Single-channel mechanical vibration signal blind separation method, system, device and medium

Abstract

The application provides a single-channel mechanical vibration signal blind separation method, system, device and medium, a single vibration sensor is used to obtain a mechanical vibration signal under a running condition as a single-channel mechanical vibration signal to be analyzed, the single-channel mechanical vibration signal is decomposed by using symplectic geometry modal decomposition to obtain a single-component signal, the single-component signal is reconstructed according to a frequency spectrum correlation coefficient to obtain a symplectic geometry modal component, the single-channel mechanical vibration signal, the symplectic geometry modal component and a residual term are used to form a multi-dimensional observation signal, a singular value decomposition is performed on a self-correlation matrix of the multi-dimensional observation signal, a Bayesian information criterion is used to obtain an estimation of the number of signal sources, the symplectic geometry modal component is selected according to a maximum time-domain correlation coefficient principle to form a new multi-dimensional signal observation matrix together with the original observation signal, and a blind source separation method based on time-frequency analysis is used to separate the multi-dimensional signal observation matrix. The application can separate multiple mechanical vibration source signals from a single-channel vibration signal.

Description

Technical Field

[0001] This invention relates to the field of blind source separation technology for mechanical vibration signals, specifically to a blind separation method for single-channel mechanical vibration signals based on Symptotic Geometry Mode Decomposition (SGMD), and more particularly to a blind separation method, system, device, and medium for single-channel mechanical vibration signals. Background Technology

[0002] Blind signal separation (BSS) is a signal processing technique that involves blindly separating source signals from sensor-observed signals. In recent years, BSS has been increasingly used in mechanical vibration signal processing, playing a significant role in fault diagnosis and condition monitoring of mechanical equipment. Based on the relationship between the number of source signals and the number of observed signals, blind source separation can be classified into three categories: overdetermined blind source separation, well-determined blind source separation, and underdetermined blind source separation. Single-channel blind source separation is an extreme case of underdetermined blind source separation. Single-channel blind source separation (SCBSS) refers to the process of using a single receiving sensor to receive the observed signal and recovering all source signals using only this single observed signal.

[0003] Traditional blind source separation methods require the number of sensors to be greater than or equal to the number of features to be separated. In practice, without prior knowledge of the vibration source, it is difficult to determine the number of sensors to be installed in advance. At the same time, due to limitations in installation space and other objective factors, it is also impossible to install a sufficient number of vibration sensors on mechanical equipment. Therefore, it is also impossible to obtain sensor signals greater than or equal to the number of signals to be separated, causing blind source separation to become underdetermined blind source separation or even single-channel blind source separation. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method, system, device, and medium for blind separation of single-channel mechanical vibration signals.

[0005] According to the present invention, a method, system, device, and medium for blind separation of single-channel mechanical vibration signals are provided, the solution of which is as follows:

[0006] In a first aspect, a blind separation method for single-channel mechanical vibration signals is provided, the method comprising:

[0007] Step S1: Use a single vibration sensor to acquire the mechanical vibration signal under operating conditions as the single-channel mechanical vibration signal to be analyzed;

[0008] Step S2: The single-channel mechanical vibration signal to be analyzed is decomposed using symplectic geometric mode decomposition to obtain several single-component signals;

[0009] Step S3: Calculate the spectral correlation coefficients between the several single-component signals and construct the spectral correlation matrix of the single-component signals;

[0010] Step S4: Using the maximum spectral correlation coefficient as the selection criterion, reconstruct the single-component signal to obtain several symplectic geometric mode components;

[0011] Step S5: Combine the single-channel mechanical vibration signal to be analyzed with the symplectic geometric modal components and their residual terms to form a multidimensional observation signal;

[0012] Step S6: Perform singular value decomposition on the autocorrelation matrix of the multidimensional observed signal, and use the Bayesian information criterion to estimate the number of signal sources;

[0013] Step S7: Based on the principle of maximizing the temporal correlation coefficient, select the symplectic geometric mode components and the original observation signal to form a new multidimensional signal observation matrix;

[0014] Step S8: Use a time-frequency analysis-based blind source separation method to separate the multidimensional signal observation matrix to obtain the signal source of the single-channel mechanical vibration signal to be analyzed.

[0015] Preferably, the spectral correlation coefficient is defined as:

[0016]

[0017] In the formula, Let represent the modulus of the Fourier transform of the two signals, respectively. N It is a discrete value sequence number in the frequency domain; Indicates frequency; i Indicates the first i A single-component signal, 1≤ i ≤ d -1; j Indicates the first j A single-component signal, 2≤ j ≤ d and i < j , d The number of single-component signals; Indicates the first i The single-component signal and the first j The spectral correlation coefficient of a single-component signal.

[0018] Preferably, the spectral correlation matrix Defined as:

[0019]

[0020] in, Indicates the first d A single-component signal and d-1 The spectral correlation coefficient of a single-component signal.

[0021] Secondly, a single-channel mechanical vibration signal blind separation system is provided, the system comprising:

[0022] Module M1: Uses a single vibration sensor to acquire mechanical vibration signals under operating conditions as the single-channel mechanical vibration signal to be analyzed;

[0023] Module M2: The single-channel mechanical vibration signal to be analyzed is decomposed using symplectic geometric mode decomposition to obtain several single-component signals;

[0024] Module M3: Calculates the spectral correlation coefficients between several of the single-component signals and constructs the spectral correlation matrix of the single-component signals;

[0025] Module M4: Using the maximum spectral correlation coefficient as the selection criterion, the single-component signal is reconstructed to obtain several symplectic geometric mode components;

[0026] Module M5: Combines the single-channel mechanical vibration signal to be analyzed with the symplectic geometric modal components and their residual terms to form a multidimensional observation signal;

[0027] Module M6: Perform singular value decomposition on the autocorrelation matrix of the multidimensional observed signal and use the Bayesian information criterion to estimate the number of signal sources;

[0028] Module M7: Based on the principle of maximizing the time-domain correlation coefficient, the symplectic geometric mode components are selected and combined with the original observation signal to form a new multidimensional signal observation matrix;

[0029] Module M8: The multidimensional signal observation matrix is ​​separated using a blind source separation method based on time-frequency analysis to obtain the signal source of the single-channel mechanical vibration signal to be analyzed.

[0030] Preferably, the spectral correlation coefficient is defined as:

[0031]

[0032] In the formula, Let represent the modulus of the Fourier transform of the two signals, respectively. N It is a discrete value sequence number in the frequency domain; Indicates frequency; i Indicates the first i A single-component signal, 1≤ i ≤ d -1;j Indicates the first j A single-component signal, 2≤ j ≤ d and i < j , d The number of single-component signals; Indicates the first i The single-component signal and the first j The spectral correlation coefficient of a single-component signal.

[0033] Preferably, the spectral correlation matrix Defined as:

[0034]

[0035] in, Indicates the first d A single-component signal and d-1 The spectral correlation coefficient of a single-component signal.

[0036] Thirdly, an apparatus is provided, the apparatus comprising:

[0037] One or more processors;

[0038] Storage device for storing one or more programs.

[0039] When the one or more programs are executed by the one or more processors, the one or more processors perform the steps in the method.

[0040] Fourthly, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps of the method.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. The method of this invention combines the advantages of symplectic geometric mode decomposition and time-frequency blind separation, overcoming the limitations of blind source separation under single-channel signal conditions;

[0043] 2. This invention uses symplectic geometric mode decomposition to decompose single-channel vibration signals, effectively avoiding the shortcomings of VMD, EMD and LMD methods in signal processing, such as parameter settings and mode mixing.

[0044] 3. The symplectic geometric mode components selected in this invention based on the principle of maximizing the spectral correlation coefficient are beneficial for the reconstruction and separation of vibration signal sources.

[0045] Other beneficial effects of the present invention will be explained in detail through the introduction of specific technical features and technical solutions in specific embodiments. Those skilled in the art should be able to understand the beneficial technical effects brought about by these technical features and technical solutions through the introduction of these technical features and technical solutions. Attached Figure Description

[0046] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0047] Figure 1 This is a flowchart of the method of the present invention;

[0048] Figure 2 This is a time-domain plot of a single-channel mechanical vibration signal.

[0049] Figure 3 Frequency domain diagram of a single-channel mechanical vibration signal;

[0050] Figure 4 Time-domain plot of independent vibration source s1;

[0051] Figure 5 Time-domain plot of independent vibration source s2;

[0052] Figure 6 Time-domain plot of independent vibration source s3;

[0053] Figure 7 The spectrum of independent vibration source s1;

[0054] Figure 8 The spectrum of the independent vibration source s2;

[0055] Figure 9 The spectrum of independent vibration source S3;

[0056] Figure 10 Spectral correlation matrix A 32x32 square matrix diagram;

[0057] Figure 11 To separate the vibration source signal y 1( t The time-domain waveform of ).

[0058] Figure 12 To separate the vibration source signal y 2( t The time-domain waveform of ).

[0059] Figure 13 To separate the vibration source signal y 3( t The time-domain waveform of ). Detailed Implementation

[0060] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0061] This invention provides a blind separation method for single-channel mechanical vibration signals, referring to... Figure 1 As shown, the specific content of this method is as follows:

[0062] Step 1: Use a vibration acceleration sensor to acquire mechanical vibration signals under operating conditions as the signals to be analyzed. y ( t ), t=1,2,3,…,T, where T is the signal length.

[0063] Step 2: Use symplectic geometric mode decomposition to analyze the single-channel mechanical vibration signal. y ( t Decompose to obtain d A single-component signal Y 1 、Y 2 、…、Y d-1 、Y d .

[0064] Step 3: Calculate the above d A single-component signal Y 1 、Y 2 、…、Y d-1 、Y d The spectral correlation coefficients between them ρ 1、2 、 ρ 1、3 、…、ρ 1、d-1 ρ 1、d ; ρ 2、3 ρ 2、4 、…、ρ 2、d-1 ρ 2、d ; …、ρ d-1、d .

[0065] The spectral correlation coefficient is defined as:

[0066]

[0067] In the formula, Let represent the modulus of the Fourier transform of the two signals, respectively. N It is a discrete value sequence number in the frequency domain; Indicates frequency; i Indicates the first i A single-component signal, 1≤ i ≤ d -1; j Indicates the first j A single-component signal, 2≤ j ≤ d and i < j , d The number of single-component signals; Indicates the first i The single-component signal and the first j The spectral correlation coefficient of a single-component signal.

[0068] Among them, the spectral correlation matrix Defined as:

[0069]

[0070] in, Indicates the first d A single-component signal and d-1 The spectral correlation coefficient of a single-component signal.

[0071] Step 4: Using the highest spectral correlation coefficient as the selection criterion, apply the above... d A single-component signal Y 1 、Y 2 、…、Y d-1 、 Y d Reconstruction is performed to obtain m Symplectic geometric modal components SGC 1 SGC 2 ... SGC m .

[0072] Step 5: Then, the original single-channel mechanical vibration signal y ( t )and m Each symplectic geometric modal component SGC 1 SGC 2 、…、 SGC m Together m +1D observation signal Z .

[0073] Step 6: Perform singular value decomposition on the autocorrelation matrix of the above multidimensional observed signals, and use the Bayesian information criterion to estimate the number of signal sources. l ,and l≤m .

[0074] Step 7: Select according to the principle of maximizing energy l-1 The individual symplectic geometric mode components, together with the original observation signal, form a new... l 3D signal observation matrix X .

[0075] Step 8: Use a blind source separation method based on time-frequency analysis to separate the above multidimensional signal observation matrix X to obtain the signal to be analyzed. y ( t )of l One signal source y 1( t ), y 2( t ), … 、y l ( t ).

[0076] The present invention also provides a single-channel mechanical vibration signal blind separation system. Those skilled in the art can understand the single-channel mechanical vibration signal blind separation method provided by the present invention as a specific implementation of the single-channel mechanical vibration signal blind separation system, that is, the single-channel mechanical vibration signal blind separation system can be implemented by executing the steps of the single-channel mechanical vibration signal blind separation method.

[0077] The invention will now be described in more detail through examples.

[0078] Step 1: Simulate the single-channel mechanical vibration signal under operating conditions acquired by the vibration acceleration sensor through simulation. S , S The expression is as follows:

[0079]

[0080] in s 1. s 2. s 3 represents three independent vibration sources, expressed as follows:

[0081]

[0082] in, Indicates the frequency of the vibration source. f 1 = 100Hz f 2 = 150Hz f 3 = 180Hz f 4 = 500Hzf 4 = 1000Hz; Represents time, 0≤ t ≤0.5;

[0083] A The matrix represents the mixture of three vibration sources. Set the sampling frequency to 6000Hz. S The waveform and spectrum are as follows Figure 2 and Figure 3 As shown; the time-domain plots of independent vibration sources s1, s2, and s3 are referenced. Figure 4 , Figure 5 and Figure 6 The spectrum diagrams of independent vibration sources s1, s2, and s3 are referenced. Figure 7 , Figure 8 and Figure 9 .

[0084] Step 2: Use Symptotic Geometry Mode Decomposition (SGMD) to analyze the single-channel mechanical vibration signal. S The decomposition yielded 33 single-component signals. Y 1 、Y 2 、…、Y 32 、Y 33 ;

[0085] Step 3: Calculate the above 33 single-component signals Y 1 、Y 2 、…、Y 32 、Y 33 The spectral correlation coefficients between them ρ 1、2 、 ρ 1、3 、…、ρ 1、32 ρ 1、33 ; ρ 2、3 ρ 2、4 、…、ρ 2、32 ρ 2、33 ; …、ρ 32、33 Single-component signals are constructed from the calculated spectral correlation coefficients. Y 1 、Y 2 、…、Y 32 、Y 33 Spectral correlation matrix , It is a 32×32 square matrix, see reference. Figure 10 As shown.

[0086] Step 4: Based on the calculated spectral correlation matrix, use the largest spectral correlation coefficient as the selection criterion to apply the above... 33 A single-component signal Y 1 、Y 2 、…、Y 32 、Y 33 Reconstruction is performed to obtain 5 Symplectic geometric modal components SGC 1 、 SGC 2 ... SGC 5;

[0087] SGC 1= Y 1 +Y 2 ；

[0088] SGC 2= Y 3 +Y 4 ；

[0089] SGC 3= Y 5 +Y 6 ；

[0090] SGC 4= Y 7 +Y 8 +Y 9 +Y 10 +Y 11 +Y 12 +Y 13 +Y 14 +Y 15 +Y 16 ；

[0091] SGC 5= Y 17 +Y 18 +Y 19 +Y 20 +Y 21+Y 22 +Y 23 +Y 24 +Y 25 +Y 26 +Y 27 +Y 28 +Y 29 +Y 30 +Y 31 +Y 32 +Y 33 ;

[0092] Step 5: Then, the original single-channel mechanical vibration signal... y ( t )and m Each symplectic geometric modal component SGC 1 SGC 2 、…、 SGC m Together they form a 6-dimensional observation signal Z :

[0093] ;

[0094] Step 6: Analyze the above 6-dimensional observation signals. Z Singular value decomposition is performed on the autocorrelation matrix, and the number of signal sources is estimated by using the Bayesian information criterion.

[0095] Step 7: Calculation SGC 1 SGC 2 ... SGC Energy of 5:

[0096] E1=16657.9, E1=492.3, E1=234.8, E1=292.3, E1=1.3×10 -24 ;

[0097] Selected based on the principle of maximizing energy 2 Each symplectic geometric modal component SGC 1 SGC 2. Together with the original observed signal, they form a new 3D signal observation matrix. X :

[0098] ;

[0099] Step 8: Use a blind source separation method based on time-frequency analysis to separate the above multidimensional signal observation matrix X to obtain the signal to be analyzed. y ( t The three source signals of ) y 1( t ), y 2( t ), y 3( t Three separate vibration source signals y 1( t ), y 2( t ), y 3( t Time-domain waveform reference Figure 11 , Figure 12 and Figure 13 ;

[0100] The present invention provides a method, system, device and medium for blind separation of single-channel mechanical vibration signals, which overcomes the bottleneck of insufficient observation signals for underdetermined blind source separation and achieves the purpose of separating multiple mechanical vibration source signals from a single-channel vibration signal; effectively extracting multiple mechanical vibration source signals from a one-dimensional vibration signal is of great significance for mechanical equipment system identification, fault diagnosis and condition monitoring.

[0101] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0102] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A blind separation method for single-channel mechanical vibration signals, characterized in that, include: Step S1: Use a single vibration sensor to acquire the mechanical vibration signal under operating conditions as the single-channel mechanical vibration signal to be analyzed; Step S2: The single-channel mechanical vibration signal to be analyzed is decomposed using symplectic geometric mode decomposition to obtain several single-component signals; Step S3: Calculate the spectral correlation coefficients between the several single-component signals and construct the spectral correlation matrix of the single-component signals; Step S4: Using the maximum spectral correlation coefficient as the selection criterion, reconstruct the single-component signal to obtain several symplectic geometric mode components; Step S5: Combine the single-channel mechanical vibration signal to be analyzed with the symplectic geometric modal components and their residual terms to form a multidimensional observation signal; Step S6: Perform singular value decomposition on the autocorrelation matrix of the multidimensional observed signal, and use the Bayesian information criterion to estimate the number of signal sources; Step S7: Based on the principle of maximizing the temporal correlation coefficient, select the symplectic geometric mode components and the original observation signal to form a new multidimensional signal observation matrix; Step S8: Use a time-frequency analysis-based blind source separation method to separate the multidimensional signal observation matrix to obtain the signal source of the single-channel mechanical vibration signal to be analyzed.

2. The blind separation method for single-channel mechanical vibration signals according to claim 1, characterized in that, The spectral correlation coefficient is defined as: In the formula, Let represent the modulus of the Fourier transform of the two signals, respectively. N It is a discrete value sequence number in the frequency domain; Indicates frequency; i Indicates the first i A single-component signal, 1≤ i ≤ d -1; j Indicates the first j A single-component signal, 2≤ j ≤ d and i < j , d The number of single-component signals. Indicates the first i The single-component signal and the first j The spectral correlation coefficient of a single-component signal.

3. The blind separation method for single-channel mechanical vibration signals according to claim 1, characterized in that, The spectral correlation matrix Defined as: in, Indicates the first d A single-component signal and d-1 The spectral correlation coefficient of a single-component signal.

4. A blind separation system for single-channel mechanical vibration signals, characterized in that, include: Module M1: Uses a single vibration sensor to acquire mechanical vibration signals under operating conditions as the single-channel mechanical vibration signal to be analyzed; Module M2: The single-channel mechanical vibration signal to be analyzed is decomposed using symplectic geometric mode decomposition to obtain several single-component signals; Module M3: Calculates the spectral correlation coefficients between several of the single-component signals and constructs the spectral correlation matrix of the single-component signals; Module M4: Using the maximum spectral correlation coefficient as the selection criterion, the single-component signal is reconstructed to obtain several symplectic geometric mode components; Module M5: Combines the single-channel mechanical vibration signal to be analyzed with the symplectic geometric modal components and their residual terms to form a multidimensional observation signal; Module M6: Perform singular value decomposition on the autocorrelation matrix of the multidimensional observed signal and use the Bayesian information criterion to estimate the number of signal sources; Module M7: Based on the principle of maximizing the time-domain correlation coefficient, the symplectic geometric mode components are selected and combined with the original observation signal to form a new multidimensional signal observation matrix; Module M8: The multidimensional signal observation matrix is ​​separated using a blind source separation method based on time-frequency analysis to obtain the signal source of the single-channel mechanical vibration signal to be analyzed.

5. The single-channel mechanical vibration signal blind separation system according to claim 4, characterized in that, The spectral correlation coefficient is defined as: In the formula, Let represent the modulus of the Fourier transform of the two signals, respectively. N It is a discrete value sequence number in the frequency domain; Indicates frequency; i Indicates the first i A single-component signal, 1≤ i ≤ d -1; j Indicates the first j A single-component signal, 2≤ j ≤ d and i < j , d The number of single-component signals; Indicates the first i The single-component signal and the first j The spectral correlation coefficient of a single-component signal.

6. The single-channel mechanical vibration signal blind separation system according to claim 4, characterized in that, The spectral correlation matrix Defined as: in, Indicates the first d A single-component signal and d-1 The spectral correlation coefficient of a single-component signal.

7. A device, characterized in that, The device includes: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors perform the steps of the method as described in any one of claims 1 to 3.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 3.

Images