Structure model search method, structure model search system, and structure model search program

Abstract

A structure model search method, a structure model search system, and a structure model search program are provided that are capable of searching for a more appropriate structure model. [Solution] A structure model search method includes a search process using a genetic algorithm, which generates a plurality of structure models corresponding to a structure and searches for a structure model from the plurality of generated structure models that has a fitness superior to other structure models, wherein the structure models include information indicating the shape and position of each of a plurality of components that make up the corresponding structure, and the fitness is calculated based on a first target variable of the corresponding structure model and a design constraint condition penalty of the corresponding structure model, and the design constraint condition penalty includes a per-model penalty calculated for each of the plurality of structure models generated in the search process.

Description

[Technical Field]

[0001] The present disclosure relates to a structure model search method, a structure model search system, and a structure model search program. [Background technology]

[0002] BACKGROUND ART Frame structures have been designed in the past. Patent Document 1 discloses the design of a building frame having structural members including at least columns, beams, and bearing walls using a genetic algorithm. Patent Document 1 also discloses a first target variable that is the cost of the frame, a second target variable that is the adaptability to design constraints, and adding the second target variable to the first target variable as a penalty. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2014-153994 Summary of the Invention [Problem to be solved by the invention]

[0004] However, Patent Document 1 leaves room for improvement in optimization.

[0005] An object of the present disclosure is to provide a structure model search method, a structure model search system, and a structure model search program that are capable of searching for a more appropriate structure model. [Means for solving the problem]

[0006] A structure model search method according to one aspect of the present disclosure includes a search step using a genetic algorithm, generating a plurality of structure models corresponding to a structure, and selecting a structure model having a fitness from among the plurality of generated structure models. a search step for searching for the structure model that is superior to other structure models, the structure model including information indicating the shape and position of each of a plurality of members that constitute the corresponding structure, the fitness calculated based on a first target variable of the corresponding structure model and a design constraint condition penalty of the corresponding structure model, and the design constraint condition penalty includes a per-model penalty calculated for each of the plurality of structure models generated in the search step. [Effects of the Invention]

[0007] According to the present disclosure, it is possible to provide a structure model search method, a structure model search system, and a structure model search program that are capable of searching for a more appropriate structure model. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 illustrates a jacket structure model search system according to an embodiment of the present disclosure. [Figure 2] FIG. 10 is a diagram showing an example of a jacket. [Figure 3] FIG. 1 is a diagram illustrating a genetic algorithm. [Figure 4] FIG. 1 is a diagram for explaining uniform crossover in a genetic algorithm. [Figure 5] FIG. 10 is a flowchart showing an example of the operation of the jacket structure model search system according to the present embodiment. [Figure 6] FIG. 10 is a flowchart showing an example of the operation of the jacket structure model search system according to the present embodiment. [Figure 7] FIG. 10 is a flowchart showing an example of the operation of the jacket structure model search system according to the present embodiment. [Figure 8] FIG. 10 is a flowchart showing an example of the operation of the jacket structure model search system according to the present embodiment. [Figure 9] FIG. 10 is a diagram showing a jacket structure model searching system according to a first modified example of the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] Next, a jacket structure model search method (structure model search method), a jacket structure model search system (structure model search system), and a jacket structure model search program (structure model search program) according to the present embodiment will be described with reference to the drawings. The embodiments described below are merely examples, and embodiments to which the present disclosure is applied are not limited to the following embodiments. In all the drawings for explaining the embodiments, the same reference numerals are used for components having the same functions, and repeated explanations will be omitted. Furthermore, in this application, "based on XX" means "based on at least XX," and includes cases where it is based on other elements in addition to XX. Furthermore, "based on XX" is not limited to cases where XX is used directly, but also includes cases where it is based on XX that has been calculated or processed. "XX" is any element (for example, any information).

[0010] (Embodiment) (Outline of the jacket structure model search system) FIG. 1 is a diagram illustrating a jacket structure model search system according to an embodiment of the present disclosure. FIG. 2 is a diagram illustrating an example of a jacket targeted by the jacket structure model search system. An example of a jacket structure is a jacket used in a port pier. An example of a jacket structure is a structure in which an upper girder UG is supported by piles stiffened with legs L and braces B. The jacket structure is constructed by inserting steel pipe piles driven into the seabed into the legs (legs L) of a space truss structure assembled from steel pipes, and integrating the piles and legs L with welding or grout. Legs L include thickened legs (cans C) to resist punching shear at the joints with braces B, and other legs (general sections). Cans C are sections at steel pipe joints that resist punching shear from intersecting members. Cans C are often thicker than the general sections. Generally, legs L and piles are integrated on-site by filling the gaps between them with grout, such as cement paste. Of the piles fixed to the seabed, the portion underground is called an underground pile SP, and the portion underwater is called an underwater pile UP. However, the structures targeted by the structure model search method, structure model search system, and structure model search program according to this embodiment are not limited to jacket structures. Examples of target structures include buildings with frameworks (framed structures). Examples of framework structures include high-rise buildings, pile-type piers, bridges, and steel towers (e.g., power transmission towers). Returning to Figure 1, we will continue the explanation.

[0011] The jacket structure model search system 100 according to this embodiment automatically optimizes a jacket structure model. The jacket structure model includes information about multiple components that make up the jacket structure. The jacket structure model includes information indicating the shape and position of each of the multiple components that make up the corresponding jacket structure. Furthermore, in this embodiment, the jacket structure model includes information indicating the material of each of the multiple components that make up the corresponding jacket structure. Note that the jacket structure model may also include information indicating at least one of the shape, position, and material of each of the multiple components. For example, the jacket structure model search system 100 optimizes the shape and material of the components, i.e., the jacket structure model.

[0012] Here, the information about the multiple members included in the jacket structure model is, for example, <1> Information about the type of material involved; <2> Information about the location of the relevant components; <3> Information about the length of the relevant member; <4> Information about other components to which the component is connected; <5> It also includes information about the steel materials (components) used for the relevant components.

[0013] <1> The information about the type is information that identifies whether the relevant member is, for example, a pile, an upper girder, a brace, or a leg (can). <2> The position information is information that identifies the position of the corresponding component, and may be, for example, the position coordinates of both end edges of the corresponding component. <3> The length information is information that specifies the length of the corresponding component. The length information may be, for example, the distance between both ends of the corresponding component calculated from the position coordinates of the both ends. <4> The information about the other connected components is information for identifying the other components connected to the end of the relevant component. <5> The information about the steel material is information that identifies the type of steel material used for the corresponding component. The information about the steel material may be, for example, a component number in a component list described below. The information about the steel material may include, for example, shape information, material quality information, and cost information.

[0014] Optimizing the jacket structure model as described above and optimizing the shapes and materials of the members can be achieved by, for example, optimizing the following information about the plurality of members: <5> The goal is to optimize information about steel materials.

[0015] Here, the above <5> For example, a component list such as that shown below is preferably used for the information on steel materials. The component list is a database of steel materials (components) used in the jacket structure. The component list includes component numbers, shape information, material information, and cost information. The component numbers are, for example, identification numbers assigned to steel materials (components) registered in the component list. The shape information, material information, and cost information are linked to the component numbers. The shape information is, for example, information about the cross section of the corresponding steel material (component). The material information is, for example, information about the material of the corresponding steel material (component). The cost information is, for example, information about the cost per unit weight of the corresponding steel material (component) (unit price of steel material) and information about the processing costs by part (processing labor, unit processing cost).

[0016] By optimizing the jacket structure model described above, <5> Optimizing the information about steel materials means, for example, optimizing the steel materials selected from the material list for each of a plurality of components.

[0017] An example of optimization theory is the genetic algorithm (GA). A genetic algorithm is an algorithm that searches for an approximate solution. It involves preparing multiple individuals that represent data using genes and repeatedly crossing them over (for example, preferentially selecting individuals with low fitness (good evaluation values ​​of the objective function) for crossover). Whether a genetic algorithm is used in crossover is decided by a roulette wheel. A genetic algorithm can also randomly rewrite genes through mutation, and the probability of mutation occurring can be set arbitrarily. Genetic algorithms are mathematically highly versatile because they can be applied to problems that are not differentiable. Furthermore, genetic algorithms are versatile in their range of application because they can be applied to any original problem, such as combinatorial optimization problems.

[0018] The genetic algorithm includes the following search conditions and constraints: The search condition is, for example, a method for calculating fitness, and the number of generations for which crossover is performed (the number of times a search step, which will be described later, is performed). An example of the constraints is a physical constraint imposed on the jacket structure. The constraints include, for example, the stress occurring in each of one or more members, the displacement occurring in each of one or more members, and the bearing capacity of each pile among the one or more members. These constraints are required for calculating the constraint penalty, which will be described later.

[0019] (Configuration of jacket structure model search system) The jacket structure model search system 100 includes an input unit 102, a first jacket structure model generation unit 104, a second jacket structure model generation unit 106, ..., an Nth (N is an integer greater than 1) jacket structure model generation unit, a search unit 108, a storage unit 110, and an output unit 112. In Fig. 1, as an example, the first jacket structure model generation unit 104 and the second jacket structure model generation unit 106 are shown among the first jacket structure model generation unit 104 to the Nth jacket structure model generation unit.

[0020] The input unit 102 inputs information. As an example, the input unit 102 may have an operation unit such as a keyboard and a mouse. In this case, the input unit 102 inputs information according to an operation performed by a user on the operation unit. As another example, the input unit 102 may input information from an external device. The external device may be, for example, a portable storage medium. Some or all of the search conditions, the component list, and the constraints may be input to the input unit 102.

[0021] The storage unit 110 is realized by a hard disk drive (HDD), flash memory, random access memory (RAM), read only memory (ROM), etc. The storage unit 110 stores a jacket structure model. The storage unit 110 may store some or all of a component list, search conditions, and constraint conditions.

[0022] The first jacket structure model generation unit 104 acquires one or more component lists from the input unit 102 or the storage unit 110. Using the acquired one or more component lists, the first jacket structure model generation unit 104 generates a first jacket structure model including a value of the shape of each of the one or more components and a value of the material of each of the one or more components. The second jacket structure model generation unit 106 and the Nth jacket structure model generation unit perform the same processing as the first jacket structure model generation unit 104.

[0023] Here, the member list includes initial values ​​corresponding to the shapes of each of the plurality of members constituting the jacket structure, and initial values ​​corresponding to the materials of each of the plurality of members.

[0024] Examples of components are piles, upper girders, braces, and legs (cans).

[0025] An example of the initial values ​​corresponding to the shape is shown below for each member. An example of an initial value corresponding to the shape is information indicating the cross-sectional specifications of a pile, which includes diameter information, plate thickness information, and shape information. The shape information is information that specifies whether the pile is a hollow circle or an H-shaped steel, for example. An example of an initial value corresponding to a shape is information indicating the cross-sectional specifications of a girder, and the information indicating the cross-sectional specifications of a girder includes girder height information, flange width information, and plate thickness information. An example of an initial value corresponding to a shape is, for a brace, information indicating the cross-sectional specifications of the brace, and the information indicating the cross-sectional specifications of the brace includes diameter information and plate thickness information. An example of the initial value corresponding to the shape is information indicating the cross-sectional specifications of a leg (can), and the information indicating the cross-sectional specifications of a brace includes diameter information and plate thickness information.

[0026] An example of the initial value corresponding to the material is information indicating the steel type of the member.

[0027] The search unit 108 acquires search conditions and constraint conditions from the input unit 102 or the storage unit 110. The search unit 108 acquires one first jacket structure model, one second jacket structure model, one third jacket structure model, one fourth jacket structure model, ..., one Nth jacket structure model from the first jacket structure model generation unit 104, the second jacket structure model generation unit 106, ..., the Nth jacket structure model generation unit. The search unit 108 searches for a jacket structure model that is more economical than the other jacket structure models from the acquired first to Nth jacket structure models.

[0028] Being more economical than other jacket structure models means, for example, that the total cost, which is determined based on the quantity of multiple components, is lower than that of other jacket structure models. The total cost may be calculated, for example, by adding up the steel cost and the processing cost. The steel cost may be calculated, for example, by adding up the product of the unit steel price of each component and the quantity of the components for each jacket structure model. The processing cost may be calculated, for example, by adding up the product of the weight of each component for each part, the processing labor per unit weight for each part, and the processing cost per unit labor for each part for each jacket structure model. The steel unit price of each component and the processing cost per unit weight for each part (processing labor, processing cost) may be included as cost information in the component list, for example. Being more economical than other jacket structure models may mean that the steel weight is lighter than that of other jacket structure models. The total cost (cost) of a jacket structure model and the total steel weight (steel weight) of a jacket structure model can be said to be a first target variable of the jacket structure model. The total cost may be calculated based on information indicating the material of each of the multiple components constituting the corresponding jacket structure model. In this embodiment, for example, the cost of each component can be calculated based on the component list, information indicating the shape of each component, information indicating the length of each component, and information indicating the material of each component, and the total cost is the sum of the costs of each component. Similarly, the total weight of the jacket structure model can be calculated as the sum of the weights of each component.

[0029] <Adaptability> The search unit 108 calculates the fitness of each of the first to Nth jacket structure models. The fitness indicates the degree of optimization. For example, the lower the fitness value, the better the individual jacket structure model is. The fitness is calculated based on a first target variable of the corresponding jacket structure model and a design condition penalty (design constraint condition penalty) of the corresponding jacket structure model. The fitness is calculated based on, for example, the sum, difference, quotient, or product of the first target variable and the design condition penalty, or a combination thereof. In the present embodiment, the fitness is calculated based on, for example, the sum of the first target variable and the design condition penalty. In the present embodiment, the lower the fitness of a jacket structure model, the better the individual jacket structure model is. However, the fitness is not limited to this. For example, it does not have to be calculated based on the sum of the first target variable and the design condition penalty. Furthermore, for example, the higher the fitness of a jacket structure model, the better the individual jacket structure model is. For example, fitness can be calculated as follows, The evaluation value required for each section included in each component, The calculation may be based on the constraint condition penalty calculated for each jacket structure model. Note that the section refers to the cross-sectional shape of the member. The evaluation value of each component related to the above is calculated based on the constraint penalty set for each component and the cost of each component. For example, the evaluation value of each component is calculated as the sum of the cost of each component and the constraint penalty of each component. However, the evaluation value of each component may also be calculated based on the difference, quotient, or product of the cost of each component and the constraint penalty of each component.

[0030] <Constraint penalty calculation> The search unit 108 calculates the constraint condition penalty (design constraint condition penalty) of the selected jacket structure model. The constraint condition penalty includes: <1> The component constraint penalty (pbi(x)) and <2> Displacement constraint penalty (pdsp), <3> Push-pull bearing capacity constraint penalty (paxf) and <4> In this embodiment, these <1> ~ <4> Of the penalties, <1> The member constraint penalty (pbi(x)) is set for each section (cross-sectional shape) of the member, and the remaining <2> ~ <4> Each penalty is set for each jacket structure model. <1> It can be said that the member constraint penalty is a member-specific penalty determined for each of the multiple members that make up the corresponding jacket structure model. <2> ~ <4> Each of the penalties can be said to be a per-model penalty determined for each structure model. <2> ~ <4> Each penalty (i.e., per-model penalty) relates to stress, displacement, or bearing capacity. <1> ~ <4> The value of each penalty is smaller, for example, as the component satisfies the constraint conditions. In this embodiment, the fitness is calculated based on the sum of the first target variable and the design condition penalty, and the lower the fitness of a jacket structure model, the better the jacket structure model is. Therefore, each penalty is set so that the smaller the penalty, the better the evaluation based on that penalty. However, for example, each penalty may be set so that the larger the penalty, the better the evaluation based on that penalty.

[0031] <1> Material Constraint Penalty The component constraint penalty is a penalty regarding the feasibility of connecting the target components. For example, the larger the component constraint penalty, the more difficult it is to connect the target components. The component constraint penalty is calculated using a different formula depending on whether the target component is a pile, leg, can, or girder. The component constraint penalty includes <1-1> pile diameter constraint penalty, <1-2> leg diameter constraint penalty, <1-3> can diameter constraint penalty, and <1-4> girder height constraint penalty. As the component constraint penalty, one of the penalties <1-1> to <1-4> is calculated depending on the type of component. In this way, the component constraint penalty may be set for each type of component.

[0032] <1-1> Pile diameter restriction penalty The search unit 108 calculates a pile diameter constraint penalty as a member constraint penalty for the piles among the members of each of the first to Nth jacket structure models. The pile diameter constraint penalty is calculated based on whether the target pile can be substantially connected to other piles connected to it. For example, the pile diameter constraint penalty is a penalty when the pile diameter of the target pile does not match the pile diameter of the other piles connected to it.

[0033] <1-2> Leg diameter restriction penalty The search unit 108 calculates a leg diameter constraint penalty as a component constraint penalty for the legs of each of the first to Nth jacket structure models. The leg diameter constraint penalty is calculated based on whether the target leg can be substantially connected to the pile connected to that leg. For example, the leg diameter constraint penalty is a penalty when the inner pipe diameter of the target leg does not match the pile diameter of the pile connected to that leg.

[0034] <1-3> Can diameter restriction penalty The search unit 108 calculates a can diameter constraint penalty as a component constraint penalty for the can of each of the first to Nth jacket structure models. The can diameter constraint penalty is calculated based on whether the target can can be substantially connected to the brace connected to that can. For example, the can diameter constraint penalty is a penalty when the ratio of the diameter of the brace connected to the target can to the outer tube diameter of the target can is less than α or greater than β. α is a coefficient that determines the minimum diameter of the brace connected to the target can. The minimum brace diameter is calculated by α × can diameter. For example, α is typically set to 0.2. β is a coefficient that determines the maximum diameter of the brace connected to the target can. The maximum brace diameter is calculated by β × can diameter. For example, β is typically set to 1.0.

[0035] <1-4> Digit height restriction penalty The search unit 108 calculates a girder height constraint penalty as a component constraint penalty for the girder of each of the components of the first to Nth jacket structure models. The girder height constraint penalty is calculated based on whether the target girder can be substantially connected to other girder connected to it. For example, the girder height constraint penalty is a penalty when the girder height of the target component girder does not match the girder height of other girder connected to it.

[0036] <2> Displacement Constraint Penalty The search unit 108 calculates the displacement constraint penalty for each of the first to Nth jacket structure models. The displacement constraint penalty is a penalty for the displacement at each node (joint) included in the target jacket structure model. The displacement constraint penalty indicates the degree to which the displacement at each node of the corresponding jacket structure model exceeds the allowable range. The node is, for example, the end of a member. The displacement constraint penalty becomes larger as the displacement of the node exceeds a predetermined allowable displacement. The displacement of each node can be obtained, for example, by the search unit 108 simulating (structural analysis) the displacement of the end of each member. The larger the displacement constraint penalty, the more the displacement of the target node exceeds the allowable displacement. The allowable variation (allowable range) is input, for example, from the input unit 102 as a constraint condition.

[0037] <3> Push-pull bearing capacity constraint penalty The search unit 108 calculates the push-pull bearing capacity constraint penalty for each of the first through Nth jacket structure models. The push-pull bearing capacity constraint penalty is a penalty for the bearing capacity of the piles of each jacket structure model. The push-pull bearing capacity constraint penalty indicates the degree to which the push-pull force generated in each pile included in the corresponding jacket structure model exceeds the allowable range. The push-pull bearing capacity constraint penalty becomes larger as the push-pull force generated in the pile exceeds the predetermined allowable bearing capacity (allowable value). The push-pull force generated in each pile can be determined, for example, by the search unit 108 performing a simulation (structural analysis). The larger the push-pull bearing capacity constraint penalty, the greater the push-pull force generated in each pile exceeds the allowable bearing capacity. For example, the push-pull bearing capacity constraint penalty is a penalty when the push-pull force generated in a specified pile i exceeds the constraint range. The allowable bearing capacity and constraint range are input, for example, from the input unit 102 as constraint conditions.

[0038] <4> Member stress constraint penalty The search unit 108 calculates the member stress constraint penalty (puc) for each of the first to Nth jacket structure models. The member stress constraint penalty is a penalty for the stress generated in the target member. The member stress constraint penalty indicates, for example, the degree to which the stress (and strength) in each member constituting the corresponding jacket structure model exceeds the allowable range. The member stress constraint penalty becomes larger when the stress generated in a member exceeds the allowable value. The stress generated in each member can be determined, for example, by the search unit 108 performing a simulation (structural analysis). The larger the member stress constraint penalty, the more the stress generated in each member exceeds the allowable stress (allowable value). The allowable stress (allowable value) is input, for example, from the input unit 102 as a constraint condition.

[0039] <Calculating material costs> The search unit 108 calculates the material cost. The material cost of each material is calculated by multiplying the weight of each material by the unit price per unit weight. For example, the material cost Costi of section i can be calculated using the following formula (1). Note that section i refers to the cross-sectional shape of the material.

[0040] Costi=Wi·Uniti (1)

[0041] In equation (1), Wi [ton] is the weight of steel, and Uniti [yen / ton] is the unit cost of the material in section i.

[0042] <Weight of components> The weight of each member is calculated by multiplying the volume of each member by the weight per unit volume. The volume of each member is calculated by multiplying the cross-sectional area of ​​each member by the length of each member. Note that the weight per unit volume may be constant regardless of the steel type.

[0043] <Calculation of evaluation value and fitness> The search unit 108 calculates fitness based on the cost of the components and the results of calculating component constraint penalties (pile diameter constraint penalty, leg diameter constraint penalty, can diameter constraint penalty, girder height constraint penalty), displacement constraint penalty, push-pull / pull-out bearing capacity constraint penalty, and component stress constraint penalty. Fitness indicates the degree of optimization.

[0044] In this embodiment, the search unit 108 calculates the fitness based on, for example, a total cost (first target variable) based on the sum of the costs of all components included in the corresponding jacket structure model and the design condition penalty of the corresponding jacket structure model. The search unit 108 implements an evaluation function, and it can be said that the fitness, which is an evaluation value, is calculated based on the evaluation function. The fitness may be calculated based on, for example, the sum, difference, quotient, or product of the total cost and the design condition penalty, or a combination thereof. The fitness represents the level of the total cost (first target variable) (e.g., how low the total cost is) and the level of the design condition penalty (e.g., the degree to which the design conditions are satisfied). As a result, it is possible to evaluate jacket structure models that have a low total cost and satisfy the design conditions based on the fitness (magnitude of the fitness). As long as the fitness is as described above, there is no limitation on the method of calculating the fitness. For example, when the cost of a component and the component constraint penalty are calculated as in this embodiment, an evaluation value set for each component may be calculated based on the cost of the component and the component constraint penalty. In this case, the sum of the evaluation values ​​of all components included in the corresponding jacket structure model may be used as the evaluation value of the corresponding jacket structure model itself. Such an evaluation value of the jacket structure model itself can also be said to represent the total cost (first target variable) of the jacket structure model. Therefore, the fitness may be calculated based on the evaluation value (first target variable) of the jacket structure model itself and the design condition penalty of the corresponding jacket structure model.

[0045] The search unit 108 optimizes the jacket structure model using a genetic algorithm based on the calculation results of the fitness of each of the multiple jacket structure models. The genetic algorithm prepares an arbitrary number of seed jacket structure models, prioritizes the selection of individuals with low fitness, and repeats the process of crossover (recombination) to search for the jacket with the optimal structure.

[0046] FIG. 3 is a diagram for explaining the genetic algorithm. As shown in Fig. 3, the jacket structure model is represented by genes. For example, the first five digits represent the cross-sectional specifications of the pile, the next five digits represent the cross-sectional specifications of the girder, and so on, and the last five digits represent the cross-sectional specifications of the brace. There are many variations of genetic algorithms, but in this embodiment, next-generation individuals are generated by (1) an elite strategy and (2) uniform crossover.

[0047] (1) Elite Strategy The n individuals are sorted in descending order of fitness, and the top e individuals are selected as elites to generate the next generation of individuals. The top e individuals unconditionally become the next generation of individuals. The remaining ne individuals may be generated by, for example, randomly crossing over the current generation of individuals (uniform crossover as described below), or by mutating the current generation of individuals.

[0048] (2) Uniform crossover FIG. 4 is a diagram for explaining uniform crossover in a genetic algorithm. Uniform crossover is performed based on the combination of parent 1 and parent 2 obtained by the elite strategy. Crossover involves converting the component number list held by each parent into binary (binary) numbers and randomly selecting which parent each bit will be inherited from. In Figure 4, the bolded bits are the selected bits. The component number list is a list of steel materials used in the parent jacket structure model. In the component number list, steel materials are represented by, for example, the identification numbers included in the component list.

[0049] In this embodiment, the elite strategy is used as the gene selection method (gene selection method) in the genetic algorithm, but methods other than the elite strategy can also be used. For example, the gene selection method may be a roulette method or a tournament selection method. In the roulette method, for example, a roulette wheel based on a ratio inversely proportional to fitness is created, and individuals are randomly selected using the roulette wheel. By executing the roulette wheel e times, e individuals are selected. The selected e individuals become the next-generation individuals. The remaining ne next-generation individuals may be generated, for example, by randomly uniformly crossing the current-generation individuals, or by mutation of the current-generation individuals. In the tournament selection method, for example, m individuals (for example, m is any value less than n) are randomly selected from a group of n individuals, and the individual with the lowest fitness is selected. By holding the tournament e times, e individuals are selected. The selected e individuals become the next-generation individuals. The remaining ne next-generation individuals may be generated, for example, by randomly uniformly crossing the current-generation individuals, or by mutation of the current-generation individuals.

[0050] In this embodiment, uniform crossover is used as the crossover of genes in the genetic algorithm, but it is also possible to use a method other than uniform crossover. For example, the crossover may be a single-point crossover or a multi-point crossover. Returning to Figure 1, we continue the explanation.

[0051] The search unit 108 performs the above-mentioned (1) elite strategy and (2) uniform crossover when generating the second generation or later jacket structure models. The search unit 108 extracts elites from the jacket structure model of the parent generation. The search unit 108 generates one or more jacket structure models by, for example, crossing any individuals of the parent generation (current generation) and / or mutating them at a predetermined probability. The search unit 108 newly selects one or more generated jacket structure models.

[0052] The search unit 108 determines whether the fitness of the acquired one or more jacket structure models is lower than the fitness of the jacket structure models stored in the storage unit 110. If the fitness of the one or more jacket structure models is lower than the fitness of the jacket structure models stored in the storage unit 110, the search unit 108 deletes the stored jacket structure models and newly stores in the storage unit 110 the jacket structure model with the lowest fitness among the acquired one or more jacket structure models.

[0053] The output unit 112 is configured by a display device such as a liquid crystal display (LCD) or an organic electro-luminescence display (OLED), and displays various information. The output unit 112 acquires the jacket structure model from the search unit 108 and displays the acquired jacket structure model. Furthermore, for example, the output unit 112 may output the acquired jacket structure model to another device.

[0054] All or part of the first jacket structure model generation unit 104, the second jacket structure model generation unit 106, ..., the Nth jacket structure model generation unit, and the search unit 108 are functional units (hereinafter referred to as software functional units) that are realized by a processor such as a CPU (Central Processing Unit) executing a program stored in the memory unit 110. In addition, all or part of the first jacket structure model generation unit 104, the second jacket structure model generation unit 106, ... the Nth jacket structure model generation unit and the search unit 108 may be realized by hardware such as an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), or an FPGA (Field-Programmable Gate Array), or may be realized by a combination of software functional units and hardware.

[0055] (Operation of jacket structure model search system) Figures 5 to 8 are flow diagrams showing an example of the operation of the jacket structure model search system according to this embodiment. The process of searching for a jacket structure model with low adaptability by the jacket structure model search system 100 will be described with reference to Figures 5 to 8. Figure 5 shows the overall flow, and Figures 6 to 8 are partial flows showing part of the process of the overall flow.

[0056] (Step S1-1) In the jacket structure model search system 100, search conditions, a component list, and constraint conditions are input to the input unit 102. The search unit 108 acquires the search conditions and constraint conditions from the input unit 102 (or the storage unit 110). The search unit 108 sets the acquired search conditions and constraint conditions.

[0057] (Step S2-1) In the jacket structure model search system 100, the search unit 108 performs a process of generating a first-generation jacket structure model. The first-generation jacket structure model includes N jacket structure models from the first jacket structure model to the Nth jacket structure model. Similarly, second-generation and subsequent jacket structure models also include N jacket structure models. Details of this process will be described later.

[0058] (Step S3-1) In the jacket structure model search system 100, the search unit 108 sets a variable k (hereinafter also referred to as the number of generations k) representing the number of generations of the jacket structure model generation process to 2.

[0059] (Step S4-1) In the jacket structure model search system 100, the search unit 108 performs processing to generate a k-th generation jacket structure model. Details of this processing will be described later.

[0060] (Step S5-1) In the jacket structure model search system 100, the search unit 108 determines whether or not a termination condition is satisfied. The termination condition can be set as appropriate. For example, if the number of generations k exceeds a predetermined number of generations, the search unit 108 may determine that the termination condition is met. For example, if the change in fitness is within a predetermined range even when tracing back from the jacket structure model of the kth generation to the jacket structure model of a predetermined number of generations before, the search unit 108 may determine that the termination condition is satisfied. In this case, it may be determined whether the change in fitness between the jacket structure model with the lowest fitness among the jacket structure models of each generation is within a predetermined range. The predetermined number of generations may be, for example, about 100 generations. If the search unit 108 determines that the termination condition is not satisfied, the process proceeds to step S6-1. If the search unit 108 determines that the termination condition is satisfied, the process proceeds to step S7-1.

[0061] (Step S6-1) In the jacket structure model searching system 100, the searching unit 108 increases the number of generations k by 1. The searching unit 108 then returns to step S4-1.

[0062] (Step S7-1) In the jacket structure model search system 100, the search unit 108 adopts the jacket structure model with the lowest adaptability among the k-th generation jacket structure models as the search result, and ends the process.

[0063] The process of generating a first-generation jacket structure model will be described with reference to FIG.

[0064] (Step S1-2) In the jacket structure model search system 100, the search unit 108 sets a variable n, which represents the identification number of the model, to 1.

[0065] (Step S2-2) In the jacket structure model search system 100, an nth jacket structure model generation unit (for example, the first jacket structure model generation unit 104 or the second jacket structure model generation unit 106) acquires one or more component lists from the input unit 102 (or the storage unit 110). The nth jacket structure model generation unit uses the acquired one or more component lists to generate an nth jacket structure model including a shape value of each of the one or more components and a material value of each of the one or more components. The nth jacket structure model generation unit, for example, randomly generates a first jacket structure model from the component list.

[0066] (Step S5-2) In the jacket structure model search system 100, the search unit 108 determines whether the identification number n is N. If the search unit 108 determines that the identification number n is N, for example, the search unit 108 stores the first jacket structure model to the Nth jacket structure model in the storage unit 110 as first-generation jacket structure models, and then ends the generation of the first-generation jacket structure models. If the search unit 108 determines that the identification number n is not N, the search unit 108 proceeds to step S6-2.

[0067] (Step S6-2) In the jacket structure model search system 100, the search unit 108 increments the identification number n by 1. The search unit 108 then returns to step S2-2.

[0068] A process for generating a k-th generation jacket structure model (k≧2) will be described with reference to Fig. 7. The process for generating a first generation jacket structure model is different from that for second generation and later jacket structure models.

[0069] (Step S1-3) In the jacket structure model search system 100, the search unit 108 calculates the fitness of each of the (k-1)th generation jacket structure models. The search unit 108, for example, acquires the first jacket structure model, second jacket structure model, third jacket structure model, fourth jacket structure model, ... Nth jacket structure model of the (k-1)th generation from the storage unit 110. The search unit 108 calculates the fitness of each of the acquired first to Nth jacket structure models.

[0070] (Step S2-3) In the jacket structure model search system 100, the search unit 108 sorts the first to Nth jacket structure models in ascending order of fitness, and extracts elite models from the results of sorting the first to Nth jacket structure models in descending order of fitness. In this embodiment, the top e models with the lowest fitness are designated as elite models. As described above, instead of the elite strategy, a roulette system or a tournament system can also be used as a gene selection method.

[0071] (Step S3-3) In the jacket structure model search system 100, the search unit 108 stores the top e elites in the storage unit 110 as the first to e-th jacket structure models of the k-th generation.

[0072] (Step S4-3) In the jacket structure model search system 100, the search unit 108 sets a variable n, which represents the identification number of the model, to (e+1).

[0073] (Step S5-3) In the jacket structure model search system 100, the search unit 108 generates the n-th jacket structure model by, for example, crossing any individuals of the parent generation (current generation) and / or mutating them at a predetermined probability. Note that the method for generating the n-th jacket structure model is not limited to this.

[0074] (Step S8-3) The search unit 108 determines whether the identification number n is N. If the search unit 108 determines that the identification number n is N, for example, the search unit 108 stores the first jacket structure model to the Nth jacket structure model in the storage unit 110 as first-generation jacket structure models, and then ends the generation of the first-generation jacket structure models. If the search unit 108 determines that the identification number n is not N, the search unit 108 proceeds to step S9-3.

[0075] (Step S9-3) In the jacket structure model search system 100, the search unit 108 increments the identification number n by 1. The search unit 108 then returns to step S5-3.

[0076] Referring to FIG. 8, a process for calculating the fitness of each jacket structure model of the (k-1)th generation will be described.

[0077] (Step S1-4) In the jacket structure model search system 100, the search unit 108 newly selects one jacket structure model from the (k-1)th generation jacket structure models.

[0078] (Step S2-4) In the jacket structure model search system 100, the search unit 108 calculates the pile diameter constraint penalty of the selected jacket structure model. (Step S3-4) In the jacket structure model search system 100, the search unit 108 calculates the leg diameter constraint penalty of the selected jacket structure model.

[0079] (Step S4-4) In the jacket structure model search system 100, the search unit 108 calculates the can diameter constraint penalty of the selected jacket structure model. (Step S5-4) In the jacket structure model search system 100, the search unit 108 calculates the order height constraint penalty of the selected jacket structure model.

[0080] (Step S6-4) In the jacket structure model search system 100, the search unit 108 calculates the displacement constraint penalty of the selected jacket structure model.

[0081] (Step S7-4) In the jacket structure model search system 100, the search unit 108 calculates the push-pull support force constraint penalty of the selected jacket structure model.

[0082] (Step S8-4) In the jacket structure model search system 100, the search unit 108 calculates the member stress constraint penalty of the selected jacket structure model.

[0083] (Step S9-4) In the jacket structure model search system 100, the search unit 108 calculates the total cost of the selected jacket structure model. The total cost can be calculated, for example, by adding up the steel material cost and the processing cost, as described above.

[0084] (Step S10-4) In the jacket structure model search system 100, the search unit 108 calculates the component costs. The search unit 108 calculates the fitness of the selected jacket structure model based on the total component costs (first target variable) and the results of calculating the pile diameter constraint penalty, leg diameter constraint penalty, can diameter constraint penalty, girder height constraint penalty, (per-component penalty), displacement constraint penalty, push-pull bearing capacity constraint penalty, and component stress constraint penalty (per-model penalty).

[0085] (Step S11-4) In the jacket structure model search system 100, the search unit 108 determines whether or not all of the (k-1)th generation jacket structure models have been selected. The process ends when all of the generated jacket structure models have been selected. If there is any jacket structure model that has not been selected, the process proceeds to step S1-4.

[0086] As described above, the jacket structure model search method is a search process using a genetic algorithm, and includes a search process of generating a plurality of structure models corresponding to a structure and searching for a structure model having a higher fitness than other structure models from among the plurality of generated structure models. While the present embodiment describes a case in which the search method of the present embodiment is applied to a jacket structure, the present invention is not limited to this. The search method of the present embodiment can be applied to structures in general (e.g., marine structures and land structures). In this case, the member constraint penalty may be, for example, a penalty related to a beam or a penalty related to a column. If the member constraint penalty is a penalty related to a beam, the magnitude of the member constraint penalty when a specific member is used for the beam may vary, for example, depending on the difficulty of using the specific member for the beam. If the member constraint penalty is a penalty related to a column, the magnitude of the member constraint penalty when a specific member is used for the column may vary, for example, depending on the difficulty of using the specific member for the column. The member constraint penalty may be set for each member (e.g., a frame structural member) constituting a frame structure.

[0087] (Modification 1 of the embodiment) (Jacket structure model search system) FIG. 9 is a diagram showing a jacket structure model searching system according to the first modification of the embodiment. The jacket structure model search system 100a and the jacket structure model search method according to the first modification of the embodiment automatically optimize the jacket structure model. For example, the jacket structure model search system 100a and the jacket structure model search method optimize the arrangement of components, the shapes of components, and the materials of components. That is, in this first modification, in addition to the shapes and materials of components, the arrangement of components is also optimized.

[0088] The jacket structure model search system 100a includes an input unit 102, a first jacket structure model generation unit 104a, a second jacket structure model generation unit 106a, a search unit 108, a storage unit 110, and an output unit 112.

[0089] Search conditions, a component list, constraints, and initial placement information are input to the input unit 102. The initial placement information includes initial values ​​corresponding to the placement of each of the multiple components. Examples of the initial values ​​corresponding to the placement of the components include information indicating the position of the component and information indicating the orientation of the component.

[0090] The first jacket structure model generation unit 104a acquires one or more component lists from the input unit 102. The first jacket structure model generation unit 104a uses each of the acquired one or more component lists to generate a first jacket structure model including a value of the shape of each of the one or more components, a value of the material of each of the one or more components, and a value of the arrangement of each of the one or more components. The second jacket structure model generating unit 106a and the Nth jacket structure model generating unit perform the same processing as the first jacket structure model generating unit 104a.

[0091] All or part of the first jacket structure model generating unit 104a and the second jacket structure model generating unit 106a are functional units (hereinafter referred to as software functional units) that are realized by, for example, a processor such as a CPU executing a program stored in the storage unit 110. Note that all or part of the first jacket structure model generating unit 104a, the second jacket structure model generating unit 106a, and the searching unit 108 may be realized by hardware such as an LSI, an ASIC, or an FPGA, or may be realized by a combination of a software functional unit and hardware.

[0092] Although the embodiments have been described above, these embodiments are presented as examples and are not intended to limit the scope of the disclosure. These embodiments can be implemented in various other forms, and various omissions, substitutions, modifications, and combinations can be made without departing from the spirit of the disclosure. For example, Modification 1 of Embodiment 1 may be combined with Modification 2 of Embodiment 2. These embodiments are within the scope and spirit of the disclosure, as well as within the scope of the claims and their equivalents.

[0093] The above-described jacket structure model search systems 100, 100a, and 100b may be realized by a computer. In this case, a program for realizing the functions of each functional block is recorded on a computer-readable recording medium. The program recorded on this recording medium may be read into a computer system and executed by a CPU to realize the system. The term "computer system" as used herein includes hardware such as an OS (Operating System) and peripheral devices. Furthermore, "computer-readable recording media" refers to portable media such as flexible disks, optical magnetic disks, ROMs, CD-ROMs, etc. "Computer-readable recording media" also includes storage systems such as hard disks built into computer systems.

[0094] Furthermore, the term "computer-readable recording medium" may include a medium that dynamically stores a program for a short period of time, such as a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. Furthermore, the term "computer-readable recording medium" may also include a storage medium that stores a program for a certain period of time, such as a volatile memory within a computer system that is a server or client. The program may also be for implementing part of the functions described above. The program may also be capable of implementing the functions described above in combination with a program already stored in the computer system. The program may also be implemented using a programmable logic device. An example of a programmable logic device is an FPGA (Field Programmable Gate Array).

[0095] The above-mentioned jacket structure model search systems 100, 100a, and 100b each have a computer therein. The processes of the above-mentioned jacket structure model search systems 100, 100a, and 100b are stored in a computer-readable recording medium in the form of a program, and the above-mentioned processes are performed by the computer reading and executing the program. Here, computer-readable recording media refers to magnetic disks, magneto-optical disks, CD-ROMs, DVD-ROMs, semiconductor memories, etc. Also, this computer program may be distributed to a computer via a communication line, and the computer that receives this distribution may execute the program. The program may also be for realizing part of the above-mentioned functions. Furthermore, the above-mentioned functions may be realized in combination with a program already recorded in the computer system, that is, a so-called differential file (differential program).

[0096] (Appendix 1) The above embodiment may be understood as follows, for example.

[0097] <1> A structure model search method according to one embodiment of the present disclosure includes a search process using a genetic algorithm, in which a plurality of structure models corresponding to a structure are generated, and a search process for a structure model having a fitness superior to other structure models from among the plurality of generated structure models is performed, wherein the structure models include information indicating the shape and position of each of a plurality of components constituting the corresponding structure, the fitness is calculated based on a first target variable of the corresponding structure model and a design constraint condition penalty of the corresponding structure model, and the design constraint condition penalty includes a per-model penalty calculated for each of the plurality of structure models generated in the search process.

[0098] For example, even if two structures use the same components in the same location, the stresses and other factors acting on those components will differ depending on the placement of other components, etc. Therefore, it may be preferable to impose penalties on each model rather than on each component. Here, the design constraint condition penalty includes a per-model penalty calculated for each of the multiple structure models generated in the search process, and therefore, it becomes possible to design a structure using a genetic algorithm according to fitness taking into account the per-model penalty.

[0099] <2> the above <1> In the structure model search method according to the present invention, a configuration may be adopted in which the per-model penalty includes a member stress constraint penalty indicating the degree to which the stress in each member constituting the corresponding structure model exceeds the allowable range.

[0100] For example, even if the arrangement of components that make up a structure model is the same between structure A and structure B, the stresses in the components will differ depending on the arrangement of other components of structure A and the arrangement of other components of structure B, etc. Therefore, by designing a structure using a genetic algorithm according to fitness taking into account penalties for each model, including member stress constraint penalties, it is possible to more reliably design an appropriate structure.

[0101] <3> the above <1> or <2> In the structure model search method according to the above, a configuration may be adopted in which the per-model penalty includes a displacement constraint penalty indicating the degree to which the displacement at each node of the corresponding structure model exceeds an allowable range.

[0102] For example, even if the position of a node in a structure model is the same between structure A and structure B, the displacement will differ depending on the arrangement of other components of structure A and the arrangement of other components of structure B. Therefore, by designing a structure using a genetic algorithm according to fitness taking into account penalties for each model, including the displacement constraint penalty, it is possible to more reliably design an appropriate structure.

[0103] <4> the above <1> ~ <3> In the structure model search method according to any one of the above aspects, a configuration may be adopted in which the per-model penalty includes a push-pull support capacity constraint penalty indicating the degree to which the push-pull force generated in each pile included in the corresponding structure model exceeds the allowable range.

[0104] It is possible to design structures using a genetic algorithm by considering the bearing capacity of each pile as a penalty for each model.

[0105] <5> the above <1> ~ <4> In the structure model search method according to any one of the above aspects, the structure model may include information indicating the material of each of a plurality of components constituting the corresponding structure, the first target variable is the cost of the corresponding structure model, and the cost is calculated based on the information indicating the material of each of the plurality of components constituting the corresponding structure model.

[0106] Since the cost is calculated based on the material of each of the multiple members that make up the structure model, the cost can be determined more accurately.

[0107] <6> the above <1> ~ <5> In the structure model search method according to any one of the above aspects, a configuration may be adopted in which the fitness is calculated based on the sum, difference, quotient, or product of the first target variable and the design constraint penalty, or a combination thereof.

[0108] The fitness is calculated based on the sum, difference, quotient, product, or combination of these of the first target variable and the design constraint penalty, which allows for more flexible setting of the penalty.

[0109] <7> the above <1> ~ <6> In the structure model search method according to any one of the above aspects, a configuration may be adopted in which the design constraint condition penalty includes a member-specific penalty determined for each of the plurality of members constituting the corresponding structure model.

[0110] By taking into account design constraint penalties, including component-specific penalties, structures can be designed more appropriately.

[0111] <8> A structure model search system according to one embodiment of the present disclosure includes a search unit that uses a genetic algorithm to generate a plurality of structure models corresponding to a structure and search for a structure model from among the plurality of generated structure models that has a fitness superior to the other structure models, wherein the structure models include information indicating the shape and position of each of a plurality of components that constitute the corresponding structure, the fitness is calculated based on a first target variable of the corresponding structure model and a design constraint condition penalty of the corresponding structure model, and the design constraint condition penalty includes a per-model penalty calculated for each of the plurality of structure models generated by the search unit.

[0112] <9> A structure model search program according to one embodiment of the present disclosure is a structure model search program that causes a computer to execute a search process using a genetic algorithm to generate a plurality of structure models corresponding to a structure and search for a structure model from among the plurality of generated structure models that has a fitness superior to the other structure models, wherein the structure models include information indicating the shape and position of each of a plurality of components that constitute the corresponding structure, the fitness is calculated based on a first target variable of the corresponding structure model and a design constraint condition penalty of the corresponding structure model, and the design constraint condition penalty includes a per-model penalty calculated for each of the plurality of structure models generated in the search process.

[0113] (Appendix 2) The above embodiment may be understood as follows, for example.

[0114] (1) A jacket structure model search method according to one embodiment of the present disclosure is a jacket structure model search method including: a jacket structure model generation step of generating a plurality of jacket structure models, each including a shape value of each of a plurality of components constituting a jacket structure and a material value of each of the plurality of components, using initial values ​​corresponding to the shape of each of the plurality of components and initial values ​​corresponding to the material of each of the plurality of components; and a search step of searching for a jacket structure model that is more economical than other jacket structure models, based on the plurality of jacket structure models and search conditions including physical constraints imposed on the jacket structure.

[0115] According to this disclosure, it is possible to search for a jacket structure model that is more economical than other jacket structure models based on multiple jacket structure models and search conditions, thereby enabling the design of an economical jacket regardless of the user's skill level.

[0116] (2) A jacket structure model search method according to one aspect of the present disclosure is the jacket structure model search method according to (1) above, wherein the jacket structure model generation process uses initial values ​​corresponding to the shape of each of the plurality of components constituting the jacket structure, initial values ​​corresponding to the material of each of the plurality of components, as well as initial values ​​corresponding to the arrangement of each of the plurality of components, to generate the plurality of jacket structure models including an arrangement value of each of the plurality of components, a shape value of each of the plurality of components, and a material value of each of the plurality of components.

[0117] By configuring in this way, it is possible to optimize not only the shapes and materials of the multiple members that make up the jacket structure, but also their arrangement.

[0118] (3) A jacket structure model search method according to one embodiment of the present disclosure is a jacket structure model search method described in (1) or (2) above, wherein the information indicating the shape includes any one of diameter, plate thickness, girder height, and flange width.

[0119] By configuring in this way, any one of the diameter, plate thickness, girder height, and flange width can be included in the information indicating the shape. Therefore, in the jacket structure model generation step, an initial value corresponding to any one of the diameter, plate thickness, girder height, and flange width of each of the multiple members constituting the jacket structure can be used to generate multiple jacket structure models including values ​​of the shapes of each of the multiple members.

[0120] (4) A jacket structure model search method according to one embodiment of the present disclosure is a jacket structure model search method described in any one of (1) to (3) above, wherein the physical constraints include any one of the stress occurring in each of the plurality of members, the displacement occurring in each of the plurality of members, and the bearing capacity of each pile among the plurality of members.

[0121] With this configuration, the physical constraints can include any of the stress occurring in each of the multiple members, the displacement occurring in each of the multiple members, and the bearing capacity of each of the piles among the multiple members. Therefore, in the search process, the jacket structure model can be searched for based on search conditions that include any of the stress occurring in each of the multiple members, the displacement occurring in each of the multiple members, and the bearing capacity of each of the piles among the multiple members that the jacket structure must satisfy.

[0122] (5) A jacket structure model search method according to one embodiment of the present disclosure is a jacket structure model search method described in any one of (1) to (4) above, wherein the search process uses a genetic algorithm to search for a jacket structure model that is more economical than the other jacket structure models.

[0123] By configuring the system in this way, the search process can use a genetic algorithm to search for a jacket structure model that is more economical than other jacket structure models, thereby making it possible to obtain a jacket structure model of stable quality regardless of the skill of the user.

[0124] (6) A jacket structure model search method according to one embodiment of the present disclosure is a jacket structure model search method described in any one of (1) to (5) above, and being more economical than other jacket structure models means that the total cost determined according to the quantity of the multiple components is cheaper than other jacket structure models.

[0125] By configuring in this way, being more economical than other jacket structure models can be defined as a jacket structure model whose total cost, determined according to the quantities of a plurality of members, is lower than that of other jacket structure models. Therefore, the search process can search for a jacket structure model whose total cost, determined according to the quantities of a plurality of members, is lower than that of other jacket structure models. Being more economical than other jacket structure models may also mean that the steel weight is lighter than that of other jacket structure models.

[0126] (7) A jacket structure model search system according to one embodiment of the present disclosure is a jacket structure model search system including: a jacket structure model generation unit that uses initial values ​​corresponding to the shape of each of a plurality of components constituting a jacket structure and initial values ​​corresponding to the material of each of the plurality of components to generate a plurality of jacket structure models including shape values ​​of each of the plurality of components and material values ​​of each of the plurality of components; and a search unit that searches for a jacket structure model that is more economical than other jacket structure models based on search conditions including the plurality of jacket structure models and physical constraints imposed on the jacket structure.

[0127] According to this disclosure, it is possible to search for a jacket structure model that is more economical than other jacket structure models based on multiple jacket structure models and search conditions, thereby enabling the design of an economical jacket regardless of the user's skill level.

[0128] (8) A jacket structure model search program according to one embodiment of the present disclosure causes a computer to execute a jacket structure model generation process for generating a plurality of jacket structure models including shape values ​​of each of a plurality of components constituting a jacket structure and material values ​​of each of the plurality of components, using initial values ​​corresponding to the shape of each of the plurality of components and initial values ​​corresponding to the material of each of the plurality of components; and a search process for searching for a jacket structure model that is more economical than other jacket structure models, based on the plurality of jacket structure models and search conditions including physical constraints imposed on the jacket structure.

[0129] According to this disclosure, it is possible to search for a jacket structure model that is more economical than other jacket structure models based on multiple jacket structure models and search conditions, thereby enabling the design of an economical jacket regardless of the user's skill level. [Explanation of symbols]

[0130] 100, 100a, 100b Jacket structure model search system 102 Input section 104, 104a First jacket structure model generation unit 106, 106a Second jacket structure model generation unit 108 Search Department 110 Storage section 112, 112b output section

Claims

1. A structure model search method in which a computer executes a search step using a genetic algorithm to generate a plurality of structure models corresponding to a structure and search for a structure model having a fitness superior to other structure models from among the plurality of generated structure models, the structure model includes information indicating the shape and position of each of a plurality of components that constitute the corresponding structure; the fitness is calculated based on a first target variable of the corresponding structural model and a design constraint penalty of the corresponding structural model; the design constraint condition penalty includes a per-model penalty calculated for each of the plurality of structure models generated in the search step, The design constraint condition penalty includes a member constraint penalty, which is a penalty regarding the feasibility of connection of the target member. A structure model search method comprising:

2. The member constraint penalty includes a pile diameter constraint penalty, which is a penalty when the pile diameter of the target pile does not match the pile diameter of another pile connected to the target pile.

2. The method for searching for a structure model according to claim 1.

3. The model-specific penalty includes a member stress constraint penalty indicating the degree to which the stress in each member constituting the corresponding structure model exceeds an allowable range.

2. The method for searching for a structure model according to claim 1.

4. The per-model penalty includes a displacement constraint penalty indicating the degree to which a displacement at each node of the corresponding structural model exceeds an allowable range.

2. The method for searching for a structure model according to claim 1.

5. The structural model search method according to claim 1, characterized in that the per-model penalty includes a push-pull support capacity constraint penalty indicating the degree to which the push-pull force occurring in each pile included in the corresponding structural model exceeds the allowable range.

6. the structure model includes information indicating the material of each of a plurality of members constituting the corresponding structure, the first target variable is a cost of the corresponding structure model; The cost is calculated based on information indicating the material of each of a plurality of members constituting the corresponding structure model.

6. The method for searching for a structure model according to claim 1.

7. The fitness is calculated based on a sum, a difference, a quotient, a product, or a combination thereof, of the first target variable and the design constraint penalty.

6. The method for searching for a structure model according to claim 1.

8. the design constraint condition penalty includes a member-specific penalty calculated for each of the plurality of members constituting the corresponding structure model; 6. The method for searching for a structure model according to claim 1.

9. a search unit that uses a genetic algorithm to generate a plurality of structure models corresponding to the structure and search for a structure model that has a fitness superior to other structure models from among the plurality of generated structure models; the structure model includes information indicating the shape and position of each of a plurality of components that constitute the corresponding structure; the fitness is calculated based on a first target variable of the corresponding structural model and a design constraint penalty of the corresponding structural model; the design constraint condition penalty includes a per-model penalty calculated for each of the plurality of structure models generated by the search unit, The design constraint condition penalty includes a member constraint penalty, which is a penalty regarding the feasibility of connection of the target member. A structure model search system characterized by:

10. The member constraint penalty includes a pile diameter constraint penalty, which is a penalty when the pile diameter of the target pile does not match the pile diameter of another pile connected to the target pile.

10. The structure model search system according to claim 9.

11. 11. A structure model search program that causes a computer to function as the structure model search system according to claim 9 or 10.

Images