JPH04679A - Method and device for supporting coordinate grid preparation - Google Patents

Abstract

PURPOSE:To efficiently attain determining plural coordinate grids with different density by preparing a second coordinate grid whose interval is different from a first coordinate grid in the partial area of the first coordinate grid and preparing further new second coordinate grid with the display of corresponding relation. CONSTITUTION:Form data are inputted by using a general coordinate grid preparation supporting CAD system and the grid division of an area where analysis is executed in an analysis object. The end/not end of multiple grid preparation is judged and when it is judged to be not end, plural coordinate grid groups with different density are prepared and displayed. The second coordinate grid is prepared corresponding to the result of whether it is coarse or fine and it is stored in an external memory. Thus, the coordinate grid corresponding to the analysis object and an analysis purpose can be efficiently generated.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a method and an apparatus for creating a coordinate grid within an analysis area necessary for numerical analysis using a computer, and particularly relates to a multidimensional, coarse-grained It is suitable for reducing human errors when creating multiple coordinate grid groups with different values, reducing the work time required to create coordinate grids, and creating the optimal coordinate grid group for analysis depending on the analysis purpose. The present invention relates to a coordinate grid creation support method and apparatus.

[Conventional technology] There are the following two conventional examples of the multi-grid method.

1. Tadashi Arayo and 2 others, "High-speed separation method for Navier-Stokes equations using multiple grid method" (Transactions of the Japan Society of Mechanical Engineers (ed. B), Vol. 54, No. 496 (1986-2), p.

290 papers & 87-0256A).

2, Akira Fujimoto, Makoto Iida - “Analysis of two-point boundary value problems using multiple lattice methods Jl (Proceedings of the Japan Society of Mechanical Engineers (B edition)) Vol. 54, 50
No. 0 (Sho 63-4) ρ, 849 Paper NQ 87-082
9B).

With the development of computers, various numerical simulation methods have been developed in fields such as structural analysis, fluid analysis, and electromagnetic field analysis, and are being put into practical use as general-purpose numerical analysis programs using the finite element method, finite difference method, etc. . Among them, the multi-grid method shown in the above-mentioned known example is rapidly spreading especially in the field of fluid analysis because of its high convergence performance. In this method, physical quantities are not solved using a single coordinate grid as in the conventional method, but are analyzed using a plurality of coordinate grid groups with different density. In order to achieve highly accurate analysis, it is necessary to create a group of coarse and dense grids that are optimal for the analysis target, but since the method of creating coarse and dense grids is highly arbitrary, known techniques cannot efficiently There is no way to create these coordinate grids.

On the other hand, the so-called CAE (Computer A
With the development of 1ded E n-gneering) systems, it has become possible to interactively model objects with complex shapes. In order to consistently perform analysis simulations from geometric modeling, these systems allow the creation of coordinate grids for use in analysis.

[Problem to be solved by the invention] However, conventional CAE systems target the difference method, finite element method, etc., and ultimately aim to create a single coordinate grid. It is not possible to create groups, and there are limits to its application. In other words, some conventional systems have functions such as redividing and thinning out the created coordinate grid, but after the operation, the created coordinate grid is treated as multiple coordinate grids and the correspondence between the two is displayed. There is no function to record or store information. When using the multi-grid method, it is essential to display and store the correspondence between a plurality of coordinate grids of different density, and in this sense it is difficult to use a conventional CAE system.

As described above, in the conventional method described above, the purpose is to ultimately create a single coordinate grid, and it is not possible to create a plurality of groups of dense and coarse coordinate grids and display or store their correspondence. Therefore, it does not have a display function for confirming the shape of the coarse/fine coordinate grid, and it is difficult for an analysis operator to interactively and efficiently create a plurality of coarse/fine coordinate grids according to the purpose of the analysis operator. - On the other hand, as the target problem becomes multi-dimensional and the structure becomes complex, it is necessary to use a coordinate grid group (multiple grids) that fully reflects the experience of the analysis worker and is suitable for use without a CAE system.
It becomes difficult to decide efficiently. Therefore, it is necessary to construct a method and a CAE system for efficiently creating multiple grids that meet the purpose of analysis during interactive processing.

Conventional examples of creating coordinate grids of different density include Japanese Patent Application Laid-Open Nos. 46374-1982 and 170777-1987. However, in these conventional examples, there is no description of displaying or storing the correspondence between coordinate grids of different density.

An object of the present invention is to provide a method for creating multiple coordinate grids necessary for calculations using the multiple grid method, in order to reduce calculation time for target physical quantities. An object of the present invention is to provide a support method and device for efficiently making decisions.

[Means for Solving the Problems] The present invention includes a first step of dividing an analysis region to create a first coordinate grid, a second step of storing the first coordinate grid, and a second step of storing the first coordinate grid. a third step of displaying a grid on a display screen, a fourth step of creating a sub-region of the first coordinate grid, and creating a second coordinate grid with a different spacing from the first coordinate grid in the sub-region. a fifth step, a sixth step of storing the correspondence between the first coordinate grid and the second coordinate grid, and a sixth step of displaying the correspondence between the first coordinate grid and the second coordinate grid on a display screen. It consists of 7 steps.

Furthermore, the present invention includes a first means for dividing an analysis region to create a first coordinate grid, a second means for storing the first coordinate grid, and a second means for displaying the first coordinate grid on a display screen. 3, and a fourth means for creating a partial area of the first coordinate grid.
a fifth means for creating a second coordinate grid having a different interval from the first coordinate grid in the partial region; and a sixth means for storing a correspondence relationship between the first coordinate grid and the second coordinate grid. means and
and seventh means for displaying the correspondence between the first coordinate grid and the second coordinate grid on a display screen.

[Operation] According to the present invention, a second coordinate grid having a different interval from that of the first coordinate grid is created in a partial area of the first coordinate grid, and the correspondence relationship between these grids is further displayed. By displaying this correspondence relationship, a new second coordinate grid can be created.

[Example] The present invention will be explained below using preferred embodiments.

FIG. 2 shows the configuration of a coordinate grid creation support device 1 which is an embodiment of the present invention. As shown in the figure, the support device 1 of this embodiment includes display devices (for example, CRT) 2A and 2B,
It has image processing devices 3A and 3B, an arithmetic processing device (for example, an electronic computer) 4, external storage devices 5a and 5b, and an operation panel (for example, a keyboard) 6. Display devices 2A and 2B
is connected to the arithmetic processing device 4 via image processing devices 3A and 3B including one-image data storage units 3C and 3D, and the storage device 5 and operation panel 6 are also connected to the arithmetic processing device 4.

The arithmetic processing device 4 includes a calculation section 4a, a processing procedure storage section 4b,
Intermediate data storage section 4c, input section 4d, coordinate grid data storage and input/output section 4e, image data input/output sections 4f and 4g
, a shape data input/output section 4h. These devices are connected around a calculation section 4a, and an input section 4d.
, input/output section 4g, 4h.

4e is connected to the calculation section 4a, and storage sections 4c and 4b are also connected to the calculation section 4a. The processing procedure storage unit 4b stores the processing procedure shown in FIG. The operation panel 6 is connected to the input section 4d. Image display devices M3A and 3B are connected to image data input/output units 4f and 4g.

The external storage device 5a stores the correspondence between coarse coordinate grids and dense coordinate grids, and the coordinate grids. The external storage device 5b stores shape data. The external storage device 5a stores data input/output units 4e and 5.
4h is connected to b.

In FIG. 1, the processing procedure storage unit 4h of this embodiment stores,
A processing procedure (program) executed by the calculation unit 4a is shown. The analysis operator inputs shape data from the external memory 5b for the object to be analyzed using the keyboard 6 (step 7A), and then inputs the shape data for the object to be analyzed (for example, a pipe through which fluid flows).
In step 7, the area to be analyzed is divided into grids (step 7).
The processing in step 7 of creating a coordinate grid including step B) is performed using a general coordinate grid creation support CAD system (for example, Japanese Patent Application No. 60-212985, Japanese Patent Application No. 61-1064).
Technique for creating a curved coordinate grid using coordinate transformation No. 66)
This can be done using Next, in step 7C,
It is determined whether multiple grid creation is completed or not. Step 7
If C is determined to be NO (not completed), a plurality of coordinate grid groups (hereinafter referred to as multiple grids) with different density are created and displayed in step 8 using the support device of this embodiment. Step 8
Next, determine whether it is from dense to coarse (step 8A) or from coarse to dense (step 8C), and depending on the result, create a second coordinate grid, either a coarse coordinate grid or a dense coordinate grid, and display this. It is displayed on the device 2A and stored in the external memory 5a (steps 8B and 8D). Step 8A and Step 8
Processing C is performed by inputting "" from the keyboard 6 by the operator.
This is done based on the fine and coarse instruction signals. The creation of the second coordinate grid is performed within the partial area of the first coordinate grid obtained in step 7B (the area in which the grid spacing within the first coordinate grid is changed), or if the determination in step 9A is YES, step 9 The residual error is determined by the physical quantity calculation in step 9B, and the obtained residual error is displayed on the display device 2A together with the coordinate grid in step 9C, and is roughly displayed in step 8B based on the operator's instruction information via the keyboard 6. Create or modify grids.

If the determination in step 9A is No, the physical quantities are displayed on the display device 2A together with the close-packed coordinate grid through the processes of step 9D, which executes physical quantity calculations on the grid in step 9, and step 9E, which displays the physical quantities obtained by the calculation. In step 8D, a close-packed coordinate grid is created which is further refined by one step. If YES (end) is determined in step 7C, physical quantities are calculated and displayed using steps 10 and 11 using the finally obtained multiple grid.

Physical quantity calculations can be performed as follows. As an example, calculations using the Poisson equation shown below, such as potential or fluid pressure, are taken.

Eψ=ρ (1) The coordinate grid includes a plurality of coordinate lines that intersect with each other. In a two-dimensional coordinate grid, an area corresponding to one square is called a control volume, and in a three-dimensional coordinate grid, an area corresponding to a solid defined by six squares located three-dimensionally adjacent to each other is called a control volume.

If the analysis region to which equation (1) is applied in the object to be analyzed is divided into three-dimensional coordinate grids and the concept of control volume is applied, the discrete equation can be expressed as follows using Glass's divergence theorem.

Note that the coordinate lines that define one square in the aforementioned two-dimensional coordinate grid and one solid in the three-dimensional coordinate grid are referred to as a unit grid.

Here, inn is the adjacent grid point number of ijk, r is the coordinate value that defines ψ, and St is the area of the control volume surface (control volume length in a two-dimensional coordinate grid)
and the product of the medium parameter, ■ is the control volume volume (control volume area in two-dimensional coordinate grid)
So, in the end.

, inn Σ A ψ +A, ψ, =b...
・(3) inn ijk 1au1xjk l
The simultaneous-order equation Ax=b has the form jk ijk, and ψ on each lattice can be determined by inverse matrix operation.

Therefore, if one coordinate grid (single grid) is given, the physical quantity ψ can be calculated on it. The center of the control volume can be moved on or between coordinate lines in response to calculations.

However, when solving A x = b especially under the Neumann condition, convergence is extremely slow in an iterative solution method such as the SOR method, and the difference (error) between the calculated value and the solution during the iteration does not attenuate easily. High-frequency error fluctuation components whose sign changes from grid to grid attenuate quickly, but low-frequency errors in particular decay extremely slowly. Therefore, after some repetitions, a smooth error distribution remains, and the attenuation of this error distribution is slow. On the other hand, since this error distribution is smooth, the error can be reduced by calculating the above physical quantities using a coarse grid. The difference from the solution obtained can be corrected. In addition, solving with a coarse grid not only reduces the number of grid points, but also shifts the error distribution to the higher frequency side.
Convergence can be accelerated.

Difference between unconverged iterative calculation value x1 of dense grid and solution X Δx=x
-X1 is the equation A(Δx)=b-Ax.

=r is satisfied. r is called the residual. Dense lattice equation A
After determining the averaging operator C for x=b and the residual r, and the complementary operation for the coarse grid solution X' to the dense grid, we obtain the coarse grid simultaneous -
The following equation A' x ''' b ' can be created. That is, CA1x' = Cr + CAIP
A' = CAl (5) Here, P is a projection operator of physical quantities on a dense coordinate grid onto a coarse coordinate grid, and is usually a dense coordinate grid located at the center within the control volume of the coarse coordinate grid. Assign the value in the control volume. Also, normally, a weighted average using the control volume is used for C, and substitution without interpolation or an arithmetic average of adjacent grids is used for (2). Using such a simple operation, the time required to create A' can be shortened. Obtain a new initial value of A x = b on the dense grid by adding the correction amount obtained by solving the simultaneous-order equation A'X""b' on the coarse coordinate grid to the calculated value on the dense grid. I can do it. Based on this initial value, equations in the dense grid are calculated.

By repeating such calculations between the fine coordinate grid and the coarse coordinate grid, calculation time can be reduced compared to using only the fine coordinate grid. If convergence is slow even with a coarse coordinate grid, it is possible to accelerate convergence by using a coarser coordinate grid.

In order to realize the iterative solution of discrete equations using the general multi-grid method described above, it is necessary to use multiple grids, the matrix of each grid,
It is necessary to prepare a source term.

Having already described the method of determining the matrix and source terms, below we present an example of a method and apparatus for creating a grid.

Now, suppose that the coordinate grid in the target analysis area is a general coordinate grid creation support CAD system or existing technology (for example, Japanese Patent Application No. 60-212985, Japanese Patent Application No. 61-1064).
Technique for creating a curved coordinate grid using coordinate transformation No. 66)
The following steps will be described as a supporting method of the present invention. As a method for creating multiple grids, we devised a method to create coarse grids hierarchically by thinning out the coordinate lines of fine grids.

In addition, we devised a method in which the coarsest grid is first created, and then a dense grid is created in one layer by interpolating the coordinate lines. We also devised a method to create a coarse grid hierarchically based on the residual distribution in a dense grid, and a method to create a dense grid based on the distribution of physical quantities in a coarse grid. In particular, the latter is useful for creating an optimal dense grid while calculating the initial values of the dense grid. In order to make it possible to create multiple grids as described above, when a coordinate grid of a certain stage is obtained in step 7, the calculation unit 4a displays on the screen of the display device 2A the "1. A menu of 2. Fine → Coarse, 3. Coarse → Fine, 4. Residual display, 5. Physical quantity display is displayed. When the analysis operator selects this menu item and instructs it using the keyboard 6, the processes of steps 8 and 9 are executed to create a corresponding coarse or dense grid.

First, a case will be described in which an analysis operator creates a coarse coordinate grid from a dense coordinate grid in multiple stages based on experience. The first coordinate grid created is considered to be a close-packed coordinate grid.

In creating a rough coordinate grid (step 8B), the coordinate grid previously stored in the external memory 5a is displayed again on the screen of the display device 2A. This coordinate grid will be referred to as the current level coordinate grid (basic coordinate grid). A coarse grid is sequentially created by thinning out some of the coordinate lines of the current coordinate grid. FIG. 3 shows an example of the current coordinate grid 14 displayed on the apparatus of this embodiment. The procedure for creating a coarse coordinate grid will be explained. As shown in FIG. 4, the analysis operator uses the keyboard 6 or a light pen to draw one side (one coordinate line) 17 of the coordinate grid 14 at the current stage on the screen of the display device N2A.
Specify the range from 18 to 19 on that side (partial area)
is specified, and the number of coordinate lines within the range to be thinned out is specified by picking it in the parameter input table 15. Next, step 8
In process B, a coarse coordinate grid is calculated in the specified partial area based on this information (for example,
1-281773), the coarse coordinate grid 16 after thinning out the specified coordinate lines, as shown in FIG.
is displayed on the display devices i2A and 2B. Through such processing, as shown in FIG. 4, a coordinate grid in which the current coordinate grid 14 and the rough coordinate grid 16 are mixed is obtained. or
As shown in FIG. 5, the coordinate grid 16 after thinning out is displayed on the display device 2A together with the thinned out coordinate lines 14A (broken lines). next.

Specify another range on the same side, or another coordinate line that intersects with this side, and the range on this coordinate line, and similarly calculate and display the coordinate grid after thinning out the specified coordinate line. .

This operation is repeated until a satisfactory coordinate grid is obtained based on the experience of the analysis operator, and the correspondence relationship of the satisfactory coarse grid to the dense grid is stored in the external memory 5a. Alternatively, the coarse grid is also memorized. Next, by considering the obtained coarse grid as the coordinate grid at the current stage and repeating the above operations, it is possible to create a set of coarse and dense coordinate grids in multiple stages (multiple grid).

One side (or coordinate line) is specified by picking with a light pen, a mouse, etc., and a range is specified by picking a coordinate line (or coordinate plane in three dimensions) that intersects side 17.

The number of coordinate lines to be thinned out is entered by either inputting every other line to be thinned out, inputting the number of lines (planes) to be thinned out, or picking a deletion line. In the calculation of the coarse coordinate grid, lines (planes) to be thinned out are deleted, grid point numbers are assigned differently, and the correspondence with the dense coordinate grid is determined. When displaying a coarse coordinate grid, the display is emphasized using color, thick outline, blinking, highlighting, alternating display, parallel display, etc. When displaying correspondence, coarse and dense coordinate grids are displayed overlapping each other (Figure 5).
.

When an analysis worker creates a dense coordinate grid in multiple stages by subdividing the coarsest coordinate grid based on experience, the current coordinate grid is the closest-packed coordinate grid at that point (basic coordinate grid). For example, if we consider 14 in Fig. 6 to be the current coordinate grid, the procedure for creating a dense coordinate grid is to specify one side 17 of the current coordinate grid 14, as shown in Fig. 6, and then Specify the range (partial area) from 18 to 19 with , and specify the number of subdivisions within the range. Next, a dense coordinate grid is calculated based on this information (for example,
No. 466, Japanese Patent Application Publication No. 1983-263564, Patent Application No. 1987-
No. 281773.

Using the technology disclosed in Japanese Patent Application Laid-Open No. 63-136258 and Japanese Patent Application No. 1-27462), as shown in FIG. The obtained dense coordinate grid 20 is displayed on the display devices 2A and 2B. Next, by specifying another range on the same side 17 or another coordinate line that intersects with this side and the range above it, subdivided coordinate grids are similarly calculated, and the display device 2A, 2B
Display above. This operation is repeated until a satisfactory coordinate grid is obtained based on the experience of the analysis operator, and the satisfied dense grid is used as the closest coordinate grid, and the correspondence of the coarse coordinate grid to the dense coordinate grid and the fine coordinate grid are stored in an external memory. 5a. Further, the coordinate data of the coarse coordinate grid may be deleted.

Next, by considering the closest coordinate grid as the current coordinate grid and repeating the above operations, multiple grids can be created.

When inputting the number of subdivisions, enter the number of insertion divisions within the unit cell, or input the total number of insertion divisions within the specified range. When displaying a coarse coordinate grid, one color, thick outline, blinking, highlight,
The display devices 2A, 2 are highlighted by alternate display, parallel display, etc.
Display on B. In the correspondence display, coarse and dense grids are overlapped on the display device 2A.

2B (Fig. 6).

If the residual distribution for the coordinate grid is obtained from the physical quantity calculation using the current coordinate grid (step 9B), and a coarse coordinate grid is created based on this residual distribution (step 8B), the coarse coordinate grid is The residual distribution in the control volume for one unit cell is coarsened using the method described below so that it takes on a positive definite value (residuals take a positive value) or a negative definite value (residuals take a negative value) as much as possible. Set up the coordinate grid.

For example, in the current coordinate grid, the number of vertical and horizontal unit grids included in the positive residual region (region where the residual takes a positive value) 21 and the negative region (region where the residual takes a negative value) 22 must be equal. For example (FIG. 7A), the unit grids are grouped together for each adjacent positive region 21 and negative region 22 to create a rough coordinate grid as shown in FIG. 7B. In addition, if the number of unit cells included in the positive region 21 and negative region 22 of the residual in the current coordinate grid is large in one direction (Fig. 7C), the unit grid of the current stage coordinate grid can be changed for each region. A rough coordinate grid as shown in Fig. 7D is created by combining the above. That is, one unit cell in the coarse coordinate grid is created by grouping the unit cells of the current coordinate grid for each positive region (or negative region) of the residual. It is preferable that the coordinate lines of the coarse coordinate grid are located as close to the boundary between the positive region 21 and the negative region 22 as possible.

Instead of dividing the magnitude of the residual into two regions, a positive region and a negative region, it is also possible to divide the magnitude of the residual into three or more regions for each predetermined level range. In this case, unit cells of the current coordinate grid are grouped into the same level range to create a coarse coordinate grid in the same manner as described above.

The procedure for creating a multiple grid considering the residual distribution is as follows. An example will be explained in which the residual distribution is divided into a positive region 21 and a negative region 22. First, set the initial value of the physical quantity on the current coordinate grid, iteratively calculate the physical quantity, and repeat it an appropriate number of times. Canceled.

Calculate the residual (step 9B). The obtained residual distribution (
The positive area 21 and the negative area 22) are displayed on the display devices 2A and 2B together with the current coordinate grid used in the calculation as shown in FIG. 8 (step 9C).

In step 8B, first, a search is performed to determine whether a coarse coordinate grid corresponding to the displayed residual distribution is stored in the external memory 5a. If the corresponding coarse coordinate grid is stored, this coarse coordinate grid is called up and highlighted and displayed on top of the residual distribution and the current coordinate grid used for calculation. Furthermore, if the corresponding coarse coordinate grid is not stored, this fact is displayed on the display device 2A in step 8B. The analysis operator looks at the residual distribution and the display data of the coordinate grid at the current stage and specifies the coordinate line to be deleted in the same manner as described above. Upon receiving this specified information, step 8B:
A coarse coordinate grid is created from the dense coordinate grid using the thinning procedure described above, and the coarse coordinate grid is highlighted in the same way (Figure 9).
. In FIG. 9, the coordinate line 14B indicated by a dashed line is thinned out. In this way, in this embodiment, the size of the residual error can be reflected in the creation of the coarse coordinate grid, so as shown in FIG.
can be easily created. Since the residual distribution is displayed, the analysis operator can easily know which coordinate lines should be thinned out when creating such a multiple grid. Next, the analysis operator determines whether the highlighted coarse coordinate grid is appropriate based on experience (FIG. 9 shows the case where it is appropriate). If the obtained rough coordinate grid is not valid, the analysis operator again specifies the coordinate lines to be thinned out from the current coordinate grid, and performs step 8B.
Create a coarse coordinate grid using the thinning procedure. Alternatively, operations such as deletion and addition of coordinate lines on the obtained rough coordinate grid (for example, Japanese Patent Application No. 61-281773, Japanese Patent Application No. 1-2
7462). Next, the created rough coordinate grid is superimposed on the residual distribution and the current coordinate grid and highlighted on the display devices 2A and 2B (FIG. 9). When the analysis operator inputs a signal indicating that the coarse coordinate grid is valid, step 8B stores the correspondence of the coarse coordinate grid to the fine coordinate grid and the coarse coordinate grid in the external memory 5a. FIG. 10 shows the residual distribution when Poisson's equation is solved on 132 current coordinate grids 14C and the current coordinate grid 14C used for calculation. The residual distribution is contour line 30
It is shown in Further, FIG. 11 shows the rough coordinate grid (solid line) 16A created by the above method together with the residual distribution of FIG. 10 and the coordinate grid at the current stage (-dotted chain line) 14c.

When the creation of the multiple grid in consideration of the residual distribution is completed in this manner and the determination in step 7C becomes YES, the process in step 10 is executed using the obtained multiple grid. In the display example of FIG. 9, the coarse coordinate grid 16 with different density has 11 unit grids with different sizes depending on the residual distribution. A simultaneous-order equation A'x = b' is created for each unit cell, and this equation is repeatedly solved until the residual becomes less than a predetermined value. As the initial value of the equation, a value obtained by solving a predetermined simultaneous-order equation Ax=b for the unit cell of the coordinate grid at the current stage, which is the basis of the multi-grid, is used. In this embodiment, since it is possible to create coarse coordinate grids with different unit grid sizes depending on the distribution of residuals, the simultaneous-order equation A'x = b
', the influence of residuals can be eliminated in a short time. Therefore, the convergence time is short.

The obtained physical quantities will also be highly accurate.

Overlapping display and parallel display can be adopted for displaying the residual distribution, calculation grid, and coarse grid.

In addition, when creating a denser coordinate grid based on the calculated physical quantities, the physical quantities and the current coordinate grid used in the calculation are displayed superimposed, and the subdivision procedure in step 8D in Figure 1 is used for the calculation. Create a denser coordinate grid from the current coordinate grid. At this time, if the coordinate line positions of the coordinate grid used in the calculation do not match the coordinate line positions of the obtained dense coordinate grid, and the analysis operator specifies that the dense coordinate grid be adopted, the dense coordinate grid As the current coordinate grid, all the coordinate grids used for calculation and coarser coordinate grids are re-created (processing of step 8D).

By combining the multiple grid creation procedures using the residual distribution and physical quantities described above by an analysis worker, it becomes possible not only to create multiple grids based on the analysis worker's experience, but also to create the following multiple grids.

(1) Creating and changing the coarse grid based on the displayed residual (2)
Creating a multi-stage Suisho grid from the coarsest coordinate grid based on the displayed physical quantities and creating and changing the coarse grid based on the residuals (3) Changing the close-packed grid based on the physical quantities and creating a coarse grid based on the residuals An example of a multi-coordinate grid created by the method and apparatus for creating a multi-coordinate grid according to the above embodiment will be shown below.

FIG. 12 is an example of a triple coordinate grid within a circle. FIG. 13 is an example of a dual coordinate grid created for analysis of fluid flowing between a group of circular tubes.

Generally, when a support device such as the present invention is not used,
As the target object becomes more complex, it becomes more difficult to set appropriate dense coordinate grids and coarse coordinate grids, and human errors increase. This is because it is difficult to create a coarse coordinate grid that fully reflects experience unless you visually check the coarse coordinate grid in correspondence with the current coordinate grid.
This is because it is difficult to create a sufficiently appropriate coarse grid unless correspondence is taken with the residual distribution to create the coarse coordinate grid. When such problems are handled interactively as in the present invention, mistakes are eliminated and work efficiency is dramatically increased.

Note that the correspondence relationship has been described as the superposition of the first lattice coordinates and the second lattice coordinates, but in addition to this, the correspondence relationship may be replaced by a parameter and displayed. Parameters include factors indicating density, size of partial regions, and the like. Furthermore, instead of displaying them in one layer, they may be displayed separately.

[Effects of the Invention] As described above, by interactively creating a multi-coordinate grid in an analysis area using the coordinate grid creation support method and device of the present invention, analysis can be performed while incorporating the analysis operator's way of thinking. Coordinate grids that are useful for reducing calculation time can be efficiently generated depending on the target and analysis purpose. As a result,
It is possible to eliminate human error when creating a coordinate grid, reduce the work time required to create the grid, and also reduce the time required to calculate physical quantities.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the flow of the coordinate grid creation support method of the present invention, FIG. 2 is a diagram showing the configuration of the coordinate grid creation support device of the present invention, FIG. 3 is a display screen of the created coordinate grid, and FIG. The figure shows the coarse grid creation screen that displays only the initially created grid, Figure 5 shows the coarse grid creation screen that displays the operation target grid and the newly created grid superimposed, and Figure 6 shows the operation target grid and the newly created grid overlapped. The dense grid creation screen shown in Figure 7 shows various display examples,
Figure 7A is an example of residual display on the operation target grid, Figure 7B is a side view of coarse grid creation, Figure 7C is an example of residual display on the operation target grid, and Figure 7D is the side of coarse grid creation. Figure 8 is a residual display screen diagram, Figure 9 is a coarse grid creation screen diagram based on residuals,
Figure 10 is the side view of the residual distribution in the actual analysis, Figure 11 is an example of coarse grid creation based on the residual distribution in the actual analysis, and Figure 12 is the side view of the multiple grid in the circular area analysis. FIG. 13 shows the side ports of the multiple grid in the fluid analysis of the tube group region. 1... Coordinate grid creation support device, 2A, 2B... Display device, 3A, 3B... Image display control device, 3C, 3
D... Image data storage unit, 4... Arithmetic processing unit, 4
a...Arithmetic section, 4b.Processing procedure storage section, 4C..Intermediate data storage section, 4d..Input section, 4e..Coordinate data input/output section. 4f, 4g...Image data input/output unit, 4h...Shape data input/output unit, 5...External storage device, 6...Operation panel, 7...Initial coordinate grid creation step, 8... Coarse/dense multiple grid setting step, 9... Guidance display step, 10... Display screen, 14... Coordinate grid at current stage (current level, operation target), 15... Parameter input table, 16... Rough coordinates Grid, 17...Specified edge, 18...Section start point marker, 19...Cursor,
20...Dense coordinate grid, 21...Residual positive region, 22...
・Residual negative area. Agent Patent Attorney Tadashi Akimoto Figure 1, Figure 2] Figure] n Figure 7] 0 Figure, Figure, Figure, Figure, Figure, Figure

Claims (1)

[Claims] 1. In a coordinate grid creation support method used when creating a coordinate grid in a target analysis area in order to analyze a physical phenomenon using a computer, the analysis area is divided and the first step is performed. a first step of creating a coordinate grid; a second step of storing the first coordinate grid;
a third step of displaying the first coordinate grid on a display screen; a fourth step of creating a sub-region of the first coordinate grid; and a second coordinate grid having a different interval from the first coordinate grid in the sub-region. a fifth step of creating a grid; a sixth step of storing the correspondence between the first coordinate grid and the second coordinate grid; and a display screen for storing the correspondence between the first coordinate grid and the second coordinate grid. A coordinate grid creation support method comprising: a seventh step of displaying a coordinate grid; 2. First, perform the processing from the first step to the seventh step, then repeat the processing from the fourth step to the seventh step, and repeat the processing from the fourth step to the seventh step.
In the case of repeating the process up to the step, the first coordinate grid in the fourth step is defined as the previous latest second coordinate grid redefined as the first coordinate grid. Coordinate grid creation support method. 3. The coordinate grid creation support method according to claim 1 or 2, wherein the second coordinate grid having a different interval from the first coordinate grid in the fifth step is a dense coordinate grid obtained by subdividing the first coordinate grid. . 4. The coordinate grid creation support method according to claim 1 or 2, wherein the second coordinate grid having a different interval from the coordinate grid in the fifth step is a coarse coordinate grid obtained by thinning out the first coordinate grid. 5. When obtaining the physical quantity distribution in the analysis area through numerical simulation by repeated calculation using the first and second coordinate grid groups of different density finally obtained by the coordinate grid creation support method of claim 2, the display The distribution of residuals obtained from the calculation results during the iterative calculation process is displayed on each sparse and dense coordinate grid displayed on the screen, and the fourth to seventh steps are processed based on this displayed information. A coordinate grid creation support method that involves re-creating each coarse/fine coordinate grid. 6. The coordinate grid creation support method according to claim 5, wherein the coarse grid is created so that the residual value has a constant sign within the control volume of the coarse grid. 7. The coordinate grid creation support method according to claim 1, wherein the correspondence in the seventh step is a relation in which the first coordinate grid and the second coordinate grid are displayed overlappingly on the same display screen. 8. In a coordinate grid creation support device that is used to create a coordinate grid in the target analysis area in order to analyze physical phenomena using a computer, there is a first step that divides the analysis area and creates the first coordinate grid. a second means for storing the first coordinate grid; a third means for displaying the first coordinate grid on a display screen; and a fourth means for creating a partial area of the first coordinate grid. means, fifth means for creating a second coordinate grid having a different interval from the first coordinate grid in the partial area, and sixth means for storing the correspondence between the first coordinate grid and the second coordinate grid. and seventh means for displaying the correspondence between the first coordinate grid and the second coordinate grid on a display screen. 9. The coordinate grid creation support device according to claim 8, wherein the second coordinate grid having a different interval from the first coordinate grid in the fifth means is a dense coordinate grid obtained by subdividing the first coordinate grid. 10. The coordinate grid creation support device according to claim 8, wherein the second coordinate grid having a different interval from the first coordinate grid in the fifth means is a coarse coordinate grid obtained by thinning out the first coordinate grid.