CN106092496A - For the APDSMC flow field detection method flowed across yardstick - Google Patents
For the APDSMC flow field detection method flowed across yardstick Download PDFInfo
- Publication number
- CN106092496A CN106092496A CN201610411018.XA CN201610411018A CN106092496A CN 106092496 A CN106092496 A CN 106092496A CN 201610411018 A CN201610411018 A CN 201610411018A CN 106092496 A CN106092496 A CN 106092496A
- Authority
- CN
- China
- Prior art keywords
- particle
- grid
- collision
- flow field
- detection method
- 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
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M9/00—Aerodynamic testing; Arrangements in or on wind tunnels
- G01M9/06—Measuring arrangements specially adapted for aerodynamic testing
Landscapes
- Physics & Mathematics (AREA)
- Fluid Mechanics (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A kind of for the APDSMC flow field detection method flowed across yardstick, the particle mobility model of uniform motion is used by stream field after being initialized, utilize vector cross-products judge particle move after change logic within a grid, obtain the positional information after simulation particle updates and network topology information；Then particle based on grid numbering and particle based on particle numbering are mapped；Finally by simulation particle collision, the particle rapidity distribution in grid is updated, and detects by realizing flow field after stream field intelligence sample；Present invention scheme based on progressive holding processes the calculating demand of zones of different, it is possible to the more intelligent collision kernel selecting different computational fields, thus accelerates former DSMC method so that it is bigger that new algorithm can calculate yardstick span, denser example condition.
Description
Technical field
The present invention relates to the technology of a kind of field of fluid mechanics, a kind of for the progressive guarantor flowed across yardstick
Hold direct simulation Monte Carlo (DSMC) the flow field detection method of (Asymptotic Preserving).
Background technology
Along with the development of aeronautical and space technology, people have deeper and deeper demand to higher and faster aircraft.
But be on the one hand continuously increased along with height and the speed of aircraft, conventional numeric analogue technique (CFD) is difficult to competent new flight
Device early grind task under new flying condition；The research worker of the newest aircraft needs again substantial amounts of numerical simulation badly to solve
The most complicated flow field problem.In this context, DSMC algorithm has obtained flourish.DSMC algorithm can accurate simulation superb
Mobilization force under velocity of sound thin fluids and flowing heat problem, be the strong alternative solving course of new aircraft design existing issue
One of.But we have found that when flight service is in the area of pneumatic multiple dimensioned leap, tradition CFD is not still because precision problem can simultaneously
Using, DSMC then can be difficult with due to the problem of amount of calculation.
In prior art, DSMC Yu EPSM mixed method is to solve DSMC to calculate speed issue in pneumatic multiple dimensioned basin
Primary solutions.The program is recognized and can be simulated by local Maxwell distribution at stream district molecular collision continuously, thus greatly
Decrease greatly DSMC algorithm required time, therefore carry out stream field by contrast Collision Number and carry out simple subregion.But this simply
Mixed method exist and can not correctly identify zoning condition, the problem such as transition region method choice difficulty so that this mixing side
There is the problems such as series of computation precision and computational efficiency in method.
Summary of the invention
The present invention is directed to deficiencies of the prior art, propose a kind of for the APDSMC flow field inspection flowed across yardstick
Survey method, scheme based on progressive holding (AP, asymptotic preserving) thinking processes the calculating of zones of different and needs
Ask, it is possible to the more intelligent collision kernel selecting different computational fields, thus accelerate former DSMC method so that new algorithm can calculate
Yardstick span is bigger, denser example condition.
The present invention is achieved by the following technical solutions:
The present invention relates to a kind of for the APDSMC flow field detection method flowed across yardstick, adopt after being initialized by stream field
With the particle mobility model of uniform motion, utilize vector cross-products judge particle move after change logic within a grid, obtain mould
Intend the positional information after particle updates and network topology information；Then by particle based on grid numbering and particle based on particle
Numbering maps；Finally by simulation particle collision, the particle rapidity distribution in grid is updated, and passes through stream field
Flow field detection is realized after intelligence sample.
Described flow field initializes and includes: MPI (program message passing interface, Message Passing Interface)
Initialization operation and the operation of initialization flow field.
Described MPI initialization operation refers to: calls MPICH2 built-in function and starts the processor participating in calculating.
Described MPICH2 built-in function is by obtaining in MPICH official website；
Described initialization flow field operation includes: a. calculates volume and the topological structure reading in grid；B. simulation grain is dispensed
Son is also numbered；C. the initial information of calculation of boundary conditions.
Described flow field is initialized and is capable of being instructed, by MPICH2 built-in function, the computation processor proposition needed, with
Time flow field initialize after, given flow field initial value so that calculate can carry out along time orientation.
Described simulation particle collision, uses progressive holding (AP) mode to realize, and concrete steps include:
The first step, to for the governing equation across yardstick flow field problem, i.e. Boltzmann equation carries out rigidity source item and divides
Split；
Described Boltzmann equation is:Wherein:(f f) is Q
Collision term, t describes the time；X describes position；F is the velocity distribution function of the described total space；υ is the speed of the described total space
Degree；ε is the characteristic constant after equation dimensionless, i.e. Knudsen number.
Described rigidity source item, (f, f), owing to the collision between molecule is subject to be the collision term Q of Boltzmann equation
The impact of factors, and be difficult in time scale when flowing is in nonequilibrium state carry out modelling process, so can only use
Less time step resolves this source item.Due to the difficulty processed, so referred to as rigidity source item.
Described rigidity source item divides first by BGK (Bhatnagar Gross Krook model) penalty method, specifically
For:Wherein: t describes the time；F is the VELOCITY DISTRIBUTION letter of the described total space
Number；ε is the characteristic constant after equation dimensionless, i.e. Knudsen number；β is penalty coefficient；(f f) is collision term to Q；P (f) is punishment source
?；Use this division, rigidity source item can be divided into rigidity and non-rigid two parts.
Second step, uses the rigidity source item after division and penalizes method continuously, the numerical model after being optimized；
Described method of penalizing continuously refers to: the rigid element continuous penalty function of use to after division:Wherein:The coefficient of relaxation advanced for VELOCITY DISTRIBUTION, L is punishment number of times,
In calculating, penalty factor selects β ＞ σ-To guarantee to protect positivity, M**It is distributed for Maxwell based on grid internal statistical information, finally
The numerical model equation solving Boltzmann equation after the optimization obtained is:
Wherein: fnIt is the VELOCITY DISTRIBUTION in the n-th step grid, f*For mistake
Crossing VELOCITY DISTRIBUTION, υ is particle rapidity, b1, b2, b3For the separation of collision, β*For based on f*Penalty coefficient,For
Reduction collision term, fn+1It it is the VELOCITY DISTRIBUTION in the (n+1)th step grid.
Described penalty factor is based on σ-Accuracy selection, thus fundamentally achieve the adaptive function of algorithm；
3rd step, the numerical model after optimizing is converted to algorithm the collision part being applied in DSMC and lax napex
Point, wherein: collision part uses the VHS collision model of DSMC design to calculate, and lax item parts uses the EPSM of Pullin
Kernel calculates.
Described stream field intelligence sample, uses fluid macrovariable to solve with the statistics relation of molecule micro-variable,
The chaos of this relational dependence molecule it is assumed that simulation particle is regarded as true molecular by the present invention, utilize simulation particle speed and
Positional information carries out sampling statistics and obtains information of flow.
The present invention relates to a kind of system realizing said method, including: DSMC module, parallel MPI module and AP collision
Kernel module, wherein: DSMC module is connected with AP crash module with the transmission speed of simulation particle, position and topology information and changes
Becoming collision model, DSMC module is connected with MPI module and transmission grid topology information, the speed of particle, position and topology information
To realize parallel processing state.
Technique effect
Compared with prior art, the present invention is in yardstick flow field problem, and the time foreshortens to about 1/10th, counts simultaneously
Calculate precision to keep being basically unchanged.
Accompanying drawing explanation
Fig. 1 is schematic flow sheet of the present invention；
Fig. 2 is for surveying spray design conditions schematic diagram；
In figure: a-quadrant x is 0～0.060m, B region x is 0.060～0.065m, and C region x is 0.065～0.200m；
Fig. 3 is side impact flow field and bundle factor cloud atlas；
In figure: SR be Disengagement zone, RZ be recirculating zone, BaS be barrel-shaped shock wave, BoS be bowshock, MD be mach disk；
Fig. 4 is embodiment bundle factor cloud atlas；
In figure: a is coefficient b2Cloud atlas, b are coefficient b3Cloud atlas；
Fig. 5 is embodiment flat-plate compressed force coefficient line chart.
Detailed description of the invention
Cluster used in the present embodiment is that Shanghai Communications University is super to be calculated, cpu model Intel Xeon CPU E5 2670, behaviour
Making system environments Red Hat Enterprise Linux Server release 6.3, compiler selects Intel MPI
5.0.1‐Fortran Compiler.Using check figure is 8.
The present embodiment is to calculate a flat board side spray problem flowed across yardstick, is used for contrasting APDSMC and DSMC method and exists
Advantage in terms of speed.As in figure 2 it is shown, do three pieces of grids in shown computational fields, grid node number is 241 × 401,21 respectively
× 401,541 × 401.It is uniformly distributed in the middle of flow field
It is as shown in the table to be respectively provided with inlet flow conditions and jet flow condition:
Table standard example calculates parameter
The present embodiment is finally calculated motion pattern and pressure contour figure as shown in Figure 3: use this method to grab
Live flow field structure classical in the middle of a side spray flow field, swash including SR Disengagement zone, RZ recirculating zone, the barrel-shaped shock wave of BaS, BoS arch
Ripple, MD mach disk, characterize the overall accuracy of algorithm improvement not defluidization field computational methods.
The present embodiment is finally calculated bundle factor cloud atlas as shown in Figure 4: at coefficient b in the middle of yardstick flow field2With b3
Distribution, characterize new algorithm in yardstick flow field, having adaptive ability.
In the present embodiment, last calculated flat-plate compressed force coefficient line chart is as shown in Figure 5: the pressure line chart of flat board pressure,
Prove that the precision of new algorithm is not less than tradition DSMC algorithm by the contrast of more accurate lines.
It is 8.1 times of APDSMC method that the present embodiment finally calculates time tradition DSMC
This example demonstrates that in the case of accuracy guarantee is constant, APDSMC has greatly accelerated the speed of DSMC method.
Above-mentioned be embodied as can by those skilled in the art on the premise of without departing substantially from the principle of the invention and objective with difference
Mode it is carried out local directed complete set, protection scope of the present invention is as the criterion with claims and is not embodied as institute by above-mentioned
Limit, each implementation in the range of it is all by the constraint of the present invention.
Claims (7)
1. one kind for the APDSMC flow field detection method flowed across yardstick, it is characterised in that adopt after being initialized by stream field
With the particle mobility model of uniform motion, utilize vector cross-products judge particle move after change logic within a grid, obtain mould
Intend the positional information after particle updates and network topology information；Then by particle based on grid numbering and particle based on particle
Numbering maps；Finally by simulation particle collision, the particle rapidity distribution in grid is updated, and passes through stream field
Flow field detection is realized after intelligence sample.
Detection method the most according to claim 1, is characterized in that, described flow field initializes and includes: MPI initialization operation
Operate, wherein with initializing flow field:
Described MPI initialization operation refers to: calls MPICH2 built-in function and starts the processor participating in calculating；
Described initialization flow field operation includes: a. calculates volume and the topological structure reading in grid；B. simulation particle is dispensed also
It is numbered；C. the initial information of calculation of boundary conditions.
Detection method the most according to claim 1, is characterized in that, described simulation particle collision, uses progressive holding side
Formula realizes, and concrete steps include:
The first step, to for the governing equation across yardstick flow field problem, i.e. Boltzmann equation carries out rigidity source item division；
Second step, uses the rigidity source item after division and penalizes method continuously, the numerical model after being optimized；
3rd step, the numerical model after optimizing is converted to algorithm the collision part being applied in DSMC and lax item parts,
Wherein: collision part uses the VHS collision model of DSMC design to calculate, and lax item parts uses the EPSM kernel of Pullin
Calculate.
Detection method the most according to claim 3, is characterized in that, described Boltzmann equation is:Wherein:(f, f) is collision term to Q, i.e. rigidity source item, and t describes the time；x
Position is described；F is the velocity distribution function of the described total space；υ is the speed of the described total space；After ε is equation dimensionless
Characteristic constant, i.e. Knudsen number.
Detection method the most according to claim 3, is characterized in that, described rigidity source item division is punished first by BGK
Method, particularly as follows:Wherein: t describes the time；F is the speed of the described total space
Degree distribution function；ε is the characteristic constant after equation dimensionless, i.e. Knudsen number；β is penalty coefficient；(f f) is collision term to Q；P(f)
For punishment source item；Use this division, rigidity source item can be divided into rigidity and non-rigid two parts.
Detection method the most according to claim 3, is characterized in that, described method of penalizing continuously refers to: to division after firm
Property part use continuous penalty function:Wherein:Advance for VELOCITY DISTRIBUTION
Coefficient of relaxation, L is punishment number of times, and in calculating, penalty factor selects β ＞ σ-To guarantee to protect positivity, M**For believing based on grid internal statistical
The Maxwell distribution of breath, the numerical model equation solving Boltzmann equation after the optimization finally obtained is:
Wherein: fnIt is the VELOCITY DISTRIBUTION in the n-th step grid, f*For transition speed
Distribution, υ is particle rapidity, b1, b2, b3For the separation of collision, β*For based on f*Penalty coefficient,Touch for reduction
Hit item, fn+1It it is the VELOCITY DISTRIBUTION in the (n+1)th step grid.
7. the system of arbitrary described detection method in a claim 1～6, it is characterised in that including: DSMC module, parallel
MPI module and AP collide kernel module, wherein: DSMC module be connected with AP crash module with the speed transmitting simulation particle,
Position and topology information also change collision model, and DSMC module is connected with MPI module and transmission grid topology information, the speed of particle
Degree, position and topology information are to realize parallel processing state.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610411018.XA CN106092496A (en) | 2016-06-14 | 2016-06-14 | For the APDSMC flow field detection method flowed across yardstick |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610411018.XA CN106092496A (en) | 2016-06-14 | 2016-06-14 | For the APDSMC flow field detection method flowed across yardstick |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106092496A true CN106092496A (en) | 2016-11-09 |
Family
ID=57845565
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610411018.XA Pending CN106092496A (en) | 2016-06-14 | 2016-06-14 | For the APDSMC flow field detection method flowed across yardstick |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106092496A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108920872A (en) * | 2018-07-26 | 2018-11-30 | 上海交通大学 | For the BCP particle localization realization method and system of DSMC method |
| CN109377510A (en) * | 2018-09-11 | 2019-02-22 | 青岛海洋科学与技术国家实验室发展中心 | Particles track method and system on supercomputer cluster |
| CN111222240A (en) * | 2020-01-06 | 2020-06-02 | 中国人民解放军国防科技大学 | Thermochemical unbalanced flow field data calculation method and device accelerated by GPU |
| 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 |
| CN113449838A (en) * | 2021-07-05 | 2021-09-28 | 中国人民解放军国防科技大学 | Biological particle cluster construction method based on BCCA optimization model |
| CN113791912A (en) * | 2021-11-11 | 2021-12-14 | 中国空气动力研究与发展中心计算空气动力研究所 | MPI + X-based DSMC parallel computing method, equipment and medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101242619A (en) * | 2008-03-03 | 2008-08-13 | 华中科技大学 | A method for simulating motion process of mobile station based on particle motion |
| CN104376151A (en) * | 2014-10-30 | 2015-02-25 | 北京宇航***工程研究所 | Vacuum disturbance pilot flow field simulation method for rocket engine |
| CN105278346A (en) * | 2015-11-06 | 2016-01-27 | 北京航空航天大学 | Thermal fluid simulation method based on discrete lattice Boltzmann dual-distribution model |
-
2016
- 2016-06-14 CN CN201610411018.XA patent/CN106092496A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101242619A (en) * | 2008-03-03 | 2008-08-13 | 华中科技大学 | A method for simulating motion process of mobile station based on particle motion |
| CN104376151A (en) * | 2014-10-30 | 2015-02-25 | 北京宇航***工程研究所 | Vacuum disturbance pilot flow field simulation method for rocket engine |
| CN105278346A (en) * | 2015-11-06 | 2016-01-27 | 北京航空航天大学 | Thermal fluid simulation method based on discrete lattice Boltzmann dual-distribution model |
Non-Patent Citations (4)
| Title |
|---|
| BIN ZHANG ET AL: ""An asymptotic preserving Monte Carlo method for the multispecies Boltzmann equation"", 《JOURNAL OF COMPUTATIONAL PHYSICS》 * |
| WEI REN ET AL: ""An asymptotic-preserving Monte Carlo method for the Boltzmann equation"", 《JOURNAL OF COMPUTATIONAL PHYSICS》 * |
| 王保国 等: ""四种飞行器绕流的三维DSMC计算与传热分析"", 《航空动力学报》 * |
| 王沫然: ""维纳尺度气体流动和换热的Monte Carlo模拟"", 《中国优秀博硕士学位论文全文数据库（博士） 工程科技Π辑》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108920872A (en) * | 2018-07-26 | 2018-11-30 | 上海交通大学 | For the BCP particle localization realization method and system of DSMC method |
| CN109377510A (en) * | 2018-09-11 | 2019-02-22 | 青岛海洋科学与技术国家实验室发展中心 | Particles track method and system on supercomputer cluster |
| CN109377510B (en) * | 2018-09-11 | 2021-07-09 | 青岛海洋科学与技术国家实验室发展中心 | Particle tracking method and system on supercomputing cluster |
| 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 |
| CN111222240A (en) * | 2020-01-06 | 2020-06-02 | 中国人民解放军国防科技大学 | Thermochemical unbalanced flow field data calculation method and device accelerated by GPU |
| CN111222240B (en) * | 2020-01-06 | 2022-08-19 | 中国人民解放军国防科技大学 | Thermochemical unbalanced flow field data calculation method and device accelerated by GPU |
| CN113449838A (en) * | 2021-07-05 | 2021-09-28 | 中国人民解放军国防科技大学 | Biological particle cluster construction method based on BCCA optimization model |
| CN113791912A (en) * | 2021-11-11 | 2021-12-14 | 中国空气动力研究与发展中心计算空气动力研究所 | MPI + X-based DSMC parallel computing method, equipment and medium |
| CN113791912B (en) * | 2021-11-11 | 2022-02-11 | 中国空气动力研究与发展中心计算空气动力研究所 | MPI + X-based DSMC parallel computing method, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106092496A (en) | For the APDSMC flow field detection method flowed across yardstick | |
| CN103345577B (en) | Variation Bayesian probability assumed density multi-object tracking method | |
| CN102490909B (en) | Multi-body separating analogy method of aircraft | |
| CN105843073A (en) | Method for analyzing wing structure aero-elasticity stability based on aerodynamic force uncertain order reduction | |
| CN106650046B (en) | A kind of unsteady characteristic acquisition methods in Ship Air flow field | |
| CN111079859B (en) | Passive multi-station multi-target direction finding cross positioning and false point removing method | |
| RU2012102394A (en) | METHOD FOR CALCULATING PHYSICAL VALUES, METHOD FOR NUMERICAL ANALYSIS, PROGRAM FOR CALCULATING PHYSICAL VALUES, PROGRAM FOR NUMERICAL ANALYSIS, DEVICE FOR CALCULATING PHYSICAL VALUES AND DEVICES FOR NUMERICAL ANALYSIS | |
| El Khoury et al. | Crossflow past a prolate spheroid at Reynolds number of 10000 | |
| CN103729564A (en) | Pressure field calculating method and device based on particle image velocimetry technology | |
| CN103427791B (en) | A kind of signal separating method based on particle group optimizing | |
| CN103353910A (en) | Circuit partitioning method for parallel circuit simulation | |
| CN105197255B (en) | A kind of air intake duct curve design method | |
| Sutter et al. | Examining subgrid models of supermassive black holes in cosmological simulation | |
| Seleznev | Kernel density estimation in the study of star clusters | |
| CN105974471B (en) | A kind of quick forward modelling method of the more GPU of seismic data based on asynchronous flow | |
| CN116956619A (en) | Civil engineering model virtual assembly method in lightweight environment | |
| CN102073796B (en) | Lattice traveling method for simulating solute three-dimensional transport process | |
| CN105760478A (en) | Large-scale distributed data clustering method based on machine learning | |
| CN108763560A (en) | 3 d human motion search method based on graph model | |
| CN106056124A (en) | Relief extraction method based on three-dimensional curved surface loop guidance | |
| CN108090514B (en) | Infrared image identification method based on two-stage density clustering | |
| Sun et al. | Modelling Principles of Grey Matrix Incidence Analysis for Panel Data. | |
| Chen et al. | Artificial immune algorithm applied to short-term prediction for mobile object location | |
| CN103886604B (en) | Parallel image dividing method based on initial profile forecast model | |
| CN108229546A (en) | A kind of overlapping corporations detection method of feature based vector center peak value cluster |
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: 20161109 |