CN105117585B - A kind of adaptive mesh dynamic weighting error evaluation method of two-phase flow pump - Google Patents
A kind of adaptive mesh dynamic weighting error evaluation method of two-phase flow pump Download PDFInfo
- Publication number
- CN105117585B CN105117585B CN201510460747.XA CN201510460747A CN105117585B CN 105117585 B CN105117585 B CN 105117585B CN 201510460747 A CN201510460747 A CN 201510460747A CN 105117585 B CN105117585 B CN 105117585B
- Authority
- CN
- China
- Prior art keywords
- unit
- gradient
- phase flow
- error
- flow pump
- 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.)
- Active
Links
Abstract
The invention provides a kind of adaptive mesh dynamic weighting error evaluation method of two-phase flow pump, comprise the following steps：(1) generation model, initial mesh divides and carries out CFD numerical computations, (2) calculated velocity gradient value and pressure gradient, and (3) calculate the Grad and gradient difference value of each unit, and normalized, (4) error evaluation；Error evaluation method of the present invention, the velocity gradient of unit and barometric gradient is combined together elemental error is judged, so as to solve the deficiency that existing evaluation method can not consider speed and barometric gradient simultaneously；It is not to give a fixed value in the assigned error threshold value present invention, but after normalized, successively decrease error threshold every time, so as to accelerate the ciphering process of adaptive mesh；According to the present invention program, it is possible to achieve the carry out error evaluation to all kinds grid, and assess efficiency high.
Description
Technical field
The present invention relates to two-phase flow pump adaptive mesh to generate field, more particularly to a kind of adaptive error assessment side
Method.
Background technology
Adaptive polo placement technology is the feedback method with certain optimal property, and its basic thought is：With convenient value
Based on calculating, adaptive mesh is generated as core, and error distribution, the automatic larger area of bridging error are obtained according to result of calculation
Domain, and then optimize grid until error meets to require, its circulation process is as shown in Figure 1.Adaptive mesh generation includes error and commented
Estimate, Adaptive refinement and grid optimization.Error evaluation is that effective estimation is made to mesh error, and mark error compared with
Big region；Adaptive refinement is then encrypted according to the information of mark；It is difficult to ensure that grid cell is several after adaptive refinement
What characteristic, easily occur causing the distortion unit for calculating and dissipating (volume is zero or is negative unit), it is necessary to by grid optimization change
The geometry of kind grid cell.
Error evaluation is the core and key of adaptive mesh generation, directly affects the efficiency of adaptive polo placement and reliable
Property.Existing method is mostly using the rate of change of adjoint variable gradient between consecutive points come evaluated error, it is believed that graded is big
The influence of zone convection field computation precision is also big.Research shows corresponding adaptive for the violent place progress of barometric gradient change
Should encrypt can preferably catch wake vortices.But for the rotating machinery of this kind of complex geometry of two-phase flow pump, office
Portion's refinement internal flow unstable region (such as stall, flow separation), barometric gradient can not weigh error well, that is, press
Power graded is larger not to represent that Flow Field Calculation error is also big, is so likely to cause and goes out in some regions that need not be encrypted
Now cross the situation of encryption.Therefore, reasonably select flow field adjoint variable effectively to avoid the grid caused by false judgment from crossing to add
It is close.For some flow field problems, the selection of adjoint variable is very clear and definite, for example, being one more excellent in impingement flow areal pressure gradient
Standard, and for shearing flow problem, the Evaluated effect of velocity gradient is better than barometric gradient.Based on that can not characterize, more flowings are existing
The single argument of elephant is assessed, thus easily causes assessment failure.
In addition, adaptive mesh generation needs to determine to need the region refined by relative error estimate and threshold value
(being referred to as error unit mark).Fixed threshold mark mode is a kind of error unit labelling strategies more conventional at present, the plan
Keep threshold value to immobilize slightly during whole adaptive polo placement, stop meter when all elemental errors are both less than the value
Calculate.But after using fixed threshold mark mode to carry out adaptive polo placement, still there is a small amount of error more than the unit of given threshold value, be
Meet that error requirements need successive ignition to calculate, therefore efficiency of algorithm is relatively low.
The content of the invention
For Shortcomings in the prior art, the invention provides a kind of adaptive mesh dynamic weighting of two-phase flow pump mistake
Poor appraisal procedure, using the mark mode for being gradually reduced threshold value, i.e. encryption every time, the quantity of computing unit is with the reduction of threshold value
And reduce, the computations repeatedly just for some regions are so effectively prevent, thus reduce the number of iteration, Neng Gouzheng
Really and error larger unit completely is evaluated, so as to instruct being smoothed out for encrypted work.
The present invention is to realize above-mentioned technical purpose by following technological means.
A kind of adaptive mesh dynamic weighting error evaluation method of two-phase flow pump, comprises the following steps：
(1) by carrying out external characteristics experiment to two-phase flow pump, the lift and efficiency data of two-phase flow pump, generation model are obtained
Pump threedimensional model simultaneously carries out initial mesh division；
(2) CFD numerical computations are carried out；
(3) the velocity gradient value for the unit i that the CFD numerical results according to step (2) obtainWith
Pressure gradientThe Grad of the unit i isWherein λ1For the power of velocity gradient
Weight coefficient, λ2For the weight coefficient of barometric gradient, λ1+λ2=1；
(4) gradient difference value of the unit i isPlace is normalized in unit i gradient difference value
Reason, obtains normalized gradient differenceWhereinFor the maximum of all unit gradient difference values；
(5) to describedUnit more than given error threshold is demarcated and encrypted, the given mistake
Poor threshold value is between 0~1；
(6) repeat step (2)~(5), and the error threshold that step (5) gives every time is gradually reduced, until what is obtained raises
Untill journey and efficiency meet required precision compared with the lift described in step (1) and efficiency.
Further, step (3) λ1>λ2。
For two-phase flow pump, the error that flow field is characterized compared to barometric gradient velocity gradient is more effective (especially not
Steady flow region), therefore of the invention when establishing speed and barometric gradient error evaluation function, the weight coefficient of velocity gradient
More than barometric gradient.
In such scheme, the barometric gradient of the unit i is：
WhereinThe barometric gradients of respectively described unit i in the X, Y, Z direction.
In such scheme, the velocity gradient of the unit i is：
WhereinThe velocity gradients of respectively described unit i in the X, Y, Z direction.
Unit i velocity gradient and barometric gradient computational methods is in addition to the method provided in the present invention, in prior art
In also have other computational methods, do not list one by one herein.
Beneficial effects of the present invention：
(1) the drawbacks of using unitary variant assessment strategy for existing appraisal procedure, the present invention use speed and pressure
The mode that gradient is combined is assessed, therefore can preferably react the cell position being had a great influence to computational accuracy, so as to solve
Existing evaluation method of having determined can not consider the deficiency of speed and barometric gradient simultaneously.
(2) in order to solve different example situations can not accurate assigned error threshold value situation, the present invention using calculate it is each
Normalized mode is used during the gradient difference value of individual unit so that all unit gradient difference values are all between 0~1, so as to add
The fast ciphering process of adaptive mesh.
(3) unit is obtained more than set-point to gradient difference value by way of dynamic alignment error threshold value in addition to be demarcated simultaneously
Encryption, i.e., give a larger error threshold (close to maximum 1), the error threshold in subsequent iterative process in first time
Be gradually reduced until the adaptive polo placement of two-phase flow pump lift and efficiencies with experiment compared with meet given accuracy untill, enter
One step improves the efficiency and validity of the process of adaptive refinement, and efficiently avoid different examples can not accurate assigned error threshold
The problem of value.
(4) according to the present invention program, it is possible to achieve the carry out error evaluation to all kinds grid, and assess efficiency high.
Brief description of the drawings
Fig. 1 is adaptive mesh generating process flow chart of the present invention.
Embodiment
Below in conjunction with the accompanying drawings and specific embodiment the present invention is further illustrated, but protection scope of the present invention is simultaneously
Not limited to this.
Velocity gradient and barometric gradient of the present invention be in unit center of gravity (unit can be that tetrahedron can also
It is hexahedron).
If Fig. 1 is adaptive mesh generating process flow chart of the present invention.
Step 1, builds two-phase flow pump external characteristics testing stand, and the lift of pump has inlet and outlet pressure table measurement to obtain；Using electricity
The power of survey method measuring pump, the optimal operating condition point flow that external characteristics result obtains are 34.48m3/ h lifts are 4.76, and efficiency is
57.3%, three-dimensional modeling is carried out using Pro/E according to the hydraulic model of two-phase flow pump, and using ICEM generation initial mesh.
Step 2, calculating use CFX, standard k-ε turbulence model, SIMPLE algorithms, and non-coupled implicit aspect is solved.
Inlet boundary condition：Using speed import, assume to determine import axial velocity by mass conservation law and irrotationality.
Export boundary condition：Free discharge, it is assumed that flowing fully develops at outlet border, and exit region is apart from recirculating zone
Farther out.
Wall condition：Solid wall surface uses side wall non-slip condition.
Lift：
Efficiency：
In formula：PoutRepresent the pressure of pump discharge, PinPump inlet pressure is represented, ρ represents the density of liquid in pump, and Q represents pump
Flow, H represent pump lift, g is acceleration of gravity；M ' is the power of front side of vane, the back side and front and rear cover plate inner and outer surfaces
Square sum；ω represents the angular speed of pump；η ' is to include pump calculation domain forecasting efficiency value after volumetric loss, disc friction losses；
The loss of bearing and sealing takes 3%；Therefore efficiency eta=η ' × (1-3%) of pump is predicted.
Step 3, the velocity gradient value of unit is obtained according to result of calculationAnd pressure gradientAnd according to formulaCalculate the value of unit, wherein λ1For the weight system of velocity gradient
Number, λ2For the weight coefficient of barometric gradient, λ1+λ2=1, i be unit label, λ1Recommendation is 0.6, λ2Recommendation is 0.4；
1st, unit i pressure gradientComputational methods are as follows：
(1) unit i barometric gradient in the X direction：
Pi+1For the pressure of label i+1 units, Pi-1For the pressure of label i-1 units, Δ xI+1, iFor unit i+1 and unit i
In the distance (distance between unit center of gravity in X-direction) of X-direction, Δ xI, i-1For unit i and unit i-1 X-direction distance
(distance between unit center of gravity in X-direction)；
Barometric gradients of the unit i in Y, Z-direction and the barometric gradient computational methods of the unit i in the X direction
Unanimously；
(2) unit i pressure gradient is：
WhereinThe barometric gradients of respectively described unit i in the X, Y, Z direction.
2nd, unit i velocity gradient valueComputational methods are as follows：
(1) present invention defines X, and Y, the speed of Z-direction is respectively u, v, w, then unit i velocity gradient calculates in X-direction
Method is as follows：
Wherein ui+1For the speed of label i+1 unit X-directions, ui-1For the speed of label i-1 unit X-directions, Δ xi+1,iFor
Unit i+1 and unit i is in the distance (distance between unit center of gravity in X-direction) of X-direction, Δ xi,i-1For unit i and unit i-1
In the distance (distance between unit center of gravity in X-direction) of X-direction.
(2) the velocity gradient size of X-direction is：
It is consistent with X-direction computational methods for the computational methods in other directions (such as Y, Z)；
(3) it is in unit i velocity gradient size：
WhereinThe velocity gradients of respectively described unit i in the X, Y, Z direction.
3rd, on border unit barometric gradient computational methods：
Unit (such as i=0) on left margin, unit i barometric gradient computational methods are as follows in the X direction：
It is consistent with X-direction computational methods for the computational methods in other directions, therefore, in unit 0
Barometric gradient size be：
Unit (such as i=i on right marginmax, imaxRepresent largest unit number), unit i in the X directionmaxPressure ladder
It is as follows to spend computational methods,Computational methods and X-direction computational methods one for other directions
Cause, therefore, in unit imaxPoint barometric gradient size be：
4th, on border unit velocity gradient computational methods：
Unit (such as i=0) on left margin, the velocity gradient computational methods of unit 0 are as follows in the X direction：
The velocity gradient size of X-direction is：
It is consistent with X-direction computational methods for the computational methods in other directions (such as Y, Z), therefore, in the speed ladder of unit 0
Spending size is：
Unit (such as i=i on right marginmax, imaxRepresent largest unit number), unit i in the X directionmaxSpeed ladder
It is as follows to spend computational methods：
The velocity gradient size of X-direction is：
It is consistent with X-direction computational methods for the computational methods in other directions (such as Y, Z), therefore, in unit imaxSpeed
Gradient magnitude is：
Step 4：According to formulaThe gradient difference value of unit is calculated, and to the gradient of unit
Difference is normalized, i.e.,WhereinFor the maximum of all unit gradient difference values；
Step 5：The unit more than the value is demarcated according to given error threshold, if giving one every time
Fixed value can increase the time of whole adaptive process, therefore be pushed away in first time adaptive process to a larger error threshold
Recommend using 0.95, that is, demarcate all unitsValue be more than 0.95, and in subsequent calculating iterative process not
It is disconnected to reduce the value；
Step 6：The unit of all demarcation is encrypted, i.e., the midpoint on each side of demarcation unit adds an encryption
Node, then continue to carry out CFD calculating on this basis；
Step 7：Repeat step two arrives step 6, and the error threshold given in step 5 be gradually reduced (as the second time to
Fixed 0.9, third time given 0.88, by that analogy), until the CFD obtained lifts calculated and efficiency are less than 2% with test error
Untill (recommend calculation error be 2%), adaptive iteration process can be terminated.
The present embodiment is 860,000 in initial mesh, and the grid number obtained after adaptive refinement is the adaptometer after 1,830,000
Result and the external characteristics Comparative result of experiment are calculated, lift error is 1.2%, and efficiency error is 1.6%, and it is 6 hours to calculate the time
26 minutes；And if initial mesh is 1,850,000, the result of calculating is that lift error is 4.2%, and efficiency error is 5.1%, and 4 is small
When 14 minutes；And grid number reaches 3,980,000, can just obtain with adaptive mesh identical computational accuracy, and calculate the time be 16
43 minutes hours, it is seen that the method that this patent proposes can preferably improve computational efficiency and computational accuracy.
The embodiment is preferred embodiment of the invention, but the present invention is not limited to above-mentioned embodiment, not
Away from the present invention substantive content in the case of, those skilled in the art can make it is any it is conspicuously improved, replace
Or modification belongs to protection scope of the present invention.
Claims (4)
1. a kind of adaptive mesh dynamic weighting error evaluation method of two-phase flow pump, it is characterised in that comprise the following steps：
(1) by carrying out external characteristics experiment to two-phase flow pump, the lift and efficiency data of two-phase flow pump, generation model pump three are obtained
Dimension module simultaneously carries out initial mesh division；
(2) CFD numerical computations are carried out；
(3) the velocity gradient value for the unit i that the CFD numerical results according to step (2) obtainWith pressure ladder
Angle valueThe Grad of the unit i isWherein λ1For the weight coefficient of velocity gradient,
λ2For the weight coefficient of barometric gradient, λ1+λ2=1；
(4) gradient difference value of the unit i isUnit i gradient difference value is normalized, obtained
Normalized gradient differenceWhereinFor the maximum of all unit gradient difference values；
(5) to describedUnit more than given error threshold is demarcated and encrypted, the given error
Threshold value is between 0~1；
(6) repeat step (2)~(5), and given error threshold is gradually reduced step (5) every time, until obtained lift and
Untill efficiency meets required precision compared with the lift described in step (1) and efficiency.
2. the adaptive mesh dynamic weighting error evaluation method of two-phase flow pump as claimed in claim 1, it is characterised in that step
Suddenly (3) λ1>λ2。
3. the adaptive mesh dynamic weighting error evaluation method of two-phase flow pump as claimed in claim 1 or 2, its feature exist
In the pressure gradient of the unit i is：
WhereinThe barometric gradients of respectively described unit i in the X, Y, Z direction.
4. the adaptive mesh dynamic weighting error evaluation method of two-phase flow pump as claimed in claim 1 or 2, its feature exist
In the velocity gradient value of the unit i is：
WhereinThe velocity gradients of respectively described unit i in the X, Y, Z direction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510460747.XA CN105117585B (en) | 2015-07-30 | 2015-07-30 | A kind of adaptive mesh dynamic weighting error evaluation method of two-phase flow pump |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510460747.XA CN105117585B (en) | 2015-07-30 | 2015-07-30 | A kind of adaptive mesh dynamic weighting error evaluation method of two-phase flow pump |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105117585A CN105117585A (en) | 2015-12-02 |
| CN105117585B true CN105117585B (en) | 2017-12-22 |
Family
ID=54665573
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510460747.XA Active CN105117585B (en) | 2015-07-30 | 2015-07-30 | A kind of adaptive mesh dynamic weighting error evaluation method of two-phase flow pump |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105117585B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110851929B (en) * | 2019-11-15 | 2022-12-27 | 中国科学院工程热物理研究所 | Two-dimensional leaf-type optimization design method and device based on self-adaptive grid |
| CN112665820B (en) * | 2021-03-15 | 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 | R-type grid self-adaptive moving method and device based on variable difference and relative displacement |
| CN113283186A (en) * | 2021-04-21 | 2021-08-20 | 南京航空航天大学 | Universal grid self-adaption method for CFD |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102043887A (en) * | 2010-12-31 | 2011-05-04 | 中国航空工业集团公司第六三一研究所 | Grid self-adaptive method based on error estimation |
| CN102708423A (en) * | 2012-06-06 | 2012-10-03 | 上海第二工业大学 | Method for solving path planning of mobile robot by introducing partial differential equation |
| CN102750414A (en) * | 2012-06-25 | 2012-10-24 | 江苏大学 | Method for testing relationship between mesh quality of centrifugal pump and computational accuracy |
-
2015
- 2015-07-30 CN CN201510460747.XA patent/CN105117585B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102043887A (en) * | 2010-12-31 | 2011-05-04 | 中国航空工业集团公司第六三一研究所 | Grid self-adaptive method based on error estimation |
| CN102708423A (en) * | 2012-06-06 | 2012-10-03 | 上海第二工业大学 | Method for solving path planning of mobile robot by introducing partial differential equation |
| CN102750414A (en) * | 2012-06-25 | 2012-10-24 | 江苏大学 | Method for testing relationship between mesh quality of centrifugal pump and computational accuracy |
Non-Patent Citations (5)
| Title |
|---|
| "Adaptive Mesh Refinement for conservative systems: multi-dimensional efficiency evaluation";R Keppens etal;《Computer Physics Communications》;20040331;第153卷(第3期);第317-339页 * |
| "Simplification of Tetrahedral Meshes with Accurate Error Evaluation";P Cignoni etal;《IEEE Visualization Conference》;20000322;第5卷(第3期);第85-92页 * |
| "基于结点应力误差估计的自适应网格划分";李佳彬 等;《计算力学学报》;20081215;第25卷(第6期);第753-757页 * |
| "自适应有限元在注塑成型数值模拟中的应用";赵建;《中国优秀硕士论文全文数据库 工程科技Ⅰ辑》;20061115(第11期);第B016-79页 * |
| "饱和砂土动静力特性的网格自适应数值模拟";邵琪 等;《岩土力学》;20140710;第35卷(第7期);第2079-2087页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105117585A (en) | 2015-12-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Politis et al. | Modeling wake effects in large wind farms in complex terrain: the problem, the methods and the issues | |
| CN104298869B (en) | A kind of fluid structurecoupling Numerical prediction method of elastic hydrofoil | |
| Tafti | Evaluating the role of subgrid stress modeling in a ribbed duct for the internal cooling of turbine blades | |
| Derakhshan et al. | The comparison of incomplete sensitivities and Genetic algorithms applications in 3D radial turbomachinery blade optimization | |
| CN103366067B (en) | The method of simulation and prediction fluid torque-converter performance | |
| Huang et al. | Well-balanced finite volume scheme for shallow water flooding and drying over arbitrary topography | |
| Wang et al. | Multiobjective optimization design and performance prediction of centrifugal pump based on orthogonal test | |
| Cea et al. | Numerical modelling of tidal flows in complex estuaries including turbulence: An unstructured finite volume solver and experimental validation | |
| CN105117585B (en) | A kind of adaptive mesh dynamic weighting error evaluation method of two-phase flow pump | |
| Abdel-Wahab et al. | Large eddy simulation of flow and heat transfer in a 90 deg ribbed duct with rotation: Effect of Coriolis and centrifugal buoyancy forces | |
| Scillitoe et al. | Evaluation of RANS and ZDES methods for the prediction of three-dimensional separation in axial flow compressors | |
| Zhou et al. | Influence of inflow conditions on turbine loading and wake structures predicted by large eddy simulations using exact geometry | |
| Zhan et al. | Numerical investigation of flow patterns in different pump intake systems | |
| Marsh et al. | Three dimensional numerical simulations of a straight-bladed vertical axis tidal turbine | |
| CN106320256A (en) | Method for determining flux of sediment passing through turbine of pumped storage power station for heavily silt-carrying rivers | |
| Krog Kruse et al. | A two‐dimensional quantitative parametric investigation of simplified surface imperfections on the aerodynamic characteristics of a NACA 63 3‐418 airfoil | |
| Mao et al. | A brief review of numerical solving methods for internal fluid of pumped storage unit | |
| CN102141464A (en) | Construction method for establishing turbulence model with Reynolds time-mean method | |
| Rengma et al. | A novel method of optimizing the Savonius hydrokinetic turbine blades using Bezier curve | |
| CN102750414B (en) | A kind of method verifying centrifugal pump mesh quality and computational accuracy relation | |
| Pullan et al. | The influence of sweep on axial flow turbine aerodynamics in the endwall region | |
| Barnes et al. | Validation and comparison of turbulence models for predicting wakes of vertical axis wind turbines | |
| Grapsas et al. | Numerical study and optimal blade design of a centrifugal pump by evolutionary algorithms | |
| CN107491623A (en) | The determination method of tack orifice plate cavitation inception number | |
| ENOMOTO et al. | Numerical simulation of turbulence flow in a Bulb turbine |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20211209 Address after: Room 503, building 3, No. 133, development avenue, Tongxiang Economic Development Zone, Tongxiang City, Jiaxing City, Zhejiang Province Patentee after: ZHEJIANG MAIZHI NETWORK TECHNOLOGY CO.,LTD. Address before: Zhenjiang City, Jiangsu Province, 212013 Jingkou District Road No. 301 Patentee before: JIANGSU University |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20221220 Address after: No. 545, Xinxing Community, Puyuan Town, Tongxiang City, Jiaxing City, Zhejiang Province, 314500 Patentee after: Jiaxing Yanzhi Network Technology Co.,Ltd. Address before: Room 503, building 3, No. 133, development avenue, Tongxiang Economic Development Zone, Tongxiang City, Jiaxing City, Zhejiang Province Patentee before: ZHEJIANG MAIZHI NETWORK TECHNOLOGY CO.,LTD. |