Information processing method, information processing device, and information processing program

Abstract

To provide an information processing method capable of estimating a situation in a predetermined space more easily than before.SOLUTION: An information processing method comprises: receiving input of information representing a state in a predetermined space, information related to a specific region in the predetermined space and information related to an arrangement position of a fluid regulation body in the predetermined space; performing, based upon the information related to the arrangement position of the fluid regulation body in the predetermined space, conversion into one-dimensional information related to the relative position relation between the specific region and the fluid regulation body; inputting the information representing the state in the predetermined space, the information related to the specific region and the one-dimensional information to a learning model having learnt relation between information related to the relative position relation between the specific region in the predetermined space and the fluid regulation body for the specific region and information related to fluidity in the specific region in this case; and estimating the fluidity in the specific region in the predetermined space in a case that the fluid regulation body is arranged at an arrangement position indicated with the information related to the arrangement position of the fluid regulation body in the predetermined space.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

[0002] Conventionally, various simulations are performed to obtain desired results, and then an actual device is created.

[0003] For example, when creating a silicon rod as described in Patent Document 1, it is necessary to perform simulations for the arrangement of the rod and the arrangement of the gas nozzle for injecting gas into the bell jar, because the desired results cannot be obtained without appropriate arrangements. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 5859626 [Patent Document 2] Japanese Patent Application Laid-Open No. 2003-335512 Summary of the Invention [Problem to be solved by the invention]

[0005] However, the method of performing the above-mentioned simulation multiple times and then using the conditions that produce the most desirable results as the conditions for the actual device has the problem that multiple simulations take time, and there is also the problem that there may be conditions remaining that produce better results than those from the multiple simulations.

[0006] Therefore, the present invention has been made in consideration of the above problems, and aims to provide an information processing device that can obtain conditions that are as close to the optimal solution as possible in a shorter time than performing multiple simulations. [Means for solving the problem]

[0007] In order to solve the above problem, an information processing method according to one aspect of the present invention is an information processing method by an information processing device that inputs and outputs information about a flow regulator that regulates a flow within a predetermined space, the information processing method including: a storage step of storing a learning model that has learned the relationship between information about a specific region in the predetermined space, information about the specific region and the relative positional relationship between the flow regulator with respect to the specific region, and information about the flow in the specific region in that case; a reception step of receiving input of information representing the state of the predetermined space, information about the specific region, and information about the placement position of the flow regulator within the predetermined space; a conversion step of converting the information about the placement position of the flow regulator within the predetermined space into one-dimensional information about the relative positional relationship between the specific region and the flow regulator, based on the information about the placement position of the flow regulator within the predetermined space; an estimation step of inputting the information representing the state of the predetermined space, information about the specific region, and the one-dimensional information received in the reception step into the learning model, and estimating the flow in the specific region within the predetermined space when the flow regulator is placed at the placement position indicated by the information about the placement position of the flow regulator within the predetermined space; and an output step of outputting information about the flow within the predetermined space based on the estimation made in the estimation step.

[0008] In the above information processing method, the learning model is a model that has learned information regarding the relative positional relationship between a specific area in a specified space and a flow regulator with respect to the specific area, the information defining the relative positional relationship between the specific area and the flow regulator with respect to the specific area based on the specific area, and the conversion step may convert into one-dimensional information regarding the relative positional relationship between the specific area and the flow regulator with respect to the specific area based on the specific area.

[0009] In the above information processing method, the learning model is a model that defines multiple small spaces based on a specific area as information regarding the relative positional relationship between a specific area in a specified space and a flow regulator with respect to the specific area, and learns the number of flow regulators contained in the multiple small spaces, and the conversion step may define multiple small spaces based on a specific area as information regarding the relative positional relationship between a specific area in a specified space and a flow regulator with respect to the specific area, and convert the information into one-dimensional information indicating the number of flow regulators contained in each of the multiple small spaces.

[0010] In the above information processing method, the information about the specific region may include information about the absolute position of the specific region in a predetermined space.

[0011] In the above information processing method, the specified space may be the space within a container of a precipitation reactor that precipitates a compound based on a raw material onto the surface of a core material, the container having an ejection hole for ejecting the raw material, the specific region may be at least a part of the surface of the core material, and the flow regulator may be the ejection hole for ejecting the raw material.

[0012] In the above information processing method, the information about the flow within the predetermined space may be information about a heat flux occurring around the core material.

[0013] In the above information processing method, the information about the flow within the predetermined space may be information about the thickness of the compound precipitated on the core material.

[0014] In the above information processing method, the specified space may be the space within a container of a precipitation reactor that precipitates a compound based on raw materials on the surface of a core material, the container having an ejection hole for ejecting the raw materials and an exhaust hole for discharging gas within the specified space, the specific region being at least a part of the surface of the core material, and the flow regulator being the exhaust hole.

[0015] In the above information processing method, the specified space is the space within a container filled with powder particles, and the powder particles are caused to flow by supplying gas into the container, the flow regulator is at least one of an obstruction plate positioned within the specified space to obstruct the flow in a specified direction and a supply port of a nozzle that supplies gas, and bubbles contained within the specified space break up and become smaller in diameter upon contact with the obstruction plate, and the information regarding the flow within the specified space may be information regarding at least either the diameter or the number of bubbles contained in a specific region.

[0016] In the above information processing method, the predetermined space may be a space within a container filled with powder particles as a specific region, and may have a measurement point for measuring the temperature within the space; a temperature control means is disposed within the powder or granular material; gas is supplied into the container to cause the powder or granular material to flow; the flow regulator is at least one of the temperature control means for adjusting the temperature and the supply port of a nozzle for supplying gas; and the information regarding the flow within the predetermined space may be the temperature in the specific region.

[0017] In the above information processing method, the specified space has a measurement area as a specific area for measuring at least either the temperature or the concentration of a substance within the space, a heat generating / cooling means or a substance generating / reducing means may be arranged within the space, and gas may flow through the space, and the flow regulator may be an air vent or opening provided in the specified space, or, if a heat generating / cooling means is arranged, the flow regulator may be at least one of an air vent or opening and a heat generating / cooling means, or, if a substance generating / reducing means is arranged, the flow regulator may be at least one of an air vent or opening and a substance generating / reducing means, and the information regarding the flow within the specified space may be information regarding at least either the temperature or the concentration of a substance within the specific area.

[0018] In order to solve the above problem, an information processing device according to one aspect of the present invention is an information processing device that inputs and outputs information about a flow regulator that regulates a flow within a predetermined space, and includes: a memory unit that stores a learning model that has learned the relationship between information about a specific region in the predetermined space, information about the specific region and the relative positional relationship between the specific region and the flow regulator with respect to the specific region, and information about the flow in the specific region in that case; a reception unit that receives input of information representing the state of the predetermined space, information about the specific region, and information about the placement position of the flow regulator within the predetermined space; a conversion unit that converts the information about the placement position of the flow regulator within the predetermined space into one-dimensional information about the relative positional relationship between the specific region and the flow regulator, based on the information about the placement position of the flow regulator within the predetermined space; an estimation unit that inputs the information representing the state of the predetermined space, information about the specific region, and the one-dimensional information received by the reception unit into the learning model, and estimates the flow in the specific region within the predetermined space when the flow regulator is placed at the placement position indicated by the information about the placement position of the flow regulator within the predetermined space; and an output unit that outputs information about the flow in the predetermined space based on the estimation by the estimation unit.

[0019] In order to solve the above problem, a communication status information output program according to one embodiment of the present invention is a computer that inputs and outputs information about a flow regulator that regulates the flow within a specified space, and is capable of accessing a memory unit that stores a learning model that has learned the relationship between information about the relative positional relationship between a specific region in the specified space and the flow regulator with respect to the specific region, and information about the flow in the specific region in that case. The computer has the following capabilities: a reception function that receives input of information representing the state of the specified space, information about the specific region, and information about the placement position of the flow regulator within the specified space; a conversion function that converts the information about the placement position of the flow regulator within the specified space into one-dimensional information about the relative positional relationship between the specific region and the flow regulator, based on the information about the placement position of the flow regulator within the specified space; an estimation function that inputs the information representing the state of the specified space, information about the specific region, and the one-dimensional information into the learning model, and estimates the flow in a specific region within the specified space when the flow regulator is placed at the placement position indicated by the information about the placement position of the flow regulator within the specified space; and an output function that outputs information about the flow within the specified space based on the estimation by the estimation function. [Effects of the Invention]

[0020] An information processing method according to one aspect of the present invention accepts input of information representing the state of a predetermined space, information relating to a specific region, and information relating to the placement position of a flow regulator within the predetermined space, converts the information representing the state of the predetermined space, information relating to a specific region, and information relating to the placement position of a flow regulator within the predetermined space, converts the information representing the state of the predetermined space, information relating to the specific region, and the one-dimensional information into a learning model that has learned the relationship between information relating to the specific region in the predetermined space and the relative positional relationship of the flow regulator with respect to the specific region, and information relating to the flow in the specific region in that case, estimates the flow in the specific region within the predetermined space when the flow regulator is placed at the placement position indicated by the information relating to the placement position of the flow regulator within the predetermined space, and outputs information relating to the flow within the predetermined space.As a result, a user who has devised the placement of the flow regulator can recognize, without performing a simulation, whether the placement of the flow regulator that he or she devised will achieve a desired result regarding the flow in the predetermined space, based on the relative positional relationship between the specific region and the flow regulator. [Brief explanation of the drawings]

[0021] [Figure 1] FIG. 1 is a block diagram illustrating an example of the configuration of an information processing device. [Figure 2A] 1(a) and 1(b) are diagrams for explaining, in two dimensions, the relative positional relationship of a measurement point in a predetermined space and a flow regulator with respect to the measurement point. [Figure 2B] FIG. 10 is a diagram showing another example of two-dimensionally explaining the relative positional relationship of the measurement points in a predetermined space and the flow regulator with respect to the measurement points. [Figure 3] 1(a) and 1(b) are diagrams for explaining, in three dimensions, the relative positional relationship of a measurement point in a predetermined space and a flow regulator with respect to the measurement point. [Figure 4] 10 is a flowchart illustrating an example of an operation related to generation of a learning model by an information processing device. [Figure 5] 10 is a flowchart illustrating an example of an operation related to flow estimation by an information processing device. [Figure 6]1A and 1B are diagrams showing a predetermined space according to Example 1, in which (a) is a schematic perspective view of a bell jar device, and (b) is a schematic side view of the bell jar device. [Figure 7A] 1(a) is a plan view showing an example of the arrangement of a silicon rod and a gas nozzle in a bell jar apparatus as a predetermined space according to Example 1. FIG. 1(b) is a diagram illustrating an example of the relative positional relationship between the silicon rod and the gas nozzle. [Figure 7B] 10(a) and 10(b) are diagrams illustrating another example of the relative positional relationship between the silicon rod and the gas nozzle. [Figure 8A] 1(a) is a plan view showing an example of the arrangement of a silicon rod and a gas nozzle in a bell jar apparatus as a predetermined space according to Example 1. FIG. 1(b) is a diagram illustrating an example of the relative positional relationship between the silicon rod and the gas nozzle. [Figure 8B] 10(a) and 10(b) are diagrams illustrating another example of the relative positional relationship between the silicon rod and the gas nozzle. [Figure 9] 10 is a flowchart illustrating an example of an operation related to generation of a learning model by the information processing device according to the first embodiment. [Figure 10] 10 is a flowchart illustrating an example of an operation related to flow estimation by the information processing device according to the first embodiment. [Figure 11] 10(a) and 10(b) are diagrams showing an example of a fluidized bed reactor according to Example 2, illustrating an example of air bubble breakup by a baffle plate. [Figure 12] 10(a) and 10(b) are diagrams illustrating an example of a measurement point in a predetermined space and a relative positional relationship of a flow regulator with respect to the measurement point according to Example 2. FIG. [Figure 13] 10 is a flowchart illustrating an example of an operation related to generation of a learning model by the information processing device according to the second embodiment. [Figure 14] 10 is a flowchart illustrating an example of an operation related to flow estimation by the information processing device according to the second embodiment. [Figure 15] 10 is a diagram schematically illustrating an example of a room as a predetermined space and an entrance / exit as a flow regulator according to Example 3. FIG. [Figure 16]10 is a diagram illustrating an example of a measurement point in a predetermined space and a relative positional relationship of a flow regulator with respect to the measurement point according to Example 3. FIG. [Figure 17] 11 is a flowchart illustrating an example of an operation related to generation of a learning model by the information processing device according to the third embodiment. [Figure 18] 11 is a flowchart illustrating an example of an operation related to flow estimation by the information processing device according to the third embodiment. [Figure 19] FIG. 10 is a diagram showing an example of a relative positional relationship between measurement points and a flow regulator with respect to the measurement points when the cooling device according to Example 4 is set as a predetermined space. [Figure 20] 10 is a flowchart illustrating an example of an operation related to generation of a learning model by the information processing device according to the fourth embodiment. [Figure 21] 10 is a flowchart illustrating an example of an operation related to flow estimation by the information processing device according to the fourth embodiment. [Figure 22] FIG. 10 is a diagram schematically illustrating an example of a reaction apparatus according to Example 5. [Figure 23] 23(a) is a perspective view showing an example of a relative positional relationship between measurement points in a predetermined space and a flow regulator with respect to the measurement points according to Example 5. (b) is a front view of FIG. 23(a) showing an example of a relative positional relationship between measurement points in a predetermined space and a flow regulator with respect to the measurement points according to Example 5. [Figure 24] 10 is a flowchart illustrating an example of an operation related to generation of a learning model by an information processing device according to a fifth embodiment. [Figure 25] 13 is a flowchart illustrating an example of an operation related to flow estimation by the information processing device according to the fifth embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0022] <Embodiment> An information processing device according to the present invention will be described below with reference to the drawings. First, a general description of the information processing device will be given, and then a specific example of use of the information processing device will be described.

[0023] <Configuration> The information processing device 100 according to this embodiment is a computer system that performs processing to estimate results obtained under input conditions. The information processing device 100 is a device that inputs and outputs information about a flow regulator that regulates the flow of gas, liquid, heat, etc. within a predetermined space. The information processing device 100 may also generate a learning model required to execute the above-described estimation processing.

[0024] 1 is a block diagram showing an example configuration of an information processing device 100. Here, the information processing device 100 includes a communication unit 120, an input unit 130, a control unit 140, a storage unit 150, and an output unit 160. The communication unit 120, the input unit 130, the control unit 140, the storage unit 150, and the output unit 160 may be connected to each other via a connection line 110 so as to be able to communicate with each other.

[0025] The communication unit 120 is a communication interface that communicates with external devices via a network. The communication unit 120 transmits data received from the external devices to the control unit 140. Furthermore, the communication unit 120 transmits specified data to a specified external device or the like in accordance with instructions from the control unit 140. The communication unit 120 may receive, from the external device, for example, information about a specified space, position information about a measurement point where flow measurement is desired in the specified space, information indicating the position of a flow regulator, and the like, and transmit this information to the control unit 140.

[0026] The input unit 130 is an input interface that accepts input from a user of the information processing device 100. The input unit 130 may be realized by, for example, a keyboard, a mouse, a touch panel, or the like, but is not limited to these. The input unit 130 may, for example, accept input by voice. When the input unit 130 accepts input from the user, it transmits the accepted input to the control unit 140. The input unit 130 may, for example, accept input of position information of a measurement point and transmit the same to the control unit 140. Furthermore, the input unit 130 may, for example, accept input of information regarding the arrangement of a flow regulator that regulates flow within a predetermined space and transmit the same to the control unit 140. Furthermore, the input unit 130 may, for example, accept input of information regarding the predetermined space and transmit the same to the control unit 140.

[0027] The control unit 140 is a processor having a function for controlling each unit of the information processing device 100. The control unit 140 uses various programs and data stored in the storage unit 150 to execute the processing to be executed by the information processing device 100, that is, the processing to estimate the flow of gas, liquid, solid, heat, etc. within a predetermined space.

[0028] To achieve this processing, the control unit 140 includes a receiving unit 141, a converting unit 142, and an estimating unit 143. The control unit 140 may also include a learning unit 144.

[0029] The receiving unit 141 receives information transmitted from the communication unit 120 or the input unit 130 as information that will be the basis for information to be input to the learning model 151. That is, the receiving unit 141 receives information that represents the state of a predetermined space, information about a specific region in the predetermined space, and placement information about the placement of a flow regulator included in the predetermined space within the predetermined space. The receiving unit 141 also receives information that includes position information that indicates the position of a measurement point (specific region) that indicates the location where the flow is measured.

[0030] Here, a predetermined space will be described. The predetermined space is a space in which a user wishes to recognize the flow of something (e.g., liquid, gas, solid, heat, etc.). The predetermined space may have various shapes, such as a plane, a rectangular parallelepiped, a cylinder, a sphere, a hemisphere, or even a more complex shape. Furthermore, information representing the state of the predetermined space may be information such as the shape and volume of the predetermined space, the temperature at which the flow is estimated, and the composition of the gas or liquid contained in the predetermined space, or may be some or all of these, or may include other information. The information representing the state of the predetermined space can also be information indicating the conditions for estimating the flow in the predetermined space. Furthermore, information about specific regions in the predetermined space may be information representing the characteristics of each specific region. For example, when comparing specific regions, the information about the specific regions may be information that can indicate the similarities, similarities, and differences between one specific region and another specific region. For example, the information may be information that can indicate the points where information on the position of a specific region, information on the orientation (direction) of the specific region, and, if a substance is present, information on the material of the substance in the specific region, are identical, similar, or different from the information on other regions. Furthermore, the information on a specific region in a predetermined space may include information on a specific region set in the predetermined space, such as the absolute position (universal position) of a flow regulator in the predetermined space. Furthermore, the absolute position here may be information on position coordinates in a coordinate system that can define the predetermined space.

[0031] The receiving unit 141 transmits to the converting unit 142 the received position information of the measurement point and the arrangement information relating to the arrangement of the flow regulator in a predetermined space.

[0032] The conversion unit 142 converts the transmitted position information of the measurement point and the placement information regarding the placement of the flow regulator into one-dimensional information indicating the relative positional relationship of the flow regulator with respect to the measurement point. That is, the conversion unit 142 generates one-dimensional information indicating the relative positional relationship between the measurement point, which is a specific region, and the flow regulator, and transmits it to the estimation unit 143.

[0033] Here, a measurement point as a specific region in a predetermined space, a relative positional relationship of a flow regulator with respect to the measurement point, and one-dimensional information indicated by this relative positional relationship will be described with reference to Figures 2A, 2B, and 3. In this embodiment, the relative positional relationship between a measurement point and a flow regulator is information regarding the direction of the flow regulator and the distance to the flow regulator as viewed from the measurement point, and more specifically, is information indicating the distribution of how many flow regulators are present and within what range in each approximate direction as viewed from the measurement point.

[0034] 2A(a) and (b) are diagrams showing an example of the arrangement of measurement points and a flow regulator in a predetermined space 200, and are diagrams for explaining one-dimensional information on the relative positional relationship between the measurement points and the flow regulator. Note that although Figures 2A(a) and (b) show an example in which the predetermined space 200 is rectangular, the shape of the predetermined space 200 is not limited to a rectangle and may be any shape.

[0035] FIG. 2A(a) shows a plan view of a predetermined space 200, illustrating an example in which 18 flow regulators 221x to 238x are arranged within the predetermined space 200. Assume that within this predetermined space, a measurement point 210x, where a user wishes to know the state of flow, is designated as the position shown in FIG. 2A(a). Here, the relative positional relationship of the flow regulators 221x to 238x with respect to the measurement point 210x as a specific region may be information indicating the direction in which each of the flow regulators 221x to 238x is located from the measurement point 210x. For example, the flow regulator 221x can be specified as being located at 1 o'clock in the drawing from the measurement point 210x. However, it is not realistic to define the relative positional relationship of each of the flow regulators 221x to 238x with respect to the measurement point 210x and input this information into a learning model. Therefore, in this embodiment, the relative positional relationship between the measurement point 210x and the flow regulators 221x to 238x can be defined as one-dimensional information indicating the relative positional relationship between the measurement point 210x and the flow regulators 221x to 238x, based on a rough direction and the number of flow regulators present in that defined direction.

[0036] For example, as shown in FIG. 2A(a), the periphery is divided by dashed lines as shown in the figure to define multiple directions based on each measurement point. In this embodiment, the regions divided by these dashed lines in a predetermined space 200 are hereinafter defined as small spaces. Then, small spaces X1 to X8 divided by dashed lines are defined clockwise as viewed from each measurement point. Note that these small spaces are virtually defined in order to define the directions as viewed from each measurement point. For example, as viewed from measurement point 210x, flow regulator 221x can be defined as being in the direction of X1 (existing in small space X1). Similarly, as viewed from measurement point 210x, three flow regulators, 222x, 229x, and 237x, can be defined as being in the direction of X2 (existing in small space X2). Also, it can be defined that six flow regulators, namely, flow regulators 231x, 232x, 234x, 235x, 236x, and 238x, exist in the direction X3 (small space X3) when viewed from measurement point 210x.

[0037] From these facts, one-dimensional information indicating the relative positional relationship between the measurement point 210x and the flow regulator can be defined as the number of flow regulators present in each direction (small space) X1 to X8 as viewed from the measurement point 210x, and in the case of Figure 2A(a), it can be defined as {X1, X2, X3, X4, X5, X6, X7, X8} = {1, 3, 6, 3, 2, 0, 1, 2}.

[0038] 2A(a), it can be seen that, for example, flow regulators 229x, 237x, and 238x exist in direction (small space) X1, flow regulator 231x exists in direction (small space) X2, and flow regulator 232x exists in direction (small space) X3. For other directions, by counting the number of flow regulators included in each direction, one-dimensional information indicating the relative positional relationship between measurement point 211x and the flow regulators can be defined as {X1, X2, X3, X4, X5, X6, X7, X8} = {3, 1, 1, 0, 2, 2, 6, 3}.

[0039] Figure 2A(a) shows an example in which only the direction is defined as the relative positional relationship. However, as shown in Figure 2A(b), the relative positional relationship may also be defined by taking into account a predetermined range, i.e., the distance from the measurement point. The distribution of the flow regulator and the positions of the measurement points shown in Figure 2A(b) are the same as those shown in Figure 2A(a).

[0040] In the example shown in FIG. 2A(b), the definitions of multiple directions for each measurement point are the same as in FIG. 2A(a). On the other hand, an example is shown in which a region is defined from each measurement point to a predetermined range. That is, an example is shown in which the relative positional relationship is defined by the direction and distance from each measurement point. For each measurement point in FIG. 2A(b), an example is shown in which fan-shaped small spaces X21 to X28 are defined as regions for determining the relative positional relationship. In this case, one-dimensional information indicating the relative positional relationship between measurement point 210x and the flow regulator can be defined as {X21, X22, X23, X24, X25, X26, X27, X28} = {1, 1, 0, 2, 1, 0, 1, 2}. Furthermore, one-dimensional information indicating the relative positional relationship between the measurement point 211x and the flow regulator can be defined as {X21, X22, X23, X24, X25, X26, X27, X28}={2, 1, 1, 0, 2, 1, 0, 1}.

[0041] 2B, eight directions are defined for each measurement point, similar to the example shown in FIGS. 2A(a) and 2A(b), but the distance from each measurement point is further defined in two stages. That is, a total of 16 small spaces, X31 to X38 and X41 to X48, are defined to define the relative positional relationship between each measurement point and the flow regulator. In this case, one-dimensional information indicating the relative positional relationship between measurement point 210x and the flow regulator can be defined as {X31, X32, X33, X34, X35, X36, X37, X38, X41, X42, X43, X44, X45, X46, X47, X48} = {0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1}. Furthermore, one-dimensional information indicating the relative positional relationship with the flow regulator as viewed from measurement point 211x can be defined as {X31, X32, X33, X34, X35, X36, X37, X38, X41, X42, X43, X44, X45, X46, X47, X48} = {0, 0, 0, 1, 0, 0, 0, 1, 1, 1, 0, 1, 1, 0, 1, 1}.

[0042] As shown in Figures 2A(a), (b), and 2B, the definition of the directions as seen from the measurement point does not have to be eight directions as shown, and the spacing between the small spaces defined as directions does not have to be uniform. Furthermore, the distance direction from the measurement point can be divided into three, four, etc. stages instead of just two stages as shown in Figure 2B. The distances do not have to be defined uniformly, and can be defined at any distance. Multiple small spaces can also be defined so that they overlap.

[0043] 3(a) and (b) show an example of a three-dimensional predetermined space 200. Note that, although Fig. 3(a) and (b) show a rectangular parallelepiped space as the predetermined space 200, the predetermined space 200 does not have to be a rectangular parallelepiped space.

[0044] 3(a) shows an example in which ten flow regulators 221y to 230y are arranged around a measurement point 210y in a predetermined space 200. Then, when the measurement point is used as a reference and the space is divided by a plane indicated by the dotted line in the figure, directions Y1 to Y8 (small spaces) are defined. That is, eight directions (small spaces) are defined from the measurement point, such as direction Y1, direction Y2, ..., direction Y8, and the number of flow regulators contained in each direction (small space) is counted. 3(a), with respect to measurement point 210y, flow regulator 221y exists in direction (small space) Y1, flow regulators 222y and 223y in direction (small space) Y2, flow regulators 224y and 225y in direction (small space) Y3, flow regulator 226y exists in direction (small space) Y4, flow regulators 228y, 229y, and 230y in direction (small space) Y5, flow regulator 226y exists in direction (small space) Y6, and flow regulator 227y exists in direction (small space) Y7. Also, no flow regulator exists in direction (small space) Y8. Therefore, the relative positional relationship between the measurement point 210y in FIG. 3(a) and the flow regulator can be expressed by one-dimensional information {Y1, Y2, Y3, Y4, Y5, Y6, Y7, Y8}={1, 2, 2, 1, 3, 1, 1, 0}.

[0045] FIG. 3(b) shows an example in which the arrangement of the measurement point 210y and the flow regulators 221y to 230y is the same as in FIG. 3(a), but the way in which the directions (small spaces) are defined is changed. The example in FIG. 3(b) shows an example in which six directions (small spaces) Y11 to Y16 are defined based on the measurement point. That is, the six directions (small spaces) are Y11 and Y12 in the vertical directions of the measurement point and Y13 to Y16 in the circumferential directions. In the example in FIG. 3(b), flow regulators 221, 222y, 229y, and 230y exist in direction Y11, flow regulators 226y and 227y exist in direction Y12, flow regulator 228y exists in direction Y13, flow regulator 224y exists in direction Y14, and flow regulators 224y and 225y exist in direction Y15, and no flow regulator exists in direction Y16. Therefore, in the case of FIG. 3(b), the relative positional relationship between the measurement point 210y and the flow regulator can be expressed by one-dimensional information of {Y11, Y12, Y13, Y14, Y15, Y16}={4, 2, 1, 1, 2, 0}.

[0046] In the examples of Figures 3(a) and (b), the areas are not divided according to the distance from the measurement point, but as in Figure 2B, the areas for determining the relative positional relationship may be divided according to the distance from the measurement point.

[0047] 2A(a), (b), 2B, 3(a), and 3(b) are merely examples. As shown in FIGS. 2A, 2B, and 3, by counting the number of flow regulators contained in a small space defined by a direction and distance defined by a rough direction and distance with respect to a measurement point in a predetermined space 200, the relative positional relationship between each measurement point and the flow regulator can be defined using information that is easy to process, without strictly determining it. Therefore, in other words, the conversion unit 142 defines a plurality of small spaces based on a specific region as information regarding the relative positional relationship between a specific region in the predetermined space and a flow regulator with respect to the specific region, and converts the information into one-dimensional information indicating the number of flow regulators contained in each of the plurality of small spaces.

[0048] Returning to Figure 1, the estimation unit 143 inputs the information received by the reception unit 141 (information representing the state of the specified space and information regarding the measurement point) and one-dimensional information indicating the relative positional relationship between the measurement point and the flow regulator generated by conversion by the conversion unit 142 into a learning model 151 stored in the memory unit 150, and estimates the flow (flow of liquid, gas, solid, heat, etc., changes such as reaction of materials) at the measurement point within the specified space 200 formed according to the location where the flow regulator is installed at the desired measurement point within the specified space 200.

[0049] The learning unit 144 receives, as teacher data, information that associates information representing the state of the predetermined space, information relating to specific regions (measurement points) included in the predetermined space, one-dimensional information indicating the relative positional relationship between the measurement points included in the predetermined space and the flow regulator, and information relating to the flow within the predetermined space in that case, from the communication unit 120 or the input unit 130. In this case, the learning unit 144 uses the received teacher data to learn the relationship between the information representing the state of the predetermined space, information relating to specific regions (measurement points) included in the predetermined space, and one-dimensional information indicating the relative positional relationship between the measurement points included in the predetermined space and the flow regulator, and information relating to the flow within the predetermined space in that case.

[0050] Alternatively, the learning unit 144 receives information representing the state of the predetermined space, information relating to specific regions (measurement points) included in the predetermined space, information indicating the positions of the measurement points and the flow regulator included in the predetermined space, and information relating to the flow within the predetermined space in that case from the communication unit 120 or the input unit 130. In this case, the learning unit 144 first causes the conversion unit 142 to convert the information indicating the positions of the measurement points and the flow regulator included in the predetermined space into one-dimensional information indicating the relative positional relationship between the measurement points and the flow regulator, and then learns the relationship between the information representing the state of the predetermined space, information relating to specific regions (measurement points) included in the predetermined space, and the one-dimensional information indicating the relative positional relationship between the measurement points and the flow regulator included in the predetermined space, and the information relating to the flow within the predetermined space in that case.

[0051] The learning unit 144 learns the relationship between information representing the state of a predetermined space, information relating to specific regions (measurement points) included in the predetermined space, one-dimensional information indicating the relative positional relationship between the measurement points included in the predetermined space and the flow regulator, and information relating to the flow within the predetermined space in that case, and generates a learning model 151. The learning unit 144 may store the generated learning model 151 in the storage unit 150.

[0052] Furthermore, the learning unit 144 may perform re-learning by receiving input of the learning model 151 already stored and new information related to the learning model 151. The new information related to the learning model 151 is basically new training data that associates information representing the state of a predetermined space that has already been learned, information about the measurement points, and the relative positional relationship between the measurement points and the flow regulator, with information about the flow in the predetermined space in that case.

[0053] The storage unit 150 is a storage medium that stores various programs and data required by the information processing device 100. The storage unit 150 may be realized, for example, by a hard disk drive (HDD), a solid state drive (SSD), a flash memory, or the like. The storage unit 150 may also be used as a read-only memory (ROM) or a random access memory (RAM) that serves as a working area when the control unit 140 executes processing.

[0054] The storage unit 150 stores a learning model 151. This learning model 151 may be a learning model generated by the learning unit 144, or may be generated by an external device. The learning model 151 is a learning model that has learned the relationship between information representing the state of a predetermined space, information about measurement points, information indicating the relative positional relationship between the measurement points and flow regulators (one-dimensional information), and information about the flow within the predetermined space in that case. This learning model can also be said to be a model that defines multiple small spaces based on a specific region as information about the relative positional relationship between a specific region (measurement point) in the predetermined space and the flow regulator with respect to the specific region, and has learned the number of flow regulators included in the multiple small spaces.

[0055] The output unit 160 outputs information related to the flow estimated by the control unit 140. The information output by the output unit 160 may be transmitted to an external device via the communication unit 120, or may be output as images or text to a monitor (not shown) installed in or connected to the information processing device 100, or may be output as sound from a speaker.

[0056] <Example of operation> 4 is a flowchart showing an example of a process for generating the learning model 151. Here, an example is shown in which the learning unit 144 generates the learning model 151, but as described above, the learning model 151 may be generated by another device.

[0057] 4, the learning unit 144 receives input of information indicating the state of a predetermined space and information regarding a flow regulator included in the predetermined space from the communication unit 120 or the input unit 130 (step S401). The information regarding the flow regulator here may be information indicating the arrangement position of the flow regulator within the predetermined space. The communication unit 120 or the input unit 130 transmits the received information to the control unit 140.

[0058] In addition, the communication unit 120 or the input unit 130 accepts input of information regarding a specific area (measurement point) within a specified space, i.e., including location information indicating the position of the measurement point (step S402), and transmits it to the control unit 140.

[0059] In addition, the communication unit 120 or the input unit 130 accepts input of information regarding the flow at the measurement point as a specific area accepted in step S402 when the flow regulator indicated by the information indicating the state of the specified space accepted in step S401 and the information regarding the flow regulator included in the specified space is placed (step S403), and transmits this to the control unit 140.

[0060] The conversion unit 142 converts the information about the flow regulator received in step S401 into one-dimensional information indicating the relative positional relationship between the measurement point and the flow regulator, based on the positional information of the measurement point received in step S402 (step S404). Note that if the information about the flow regulator received in step S401 is one-dimensional information indicating the relative positional relationship between the measurement point as a specific region and the flow regulator, the processes of steps S402 and S404 may be omitted.

[0061] Based on the information representing the state of the specified space received in step S401, the information relating to the flow at the measurement points received in step S403, and the one-dimensional information representing the relative positional relationship between the measurement points and the flow regulator converted in step S404, the learning unit 144 learns the relationship between (i) the information representing the state of the specified space, the information relating to specific regions (measurement points) included in the specified space, and the information representing the relative positional relationship between the measurement points corresponding to the specific regions and the flow regulator (one-dimensional information), and (ii) the information relating to the flow in the specific regions (measurement points) in that case (step S405).

[0062] The learning unit 144 performs learning on the plurality of pieces of teacher data generated from step S401 to step S404. The algorithm used by the learning unit 144 for learning may be an existing algorithm for deep learning.

[0063] The learning unit 144 generates a learning model 151 based on the learning in step S404. Then, the learning unit 144 stores the learning model 151 in the storage unit 150 (step S406) and ends the process. Through this process, the learning model 151 used in the estimation process is generated and stored in the storage unit 150 of the information processing device 100.

[0064] FIG. 5 is a flowchart showing an example of operation of the flow estimation process performed by the information processing device 100.

[0065] The receiving unit 141 of the control unit 140 receives input of information representing the state of a predetermined space and information regarding a flow regulator included in the predetermined space from the communication unit 120 or the input unit 130 (step S501). The information received here is information that the user wants to know about how the flow will be when processing is performed with the state of the predetermined space and the arrangement of the flow regulator based on the input information.

[0066] Furthermore, the receiving unit 141 receives information including position information indicating the position of a measurement point as a specific area in a predetermined space from the communication unit 120 or the input unit 130 (step S502).

[0067] The conversion unit 142 of the control unit 140 converts the information about the flow regulator received in step S401 into one-dimensional information indicating the relative positional relationship of the flow regulator with respect to the measurement point received in step S502 (step S503).

[0068] The estimation unit 143 inputs the information representing the state of the predetermined space received by the receiving unit 141 and the one-dimensional information generated in step S503 into the learning model 151 stored in the storage unit 150 (step S504). As described above, the learning model 151 is a model that has learned the relationship between the information representing the state of the predetermined space, information relating to specific regions (measurement points) included in the predetermined space, information representing the relative positional relationship between the specific regions (measurement points) included in the predetermined space and the flow regulator, and information relating to the flow within the predetermined space.

[0069] As a result, the estimation unit 143 acquires information indicating the flow in a predetermined space in the case of the arrangement of the flow regulating body accepted in step S501. Then, the estimation unit 143 causes the output unit 160 to output the acquired information (step S505), and ends the process.

[0070] This allows a user to obtain information about the flow in a specific area (measurement point) within a specific space simply by inputting information that represents the state of the specific space, information about the placement of the flow regulator that the user has devised, and information including position information about the specific area (measurement point). Therefore, from the obtained flow information, it is possible to determine whether the input information about the placement of the flow regulator is desirable for the user. Furthermore, it is also possible to reverse-calculate a more ideal placement from the obtained flow information.

[0071] Furthermore, because the system utilizes a learning model that has learned the relationship between information representing the state of a predetermined space, information about specific regions (measurement points) included in the predetermined space, and information about the relative positional relationship between the specific region in the predetermined space and the flow regulator relative to the specific region, and information about the flow within the predetermined space at that time, it is possible to optimize not only the placement of the flow regulator but also the state of the predetermined space, depending on the target of information processing. In other words, by utilizing a learning model that has learned the relationship between information representing the state of a predetermined space, information about specific regions (measurement points) included in the predetermined space, and information about the relative positional relationship between the specific region and the flow regulator, and information about the flow within the predetermined space, the estimation process performed by the estimation unit 143 can determine whether the obtained flow information is desirable or not, and this can be used as a criterion for determining whether the state representing the input state of the predetermined space was appropriate. Furthermore, it is possible to reverse-calculate a more ideal state of the predetermined space from the obtained flow information.

[0072] Furthermore, as shown in the above embodiment, by using one-dimensional information on the flow regulator included in each small space when multiple small spaces included in a predetermined space are defined as information to be input to the learning model and as information on the arrangement of the flow regulator as part of the teacher data when generating the learning model, it is possible to improve the estimation function of the information processing device 100. The one-dimensional information will be explained in further detail in Examples 1 to 5 below, but it can be said that this is information obtained by converting information on the arrangement of the flow regulator included in the predetermined space.

[0073] In particular, in this embodiment, by utilizing the relative positional relationship between a specific region (measurement point) and the flow regulator, even if the positional relationship between the measurement point and the flow regulator appears different (when viewed from the same direction), there may be a similar relationship or the same information in terms of one-dimensional information, and in such cases, it can be estimated that a similar flow occurs at the measurement point.

[0074] As mentioned above, if the distribution of multiple flow regulators were to be directly expressed using information that simply indicates the coordinates of each flow regulator, such as the relative positional relationship between each flow regulator and a specific region within a specified space, the amount of data would be enormous.Even if this could be expressed, the input data for the learning model would be too complicated, and there would be no guarantee that learning could be performed for the learning model so that appropriate estimations could be made, making it unrealistic.

[0075] In contrast, the one-dimensional information can express the distribution of multiple flow regulators as one-dimensional information representing relative positional relationships seen from measurement points, and can effectively express the characteristics of the arrangement of the flow regulators in a predetermined space. Therefore, according to the information processing device 100 shown in this embodiment, the learning model can easily learn the relationship between information representing the state of the predetermined space, information about specific regions (measurement points) included in the predetermined space, information about the relative positional relationships between the specific regions included in the predetermined space and the flow regulators, and information about the flow within the predetermined space in that case, thereby improving the estimation function acquired through learning. In short, in the information processing device 100 according to this embodiment, rather than learning two-dimensional (or three-dimensional) information such as simple position coordinates of two-dimensional (or three-dimensional) flow regulators, the learning target is one-dimensional information that can be expressed in one dimension, such as the positions of the flow regulators in each direction as seen from a specific region and the number of flow regulators included in each distance range. This simplifies the learning content for the information processing device, enabling more accurate learning, and as a result, improving estimation results.

[0076] Hereinafter, the estimation by the information processing device 100 as described above will be described using a specific example.

[0077] Example 1 In Example 1, a deposition reactor for carrying out a deposition reaction will be described as an example. That is, a deposition reactor for depositing silicon on a silicon rod in a bell jar apparatus will be described as an example. Note that although an example of depositing silicon on a silicon rod is shown here, the deposition reactor may be one for other deposition reactions as long as a substance based on an injected gas is deposited on a core body.

[0078] In the bell jar apparatus, silicon is deposited on the silicon rods by flowing gas, and in this case, it is preferable that silicon is deposited as uniformly as possible on the surface of all the silicon rods.

[0079] FIG. 6(a) is a perspective view schematically illustrating a bell jar apparatus reaction vessel 200a serving as a deposition reactor. FIG. 6(b) is a schematic side cross-sectional view of the bell jar apparatus reaction vessel 200a. As shown in FIG. 6, gas nozzles 220a are provided in the bell jar apparatus reaction vessel 200a for injecting gases related to the substance to be deposited in the reaction vessel. In the example of FIG. 6, five gas nozzles 220a are shown, but the number of gas nozzles 220a may be any number. In addition, a silicon rod, which serves as a core for depositing the substance, is erected in the reaction vessel. In the example of FIG. 6, four silicon rods are shown provided in the reaction vessel 200a, but the number of silicon rods may be any number.

[0080] 7A and 7B are diagrams showing examples of small spaces included in the predetermined space in Example 1, and examples of the arrangement of flow regulators arranged therein. Using these diagrams, the input format of information input to the learning model 151 of the information processing device 100 in each case will be described. FIG. 7A(a) shows, as an example, an example in which four silicon rods 210a-213a and eight gas nozzles 220a-227a are arranged in a predetermined space 200a.

[0081] In the example shown in FIG. 7A(b), small spaces A1 to A3 are defined by dotted lines to define the distance from each silicon rod. As an example, two gas nozzles 225a and 227a are present in small space A1 as viewed from silicon rod 210a. The specific region may be the entire silicon rod or a portion of its surface. Furthermore, no gas nozzle is present in small space A2 as viewed from silicon rod 210a. Similarly, three gas nozzles 223a, 224a, and 226a are present in small space A3 as viewed from silicon rod 210a.

[0082] In the illustrated example, the dotted lines defining the boundaries of the small spaces overlap on the gas nozzles 223a, 224a, and 226a, and as viewed from the silicon rod 210a, the gas nozzles 223a, 224a are positioned across two small spaces. For example, a portion of the gas nozzles 223a and 224a is positioned within the small space A2, and another portion is positioned within the small space A3. Also, a portion of the gas nozzle 226a is positioned within the small space A3, and another portion is positioned outside the small space A3. In such cases, the gas nozzles 223a, 224a, and 226a serving as flow regulators may be positioned within the small space that occupies the larger area, or the coefficient may be determined based on the ratio of the area occupied by each of the gas nozzles 223a, 224a, and 226a. For example, the area occupied by the gas nozzles within a small space may be defined as 1 when the area of ​​one entire gas nozzle is positioned within the small space, and the area may be used as a numerical value indicating the number of flow regulators present in the small space. Therefore, in the illustrated example, if the gas nozzles are defined as being present in only one of the spaces, then it can be said that two gas nozzles exist in the small space A1, no gas nozzles exist in the small space A2, and three gas nozzles exist in the small space A3. Also, if the gas nozzles included in each small space are defined by the area of ​​the gas nozzles, then, for example, it can be defined that 2.0 gas nozzles exist in the small space A1, 0.5 gas nozzles (part of the gas nozzles 223a and 224a) exist in the small space A2, and 1.9 gas nozzles exist in the small space A3.

[0083] 7A(b), the one-dimensional information indicating the relative positional relationship between the silicon rod 210a and the gas nozzle serving as the flow regulator can be defined as {A1, A2, A3}={2, 0, 3} or {A1, A2, A3}={2.0, 0.5, 1.9}. Which one is used as the one-dimensional information depends on which aspect of the relative positional relationship was used when creating the learning model 151.

[0084] FIG. 7A(b) shows an example in which the relative positional relationship is not determined in the direction from the gas nozzle.

[0085] Fig. 7B is a diagram illustrating another example of a method for defining a small space different from that shown in Fig. 7A. Fig. 7B(a) is a diagram illustrating another example of a relative positional relationship between the silicon rod 210a and the gas nozzle.

[0086] Fig. 7B(a) shows an example in which the direction and distance from each silicon rod (defining the small spaces) are defined by the dotted lines to determine the relative positional relationship. Fig. 7B(a) shows an example in which the relative positional relationship is defined in four directions and three distances based on each silicon rod, and among these, 12 small spaces A11 to A14, A21 to A24, and A31 to A34 relative to the silicon rod 210a are shown. In the example of Fig. 8B(a), the small space A11, the small space A21, and the small space A31 have the same direction but different distances from the silicon rod 210a. The small space A12, the small space A22, the small space A32, ..., the small space A14, the small space A24, and the small space A34 have a similar relationship. In the example of Figure 7B(a), the direction defining the small spaces is determined based on the orientation of the silicon rod 210a as viewed from the center point 230a of the reaction vessel 200a, and the small spaces A11, A21, and A31 are located outside the reaction vessel 200a, while the small spaces A13, A23, and A33 are located inside the reaction vessel 200a.

[0087] As shown in FIG. 7B(a), when a small space is defined to identify the relative positional relationship based on the direction and distance from each silicon rod, in the example of FIG. 7B(a), the one-dimensional information indicating the relative positional relationship of the gas nozzles as seen from the silicon rod 210a can be defined as {A11, A12, A13, A14, A21, A22, A23, A24, A31, A32, A33, A34}={0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 2} or {A11, A12, A13, A14, A21, A22, A23, A24, A31, A32, A33, A34}={0, 1, 1, 0, 0, 0, 0.25, 0.25, 0, 0, 0.75, 1.15}. 7A and 7B show an example in which the distance from the center point 230a of the reaction vessel to each silicon rod is the same, but the distance from the center point 230a to each silicon rod may be different. If the distance from the center point 230a to each silicon rod is different, a variable indicating each distance may be prepared separately, and one-dimensional information indicating the relative positional relationship between the specific region and the flow regulator may be obtained. In this way, information on the specific region, specifically, information on the absolute position (universal position) of the specific region, can be provided.

[0088] 7B(b) is a diagram illustrating an example of the relative positional relationship of the gas nozzles as viewed from silicon rods 211a, 212a, and 213a. As shown in FIG. 7B(b), one-dimensional information indicating the relative positional relationship of the gas nozzles as viewed from silicon rod 211a can be defined as {A1, A2, A3}={0, 1, 2}. Similarly, for silicon rod 212a, {A1, A2, A3}={0, 0, 3} can be defined, and for silicon rod 213a, {A1, A2, A3}={0, 1, 2} can be defined.

[0089] In the examples of Figures 7A(b) and 7B(b), the relative positional relationship between each of the four silicon rods and the gas nozzle can be obtained. That is, by examining information regarding the flow of each of the four silicon rods, it is possible to obtain, for one space state, four types of information regarding the relative positional relationship between the silicon rod and the gas nozzle and the relationship between the precipitation reaction in the silicon rod. Therefore, more information can be obtained than by simply examining the placement position of the flow regulator in the specified space 200 and the precipitation reaction at that time.

[0090] Therefore, when one-dimensional information indicating the relative positional relationship between a specific region and a flow regulator is used, it can be said that a learning model 151 for estimation can be generated with a smaller amount of training data than when simply using the placement position of the flow regulator in the predetermined space 200 and the precipitation reaction at that time as training data. In other words, it is possible to reduce the amount of training data that needs to be prepared.

[0091] 8A and 8B are diagrams showing examples of small spaces included in the predetermined space in Example 1 and examples of the arrangement of flow regulators disposed therein, which are different from those shown in Figs. 7A and 7B. Using these diagrams, the input format of information input to the learning model 151 of the information processing device 100 in each case will be described.

[0092] FIG. 8A(a) is a plan view illustrating an example in which the space inside the reaction vessel 200a of the bell jar apparatus is regarded as a predetermined space, silicon rods 210a (210b, 210c, 210d) are used as specific regions, and gas nozzles 220a to 229a are used as flow regulators.

[0093] FIG. 8A(b) shows an example of determining the relative positional relationship between the silicon rod as a specific region and the gas nozzle as a flow regulator based on the silicon rod.

[0094] In the example shown in FIG. 8A(b), directions (small spaces) A1 to A8 are defined by dividing the area with dotted lines to define the directions as viewed from each silicon rod 213a. As an example, one gas nozzle 220a exists in the direction (small space) A1 as viewed from the silicon rod 213a. Note that the specific region may be the entire silicon rod or a part of its surface. Furthermore, one gas nozzle 221a exists in the direction (small space) A2 as viewed from the silicon rod 213a. Similarly, three gas nozzles 222a, 223a, and 229a exist in the direction (small space) A3 as viewed from the silicon rod 213a.

[0095] In the illustrated example, the gas nozzle 223a is overlapped with a dotted line that defines the boundary for defining the direction (small space). When viewed from the silicon rod 213a, a portion of the gas nozzle 223a is included in the direction (small space) A3, and another portion is included in the direction (small space) A4. In such a case, the gas nozzle 223a serving as the flow regulator may be present in the direction in which the area occupied by the gas nozzle 223a is larger, or the coefficient may be calculated based on the ratio of the area occupied by the gas nozzle 223a. For example, the area occupied by the gas nozzles included in the small space may be used as a numerical value indicating the number of flow regulators present in the direction (small space) when the area of ​​one gas nozzle in its entirety is set to 1. Therefore, in the illustrated example, when defining the gas nozzles present in only one direction, it can be said that there are three gas nozzles in the direction (small space) A3. When defining the gas nozzles included in each direction (small space) by the area of ​​the gas nozzles, it can be defined, for example, that there are 2.6 gas nozzles in the direction (small space) A3. Similarly, it may be specified that two gas nozzles, gas nozzles 224a and 228a, exist in direction (small space) A4, or that 2.4 gas nozzles, gas nozzles 224a, 228a, and 223a, exist. Furthermore, gas nozzle 225a exists in direction (small space) A5, and gas nozzle 226a exists in direction (small space) A6. Furthermore, gas nozzle 227a exists in direction (small space) A7, or that 0.55 gas nozzle 227a exists. Furthermore, it may be specified that no gas nozzle exists in direction (small space) A8, or that 0.45 gas nozzle 227a exists.

[0096] 8A(b), the one-dimensional information indicating the relative positional relationship between the silicon rod 213a and the gas nozzle serving as the flow regulator can be defined as {A1, A2, A3, A4, A5, A6, A7, A8} = {1, 1, 3, 2, 1, 1, 1, 0} or {A1, A2, A3, A4, A5, A6, A7, A8} = {1, 1, 2.6, 2.4, 1, 1, 0.55, 0.45}. Which one is used as the one-dimensional information depends on which aspect of the relative positional relationship was used when creating the learning model 151.

[0097] FIG. 8A(b) shows an example in which the relative positional relationship is not determined in the distance direction from the gas nozzle.

[0098] Fig. 8B is a diagram for explaining another example of a method for defining a small space different from that shown in Fig. 8A. Fig. 8B(a) is a diagram for explaining an example of a relative positional relationship between the silicon rod 210a and the gas nozzle as seen from the silicon rod 210a.

[0099] Fig. 8B(a) shows an example in which the direction and distance from each silicon rod are defined (small spaces are defined) using dotted lines to determine the relative positional relationship. Fig. 8B(a) shows an example in which the relative positional relationship is defined in eight directions and two distances using the silicon rod 210a as the reference, and shows 16 small spaces A11 to A18 and A21 to A28 using each silicon rod as the reference. In the example of Fig. 8B(a), the small space A11 and the small space A21 have the same direction, but the small spaces A12 and A22, the small spaces A13 and A23, ..., the small spaces A18 and A28 also have similar relationships relative to the silicon rod 210a.

[0100] As shown in FIG. 8B(a), when a small space for specifying the relative positional relationship based on the direction and distance from each silicon rod is defined, in the example of FIG. 8B(a), one-dimensional information indicating the relative positional relationship of the gas nozzles as seen from the silicon rod 210a is expressed as {A11, A12, A13, A14, A15, A16, A17, A18, A21, A22, A23, A24, A25, A 26, A27, A28} = {0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0} or {A11, A12, A13, A14, A15, A16, A17, A18, A21, A22, A23, A24, A25, A26, A27, A28} = {0, 0, 0, 0, 0, 0, 0, 0, 0.45, 0.55, 0, 0, 1, 0, 0, 0}.

[0101] FIG. 8B(b) is a diagram illustrating an example of the relative positional relationship between the silicon rod 211a and the gas nozzle. FIG. 8B(b) illustrates an example in which the relative positional relationship is defined in eight directions based on the silicon rod 211a. Assume that small spaces A1 to A8 are defined as shown. Then, one-dimensional information indicating the relative positional relationship of the gas nozzles as viewed from the silicon rod 211a can be defined as {A1, A2, A3, A4, A5, A6, A7, A8} = {1, 1, 3, 2, 1, 1, 1, 0}. This one-dimensional information is the same as the one-dimensional information indicating the relative positional relationship of the gas nozzles with respect to the silicon rod 210a shown in FIG. 8A(b). In other words, the learning results of the precipitation reaction when small spaces are defined for the silicon rod 210a as shown in FIG. 8A(b) and gas nozzles are arranged can be applied to the silicon rod 211a in FIG. 8B(b).

[0102] The relative positional relationship between the silicon rod as the specific region and the gas nozzle as the flow regulator shown in FIGS. 8A and 8B is an example, and the small space may be defined in any direction and at any distance.

[0103] FIG. 9 is a flowchart showing an example of an operation related to generation of a learning model by the information processing device 100 in which the flowchart of FIG. 4 is applied to the first embodiment.

[0104] As shown in FIG. 9, the learning unit 144 of the control unit 140 of the information processing device 100 receives input of information such as the size, volume, and shape of the reaction vessel of the deposition reactor, rod information regarding the arrangement of silicon rods, and nozzle information regarding the arrangement of gas nozzles (step S901).

[0105] Based on the received information, the learning unit 144 receives information about precipitation in the silicon rods when a precipitation reaction is carried out under predetermined conditions (step S902). Here, the information about precipitation may be information about the heat flux in the reaction vessel, the heat flux near each silicon rod, the thickness of silicon precipitated on the surface of each silicon rod, etc.

[0106] The conversion unit 142 converts the nozzle information regarding the arrangement of the gas nozzles received in step S901 into one-dimensional information indicating the relative positional relationship between the gas nozzles as flow regulators and the silicon rods as specific regions in the deposition reactor based on the rod information regarding the arrangement of the silicon rods (step S903). That is, the conversion unit 142 generates one-dimensional information that can be expressed by small spaces that define the direction and distance as seen from the silicon rods and the number of gas nozzles included in the small spaces.

[0107] The learning unit 144 learns the respective relationships by using the one-dimensional information generated in step S903, i.e., the relative positional relationship between the silicon rod and the gas nozzle, as information regarding the relative positional relationship with the flow regulator for a specific region, and the information regarding precipitation in the silicon rod received in step S902 as information regarding the flow within the specified space 200 (step S904).

[0108] The learning unit 144 learns a plurality of relationships between information representing the state of the predetermined space, information about each silicon rod (including position information indicating the position of the silicon rod in the predetermined space), information indicating the relative positional relationship of the gas nozzles as seen from the silicon rod, and information about precipitation on the silicon rod, and generates a learning model. That is, learning is performed using a combination of information representing the state of the predetermined space, information indicating the number of gas nozzles included in a small space that defines the direction and distance as seen from the silicon rod, and information about precipitation on the silicon rod as training data.

[0109] The learning unit 144 stores the generated learning model 151 in the storage unit 150 (step S905), and ends the process.

[0110] FIG. 10 is a flowchart showing an example of an estimation operation performed by the information processing device 100 to obtain information about precipitation using the learning model 151 stored in the storage unit 150.

[0111] 10, the receiving unit 141 of the control unit 140 receives information indicating the state inside the deposition reactor and placement information for placing gas nozzles in the deposition reactor as information of the flow regulator (step S1001). The receiving unit 141 receives, for example, placement information indicating the placement of gas nozzles input by the user to the input unit 130.

[0112] The receiving unit 141 also receives rod information indicating the placement position of the silicon rod to be placed in the deposition reactor as information on the specific region (step S1002).

[0113] The conversion unit 142 of the control unit 140 converts the arrangement information indicating the arrangement of the gas nozzles received by the reception unit 141 into one-dimensional information indicating the relative positional relationship of the gas nozzles with respect to the silicon rod, based on the position of the silicon rod indicated in the rod information received by the reception unit 141 (step S1003). That is, the conversion unit 142 defines small spaces that specify the direction and distance for determining the relative positional relationship as seen from the silicon rod, counts the number of gas nozzles included in the small spaces, and generates a set indicating the number of gas nozzles included in each small space as one-dimensional information indicating the relative positional relationship between the silicon rod and the gas nozzles. Note that the definition of the multiple small spaces at this time is the same as the definition of the multiple small spaces used when training the learning model 151. In other words, for example, when the small spaces shown in Figures 7A(b) and 7B(a) are used to create the learning model 151, the small spaces A1 to A3, A11 to A14, A21 to A24, and A31 to A34 are also used at the stage of estimation by the estimation unit 143, and the number of gas nozzles contained in each of the small spaces A1 to A3, A11 to A14, A21 to A24, and A31 to A34 is counted.

[0114] The estimation unit 143 inputs information representing the state of the predetermined space, identification information for identifying the small space, and one-dimensional information showing the relative positional relationship between the silicon rod and the gas nozzle, i.e., one-dimensional information associating the identification information of the small space with the number of gas nozzles included in the small space, into the learning model 151 read from the storage unit 150 (step S1004). By inputting the one-dimensional information into the learning model 151, information regarding the precipitation reaction can be obtained.

[0115] The estimation unit 143 instructs the output unit 160 to output information about the heat flux around the silicon rod as information about the precipitation reaction obtained from the learning model 151 (step S1005), and ends the process.

[0116] This allows the user to recognize whether the input gas nozzle arrangement is an arrangement that can obtain desirable results, without actually performing a simulation.

[0117] Note that the information about the flow regulator included in each small space received by the learning unit 144 may be the sum of the gas flow rates flowing in or out of the gas nozzles included in each small space, instead of the number of gas nozzles included in each small space, or may be both (both the number of gas nozzles and the inflow rates, both the number of gas nozzles and the outflow rates, both the gas inflow rates and the gas outflow rates, or all of them). For example, in the case of FIG. 8B(b), the gas flow rates flowing into a predetermined space from gas nozzles 220a, 221a, 222a, 223a, 224a, 225a, 226a, 227a, 228a, and 229a are assumed to be G0, G1, G2, G3, G4, G5, G6, G7, G8, and G9, respectively. Then, if the number of gas nozzles included in each small space is {A1, A2, A3, A4, A5, A6, A7, A8} = {1, 1, 3, 2, 1, 1, 1, 0} as described above, the one-dimensional information based on the sum of the gas flow rates flowing in from the gas nozzles can be expressed as {A1, A2, A3, A4, A5, A6, A7, A8} = {G0, G1, G2 + G3 + G4, G5 + G6, G7, G8, G9, 0}. Furthermore, when using the number of gas nozzles and the gas inflow rate, if the small spaces for defining the relative positional relationship based on the number of gas nozzles are assumed to be AN1 to AN8 (corresponding to A1 to A8 in Figure 8B(b)), and the small spaces for defining the relative positional relationship based on the gas inflow rate in the gas nozzles are assumed to be AG1 to AG8 (corresponding to A1 to A8 in Figure 8B(b)), then it can also be expressed as {AN1, AN2, AN3, AN4, AN5, AN6, AN7, AN8, AG1, AG2, AG3, AG4, AG5, AG6, AG7, AG8} = {1, 1, 3, 2, 1, 1, 1, 0, G0, G1, G2+G3+G4, G5+G6, G7, G8, G9, 0}.

[0118] In FIG. 10, the conversion unit 142 counts the number of gas nozzles contained in each small space to generate one-dimensional information, but in step S1001, it may also be possible to receive information indicating the small space and information indicating the number of gas nozzles contained in that small space, in which case the processing of step S1002 can be omitted.

[0119] According to this Example 1, information about the deposition reaction can be obtained by inputting information representing the state inside the deposition reactor, information about each silicon rod, and information indicating the arrangement of the gas nozzles or the relative positional relationship between the silicon rods and the gas nozzles, i.e., the small spaces and the number of gas nozzles contained in the small spaces.

[0120] In Example 1, the inventors created a learning model by inputting multiple pieces of randomly generated information (randomly generated information representing the state of a given space, information about the arrangement of gas nozzles in the small space, and information about the convective heat flux on the surface of the silicon rod at that time). Then, they evaluated the estimation results obtained by inputting different random information (randomly generated information representing the state of a given space and information about the arrangement of gas nozzles in the small space) to the created learning model. Specifically, when the results of an experiment conducted under the same conditions as the information input to the learning model were compared with the estimation results of the learning model, the coefficient of determination for the estimation results by the learning model was found to be 0.7 or higher. In other words, the accuracy (correctness) of the estimation results by the learning model was found to be 70% or higher. This coefficient of determination can be further improved by increasing the number of training data (the number of training data) or by selecting the training data.

[0121] Although the above example describes randomly generated information, it goes without saying that similar results can be obtained by learning pre-specified information in the same way, rather than randomly generated information.

[0122] On the other hand, the inventors also created and evaluated a learning model in which the information was handled using x-y coordinates rather than information indicating the relative positional relationship of the gas nozzles as seen from the silicon rod. The coefficient of determination was found to be less than 0.4. In other words, it can be said that the performance of predicting a flow is significantly improved by using a learning model that learns the relative positional relationship of the gas nozzles as seen from the silicon rod, rather than by using a learning model in which the flow regulator in a predetermined space is defined by coordinates indicating its placement position. Therefore, in a case where the flow regulator is disposed in a predetermined space, in a mode in which a flow is predicted, the estimation result is expected to be improved more than by using the position coordinates of the flow regulator in the predetermined space. This also applies to Examples 2 to 5 described below.

[0123] <Example 2> The above Example 1 shows an example of estimation in a precipitation reactor. In Example 2, an example will be described in which gas bubbles are broken down as finely as possible when gas is mixed into powder or granular material in a fluidized bed reactor.

[0124] FIG. 11 is a schematic diagram of a fluidized-bed reactor. A fluidized-bed reactor 200b is filled with powder and granular material, and gas is mixed in. It is desirable that the gas bubbles be broken down into as small pieces as possible before being mixed in. However, as the bubbles rise inside the reactor, they coalesce and grow larger, so they must be broken down again by a baffle plate. Bubbles 1001 are broken down by colliding with the baffle plate 222b. Therefore, it is desirable to provide the baffle plate 222b so that the gas flowing in from the gas inlet (see the arrow in the drawing) can be broken down through collisions as much as possible, and the size of the bubbles becomes as uniform as possible. As shown in FIG. 11(a), bubbles 1001 coalesce and grow larger as they move toward the upper layer of the fluidized-bed reactor. However, after colliding with the baffle plate 222b, as shown in FIG. 11(b), the bubbles 1001 are broken down again.

[0125] In the second embodiment, the information processing device 100 estimates the size or particle size of the bubbles based on the arrangement of the baffle plate 222b serving as the flow regulator. That is, it estimates whether the bubbles are in a desired state (such as the particle size of the bubbles at the measurement point or the ratio of the bubbles to the powder or granular material within a predetermined range including the measurement point) at a desired location (measurement point) in the fluidized-bed reactor 200b.

[0126] FIG. 12 shows an example of defining small spaces for defining the relative positional relationship between the specific region and the flow regulator in Example 2, i.e., the relative positional relationship of the baffle plate 222b from the measurement point. FIG. 12(a) shows an example in which directions (small spaces) B1 to B6 are defined for a fluidized-bed reactor 200b as viewed from an arbitrary measurement point 210b within the fluidized-bed reactor 200b, as shown by the dotted line in the figure. When the small spaces are defined as shown in FIG. 12(a), one baffle plate is disposed in each of the directions (small spaces) B1, B2, B4, B5, and B6. In this case, one-dimensional information indicating the relative positional relationship between the measurement point 210b (the specific region) and the baffle plate (the flow regulator) can be expressed as {B1, B2, B3, B4, B5, B6} = {1, 1, 0, 1, 1, 1}, for example.

[0127] Also, Figure 12(b) is a diagram for explaining how the relative positional relationship with the obstruction plate is expressed as one-dimensional information when the position of measurement point 210b is changed from the position shown in Figure 12(a).

[0128] As shown in Figure 12(b), when the position of measurement point 210b is determined, and directions (small spaces) B1 to B6 viewed from measurement point 210b are defined based on measurement point 210b, the one-dimensional information indicating the relative positional relationship between measurement point 210 and obstruction plate 222b can be expressed as {B1, B2, B3, B4, B5, B6} = {0, 1, 3, 1, 0, 0}, for example.

[0129] FIG. 13 is a flowchart showing an example of an operation related to generation of a learning model by the information processing device 100 in which the flowchart of FIG. 4 is applied to the second embodiment.

[0130] As shown in FIG. 13, the learning unit 144 of the control unit 140 of the information processing device 100 receives input of information such as the size, volume, and shape of the fluidized bed reactor, information on the placement positions of the gas inlet holes, and placement information on the placement of the baffle plates (step S1301).

[0131] Based on the received information, the learning unit 144 receives information about changes in bubbles in the fluidized bed reactor when gas is introduced through the gas inlet under predetermined conditions (step S1302). Here, the information about changes in bubbles may be information about the size and particle size of bubbles at a desired measurement position.

[0132] The learning unit 144 receives input of information indicating the positions of measurement points at which the state of bubbles is measured as information indicating a specific region (step S1303).

[0133] The conversion unit 142 converts the information on the position of the baffle plates received in step S1301 into one-dimensional information indicating the relative positional relationship between the baffle plates as flow regulators and the measurement points as specific regions in the deposition reactor based on the information on the position of the measurement points received in step S1303 (step S1304). That is, the conversion unit 142 generates one-dimensional information that can be expressed by small spaces that regulate the direction and distance as seen from the measurement points and the number of baffle plates included in the small spaces.

[0134] The learning unit 144 uses the one-dimensional information generated in step S1304, i.e., information indicating the relative positional relationship between the measurement point and the baffle plate, as information regarding the relative positional relationship with the flow regulator for a specific region, and uses the information regarding the state of bubbles at the measurement point received in step S1302 as information regarding the flow within a predetermined space, and learns the respective relationships (step S1305). The learning unit 144 learns, as training data, relationships between information indicating the state within the fluidized bed reactor, information indicating the position of the measurement point within the fluidized bed reactor, the number of baffle plates included in each small space for specifying the relative positional relationship of the baffle plate as seen from the measurement point, and information regarding the state of bubbles at the measurement point in the fluidized bed reactor in that case, and generates a learning model.

[0135] The learning unit 144 stores the generated learning model 151 in the storage unit 150 (step S1306), and ends the process.

[0136] FIG. 14 is a flowchart showing an example of an estimation operation performed by the information processing device 100 to obtain information about the state of bubbles using the learning model 151 stored in the storage unit 150.

[0137] 14, the receiving unit 141 of the control unit 140 receives placement information for placing baffle plates in the fluidized bed reactor as information on the flow regulator (step S1401). The receiving unit 141 receives, for example, placement information indicating the placement of the baffle plates input by the user to the input unit 130.

[0138] Furthermore, the receiving unit 141 receives position information of the measurement point in the fluidized bed reactor as information on the specific region (step S1402).

[0139] The conversion unit 142 of the control unit 140 converts the placement information indicating the placement of the obstruction boards received by the reception unit 141 into one-dimensional information indicating the relative positional relationship of the obstruction boards with respect to the measurement points, based on the positional information of the measurement points received by the reception unit 141 (step S1403). That is, the conversion unit 142 defines small spaces that specify the direction and distance for determining the relative positional relationship as seen from the measurement points, counts the number of obstruction boards included in the small spaces, and generates a set indicating the number of obstruction boards included in each small space as one-dimensional information indicating the relative positional relationship between the measurement points and the obstruction boards. Note that the definition of the multiple small spaces at this time is the same as the definition of the multiple small spaces used when training the learning model 151.

[0140] The estimation unit 143 inputs information representing the state of a predetermined space, information relating to the position of the measurement point in the fluidized bed reactor, identification information for identifying the small space, and one-dimensional information representing the relative positional relationship between the measurement point and the baffle plates, i.e., one-dimensional information associating the identification information of the small space with the number of baffle plates included in the small space, into the learning model 151 read from the storage unit 150 (step S1404). By inputting the one-dimensional information into the learning model 151, information relating to the state of bubbles at the measurement point can be obtained.

[0141] The estimation unit 143 instructs the output unit 160 to output information about the state of the bubbles obtained from the learning model 151 (step S1404), and ends the process.

[0142] As shown in Example 2, the user can recognize whether the arrangement of the baffle plates that he or she has placed in the fluidized bed reactor will result in the desired size and particle size of the bubbles. Furthermore, by back-calculating from the obtained results, the user can estimate the arrangement of the baffle plates that will likely result in the desired state of the bubbles.

[0143] The information about the flow regulator included in each small space received by the learning unit 144 may be the sum of the areas of the obstruction plates included in each small space, or both, in addition to the number of obstruction plates included in each small space. That is, as an example, if the obstruction plates are arranged as shown in Figure 12(a) and the areas of the obstruction plates arranged in the small spaces B1, B2, B4, B5, and B6 are BA1, BA2, BA4, BA5, and BA6, respectively, the teacher data of the learning model when using the sum of the areas of the obstruction plates or the information to be input to the learning model can be expressed as {B1, B2, B3, B4, B5, B6} = {BA1, BA2, 0, BA4, BA5, BA6}. Furthermore, when both are used, if the small space defined by the number of obstruction plates is Ba, and the small space defined by the area of ​​the obstruction plates is Bb, then it can be expressed as {Ba1, Ba2, Ba3, Ba4, Ba5, Ba6, Bb1, Bb2, Bb3, Bb4, Bb5, Bb6} = {1, 1, 0, 1, 1, 1, BA1, BA2, 0, BA4, BA5, BA6}. Note that here, Ba1 and Bb1 are the same space as the small space b1 in Figure 12(a) as a small space, and the same applies to the other small spaces.

[0144] Furthermore, since the supply port through which gas flows into a specified space also affects the diameter and particle size of the bubbles at the measurement point, the relative positional relationship of the gas supply port (indicated by the arrow in the drawing) shown in Figure 12 to the measurement point 210b may be used instead of or in addition to the relative positional relationship of the measurement point to the baffle plate. Regarding the gas supply ports, the number of supply ports contained in the small space based on the direction and distance specified with respect to the measurement point, the amount of gas supplied from the supply ports per unit time, etc. may also be learned or estimated as one-dimensional information indicating the relative positional relationship of the flow regulator (gas supply port) to the measurement point.

[0145] In this embodiment, the size (diameter) and particle size of the bubbles are output as information (information to be evaluated), but this may also be the number of bubbles contained within the desired measurement range. For example, a large number of bubbles indicates that the diameter of the bubbles contained within the desired measurement range is small.

[0146] Example 3 In the third embodiment, an example of estimating the temperature at a specific location in a desired space will be described.

[0147] FIG. 15 is a diagram illustrating an example of a desired space according to Example 3. FIG. 15 is a plan view illustrating an example indoors that includes a heating element 1600 and ventilation openings 220c, 221b, 221c, 221d, and 221e. The heating element 1600 may be, for example, a device that becomes hot through some process or may be a heater. Since some heating elements 1600 are required to maintain a certain temperature range for processing purposes, it is desirable to know in advance the conditions for maintaining the desired temperature range. In the figure, ventilation openings 220c and 221c are openings for taking in outside air, and ventilation openings 222c, 223c, and 224c are ventilation fans for exhausting indoor air to the outside.

[0148] In the third embodiment, an estimation will be described for a case where temperature control is performed by providing a ventilation opening or the like for ventilation in an object such as the heat generating body 1600, which has a variable temperature.

[0149] Fig. 16 is a perspective view schematically showing a desired space according to Example 3, which is a room in which a heating element 1600 is installed. Fig. 16 is also a diagram for explaining an example of defining a plurality of small spaces in Example 3.

[0150] As shown in Fig. 16, a heating element 1600 is provided in a room 200c. Ventilation is performed to prevent the temperature of this heating element 1600 from rising above a predetermined level. To this end, in the example shown in Fig. 16, ventilation openings 220c and 221c are provided in the upper wall of the room 200c, two ventilation openings 222c and 223c are provided in the lower wall, and one ventilation opening 224c is provided in the wall on the right side of the room. Consider the relative positional relationship between a specific region in a predetermined space 200c arranged as shown in Fig. 16 and the flow regulator.

[0151] As an example, as shown in the figure, measurement point 210c on heating element 1600 is set as a specific area, and with measurement point 210c as the reference, directions (small spaces) as seen from measurement point 210c are defined by the dashed lines shown in the figure, and small spaces C1 to C8 are defined. In the figure, an example is shown in which each direction is evenly distributed, but as mentioned above, the angular width of each direction is arbitrary.

[0152] Furthermore, the measurement point may be anywhere where the temperature is desired to be determined, and does not have to be on the heating element 1600, but may be any location within the predetermined space 200.

[0153] In the example of Figure 16, ventilation openings 220c to 224c are involved in the air flow within room 200c, taking in outside air and expelling inside air, and therefore are involved in the temperature fluctuations caused by heating element 1600 and the temperature fluctuations within specified space 200c, and can therefore be defined as flow regulators.

[0154] 16, the relative positional relationship between the ventilation openings 220c to 224c, which are flow regulators, seen from the measurement point 210c can be expressed as one-dimensional information, {C1, C2, C3, C4, C5, C6, C7, C8}={1, 0, 1, 1, 1, 0, 0, 1}, which is information on the number of flow regulators contained in each direction (small space). Also, if the ventilation opening 1402a is an air intake opening (the side that takes in air (in)) and the ventilation opening 1402b is an exhaust opening (the side that exhausts air (out)), their respective positions can be treated as separate information. Therefore, by defining the side that takes in outside air as in and the side that expels inside air as out for the small space to be input into the learning model 151, the information on the number of flow defining bodies contained in the small space can be expressed as one-dimensional information {C1_in, C2_in, C3_in, C4_in, C5_in, C6_in, C7_in, C8_in, C1_out, C2_out, C3_out, C4_out, C5_out, C6_out, C7_out, C8_out} = {1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0}.

[0155] Furthermore, from the viewpoint of contributing to the fluctuation of temperature within the predetermined space 200c, the heat generated by the heating element 1600 or the amount of heat generated also has an influence, and therefore the heating element 1600 can be defined as a flow regulator. Therefore, as the one-dimensional information described above, the relative positional relationship with the heating element as seen from a measurement point other than the ventilation opening may also be used as one-dimensional information for learning or estimation as the relative positional relationship between a specific region and the flow regulator.

[0156] FIG. 17 is a flowchart showing an example of an operation related to generation of a learning model by the information processing device 100 in which the flowchart of FIG. 4 is applied to the third embodiment.

[0157] 17, the learning unit 144 of the control unit 140 of the information processing device 100 receives input of information such as the size, volume, and shape of the room in which the heat generating element 1600 is installed, information on the position of the heat generating element 1600, and placement information on the placement of ventilation openings (step S1701). Information on the position of the heat generating element 1600 may also be received.

[0158] The receiving unit 141 of the control unit 140 receives, via the communication unit 120 or the input unit 130, information on the position of a measurement point where a temperature change is to be measured within a predetermined space (step S1702).

[0159] Based on the received information, the receiving unit 141 also receives information about temperature changes in the room when air is allowed to flow in and / or out through the ventilation openings 220c to 224c under predetermined conditions (step S1703). Here, the information about temperature changes is information about the temperature at a desired measurement point, and may be information about temperature rise or fall per unit time (for example, a desired time interval such as every 5 seconds or every minute).

[0160] The conversion unit 142 converts the arrangement information regarding the arrangement of the ventilation openings received in step S1701 into one-dimensional information indicating the relative positional relationship with respect to the measurement point received in step S1702 (step S1704). That is, one-dimensional information is generated that defines the number of ventilation openings included in small spaces that specify the direction and distance from the specified measurement point, using the specified measurement point as a reference. In the example of Fig. 16, information indicating the number of ventilation openings included in each of the small spaces C1 to C8 is generated as one-dimensional information indicating the relative positional relationship of the flow defining body with respect to a specific region.

[0161] The learning unit 144 learns the relationship between information relating to the predetermined space, information indicating the positions of the measurement points within the predetermined space, information indicating the relative positional relationship between the measurement points and the flow regulator generated by conversion by the conversion unit 142, and information associating the information on temperature changes at the measurement points as information relating to the flow at the measurement points received in step S1703 (step S1705). The information relating to the predetermined space here is information relating to room 200c, and may include the volume and shape of the room, as well as the heat generation amount of heating element 1600.

[0162] That is, the learning unit 144 learns multiple relationships between the number of ventilation openings (and / or the number or area or volume of heating elements) contained in the small space that defines the directions and distances based on the measurement points as information about the specified space, information indicating the position of the measurement points within the specified space, and information indicating the relative positional relationship of the flow regulator with respect to a specific area, and information about temperature changes in the room, and generates a learning model.

[0163] The learning unit 144 stores the generated learning model 151 in the storage unit 150 (step S1706), and ends the process.

[0164] FIG. 18 is a flowchart showing an example of an estimation operation performed by the information processing device 100 to obtain information about temperature changes at a desired location using the learning model 151 stored in the storage unit 150.

[0165] 18, the receiving unit 141 of the control unit 140 receives, as information on the flow defining body, information on the room 200c as a predetermined space and placement information for placing ventilation openings in the room (step S1801). The receiving unit 141 receives, for example, the placement information indicating the placement of ventilation openings input by the user to the input unit 130.

[0166] Furthermore, the receiving unit 141 receives, via the communication unit 120 or the input unit 130, input of position information of a measurement point, which is a location in the room 200c where the user wants to know the temperature change (step S1802).

[0167] The conversion unit 142 converts the placement information for arranging the ventilation openings received in step S1801 into one-dimensional information indicating the relative positional relationship of the ventilation openings with respect to the measurement point received in step S1802 (step S1803). That is, the conversion unit 142 generates one-dimensional information that defines multiple small spaces that define each direction and distance as seen from the measurement point, using the measurement point as a reference, and defines the number of ventilation openings included in each small space. Note that the definitions of the multiple small spaces at this time are the same as the multiple small spaces used when training the learning model 151.

[0168] The estimation unit 143 inputs information about the predetermined space, information indicating the positions of the measurement points within the predetermined space, and one-dimensional information indicating the relative positional relationship between the measurement points and the ventilation openings into the learning model 151 read from the storage unit 150 (step S1803). By inputting the one-dimensional information into the learning model 151, information about temperature changes at a desired location within the room can be obtained.

[0169] The estimation unit 143 instructs the output unit 160 to output information about the temperature change obtained from the learning model 151 (step S1804), and ends the process. The information about the temperature change may be the final room temperature or the amount of temperature change.

[0170] As shown in Example 3, the user can estimate whether the arrangement of the ventilation openings that the user has placed in the room will result in a desired temperature at a certain location in the room, for example, at heating element 1600. Furthermore, by working backward from the obtained results, the user can estimate the arrangement of the ventilation openings that will likely result in a desired temperature in the room.

[0171] The information about the flow regulator included in each small space received by the learning unit 144 may be the number of ventilation openings included in each small space, the sum of the opening areas of the ventilation openings included in each small space, or both. Alternatively, the amount of gas discharged from each ventilation opening per unit time may be used instead of the area, or the number of ventilation openings and the sum of the areas may be combined. For example, if the opening areas of ventilation openings 220c to 224c in FIG. 16 (FIG. 17) are CA0, CA1, CA2, CA3, and CA4, respectively, the one-dimensional information used when the sum of the opening areas is used may be {C1, C2, C3, C4, C5, C6, C7, C8} = {CA1, 0, CA4, CA3, CA2, 0, 0, CA0}. Furthermore, if the small spaces C1, C2, C3, C4, C5, C6, C7, and C8 in Figure 16 (Figure 17) are defined as Ca1, Ca2, Ca3, Ca4, Ca5, Ca6, Ca7, and Ca8 when specifying the number of ventilation openings corresponding to them, and the small spaces Cb1, Cb2, Cb3, Cb4, Cb5, Cb6, Cb7, and Cb8 when specifying the sum of the opening areas, the one-dimensional information can be expressed as {Ca1, Ca2, Ca3, Ca4, Ca5, Ca6, Ca7, Ca8, Cb1, Cb2, Cb3, Cb4, Cb5, Cb6, Cb7, Cb8} = {1, 0, 1, 1, 1, 0, 0, 1, CA1, 0, CA4, CA3, CA2, 0, 0, CA0}.

[0172] Furthermore, in addition to or separately from the arrangement of the ventilation openings, the relative positional relationship between the measurement point 210c and the heating element 1600c may be used as one-dimensional information. For example, in the example shown in FIG. 16, if the heating element 1600 is simply assigned a value of 1 if it is included in a small space, the one-dimensional information indicating the relative positional relationship between the measurement point and the heating element serving as the flow regulator can be expressed as {C1, C2, C3, C4, C5, C6, C7, C8} = {1, 1, 1, 1, 1, 1, 1, 1}. Alternatively, instead of indicating whether the heating element is included in a small space, the area (volume) of the heating element included in each small space or the ratio of the heating element to the total area (volume) of the heating element may be used. Furthermore, the one-dimensional information may be integrated with the number of ventilation openings included in the small space.

[0173] In FIG. 17, the temperature change at measurement point 210c on heating element 1600 is estimated. However, as described above, this may be the surface temperature of heating element 1600, the internal temperature, or even the temperature of a specific location on heating element 1600. Alternatively, the temperature of a specific location in a room may be estimated. That is, the presence of heating element 1600 may cause the room to become hot, potentially causing inconvenience to people passing through. Therefore, the temperature change at a point along the path of people passing through the room may be estimated. Furthermore, while Example 3 illustrates the case where air cooling is performed, temperature control may be performed by filling room 200c with water or the like and allowing the water to enter and exit through ventilation openings 220c to 224c. Furthermore, a cooling element that lowers temperature may be used instead of heating element 1600. In this case, ventilation may be performed to prevent the temperature in a specified space from dropping too low.

[0174] Example 4 In this Example 4, an example of temperature estimation different from that in Example 3 will be described. In the above Example 3, examples of temperature regulation (cooling) by a ventilation port and temperature regulation (heating) by a heating element have been described, but in this Example 4, estimation of temperature change in the case of direct cooling by a cooling tube will be described.

[0175] FIG. 19 is a plan view schematically illustrating a cooling device as a predetermined space 200d. As shown in FIG. 19, cooling tubes 222d are arranged and filled with a coolant to maintain the temperature inside the cooling device (predetermined space 200d) below a predetermined temperature. As a result, the inside of the cooling device (predetermined space 200d) is cooled. At this time, the manner in which the cooling tubes 222d are arranged varies within the cooling device 200d (cooling may be faster or slower depending on the location). Therefore, the cooling tubes 222d can be defined as a flow regulator that regulates the flow of temperature inside the cooling device 200d.

[0176] Therefore, in this fourth embodiment, the interior of the cooling device 200d is set as a predetermined space, and it is estimated how cooling is achieved at a desired measurement point (for example, measurement point 210d) depending on how the cooling tube 222d is arranged.

[0177] 19 shows an example of defining small spaces for defining the relative positional relationship between the measurement point and the flow regulator in Example 4. FIG. 19 shows an example of defining small spaces for defining the relative positional relationship in four directions and two distance ranges with measurement point 210d as the reference when viewing the cooling device (predetermined space 200d) in a plan view, and shows an example of defining small spaces D1 to D8. When the small spaces are defined as in FIG. 19, in the example of FIG. 19, one cooling tube 222d (the cooling tube on the right side of the drawing) is arranged across small spaces D5, D2, D6, and D7, and one cooling tube (the cooling tube on the left side of the drawing) is arranged across small spaces D5 and D8. Therefore, in this case, the information about the flow regulator input to the learning model 151 during estimation can be expressed as one-dimensional information, such as {D1, D2, D3, D4, D5, D6, D7, D8} = {1, 0, 1, 1, 2, 1, 1, 1}, based on whether or not a cooling tube is present in the small space, and can be used as input. As described above, each small space D1-D8 defines the direction and distance based on measurement point 210d, and determines the relative positional relationship between the measurement point and the cooling tube serving as the flow regulator. Therefore, the one-dimensional information can be said to be information indicating the relative positional relationship between measurement point 210d and cooling tube 222d. Note that, here, a cooling tube contained in one small space is counted as one, but the proportion of the cooling tube in the small space or a numerical value representing the area or volume of the cooling tube within the small space may also be used as the number of flow regulators.

[0178] FIG. 20 is a flowchart showing an example of an operation related to generation of a learning model by the information processing device 100 in which the flowchart of FIG. 4 is applied to the fourth embodiment.

[0179] As shown in FIG. 20, the receiving unit 141 of the control unit 140 of the information processing device 100 receives input of information such as the size, volume, and shape of the cooling device, and arrangement information regarding the arrangement of the cooling tubes (step S2001).

[0180] Furthermore, the receiving unit 141 receives, via the communication unit 120 or the input unit 130, input of position information of a measurement point indicating a position where the user wants to measure a temperature change (step S2002).

[0181] Furthermore, the receiving unit 141 receives information about a temperature change at a measurement point in the cooling device when the cooling tube is filled with a coolant under predetermined conditions (step S2003). Here, the information about the temperature change may be information about the degree of temperature increase or decrease at a desired measurement position, the temperature after the change, etc.

[0182] The conversion unit 142 converts the arrangement information indicating the arrangement of the cooling tubes received in step S2001 into one-dimensional information indicating the relative positional relationship of the cooling tubes, which are flow regulators, with respect to the measurement points received in step S2002 (step S2004). That is, the conversion unit 142 defines small spaces in any direction and at any distance, as shown in FIG. 19 as an example, based on the measurement points received in step S2002. Then, the conversion unit 142 counts the number of cooling tubes (or their area or volume) included in each defined small space to generate one-dimensional information. The term "small space" here refers to determining the range of a small space in a cooling device as data. In the example of FIG. 19, it refers to small spaces D1 to D8 obtained by dividing the top surface of the cooling device 200d by dotted lines based on the measurement point 210d. The definition of the small spaces is not limited to the example of FIG. 19, and the small spaces may be defined so as to be divided in any direction and at any distance.

[0183] The learning unit 144 learns the relationship between each piece of information, using the relative positional relationship between the measurement point and the cooling tube as information on the relative positional relationship with the flow regulator for a specific region, and information on temperature changes at the measurement point in the cooling device as information on the flow within a specified space (step S2005). The learning unit 144 learns multiple relationships between information on the cooling device, information on the position of the measurement point in the cooling device, and information indicating the relative positional relationship with the cooling tube as the flow regulator based on the measurement point in the cooling device, and information on temperature changes at the measurement point, and generates a learning model.

[0184] The learning unit 144 stores the generated learning model 151 in the storage unit 150 (step S2006), and ends the process.

[0185] FIG. 21 is a flowchart showing an example of an estimation operation by the information processing device 100 for obtaining information about temperature changes at measurement points in the cooling device (predetermined space 200d) using the learning model 151 stored in the storage unit 150.

[0186] 21, the receiving unit 141 of the control unit 140 receives, as information on the flow regulating body, placement information indicating positions at which cooling tubes are to be placed relative to the cooling device (step S2101). The receiving unit 141 receives, for example, placement information indicating the placement of the cooling tubes input by the user to the input unit 130.

[0187] Furthermore, the receiving unit 141 receives position information indicating the positions of the measurement points in the cooling device as information on the specific area (step S2102). The receiving unit 141 receives, for example, position information indicating the positions of the measurement points input by the user to the input unit 130.

[0188] The conversion unit 142 converts the cooling tube arrangement information received in step S2101 into one-dimensional information indicating the relative positional relationship of the cooling tube with respect to the measurement point, based on the positional information indicating the position of the measurement point received in step S2102 (step S2103). That is, the conversion unit 142 defines multiple directions and distances around the measurement point, using the position of the measurement point received in step S2102 as a reference, and defines multiple small spaces. Then, the conversion unit 142 counts the number of cooling tubes included in each small space (if included, it is set to 1; if not, it is set to 0). If multiple different cooling tubes are included in the same small space, the conversion unit 142 counts the number of different cooling tubes in the small space and generates the small space and the number of cooling tubes included in the small space as one-dimensional information indicating the relative positional relationship of the cooling tube with respect to the measurement point. Note that the definition of the multiple small spaces at this time is the same as the definition of the multiple small spaces used when training the learning model 151.

[0189] The estimation unit 143 inputs the one-dimensional information indicating the relative positional relationship between the measurement point and the cooling tube generated by the conversion unit 142 into the learning model 151 read from the storage unit 150 (step S2104). By inputting the one-dimensional information into the learning model 151, information regarding the temperature change at a desired location (measurement point) in the cooling device can be obtained.

[0190] The estimation unit 143 instructs the output unit 160 to output information about the temperature change obtained from the learning model 151 (step S2105), and ends the process.

[0191] As shown in Example 4, the user can recognize whether the cooling tube arrangement he or she has arranged in the cooling device will result in the desired temperature inside the cooling device, or whether the desired temperature change will occur. Furthermore, by working backwards from the obtained results, the user can estimate the cooling tube arrangement that is likely to result in the desired temperature state.

[0192] The information about the flow regulator included in each small space, which is received by the learning unit 144, may be the number of cooling tubes included in each small space, or the sum of the heat removal amounts per unit time of the cooling tubes included in each small space, or both. For example, if the heat removal amount (e.g., the temperature reduction or the amount of heat change) of the small space cooled within a unit time by the cooling tube on the right side of Fig. 19 is defined as DA1, and the heat removal amount by the cooling tube on the left side is defined as DA2, the one-dimensional information can be expressed as {D1, D2, D3, D4, D5, D6, D7, D8} = {0, 0, DA1, 0, DA1 + DA2, DA1, DA1, DA2}. Furthermore, when using both the number of cooling tubes and the amount of heat removal, if the small spaces corresponding to the number of cooling tubes are defined as Da1, Da2, Da3, Da4, Da5, Da6, Da7, and Da8, and the small spaces corresponding to the amount of heat removed by the cooling tubes are defined as Db1, Db2, Db3, Db4, Db5, Db6, Db7, and Db8, the one-dimensional information can be expressed as {Da1, Da2, Da3, Da4, Da5, Da6, Da7, Da8, Db1, Db2, Db3, Db4, Db5, Db6, Db7, Db8} = {0, 0, 1, 0, 2, 1, 1, 1, 0, 0, DA1, 0, DA1+DA2, DA1, DA1, DA2}.

[0193] <Example 5> In this Example 5, similar to Example 1, an example of estimating a reaction in a container will be described, but an example different from Example 1 will be described.

[0194] FIG. 22 is a schematic perspective view showing an example in which a reactor is used as a predetermined space 200e in which a gas and a substance are reacted. The reactor may be a deposition reactor. For example, a gas is introduced into the reactor 200e through a plurality of gas inlet holes 221e provided on the upper surface 2201 of the reactor. The gas is then reacted with the target substance. The gas in the reactor 200e is then appropriately discharged through outlet holes 222e provided on the lower surface 2105. In this embodiment, the gas inlet holes 221e and outlet holes 222e determine the flow within the reactor 200e, i.e., the gas flow or the reaction in the target substance, and therefore can be defined as a flow regulator.

[0195] FIG. 23 is a diagram illustrating an example of defining the relative positional relationship between a specific region and a flow regulator in Example 5. In FIG. 23(a), directions (small spaces) E1 to E5 are defined as indicated by dotted lines to identify the relative positional relationship between a measurement point 210e on the object and gas inlet hole 221e and outlet hole 222e, which are flow regulators. FIG. 23(b) is a front view of a predetermined space 200e in FIG. 23(a) viewed from the surface. In this case, as shown in FIG. 23(a), there are two gas inlet holes in direction (small space) E1, two gas outlet holes in direction (small space) E3, and five gas inlet holes in direction (small space) E5. Therefore, in this case, information about the flow regulator input to the learning model 151 during estimation can be expressed as one-dimensional information, such as {E1, E2, E3, E4, E5}={2, 0, 2, 0, 5}, and can be used as input.

[0196] Furthermore, since gas inflow and outflow express mutually opposing properties, if the small space when specifying gas inflow is defined as Ein and the small space when specifying gas outflow is defined as Eout, the one-dimensional information may be expressed as {E1in, E2in, E3in, E4in, E5in, E1out, E2out, E3out, E4out, E5out} = {2, 0, 0, 0, 5, 0, 0, 2, 0, 0}.

[0197] FIG. 24 is a flowchart showing an example of an operation related to generation of a learning model by the information processing device 100 in which the flowchart of FIG. 4 is applied to the fifth embodiment.

[0198] As shown in FIG. 24, the learning unit 144 of the control unit 140 of the information processing device 100 receives input of information such as the size, volume, and shape of the reaction vessel, and arrangement information regarding the arrangement of gas inlet holes (exhaust holes) (step S2401).

[0199] Furthermore, the learning unit 144 receives input of position information indicating the position of the measurement point on the object in the reaction vessel (step S2402).

[0200] Based on the received information, the learning unit 144 receives information about the reaction in the target object at the measurement point in the reaction vessel when gas is allowed to flow in through the gas inlet under predetermined conditions (step S2403). Here, the information about the reaction may be information about whether the desired reaction is occurring in the target object, whether the reaction is occurring evenly, etc. Furthermore, if the reaction is a precipitation reaction, the information may be information about the amount (thickness) of the precipitation. Furthermore, if the reaction generates (or consumes) a specific substance, the information may be information about the concentration of the specific substance within a predetermined range of space (measurement point).

[0201] The conversion unit 142 converts the arrangement information regarding the arrangement of gas inlet holes (exhaust holes) received in step S2401 into one-dimensional information indicating the relative positional relationship of the gas inlet holes with respect to the measurement point received in step S2402 (step S2404). That is, the learning unit 144 defines multiple small spaces according to multiple directions and distances from the measurement point of the reaction vessel as a reference. Then, the number of gas inlet holes (exhaust holes) included in each small space is counted to generate one-dimensional information. The definition of a small space here refers to determining the range of the small space in the reaction vessel as data, and in the example of FIG. 23, refers to each area obtained by dividing a predetermined space 200e by planes indicated by dotted lines.

[0202] The learning unit 144 learns the relative positional relationship between the measurement point and the gas inlet (exhaust hole), using the relative positional relationship information with the flow regulator for a specific region as relative positional relationship information with the flow regulator, and the reaction information at the measurement point of the object as information about the flow (step S2405). The learning unit 144 learns multiple relationships between the information indicating the relative positional relationship between the measurement point and the gas inlet (exhaust hole), information about the position of the measurement point in the reaction vessel, and information about the reaction in the reaction vessel, and generates a learning model.

[0203] The learning unit 144 stores the generated learning model 151 in the storage unit 150 (step S2406), and ends the process.

[0204] FIG. 25 is a flowchart showing an example of an estimation operation performed by the information processing device 100 to obtain information about a reaction in a target object (measurement point) using the learning model 151 stored in the storage unit 150.

[0205] 25, the receiving unit 141 of the control unit 140 receives arrangement information for arranging gas inlet holes (discharge holes) in the reaction vessel as information of the flow regulator (step S2501). The receiving unit 141 receives, for example, arrangement information indicating the arrangement of the gas inlet holes (discharge holes) input by the user to the input unit 130.

[0206] Furthermore, the receiving unit 141 receives position information relating to the position of the measurement point on the object in the reaction vessel as information on the specific region.

[0207] The conversion unit 142 converts the arrangement information of the gas inlet holes (exhaust holes) received in step S2501 into one-dimensional information indicating the relative positional relationship of the gas inlet holes (exhaust holes) with respect to the measurement point received in step S2502 (step S2503). That is, the conversion unit 142 defines multiple directions and distances based on the measurement point and defines multiple small spaces for defining the relative positional relationship with the measurement point. Then, the conversion unit 142 counts the number of gas inlet holes (exhaust holes) included in each of the multiple defined small spaces. The conversion unit 142 generates one-dimensional information that specifies identification information indicating each small space and the number of gas inlet holes (exhaust holes) included in the corresponding small space as one-dimensional information indicating the relative positional relationship between the measurement point and the gas inlet holes (exhaust holes). Note that the definitions of the multiple small spaces at this time are the same as the multiple small spaces used when training the learning model 151.

[0208] The estimation unit 143 inputs the one-dimensional information indicating the relative positional relationship between the measurement point and the gas inlet (exhaust) generated by the conversion unit 142 into the learning model 151 read from the storage unit 150 (step S2503). By inputting the one-dimensional information into the learning model 151, information regarding the reaction of the target at a desired location in the reaction vessel can be obtained.

[0209] The estimation unit 143 instructs the output unit 160 to output information about the reaction in the object obtained from the learning model 151 (step S2504), and ends the process.

[0210] As shown in Example 5, the user can recognize whether the arrangement of the gas inlet (exhaust) holes that the user has arranged in the reaction vessel will result in a desired reaction. Furthermore, by working backward from the obtained results, the user can estimate the arrangement of the gas inlet (exhaust) holes that will likely result in a desired reaction in the target substance.

[0211] The information about the flow regulator included in each small space received by the learning unit 144 may be the number of gas inlets (outlets) included in each small space, the sum of the amounts of gas flowing in (outlets) from the gas inlets (outlets) included in each small space, or both. Here, "both" may mean specifying both the amount of gas inlet and the amount of gas outlet, or specifying both the number of gas inlets (outlets) and the amount of gas inlet (outlet), or all of these. That is, instead of the number of gas inlets (outlets) shown in Example 5, the sum of the amount of gas flowing in per unit time (or a predetermined length of time) from the gas inlets (outlets) included in each small space may be used. As an example, referring to FIG. 23, let us assume that the amount of gas flowing in per unit time through the gas inlet holes in small space E1 is E1 liters and E2 liters, and the amount of gas flowing in per unit time through the gas inlet holes in small space E5 is E3 liters, E4 liters, E5 liters, E6 liters, and E7 liters. In this case, the one-dimensional information input to the learning model is {E1, E2, E3, E4, E5} = {E1 + E2, 0, 0, 0, E3 + E4 + E5 + E6 + E7}. In this way, the learning model may be defined and estimation may be performed based on the gas inflow (exhaust) amount rather than the number of gas inlet holes (exhaust holes). Defining the learning model and the one-dimensional information to be input based on the number of gas inlet holes (exhaust holes) is effective when the amount of gas flowing in (exhaust) from each gas inlet hole (exhaust hole) is constant for each hole. In addition, defining the learning model and input one-dimensional information based on the amount of gas flowing in (out) is effective when the amount of gas flowing in (out) from each gas inlet (outlet) is not constant.

[0212] The measurement point in the above embodiment does not have to be a single point, but may be a predetermined range within which the user wants to know the state of a predetermined space.

[0213] The program of each embodiment of the present disclosure may be provided in a state stored in a computer-readable storage medium. The storage medium can store the program in a "non-transitory tangible medium." The storage medium can include any suitable storage medium, such as an HDD or SSD, or a suitable combination of two or more of these. The storage medium may be volatile, non-volatile, or a combination of volatile and non-volatile. The storage medium is not limited to these examples and may be any device or medium capable of storing the program.

[0214] The information processing device 100 can realize the functions of the multiple functional units shown in each embodiment by, for example, reading a program stored in a storage medium and executing the read program. The program may also be provided to the information processing device 100 via any transmission medium (such as a communication network or broadcast waves). The information processing device 100 realizes the functions of the multiple functional units shown in each embodiment by, for example, executing a program downloaded via the Internet or the like.

[0215] The program can be implemented using, for example, a scripting language such as ActionScript or JavaScript (registered trademark), an object-oriented programming language such as Objective-C or Java (registered trademark), or a markup language such as HTML5.

[0216] At least a part of the processing in the information processing device 100 may be realized by cloud computing consisting of one or more computers. Furthermore, each functional unit of the information processing device 100 may be realized by one or more circuits that realize the functions described in the above embodiments, and the functions of multiple functional units may be realized by one circuit.

[0217] Furthermore, although the embodiments of the present disclosure have been described based on various drawings and examples, it should be noted that those skilled in the art would easily be able to make various modifications and alterations based on the present disclosure. Therefore, it should be noted that these modifications and alterations are within the scope of the present disclosure. For example, the functions included in each means, step, etc. may be rearranged so as not to be logically contradictory, and multiple means, steps, etc. may be combined into one or divided into separate units. Furthermore, the configurations shown in each embodiment may be combined as appropriate. [Explanation of symbols]

[0218] 100 Information processing device 110 connecting wire 120 Communications Department 130 Input section 140 Control Unit 141 Reception Department 142 Conversion Unit 143 Estimation Department 144 Learning Department 150 Storage section 160 Output section

Claims

1. An information processing method by an information processing device that inputs and outputs information about a flow regulator that regulates a flow in a predetermined space, the method comprising: a storage step of storing a learning model that has learned the relationship between information representing a state of the predetermined space, information relating to a specific region in the predetermined space, one-dimensional information relating to a relative positional relationship between the specific region and the flow regulator with respect to the specific region, and information relating to the flow in the specific region in that case; a receiving step of receiving input of information representing a state of the predetermined space, information regarding the specific region, and information regarding an arrangement position of the flow regulator within the predetermined space; a conversion step of converting information about the specific region and information about the arrangement position of the flow regulator in the predetermined space into one-dimensional information about a relative positional relationship between the specific region and the flow regulator; an estimation step of inputting the information representing the state of the predetermined space and the information relating to the specific region received in the receiving step, and the one-dimensional information converted in the converting step, into the learning model, and estimating a flow in the specific region within the predetermined space when the flow regulator is placed at an arrangement position indicated by the information relating to the arrangement position of the flow regulator within the predetermined space; an output step of outputting information about the flow within the predetermined space based on the estimation by the estimation step; An information processing method including:

2. the learning model is a model that has learned information that defines the relative positional relationship between the flow regulator and a specific region, using the specific region as a reference, as one-dimensional information regarding the relative positional relationship between the specific region in the predetermined space and the flow regulator with respect to the specific region, The conversion step converts the specific region into one-dimensional information regarding the relative positional relationship between the specific region and the flow regulator, with the specific region as a reference.

2. The information processing method according to claim 1,

3. the learning model is a model that defines a plurality of small spaces with the specific area as a reference, as one-dimensional information relating to a relative positional relationship between a specific area in the predetermined space and the flow governing body with respect to the specific area, and learns the number of the flow governing bodies included in the plurality of small spaces, The conversion step defines a plurality of small spaces based on the specific area as one-dimensional information relating to the relative positional relationship between the specific area in the predetermined space and the flow regulator with respect to the specific area, and converts the information into one-dimensional information indicating the number of the flow regulators included in each of the plurality of small spaces.

3. The information processing method according to claim 1 or 2.

4. 4. The information processing method according to claim 1, wherein the information about the specific area includes information about the absolute position of the specific area in a predetermined space.

5. the predetermined space is a space within a vessel of a precipitation reactor that precipitates a compound based on a raw material on a surface of a core material, the vessel having a jetting hole for jetting the raw material; the specific region is at least a part of the surface of the core material, The flow regulator is a jet hole for jetting the raw material.

5. The information processing method according to claim 1, wherein the first and second information processing units are connected to each other.

6. 6. The information processing method according to claim 5, wherein the information about the flow in the predetermined space is information about a heat flux occurring around the core material.

7. 6. The information processing method according to claim 5, wherein the information about the flow in the predetermined space is information about the thickness of the compound precipitated on the core material.

8. the predetermined space is a space within a vessel of a precipitation reactor that precipitates a compound based on a raw material on a surface of a core material, the container has an ejection hole for ejecting the raw material and an exhaust hole for exhausting gas within the predetermined space, the specific region is at least a part of the surface of the core material, The flow regulator is the discharge hole.

5. The information processing method according to claim 1, wherein the first and second information processing units are connected to each other.

9. the predetermined space is a space within a container filled with powder particles, By supplying gas into the container, the powder and granular material become fluidized. The flow regulator is at least one of an obstruction plate provided in a position to obstruct the flow in a predetermined direction within the predetermined space and a supply port of a nozzle that supplies the gas, The bubbles contained in the predetermined space are broken up by contact with the baffle plate and become smaller in diameter, The information about the flow in the predetermined space is information about at least one of the diameter and number of bubbles contained in the specific area.

5. The information processing method according to claim 1, wherein the first and second information processing units are connected to each other.

10. the predetermined space is a space within a container filled with powder particles as the specific region, It has a measurement point that measures the temperature within the space, A temperature control means is disposed in the powder or granular material, and gas is supplied into the container to cause the powder or granular material to flow. the flow regulator is at least one of the temperature control means for adjusting a temperature and a supply port of a nozzle for supplying the gas, The information about the flow in the predetermined space is the temperature in the specific area.

5. The information processing method according to claim 1, wherein the first and second information processing units are connected to each other.

11. the predetermined space has, as the specific region, a measurement region for measuring at least one of a temperature and a concentration of a substance within the space; A heating / cooling means or a substance generating / reducing means may be disposed in the space; The gas flows, the flow regulator is a vent or an opening provided in the predetermined space, When the heating / cooling means is disposed, the flow regulator is at least one of the vent or opening and the heating / cooling means; When the substance generating / reducing means is disposed, the flow regulator is at least one of the vent or opening and the substance generating / reducing means, The information about the flow in the predetermined space is information about at least one of the temperature and the concentration of a substance in the specific region.

5. The information processing method according to claim 1, wherein the first and second information processing units are connected to each other.

12. An information processing device that inputs and outputs information about a flow regulator that regulates a flow in a predetermined space, the information processing device comprising: a storage unit that stores a learning model that has learned the relationship between information representing a state of the predetermined space, one-dimensional information regarding a relative positional relationship between a specific region in the predetermined space and the flow governing body with respect to the specific region, and information regarding the flow in the specific region in that case; a receiving unit that receives input of information representing a state of the predetermined space, information about the specific region, and information about an arrangement position of the flow regulator within the predetermined space; a conversion unit that converts information about the specific region and information about the arrangement position of the flow regulator in the predetermined space into one-dimensional information about a relative positional relationship between the specific region and the flow regulator; an estimation unit that inputs the information representing the state of the predetermined space and the information related to a specific region received by the reception unit, and the one-dimensional information converted by the conversion unit, into the learning model, and estimates a flow in the specific region within the predetermined space when the flow regulator is placed at an arrangement position indicated by the information related to the arrangement position of the flow regulator within the predetermined space; an output unit that outputs information about the flow within the predetermined space based on the estimation by the estimation unit; An information processing device comprising:

13. a computer that inputs and outputs information about a flow regulator that regulates a flow in a predetermined space, the computer having access to a storage unit that stores a learning model that has learned the relationship between information that represents the state of the predetermined space, one-dimensional information about a relative positional relationship between a specific region in the predetermined space and the flow regulator with respect to the specific region, and information about the flow in the specific region; a receiving function for receiving input of information representing a state of the predetermined space, information relating to the specific region, and information relating to an arrangement position of the flow regulator within the predetermined space; a conversion function for converting information about the specific region and information about the arrangement position of the flow regulator in the predetermined space into one-dimensional information about the relative positional relationship between the specific region and the flow regulator; an estimation function that inputs the information representing the state of the predetermined space and the information relating to a specific region received by the reception function, and the one-dimensional information converted by the conversion function, into the learning model, and estimates the flow in the specific region within the predetermined space when the flow regulator is placed at an arrangement position indicated by the information relating to the arrangement position of the flow regulator within the predetermined space; an output function that outputs information about the flow within the predetermined space based on the estimation by the estimation function; An information processing program that makes this possible.

Images