Structural design support method and design support system

Abstract

To provide a structural body design support method and system that can clarify priority for the performance and components of a structural body, and subsequently take effective measures, thereby improving development efficiency.SOLUTION: Provided is a method for supporting design of a structural body that includes a plurality of components and has requirements to be satisfied for a plurality of types of performance. The method includes: a target setting step S1 of setting a target related to one or more optimization purposes; a bottleneck extraction step S2 of extracting a bottleneck by performing analysis using MDO; a mechanism analysis step S3 of analyzing a mechanism of the generation of the bottleneck; a structure examination step S4 of examining one or more design change plans for the structural body so that the degree of inhibiting the achievement of the target in the bottleneck is reduced; a measures structure plan derivation step S5 of deriving a measures structure plan based on the one or more design change plans; and a verification step of verifying the measures structure plan to determine the structure.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] This disclosure relates to a method and system for supporting the design of structural elements. [Background technology]

[0002] Traditionally, computer simulations (CAE, Computer-Aided Engineering) have been used when designing structures (see, for example, Patent Document 1).

[0003] Patent Document 1 discloses a design method for finding the optimal solution for interrelated components by iteratively calculating the optimal solution by prioritizing components with higher priority according to their importance. [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2003-141192 [Overview of the project] [Problems that the invention aims to solve]

[0005] Patent Document 1 provides an example of a method for determining the priority of components, in which the user can arbitrarily set the priority, but it does not specifically describe any other methods. For example, when there are many interrelated components or when there are many performance requirements to be met, it is difficult to determine the priority of the components in the first place.

[0006] Therefore, this disclosure aims to provide a design support method and system for structures that can improve development efficiency by clearly defining priorities for the performance and components of the structure and then taking effective countermeasures. [Means for solving the problem]

[0007] To solve the above problems, one aspect of the structural design support method disclosed herein is a structural design support method comprising a plurality of parts and having requirements to satisfy a plurality of performance characteristics, comprising: a goal setting step of setting a goal for one or more optimization objectives; a bottleneck extraction step of extracting one or more performance characteristics and at least one of one or more parts that hinder the achievement of the goal by performing an analysis using multi-domain optimization (MDO) with the one or more optimization objectives as the objective function, the specifications of the plurality of parts as design variables, and the requirements for at least one of the plurality of performance characteristics as constraints, thereby extracting one or more performance characteristics and at least one of one or more parts that hinder the achievement of the goal as a bottleneck; and the extracted bottleneck The present invention is characterized by comprising: a mechanism analysis step of analyzing the mechanism by which the bottleneck occurs based on the information; a structural analysis step of considering one or more design modification proposals for the structure so as to reduce the degree to which the bottleneck hinders the achievement of the objective; a countermeasure structural proposal derivation step of deriving a countermeasure structural proposal based on the one or more design modification proposals; and a verification step of verifying the countermeasure structural proposal and determining the countermeasure structural proposal as the structure of the structure when it is determined that the multiple performances of the countermeasure structural proposal satisfy the requirements and the countermeasure structural proposal achieves the objective.

[0008] Figure 9 shows an example of a conventional structural design process. A conventional design process includes, for example, a step S101 of setting goals related to optimization objectives, a step S102 of considering various design modification options while taking into account the relationships between multiple performance aspects in order to achieve the goals, a step S103 of reconciling performance aspects that are in a trade-off relationship, a step S104 of deriving a countermeasure structural option from among the various design modification options based on the results of reconciling the performance aspects, a step S105 of evaluating the performance of the countermeasure structural option, and a step S106 of determining whether the countermeasure structural option has achieved the goals. If it is determined in step S106 that the countermeasure structural option has achieved the goals, the countermeasure structural option is decided as the structure of the structural element (step S107), and the design process ends.

[0009] Traditionally, MDOs have been used in the downstream stages of the design process, such as processes S104, S105, and S106, for the purpose of verifying and evaluating proposed structural solutions. This use of MDOs in the downstream design process has yielded some success in terms of improving development efficiency. However, due to social demands stemming from environmental issues, diversifying customer needs, and the need to differentiate product strengths, the structure of structures and control / safety systems are becoming increasingly complex, and the resulting increase in development costs and time remains a significant problem. In particular, for structures with a large number of parts and a large number of performance requirements, reconciling the trade-offs between these performances and deriving structural solutions that achieve both optimization objectives and the satisfaction of each performance requirement takes considerable time, leading to prolonged development periods.

[0010] In this configuration, bottlenecks are identified through analysis using MDO (Model-Driven Architecture) in the early stages of the design process, before specific structural analysis of the structure is performed. In other words, bottlenecks that represent engineering weaknesses in achieving the optimization objectives are identified in advance of structural analysis. Then, the priority of components and performance that require focused countermeasures is clearly defined, and mechanism analysis, structural analysis, and derivation of countermeasure structural proposals are carried out starting with the highest priority components and performance. This effectively reduces the degree to which bottlenecks hinder the achievement of objectives, and ultimately improves the efficiency of development.

[0011] The aforementioned Using MDO The analysis preferably includes a sensitivity analysis that calculates the degree to which changes in the specifications of each component affect each of the performance characteristics as sensitivity.

[0012] This configuration allows for the efficient identification of bottlenecks by narrowing down the components that have a trade-off relationship with the optimization objective and those that significantly affect that performance. Note that sensitivity analysis can be performed either before or after MDO (Model-Driven Objective).

[0013] Preferably, the bottleneck extraction step includes: a first step of performing an analysis using the MDO under single-performance constraints for each of the multiple performances; a second step of performing an analysis using the MDO under multiple performance constraints for all of the multiple performances; and a third step of extracting as the bottleneck a component in which both the sensitivity obtained in the analysis of the first step and the sensitivity obtained in the analysis of the second step are greater than or equal to a predetermined value.

[0014] For a given component, if its sensitivity before and after the first analysis is above a predetermined value, it can be determined to be a bottleneck component. By comparing the sensitivity of such a component with the sensitivity obtained in the second analysis, components where both are above a predetermined value indicate that changes in specifications will have a significant impact on both the bottleneck performance and overall performance, and are therefore bottleneck components that require particular attention. Consequently, this configuration allows for the efficient identification of bottlenecks.

[0015] In the bottleneck extraction step, it is preferable to perform an analysis using the MDO under single-performance constraints for each of the multiple performances, and to extract the component having the performance for which the sign of the sensitivity obtained in the analysis is inverse as the bottleneck.

[0016] Performance characteristics with opposite signs in sensitivity are in a trade-off relationship. In components with performance characteristics in a trade-off relationship, it is thought that multiple performance characteristics are functionally distributed in a complex relationship. Therefore, components with performance characteristics in a trade-off relationship can be identified as bottlenecks.

[0017] Preferably, the bottleneck extraction step includes the steps of: performing an analysis using the MDO under single-performance constraints for each of the multiple performances, calculating a value related to the specification after the analysis or a value related to the difference in the specification before and after the analysis as an inhibition degree that hinders the achievement of the target for each performance; and extracting as a bottleneck the component whose sensitivity to the performance for which the inhibition degree is above a predetermined value is above a predetermined value.

[0018] Performance with an inhibition degree equal to or higher than a predetermined value means that the degree of inhibiting the achievement of the target is large, and thus corresponds to bottleneck performance. Components with a sensitivity in the bottleneck performance equal to or higher than a predetermined value have a large impact on the bottleneck performance, and thus can be efficiently extracted as bottleneck components.

[0019] In the bottleneck extraction step, it is preferable to divide the sensitivity by the area of the component to convert it into the sensitivity per unit area, and extract a component with a sensitivity per unit area equal to or higher than a predetermined value as the bottleneck.

[0020] By converting the sensitivity into the sensitivity per unit area, the influence of the component specifications on the performance can be evaluated regardless of the area of the component, and thus bottleneck components can be extracted more effectively.

[0021] The optimization objective is preferably minimization of mass.

[0022] According to this configuration, a structure of a more lightweight structure can be proposed.

[0023] Moreover, one aspect of the structural design support system disclosed herein is for supporting the design of a structure having a plurality of components and requirements to be satisfied for a plurality of performances. Design supportA system comprising: a goal setting unit that sets goals related to one or more optimization objectives; a bottleneck extraction unit that extracts one or more performance characteristics and at least one of one or more components that hinder the achievement of the goals by performing an analysis using multi-domain optimization (MDO) with the one or more optimization objectives as the objective function, the specifications of the plurality of components as design variables, and the requirements in at least one of the plurality of performance characteristics as constraints; a mechanism analysis unit that analyzes the mechanism by which the bottleneck occurs based on the extracted bottleneck information; and a mechanism analysis unit that analyzes the mechanism obtained as a result of the analysis and identifies the bottleneck The system is characterized by comprising: a structural examination unit that examines one or more design modification proposals for the structure in such a way that the degree to which the achievement of the objective in the system is hindered is reduced; a countermeasure structural proposal derivation unit that derives a countermeasure structural proposal based on the one or more design modification proposals; a determination unit that verifies the countermeasure structural proposal and determines whether the multiple performances in the countermeasure structural proposal satisfy the requirements and whether the countermeasure structural proposal achieves the objective; and a structural determination unit that, when the determination unit determines that the multiple performances in the countermeasure structural proposal satisfy the requirements and the countermeasure structural proposal achieves the objective, determines the countermeasure structural proposal as the structure of the structure.

[0024] In this configuration, bottlenecks that represent engineering weaknesses in achieving optimization objectives are identified in advance of structural analysis. Then, prioritizing the components and performance aspects that require focused attention is clearly defined, and mechanism analysis, structural analysis, and derivation of countermeasures are carried out starting with the highest-priority components and performance aspects. This effectively reduces the degree to which bottlenecks hinder the achievement of objectives, and ultimately improves development efficiency. [Effects of the Invention]

[0025] As described above, according to this disclosure, in a structural design support method and system, development efficiency can be improved by clearly prioritizing the performance and components of the structure and then taking effective countermeasures. [Brief explanation of the drawing]

[0026] [Figure 1] A diagram showing the hardware configuration of a structural design support system. [Figure 2] A diagram showing the software configuration of a structural design support system. [Figure 3] A flowchart illustrating the steps for supporting the design of a structure. [Figure 4] A flowchart illustrating the steps for supporting the design of a structure. [Figure 5] A schematic diagram showing a portion of the vehicle body analyzed in the experimental example. [Figure 6] A figure showing some of the results of the sensitivity analysis of experimental examples. [Figure 7] A graph showing some of the inhibition levels calculated from experimental examples. [Figure 8] A diagram showing some of the proposed countermeasures for the experimental example and some of the results of the load transfer analysis. [Figure 9] A flowchart illustrating the steps of a conventional structural design support method. [Modes for carrying out the invention]

[0027] Embodiments of the present disclosure will be described in detail below with reference to the drawings. The following description of preferred embodiments is illustrative in nature and is not intended to limit the present disclosure, its applications, or its uses in any way.

[0028] <Structure> In this disclosure, the structure under design is not particularly limited, as long as it comprises multiple components and has requirements to satisfy multiple performance aspects. Specific examples of structures include, for example, the bodies of various vehicles such as automobiles, motorcycles, trucks, tractors, heavy machinery, and aircraft; the hulls of ships; and various structures such as the whole or parts of buildings.

[0029] The number and types of parts, as well as the number and types of performance, will vary depending on the type of structure, generation, etc., and are not particularly limited.

[0030] Taking the example of a car body as the structure, a car body consists of, for example, more than 50 parts. Specifically, these parts include, for example, the front frame, side frames, rear side frames, roof rails, tunnel reinforcement, cross members, floor panels, front panels, rear panels, engine covers, hoods, rear fenders, roofs, doors, liftgates, rear end panels, rear wheelhouse outers, rear end members, rear pillar outers, rear pillar inners, rear pillar reinforcements, upper reinforcements, rear roof rails, and other body components.

[0031] Furthermore, an automobile body possesses multiple performance characteristics, specifically, for example, more than 50 performance characteristics, and each performance characteristic has requirements that must be met. These requirements may be permissible ranges and / or permissible values, or they may be preferred ranges and / or preferred values ​​that are targets for improvement. Specific performance characteristics include, for example, collision performance (a performance index indicating resistance to deformation during a collision, specifically frontal collision performance, side collision performance, and rear collision performance), noise insulation performance (hereinafter sometimes referred to as "NVH"), body rigidity (a performance index indicating resistance to deformation), body frame resonance (a performance index expressed by low-frequency eigenvalues ​​related to handling stability), reliability, and the like.

[0032] <Design support system> Figure 1 schematically shows an example of the hardware configuration of the structural design support system (more precisely, the computer system 201 that implements the system) related to this disclosure. Note that computer system 201 (hereinafter referred to as computer 201) is merely one example of a structural design support system related to this disclosure, and the configuration of the design support system is not limited to this example.

[0033] Computer 201 is a CAE (Computer-Aided Engineering) system. Computer 201 includes a CPU 203 that controls the entire system, a ROM 205 that stores boot programs and other data, a RAM 207 that functions as main memory, and a hard disk 209 (hereinafter referred to as HDD 209) as secondary storage. Computer 201 also includes a display 211 as a display device and a VRAM 213 that functions as memory for storing image data to be displayed on the display 211. Furthermore, Computer 201 includes a keyboard 215 and a mouse 217 as input devices. This computer 201 is configured to communicate with external devices via an interface 221.

[0034] As shown in Figure 2, the HDD209 stores various programs in its program memory, including the operating system (OS) 219, MDO analysis program 229, sensitivity analysis program 239, and application program 249. Meanwhile, the HDD209 also stores various data, such as model data 259, in its data memory. This model data 259 includes information that visualizes the structure, specifications for each component of the structure, information on requirements that must be satisfied for each performance aspect, and approximate models. The information stored in the model data 259 can be arbitrarily modified and added by the user. Furthermore, the data memory of the HDD209 also stores various calculation results generated by the execution of various programs.

[0035] In the above configuration, various programs are launched in response to specific commands input from the keyboard 215 or mouse 217. At that time, the various programs are loaded from the HDD 209 into the RAM 207 and executed by the CPU 203, thereby enabling the computer 201 to function as a design support system. In this embodiment, the CPU 203 corresponds to the calculation unit (target setting unit, bottleneck extraction unit, mechanism analysis unit, structural examination unit, countermeasure structural proposal derivation unit, determination unit, and structural determination unit) as referred to in this disclosure.

[0036] <Design support method> As shown in Figure 3, an example of a structural design support method according to this disclosure comprises a target setting step S1, a bottleneck extraction step S2, a mechanism analysis step S3, a structural examination step S4, a countermeasure structural proposal derivation step S5, a performance evaluation step S6 (verification step), a determination step S7 (verification step), and a structural determination step S8 (verification step).

[0037] [Goal setting process] The goal-setting process S1 is the process of setting goals related to the optimization objective.

[0038] An optimization objective is an event set in the objective function of MDO, which is an objective to be minimized or maximized. Specific examples of optimization objectives include mass minimization (also called "lightweighting"), cost minimization, CO2 minimization during operation, CO2 minimization during manufacturing, and common parts maximization. Note that there may be one or more optimization objectives; that is, one or two or more. From the viewpoint of proposing a lighter structure, it is preferable that the optimization objective include mass minimization.

[0039] The objectives for optimization are, for example, numerical targets, and vary depending on the type of optimization objective. If the optimization objective is mass minimization, the target is given, for example, the lightweight efficiency η[%] of the mass of the structure represented by the following equation (1).

[0040] η = (W0 - W1) / W0 × 100 ... (1) However, in equation (1), W0 is the initial mass of the vehicle body [kg], and W1 is the mass of the vehicle body after design [kg].

[0041] A positive value for η indicates that the mass of the vehicle body has decreased, achieving weight reduction; a negative value indicates that the mass of the vehicle body has increased, failing to achieve weight reduction. In the design support method described herein, when the optimization objective is weight reduction, the final proposed countermeasure structure only needs to satisfy a predetermined weight reduction efficiency η while also meeting the requirements for multiple performance aspects. The predetermined weight reduction efficiency η can be appropriately changed depending on the weight reduction objective.

[0042] [Bottleneck extraction process] The bottleneck extraction step S2 is a process in which, by performing analysis using multi-domain optimization (MDO), at least one of one or more performance characteristics and at least one of one or more components that hinder the achievement of the target is identified as a bottleneck.

[0043] MDO is an analysis technique that efficiently optimizes multiple performance evaluation items. MDO is an optimization calculation in which one or more optimization objectives are defined as the objective function, the specifications of multiple components are defined as design variables, and the requirements for at least one of multiple performance items are defined as constraints. A concrete example of this optimization calculation is the method described in "Takeo Kodaira, Kohei Amano, "Extraction of Design Knowledge for Lightweighting of Vehicle Body Structures by Multi-Domain Optimization and Trade-Off Analysis," Transactions of the Institute of Electrical Engineers of Japan, Vol. 134 No. 9 (2014), pp. 1348-1354." (hereinafter sometimes referred to as "Kodaira et al. 2014"). The underlying tools for MDO are not limited and commercially available tools can be used. Specifically, for example, general-purpose automated integrated optimization software such as Isight and modeFRONTIER can be used.

[0044] According to Kodaira et al. (2014), the MDO can be formulated, for example, by the following equations (2.1) to (2.4). In equations (2.1) to (2.4), D is the number of design variables, T is the matrix transpose, and N is the matrix transpose. con x is the number of constraints (performance evaluation items). i L , x i U (i=1,2,...,D) are the lower and upper limits of the design variables.

[0045]

number

[0046] Based on equations (2.1) to (2.4), MDO can be performed as a computational process combining experimental design and approximate models.

[0047] Experimental design is a data sampling method for efficiently analyzing the effects of variations in design variables on performance indicators and mass. Specific examples of experimental design include factorial design, partial factorial design, and optimal Latin hypersquare design, with the optimal Latin hypersquare design being particularly preferred. The optimal Latin hypersquare design can handle both continuous (real) and discrete (integer) design variables and can generate multi-level, uniformly distributed sample data within the design space.

[0048] Next, based on the generated sampling data, the relationship between the specifications of each component and each performance indicator is modeled using an approximate formula, employing methods such as response surface theory.

[0049] Response surface approximation is a method for approximating discrete sampled data to a continuous surface, allowing for the visualization of the characteristics of the objective function and enabling analysis such as identifying important design variables that have a significant impact on the objective function. Specific examples of response surface approximation include polynomial approximation using multiple regression analysis with least squares method, RBF (Radial Basis Function) interpolation, and the Kriging method.

[0050] After creating an approximate model, optimization calculations are performed. The process of creating an approximate model and performing optimization calculations is repeated until the prediction accuracy of the optimal solution is within an acceptable range, and the final optimal solution is obtained.

[0051] The analysis preferably includes a sensitivity analysis that calculates, as sensitivity, the degree of influence of changes in the specifications of each component on each performance prior to and / or based on the results of MDO. Specific examples of such sensitivity analysis include the methods described in the above-mentioned Kojima et al. 2014, "Hiroshi Kojima, Takehisa Kojima, "Development of Design Support Technology for Automobile Body Structures", Mazda Technical Report, No. 36 (2019), pp. 272-276", Japanese Unexamined Patent Application Publication No. 2020-071725, "Toshiki Kondo, Takehisa Kojima, Hiroshi Kojima, "Development of Design Support Technology for Efficient Discovery of Structural Knowledge in Automobile Bodies (1) Proposal of Nonlinear Sparse Modeling Using Evolutionary Factor Extraction and Factor Selection Probability", The Japan Society of Mechanical Engineers, Proceedings of the 31st Design Engineering and System Division Conference, 2021" (hereinafter sometimes referred to as "Kondo et al. 2021").

[0052] Specifically, for example, according to the above-mentioned Kojima et al. 2014, the performance approximation function can be expressed, for example, by the following formula (3) using only the linear terms in the polynomial regression model (PRM) of the response surface method. However, in formula (3), β0, β j are regression coefficients.

[0053]

Equation

[0054] When the specifications of each component are changed, that is, when the design variable x j is changed, the degree of influence of the change in the specifications on the performance index is obtained as the sensitivity of the specifications of each component to each performance index.

[0055] Specifically, as shown in the following formula (4), the sensitivity β j is obtained by partially differentiating the response value y of the polynomial approximation formula of the above formula (3) with respect to the design variable x j That is, the sensitivity β j is the partial regression coefficient of formula (3).

[0056]

Equation

[0057] In this way, by adopting the partial regression coefficients of the polynomial approximation formula as the sensitivity, it is possible to provide a design support method that is intuitive and empirically easy for designers to understand.

[0058] Furthermore, by using sensitivity analysis, it is possible to narrow down the performance aspects that have a trade-off relationship with the optimization objective and the components that have a significant impact on those performance aspects, thereby efficiently identifying bottlenecks.

[0059] Figure 4 shows an example of the bottleneck extraction process S2. As shown in Figure 4, the bottleneck extraction process comprises a first step S91 in which MDO is performed under single-performance constraints for each of the multiple performances, a second step S92 in which MDO is performed under multiple performance constraints for at least two or more, preferably all, of the multiple performances, and a first and second step S93 in which sensitivity analysis is performed based on the results of steps S91 and S92.

[0060] Then, in step S95 (third step), components whose sensitivity obtained from the analysis in steps S91 and S93 and the sensitivity obtained from the analysis in steps S92 and S93 are both above a predetermined value can be identified as bottlenecks.

[0061] For a given component, if the sensitivity obtained from the analysis in processes S91 and S93 is above a predetermined value, it can be determined to be a bottleneck component. By comparing the sensitivity of such a component with the sensitivity obtained from the analysis in processes S92 and S93, components where both are above a predetermined value can be said to be bottleneck components where changes in specifications have a significant impact on both the bottleneck performance and overall performance, and where countermeasures should be taken with particular emphasis. Therefore, this configuration allows for the efficient identification of bottlenecks.

[0062] The predetermined sensitivity value can be set appropriately depending on the type of structure, components, performance, optimization objectives, etc.

[0063] Once all components and their performance have been examined and the bottlenecks have been identified (process S96), the process can proceed to the next mechanism analysis process S3.

[0064] The bottleneck extraction step S2 may include a step S94 for calculating the degree of inhibition. The degree of inhibition is the degree to which the achievement of the target is hindered. The degree of inhibition is obtained by performing an analysis using MDO under single-performance constraints for each of the multiple performances, and organizing the values ​​related to the specifications after the analysis, or the values ​​related to the difference in specifications before and after the analysis, for each performance. If the optimization objective is weight reduction, the values ​​related to the specifications after the analysis are given by, for example, the plate thickness after the analysis, the weight of the part, etc., and the values ​​related to the difference in specifications before and after the analysis are given by, for example, the increase in plate thickness, the increase in the weight of the part or vehicle body (also called "weight difference"), the reciprocal of the weight difference of the part or vehicle body, the weight difference per unit area of ​​the part or vehicle body, the reciprocal of the weight difference per unit area of ​​the part or vehicle body, etc. Furthermore, if a sensitivity analysis is performed, the values ​​related to the difference in specifications before and after the analysis may be sensitivity.

[0065] Then, in step S95, components whose sensitivity to the performance of the above-mentioned inhibition level is above a predetermined value are identified as bottlenecks.

[0066] Performance that exhibits an inhibitory level above a predetermined value indicates a significant degree of impediment to achieving the target, and therefore corresponds to a bottleneck performance. Components with a sensitivity to bottleneck performance above a predetermined value have a significant impact on bottleneck performance and can therefore be efficiently identified as bottleneck components.

[0067] The predetermined value of the inhibition level can be set appropriately depending on the type of structure, components, performance, optimization objectives, etc.

[0068] In the bottleneck extraction step S2, an analysis using MDO under single-performance constraints may be performed for each of the multiple performances, and components with performance where the sign of the sensitivity obtained in the analysis is inverse may be extracted as bottlenecks.

[0069] Performance characteristics with opposite signs in sensitivity are in a trade-off relationship. In components with performance characteristics in a trade-off relationship, it is thought that multiple performance characteristics are functionally distributed in a complex relationship. Therefore, components with performance characteristics in a trade-off relationship can be identified as bottlenecks.

[0070] In the bottleneck extraction step S2, the sensitivity may be converted to sensitivity per unit area by dividing it by the area of ​​the component, and components whose sensitivity per unit area is equal to or greater than a predetermined value may be extracted as bottlenecks.

[0071] By converting sensitivity to sensitivity per unit area, it becomes possible to evaluate the impact of component specifications on performance regardless of the component's area, thus enabling more effective identification of bottleneck components.

[0072] [Mechanism Analysis Process] The mechanism analysis step S3 is a step in which the mechanism by which the bottleneck occurs is analyzed based on the extracted bottleneck information.

[0073] The analysis method for the mechanism can be appropriately selected depending on the performance and type of component that are the bottlenecks.

[0074] One example of a mechanism analysis method is load transfer path analysis, which uses the finite element method ("Hiroshi Kenmochi, Takeo Kodaira, Sadayoshi Okamoto, "Development of Design Support Technology for Efficient Discovery of Structural Knowledge in Automobile Bodies (2) Proposal of Dynamic Visualization Analysis Method Based on Load Transfer Index (U*)", Japan Society of Mechanical Engineers, Proceedings of the 31st Conference of the Design Engineering and Systems Division, 2021" (hereinafter sometimes referred to as "Kenmochi et al. 2021"), JP 2020-013354). It is desirable that the load is transferred in such a way that the flow of mechanical energy is uniform throughout the entire structure. In other words, in bottleneck parts and their surroundings, the load transfer path may not be sufficiently secured, which may greatly hinder the achievement of the target.

[0075] [Structural analysis process and derivation of proposed structural solutions] Structural analysis step S4 is a step in which, based on the mechanism obtained from the analysis, one or more, preferably multiple, design modification options for the structure are considered so as to reduce the degree to which the bottleneck hinders the achievement of the goal.

[0076] Specifically, for example, the results of the load transfer path analysis described above may reveal that the flow of mechanical energy in and around the bottleneck component is non-uniform, and that the load transfer path is not adequately secured. In such cases, various design modifications may be considered, such as changing the joining method of the connection parts of the bottleneck component and its surroundings, changing multiple separately molded parts into a single molded part, or changing the shape of the part itself.

[0077] Then, in the S5 process for deriving countermeasure structural proposals, from among the various design modification proposals considered, a countermeasure structural proposal that is expected to improve performance more effectively and has minimal adverse effects on objective functions such as mass is derived using methods such as MDO.

[0078] In the structural analysis step S4, one design change proposal may be considered, but it is preferable to consider multiple proposals. Furthermore, if there are multiple types of bottlenecks, i.e., if there are multiple bottleneck performance characteristics and multiple bottleneck components, it is preferable to consider multiple design change proposals for each of the multiple types of bottlenecks. By considering multiple design change proposals in the structural analysis step S4 and comparing them in the subsequent countermeasure structural proposal derivation step S5, it is possible to derive a countermeasure structural proposal that is more advantageous in improving performance and achieving the target.

[0079] If only one design change proposal was considered in the structural analysis process S4, then that design change proposal should be used as the countermeasure structural solution in the countermeasure structural solution derivation process S5.

[0080] [Performance evaluation process, judgment process, and structural determination process] In the performance evaluation process S6, the proposed countermeasure structure is verified, and it is evaluated whether the proposed countermeasure structure satisfies the requirements for each performance aspect.

[0081] Specifically, for example, the mechanism analysis performed in step S3 above is repeated for the proposed countermeasure structure to confirm whether the degree to which the bottleneck hinders the achievement of the target has been reduced. In addition, MDO, etc., performed in step S2 above is carried out for the proposed countermeasure structure to confirm whether the proposed countermeasure structure has achieved the target. Furthermore, a sensitivity analysis may be performed to confirm whether the degree to which the bottleneck hinders the achievement of the target has been reduced by comparing the sensitivity of each component and performance of the proposed countermeasure structure with the sensitivity of each component and performance of the base.

[0082] Based on the results obtained in the performance evaluation step S6, it is determined whether multiple performance aspects of the proposed countermeasure structure satisfy the requirements and whether the proposed countermeasure structure achieves the objective (determination step S7).

[0083] Then, when it is determined that each performance aspect of the proposed countermeasure structure satisfies the requirements and the proposed countermeasure structure achieves the objective, the proposed countermeasure structure is decided as the structure of the structure (structure determination step S8).

[0084] Once the structure is determined, the structural drawings may be output as 3D CAD data or similar. The manufacturing department will then manufacture the structure based on this drawing data.

[0085] Furthermore, if in the determination step S7 it is determined that each performance aspect of the proposed countermeasure structure does not satisfy the requirements or that the proposed countermeasure structure does not achieve the objective, then the process can be repeated, for example, by returning to the mechanism analysis step S3 (or possibly the objective setting step S1 or bottleneck extraction step S2) and repeating the steps of each step.

[0086] <Effects and Effects> Unlike the conventional design process shown in Figure 9, this configuration is characterized by its upstream approach to the design process, specifically by extracting bottlenecks through analysis using MDO before conducting concrete structural studies of the structure. In other words, bottlenecks that represent engineering weaknesses in achieving the optimization objectives are identified in advance of structural studies. Then, the priority of components and performance that require focused countermeasures is clearly defined, and mechanism analysis and structural studies are performed prioritizing the components and performance with the highest priority. This effectively reduces the degree to which bottlenecks hinder the achievement of objectives, and ultimately improves development efficiency.

[0087] <Example of experiment> For the vehicle body 100 of the current mass-produced vehicle shown in Figure 5, MDO and sensitivity analysis were performed under the following conditions, and sensitivity β j The result was calculated.

[0088] Objective function: Mass minimization Design variables: Maximum 48 variables for part thickness (407 parts in total, mainly including body frame components such as front frame, side frame, rear side frame, and roof rails) Constraints: Under individual performance constraints and overall performance constraints for 71 performance indicators (crash performance, handling stability, NVH, etc.) Experimental Design: Optimal Latin Hypersquare Method Sample size: 95-110 Response surface methodology: polynomial approximation, RBF Variable selection: GA + mean p-value Initial generation population: 200 individuals Number of evolutionary generations: 200 generations Figure 6 shows the obtained sensitivity β for some components. j The values ​​are shown as a bar graph. In Figure 6, the longer the bar, the greater the sensitivity β. j This indicates that the absolute value of the value is large. In Figure 6, sensitivity β jA positive sign for the value (indicated by coarse hatched bars) means that performance deteriorates with reduced plate thickness. Conversely, a negative sign (indicated by fine hatched bars) means that performance does not deteriorate with reduced plate thickness. In Figure 6, "OUT" and "IN" in the part names represent the outer and inner parts, respectively.

[0089] Furthermore, Figure 7 shows the results of calculating the degree of impediment to weight reduction (the reciprocal of the weight difference before and after MDO in the total weight of all components set as design variables) for each major performance based on the results of single-performance constraint MDO and sensitivity analysis.

[0090] As shown in Figure 7, it was found that side impact performance was the most detrimental and the most difficult to reduce in terms of weight. In other words, side impact performance can be considered a bottleneck performance characteristic.

[0091] Considering the results in Figure 7, it was found in Figure 6 that the parts with a positive sign in sensitivity for side impact performance are the A-pillar outer 4, B-pillar outer 5, side sill inner 6, side frame outer 9, #4.5 cloth 16, F-door outer 18, and F-door inner 19. Since reducing the thickness of these parts worsens side impact performance, they can be considered bottleneck parts. In particular, the A-pillar outer 4, B-pillar outer 5, side sill inner 6, side frame outer 9, and F-door inner 19 have higher sensitivity values ​​compared to other parts, indicating that the degree to which side impact performance deteriorates with reduced thickness is greater, and they can be judged as high-priority bottleneck parts. For reference, the three parts A-pillar outer 4, B-pillar outer 5, and side sill inner 6 are shown with hatching in Figure 5.

[0092] Furthermore, in Figure 6, the sensitivity obtained from the MDO under single-performance constraints and the MDO under overall performance constraints for both the B-pillar outer 5 and the side sill inner 6 is very large and the same value. This indicates that reducing the thickness of the B-pillar outer 5 and the side sill inner 6 greatly deteriorates not only the side impact performance but also the overall performance. In other words, reducing the thickness of the B-pillar outer 5 and the side sill inner 6 has a significant impact on the bottleneck performance, which is side impact performance and overall performance, and it was found that these are bottleneck components that should be addressed with particular emphasis.

[0093] Furthermore, as mentioned above, performance characteristics with opposite signs of sensitivity are in a trade-off relationship. In Figure 6, for example, the cab side region components, such as the A-pillar outer 4, B-pillar outer 5, side sill inner 6, side frame outer 9, R-suspension housing inner 13, R-suspension housing outer 14, #4 cross UP 15, F-door outer 18, F-door inner 19, and R-door outer 20, have many performance characteristics in such a trade-off relationship, indicating that multiple performance characteristics are functionally distributed in a complex relationship. From this, it can be said that the cab side region is a bottleneck.

[0094] Furthermore, Figure 6 shows the area of ​​the part, assuming the part is a curved surface without considering the plate thickness. In the "Area" column of Figure 6, the area is shown as a bar graph, with longer bars indicating a larger area. The sensitivity value can be converted to sensitivity per unit area by dividing it by the area of ​​the part, and bottleneck extraction can be performed based on this sensitivity per unit area. For example, comparing the B-pillar outer 5 and the side sill inner 6, which have almost the same sensitivity, the sensitivity per unit area in side impact performance is about twice that of the former compared to the latter, indicating that the former has a greater impact on side impact performance. In other words, although the B-pillar outer 5 is a smaller part in area than the side sill inner 6, it has a greater impact on side impact performance, so it may be extracted as a higher-priority bottleneck part.

[0095] In this way, all components and their performance were examined, and bottleneck components and bottleneck products were identified.

[0096] Next, we focused on elucidating the mechanism by which the bottleneck occurs. As a specific example, we investigated the flow of mechanical energy in the cabside region using the load transfer path analysis described above by Kenmochi et al. (2021).

[0097] As a result, it was found that in the base structure, for example, the flow of mechanical energy becomes non-uniform at the connection point 101 between the rear end of the side sill inner 6 and the rear wheel arch 21, as shown in Figures 5 and 8(a) (Figure 8(b)).

[0098] Therefore, we considered a countermeasure structural design that involves applying a strength-enhancing filler to the connection part 101 to reinforce the ridge line 31. When we performed a load transmission path analysis on the countermeasure structural design, we found that by performing the above-mentioned ridge line reinforcement at the connection part 101, a path is formed through which the load is transmitted to the inner part via the side frame outer 9, and the flow of mechanical energy at the connection part 101 is made uniform (Figure 8(c)). It is thought that the ridge line reinforcement promotes load transmission and improves performance by acting as a link between the inner part and the outer part.

[0099] Regarding other bottlenecks, various countermeasures were implemented, and the proposed structural solutions were subjected to the aforementioned Model Development Experiment (MDO) and verified. It was found that the performance requirements and targets (3% vehicle weight reduction efficiency) were met. Specifically, comparing the base weight and the weight of the proposed structural solutions before and after the MDO, it was found that a weight reduction of 9 kg could be achieved. [Industrial applicability]

[0100] This disclosure is extremely useful because, in a structural design support method and system, development efficiency can be improved by clearly prioritizing the performance and components of a structure and then taking effective countermeasures. [Explanation of Symbols]

[0101] 1. Cowl side up (part) 2 Cowl (parts) 3. Lower Dash (Part) 4. A-pillar outer (part) 5. B-pillar outer (part) 6. Side sill inner (part) 7 F Floor Panel (Part) 8 F-frame rear (parts) 9. Side frame outer (part) 10#3 Cross Up (Parts) 11 R Header (Part) 12 R corner inner (part) 13. Rear suspension housing inner (part) 14. Rear suspension housing outer (part) 15 #4 Cross Up (Parts) 16 #4.5 Cross (Parts) 17 R End Panel (Part) 18 Front Door Outer (Part) 19 Front Door Inner (Part) 20 R Door Outer (Part) 21 Rear wheel arch (part) 31 Ridge reinforcement 100 Vehicle body (structure) 101 Connection part 201 Computer Systems (Structural Design Support Systems) 203 CPU (Target setting unit, bottleneck extraction unit, mechanism analysis unit, structural study unit, countermeasure structural proposal derivation unit, judgment unit, structural determination unit)

Claims

1. A method for supporting the design of a structure having multiple components and having requirements to satisfy multiple performance aspects, A goal setting process involves setting goals related to one or more optimization objectives, A bottleneck extraction step is performed by conducting an analysis using multi-domain optimization (MDO), in which the one or more optimization objectives are defined as the objective function, the specifications of the multiple components are defined as design variables, and the requirements for at least one of the multiple performances are defined as constraints, thereby extracting one or more performances and at least one of the one or more components that hinder the achievement of the objective as a bottleneck. A mechanism analysis step is performed to analyze the mechanism by which the bottleneck occurs, based on the extracted bottleneck information. A structural analysis step involves considering one or more design modifications to the structure based on the mechanism obtained from the analysis, so as to reduce the degree to which the bottleneck hinders the achievement of the objective. A process for deriving a countermeasure structure proposal based on one or more design change proposals, The system includes a verification step of verifying the proposed countermeasure structure and determining the proposed countermeasure structure as the structure of the structure when it is determined that the multiple performance characteristics of the proposed countermeasure structure satisfy the requirements and the proposed countermeasure structure achieves the objective. A method for supporting the design of a structure characterized by the following features.

2. In claim 1, The analysis using the aforementioned MDO includes a sensitivity analysis that calculates the degree to which changes in the specifications of each component affect each of their respective performance characteristics as sensitivity. A method for supporting the design of a structure characterized by the following features.

3. In claim 2, The aforementioned bottleneck extraction process is: A first step involves performing an analysis using the MDO under single-performance constraints for each of the aforementioned multiple performances, A second step involves performing an analysis using the MDO under multiple performance constraints for all of the aforementioned performance levels, A method for supporting the design of a structure, characterized by including a third step of extracting a component as the bottleneck in which both the sensitivity obtained in the analysis of the first step and the sensitivity obtained in the analysis of the second step are greater than or equal to a predetermined value.

4. In claim 2, In the bottleneck extraction step, an analysis is performed using the MDO under single-performance constraints for each of the multiple performances, and the component having the performance for which the sign of the sensitivity obtained in the analysis is reversed is extracted as the bottleneck. A method for supporting the design of a structure characterized by the following features.

5. In claim 2, The aforementioned bottleneck extraction process is: A step of performing an analysis using the MDO under single performance constraints for each of the aforementioned multiple performances, and calculating the degree of impediment to achieving the target for each performance, based on the value related to the specification after the analysis or the value related to the difference in the specification before and after the analysis. The process includes a step of identifying a component as a bottleneck whose sensitivity to the performance is above a predetermined value and whose inhibition level is above a predetermined value. A method for supporting the design of a structure characterized by the following features.

6. In claim 2, In the bottleneck extraction step, the sensitivity is converted to sensitivity per unit area by dividing the sensitivity by the area of ​​the component, and components whose sensitivity per unit area is equal to or greater than a predetermined value are extracted as the bottleneck. A method for supporting the design of a structure characterized by the following features.

7. In claim 2, The aforementioned optimization objective is to minimize mass. A method for supporting the design of a structure characterized by the following features.

8. A design support system for assisting in the design of a structure having multiple components and having requirements to satisfy multiple performance aspects, A goal setting unit that sets goals related to one or more optimization objectives, A bottleneck extraction unit performs an analysis using multi-domain optimization (MDO), where the one or more optimization objectives are defined as the objective function, the specifications of the multiple components are defined as design variables, and the requirements for at least one of the multiple performances are defined as constraints, thereby extracting one or more performances and at least one of the one or more components that hinder the achievement of the objective as bottlenecks. Based on the extracted bottleneck information, a mechanism analysis unit analyzes the mechanism by which the bottleneck occurs, A structural analysis unit considers one or more design modification proposals for the structure based on the mechanism obtained from the analysis, so as to reduce the degree to which the bottleneck hinders the achievement of the objective, A countermeasure structure deriving unit that derives a countermeasure structure based on one or more of the above design change proposals, A determination unit that verifies the proposed countermeasure structure and determines whether the multiple performances in the proposed countermeasure structure satisfy the requirements and whether the proposed countermeasure structure achieves the objective, The system comprises a structure determination unit that determines the proposed countermeasure structure as the structure of the structure when the determination unit determines that the multiple performance characteristics of the proposed countermeasure structure satisfy the requirements and that the proposed countermeasure structure achieves the objective. A structural design support system characterized by the following features.

Images