JPH03224063A - Optimum designing system - Google Patents

Abstract

PURPOSE:To output an optimum model by providing an interface between an analyzing part and a modeling part and performing automatically the change of the design parameter of a structure and a remodeling job as necessary based on the model analyzing result. CONSTITUTION:A design parameter setting part 102 sets the initial value based on a range of the input design parameter by reference to an optimization rule data base. An analyzing data production part 105 produces an element model based on the design parameter to be applied and the initial form of a structure. An analyzing part 106 deforms and analyzes the element model, and a convergence deciding part 107 decides the deforming/analyzing result of the part 106. Receiving a result for a design parameter which is newly applied, the part 102 refers again to the optimization rule data base to output a new design parameter. The part 107 compares the old and new analyzing results with each other to hold the analyzing result more approximate to the design specifications. Then the part 107 ends its operation when the convergence is confirmed. Thus an optimum model is automatically obtained.

Description

[Detailed description of the invention] [Industrial application field] The present invention relates to an optimal design system using a computer.

[Conventional technology]

When creating an optimal model of a structure that meets design specifications using a conventional design system, modeling to create a model of the structure and analysis of the created model (e.g. static strength analysis and dynamic vibration analysis) are not automated. However, the user of the design system had to repeatedly enter the input for modeling the structure and judge the analysis results of the model. When modeling structures,
The user sets the design parameters such as the thickness, length, and material of the structural members one by one, but in order to satisfy the design specifications such as the strength of the specified part of the structure, the total weight, and the manufacturing cost, it is necessary to Design parameters must be determined taking into account trade-offs (reshuffling the priorities of design specification items), which places a heavy burden on the user.

[Problem to be solved by the invention]

In the above-mentioned conventional technology, since the part for modeling the structure and the part for analyzing the created model are separate, the user is responsible for all procedures for reflecting the analysis results in the modeling. Therefore, in a design system that requires repeating modeling and analysis until a model that satisfies the design specifications is obtained, there is a problem in that design parameter setting and modeling each time requires considerable effort.

The present invention provides an interface between the part that performs analysis and the part that performs modeling, automatically changes the design parameters of the structure and remodels it as necessary based on the model analysis results, and designs the structure after repeated calculations. The purpose is to provide an optimal design system that outputs an optimal model that satisfies specifications.

[Means to solve the problem]

The above purpose is to be used by the user to input the design specifications of the structure and the priority order of the design specifications in an optimal design system that uses a computer to create an element model of the structure and analyzes the element model. a design specification input section; an initial data input section used by the user to input the initial shape of the structure, the application range of design parameters, and boundary conditions; and creation of design parameters to be applied sequentially from the application range of the design parameters. a design parameter setting section that performs and outputs the
an optimization rule database that sequentially instructs changes in design parameters to be applied based on the priority order of the design specifications when creating the design parameters; and an optimization rule database that outputs the initial shape and the design parameter setting unit each time the design parameter setting unit outputs the optimization rule database. an analysis data creation section that creates an element model of the structure using the design parameters created by the analysis data creation section; an analysis section that analyzes the element model created by the analysis data creation section; an optimal design system, comprising: a convergence determination unit that determines convergence of the calculated analysis results based on the design specifications, and requests the design parameter setting unit to output new design parameters until convergence is achieved. This is achieved by

[Effect]

In the optimal design system of the present invention, the user uses the design specification input section to specify physical aspects such as strength, weight, deformation, etc. of the target structure as set specifications, and also uses the initial data input section to specify the physical aspects of the target structure such as strength, weight, deformation, etc. Input the initial shape, applicable range of design parameters such as plate thickness, dimensions, material, etc., and boundary conditions. The design parameter setting section first sets initial values from the input design parameter range by referring to the optimization rule database, and outputs them as design parameters to be applied, and the analysis data creation section sets the design parameters to be applied. An element model is created using the initial shape of the structure, the analysis section performs a deformation analysis on the element model, and the convergence judgment section uses the deformation analysis results to determine whether the element model passes or fails based on the priorities of the design specifications. In this case, since there are no previously analyzed analysis results, the design parameter setting unit is requested to output design parameters to be newly applied. Upon receiving the request, the design parameter setting section refers to the optimization rule database again, receives instructions for the amount of change in design parameters according to the optimization rule, outputs new design parameters according to the amount of change, and performs analysis. The data creation section and the analysis section operate as described above, and the convergence determination section compares the analysis results obtained previously and holds the analysis results that are closer to the design specifications.

As described above, the optimal design system repeatedly sets design parameters to be applied, creates analysis data, and performs analysis, and the operation of the optimal design system ends when the analysis results converge according to the design specifications.

〔Example〕

Embodiments of the present invention will be described below with reference to FIGS. 1 to 10. FIG. 1 is an overall configuration diagram of an optimal design system according to an embodiment. According to this embodiment, the user inputs the initial shape of the target structure, design parameters such as the dimensions and materials of the structure to be applied first, the allowable range of the design parameters, etc., and then designs the structure. Once specifications are input, the optimal design system creates an analysis model of the initial structure, performs, for example, static strength analysis, determines whether the analysis results satisfy the design specifications, and then... Using optimization rules based on the analysis results already obtained and the design specifications entered by the user, the values of design parameters are automatically corrected at a pitch within the allowable range, and then the analysis model (analysis conditions such as boundary conditions) is This method generates a mesh model (added mesh model), performs reanalysis, and outputs the optimal model that satisfies the design specifications when the analysis results finally converge, that is, when there is no need to modify the design parameters beyond the current stage. It is.

The initial data input unit 104 in FIG. 1 has a shape modeling function for the user to create an initial shape of the structure to be designed, and a function for interactively setting analysis conditions such as material properties and boundary conditions.

These have a structure that allows initial values and change ranges of design parameters (items to be modified such as dimensions and material property values) to be given.

The analysis data creation unit 105 automatically generates an analysis model (mesh model with analysis conditions added) when the shape of the structure and analysis conditions are given. First, the data created in the initial data input unit 104 is For initial analysis, the data is input to the analysis data creation unit 105, and an initial analysis model is created.

The analysis unit 106 performs, for example, deformation analysis on the initial analysis model created by the analysis data creation unit 105.

The convergence determination unit 107 performs a convergence determination between the analysis result obtained by the analysis unit 106 and the previously obtained analysis result, and checks whether the optimization has been completed. If the analysis results have not converged, or in the case of an initial analysis in which there is no comparison target for convergence determination, processing is transferred to the design parameter setting unit 102 to optimize the design parameters.

The design specification input unit 101 inputs constraint conditions for optimization,
This is for inputting the allowable range (for deformation amount, stress, etc.) and target values for the analysis results, as well as the priority order of design specifications.

After the first analysis model, the design parameter setting section 102 uses the analysis results of the analysis section 106 or previous analysis results and the design specification input section 1 to create the next analysis model.
Based on the design specifications input by the user in step 01, the optimization rule database 103 is searched to find countermeasures for the current analysis results, and the values of the design parameters specified in the initial data input section 104 are modified within the change range. It is.

Based on the design parameters modified by the design parameter setting unit 102 and the initial data (shape data, analysis conditions) input by the initial data input unit 104, the analysis data creation unit 105
creates a new analysis model, the analysis unit 106 performs the analysis again, and the convergence determination unit 107 determines the convergence of the analysis results.

The cycle from design parameter setting to convergence determination is repeated until the analysis results converge, and when the analysis results finally converge, the optimal model that satisfies the design specifications is displayed on the result display section 108.

Next, specific operations of each part will be explained. FIG. 2 shows an example of the initial shape of the model to be designed. FIG. 3 shows the equipment configuration and screen example for input in the design specification input section 101. Since input of design specifications is performed interactively with the system, a graphic display 301 is used as a display unit, and a pen 302 and a doublet 303 are used as input means. This can also be replaced with another input means such as a mouse. In the figure, graphic display 30
1, a design specification table 304 for setting the range of displacement of a structure is displayed as an example of design specification items, and a design target model 305 is displayed above it. In addition to the displacement amount shown in the example, the design specification table 304 includes velocity, acceleration, pressure, contact force, stress, strain, and temperature.

It is possible to set the range (maximum and minimum amount) of heat flux, etc. Furthermore, the design target model has a structure in which the user can arbitrarily specify each component, or each component surface or line.

FIG. 4 is a diagram showing a design specification table in which displacement is set as a design specification item. One side 401 of the model to be designed
To set the range of deformation to ±51m in the three-dimensional Z direction,
5 in the Z direction/max column in the row with displacement number 1 in the design specification table 402. After inputting -5 in the win column and, in this case, 0 in the target value column, specifying the side 401 and assigning the displacement amount of displacement number 1. The displacement number will be displayed near the specified line on the screen. This operation is similar to other design specifications including design specifications such as surface area and weight that are not related to the analysis results. Note that if a target value is not input, it is treated as a constraint condition for optimization.

FIG. 5 is a diagram showing a design specification priority setting table for setting the priority of each design specification input by the user. A list of design specifications in which target values have been input is displayed together with serial numbers assigned to each design specification table, so enter the priority order of the design specifications in the priority column. When performing optimization, this is to first perform processing such as changing the shape so that it approaches the target value of the design specification with a high priority.

The initial data input unit 104 is an interactive modeler for providing an initial shape model and analysis conditions such as boundary conditions. In order to input the dimensional range of the part related to the shape change in advance, the dimensional range as shown in FIG. Use a configuration table.

When setting this dimension range, consider models that require common stretching operations on multiple line elements located symmetrically to each other, such as symmetrical structures, and group line elements that will be symmetrical for dimension changes. . In the dimension range setting table, the "line group number" refers to a number uniquely assigned to grouped line elements. Symbols ″- (minus) and lt
I+ (plus) indicates the direction of dimension change, and a shape change in the direction in which the area (or volume) decreases due to dimension change is negative, and a shape change in the direction in which it increases is positive. If the path step value is specified within this range, the line element will be expanded or contracted by adjusting the increase or decrease in dimension to be an integral multiple of the step value.When specifying the line element to be changed in dimension, is an input/output means that can determine which end point to move among the two end points that make up the line element.Also, if the specified line element intersects the center line, the line element Both end points of are used as moving points, and the center is fixed and the expansion/contraction process is performed symmetrically.

Figure 7 shows an example of a model that is subject to shape change, and shows 70
1 and 702 are grouped line elements, and black circles on each line element indicate moving end points. On the other hand, a maximum shape 801 and a minimum shape 802 when a certain deformation range is given are shown in FIGS. 8 and 9, respectively. However, in these figures, the broken line portion represents the initial shape.

The design parameter setting unit 102 is for changing and setting the dimensions and materials of the structure based on the analysis results.
An optimization rule database is used for this determination.

The optimization rule database 103 includes both a knowledge base that stores empirical rules and a nonlinear programming program that performs optimization mathematically based on mathematical programming, which can be used depending on the problem. .

First, an outline of optimization using the knowledge-based method will be explained with reference to FIG. 1o. Search for optimization rules is performed according to the priority order of each design specification input in the design specification input unit 101. For example, in a certain optimal design, the surface area of the structure is minimum (tolerable limit α) and the maximum principal stress β is below (target value 0
), these require conflicting solutions for optimization, since stress generally increases as surface area decreases. Therefore, operations such as shape changes are performed so that the design specifications with higher priority are satisfied first. Finally, the design parameter setting section 10
In step 2, the specific numerical value (amount of change) or material (material characteristic value) of the portion for which the dimensional change range is specified is determined and output to the analysis data creation unit 105.

Next, the nonlinear planning method will be explained.

A nonlinear programming program is a program that automatically finds design parameters that minimize (or maximize) the objective function while satisfying the 9 constraints when the objective function and constraints are expressed as nonlinear functions of design parameters. . now,
For the specification with the first priority, the absolute value of the difference between the target value and the current value may be used as the objective function, and the other specifications may be input into the nonlinear programming program as constraints. At this time, the rate of change of the objective function and constraint function with respect to changes in design parameters is determined. Calculations can be carried out efficiently by using the so-called sensitivity analysis function.

The analysis section data creation section 105 includes the initial data input section 104
An element model suitable for analysis is automatically created from the initial shape data and analysis conditions input in , or data output from the design parameter section 102 . Analysis conditions such as boundary conditions are added to this, but in the finite element method, analysis conditions are given to each element or each node that constitutes an element. The analysis condition input in the system of this example is given to the block (solid), surface, line, and node unit of the shape model, and the geometric structure of the shape model (the number of surfaces and lines) is , bond state) do not change, the analysis conditions for the initial shape given by the initial data input unit 104 are also valid for the changed shape. Note that, for example, when applying a load to the center of a surface and changing the size of the surface, the surface is divided into four parts in advance, the line elements constituting these parts are grouped together, and a load condition is applied to the intersection of the four surfaces.

Initial shape model 1001 with analysis conditions shown in FIGS. 11, 12, and 13, respectively. Initial shape model 10
An element model 1002 obtained by decomposing 01 into finite elements, and an element model 1003 whose shape has been changed through the design parameter setting unit are shown. If the shape is different, the structure of the element model (the number of divisions of each surface, the total number of elements) will also change, but the analysis conditions are set to the part intended by the user.

The analysis data generated through the above procedure is stored in the analysis unit 106.
The analysis unit 106 performs static analysis, vibration analysis, thermal analysis, fluid analysis, etc. Then, by inputting the analysis results, the convergence determination unit 107 checks whether there are any changes from the previous analysis results. If there are no changes, the optimization has been completed, and the final structure and material of the model are determined. Display section 1
Display it on 08. If the analysis results have not converged,
Since there is a possibility of further optimization, the decision convergence unit 10
Step 7 returns the processing to the design parameter design unit 102, where optimization is again attempted. Ultimately, this system provides an optimal model that satisfies the design specifications entered by the user.

As explained above, in the optimal design system of this embodiment, the analysis section and the modeling section, which conventionally operated separately, are integrated, so that the analysis results are reflected in the modeling. The modeling part of this system is an interactive modeler (this corresponds to the initial data input part) that allows the user to input the initial shape data of the structure and analysis conditions such as boundary conditions.
It consists of an internal modeler (corresponding to the analysis data creation section) in which a computer automatically performs modeling such as shape changes based on design parameters.

Perform analysis using the data created in the modeling section, and if there are any items in the analysis results that do not meet the design specifications,
Perform trade-offs with other items according to optimization rules, change design parameters such as material reinforcement, cutting, and material changes, generate input data for analysis, and perform analysis again until the design specifications are satisfied. It is something to do.

In addition, the design parameter setting values are set by the user, such as the part that requires shape change or the range of dimensional change.
A value completely unintended by the user is never set. Modify the shape according to the design parameters, generate data in the same format as that created with the interactive modeler, generate a model suitable for the analysis method based on this data, perform analysis, and compare with the history of analysis results. Since convergence is determined by , an optimal model that finally satisfies the design specifications can be obtained.

〔Effect of the invention〕

According to the optimal design system of the present invention, when a user inputs the design specifications, initial shape, application range of design parameters, and boundary conditions of a structure, an optimal model that satisfies the design specifications is automatically obtained using a computer. This has the effect of significantly shortening data input time and calculation time compared to when the user performs modeling and analysis for each case. Furthermore, the analysis results are judged as objective points and optimization is performed until convergence, which has the effect of improving the reliability of the optimal model.

[Brief explanation of drawings]

Figure 1 is an overall configuration diagram of the optimal design system of the present invention;
The figure shows an example of the initial shape of the model to be designed, Figure 3 shows the equipment configuration and screen example in the design specification input section, and Figure 4 shows an example of setting the amount of displacement in the design specification input section. , Fig. 5 shows a table for inputting the priority order of each design specification, Fig. 6 shows a table for setting the dimension range in the initial data input section, and Fig. 7 shows a table for setting the dimensional range in the initial data input section. Figures 8 and 9 are diagrams showing an example, respectively, showing a model in which the shape has been changed from the model in Figure 7, Figure 1o is a simplified diagram of the optimization rule database, and Figures 11 and 12 are diagrams showing an example. The figure and FIG. 13 are diagrams showing an initial shape model given analysis conditions, its element model, and a modified element model, respectively. 101... Design specification input section, 102... Design parameter setting section, 103... Optimization rule database section, 104... Initial data input section, 105... Analysis data creation section, 106... - Analysis section, 107... Convergence judgment section, 108... Result display section, 402... Design specification table.

Claims (1)

[Claims] 1. In an optimal design system that uses a computer to create an elemental model of a structure and analyzes the elemental model, a design specification used by the user to input the design specifications of the structure and the priority order of the design specifications. an input section, an initial data input section used by the user to input the initial shape of the structure, the application range of design parameters, and boundary conditions; an optimization rule database that sequentially instructs changes in design parameters to be applied based on the priority order of the design specifications when creating the design parameters; an analysis data creation unit that creates an element model of a structure using the design parameters output by the shape and design parameter setting unit; an analysis unit that analyzes the element model created by the analysis data creation unit; and a convergence determination unit that determines convergence between the analysis results analyzed by and the analysis results that have already been analyzed based on the design specifications, and requests the design parameter setting unit to output new design parameters until convergence is achieved. An optimal design system characterized by: