JP3289230B2 - Generator of finite element mesh distribution - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of preparing analysis input data for executing an analysis simulation by the finite element method, and more particularly to a method of specifying a finite element mesh coarse / dense distribution for improving analysis accuracy.

[0002]

2. Description of the Related Art There are the following two conventional methods for specifying a dense / dense distribution of a finite element mesh model. One method is to control the density of the finite element mesh model by dividing the shape to be analyzed into blocks such as a hexahedron in advance and inputting the blocks to a computer, and specifying the number of divisions and the division ratio of the ridge lines of the blocks. . For example, to generate a partially fine mesh, the area is defined in advance as a block,
Generate a fine mesh by specifying a large number of edges. When a cube is targeted, by specifying the number of divisions for each of the three ridges of length, width, and height, nodes defined by the intersections of the three-way division planes are studded in the cube. The mesh is divided based on the studded nodes.

Another method is a system that inputs a shape to be analyzed as a shape model to a computer and generates a finite element mesh model by automatic meshing.
This method controls the density of a finite element mesh model by designating a reference element size (interval between nodes) and the number of divisions of an edge (or an element size on an edge).

[0004] As described above, the conventional method of designating the dense / dense distribution of a finite element mesh model is described, for example, in the pre-
It is described in the post-processor HICAD / CADAS / W command manual (Hitachi, February 1990).

[0005]

In the conventional method for specifying mesh density distribution, it is necessary to divide a region for specifying mesh density into blocks such as a hexahedron in advance. Not only has a great deal of labor been spent on input work, but it has been difficult to perform the intended mesh division. When the density of the mesh is changed, the block must be divided again.

On the other hand, when a finite element mesh model is created using an automatic meshing function, the target for which the mesh density can be specified is limited to the ridge lines and vertices of the shape model, and the finite element mesh model intended by the user is specified. There was a problem that it could not be created freely.

It is an object of the present invention to provide a method of specifying a mesh coarse / dense distribution such that a finite element mesh model intended by a user can be generated accurately.

[0008]

In order to achieve the above object, vertices, ridges, surfaces, and the like constituting a shape model to be analyzed are described.
For a three-dimensional part (hereinafter, these are called shape elements), a function (hereinafter, this function is called a mesh coarse-dense distribution function) of the distance from the shape elements and the distribution density of nodes is designated, and the designated mesh is designated. Nodes on the surface and / or inside of the shape model are defined based on the density distribution function. This facilitates control of the element size not only for the shape element itself but also for the vicinity of the shape element.

[0009]

[0010]

According to the method of specifying the mesh density distribution of the present invention, the user selects a shape element from the shape model to be analyzed displayed on the display using an input device such as a mouse,
The mesh density distribution function is input for the selected shape element. Next, when the user inputs a mesh division command, nodes are generated in the shape model and / or on the surface by the mesh density distribution function specified for the shape element. Then, nodes constituting the finite element are extracted so that the elements do not overlap, and a finite element mesh model is automatically generated.

[0011]

[0012]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic operation procedure from a shape definition to a mesh generation when a shape to be analyzed is represented by a single shape block, and FIG. 2 is a shape when a shape to be analyzed is represented by superposition of a plurality of shape blocks. The schematic operation procedure from definition to mesh generation is shown.

First, the former case will be described with reference to FIG.
Example) will be described. When the user inputs a shape model to be analyzed (in this case, a single shape block) to the computer using a data input device such as a keyboard or a mouse, the input shape model is displayed on a display (101). . Next, a shape element of the shape model displayed on the display is designated (102), and a mesh coarse / dense distribution function is input (103). Specifically, as shown in FIG. 3, when a shape element (a ridge line in this case) of a shape model is designated using a mouse (301), the vertical axis represents the distribution density ρ of the nodes, and the horizontal axis represents the shape density from the designated shape element. A graph showing the mesh density distribution function for the shape element by taking a distance is displayed on the display (302), and when a combination of the distance from the specified shape element and the distribution density of the nodes is input (302), the input is performed. The mesh coarse / dense distribution function connecting the points represented by the selected combination is displayed (30).
3). At 303 in FIG. 3, the points of the input combination are displayed as black circles. At this time, when there is a mesh coarse / dense distribution function already specified (input) for another shape element, the priority order of the mesh coarse / dense distribution function inputted this time is specified for the mesh coarse / dense distribution function. With this priority order, the mesh coarse / dense distribution function of the area where the mesh coarse / dense distribution functions overlap is determined. Judge whether the designation of mesh density distribution in the block is completed (1
04) If not completed, steps 102 to 104 are repeated. When the mesh density distribution within the block is specified, nodes are generated in the shape block based on the specified mesh density distribution function according to the user's instruction for mesh generation (105), and the shape model is generated based on the generated nodes. , A finite element mesh model is automatically generated (106).

On the other hand, in the case of the second embodiment, FIG.
As shown in (1), the operation from the definition of the shape block to the designation of the mesh coarse / dense distribution is repeated by the number of blocks to be superimposed (that is, for each defined shape block) (201, 202, 203, 204). When the definition of the shape block and the designation of the mesh coarse / dense distribution for the defined shape block are completed (205), the position of the shape block is superimposed (206), so that the shape model formed by combining a plurality of shape models is obtained. Specify the mesh density as At this time, it is specified whether to add the mesh density distribution functions of the overlapping shape blocks or to give priority to the mesh density distribution function of one of the shape blocks. Finally, a mesh generation instruction (2)
07), nodes are generated in the shape model (208), and a finite element mesh model is automatically generated (20).
9).

Next, the specification of mesh density distribution and automatic generation of a finite element mesh will be described in detail.

A shape model to be analyzed is composed of a set of shape elements such as a solid, a plane, a line, and a point. The mesh density distribution function can be specified for each shape element. FIG. 4-
FIG. 7 shows the relationship between the shape element and the mesh distribution function. For example , as in the reference example shown in FIG.
In the case of, the distance r from the point p and the radius r around the point p
Specifies the distribution density of nodes on the spherical surface of (or element size, that is, the interval between nodes). Similarly, when the shape element is a line, a plane, or a solid, the distribution density (or element size) of the nodes on the closed curved surface having the same distance from the shape element and the distance from the shape element is specified. FIG. 8 shows an example of a data structure representing the mesh density distribution function. One row of the data table in FIG. 8 represents a mesh coarse / dense distribution function for one shape element. Each mesh coarse / dense distribution function has a priority order, and a mesh coarse / dense distribution function with a high priority order is adopted for an overlapping area. When the priority order is the same, the mesh coarse / dense distribution function is determined by adding both the mesh coarse / dense distribution functions, or one of them is selected.

FIG. 9 shows an algorithm of a process for generating a finite element mesh model from a shape model in which a mesh distribution function is specified. First, nodes are generated according to each mesh coarse / dense distribution function for each shape block (901), and nodes located outside the shape model that is to be element-divided are removed (902). For example, as shown in FIG. 10, the mesh coarse / dense distribution function (1001) has a line (1001).
When defined for 2), spacing the lines h _0, distance h ₁ is at a position r ₁ away from the line, r ₂ from the line
The remote location generating a node at intervals of h _2. next,
If there is an area where the mesh density distribution functions overlap, nodes are removed according to the priority order of the mesh density distribution functions (903).

At this time, if the priority order is the same, a selection is made as to whether to leave nodes generated from both mesh coarse / dense distribution functions as they are or to leave only nodes generated from one mesh coarse / dense distribution function. FIG. 11 shows the generation of nodes when the shape block 2 has a higher priority of the mesh density distribution function than the shape block 1. On the other hand, FIG. 12 shows the generation of nodes when the selection is made to leave the nodes generated from the mesh coarse / dense distribution functions of both the shape blocks when the priority order of the distribution functions of the shape blocks 1 and 2 is the same. Finally, the nodes are connected using the Delaunay method or the like to generate a finite element mesh (904).

By the mesh coarse / dense distribution specifying method and the finite element mesh generation processing as described above, it is possible to easily specify the mesh coarse / dense distribution even in the vicinity of a shape element that requires mesh density control in analysis. Become so.

[0020]

According to the present invention, the user can freely control the mesh density of the finite element mesh model simply by specifying the shape element and inputting the mesh density distribution function to the shape model to be analyzed. You. Thus, conventionally, it has been necessary to divide the shape into blocks in order to control the density of the mesh, but the labor and time for these can be greatly reduced. Further, since the mesh density distribution can be designated by superimposing the shape models, the mesh density can be easily changed. Further, since the distribution of nodes is specified by the mesh density distribution function according to the distance from the shape element, accurate mesh density control can be performed even in the vicinity of the shape element.

[Brief description of the drawings]

FIG. 1 is a flowchart showing a first embodiment of the present invention.

FIG. 2 is a flowchart showing a first embodiment of the present invention.

FIG. 3 is an explanatory diagram of a method for specifying a mesh coarse / dense distribution to a shape element according to the present invention.

FIG. 4 is a perspective view showing a mesh density distribution control area in a method of specifying a mesh density distribution for a shape element according to a reference example of the present invention.

FIG. 5 is a perspective view showing a mesh coarse / dense distribution control area in a line of the method for specifying a mesh coarse / dense distribution to a shape element according to the present invention.

FIG. 6 is a perspective view showing a mesh density distribution control area in a method of specifying a mesh density distribution for a shape element according to the present invention.

FIG. 7 is a perspective view showing a mesh coarse / dense distribution control area in a three-dimensional object in the method of designating a mesh coarse / dense distribution to a shape element according to the present invention.

FIG. 8 is a diagram showing an example of a data structure representing a mesh distribution function according to the present invention.

FIG. 9 is a flowchart showing an algorithm for generating a finite element mesh according to the present invention.

FIG. 10 is an explanatory diagram of a node generation method using a mesh coarse / dense distribution function according to the present invention.

FIG. 11 is a plan view showing an example of occurrence of nodes when the priority order of two mesh coarse / dense distribution functions is different.

FIG. 12 is a plan view showing an example of occurrence of nodes when the priority order of two mesh coarse / dense distribution functions is the same.

[Explanation of symbols]

 1001 mesh coarse-dense distribution function 1002 shape element (line) r distance from shape element ρ distribution density of node h element size

Continuation of front page (56) References JP-A-5-46712 (JP, A) C. Lin et al. , An expert system for forming quadrilatal eternal fineelements, Engineering Computers, UK, Pineridge Press, vol. 7, no. 3, September 1990, pp. 249-257 Hideomi Ohtsubo et al., Journal of the Shipbuilding Society of Japan, Recent trends in the finite element method (Part 4) Two-dimensional automatic element segmentation algorithm and preprocessing, Japan, No. 702, December 1987 Pages 819-830 A. Fernandez et al. , A Simple Adaptive Mesh Generator for 2-D Finite Elemental Calculations, IEEE Transactions on magnetics, USA, vol. 29, no. 2, March 1993, pp. 182-1885 Hidemi Nagai et al., Adaptive automatic element segmentation using the technique of randomly distributed random numbers (effectiveness of using random numbers), Proceedings of the 70th Annual Meeting of the Japan Society of Mechanical Engineers, A , Japan, September 30, 1992, pp. 208-210 (58) Fields investigated (Int. Cl. ⁷ , DB name) G06F 17/50 612 JICST file (JOIS)

Claims (2)

(57) [Claims] 1. An analysis method using a computer which inputs a shape model to be analyzed, defines nodes in the shape model, and automatically generates a finite element mesh model from the shape model using the defined nodes. In the input data creation device, one of ridge lines , faces, and solids, which are the shape elements of the shape model, is specified, and a function of the distance from the specified shape element and the distribution density of the nodes is interactively input. An apparatus for generating a finite element mesh density distribution, comprising: a mesh density information input unit; and a node generation unit that defines a node in the shape model from the input mesh density information. 2. A finite element mesh coarse / dense distribution generating apparatus according to claim 1 , wherein a finite element mesh model of the shape model is generated from a node group generated based on mesh coarse / dense information input by the apparatus. An input data creation device for analysis characterized by automatic generation.