes a second estimation step of inputting second pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part into a learned model that has machine-learned the correlation between first pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface, and using the basic data set associated therewith as teacher data, and obtaining second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to the second pressure data.

[0012] A method for generating a learned model according to one aspect of the present invention is a generation method executed by a computer. Using, as teacher data, a basic dataset in which first pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part is associated with first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface, a correlation between the first pressure data and the first heat transfer coefficient data is machine-learned. A first estimation step of generating a learned model for outputting second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to second pressure data by inputting the second pressure data indicating a pressure value acting on an interface between a metal part of the mold and a melt filled in a space surrounded by the metal part is included.

[0013] A method for generating teacher data according to one aspect of the present invention is a generation method executed by a computer, and includes a generation step of generating teacher data used for machine-learning a correlation between first pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface.

[0014] An estimation device according to each aspect of the present invention may be realized by a computer. In this case, a control program for the estimation device that realizes the estimation device by operating the computer as each part (software element) included in the estimation device, and a computer-readable recording medium on which the control program is recorded also fall within the scope of the present invention.

Effects of the Invention

[0015] According to one aspect of the present invention, estimation of a heat transfer coefficient can be facilitated.

Brief Description of the Drawings

[0016]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

Figure 23

Figure 24

Figure 25

Mode for Carrying Out the Invention

[0017] 〔Embodiment 1〕 <Outline of the Simulation Device> The simulation device 100 according to Embodiment 1 of the present invention is a device that simulates the flow state and solidification behavior of molten metal 50 (melt) during die casting. In this embodiment, a tablet terminal will be described as an example of the simulation device 100. The simulation device 100 may be, for example, a smartphone or a desktop personal computer in addition to the tablet terminal. Details of the molten metal 50 will be described later.

[0018] As shown in FIG. 1, the simulation device 100 includes an input unit 1, an output unit 2, a storage unit 3, a control device 4, a first sensor 10, and a second sensor 20. The input unit 1 receives operation inputs from a user or the like. The output unit 2 outputs the results of various processes by the simulation device 100. In this embodiment, the simulation device 100 includes a touch panel in which the input unit 1 and the display unit as the output unit 2 are integrated.

[0019] Note that the input unit 1 and the display unit (output unit 2) may be physically separated. Also, the input unit 1 may be, for example, a keyboard or a pointing device. The output unit 2 may be, for example, a communication unit that wirelessly transmits (or wired transmits) the results of various processes by the simulation device 100 to an external communication device (not shown), or may be a printer.

[0020] The first sensor 10 is a component of the simulation device 100 that measures the temperatures Tmp1 to Tmp6 (first temperature), the temperature Tcp (second temperature), and the contact pressure Fp. Specifically, the temperatures Tmp1 to Tmp6 are six temperatures, namely, the temperatures Tmp1, Tmp2, Tmp3, Tmp4, Tmp5, and Tmp6.

[0021] The first sensor 10 includes a first temperature measurement unit 11, a first pressure measurement unit 12, and a first sensor body 13 described later. The first temperature measurement unit 11 measures the temperatures Tmp1 to Tmp6 and Tcp, and the first pressure measurement unit 12 measures the contact pressure Fp. Details of the temperatures Tmp1 to Tmp6 and Tcp, as well as the contact pressure Fp, will be described later.

[0022] The second sensor 20 is a component of the simulation device 100 that measures the temperatures Tmd1 to Tmd6 (first temperature), the temperature Tcd (second temperature), and the contact pressure Fd. Specifically, the temperatures Tmd1 to Tmd6 are six temperatures, namely, the temperatures Tmd1, Tmd2, Tmd3, Tmd4, Tmd5, and Tmd6.

[0023] The second sensor 20 includes a second temperature measurement unit 21, a second pressure measurement unit 22, and a second sensor body 23 described later. The second temperature measurement unit 21 measures the temperatures Tmd1 to Tmd6 and Tcd, and the second pressure measurement unit 22 measures the contact pressure Fd. Details of the temperatures Tmd1 to Tmd6 and Tcd, as well as the contact pressure Fd, will be described later.

[0024] In this embodiment, the first and second temperature measurement units 11 and 21 are thermocouples. The first and second temperature measurement units 11 and 21 may be, for example, radiation thermometers or thermometers using optical fibers, but it is preferable to use thermocouples in consideration of measurement accuracy. Also, in this embodiment, the first and second pressure measurement units 12 and 22 are strain gauges. The first and second pressure measurement units 12 and 22 may be, for example, piezoelectric elements.

[0025] The storage unit 3 is a storage device that stores various data used by the simulation device 100. The control device 4 is, for example, a CPU (Central Processing Unit), and comprehensively controls each part constituting the simulation device 100. The control device 4 has a first estimation device 30 (estimation device) and a second estimation device 60 (behavior estimation device).

[0026] The first estimation device 30 estimates the second heat transfer coefficient data hd-2, hpl-2, and hpu-2 described later. Details of the estimation process of the first estimation device 30 will be described later. The second estimation device 60 estimates the respective temperature distributions of the mold 40 and the molten metal 50 during mold forming using the second heat transfer coefficient data hd-2, hpl-2, and hpu-2 estimated by the first estimation device 30. Also, the second estimation device 60 estimates the solidification behavior of the molten metal 50 during mold forming from the respective temperature distributions estimated by the first estimation device 30.

[0027] Specifically, the second estimation device 60 defines a solidification temperature according to the type of the molten metal 50, a solid fraction corresponding to the solidification temperature, and the like. Next, the second estimation device 60 estimates the solidification behavior of the molten metal 50 for each time step based on each temperature distribution for each time step described later. The method for estimating the solidification behavior is not particularly limited. For example, the law of conservation of energy may be used as a basic equation for heat transfer estimation, or the forward Euler method, the Crank-Nicolson method, or the like may be used as a numerical solution method for differential equations. Also, as a method for handling solidification, the temperature recovery method, the equivalent specific heat method, the enthalpy method, or the like may be used.

[0028] Note that the first and second estimation devices 30 and 60 do not necessarily have to be built into the control device 4. It suffices if the simulation device 100 includes these devices in some way. Also, the simulation device 100 does not necessarily have to include the first and second estimation devices 30 and 60. For example, it may function as a server that stores the data of each of the first and second estimation devices 30 and 60. In this case, a simulation system may be constructed using the first and second sensors 10 and 20, the first and second estimation devices 30 and 60, and the simulation device 100.

[0029] The first estimation device 30 includes a temperature acquisition unit 31, a pressure acquisition unit 32, a generation unit 33, a first estimation unit 34, and a second estimation unit 35. The temperature acquisition unit 31 acquires the temperature values of the temperatures Tmp1 to Tmp6 and Tcp measured by the first sensor 10 from the first sensor 10, and the temperature values of the temperatures Tmd1 to Tmd6 and Tcd measured by the second sensor 20 from the second sensor 20. Hereinafter, each temperature value acquired by the temperature acquisition unit 31 from the first sensor 10 is referred to as "temperature data Tmp1-1 to Tmp6-1 and Tcp-1 (first temperature data, second temperature data)". Also, each temperature value acquired by the temperature acquisition unit 31 from the second sensor 20 is referred to as "temperature data Tmd1-1 to Tmd6-1 and Tcd-1 (first temperature data, second temperature data)".

[0030] The pressure acquisition unit 32 acquires the pressure value of the contact pressure Fp measured by the first sensor 10 from the first sensor 10, and the pressure value of the contact pressure Fd measured by the second sensor 20 from the second sensor 20. Hereinafter, the pressure value acquired by the pressure acquisition unit 32 from the first sensor 10 is referred to as "contact pressure data Fp-1 (pressure data)". Also, the pressure value acquired by the pressure acquisition unit 32 from the second sensor 20 is referred to as "contact pressure data Fd-1 (pressure data)". Each data acquired by each of the temperature acquisition unit 31 and the pressure acquisition unit 32 may be temporarily stored in the storage unit 3.

[0031] In this embodiment, each of the temperature acquisition unit 31 and the pressure acquisition unit 32 acquires data from the first and second sensors 10 and 20 at a plurality of time steps at regular intervals. Also, each of the temperature acquisition unit 31 and the pressure acquisition unit 32 acquires data from the first and second sensors 10 and 20 by receiving data through wireless communication. This data reception may be performed by wired communication.

[0032] Note that the first estimation device 30 may calculate each of the contact pressure data Fd-1 and Fp-1 using each temperature data acquired by the temperature acquisition unit 31. In this case, the first estimation device 30 may not have the pressure acquisition unit 32.

[0033] The generation unit 33 generates an upper punch side basic data set (basic data set) by determining first heat transfer coefficient data hpu-1 using a plurality of temperature data acquired by the temperature acquisition unit 31. The first heat transfer coefficient data hpu-1 is a value indicating the first heat transfer coefficient between the upper punch 42 and the molten metal 50 at the first interface Si1. The first heat transfer coefficient on the upper punch side is a value representing the ease of heat transfer at the first interface Si1. In this embodiment, the generation unit 33 acquires data from the temperature acquisition unit 31 at a plurality of time steps at regular intervals, and determines the first heat transfer coefficient data hpu-1 for each time step.

[0034] The upper punch side basic data set is a data set in which the contact pressure data Fp-1 and the first heat transfer coefficient data hpu-1 are associated with each other. In this embodiment, the generation unit 33 determines the first heat transfer coefficient data hpu-1 and acquires the contact pressure data Fp-1 from the pressure acquisition unit 32, thereby generating the upper punch side basic data set. Also in this embodiment, the generation unit 33 acquires data from the pressure acquisition unit 32 at a plurality of time steps at regular intervals. Therefore, the upper punch side basic data set of this embodiment has a configuration including a plurality of combinations of the contact pressure data Fp-1 and the first heat transfer coefficient data hpu-1 in one time step.

[0035] The generation unit 33 generates a die-side basic data set (basic data set) by the same generation method as the upper punch-side basic data set. The die-side basic data set is a data set in which contact pressure data Fd-1 and first heat transfer coefficient data hdu-1 are associated with each other.

[0036] The generation unit 33 generates a lower punch-side basic data set (basic data set) by the same generation method as the upper punch-side basic data set. The lower punch-side basic data set is a data set in which contact pressure data Fp-1 and first heat transfer coefficient data hpl-1 are associated with each other. In the present embodiment, the generation unit 33 determines the first heat transfer coefficient data hpl-1, and generates the lower punch-side basic data set by regarding the contact pressure data Fp-1 acquired from the pressure acquisition unit 32 as the pressure value of the contact pressure acting on the third interface Si3.

[0037] Note that each of the foregoing basic data sets does not necessarily have to be generated by the generation unit 33. For example, each of the foregoing basic data sets may be stored in advance in the storage unit 3 or an external server or the like, and may be read from the storage unit 3 or the like when the first estimation unit 34 performs an estimation process. Further, for example, an external information processing device may generate each of the foregoing basic data sets, and the first estimation unit 34 may acquire the basic data sets from the information processing device during the estimation process. Details of the first to third interfaces Si1 to Si3 and the determination processes of the first heat transfer coefficient data hd-1, hpl-1, and hpu-1 by the generation unit 33 will be described later.

[0038] The first estimation unit 34 estimates the correlation between the contact pressure data Fp-1 and the first heat transfer coefficient data hpu-1 using the upper punch-side basic data set generated by the generation unit 33. The first estimation unit 34 estimates the correlation between the contact pressure data Fd-1 and the first heat transfer coefficient data hd-1 using the die-side basic data set generated by the generation unit 33. The first estimation unit 34 estimates the correlation between the contact pressure data Fp-1 and the first heat transfer coefficient data hpl-1 using the lower punch-side basic data set generated by the generation unit 33. Details of the estimation process by the first estimation unit 34 will be described later.

[0039] The second estimation unit 35 estimates the second heat transfer coefficient data hd-2 from the estimated contact pressure data Fd-2 (specific pressure data) using the die-side estimated correlation relationship (estimated correlation relationship). The die-side estimated correlation relationship is the correlation between the contact pressure data Fd-1 estimated by the first estimation unit 34 and the first heat transfer coefficient data hd-1. The second heat transfer coefficient data hd-2 is the estimated value of the first heat transfer coefficient data hd-1 corresponding to the estimated contact pressure data Fd-2.

[0040] The second estimation unit 35 may select the estimated contact pressure data Fd-2 from among a plurality of contact pressure data Fd-1 included in the die-side basic data set. The second estimation unit 35 may use the value received by the input unit 1 as the estimated contact pressure data Fd-2 when the input unit 1 receives an input operation of the estimated contact pressure data Fd-2. The second estimation unit 35 may use the contact pressure data Fd-1 acquired from the pressure acquisition unit 32 as the estimated contact pressure data Fd-2. The same applies to the estimated contact pressure data Fp-2 described below.

[0041] Also, the second estimation unit 35 estimates the second heat transfer coefficient data hpu-2 from the estimated contact pressure data Fp-2 (specific pressure data) using the upper punch-side estimated correlation relationship (estimated correlation relationship). The upper punch-side estimated correlation relationship is the correlation between the contact pressure data Fp-1 estimated by the first estimation unit 34 and the first heat transfer coefficient data hpu-1. The second heat transfer coefficient data hpu-2 is the estimated value of the first heat transfer coefficient data hpu-1 corresponding to the estimated contact pressure data Fp-2.

[0042] Furthermore, the second estimation unit 35 estimates the second heat transfer coefficient data hpl-2 from the estimated contact pressure data Fp-2 using the lower punch-side estimated correlation relationship (estimated correlation relationship). The lower punch-side estimated correlation relationship is the correlation between the contact pressure data Fp-1 estimated by the first estimation unit 34 and the first heat transfer coefficient data hpl-1. The second heat transfer coefficient data hpl-2 is the estimated value of the first heat transfer coefficient data hpl-1 corresponding to the estimated contact pressure data Fp-2. Details of the estimation process by the second estimation unit 35 will be described later.

[0043] <Die and Contact Pressure> (Die) The die 40 is a metal mold used for manufacturing products by press working. In this embodiment, the forming material of the die 40 is carbon steel. As shown in FIG. 2, the die 40 includes a die 41, an upper punch 42 (metal part), and a lower punch 43 (metal part).

[0044] The die 41 is a hollow cylindrical mold, and the hollow part of the die 41 is cylindrical. On the inner side surface of the die 41 forming the hollow part, any one of various release agents such as a water-soluble release agent, an oil-based release agent, a BN (boron nitride) spray, and a black body spray is applied as necessary. A hole 41Y is formed in the metal part 41X of the die 41 in a direction perpendicular to the central axis of the die 41. The hole 41Y is a cylindrical hole having a size that allows the second sensor 20 to be inserted and removed, and penetrates from the outer side surface of the metal part 41X to the hollow part. In a state where the second sensor 20 is housed in the hole 41Y and set, as shown by reference numeral 202 in FIG. 2, the surface of the second sensor 20 (specifically, the second sensor body 23 described later) facing the molten metal 50 is flush with the inner side surface of the upper punch 42.

[0045] The upper punch 42 is a cylindrical mold having a size that can be inserted into and removed from the hollow part of the die 41. When setting the upper punch 42 in the die 41, the upper punch 42 is inserted through the upper opening in the hollow part. A hole 42X is formed in the upper punch 42 in a direction parallel to the central axis of the upper punch 42. The hole 42X is a cylindrical hole having a size that allows the first sensor 10 to be inserted and removed, and penetrates from the upper end surface to the lower end surface of the upper punch 42. Also, the central axis of the hole 42X coincides with the central axis of the upper punch 42. In a state where the first sensor 10 is housed in the hole 42X and set, as shown by reference numeral 202 in FIG. 2, the surface of the first sensor 10 (specifically, the first sensor body 13 described later) facing the molten metal 50 is flush with the lower end surface of the upper punch 42.

[0046] Note that the holes 41Y in the die 41 and the holes 42X in the upper punch 42 do not necessarily have to penetrate through. In this case, the thickness from the bottom surface of the non-penetrating hole 41Y to the inner side surface of the die 41 may be a thickness that can obtain the same measurement accuracy as that of the second sensor 20 when the second sensor 20 is set in the hole 41Y in the embodiment. Similarly, the thickness from the bottom surface of the non-penetrating hole 42X to the lower end surface of the upper punch 42 may be a thickness that can obtain the same measurement accuracy as that of the first sensor 10 when the first sensor 10 is set in the hole 42X in the embodiment.

[0047] The lower punch 43 is also a cylindrical mold with a size that can be inserted into and removed from the hollow portion of the die 41, similar to the upper punch 42. When setting the lower punch 43 in the die 41, the lower punch 43 is inserted from the lower opening in the hollow portion. Then, as shown by reference numeral 201 in FIG. 2, the lower punch 43 is accommodated in the hollow portion so that the lower end surface of the lower punch 43 is flush with the lower end surface of the die 41.

[0048] Casting (pressure casting) by press working using the mold 40 is performed as follows. First, the die 41 with the lower punch 43 accommodated in the hollow portion is set in a hydraulic press (not shown). Next, the molten metal 50 is poured into the hollow portion from the upper opening in the hollow portion. In this embodiment, the molten metal 50 is a molten metal of an aluminum alloy die-cast such as ADC12. The molten metal 50 may be a molten metal of other die-cast alloys such as a zinc alloy die-cast.

[0049] Next, insert the upper punch 42 from the upper opening in the hollow part, bring the upper punch 42 into contact with the molten metal 50, and then press the upper punch 42 vertically downward with a hydraulic press. That is, in the pressing process by the upper punch 42, the space 40X surrounded by the metal part 41X of the die 41, the upper punch 42, and the lower punch 43 is filled with the molten metal 50. Also, in this process, the central axes of the die 41, the upper punch 42, and the lower punch 43 all coincide. This coincident central axis becomes the central axis AX of the mold 40. Finally, by solidifying the molten metal 50 while pressing it with the upper punch 42, a pre-finished molded body (not shown) is completed.

[0050] Note that the "orthogonal", "parallel", "coincident", and "flush" in the aforementioned "orthogonal to the central axis", "parallel to the central axis", "coincident with the central axis", and "flush with the inner side surface / lower end surface" do not strictly require "orthogonal", "parallel", "coincident", and "flush" in the exact sense. It is sufficient that they are "orthogonal", "parallel", "coincident", and "flush" at the visual recognition level, and it is a concept that includes dimensional errors and the like. This also applies to the following explanations.

[0051] The mold 40 does not necessarily have to be for press working. For example, when manufacturing a product with a molten resin (melt) instead of the molten metal 50, an injection molding die may be used as the mold 40. Further, the mold 40 does not have to be composed of the die 41, the upper punch 42, and the lower punch 43. The mold 40 may be of any shape and structure as long as each of the first and second estimation devices 30 and 60 can execute the estimation process.

[0052] (Contact pressure) As described above, the first pressure measurement unit 12 of the first sensor 10 measures the contact pressure Fp, and the second pressure measurement unit 22 of the second sensor 20 measures the contact pressure Fd. The contact pressure Fp is the pressure acting on the first interface Si1 indicated by reference numeral 202 in FIG. 2. The contact pressure Fd is the pressure acting on the second interface Si2 indicated by reference numeral 202 in FIG. 2. Note that in reference numeral 202 in FIG. 2, a gap is formed between the die 41, the upper punch 42, the lower punch 43, and the molten metal 50. This is for convenience of explanation, and actually, the die 41, the upper punch 42, and the lower punch 43 are in contact with the molten metal 50. The same applies to FIG. 13.

[0053] The first interface Si1 is a concept composed of a surface Sp1 in contact with the molten metal 50 in each of the first sensor 10 and the upper punch 42, and a surface Sc1 in contact with each of the first sensor 10 and the upper punch 42 in the molten metal 50. The "surface Sp1 in contact with the molten metal 50 in each of the first sensor 10 and the upper punch 42" is specifically composed of the surface facing the molten metal 50 in the first sensor 10 (the first sensor main body 13 described later) and the lower end surface of the upper punch 42. In the following description, this surface is referred to as the "first contact surface Sp1". The "surface Sc1 in contact with each of the first sensor 10 and the upper punch 42 in the molten metal 50" is specifically the upper end surface of the molten metal 50, and is referred to as the "second contact surface Sc1" in the following description.

[0054] From the above definition of the first interface Si1, the "contact pressure Fp" refers to two pressures, namely, the contact pressure Fp acting from the first contact surface Sp1 toward the second contact surface Sc1, and the contact pressure Fp acting from the second contact surface Sc1 toward the first contact surface Sp1. Here, the pressure Fpx acting from the second contact surface Sc1 on the surface facing the molten metal 50 in the first sensor 10 also becomes the contact pressure Fp. In the present embodiment, since the contact pressure Fp is measured based on the distortion of the first sensor 10 (specifically, the first sensor main body 13 described later) caused by the pressure acting on the first sensor 10, the pressure Fpx is referred to as the "contact pressure Fp" in the following description.

[0055] The second interface Si2 is a concept composed of a surface Sd that comes into contact with the molten metal 50 in each of the second sensor 20 and the die 41, and a surface Sc2 that comes into contact with each of the second sensor 20 and the die 41 in the molten metal 50. Specifically, the "surface Sd that comes into contact with the molten metal 50 in each of the second sensor 20 and the die 41" is composed of the surface facing the molten metal 50 in the second sensor 20 (the second sensor body 23 described later) and the inner side surface of the die 41. In the following description, this surface is referred to as the "third contact surface Sd". The "surface Sc2 that comes into contact with each of the second sensor 20 and the die 41 in the molten metal 50" is specifically the side surface of the molten metal 50, and is referred to as the "fourth contact surface Sc2" in the following description.

[0056] From the definition of the aforementioned second interface Si2, simply the "contact pressure Fd" refers to two pressures, namely the contact pressure Fd acting from the third contact surface Sd towards the fourth contact surface Sc2, and the contact pressure Fd acting from the fourth contact surface Sc2 towards the third contact surface Sd. Here, the pressure Fdx acting from the fourth contact surface Sc2 on the surface facing the molten metal 50 in the second sensor 20 also becomes the contact pressure Fd. In this embodiment, since the contact pressure Fd is measured based on the distortion of the second sensor 20 (specifically, the second sensor body 23 described later) caused by the pressure acting on the second sensor 20, the pressure Fdx is referred to as the "contact pressure Fd" in the following description.

[0057] Note that the third interface Si3 is composed of a surface Sp2 that comes into contact with the molten metal 50 in the lower punch 43 and a surface Sc3 that comes into contact with the lower punch 43 in the molten metal 50. Specifically, the "surface Sp2 that comes into contact with the molten metal 50 in the lower punch 43" is the upper end surface of the lower punch 43, and is referred to as the "fifth contact surface Sp2" in the following description. The "surface Sc3 that comes into contact with the lower punch 43 in the molten metal 50" is specifically the lower end surface of the molten metal 50, and is referred to as the "sixth contact surface Sc3" in the following description.

[0058] Then, at the first interface Si1, the second interface Si2, and the third interface Si3, an interface between the entire metal part of the mold 40 and the molten metal 50 is formed. In the present embodiment, the simulation device 100 performs various processes on the assumption that a contact pressure Fp (specifically, pressure Fpx) also acts on the third interface Si3 without measuring the contact pressure acting on the third interface Si3.

[0059] <Specific structures of the first and second sensors> (Specific structure of the first sensor) The first sensor 10 has a first sensor body 13 as shown in FIG. 3. The first sensor body 13 is a solid bar-shaped member and is composed of a plurality of cylindrical portions with different outer diameters. The central axes of the plurality of cylindrical portions coincide, and this coincident central axis becomes the central axis AX1 of the first sensor body 13. In the present embodiment, the first sensor body 13 is formed of the same carbon steel as the mold 40.

[0060] Also, in the present embodiment, as indicated by reference numeral 301 in FIG. 3, a pair of first pressure measurement portions 12 are provided on the first sensor body 13. Specifically, the two first pressure measurement portions 12 constituting this pair are arranged at symmetric positions with respect to the central axis AX1 in the first sensor body 13 when viewed from the extending direction of the central axis AX1.

[0061] The aforementioned "symmetric positions with respect to the central axis AX1 in the first sensor body 13" are positions that satisfy the following two conditions (i) and (ii) when the first sensor body 13 is viewed from the extending direction of the central axis AX1. Condition (i) is that a straight line (hereinafter, "virtual line") connecting the centers of gravity of the two first pressure measurement portions 12 constituting the pair intersects the central axis AX1. Condition (ii) is that the intersection point of the virtual line of condition (i) and the central axis AX1 is the midpoint of the virtual line.

[0062] Further, these two first pressure measurement units 12 are arranged at symmetric positions with respect to the central axis AX1 in the first sensor body 13 even when viewed from a direction orthogonal to the extending direction of the central axis AX1. The "symmetric positions with respect to the central axis AX1 in the first sensor body 13" means positions that satisfy the aforementioned condition (ii) and the following condition (iii) when the first sensor body 13 is viewed from a direction orthogonal to the extending direction of the central axis AX1. Condition (iii) is that the virtual line of condition (i) is orthogonal to the central axis AX1.

[0063] Furthermore, these two first pressure measurement units 12 are arranged at positions separated by a predetermined distance from the end face of the first sensor body 13 facing the molten metal 50. The "predetermined distance" varies depending on the sizes of the first sensor body 13 and the mold 40, the measurement accuracy of the first pressure measurement unit 12, etc., but it suffices that a distance is ensured such that the influence of the temperature change of the molten metal 50 is not significantly affected. The same applies to the second pressure measurement unit 22 for such an arrangement.

[0064] The number and arrangement mode of the first pressure measurement units 12 are not limited to the examples of this embodiment. For example, only one first pressure measurement unit 12 may be provided in the first sensor body 13, or a plurality of pairs of first pressure measurement units 12 may be provided. Furthermore, the two first pressure measurement units 12 constituting a pair may not be arranged at symmetric positions with respect to the central axis AX1 in the first sensor body 13. The same applies to the second pressure measurement unit 22.

[0065] Furthermore, in the present embodiment, seven first temperature measurement units 11 are provided in the first sensor 10, and the first sensor 10 measures the temperature at seven locations shown by reference numeral 302 in FIG. 3. Specifically, the first sensor 10 measures the temperature at three locations on the central axis AX1. These three locations are provided in the order of measurement locations Pcp, Pmp1, and Pmp2 from the position on the molten metal 50 side. The measurement location Pcp is provided in the molten metal 50 and is provided near the end face of the first sensor body 13 that contacts the molten metal 50. Hereinafter, the temperature of the molten metal 50 measured at the measurement location Pcp is defined as temperature Tcp. The measurement location Pmp1 is provided near the boundary between the cylindrical portion (hereinafter, “first cylindrical portion”) including the end face that contacts the molten metal 50 and the cylindrical portion adjacent to the first cylindrical portion (hereinafter, “second cylindrical portion”). Hereinafter, the temperature of the first sensor body 13 measured at the measurement location Pmp1 is defined as temperature Tmp1.

[0066] Here, since the first sensor body 13 is formed of the same carbon steel as the mold 40, in the present embodiment, the temperature of the first sensor body 13 is regarded as the temperature of the upper punch 42 (that is, the mold 40). Therefore, the temperature Tmp1 is the temperature of the upper punch 42. The measurement location Pmp2 is provided closer to the central portion side of the first sensor body 13 in the direction of the central axis AX1 than the measurement location Pmp1. Hereinafter, the temperature of the upper punch 42 measured at the measurement location Pmp2 is defined as temperature Tmp2.

[0067] In addition, the first sensor 10 measures the temperature at four locations near the boundary between the first cylindrical portion and the second cylindrical portion. All of these four locations are provided on the side surface of the second cylindrical portion. And when the first sensor body 13 is viewed from the extending direction of the central axis AX1, the angle formed by a straight line that virtually connects one of two adjacent locations and the central axis AX1 and a straight line that virtually connects the other and the central axis AX1 is 90°.

[0068] The above four locations are provided in the clockwise order of measurement locations Pmp3, Pmp6, Pmp4, and Pmp5 when viewed from the extending direction of the central axis AX1 of the first sensor body 13. Hereinafter, the temperatures of the upper punch 42 measured at the measurement locations Pmp3, Pmp4, Pmp5, and Pmp6 are defined as temperatures Tmp3, Tmp4, Tmp5, and Tmp6, respectively.

[0069] (Specific Structure of the Second Sensor) The second sensor 20 has a second sensor body 23 as shown in FIG. 4. The second sensor body 23 is a solid bar-shaped member, and both ends are cylindrical. On the other hand, the body portion sandwiched between both ends is an elongated flat plate shape, and the shape of the end face in the thickness direction of the body portion is a curved surface shape corresponding to the outer shape of both ends. The central axes of both ends and the body portion coincide, and this coincident central axis becomes the central axis AX2 of the second sensor body 23. Among the end faces of both ends of the second sensor body 23, the end face that contacts the molten metal 50 has a curved surface shape corresponding to the shape of the inner side surface of the die 41. In the present embodiment, the second sensor body 23 is also formed of the same carbon steel as the mold 40, similar to the first sensor body 13.

[0070] Also, in the present embodiment, as shown by reference numeral 401 in FIG. 4, a pair of second pressure measurement units 22 are provided on the second sensor body 23. Specifically, the two second pressure measurement units 22 constituting this pair are arranged at symmetric positions with respect to the central axis AX2 of the second sensor body 23 when viewed from either the extending direction of the central axis AX2 of the second sensor body 23 or the direction orthogonal to the extending direction. The meaning of "symmetric positions with respect to the central axis AX2 of the second sensor body 23" is the same as that of "symmetric positions with respect to the central axis AX1 of the first sensor body 13" described above.

[0071] Furthermore, in the present embodiment, seven second temperature measurement units 21 are provided in the second sensor 20, and the second sensor 20 measures the temperature at seven locations shown by reference numeral 402 in FIG. 4. Specifically, the second sensor 20 measures the temperature at three locations on the central axis AX2. These three locations are provided in the order of measurement locations Pcd, Pmd1, and Pmd2 from the position on the molten metal 50 side. The measurement location Pcd is provided in the molten metal 50 and is provided near the end face of the second sensor body 23 that contacts the molten metal 50. Hereinafter, the temperature of the molten metal 50 measured at the measurement location Pcd is defined as temperature Tcd. The measurement location Pmd1 is provided near the boundary between the end portion including the end face that contacts the molten metal 50 (hereinafter, “molten metal side end portion”) and the main body portion. Hereinafter, the temperature of the second sensor body 23 measured at the measurement location Pmd1 is defined as temperature Tmd1.

[0072] Here, since the second sensor body 23 is also formed of the same carbon steel as the mold 40, in the present embodiment, the temperature of the second sensor body 23 is regarded as the temperature of the die 41 (that is, the mold 40). Therefore, the temperature Tmd1 is the temperature of the die 41. The measurement location Pmd2 is provided closer to the central portion side of the second sensor body 23 in the direction of the central axis AX2 than the measurement location Pmd1. Hereinafter, the temperature of the die 41 measured at the measurement location Pmd2 is defined as temperature Tmp2.

[0073] In addition, the second sensor 20 measures the temperature at four locations near the boundary between the molten metal side end portion and the main body portion. Two of these four locations are provided on one plane of the main body portion, and the remaining two locations are provided on the other plane. And when the second sensor body 23 is viewed from the extending direction of the central axis AX2, the angle formed by a straight line that virtually connects one of two adjacent locations and the central axis AX2 and a straight line that virtually connects the other and the central axis AX2 is 90°.

[0074] The above four locations are provided in the clockwise order of measurement locations Pmd3, Pmd5, Pmd4, and Pmd6 when the molten-metal-side end of the second sensor body 23 is viewed in the extending direction of the central axis AX2. Hereinafter, the temperatures of the die 41 measured at the measurement locations Pmd3, Pmd4, Pmd5, and Pmd6 are defined as temperatures Tmd3, Tmd4, Tmd5, and Tmd6, respectively.

[0075] <Temperature Measurement and Pressure Measurement> (Temperature Measurement) In the present embodiment, as shown by reference numerals 501 and 502 in FIG. 5, the simulation device 100 measures temperatures at a total of 21 locations, including 14 locations on the mold 40 side and 7 locations on the molten-metal 50 side. The arrangement of the total 21 measurement locations is such that three measurement locations Pcn, Pmn-1, and Pmn-2 (n: natural number from 1 to 7) shown by reference numeral 503 in FIG. 5 are provided linearly in both the plan view and the front view. And a total of 7 groups of measurement locations each consisting of these three measurement locations are provided.

[0076] Hereinafter, the three measurement locations Pcn, Pmn-1, and Pmn-2 are abbreviated as "measurement locations Pcn, ~Pmn-2". Also, the temperature measured at the measurement location Pcn is defined as temperature Tcn, the temperature measured at the measurement location Pmn-1 is defined as temperature Tmn-1, and the temperature measured at the measurement location Pmn-2 is defined as temperature Tmn-2.

[0077] In the present embodiment, as shown by reference numeral 503 in FIG. 5, the measurement location Pcn is provided at a position 2 mm from the first mold surface toward the molten-metal 50 side. The measurement location Pmn-1 is provided at a position 2 mm from the first mold surface toward the mold 40 side. The measurement location Pmn-2 is provided at a position 6 mm from the first mold surface toward the mold 40 side. Here, the "first mold surface" refers to any one of the first contact surface Sp1, the third contact surface Sd, or the fifth contact surface Sp2.

[0078] The measurement points Pc1 to Pm1-2 are provided on the central axis AX in both the plan view and the front view. As shown by reference numeral 502 in FIG. 5, the measurement point Pc1 is provided near the first interface Si1 inside the molten metal 50, and the measurement points Pm1-1 and Pm1-2 are provided inside the upper punch 42.

[0079] The measurement point Pc1 is the measurement point Pcp at which the first sensor 10 measures temperature, and the temperature Tc1 measured at the measurement point Pc1 becomes the temperature Tcp. The measurement point Pm1-1 is the measurement point Pmp1 at which the first sensor 10 measures temperature, and the temperature Tm1-1 measured at the measurement point Pm1-1 becomes the temperature Tmp1. The measurement point Pm1-2 is the measurement point Pmp2 at which the first sensor 10 measures temperature, and the temperature Tm1-2 measured at the measurement point Pm1-2 becomes the temperature Tmp2.

[0080] The measurement points Pc2 to Pm2-2 are provided on a straight line (not shown) parallel to the central axis AX in both the plan view and the front view, as shown by reference numerals 501 and 502 in FIG. 5. Also, as shown by reference numeral 502 in FIG. 5, the vertical positions of the measurement points Pc2 to Pm2-2 in the front view are the same as those of the measurement points Pc1 to Pm1-2, and they are provided closer to the die 41 side than the measurement points Pc1 to Pm1-2.

[0081] The measurement points Pc3 to Pm3-2 are provided on a straight line (not shown) parallel to the central axis AX in the plan view and perpendicular to the central axis AX in the front view, as shown by reference numerals 501 and 502 in FIG. 5. The measurement point Pc3 is provided near the second interface Si2 inside the molten metal 50, and the measurement points Pm1-1 and Pm1-2 are provided inside the upper punch 42. Also, the measurement point Pc3 is provided near the measurement point Pc2. The measurement points Pm1-2, Pm2-2, Pc3 to Pm3-2 are provided on the same straight line, as shown by reference numeral 502 in FIG. 5.

[0082] The measurement points Pc4 to Pm4-2 and the measurement points Pc5 to Pm5-2 are provided on a straight line that is parallel to the central axis AX in a plan view and orthogonal to the central axis AX in a front view, similar to the measurement points Pc3 to Pm3-2. As shown by reference numeral 502 in FIG. 5, the measurement points Pc4 to Pm4-2 and the measurement points Pc5 to Pm5-2 have the same horizontal positions in the front view as the measurement points Pc3 to Pm3-2. The measurement points Pc4 to Pm4-2 are provided below the measurement points Pc3 to Pm3-2 in the front view, and the measurement points Pc5 to Pm5-2 are provided below the measurement points Pc4 to Pm4-2 in the front view.

[0083] The measurement points Pc3 to Pm3-2 and the measurement points Pc5 to Pm5-2 are provided at the same position in a plan view as shown by reference numeral 501 in FIG. 5. On the other hand, the measurement points Pc4 to Pm4-2 are provided at a position where the angle formed by the straight line including the measurement points Pc4 to Pm4-2 and the straight line including the measurement points Pc3 to Pm3-2 is 45° in the plan view.

[0084] The measurement point Pc4 is the measurement point Pcd where the second sensor 20 measures the temperature, and the temperature Tc4 measured at the measurement point Pc4 becomes the temperature Tcd. The measurement point Pm4-1 is the measurement point Pmd1 where the second sensor 20 measures the temperature, and the temperature Tm4-1 measured at the measurement point Pm4-1 becomes the temperature Tmd1. The measurement point Pm4-2 is the measurement point Pmd2 where the second sensor 20 measures the temperature, and the temperature Tm4-2 measured at the measurement point Pm4-2 becomes the temperature Tmd2.

[0085] The measurement points Pc6 to Pm6-2 are provided on a straight line parallel to the central axis AX in both the plan view and the front view, similar to the measurement points Pc2 to Pm2-2, as shown by reference numerals 501 and 502 in FIG. 5. The measurement points Pc2 to Pm2-2 and the measurement points Pc6 to Pm6-2 are provided at the same position in the plan view as shown by reference numeral 501 in FIG. 5. As shown by reference numeral 502 in FIG. 5, the measurement point Pc6 is provided near the third interface Si3 inside the molten metal 50, and the measurement points Pm6-1 and Pm6-2 are provided inside the lower punch 43.

[0086] The measurement points Pc7 to Pm7-2 are provided on the central axis AX in both the plan view and the front view, as shown by reference numerals 501 and 502 in Fig. 5, similar to the measurement points Pc1 to Pm1-2. The measurement points Pc1 to Pm1-2 and the measurement points Pc7 to Pm7-2 are provided at the same positions in the plan view as shown by reference numeral 501 in Fig. 5. Also, as shown by reference numeral 502 in Fig. 5, the vertical positions of the measurement points Pc7 to Pm7-2 in the front view are the same as those of the measurement points Pc6 to Pm6-2.

[0087] In each of the die 41, the upper punch 42, and the lower punch 43, through-holes and non-through holes (both not shown) for measuring temperatures other than the temperatures Tc1 to Tm1-2 and Tc4 to Tm4-2 among the temperatures measured at a total of 21 measurement points are formed. Then, thermocouples (not shown) inserted into these holes measure temperatures other than the temperatures Tc1 to Tm1-2 and Tc4 to Tm4-2. The measurement results of these thermocouples, that is, the temperature data of the temperatures other than the temperatures Tc1 to Tm1-2 and Tc4 to Tm4-4 among the temperatures measured at a total of 21 measurement points, are transmitted to the temperature acquisition unit 31 by wireless communication or wired communication.

[0088] Hereinafter, the temperature values of the temperatures Tc1 to Tc7, Tm1-1 to Tm7-1, and Tm1-2 to Tm7-2 acquired by the temperature acquisition unit 31 are referred to as "temperature data Tc1' to Tc7' (second temperature data), Tm1-1' to Tm7-1', and Tm1-2' to Tm7-2' (first temperature data)".

[0089] Here, the temperature data Tc1' is the same as the temperature data Tcp-1, the temperature data Tm1-1' is the same as the temperature data Tmp1-1, and the temperature data Tm1-2' is the same as the temperature data Tmp2-1. In the following description, the names will be unified as "temperature data Tc1', Tm1-1' and Tm1-2'". Also, the temperature data Tc4' is the same as the temperature data Tcd-1, the temperature data Tm4-1' is the same as the temperature data Tmd1-1, and the temperature data Tm4-2' is the same as the temperature data Tmd2-1. In the following description, the names will be unified as "temperature data Tc4', Tm4-1' and Tm4-2'".

[0090] (Pressure measurement) The first and second sensors 10 and 20 convert the contact pressures Fd and Fp from the strain values ε (unit: μST) measured by strain gauges as the first and second pressure measurement units 12 and 22. As shown in FIG. 6, the strain gauge also changes the gauge factor (a coefficient representing the sensitivity of the strain gauge) according to the temperature change. Therefore, in the present embodiment, in order to accurately convert the contact pressures Fd and Fp from the strain value ε measured by the strain gauge, the first and second sensors 10 and 20 are compressed in advance at a plurality of temperatures.

[0091] Specifically, each of the first and second sensors 10 and 20 is compressed at room temperature (293K in this embodiment) and at a plurality of temperatures higher than room temperature (the increase in temperature is constant), and the compression characteristics at each temperature are specified. Then, based on the compression characteristics at a plurality of temperatures specified in this preliminary process, the correlation between the strain generated in each of the first and second sensors 10 and 20 and the compression stress σ (unit: MPa) acting on these sensors is calibrated. The first and second sensors 10 and 20 use the compression stress σ converted from the strain value ε after the above calibration as the contact pressures Fd and Fp.

[0092] (Determination of the first heat transfer coefficient data using a three-dimensional unsteady heat transfer model) The generation unit 33 determines the first heat transfer coefficient data hd-1 that constitutes the die-side basic data set by representing the heat transfer in the vicinity of each of the first to third interfaces Si1 to Si3 with a three-dimensional unsteady heat transfer model. The generation unit 33 determines the first heat transfer coefficient data hpu-1 that constitutes the upper punch-side basic data set. The generation unit 33 determines the first heat transfer coefficient data hpl-1 that constitutes the lower punch-side basic data set.

[0093] Specifically, the generation unit 33 determines the first heat transfer coefficient data h3-1, h4-1, and h5-1 as the first heat transfer coefficient data hd-1. The generation unit 33 determines the first heat transfer coefficient data h1-1 and h2-1 as the first heat transfer coefficient data hpu-1. The generation unit 33 determines the first heat transfer coefficient data h6-1 and h7-1 as the first heat transfer coefficient data hpl-1.

[0094] The first heat transfer coefficient data h1-1 is the first heat transfer coefficient data in the interface region between the measurement locations Pc1 and Pm1-1. The first heat transfer coefficient data h2-1 is the first heat transfer coefficient data in the interface region between the measurement locations Pc2 and Pm2-1. The first heat transfer coefficient data h3-1 is the first heat transfer coefficient data in the interface region between the measurement locations Pc3 and Pm3-1. The first heat transfer coefficient data h4-1 is the first heat transfer coefficient data in the interface region between the measurement locations Pc4 and Pm4-1. The first heat transfer coefficient data h5-1 is the first heat transfer coefficient data in the interface region between the measurement locations Pc5 and Pm5-1. The first heat transfer coefficient data h6-1 is the first heat transfer coefficient data in the interface region between the measurement locations Pc6 and Pm6-1. The first heat transfer coefficient data h7-1 is the first heat transfer coefficient data in the interface region between the measurement locations Pc7 and Pm7-1.

[0095] (Die model) The generation unit 33 generates a three-dimensional model of the mold 40 (hereinafter referred to as the "mold model") as shown in FIG. 7, and determines the first heat transfer coefficient data h1-1 to h7-1 using this mold model. The mold model is an eighth model of the entire mold 40 assuming that the mold 40 is equally divided into eight parts in the direction of the central axis AX (see FIG. 2). The mold model includes an eighth model of the entire molten metal 50 (hereinafter referred to as the "molten metal model") and is assumed to be placed under air.

[0096] The mold model may be generated by the generation unit 33 or may be stored in the storage unit 3 in advance. When the mold model is stored in the storage unit 3, the generation unit 33 reads the mold model from the storage unit 3 when determining each first heat transfer coefficient data. Alternatively, the generation unit 33 may acquire the data of the mold model from an external server or the like.

[0097] The mold model is composed of a plurality of hexahedral elements as shown by reference numeral 701 in FIG. 7. In addition, the mold model is assigned an element number i in the radial direction from the central axis AX, an element number j in the circumferential direction of the mold 40, and an element number k in the direction of the central axis AX. For the entire mold model, the element numbers are i = 1 to 15, j = 1 to 6, and k = 1 to 35, and the total number of elements constituting the mold model (including the molten metal model) is 3240.

[0098] The measurement points Pcn to Pmn-2 (n: natural numbers from 1 to 7) are displayed at the respective positions shown by reference numeral 702 in FIG. 7 when using the mold model with a total of 3240 elements. Reference numeral 702 in FIG. 7 is a side view of the mold model, that is, a view showing the outer side surface of each element with j = 1 in the mold model.

[0099] In reference numeral 702 of FIG. 7, measurement locations Pc4 to Pm4-2 are shown at the position of j = 1 for convenience of explanation. Actually, the measurement locations Pc4 to Pm4-2 are located at j = 6. As shown in reference numeral 702 of FIG. 7, the mold model shall handle measurement locations Pc1 to Pm1-2 and Pc7 to Pm7-2 as positions of i = 1 for convenience of calculation. Actually, the measurement locations Pc1 to Pm1-2 and Pc7 to Pm7-2 are provided on the central axis AX.

[0100] (Details of the determination process of the first heat transfer coefficient data) Hereinafter, with reference to the flowchart shown in FIG. 8, the determination process of the first heat transfer coefficient data h1-1 to h7-1 using the mold model will be described. The generation unit 33 determines the first heat transfer coefficient data h1-1 from among the first heat transfer coefficient data h1-1 to h7-1 (S101). First, when determining the first heat transfer coefficient data h1-1, the generation unit 33 specifies only the temperature data of a plurality of elements (hereinafter, "target elements") including the molten metal side unit contact surface among the elements constituting the molten metal model in the mold model. The molten metal side unit contact surface is the surface of the element constituting a part of each of the second, fourth, and sixth contact surfaces Sc1, Sc2, and Sc3.

[0101] Specifically, the generation unit 33 acquires temperature data Tc1' to Tc7' at a certain time step (time point) from the temperature acquisition unit 31. The temperature data Tc1' to Tc7' are the temperature values of the target elements provided with the measurement locations Pc1 to Pc7 among the plurality of target elements. Further, the generation unit 33 calculates the temperature data at a certain time step of the target elements other than the target elements provided with the measurement locations Pc1 to Pc7 among the plurality of target elements (hereinafter, "unmeasured elements") using the following formula (1) (S111).

[0102]

Equation

[0103] Tcp: Temperature data [K] of unmeasured elements at a certain time step Tci: Among the plurality of target elements, the target element (natural number i = 1 to 7) provided with measurement points Pc1 to Pc7 di: Distance [mm] between each of the measurement points Pc1 to Pc7 and the centroid of the non-measured element m: Parameter representing weight (can be arbitrarily set. Preferably 0 ≤ m ≤ 1) Next, the generation unit 33 sets the first heat transfer coefficient data h1-1 to h7-1 at a certain time step. For example, when the input unit 1 receives a setting operation for the first heat transfer coefficient data h1-1 to h7-1, the generation unit 33 may use the value received by the input unit 1 as the setting value of the first heat transfer coefficient data h1-1 to h7-1. Also, for example, the generation unit 33 may set the first heat transfer coefficient data h1-1 to h7-1 by reading the setting value previously stored in the storage unit

[0104] Also, the generation unit calculates all the unit first heat transfer coefficient data other than the first heat transfer coefficient data h1-1 to h7-1 using the following formula (2) (S112).

[0105]

Equation

[0106] hip: Unit first heat transfer coefficient data [W / (m 2 ·K)] hi: First heat transfer coefficient data h1-1 to h7-1 [W / (m 2 ·K)] The unit first heat transfer coefficient data is the first heat transfer coefficient data in the unit interface area, and in S112, it includes the first heat transfer coefficient data h2-1 to h7-1. The unit interface area is composed of the surface (hereinafter, "die side unit contact surface") that constitutes a part of each of the first, third, and fifth contact surfaces Sp1, Sd, and Sp2 included in one element on the die 40 side, and the molten metal side unit contact surface facing the die side unit contact surface. Note that the first heat transfer coefficient data between the die model and the air is uniformly set to 50 W / (m 2 ·K).

[0107] Next, the generation unit 33 estimates the temperature data T for all elements (hereinafter referred to as "die-side elements") that constitute the die model other than the molten metal model in the die model, using the following equations (3) to (5). i、j、k ´ (S113). The estimated temperature data T i、j、k ´ is an estimated value of the temperature data (first temperature data) of the die-side elements at the time when a certain period of time has elapsed from a certain time step (hereinafter referred to as "the next time step").

[0108]

Equation

[0109] T i、j、k ´: Estimated temperature data of die-side elements [K] T i、j、k : Temperature data of die-side elements and target elements at a certain time step [K] i: Element number in the radial direction from the central axis AX j: Element number in the circumferential direction of the die 40 k: Element number in the direction of the central axis AX C: Δt / Cm [(sec·m 3 ) / J] Δt: Certain time (required time between time steps) [sec] Cm: Heat capacity of the die 40 [J / m 3 R: Thermal resistance in the unit interface area [(m 2 ·K) / W] Δrj: Length in the radial direction from the central axis AX [m] Δθ: Angle in the circumferential direction of the die 40 [rad] Δz: Length in the direction of the central axis AX [m]

[0110]

Equation

[0111] h: First heat transfer coefficient data h1-1 to h7-1 [W / (m 2 ·K)] ​λc: Thermal conductivity of molten metal 50 [W / (m K)] λm: Thermal conductivity of mold 40 [W / (m K)]

[0112]

number

[0113] The generating unit 33 generates the estimated temperature data T i、j、k When calculating ', if the unit interface area is formed between the mold-side element and the target element, the above-mentioned formulas (3) and (4) are used. Also, if the unit interface area is formed between mold-side elements, the generation unit 33 uses the above-mentioned formulas (3) and (5).

[0114] Next, the generation unit 33 calculates the plurality of estimated temperature data T i、j、k Of these, the generation unit 33 calculates the difference between the estimated temperature data Te1-1' at measurement point Pm1-1 and the temperature data Tm1-1' at measurement point Pm1-1 in the next time step. Similarly, the generation unit 33 calculates a total of 14 types of differences, including the difference between the estimated temperature data Te7-2' at measurement point Pm7-2 and the temperature data Tm7-2' at measurement point Pm7-2 in the next time step. The generation unit 33 then calculates the evaluation function e by summing up the absolute values ​​of these 14 types of differences.

[0115] Next, the generation unit 33 changes the value of the first heat transfer coefficient data h1-1 and calculates the evaluation function e again. The generation unit 33 repeats the calculation of the evaluation function e multiple times to identify the value of the first heat transfer coefficient data h1-1 that minimizes the value of the evaluation function e (S114).

[0116] 9, (A) the generating unit 33 calculates an evaluation function e1 using first heat transfer coefficient data h1-1 (hereinafter, "first heat transfer coefficient data h1-11") set at a certain time step. Next, (B) the generating unit 33 calculates an evaluation function e2 using first heat transfer coefficient data h1-1 (hereinafter, "first heat transfer coefficient data h1-12") changed from the first heat transfer coefficient data h1-11 by an amount of change Δh.

[0117] Regarding the change Δh in the first heat transfer coefficient data h1-12 from the first heat transfer coefficient data h1-11, the generation unit 33 may set the value received by the input unit 1 as the setting value of the change Δh when the input unit 1 receives a setting operation of the change Δh. Also, for example, the generation unit 33 may set the value of the change Δh by reading out a setting value stored in advance in the storage unit.

[0118] Next, (C) the generation unit 33 compares the values ​​of the evaluation function e1 and the evaluation function e2. If the value of the evaluation function e1 is smaller, the above-mentioned change amount Δh is multiplied by a coefficient M(-1 <M<0)を乗じた値M×Δhを第1熱伝達係数データh1-11に加算して、新たな第1熱伝達係数データh1-12´とする。そして、この第1熱伝達係数データh1-12´を用いて評価関数e2´を算出し、評価関数e1と評価関数e2´との値の大小を比較する。一方、評価関数e2の方が値が小さければ、第1熱伝達係数データh1-12に前述の変化量Δhを加算して第1熱伝達係数データh1-13とする。そして、この第1熱伝達係数データh1-13を用いて評価関数e3を算出し、評価関数e2と評価関数e3との値の大小を比較する。

[0119] Thereafter, the generation unit 33 repeats the above-mentioned processes (B) and (C), and ends this repeated process when the value of M×Δh becomes equal to or smaller than a predetermined threshold value (which can be set arbitrarily).

[0120] Also, for example, the generating unit 33 may calculate a plurality of evaluation functions e in advance and extract the evaluation function e having the smallest value from the calculated plurality of evaluation functions e. Specifically, as shown by reference numeral 1001 in Fig. 10, (D) the generating unit 33 determines the number of evaluation functions e11 to be subjected to one extraction process, and repeats each process of S112 to S114 until the determined number of evaluation functions e11 are obtained. Each time the generating unit 33 repeats each process of S112 to S114, it increases the first heat transfer coefficient data h1-1 by the amount of change Δh1.

[0121] Regarding the determination of the number of evaluation functions e11 to be subjected to one extraction process, the generation unit 33 may set the number of evaluation functions e11 to the value accepted by the input unit 1 when the input unit 1 accepts a setting operation of the number. Also, for example, the generation unit 33 may set the number of evaluation functions e11 by reading out a numerical value of the number stored in advance in the storage unit.

[0122] Next, as shown by reference numeral 1002 in Fig. 10, (E) the generating unit 33 extracts the evaluation function emin1 having the smallest value from among the multiple evaluation functions e11 obtained in the process of (D), and calculates the same number of evaluation functions e22 as the evaluation functions e11 to be subjected to the second extraction process. The generating unit 33 obtains a predetermined number of evaluation functions e22 by sequentially increasing or decreasing the first heat transfer coefficient data h1-1 corresponding to the evaluation function emin1 by the amount of change Δh2, with the value of the evaluation function emin1 set as the median. Here, the absolute value of the amount of change Δh2 is smaller than the absolute value of the amount of change Δh1.

[0123] Next, (F) the generating unit 33 extracts the evaluation function emin2 with the smallest value from among the multiple evaluation functions e22 obtained in the process of (E), and calculates the same number of evaluation functions e33 as the evaluation functions e11 and e22 to be subjected to the third extraction process. The generating unit 33 obtains a predetermined number of evaluation functions e33 by sequentially increasing or decreasing the first heat transfer coefficient data h1-1 corresponding to the evaluation function emin2 by the change amount Δh3, taking the value of the evaluation function emin2 as the median value. Here, the absolute value of the change amount Δh3 is smaller than the absolute value of the change amount Δh2.

[0124] Thereafter, the generation unit 33 repeats each process of (F) described above, and ends this repeated process when the value of Δh3 becomes equal to or smaller than a predetermined threshold value (which can be set arbitrarily).

[0125] The generating unit 33 specifies the first heat transfer coefficient data h1-1 by repeating the processes of S112 to S114, and temporarily stores the specified first heat transfer coefficient data h1-1 (hereinafter, "first heat transfer coefficient data h1-1'") in the storage unit 3. The generating unit 33 may store these data in a memory (not shown) in the generating unit 33.

[0126] Similarly, the generation unit 33 specifies the first heat transfer coefficient data h2-1 in the process steps S111 to S115. At this time, the estimated temperature data T i、j、k The generation unit 33 temporarily stores the specified first heat transfer coefficient data h2-1 (hereinafter, “first heat transfer coefficient data h2-1′”) in the storage unit 3.

[0127] Similarly, the generation unit 33 specifies the first heat transfer coefficient data h3-1 in the process steps S111 to S115. At this time, the estimated temperature data T i、j、k The calculation of "first heat transfer coefficient data h3-1" uses the previously specified first heat transfer coefficient data h1-1' and h2-1' and the first heat transfer coefficient data h4-1 to h7-1. The generation unit 33 temporarily stores the specified first heat transfer coefficient data h3-1 (hereinafter, "first heat transfer coefficient data h3-1'") in the storage unit 3.

[0128] In this way, the generating unit 33 sequentially identifies the first heat transfer coefficient data up to h7-1 by using the first heat transfer coefficient data identified in the previous process to identify the next first heat transfer coefficient data as needed. In the following description, the first heat transfer coefficient data h7-1 identified by the generating unit 33 is referred to as "first heat transfer coefficient data h7-1'". The generating unit 33 temporarily stores the identified seven pieces of first heat transfer coefficient data in the storage unit 3.

[0129] Next, the generation unit 33 again specifies the first heat transfer coefficient data h1-1 in the process procedure of S111 to S115. At this time, the estimated temperature data T i、j、kThe previously determined first heat transfer coefficient data h2-1' to h7-1' are used to calculate "first heat transfer coefficient data h1-1". The generation unit 33 temporarily stores the determined first heat transfer coefficient data h1-1 (hereinafter, "first heat transfer coefficient data h1-1"") in the storage unit 3. Similarly, the generation unit 33 re-determines up to the first heat transfer coefficient data h7-1 in the process steps of S111 to S115.

[0130] Next, the generating unit 33 judges whether the first heat transfer coefficient data h1-1 to h7-1 have converged (S108). Specifically, the generating unit 33 compares the first heat transfer coefficient data h1-1' with h1-1" to identify the amount of change in the first heat transfer coefficient data h1-1. If the identified amount of change in the first heat transfer coefficient data h1-1 is equal to or smaller than a threshold value (which can be set arbitrarily), the generating unit 33 judges that the first heat transfer coefficient data h1-1 has converged. In a similar manner, the generating unit 33 judges whether each of the first heat transfer coefficient data h2-1 to h7-1 has converged.

[0131] Next, the generation unit 33 acquires the contact pressure data Fp-1 at a certain time step from the pressure acquisition unit 32, and sets it as constituent data of the upper punch side foundation data set. Hereinafter, the upper punch side foundation data set including the first heat transfer coefficient data h1-1 as constituent data will be referred to as the "first upper punch side foundation data set."

[0132] Similarly, the generating unit 33 selects the contact pressure data Fp-1 and the first heat transfer coefficient data h2-1 as constituent data of the "second upper punch side basic data set". The generating unit 33 selects the contact pressure data Fd-1 and the first heat transfer coefficient data h3-1 as constituent data of the "first die side basic data set". The generating unit 33 selects the contact pressure data Fd-1 and the first heat transfer coefficient data h4-1 as constituent data of the "second die side basic data set". The generating unit 33 selects the contact pressure data Fd-1 and the first heat transfer coefficient data h5-1 as constituent data of the "third die side basic data set". The generating unit 33 selects the contact pressure data Fp-1 and the first heat transfer coefficient data h6-1 as constituent data of the "first lower punch side basic data set". The generating unit 33 selects the contact pressure data Fp-1 and the first heat transfer coefficient data h7-1 as constituent data of the "second lower punch side basic data set."

[0133] If the answer is No in S108, the generation unit 33 performs the processes of S101 to S107 again. On the other hand, if the answer is Yes in S108, the generation unit 33 updates the time step once (S109). If the updated time step is not the final time step (No in S110), the generation unit 33 performs the processes of S101 and onwards again to determine the first heat transfer coefficient data h1-1 to h7-1 in the updated time step. If the updated time step is the final time step (Yes in S110), the generation unit 33 ends the process of determining the first heat transfer coefficient data h1-1 to h7-1.

[0134] <Estimation of correlation between contact pressure data and first heat transfer coefficient data> The first estimation unit 34 estimates the correlation between the contact pressure data Fp-1 and the first heat transfer coefficient data h1-1 and h2-1 (hereinafter, "first and second upper punch side correlation") using the first and second upper punch side basic data sets generated by the generation unit 33. The first estimation unit 34 estimates the correlation between the contact pressure data Fd-1 and the first heat transfer coefficient data h3-1 to h5-1 (hereinafter, "first to third die side correlation") using the first to third die side basic data sets. The first estimation unit 34 estimates the correlation between the contact pressure data Fp-1 and the first heat transfer coefficient data h7-1 and h6-1 (hereinafter, "first and second lower punch side correlation") using the first and second lower punch side basic data sets.

[0135] The estimation of the seven correlations by the first estimation unit 34 results in substantially the same result. Therefore, in the following description, the first upper punch side correlation is taken as an example, and the description of the other correlations is omitted. The first estimation unit 34 estimates the first upper punch side correlation by dividing it into three periods as shown in Fig. 11. The estimation process of the first estimation unit 34 described below corresponds to the process performed next when Yes is determined in S110 using the flowchart in Fig. 8, and is an example of a first estimation step according to one aspect of the present invention.

[0136] First, the first estimation unit 34 estimates the first upper punch side correlation during the period from when the molten metal 50 is filled into the hollow portion of the die 41 in which the lower punch 43 is set until before the upper punch 42 starts to pressurize the molten metal 50 ("period I" in FIG. 11). During period I, the upper punch 42 does not pressurize the molten metal 50, so the contact pressure data Fp-1 is approximately 0. On the other hand, since the inner side surface of the die 41 and the molten metal 50 are in contact with each other, heat transfer occurs between the die 41 and the molten metal 50, so that the value of the temperature data Tc1' decreases and the value of the temperature data Tm1-1' increases. Therefore, even during period I, the first heat transfer coefficient data h1-1 does not become 0.

[0137] From these analysis results, the first estimation unit 34 estimates that the first upper punch side correlation in period I will be the relationship of the intercepts in the graph in Fig. 11. The relationship of the intercepts in the graph in Fig. 11 estimated by the first estimation unit 34 is referred to as the "first estimated correlation (estimated correlation)."

[0138] Next, the first estimation unit 34 estimates the first upper punch side correlation during a period from when the upper punch 42 starts to pressurize the molten metal 50 until the surface temperature of the molten metal 50 reaches the solidification temperature of the molding material of the molten metal 50 (hereinafter, "period II"). The "surface temperature of the molten metal 50" specifically refers to the temperature of the second contact surface Sc1 of the molten metal 50. During period II, the upper punch 42 continues to pressurize the molten metal 50 with a constant pressure force.

[0139] In the period II, the first heat transfer coefficient data h1-1 depends on the contact pressure data Fp-1 and is substantially independent of the solidification temperature. The first heat transfer coefficient data h1-1 and the contact pressure data Fp-1 are substantially proportional to each other. When the relationship between the first heat transfer coefficient data h1-1 and the contact pressure data Fp-1 is regarded as a linear function, the slope of the relationship is a positive value and the absolute value varies depending on the type of release agent and the surface roughness of the mold 40.

[0140] From these analysis results, the first estimation unit 34 estimates that the first upper punch side correlation in period II will be a linear function (with a positive slope) in the graph of Fig. 11. The first estimation unit 34 also estimates that the slope of the linear function varies depending on the type of release agent. The linear function (with a positive slope) in the graph of Fig. 11 estimated by the first estimation unit 34 is referred to as the "second estimated correlation (estimated correlation)."

[0141] Next, the first estimation unit 34 estimates the first upper punch side correlation during the period (hereinafter, "period III") from when the surface temperature of the molten metal 50 becomes lower than the solidification temperature of the molding material of the molten metal 50 until the solidified molded body is removed from the die 40. During period III, the upper punch 42 continues to pressurize the molten metal 50 with a constant pressure force, as in period II.

[0142] In period III, the first heat transfer coefficient data h1-1 depends on both the contact pressure data Fp-1 and the solidification temperature described above. The relationship between the first heat transfer coefficient data h1-1 and the contact pressure data Fp-1 is expressed by a linear function h1-1=a×Fp-1+b×Ts+c (a, b, c: coefficients, Ts: surface temperature of the molten metal 50).

[0143] From these analysis results, the first estimation unit 34 estimates that the first upper punch side correlation in period III will be a linear function (with a negative slope) relationship in the graph of Fig. 11. The linear function (with a negative slope) relationship in the graph of Fig. 11 estimated by the first estimation unit 34 is referred to as the "third estimated correlation (estimated correlation)."

[0144] The first to third estimated correlations derived by the first estimation unit 34 as estimation results are collectively referred to as the “first upper punch side estimated correlations.” Similarly, the first estimation unit 34 derives the “second upper punch side estimated correlations,” “first to third die side estimated correlations,” and “first and second lower punch side estimated correlations” as estimation results.

[0145] The first upper punch side estimated correlation and the second upper punch side estimated correlation constitute an upper punch side estimated correlation. The first die side estimated correlation, the second die side estimated correlation and the third die side estimated correlation constitute a die side estimated correlation. The first lower punch side estimated correlation and the second lower punch side estimated correlation constitute a lower punch side estimated correlation.

[0146] <Estimation of the second heat transfer coefficient data using the first to third estimated correlations> The second estimation unit 35 estimates the second heat transfer coefficient data h1-2 to h7-2 using any one of the first and second upper punch side estimated correlation, the first to third die side estimated correlation, and the first and second lower punch side estimated correlation, as appropriate. The second heat transfer coefficient data h1-2 and h2-2 are estimated values ​​of the first heat transfer coefficient data h1-1 and h2-1 corresponding to the contact pressure data Fp-2 for estimation. The second heat transfer coefficient data h1-2 and the second heat transfer coefficient data h2-2 constitute the second heat transfer coefficient data hpu-2.

[0147] The second heat transfer coefficient data h3-2 to h5-2 are estimates of the first heat transfer coefficient data h3-1 to h5-1 corresponding to the contact pressure data Fd-2 for estimation. The second heat transfer coefficient data h3-2, the second heat transfer coefficient data h4-2, and the second heat transfer coefficient data h5-2 constitute the second heat transfer coefficient data hd-2. The second heat transfer coefficient data h6-2 and h7-2 are estimates of the first heat transfer coefficient data h6-1 and h7-2 corresponding to the contact pressure data Fd-2 for estimation. The second heat transfer coefficient data h6-2 and the second heat transfer coefficient data h7-2 constitute the second heat transfer coefficient data hpl-2.

[0148] For example, when estimating the heat transfer coefficient data in the interface region between the measurement points Pc1 and Pm1-1 from the contact pressure data for estimation Fp-2 for the above-mentioned period II, the second estimation unit 35 uses the second estimated correlation of the first upper punch side estimated correlation. Then, the second estimation unit 35 applies the contact pressure data for estimation Fp-2 to this second estimated correlation to estimate the second heat transfer coefficient data h1-2 corresponding to the contact pressure data for estimation Fp-2 for the period II.

[0149] For example, when estimating heat transfer coefficient data in the interface region between measurement points Pc4 and Pm4-1 from the contact pressure data for estimation Fd-2 for the aforementioned period III, the second estimation unit 35 uses the third estimated correlation of the second die-side estimated correlation. Then, the second estimation unit 35 applies the contact pressure data for estimation Fd-2 to the third estimated correlation to estimate the second heat transfer coefficient data h4-2 corresponding to the contact pressure data for estimation Fd-2 for the period III.

[0150] Using the flowchart of FIG. 8, the estimation process of the second estimation unit 35 exemplified above corresponds to the process performed following the estimation process of the first estimation unit 34 which is performed following the Yes in S110, and is an example of a second estimation step according to one embodiment of the present invention.

[0151] In addition, the second estimation unit 35 may not use only the seven estimated correlations estimated by the first estimation unit 34 during the estimation process. Taking the first heat transfer coefficient data h1-1 as an example, the second estimation unit 35 may perform the estimation process using a data set including the first heat transfer coefficient data h1-1, the contact pressure data Fp-1, and the first upper punch side estimated correlation estimated by the first estimation unit 34.

[0152] [Embodiment 2] A second embodiment of the present invention will be described below. For ease of explanation, the same reference numerals are used to designate members having the same functions as those described in the first embodiment, and the description thereof will not be repeated. This also applies to a third embodiment described below.

[0153] The simulation device 200 according to the second embodiment of the present invention differs from the simulation device 100 according to the first embodiment of the present invention in that it estimates the temperature distribution of the die 40a instead of the die 40. The simulation device 200 also differs from the simulation device 100 in that it is provided with a third sensor 80 instead of the first and second sensors 10 and 20.

[0154] <Mold and temperature measurement> (Mold) The mold 40a is a metal mold used for manufacturing products by gravity casting. The mold 40a is formed of carbon steel, similar to the mold 40. As shown by reference numerals 1201 and 1202 in FIG. 12, the mold 40a is composed of only a die 41a. The die 41a is a hollow cylindrical mold, and the hollow part of the die 41a is cylindrical, similar to the die 41 of the mold 40. Also, similar to the die 41, various release agents are applied to the inner side surface of the die 41a as necessary.

[0155] As shown by reference numeral 1202 in Fig. 12, a first hole 41Ya, a second hole 41Yb, a third hole 41Yc, and a fifth hole 41Ye are formed in a side wall 41Xa-1 of the metal portion 41Xa of the die 41a in a front view in a direction perpendicular to a central axis AX-1 of the die 41a (i.e., the metal mold 40a). The first to third holes 41Ya to 41Yc are cylindrical holes large enough to insert and remove a third sensor body 83 described below, and penetrate from the outer side surface of the side wall 41Xa-1 to the hollow portion of the die 41a. In this embodiment, the diameters of the first to third holes 41Ya to 41Yc are all 20 mm.

[0156] Of the first to third holes 41Ya to 41Yc, the first hole 41Ya is formed at the top and the third hole 41Yc is formed at the bottom. The second hole 41Yb is formed between the first hole 41Ya and the third hole 41Yc in the vertical direction. In the vertical direction, the shortest distance between the first hole 41Ya and the second hole 41Yb is approximately the same as the shortest distance between the second hole 41Yb and the third hole 41Yc.

[0157] The fifth hole 41Ye is a cylindrical hole with a diameter of 8 mm, which is smaller than the diameters of the first to third holes 41Ya to 41Yc, and an opening is formed on the outer side surface of the side wall 41Xa-1. The fifth hole 41Ye does not penetrate to the hollow part of the die 41a, and the bottom 41Ye-1 of the fifth hole 41Ye is formed by a part of the side wall 41Xa-1. In this embodiment, the thickness of the bottom 41Ye-1 is 2 mm. The fifth hole 41Ye is formed in a portion of the side wall 41Xa-1 between the second hole 41Yb and the third hole 41Yc and in the vicinity of the second hole 41Yb when viewed from the front.

[0158] The central axes of the first to third holes 41Ya to 41Yc intersect on a central axis AX-1 in a plan view, as indicated by reference numeral 1201 in Fig. 12. In a plan view, the angle between the central axis of the first hole 41Ya and the central axis of the second hole 41Yb is 30°, the angle between the central axis of the second hole 41Yb and the central axis of the third hole 41Yc is 30°, and the angle between the central axis of the first hole 41Ya and the central axis of the third hole 41Yc is 60°. In addition, the central axis of the fifth hole 41Ye coincides with the central axis of the second hole 41Yb in a plan view.

[0159] As shown by reference numeral 1202 in FIG. 12, a fourth hole 41Yd, a sixth hole 41Yf, and a through hole 41Yg are formed in the bottom wall 41Xa-2 of the metal portion 41Xa in a direction parallel to the central axis AX-1 in a front view. The central axis of the fourth hole 41Yd coincides with the central axis AX-1. The fourth hole 41Yd is also a cylindrical hole of a size that allows the third sensor body 83 to be inserted and removed, similar to the first to third holes 41Ya to 41Yc, and penetrates from the outer bottom surface of the bottom wall 41Xa-2 to the hollow part of the die 41a. The diameter of the fourth hole 41Yd is 20 mm, similar to the first to third holes 41Ya to 41Yc.

[0160] The sixth hole 41Yf is a cylindrical hole with a diameter of 8 mm, which is smaller than the diameter of the first to fourth holes 41Ya to 41Yd, and an opening is formed on the outer bottom surface of the bottom wall 41Xa-2. The sixth hole 41Yf does not penetrate to the hollow part of the die 41a, and a bottom 41Yf-1 of the sixth hole 41Yf is formed by a part of the bottom wall 41Xa-2. The thickness of the bottom 41Yf-1 is 2 mm, which is the same as the bottom 41Ye-1 in this embodiment. The sixth hole 41Yf is formed in the vicinity of the sixth hole 41Yf on the side where the holes (the first to third holes 41Ya to 41yc and the fifth hole 41Ye) are formed in the bottom wall 41Xa-2 in the front view. The central axis of the sixth hole 41Yf intersects with the central axis of the first hole 41Ya in the plan view, as indicated by reference numeral 1201 in FIG.

[0161] The through hole 41Yg is a cylindrical hole with a diameter of 1.05 mm as indicated by reference numeral 1202 in Fig. 12, penetrating from the outer bottom surface of the bottom wall 41Xa-2 to the hollow portion of the die 41a. The through hole 41Yg is formed near the side wall 41Xa-1 in front view, and is formed near the first hole 41Ya in plan view as indicated by reference numeral 1201 in Fig. 12. Of course, the formation positions, shapes, sizes, etc. of the first to sixth holes 41Ya to 41Yf and the through hole 41Yg are not limited to the example of this embodiment.

[0162] In this embodiment, gravity casting using the mold 40a is performed as follows. First, the mold 40a is preheated to 573.15K (300°C). Next, as shown in FIG. 13, molten metal 50 melted at 1023.15K (750°C) is poured into the hollow portion from the upper opening of the hollow portion, and the molten metal 50 is solidified in this state. Through these steps, a molded body (not shown) before finishing processing is completed. In addition, a release agent is applied to the inner side surface of the die 41a of the mold 40a before gravity casting. Examples of the release agent applied to the mold 40a include BN spray and black body spray.

[0163] (temperature measurement) The simulation device 200 includes a third sensor 80 (details will be described later) as shown in FIG. 1. In this embodiment, the simulation device 200 performs temperature measurement using the third sensor 80 at a total of 12 locations, including a total of eight locations on the mold 40a side and a total of four locations on the molten metal 50 side, as shown by reference numeral 1202 in FIG. 12. The total of 12 measurement locations are arranged such that three measurement locations Pdn, Pen-1, and Pen-2 (n: a natural number from 1 to 4) shown by reference numerals 1201 and 1202 in FIG. 12 are linearly arranged in both plan and front views. A total of four groups of measurement locations, each consisting of these three measurement locations, are provided.

[0164] Hereinafter, the three measurement points Pdn, Pen-1, and Pen-2 will be abbreviated as "measurement points Pdn, ~Pen-2." In addition, the temperature value of temperature Tdn measured at measurement point Pdn will be temperature data Tdn' (second temperature data). The temperature value of temperature Ten-1 measured at measurement point Pen-1 will be temperature data Ten-1' (first temperature data). The temperature value of temperature Ten-2 measured at measurement point Pen-2 will be temperature data Ten-2' (first temperature data). In addition, the three temperatures Tdn, Ten-1, and Ten-2 will be abbreviated as "temperature Tdn, ~Ten-2," and the three temperature data Tdn', Ten-1', and Ten-2' will be abbreviated as "temperature data Tdn', ~Ten-2'."

[0165] In this embodiment, as shown by reference numeral 1203 in Fig. 12, the measurement points Pdn are provided at positions 2 mm toward the molten metal 50 from the inner side surface of the side wall 41Xa-1 or the inner bottom surface of the bottom wall 41Xa-2 (collectively referred to below as "second mold surface"). The measurement point Pen-1 is provided at a position 2 mm toward the mold 40a from the second mold surface. The measurement point Pen-2 is provided at a position 6 mm toward the mold 40a from the second mold surface.

[0166] As shown by reference numerals 1201 and 1202 in FIG. 12, the measurement point Pd1 is provided inside the molten metal 50 and in the vicinity of the measurement point Pe1-1. Pe1-1 and Pe1-2 are provided on the wall surface of the metal part 41Xa forming the first hole 41Ya. The measurement point Pd2 is provided inside the molten metal 50 and in the vicinity of the measurement point Pe2-1. Pe2-1 and Pe2-2 are provided on the wall surface of the metal part 41Xa forming the second hole 41Yb. The measurement point Pd3 is provided inside the molten metal 50 and in the vicinity of the measurement point Pe3-1. Pe3-1 and Pe3-2 are provided on the wall surface of the metal part 41Xa forming the third hole 41Yc.

[0167] The measurement points Pd4, ~Pe4-2 are provided on a straight line parallel to the central axis AX-1 in both plan and front views. This straight line parallel to the central axis AX-1 is located closer to the side wall 41Xa-1 side than the central axis AX-1. Furthermore, the measurement points Pd3, ~Pe3-2, and Pd4 are provided on the same straight line (i.e., the central axis of the third hole 41Yc) in plan view, as indicated by reference numeral 1201 in FIG. 12. Furthermore, the measurement point Pd4 is provided inside the molten metal 50 and in the vicinity of the measurement point Pe4-1, as indicated by reference numeral 1202 in FIG. The measurement points Pe4-1 and Pe4-2 are provided on the wall surface of the metal portion 41Xa that forms the fourth hole 41Yd.

[0168] In this embodiment, the third sensor 80 is set in the mold 40a by accommodating the third sensor main bodies 83 in the first to fourth holes 41Ya to 41Yd, respectively, as shown in Fig. 13. Then, with the third sensor 80 set in the mold 40a, the third temperature measurement section 81 (temperature measurement section; details will be described later) of the third sensor 80 measures temperatures Td1 to Td4, Te1-1 to Te4-1, and Te1-2 to Te4-2.

[0169] The temperature acquisition section 31 of the simulation apparatus 200 shown in FIG. 1 acquires from the third temperature measurement section 81 the temperature data Td1' to Td4', Te1-1' to Te4-1', and Te1-2' to Te4-2' that are the measurement results of the third temperature measurement section 81.

[0170] The simulation device 200 includes a thermocouple (hereinafter, "fourth temperature measurement unit") not shown, and measures temperature at three locations inside and near the through-hole 41Yg by the fourth temperature measurement unit, as indicated by reference numeral 1202 in Fig. 12. Regarding the arrangement of the three measurement locations, the three measurement locations Pd5, Pe5-1, and Pe5-2 indicated by reference numerals 1201 and 1202 in Fig. 12 are provided linearly in both plan and front views.

[0171] The measurement points Pd5, Pe5-1, and Pe5-2 are all provided on the central axis of the through hole 41Yg. The positions of the measurement points Pd5, Pe5-1, and Pe5-2 coincide with the position of the measurement point Pd1 in a plan view as shown by reference numeral 1201 in FIG. 12. The measurement point Pd5 is provided at a position 2 mm from the inner bottom surface of the bottom wall 41Xa-2 toward the molten metal 50. The measurement point Pe5-1 is provided at a position 2 mm from the inner bottom surface of the bottom wall 41Xa-2 toward the mold 40a. The measurement point Pe5-2 is provided at a position 6 mm from the inner bottom surface of the bottom wall 41Xa-2 toward the mold 40a.

[0172] A fourth temperature measurement unit arranged at measurement point Pd5 measures temperature Te5, a fourth temperature measurement unit arranged at measurement point Pe5-1 measures temperature Te5-1, and a fourth temperature measurement unit arranged at measurement point Pe5-2 measures temperature Te5-2. The temperature acquisition unit 31 of the simulation device 200 acquires temperature data Te5', Te5-1', and Te5-2' of the measurement results of the fourth temperature measurement unit from the fourth temperature measurement unit. The temperature data Te5' is the temperature value of temperature Te5, the temperature data Te5-1' is the temperature value of temperature Te5-1, and the temperature data Te5-2' is the temperature value of temperature Te5-2.

[0173] For the sake of simplicity, the die 40a in Fig. 13 is shown as if the central axes of the first to sixth holes 41Ya to 41Yf and the through hole 41Yg are all on the cross section of the metal portion 41Xa of the die 41a. In reality, the second hole 41Yb, the third hole 41Yc, and the fifth hole 41Ye are formed at the locations indicated by the reference characters 1201 and 1202 in Fig. 12.

[0174] <Specific structure of the third sensor and pressure measurement> (Specific structure of the third sensor) The third sensor 80 is a component of the simulation device 200 that measures temperatures Tdn, Ten-1 and Ten-2 (n: a natural number from 1 to 4) and contact pressures Fd1 to Fd4. The contact pressures Fd1 to Fd4 are four pressures: contact pressures Fd1, Fd2, Fd3 and Fd4.

[0175] Specifically, the third sensor 80 set in the first hole 41Ya measures temperatures Td1, ~Te1-2, and contact pressure Fd1. The third sensor 80 set in the second hole 41Yb measures temperatures Td2, ~Te2-2, and contact pressure Fd2. The third sensor 80 set in the third hole 41Ya measures temperatures Td3, ~Te3-2, and contact pressure Fd3. The third sensor 80 set in the fourth hole 41Ya measures temperatures Td4, ~Te4-2, and contact pressure Fd4.

[0176] As shown by reference numeral 1401 in Fig. 14, the third sensor 80 has a third temperature measurement unit 81, a third pressure measurement unit 82 (pressure measurement unit), and a third sensor body 83 (single member). The third temperature measurement unit 81 measures temperatures Td1 to Td4, Te1-1 to Te4-1, and Te1-2 to Te4-2. In this embodiment, the third temperature measurement unit 81 is the same thermocouple as the first and second temperature measurement units 11 and 21. Moreover, each of the temperatures Te1-1 to Te4-1 and Te1-2 to Te4-2 measured by the third temperature measurement unit 81 is regarded as the temperature of the die 41a (i.e., the metal mold 40a).

[0177] The third sensor body 83 is a hollow cylindrical member. In this embodiment, the third sensor body 83 is made of the same carbon steel as the mold 40a, and has an outer diameter of 20 mm and an inner diameter of 10 mm. As shown by reference numeral 1402 in FIG. 14, the length of the third sensor body 83 in the direction of the central axis AX3 is approximately the same as the thickness of the side wall 41Xa-1. Of course, the material, shape, and size of the third sensor body 83 are not limited to the example of this embodiment.

[0178] The third sensor body 83 is provided with three third temperature measurement parts 81, and the third sensor 80 measures temperature at three locations. The three locations are measurement locations Pdn, Pen-1, and Pen-2 (n: a natural number from 1 to 4) in this order from the position on the molten metal 50 side. Specifically, the measurement locations Pdn-1 and Pen-2 are provided on the wall surfaces of the metal portion 41Xa that form the first to fourth holes 41Ya to 41Yd, respectively, as shown by reference numerals 1201 and 1202 in Fig. 12. The measurement location Pdn is provided inside the molten metal 50 and in the vicinity of the measurement location Pen-1.

[0179] Moreover, the measurement points Pdn and ~Pen-2 (n: natural number from 1 to 4) are provided on the same axis. The axis on which the measurement points Pdn and ~Pen-2 (n: natural number from 1 to 3) are provided is perpendicular to the central axis AX-1, and is at the same height and parallel to the central axes of the first to third holes 41Ya to 41Yc in a front view. The axis (not shown) on which the measurement points Pd4 and ~Pe4-2 are provided is parallel to the central axis AX-1 in both a plan view and a front view. Note that, when the third sensor main body 83 is housed in the first to fourth holes 41Ya to 41Yd, the central axis AX3 of the third sensor main body 83 coincides with the central axis of the first to fourth holes 41Ya to 41Yd.

[0180] The third pressure measuring unit 82 measures contact pressures Fd1 to Fd4. The contact pressures Fd1 to Fd3 are pressures acting on the fourth interface Si4. Specifically, the contact pressure Fd1 is a pressure acting on a region of the fourth interface Si4 facing the first hole 41Ya as shown in FIG. 13. The contact pressure Fd2 is a pressure acting on a region of the fourth interface Si4 facing the second hole 41Yb. The contact pressure Fd3 is a pressure acting on a region of the fourth interface Si4 facing the third hole 41Yc.

[0181] The fourth interface Si4 is a concept that includes a surface Sd1 of the side wall 41Xa-1 that contacts the molten metal 50 and a surface Sc4 of the molten metal 50 that contacts the side wall 41Xa-1. Here, the "surface Sd1 of the side wall 41Xa-1 that contacts the molten metal 50" corresponds to the inner side surface of the side wall 41Xa-1. In the following description, this surface is referred to as the "fifth contact surface Sd1." The fifth contact surface Sd1 includes the surfaces 82a-1 of the three films 82a (details will be described later) that close the first to third holes 41Ya to 41Yc, respectively, and the surface of the third sensor body 83 that contacts the molten metal 50. In addition, the "surface Sc4 of the molten metal 50 that contacts the side wall 41Xa-1" corresponds to the side surface of the molten metal 50. In the following description, this surface is referred to as the "sixth contact surface Sc4."

[0182] From the definition of the fourth interface Si4 described above, "contact pressures Fd1 to Fd3" refers to two pressures: contact pressures Fd1 to Fd3 acting from the fifth contact surface Sd1 toward the sixth contact surface Sc4, and contact pressures Fd1 to Fd3 acting from the sixth contact surface Sc4 toward the fifth contact surface Sd1.

[0183] The contact pressure Fd4 is a pressure acting on the fifth interface Si5 shown in Fig. 13. Specifically, the contact pressure Fd4 is a pressure acting on a region of the fifth interface Si5 facing the fourth hole 41Yd. The fifth interface Si5 is conceptually constituted by a surface Sd2 of the bottom wall 41Xa-2 in contact with the molten metal 50 and a surface Sc5 of the molten metal 50 in contact with the bottom wall 41Xa-2.

[0184] Here, "surface Sd2 of bottom wall 41Xa-2 in contact with molten metal 50" corresponds to the inner bottom surface of bottom wall 41Xa-2. In the following description, this surface is referred to as "seventh contact surface Sd2." The seventh contact surface Sd2 includes surface 82a-1 and the surface of the third sensor body 83 on the side in contact with molten metal 50. Also, "surface Sc5 of molten metal 50 in contact with bottom wall 41Xa-2" corresponds to the lower end surface of molten metal 50. In the following description, this surface is referred to as "eighth contact surface Sc5."

[0185] From the definition of the fifth interface Si5 mentioned above, "contact pressure Fd4" refers to two pressures: the contact pressure Fd4 acting from the seventh contact surface Sd2 toward the eighth contact surface Sc5, and the contact pressure Fd4 acting from the eighth contact surface Sc5 toward the seventh contact surface Sd2.

[0186] The third pressure measurement unit 82 has a membrane 82a and a laser displacement meter 82b, as indicated by reference numeral 1401 in Fig. 14. The membrane 82a is attached to the third sensor body 83 and closes the opening of the third sensor body 83 on the molten metal 50 side. When the third sensor body 83 is accommodated in the first to fourth holes 41Ya to 41Yd, the membrane 82a comes into contact with the molten metal 50 poured into the die 40a.

[0187] Before the third sensor 80 is set in the mold 40a, the film 82a has a flat plate shape indicated by the broken line of reference numeral 1402 in FIG. 14, and the surface 82a-1 on the side in contact with the molten metal 50 is flush with the surface on the side in contact with the molten metal 50 in the third sensor body 83. As shown by reference numeral 1402 in FIGS. 13 and 14, when the film 82a comes into contact with the molten metal 50, the surface 82a-1 is pressed by the molten metal 50 and bends so as to protrude toward the outside of the mold 40a.

[0188] Hereinafter, the force with which the molten metal 50 in contact with the film 82a presses the film 82a is referred to as the "pressing force". Also, as shown in FIG. 13, the pressing force acting on the film 82a of the third pressure measurement unit 82 disposed in the first hole 41Ya is defined as the "pressing force Fdy-1". The pressing force acting on the film 82a of the third pressure measurement unit 82 disposed in the second hole 41Yb is defined as the "pressing force Fdy-2". The pressing force acting on the film 82a of the third pressure measurement unit 82 disposed in the third hole 41Yc is defined as the "pressing force Fdy-3". The pressing force acting on the film 82a of the third pressure measurement unit 82 disposed in the fourth hole 41Yd is defined as the "pressing force Fdy-4".

[0189] In order to avoid damage to the film 82a during gravity casting, it is necessary for the film 82a not to react with the molten metal 50. In this embodiment, since the molten metal 50 is an aluminum alloy die-cast, it is necessary to select, as the material for forming the film 82a, a material that does not react with aluminum and has a melting point higher than the temperature of the molten metal 50 during casting.

[0190] Examples of metal materials as the material for forming the film 82a include Ti (titanium), Zr (zirconium), Ta (tantalum), Ag (silver), Au (gold), Cr (chromium), Co (cobalt), Ni (nickel), and Pt (platinum). Examples of non-metals include woven fabrics of carbon fibers and ceramic papers. In particular, Mo (molybdenum), Nb (niobium), and W (tungsten) are preferable as the material for forming the film 82a. On the other hand, when using Fe (iron) and its alloys or Cu (copper) that easily react with aluminum, it is preferable to coat the surface of the film 82a with a release agent or to apply a wear-resistant coating with a ceramic-based material.

[0191] Regarding the thickness of the film 82a, the thinner it is, the greater the amount of deflection when the same pressing force is applied, and thus higher resolution can be obtained. On the other hand, the thinner the film 82a becomes, the easier it is to be damaged. Therefore, it is preferable that the film 82a has a thickness that enables both resolution and resistance to damage to be within an acceptable range. In this embodiment, in comprehensive consideration of the above-mentioned considerations, a metal film made of Mo with a thickness of 10 μm or more is used as the film 82a.

[0192] The laser displacement meter 82b is a measuring device that non - contact measures the amount of deflection of the film 82a caused by the action of the pressing force. Specifically, the light from the light - emitting element is condensed by the light - projecting lens and projected onto the film 82a. Then, a part of the light reflected by the film 82a reaches the linear image sensor through the light - receiving lens, and the linear image sensor detects the deflection of the film 82a. As the amount of deflection of the film 82a increases or decreases, the light spot on the linear image sensor moves. Therefore, the laser displacement meter 82b detects this movement amount as the amount of deflection of the film 82a. The light - emitting element, the light - projecting lens, the light - receiving lens, and the linear image sensor are all components (not shown) of the laser displacement meter 82b. There is no limitation on the type of the laser displacement meter 82b, and a known laser displacement meter can be used.

[0193] (Pressure Measurement) First, insert the third sensor body 83 into the first to fourth holes 41Ya to 41Yd. Then, as shown by reference numeral 1402 in FIG. 14, the third sensor body 83 is accommodated in the first to fourth holes 41Ya to 41Yd so that the surface 82a - 1 of the film 82a is flush with the fifth contact surface Sd1 (the inner side surface of the side wall 41Xa - 1).

[0194] Next, when the molten metal 50 is poured into the mold 40a, as shown in FIG. 13, a pressing force acts on the film 82a, causing the film 82a to bend convexly toward the outside of the mold 40a. The amount of this bending is measured by the third sensor 80. Specifically, the amount of displacement of the film 82a caused by the action of the pressing force is measured by the laser displacement meter 82b of the third pressure measurement unit 82. At the same time, the amount of displacement of each of the bottoms 41Ye-1 and 41Yf-1 caused by the thermal expansion of the metal part 41Xa is measured by the laser displacement meter 82b. The laser displacement meter 82b for measuring the amount of displacement described above is not provided in the third pressure measurement unit 82 but is used alone.

[0195] Then, the third sensor 80 measures the amount of bending of the film 82a by regarding the difference obtained by subtracting the amount of displacement of each of the bottoms 41Ye-1 and 41Yf-1 from the amount of displacement of the film 82a as the amount of bending of the film 82a. Note that it is not essential to consider the thermal expansion of the metal part 41Xa when measuring the amount of bending of the film 82a, and the amount of displacement of the film 82a caused by the action of the pressing force may be directly regarded as the amount of bending of the film 82a.

[0196] Next, the third sensor 80 converts the pressing forces Fdy-1 to Fdy-4 acting on the film 82a of each of the third sensor bodies 83 accommodated in the first to fourth holes 41Ya to 41Yd from the obtained amount of bending of the film 82a. As conversion methods, at least the following two types of methods can be mentioned.

[0197] As a first conversion method, a method using the following formulas (6) and (7) can be mentioned.

[0198]

Equation

[0199] W: Amount of bending of film 82a [μm] Po: Pressing force [Pa] E: Elastic modulus of film 82a [Pa] h: Thickness of film 82a [μm] ν: Poisson's ratio a: Radius of membrane 82a [mm] r: Radial linear distance between the center of membrane 82a and the measurement position (the position where the amount of deflection of membrane 82a is measured) [mm] ρ: r / a A, B, C, D: Coefficients

[0200]

Equation

[0201] When the conversion unit (not shown) of the third sensor 80 calculates the pressing forces Fdy-1 to Fdy-4 acting on the membrane 82a of each of the third pressure measurement units 82 arranged in the first to third holes 41Ya to 41Yc, the above-mentioned formula (6) is used. When calculating the pressing force Fdy-4 acting on the membrane 82a of the third pressure measurement unit 82 arranged in the fourth hole 41Yd, the conversion unit uses the above-mentioned formula (7). The conversion unit is, for example, a CPU and is provided in the laser displacement meter 82b.

[0202] As a second conversion method, there is a method using the calibration graph shown in FIG. 15. Specifically, before pouring the molten metal 50, the low-melting-point metal molten metal 50a is poured into the calibration mold 40b shown in FIG. 16 in advance and gravity casting is performed.

[0203] The reason for using the low-melting-point metal molten metal 50a is as follows. That is, since the fluid in contact with the membrane 82a is the molten metal, and the surface shape of the fluid becomes unchanged when it solidifies. Therefore, in the case of the molten metal, there is a possibility that the amount of deflection of the membrane 82a cannot be accurately measured due to solidification. In that regard, when the molten metal 50a is a high-melting-point metal such as an aluminum alloy (913.15 K (640 °C) in the case of an aluminum alloy), the portion of the molten metal 50a in contact with the membrane 82a immediately after pouring becomes solid, and the measurement accuracy decreases at an early stage. Therefore, it is preferable to use the molten metal of a low-melting-point metal as the molten metal 50a. In this embodiment, the molten metal of tin (melting point; 503.15 K (230 °C)) is used as the molten metal 50a.

[0204] The temperature of the molten metal 50a is the same as that of the molten metal 50, which is 1023.15 K. The mold 40b is the same as the mold 40a except that the first to third holes 41Ya to 41Yc, the fifth hole 41Ye, and the sixth hole 41Yf are not formed. Also, a third sensor 80 is set in the fourth hole 41Yd of the mold 40b to measure the amount of deflection of the film 82a caused by the pressing of the molten metal 50a and to measure the temperature of the film 82a. Regarding the temperature of the film 82a, the temperature Te4-1 at the measurement point Pe4-1 measured by the third temperature measurement unit 81 is regarded as the temperature of the film 82a. Then, while changing the pouring amount of the molten metal 50a, this gravity casting is performed multiple times.

[0205] Here, the pressing force Pi acting on the film 82a set in the mold 40b is Pi = l×g×hi (l: specific gravity of the molten metal 50a [kg / m 3 , g: gravitational acceleration [m / s 2 , hi: height of the molten metal 50a [m]). When the pouring amount of the molten metal 50a is changed, as in the example of FIG. 16, the height hi changes to h1 to h3 (h1 > h2 > h3), so the pressing force Pi acting on the film 82a also changes like the pressing forces P1 to P3 (P1 > P2 > P3). Also, the higher the temperature of the film 82a, the lower the elastic modulus of the film 82a. Therefore, as in the example of FIG. 15, even with the same pressing force Pi, the higher the temperature of the film 82a, the greater the amount of deflection of the film 82a.

[0206] From the above, by performing the gravity casting of the molten metal 50a using the mold 40b multiple times, for a plurality of different pressing forces Pi, a calibration graph showing the relationship between the amount of deflection of the film 82a and the temperature of the film 82a as shown in FIG. 15 can be generated. Regarding the data of the generated calibration graph, for example, it may be stored in a storage unit (not shown) of the third sensor 80, or it may be stored in a storage unit (not shown) built into the simulation device 200.

[0207] The conversion unit reads out the calibration graph from any of the aforementioned storage units, and compares the amount of deflection of the molten metal 50 measured by the third pressure measurement unit 82 and the temperature of the film 82a measured by the third temperature measurement unit 81 (i.e., the temperature Te4-1) with the aforementioned graph. Then, the conversion unit converts the pressing forces Fdy-1 to Fdy-4 from the collation result.

[0208] Finally, the third sensor 80 measures the contact pressures Fd1 to Fd4 by regarding the obtained converted pressing forces Fdy-1 to Fdy-4 as the contact pressures Fd1 to Fd4. Specifically, the third sensor 80 regards the pressing force Fdy-1 as the contact pressure Fd1, the pressing force Fdy-2 as the contact pressure Fd2, the pressing force Fdy-3 as the contact pressure Fd3, and the pressing force Fdy-4 as the contact pressure Fd4. The pressure acquisition unit 32 of the simulation device 200 shown in FIG. 1 acquires the contact pressure data Fd1' to Fd4' of the contact pressures Fd1 to Fd4 measured by the third sensor 80 from the third sensor 80.

[0209] <Example of Processing Results by Simulation Device> Hereinafter, an example of the processing results by the simulation device 200 will be described with reference to FIGS. 12, 17 to 19. Note that each graph example shown in FIGS. 17 to 19 is an example when the molten metal 50a is gravity cast using the mold 40a coated with the BN spray as the mold release agent.

[0210] First, the measurement result of the third pressure measurement unit 82 is as shown in the graph of FIG. 17. The horizontal axis of the graph shown in FIG. 17 represents the suitable elapsed time. The suitable elapsed time is the elapsed time from when the molten metal 50a starts to be poured into the mold 40a until before the molten metal 50a solidifies, and within this elapsed time range, the amount of deflection of the film 82a can be measured with high accuracy.

[0211] The suitable elapsed time varies depending on the type of the mold 40a, the constituent metal of the molten metal 50a, the casting conditions, and the like. In the present embodiment, the end point of the suitable elapsed time is set to 20 seconds from when the molten metal 50a starts to be poured into the mold 40a as shown in the graph of FIG. 17.

[0212] When the pouring of the molten metal 50a starts, it comes into contact with the molten metal 50a in the order of "the film 82a set in the fourth hole 41Yd → the film 82a set in the third hole 41Yc → the film 82a set in the second hole 41Yb → the film 82a set in the first hole 41Ya". Since the contact pressures Fd1 to Fd4 start to increase when the contact between the molten metal 50a and the film 82a begins, as shown in the graph of Fig. 17, the values start to increase in the order of approximately "contact pressure Fd4 → contact pressure Fd3 → contact pressure Fd2 → contact pressure Fd1".

[0213] The value that becomes the largest throughout the entire elapsed time is the contact pressure Fd4 which is the farthest from the upper surface of the molten metal 50a (= the height hi of the molten metal 50a). However, the maximum value of the contact pressure Fd4 is about 0.003 MPa in the example of Fig. 17, which is on the order of 1 / 1000 of the contact pressure in pressure casting. This is because in pressure casting, it is pressurized with a hydraulic press, while in gravity casting, it is only pressurized with the pressure corresponding to the weight of the molten metal.

[0214] Next, the first heat transfer coefficient data hd1 to hd5 continuously determined by the generation unit 33 of the simulation device 200 for each time step are as shown in the graph of Fig. 18. The graph shown in Fig. 18 shows the determination results until 50 seconds have elapsed since the pouring of the molten metal 50a into the mold 40a started.

[0215] As shown by reference numeral 1202 in FIG. 12, the first heat transfer coefficient data hd1 is a value indicating the heat transfer coefficient of the region facing the first hole 41Ya at the fourth interface Si4. The first heat transfer coefficient data hd2 is a value indicating the heat transfer coefficient of the region facing the second hole 41Yb at the fourth interface Si4. The first heat transfer coefficient data hd3 is a value indicating the heat transfer coefficient of the region facing the third hole 41Yc at the fourth interface Si4. The first heat transfer coefficient data hd4 is a value indicating the heat transfer coefficient of the region facing the fourth hole 41Yd at the fifth interface Si5. The first heat transfer coefficient data hd5 is a value indicating the heat transfer coefficient of the region facing the through hole 41Yg at the fifth interface Si5. The method for determining the first heat transfer coefficient data hd1 to hd5 is the same as that in Embodiment 1. As shown in the graph of FIG. 18, each first heat transfer coefficient data shows different behaviors depending on the calculation location.

[0216] Next, the correlation between the contact pressure data Fd1' to Fd4' estimated by the first estimation unit 34 of the simulation device 200 and the first heat transfer coefficient data hd1 to hd4 is as shown in the graph of FIG. 19. As shown in the graph of FIG. 19, even when minute contact pressures Fd1 to Fd4 act on the fourth and fifth interfaces Si4 and Si5 under gravity casting, the simulation device 200 can analyze the correlation between the contact pressure data Fd1' to Fd4' and the first heat transfer coefficient data hd1 to hd4.

[0217] 〔Embodiment 3〕 Embodiment 3 of the present invention will be described below. The simulation device 300 according to Embodiment 3 of the present invention is different from the simulation devices 100 and 200 in that the first estimation unit 34 constructs estimation models 34a, 34b, and 34c. Further, the simulation device 300 is also different from the simulation devices 100 and 200 in that the second estimation unit 35 estimates the second heat transfer coefficient data hd-2, hpu-2, and hpl-2 using the above-described respective estimation models.

[0218] Next, a series of processes until the simulation device 300 estimates the correlation between the contact pressure data and the first heat transfer coefficient data will be specifically described with reference to FIG. 1. A series of processes until the generation unit 33 of the simulation device 300 shown in FIG. 1 generates each basic data set is the same as in Embodiments 1 and 2.

[0219] The first estimation unit 34 that has received each basic data set from the generation unit 33 of the simulation device 300 constructs estimation models 34a, 34b, and 34c by machine learning. The estimation model 34a is a learned model obtained by machine learning using the die-side basic data set as teacher data. Using the estimated contact pressure data Fd-2 as input data, it outputs the second heat transfer coefficient data hd-2. The estimation model 34b is a learned model obtained by machine learning using the upper punch-side basic data set as teacher data. Using the estimated contact pressure data Fp-2 as input data, it outputs the second heat transfer coefficient data hpu-2. The estimation model 34c is a learned model obtained by machine learning using the lower punch-side basic data set as teacher data. Using the estimated contact pressure data Fp-2 as input data, it outputs the second heat transfer coefficient data hpl-2. The first estimation unit 34 of the simulation device 300 temporarily stores the constructed estimation models 34a to 34c in the storage unit 3.

[0220] The second estimation unit 35 of the simulation device 300 reads the estimation model 34a from the storage unit 3, inputs the estimated contact pressure data Fd-2 to the estimation model 34a, and obtains the second heat transfer coefficient data hd-2, thereby estimating the second heat transfer coefficient data hd-2. The aforementioned second estimation unit 35 reads the estimation model 34b from the storage unit 3, inputs the estimated contact pressure data Fp-2 to the estimation model 34b, and obtains the second heat transfer coefficient data hpu-2, thereby estimating the second heat transfer coefficient data hpu-2. The aforementioned second estimation unit 35 reads the estimation model 34c from the storage unit 3, inputs the estimated contact pressure data Fp-2 to the estimation model 34c, and obtains the second heat transfer coefficient data hpl-2, thereby estimating the second heat transfer coefficient data hpl-2.

[0221] There is no particular limitation on the method of machine learning performed by the first estimation unit 34 of the simulation device 300, and a known method such as a convolutional neural network can be adopted. Further, each basic dataset used when the first estimation unit 34 of the simulation device 300 performs machine learning is not limited to only those generated immediately before the machine learning. For example, the first estimation unit 34 of the simulation device 300 may read out each basic dataset generated in the past from the storage unit 3 and use it for machine learning. Also, for example, the first estimation unit 34 of the simulation device 300 may use, for the machine learning, a combination of each basic dataset generated in the past and each basic dataset generated immediately before the machine learning.

[0222] The estimation models 34a to 34c are learned models that output second heat transfer coefficient data based on the estimated correlation relationship (i.e., the estimated correlation) between the contact pressure data and the first heat transfer coefficient data. Therefore, the estimation models 34a to 34c are the estimation results by the first estimation unit 34 of the simulation device 300. Thus, the simulation device 300 can estimate the correlation relationship between the contact pressure data and the first heat transfer coefficient data by machine learning using each basic dataset. Also, the simulation device 300 can estimate each second heat transfer coefficient data by using the estimation models 34a to 34c.

[0223] 〔Example of implementation by software〕 The functions of the simulation device 100 (hereinafter, "device 100") can be realized by a program for causing a computer to function as the device 100. This program is a program for causing a computer to function as each control block of the device 100 (particularly, each part included in the first estimation device 30).

[0224] In this case, the apparatus 100 includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the above-described program. By executing the above-described program with this control device and storage device, each function described in the above-described embodiment is realized.

[0225] The above-described program may be recorded on one or more computer-readable recording media, rather than temporarily. This recording medium may or may not be provided in the apparatus 100. In the latter case, the above-described program may be supplied to the apparatus 100 via any wired or wireless transmission medium.

[0226] Also, part or all of the functions in each control block of the apparatus 100 can also be realized by a logic circuit. For example, an integrated circuit in which a logic circuit functioning as each of the above-described control blocks is formed is also included in the scope of the present invention. In addition to this, for example, it is also possible to realize the functions of each of the above-described control blocks by a quantum computer.

[0227] 〔Example〕 <Apparatus> The apparatus used in this example is as follows. As the mold 40, a mold composed of a die 41 having a height of 100 mm and an outer diameter of 100 mm, an upper punch 42 having an outer diameter of 40 mm, and a lower punch 43 having a height of 50 mm and an outer diameter of 40 mm was used.

[0228] As the first sensor 10, one having a first sensor body 13 with an outer diameter of 12 mm at the first cylindrical portion, an outer diameter of 18 mm at the cylindrical portion with the largest outer diameter, and a total length of 150 mm was used. As the second sensor 20, one having a second sensor body 23 with an outer diameter of 12 mm at the molten metal side end and an outer diameter of 20 mm at the end opposite to the molten metal side end and a total length of 128 mm was used. Also, strain gauges as the first and second pressure measurement units 12 and 22 were welded and attached at positions more than 60 mm away from the end faces of the first cylindrical portion and the end face of the molten metal side end, respectively.

[0229] In this embodiment, for each of the first and second sensors 10 and 20, the compression characteristics were specified every 50 K from room temperature to about 673 K. Then, the correlation between the strain generated in each of the first and second sensors 10 and 20 and the compression stress acting on these sensors was calibrated.

[0230] For each measurement location where temperature is not measured by the first and second sensors 10 and 20, through-holes or non-through holes were formed at each location of the die 41, the upper punch 42, and the lower punch 43 corresponding to each measurement location. Then, a thermocouple similar to the first and second temperature measurement units 11 and 21 was inserted into these holes, and temperature measurement was performed by arranging the thermocouple at each measurement location.

[0231] <Casting method> In this embodiment, the die 41 with the lower punch 43 set in the hollow portion was set in a hydraulic press, and the molten metal 50 (ADC12) at about 1023 K was poured into the hollow portion. Next, while pressurizing the upper punch 42 with the hydraulic press, it was inserted into the hollow portion to pressurize the molten metal 50. Then, while maintaining the state where the molten metal 50 was pressurized by the upper punch 42, the molten metal was solidified. In addition, except for the verification of the estimation result of the first estimation unit 34 described later, a black body spray was sprayed as a mold release agent on the inner side surface of the die 41. Also, the press load of the hydraulic press was set to 7.6 t.

[0232] <Investigation results and evaluation results> (Relationship between the first heat transfer coefficient data and the elapsed time) In this embodiment, the relationship between each of the first heat transfer coefficient data h1-1 to h7-1 determined by the generation unit 33 and the elapsed time was investigated. The "elapsed time" refers to the elapsed time since the molten metal 50 started to be poured into the die 41. As a result of the investigation, a graph as shown in FIG. 20 was obtained. As shown in the graph of FIG. 20, the values of the first heat transfer coefficient data h1-1 to h7-1 were different at any elapsed time. From this, it was found that the behavior of the first heat transfer coefficient data with respect to the change over time differed depending on the measurement location of the temperature data. This is presumably due to the fact that the contact state between the mold 40 and the molten metal 50 differs depending on the measurement location of the temperature data.

[0233] (Comparison with the first heat transfer coefficient data determined using the one-dimensional unsteady heat conduction model) In this embodiment, for each of the first heat transfer coefficient data h1-1 to h3-1, the relationship between the value determined by the generation unit 33 by the determination method of this embodiment and the value determined using the one-dimensional unsteady heat conduction model and the elapsed time was investigated. The generation unit 33 calculated the first heat transfer coefficient data hm1-1 to hm3-1 using the one-dimensional unsteady heat conduction model by the method shown below. The first heat transfer coefficient data hm1-1 is the first heat transfer coefficient data h1-1 calculated using the one-dimensional unsteady heat conduction model. The first heat transfer coefficient data hm2-1 is the first heat transfer coefficient data h2-1 calculated using the one-dimensional unsteady heat conduction model. The first heat transfer coefficient data hm3-1 is the first heat transfer coefficient data h3-1 calculated using the one-dimensional unsteady heat conduction model.

[0234] In the following description, for the sake of simplicity of explanation, only the determination method of the first heat transfer coefficient data h1-1 will be described. The determination methods of the first heat transfer coefficient data h2-1 and h3-1 are substantially the same as the determination method of the first heat transfer coefficient data h1-1.

[0235] First, the generation unit 33 replaced the state in the vicinity of the interface region corresponding to the first heat transfer coefficient data h1-1 with an equivalent circuit 70 as shown in FIG. 21. The measurement location Pc is a general term for the measurement locations Pc1 to Pc3, and the temperature data Tc´ is a general term for the temperature data Tc1´ to Tc3´. The measurement location Pm1 is a general term for the measurement locations Pm1-1 to Pm3-1, and the temperature data Tm1´ is a general term for the temperature data Tm1-1´ to Tm3-1´. The measurement location Pm2 is a general term for the measurement locations Pm1-2 to Pm3-2, and the temperature data Tm2´ is a general term for the temperature data Tm1-2´ to Tm3-2´.

[0236] The thermal resistance RHTC is a general term for the thermal resistances RHTC1 to RHTC3. The thermal resistances RHTC1 to RHTC3 are the thermal resistances of the interface regions corresponding to the first heat transfer coefficient data h1-1 to h3-1, respectively. The thermal resistance R1 is a general term for the thermal resistances R11 to R13. The thermal resistance R11 is the thermal resistance between the measurement locations Pm1-1 and Pm1-2. The thermal resistance R12 is the thermal resistance between the measurement locations Pm2-1 and Pm2-2. The thermal resistance R13 is the thermal resistance between the measurement locations Pm3-1 and Pm3-2.

[0237] Next, the generation unit 33 acquired the temperature data Tc1´, Tm1-1´, and Tm1-2´ from the temperature acquisition unit 31 at a certain time step. Next, the generation unit 33 set the value of the thermal resistance RHTC1. In this embodiment, the generation unit 33 set the value received by the input unit 1 as the set value of the thermal resistance RHTC1 when the input unit 1 received the setting operation of the value of the thermal resistance RHTC1.

[0238] Next, the generation unit 33 calculated the estimated temperature data Tmpe1´ using the following formula (8) in which each temperature data acquired from the temperature acquisition unit 31, the thermal resistance RHTC1, and the one-dimensional unsteady heat conduction equation were applied to the equivalent circuit 70. The estimated temperature data Tmpe1´ is the estimated temperature data at the measurement location Pm1-1. The estimated temperature data Tmpe1´, the estimated temperature data Tmpe2´ at the measurement location Pm2-1, and the estimated temperature data Tmde1´ at the measurement location Pm3-1 are collectively referred to as "estimated temperature data Tme´".

[0239]

Number

[0240] Tme´: Estimated temperature data Tme´ [K] Tm1´: Temperature data Tm1´ at a certain time step Tm2´: Temperature data Tm2´ [K] at a certain time step Tc´: Temperature data Tc´ [K] at a certain time step RHTC: Thermal resistances RHTC1, RHTC2, and RHTC3 [(m 2 ·K) / W] β0, β1: Constants R1: Δx / λm [(m 2 ·K) / W] Δx: Distance between measurement points Pc and Pm1, distance between measurement points Pm1 and Pm2 [m] In this embodiment, the generation unit 33 used the equation obtained by applying the unsteady heat conduction equation in the one-dimensional orthogonal coordinate system to the equivalent circuit 70 as the aforementioned equation (8) to calculate the estimated temperature data Tmpe1´. When calculating the estimated temperature data Tmpe2´, the generation unit 33 also used the equation obtained by applying the unsteady heat conduction equation in the one-dimensional orthogonal coordinate system to the equivalent circuit 70 as the aforementioned equation (8). On the other hand, when calculating the estimated temperature data Tmde1´, the generation unit 33 used the equation obtained by applying the unsteady heat conduction equation in the one-dimensional cylindrical coordinate system to the equivalent circuit 70 as the aforementioned equation (8). Further, the generation unit 33 made the values of β0 and β1 different when calculating the estimated temperature data Tmde1´ and when calculating the estimated temperature data Tmpe1´ and Tmpe2´.

[0241] Next, the generation unit 33 calculated the difference between the calculated estimated temperature data Tmpe1' and the temperature data Tm1-1' (measured value) at the next time step. Then, the generation unit 33 determined whether the calculated difference was the smallest among the differences between the plurality of estimated temperature data Tmpe1' obtained by repeatedly performing the setting process of the thermal resistance RHTC1 and the calculation process using the above-described formula (8) a plurality of times at this time step and the above-described temperature data Tm1-1'. When it was determined that it was not the smallest, the generation unit 33 recalculated the estimated temperature data Tmpe1', calculated the above-described difference, and determined whether the difference was the smallest again.

[0242] In this embodiment, the generation unit 33 recalculated the estimated temperature data Tmpe1' and the like by comparing the magnitudes of the values of the plurality of evaluation functions e = |Tmpe1' - Tm1-1'| calculated at a certain time step among the plurality of evaluation functions e. The comparison method of the evaluation function e is the same as the comparison method of the evaluation function e described in this embodiment except that Δh is replaced with ΔRHTC1 (change amount of thermal resistance).

[0243] When it was determined that it was the smallest, the generation unit 33 calculated the first heat transfer coefficient data hm1-1 by substituting the value of the thermal resistance RHTC1 corresponding to the estimated temperature data Tmpe1' with the smallest difference into the following formula (9). The generation unit 33 calculated the first heat transfer coefficient data hm2-1 and hm3-1 in the same manner.

[0244]

Equation

[0245] h: First heat transfer coefficient data hm1-1, hm2-1, and hm3-1 [W / (m 2 ·K)] Rc: Thermal resistance of the molten metal 50 [(m 2 ·K) / W] Rm: Thermal resistance of the mold 40 [(m 2 ·K) / W] When the calculation process of the first heat transfer coefficient data hm1-1 to hm3-1 at a certain time step is completed, the generation unit 33 updates the time step once and determines whether the updated time step is the final time step. If the updated time step is not the final time step, the generation unit 33 calculates the first heat transfer coefficient data hm1-1 to hm3-1 at the updated time step. In this way, the calculation of the first heat transfer coefficient data hm1-1 to hm3-1 is repeated for each time step, and the generation unit 33 continues to calculate the first heat transfer coefficient data hm1-1 to hm3-1 until the final time step is reached.

[0246] As a result of the investigation, each graph shown in FIG. 22 was obtained. In the figure, "1D" indicates a value determined using a one-dimensional unsteady heat conduction model, and "3D" indicates a value determined by the determination method of the present embodiment. In each graph regarding the measurement points Pc2, ~Pm2-2 and Pc3, ~Pm3-2 in FIG. 22, the value of the first heat transfer coefficient data is generally larger for "1D" than for "3D". On the other hand, in the graph regarding the measurement points Pc1, ~Pm1-2 in FIG. 22, there was no significant difference in the value of the first heat transfer coefficient data between "1D" and "3D" regardless of the elapsed time.

[0247] This is presumably due to the fact that the heat transfer from the corner portion of the molten metal 50 including the measurement points Pc2 and Pc3 to the mold 40 was larger than the heat transfer from the other portions of the molten metal 50. That is, since the dies 41 and the upper punch 42 are present in the vicinity above and laterally of the aforementioned corner portion, it is presumed that heat transfer from the aforementioned corner portion to the dies 41 and the upper punch 42 was more likely to occur than in the other portions of the molten metal 50.

[0248] (Determination accuracy of the first heat transfer coefficient data) In this embodiment, by calculating the error between the measured value of the temperature data at a specific location and the specific temperature data Tms´ using the first heat transfer coefficient data h1-1 to h7-1 determined by the generation unit 33, the determination accuracy of the generation unit 33 was evaluated. The specific temperature data Tms´ is the estimated temperature data Tme´ among a plurality of estimated temperature data Tme´ at a certain time step, for which the difference between the estimated temperature data Tme´ and the temperature data at the next time step is minimized. The "temperature data" in this case refers to the measured value of the temperature at the measurement location where the estimated temperature data Tme´ is estimated.

[0249] Hereinafter, the specific temperature data Tms´ at the measurement locations Pm1-1 to Pm7-1 will be referred to as "specific temperature data Tms1-1´ to Tms7-1´", and the specific temperature data Tms´ at the measurement locations Pm1-2 to Pm7-2 will be referred to as "specific temperature data Tms1-2´ to Tms7-2´".

[0250] Specifically, for a total of 14 locations including the measurement locations Pm1-1 to Pm7-1 and Pm1-2 to Pm7-2 shown in FIG. 5, the error between the measured value and the specific temperature data Tms´ was calculated. Also, in this embodiment, a series of processes related to the evaluation of the determination accuracy were performed using the simulation device 100.

[0251] In this embodiment, for the first heat transfer coefficient data h1-1, (i) the value when using the method (3D unsteady heat transfer model) of this embodiment was determined. Also, (ii) the value when using the above-described 1D unsteady heat transfer model was determined. Hereinafter, the first heat transfer coefficient data h1-1 in case (i) will be referred to as "first heat transfer coefficient data h11", and the first heat transfer coefficient data h1-1 in case (ii) will be referred to as "first heat transfer coefficient data h12". Further, as a comparative example, first heat transfer coefficient data with a constant value was used for calculating each specific temperature data. As the constant values, values of 2000, 5000, 10000, 20000, and 30000 W / (m 2 ·K) were set.

[0252] In this embodiment, as the measured values of the temperature data Tm1-1´ to Tm7-1´ and Tm1-2´ to Tm7-2´, the respective values acquired by the first and second sensors 10 and 20 were used. Also, in the cases of the aforementioned (i), (ii), and each comparative example, the specific temperature data Tms1-1´ to Tms7-1´ and Tms1-2´ to Tms7-2´ were calculated using the aforementioned formulas (3) to (5). In the case of the aforementioned (ii), each specific temperature data was calculated using the aforementioned formulas (8) and (9).

[0253] Next, by calculating seven types of average errors E1 to E7, the determination accuracy of the generation unit 33 in each of the aforementioned cases (i) and (iv), as well as each comparative example, was evaluated. The average error E1 is obtained by calculating the error between the measured value and the specific temperature data Tms´ in the case of the aforementioned (i) for each temperature data, and taking the average value of the calculated errors. The average error E2 is obtained by calculating the error between the measured value and the specific temperature data Tms´ in the case of the aforementioned (ii) for each temperature data, and taking the average value of the calculated errors.

[0254] The average errors E3 to E7 are obtained by calculating the error between the measured value and the specific temperature data Tms´ when the first heat transfer coefficient data is each of the aforementioned constant values for each temperature data, and taking the average value of the calculated errors. Specifically, the average errors E1 to E7 were calculated using the following formula (10).

[0255]

Equation

[0256] E: Average errors E1 to E7 [K] Tmes: Measured values of the temperature data Tm1-1´ to Tm7-1´ and Tm1-2´ to Tm7-2´ [K] Tcal: Specific temperature data Tms1-1´ to Tms7-1´ and Tms1-2´ to Tms7-2´ [K] In this embodiment, for the evaluation of the calculation accuracy, the average errors E1 to E7 for each time step were calculated, and the calculation results were plotted to create a graph. As a result, a graph as shown in FIG. 23 was obtained. The horizontal axis of the graph shown in FIG. 23 is the elapsed time since the molten metal 50 started to be poured into the die 41.

[0257] As shown in FIG. 23, the average error E1 had the smallest error among all the average errors regardless of the elapsed time. Also, for the average error E1, the error became maximum at the time when the elapsed time was about 11 seconds. However, the maximum value of the average error E1 was extremely small, about 3K, and it was found that according to the method of this embodiment, the first heat transfer coefficient data can be determined with extremely high accuracy.

[0258] <Verification of Estimation Results> In this embodiment, the estimated correlation relationship which is the estimation result of the first estimator 34 was verified. Specifically, first, for each of the first upper punch side estimated correlation relationship and the second die side estimated correlation relationship, the second estimated correlation relationship in period II was verified. As a premise, the first estimator 34 performed molding processing for three cases: (iii) when no release agent was applied to the inner side surface of the die 41, (iv) when a BN spray was sprayed, and (v) when a black body spray was sprayed. Specifically, for each of the above three cases, the load of the hydraulic press was set to five values of about 3t, about 4t, about 5t, about 6t, and about 7.5t for molding processing. Then, the above-described two second estimated correlation relationships were derived.

[0259] The estimation result of the first estimation unit 34 was obtained in the form of each graph in FIG. 24. As shown in FIG. 24, in any of the above-mentioned cases (iii) to (v), it was found that the second estimated correlation relationship can be approximated by a linear function with a positive slope. Also, the slope of this linear function was approximately 3251 in the case of (iii) above, approximately 1216 in the case of (iv) above, and approximately 696 in the case of (v) above. From this, it was found that the slope of the linear function approximately representing the second estimated correlation relationship changes according to the presence or absence of the mold release agent and the type of the mold release agent. Also, it was found that the slope of this linear function becomes the largest when the mold release agent is not applied to the inner side surface of the die 41 (the case of (iii) above).

[0260] Next, the third estimated correlation relationship in period III in the first upper punch side estimated correlation relationship was verified. As a premise, BN spray was sprayed on the inner side surface of the die 41. Also, the first estimation unit 34 derived the third estimated correlation relationship for each of the four cases of the numerical range of the contact pressure data Fp-1. The estimation result of the first estimation unit 34 was obtained in the form of the graph in FIG. 25. Note that the "molten metal surface temperature" on the horizontal axis of this graph represents the temperature at the measurement location Pc1.

[0261] As shown in FIG. 25, the first heat transfer coefficient data h1-1 decreased in value as the molten metal surface temperature decreased in any case of the numerical range of the contact pressure data Fp-1. Also, the slope (degree of decrease in value) of the first heat transfer coefficient data h1-1 increased as the value of the contact pressure data Fp-1 decreased. From these facts, it was found that the first heat transfer coefficient data is affected by the contact pressure data and the molten metal surface temperature.

[0262] Here, as an example, the case of performing molding processing by setting the load of the hydraulic press to five values of 3.3 t, 4.3 t, 5.3 t, 6.4 t, and 7.6 t using the mold 40 with BN spray sprayed on the inner side surface of the die 41 is given. In this case, in the above-mentioned linear function h1-1 = a × Fp-1 + b × Ts + c, a = approximately 1362.02, b = approximately 3.86, c = approximately -1475.31, and the average error is approximately 3500 W / (m 2The third estimated correlation could be approximated by (K).

[0263] [First Summary] The estimation device according to Embodiment 1 of the present invention uses, as teacher data, a basic data set in which first pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part is associated with first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface. A learned model that has learned the correlation between the first pressure data and the first heat transfer coefficient data is input with second pressure data indicating a pressure value acting on an interface between the metal part of the mold and a melt filled in a space surrounded by the metal part, and a second estimation unit that acquires second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to the second pressure data is provided.

[0264] The estimation device according to Embodiment 2 of the present invention further includes, in Embodiment 1, a generation unit that generates the first heat transfer coefficient data. The generation unit calculates, for a target element including the interface among a plurality of elements when at least a part of the mold is virtually divided into the plurality of elements, temperature data of a second element not including the temperature measurement unit at a first time point using temperature data of a first element including the temperature measurement unit at the first time point. Using the obtained temperature data of the first element and the second element and a plurality of the first heat transfer coefficients set for each target element, for a third element other than the target element among the plurality of elements, estimated temperature data at a second time point after a lapse of a predetermined time from the first time point is calculated. The first heat transfer coefficient at which the value of an evaluation function indicating the difference between the temperature data of a fourth element including the temperature measurement unit among the third elements at the second time point and the estimated temperature data of the fourth element becomes minimum may be specified as the value of the first heat transfer coefficient data.

[0265] The generation device of the learned model according to aspect 3 of the present invention uses, as teacher data, a basic data set in which first pressure data indicating a pressure value acting on the interface between the metal part of the mold and the melt filled in the space surrounded by the metal part, and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface are associated with each other, machine-learns the correlation between the first pressure data and the first heat transfer coefficient data, and uses, as an input, second pressure data indicating a pressure value acting on the interface between the metal part of the mold and the melt filled in the space surrounded by the metal part, and includes a first estimation unit that generates a learned model for outputting second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to the second pressure data.

[0266] The generation device of teacher data according to aspect 4 of the present invention includes first pressure data indicating a pressure value acting on the interface between the metal part of the mold and the melt filled in the space surrounded by the metal part, and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface, and includes a generation unit that generates teacher data used for machine-learning the correlation between the first pressure data and the first heat transfer coefficient data.

[0267] The control program according to aspect 5 of the present invention is a control program for causing a computer to function as the estimation device of aspect 1, and is a control program for causing a computer to function as the second estimation unit.

[0268] The control program according to aspect 6 of the present invention is a control program for causing a computer to function as the generation device of the learned model of aspect 3, and is a control program for causing a computer to function as the first estimation unit.

[0269] The control program according to aspect 7 of the present invention is a control program for causing a computer to function as the generation device of teacher data of aspect 4, and is a control program for causing a computer to function as the generation unit.

[0270] The estimation method according to aspect 8 of the present invention is an estimation method executed by a computer. Using a basic data set in which first pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface are associated as teacher data, a learned model that has learned the correlation between the first pressure data and the first heat transfer coefficient data is used. The method includes a second estimation step of inputting second pressure data indicating a pressure value acting on an interface between a metal part of the mold and a melt filled in a space surrounded by the metal part, and obtaining second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to the second pressure data.

[0271] The method for generating a learned model according to aspect 9 of the present invention is a generation method executed by a computer. Using a basic data set in which first pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface are associated as teacher data, the correlation between the first pressure data and the first heat transfer coefficient data is learned by machine learning. The method includes a first estimation step of generating a learned model for inputting second pressure data indicating a pressure value acting on an interface between a metal part of the mold and a melt filled in a space surrounded by the metal part and outputting second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to the second pressure data.

[0272] The method for generating teacher data according to aspect 10 of the present invention is a generation method executed by a computer, and includes a generation step of generating teacher data used for machine learning of the correlation between first pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface.

[0273] 〔Second Summary〕 The estimation device according to Aspect 11 of the present invention uses a basic data set in which pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface are associated with each other, and includes a first estimation unit that estimates a correlation between the pressure data and the first heat transfer coefficient data, and a second estimation unit that estimates second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to specific pressure data from the specific pressure data using an estimated correlation that is the correlation estimated by the first estimation unit.

[0274] According to the above configuration, if the pressure data acting on the interface can be specified, the second heat transfer coefficient data corresponding to the pressure specified using the estimated correlation can be estimated. Thereby, the heat transfer coefficient of the interface can be easily estimated, and furthermore, the temperature distributions of the metal part of the mold and the melt can be easily estimated.

[0275] The estimation device according to Aspect 12 of the present invention, in the above Aspect 11, the first estimation unit may estimate, as the estimated correlation, a first estimated correlation in a period from when the melt is filled in the space until a molding pressure for molding the melt is applied to the melt, a second estimated correlation in a period from when the molding pressure is applied to the melt until the surface temperature of the melt reaches the solidification temperature of the melt, and a third estimated correlation after the action of the molding pressure on the melt continues and the surface temperature becomes lower than the solidification temperature.

[0276] According to the above configuration, for example, compared with the case where the estimated correlation is estimated without considering at least one of the molding pressure and the solidification temperature, an estimated correlation closer to the actual state of the molding of the melt can be estimated. Thereby, the estimation accuracy of the second heat transfer coefficient data using the estimated correlation can be improved.

[0277] In the estimation device according to Aspect 13 of the present invention, in the above Aspect 11 or 12, the first estimation unit constructs an estimation model by machine learning using the basic data set as teacher data, and the second estimation unit inputs the specific pressure data into the estimation model, and obtains the second heat transfer coefficient data output from the estimation model, thereby estimating the second heat transfer coefficient data.

[0278] According to the above configuration, the correlation between the contact pressure data and the first heat transfer coefficient data can be estimated by machine learning using the basic data set as teacher data. Further, the second heat transfer coefficient data can be estimated by using the estimation model.

[0279] The estimation device according to Aspect 14 of the present invention may further include, in any one of the above Aspects 11 to 13, a temperature acquisition unit that acquires first temperature data indicating a temperature value of the first temperature of the metal part and second temperature data indicating a temperature value of the second temperature of the melt, and a generation unit that generates the basic data set by determining the first heat transfer coefficient using the first temperature data and the second temperature data acquired by the temperature acquisition unit.

[0280] According to the above configuration, compared with the case where, for example, the actual value of the first heat transfer coefficient data determined by the generation unit in the past is used as the first heat transfer coefficient data constituting the basic data set, the first heat transfer coefficient data constituting the basic data set becomes a value that reflects the timely state of the interface. Thereby, the estimation accuracy of the estimation correlation can be improved, and consequently, the estimation accuracy of the second heat transfer coefficient data can be improved.

[0281] In the estimator according to Aspect 15 of the present invention, in the above Aspect 14, the temperature acquisition unit acquires the first temperature data at a plurality of locations in the metal part and the second temperature data at a plurality of locations in the melt at regular intervals. The generation unit, for each of a plurality of locations in the metal part, uses the plurality of first temperature data and the plurality of second temperature data acquired by the temperature acquisition unit at a certain point in time, and while changing the value of the first heat transfer coefficient data, calculates a plurality of estimated temperature data which are estimated values of the first temperature data at the time when the certain time has elapsed from the certain point in time. For each of the plurality of estimated temperature data, calculates the difference between the estimated temperature data and the first temperature data at the time when the certain time has elapsed from the certain point in time. Using the plurality of differences calculated for each of the plurality of locations in the metal part, select the first heat transfer coefficient data that constitutes the basic data set from among the plurality of first heat transfer coefficient data whose values have been changed.

[0282] According to the above configuration, it is possible to determine the first heat transfer coefficient data that constitutes the basic data set in consideration of the plurality of differences calculated for each of the plurality of locations in the metal part. As a result, the first heat transfer coefficient data that constitutes the basic data set becomes a value that not only reflects the timely state of the interface but also accurately reflects the temperature distribution of the metal part. Therefore, the estimation accuracy of the second heat transfer coefficient data can be further improved.

[0283] The simulation device according to Aspect 16 of the present invention includes the estimator according to Aspect 11 and a behavior estimator that estimates the solidification behavior of the melt using the second heat transfer coefficient data estimated by the estimator.

[0284] As Aspect 17 of the present invention, the estimator according to each aspect of the present invention may be realized by a computer. In this case, a control program for the estimator that realizes the estimator on a computer by operating the computer as each part (software element) included in the estimator, and a computer-readable recording medium on which it is recorded also fall within the scope of the present invention.

[0285] The data set according to Embodiment 18 of the present invention includes pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part, first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface, and an estimated correlation relationship obtained by estimating a correlation relationship between the pressure data and the first heat transfer coefficient data. When an estimation device that performs an estimation process for estimating second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to specific pressure data from the specific pressure data using the estimated correlation relationship performs the estimation process, the data set is used.

[0286] According to the above configuration, the heat transfer coefficient of the interface can be easily estimated using the pressure data, the first heat transfer coefficient data, and the estimated correlation relationship included in the data set. As a result, the temperature distributions of the metal part of the mold and the melt can be easily estimated.

[0287] The estimation method according to Embodiment 19 of the present invention includes a first estimation step of estimating a correlation relationship between pressure data indicating a pressure value acting on an interface between a metal part of a mold and a melt filled in a space surrounded by the metal part and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal part and the melt at the interface, using a basic data set in which the pressure data and the first heat transfer coefficient data are associated with each other, and a second estimation step of estimating second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to specific pressure data from the specific pressure data, using the estimated correlation relationship between the pressure data and the first heat transfer coefficient data estimated in the first estimation step.

Explanation of Signs

[0288] 30 First Estimation Device (Estimation Device) 31 Temperature Acquisition Unit 32 Pressure Acquisition Unit 33 Generation Unit 34 First Estimation Unit 35 Second Estimation Unit 40, 40a Mold 40X Space 41X and 41Xa metal parts 41Y and 42X holes 41Ya First hole 41Yb Second hole 41Yc Third hole 41Yd Fourth hole 42 Upper punch (metal part) 43 Lower punch (metal part) 50 Molten metal (melt) 60 Second estimation device (behavior estimation device) 81 Third temperature measurement part (temperature measurement part) 82 Third pressure measurement part (pressure measurement part) 82a Membrane 83 Third sensor body (single member) 100, 200, 300 Simulation devices hd-1, hpl-1, hpu-1, hd1, hd2, hd3, hd4, hd5 First heat transfer coefficient data hd-2, hpl-2, hpu-2 Second heat transfer coefficient data Fd-1, Fd1´, Fd4´, Fp-1 Contact pressure data (pressure data) Fd-2, Fp-2 Estimated contact pressure data (specific pressure data) Si1 First interface Si2 Second interface Te1-1´, Te4-1´, Te1-2´, Te4-2´, Tm1-1´, Tm2-1´, Tm3-1´, Tm4-1´, Tm5-1´, Tm6-1´, Tm7-1´, Tm1-2´, Tm2-2´, Tm3-2´, Tm4-2´, Tm5-2´, Tm6-2´, Tm7-2´ Temperature data (first temperature data) Tc1´, Tc2´, Tc3´, Tc4´, Tc5´, Tc6´, Tc7´, Td1´, Td4´ Temperature data (second temperature data) T i、j、k ´ Estimated temperature data

Claims

1. an estimation device comprising: a second estimation unit that inputs second pressure data indicating a pressure value acting on an interface between a metal portion of a mold and a molten material filled in a space surrounded by the metal portion into a trained model that has been machine-learned to determine a correlation between the first pressure data and the first heat transfer coefficient data, using a basic data set in which first pressure data indicating a pressure value acting on an interface between a metal portion of the mold and a molten material filled in a space surrounded by the metal portion and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal portion and the molten material at the interface as training data; and acquires second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to the second pressure data.

2. A generating unit that generates the first heat transfer coefficient data, The generation unit is When at least a part of the mold is virtually divided into a plurality of elements, for a target element including the interface among the plurality of elements, calculating temperature data at the first time point of a second element not including the temperature measurement unit using temperature data at the first time point of a first element including the temperature measurement unit; Using the obtained temperature data of the first element and the second element and a plurality of the first heat transfer coefficients set for each of the target elements, calculate estimated temperature data for a third element other than the target element among the plurality of elements at a second time point after a predetermined time has elapsed from the first time point; 2. The estimation device according to claim 1, wherein the first heat transfer coefficient that minimizes a value of an evaluation function indicating a difference between temperature data at the second time point of a fourth element including the temperature measurement unit among the third elements and the estimated temperature data of the fourth element is identified as the value of the first heat transfer coefficient data.

3. a first estimation unit that uses a basic data set in which first pressure data indicating a pressure value acting on an interface between a metal portion of a mold and a molten material filled in a space surrounded by the metal portion and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal portion and the molten material at the interface as training data, machine learning a correlation between the first pressure data and the first heat transfer coefficient data, and generates a trained model for using second pressure data indicating a pressure value acting on an interface between the metal portion of the mold and the molten material filled in the space surrounded by the metal portion as input, and outputting second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to the second pressure data.

4. A training data generation device comprising: first pressure data indicating a pressure value acting on an interface between a metal portion of a mold and a molten material filled in a space surrounded by the metal portion; and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal portion and the molten material at the interface; and a generation unit that generates training data used for machine learning of a correlation between the first pressure data and the first heat transfer coefficient data.

5. A control program for causing a computer to function as the estimation device according to claim 1, the control program causing a computer to function as the second estimation unit.

6. A control program for causing a computer to function as the trained model generation device according to claim 3, the control program causing a computer to function as the first estimation unit.

7. A control program for causing a computer to function as the teacher data generating device according to claim 4, the control program causing a computer to function as the generating unit.

8. a second estimation step of inputting second pressure data indicating a pressure value acting on an interface between a metal portion of a mold and a molten material filled in a space surrounded by the metal portion into a trained model that has been machine-learned to determine a correlation between the first pressure data and the first heat transfer coefficient data, using a basic data set in which first pressure data indicating a pressure value acting on an interface between a metal portion of the mold and a molten material filled in a space surrounded by the metal portion and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal portion and the molten material at the interface as training data, and acquiring second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to the second pressure data.

9. a first estimation step of using a basic data set in which first pressure data indicating a pressure value acting on an interface between a metal portion of a mold and a molten material filled in a space surrounded by the metal portion and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal portion and the molten material at the interface as training data, and machine learning a correlation between the first pressure data and the first heat transfer coefficient data, and generating a trained model for using second pressure data indicating a pressure value acting on an interface between a metal portion of the mold and a molten material filled in a space surrounded by the metal portion as input and outputting second heat transfer coefficient data indicating a second heat transfer coefficient corresponding to the second pressure data.

10. A computer-executed method for generating teacher data, the method including: first pressure data indicating a pressure value acting on an interface between a metal portion of a mold and a molten material filled in a space surrounded by the metal portion; and first heat transfer coefficient data indicating a first heat transfer coefficient between the metal portion and the molten material at the interface; and a generation step of generating teacher data used for machine learning of a correlation between the first pressure data and the first heat transfer coefficient data.

Images