Leakage sound signal denoising method based on combination of optimized VMD and improved wavelet threshold

Abstract

The invention discloses a leakage sound signal denoising method based on an optimized VMD combined with an improved wavelet threshold, and aims to solve the problem of low recognition accuracy when an existing noise processing method is used for processing leakage sound signals of a water supply network. According to the method, the noise component and range in the leakage sound signal are analyzed by collecting the leakage sound signal of the actual water supply network; performing parameter optimization on the decomposition modal number K and the penalty factor alpha of the VMD based on an eagle optimization algorithm to obtain optimal parameters, and constructing a variational model by adopting the optimal parameters; decomposing the leakage sound signal by using a variational model to obtain a plurality of intrinsic mode components; screening the plurality of intrinsic mode components by using a correlation coefficient method to obtain high-frequency components and low-frequency components; performing wavelet threshold de-noising processing on the high-frequency component to obtain a de-noised high-frequency component; and reconstructing the low-frequency component and the denoised high-frequency component to obtain a denoised leakage sound signal. The beneficial effects are that the signal-to-noise ratio of denoising processing is improved, and the recognition accuracy is improved.

Description

Technical Field

[0001] The invention relates to a method for denoising a leakage sound signal of a water supply network. Background Art

[0002] The collection of leakage acoustic emission signals will inevitably introduce some noise interference, including environmental noise, noise inside the pipe and noise inside the circuit; the presence of noise will reduce the clarity and accuracy of the leakage acoustic signal, making detection and analysis more difficult, thereby reducing detection efficiency; noise will also interfere with the accurate identification of leakage sounds, thereby affecting the location and repair of the leakage; and the above factors will significantly increase the time and economic cost of pipeline maintenance.

[0003] Existing noise processing methods include Fourier transform, wavelet transform, adaptive filter, spectral subtraction and neural network, but these methods have limitations. Fourier transform cannot directly process non-stationary signals, and has limited effect on noise in frequency segments. The selection of basis functions of wavelet transform has a great influence on the denoising effect. The convergence speed of adaptive filter may be slow, and the performance is unstable when processing non-stationary noise. Spectral subtraction requires accurate noise spectrum estimation, and inaccurate estimation of noise frequency may cause signal distortion. Neural network requires a large amount of labeled data to train the network, and the model is complex and computationally intensive. Therefore, the existing noise processing methods have low recognition accuracy when processing water supply network leakage sound signals, and cannot meet pipeline maintenance requirements. Summary of the invention

[0004] The purpose of the present invention is to solve the problem of low recognition accuracy of existing noise processing methods when processing leakage sound signals of water supply pipe networks, and propose a leakage sound signal denoising method based on optimized VMD combined with improved wavelet threshold.

[0005] The present invention discloses a method for denoising a leakage sound signal based on optimizing VMD and combining improved wavelet threshold, the denoising method comprising:

[0006] Step 1, obtaining the leakage sound signal of the leakage point;

[0007] Step 2: Based on the Goshawk optimization algorithm, the decomposition mode number K and penalty factor α of VMD are optimized to obtain the optimal parameters, and the variational model is constructed using the optimal parameters;

[0008] Step 3: Decompose the leakage sound signal obtained in step 1 by using the variational model constructed in step 2 to obtain multiple intrinsic mode components;

[0009] Step 4: Use the correlation coefficient method to screen the multiple natural modal components obtained in step 3 to screen out high-frequency components and low-frequency components;

[0010] Step 5: Perform wavelet threshold denoising on the high-frequency components screened out in step 4 to obtain denoised high-frequency components;

[0011] Step 6: Reconstruct the low-frequency components screened out in step 4 and the denoised high-frequency components obtained in step 5 by using inverse wavelet transform to obtain the denoised leakage sound signal.

[0012] Furthermore, the specific method for obtaining the leakage sound signal of the leakage point in step 1 is:

[0013] Vibration sensors are used to collect leakage acoustic signals from pipelines near the leakage point. The collection interval is 5 minutes. 72 leakage and non-leakage data are collected from each leakage point during the day and night respectively. A total of 2×2×72=288 data are collected from each leakage point. Data from 7 leakage points are collected, with a total of 7×288=2016 acoustic emission data. The sampling frequency of each data is 50kHz, the sampling time is 0.5 seconds, and the length of each data is 25k.

[0014] Furthermore, in step 2, the decomposition mode number K and penalty factor α of VMD are optimized based on the Goshawk optimization algorithm. The specific process of obtaining the optimal parameters includes: initialization stage, prey identification stage, and pursuit and escape stage;

[0015] The initialization stage refers to: before optimizing the parameters using the iterative algorithm, a certain number of initial solutions are randomly generated, the initial solutions represent the positions of the goshawks, and the initial solutions constitute the search space, and the quality of each initial solution is evaluated by the objective function of the goshawk population;

[0016] The prey identification stage refers to: the goshawk population performs a global search of the search space to select the optimal area where the target function is located;

[0017] The chasing and escaping phases refer to: as the number of iterations increases, the radius of the optimal area will gradually decrease as the number of iterations increases, so that the sum of the bandwidths of each mode is minimized and the optimal parameters are output.

[0018] Furthermore, the position expression of the goshawk population is:

[0019]

[0020] Among them, X represents the position matrix of the goshawk population, X i is the initial solution of the ith individual, x i,j is the value of the jth dimension of the i-th individual, N is the number of members of the goshawk population, and m is the dimension of the problem space.

[0021] Furthermore, the objective function of the goshawk population is expressed as an objective function value vector:

[0022]

[0023] Among them, F(X) is the necklace of the objective function obtained, F i is the objective function value obtained for the ith solution.

[0024] Furthermore, the expression for the goshawk population to perform a global search in the search space to select the optimal region where the objective function is located is:

[0025] P i =X k ,i=1,2,…,N,k=1,2,…,i-1,i+1,…,N

[0026]

[0027]

[0028] Among them, P i is the prey position of the i-th goshawk, F Pi is the target value of the i-th goshawk, is the new position of the ith solution, is the value of the jth dimension of the new position of the ith solution, F i (new,P1) It is the objective function value of the first stage of Goshawk algorithm optimization. I is a constant, and the value of I is 1 or 2.

[0029] Furthermore, the formula for outputting the optimal parameters is:

[0030]

[0031]

[0032]

[0033] Among them, R is the parameter optimization range, t represents the number of current iterations, T is the maximum number of iterations, and F i new,P2 It is the objective function value of the second stage of Goshawk algorithm optimization.

[0034] Furthermore, after the final optimization in step 2, it is obtained that the value range of the decomposition mode number K is: 11~13, and the value range of the penalty factor α is: 525~527.

[0035] Furthermore, the wavelet threshold function expression of wavelet threshold denoising in step 5 is:

[0036]

[0037] Where W j,k is the wavelet coefficient before denoising, is the denoised wavelet coefficient, λ is the wavelet threshold, sign() is the sign function, θ is the smoothness adjustment factor, β is the approaching waveform speed adjustment factor, θ∈[1,2]; β∈(0,1000].

[0038] Furthermore, the inverse wavelet transform formula in step six is:

[0039]

[0040] Among them, WT f (j,k) is the wavelet coefficient, ψ j,k (n) is the wavelet basis function and A is the scale factor.

[0041] Compared with the prior art, the present invention has the following beneficial effects: the denoising method improves the threshold function of wavelet threshold denoising and optimizes the parameters of VMD using an optimization algorithm, thereby improving the denoising effect of wavelet threshold and VMD, and combines the two to denoise the leakage sound signal, thereby improving the signal-to-noise ratio, improving the denoising effect of identifying the leakage sound signal of the water supply network, and improving the accuracy of identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A flow chart of a method for denoising a leakage sound signal based on optimizing VMD and combining improved wavelet threshold as described in the first embodiment;

[0043] Figure 2 This is a schematic diagram of a leakage sound signal denoising method based on optimized VMD combined with improved wavelet threshold described in a specific implementation mode 1;

[0044] Figure 3 The figures are the time domain and frequency domain diagrams of the leakage signal in different time periods in the second specific implementation mode; wherein, (a) is the time domain diagram of the leakage signal at night; (b) is the frequency domain diagram of the leakage signal at night; (c) is the time domain diagram of the leakage signal during the day; (d) is the frequency domain diagram of the leakage signal during the day; (e) is the time domain diagram of the non-leakage signal; (f) is the frequency domain diagram of the non-leakage signal;

[0045] Figure 4 A comparison diagram of the improved threshold value and the soft and hard threshold values ​​in the first specific implementation mode;

[0046] Figure 5 This is a standard signal decomposition diagram of VMD denoising in the first specific implementation mode;

[0047] Figure 6 The denoising effect diagram of the standard signal of VMD denoising in the first specific implementation mode; wherein (a) is the standard signal with noise; (b) is the denoising signal without parameter optimization; (c) is the denoising signal with parameter optimization;

[0048] Figure 7 This is a diagram showing the denoising effect of the standard signal using the improved wavelet threshold denoising in the first specific implementation mode; wherein (a) is the noisy standard signal; (b) is the denoised signal;

[0049] Figure 8 The figure is a standard signal denoising effect diagram of the joint denoising algorithm in the specific implementation mode 1; wherein (a) is the noisy standard signal; (b) is the signal after joint denoising;

[0050] Fig. 9 It is a diagram of the actual leakage sound signal denoising effect of the joint denoising algorithm in the specific implementation method one, wherein (a) is the leakage signal before denoising; (b) is the leakage signal after denoising; (c) is the leakage signal before denoising; and (d) is the denoised leakage signal. DETAILED DESCRIPTION

[0051] Specific implementation method 1. Combination Figures 1 to 9 This embodiment describes a leakage sound signal denoising method based on optimized VMD combined with improved wavelet threshold, the denoising method comprising:

[0052] Step 1, obtaining the leakage sound signal of the leakage point;

[0053] Step 2: Based on the Goshawk optimization algorithm, the decomposition mode number K and penalty factor α of VMD are optimized to obtain the optimal parameters, and the variational model is constructed using the optimal parameters;

[0054] Step 3: Decompose the leakage sound signal obtained in step 1 by using the variational model constructed in step 2 to obtain multiple intrinsic mode components;

[0055] Step 4: Use the correlation coefficient method to screen the multiple natural modal components obtained in step 3 to screen out high-frequency components and low-frequency components;

[0056] Step 5: Perform wavelet threshold denoising on the high-frequency components screened out in step 4 to obtain denoised high-frequency components;

[0057] Step 6: Reconstruct the low-frequency components screened out in step 4 and the denoised high-frequency components obtained in step 5 by using inverse wavelet transform to obtain the denoised leakage sound signal.

[0058] In this embodiment, in step one, the actual water supply network is manually checked at night, and after a leak is found, a vibration sensor is used to collect leakage sound signals from the pipes near the leak. After the collection is completed, the leak is repaired and non-leaking sound signals are collected, and the noise components and ranges in the leakage sound signals and non-leaking sound signals are analyzed, including environmental noise, noise inside the pipe, and noise inside the circuit.

[0059] In this embodiment, the specific method of obtaining the leakage sound signal of the leakage point in step 1 is:

[0060] Vibration sensors are used to collect leakage acoustic signals from pipelines near the leakage point. The collection interval is 5 minutes. 72 leakage and non-leakage data are collected from each leakage point during the day and night respectively. A total of 2×2×72=288 data are collected from each leakage point. Data from 7 leakage points are collected, with a total of 7×288=2016 acoustic emission data. The sampling frequency of each data is 50kHz, the sampling time is 0.5 seconds, and the length of each data is 25k.

[0061] In this embodiment, the leakage and non-leakage acoustic emission signals collected at the same leakage location and different time periods are compared in the time domain and frequency domain; the comparison results are as follows: Figure 3 shown; according to Figure 3 It can be seen that there is a difference in the amplitude of the two. The amplitude of the leakage acoustic emission signal at night is slightly lower than that of the acoustic emission signal collected during the day. By comparing the spectrum of the leakage acoustic emission signal and the non-leakage acoustic emission signal, it is found that the useful frequency band of the leakage signal is mainly concentrated in the range of 1 to 2.5 kHz. For the noise component, the wavelet threshold denoising algorithm is combined with VMD to denoise the leakage acoustic signal, and the denoising algorithm is improved and optimized to improve the signal-to-noise ratio of the leakage acoustic signal.

[0062] VMD decomposition is to convert the decomposition problem of the original signal into a variational problem, and the key lies in the construction of the variational problem;

[0063] Constructing the variational problem:

[0064]

[0065]

[0066] Where K is the number of decomposed modes, δ(t) is the Dirac function, and u k is a set of K modes, ω k is the center frequency of the corresponding mode.

[0067] Solve the variational problem:

[0068] VMD decomposition is actually to solve the optimal solution of the variational problem. By transforming the constrained variational problem into an unconstrained variational problem, taking advantage of the quadratic penalty term and the Lagrange multiplier method, and introducing the augmented Lagrangian function, the problem can be transformed into the following function:

[0069]

[0070] Among them, u k(t) is the kth modal component, α is the penalty factor, λ(t) is the Lagrange multiplier, and f(t) is the signal to be decomposed. The function is solved by fixing the values ​​of the other two variables and updating the value of another variable. First, for all ω ≥ 0, iteratively update Value, the specific mathematical expression is:

[0071]

[0072] Then iteratively update ω k The specific iteration expression is:

[0073]

[0074] Then, for all cases where ω≥0, double lifting is performed, and the expression is:

[0075]

[0076] Finally, repeat the above steps until the constraints are met:

[0077]

[0078] For each modal component after decomposition, in order to filter out useful signals and remove the noise contained in the original signal, the Pearson Correlation Coefficient (PCCs) of each mode is calculated, the threshold is set to filter the modal components, and the filtered modal components are reconstructed to obtain the denoised signal.

[0079] When the decomposition mode number K is too large, false IMF components may appear after VMD algorithm decomposition;

[0080] When its value is too small, VMD may not decompose the IMF components sufficiently and fail to fully extract the features of the original signal.

[0081] The size of the penalty factor affects the bandwidth of the IMF component. When α is too large, the bandwidth of the IMF component is too small, which may cause the characteristics of the original signal to be lost, resulting in the IMF component being unable to describe the characteristics of the original signal.

[0082] When α is too small, the bandwidth of the IMF component is too large, which may cause modal aliasing of the IMF component.

[0083] In order to reduce the situation where the K and α parameters are set too large or too small, an optimization algorithm is used to optimize the variational mode decomposition parameters. By setting a certain threshold range, K and α are optimized to find the optimal parameters, thereby improving the decomposition effect of variational mode decomposition.

[0084] In this embodiment, in step 2, the decomposition mode number K and the penalty factor α of VMD are optimized based on the Goshawk optimization algorithm, and the specific process of obtaining the optimal parameters includes: initialization stage, prey identification stage, and pursuit and escape stage;

[0085] The initialization stage refers to: before optimizing the parameters using the iterative algorithm, a certain number of initial solutions are randomly generated, the initial solutions represent the positions of the goshawks, and the initial solutions constitute the search space, and the quality of each initial solution is evaluated by the objective function of the goshawk population;

[0086] The prey identification stage refers to: the goshawk population performs a global search of the search space to select the optimal area where the target function is located;

[0087] The chasing and escaping phases refer to: as the number of iterations increases, the radius of the optimal area will gradually decrease as the number of iterations increases, so that the sum of the bandwidths of each mode is minimized and the optimal parameters are output.

[0088] In this embodiment, the goshawk optimization (NGO) algorithm simulates the behavior of the goshawk during the hunting process, specifically, it simulates the goshawk's recognition, attack, pursuit and escape behaviors, and adopts an adaptive learning rate gradient descent method. In the goshawk algorithm, each solution is regarded as a "goshawk" and is evaluated and compared according to its fitness value. The solution is continuously updated through an iterative algorithm to find the optimal solution.

[0089] In this embodiment, the position expression of the goshawk population is:

[0090]

[0091] Among them, X represents the position matrix of the goshawk population, X i is the initial solution of the ith individual, x i,j is the value of the jth dimension of the i-th individual, N is the number of members of the goshawk population, and m is the dimension of the problem space.

[0092] In this embodiment, the objective function of the goshawk population is expressed as an objective function value vector:

[0093]

[0094] Among them, F(X) is the necklace of the objective function obtained, F i is the objective function value obtained for the ith solution.

[0095] In this embodiment, F(X) is the necklace of the obtained objective function, F i is the objective function value obtained by the ith solution, the objective function value F iis a key indicator used to evaluate the quality of each solution. This objective function value is used to guide the optimization process to find the optimal parameter combination. i It is determined by the signal reconstruction error and the sum of the modal bandwidths. The signal reconstruction error measures the gap between the decomposed mode and the original signal. Ideally, the modal components and residuals of VMD should reconstruct the original signal as accurately as possible; the sum of the modal bandwidths measures the total bandwidth of all modes in the frequency domain. The bandwidth of each mode reflects the width of its spectrum, and the goal of the sum of the bandwidths is to make it as small as possible to ensure the validity of the mode.

[0096] The expression of the goshawk population performing a global search in the search space to select the optimal area where the objective function is located is:

[0097] P i =X k ,i=1,2,…,N,k=1,2,…,i-1,i+1,…,N

[0098]

[0099]

[0100] Among them, P i is the prey position of the i-th goshawk, F Pi is the target value of the i-th goshawk, is the new position of the ith solution, is the value of the jth dimension of the new position of the ith solution, F i (new,P1) It is the objective function value of the first stage of Goshawk algorithm optimization. I is a constant, and the value of I is 1 or 2.

[0101] In this embodiment, during the chasing and escaping stage, when the goshawk attacks the prey, the prey will avoid the goshawk's pursuit in a certain area, but the goshawk will eventually complete the hunting of the prey after a period of pursuit; by simulating this hunting behavior, the search ability of the optimization algorithm in the local space is improved, and the formula for outputting the optimal parameters is:

[0102]

[0103]

[0104]

[0105] Among them, R is the parameter optimization range, t represents the number of current iterations, T is the maximum number of iterations, and F i new,P2 It is the objective function value of the second stage of Goshawk algorithm optimization.

[0106] In this embodiment, the Doppler noise signal is used to verify the denoising effect of the optimized variational mode decomposition algorithm, and the parameters K and α of the initial variational mode decomposition are set to 10 and 5000;

[0107] The parameters K and α are optimized by NGO algorithm, and the optimization ranges of K and α are [4,15] and [100,8000] respectively.

[0108] In this embodiment, the value range of the decomposition mode number K is 11 to 13, and the value range of the penalty factor α is 525 to 527. The best effect is achieved when the decomposition mode number K is 12 and the penalty factor α is 526.

[0109] In this embodiment, the wavelet threshold function expression of the wavelet threshold denoising in step 5 is:

[0110]

[0111] Where W j,k is the wavelet coefficient before denoising, is the denoised wavelet coefficient, λ is the wavelet threshold, sign() is the sign function, θ is the smoothness adjustment factor, β is the approaching waveform speed adjustment factor, θ∈[1,2]; β∈(0,1000].

[0112] In this embodiment, the decomposed modal components are denoised by the correlation coefficient method, and the noisy standard Doppler signal for verifying the denoising effect of the wavelet threshold denoising algorithm is used for testing. The signal-to-noise ratio (SNR) and root mean square error (RSME) after denoising are 20.275 and 0.0982, respectively. The SNR and RSME before parameter optimization are 24.653 and 0.0593, respectively. The SNR and RSME after parameter optimization are 27.351 and 0.0435, respectively.

[0113] The denoising effect of the wavelet threshold denoising algorithm is greatly affected by the wavelet basis function and the wavelet threshold function. The characteristics of different wavelet basis functions and the denoising effects of different wavelet threshold functions are compared and analyzed.

[0114] The selection of wavelet basis functions usually needs to consider the following properties:

[0115] Orthogonality: Whether the wavelet basis function is orthogonal determines whether the signal can be decomposed into mutually orthogonal subspaces;

[0116] Compact support: The compactly supported wavelet has the characteristic of localization, which can accurately locate the features in the signal. The compactly supported wavelet can usually better localize the signal in time and frequency.

[0117] Symmetry: Symmetrical wavelet basis functions can avoid phase distortion of signals when decomposing and transforming them, and help maintain the phase information of signals;

[0118] Regularity: Regularity is a measure of the reconstruction effect of the wavelet basis function. The wavelet basis function with regularity can effectively reduce errors and distortions in the process of signal reconstruction.

[0119] Vanishing moment order: The vanishing moment order reflects the adaptability of the wavelet basis function to the smoothness of the signal. The larger the vanishing moment order, the better the wavelet basis function can compress the signal and suppress noise when changing the signal.

[0120] Support length: The support length is generally proportional to . The longer the support length, the longer the calculation time of the wavelet transform will be, and boundary problems will occur. The shorter the support length, the more difficult it is to concentrate the signal energy. Generally, a wavelet with a support length between 5 and 9 is selected;

[0121] Compare and analyze the orthogonality, compact support, support length, symmetry, regularity, and vanishing moment order of several common wavelet basis functions;

[0122] Among them, the mortlet wavelet, mexh wavelet and meyer wavelet do not have compact support. When these three wavelets are used to transform the signal, the time-frequency characteristics of the signal cannot be effectively extracted. The symN wavelet has compact support and orthogonality, and has better symmetry than the dbN wavelet, and has better vanishing moment order and support length than the Haar wavelet.

[0123] In addition, considering the influence of the support length of the wavelet basis function on the computational complexity and the data compression and denoising effects of the vanishing moment order, the sym4 wavelet basis function is selected to decompose the leakage acoustic emission signal.

[0124] In addition to the fact that the selection of wavelet basis function will affect the effect of wavelet threshold denoising, the selection of threshold function is also crucial. Only by selecting a suitable threshold function can the signal have good continuity in the reconstruction process and no distortion will occur; both hard threshold and soft threshold will have some signal precision loss and distortion when processing wavelet coefficients. In order to improve this phenomenon, a wavelet threshold function for wavelet threshold denoising in step 5 is proposed; the improved wavelet threshold function can change the smoothness of the curve by adjusting θ, and can change the speed at which the curve approaches the hard threshold waveform by adjusting β. The improved wavelet threshold function retains the advantages of both soft and hard thresholds, avoiding the phenomenon of signal oscillation caused by the discontinuity of the hard threshold in the threshold interval, and also improves the phenomenon of constant deviation of the soft threshold during signal reconstruction.

[0125] In this embodiment, the inverse wavelet transform formula in step six is:

[0126]

[0127] Among them, WT f (j,k) is the wavelet coefficient, ψ j,k (n) is the wavelet basis function and A is the scale factor.

[0128] In order to verify the denoising effect of the improved wavelet threshold function, the Doppler standard signal with Gaussian white noise added is used to verify the denoising effect of different threshold functions.

[0129] The standard signal-to-noise ratio (SNR) and root mean square error (RSME) of the signal after noise addition are 20.275 and 0.0982 respectively.

[0130] The proposed improved wavelet threshold function is used to denoise the standard signal. When θ=1.7,β=22,the signal-to-noise ratio and root mean square error after denoising are calculated to evaluate the denoising effect. The SNR and RSME are 29.612 and 0.0345 respectively.

[0131] In order to retain the time and frequency domain characteristics of the leakage acoustic emission signal as much as possible and improve the recognition accuracy of the leakage signal.

[0132] The wavelet threshold denoising algorithm is combined with the variational mode decomposition algorithm to process the collected noisy signal and remove the noise component. The wavelet threshold denoising algorithm can effectively filter out the high-frequency noise in the leakage acoustic emission signal. The VMD decomposition denoising algorithm has good complex data decomposition accuracy and good noise resistance. The combination of the two methods can perform wavelet threshold denoising on different modal components after VMD decomposition, thereby achieving better denoising effect.

[0133] In order to ensure that the signal noise is effectively filtered out, this embodiment uses the above two improved and optimized algorithms to perform denoising experiments on the leakage acoustic emission signal, uses a standard signal to verify the denoising effect, calculates the signal-to-noise ratio and root mean square error after denoising to evaluate the denoising effect, and uses a joint denoising method to verify the denoising effect of the standard signal. The SNR and RSME after denoising are 30.580 and 0.0299 respectively. It can be seen that the joint denoising effect is better than that of the single denoising algorithm.

[0134] After verifying that the joint denoising algorithm has a better denoising effect, denoising is performed on the collected real leakage sound signal;

[0135] The VMD parameters were optimized by the optimization algorithm, and the optimal K and α were 12 and 526 respectively;

[0136] This parameter is used in conjunction with the improved wavelet threshold to denoise the leakage acoustic emission signal, ultimately achieving denoising of the leakage acoustic signal, thereby improving the signal-to-noise ratio of the leakage acoustic signal.

[0137] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A leakage sound signal denoising method based on optimized VMD and improved wavelet threshold. It is characterized in that The denoising method comprises: Step 1, obtaining the leakage sound signal of the leakage point; Step 2: Based on the Goshawk optimization algorithm, the decomposition mode number K and penalty factor α of VMD are optimized to obtain the optimal parameters, and the variational model is constructed using the optimal parameters; Step 3: Decompose the leakage sound signal obtained in step 1 by using the variational model constructed in step 2 to obtain multiple intrinsic mode components; Step 4: Use the correlation coefficient method to screen the multiple natural modal components obtained in step 3 to screen out high-frequency components and low-frequency components; Step 5: Perform wavelet threshold denoising on the high-frequency components screened out in step 4 to obtain denoised high-frequency components; Step 6: Reconstruct the low-frequency components screened out in step 4 and the denoised high-frequency components obtained in step 5 by using inverse wavelet transform to obtain the denoised leakage sound signal.

2. According to claim 1, a method for denoising a leakage sound signal based on optimizing VMD and combining improved wavelet threshold, It is characterized in that The specific method for obtaining the leakage sound signal of the leakage point in step 1 is: Vibration sensors are used to collect leakage acoustic signals from pipelines near the leakage point. The collection interval is 5 minutes. 72 leakage and non-leakage data are collected from each leakage point during the day and night respectively. A total of 2×2×72=288 data are collected from each leakage point. Data from 7 leakage points are collected, with a total of 7×288=2016 acoustic emission data. The sampling frequency of each data is 50kHz, the sampling time is 0.5 seconds, and the length of each data is 25k.

3. According to claim 1, a method for denoising a leakage sound signal based on optimizing VMD and combining improved wavelet threshold, It is characterized in that In step 2, the decomposition mode number K and penalty factor α of VMD are optimized based on the Goshawk optimization algorithm. The specific process of obtaining the optimal parameters includes: initialization stage, prey identification stage, and pursuit and escape stage; The initialization stage refers to: before optimizing the parameters using the iterative algorithm, a certain number of initial solutions are randomly generated, the initial solutions represent the positions of the goshawks, and the initial solutions constitute the search space, and the quality of each initial solution is evaluated by the objective function of the goshawk population; The prey identification stage refers to: the goshawk population performs a global search of the search space to select the optimal area where the target function is located; The chasing and escaping phases refer to: as the number of iterations increases, the radius of the optimal area will gradually decrease as the number of iterations increases, so that the sum of the bandwidths of each mode is minimized and the optimal parameters are output.

4. According to claim 3, a method for denoising a leakage sound signal based on optimizing VMD and combining improved wavelet threshold, It is characterized in that The position expression of the goshawk population is: Among them, X represents the position matrix of the goshawk population, X i is the initial solution of the ith individual, x i,j is the value of the jth dimension of the i-th individual, N is the number of members of the goshawk population, and m is the dimension of the problem space.

5. According to claim 4, a method for denoising a leakage sound signal based on optimizing VMD and combining improved wavelet threshold, It is characterized in that The objective function of the goshawk population is expressed as an objective function value vector: Among them, F(X) is the necklace of the objective function obtained, F i is the objective function value obtained for the ith solution.

6. The method for denoising a leakage sound signal based on optimizing VMD and combining improved wavelet threshold according to claim 5, It is characterized in that The expression of the goshawk population performing a global search in the search space to select the optimal area where the objective function is located is: P i =X k ,i=1,2,…,N,k=1,2,…,i-1,i+1,…,N r∈[0,1] Among them, P i is the prey position of the i-th goshawk, F Pi is the target value of the i-th goshawk, is the new position of the ith solution, is the value of the jth dimension of the new position of the ith solution, F i (new,P1) It is the objective function value of the first stage of Goshawk algorithm optimization. I is a constant, and the value of I is 1 or 2.

7. The method for denoising a leakage sound signal based on optimizing VMD and combining improved wavelet threshold according to claim 6, It is characterized in that The formula for outputting the optimal parameters is: Among them, R is the parameter optimization range, t represents the number of current iterations, T is the maximum number of iterations, and F i new,P2 It is the objective function value of the second stage of Goshawk algorithm optimization.

8. The method for denoising a leakage sound signal based on optimizing VMD and combining improved wavelet threshold according to claim 7, It is characterized in that After the final optimization in step 2, the value range of the decomposition mode number K is: 11~13, and the value range of the penalty factor α is: 525~527.

9. The method for denoising a leakage sound signal based on optimizing VMD and combining improved wavelet threshold according to claim 1, It is characterized in that The wavelet threshold function expression of wavelet threshold denoising in step 5 is: Where W j,k is the wavelet coefficient before denoising, is the denoised wavelet coefficient, λ is the wavelet threshold, sign() is the sign function, θ is the smoothness adjustment factor, β is the approaching waveform speed adjustment factor, θ∈[1,2]; β∈(0,1000]。 10. The method for denoising a leakage sound signal based on optimizing VMD and combining improved wavelet threshold according to claim 1, It is characterized in that The inverse wavelet transform formula in step six is: Among them, WT f (j,k) is the wavelet coefficient, ψ j,k (n) is the wavelet basis function and A is the scale factor.

Images