CN108153706A - A kind of ultra-large calculating grid reconstruction method - Google Patents
A kind of ultra-large calculating grid reconstruction method Download PDFInfo
- Publication number
- CN108153706A CN108153706A CN201711444253.8A CN201711444253A CN108153706A CN 108153706 A CN108153706 A CN 108153706A CN 201711444253 A CN201711444253 A CN 201711444253A CN 108153706 A CN108153706 A CN 108153706A
- Authority
- CN
- China
- Prior art keywords
- grid
- block
- ultra
- point
- calculating
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F15/00—Digital computers in general; Data processing equipment in general
- G06F15/76—Architectures of general purpose stored program computers
- G06F15/78—Architectures of general purpose stored program computers comprising a single central processing unit
- G06F15/7867—Architectures of general purpose stored program computers comprising a single central processing unit with reconfigurable architecture
- G06F15/7871—Reconfiguration support, e.g. configuration loading, configuration switching, or hardware OS
Landscapes
- Engineering & Computer Science (AREA)
- Computer Hardware Design (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of ultra-large calculating grid reconstruction methods, and the index point of each piece of grid, the key message of completely true reduction original mesh are obtained in a manner that coupled room grid and boundary condition carry out whole audience joint parsing；And each piece of independent reconstruct is carried out using dynamic memory distribution technology on this basis, avoid excessive this business software common fault that memory is caused to overflow of grid so that obtained using PC ultra-large（10000000000 grades or more）Grid is calculated to become possible to；By the related information for being managed collectively whole audience grid, mainframe computer cluster just obtainable ultra-large calculating grid can be realized on PC, this technology inventor obtains the high quality calculating grid of hundred billion grades of complex appearances on common computer at present, and business software can only obtain 24 hundred million grades of calculating grids on same computer.
Description
Technical field
The present invention relates to calculating field more particularly to a kind of ultra-large calculating grid reconstruction methods.
Background technology
For current computation fluid dynamics (Computational Fluid Dynamics), although various non-
Structure and Mixed grid method development are swift and violent, but traditional structured grid method is still in dominant position.However Structure Network
The generation that grid is calculated in lattice method is an extremely important and relatively time consuming job, from the point of view of actual use, calculates net
The result of the density convection current Numerical Simulation of grid has very important influence in the scale of lattice and near wall boundary layer, i.e.,
Just mesh topology is appropriate, if without enough scale and density, is also likely to cause result of calculation poor with actual conditions
Away from larger, for the Reynold's average NS equation being widely used at present（RANS）, need great grid scale that can just obtain on earth
The problem of widely must being discussed and studied closer to the analog result in true flow field and one, CFD fields are also special thus
There is a special verification content for being referred to as " grid convergence "：Using series of identical topological structure but different scales
Same flow field state is simulated with the calculating grid of density, and gradually grid is added very fine and closely woven, so that RANS
Governing equation can not obtain this state finer flow field structure, and last Flow Field Solution tends to restrain.But to obtain this
A convergence process is not easy to, and particularly in the case where mesh topology is constant, how not to change original mesh orthogonality
It is a urgent problem to be solved that equably grid, which is encrypted, simultaneously with slickness.
Existing calculating grid generation work is realized generally by the business Grid Generation Software with graphical interfaces
, typical business software includes pointwise, ICEM, gridpro etc..However all there are one common for such business software
The defects of be lead to not when needing to generate ultra-large calculating grid or since memory requirements is excessive generation or because
Not applied to for algorithm leads to software crash, and for business grid software, the calculating grid of several hundred million grid cell magnitudes is also exerted oneself
It can export by force, but when scale reaches 1,000,000,000,10,000,000,000 or even when hundred billion, business software is completely helpless.This is just
An apparent bottleneck has been manufactured to grid generation work, has also been brought to ultra-large CFD numerical simulations work huge
Puzzlement.
Invention content
The present invention is based on the boundary condition forms of mainstream business software（Compatible pointwise and ICEM）, to structured grid
A kind of calculating grid reconstruction method is proposed, its object is to：
First：Can carry out space and surface reconstruction to calculating grid in the case where not changing network topology, uniformly reduce or
Person increases grid scale, and the flow field calculation for CFD provides a series of densitys different calculating grids；
Second：The workload of the grid generation of grid convergence calculating is substantially reduced, as long as generating a set of benchmark grid, is recycled
This method zooms in and out, it is possible to obtain a series of calculating grid for meeting the requirement of grid convergence of different scales；
Third：Passing through the memory management of independent block so that the super large that hundred billion ranks can be also generated on personal PC calculates grid,
It solves the problems, such as the puzzlement that traditional commerce software generates super large grid, CFD calculation scales and computational accuracy is made more to step to a new level.
To achieve the above object, the present invention adopts the following technical scheme that：
A kind of ultra-large calculating grid reconstruction method, includes the following steps：
Step 1：Read original calculation mesh space point coordinates and boundary condition；
Step 2：Establish the boundary segmentation index of each grid block；
Step 3：According to the boundary condition of this block grid, the inner mark point information in each window side of this block grid is obtained；
Step 4：Each grid block crest line coordinate is filled, obtains the internal informations such as coordinate extreme value；
Step 5：It is scanned for calculating grid total space domain, obtains this block grid trappings point associated with other blocks；
Step 6：According to encryption coefficient, in the case where ensureing both ends apart from equal proportion amplification, weight is carried out to the crest line of grid block
Distribution；
Step 7：Using all index points as waypoint, according to encryption coefficient and the original line segments regularity of distribution to edge of window boundary line into
Row redistribution；
Step 8：Two dimension TFI is carried out, reconstruction window face is simultaneously projected using fuzzy normal direction projection method, is ensureing to throw
Original grid is kept to be distributed while shadow precision；
Step 9：Carry out three-dimensional standard TFI, reconstruction attractor grid block；
Step 10：Encrypted space lattice block is checked, exports relevant information.
In the above-mentioned technical solutions, the boundary point of all inner windows, all adjacent segments are bent on six faces in grid block
Rate variation is more than 30% or length ratio is more than 40% point, is denoted as inner mark point.
In the above-mentioned technical solutions, the inner mark point of the inner mark point grid block adjacent thereto in grid block is carried out
Compare, if any misaligned index point, be denoted as this block grid block and the relevant trappings point of other grid blocks.
In the above-mentioned technical solutions, in grid block information process is obtained, the ginseng for projecting can be read in if necessary
Number compound veil lattice file, and grid block and window belonging to positioning.
In the above-mentioned technical solutions, the parametrization object plane grid file is to parameterize object plane, in Grid Generation Software
In individually export a set of highdensity surface mesh.
In the above-mentioned technical solutions, the fuzzy normal direction projection method being averaged using one piece of region of view field
Projecting direction of the normal as some object plane unit, when projection object plane unit moves in both directions, view field is also therewith
Slightly translation, even if can thus so that object plane unit Curvature varying is larger, projecting direction is not in violent variation.
In conclusion by adopting the above-described technical solution, the beneficial effects of the invention are as follows：
The index point of each piece of grid is obtained in a manner that coupled room grid and boundary condition carry out whole audience joint parsing, completely
The key message of true reduction original mesh；And each piece of independent reconstruct is carried out using dynamic memory distribution technology on this basis,
Avoid excessive this business software common fault that memory is caused to overflow of grid so that obtained using PC ultra-large（Hundred
Hundred million grades or more）Grid is calculated to become possible to；By being managed collectively the related information of whole audience grid, can be realized on PC
Mainframe computer cluster just obtainable ultra-large calculating grid, at present this technology inventor obtained on common computer
The high quality of hundred billion grades of complex appearances calculates grid, and business software can only obtain hundred million grades of calculating of 2-4 on same computer
Grid；
External figurate number mould is parameterized by the way of identical with original wall surface grid distribution, and this method can not increase
Original configuration is described as accurately as possible so that finally encrypt obtained ultra-large calculating grid outline in the case of workload
With true profile close to；
Wall surface grid is projected using fuzzy normal direction projection technology, it is ensured that projecting direction is slow with wall surface molded line
Variation, had both been avoided that excessive normal direction difference caused the projection line of adjacent mesh to intersect, and grid negative twist had occurred；Monolithic can be solved again
Wall surface Curvature varying is excessive so that situation unevenly distributed occurs in fixed-direction projection pattern.This technology not only greatly improves
The adaptability of wall surface Grid Projection, and for keeping the original mesh regularity of distribution that there is obviously effect.This point pair
Also it is most important for computational fluid dynamics.
Description of the drawings
Examples of the present invention will be described by way of reference to the accompanying drawings, wherein：
Fig. 1 is fuzzy normal direction projection technology schematic diagram；
Fig. 2 is the original mesh spatial distribution before reconstructing；
Fig. 3 is the refined net spatial distribution after reconstruct；
Fig. 4 is the original mesh cross-sectional distribution before reconstructing；
Fig. 5 is the refined net cross-sectional distribution after reconstruct；
Fig. 6 is the surface mesh and partial enlarged view of certain 111,400,000,000 unit grid of cruise configuration mark mould；
Wherein：1 is fixed-direction projection, and 2 be local normal projection, and 3 be fuzz method line projection, has different colours in Fig. 1
Lines are only to play the role of difference, and color is not offered as any specific meanings in figure.
Specific embodiment
All features or disclosed all methods disclosed in this specification or in the process the step of, in addition to mutually exclusive
Feature and/or step other than, can combine in any way.
The memory consumption faced when the present invention is for business software generation super large grid is excessive and the tediously long complexity of algorithm causes
The problem of time loss is excessive using the topological structure of generated prototype structure grid, develops a set of lossless reconstruction method, leads to
The key message of parsing original mesh is crossed, including each 8 vertex positions of grid block, the shape of 12 crest lines and distribution, each window
Distribution in face etc., using spline interpolation to space curve redistribution, while in the feelings for completely retaining original mesh key message
Under condition, using transfinite interpolation method (TFI, TransFinite Interpolation) and wall surface fuzz method line projection technology pair
Original mesh is reconstructed, and finally obtains ultra-large calculating grid.Since this method is first to extract the association of each grid block
After information, piecemeal carries out original mesh independent reconstructed operation, so by reasonable piecemeal, the calculating grid of this method acquisition
Scale does not have the upper limit, can theoretically arrive infinity.This reconfiguration technique can realize hundred billion to trillion on common PC computers
The CFD of grade calculates grid generation.
Idiographic flow is：
Read original calculation mesh space point coordinates and boundary condition；
Establish the boundary segmentation index of each grid block；
According to the boundary condition of this block grid, the inner mark point information in each window side of this block grid is obtained；
Each grid block crest line coordinate is filled, obtains the internal informations such as coordinate extreme value；
It is scanned for calculating grid total space domain, obtains this block grid trappings point associated with other blocks；
According to encryption coefficient, in the case where ensureing both ends apart from equal proportion amplification, redistribution is carried out to the crest line of grid block；
Using all index points as waypoint, edge of window boundary line is divided again according to encryption coefficient and the original line segments regularity of distribution
Cloth；
Two dimension TFI is carried out, reconstruction window face is simultaneously projected using fuzzy normal direction projection technology, is ensureing projection accuracy
Original grid is kept to be distributed simultaneously；
Carry out three-dimensional standard TFI, reconstruction attractor grid block；
Encrypted space lattice block is checked, exports relevant information.
In above-mentioned flow, it is worth noting the following：
Firstth, the search of this block grid inner mark point, in grid block on six faces all inner windows boundary point, Suo Youxiang
Adjacent line section Curvature varying is more than 30% or point of the length ratio more than 40%, is inner mark point.
Secondth, the search of block grid and the relevant trappings point of other block grids.By the index point in this block grid with
Index point in its adjacent mesh block is compared, and if any misaligned index point, that is, is defined as this block grid and other grids
Relevant trappings point.
Third, the acquisition for parameterizing object plane grid file.It is true to object plane since the surface mesh of original mesh is not fine enough
Real shape describe it is not accurate enough, encrypt amplitude it is little when problem it is not serious, but when encryption amplitude it is larger, such as 6 times, 10
Times（Corresponding grid cell can increase to 216 to 1000 times）When, the problem of wall surface precision of original mesh is inadequate, can not just be ignored.
6 times of later surface topographies of conventional encryption, it can be seen that at fuselage bulge intersection, there are a large amount of sawtooth, are retouched to original configuration
It states not accurate enough.The solution taken herein is to parameterize object plane, is individually exported in Grid Generation Software a set of highly dense
The surface mesh of degree, since surface mesh is small for space lattice, memory consumption is little, therefore soft in business
It is feasible, and workload very little that a set of highdensity surface mesh is individually exported in part.Then when needing surface projection,
Projection object plane is replaced with to the high density object plane individually exported, it is possible to obtain the object plane shape of a high-fidelity.Increase parameter
The surface topography of compound face shadow casting technique, it can be seen that compared with original processing mode, parameterize the table obtained after object plane projection
Facial contour fidelity is much higher, can meet calculating needs completely.
4th, normal direction projection technology is obscured.Fig. 1 gives fuzzy normal direction projection technology signal proposed in this paper
Figure.Conventional wall surface projection pattern includes fixed-direction projection, local normal projection, cylindrical surface projecting, spherical projection and closest approach
Projection etc..Cylindrical surface projecting, spherical projection and closed point projection have its respective using limitation, when shape is very irregular this
Several ways cannot be applicable in completely.And when projection wall surface Curvature varying is larger, there is also ask for most common fixed-direction projection
Topic, as shown below, drop shadow green's line represents fixed-direction projection（It is typically chosen the average normal side of all units of the object plane
To）, at the both ends on perspective plane since surface normal direction and projecting direction gap are larger, can projection accuracy be reduced, and
It even will appear the phenomenon that can not projecting under extreme case；And if using local normal direction projection, i.e., the blue in figure is thrown
Hachure, this projection pattern, it is possible that projection line intersects, are caused in the case where adjacent object plane unit Curvature varying is very big
Projection failure, this causes this method, and there is also larger limitations.It is and orange in fuzz method line projection proposed in this paper, i.e. figure
Line, using a bulk of region of view field（Red color range）Projecting direction of the average normal as some object plane unit, when
When projection object plane unit moves in both directions, view field's also slightly translation therewith, even if can thus cause object plane list
First curvature changes greatly, and violent variation will not occur in projecting direction.Fuzz method line projection method combines fixed-direction throwing
The advantages of shadow and local normal project, while avoided the respective limitation of the two again, have to various irregularity faces very strong
Adaptability so that it is more accurate reliable that object plane projects this process.
Based on above-mentioned method, following embodiment is provided.
Embodiment one：
Such as Fig. 2, Fig. 3, give according to the front and rear space lattice profiles versus of this method reconstruct.It can be seen that each piece of grid
Space crest line maintains the molded line shape and distribution characteristics of original mesh, in the situation using spline-fit original space curve
Under, the density regularity of distribution of primitive curve is maintained, before the ratio and the reconstruct that also ensure two adjacent curve end distances
It is completely the same, be conducive to the numerical simulation of computational fluid dynamics.
Embodiment two：
As Fig. 4, Fig. 5 give the section grid comparison obtained using this method.It can be seen that, the topological structure of refined net with
Original mesh is identical, and the regularities of distribution of each sub- face four edges is similar to original mesh also height, this illustrate this method for
Two-dimensional section is completely applicable.
Embodiment three：
Example, respectively original mesh are reconstructed with the three-dimensional space grid of certain lift-rising configuration mark mould, encrypt 2 times of grids and encryption 8
Times grid, grid scale is respectively 4,310,000,34,550,000 and 22.1 hundred million.
Fig. 6 gives the surface mesh drawn game of certain 111,400,000,000 unit grid of cruise configuration mark mould obtained using this paper technologies
Portion's enlarged drawing.The calculating grid single file reaches 2694745581792 bytes, about 2.7TB.It can be seen that encrypted net
Lattice not only highly maintain original configuration type face, while also have very high mesh quality, in terms of slickness and orthogonality
It is completely the same with original mesh.At present in computational fluid dynamics circle, do not occurred also calculating such large-scale complex shape
The generation of grid and calculating simulation work, this application example also fully demonstrate the reliability and scalability of this paper technologies.
The invention is not limited in aforementioned specific embodiments.The present invention, which expands to, any in the present specification to be disclosed
The step of new feature or any new combination and any new method or process disclosed or any new combination.
Claims (6)
1. a kind of ultra-large calculating grid reconstruction method, it is characterised in that include the following steps：
Step 1：Read original calculation mesh space point coordinates and boundary condition；
Step 2：Establish the boundary segmentation index of each grid block；
Step 3：According to the boundary condition of this block grid, the inner mark point information in each window side of this block grid is obtained；
Step 4：Each grid block crest line coordinate is filled, obtains the internal informations such as coordinate extreme value；
Step 5：It is scanned for calculating grid total space domain, obtains this block grid trappings point associated with other blocks；
Step 6：According to encryption coefficient, in the case where ensureing both ends apart from equal proportion amplification, weight is carried out to the crest line of grid block
Distribution；
Step 7：Using all index points as waypoint, according to encryption coefficient and the original line segments regularity of distribution to edge of window boundary line into
Row redistribution；
Step 8：Two dimension TFI is carried out, reconstruction window face is simultaneously projected using fuzzy normal direction projection method, is ensureing to throw
Original grid is kept to be distributed while shadow precision；
Step 9：Carry out three-dimensional standard TFI, reconstruction attractor grid block；
Step 10：Encrypted space lattice block is checked, exports relevant information.
2. a kind of ultra-large calculating grid reconstruction method according to claim 1, it is characterised in that six in grid block
The boundary point of all inner windows, all adjacent segments Curvature varyings are more than 30% on face or length ratio is more than 40% point,
It is denoted as inner mark point.
3. a kind of ultra-large calculating grid reconstruction method according to claim 1 or 2, it is characterised in that will be in grid block
The inner mark point of inner mark point grid block adjacent thereto be compared, if any misaligned index point, be denoted as this block
Grid block and the relevant trappings point of other grid blocks.
4. a kind of ultra-large calculating grid reconstruction method according to claim 1, it is characterised in that obtaining grid block
In information process, the parametrization object plane grid file for projecting, and grid block and window belonging to positioning can be read in if necessary
Mouthful.
A kind of 5. ultra-large calculating grid reconstruction method according to claim 4, it is characterised in that the parameter compound
Veil lattice file is to parameterize object plane, and a set of highdensity surface mesh is individually exported in Grid Generation Software.
A kind of 6. ultra-large calculating grid reconstruction method according to claim 1, it is characterised in that the fuzzy normal
Direction projection method works as projection using projecting direction of the average normal in one piece of region of view field as some object plane unit
When object plane unit moves in both directions, view field's also slightly translation therewith, even if can thus so that object plane unit is bent
Rate changes greatly, and projecting direction is not in violent variation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711444253.8A CN108153706A (en) | 2017-12-27 | 2017-12-27 | A kind of ultra-large calculating grid reconstruction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711444253.8A CN108153706A (en) | 2017-12-27 | 2017-12-27 | A kind of ultra-large calculating grid reconstruction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108153706A true CN108153706A (en) | 2018-06-12 |
Family
ID=62463256
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711444253.8A Pending CN108153706A (en) | 2017-12-27 | 2017-12-27 | A kind of ultra-large calculating grid reconstruction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108153706A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109783965A (en) * | 2019-01-25 | 2019-05-21 | 中国空气动力研究与发展中心计算空气动力研究所 | A kind of structured grid automatic Partitioning encryption method |
| CN109920059A (en) * | 2019-03-14 | 2019-06-21 | 空气动力学国家重点实验室 | Ultra-large overlapping grid concurrent assembly method based on auxiliary grid |
| CN111354086A (en) * | 2018-12-24 | 2020-06-30 | 中国空气动力研究与发展中心超高速空气动力研究所 | Bidirectional three-dimensional scanning method suitable for grid position attribute judgment of DSMC (distributed data center) method |
| CN112906314A (en) * | 2021-02-01 | 2021-06-04 | 中国空气动力研究与发展中心计算空气动力研究所 | Depth compression storage and decompression method for multi-block structure grid data |
-
2017
- 2017-12-27 CN CN201711444253.8A patent/CN108153706A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111354086A (en) * | 2018-12-24 | 2020-06-30 | 中国空气动力研究与发展中心超高速空气动力研究所 | Bidirectional three-dimensional scanning method suitable for grid position attribute judgment of DSMC (distributed data center) method |
| CN111354086B (en) * | 2018-12-24 | 2023-04-14 | 中国空气动力研究与发展中心超高速空气动力研究所 | Bidirectional three-dimensional scanning method suitable for grid position attribute judgment of DSMC (distributed multi-media card) method |
| CN109783965A (en) * | 2019-01-25 | 2019-05-21 | 中国空气动力研究与发展中心计算空气动力研究所 | A kind of structured grid automatic Partitioning encryption method |
| CN109783965B (en) * | 2019-01-25 | 2022-08-02 | 中国空气动力研究与发展中心计算空气动力研究所 | Automatic block encryption method for structural grid |
| CN109920059A (en) * | 2019-03-14 | 2019-06-21 | 空气动力学国家重点实验室 | Ultra-large overlapping grid concurrent assembly method based on auxiliary grid |
| CN112906314A (en) * | 2021-02-01 | 2021-06-04 | 中国空气动力研究与发展中心计算空气动力研究所 | Depth compression storage and decompression method for multi-block structure grid data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108153706A (en) | A kind of ultra-large calculating grid reconstruction method | |
| Minetto et al. | An optimal algorithm for 3D triangle mesh slicing | |
| CN100561523C (en) | A kind of method for re-establishing three-dimensional model gridding | |
| Frey et al. | Mesh relaxation: a new technique for improving triangulations | |
| CN105787226A (en) | Reconstruction of parameterization model of quadrilateral finite element grid model | |
| CN101512633B (en) | A hair motion compositor system and optimization techniques for use in a hair/fur pipeline | |
| CN105678683A (en) | Two-dimensional storage method of three-dimensional model | |
| CN103810756B (en) | The method for drafting of self adaptive Loop subdivision curved surface based on irregular area | |
| Bronson et al. | Lattice cleaving: A multimaterial tetrahedral meshing algorithm with guarantees | |
| CN1877640A (en) | Geometric data subdivision method based on triangle interpolation surface subdivision | |
| CN101587596A (en) | Three-dimensional visual simulation method of lithography development process | |
| Luu et al. | Efficient slicing of Catmull–Clark solids for 3D printed objects with functionally graded material | |
| CN103325086B (en) | A kind of deformation method of the three-dimensional picture based on tetrahedral coordinate system | |
| CN109685844B (en) | Method and device for optimizing marking and direction calibration of woven grids | |
| CN111028349B (en) | Hierarchical construction method suitable for rapid visualization of massive three-dimensional live-action data | |
| CN103886635A (en) | Self-adaption LOD model establishing method based on face clustering | |
| CN104537704B (en) | A kind of feather real-time dynamic creation method of birds human trunk model | |
| CN104574479B (en) | The rapid generation of the single feather of birds in a kind of three-dimensional animation | |
| Wang et al. | A novel slicing-based regularization method for raw point clouds in visible IoT | |
| CN109783965A (en) | A kind of structured grid automatic Partitioning encryption method | |
| Lv et al. | Adaptively isotropic remeshing based on curvature smoothed field | |
| Leng et al. | Rapid simplification of 3D geometry model of mechanisms in the digital twins-driven manufacturing system design | |
| CN104794747A (en) | Three-dimensional point cloud data simplification algorithm based on ray theory | |
| CN101739723A (en) | Intelligent layout method for free curved surface network | |
| CN113470177A (en) | Three-dimensional model geometric self-adaptive simplification method in GIS system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |