Guided wave acoustic time measuring method and device based on multi-resolution singular value decomposition
Through multi-resolved singular value decomposition and sliding window processing, the singular correlation value sequence during the guide sound is extracted, and the problem of low signal-to-noise ratio measured during the guide sound in complex waveguide structures is solved, and the waveguide sound in complex structures is realized Accurate determination.
Patent Information
- Application Number
- CN202411173252.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-26
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2044-08-26
AI Technical Summary
In complex waveguide structures, it is difficult to accurately obtain ultrasonic waveguide sound measurement, because the echo signal is affected by noise and multimodal aliasing, and the signal-to-noise ratio is low, making it difficult to measure waveguide sound.
Using the wave guide sound time measurement method based on multi-resolution singular value decomposition, the singular correlation value sequence is extracted through sliding window processing and binary recursive matrix construction, and the time difference between the direct wave and the singular correlation peak at the end echo is determined as When the sound of the guided wave.
When effectively measuring the guided sound in a complex structure, the signal-to-noise ratio is improved, the acoustic measurement of low signal-to-noise ratio guided signals is realized, and the detection capability of the guided sound is enhanced.
Smart Images

Figure CN119023053A_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of signal processing technology, and more specifically, relates to a method and device for determining guided wave acoustic timing based on multi-resolution singular value decomposition. Background Art
[0002] Ultrasonic guided waves based on the acoustic-elastic effect have the advantages of wide coverage, long propagation distance, and propagation parallel to the medium boundary. While detecting medium defects, they can truly reflect the overall stress of the medium. They are a very promising non-destructive testing stress identification method. However, in the process of stress detection using ultrasonic guided waves, the end echo is affected by noise and multi-modal aliasing, and its echo arrival time is difficult to obtain accurately.
[0003] When ultrasonic guided waves propagate in a waveguide under different stress levels, the propagation velocity will change, which is the acoustic elastic effect of the guided wave. The key to applying the acoustic elastic effect to stress detection is to obtain the acoustic time of the guided wave, and calculate the guided wave velocity based on the obtained acoustic time, and reflect the overall stress change of the medium through the change of the guided wave velocity. In complex waveguide structures, the guided wave is affected by multi-mode and noise, and its echo signal is severely aliased and attenuated, which makes it difficult to measure the echo acoustic time, posing a huge challenge to stress detection. There are many methods for acoustic time measurement, such as amplitude method, zero-crossing detection method, cross-correlation method, characteristic structure method and time-frequency measurement method. The cross-correlation method, amplitude method and zero-crossing detection method are easily affected by noise, the time-frequency measurement method has the problem of time-frequency resolution, and the characteristic structure method requires multiple sets of independent signals. For the problem of guided wave acoustic time measurement of complex structures, the above methods all have their own limitations. Summary of the invention
[0004] In view of the defects existing in the related art, the embodiments of the present application provide a method and device for determining the guided wave acoustic timing based on multi-resolution singular value decomposition, aiming to solve the problem of detecting the guided wave acoustic timing of complex structures.
[0005] In a first aspect, an embodiment of the present application provides a method for determining guided wave acoustic timing based on multi-resolution singular value decomposition, comprising:
[0006] Preprocessing the waveguide signal S, extracting the direct wave signal SD from the preprocessed waveguide signal SP, and intercepting the reference signal Y from the direct wave signal SD;
[0007] The pre-processed waveguide signal SP is subjected to sliding window processing, and based on the signal X in the i-th window during the sliding window processing, i and the reference signal Y, construct the binary recursive matrix H 1,i , based on the binary recursive matrix H 1,i The window signal X in the sliding window processing process i Perform j times of multi-resolution singular value decomposition to obtain j times of singular correlation values, based on all window signals X in the sliding window processing processi The j-order singular correlation values are calculated respectively, and a j-order singular correlation value sequence is constructed;
[0008] In the waveguide singular correlation value spectrum corresponding to the sub-singular correlation value sequence, the time difference between the singular correlation spectrum peaks at the direct wave and the end echo is determined as the acoustic time of the waveguide.
[0009] In some embodiments, extracting the reference signal Y from the direct wave signal SD comprises:
[0010] The direct wave signal SD is processed by sliding window, and the x-th window signal Y is obtained based on the sliding window processing process. x , construct the binary recursive matrix H x , based on the binary recursive matrix H x Perform singular value decomposition to obtain a singular correlation value; based on all window signals Y in the sliding window processing process x The first-order singular correlation values are calculated respectively, and a first-order singular correlation value sequence ψ1 is constructed;
[0011] The first-order singular correlation value sequence ψ1 is normalized, and the normalized first-order singular correlation value sequence is plotted into a direct wave singular correlation value spectrum. The signal intercepted when the center of the sliding window is located at the peak of the direct wave singular correlation value spectrum during the sliding window processing is determined as the reference signal Y.
[0012] In some embodiments, constructing a singular correlation value sequence ψ1 comprises:
[0013] Perform sliding window processing on the direct wave signal SD to obtain the window signal Y in the sliding window processing process x =(y x,1 ,y x,2 , …, y x,L ), L is the sliding window length;
[0014] Based on Y x =(y x,1 ,y x,2 , …, y x,L ), construct the binary recursive matrix H x And perform singular value decomposition to obtain Get the first singular value correlation value ψ 1,x =ρσ x,1 (σ x,1 -σ x,2 );ρ is the zero point correction coefficient, ρ=a max -|a m |, a max is the maximum value in the time series, a m is the value at the middle position of the time series; U x and V xis the orthogonal matrix generated during the singular value decomposition process;
[0015] Based on all the in-window signals Y during the sliding window processing, x calculate the first-order singular correlation values respectively to obtain the first-order singular correlation value sequence ψ1 = (ψ 1,1 , ψ 1,2 , …, ψ 1,M ), where M is the total number of sliding windows for the sliding window processing of the direct wave signal SD.
[0016] In some embodiments, constructing the j-th order singular correlation value sequence includes:
[0017] Perform sliding window processing on the preprocessed guided wave signal SP to obtain the in-window signal X i =(x i,1 , x i,2 , …, x i,Q ); Q is the signal length of the in-window signal X i , which is the same as the length of the reference signal Y;
[0018] Based on X i =(x i,1 , x i,2 , …, x i,Q ) and the reference signal Y = (y1, y2, …, y Q ), construct the binary recursive matrix
[0019] Based on the binary recursive matrix H 1,i , perform j times of multi-resolution singular value decomposition to obtain the j-th order singular correlation value sequence; among them, in the j-th order multi-resolution singular value decomposition, the first multi-resolution singular value decomposition is to decompose X i into the approximate signal A 1,i and the detail signal D 1,i , and the n-th multi-resolution singular value decomposition is to further decompose the approximate signal A n-1,i obtained from the (n - 1)-th multi-resolution singular value decomposition into the approximate signal A n,i and the detail signal D n,i ; 1 < n ≤ j, and n is a positive integer.
[0020] In some embodiments, the process of the first multi-resolution singular value decomposition includes:
[0021] Perform multi-resolution singular value decomposition on the binary recursive matrix H 1,i to obtain where the singular values include σ 1,i,1 and σ 1,i,2 , σ 1,i,1 ≥ σ 1,i,2 ; U 1,i and V1,i is the orthogonal matrix generated in the process of multi-resolution singular value decomposition;
[0022] H 1,i Expand to get U 1,i =(u 1,1,i ,u 1,2,i ), U 1,i ∈R 2×2 , V j,i =(v 1,1,i ;v 1,2,i ;…;v 1,Q,i ), V j,i ∈R Q×Q ;H 1,a,i is an approximate matrix, H 1,d,i is the detail matrix,
[0023] Based on the approximation matrix and the detail matrix, the approximate signal A is obtained 1,i =L 1,a,1 , and detail signal D 1,i =L 1,d,1 , where X i =L 1,a,1 +L 1,d,1 =A 1,i +D 1,i .
[0024] In some embodiments, based on the binary recursive matrix H 1,i , perform j times of multi-resolution singular value decomposition, and obtain j times singular correlation value sequences, including:
[0025] Based on the binary recursive matrix H 1,i , perform j times of multi-resolution singular value decomposition, and obtain the approximate signal A obtained by the j-th multi-resolution singular value decomposition j,i =(a j,1,i , a j,2,i , …, a j,Q,i );
[0026] Based on A j,i Constructing a binary recursive matrix
[0027] For H j,i Perform singular value decomposition and get H j,i =U j,i Λ j,i V j,i T Among them, U j,i =(u j,1,i ,u j,2,i ), U j,i ∈R2×2 , V j,i =(v j,1,i ;v j,2,i ;…;v j,Q,i ), V j,i ∈R Q×Q , Λ j,i =(diag(σ j,1,i , σ j,2,i ), O), Λ j,i ∈R Q×Q ;
[0028] Get the j-th singular correlation value ψ j,i =ρσ j,1,i (σ j,1,i -σ j,2,i ), where ρ is the zero point correction coefficient, ρ = a max -|a m |, a max is the maximum value in the time series, a m is the value at the middle position of the time series;
[0029] Based on the j-order singular correlation values calculated from all window signals in the sliding window processing process of the preprocessed waveguide signal SP, the j-order singular correlation value sequence ψ is obtained. j =(ψ 1,1 , ψ 1,2 ,…,ψ 1,K ), K is the total number of sliding windows for performing sliding window processing on the pre-processed waveguide signal SP.
[0030] In some embodiments, the waveguide signal S of the waveguide is preprocessed, including:
[0031] The SG filtering method is used to remove the noise in the waveguide signal S;
[0032] The linear interpolation method is used to process the waveguide signal after noise removal.
[0033] In a second aspect, an embodiment of the present application further provides a guided wave acoustic timing determination device based on multi-resolution singular value decomposition, comprising:
[0034] A preprocessing module, used to preprocess the waveguide signal S, extract the direct wave signal SD from the preprocessed waveguide signal SP, and intercept the reference signal Y from the direct wave signal SD;
[0035] The sliding window processing module is used to perform sliding window processing on the pre-processed waveguide signal SP, based on the i-th window signal X i and the reference signal Y, construct the binary recursive matrix H 1,i , based on the binary recursive matrix H 1,iThe window signal X in the sliding window processing process i Perform j times of multi-resolution singular value decomposition to obtain j times of singular correlation values, based on all window signals X in the sliding window processing process i The j-order singular correlation values are calculated respectively, and a j-order singular correlation value sequence is constructed;
[0036] The acoustic time determination module is used to determine the time difference between the singular correlation spectrum peaks at the direct wave and the end echo in the waveguide singular correlation value spectrum corresponding to the sub-singular correlation value sequence as the acoustic time of the waveguide.
[0037] In a third aspect, an embodiment of the present application further provides an electronic device, comprising: at least one memory for storing programs; and at least one processor for executing the programs stored in the memory. When the program stored in the memory is executed, the processor is used to execute the method described in the first aspect or any possible implementation of the first aspect.
[0038] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program runs on a processor, the processor executes the method described in the first aspect or any possible implementation of the first aspect.
[0039] In a fifth aspect, an embodiment of the present application further provides a computer program product, which, when executed on a processor, enables the processor to execute the method described in the first aspect or any possible implementation of the first aspect.
[0040] The embodiment of the present application provides a method and device for determining the acoustic time of a waveguide based on multi-resolution singular value decomposition. Based on the principle of binary recursion, a new binary recursive matrix is proposed. In combination with multi-resolution singular value decomposition, a new echo acoustic time positioning feature, namely the singular correlation value, is extracted. The sub-singular correlation value spectrum of the waveguide signal is calculated, and the singular correlation value spectrum is used to determine the acoustic time of the waveguide. The time difference between the peak values of the singular correlation spectrum at the direct wave and the end echo is extracted as the acoustic time. Compared with the related art, in a complex waveguide structure, the echo signal-to-noise ratio is low, which makes the echo acoustic time difficult to determine. The present application can determine the acoustic time from the waveguide signal with a low signal-to-noise ratio, and can realize the detection of the acoustic time of the waveguide in a complex structure. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] In order to more clearly illustrate the technical solutions in the present application or related technologies, the following is a brief introduction to the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0042] Figure 1 It is one of the flow charts of the method for determining the acoustic timing of guided waves based on multi-resolution singular value decomposition provided in the embodiment of the present application;
[0043] Figure 2 This is the second flow chart of the method for determining the acoustic timing of guided waves based on multi-resolution singular value decomposition provided in an embodiment of the present application;
[0044] Figure 3 is a schematic diagram of a measured cable waveguide signal to be analyzed provided in an embodiment of the present application;
[0045] Figure 4 is a schematic diagram of a preprocessed waveguide signal provided in an embodiment of the present application;
[0046] Figure 5 It is a direct wave signal extracted from the pre-processed waveguide signal and a first-order singular correlation value spectrum calculated therefrom provided in an embodiment of the present application;
[0047] Figure 6 is a schematic diagram of a multi-resolution singular value decomposition process provided by an embodiment of the present application;
[0048] Figure 7 Schematic diagram of the 5th singular correlation value spectrum of the waveguide signal provided in the embodiment of the present application;
[0049] Figure 8 It is a structural schematic diagram of a guided wave acoustic timing measurement device based on multi-resolution singular value decomposition provided in an embodiment of the present application;
[0050] Fig. 9 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0051] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0052] Figure 1 It is one of the flow charts of the method for determining acoustic timing of guided waves based on multi-resolution singular value decomposition provided in the embodiment of the present application, such as Figure 1 As shown, the method includes the following steps 101 to 103:
[0053] Step 101, pre-process the waveguide signal S of the waveguide, extract the direct wave signal SD from the pre-processed waveguide signal SP, and intercept the reference signal Y from the direct wave signal SD.
[0054] In some embodiments, the waveguide signal S of the waveguide is preprocessed in step 101, specifically including:
[0055] The SG filtering method is used to remove the noise in the waveguide signal S;
[0056] The linear interpolation method is used to process the waveguide signal after noise removal.
[0057] Specifically, the SG filtering method is first used to remove high-frequency noise on the signal surface. This method has little effect on the echo acoustic time. Then, the linear interpolation method is used to interpolate between two sampling points through linear fitting to generate more sampling points to increase the sampling rate, thereby improving the acoustic time measurement accuracy. The computational complexity will also increase with the increase in sampling points.
[0058] In some embodiments, the step 101 of extracting the reference signal Y from the direct wave signal SD specifically includes the following steps (1a) to (2a):
[0059] Step (1a), perform sliding window processing on the direct wave signal SD, based on the x-th window signal Y in the sliding window processing process x , construct the binary recursive matrix H x , based on the binary recursive matrix H x Perform singular value decomposition to obtain a singular correlation value based on all window signals Y in the sliding window processing process x The first-order singular correlation values are calculated respectively, and a first-order singular correlation value sequence ψ1 is constructed.
[0060] Among them, for the binary recursive matrix H x Introduction: According to the direct wave signal SD, the signal in the sliding window is Y x =(y x,1 ,y x,2 , …, y x,L ), construct the matrix H x for:
[0061]
[0062] For the matrix H x Perform SVD decomposition, where the singular values are λ1 and λ2, λ1 ≥ λ2, U x and V x is the orthogonal matrix generated in the singular value decomposition process. The waveguide signal is a centrally symmetric multi-period sinusoidal signal. When the center of the sliding window is located at the central symmetric point of the waveguide signal, the smaller singular value λ2 tends to 0.
[0063] Step (2a), normalize the first-order singular correlation value sequence ψ1, plot the normalized first-order singular correlation value sequence into a direct wave singular correlation value spectrum, and determine the signal intercepted when the center of the sliding window is located at the peak of the direct wave singular correlation value spectrum during the sliding window processing as the reference signal Y.
[0064] In some embodiments, step (1a) specifically includes the following steps (1a.1) to (1a.3):
[0065] Step (1a.1), perform sliding window processing on the direct wave signal SD to obtain the window signal Y in the sliding window processing process x =(y x,1 ,y x,2 , …, y x,L ), L is the sliding window length.
[0066] Step (1a.2), based on Y x =(y x,1 ,y x,2 , …, y x,L ), construct a binary recursive matrix H with 2 rows x And perform singular value decomposition (SVD decomposition) to obtain:
[0067]
[0068] Then we get the first singular value correlation value ψ 1,x =ρσ x,1 (σ x,1 -σ x,2 );U x and V x is the orthogonal matrix generated in the singular value decomposition process, ρ is the zero point correction coefficient, ρ = a max -|a m |, a max is the maximum value in the time series, a m is the value at the middle of the time series.
[0069] Step (1a.3), based on all window signals Y in the sliding window processing process x The first singular correlation values are calculated respectively, and the first singular correlation value sequence ψ1=(ψ 1,1 , ψ 1,2 ,…,ψ 1,M ), M is the total number of sliding windows for sliding window processing of the direct wave signal SD.
[0070] Step 102: Perform sliding window processing on the pre-processed waveguide signal SP, and calculate the value of the signal X in the i-th window during the sliding window processing. i and the reference signal Y, construct the binary recursive matrix H 1,i , based on the binary recursive matrix H 1,i The window signal X in the sliding window processing process i Perform j times of multi-resolution singular value decomposition to obtain j times of singular correlation values, based on all window signals X in the sliding window processing process iThe j-th singular correlation values obtained by separate calculations are used to construct the j-th singular correlation value sequence.
[0071] In some embodiments, step 102 specifically includes the following steps (1b) - step (3b):
[0072] Step (1b): Perform a sliding window process on the preprocessed guided wave signal SP to obtain the in-window signal X during the sliding window process i =(x i,1 , x i,2 , …, x i,Q ); Q is the signal length of the in-window signal X i , which is the same as the length of the reference signal Y.
[0073] Step (2b): Based on X i =(x i,1 , x i,2 , …, x i,Q ) and the reference signal Y = (y1, y2, …, y Q ), construct a binary recursive matrix
[0074] Step (3b): Based on the binary recursive matrix H 1,i , perform j times of multi-resolution singular value decomposition to obtain the j-th singular correlation value sequence; among them, in the j times of multi-resolution singular value decomposition, the first multi-resolution singular value decomposition decomposes X i into an approximation signal A 1,i and a detail signal D 1,i , and the n-th multi-resolution singular value decomposition decomposes the approximation signal A n-1,i obtained from the (n - 1)-th multi-resolution singular value decomposition into an approximation signal A n,i and a detail signal D n,i ; 1 < n ≤ j, and n is a positive integer, and the value of j is determined according to actual needs.
[0075] Among them, the j times of multi-resolution singular value decomposition is introduced: According to the original signal A0 = (a 0,1 , a 0,2 , …, a 0,N )(that is, each in-window signal X i in the sliding window process in the embodiments of the present application) and the reference signal Y = (y1, y2, …, y N ), construct the following matrix:
[0076]
[0077] Perform singular value decomposition on H ∈ R 2×N to obtain:
[0078] H = UΛVT
[0079] Where U = (u1, u2), U∈R 2×2 , V=(v1;v2;…;v N ), V∈R N×N ,Λ=(diag(σ1,σ2),O),Λ∈R 2×N .
[0080] σ1 is the approximate singular value, corresponding to the larger of the two singular values. σ2 is the detail singular value, corresponding to the smaller of the two singular values. By reconstructing the signal by the approximate and detail singular values, two component signals can be obtained, namely the approximate signal A1 and the detail signal D1. Similarly, the approximate signal A1 can be further constructed into the above matrix form for SVD decomposition to obtain the approximate signal A2 and the detail signal D2. Therefore, by continuously constructing the approximate signal into the above matrix form for decomposition, detail signals and approximate signals at different levels can be obtained, namely:
[0081] A j-1 =A j +D j
[0082]
[0083] Among them, A j-1 is the approximate signal after j-1 decompositions, A j is the approximate signal after j-time decomposition, D j is the detail signal after j times decomposition. By decomposing the approximate signal N times, the original signal A0 can be decomposed into the approximate signal A N and a series of detail signals.
[0084] In some embodiments, the first multi-resolution singular value decomposition process in step (3b) may include the following steps (3b.1) to (3b.3):
[0085] Step (3b.1), divide the recursive matrix H into two parts 1,i Perform multi-resolution singular value decomposition to obtain:
[0086]
[0087] Among them, the singular values include σ 1,i,1 and σ 1,i,2 , σ 1,i,1 ≥σ 1,i,2 ; When the sliding window center is located at the center of the echo signal, the smaller singular value σ 1,i,2 Tends to 0.
[0088] Step (3b.2), expand H1,i to obtain:
[0089]
[0090] U 1,i =(u 1,1,i ,u 1,2,i ), U 1,i ∈R 2×2 , V 1,i =(v 1,1 ,i;v 1,2,i ;…;v 1,Q,i ), V 1,i ∈R Q×Q ;H 1,a,i is an approximate matrix:
[0091]
[0092] H 1,d,i For the detail matrix:
[0093]
[0094] Step (3b.3), based on the approximate matrix and the detail matrix, obtain the approximate signal A 1,i =L 1,a,1 , and detail signal D 1,i =L 1,d,1 , where X i =L 1,a,1 +L 1,d,1 =A 1,i +D 1,i .
[0095] In some embodiments, after performing the first multi-resolution singular value decomposition in step (3b), the following steps may also be included:
[0096] Step (3b.4), approximate signal A 1,i Continue to construct the above matrix form and perform SVD decomposition to obtain the approximate signal A 2,i and detail signal D 2,i , and so on, the approximate signal is continuously constructed into the above matrix form for decomposition, and different levels of detail signals and approximate signals can be obtained until j times of multi-resolution singular value decomposition is completed. The signal after j-1 times of decomposition is:
[0097] A j-1,i =(a j-1,1,i , a j-1,2,i , …, a j-1,Q,i )
[0098] In the jth multi-resolution singular value decomposition, A j-1,i The construction matrix is:
[0099]
[0100] Perform SVD decomposition on it and get:
[0101] H j,i =U j,i Λ j,i V j,i T
[0102] Where U j,i =(u j,1,i ,u j,2,i ), V j,i ∈R 2×2 , V j,i =(v j,1,i ;v j,2,i ;…;v j,Q,i ), V j,i ∈R (Q-1) × (Q-1) , Λ j,i =(diag(σ j,1,i , σ j,2,i ), O), Λ j,i ∈R Q×Q ;
[0103] A j-1,i Decomposed into:
[0104] A j-1,i =A j,i +D j,i
[0105] Get the approximate signal A obtained by the j-th multi-resolution singular value decomposition j,i =(a j,1,i , a j,2,i , …, a j,Q,i ); complete the multi-resolution singular value decomposition j times.
[0106] In some embodiments, in step (3b), based on the binary recursive matrix H 1,i , perform j times of multi-resolution singular value decomposition, and obtain the approximate signal A obtained by the j-th multi-resolution singular value decomposition j,i =(a j,1,i , a j,2,i , …, a j,Q,i ) can also include the following steps (3b.5) to (3b.8):
[0107] Step (3b.5), based on the j-th multi-resolution singular value decomposition, the approximate signal A j,i Construct a bisection recursion matrix:
[0108]
[0109] Step (3b.6), H j,i Perform singular value decomposition and get:
[0110] H j,i =U j,i Λ j,i V j,i T
[0111] Among them, U j,i =(u j,1,i ,u j,2,i ), U j,i ∈R 2×2 , V j,i =(v j,1,i ;v j,2,i ;…;v j,Q,i ), V j,i ∈R Q×Q , Λ j,i =(diag(σ j,1,i , σ j,2,i ), O), Λ j,i ∈R Q×Q .
[0112] Step (3b.7) obtains the subsingular correlation value:
[0113] ψ j,i =ρσ j,1,i (σ j,1,i -σ j,2,i )
[0114] Among them, the approximate singular value σ j,1,i Mainly reflects the signal energy, approximate singular value σ j,1,i and the detail singular value σ j,2,i The difference between them reflects the correlation between the echo signal and the reference signal. When the sliding window center is located at the center of the echo signal, the smaller singular value σ 1,i,2 tends to 0, ρ is the zero point correction coefficient, ρ = a max -|a m |, a max is the maximum value in the time series, a m is the value at the middle of the time series.
[0115] Step (3b.8), based on all window signals X in the sliding window processing process i The j-th singular correlation values are calculated respectively to obtain the j-th singular correlation value sequence:
[0116] ψ j =(ψ 1,1 , ψ 1,2 ,…,ψ 1,K )
[0117] Where K is the total number of sliding windows for sliding window processing of the pre-processed waveguide signal SP, that is, the ψ calculated in all sliding windows i,j , we get the singular value sequence ψ j =(ψ j,1 , ψ j,2 ,…,ψ j,K ).
[0118] Step 103: In the waveguide singular correlation value spectrum corresponding to the jth singular correlation value sequence, determine the time difference between the singular correlation spectrum peaks at the direct wave and the end echo as the acoustic time of the waveguide.
[0119] Specifically, according to the j-th singular correlation value sequence ψ j =(ψ 1,1 , ψ 1,2 ,…,ψ 1,K ) and the time position of the sliding window center, and the waveguide singular correlation value spectrum is plotted. For the echo signal and the direct wave signal with a high correlation with the reference signal, according to the singular value decomposition principle, the approximate singular value σ j,1 Larger, more detailed singular values σ j,2 tends to 0. When the sliding window center is located at the center of the echo or direct wave signal, the singular correlation value spectrum will have a peak here. Therefore, from the singular value correlation spectrum, the time difference between the singular correlation spectrum peaks at the direct wave and the guided wave echo can be extracted as the acoustic time of the guided wave.
[0120] The method for determining the acoustic time of a waveguide based on multi-resolution singular value decomposition provided in the embodiment of the present application proposes a novel binary recursive matrix based on the binary recursive principle, and combines the multi-resolution singular value decomposition to extract a new echo acoustic time positioning feature, namely the singular correlation value, calculate the j-th singular correlation value spectrum of the waveguide signal, and use the singular correlation value spectrum to determine the acoustic time of the waveguide, and extract the time difference between the singular correlation spectrum peaks at the direct wave and the end echo as the acoustic time. Compared with the related art, in a complex waveguide structure, the echo signal-to-noise ratio is low, which makes the echo acoustic time difficult to determine. The present application can determine the acoustic time from the waveguide signal with a low signal-to-noise ratio, and can realize the detection of the acoustic time of the waveguide in a complex structure.
[0121] The technical solution provided in the embodiments of the present application is further illustrated below through specific examples.
[0122] In one embodiment, Figure 2 This is a second flow chart of a method for determining acoustic timing of guided waves based on multi-resolution singular value decomposition provided in an embodiment of the present application, such as Figure 2 As shown, the main steps include:
[0123] Step 1: Preprocess the original waveguide signal to be analyzed. The preprocessing process mainly includes SG filtering method {Savitzy-Golay filter) to remove high-frequency noise and use linear interpolation method to increase the signal sampling rate, so as to obtain the preprocessed waveguide signal.
[0124] Step 2: Extract the direct wave signal from the preprocessed waveguide signal; then intercept the reference signal from the direct wave signal.
[0125] Step 3: Perform sliding window processing on the entire processed waveguide signal, calculate the singular correlation values of the reference signal and the signal in each window, and obtain a singular correlation value sequence composed of a series of singular correlation values.
[0126] Step 4, plot the singular correlation value sequence into a singular correlation value spectrum (i.e., a waveguide singular correlation value spectrum), and use the time difference between the peaks of the singular correlation spectrum at the direct wave and the end echo as the acoustic time. In a complex waveguide structure, the waveguide is affected by multi-modality and noise, and its echo signal has a low signal-to-noise ratio, which makes it difficult to measure the echo acoustic time. The embodiment of the present application can achieve acoustic time measurement in a waveguide signal with a low signal-to-noise ratio.
[0127] For example, the waveguide signal to be analyzed can be loaded first, and the waveguide signal can be preprocessed. SG filtering can be used to reduce high-frequency noise. Linear interpolation can be performed on the filtered signal to increase the signal sampling rate.
[0128] Furthermore, the direct wave signal in the waveguide to be analyzed is extracted as the reference signal. The waveguide signal is a centrally symmetrical multi-period sinusoidal signal. The central symmetric point position of the waveguide signal can be extracted by SVD decomposition. The direct wave signal is extracted from the received signal, and the intercepted signal is subjected to sliding window processing, with a window length of L and a step length of 1, and the signal in each window is subjected to detrending term processing. Let the processed signal in the xth sliding window be Y x =(y x,1 ,y x,2 , …, y x,L ), and construct the matrix H x Perform SVD decomposition:
[0129]
[0130] Compute a singular correlation value:
[0131] ψ 1,x =ρσ x,1 (σ x,1 -σ x,2 )
[0132] Where ρ is the zero point correction coefficient, ρ = a max -|a m |, amax is the maximum value in the time series, a m is the value at the middle position of the time series. Combining the first singular correlation values in all windows, we get the first singular correlation value sequence ψ1=(ψ 1,1 , ψ 1,2 ,…,ψ 1,M ), M is the total number of sliding windows. The singular correlation value sequence is normalized and the normalized singular correlation value spectrum is plotted. The intercepted signal when the sliding window center is located at the singular correlation spectrum peak is selected as the reference signal Y.
[0133] Furthermore, sliding window processing is performed on the entire received signal, with a window length of Q, and steps are performed according to the unit length, and detrending items are processed for the signal in each window. Let the processed signal in the i-th sliding window be X i =(x i,1 , x i,2 , …, x i,Q ), combined with the reference signal Y, construct the matrix H 1,i :
[0134]
[0135] Where U 1,i and V 1,i is the orthogonal matrix generated in the process of multi-resolution singular value decomposition, U 1,i =(u 1,1,i ,u 1,2,i ), U 1,i ∈R 2×2 , V 1,i =(v 1,1,i ;v 1,2,i ;…;v 1,Q,i ), V 1,i ∈R Q×Q . The matrix H 1,i Expand and decompose into the approximate matrix H 1,a,i and the detail matrix H 1,d,i ,get:
[0136]
[0137]
[0138]
[0139] Let L 1,a,1 =(a 1,1,i , a 1,2,i , …, a 1,Q,i ), L 1,a,2 =(a 2,1,i , a 2,2,i , …, a 2,Q,i ), L1,a,1 The subvector corresponding to the first row in the approximation matrix, L 1,a,2 The subvector corresponding to the second row in the approximation matrix. L 1,d,1 =(d 1,1,i , d 1,2,i , …, d 1,Q,i ), L 1,d,2 =(d 2,1,i , d 2,2,i , …, d 2,Q,i ), L 1,d,1 The subvector corresponding to the first row vector in the detail matrix, L d,2 The subvector corresponding to the second row vector in the detail matrix. Let the approximation signal A 1,i and detail signal D 1,i for:
[0140]
[0141] Therefore, the signal X in the window can be i Decomposed into approximate signal A 1,i and detail signal D 1,i :
[0142] X i =L 1,a,1 +L 1,d,1 =A 1,i +D 1,i
[0143] Similarly, the approximate signal A can be 1,i Continue to construct the above matrix form and perform SVD decomposition to obtain the approximate signal A 2,i and detail signal D 2,i Therefore, by continuously constructing the approximate signal into the above matrix form for decomposition, we can obtain detail signals and approximate signals at different levels, namely:
[0144] A j-1,i =A j,i +D j,i
[0145] Let the signal after j times decomposition be A j,i =(a j,1,i , a j,2,i , …, a j,Q,i ), the constructed 2-row matrix is:
[0146]
[0147] Perform SVD decomposition on it and get:
[0148] H j,i =U j,i Λ j,i Vj,i T
[0149] Among them, U j,i =(u j,1,i ,u j,2,i ), U j,i ∈R 2×2 , V j,i =(v j,1,i ;v j,2,i ;…;v j,Q,i ), V j,i ∈R Q×Q , Λ j,i =(diag(σ j,1,i , σ j,2,i ), O), Λ j,i ∈R Q×Q .
[0150] Calculate the j-th singular correlation value:
[0151] ψ j,i =ρσ j,1,i (σ j,1,i -σ j,2,i )
[0152] Where ρ is the zero point correction coefficient, ρ = a max -|a m |, a max is the maximum value in the time series, a m is the value at the middle position of the time series. Combining the j-order singular correlation values in all windows, we get the j-order singular correlation value sequence ψ j =(ψ 1,1 , ψ 1,2 ,…,ψ 1,K ), K is the total number of sliding windows.
[0153] Furthermore, a singular correlation value spectrum is plotted according to the j-th singular correlation value sequence, and the time difference between the singular correlation spectrum peaks at the direct wave and the end echo is extracted as the acoustic time.
[0154] In one embodiment, the cable waveguide is used as the waveguide for the sound to be measured. Figure 3 is a schematic diagram of a measured cable waveguide signal to be analyzed provided in an embodiment of the present application, Figure 4 is a schematic diagram of the preprocessed waveguide signal provided in the embodiment of the present application, which is loaded as follows Figure 3 The measured cable waveguide signal to be analyzed is shown in the figure. The waveguide signal is preprocessed. First, the SG filtering method is used to reduce the high-frequency noise. Then the linear interpolation method is used to increase the signal sampling rate. Figure 4 The preprocessed waveguide signal shown in the figure extracts the direct wave signal in the waveguide to be analyzed as the reference signal.
[0155] Figure 5 The direct wave signal extracted from the pre-processed waveguide signal and the first-order singular correlation value spectrum calculated by the embodiment of the present application are intercepted from the received signal as shown in Figure 5 The direct wave signal is shown in black. The intercepted signal is processed by sliding window, the window length is L, the step length of the sliding window is 1, and the signal in each window is detrended. Let the processed signal in the xth sliding window be Y x =(y x,1 ,y x,2 , …, y x,L ), and construct the matrix H x Perform SVD decomposition:
[0156] The signal in each window is decomposed only once and the singular correlation value is calculated once:
[0157] ψ 1,i =ρσ i,1 (σ i,1 -σ i,2 )
[0158] Where ρ is the zero point correction coefficient, ρ = a max -|a m |, a max is the maximum value in the time series, a m is the value at the middle position of the time series. The singular correlation value sequence ψ1=(ψ 1,1 , ψ 1,2 ,…,ψ 1,M ), M is the total number of sliding windows. The singular correlation value sequence is normalized, and the normalized singular correlation spectrum is as follows: Figure 5 As shown by the red line in the middle. Taking the singular correlation spectrum peak as the sliding window center, the signal in the sliding window is selected as the reference signal Y, as shown in Figure 5 Shown as a thick solid blue line.
[0159] Sliding window processing is performed on the entire received signal, with a window length of Q, and steps are performed according to the unit length, and detrending items are processed for the signal in each window. Let the signal after processing in the i-th sliding window be X i =(x i,1 , x i,2 , …, x i,Q ), combined with the reference signal Y, construct the matrix H 1,i :
[0160]
[0161] Figure 6 is a schematic diagram of the multi-resolution singular value decomposition process provided in the embodiment of the present application, according to Figure 6The multi-resolution singular value decomposition process shown in the figure is used to decompose the signal X i Perform multi-resolution singular value decomposition, and the signal after j decompositions is A j,i =(a j,1,i , a j,2,i , …, a j,Q,i ), construct it as a 2-row matrix:
[0162]
[0163] Perform SVD decomposition on it and get:
[0164]
[0165] Calculate the singular correlation value ψ after j decompositions j,i :
[0166] ψ j,i =ρσ j,1,i (σ j,1,i -σ j,2,i )
[0167] Combine all the ψ calculated in the sliding window i,j , and obtain the j-th singular correlation value sequence ψ j =(ψ j,1 , ψ j,2 ,…,ψ j,K ), where K is the total number of sliding windows on the received signal.
[0168] Figure 7 Schematic diagram of the 5th singular correlation value spectrum of the waveguide signal provided in the embodiment of the present application. According to the 5th singular correlation value sequence, the following is drawn: Figure 7 The 5th singular correlation value spectrum is shown. The peak point of the singular correlation value spectrum of the direct wave is located at 0.1611ms, the peak point of the singular correlation value spectrum of the echo signal is located at 0.4575ms, and the extracted sound time is 0.2964ms.
[0169] Figure 8 is a schematic diagram of the structure of a guided wave acoustic timing measurement device based on multi-resolution singular value decomposition provided in an embodiment of the present application, such as Figure 8 As shown, the device comprises:
[0170] A preprocessing module 801 is used to preprocess the waveguide signal S, extract the direct wave signal SD from the preprocessed waveguide signal SP, and intercept the reference signal Y from the direct wave signal SD;
[0171] The sliding window processing module 802 is used to perform sliding window processing on the pre-processed waveguide signal SP, based on the i-th window signal X i and the reference signal Y, construct the binary recursive matrix H 1,i, based on the binary recursive matrix H 1,i The window signal X in the sliding window processing process i Perform j times of multi-resolution singular value decomposition to obtain j times of singular correlation values, based on all window signals X in the sliding window processing process i The j-order singular correlation values are calculated respectively, and a j-order singular correlation value sequence is constructed;
[0172] The acoustic time determination module 803 is used to determine the time difference between the singular correlation spectrum peaks at the direct wave and the end echo in the waveguide singular correlation value spectrum corresponding to the sub-singular correlation value sequence as the acoustic time of the waveguide.
[0173] In some embodiments, extracting the reference signal Y from the direct wave signal SD comprises:
[0174] The direct wave signal SD is processed by sliding window, and the x-th window signal Y is obtained based on the sliding window processing process. x , construct the binary recursive matrix H x , based on the binary recursive matrix H x Perform singular value decomposition to obtain a singular correlation value; based on all window signals Y in the sliding window processing process x The first-order singular correlation values are calculated respectively, and a first-order singular correlation value sequence ψ1 is constructed;
[0175] The first-order singular correlation value sequence ψ1 is normalized, and the normalized first-order singular correlation value sequence is plotted into a direct wave singular correlation value spectrum. The signal intercepted when the center of the sliding window is located at the peak of the direct wave singular correlation value spectrum during the sliding window processing is determined as the reference signal Y.
[0176] In some embodiments, constructing a singular correlation value sequence ψ1 comprises:
[0177] Perform sliding window processing on the direct wave signal SD to obtain the window signal Y in the sliding window processing process x =(y x,1 ,y x,2 , …, y x,L ), L is the sliding window length;
[0178] Based on Y x =(y x,1 ,y x,2 , …, y x,L ), construct the binary recursive matrix H x And perform singular value decomposition to obtain Get the first singular value correlation value ψ 1,x =ρσ x,1 (σ x,1 -σ x,2 );ρ is the zero point correction coefficient, ρ=a max -|am |, a max is the maximum value in the time series, a m is the value at the middle position of the time series; U x and V x are orthogonal matrices generated during the singular value decomposition process;
[0179] Based on the singular correlation values calculated respectively from all the signals within the windows during the sliding window processing of the direct wave signal SD, obtain the first singular correlation value sequence ψ1 = (ψ 1,1 , ψ 1,2 , …, ψ 1,M ), where M is the total number of sliding windows for the sliding window processing of the direct wave signal SD.
[0180] In some embodiments, the sliding window processing module 802 is specifically configured to:
[0181] Perform sliding window processing on the preprocessed guided wave signal SP to obtain the signal X within the window during the sliding window processing i =(x i,1 , x i,2 , …, x i,Q ); Q is the signal length of the signal X within the window, which is the same as the length of the reference signal Y; i Based on X
[0182] =(x i , x i,1 , …, x i,2 , …, x i,Q ) and the reference signal Y = (y1, y2, …, y Q ), construct a binary recursive matrix
[0183] Based on the binary recursive matrix H 1,i , perform j - order multi - resolution singular value decomposition to obtain the j - order singular correlation value sequence; where, in the j - order multi - resolution singular value decomposition, the first multi - resolution singular value decomposition decomposes X i into an approximation signal A 1,i and a detail signal D 1,i , and the nth multi - resolution singular value decomposition further decomposes the approximation signal A n-1,i obtained from the (n - 1)th multi - resolution singular value decomposition into an approximation signal A n,i and a detail signal D n,i ; 1 < n ≤ j, and n is a positive integer.
[0184] In some embodiments, the process of the first multi - resolution singular value decomposition includes:
[0185] Perform multi - resolution singular value decomposition on the binary recursive matrix H 1,i to obtain Among them, the singular values include σ 1,i,1 and σ 1,i,2 , σ 1,i,1 ≥σ 1,i,2 ; U 1,i and V 1,i is the orthogonal matrix generated in the process of multi-resolution singular value decomposition;
[0186] H 1,i Expand to get U 1,i =(u 1,1,i ,u 1,2,i ), U 1,i ∈R 2×2 , V 1,i =(v 1,1,i ;v 1,2,i ;…;v 1,Q,i ), V 1,i ∈R Q×Q ;H 1,a,i is an approximate matrix, H 1,d,i is the detail matrix,
[0187] Based on the approximation matrix and the detail matrix, the approximate signal A is obtained 1,i =L 1,a,1 , and detail signal D 1,i =L 1,d,1 , where X i =L 1,a,1 +L 1,a,1 =A 1,i +D 1,i .
[0188] In some embodiments, based on the binary recursive matrix H 1,i , perform j times of multi-resolution singular value decomposition to obtain j times of singular correlation value sequences, including:
[0189] Based on the binary recursive matrix H 1,i , perform j times of multi-resolution singular value decomposition, and obtain the approximate signal A obtained by the j-th multi-resolution singular value decomposition j,i =(a j,1,i , a j,2,i , …, a j,Q,i );
[0190] Based on A j,i Construct a binary recursive matrix
[0191] For H j,i Perform singular value decomposition and get H j,i =U j,i Λ j,i V j,iT Among them, U j,i =(u j,1,i ,u j,2,i ), U j,i ∈R 2×2 , V j,i =(v j,1,i ;v j,2,i ;…;v j,Q,i ), V j,i ∈R Q×Q , Λ j,i =(diag(σ j,1,i , σ j,2,i ), O), Λ j,i ∈R Q×Q ;
[0192] Get the j-th singular correlation value ψ j,i =ρσ j,1,i (σ j,1,i -σ j,2,i ), where ρ is the zero point correction coefficient, ρ = a max -|a m |, a max is the maximum value in the time series, a m is the value at the middle position of the time series;
[0193] Based on the j-order singular correlation values calculated from all window signals in the sliding window processing process of the preprocessed waveguide signal SP, the j-order singular correlation value sequence ψ is obtained. j =(ψ 1,1 , ψ 1,2 ,…,ψ 1,K ), K is the total number of sliding windows for performing sliding window processing on the pre-processed waveguide signal SP.
[0194] In some embodiments, the waveguide signal S of the waveguide is preprocessed, including:
[0195] The SG filtering method is used to remove the noise in the waveguide signal S;
[0196] The linear interpolation method is used to process the waveguide signal after noise removal.
[0197] The waveguide acoustic time measurement device based on multi-resolution singular value decomposition provided in the embodiment of the present application proposes a new type of binary recursive matrix based on the binary recursive principle, and combines with multi-resolution singular value decomposition to extract a new echo acoustic time positioning feature - singular correlation value, calculate the j-th singular correlation value spectrum of the waveguide signal, and use the singular correlation value spectrum to measure the waveguide acoustic time, and extract the time difference between the singular correlation spectrum peaks at the direct wave and the end echo as the acoustic time. Compared with the related art, in a complex waveguide structure, the echo signal-to-noise ratio is low, which makes the echo acoustic time difficult to measure. The present application can measure the acoustic time from the waveguide signal with a low signal-to-noise ratio, and can realize the detection of the waveguide acoustic time of complex structures.
[0198] It can be understood that the detailed functional implementation of each of the above-mentioned units / modules can be found in the introduction of the aforementioned method embodiment, and will not be repeated here.
[0199] It should be understood that the above-mentioned device is used to execute the method in the above-mentioned embodiment. The implementation principle and technical effect of the corresponding program module in the device are similar to those described in the above-mentioned method. The working process of the device can refer to the corresponding process in the above-mentioned method, which will not be repeated here.
[0200] Based on the method in the above embodiment, an embodiment of the present application provides an electronic device. The device may include: at least one memory for storing programs and at least one processor for executing the programs stored in the memory. When the program stored in the memory is executed, the processor is used to execute the method described in the above embodiment.
[0201] Fig. 9 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application, such as Fig. 9 As shown, the electronic device may include: a processor (processor) 901, a communication interface (Communications Interface) 902, a memory (memory) 903 and a communication bus 904, wherein the processor 901, the communication interface 902, and the memory 903 communicate with each other through the communication bus 904. The processor 901 can call the software instructions in the memory 903 to execute the method described in the above embodiment.
[0202] In addition, the logic instructions in the above-mentioned memory 903 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application, or the part that contributes to the relevant technology or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present application.
[0203] Based on the method in the above embodiment, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program runs on a processor, the processor executes the method in the above embodiment.
[0204] Based on the method in the above embodiment, an embodiment of the present application provides a computer program product. When the computer program product runs on a processor, the processor executes the method in the above embodiment.
[0205] It is understandable that the processor in the embodiment of the present application may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. The general-purpose processor may be a microprocessor or any conventional processor.
[0206] The method steps in the embodiments of the present application can be implemented by hardware, or by a processor executing software instructions. The software instructions can be composed of corresponding software modules, and the software modules can be stored in random access memory (Random Access Memory, RAM), flash memory, read-only memory (Read-Only Memory, ROM), programmable read-only memory (Programmable ROM, PROM), erasable programmable read-only memory (Erasable PROM, EPROM), electrically erasable programmable read-only memory (Electrically EPROM, EEPROM), registers, hard disks, mobile hard disks, CD-ROMs or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor so that the processor can read information from the storage medium and can write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an ASIC.
[0207] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loading and executing computer program instructions on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted by a computer-readable storage medium. The computer instructions can be transmitted from a website site, a computer, a server or a data center to another website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server, a data center, etc. that contains one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)), etc.
[0208] It should be understood that the various numerical numbers involved in the embodiments of the present application are only used for the convenience of description and are not used to limit the scope of the embodiments of the present application.
[0209] It is easy for those skilled in the art to understand that the above are only preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the scope of protection of the present application.
Claims
1. A method for determining acoustic timing of guided waves based on multi-resolution singular value decomposition, characterized in that: include: Preprocessing the waveguide signal S, extracting the direct wave signal SD from the preprocessed waveguide signal SP, and intercepting the reference signal Y from the direct wave signal SD; The pre-processed waveguide signal SP is subjected to sliding window processing, and based on the i-th window signal X i and the reference signal Y, construct a binary recursive matrix H 1,i , based on the binary recursive matrix H 1,i The window signal X in the sliding window processing process i Perform j times of multi-resolution singular value decomposition to obtain j times of singular correlation values, based on all window signals X in the sliding window processing process i The j-order singular correlation values are calculated respectively, and a j-order singular correlation value sequence is constructed; In the waveguide singular correlation value spectrum corresponding to the j-th singular correlation value sequence, the time difference between the singular correlation spectrum peaks at the direct wave and the end echo is determined as the acoustic time of the waveguide.
2. The method for determining the acoustic timing of guided waves based on multi-resolution singular value decomposition according to claim 1, characterized in that: The step of extracting a reference signal Y from the direct wave signal SD comprises: The direct wave signal SD is subjected to sliding window processing, and based on the x-th window signal Y in the sliding window processing process x , construct the binary recursive matrix H x , based on the binary recursive matrix H x Perform singular value decomposition to obtain a singular correlation value; based on all window signals Y in the sliding window processing process x The first-order singular correlation values are calculated respectively, and a first-order singular correlation value sequence ψ1 is constructed; The first-order singular correlation value sequence ψ1 is normalized, and the normalized first-order singular correlation value sequence is plotted into a direct wave singular correlation value spectrum, and the signal intercepted when the sliding window center is located at the direct wave singular correlation value spectrum peak during the sliding window processing is determined as the reference signal Y.
3. The guided wave acoustic timing determination method based on multi-resolution singular value decomposition according to claim 2, characterized in that: The step of constructing a singular correlation value sequence ψ1 comprises: The direct wave signal SD is subjected to sliding window processing to obtain the window signal Y in the sliding window processing process. x =(y x,1 ,y x,2 ,…,y x,L ), L is the sliding window length; Based on Y x =(y x,1 ,y x,2 ,…,y x,L ), construct the binary recursive matrix H x And perform singular value decomposition to obtain Get the first singular value correlation value ψ 1,x =ρσ x,1 (σ x,1 -σ x,2 );ρ is the zero point correction coefficient, ρ=a max -|a m |, a max is the maximum value in the time series, a m is the value at the middle position of the time series; U x and V x is the orthogonal matrix generated in the singular value decomposition process; Based on all the window signals Y in the sliding window processing process x The first singular correlation values are calculated respectively, and the first singular correlation value sequence ψ1=(ψ 1,1 ,ψ 1,2 ,…,ψ 1,M ), M is the total number of sliding windows for sliding window processing of the direct wave signal SD.
4. The method for determining the acoustic timing of guided waves based on multi-resolution singular value decomposition according to claim 1, characterized in that: The step of constructing a j-time singular correlation value sequence comprises: The pre-processed waveguide signal SP is subjected to sliding window processing to obtain the window signal X in the sliding window processing process. i =(x i,1 ,x i,2 ,…,x i,Q ); Q is the signal X in the window i The signal length is consistent with the length of the reference signal Y; X-based i =(x i,1 ,x i,2 ,…,x i,Q ) and the reference signal Y=(y1,y2,…,y Q ), construct a binary recursive matrix Based on the binary recursive matrix H 1,i , perform j times of multi-resolution singular value decomposition to obtain j times of singular correlation value sequences; wherein, in the j times of multi-resolution singular value decomposition, the first multi-resolution singular value decomposition is to decompose X i Decomposed into approximate signal A 1,i and detail signal D 1,i The nth multi-resolution singular value decomposition is the approximate signal A obtained by the n-1th multi-resolution singular value decomposition n-1,i Further decomposed into approximate signal A n,i and detail signal D n,i ; 1<n≤j, and n is a positive integer.
5. The method for determining guided wave acoustic timing based on multi-resolution singular value decomposition according to claim 4, characterized in that: The process of the first multi-resolution singular value decomposition includes: For the binary recursive matrix H 1,i Perform multi-resolution singular value decomposition to obtain Among them, the singular values include σ 1,i,1 and σ 1,i,2 , σ 1,i,1 ≥σ 1,i,2 ; U 1,i and V 1,i is an orthogonal matrix generated in the process of multi-resolution singular value decomposition; H 1,i Expand to get U 1,i =(u 1,1,i ,u 1,2,i ), U 1,i ∈R 2×2 , V 1,i =(v 1,1,i ;v 1,2,i ;…;v 1,Q,i ), V 1,i ∈R Q×Q ;H 1,a,i is an approximate matrix, H 1,d,i is the detail matrix, Based on the approximation matrix and the detail matrix, the approximation signal A is obtained. 1,i =L 1,a,1 , and the detail signal D 1,i =L 1,d,1 , where X i =L 1,a,1 +L 1,d,1 =A 1,i +D 1,i .
6. The guided wave acoustic timing determination method based on multi-resolution singular value decomposition according to claim 4 or 5, characterized in that: Based on the binary recursive matrix H 1,i , perform j times of multi-resolution singular value decomposition to obtain j times of singular correlation value sequences, including: Based on the binary recursive matrix H 1,i , perform j times of multi-resolution singular value decomposition, and obtain the approximate signal A obtained by the j-th multi-resolution singular value decomposition j,i =(a j,1,i ,a j,2,i ,…,a j,Q,i ); Based on A j,i Construct a binary recursive matrix For H j,i Perform singular value decomposition and get H j,i =U j,i Λ j,i V j,i T Among them, U j,i =(u j,1,i ,u j,2,i ), U j,i ∈R 2×2 , V j,i =(v j,1,i ;v j,2,i ;…;v j,Q,i ), V j,i ∈R Q×Q , Λ j,i =(diag(σ j,1,i ,σ j,2,i ),O),Λ j,i ∈R Q×Q ; Get the j-th singular correlation value ψ j,i =ρσ j,1,i (σ j,1,i -σ j,2,i ), where ρ is the zero point correction coefficient, ρ = a max -|a m |, a max is the maximum value in the time series, a m is the value at the middle position of the time series; Based on the j-order singular correlation values respectively calculated from all the signals in the window during the sliding window processing of the preprocessed waveguide signal SP, the j-order singular correlation value sequence ψ is obtained. j =(ψ 1,1 ,ψ 1,2 ,…,ψ 1,K ), K is the total number of sliding windows for performing sliding window processing on the preprocessed waveguide signal SP.
7. The method for determining the acoustic timing of guided waves based on multi-resolution singular value decomposition according to claim 1, characterized in that: The preprocessing of the waveguide signal S of the waveguide includes: The SG filtering method is used to remove the noise in the waveguide signal S; The linear interpolation method is used to process the waveguide signal after noise removal.
8. A guided wave acoustic timing measurement device based on multi-resolution singular value decomposition, characterized in that: include: A preprocessing module, used to preprocess the waveguide signal S, extract the direct wave signal SD from the preprocessed waveguide signal SP, and intercept the reference signal Y from the direct wave signal SD; The sliding window processing module is used to perform sliding window processing on the pre-processed waveguide signal SP, based on the i-th window signal X in the sliding window processing process. i and the reference signal Y, construct a binary recursive matrix H 1,i , based on the binary recursive matrix H 1,i The window signal X in the sliding window processing process i Perform j times of multi-resolution singular value decomposition to obtain j times of singular correlation values, based on all window signals X in the sliding window processing process i The j-order singular correlation values are calculated respectively, and a j-order singular correlation value sequence is constructed; The acoustic time determination module is used to determine the time difference between the singular correlation spectrum peaks at the direct wave and the end echo in the waveguide singular correlation value spectrum corresponding to the j-th singular correlation value sequence as the acoustic time of the waveguide.
9. An electronic device, characterized in that: include: at least one memory for storing a computer program; At least one processor is used to execute the program stored in the memory. When the program stored in the memory is executed, the processor is used to execute the method according to any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program runs on a processor, the processor is caused to execute the method according to any one of claims 1 to 7.
Citation Information
Patent Citations
SVD-based denoising algorithm for storage tank bottom plate ultrasonic guided wave detection signals
CN106290587A
