JP2000172883A - Fluid simulator - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve a fluid behavior analyzing efficiency of a fluid simulator which simulates a fluid flowing around an object by shortening the grid generating time of the simulator by improving the functions of the simulator for recognizing shape of object and for discriminating fluid and nonfluid area in a calculated area. SOLUTION: An analytical data generating section 3 is provided with a function which projects a hierarchical grid generated by adjusting roughness and fineness in a three-dimensional space upon a plane in the direction of each coordinate axis and defines the intersections of shape surface elements and a Descartes grid by sweeping grid points on the projected plane in the direction of a coordinate axis perpendicular to the projected plane, and based on the generated grid points, scans the intersection of the object surface most adjacent to the grid point in the directions of all coordinate axes and discriminates a fluid area when the inner product values of the normal vectors of the surface elements to which each intersection belongs and the distance vectors toward the grid points from the intersections are not negative and a nonfluid area in the other case. Thus, the grid forming time can be shortened and the fluid behavior analyzing efficiency of the simulator can be improved.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fluid simulation apparatus for numerically analyzing the behavior of a fluid flowing around an object using a three-dimensional grid.

[0002]

2. Description of the Related Art Conventionally, in the field of fluid dynamics in which the flow state of a fluid is handled by this kind of fluid simulation apparatus, research on numerical simulation by CAE has been dramatically developed with the rapid progress of computers, and the amount of calculation has been increased. It is possible to analyze even a three-dimensional large-scale problem that would be practically impossible to analyze because of its enormous volume. As this kind of fluid simulation, Japanese Patent Application No. 9-230 is disclosed.
No. 901 is disclosed.

Hereinafter, a conventional numerical simulation apparatus and processing means will be described with reference to the drawings.

Here, a two-dimensional mesh such as a triangle created on the surface of the object to represent the shape of the object is referred to as a “plane element”, and a three-dimensional mesh such as a hexahedron created for actual use in calculation is referred to as a “grid”. To distinguish them.

As shown in FIG. 8, object surface data created in 3D-CAD1 and output in DXF format is converted to UCD data.
I / F unit 2 for converting to a format (corresponding to data conversion means)
Using the UCD format object surface data output by the I / F unit 2, a Cartesian grid is created in a space within a predetermined analysis target range, and other analysis data (various analysis conditions, etc.) is created. Using the analysis data creation unit 3 (corresponding to a grid creation unit) and the analysis data created by the analysis data creation unit 3, numerical calculation is performed by the Cartesian grid local collocation method (described later), and the result is expressed in UCD format. Analysis unit 4 (corresponding to analysis means and analysis result output means), and display unit 5 (corresponding to display means) for visualizing analysis results using the analysis result data in UCD format output from the analysis unit 4. It is composed of

[0006] As the 3D-CAD1, a general commercially available three-dimensional CAD system can be used, and the shape of an object within the analysis target range is created by a three-dimensional solid model or the like. Then, a triangular surface element is generated so as to cover the surface of the created solid model or the like. This is because 3D-CAD1 and the fluid simulation device A1 are represented by unifying the surface shape of the object into a triangle which is a simple minute plane.
This is for ensuring the reliability of the data between them and simplifying the grid creation processing in the analysis data creation unit 3. The free-form surface data or solid data constituting the three-dimensional solid model may be directly applied to the fluid simulation device A1, but the reliability of the data is different due to the difference in interpolation accuracy between the 3D-CAD1 and the fluid simulation device A1. May be missing, and the grid creation processing becomes very complicated. In this triangular plane element automatic generation function, the approximate size of the generated triangle can be set, so that an appropriate size can be set. However, the generated plane element is represented as an aggregate of triangles (planes). Because
If the size of the generated surface element is too large, the shape of the object cannot be accurately represented, affecting analysis accuracy.On the other hand, if the size of the surface element is too small, not only does it take much time to generate the surface element, Unnecessary time is required during the grid generation processing. Therefore, it is necessary to set a large surface element size so as not to affect the analysis accuracy. Also,
There are many CAD systems that can select a quadrangle other than a triangle as a surface element automatically generated on the surface of a solid model,
In the case of a quadrangle, the normal direction is not determined if the four points are in a twisted state that is not on the same plane, and it is necessary to divide it into two triangles at the time of creating a grid that requires the subsequent normal direction, and eventually process it. Here, a triangular surface element is used.

The generated shape data of the triangular surface elements (the coordinate values of the constituent points and the connection data of each triangle) are output to a DXF format file. Here, DXF is one of the standard data formats in the CAD field.

Further, a general three-dimensional CAD system is often provided with a function of grouping each shape at the time of shape creation, and the generated triangular surface element is subjected to, for example, a difference in boundary conditions and other conditions. Grouped by
Some have a function of outputting the group information (corresponding to the classification information) together with the shape data to the DXF format file. If such a function is provided, the work can be performed easily and in a short time by utilizing the group information when the analysis data such as the boundary condition is created by the analysis data creation unit 3 thereafter.

Subsequently, the I / F unit 2 converts the triangular plane element data (including group information) in DXF format into UCD
Convert to format. Here, UCD (Unstructured)
The red cell data (AV) format is an application program for visualizing scientific calculation results and the like.
S and VIZ (both are developed by Advanced Visual
1 Systems, Inc.), which can be expressed in a form that includes all data of shapes, grids, and calculation results. In the fluid simulation device A1, all data are unified into this UCD format. Of course, the data format used here is U
The data format is not limited to the CD format, and other similar data formats can be used.

In the analysis data creation section 3, the I / F section 2
A lattice of hexahedral cells is created in a space (fluid area) within the analysis target range using the triangular surface element data in the UCD format output from. This grid creation procedure will be described.

In the case of the conventional method using an unstructured grid, a space (fluid region) is usually used by directly using a triangular (or square) plane element generated on the object surface as shown in FIG. 9), or a grid in which the grid width gradually decreases as approaching the object from the space (fluid region) toward the object as shown in FIG. 9B. Specifically, FIG. 10 (for simplicity, 2
(Represented in two dimensions), the entire fluid region is first divided by an orthogonal lattice (cubic cell) having a size of c1. Subsequently, among the cells of the c1 size, the cells within a range a1 of a predetermined distance from the object in the normal direction thereof are re-divided into 1/2 c2 size. Further, cells in the range a2 are reduced to 1/2 c.
Subdivide into three sizes. By repeating these processes a predetermined number of times, the intersection between the smallest cell and the object, that is, the surface formed by the triangular surface elements is obtained.

The subdivision ranges a1, a2,... Are set by distances in the normal direction from each triangle constituting the object surface, as shown in FIG. For example, FIG.
As shown in (a), the unit vector n in the normal direction of each triangle is defined, and the inner product distance between the vector R and the vector R from the vector n definition point (open circle in FIG. 11A) to each cell center point. D (= R · n) is the distance Dnes specified in advance
Select cells that are less than or equal to t. That is, FIG.
Broken line in (a) and (b) (distance Dnes from object)
t) corresponds to the subdivision ranges a1, a2,.

Further, there is a method of setting the distance from each point instead of the normal direction distance from each triangle constituting the object surface as shown in FIG. 12 (a). If the distance between two points is partially large as shown in parentheses), that is, if the triangle is large, it may occur that cells near the object cannot be partially re-divided. The reason is,
This is because the surface element size is often not constant in CAD automatic surface element division. In such a case, 3D again
It is necessary to return to -CAD1 and change the size parameter at the time of automatic generation of the triangular surface element to a smaller value, and to perform the processing again, so that the loss in the processing operation is large. On the other hand, the method of setting the distance in the normal direction does not require reprocessing.

The position of the subdivision ranges a1, a2,... (Distance Dnest from the object) is set by first changing the cell width in the normal direction from the surface of the object, as shown in FIG. Assuming a cubic cell, half the distance between the minimum cell cs and the cell c1 having a cell width twice that of the minimum cell cs is a distance D.
nest. Of course, the cell width of the assumed cubic cell is not limited to a geometric change,
The increase rate of the cell width is smaller near the object surface, and the increase rate of the cell width can be increased as the distance from the object surface increases.Therefore, the grid near the object surface is made denser while suppressing the increase in the number of grids. Processing such as roughening of the grid can be easily performed, and the degree of freedom of grid generation can be increased.

When the subdivision is performed several times by the above-described procedure until the set minimum cell is divided, each side of the cell along the object and the triangular surface element formed on the surface of the object are compared with each other. The cell is fitted to the surface of the object by finding the intersection and leaving only the cell outside the object. That is, processing for cutting cells along the object at the object surface position is performed. Therefore, in most cases, cells along the object do not become cubic cells, but in the numerical calculation method used in the conventional configuration, the cells near the object need not be cubic cells, and may have any shape. Analyze with high accuracy.

When the grid is created, the analysis data creating unit 3
Creates analysis data such as boundary conditions necessary for analysis. This work is performed interactively using a mouse or the like on the grid displayed three-dimensionally on the screen of the display unit 5. In this case, if the group information created by the 3D-CAD1 is used, conditions can be easily set for each group, and the state of a triangular surface element generated on the surface of an object, the state of a grid being created, the setting state of analysis data, and the like. Can be performed while confirming on the display of the display unit 5.

When the creation of the grid and other analysis data is completed, the analysis data creation unit 3 outputs all the analysis data including the grid data to the analysis unit 4 in UCD format. The analysis unit 4 performs a numerical calculation (actual calculation) by the Cartesian grid local collocation method using the analysis data.

Here, the Cartesian lattice local collocation method is a thermofluid simulation method newly developed by the present inventors, and has already been presented at an academic meeting (Transactions of the Japan Society of Mechanical Engineers, Vol. 61, No. 592 (1995-)). 12) etc.). The present method uses a grid similar to the conventional Cartesian grid method, overcomes the problems of the calculation time and accuracy in the Cartesian grid method, and makes it applicable to the viscosity problem.

When the numerical calculation is completed in the analysis unit 4, the UC including the analysis result in addition to the grid data used in the analysis is obtained.
The file in the D format is output, and the analysis result is visualized and displayed on the display unit 5 as a color contour diagram or a vector line diagram as shown in FIG.

[0020]

However, in the above-described conventional configuration, a method for detecting the intersection of the shape surface element and the Cartesian grid is not defined in the grid creation procedure.

In addition, it is not determined whether the grid point created by the grid creation means belongs to a fluid region or a non-fluid region related to the grid creation time.

The present invention is to solve such a conventional problem, and provides a fluid simulation apparatus with high analysis efficiency by defining a method of detecting an intersection of a shape surface element and a Cartesian grid in a grid creation procedure. The purpose is to:

[0023]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, an object of the present invention is to analyze the behavior of a fluid flowing around an object numerically using a three-dimensional grid. It is provided with data conversion means for converting into external form data, grid creation means for creating a grid, and analysis means for analyzing the behavior of a fluid, and a hierarchical grid created by adjusting the three-dimensional spatial density. It is provided with a function of defining the intersection of the shape surface element and the Cartesian grid by projecting on the plane of the direction and sweeping the grid points on the projection plane in the coordinate axis direction perpendicular to the projection plane. By providing the function of defining the intersection of the shape surface element and the Cartesian grid in this manner, it is possible to shorten the conventional grid creation time and improve the analysis efficiency.

Further, the intersection point between the lattice point and the nearest object plane in all coordinate axis directions is scanned, and the inner product value of the normal vector of the surface element to which each scanned intersection point belongs and the distance vector from the intersection point to the lattice point is obtained. If all are not negative values, it has a function of judging that it is a fluid region, and in other cases, it judges that it is a non-fluid region. By not dividing or deleting the grid in the non-fluid region by such a process, the calculation area can be set more efficiently and the grid creation time can be reduced.

[0025]

DESCRIPTION OF THE PREFERRED EMBODIMENTS In order to solve the above-mentioned problems, the present invention provides a fluid simulation apparatus which numerically analyzes the behavior of a fluid flowing around an object using a three-dimensional grid, using a three-dimensional CAD. Data conversion means for converting the first object outer shape data representing the outer shape of the object created as described above into second object outer shape data in a predetermined data format, and the second object obtained by the data conversion means A grid creating means for creating a grid such that the grid width is gradually reduced at arbitrary intervals as approaching the object in a space within a predetermined analysis target range using the external shape data; Analysis means for analyzing the behavior of the fluid by using the local selection method at grid points within a predetermined range near the object using the grid and using the finite difference method at other grid points. Then, the hierarchical grid created by adjusting the three-dimensional spatial density is projected onto a plane in each coordinate direction, and the grid points on the projection plane are swept in the coordinate axis direction perpendicular to the projection plane to form the shape plane. It has the function of defining the intersection of the element and the Cartesian grid. Therefore, the grid creation time can be reduced, and efficient analysis can be performed.

The grid point created by the grid creating means scans an intersection point between the grid point and the nearest object plane in all coordinate axis directions, and intersects the normal vector of the surface element to which each scanned intersection point belongs with the intersection point. If the inner product of the distance vector from the to the grid point is not a negative value, it is determined to be a fluid region, otherwise it is determined to be a non-fluid region, and the grid of the non-fluid region is not divided, or It has a function to delete. Therefore, it is possible to improve the efficiency of setting the calculation area, shorten the grid creation time, and improve the analysis accuracy.

[0027]

Embodiments of the present invention will be described below with reference to the drawings. The same components as those in the related art are denoted by the same reference numerals, and detailed description thereof is omitted.

(Embodiment 1) As shown in FIG. 1, a fluid simulation apparatus A2 is constructed around a 3D-CAD1 and outputs object surface data output in DXF format to a UC.
An analysis for creating a Cartesian grid in a space within a predetermined analysis target area based on the I / F unit 2 for converting into the D format and the object surface data output from the I / F unit 2 and other analysis data Numerical calculations are calculated by the data creation unit 3 and the analysis data by the Cartesian grid local collocation method,
An analysis unit 4 that outputs the result and a display unit 5 that visualizes and displays the analysis result data are configured. As for a specific analysis procedure, as shown in FIG. 2, in the analysis data creation unit 3, first, in step 101, the object outline data created by the 3D-CAD 1 was converted into UCD format data by the I / F unit 2. Read the shape data file.

In step 102, a basic grid is created. In step 103, a range for dividing the cell is determined.

In step 104, the cell is divided and its boundary surface is defined. In step 105, an intersection point between the grid point whose object area is re-divided and the object plane is determined.

In step 106, the continuation of the cell division is determined. That is, if the cell division is sufficient, the process proceeds to step 107, and the determined grid is output to the next analysis unit. If not, go to step 108,
The process returns to the step 103 for determining the cell division range. The emphasis of this embodiment lies in the method of detecting the intersection between the mirror interface and the grid line in steps 103 to 106, which will be described with reference to FIG. FIG. 3 is shown in two dimensions for simplicity.

First, the object shape is projected on each reference plane.
In this case, since the image is a two-dimensional display, the image is projected on the axis. That is, each reference coordinate plane is swept in the direction of the object, and a point at which the lattice point on the plane intersects with the object shape becomes a boundary point for recognizing the shape of the object. This is performed in each reference plane direction. Therefore, in the case of three-dimensional display, there are three directions. As described above, according to this method, the object shape can be finely recognized in a short time, and there is an effect of shortening the grid formation time and accurately recognizing the object shape.

(Embodiment 2) As shown in FIG. 4, in the analysis data creating section 3, first, in step 201, the 3D-CA
The I / F unit 2 converts the object shape data created in D1 into UCD
Read the shape data file converted to the format data.

In step 202, a basic grid is created. In step 203, a range for dividing the cell is determined.

In step 204, a cell boundary surface is defined.
In step 205, an intersection point between the grid point re-divided in the division range and the object plane is obtained.

With respect to the grid points formed in step 206, it is determined whether each grid point exists in a fluid area or a non-fluid area.

At step 207, the continuation of the cell division is determined. That is, if the cell division is sufficient, the process proceeds to step 208, and the determined grid is output to the next analysis unit. If not, proceed to step 209 and step 2
Returning to the procedure from 03 to 207, the range of cell division is determined again. The emphasis of this embodiment is a method of judging in step 206 whether each grid point exists, that is, whether it exists in a fluid area or a non-fluid area, which will be described with reference to FIG.

In FIG. 5, a point A is an arbitrary lattice point. The intersection from point A to the element on the nearest object plane in the reference coordinate axis direction is scanned. The surface elements belonging to the scanned intersection point B have normal vector data in the conventional configuration. Therefore, when the inner product of the normal vector and the distance vector from intersection B to grid point A is a positive value in all reference axis directions as shown in FIG. 6, the grid point exists in the fluid region. Then, if there is at least one negative value as shown in FIG. Thus, by not dividing or deleting the grid in the non-fluid area, the calculation area can be set more efficiently and the grid creation time can be reduced.

[0039]

As described above, according to the present invention, a fluid simulation apparatus for numerically analyzing the behavior of a fluid flowing around an object using a three-dimensional grid is created using three-dimensional CAD. Data conversion means for converting first object outline data representing the outline of the object into second object outline data in a predetermined data format, and using the second object outline data obtained by the data conversion means hand,
In a space within a predetermined analysis target range, a grid creating means for creating a grid in which the grid width gradually decreases at arbitrary intervals as approaching the object, and an object using the grid obtained by this means. By using the local collocation method at grid points within a predetermined range in the vicinity and using the finite difference method at other grid points, analysis means for analyzing the behavior of the fluid is provided, and the intersection of the shape surface element and the Cartesian grid is provided. Is defined, the hierarchical grid created by adjusting the three-dimensional spatial density is projected onto a plane in each coordinate direction, and the grid points on the projection plane are swept in the direction of the coordinate axes perpendicular to the projection plane. Is defined. Therefore, it is possible to shorten the time required for forming a conventional grid and to perform highly accurate analysis.

When it is determined whether the grid point created by the grid creating means belongs to a fluid region or a non-fluid region, the intersection of the lattice point with the nearest object plane in all coordinate axis directions is scanned. If all the inner product values of the normal vector of the surface element to which each scanned intersection belongs and the distance vector from the intersection to the lattice point are all positive values, it is determined that they belong to the fluid region,
If there is at least one negative value, the grid in the non-fluid area is not divided or deleted, so that the calculation area can be set more efficiently and the grid formation time can be reduced.

[Brief description of the drawings]

FIG. 1 is a schematic block diagram of a fluid simulation apparatus according to a first embodiment of the present invention.

FIG. 2 is a float chart of grid creation in the analysis data creation unit.

FIG. 3 is an explanatory view of an intersection detection method for a boundary surface.

FIG. 4 is a flow chart of grid creation in an analysis data creation unit of a fluid simulation apparatus according to a second embodiment of the present invention.

FIG. 5 is an explanatory diagram of a method of determining the existence area (fluid area or non-fluid area) of a grid point created by the grid creating means.

FIG. 6 is a judgment diagram when the lattice point exists in the fluid region.

FIG. 7 is a determination diagram when a grid point exists in a non-fluid region.

FIG. 8 is a block diagram showing a schematic configuration of a conventional fluid simulation apparatus.

FIG. 9 (a) A conventional grid creation diagram using an unstructured grid (b) A grid creation diagram in a conventional fluid simulation device

FIG. 10 is a schematic diagram of a grid creation procedure in a conventional fluid simulation apparatus.

FIG. 11A is an analysis diagram for defining a unit vector in the normal direction from a triangular surface element constituting the outer surface of an object and setting a re-division range of a grid.

FIG. 12 (a) is an analysis diagram in which a grid subdivision range is set based on a distance between definition points constituting a triangular surface element on the outer surface of an object; Analysis diagram

FIG. 13 is an analysis diagram for setting a subdivision range of a grid by changing a grid width at an equal ratio;

FIG. 14 is a diagram showing visualization of analysis results.

[Explanation of symbols]

 1 3D-CAD 2 I / F section 3 Analysis data creation section 4 Analysis section 5 Display section

Continuation of front page (72) Inventor Nobuo Shimomura 4-2-5 Takaida Hondori, Higashiosaka-shi, Osaka Matsushita Refrigerator Co., Ltd. F-term (reference) 2G023 AA10 AB30 AC03 5B046 JA09 5B049 AA00 AA04 EE01 EE41 GG07 GG09

Claims (2)

[Claims] 1. The method according to claim 1, wherein the behavior of the fluid flowing around the object is converted from first object outline data, which is created by using three-dimensional CAD, representing the outline of the object, to second object outline data in a predetermined data format. Data conversion means, and using the second object outline data obtained by the data conversion means, in a space within a predetermined analysis target range, the grid width gradually decreases at arbitrary intervals as approaching the object. Grid creation means for creating such a grid, and using the grid obtained by the grid creation means, a local collocation method for grid points within a predetermined range near the object, and a finite difference method for other grid points. And a projection means for projecting a hierarchical grid created by adjusting the intersection of the shape surface element and the Cartesian grid spatially and densely on a plane in each coordinate direction. A fluid simulation device capable of numerical analysis using a three-dimensional grid, wherein the upper grid point is defined by sweeping in a coordinate axis direction perpendicular to the projection plane. 2. When determining whether a grid point belongs to a fluid region or a non-fluid region, the grid point is scanned from the intersection with the nearest object plane in all coordinate axis directions, and each scanned intersection belongs to If the inner product of the normal vector of the surface element and the distance vector from the intersection to the grid point is not a negative value, it is determined that the area is a fluid area, otherwise it is determined to be a non-fluid area. Item 7. The fluid simulation device according to Item 1.