CN104991982B - A kind of flight vehicle aerodynamic elastic lag sensor placement method - Google Patents
A kind of flight vehicle aerodynamic elastic lag sensor placement method Download PDFInfo
- Publication number
- CN104991982B CN104991982B CN201510182075.0A CN201510182075A CN104991982B CN 104991982 B CN104991982 B CN 104991982B CN 201510182075 A CN201510182075 A CN 201510182075A CN 104991982 B CN104991982 B CN 104991982B
- Authority
- CN
- China
- Prior art keywords
- finite element
- dynamic response
- clustering
- structure dynamic
- data set
- 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
Landscapes
- Aerodynamic Tests, Hydrodynamic Tests, Wind Tunnels, And Water Tanks (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of flight vehicle aerodynamic elastic lag sensor placement methods, this method is on the basis of using finite element method Flight Vehicle Structure dynamic response, finite element model is modified by fluid structurecoupling method, carries out the wind-structure interaction of traversal all flight conditions and disturbance factor；Relatively accurate structure dynamic response is calculated, the response of each node in finite element model is obtained；Data set is standardized, elastic wing structure dynamic response data set is generated；Finally the clustering method based on distance metric in pattern-recognition is used to cluster data, optimization obtains final sensing station.The present invention turns to overall goal with sensor measurement information maximum, and balance is optimized between number of sensors and optimal location position, it can be ensured that the position of each sensor can collect valuable aeroelasticity characteristic value.
Description
Technical field：
The present invention relates to flight vehicle aerodynamic elastic lag sensor placement field, the dynamic response based on Flight Vehicle Structure
Analysis and dynamic cluster method.
Background technology：
Since the end of last century, NASA starts the research about flexible flier intelligence structure control technology, grinds
Study carefully plan to be intended to optimize layout configuration for Embedded sensor and actuator, to realize wing shapes, aircraft
The active control of construct noise and structural vibration.From large-scale spacecraft to minute vehicle, relevant achievement in research quilt
In aircraft applied to various scales, safety, reliability and environment that this technology can be used for improving aircraft adapt to
Property.The sensor placement technical research of optimization is a vital research contents of the research plan.
The initial stage of sensor placement technical research, domain experts often carry out cloth using artificial optimization's technology to sensor
Office, the layout into line sensor and actuator is carried out with abundant engineering experience and intuition.Sensing station location problem can retouch
State for：In given M location arrangements sensor of N number of possible installation position selection, to obtain optimal performance indicator.One
As for, the optimization for performance indicator is very complicated.In addition, the geometry that placement scheme also suffers from structure restricts,
Placement scheme also suffers from the restriction of the physical limit and control energy of sensor.When performance indicator cannot lean on experience intuition to obtain
, when objective constraints are most important, or when the position M numbers that may be installed are more than the quantity of desk checking, then more
Need a kind of optimization sensor placement technology of science.Sensor optimization placement scheme often rule of thumb tries mostly at present
It gathers, not yet retrieves general flight vehicle aerodynamic elastic lag sensor placement optimization method.
Invention content：
The purpose of the present invention is in view of the above shortcomings of the prior art, in conjunction with flight vehicle aerodynamic elasticity own characteristic, provide
A kind of dynamic clustering flight vehicle aerodynamic elastic lag sensor placement method, it is a kind of general purpose transducer layout optimization criterion,
Overall goal is turned to sensor measurement information maximum, balance is optimized between number of sensors and optimal location position,
It can ensure that the position of each sensor can collect valuable aeroelasticity characteristic value, and support domain expert people
The quantity of work specific arrangements sensor can effectively reduce control system cost.
The present invention is achieved by the following technical solutions, specifically comprises the following steps：
Step 1：On the basis of using finite element method Flight Vehicle Structure dynamic response, pass through fluid structurecoupling
Method is modified finite element model, carries out the wind-structure interaction of traversal all flight conditions and disturbance factor；
Step 2：Relatively accurate structure dynamic response is calculated, the response of each node in finite element model is obtained；
Step 3：Data set is standardized, elastic wing structure dynamic response data set is generated；
Step 4：Data are clustered using the clustering method based on distance metric in pattern-recognition, optimization obtains
Final sensing station.
The fluid structurecoupling method is fluid structurecoupling MPCCI methods.
The relatively accurate structure dynamic response is calculated using the method for direct integral or modal superposition.
Described to be standardized to data set, generating elastic wing structure dynamic response data set is specially：Instead
Multiple repeatedly transformation flow field and disturbing load frequency, reproduce the various state of flights being likely to occur in practical flight as much as possible,
The deflection angle being likely to occur including traversing all rudder faces traverses plug-in device each possible installation site, obtains each knot
Field output is carried out after the response of point, obtains structure dynamic response data set.
The clustering method based on distance metric using in pattern-recognition clusters data, and optimization obtains final
Sensing station is specially：(1) rational cluster granularity, inter- object distance and class are determined by way of debugging cluster threshold value repeatedly
Between distance；(2) it is reasonably clustered using based on the characteristic value means clustering algorithm for improving Euclidean distance measurement；(3)
The number of each node and its in finite element model in being clustered by finite element Knot Insertion fitting algorithm backwards calculation
Position, carry out redundancy resolution, obtain sensor final layout scheme.
The characteristic value means clustering algorithm for improving Euclidean distance measurement is by each point to mean value error square
It is used as measurement criterion with maximum, it is measured to the distance at the center of clustering to each example, it is grouped into when meeting specified threshold
The class of barycenter, through iterate while adjusting in final cluster have in minimum class away from, there is maximum kind spacing between clustering,
Reach iteration ends after constraints.
The finite element Knot Insertion fitting algorithm refers to being obtained into row interpolation to the result of calculation of modal superposition in finite element
To the aeroelasticity structure dynamic response of all nodes of smooth finite element model.
Conventional method is laid out to sensor using artificial optimization's technology, is come with abundant engineering experience and intuition
Into the layout of line sensor and actuator, when objective constraints are most important, the application of experience sensor placement just has one
Fixed limitation.
This method has the following advantages that：The maximum aeroelasticity response of each node is obtained using fluid structurecoupling method, then
Optimize to obtain final sensing station by the method for dynamic clustering；Repeated multiple times transformation flow field and disturbing load frequency, to the greatest extent may be used
The various state of flights being likely to occur in reproduction practical flight more than energy, including traverse the deflection angle that all rudder faces are likely to occur
Degree traverses plug-in device each possible installation site, carries out field output after obtaining the response of each node, obtains structural dynamic
Learn response data sets；Finite element model is modified by fluid structurecoupling MPCCI methods or other fluid and structural simulation methods,
Carry out the wind-structure interaction of traversal all flight conditions and disturbance factor.
Compared with prior art, the present invention has the advantages that：It can ensure the position of each sensor
Can collect valuable aeroelasticity characteristic value, and can artificial specific arrangements sensor quantity, reduce control system at
This.Finite element model is modified by fluid structurecoupling MPCCI methods or other fluid and structural simulation methods, can be carried out time
Go through all flight conditions and the wind-structure interaction of disturbance factor.It is laid out using the Flight Vehicle Structure kinetic sensors of the present invention
Method can approach optimal sensor placement by careful parameter regulation.It, can be as much as possible real again using this method
The various state of flights that border is in-flight likely to occur, including the deflection angle that all rudder faces are likely to occur is traversed, traversal is outer to be hung
Standby each possible installation site, carries out field output after obtaining the response of each node, finally obtains structure dynamic response number
According to collection.
Description of the drawings：
Fig. 1 is flight vehicle aerodynamic elastic lag sensor placement method schematic；
Fig. 2 is analysis 4 figure of granularity；
Fig. 3 is analysis 5 figure of granularity；
Fig. 4 is analysis 6 figure of granularity；
Fig. 5 is analysis 7 figure of granularity；
Fig. 6 is analysis 8 figure of granularity；
Fig. 7 is analysis 9 figure of granularity；
Fig. 8 is analysis 10 figure of granularity；
Fig. 9 is 4 to cluster spacing figure；
Figure 10 is 5 to cluster spacing figure；
Figure 11 is 6 to cluster spacing figure；
Figure 12 is 7 to cluster spacing figure；
Figure 13 is 8 to cluster spacing figure；
Figure 14 is 9 to cluster spacing figure；
Figure 15 is 10 to cluster spacing figure；
Figure 16 is 5 point sensor layouts；
Figure 17 is 6 point sensor layouts；
Figure 18 is 9 point sensor layouts；
Specific implementation mode：
The invention will be further described with reference to the accompanying drawings and examples.
Flexible flier housing construction has larger flexibility, and adequately to measure the flexural property of flexible wing, fitful wind is rung
Needs should be waited, each partial accession inertial sensor of full machine is coped with, multiple sensors form sensor array and provide complementary letter
Breath, so as to more effectively carry out active control to aeroelasticity inertia.
The present invention provides dynamic clustering flight vehicle aerodynamic elastic lag sensor placement method, and system principle diagram is shown in attached drawing
1, it can be ensured that placement position approaches optimal solution to greatest extent.Dynamic clustering is substantially a kind of iteration cluster, first gives one
It is coarse it is initial cluster, and define and the distance between cluster, then modification is iterated with distance metric principle, until the ratio that clusters
Relatively rationally until, the clustering method using this thought is called dynamic state clustering.On the basis of initial clustering, calculate first just
The object function for the cluster that begins, the individual for adjusting initial classes realize transposition to new class if object function reduces；Otherwise individual
It stays in former group.One wheel cycle terminates, then adjusts intermediate individual, until the reorganization transposition of any individual cannot make target
Function becomes smaller, and cluster terminates.Initial clustering determines that method has and by virtue of experience determines that initial clustering, the number that will cluster are randomly divided into N
Class artificially determines classification number M.
Embodiment one：The present embodiment includes the following steps：
1. on the basis of using finite element method Flight Vehicle Structure dynamic response, pass through fluid structurecoupling MPCCI
Or other fluid structurecoupling methods are modified finite element model, the stream for carrying out traversal all flight conditions and disturbance factor consolidates coupling
Close analysis.The analysis of flight vehicle aerodynamic elastic lag includes harmonic responding analysis, random vibration and instantaneous response analysis, and carries out phase
Harmony response load, PSD response, the transient response experiment answered.
2. calculating relatively accurate structure dynamic response, the response of each node in finite element model is obtained：Practical work
Because computing resource is limited in journey calculating, the response analysis data that progress fluid and structural simulation is obtained are often not comprehensive enough, still
It can generally reflect aeroelasticity response of the aircraft under typical disturb.It is folded that direct integral, mode can be used in derivation algorithm
The methods of add to calculate and obtain.
3. a pair data set is standardized, elastic wing structure dynamic response data set is generated：Repeated multiple times change
Change of current field and disturbing load frequency, it is as much as possible to reproduce the various state of flights being likely to occur in practical flight, including traversal
The deflection angle that all rudder faces are likely to occur traverses plug-in device each possible installation site, obtains the response of each node
Field output is carried out afterwards, obtains structure dynamic response data set
4. being clustered to data using the clustering method based on distance metric in pattern-recognition：Artificial specified cluster grain
Degree, inter- object distance and between class distance.After obtaining the characteristic value data collection of typical response, the number that rationally clusters is further determined that,
Accurate node can be obtained by specified rational analysis granularity to respond；Redundancy is cleared up by specified reasonable class spacing to cluster；It is logical
It crosses and rational inter- object distance is specified to approach optimal sensor placement position.Using based on the feature for improving Euclidean distance measurement
It is worth means clustering algorithm, after rationally being clustered, in being clustered by finite element Knot Insertion fitting algorithm backwards calculation
Each node numbering, and its position in finite element model again and carry out redundancy resolution, this process data amount is small, model
Intuitively, Flight Vehicle Structure dynamics field expertise, corresponding actual physical meaning can fully be used.
Improve Euclidean distance measurement characteristic value means clustering algorithm be by each point to mean value error quadratic sum most
Wonderful works measurement criterion measures it to the distance at the center of clustering to each example, it is grouped into barycenter when meeting specified threshold
Class, have in minimum class away from there is maximum kind spacing between clustering, reach through iterating while adjusting in final cluster
Iteration ends after constraints.
Finite element Knot Insertion fitting algorithm refers to being put down into row interpolation to the result of calculation of modal superposition in finite element
The aeroelasticity structure dynamic response of all nodes of sliding finite element model.
For example, export structure dynamic response calculates field data, wing finite element node is obtained after being standardized
1457, and each Y-axis shift value data set of node structure dynamic response.Using Dynamic Clustering Algorithm to data set into
Row clustering specifies clustering granularity from 4 to 10, and the specified interior distance metric threshold value that clusters is 0.9, between random generation clusters
Away from.Cluster result is as shown in attached drawing 2-8.
Distance is as shown in Fig. 9-15 between clustering under difference analysis granularity.
It can be obtained according to the above analysis, the method that the spacing that clusters is randomly assigned can meet algorithm requirements substantially, cluster
Between can be effectively isolated, can be taking human as specified cluster spacing if having the relevant parameter of elastic wing structure so that poly-
Class effect is more excellent.
Each node numbering in being clustered by finite element Knot Insertion fitting algorithm backwards calculation, obtains sensor most
Whole placement scheme selects representative configuration scheme as shown in figs. 16-18.
Claims (3)
1. a kind of flight vehicle aerodynamic elastic lag sensor placement method, it is characterised in that：Described method includes following steps：
Step 1：On the basis of using finite element method Flight Vehicle Structure dynamic response, pass through fluid structurecoupling method
Finite element model is modified, the wind-structure interaction of traversal all flight conditions and disturbance factor is carried out；
Step 2：Relatively accurate structure dynamic response is calculated, the response of each node in finite element model is obtained；
Step 3：Data set is standardized, elastic wing structure dynamic response data set is generated；
Step 4：Data are clustered using the clustering method based on distance metric in pattern-recognition, optimization obtains final
Sensing station；
The fluid structurecoupling method is fluid structurecoupling MPCCI methods；
The relatively accurate structure dynamic response is calculated using the method for direct integral or modal superposition；
Described to be standardized to data set, generating elastic wing structure dynamic response data set is specially：It is repeatedly more
Secondary transformation flow field and disturbing load frequency reproduce the various state of flights being likely to occur in practical flight as much as possible, including
The deflection angle that all rudder faces are likely to occur is traversed, plug-in device each possible installation site is traversed, obtains each node
Field output is carried out after response, obtains structure dynamic response data set；
The clustering method based on distance metric using in pattern-recognition clusters data, and optimization is finally sensed
Device position is specially：(1) rational cluster granularity, inter- object distance and class spacing are determined by way of debugging cluster threshold value repeatedly
From；(2) it is reasonably clustered using based on the characteristic value means clustering algorithm for improving Euclidean distance measurement；(3) pass through
Finite element Knot Insertion fitting algorithm the backwards calculation number of each node and its position in finite element model in being clustered
It sets, carries out redundancy resolution, obtain sensor final layout scheme.
2. flight vehicle aerodynamic elastic lag sensor placement method according to claim 1, it is characterized in that：The improvement Europe
In several distance metric characteristic value means clustering algorithm be using each point to mean value error quadratic sum maximum as measurement criterion,
It is measured to the distance at the center of clustering to each example, it is grouped into when meeting specified threshold the class of barycenter, through iterating
Adjusting simultaneously has in minimum class in finally clustering away from having maximum kind spacing between clustering, reach iteration after constraints
It terminates.
3. flight vehicle aerodynamic elastic lag sensor placement method according to claim 1, feature are being：It is described limited
First Knot Insertion fitting algorithm refers to obtaining smooth finite element model into row interpolation to the result of calculation of modal superposition in finite element
The aeroelasticity structure dynamic response of all nodes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510182075.0A CN104991982B (en) | 2015-04-16 | 2015-04-16 | A kind of flight vehicle aerodynamic elastic lag sensor placement method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510182075.0A CN104991982B (en) | 2015-04-16 | 2015-04-16 | A kind of flight vehicle aerodynamic elastic lag sensor placement method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104991982A CN104991982A (en) | 2015-10-21 |
| CN104991982B true CN104991982B (en) | 2018-09-25 |
Family
ID=54303797
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510182075.0A Active CN104991982B (en) | 2015-04-16 | 2015-04-16 | A kind of flight vehicle aerodynamic elastic lag sensor placement method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104991982B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106919584B (en) * | 2015-12-26 | 2020-07-07 | 华为技术有限公司 | Topological graph layout method and device |
| CN105843076A (en) * | 2016-03-31 | 2016-08-10 | 北京理工大学 | Flexible aircraft aeroelasticity modeling and controlling method |
| CN108092751A (en) * | 2016-11-22 | 2018-05-29 | 北京空间技术研制试验中心 | Pneumatic gauging parameter information processing method |
| CN106557633A (en) * | 2016-11-29 | 2017-04-05 | 上海卫星工程研究所 | Satellite sun wing sensor placement method is realized based on EI methods |
| CN107330133B (en) * | 2017-04-01 | 2019-03-29 | 中国商用飞机有限责任公司北京民用飞机技术研究中心 | A kind of optimizing layout method based on virtual test |
| CN109711036B (en) * | 2018-12-24 | 2023-05-23 | 中国航空工业集团公司西安飞机设计研究所 | Evaluation method of flight control system test result |
| CN109444350B (en) * | 2018-12-27 | 2021-09-24 | 中山大学 | Layout method of atmospheric pollutant monitoring sensor based on unmanned aerial vehicle |
| CN111595433B (en) * | 2019-02-20 | 2022-07-08 | 中国航发商用航空发动机有限责任公司 | Position determination method and system for vibration sensor of whole aircraft engine |
| CN110532607B (en) * | 2019-07-24 | 2021-06-22 | 北京航空航天大学 | Sensor layout method for identifying distributed load of hypersonic aircraft control surface structure |
| CN117169900B (en) * | 2023-11-03 | 2024-02-20 | 威科电子模块（深圳）有限公司 | Accurate sensing system, method, equipment and medium based on thick film circuit |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102866637A (en) * | 2012-10-07 | 2013-01-09 | 西北工业大学 | Quadratic order-reduction based method for simulating unsteady aerodynamic force of aerofoil with operation surface |
-
2015
- 2015-04-16 CN CN201510182075.0A patent/CN104991982B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102866637A (en) * | 2012-10-07 | 2013-01-09 | 西北工业大学 | Quadratic order-reduction based method for simulating unsteady aerodynamic force of aerofoil with operation surface |
Non-Patent Citations (4)
| Title |
|---|
| 基于聚类算法和自由度集结的柔性结构模型降阶研究;李成涛 等;《计算力学学报》;20120430;第29卷(第2期);摘要、第237-238页 * |
| 弹性飞行器的传感器位置设计;袁建平 等;《宇航学报》;19870131(第1期);第81-88页 * |
| 柔性翼微型飞行器流固耦合数值模拟;孟令兵 等;《南京航空航天大学》;20131030;第45卷(第5期);摘要、第624-626页、图5-9 * |
| 直升机中减速器谐响应分析与传感器优化布局;苏勋文 等;《北京航空航天大学学报》;20110930;第37卷(第9期);第1051-1053页、图4-7 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104991982A (en) | 2015-10-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104991982B (en) | A kind of flight vehicle aerodynamic elastic lag sensor placement method | |
| Liu et al. | Approaches to quadratic stability conditions and h/sub/spl infin//control designs for ts fuzzy systems | |
| CN102377180B (en) | Power system load modeling method based on electric energy quality monitoring system | |
| CN103401236A (en) | Wind power farm generator unit grouping method based on flow correlation of wind power farm | |
| CN103268366A (en) | Combined wind power prediction method suitable for distributed wind power plant | |
| CN108233400A (en) | A kind of more feed-in interaction factor computational methods of meter and hvdc control mode | |
| CN110165709A (en) | Consider the virtual synchronous machine grid-connected inverting system stability method for improving of sampling time delay | |
| CN109726437A (en) | A kind of hatch door aerodynamic loading equivalent nodal force processing method | |
| CN102427229B (en) | Zero-injection-constraint electric power system state estimation method based on modified Newton method | |
| CN102880804B (en) | A kind of aircraft rib dynamics defining method based on mode superposition method | |
| CN114065662B (en) | Method suitable for rapidly predicting airfoil flow field with variable grid topology | |
| CN103729570B (en) | The matching process of the power system oscillation pattern based on Matrix Perturbation | |
| CN110119593A (en) | A kind of visualization tide and tidal current forecast method based on FVCOM model | |
| CN103217896B (en) | Many FACTS based on free-form curve and surface method anti-time lag control method for coordinating | |
| CN107657116B (en) | Method for affine modeling of power curve of wind power plant | |
| CN104462608A (en) | Wireless sensor network data clustering method based on fuzzy C-mean clustering algorithm | |
| CN102709955A (en) | Multi-section-based power flow control method | |
| CN107977730A (en) | A kind of wind measurement method of multisensor Data Fusion technology | |
| CN112665820A (en) | R-type grid self-adaptive moving method and device based on variable difference and relative displacement | |
| CN114692455B (en) | Modeling method of composite material laminated plate based on segmented equidistant recombination curve | |
| CN106124933A (en) | A kind of power system failure diagnostic method based on input nonlinearities method of equal value | |
| CN105977968A (en) | Optimal configuration method for power quality monitor of annular multi-source power distribution network | |
| CN106873363B (en) | A kind of modeling method of aircraft angle of attack signal | |
| CN110137967A (en) | A kind of large-scale electrical power system trend convergence method of adjustment for key node | |
| Ivanell et al. | Validation of methods using EllipSys3D |
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 |