Structural design support system, method, and program

Abstract

The present invention provides a structural design support system, method, and program that can optimize multiple parameters more efficiently. [Solution] The present invention is a structural design support system for determining the design values ​​of multiple design items in a structure in order to satisfy multiple required performance characteristics of the structure, using a first model in which at least one of the multiple required performance characteristics is set as the first objective variable and one or more of the multiple design items are set as the first explanatory variables, the required performance characteristics set as the first objective variable are expressed using the constraints of the first explanatory variables, and under the constraints of the expressed first explanatory variables, the remaining required performance characteristics obtained by subtracting the required performance characteristics set as the first objective variable are set as the second objective variables and the multiple design items are set as the second explanatory variables, and the design values ​​of the multiple design items in the structure are determined and output based on a second model.

Description

[Technical Field]

[0001] The present invention relates to a structural design support system, a structural design support method, and a structural design support program that assist in the design of structural structures. [Background technology]

[0002] In recent years, with advances in simulation technology, structures are sometimes designed by determining through simulation whether or not they meet required performance (see, for example, Patent Document 1). In such cases, for example, a model of the structure is generated from physical knowledge, empirical knowledge, or by machine learning, and the design value of the structure that meets the required performance is obtained from the generated model of the structure. If there are design conditions, these design conditions are used as constraints on explanatory variables, and the design value of the structure is obtained based on the model of the structure as an optimization problem. Alternatively, for example, a model of the structure is generated in the same manner as described above, and multiple design values ​​of the structure that meet the required performance are searched for (first simulation) based on the generated model of the structure, and the optimal design value is searched for in more detail (second simulation) from the multiple searched design values, for example, using the finite element method, and selected. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Patent No. 6544006 [Overview of the project] [Problems that the invention aims to solve]

[0004] By the way, when there are multiple performance requirements, optimization must be performed on multiple items, which degrades the efficiency of the optimization process.

[0005] This invention was made in view of the above circumstances, and its purpose is to provide a structural design support system, a structural design support method, and a structural design support program that can optimize multiple items more efficiently. [Means for solving the problem]

[0006] As a result of various studies, the inventors have found that the above objective can be achieved by the present invention as described below. That is, a structural design support system according to one aspect of the present invention is a system for determining the design values ​​of a plurality of design items in a structure in order to satisfy a plurality of required performance characteristics required for the structure, and comprises a constraint processing unit that expresses the required performance characteristics of the first required performance characteristics as constraints of the first explanatory variable by using a first model in which at least one of the plurality of required performance characteristics is set as a first objective variable and one or more of the plurality of design items as first explanatory variables, a design processing unit that determines the design values ​​of a plurality of design items in the structure based on a second model in which, under the constraints of the first explanatory variable expressed by the constraint processing unit, each of the remaining required performance characteristics obtained by removing the required performance characteristics of the first objective variable from the plurality of required performance characteristics is set as a second objective variable and the plurality of design items as second explanatory variables, and an output unit that outputs the design values ​​of the plurality of design characteristics obtained by the design processing unit.

[0007] Such structural design support systems transform required performance into constraints, reducing the number of items to be optimized, thus enabling more efficient optimization of multiple items.

[0008] Furthermore, as described above, the required performance constraints are not limited to one; therefore, the third objective variable described later is the second objective variable in a specific case where there are multiple required performance constraints, and the third model described later is the second model in the aforementioned specific case.

[0009] In another embodiment, the structural design support system described above further includes an input unit that receives input of the first model.

[0010] Since such a structural design support system has an input section, if the first model can be generated based on physical or empirical knowledge, the required performance can be easily converted into constraints.

[0011] In another embodiment, the structural design support system described above further comprises a first model generation unit that generates the first model.

[0012] Such a structural design support system can generate the first model even when it is not possible to generate the first model based on physical knowledge or empirical knowledge, because it includes a first model generation unit.

[0013] In another embodiment, in the above-described structural design support systems, the first objective variable is the item of the multiple required performance characteristics that requires the least amount of information processing. Preferably, in the above-described structural design support systems, the first objective variable is the item of the multiple required performance characteristics that, when sorted in order of least amount of information processing, accounts for less than half of the required performance characteristics in that sorting order. Preferably, in the above-described structural design support systems, the first objective variable is the item of the multiple required performance characteristics that requires the least amount of information processing.

[0014] The amount of information processing required for each of the multiple performance requirements is not necessarily equal and is usually different. That is, some performance requirements require a relatively large amount of information processing to solve, while others require a relatively small amount of information processing to solve. The above structural design support system can optimize multiple items more efficiently by selecting the performance requirement requiring the least amount of information processing as the first objective variable.

[0015] In another embodiment, in the above-described structural design support system, the required performance items designated as the first objective variable are multiple, the first model is multiple, each corresponding to one of the multiple first objective variables, the constraint processing unit expresses the required performance items of the first objective variable as constraints of the first explanatory variable by using the first model corresponding to each of the multiple first objective variables, and the design processing unit, under the multiple constraints of the first explanatory variable expressed by the constraint processing unit, determines the design values ​​of each of the multiple design items in the structure based on a third model in which each of the remaining required performance items obtained by subtracting the multiple required performance items designated as the multiple first objective variables from the multiple required performance items are designated as third objective variables and the multiple design items are designated as third explanatory variables.

[0016] Such structural design support systems can transform multiple performance requirements into multiple constraints, thereby reducing the number of items to be optimized, and thus enabling more efficient optimization of multiple parameters.

[0017] Another aspect of the present invention relates to a structural design support method, which is a method for determining the design values ​​of a plurality of design items in a structure so as to satisfy a plurality of required performance characteristics required for the structure, comprising: a constraint setting process step in which a first model is used in which at least one of the plurality of required performance characteristics is set as a first objective variable and one or more of the plurality of design items are set as first explanatory variables, thereby expressing the required performance characteristics set as the first objective variable with constraints of the first explanatory variable; a design processing step in which, under the constraints of the first explanatory variable expressed in the constraint setting process, the remaining required performance characteristics obtained by removing the required performance characteristics set as the first objective variable from the plurality of required performance characteristics are set as second objective variables and the plurality of design items are set as second explanatory variables, thereby determining the design values ​​of the plurality of design items in the structure based on a second model; and an output step in which the design values ​​of the plurality of design items determined in the design processing step are output.

[0018] Such a structural design support method can change the required performance into a constraint condition and reduce the number of items to be optimized, so that multi-item optimization can be achieved more efficiently.

[0019] A structural design support program according to another aspect of the present invention is a program for causing a computer to function as any of the above-described structural design support systems.

[0020] According to this, a structural design support program can be provided, and this structural design support program exhibits the same operational effects as those of the above-described structural design support systems.

Advantages of the Invention

[0021] The structural design support system, structural design support method, and structural design support program according to the present invention can achieve multi-item optimization more efficiently.

Brief Description of the Drawings

[0022] [Figure 1] It is a block diagram showing the configuration of the structural design support system in the embodiment. [Figure 2] It is a diagram for explaining the constraint condition conversion process of representing the required performance as the first objective variable by the constraint conditions of the first explanatory variable. [Figure 3] It is a flowchart showing the operation of the structural design support system. [Figure 4] As an example, it is a diagram for explaining the design variables (explanatory variables), objective variables, and required performance in the examples. [Figure 5] It is a diagram for explaining the first model (regression model) used in the constraint condition conversion process in the above example. [Figure 6] It is a diagram for explaining the dataset of the first explanatory variable and the first objective variable used for generating the first model (regression model). [Figure 7] It is a diagram for explaining the dataset of the second explanatory variable and the second objective variable in the above example. [Figure 8] As an example, this figure shows the optimization calculation process for the blank area. [Figure 9] As an example, this figure shows the results of comparing the example and the comparative example. [Modes for carrying out the invention]

[0023] Hereinafter, one or more embodiments of the present invention will be described with reference to the drawings. However, the scope of the invention is not limited to the disclosed embodiments. In each figure, components denoted by the same reference numerals are identified as identical components, and their descriptions are omitted where appropriate. In this specification, general reference numerals are used without subscripts, while individual components are indicated by subscripts.

[0024] The structural design support system in this embodiment is a system for determining the design values ​​of multiple design items in a structure in order to satisfy multiple performance requirements for the structure. This structural design support system comprises a constraint processing unit, a design processing unit, and an output unit. The constraint processing unit uses a first model in which at least one of the multiple performance requirements is set as the first objective variable and one or more of the multiple design items are set as the first explanatory variables, thereby expressing the performance requirement set as the first objective variable in terms of constraints of the first explanatory variable. The design processing unit, under the constraints of the first explanatory variable expressed by the constraint processing unit, determines the design values ​​of multiple design items in the structure based on a second model in which each of the remaining performance requirements obtained by subtracting the performance requirement set as the first objective variable from the multiple performance requirements is set as the second objective variable and the multiple design items are set as the second explanatory variables. The output unit outputs the design values ​​of the multiple design values ​​determined by the design processing unit.

[0025] The following provides a more detailed description of such a structural design support system, as well as the structural design support method and structural design support program implemented therein. The structural design support system may be configured by connecting an input / output terminal device for inputting and outputting data, one or more processing units (e.g., a server device) for performing various calculations, and one or more database devices for storing (managing) various data, so that they can communicate with each other. At least some of these input / output terminal devices, one or more processing units, and one or more database devices may be configured as a single unit and connected to the rest so that they can communicate with each other. Here, however, the structural design support system will be described using a structural design support system in which all components are integrated as an example.

[0026] Figure 1 is a block diagram showing the configuration of a structural design support system (a structural design support device as an example) in an embodiment. Figure 2 is a diagram illustrating the constraint processing in which the required performance, which is the first objective variable, is expressed using the constraints of the first explanatory variable. In Figure 2, the horizontal axis represents the first explanatory variable, and the vertical axis represents the first objective variable.

[0027] The structural design support system (a structural design support device as an example) 1000 in the embodiment is a device that determines the design values ​​of multiple design items in a structure in order to satisfy multiple performance requirements for the structure, and for example, as shown in Figure 1, it comprises a control processing unit 1, an input unit 2, an output unit 3, an interface unit (IF unit) 4, and a storage unit 5.

[0028] A structure is an article that is composed of multiple components, such as extruded parts, pressed parts, forged parts, and machined parts, which are parts made by processing a single component.

[0029] The input unit 2 is connected to the control processing unit 1 and is a device that inputs various commands, such as a command to instruct the start of design, and various data necessary for operating the structural design support device 1000, such as the design name, design conditions (original constraints), and required performance. For example, it may be a keyboard, mouse, or multiple input switches assigned to predetermined functions. The output unit 3 is connected to the control processing unit 1 and is a device that outputs commands, data, and calculation results input from the input unit 2 according to the control of the control processing unit 1. For example, it may be a display device such as a CRT display, LCD (liquid crystal display device), or organic EL display, or a printing device such as a printer.

[0030] The input unit 2 and output unit 3 may be configured as touch panels. In this configuration, the input unit 2 is a position input device that detects and inputs the operating position, such as a resistive or capacitive touch panel, and the output unit 3 is a display device. In this touch panel, a position input device is provided on the display surface of the display device, and one or more candidate input contents that can be input to the display device are displayed. When the user touches the display position that displays the input content they want to input, the position input device detects that position, and the display content displayed at the detected position is input to the structural design support device 1000 as the user's operation input. With such a touch panel, the user can easily understand the input operation intuitively, thus providing a structural design support device 1000 that is easy for the user to use.

[0031] The IF unit 4 is connected to the control processing unit 1 and, in accordance with the control of the control processing unit 1, is a circuit that inputs and outputs data to and from external devices, for example. Examples include an RS-232C serial communication interface circuit, an interface circuit using the Bluetooth® standard, and an interface circuit using the USB standard. Alternatively, the IF unit 4 may be a communication interface circuit that sends and receives communication signals to and from external devices, such as a data communication card or a communication interface circuit conforming to the IEEE 802.11 standard.

[0032] The memory unit 5 is connected to the control processing unit 1 and is a circuit that stores various predetermined programs and various predetermined data in accordance with the control of the control processing unit 1.

[0033] The various predetermined programs mentioned above include, for example, a control processing program, which includes, for example, a control program, a constraint processing program, and a design processing program. The control program is a program that controls each of the parts 2 to 5 of the structural design support device 1000 according to the function of each part. The constraint processing program is a program that expresses the required performance item of the first objective variable as a constraint of the first explanatory variable by using a first model in which at least one required performance item from a plurality of required performance items is the first objective variable and one or more design items from a plurality of design items are the first explanatory variables. The design processing program is a program that, under the constraint of the first explanatory variable expressed in the constraint processing program, determines the design value of each of the plurality of design items in the structure based on a second model in which each of the plurality of design items is the second objective variable, and each of the remaining required performance items obtained by removing the required performance item of the first objective variable from the plurality of required performance items is the second explanatory variable.

[0034] The aforementioned various predetermined data include, for example, data necessary for executing each of these programs, such as the design name, design conditions (original constraints), and required performance.

[0035] Such a storage unit 5 may include, for example, a non-volatile memory element such as ROM (Read Only Memory) or a rewritable non-volatile memory element such as EEPROM (Electrically Erasable Programmable Read Only Memory). Furthermore, the storage unit 5 includes RAM (Random Access Memory) which serves as the working memory of the control processing unit 1, storing data generated during the execution of the predetermined program. The storage unit 5 may also be configured to include a hard disk drive (HDD) or solid-state drive (SSD) with a relatively large storage capacity.

[0036] The control processing unit 1 is a circuit that controls each part 2 to 5 of the structural design support device 1000 according to the function of each part, and determines the design values ​​for each of the multiple design items in the structure in order to satisfy the multiple performance requirements of the structure. The control processing unit 1 is configured, for example, with a CPU (Central Processing Unit) and its peripheral circuits. When the control processing program is executed, the control unit 11, the constraint processing unit 12, and the design processing unit 13 are functionally configured in the control processing unit 1.

[0037] The control unit 11 controls each of the structural design support device 1000 parts 2 to 5 according to the function of each part, and is in charge of the overall control of the structural design support device 1000.

[0038] The constraint-condition processing unit 12 represents the required performance with the item as the first objective variable among at least one of the plurality of required performances using a first model with one or more design items among the plurality of design items as the first explanatory variables. The first model is generated, for example, by an operator (user) from physical knowledge, empirical knowledge, etc., input from the input unit 2 into the structural design support apparatus 1000, and stored in the storage unit 5. In one example, when the required performance Y is less than a predetermined condition value C and the first model is a function f(x) with one or more design items among the plurality of design items as the first explanatory variable x (Y < C, Y = f(x), when the set with the plurality of design items di1, di2, di3,... as elements is X (= {di1, di2, di3,...}) and x ∈ X), the inverse function f -1 (Y) of the function f(x) is obtained, and by rewriting the required performance f(x) < C as x < f -1 (C), it is represented by the constraint conditions of the first explanatory variable x.

[0039] If solving while keeping the required performance f(x) < C, as shown in FIG. 2, it is necessary to search over the entire explanatory variable X. However, by rewriting the required performance f(x) < C as x < f -1 (C), as shown in FIG. 2, it is only necessary to search with x < f -1 (C). Thus, the search range is narrowed, and multi-item optimization can be achieved more efficiently.

[0040] Here, the first objective variable is an item of the required performance with a small amount of information processing among the plurality of required performances. That is, among the plurality of required performances, an item of the required performance with a small amount of information processing is selected as the first objective variable. Preferably, the first objective variable is an item of the required performance that is less than half of the plurality of required performances in the order of sorting the required performances in ascending order of the amount of information processing. Preferably, the first objective variable is an item of the required performance with the minimum amount of information processing among the plurality of required performances.

[0041] Each amount of information processing required for a plurality of required performances is not necessarily equal and is usually different. That is, there are required performances that require a relatively large amount of information processing for solution, and there are also required performances that require a relatively small amount of information processing for solution. By selecting a required performance with a small amount of information processing as the first objective variable, the above-mentioned structure design support device 1000 can more efficiently optimize multiple items.

[0042] Based on a second model in which the design processing unit 13 uses each item of the remaining required performances obtained by removing the required performance with the item of the first objective variable from the plurality of required performances as the second objective variables and the plurality of design items as the second explanatory variables under the constraint conditions of the first explanatory variables represented by the constraint condition processing unit 12, each design value of the plurality of design items in the structure is obtained. For example, when a plurality of required performances are Y1 < C1, Y2 < C2, Y3 < C3, ···, a plurality of design items are di1, di2, di3, ···, and Y1 is the first objective variable and the first model is Y1 = f(x), the design processing unit 13, under the constraint condition x < f -1 (C1), from the plurality of required performances Y1 < C1, Y2 < C2, Y3 < C3, ···, each item Y2, Y3, ··· of the remaining required performances Y2 < C2, Y3 < C3, ··· obtained by removing the required performance Y1 < C1 with the item of the first objective variable Y1 is used as the second objective variable, and based on a second model h(X) in which a plurality of design items di1, di2, di3, ··· are used as the second explanatory variable X, each design value of the plurality of design items di1, di2, di3, ··· in the structure is obtained.

[0043] For obtaining each design value, since each design value is obtained under constraint conditions, it becomes an optimization problem, and for example, known methods such as a direct search method or a sequential approximation optimization method are used.

[0044] The aforementioned direct search method is an optimization algorithm that directly searches only the explanatory variables of multiple design items and their objective function values. Examples of metaheuristic methods include the genetic algorithm. For example, in the genetic algorithm, the number of individuals in each generation (generational population) is set in advance, and when generating individuals for the next generation from the current generation, firstly, a predetermined number of individuals (selected individuals) with high evaluation values ​​according to a predetermined evaluation function are selected from all individuals in the current generation, and these selected individuals become the individuals for the next generation (so-called tournament selection method). Secondly, a number of individuals obtained by subtracting the selected individuals from the generational population are generated with a predetermined probability through genetic operations such as crossover, mutation, and copying to satisfy the constraints, and these generated individuals become the individuals for the next generation. Thirdly, the current generation is replaced with the next generation thus generated. Fourthly, each of these first to third processes is carried out until the predetermined number of generations is reached, or until the difference between generations is substantially eliminated, and the individual with the highest evaluation value from among the individuals of the final generation is selected as the solution. Crossover is a process of selecting two individuals and exchanging some of the constituent elements (genes) between them. Mutation is a process of selecting one individual and changing some of the constituent elements (genes) that make up that individual. Copying is a process of selecting one individual and using that selected individual as is for the next generation. In addition to the tournament selection method, the genetic algorithm may also include a roulette selection method or a ranking selection method, and any of these may be used.

[0045] The aforementioned successive approximation optimization method is a method for searching for the optimal solution by first generating a proxy model of the objective function that derives the objective variable from the multiple explanatory variables from a dataset containing multiple datasets of explanatory variables and their objective variables for multiple design items, second obtaining the optimal solution of the proxy model, and third obtaining the value of the objective variable based on this obtained solution, for example by the finite element method, and adding it to the dataset, and then sequentially repeating each of these first to third processes. For generating the first dataset, for example, a simple random sampling method that randomly extracts data from the space of design variables, or the Latin hypercube sampling method, which is a type of experimental design method, can be used.

[0046] More specifically, first, a plurality of required performances required for the structure, a plurality of design items in the structure, each numerical range that each design value of the plurality of design items can take, the required performance as the item of the first objective variable, and the numerical range of the value to be achieved for each of the plurality of required performances are set by an operator (user), and these are input to the structural design support apparatus 1000 and stored in the storage unit 5. The design processing unit 13, for each item of the remaining required performances excluding the required performance as the item of the first objective variable from the plurality of required performances, and for the plurality of design items, generates data with the plurality of design items as the second explanatory variables by, for example, a simple random sampling method or a Latin hypercube sampling method, and uses the finite element method as the second model to obtain each numerical value of each item of the remaining required performances, and generates a plurality of data sets each having data with this as the second objective variable. Subsequently, the design processing unit 1 uses the data set to generate a surrogate model of the objective function for obtaining the second objective variable from the second explanatory variables, and obtains an optimal solution of the surrogate model (a candidate (numerical value of the explanatory variable) predicted to satisfy the required performance with the surrogate model) by a constrained optimization method while satisfying the constraint conditions. Examples of the constrained optimization method include an optimal problem with equality constraints such as the Lagrange multiplier method, the Lagrange dual problem, and the optimization of dual variables, and inequality constraint optimization problems such as conversion to an unconstrained optimization problem (barrier method, penalty method, projected gradient method) and the dual problem. Subsequently, the design processing unit 13 obtains the numerical value of the second objective variable by the finite element method as the second model based on the obtained candidate, and evaluates whether or not the obtained numerical value of the second objective variable satisfies the required performance. Then, as a result of the evaluation, when the design processing unit 13 satisfies the required performance, it ends the process with the obtained candidate as the solution (each design value of the plurality of design items), and when it does not satisfy the required performance, it repeats the addition of the candidate point to the data set, the generation of the second model, the search for the candidate, and the evaluation of the required performance until the required performance is satisfied.

[0047] Note that in the above description, one required performance Y1 < C1 is used as a constraint condition x < f -1Although an example of changing to (C1) has been described, the required performance that is the first objective variable may be plural. In this case, there are a plurality of the first models corresponding to each of the plurality of first objective variables. The constraint-conditioning processing unit 12 represents, for each of the plurality of first objective variables, the required performance with the item of the first objective variable in the constraint conditions of the first explanatory variable by using the first model corresponding to the first objective variable. Then, the design processing unit 13, under the plurality of constraint conditions of the first explanatory variable represented by the constraint-conditioning processing unit 12, uses, as the third objective variable, each item of the remaining required performance obtained by excluding the plurality of required performances with the items of the plurality of first objective variables from the plurality of required performances, and uses the plurality of design items as the third explanatory variables, and obtains each design value of the plurality of design items in the structure based on a third model. For example, when a plurality of required performances are Y1 < C1, Y2 < C2, Y3 < C3, ···, a plurality of design items are di1, di2, di3, ···, and the first objective variables are Y1 and Y2, and the first models are Y1 = f(x) and Y2 = j(x), the design processing unit 13 has constraint conditions x < f -1 (C1) and constraint condition x < j -1 (C2), and under the plurality of required performances Y1 < C1, Y2 < C2, Y3 < C3, ···, from the plurality of required performances, for each item Y3, ··· of the remaining required performance obtained by excluding the required performances Y1 < C1 and Y2 < C2 with the items Y1 and Y2 of the first objective variables as the third objective variables, and using the plurality of design items di1, di2, di3, ··· as the third explanatory variables X, each design value of the plurality of design items di1, di2, di3, ··· in the structure is obtained based on a third model g(x).

[0048] The control unit 11 outputs each design value of the plurality of design items obtained by the design processing unit 13 to the output unit 3, and causes the output unit 3 to output the plurality of design values to the outside. The control unit 11 may output the numerical value of the item of the required performance (the numerical value of the objective variable) to the output unit 3 and cause the output unit 3 to output it to the outside.

[0049] The control processing unit 1, the input unit 2, the output unit 3, the IF unit 4, and the storage unit 5 in such a structure design support device 1000 can be configured by, for example, a computer such as a desktop type or a notebook type.

[0050] Next, the operation of this embodiment will be described. Figure 3 is a flowchart showing the operation of the structural design support system.

[0051] When the structural design support device 1000 with this configuration is powered on, it performs the initialization of each necessary part and starts operating. The control processing unit 1 is functionally configured with a control unit 11, a constraint processing unit 12, and a design processing unit 13 through the execution of its control processing program.

[0052] In Figure 3, the user inputs the necessary data for determining the design values ​​of multiple design items in the structure, such as the multiple required performance characteristics required for the structure, the multiple design items, and the required performance characteristics of the first objective variable, into the structure design support device 1000 and stores them in the storage unit 5 (S1).

[0053] Next, the user inputs the first model into the structural design support device 1000 and stores it in the storage unit 5 (S2).

[0054] Next, the structural design support device 1000, using the constraint condition processing unit 12 of the control processing unit 1, expresses the required performance items of the first objective variable as constraint conditions of the first explanatory variable using the first model, and stores it in the storage unit 5 (S3).

[0055] Next, the structural design support device 1000, using the design processing unit 13 of the control processing unit 1, determines the design values ​​for each of the multiple design items in the structure based on a second model in which each of the remaining required performance items obtained by subtracting the required performance items of the first objective variable from the multiple required performance items is used as the second objective variable, and the multiple design items are used as the second explanatory variables, under the constraints of the first explanatory variable expressed by the constraint processing unit 12 (S4).

[0056] Then, the structural design support device 1000, via the control unit 11 of the control processing unit 1, outputs each design value of the multiple design items obtained by the design processing unit 13 to the output unit 3, and outputs the multiple design values ​​to the outside via the output unit 3 (S5), thus ending this process. The control unit 11 may also output each design value to an external device via the IF unit 4 if necessary.

[0057] As described above, the structural design support system (a structural design support device as an example) 1000 in the embodiment, as well as the structural design support method and structural design support program implemented therein, change the required performance into constraints and reduce the number of items to be optimized, thereby enabling more efficient optimization of multiple items.

[0058] Since the above-mentioned structural design support system (structural design support device) 1000, structural design support method, and structural design support program are equipped with an input unit 2, if the first model can be generated based on physical knowledge or empirical knowledge, the required performance can be easily converted into constraints.

[0059] In the above-described embodiment, the first model was generated by the user and input from the input unit 2. However, the structural design support device 1000 may further include a first model generation unit 14 for generating the first model, as shown by the dashed line in Figure 1. Such a structural design support device 1000 can generate the first model even when the first model cannot be generated based on physical knowledge or empirical knowledge, because it includes the first model generation unit 14. For example, the first model generation unit 14 generates the first model using machine learning with training data. The training data is obtained by generating data within the numerical range of each design item of the first explanatory variable using a sampling method from the perspective of statistical analysis, such as random sampling, and calculating the value of the first objective variable corresponding to the generated first explanatory variable. Since the amount of information processing required for the first objective variable is small, the computational load for obtaining the training data is reduced. When the first model is generated by machine learning, the storage unit 5 functionally further includes a learning information storage unit 51 for storing the learning dataset used for machine learning of the first model, as shown by the dashed line in Figure 1.

[0060] An example will be described. Figure 4 is a diagram illustrating the design variables (explanatory variables), objective variable, and required performance in an example. Figure 5 is a diagram illustrating the first model (regression model) used in the constraint processing in the above example. Figure 6 is a diagram illustrating the dataset of the first explanatory variable and the first objective variable used to generate the first model (regression model). Figure 7 is a diagram illustrating the dataset of the second explanatory variable and the second objective variable in the above example. Figure 8 is a diagram illustrating the optimization calculation progress for the blank area as an example. Figure 9 is a diagram illustrating the comparison results between the example and the comparative example.

[0061] In this embodiment, when manufacturing a pillar, which is a structural component of a vehicle, by press forming, a roughly curved quadrilateral blank shape (an example of a structure) is designed, as shown in Figure 4. In this design, both ends of the lower edge of the blank are fixed, and on the lower edge, four nodes Node_1 to Node_4 are set at an appropriate interval between the two ends. On the upper edge, node Node_5 is set at one end and node Node_9 is set at the other end, and three nodes Node_6 to Node_8 are set at an appropriate interval between these two nodes Node_5 and Node_9. The x-coordinates (Node_1x to Node_9x) at each node Node_1 to Node_9 are fixed values, and the y-coordinates (Node_1y to Node_9y) at each node Node_1 to Node_9 and the fillet radius size (Node_2R) at one location are design items. The objective variables are sway, camber, torsion, and blank area, and the required performance is minimization of sway, camber, and torsion, and a blank area of ​​1.4 times or less the minimum blank area. Sway refers to a defect in in-plane wobbling deformation relative to the desired part shape, camber refers to a defect in out-of-plane wobbling deformation relative to the desired part shape, and torsion refers to a defect in torsional deformation relative to the desired part shape. The minimum blank shape refers to the smallest blank shape that allows the part shape after press forming to be formed, and the area in this case is the minimum blank area. The maximum blank shape refers to the blank shape that is enlarged to the maximum extent possible from the viewpoint of material yield and mold structure, and the area in this case is the maximum blank area.

[0062] For the first model, a regression model is used. A learning dataset for learning the first model shown in FIG. 6 is prepared. The first model generation unit 14 learns using this learning data, and the first model (regression model) shown in FIG. 5 is generated. Among the items of required performance as the first objective variable, the blank area is selected because it can be relatively easily calculated based on the coordinates of each of the nine nodes Node_1 to Node_9 and the fillet R size as compared with sway, camber, and twist respectively. That is, the information processing cost (amount of information processing) for the simulation of various dimensional accuracies is higher (more) than the information processing cost (amount of information processing) for the calculation of the area of the blank before press working. The data of the learning dataset includes the y coordinates of the design items (the first explanatory variables of the first model); Node_1y to Node_9y and the fillet R size at one location; Node_2R and the blank area (the first objective variable of the first model) as elements, and these are associated with each other. The learning dataset includes 240 pieces of such data, 168 of which are used for learning the first model, and the remaining 72 pieces of data are used for verifying (testing) the first model generated by the first model generation unit 14. The generated first model has a correlation coefficient R 2 of 1, and the first objective variable can be well estimated from the first explanatory variables.

[0063] Each data of the dataset of the second explanatory variable and the second objective variable generated by the design processing unit 13 includes, for example, as shown in FIG. 7, the y coordinates of the design items (the second explanatory variables of the second model); Node_1y to Node_9y and the fillet R size at one location; Node_2R and sway, camber, and twist (the second objective variables of the second model) as elements, and these are associated with each other. The design processing unit 13 obtains the design values of sway, camber, and twist by the sequential approximation optimization method using such a dataset under the above constraints where the constraints of the second explanatory variables of the second model are replaced by the first model.

[0064] Figure 8 shows the calculation progress in the successive approximation optimization method along with the comparative example. The horizontal axis of Figure 8 represents the number of searches, and the vertical axis represents the ratio of the blank area to the minimum blank area (blank area ratio). In the comparative example, the required performance condition that the blank area is 1.4 times or less of the minimum blank area is not used as a constraint condition for the second explanatory variable, but is used directly as the required performance condition. Therefore, the second objective variables in the comparative example are sway, camber, torsion, and blank area. As can be seen from Figure 8, in the comparative example, points that do not satisfy the required performance condition that the blank area is 1.4 times or less of the minimum blank area are also considered as candidate points for the successive approximation optimization method, whereas in the example, points that almost always satisfy the required performance condition that the blank area is 1.4 times or less of the minimum blank area are considered as candidate points for the successive approximation optimization method, enabling efficient optimization of multiple items.

[0065] Figure 9 shows the comparison results of sway, camber, and torsion for each calculated design value. The horizontal axis of Figure 9 represents the evaluation index for sway, camber, and torsion, and the vertical axis represents the value of each evaluation index [mm, deg, -]. In Figure 9, each design value with equivalent blank area ratio is extracted, and the distance from the target 0 [mm] for sway (sway distance), the distance from the target 0 [deg] for camber (camber distance), the distance from the target 0 for torsion (torsion distance), and the distance from the target (0, 0, 0) for these three evaluation indices (total distance) are shown. As can be seen from Figure 9, the example is superior to the comparative example in terms of sway distance, camber distance, and total distance, and in terms of sway, camber, and the total of the three evaluation indices, the example is closer to the target and yields a good solution.

[0066] To illustrate the present invention, the embodiments have been adequately and fully described above with reference to the drawings. However, those skilled in the art should recognize that it is easy to modify and / or improve upon the embodiments described above. Therefore, unless such modifications or improvements implemented by those skilled in the art fall outside the scope of the claims, such modifications or improvements shall be considered to be included within the scope of the claims. [Explanation of symbols]

[0067] 1000 Structural Design Support System (An example of a structural design support device) 1 Control Processing Unit 2 Input section 3. Output section 4. Interface section (IF section) 5 Storage section 11 Control Unit 12. Constraint Condition Processing Unit 13 Design Processing Department 14. First Model Generation Unit 51 Learning Information Storage Unit

Claims

1. A structural design support system that determines the design values ​​for multiple design items in a structural structure in order to satisfy multiple performance requirements for the structure, A constraint processing unit is provided that uses a first model in which at least one of the multiple required performance items is the first objective variable and one or more of the multiple design items are the first explanatory variables, thereby expressing the required performance item of the first objective variable in terms of the constraints of the first explanatory variables. A design processing unit determines the design values ​​for each of the multiple design items in the structure based on a second model in which, under the constraints of the first explanatory variable expressed by the constraint processing unit, each of the remaining required performance items obtained by removing the required performance items of the first objective variable from the multiple required performance items is set as the second objective variable, and the multiple design items are set as the second explanatory variables. The system includes an output unit that outputs each of the multiple design values ​​obtained by the design processing unit. Structural design support system.

2. The first model further includes an input unit that receives input, A structural design support system according to claim 1.

3. The system further comprises a first model generation unit that generates the first model, A structural design support system according to claim 1.

4. The first objective variable is the item among the multiple required performance items that requires the least amount of information processing. A structural design support system according to claim 1.

5. The required performance, which is considered the first objective variable, is multifaceted. The first model is a plurality of models corresponding to each of the plurality of first objective variables, The constraint processing unit expresses the required performance items of the first objective variable as constraints of the first explanatory variable by using a first model corresponding to each of the plurality of first objective variables. The design processing unit, under the constraints of the first explanatory variable expressed by the constraint processing unit, determines the design values ​​for each of the multiple design items in the structure based on a third model in which each of the remaining required performance items obtained by subtracting the multiple required performance items that were set as the multiple first objective variable items from the multiple required performance items are set as the third objective variable, and each of the multiple design items are set as the first explanatory variable. A structural design support system according to claim 1.

6. A structural design support method for determining the design values ​​of multiple design items in a structural structure in order to satisfy multiple performance requirements for the structure, A constraint setting process is performed by using a first model in which at least one of the multiple required performance items is the first objective variable and one or more of the multiple design items are the first explanatory variables, thereby expressing the required performance items of the first objective variable in terms of the constraints of the first explanatory variables. A design processing step to determine the design values ​​of the multiple design items in the structure based on a second model in which, under the constraints of the first explanatory variable expressed in the constraint processing step, each of the remaining required performance items obtained by removing the required performance items designated as the first objective variable from the multiple required performance items are designated as the second objective variable, and the multiple design items are designated as the second explanatory variables. The system includes an output step that outputs each of the multiple design values ​​obtained in the design processing step. Structure design support method.

7. A structural design support program for causing a computer to function as a structural design support system according to any one of claims 1 to 5.

Images