Blasting vibration signal denoising method and system

Abstract

This invention discloses a method and system for denoising blasting vibration signals. It decomposes the original noisy blasting vibration signal using CEEMDAN, and classifies each intrinsic mode function (IMF) into effective vibration and noise using a correlation coefficient threshold, reconstructing the primary and secondary IMF components. Further, it applies wavelet packet threshold denoising at multiple decomposition levels to the primary and secondary IMF components, combining these levels to obtain the optimal decomposition levels for the primary and secondary IMF components. Based on these optimal decomposition levels, it denoises the primary and secondary IMF components, reconstructing the denoised primary and secondary IMF components into a high-precision blasting vibration denoised signal. This effectively reduces noise components in the blasting vibration signal while preserving the effective features of the original signal to the maximum extent.

Description

Technical Field

[0001] This invention relates to the field of blasting vibration signal analysis technology, and in particular to a method and system for blasting vibration signal noise reduction. Background Technology

[0002] Blasting vibration signals are typical non-stationary random signals, containing a wealth of crucial information about engineering blasting and directly reflecting the blasting operation. However, measured engineering blasting vibration signals contain significant noise interference, which masks or even eliminates the signal's inherent characteristics. This greatly reduces the reliability of signal analysis results such as vibration feature identification and vibration safety assessment, affecting the accurate judgment of actual blasting projects. Therefore, for analysts, removing unnecessary noise interference from recorded signals is crucial.

[0003] The CEEMDAN algorithm decomposes a non-stationary signal into a series of Intrinsic Mode Functions (IMFs) based on the signal's inherent vibrational characteristics. It then eliminates several IMFs with weak correlations from the original signal, treating them as noise components. While applicable to most partially stationary signals, its judgment of noise components is too absolute. When eliminating weakly correlated IMFs, it inevitably causes the loss of useful signal components, resulting in less than ideal noise reduction results.

[0004] Wavelet packet thresholding denoising algorithm decomposes a signal using a specific wavelet basis function to obtain different wavelet packet coefficients. By selecting an appropriate threshold, wavelet coefficients larger than the threshold are considered to be generated by the signal and should be retained, while those smaller than the threshold are considered to be generated by noise and set to zero, thus achieving the purpose of denoising. However, this also shows that the selection of wavelet basis function and wavelet decomposition level is crucial for different signals; improper selection can lead to significant differences.

[0005] Patent application CN 201910812168.5 describes a noise reduction method for blasting vibration signals based on EMD and VMD. First, the original blasting vibration signal is decomposed using EMD to obtain intrinsic mode functions (EMFs) and their number n. Then, the power spectral density of the original blasting vibration signal and each EMF is calculated. The maximum value and corresponding frequency of the power spectral density of each EMF are identified. The ratio of the power spectral density of the EMF to the power spectral density of the original blasting vibration signal at this frequency is determined. If the ratio is less than 10%, it is considered a noise signal, and the number of noise signals is denoted as j. Finally, the original blasting vibration signal is decomposed using VMD, with K decompositions of number nj. Components with center frequencies below 10Hz and above 200Hz are filtered out, and the remaining signal is reconstructed.

[0006] However, the EMD decomposition method used in this scheme has flaws. In the actual decomposition process, mode aliasing effects occur, which can easily lead to errors in the decomposition results and inaccurate subsequent analysis. Furthermore, it directly filters out components below 10Hz and above 200Hz. This not only makes the determination of the upper limits for low and high frequencies too absolute, but in actual blasting vibration signals, the frequency distribution is significantly correlated with the distance of the monitoring point from the blast source and the setting of blasting parameters. It cannot guarantee that components below 10Hz and above 200Hz are noise signals. Moreover, it directly deletes components with low correlation to the original signal as noise, retaining only those with high correlation as useful components, ignoring the fact that some effective components still exist in the filtered portion. This can easily lead to the loss of effective detail information distributed in the filtered portion of the original signal, affecting the results of subsequent analysis. Summary of the Invention

[0007] To overcome the problems existing in current blasting vibration signal analysis, this invention provides a blasting vibration signal noise reduction method and system, which can remove noise while preserving the effective features of the original signal to the maximum extent.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: to provide a method for noise reduction of blasting vibration signals, which includes the following steps:

[0009] The original noisy blasting vibration signal was decomposed into several intrinsic mode functions and a residual using CEEMDAN.

[0010] The correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal is calculated. Each intrinsic mode function is classified into effective vibration and noise based on the correlation coefficient threshold, and the main intrinsic mode function components and secondary intrinsic mode function components are reconstructed respectively.

[0011] Wavelet packet decomposition at multiple levels is performed on the primary intrinsic mode function (IMF) components and the secondary IMF components respectively. Denoising is then performed by setting thresholds based on fixed thresholding techniques and soft thresholding functions. The denoising results of the primary and secondary IMF components at different decomposition levels are then combined to obtain the primary IMF component decomposition level and the secondary IMF component decomposition level with the best denoising effect.

[0012] Based on the decomposition levels of the primary intrinsic mode function (IMF) components and the secondary IMF components when the noise reduction effect is optimal, wavelet packet thresholding is performed on the primary IMF components and the secondary IMF components respectively to obtain the primary IMF components and the secondary IMF components with minimized noise.

[0013] The primary and secondary intrinsic mode function components that minimize noise are summed with the residuals and reconstructed into a high-precision blasting vibration noise reduction signal.

[0014] Preferably, the step of decomposing the original noisy blasting vibration signal into several intrinsic mode functions and a residual using CEEMDAN specifically includes:

[0015] A series of Gaussian white noises are added to the original noisy blasting vibration signal to generate a series of Gaussian noise blasting vibration signals.

[0016] Find the average value of the upper and lower envelopes of each Gaussian noise-containing blasting vibration signal, and subtract this average value from the Gaussian noise-containing blasting vibration signal to obtain a new sequence. Check whether the new sequence meets the preset decomposition characteristics. If the new sequence does not meet the preset decomposition characteristics, it is used as the original noisy blasting vibration signal for re-decomposition. The new sequence that meets the preset decomposition characteristics is used as the intrinsic mode function obtained under the Gaussian noise-containing blasting vibration signal. The intrinsic mode function components are ensembled and averaged to obtain the intrinsic mode function of the original noisy blasting vibration signal.

[0017] The intrinsic mode function is subtracted from the original noisy blasting vibration signal to obtain the residual sequence. The residual sequence is checked to see if it meets the preset decomposition characteristics. The residual sequence that meets the preset decomposition characteristics is used as the original noisy blasting vibration signal for re-decomposition. The residual sequence that does not meet the preset decomposition characteristics is used as the final residual. Finally, all obtained intrinsic mode functions and the final residual are output.

[0018] Preferably, the preset decomposition features include:

[0019] (1) In the entire dataset, the number of extreme values ​​and the number of zero crossings must be equal to or at most differ by one;

[0020] (2) At any point, the average of the envelope defined by the local maximum and the envelope defined by the local minimum is zero.

[0021] Preferably, the step of calculating the correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal, classifying each intrinsic mode function for effective vibration and noise using a correlation coefficient threshold, and reconstructing the primary intrinsic mode function components and secondary intrinsic mode function components respectively; specifically includes:

[0022] Calculate the correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal. Intrinsic mode functions with correlation coefficients greater than or equal to the correlation coefficient threshold are classified as effective vibration mode function components, and intrinsic mode functions with correlation coefficients less than the correlation coefficient threshold are classified as noise mode function components. Among them, the proportion of effective vibration components in effective vibration mode function components is greater than the proportion of noise components, and the proportion of noise components in noise mode function components is greater than the proportion of effective vibration components.

[0023] The effective vibration modal function components and the noise modal function components are summed and reconstructed respectively to obtain the main modal function components in which the proportion of effective vibration components is greater than that of noise components, and the secondary modal function components in which the proportion of noise components is greater than that of effective vibration components.

[0024] Preferably, the formula for calculating the correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal is as follows:

[0025]

[0026] Where k represents the order of the intrinsic mode function, n is the number of data points in the data sequence, and X i This represents the i-th data point of the original noisy blasting vibration signal. IMFk is the average value of each data point in the original noisy blasting vibration signal. i For the i-th data point of the k-th intrinsic mode function component, is the average value of each data point of the k-th order intrinsic mode function component.

[0027] Preferably, the step involves performing wavelet packet decomposition at multiple decomposition levels on the primary and secondary intrinsic mode function (IMF) components, applying thresholds based on fixed thresholding techniques and soft thresholding methods for denoising, and then arranging and combining the denoising results of the primary and secondary IMF components at different decomposition levels to obtain the decomposition levels of the primary and secondary IMF components that achieve the optimal denoising effect. Specifically, this includes:

[0028] Wavelet packet decomposition of the main intrinsic mode function components at multiple decomposition levels is performed to obtain principal component wavelet packet coefficient sequences at multiple decomposition levels. Denoising processing of the principal component wavelet packet coefficient sequences at multiple decomposition levels is performed by setting thresholds based on fixed threshold technology and soft threshold function method, and the denoising results of the principal component wavelet packet coefficient sequences at multiple decomposition levels are obtained.

[0029] Wavelet packet decomposition of the secondary intrinsic mode function components at multiple decomposition levels is performed to obtain wavelet packet coefficient sequences of the secondary components at multiple decomposition levels. Wavelet packet threshold denoising techniques based on fixed threshold and soft threshold functions are applied to denoise the wavelet packet coefficient sequences of the secondary components at multiple decomposition levels to obtain the denoising results of the wavelet packet coefficient sequences of the secondary components at multiple decomposition levels.

[0030] The denoising results of the main component wavelet packet coefficient sequence and the secondary component wavelet packet coefficient sequence under multiple decomposition levels are arranged and combined, and added with the residual to obtain the set of denoised blasting vibration signals under multiple decomposition levels for the main intrinsic mode function component and the secondary intrinsic mode function component respectively.

[0031] Calculate the signal-to-noise ratio (SNR) of the denoised blasting vibration signal and the original noisy blasting vibration signal under various permutations and combinations in the denoised blasting vibration signal set. The decomposition levels of the primary and secondary intrinsic mode functions (IMFs) corresponding to the maximum SNR are the decomposition levels of the primary and secondary IMFs when the denoising effect is optimal, respectively.

[0032] Preferably, the decomposition levels of the primary and secondary intrinsic mode functions (IMFs) based on the optimal denoising effect are used to perform wavelet packet thresholding denoising on the primary and secondary IMFs respectively, to obtain the primary and secondary IMFs with minimized noise. Specifically, this includes:

[0033] Based on the decomposition levels of the primary and secondary intrinsic mode functions (IMFs) at which the noise reduction effect is optimal, the primary and secondary IMFs are transformed into the wavelet packet domain. The optimal wavelet packet coefficients of the primary IMFs at the decomposition level with the optimal noise reduction effect and the optimal wavelet packet coefficients of the secondary IMFs at the decomposition level with the optimal noise reduction effect are then obtained.

[0034] The optimal wavelet packet coefficients of the principal component and the optimal wavelet packet coefficients of the secondary component are subjected to noise reduction processing based on fixed thresholding and soft thresholding methods, and then inversely transformed to obtain the primary intrinsic mode function components and secondary intrinsic mode function components with minimized noise.

[0035] Preferably, the set of noise-reduced blasting vibration signals for the primary and secondary intrinsic mode function components under multiple decomposition scales is as follows:

[0036]

[0037] Where X′ represents the original noisy blasting vibration signal X reduced to a denoised blasting vibration signal at different decomposition scales; D = [D1, D2, D3, ..., Dm], which is the denoising result of the principal component wavelet packet coefficient sequence at different decomposition scales; S = [S1, S2, S3, ..., Sm]. m ], which is the denoising result of the subcomponent wavelet packet coefficient sequence at different decomposition scales; r is the residual.

[0038] Preferably, the formula for calculating the signal-to-noise ratio of the noise-reduced blasting vibration signal to the original noisy blasting vibration signal under various permutations and combinations in the noise-reduced blasting vibration signal set is as follows:

[0039]

[0040] The present invention also provides a blasting vibration signal noise reduction system, which includes the following functional modules:

[0041] The CEEMDAN decomposition module is used to decompose the original noisy blasting vibration signal into several intrinsic mode function components and a residual using CEEMDAN.

[0042] The signal classification and reconstruction module is used to calculate the correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal. The module classifies each intrinsic mode function into effective vibration and noise based on the correlation coefficient threshold, and reconstructs the main intrinsic mode function components and secondary intrinsic mode function components respectively.

[0043] The decomposition and combination module is used to perform wavelet packet decomposition at multiple decomposition levels on the primary intrinsic mode function (IMF) components and the secondary IMF components, and apply threshold setting based on fixed threshold technology and soft threshold function method for noise reduction. The module arranges and combines the noise reduction results of the primary IMF components and the secondary IMF components at different decomposition levels to obtain the decomposition level of the primary IMF components and the decomposition level of the secondary IMF components with the optimal noise reduction effect.

[0044] The optimal level denoising module is used to perform wavelet packet threshold denoising on the primary intrinsic mode function component decomposition level and the secondary intrinsic mode function component decomposition level when the denoising effect is optimal, respectively, to obtain the primary intrinsic mode function component and the secondary intrinsic mode function component with minimized noise.

[0045] The summation and reconstruction module is used to sum the primary and secondary intrinsic mode function components and residuals that minimize noise, and reconstruct them into a high-precision blasting vibration noise-reduced signal.

[0046] The beneficial effects of this invention are:

[0047] After classifying the primary mode function components based on effective vibration mode function components and the secondary mode function components based on noise mode function components, this invention does not directly delete the secondary mode function components based on noise mode function components. Instead, it reintroduces the analysis of the low-correlation functions of the original signal, considering them to be effective signals with high noise levels. Through effective noise reduction methods, the noise components are eliminated while retaining the effective components. Specifically, the primary and secondary intrinsic mode function components are further decomposed and classified at multiple levels, and different degrees of noise reduction are applied according to the different levels of noise in the classification results. While removing noise, the effective features in the original signal are preserved to the maximum extent.

[0048] The CEEMDAN algorithm used in this invention can effectively eliminate the mode aliasing phenomenon in the EMD decomposition method, ensure the accuracy of the decomposition results, and avoid affecting subsequent processes. Furthermore, this invention does not set upper or lower frequency limits, but rather uses the actual decomposition situation as the standard, thereby preserving effective features to the greatest extent. Attached Figure Description

[0049] Figure 1 This is a flowchart of the blasting vibration signal noise reduction method according to an embodiment of the present invention.

[0050] Figure 2 This is a flowchart illustrating the steps of the blasting vibration signal noise reduction method according to an embodiment of the present invention.

[0051] Figure 3 This is a block diagram of the blasting vibration signal noise reduction system according to an embodiment of the present invention.

[0052] The components in the attached diagram are labeled as follows:

[0053] 10. CEEMDAN decomposition module; 20. Signal classification and reconstruction module; 30. Decomposition and combination module; 40. Optimal level noise reduction module; 50. Summation and reconstruction module. Detailed Implementation

[0054] The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby providing a clearer and more explicit definition of the scope of protection of the present invention.

[0055] This invention provides a method for noise reduction of blasting vibration signals, such as... Figure 1 and Figure 2 As shown, it includes the following steps:

[0056] S1. The original noisy blasting vibration signal is decomposed into several intrinsic mode functions and a residual using CEEMDAN.

[0057] Specifically, the original blasting vibration signal was acquired using a NUBOX-8016 blasting vibration meter to obtain the original noisy blasting vibration signal X. Let k be the number of intrinsic mode function components (IMFs), initialized as k = 1; and determine the intensity (ε) and quantity (n) of Gaussian white noise added.

[0058] A series of Gaussian white noises are added to the original noisy blasting vibration signal to generate a series of Gaussian noise blasting vibration signals, which are represented as follows:

[0059] X i =X+εv,i=1,2,3,...,n (1)

[0060] In equation (1), X i It is the i-th signal to which Gaussian white noise has been added, and εv represents Gaussian white noise with intensity ε.

[0061] Find each blasting vibration signal X containing Gaussian noise i The average value m of the upper and lower envelopes i And from the blasting vibration signal X containing Gaussian noise i Subtract the average value m from the middle i A new sequence a is obtained. i :

[0062] a i =X i -m i ,i=1,2,3,...,n (2)

[0063] Test the new sequence a i Whether it meets the preset decomposition characteristics, the preset decomposition characteristics include:

[0064] (1) In the entire dataset, the number of extreme values ​​and the number of zero crossings must be equal to or at most differ by one;

[0065] (2) At any point, the average of the envelope defined by the local maximum and the envelope defined by the local minimum is zero.

[0066] The new sequence a that does not conform to the preset decomposition characteristics i The original noisy blasting vibration signal is re-decomposed; a new sequence a that conforms to the preset decomposition characteristics is obtained. i As an intrinsic mode function component, IMFk i .

[0067] For the intrinsic mode function components IMFk i The k-th IMF is obtained by averaging the sets as follows:

[0068]

[0069] Where k represents the order of the intrinsic mode function, n is the number of groups of blasting vibration signals containing Gaussian noise, and IMF is the k-th intrinsic mode function, IMFk. i Let be the k-th natural mode function generated from the i-th group of blasting vibration signals containing Gaussian noise.

[0070] Subtracting the average value IMFK from the original noisy blasting vibration signal X yields the residual sequence rk:

[0071] rk=X-IMFk (4)

[0072] Check whether the residual sequence rk conforms to the preset decomposition characteristics; use the residual sequence rk that conforms to the preset decomposition characteristics as the original noisy blasting vibration signal for re-decomposition; use the residual sequence rk that does not conform to the preset decomposition characteristics as the final residual, that is, let the residual r = rk and output all IMFk and residual r.

[0073] At the end of all the above steps, the original noise signal X can be decomposed into k IMFs and a residual r, that is:

[0074]

[0075] Where IMFi is the i-th intrinsic mode function.

[0076] S2. Calculate the correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal. Classify each intrinsic mode function into effective vibration and noise based on the correlation coefficient threshold, and reconstruct the main intrinsic mode function components and secondary intrinsic mode function components respectively.

[0077] Specifically, the correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal is calculated. Intrinsic mode functions with correlation coefficients greater than or equal to the correlation coefficient threshold are classified as effective vibration mode function components, and intrinsic mode function components with correlation coefficients less than the correlation coefficient threshold are classified as noise mode function components. Among them, the proportion of effective vibration components in effective vibration mode function components is greater than the proportion of noise components, and the proportion of noise components in noise mode function components is greater than the proportion of effective vibration components.

[0078] The effective vibration modal function components and the noise modal function components are summed and reconstructed respectively to obtain the main modal function components in which the proportion of effective vibration components is greater than that of noise components, and the secondary modal function components in which the proportion of noise components is greater than that of effective vibration components.

[0079] The formula for calculating the correlation coefficient between each intrinsic mode function component (IMF) and the original noisy blasting vibration signal X is as follows:

[0080]

[0081] In equation (6), k represents the order of the intrinsic mode function, n is the number of data points in the data sequence, and X i This represents the i-th data point of the original noisy blasting vibration signal. IMFk is the average value of each data point in the original noisy blasting vibration signal. i For the i-th data point of the k-th intrinsic mode function component, is the average value of each data point of the k-th order intrinsic mode function component.

[0082] S3. Perform wavelet packet decomposition at multiple decomposition levels on the primary intrinsic mode function (IMF) components and the secondary IMF components respectively, and apply threshold setting based on fixed threshold technology and soft threshold function method for noise reduction. Then, arrange and combine the noise reduction results of the primary IMF components and the secondary IMF components at different decomposition levels to obtain the decomposition level of the primary IMF components and the decomposition level of the secondary IMF components with the best noise reduction effect.

[0083] Specifically, wavelet packet decomposition of the main intrinsic mode function components at multiple levels is performed to obtain wavelet packet coefficient sequences of the principal components at multiple levels.

[0084] A wavelet packet thresholding denoising technique based on fixed thresholding and soft thresholding function is applied to denoise the principal component wavelet packet coefficient sequences of various decomposition levels.

[0085] A threshold is set for wavelet packet thresholding. Wavelet packet coefficients smaller than this threshold are considered noise components and are removed. The threshold function is chosen as the soft threshold function S(t).

[0086]

[0087] In equation (7), λ is a fixed threshold. Where σ is the noise standard deviation and N is the signal length; noise standard deviation Median(Wj) is the median of the wavelet packet multi-resolution decomposition coefficients.

[0088] The denoising results of the principal component wavelet packet coefficient sequence at various scales after wavelet packet thresholding are as follows:

[0089] D=[D1, D2, D3,..., Dm] (8)

[0090] In equation (8), m is the decomposition scale.

[0091] Similarly, wavelet packet decomposition at multiple levels is performed on the secondary intrinsic mode function components to obtain wavelet packet coefficient sequences of various scales. Wavelet packet thresholding denoising techniques based on fixed thresholding and soft thresholding are then applied to denoise these sequences at various scales. The denoising results for the secondary wavelet packet coefficient sequences at various scales are as follows:

[0092] S=[S1, S2, S3,..., Sm] (9)

[0093] In equation (9), m is the decomposition scale.

[0094] The denoising results of the principal component wavelet packet coefficient sequence and the secondary component wavelet packet coefficient sequence at multiple scales are arranged and combined, and added to the residuals to obtain the set X″ of the denoised blasting vibration signals X′ of the principal intrinsic mode function components and the secondary intrinsic mode function components under multiple decomposition scales:

[0095]

[0096] Calculate the signal-to-noise ratio (SNR) of the denoised blasting vibration signal X′ and the original noisy blasting vibration signal X under various permutations and combinations in the denoised blasting vibration signal set. The decomposition scale j1 and the decomposition scale j2 of the primary intrinsic mode function (TEM) components corresponding to the maximum value of the SNR are the decomposition levels of the primary and secondary TEM components when the denoising effect is optimal, respectively.

[0097]

[0098] S4. Based on the decomposition levels of the primary intrinsic mode function components and the secondary intrinsic mode function components when the noise reduction effect is optimal, wavelet packet thresholding is performed on the primary intrinsic mode function components and the secondary intrinsic mode function components respectively to obtain the primary intrinsic mode function components and the secondary intrinsic mode function components with minimized noise.

[0099] Specifically, based on the decomposition levels of the primary and secondary intrinsic mode functions (IMFs) at which the noise reduction effect is optimal, the primary and secondary IMFs are transformed into the wavelet packet domain. The optimal wavelet packet coefficients Wj1 of the primary IMF at its optimal noise reduction decomposition level j1, and the optimal wavelet packet coefficients Wj2 of the secondary IMF at its optimal noise reduction decomposition level j2 are obtained. The specific formulas are as follows:

[0100] Wj1 = W (Main IMF);

[0101] Wj2 = W (Secondary IMF); (12)

[0102] In equation (12), W(·) represents wavelet packet transform.

[0103] Then, the optimal wavelet packet coefficients of the principal component and the optimal wavelet packet coefficients of the secondary component are subjected to noise reduction processing based on fixed threshold technology and soft threshold function method, and inverse transformation is performed to obtain the main intrinsic mode function components and secondary intrinsic mode function components with minimized noise.

[0104] S5. The primary and secondary intrinsic mode function components and residuals that minimize noise are summed and reconstructed into a high-precision blasting vibration noise reduction signal.

[0105] Based on the above-mentioned method for denoising blasting vibration signals, this invention also provides a system for denoising blasting vibration signals, such as... Figure 3 As shown, it includes the following functional modules:

[0106] CEEMDAN decomposition module 10 is used to decompose the original noisy blasting vibration signal into several intrinsic mode functions and a residual using CEEMDAN.

[0107] The signal classification and reconstruction module 20 is used to calculate the correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal, classify each intrinsic mode function into effective vibration and noise based on the correlation coefficient threshold, and reconstruct the main intrinsic mode function components and secondary intrinsic mode function components respectively.

[0108] The decomposition and combination module 30 is used to perform wavelet packet decomposition at multiple decomposition levels on the main intrinsic mode function components and the secondary intrinsic mode function components, and apply threshold setting based on fixed threshold technology and soft threshold function method for noise reduction processing. The module also obtains the decomposition level of the main intrinsic mode function components and the decomposition level of the secondary intrinsic mode function components with the best noise reduction effect by comparing the noise reduction results of the main intrinsic mode function components and the secondary intrinsic mode function components at different decomposition levels.

[0109] The optimal level denoising module 40 is used to perform wavelet packet threshold denoising on the main intrinsic mode function component decomposition level and the secondary intrinsic mode function component decomposition level based on the optimal denoising effect, respectively, to obtain the main intrinsic mode function component and the secondary intrinsic mode function component with minimized noise.

[0110] The summation and reconstruction module 50 is used to sum the primary and secondary intrinsic mode function components and residuals that minimize noise, and reconstruct them into a high-precision blasting vibration noise reduction signal.

[0111] The execution method of the blasting vibration signal noise reduction system described in this embodiment is basically the same as the above-mentioned blasting vibration signal noise reduction method, so it will not be described in detail.

[0112] After classifying the primary mode function components based on effective vibration mode function components and the secondary mode function components based on noise mode function components, this invention does not directly delete the secondary mode function components based on noise mode function components. Instead, it reintroduces the analysis of the low-correlation functions of the original signal, considering them to be effective signals with high noise levels. Through effective noise reduction methods, the noise components are eliminated while retaining the effective components. Specifically, the primary and secondary intrinsic mode function components are further decomposed and classified at multiple levels, and different degrees of noise reduction are applied according to the different levels of noise in the classification results. While removing noise, the effective features in the original signal are preserved to the maximum extent.

[0113] Meanwhile, the CEEMDAN algorithm used in this invention can effectively eliminate the mode aliasing phenomenon in the EMD decomposition method, ensuring the accuracy of the decomposition results and avoiding the impact on subsequent processes. Furthermore, this invention does not set upper or lower frequency limits, but rather uses the actual decomposition situation as the standard, thereby preserving effective features to the greatest extent.

[0114] The above are merely embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for noise reduction of blasting vibration signals, characterized in that, Includes the following steps: The original noisy blasting vibration signal was decomposed into several intrinsic mode functions and a residual using CEEMDAN. The correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal is calculated. Each intrinsic mode function is classified into effective vibration and noise based on the correlation coefficient threshold, and the main intrinsic mode function components and secondary intrinsic mode function components are reconstructed respectively. Wavelet packet decomposition of the main intrinsic mode function components at multiple decomposition levels is performed to obtain principal component wavelet packet coefficient sequences at multiple scales. Wavelet packet thresholding denoising techniques based on fixed thresholding and soft thresholding are applied to denoise the principal component wavelet packet coefficient sequences at multiple decomposition levels, resulting in denoising results of principal component wavelet packet coefficient sequences at multiple decomposition levels. Wavelet packet decomposition of the secondary intrinsic mode function components at multiple decomposition levels is performed to obtain wavelet packet coefficient sequences of the secondary components at multiple scales. Denoising processing of the wavelet packet coefficient sequences of the secondary components at multiple decomposition levels is performed by setting thresholds based on fixed threshold technology and soft threshold function method, and the denoising results of the wavelet packet coefficient sequences of the secondary components at multiple decomposition levels are obtained. The denoising results of the main component wavelet packet coefficient sequence and the secondary component wavelet packet coefficient sequence under multiple decomposition levels are arranged and combined, and added with the residual to obtain the set of denoised blasting vibration signals under multiple decomposition levels for the main intrinsic mode function component and the secondary intrinsic mode function component respectively. Calculate the signal-to-noise ratio (SNR) of the denoised blasting vibration signal and the original noisy blasting vibration signal under various permutations and combinations in the denoised blasting vibration signal set. The decomposition levels of the main intrinsic mode function (NEM) components and the secondary intrinsic mode function (NEM) components corresponding to the maximum SNR are the decomposition levels of the main intrinsic mode function (NEM) components and the secondary intrinsic mode function (NEM) components when the denoising effect is optimal, respectively. Based on the decomposition levels of the primary intrinsic mode function (IMF) components and the secondary IMF components when the noise reduction effect is optimal, wavelet packet thresholding is performed on the primary IMF components and the secondary IMF components respectively to obtain the primary IMF components and the secondary IMF components with minimized noise. The primary and secondary intrinsic mode function components with minimized noise are summed and the residuals are reconstructed into a high-precision blasting vibration noise reduction signal.

2. The method for noise reduction of blasting vibration signals according to claim 1, characterized in that, The process of decomposing the original noisy blasting vibration signal into several intrinsic mode functions and a residual using CEEMDAN specifically includes: A series of Gaussian white noises are added to the original noisy blasting vibration signal to generate a series of Gaussian noise blasting vibration signals. Find the average value of the upper and lower envelopes of each Gaussian noise-containing blasting vibration signal, and subtract this average value from the Gaussian noise-containing blasting vibration signal to obtain a new sequence. Check whether the new sequence meets the preset decomposition characteristics. If the new sequence does not meet the preset decomposition characteristics, use it as the original noisy blasting vibration signal for re-decomposition. Use the new sequence that meets the preset decomposition characteristics as the intrinsic mode function obtained under the Gaussian noise-containing blasting vibration signal. Perform a ensemble average on the obtained series of intrinsic mode functions to obtain the intrinsic mode function of the original noisy blasting vibration signal. The intrinsic mode function is subtracted from the original noisy blasting vibration signal to obtain the residual sequence. The residual sequence is checked to see if it meets the preset decomposition characteristics. The residual sequence that meets the preset decomposition characteristics is used as the original noisy blasting vibration signal for re-decomposition. The residual sequence that does not meet the preset decomposition characteristics is used as the final residual. Finally, all obtained intrinsic mode functions and the final residual are output.

3. The method for noise reduction of blasting vibration signals according to claim 2, characterized in that, The preset decomposition features include: (1) In the entire dataset, the number of extreme values ​​and the number of zero crossings must be equal or differ by at most one; (2) At any point, the average of the envelope defined by the local maximum and the envelope defined by the local minimum is zero.

4. The method for noise reduction of blasting vibration signals according to claim 1, characterized in that, The correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal is calculated. Each intrinsic mode function is classified into effective vibration and noise based on the correlation coefficient threshold, and the main intrinsic mode function components and secondary intrinsic mode function components are reconstructed respectively. Specifically, it includes: Calculate the correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal. Intrinsic mode functions with correlation coefficients greater than or equal to the correlation coefficient threshold are classified as effective vibration mode function components, and intrinsic mode functions with correlation coefficients less than the correlation coefficient threshold are classified as noise mode function components. Among them, the proportion of effective vibration components in effective vibration mode function components is greater than the proportion of noise components, and the proportion of noise components in noise mode function components is greater than the proportion of effective vibration components. The effective vibration modal function components and the noise modal function components are summed and reconstructed respectively to obtain the main modal function components in which the proportion of effective vibration components is greater than that of noise components, and the secondary modal function components in which the proportion of noise components is greater than that of effective vibration components.

5. The method for noise reduction of blasting vibration signals according to claim 1, characterized in that, The formula for calculating the correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal is as follows: Where k represents the order of the intrinsic mode function, and n is the number of data points in the data sequence. This represents the i-th data point of the original noisy blasting vibration signal. This represents the average value of each data point in the original noisy blasting vibration signal. For the i-th data point of the k-th order intrinsic mode function component, The average value of each data point of the k-th order intrinsic mode function component. .

6. The method for noise reduction of blasting vibration signals according to claim 1, characterized in that, The optimal decomposition levels for the primary and secondary intrinsic mode functions (IMFs) are described. Wavelet packet thresholding is then applied to both the primary and secondary IMFs to obtain the primary and secondary IMFs with minimized noise. Specifically, this includes: Based on the decomposition levels of the primary and secondary intrinsic mode functions (IMFs) at which the noise reduction effect is optimal, the primary and secondary IMFs are transformed into the wavelet packet domain. The optimal wavelet packet coefficients of the primary IMFs at the decomposition level with the optimal noise reduction effect and the optimal wavelet packet coefficients of the secondary IMFs at the decomposition level with the optimal noise reduction effect are then obtained. The optimal wavelet packet coefficients of the principal component and the optimal wavelet packet coefficients of the secondary component are subjected to noise reduction processing based on fixed thresholding and soft thresholding methods, and then inversely transformed to obtain the primary intrinsic mode function components and secondary intrinsic mode function components with minimized noise.

7. The method for noise reduction of blasting vibration signals according to claim 1, characterized in that, The sets of denoised blasting vibration signals for the main and secondary intrinsic mode function components under various decomposition levels are as follows: ； Wherein, Xʹ is the original noisy blasting vibration signal X reduced to a denoised blasting vibration signal at different decomposition scales; The denoising results of the main component wavelet packet coefficient sequence at different decomposition scales; , which is the denoising result of the subcomponent wavelet packet coefficient sequence at different decomposition scales; r is the residual.

8. The method for noise reduction of blasting vibration signals according to claim 7, characterized in that, The formula for calculating the signal-to-noise ratio of the denoised blasting vibration signal to the original noisy blasting vibration signal under various permutations and combinations in the denoised blasting vibration signal set is as follows: 。 9. A blasting vibration signal noise reduction system, characterized in that, Includes the following functional modules: The CEEMDAN decomposition module is used to decompose the original noisy blasting vibration signal into several intrinsic mode functions and a residual using CEEMDAN. The signal classification and reconstruction module is used to calculate the correlation coefficient between each intrinsic mode function and the original noisy blasting vibration signal. The module classifies each intrinsic mode function into effective vibration and noise based on the correlation coefficient threshold, and reconstructs the main intrinsic mode function components and secondary intrinsic mode function components respectively. The decomposition and combination module is used to perform wavelet packet decomposition on the main intrinsic mode function components at multiple decomposition levels to obtain the principal component wavelet packet coefficient sequences at multiple scales. Wavelet packet thresholding denoising technology based on fixed thresholding and soft thresholding function is applied to denoise the principal component wavelet packet coefficient sequences at multiple decomposition levels to obtain the denoising results of the principal component wavelet packet coefficient sequences at multiple decomposition levels. Wavelet packet decomposition of the secondary intrinsic mode function components at multiple decomposition levels is performed to obtain wavelet packet coefficient sequences of the secondary components at multiple scales. Denoising processing of the wavelet packet coefficient sequences of the secondary components at multiple decomposition levels is performed by setting thresholds based on fixed threshold technology and soft threshold function method, and the denoising results of the wavelet packet coefficient sequences of the secondary components at multiple decomposition levels are obtained. The denoising results of the main component wavelet packet coefficient sequence and the secondary component wavelet packet coefficient sequence under multiple decomposition levels are arranged and combined, and added with the residual to obtain the set of denoised blasting vibration signals under multiple decomposition levels for the main intrinsic mode function component and the secondary intrinsic mode function component respectively. Calculate the signal-to-noise ratio (SNR) of the denoised blasting vibration signal and the original noisy blasting vibration signal under various permutations and combinations in the denoised blasting vibration signal set. The decomposition levels of the main intrinsic mode function (NEM) components and the secondary intrinsic mode function (NEM) components corresponding to the maximum SNR are the decomposition levels of the main intrinsic mode function (NEM) components and the secondary intrinsic mode function (NEM) components when the denoising effect is optimal, respectively. The optimal level denoising module is used to perform wavelet packet threshold denoising on the primary intrinsic mode function component decomposition level and the secondary intrinsic mode function component decomposition level when the denoising effect is optimal, respectively, to obtain the primary intrinsic mode function component and the secondary intrinsic mode function component with minimized noise. The summation and reconstruction module is used to sum the primary and secondary intrinsic mode function components and residuals that minimize noise, and reconstruct them into a high-precision blasting vibration noise-reduced signal.

Images