CN110110619A - A kind of satellite micro-vibration source quantitative identification method based on sparse blind source separating - Google Patents
A kind of satellite micro-vibration source quantitative identification method based on sparse blind source separating Download PDFInfo
- Publication number
- CN110110619A CN110110619A CN201910323968.0A CN201910323968A CN110110619A CN 110110619 A CN110110619 A CN 110110619A CN 201910323968 A CN201910323968 A CN 201910323968A CN 110110619 A CN110110619 A CN 110110619A
- Authority
- CN
- China
- Prior art keywords
- signal
- source
- vibration
- vibration source
- satellite
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H17/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves, not provided for in the preceding groups
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
- G06F2218/10—Feature extraction by analysing the shape of a waveform, e.g. extracting parameters relating to peaks
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention discloses a kind of satellite micro-vibration source quantitative identification methods of sparse blind source separating, first, at satellite cargo tank structure model sensitive load and model surface different location arranges acceleration transducer, acquire vibration signal when each vibration source works normally, it is ensured that observation signal number is greater than the number in source；Then, L is utilized1Reference L of the norm construction with reference to sparse blind deconvolution algorithm1Norm objective function is constructed reference signal according to the time and frequency domain characteristics prior information of vibration source, the optimal solution of separation signal is found using gradient descent method iteration optimization objective function, realizes the extraction of single signal of vibrating.Finally, calculating single source response signal of each vibration source at sensitive load using frequency domain list source response signal method for solving；Each vibration source contribution amount at sensitive load is calculated using the contribution amount characterizing method based on vector projection.Contribution amount evaluation index is the index of satellite micro-vibration source quantitative judge, can inhibit to provide foundation for micro-vibration.
Description
Technical field
The present invention relates to mechanical equipment vibration source quantitative identification methods, and in particular to a kind of defending based on sparse blind source separating
Star micro-vibration source quantitative identification method.
Background technique
Satellite plays a significant role in the construction of national military and national defense and the national economic development as crucial Space Equipment.
High-resolution satellite receives countries in the world extensive concern as satellite future thrust.However satellite micro-vibration is seriously made
The about promotion of the performances such as satellite resolution ratio, therefore carry out satellite micro-vibration source quantitative judge, it assesses main vibration source and sensitivity is carried
The contribution amount of He Chu can inhibit work to provide basis and foundation, have significant engineering application value for satellite micro-vibration
Since structure is complicated for satellite system, signal of vibrating is difficult to directly measure in the process in orbit, even vibration source is attached
Close measured signal is also and closes on the mixed signal of vibration source, and impure source signal.It is distributed in the micro- of satellite different location
Vibration source is coupled to sensitive load area with different paths and transmitting, therefore even if each vibration source vibration level is identical, sensitivity is carried
The contribution of He Qu also can be different.And structure is complicated for real satellite, main micro- vibration source is the rotation sections such as control-moment gyro and flywheel
Part, causes signal of vibrating harmonic components more and frequency band is overlapping strong between each signal of vibrating, so that the identification of micro-vibration source signal is more
Difficulty, also cause each vibration source branch operate when each response signal of observation point the sum of energy, when being operated simultaneously with each vibration source
The energy of observation point hybrid response signal is unequal, so that being difficult to accurately reflect vibration source at observation with the contribution amount of energy characterization
True contribution.
Summary of the invention
The satellite micro-vibration source quantitative identification method based on sparse blind source separating that the purpose of the present invention is to provide a kind of, with
The shortcomings that overcoming the prior art, the method for the present invention is high-efficient, at low cost, accuracy is high, can inhibit to provide base for satellite micro-vibration
Plinth and foundation can satisfy practical engineering application demand.
In order to achieve the above objectives, the present invention adopts the following technical scheme:
A kind of satellite micro-vibration source quantitative identification method based on sparse blind source separating, comprising the following steps:
(1) acquisition of vibration signal
(camera simulating piece) and model surface different location arrange acceleration at satellite cargo tank structure model sensitive load
Sensor acquires vibration signal when each vibration source works normally, i.e. observation signal:
X (t)=[x1(t),x2(t),...,xM(t),]T
Wherein, M is observation signal number, xmIt (t) is t moment in the collected vibration response signal of m-th of observation point, m
=1~M；
Ensure that observation signal number is greater than the number N of vibration source；
(2) extraction based on single signal of vibrating with reference to sparse blind deconvolution algorithm
Utilize L1Reference L of the norm construction with reference to sparse blind deconvolution algorithm1Norm objective function, according to vibration source when
Frequency domain character prior information constructs reference signal, finds the optimal of separation signal using gradient descent method iteration optimization objective function
Solution realizes the extraction of single signal of vibrating；
(3) each vibration source is calculated to contribution amount at sensitive load using frequency-domain sparse contribution amount estimation method
Single source response signal of each vibration source at sensitive load is calculated using frequency domain list source response signal method for solving；Benefit
Each vibration source contribution amount at sensitive load is calculated with the contribution amount characterizing method based on vector projection.
Further, step (2) specifically includes:
(2.1) L is utilized1The reference L of norm construction blind deconvolution algorithm1Norm objective function J (y, r) are as follows:
J (y, r)=| | F (y) | |1+λE{(y-r)2}
Wherein, y is separation signal；R is reference signal；F () is the discrete Fourier transform of variable；λ is scale factor；
E { } is the mathematic expectaion of variable；
(2.2) reference signal r is constructed according to the prior information in vibration equipment source；
(2.3) separation filter length L is set, x (t) obtained in step (1) is rewritten as time lag form
Then blind solution deconvolution model conversation is the ICA model of standard:
Wherein,For the separation filter of 1 × NL dimension；
At this point, the L about y1Norm objective function be converted into aboutObjective function:
Above formula can be expressed by Fourier transformation theory are as follows:
Wherein, Re isThe matrix that real part after being fourier transformed is constituted；Im isImaginary part after being fourier transformed
The matrix of composition；
(2.4) L is referred to using gradient descent method iteration optimization1Norm objective function, by separating vectorIteration is more
The new optimal solution for finding separation signal y corresponding when objective function minimum,Iteration it is updated be known as:
Wherein, k represents kth time iteration and updates, and g's is that objective function J (y, r) is rightInclined gradient, α is the iteration of algorithm
Step-length；
Each iteration obtains newAfterwards, it needs that it is normalized:
Iteration is until convergence obtains separating vectorOptimal solution, then separating signal isRealize single vibration source
The extraction of signal.
The extraction of not homologous signal is realized by constructing different reference signals.
Further, step (3) specifically:
(3.1) y (t) is set as the separation signal that extracts, can be obtained by blind deconvolution theory, and it is full that there are a filter h (τ)
The following relational expression of foot:
X (t)=h (τ) * y (t)
Wherein, * is convolution algorithm；
Based on convolution theory, convolution algorithm form is expressed as the product form in frequency domain, each available frequency point
Matrix between the frequency domain value X (ω) of observation signal at ω, filter frequency domain value H (ω) and separation signal frequency domain value Y (ω) closes
System:
X (ω)=H (ω) Y (ω)
A hybrid matrix H (ω) can be obtained at each frequency point in a frequency domain, since signal is sparse in frequency domain
Signal observation signal value or is zero or close to zero or is exactly to be generated by some signal of vibrating at frequency point ω
Response calculates in each source signal ingredient at frequencies omega so largely value is 0 in matrix H (ω) each frequency point ω at
Energy accounts for the specific gravity of the source signal gross energy, and the maximum source signal of specific gravity is exactly the generation source of observation signal at frequency point ω.
Assuming that frequency point ω0Locate observation signal and comes from Yn, then amplitude is 0 to remaining observation signal at this point, in frequency point ω0Place
It can obtain:
Above formula is solved up to frequency point ω0Hybrid matrix H (the ω at place0)；
Hybrid matrix H (ω) at each frequency point is solved using identical method, to can be in frequency domain after hybrid matrix
Solve single source response of each vibration source at camera simulating piece:
Wherein,For response signal of n-th of source at m-th of observation；Hmn(ω) is in hybrid matrix H (ω)
The n-th column element of m row；
(3.2) by single source response signalThe observation signal X at sensitive loadmOn projection ZmnAs n-th of vibration source
The contribution that m-th is observed, ZmnIn XmMiddle proportion is contribution amount evaluation index cmn:
The size of contribution amount evaluation index reflects each vibration source to the size that vibration is contributed and influenced at sensitive load, is
The index of satellite micro-vibration source quantitative judge.
Compared with prior art, the invention has the following beneficial technical effects:
The present invention uses the extraction that signal of vibrating is realized with reference to sparse blind deconvolution algorithm, is estimated by frequency-domain sparse contribution amount
Meter method calculates each vibration source to contribution amount at sensitive load, realizes the quantitative judge in satellite micro-vibration source, is micro-vibration
Inhibit to provide foundation.Its advantage is that in terms of the extraction of signal of vibrating, the L of selected metric signal sparsity1Norm is as target
Function accuracy when seeking comprising the overlapping source signal of harmonic wave, frequency band is higher, and the introducing of reference signal further improves
Precision；In terms of single source responds solution, is solved using frequency domain list source response signal extracting method, just knowing that signal spectrum
When be also able to achieve effective calculating；In terms of contribution amount calculating, the contribution amount characterizing method based on vector projection can be more realistically anti-
Reflect contribution amount of each vibration source at sensitive load.To sum up, the present invention is adopted using at satellite cargo tank structure sensitive load and surface
Each vibration source of vibration signal quantitative judge of collection has that simple, efficient, easy, accuracy is high to the contribution amount of sensitive load point
The features such as, work can be inhibited to provide basis and foundation for satellite micro-vibration, there is important engineering practical value.
Detailed description of the invention
Fig. 1 is that the present invention is based on the satellite micro-vibration source quantitative judge flow charts of sparse blind source separating；
Fig. 2 is equipment operational shock experimental provision；Wherein, 1, acceleration transducer；2, data collection system；3, it calculates
Machine；4, governor；5, satellite cargo tank structure model；6, the first vibration excitor；7, the second vibration excitor；8, the first power amplifier；9,
Second power amplifier；10, signal generator.
Fig. 3 is the time-domain diagram of the direction the x vibration response signal acquired by Fig. 2 device；Wherein, (a) is camera simulating piece top
The time domain waveform of portion's response signal；It (b) is the time domain waveform of camera simulating piece root response signal；It (c) is certain on cabinet
The time domain waveform of the collected response signal of one sensor；Abscissa indicates time, unit s in figure；Ordinate indicates vibration
Dynamic amplitude, unit mg.
Fig. 4 is the spectrogram of Fig. 3 vibration response signal；Abscissa indicates frequency, unit Hz in figure；Ordinate indicates width
Value, unit mg.
Fig. 5 is to utilize the time domain waveform with reference to the isolated source signal of sparse blind deconvolution algorithm；Wherein, (a) and
(b) single signal of vibrating of two vibration excitors is respectively corresponded；(c) single signal of vibrating of corresponding motor；Abscissa indicates the time in figure,
Unit is s；Ordinate indicates vibration amplitude, unit mg.
Fig. 6 is the spectrogram of Fig. 5 source signal；Abscissa indicates frequency, unit Hz in figure；Ordinate indicates amplitude, single
Position is mg.
Fig. 7 is each vibration source for being acquired by frequency domain list source response signal method for solving response signal at sensitive load point
Spectrogram；Wherein, (a) and (b) respectively corresponds single source response signal frequency spectrum of two vibration excitors；(c) single source of corresponding motor
Response signal frequency spectrum；Abscissa indicates frequency, unit Hz in figure；Ordinate indicates amplitude, unit mg.
Specific embodiment
The following describes the present invention in detail with reference to the accompanying drawings and specific embodiments:
It is satellite micro-vibration source quantitative judge flow chart shown in referring to Fig.1, each vibration source of satellite cargo tank structure model is normal
Work acquires the vibration signal at sensitive load and at model surface different location, it is ensured that observation signal number is greater than source signal
Number；Utilize L1Reference L of the norm construction with reference to sparse blind deconvolution algorithm1Norm objective function, according to the priori of vibration source
Information structuring reference signal refers to L using gradient descent method iteration optimization1Norm objective function, when finding objective function minimum
The optimal solution of corresponding separation signal y, realizes the extraction of single signal of vibrating；Pass through frequency domain list source response signal method for solving point
Response of each vibration source at sensitive load point is not calculated；Each micro- vibration is calculated using the contribution amount characterizing method based on vector projection
The contribution amount in dynamic source.The size of contribution amount evaluation index reflects each vibration source and vibration at sensitive load is contributed and influenced big
It is small, basis and foundation can be provided for vibration suppression.
The present invention is based on sparse blind source separatings to realize that satellite micro-vibration source quantitative judge is implemented by step in detail below:
(1) acquisition of vibration signal
With reference to Fig. 2, experimental subjects is satellite cargo tank structure model.Cavity material is aluminum honeycomb compound material, and model bottom is logical
It crosses four rubber air spring supports and is placed on the influence for eliminating ground vibration in optics vibration isolation table, top is camera simulation
Part is at the sensitive load of the present apparatus.Vibration source includes 4 vibrating motors and two vibration excitors.Operating condition of each vibration source to set
Operation, at sensitive load (at camera), tank surface with acceleration transducer acquires vibration response signal, and ensures to observe letter
Number mesh is greater than the number in source.
(2) extraction based on single signal of vibrating with reference to sparse blind deconvolution algorithm
Firstly, utilizing L1The reference L of norm construction blind deconvolution algorithm1Norm objective function J (y, r) are as follows:
J (y, r)=| | F (y) | |1+λE{(y-r)2}
Wherein, y is separation signal；R is reference signal；F () is the discrete Fourier transform of variable；λ is scale factor；
E { } is the mathematic expectaion of variable；
It is frequency single week for signal of vibrating secondly, constructing reference signal according to the prior information in vibration equipment source
Phase signal, using fundamental frequency is the square-wave signal of source signal frequency to be extracted as reference signal, for the source signal of frequency complexity,
The time-frequency characteristics that vibration source is extracted from observation signal collected from sensitive load, using the frequency and phase of main component come structure
Reference signal is made to get reference signal r is arrived；
Then, separation filter length L is set, x (t) obtained in step (1) is rewritten as time lag form
Then blind solution deconvolution model conversation is the ICA model of standard:
Wherein,For the separation filter of 1 × NL dimension；
At this point, the L about y1Norm objective function be converted into aboutObjective function:
Above formula can be expressed by Fourier transformation theory are as follows:
Wherein, Re isThe matrix that real part after being fourier transformed is constituted；Im isImaginary part structure after being fourier transformed
At matrix；
Finally, referring to L using gradient descent method iteration optimization1Norm objective function, by separating vectorIteration is more
The new optimal solution for finding separation signal y corresponding when objective function minimum,Iteration it is updated be known as:
Wherein, k represents kth time iteration and updates, and g's is that objective function J (y, r) is rightInclined gradient, α is the iteration of algorithm
Step-length；
Each iteration obtains newAfterwards, it needs that it is normalized:
Iteration is until convergence obtains separating vectorOptimal solution, then separating signal isRealize single vibration source
The extraction of signal.
The extraction of not homologous signal is realized by constructing different reference signals.
(3) each vibration source is calculated to contribution amount at sensitive load using frequency-domain sparse contribution amount estimation method.
Y (t) is the separation signal extracted, can be obtained by blind deconvolution theory, and there are filter h (τ) satisfaction is as follows
Relational expression:
X (t)=h (τ) * y (t)
Wherein, * is convolution algorithm；
Based on convolution theory, convolution algorithm form is expressed as the product form in frequency domain, each available frequency point
Matrix between the frequency domain value X (ω) of observation signal at ω, filter frequency domain value H (ω) and separation signal frequency domain value Y (ω) closes
System:
X (ω)=H (ω) Y (ω)
A hybrid matrix H (ω) can be obtained at each frequency point in a frequency domain, since signal is sparse in frequency domain
Signal observation signal value or is zero or close to zero or is exactly to be generated by some signal of vibrating at frequency point ω
Response calculates in each source signal ingredient at frequencies omega so largely value is 0 in matrix H (ω) each frequency point ω at
Energy accounts for the specific gravity of the source signal gross energy, and the maximum source signal of specific gravity is exactly the generation source of observation signal at frequency point ω.
Assuming that frequency point ω0Locate observation signal and comes from Yn, then amplitude is 0 to remaining observation signal at this point, in frequency point ω0Place
It can obtain:
Above formula is solved up to frequency point ω0Hybrid matrix H (the ω at place0)；
Hybrid matrix H (ω) at each frequency point is solved using identical method, to can be in frequency domain after hybrid matrix
Solve single source response of each vibration source at camera simulating piece:
Wherein,For response signal of n-th of source at m-th of observation；Hmn(ω) is in hybrid matrix H (ω)
The n-th column element of m row；
By single source response signalThe observation signal X at sensitive loadmOn projection ZmnAs n-th of vibration source to m
The contribution of a observation, ZmnIn XmMiddle proportion is contribution amount evaluation index cmn:
The size of contribution amount evaluation index reflects each vibration source to the size that vibration is contributed and influenced at sensitive load, is
The index of satellite micro-vibration source quantitative judge can inhibit to provide foundation for micro-vibration.
The present invention utilizes the sparse blind deconvolution algorithm of reference from satellite cargo tank structure model sensitive load and tank surface
Each single signal of vibrating is extracted in collected vibration signal, contributes amount estimation method to calculate each micro-vibration source pair by frequency-domain sparse
The contribution amount of sensitive load point, easily and effectively provides the foundation for the vibration suppression of satellite and foundation.Therefore, based on sparse blind
Source separation and frequency-domain sparse contribution amount estimation method realize that the quantitative judge in satellite micro-vibration source is a kind of effective technological approaches.
A specific application example process is given below, while verifying validity of the present invention in engineer application:
Experimental provision is as shown in Fig. 2, experimental subjects is satellite cargo tank structure model 5.Cavity material is aluminum honeycomb compound material,
Model bottom is placed on the influence that ground vibration is eliminated in optics vibration isolation table by four rubber air spring supports, and top is
Camera simulating piece is at the sensitive load of the present apparatus.Vibration source includes two vibration excitors and 4 vibrating motors, two vibration excitors
Output signal controlled respectively with power amplifier 8 and power amplifier 9 by signal generator 10, motor speed passes through speed regulation
Device 4 controls.In this example, two vibration excitors and a vibrating motor are opened as vibration source, the output of the first vibration excitor 6 is believed
Number for comprising 30Hz, 60Hz, the harmonic signal of 90Hz frequency content, the output signal of the second vibration excitor 7 be comprising 40Hz,
The harmonic signal of 80Hz, 120Hz frequency content, the face A motor are run with the revolving speed of 2000r/min.It is measured with acceleration transducer 1
(for the sensitive load position of the present apparatus), the vibration of camera simulating piece root, the face cabinet A, the face cabinet C are rung at the top of camera simulating piece
Induction signal, data collection system 2 acquire vibration acceleration signal, computer 3 by collected vibration acceleration signal data into
It is 5000Hz, sampling time 6s that row, which saves sample frequency, preferably to show that time domain signal characteristics, time domain waveform are only shown
The data of 0.5s.Acceleration transducer can acquire the vibration response signal in three directions, wherein the direction in the vertical face A is the side x
Be the direction y to the direction in, the vertical face B, direction perpendicular to the ground be z to.First carry out the micro-vibration source quantitative judge in the direction x.It adopts
The direction the x vibration response signal time domain waveform and its spectrogram of collection are as shown in Figure 3 and Figure 4.Using with reference to sparse blind deconvolution
The time domain waveform and spectrogram of the isolated single signal of vibrating of algorithm are as shown in Figure 5 and Figure 6.It can be seen that each vibration source
Vibration signal be successfully separated.Then, each vibration source is acquired in sensitive load by frequency domain unit response signal method for solving
The spectrogram of response signal is as shown in Figure 7 at point.Finally, calculating each micro- vibration using the contribution amount characterizing method based on vector projection
Dynamic contribution amount of the source at sensitive load.The direction y and each micro-vibration source in the direction z are calculated at sensitive load with same method
Indeed vibrations response signal of each vibration source at sensitive load can be obtained by each vibration source isolated operation, to acquire in contribution amount
Each contribution amount true value, contribution amount estimated value and true value of each micro-vibration source of different directions at sensitive load are as shown in table 1,
Pass through the validity of contrast verification method.The size of contribution amount evaluation index reflects each vibration source and vibrates at sensitive load
Contribution and the size influenced, are the indexs of satellite micro-vibration source quantitative judge, can inhibit to provide foundation for micro-vibration.
Contribution amount calculated result of each micro-vibration source of 1 different directions of table at sensitive load
Claims (4)
1. a kind of satellite micro-vibration source quantitative identification method based on sparse blind source separating, which comprises the following steps:
(1) acquisition of vibration signal
At satellite cargo tank structure model sensitive load and model surface different location arranges acceleration transducer, acquires each vibration
Vibration signal when source works normally, i.e. observation signal:
X (t)=[x1(t),x2(t),...,xM(t),]T
Wherein, M is observation signal number, xmIt (t) is t moment in the collected vibration response signal of m-th of observation point, m=1~
M；
Ensure that observation signal number is greater than the number N of vibration source；
(2) extraction based on single signal of vibrating with reference to sparse blind deconvolution algorithm
Utilize L1Reference L of the norm construction with reference to sparse blind deconvolution algorithm1Norm objective function, according to the time-frequency domain of vibration source
Feature prior information constructs reference signal, and the optimal solution of separation signal is found using gradient descent method iteration optimization objective function,
Realize the extraction of single signal of vibrating；
(3) each vibration source is calculated to contribution amount at sensitive load using frequency-domain sparse contribution amount estimation method
Single source response signal of each vibration source at sensitive load is calculated using frequency domain list source response signal method for solving；Utilize base
Each vibration source contribution amount at sensitive load is calculated in the contribution amount characterizing method of vector projection.
2. a kind of satellite micro-vibration source quantitative identification method based on sparse blind source separating according to claim 1, special
Sign is that step (2) specifically includes:
(2.1) L is utilized1The reference L of norm construction blind deconvolution algorithm1Norm objective function J (y, r) are as follows:
J (y, r)=| | F (y) | |1+λE{(y-r)2}
Wherein, y is separation signal；R is reference signal；F () is the discrete Fourier transform of variable；λ is scale factor；E
{ } is the mathematic expectaion of variable；
(2.2) reference signal r is constructed according to the prior information in vibration equipment source；
(2.3) separation filter length L is set, x (t) obtained in step (1) is rewritten as time lag form
Then blind solution deconvolution model conversation is the ICA model of standard:
Wherein,For the separation filter of 1 × NL dimension；
At this point, the L about y1Norm objective function be converted into aboutObjective function:
Above formula is expressed by Fourier transformation theory are as follows:
Wherein, Re isThe matrix that real part after being fourier transformed is constituted；Im isWhat the imaginary part after being fourier transformed was constituted
Matrix；
(2.4) L is referred to using gradient descent method iteration optimization1Norm objective function, by separating vectorIteration, which updates, to be found
The optimal solution of corresponding separation signal y when objective function minimum.
3. a kind of satellite micro-vibration source quantitative identification method based on sparse blind source separating according to claim 2, special
Sign is, in step (2.4)Iteration it is updated be known as:
Wherein, k represents kth time iteration and updates, and g is that objective function J (y, r) is rightInclined gradient, α is the iteration step length of algorithm；
Each iteration obtains newAfterwards, it is normalized:
Iteration is until convergence obtains separating vectorOptimal solution, then separating signal isRealize single signal of vibrating
Extraction；The extraction of not homologous signal is realized by constructing different reference signals.
4. a kind of satellite micro-vibration source quantitative identification method based on sparse blind source separating according to claim 2, special
Sign is that step (3) specifically includes:
(3.1) y (t) is set as the separation signal that extracts, and by blind deconvolution theory, there are a filter h (τ) to meet such as ShiShimonoseki
It is formula:
X (t)=h (τ) * y (t)
Wherein, * is convolution algorithm；
Based on convolution theory, convolution algorithm form is expressed as the product form in frequency domain and is seen to get at each frequency point ω
The frequency domain value X (ω) of signal is surveyed, the matrix relationship between filter frequency domain value H (ω) and separation signal frequency domain value Y (ω):
X (ω)=H (ω) Y (ω)
A hybrid matrix H (ω) can be obtained at each frequency point in a frequency domain, calculates in each source signal ingredient at frequencies omega
Energy account for the specific gravity of the source signal gross energy, the maximum source signal of specific gravity is exactly the generation source of observation signal at frequency point ω；
Assuming that frequency point ω0Locate observation signal and comes from Yn, then amplitude is 0 to remaining observation signal at this point, in frequency point ω0Place can obtain:
Above formula is solved up to frequency point ω0Hybrid matrix H (the ω at place0)；
Hybrid matrix H (ω) at each frequency point is solved using identical method, obtaining can be in strip method after hybrid matrix
Single source response of each vibration source at camera simulating piece:
Wherein,For response signal of n-th of source at m-th of observation；Hmn(ω) is m row in hybrid matrix H (ω)
N-th column element；
(3.2) by single source response signalThe observation signal X at sensitive loadmOn projection ZmnAs n-th of vibration source to m
The contribution of a observation, ZmnIn XmMiddle proportion is contribution amount evaluation index cmn:
The size of contribution amount evaluation index reflects each vibration source to the size that vibration is contributed and influenced at sensitive load, is satellite
The index of micro-vibration source quantitative judge.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910323968.0A CN110110619B (en) | 2019-04-22 | 2019-04-22 | Satellite micro-vibration source quantitative identification method based on sparse blind source separation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910323968.0A CN110110619B (en) | 2019-04-22 | 2019-04-22 | Satellite micro-vibration source quantitative identification method based on sparse blind source separation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110110619A true CN110110619A (en) | 2019-08-09 |
| CN110110619B CN110110619B (en) | 2021-02-09 |
Family
ID=67486218
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910323968.0A Active CN110110619B (en) | 2019-04-22 | 2019-04-22 | Satellite micro-vibration source quantitative identification method based on sparse blind source separation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110110619B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113465850A (en) * | 2021-02-07 | 2021-10-01 | 西北工业大学 | Method for identifying mechanical vibration signal path, testing device and testing method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1936926A (en) * | 2006-09-28 | 2007-03-28 | 上海大学 | Image blind separation based on sparse change |
| US20090268962A1 (en) * | 2005-09-01 | 2009-10-29 | Conor Fearon | Method and apparatus for blind source separation |
| CN103674230A (en) * | 2013-11-26 | 2014-03-26 | 西安交通大学 | Denoising source separation based naval vessel vibration noise source separation and recognition method |
| CN102854504B (en) * | 2011-06-30 | 2014-08-13 | 中国科学院电子学研究所 | Method for sparse synthetic aperture radars imaging on basis of echo simulation operators |
| CN108009584A (en) * | 2017-12-01 | 2018-05-08 | 西安交通大学 | Deficient based on the detection of single source point determines blind source separation method |
| CN109409338A (en) * | 2018-12-03 | 2019-03-01 | 西安交通大学 | A kind of satellite micro-vibration source quantitative identification method based on blind source separate technology |
-
2019
- 2019-04-22 CN CN201910323968.0A patent/CN110110619B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090268962A1 (en) * | 2005-09-01 | 2009-10-29 | Conor Fearon | Method and apparatus for blind source separation |
| CN1936926A (en) * | 2006-09-28 | 2007-03-28 | 上海大学 | Image blind separation based on sparse change |
| CN102854504B (en) * | 2011-06-30 | 2014-08-13 | 中国科学院电子学研究所 | Method for sparse synthetic aperture radars imaging on basis of echo simulation operators |
| CN103674230A (en) * | 2013-11-26 | 2014-03-26 | 西安交通大学 | Denoising source separation based naval vessel vibration noise source separation and recognition method |
| CN108009584A (en) * | 2017-12-01 | 2018-05-08 | 西安交通大学 | Deficient based on the detection of single source point determines blind source separation method |
| CN109409338A (en) * | 2018-12-03 | 2019-03-01 | 西安交通大学 | A kind of satellite micro-vibration source quantitative identification method based on blind source separate technology |
Non-Patent Citations (1)
| Title |
|---|
| 成玮 等: "基于独立分量分析的潜艇振源贡献量定量计算方法", 《机械工程学报》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113465850A (en) * | 2021-02-07 | 2021-10-01 | 西北工业大学 | Method for identifying mechanical vibration signal path, testing device and testing method |
| CN113465850B (en) * | 2021-02-07 | 2023-09-08 | 西北工业大学 | Method, test device and test method for identifying mechanical vibration signal path |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110110619B (en) | 2021-02-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102879014B (en) | Optical imaging autonomous navigation semi-physical simulation testing system for deep space exploration proximity process | |
| CN105353789B (en) | Continuous shaking signal time course reproduction control method | |
| CN107543672B (en) | Multiple degrees of freedom micro-vibration environmental simulation method | |
| CN104950169B (en) | A kind of method of testing and system of high speed fibre gyro frequency characteristic | |
| CN106844880A (en) | A kind of satellite momentum wheel disturbs shake experiment and data interpretation method | |
| CN103466109A (en) | Space microgravity environment ground simulation experiment device | |
| CN105808792B (en) | A kind of numerical computation method of tank slosh mass | |
| CN103618492A (en) | Time-frequency transform based method for identifying parameters of synchronous generator | |
| CN107356265B (en) | Method for identifying even unbalance amount based on gyro flywheel torque coil current | |
| CN107843255A (en) | Towards the engineering truck driving posture measuring system and method for motion reappearance | |
| CN109753689A (en) | A kind of online identifying approach of electric system electromechanical oscillations modal characteristics parameter | |
| CN110110619A (en) | A kind of satellite micro-vibration source quantitative identification method based on sparse blind source separating | |
| CN107037739A (en) | A kind of used group analogy method of carrier rocket Hardware-in-loop Simulation Experimentation | |
| Yamazaki | Stratified turbulence near a critical dissipation rate | |
| CN107941441A (en) | Determine the method that the in-orbit border of simulation influences the in-orbit dynamics of spacecraft | |
| Le Treut et al. | Isotopic modeling of climatic oscillations: Implications for a comparative study of marine and ice core records | |
| CN104062072B (en) | A kind of shafting dynamic balance Multipurpose Optimal Method based on differential searching algorithm | |
| Klein et al. | Mesoscale heterogeneity of the wind-driven mixed layer: Influence of a quasigeostrophic flow | |
| CN109409338A (en) | A kind of satellite micro-vibration source quantitative identification method based on blind source separate technology | |
| CN113484911B (en) | Earthquake motion evolution power spectrum-based earthquake displacement response calculation method | |
| CN112327957B (en) | Method and device for controlling low-frequency vibration multi-order line spectrum of rotor | |
| CN108120439B (en) | Three-component induction coil attitude measurement method and device | |
| CN108573084B (en) | Environmental vibration test method and system | |
| Collette et al. | Seismic response of linear accelerators | |
| Bonanno et al. | Seismic inference of differential rotation in Procyon A |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |