CN103294854A - Design space division and recombination technique - Google Patents
Design space division and recombination technique Download PDFInfo
- Publication number
- CN103294854A CN103294854A CN2013101783428A CN201310178342A CN103294854A CN 103294854 A CN103294854 A CN 103294854A CN 2013101783428 A CN2013101783428 A CN 2013101783428A CN 201310178342 A CN201310178342 A CN 201310178342A CN 103294854 A CN103294854 A CN 103294854A
- Authority
- CN
- China
- Prior art keywords
- design space
- simulation
- space
- subspace
- model
- 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
Images
Abstract
The invention relates to a division and recombination technique of design space for high-speed simulation of a large linear system. According to the basic principle, the whole design space is divided first by full utilization of superposition of the linear system, and subspaces are superposed and recombined for secondary simulation. The technique includes: building a simulation object geometric model by three-dimensional modeling software; dividing the design space for the geometric model into subspaces; subjecting each subspace to primary simulation calculation; calculating to obtain complete area models for the subspaces; recombing the space by the subspace complete area models; and using the recombined space as a new equivalent design space for secondary simulation calculation. The starting point of the design space division and recombination technique is far higher than that of the existing simulation technique in the market, and the technique tends to lead the upgrade of electromagnetic simulation technologies.
Description
Technical field
The present invention relates to a kind of design space partition combination technique, be applicable to but be not limited to the design space partition combination technique in Electromagnetic Simulation field.
Background technology
Existing Electromagnetic Simulation technology on the domestic and international market, adopt method of moment, boundary element method, finite element method, method of difference or finite volume method, Maxwell's partial differential equation is carried out discretize, it is approximated to a hundreds of thousands to the huge linear system of millions of unknown numbers.Calculate the result of this system, required time is often counted in the sky, and even week counts even month meter; And the required hardware resource of computing, also non-general user can bear.Therefore, the electromagnetic field simulation software in the market now, needs obtain result's characteristics in real time in the time of can't adapting to design at all, generally can only be for to insensitive checking of time and analytic activity, namely they belong to the analytic type simulation software.
This emulation technology bottleneck, the development that has hindered Electromagnetic Simulation calculating is nearly 20 years.Even the research of current forefront also just is divided into submatrix with this super large system of linear equations, then by synchronous use multinuclear multithreading computing technique, calculated sub-matrix is accelerated arithmetic speed respectively.Problem is, all submatrixs must calculate synchronously, therefore comes from reality, and this method is still with large resource, the big input and exchanges the short time for, does not fundamentally break through the predicament that prior art faces.At present, academia is the seldom research activities of this respect, and industry member is stuck to the old path substantially, and the expectation hardware advances can continue to follow " Moore's Law " speed-raising and then solve the slow problem of simulation calculating by hardware.Even be in the HFSS software of electromagnetic-field simulation forefront, also fail free from vulgarity, the design mode simulation software product that never provides a market to need badly for many years.
Summary of the invention
In order to overcome above deficiency, the present invention has designed a kind of design space partition combination technique that is applicable to large-scale linear system high-speed simulation according to the principle of the superposability of linear system.
Major technique of the present invention thes contents are as follows:
A kind of design space partition combination technique may further comprise the steps:
(1), set up the geometric model of simulation object with 3 d modeling software, all elements are included in the design space limited, that can represent in computing machine;
(2), the residing design space of geometric model model is divided into a plurality of less subspaces, the basic demand of division is to divide the subspace surface configuration comparison rule that the back forms, and is easy to connect;
(3), the ready-portioned design space of step (2) model is carried out the simulation calculating first time respectively;
(4), behind the simulation calculating, extract the complete area model of subspace, comprise parameters such as how much, material, border, coupling, mathematics;
(5), with subspace complete area model stack and reorganization space that step (4) obtains, obtain one with the luv space equivalence but the simpler new design space that is easier to calculate;
(6), the new design space after will recombinating carries out the secondary simulation calculating as new simulation space, obtains the solution of a physical quantity in the original design space.
The algorithm of the simulation calculating of above-mentioned steps (3) and step (6) is applicable to any in the traditional simulation algorithms such as finite element method, method of moment, boundary element method, method of difference, finite volume method.
The complete area model of the subspace of above-mentioned steps (4) comprises parameters such as how much, material, border, coupling, mathematics.
The equivalent design space of above-mentioned steps (5) reorganization, all physical phenomenons that can reappear the original design space.
By technique scheme, the present invention has following advantage at least:
Compare this project innovation part with the technology that adopt existing Electromagnetic Simulation software on the market and they:
The brand-new simulation algorithm of independent research.By the split and reorganization to the design space, avoided the large linear system that produces after Maxwell's partial differential equation discretize, reduce the demand to computational resource when improving simulation velocity, broken through first in the world and hindered the nearly 20 years technical bottleneck of Electromagnetic Simulation calculating development.Compare with mark post product on the market, speed has improved several magnitude; And gained result " complete area model " can be used repeatedly, has further improved work efficiency.
The specific embodiment of the present invention is provided in detail by following examples and accompanying drawing.
Description of drawings
Fig. 1 is the primal system C among the embodiment;
Fig. 2 is the subsystem A of the decomposition among the embodiment;
Fig. 3 is the subsystem B of the decomposition among the embodiment.
Embodiment
Reach technological means and the effect that predetermined goal of the invention is taked for further setting forth the present invention, below in conjunction with accompanying drawing and preferred embodiment, to its embodiment, structure, feature and the effect thereof that foundation the present invention proposes, describe in detail as after.
A kind of design space partition combination technique, particularly Electromagnetic Simulation field may further comprise the steps:
(1), set up the geometric model of simulation object with 3 d modeling software, all elements are included in the design space limited, that can represent in computing machine;
(2), the residing design space of geometric model model is divided into a plurality of less subspaces, the basic demand of division is to divide the subspace surface configuration comparison rule that the back forms, and is easy to connect;
(3), the ready-portioned design space of step (2) model is carried out the simulation calculating first time respectively;
(4), behind the simulation calculating, extract the complete area model of subspace, comprise parameters such as how much, material, border, coupling, mathematics;
(5), with subspace complete area model stack and reorganization space that step (4) obtains, obtain one with the luv space equivalence but the simpler new design space that is easier to calculate;
(6), the new design space after will recombinating carries out the secondary simulation calculating as new simulation space, obtains the solution (as electromagnetic field distribution etc.) of a physical quantity in the original design space.
The algorithm of the simulation calculating of above-mentioned steps (3) and step (6) is applicable to any in the traditional simulation algorithms such as finite element method, method of moment, boundary element method, method of difference, finite volume method.
The complete area model of the subspace of above-mentioned steps (4) comprises parameters such as how much, material, border, coupling, mathematics.
The equivalent design space of above-mentioned steps (5) reorganization, all physical phenomenons that can reappear the original design space.
Embodiment:
A system can come the characteristic of this system of equivalents with transmitting parameter T, scattering parameter S, admittance parameter Y, impedance parameter Z, and other parameter that similarly, can be converted by above-mentioned parameter.At this with the general representative system parameter of Q.
We are example with the C system.Primal system is for being illustrated in figure 1 as V2=[C] U1: wherein U and V represent two kinds of physical quantitys respectively, and numeral 1 and 2 represents the port label.
Through after the partition (as shown in phantom in Figure 1) once, this system can resolve into A and two subsystems of B, as Fig. 2 and as shown in Figure 3, and (notes: in theory, cutting is repeatedly arbitrarily in the C system):
Emulation for the first time to the carrying out of subsystem A and B, obtains to represent the parameter Q of their system performances:
U2 = [QA] U1 V2 = [QB] V1
Wherein U and V represent two kinds of physical quantitys respectively, and numeral 1 and 2 represents the port label.
Different with the traditional Q parameter is, the Q parameter among the present invention is at discrete point definition one by one on the cut surface, the corresponding port of each discrete point; And classic method defines at whole cut surface, the corresponding port of face.The port number N of the cut surface among the present invention is to be defined according to accuracy requirement by the user.
After obtaining the Q parameter of each subsystem, they and the complete area model of synthetic each subspace of parameter group such as separately how much, material, border, coupling, mathematics (Complete Domain Model-CDM).
The complete area model that calculated in advance is good is as the assembly unit that plays with building blocks.Then, check whether each interface mates mutually at aspects such as shape, size, material, grids.In this way, then two sub spaces are shared and are mated the face that connects, and can connect face combination refitting by this.All connecing that circulate faced, as all passing through matching check, new space that is equivalent to the original design space then, can be fully by complete area model (CDM) combined copy of subspace; As not by the subspace of matching check, its internal physical phenomenon can be calculated by the traditional simulation method.
During the connect face combination of two subspaces by a coupling, employing be the processing mode different with traditional excitation-response.For high speed processing connects ports a large amount of on the face, will encourage with response and make equivalent processes, traditional representation:
U2 = [QA] U1 V2 = [QB] V1
Just can be rewritten into:
0 = [QA，IA][U1，U2]T 0 = [QB, IB][V1, V2]T
Wherein I represents unit matrix, and T represents transpose of a matrix.The complete area model combination of two sub spaces, above-mentioned two equations are softened lumps together exactly, eliminates the influence of coupling port U2=V1 then, after the rewrite equation:
V2 = [ f(QA, IA, QB, IB)] U1
Wherein f represents that the characterisitic parameter of this combined system is the function of the special new argument of two subsystems and unit matrix, as long as obtained the parameter of subsystem in advance, the parameter of combined system just can be calculated by them, and is equal to the primal system before the partition:
V2 = [C] U1。
The above, it only is preferred embodiment of the present invention, be not that the present invention is done any pro forma restriction, though the present invention discloses as above with preferred embodiment, yet be not in order to limit the present invention, any those skilled in the art, in not breaking away from the technical solution of the present invention scope, when the technology contents that can utilize above-mentioned announcement is made a little change or is modified to the equivalent embodiment of equivalent variations, in every case be the content that does not break away from technical solution of the present invention, any simple modification that foundation technical spirit of the present invention is done above embodiment, equivalent variations and modification all still belong in the scope of technical solution of the present invention.
Claims (4)
1. design space partition combination technique is characterized in that may further comprise the steps:
(1), set up the geometric model of simulation object with 3 d modeling software, all elements are included in the design space limited, that can represent in computing machine;
(2), the residing design space of geometric model model is divided into a plurality of less subspaces, the basic demand of division is to divide the subspace surface configuration comparison rule that the back forms, and is easy to connect;
(3), the ready-portioned design space of step (2) model is carried out the simulation calculating first time respectively;
(4), behind the simulation calculating, extract the complete area model of subspace, comprise parameters such as how much, material, border, coupling, mathematics;
(5), with subspace complete area model stack and reorganization space that step (4) obtains, obtain one with the luv space equivalence but the simpler new design space that is easier to calculate;
(6), the new design space after will recombinating carries out the secondary simulation calculating as new simulation space, obtains the solution of a physical quantity in the original design space.
2. design space according to claim 1 partition combination technique, it is characterized in that: the algorithm of the simulation calculating of described step (3) and step (6) is applicable to any in the traditional simulation algorithms such as finite element method, method of moment, boundary element method, method of difference, finite volume method.
3. design space according to claim 1 partition combination technique is characterized in that: the complete area model of the subspace of described step (4) comprises parameters such as how much, material, border, coupling, mathematics.
4. partition combination technique in design space according to claim 1 is characterized in that: the equivalent design space of described step (5) reorganization, all physical phenomenons that can reappear the original design space.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013101783428A CN103294854A (en) | 2013-05-14 | 2013-05-14 | Design space division and recombination technique |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013101783428A CN103294854A (en) | 2013-05-14 | 2013-05-14 | Design space division and recombination technique |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103294854A true CN103294854A (en) | 2013-09-11 |
Family
ID=49095711
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2013101783428A Pending CN103294854A (en) | 2013-05-14 | 2013-05-14 | Design space division and recombination technique |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103294854A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030137511A1 (en) * | 2001-04-23 | 2003-07-24 | Naruki Aruga | Electromagnetic wave simulation device and method, amd three-dimensional model data creation apparatus |
| CN102033985A (en) * | 2010-11-24 | 2011-04-27 | 南京理工大学 | High-efficiency time domain electromagnetic simulation method based on H matrix algorithm |
| CN102722618A (en) * | 2012-06-04 | 2012-10-10 | 北京航空航天大学 | Method for building and paralleling quasi three-dimensional electromagnetic environmental model on basis of parabolic equation |
-
2013
- 2013-05-14 CN CN2013101783428A patent/CN103294854A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030137511A1 (en) * | 2001-04-23 | 2003-07-24 | Naruki Aruga | Electromagnetic wave simulation device and method, amd three-dimensional model data creation apparatus |
| CN102033985A (en) * | 2010-11-24 | 2011-04-27 | 南京理工大学 | High-efficiency time domain electromagnetic simulation method based on H matrix algorithm |
| CN102722618A (en) * | 2012-06-04 | 2012-10-10 | 北京航空航天大学 | Method for building and paralleling quasi three-dimensional electromagnetic environmental model on basis of parabolic equation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Pace | BSP vs MapReduce | |
| CN107069696B (en) | A kind of parallel calculating method of Power system state estimation | |
| CN104570081A (en) | Pre-stack reverse time migration seismic data processing method and system by integral method | |
| Luo et al. | Study on multi-view model for cloud manufacturing | |
| CN108427776A (en) | Distributed all digital combined simulation system and method | |
| She et al. | Front-tracking methods | |
| JP2024501579A (en) | Spatial connection query method, apparatus, electronic device and storage medium | |
| CN106250101A (en) | Migration before stack method for parallel processing based on MapReduce and device | |
| Fan et al. | Simulating brittle fracture with material points | |
| CN104299170B (en) | Intermittent energy source mass data processing method | |
| CN106055768A (en) | High-safety number simulation method, solver and system based on cloud computing | |
| Tishkin et al. | Justification of Godunov’s scheme in the multidimensional case | |
| CN105825348A (en) | Method for quantitatively analyzing influence exerted by bilateral trade on power grid power transmission cost | |
| CN103294854A (en) | Design space division and recombination technique | |
| JP2017004305A (en) | Analytic model creation support system, analytic model creation support device, and analytic model creation support program | |
| CN107808021B (en) | CFD-based fluid device resistance calculation method | |
| CN104965846A (en) | Virtual human establishing method on MapReduce platform | |
| Zhao et al. | A large‐scale parallel hybrid grid generation technique for realistic complex geometry | |
| CN103942376A (en) | Collision correction method for real-time hair motion simulation | |
| CN104978382A (en) | Clustering method based on local density on MapReduce platform | |
| CN104090813A (en) | Analysis modeling method for CPU (central processing unit) usage of virtual machines in cloud data center | |
| CN103246793A (en) | Method for drawing tire grounding cloud picture based on ABAQUS analysis result | |
| CN103164557B (en) | A kind of Meta Materials design apparatus and method for designing thereof | |
| CN105512361A (en) | Substructure analysis method based on ANSYS software | |
| CN107609266A (en) | Fitting-out work management information system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20130911 |