Porous body design methods and porous body manufacturing methods

Abstract

A method for designing a porous body using a computer, wherein the computer is made to execute, a plurality of times each, feature prediction / calculation processing to predict or calculate a feature of the porous body, which has a prescribed three-dimensional structure, evaluation processing to evaluate the feature of the porous body that was predicted or calculated through the feature prediction / calculation processing, and optimization processing to change the connectivity between pores in the three-dimensional structure and search for the optimal connectivity, and the three-dimensional structure of the porous body is determined on the basis of the results of the evaluation of the feature of the porous body from the evaluation processing.

Description

[Technical Field]

[0001] The present invention relates to a method for designing and manufacturing a porous body. [Background technology]

[0002] Conventionally, porous materials, which have a large number of pores arranged three-dimensionally within the material, have been widely used in a variety of applications, such as purifying exhaust gases emitted from internal combustion engines such as gasoline and diesel engines. When designing such porous materials, it is necessary to determine the three-dimensional structure of the porous material so that appropriate properties can be obtained for each application.

[0003] Conventional porous body design methods include, for example, creating several prototypes of porous bodies based on the designer's experience and intuition, measuring the three-dimensional structures and properties of each of these prototypes, and then determining the final three-dimensional structure of the porous body. However, this design method requires a lot of time and effort to create the prototypes and measure their properties, which limits the scope of the relationship between the three-dimensional structure and properties that can be explored. As a result, there is a problem in that it is not always possible to determine a three-dimensional structure of a porous body that will achieve the desired properties.

[0004] Meanwhile, in recent years, methods using simulation software have also become known, as described in, for example, Patent Document 1. With such simulation software, a three-dimensional model of a porous body is created, and the properties of the porous body can be calculated from this three-dimensional model. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent No. 6940786 Summary of the Invention [Problem to be solved by the invention]

[0006] In the method described in Patent Document 1, a sample is created using a partition piece cut out from an exhaust gas purification filter, and the sample is photographed using an X-ray CT scanner to obtain continuous tomographic images, which are then loaded into simulation software to create a three-dimensional model, thereby deriving the number of communicating holes in the porous exhaust gas purification filter. Therefore, the exhaust gas purification filter needs to be created in advance, which still poses the problem of the time and effort required for prototyping.

[0007] In view of the above problems, an object of the present invention is to provide a useful technique that enables an optimum structure to be determined early in the design of a porous body. [Means for solving the problem]

[0008] The porous body design method according to the present invention is a method for designing a porous body using a computer, which causes the computer to execute a property prediction / calculation process for predicting or calculating the properties of the porous body having a predetermined three-dimensional structure, an evaluation process for evaluating the properties of the porous body predicted or calculated by the property prediction / calculation process, and an optimization process for changing the connectivity between pores in the three-dimensional structure to search for the optimal connectivity, multiple times each, and determines the three-dimensional structure of the porous body based on the evaluation results of the properties of the porous body obtained by the evaluation process. The porous body manufacturing method according to the present invention manufactures the porous body using three-dimensional shape data based on the three-dimensional structure of the porous body determined by the porous body design method. [Effects of the Invention]

[0009] According to the present invention, a useful technique can be provided that enables an optimum structure to be determined early in the design of a porous body. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a block diagram showing the configuration of a porous body design apparatus according to an embodiment of the present invention. [Figure 2] 1 is a flowchart showing the processing flow of the porous body design apparatus according to the first embodiment of the present invention. [Figure 3]4 is a flowchart showing details of an initial learning model generation process according to the first embodiment of the present invention. [Figure 4] FIG. 1 shows an example of a PN model. [Figure 5] 1 is a flowchart showing the flow of a manufacturing process for a porous body using a 3D printer. [Figure 6] 10 is a flowchart showing the processing flow of a porous body design apparatus according to a second embodiment of the present invention. [Figure 7] 10 is a flowchart showing details of an initial learning model generation process according to a second embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the following embodiment, a porous body design device will be described that can be operated by a designer who designs a porous body, and that can determine the three-dimensional structure of a porous body having desired properties in a relatively short time based on the existing three-dimensional structure of a porous body.

[0012] (First embodiment) Fig. 1 is a block diagram showing the configuration of a porous body design apparatus according to one embodiment of the present invention. The porous body design apparatus 100 shown in Fig. 1 is a computer equipped with a control unit 1, a storage unit 2, a memory 3, an operation input device 4, and a display device 5, and is configured by connecting these devices to each other via a bus 6.

[0013] The control unit 1 is configured using, for example, a CPU (Central Processing Unit) and performs various processes and calculations to operate the porous body design apparatus 100. The control unit 1 executes programs stored in the memory unit 2 to realize the following functional blocks: a learning model creation unit 11, a PN model conversion unit 12, a physical simulation unit 13, an AI calculation unit 14, a property evaluation unit 15, a gene evolution unit 16, a learning model update unit 17, and a data conversion unit 18. Details of these functional blocks will be described later. Note that some or all of the functions of the control unit 1 may be realized using devices other than a CPU, such as a GPU (Graphic Processing Unit), FPGA (Field Programmable Gate Array), or ASIC (Application Specific Integrated Circuit).

[0014] The storage unit 2 is configured using a large-capacity, non-volatile storage device, such as a magnetic storage device such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive), and stores programs executed by the control unit 1 and various information used in the processing of the control unit 1. The information stored in the storage unit 2 includes porous body structure data 21, PN model data 22, learning model data 23, input parameter data 24, property data 25, property evaluation data 26, and three-dimensional shape data 27. Details of these data will be described later.

[0015] The memory 3 is configured using a high-speed and volatile storage device such as a DRAM (Dynamic Random Access Memory), and is used as a work area when the control unit 1 executes a program.

[0016] The operation input device 4 is a device for detecting operation input by the user, and is configured using, for example, a keyboard or a mouse. The display device 5 is a device for displaying the processing results of the porous body design device 100 on a screen to present them to the user, and is configured using, for example, a liquid crystal display or an organic EL display. Note that other computers, smartphones, etc. that can communicate with the porous body design device 100 may also be used as the operation input device 4 or the display device 5.

[0017] 1 may be physically constructed on a single computer, or may be distributed across multiple computers. The porous body design system 100 may also be realized by a cloud computer installed on a cloud, a virtual machine operating in a virtual environment, or the like.

[0018] Next, we will explain each of the functional blocks in the control unit 1: the learning model creation unit 11, the PN model conversion unit 12, the physical simulation unit 13, the AI ​​calculation unit 14, the property evaluation unit 15, the gene evolution unit 16, the learning model update unit 17, and the data conversion unit 18. In the porous body design device 100, the control unit 1 operates as these functional blocks, making it possible to determine the three-dimensional structure of a porous body having desired properties.

[0019] The learning model creation unit 11 creates a learning model capable of predicting characteristics corresponding to the three-dimensional structure of a porous body. A learning model is a prediction model that learns the relationship between the three-dimensional structure and properties of a porous body and is also called a surrogate model. For example, a support vector machine (SVM) can be used to generate a learning model using support vector regression, which can predict the properties of a porous body from the features of the three-dimensional structure of the porous body. Methods for building learning models using SVM are well known and are described, for example, in "Machine learning and data-driven characterization framework for porous materials: Permeability prediction and channeling defect detection," Chemical Engineering Journal 420 (2021) 130069, DOI: 10.1016 / j.cej.2021.130069, by Tomoki Yasuda, Shinichi Ookawara, Shiro Yoshikawa, and Hideyuki Matsumoto. Information about the learning model created by the learning model creation unit 11 is stored in the memory unit 2 as learning model data 23. Note that a learning model acquired from an external source may also be stored in the memory unit 2 as learning model data 23.

[0020] When creating a learning model, the learning model creation unit 11 acquires a predetermined number of samples of the three-dimensional structure of the porous body in advance. The various structural features of each sample are then expressed by the values ​​of multiple predetermined types of structural descriptors. Structural descriptors that represent the three-dimensional structure of the porous body include, for example, porosity, solid fraction, specific surface area, pore diameter, particle diameter, pore or solid structural uniformity, pore code length, solid code length, Pore-Throat ratio, and Pore-Throat coordination number. More specifically, the structural uniformity of the pores or solid refers to the variance of values ​​related to the pores or solid calculated for each region by dividing the porous body into multiple regions. Examples of values ​​related to the pores or solid calculated for each region include the porosity, solid fraction, specific surface area, pore diameter, and particle diameter described above. The pore chord length is the continuous length of the pores in a given direction in the porous body, and the solid chord length is the continuous length of the solid in a given direction in the porous body. The pore-throat ratio is the ratio of the pore diameter to the throat diameter (described later), and the pore-throat coordination number is the average number of throats between adjacent pores for each pore. The learning model creation unit 11 selects, from these structural descriptors, multiple structural descriptors that contribute highly to the characteristics of the porous body as input parameters for the learning model. Then, using the values ​​of each input parameter of each sample and the porous body characteristics obtained for each sample by the following physical simulation performed by the physical simulation unit 13, a learning model is created that predicts characteristics from the input parameters.

[0021] The PN model conversion unit 12 converts the three-dimensional structure of a porous body into a PN (Pore-Throat Network) model. The PN model is a model of the three-dimensional structure of a porous body, combining pores, which represent the pores, with throats, which represent the connections between the pores. In other words, the PN model expresses the three-dimensional structure of a porous body using pores, which represent the position and size of each pore formed inside the porous body, throats, which represent the pore network related to the connectivity between the pores, and the thickness (diameter) of each throat. Here, in the porous body design device 100 of this embodiment, as described in detail later, information about each throat between pores in the PN model is treated as vector information called a network gene vector. This allows the connectivity between pores in the three-dimensional structure of a porous body to be easily changed to a value represented by any vector information. The PN model information created by converting the three-dimensional structure of the porous body by the PN model conversion unit 12 is stored in the storage unit 2 as PN model data 22.

[0022] The PN model conversion unit 12 can also read out the PN model data 22 stored in the storage unit 2 and inversely convert the PN model represented by the PN model data 22 into the three-dimensional structure of the porous body, thereby reconstructing the three-dimensional structure of the porous body corresponding to the PN model. The PN model conversion unit 12 can perform inverse conversion from the PN model to the three-dimensional structure of the porous body using a well-known algorithm such as pix2pix. Information on the three-dimensional structure of the porous body obtained by inversely converting the PN model by the PN model conversion unit 12 is stored in the storage unit 2 as porous body structure data 21.

[0023] The physical simulation unit 13 performs physical calculations based on the three-dimensional structure of the porous body represented by the porous body structure data 21, and performs a physical simulation of a case in which a predetermined substance is passed through the porous body. As a result, when a fluid, such as exhaust gas from an internal combustion engine, is passed through the porous body, values ​​such as the permeability of the fluid and the filtration efficiency (capture efficiency) of particulate matter (PM) contained in the fluid can be calculated as the properties of the porous body. The properties of the porous body to be calculated may also include mechanical strength properties, electrochemical properties, thermal conductivity properties, heat exchange properties, electrical conductivity properties, gas adsorption properties, gas purification performance, catalyst coating properties, and removal efficiency of substances captured in the porous body. The results of the calculation of the properties of the porous body by the physical simulation unit 13 are stored in the storage unit 2 as property data 25. The physical simulation unit 13 can be realized, for example, by using Ansys Fluent, a thermal fluid analysis software from ANSYS, Inc. The transmittance and filtration efficiency can also be calculated using the FlowDict module and FilterDict module, respectively, included in GeoDict, a microstructure simulation software developed by Math2Market GmbH.

[0024] The AI ​​calculation unit 14 uses the learning model created by the learning model creation unit 11 to perform calculations related to the prediction of porous body properties using AI (Artificial Intelligence). By inputting the values ​​of each of the above-mentioned input parameters determined for the three-dimensional structure of the porous body into the learning model, the AI ​​calculation unit 14 can perform AI calculations using well-known AI methods such as support vector regression, and predict the properties of the porous body. The values ​​of each input parameter used in the AI ​​calculation unit 14 and the obtained property prediction results of the porous body are stored in the memory unit 2 as input parameter data 24 and property data 25, respectively.

[0025] The property evaluation unit 15 performs an evaluation process to evaluate the properties of the porous body calculated or predicted by the physical simulation unit 13 or the AI ​​calculation unit 14. The property evaluation unit 15 performs an evaluation process for the properties of the porous body, for example, by plotting each calculated or predicted property value on a graph. Note that other methods may be used as long as they can appropriately evaluate the properties of the porous body. For example, an evaluation score for the properties of the porous body may be calculated from each property value, or each property value may be determined to satisfy a predetermined standard value, and the properties of the porous body may be judged as OK / NG based on the judgment result. The property evaluation results of the porous body by the property evaluation unit 15 are stored in the memory unit 2 as property evaluation data 26.

[0026] The gene evolution unit 16 performs a process to change the connectivity between pores in the PN model. The gene evolution unit 16 performs a gene evolution process using a well-known calculation method such as a genetic algorithm, thereby changing the network gene vector in the PN model data 22 so as to improve the characteristics of the porous body having a three-dimensional structure with connectivity corresponding to the network gene vector. By repeating this gene evolution process by the gene evolution unit 16, the contents of the PN model represented by the PN model data 22 can be successively updated, and optimal connectivity can be searched for.

[0027] The learning model update unit 17 performs an update process for the learning model based on the results of the physical simulation performed by the physical simulation unit 13. As will be described later, the physical simulation unit 13 performs a physical simulation based on the three-dimensional structure of the porous body each time the AI ​​calculation unit 14 performs a predetermined number of AI calculations. When the physical simulation unit 13 performs a physical simulation in this manner, the learning model update unit 17 uses the results to perform an update process for the learning model and update the contents of the learning model data 23. This makes it possible to reflect the results of the physical simulation in the learning model and update the learning model data 23 so that the values ​​of the properties calculated by the AI ​​calculation from the input parameter values ​​corresponding to the three-dimensional structure of the porous body become more accurate.

[0028] The data conversion unit 18 converts the porous body structure data 21, which is based on the three-dimensional structure of the porous body having the finally determined connectivity between pores, into three-dimensional shape data 27. The three-dimensional shape data 27 generated by conversion from the porous body structure data 21 by the data conversion unit 18 is stored in the storage unit 2 and, as necessary, is read from the storage unit 2 and provided to an external device outside the porous body design device 100. For example, by inputting the three-dimensional shape data 27 into a 3D printer, the 3D printer is operated according to the three-dimensional shape data 27 to produce a porous body that reproduces the three-dimensional structure of the porous body represented by the porous body structure data 21. This makes it possible to create a prototype of the porous body designed by the porous body design device 100. The method for producing a porous body using a 3D printer and the three-dimensional shape data 27 will be described in detail below.

[0029] Next, details of the processing performed by the control unit 1 will be described with reference to Fig. 2. Fig. 2 is a flowchart showing the processing flow of the porous body design device according to the first embodiment of the present invention. The porous body design device 100 of this embodiment searches for a three-dimensional structure of a porous body having desired properties by repeatedly executing the processing shown in the flowchart of Fig. 2 by the control unit 1 in response to operational inputs from the user. This supports the user in designing a porous body.

[0030] When the user instructs the porous body design device 100 to start designing a porous body via the operation input device 4, in step S10 the control unit 1 sets the initial value 0 to each of the variables i and j. The variable i is a variable for counting the number of times the AI ​​calculation unit 14 has performed AI calculations between the time the physical simulation unit 13 performed one physical simulation and the time the next physical simulation is performed. The variable j is a variable for counting the number of times the gene evolution unit 16 has performed gene evolution processing, i.e., the number of generations of network gene vectors generated so far by the gene evolution processing. While the processing shown in the flowchart of FIG. 2 is being executed, the values ​​of the variables i and j are stored in the memory 3.

[0031] In step S20, the control unit 1 performs an initial learning model generation process to set the initial state of the learning model data 23. Here, using the learning model creation unit 11 and the physical simulation unit 13, a large number of samples of the three-dimensional structure of the porous body are obtained, and the relationship between the values ​​of input parameters selected from the structural descriptors of each sample and the properties is determined. Then, a learning model can be created based on the determined relationship, thereby setting the initial state of the learning model data 23. Details of the initial learning model generation process in this embodiment will be described later with reference to FIG. 3.

[0032] In step S30, the control unit 1 acquires the three-dimensional structure of the porous body of interest. The three-dimensional structure of the porous body of interest refers to the three-dimensional structure of a porous body that is selected as the starting point for the search by the porous body design system 100 when searching for a three-dimensional structure of a porous body having desired properties. For example, an existing porous body can be selected as the porous body of interest, and the three-dimensional structure of the porous body of interest can be acquired by photographing the porous body of interest using a computed tomography (CT) device, or a three-dimensional structure virtually generated in the porous body design system 100 using the aforementioned GeoDict. Once the three-dimensional structure of the porous body of interest is acquired, the information is stored in the storage unit 2 as porous body structure data 21, and the process proceeds to step S40.

[0033] In step S40, the control unit 1 converts the three-dimensional structure of the porous body of interest acquired in step S30 into a PN model using the PN model conversion unit 12. Here, the three-dimensional structure of the porous body of interest can be converted into a PN model, for example, as described below.

[0034] Fig. 4 is a diagram showing an example of a PN model. In Fig. 4, the circles denoted by reference numerals 41 to 45 represent pores contained in the porous body of interest. The line segments denoted by reference numerals 51 to 57 represent throats or unconnected throats present between the pores 41 to 45. In Fig. 4, the line segments 51, 53, and 56 shown by solid lines represent throats that actually connect pores, while the line segments 52, 54, 55, and 57 shown by dashed lines represent unconnected throats that do not actually connect pores.

[0035] When the PN model conversion unit 12 detects the positions and sizes of the pores 41-45 from the three-dimensional structure of the porous body of interest, it extracts combinations of the pores 41-45 whose mutual distances are within a predetermined range, and provisionally sets throat candidates 51-57 that connect the pores for each combination. Of these throat candidates 51-57, those that actually connect the pores are set as throats, and those that do not actually connect the pores are set as unconnected throats. As a result, throats 51, 53, and 56 and unconnected throats 52, 54, 55, and 57 are set, and the three-dimensional structure of the porous body of interest is converted into the PN model shown in FIG. 4.

[0036] After converting the three-dimensional structure of the porous body of interest into a PN model as described above, the control unit 1 generates initial values ​​for a network gene vector to which a genetic algorithm is applied based on the network gene vector corresponding to the PN model in the following step S50. As described above, a network gene vector is a vector that represents the information about each throat between pores in the PN model, and each element of the vector represents the connection information for each throat in the PN model. Here, for the network gene vector representing the PN model of the porous body of interest created in step S40, a combination of element values ​​consisting of the same number of vector elements as the network gene vector representing the PN model of the porous body of interest, and which are the same as or different from the element values ​​of the network gene vector, is generated as the initial value of the network gene vector. For example, for the PN model of Figure 4 corresponding to the three-dimensional structure of the porous body of interest, the values ​​of each vector element corresponding to throats 51, 53, and 56 are set to "1" and the values ​​of each vector element corresponding to unconnected throats 52, 54, 55, and 57 are set to "0" to create a network gene vector A for the porous body of interest. The initial values ​​of the network gene vector can then be generated by using the element values ​​of this network gene vector A as is or by arbitrarily changing them.

[0037] In step S50, the values ​​of each element of the network gene vector A are randomly changed by a predetermined population number (number of individuals in a generation) P, and these values ​​are set as the initial values ​​of the network gene vector. For example, if P=100, 100 combinations of vector element values, each consisting of the same number of vector element values ​​as the network gene vector A of the porous body of interest, are set as the initial values ​​of the network gene vector. Note that each porous body represented by this initial value is used as the parent generation for the first-generation porous body in the gene evolution process of step S160, which will be described later. Therefore, in the following explanation, the porous body represented by the initial value of the network gene vector will be referred to as the 0th-generation porous body.

[0038] The control unit 1 associates information about the position and size of each pore in the PN model created in step S40 with information about the initial values ​​of the network gene vector created in step S50, and stores these in the storage unit 2 as PN model data 22. As a result, each PN model data 22 representing the three-dimensional structure of the 0th generation porous body and having a different throat structure is stored in the storage unit 2 for the population number P. Then, the process proceeds to step S55.

[0039] In step S55, the control unit 1 causes the PN model conversion unit 12 to reconstruct the three-dimensional structure of each zeroth-generation porous body from the initial values ​​of the network gene vector created in step S50. Here, the P PN model data 22 recorded in step S50 are read from the storage unit 2, and the position and size of each pore in the PN model represented by each PN model data 22, as well as the combination of throats connecting each pore, are reproduced on the porous body design device 100. This process can be realized using a well-known algorithm, such as pix2pix. As a result, each PN model represented by the initial values ​​of the network gene vector is inversely converted into the three-dimensional structure of each zeroth-generation porous body. The three-dimensional structure of each zeroth-generation porous body, which has different pore connectivity from the three-dimensional structure of the target porous body, can be virtually generated on the porous body design device 100. After recording the information on the three-dimensional structure of each reconstructed zeroth-generation porous body in the porous body structure data 21, the process proceeds to step S60.

[0040] In step S60, the control unit 1 calculates the values of the input parameters corresponding to the three-dimensional structure of each porous body of the j-th generation obtained in step S55 (when j = 0) or step S170 (when j ≥ 1) described later. Here, as described above, for example, among the structure descriptors such as porosity, solid fraction, specific surface area, pore diameter, particle diameter, structural uniformity of pores or solids, code length of pores, code length of solids, Pore-Throat ratio, coordination number of Pore-Throat, etc., the learning model creation unit 11 pre-selects each structure descriptor with a high contribution to the characteristics of the porous body as the input parameter of the learning model, and calculates the combination of the input parameter values for the three-dimensional structure of the j-th generation porous body obtained in step S55 or step S170. After storing the input parameter values calculated for the three-dimensional structure of the j-th generation porous body as input parameter data 24 in the storage unit 2, the process proceeds to step S70.

[0041] In step S70, the control unit 1 determines whether the value of the current variable i is equal to a predetermined value M. If i < M, the process proceeds to step S80, and if i = M, the process proceeds to step S100. The value of M represents the number of settings for performing the physical simulation and updating the learning model each time the AI operation is executed, and an arbitrary number of 1 or more, for example, M = 20 can be set.

[0042] In step S80, the control unit 1 performs an AI operation for characteristic prediction using the learning model based on the input parameter values calculated in step S60 by the AI operation unit 14. Here, the learning model data 23 is read out to obtain the learning model, and an AI operation regarding the characteristic prediction of the porous body is performed by inputting the values of the input parameters obtained from the three-dimensional structure of the j-th generation porous body to this learning model. As a result, a characteristic prediction value corresponding to the combination of the input parameters is obtained, and the characteristic prediction of the j-th generation porous body is performed. After storing the characteristic prediction value obtained for the j-th generation porous body as characteristic data 25 in the storage unit 2, the process proceeds to step S90.

[0043] In step S90, the control unit 1 adds 1 to the value of the variable i, and the process proceeds to step S130.

[0044] In step S100, the control unit 1 performs property calculations for the jth generation porous body and calculates its property values ​​by carrying out a physical simulation based on the three-dimensional structure of the jth generation porous body acquired in step S170 using the physical simulation unit 13. After storing the calculation results of the property values ​​obtained for the jth generation porous body in the storage unit 2 as property data 25, the process proceeds to step S110.

[0045] In step S110, the control unit 1 causes the learning model update unit 17 to update the learning model represented by the learning model data 23 based on the results of the physical simulation performed in step S100. Here, the control unit 1 reads from the memory unit 2 the input parameter data 24 stored in step S60 and the property data 25 stored in step S100 for the porous bodies that have been the subject of the physical simulations performed up to that point, and updates the contents of the learning model (weight values) based on the relationship between the input parameter values ​​and property values ​​of the jth generation porous body in these data. After recording the contents of the updated learning model in the learning model data 23, the process proceeds to step S120.

[0046] In step S120, the control unit 1 resets the value of the variable i to 0, and the process proceeds to step S130.

[0047] In step S130, the control unit 1 causes the characteristic evaluation unit 15 to plot the characteristic values ​​predicted or calculated in step S80 or S100 on a graph, thereby evaluating the predicted / calculated results of the jth generation porous body. Here, for example, a two-dimensional graph is prepared, with one axis representing fluid permeability, one of the porous body's characteristics, and the other axis representing filtration efficiency, another of the porous body's characteristics, and the predicted or calculated characteristic values ​​for the jth generation porous body are plotted on this two-dimensional graph. This allows the predicted / calculated results of the jth generation porous body to be evaluated based on the position of each point plotted on the two-dimensional graph. Note that, as long as the predicted / calculated results can be evaluated, it is not necessary to actually plot the points on the graph; simply storing data indicating the positions of the points on the graph is sufficient. Alternatively, as described above, the predicted / calculated results of the properties may be evaluated using methods other than plotting on a graph.

[0048] In step S135, the control unit 1 selects parent porous bodies for the next generation based on the evaluation results of the property prediction / calculation results obtained in step S130. Here, for example, the property values ​​predicted or calculated in the previous and current steps S80 or S100 are relatively evaluated based on the positions of each point plotted on the graph in step S130. Then, a predetermined percentage of the porous bodies corresponding to the most suitable property values ​​are selected from these, for example, P porous bodies (where P is the population number) having property values ​​corresponding to the top 50% of the properties, as parent porous bodies for the next generation, i.e., parent porous bodies for the (j+1)th generation. However, when the processing of step S135 is performed for the first time, i.e., when j=0, there are no plotted points from the previous generation. Therefore, in this case, all of the porous bodies of the 0th generation represented by the initial values ​​of the network gene vector generated in step S50 can be selected as parent porous bodies for the first generation.

[0049] In step S140, the control unit 1 determines whether the value of the current variable j is equal to a predetermined value N. If j = N, the process proceeds to step S150, and if j < N, the process proceeds to step S160. The value of N represents the number of times the gene evolution unit 16 executes the gene evolution process, and any number greater than or equal to 1 can be set, for example, N = 100.

[0050] In step S150, the control unit 1 determines the three-dimensional structure of the final target porous body based on the evaluation results for the characteristics of the porous bodies of each generation obtained so far. Here, for example, from the positions of the points plotted on the graph in the immediately preceding step S130, the characteristic values predicted or calculated in this step S80 or S100 are relatively evaluated, and the characteristic value with the highest evaluation is specified. For example, if the transmittance and the filtration efficiency are each within a predetermined range and a point that exists at the position farthest from the origin of the graph in the upper right direction is extracted, the characteristic value represented by that point can be specified as the characteristic value with the highest evaluation.

[0051] If the characteristic value with the highest evaluation can be specified in the process of step S150, the control unit 1 determines the three-dimensional structure corresponding to that characteristic value as the three-dimensional structure of the final target porous body and ends the process shown in the flowchart of FIG. 2.

[0052] In step S160, the control unit 1 generates a next-candidate network gene vector using the gene evolution unit 16. Here, gene evolution processing is performed based on the network gene vectors of each porous body selected as the next-generation parent porous body in the immediately preceding step S135, and the network gene vectors of child porous bodies for those parent porous bodies are created to generate the next-candidate network gene vector. Specifically, using a well-known computational method such as a genetic algorithm, the network gene vectors of the j-th generation porous bodies selected in step S135 are multiplied together to generate a network gene vector of a new porous body corresponding to that child as the network gene vector of the (j+1)-th generation porous body. By repeating this operation for multiple combinations of parent porous bodies, any number of network gene vectors of the next-generation child porous bodies are generated. For example, P pairs of parent porous bodies in the j-th generation (P is the population number) are set, and P network gene vectors of the (j+1)-th generation child porous bodies are created from these pairs as next-candidate network gene vectors. As a result, it is possible to obtain a population number P of three-dimensional structures of (j+1)-generation porous bodies in which the pore positions and sizes remain the same and only the connectivity between pores differs from the three-dimensional structure of the j-th generation porous body. Once the values ​​of each element of the network gene vector corresponding to the (j+1)-generation porous body newly obtained by this gene evolution process are recorded in the PN model data 22, the process proceeds to step S170.

[0053] In step S170, the control unit 1 causes the PN model conversion unit 12 to reconstruct the three-dimensional structure of each porous body of the (j+1)th generation from the next candidate network gene vector generated by the gene evolution process in step S160. Here, the P PN model data 22 recorded in the previous step S160 are read from the storage unit 2, and the position and size of each pore in the PN model represented by each PN model data 22, as well as the combination of throats connecting each pore, are reproduced on the porous body design system 100. This process, like step S55, can be realized using a well-known algorithm such as pix2pix. As a result, each PN model represented by the network gene vector after the gene evolution process is inversely converted into the three-dimensional structure of each porous body of the (j+1)th generation. The three-dimensional structure of each porous body of the (j+1)th generation, which has different pore connectivity from the three-dimensional structure of the porous body of interest, can be virtually generated on the porous body design system 100. Once the information on the three-dimensional structure of each porous body of the (j+1)th generation thus reconstructed is recorded in the porous body structure data 21, the process proceeds to step S180.

[0054] In step S180, the control unit 1 adds 1 to the value of the variable j, and returns the process to step S60.

[0055] After returning from step S180 to step S60, the control unit 1 repeats the processing from step S60 onwards. Based on the network gene vectors obtained after the gene evolution processing, a new three-dimensional structure of a porous body with pore connectivity different from that of the target porous body is generated, and characteristic values ​​are predicted or calculated using AI calculations or physical simulations. By plotting the obtained characteristic prediction / calculation results on a graph, new evaluation results for the characteristic values ​​of each next-generation porous body are obtained in addition to the evaluation results for the characteristic values ​​of each previous-generation porous body obtained in the previous processing. Based on the obtained evaluation results, the next-generation parent porous body to be used in the gene evolution processing is selected. This series of processing realizes an optimization process that changes the pore connectivity in the three-dimensional structure of the porous body to search for the optimal pore connectivity.

[0056] FIG. 3 is a flowchart showing details of the initial learning model generation process according to the first embodiment of the present invention, which is executed in step S20 of FIG.

[0057] In step S210, the control unit 1 determines the three-dimensional structure of the porous body to be used in generating the initial learning model by the learning model creation unit 11. Here, for example, a three-dimensional structure of a porous body set in advance as a sample is determined as the three-dimensional structure for generating the initial learning model.

[0058] In step S220, the control unit 1 calculates values ​​of structural descriptors corresponding to the three-dimensional structure determined in step S210 using the learning model creation unit 11. Here, the control unit 1 calculates values ​​of various predetermined structural descriptors, such as porosity, solid fraction, specific surface area, pore diameter, particle diameter, pore or solid structural uniformity, pore chord length, solid chord length, Pore-Throat ratio, and Pore-Throat coordination number, for the three-dimensional structure of the porous body determined in step S210.

[0059] In step S230, the control unit 1 performs a physical simulation based on the three-dimensional structure determined in step S210 using the physical simulation unit 13. Here, by performing processing similar to that of step S100 in Fig. 2, the properties of the porous body having a three-dimensional structure for generating the initial learning model are calculated by physical simulation, and the property values ​​obtained as the simulation results are obtained.

[0060] In step S240, the control unit 1 determines whether each of the processes in steps S210 to S230 has been performed for a predetermined number of samples. If the processes in steps S210 to S230 have been performed for the predetermined number of porous bodies and the characteristic values ​​for each sample have been calculated, the process proceeds to step S250. On the other hand, if the number of three-dimensional structures for which characteristic values ​​have been calculated by the processes in steps S210 to S230 is less than the predetermined number of samples and the characteristic values ​​acquired so far do not satisfy the number of samples, the process returns to step S210 and the processes from step S210 onwards are performed again.

[0061] In step S250, the control unit 1 causes the learning model creation unit 11 to select, from the structural descriptors of each sample calculated in step S220, structural descriptors that are highly correlated with the characteristic values ​​of the porous body calculated by physical simulation in step S230 as input parameters of the learning model. Here, for example, the correlation between the structural descriptor value calculated for each sample and the characteristic value is calculated for each structural descriptor, and a predetermined number of structural descriptors are selected in descending order of the obtained correlation values ​​as input parameters of the learning model. This allows structural descriptors that contribute most to the characteristics of the porous body to be selected as input parameters of the learning model.

[0062] In step S260, the control unit 1 creates an initial learning model based on the input parameters selected in step S250 and the characteristic values ​​of the porous body determined in step S230, using the learning model creation unit 11. Here, the initial learning model can be created from the relationship between the input parameters and characteristic values ​​of each sample by performing a model learning process using, for example, a well-known method.

[0063] If the initial learning model is created in step S260, the control unit 1 records the contents of the initial learning model in the learning model data 23 and ends the initial learning model generation process shown in the flowchart of Fig. 3. Thereafter, the process proceeds to step S30 in Fig. 2.

[0064] In the flowchart of Figure 2, the process includes step S130, in which the characteristic values ​​predicted or calculated by AI calculation or physical simulation are evaluated, for example by plotting them on a graph; step S160, which is a gene evolution process; and step S170, which is a process in which the 3D structure of the porous body is reconstructed from a PN model of the network gene vector after the gene evolution process. By repeating these steps, a large number of plot points representing the characteristic values ​​of porous bodies having various 3D structures with different pore connectivity are finally obtained. In step S150, the point exhibiting the best characteristic among these plot points is selected as the most highly evaluated characteristic value. The 3D structure corresponding to this characteristic value is then identified to determine the final 3D structure of the porous body of interest. Even if the points are not actually plotted on a graph as described above but data indicating the positions of the points on the graph is stored, it is possible to determine the final 3D structure of the porous body of interest by selecting the point exhibiting the best characteristic using a similar method and identifying the 3D structure corresponding to that point.

[0065] Here, the pore connectivity index and its characteristic value in the original 3D structure of the porous body of interest are compared with the pore connectivity index and its characteristic value in the 3D structure of the porous body of interest after optimization, which is finally determined when the process shown in the flowchart of FIG. 2 is repeated a predetermined number of times, for example, 100 times. The comparison results are, for example, as follows. In the following explanation, the pore connectivity index and its characteristic values ​​(permeability, filtration efficiency) in the original 3D structure of the porous body of interest are each set to 100, and the 3D structure of the porous body of interest after optimization and its characteristic values ​​are normalized. Here, a larger value of the connectivity index indicates a larger number of throats connecting pores, indicating a larger proportion of connected throats among the throat candidates in the PN model. Larger values ​​of the permeability and filtration efficiency indicate better characteristics, respectively. Original porous body of interest: Connectivity index = 100, permeability = 100, filtration efficiency = 100 Optimized porous body: connectivity index = 108, permeability = 122, filtration efficiency = 105

[0066] The above comparison results show that the connectivity index value of the optimized porous body of interest is higher than that of the original porous body of interest, thereby enabling good properties to be obtained. Therefore, by adopting the porous body design method using the porous body design system 100 of this embodiment, it is possible to adjust the connectivity between pores to an optimal value and determine the three-dimensional structure of a porous body with desired properties.

[0067] Next, the method for manufacturing a porous body by a 3D printer using the three-dimensional shape data 27 will be described in detail with reference to Fig. 5. Fig. 5 is a flowchart showing the flow of the manufacturing process for a porous body using a 3D printer.

[0068] When the user instructs the porous body design device 100 to start manufacturing a porous body via the operation input device 4, in step S310 the control unit 1 creates three-dimensional shape data 27 using the data conversion unit 18. Here, the porous body structure data 21 is read from the storage unit 2 and subjected to a predetermined data conversion process to convert the porous body structure data 21 into three-dimensional shape data 27. Note that the porous body structure data 21 may be used as the three-dimensional shape data 27 as is. In this case, the control unit 1 does not need to have the data conversion unit 18.

[0069] In step S320, the three-dimensional shape data 27 created in step S310 is input to a 3D printer (not shown). Here, the three-dimensional shape data 27 is input to the 3D printer by, for example, connecting the porous body design device 100 to the 3D printer wirelessly or via a wire, and transmitting the three-dimensional shape data 27 from the porous body design device 100 to the 3D printer. Alternatively, the three-dimensional shape data 27 may be input to the 3D printer by transferring the three-dimensional shape data 27 from the porous body design device 100 to the 3D printer via a storage medium such as a USB memory. In addition to this, the three-dimensional shape data 27 can be input to the 3D printer by any other method.

[0070] In step S330, the 3D printer starts injecting material to form a porous body based on the three-dimensional shape data 27 input in step S320. Then, in step S340, the head of the 3D printer is moved to the coordinate position represented by the three-dimensional shape data 27 input in step S320, and then material is injected.

[0071] In step S350, it is determined whether or not material has been injected from the 3D printer for all three-dimensional structures of the porous body represented by the three-dimensional shape data 27 input in step S320. If there are any three-dimensional structure portions for which material has not yet been injected, the process returns to step S340 to continue injecting material, and if material has been injected for all three-dimensional structures, the process proceeds to step S360.

[0072] In step S360, the injection of material in the 3D printer is completed. This completes the production of a porous body using the 3D printer, and a porous body that reproduces the three-dimensional structure designed by the porous body design device 100 can be created. After the processing of step S360 has been executed, the processing shown in the flowchart of FIG. 5 is completed.

[0073] Although the above describes an example of a method for producing a porous body using a 3D printer and the three-dimensional shape data 27, a porous body that reproduces the three-dimensional structure designed by the porous body design device 100 may be produced by other methods. Alternatively, a porous body may be produced using something other than a 3D printer. As long as a porous body that reproduces the three-dimensional structure designed by the porous body design device 100 can be properly produced using the three-dimensional shape data 27, any method can be selected from a variety of well-known methods and used to produce the porous body.

[0074] According to the first embodiment of the present invention described above, the following advantageous effects are achieved.

[0075] (1) A method for designing a porous body using the porous body design device 100 involves having the porous body design device 100 execute a property prediction / calculation process (steps S80, S100) that predicts or calculates the properties of a porous body having a predetermined three-dimensional structure, an evaluation process (step S130) that evaluates the properties of the porous body predicted or calculated by the property prediction / calculation process, and an optimization process (steps S135, S160) that changes the connectivity between pores in the three-dimensional structure to search for optimal connectivity.The three-dimensional structure of the porous body is then determined based on the evaluation results of the properties of the porous body obtained by the evaluation process (step S150).This provides a useful technology that enables early determination of an optimal structure in the design of a porous body.

[0076] (2) In the porous body design method, the porous body design device 100 executes a PN modeling process (step S40) to generate a PN model that represents a three-dimensional structure using a plurality of pores and a plurality of throats that connect the pores. In the process of step S160, the connectivity is changed by adding or deleting throats in the PN model generated in the PN modeling process of step S40. In this way, the connectivity between pores in the three-dimensional structure can be easily changed.

[0077] (3) In the property prediction / calculation process, the properties of the porous body are predicted based on the PN model generated in the PN modeling process of step S40. Specifically, the porous body design device 100 is caused to execute a structure generation process (step S170) in which a three-dimensional structure of the porous body is virtually generated on the porous body design device 100 based on the PN model to which a throat has been added or deleted in the process of step S160. In the property prediction process of step S80, the properties of the porous body having the three-dimensional structure generated by this structure generation process are predicted. This allows reliable prediction of the properties of a porous body whose pore connectivity has been changed by the optimization process.

[0078] (4) In addition, the property prediction / calculation process involves at least one of a physical simulation (step S100) that calculates the properties of the porous body by physical calculation based on the three-dimensional structure generated by the structure generation process in step S170, and an AI calculation (step S80) that predicts the properties of the porous body using a learning model of the three-dimensional structure. This allows for appropriate physical simulation and AI calculation to be performed, making it possible to determine the three-dimensional structure of the porous body that will provide better properties.

[0079] (5) It is also possible to manufacture a porous body (steps S330 to S360) using three-dimensional shape data based on the three-dimensional structure of the porous body determined by the porous body design method described above. In this way, a prototype of the porous body designed by the porous body design device 100 can be quickly and easily produced.

[0080] (Second embodiment) Next, a second embodiment of the present invention will be described. In the first embodiment, an example was described in which the optimal connectivity of a PN model representing the three-dimensional structure of a porous body was searched for using a learning model in which structural descriptors representing the three-dimensional structure of the porous body that have a high contribution to the properties of the porous body were used as input parameters. Therefore, the three-dimensional structure of the porous body was reconstructed from a PN model corresponding to a network gene vector after gene evolution processing, and input parameters calculated based on that three-dimensional structure were input into the learning model to predict the properties of the porous body using AI calculations. In contrast, in the present embodiment, an example will be described below in which the properties of a porous body are predicted directly using AI calculations from a PN model representing the three-dimensional structure of the porous body.

[0081] The porous body design system according to this embodiment has the same configuration as that described in the first embodiment in Fig. 1. Therefore, the porous body design system according to this embodiment will be described below using the configuration of the porous body design system 100 shown in Fig. 1.

[0082] Fig. 6 is a flowchart showing the processing flow of a porous body design apparatus according to a second embodiment of the present invention. The porous body design apparatus 100 of this embodiment searches for a three-dimensional structure of a porous body having desired properties by repeatedly executing the processing shown in the flowchart of Fig. 6 using the control unit 1 in response to operational inputs from the user. This supports the user in designing a porous body.

[0083] In the flowchart of Fig. 6, the parts that perform the same processes as those in the flowchart of Fig. 2 explained in the first embodiment are given the same step numbers as in Fig. 2. Therefore, the following description will focus on the processing contents of the step numbers that are different from Fig. 2, and a description of the common processes will be omitted unless particularly necessary.

[0084] In step S20A, the control unit 1 performs an initial learning model generation process to set the initial state of the learning model data 23. Here, the learning model creation unit 11 and the physical simulation unit 13 are used to obtain a large number of samples of the three-dimensional structure of the porous body, and the relationship between the input parameters and characteristics that can be directly calculated from the PN model obtained from each sample is determined. Then, a learning model is created based on the determined relationship, thereby setting the initial state of the learning model data 23. Details of the initial learning model generation process in this embodiment will be described later with reference to FIG. 7.

[0085] In step S50A, the control unit 1 generates initial values ​​for the PN model to which the genetic algorithm is applied. Here, the initial values ​​for the PN model are generated by, for example, adding or deleting throats or changing the number and positions of pores in the PN model of the porous body of interest created in step S40. At this time, as in step S50 of FIG. 2, the PN model of the porous body of interest is randomly changed by a predetermined population number (number of individuals in a generation) P, and this is set as the initial value of the PN model. In the following explanation, the porous body represented by the initial values ​​of the PN model will be referred to as the porous body of the 0th generation.

[0086] In this embodiment, when the initial value of the PN model is set in step S50A and the information is stored in the storage unit 2 as the PN model data 22, the process proceeds to step S60A without performing the process of step S55 in FIG. 2.

[0087] In step S60A, the control unit 1 calculates the values of the input parameters that can be calculated from the PN model of each porous body of the j-th generation obtained in step S50A (when j = 0) or step S160A (when j ≥ 1) described later. Here, among the aforementioned structure descriptors related to the PN model, for example, those with a high contribution to the characteristics of the porous body, such as the Pore-Throat ratio and the coordination number of Pore-Throat, etc., are selected in advance by the learning model creation unit 11, and the combination of the input parameter values that can be calculated from the PN model of the porous body of the j-th generation obtained in step S50A or step S160A is calculated as the input parameters of the learning model. After storing the input parameter values calculated from the PN model of the porous body of the j-th generation in the storage unit 2 as the input parameter data 24, the process proceeds to step S70. Then, in step S70, it is determined whether the value of the current variable i is the predetermined value M. If i < M, the process proceeds to step S80A, and if i = M, the process proceeds to step S95.

[0088] In step S80A, the control unit 1 performs AI calculation for characteristic prediction using the learning model based on the input parameter values calculated in step S60A by the AI calculation unit 14. Here, the learning model data 23 is read out to obtain the learning model, and by inputting the values of the input parameters calculated from the PN model of the porous body of the j-th generation to this learning model, an AI calculation related to the characteristic prediction of the porous body is performed. As a result, the characteristic prediction value corresponding to the input parameters calculated from the PN model is obtained, and the characteristic prediction of the porous body of the j-th generation is performed. After storing the characteristic prediction value obtained for the porous body of the j-th generation in the storage unit 2 as the characteristic data 25, the process proceeds to step S90.

[0089] In step S95, the control unit 1 reconstructs the three-dimensional structure of the j-th generation porous body from the PN model by the PN model conversion unit 12. Here, the PN model data 22 representing the PN model of the j-th generation porous body generated by the gene evolution process performed in step S160A in the previous process is read from the storage unit 2, and the positions and sizes of each pore in the PN model and the throats connecting between the pores are reproduced on the porous body design device 100. This process can be realized by using a well-known algorithm such as pix2pix as in steps S55 and S170 of FIG. 2 described in the first embodiment. As a result, the PN model after the gene evolution process is inversely converted into the three-dimensional structure of each porous body of the j-th generation, and the three-dimensional structure of the j-th generation porous body with different pore-to-pore connectivity from the three-dimensional structure of the target porous body can be virtually generated on the porous body design device 100. After recording the information on the three-dimensional structure of each porous body of the j-th generation thus reconstructed in the porous body structure data 21, the process proceeds to step S100.

[0090] When it is determined in step S140 that j < N, in step S160A, the control unit 1 generates a next candidate PN model by the gene evolution unit 16. Here, a gene evolution process based on the PN models of each porous body selected as the next-generation parent porous body in the immediately preceding step S135 is performed, and the PN model of the child porous body for the parent porous body is created, thereby generating a next candidate PN model. Specifically, for example, by a well-known arithmetic method such as a genetic algorithm, the PN models of the j-th generation porous bodies selected in step S135 are multiplied together to generate a PN model of a new porous body corresponding to the child as the PN model of the (j + 1)-th generation. By repeating such an operation for a plurality of combinations of parent porous bodies, an arbitrary number of PN models of the next-generation child porous bodies can be generated. For example, P pairs (P is the population number) of the j-th generation parent porous bodies are set, and P PN models of the (j + 1)-th generation child porous bodies are created as the next candidate PN models from these pairs. As a result, with respect to the three-dimensional structure of the j-th generation porous body, the three-dimensional structures of the (j + 1)-th generation porous bodies with different pore-to-pore connectivity can be obtained in the number of population P.

[0091] In this embodiment, after performing the process of step S160A and recording the PN model after the gene evolution process in the PN model data 22, the process proceeds to step S180 without performing the process of step S170 in FIG. 2. Then, after adding 1 to the value of variable j in step S180, the process returns to step S60A and repeats the processes from step S60A onward. This allows prediction or calculation of property values ​​by AI calculation or physical simulation based on the PN model after the gene evolution process. Then, by adding a plot of the obtained property prediction / calculation results to a graph, new evaluation results for the property values ​​of each porous body of the next generation are obtained in addition to the evaluation results for the property values ​​of each porous body of the previous generation obtained in the previous process. The next-generation parent porous body to be used in the gene evolution process is selected from the obtained evaluation results. This series of processes realizes an optimization process that changes the connectivity between pores in the three-dimensional structure of the porous body to search for the optimal connectivity between pores.

[0092] FIG. 7 is a flowchart showing details of the initial learning model generation process according to the second embodiment of the present invention, which is executed in step S20A of FIG.

[0093] In the flowchart of Fig. 7, the parts that perform the same processes as those in the flowchart of Fig. 3 described in the first embodiment are given the same step numbers as in Fig. 3. Therefore, the following description will focus on the processing content of the step numbers that are different from Fig. 3, and a description of the common processes will be omitted unless particularly necessary.

[0094] In step S210A, the control unit 1 converts the three-dimensional structure determined in step S210 into a PN model using the PN model conversion unit 12. Here, the three-dimensional structure of a porous body set in advance as a sample can be converted into a PN model using a method similar to that of step S40 in Fig. 6.

[0095] In step S220A, the control unit 1 calculates values ​​of structural descriptors corresponding to the PN model obtained by converting the three-dimensional structure in step S210 using the learning model creation unit 11. As described above, the control unit 1 calculates values ​​of various structural descriptors related to the PN model, such as the Pore-Throat ratio and the Pore-Throat coordination number, for the PN model obtained in step S210.

[0096] In step S260A, the control unit 1 creates an initial learning model using the learning model creation unit 11 based on the input parameters of the PN model selected in step S250 from the structural descriptors calculated in step S220A and the predicted property values ​​of the porous body obtained in step S230. Here, the initial learning model can be created by performing a model learning process using, for example, a well-known method, based on the relationship between the input parameters and the predicted property values ​​corresponding to the PN model of each sample.

[0097] If the initial learning model can be created in step S260A, the control unit 1 records the contents of the initial learning model in the learning model data 23 and ends the initial learning model generation process shown in the flowchart of Fig. 7. Thereafter, the process proceeds to step S30 in Fig. 6.

[0098] According to the second embodiment of the present invention described above, in the property prediction process of step S80A, the properties of a porous body are predicted based on the initial values ​​of the PN model set in step S50A, or on the PN model to which a throat has been added or deleted in the previous process of step S160A. This makes it possible to predict the properties of a porous body in which the connectivity between pores has been changed by the optimization process in a shorter processing time.

[0099] (Third embodiment) Next, a third embodiment of the present invention will be described. In this embodiment, an example of predicting the properties of a porous body by calculation using mathematical formulas from a PN model that represents the three-dimensional structure of the porous body will be described below.

[0100] In the PN model, when considering the flow rate between two pores, assuming saturated flow and constant fluid viscosity μ, the conductance G can be expressed by the following equation (1) using the flow resistance R and fluid viscosity μ. G = 1 / (Rμ) (1)

[0101] From the above equation (1), the pressure vector of one pore is P i , the pressure vector of the other pore is P j Then, the flow rate between these two pores can be expressed by the following equation (2). Q ij =G ij (P i -P j ) ···(2)

[0102] In addition, in the three-dimensional structure of a porous material represented by the PN model, the pressure in the pores can be calculated by solving the equation expressed by the following equation (3). GP=b (3)

[0103] In equation (3), G is a modified diagonal element of the adjacency matrix weighted by conductance, and can be calculated by replacing the diagonal element of the boundary pore with 1 and the other elements with the sum of that row multiplied by -1. This allows the pressure boundary condition and mass conservation equation to be incorporated. Furthermore, P is the pore pressure vector, and b is a vector in which the row of the boundary pore has a specified pressure and the rest has a value of 0. The pore pressure can be determined by solving equation (3), and by calculating the flow rate using equation (2) from that, the permeability, one of the properties of porous materials, can be calculated from the PN model.

[0104] Although the example above describes how to calculate the transmittance, other properties can also be calculated from the PN model by using a calculation formula appropriate for each property. By combining this with the optimization process for the PN model described in the second embodiment, it becomes possible to determine the three-dimensional structure of a porous body that will provide good properties.

[0105] According to the third embodiment of the present invention described above, by performing calculations using a predetermined formula, it is possible to provide a useful technology that can determine the optimal structure early in the design of a porous body without performing AI calculations using a learning model.

[0106] The present invention is not limited to the above-described embodiment, and can be implemented using any components without departing from the spirit of the present invention.

[0107] The above-described embodiments and modifications are merely examples, and the present invention is not limited to these details as long as the features of the invention are not impaired. Furthermore, although various embodiments and modifications have been described above, the present invention is not limited to these details. Other aspects that can be considered within the scope of the technical idea of ​​the present invention are also included within the scope of the present invention. [Explanation of symbols]

[0108] 1...control unit, 2...storage unit, 3...memory, 4...operation input device, 5...display device, 6...bus, 11...learning model creation unit, 12...PN model conversion unit, 13...physical simulation unit, 14...AI calculation unit, 15...characteristic evaluation unit, 16...gene evolution unit, 17...learning model update unit, 18...data conversion unit, 21...porous body structure data, 22...PN model data, 23...learning model data, 24...input parameter data, 25...characteristic data, 26...characteristic evaluation data, 27...3D shape data, 100...porous body design device

Claims

1. A method for designing a porous body using a computer, comprising: a property prediction / calculation process for predicting or calculating the properties of the porous body having a predetermined three-dimensional structure; an evaluation process for evaluating the properties of the porous body predicted or calculated by the property prediction / calculation process; an optimization process for changing the connectivity between pores in the three-dimensional structure to search for an optimal connectivity; A porous body design method for determining a three-dimensional structure of the porous body based on the evaluation results of the properties of the porous body obtained by the evaluation process.

2. The porous body design method according to claim 1, causing the computer to execute a PN modeling process for generating a PN model that represents the three-dimensional structure using a plurality of pores and a plurality of throats that respectively connect the pores; A porous body design method in which the optimization process changes the connectivity by adding or deleting the throat in the PN model generated in the PN modeling process.

3. The porous body design method according to claim 2, In the property prediction / calculation process, the properties of the porous body are predicted or calculated based on the PN model.

4. The porous body design method according to claim 2 or 3, causing the computer to execute a structure generation process for virtually generating, on the computer, a three-dimensional structure of the porous body based on the PN model to which the throat has been added or deleted by the optimization process; The property prediction / calculation process is a porous body design method in which the properties of the porous body having the three-dimensional structure generated by the structure generation process are predicted or calculated.

5. The porous body design method according to claim 4, The property prediction / calculation process is a porous body design method that performs at least one of a physical simulation that calculates the properties of the porous body by physical calculation based on the three-dimensional structure generated by the structure generation process, and an AI calculation that predicts the properties of the porous body using a learning model of the three-dimensional structure.

6. A porous body manufacturing method for manufacturing a porous body using three-dimensional shape data based on the three-dimensional structure of the porous body determined by the porous body design method according to any one of claims 1 to 3.

7. A porous body manufacturing method for manufacturing the porous body using three-dimensional shape data based on the three-dimensional structure of the porous body determined by the porous body design method according to claim 4.

8. A porous body manufacturing method for manufacturing the porous body using three-dimensional shape data based on the three-dimensional structure of the porous body determined by the porous body design method according to claim 5.

Images