Blasting vibration signal noise reduction method, computer equipment and medium

Abstract

The invention relates to the technical field of tunnel construction, in particular to a blasting vibration signal noise reduction method, computer equipment and a medium, and the noise reduction method comprises the steps: carrying out the adaptive optimization of key parameters of a VMD through a frost ice optimization algorithm, and obtaining an optimized VMD; using the optimized VMD to decompose the actually measured signal of the tunnel blasting vibration to obtain a plurality of intrinsic mode components; calculating the variance contribution rate of each intrinsic mode component, and removing the components of which the variance contribution rates do not exceed 1%; using a wavelet threshold to carry out noise reduction on the residual intrinsic mode components; and reconstructing the intrinsic mode component after noise reduction to obtain a tunnel blasting vibration signal after noise reduction. According to the method, the VMD parameters are optimized through the frost ice optimization algorithm, the tunnel blasting vibration signals are subjected to noise reduction in combination with the wavelet threshold value, the good signal noise reduction effect is achieved, meanwhile, the modal aliasing phenomenon during signal decomposition is effectively restrained, and the problem that key parameters need to be valued depending on experience in the VMD is solved.

Description

Technical Field

[0001] The invention relates to the technical field of tunnel construction, and in particular to a blasting vibration signal noise reduction method, computer equipment and medium. Background Art

[0002] It is difficult to avoid the problem of high-speed rail being close to existing building structures when planning new lines. Especially in mountainous and hilly areas, railway tunnels are often built to shorten the travel distance and improve transportation capacity, so it is easy for tunnels to be close to existing building structures. At present, most tunnel construction still adopts the drilling and blasting construction method. The vibration generated by the blasting process will inevitably have an adverse effect on the existing structures nearby. Monitoring and analyzing the structural vibration caused by blasting is crucial to the assessment of structural safety. However, the nonlinear characteristics of tunnel blasting vibration signals and the interference of complex factors in the construction environment make it difficult to extract effective signal components from the actual signal containing noise, which has a great impact on the subsequent signal analysis.

[0003] Common signal denoising methods include Fourier transform, wavelet transform, wavelet packet transform, empirical mode decomposition (EMD) and variational mode decomposition (VMD). Fourier transform is a mathematical transformation method that converts signals from time domain to frequency domain, which can reveal the frequency components of signals. Although it is relatively simple and computationally efficient, it is limited in time resolution and has specific requirements for signal length. Wavelet transform can capture the local characteristics of signals, especially non-stationary signals, but the effect of its analysis depends largely on the selection of wavelet functions, which requires a high level of prior knowledge from the operator, and the frequency resolution of wavelet analysis in the high-frequency part is limited. Wavelet packet analysis can more comprehensively analyze the spectral components of signals, especially in the high-frequency part, but its computational complexity and the difficulty of interpreting the results also increase accordingly. Although EMD is widely used, it has problems of mode aliasing and pseudo-oscillation of boundaries. Some scholars have improved the EMD method and proposed the ensemble empirical mode decomposition method (EEMD) and the complete population empirical mode decomposition method (CEEMD). The IMF components obtained by these two methods are clearer, but the computational efficiency is relatively low. Moreover, although the noise reduction capability of the improved EMD method has been improved, the modal aliasing phenomenon has not been truly solved.

[0004] In recent years, the variational mode decomposition algorithm (VMD) has emerged. This algorithm can effectively separate the IMF components and divide the signal in the frequency domain, overcoming the problems of modal aliasing and endpoint effects of EMD. However, the VMD method usually requires experience to determine the penalty parameters and the number of modal decompositions, which increases the difficulty of calculation and affects the accuracy of the results.

[0005] In summary, there is an urgent need for a method that can adaptively optimize the penalty parameter and the number of modal decompositions to solve the problems existing in the prior art. Summary of the invention

[0006] The present invention aims to provide a method for reducing noise of blasting vibration signals, a computer device and a medium, so as to solve the technical problem that the nonlinear characteristics of tunnel blasting vibration signals and the interference of complex factors of the construction environment in the prior art make it difficult to extract effective signal components from actual signals containing noise, which has a great impact on subsequent signal analysis. The specific technical solution is as follows: The present invention provides a method for reducing noise of blasting vibration signals, comprising the following steps: S1, collecting the measured signal of tunnel blasting vibration containing noise; S2. Use the frost optimization algorithm to adaptively optimize the parameters of VMD to obtain the optimized VMD. The VMD parameters include the penalty factor and the number of decomposition levels ; S3, using the optimized VMD obtained in S2 to decompose the measured signal of tunnel blasting vibration, we get The intrinsic modal components; S4. Calculation The variance contribution rate of the intrinsic modal components is calculated, and the components with variance contribution rate not exceeding 1% are removed, and the remaining intrinsic modal components are retained; S5, using the wavelet threshold to reduce the noise of the retained intrinsic mode components to obtain the denoised IMF components; S6. Reconstruct the denoised IMF components obtained in S5 to obtain the tunnel blasting vibration signal after denoising.

[0007] A further improvement of the blasting vibration signal noise reduction method of the present invention is that S2 specifically includes the following steps: S201, first determine the number of individuals in the population in the RIME algorithm , maximum number of iterations And VMD parameters, then assign VMD parameters to each population individual and normalize them to the [0,1] interval, and then randomly generate the initial population; S202, performing VMD decomposition on the measured signal of tunnel blasting vibration containing noise according to the VMD parameters corresponding to each individual; S203, simulating the random movement of soft frost particles through a soft frost search strategy; and simulating the rapid growth and crossover behavior of hard frost particles in the same direction through a hard frost puncture mechanism; S204, calculate and compare the fitness values ​​of the particles before and after the update. If the fitness value of the particle after the update is better than that before the update, the individual is replaced, otherwise, it remains unchanged; after each population update, the current optimal fitness value and the corresponding parameter combination are recorded; as the number of iterations increases, the population continues to evolve in a better direction, and finally outputs the penalty factor for the decomposition of the measured signal of the tunnel blasting vibration containing noise and the number of decomposition levels The optimal combination of is obtained to obtain the optimized VMD.

[0008] The further improvement of the blasting vibration signal denoising method of the present invention is that after VMD decomposition of the measured signal of the tunnel blasting vibration containing noise is performed according to the VMD parameters corresponding to each individual, the IMF component is obtained, and the minimum envelope entropy of the IMF component after VMD decomposition is selected as the fitness function to calculate the fitness value. The expression of the envelope entropy is as follows: 1); 2); In the formula, is the envelope entropy; is the envelope value of the signal; is the normalized envelope value, i.e. The ratio of the amplitude of a frequency component to the total amplitude; is the total number of frequency components of the signal after Hirt transform.

[0009] A further improvement of the blasting vibration signal denoising method of the present invention is that, when simulating the random motion of soft frost particles by the soft frost search strategy, the new position of the updated particles is calculated by the following formula: 3); Where: is the new position of the particle after the update; It is a frost population The best frost agent Particles; parameters is a random number in the range (-1,1), parameter is a random number in the range (0,1); It is an environmental factor; is the adhesion; It is The first of the frost The upper bound of the escape space of particles; It is The first of the frost The lower bound of the escape space of particles; is the adhesion coefficient; is the direction change factor during the iteration process, simulating the change of wind direction; When simulating the rapid growth and crossover behavior of frost particles in the same direction through the frost penetration mechanism, the new position of the updated particle is calculated by the following formula: 4); In the formula, It is The normalized value of the fitness value of each frost agent; is a random number in the range (-1,1).

[0010] A further improvement of the blasting vibration signal noise reduction method of the present invention is that S3 specifically includes the following steps: S301, the decomposed IMF components are expressed as: 5); Where: For the IMF components; For the The amplitude of the IMF component; For the The instantaneous frequency of the IMF components; For time; S302, constructing an optimization problem using a variational method; S303, introduce penalty factor , and the alternating direction multiplier method is used to iteratively solve the optimization problem obtained in S302 to obtain the corresponding modal components and its center frequency ; S304, in each iteration, first fix the center frequency and optimize the mode of the IMF, then fix the mode of the IMF and optimize the center frequency; continue to iterate in this alternating manner, and the solution change between the current and subsequent iterations is less than , it is judged to have converged; after iteration, the modal component solution that achieves the optimization goal is the desired inherent modal component. Each inherent modal component represents a main frequency component in the measured signal of the tunnel blasting vibration containing noise, and is output The intrinsic mode components.

[0011] The further improvement of the blasting vibration signal noise reduction method of the present invention is that the specific steps of S302 are as follows: Combining the classical Wiener filter, Hilbert-Huang transform and spectrum modulation, the expression of the optimization problem constructed by the variational method is obtained as follows: 6); Where: is the decomposed IMF component signal; is the center frequency of the corresponding signal; is the impulse function, and its expression is ; is the single side frequency corresponding to the IMF component; is the original signal before decomposition; For Find the partial score, that is, In To conduct partial division; is frequency shift; Then by introducing the penalty factor and Lagrange multipliers as the constraints for signal reconstruction, the expression of the optimization problem constructed by the variational method becomes: 7); Where: is the Lagrange multiplier; is the penalty factor; is the Lagrangian function; represents the inner product operation; No. The modal components in the iteration and center frequency The expression is as follows: 8); 9); Where: for Frequency domain expression after Fourier transformation; For the In the iteration Frequency domain expressions of the natural mode components; For the The Lagrange multiplier in the iteration; For the In the iteration The center frequency of the modal components; Indicates the number of iterations; is the frequency.

[0012] A further improvement of the blasting vibration signal noise reduction method of the present invention is that S5 specifically includes the following steps: S501, determining the wavelet basis function and the number of decomposition layers, and performing wavelet decomposition on the intrinsic mode components retained in S4 accordingly; S502, determining a threshold function and a threshold selection principle, specifically: determining a wavelet threshold by a fixed threshold criterion, and performing soft threshold processing on the wavelet coefficients according to the determined threshold; S503 , using the wavelet coefficients after threshold processing, reconstruct the signal through inverse wavelet transform to obtain the denoised IMF component.

[0013] A further improvement of the blasting vibration signal noise reduction method of the present invention is that the tunnel blasting vibration signal after noise reduction is The expression is as follows: 12); In the formula, For the processed modal components; is the number of IMF components after screening.

[0014] The present invention also provides a computer device, comprising a memory and a processor; The memory is used to store a computer program executable on the processor; The processor is used to implement the steps of the blasting vibration signal noise reduction method as described above when executing the computer program.

[0015] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the blasting vibration signal noise reduction method as described above are implemented.

[0016] The application of the technical solution of the present invention has the following beneficial effects: The blasting vibration signal denoising method of the present invention adaptively optimizes VMD parameters through a frost optimization algorithm, and denoises the tunnel blasting vibration signal in combination with a wavelet threshold, thereby solving the technical problem in the prior art that the nonlinear characteristics of the tunnel blasting vibration signal and the complex factors of the construction environment interfere with the extraction of effective signal components from the actual signal containing noise, which has a great impact on the subsequent signal analysis; the denoising method can effectively suppress the modal aliasing phenomenon during signal decomposition, and solves the problem that the key parameter values ​​usually need to be determined based on experience in the variational mode decomposition algorithm.

[0017] In addition to the above-described purposes, features and advantages, the present invention has other purposes, features and advantages. The present invention will be further described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings constituting a part of this application are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings: Figure 1 is a flow chart of the blasting vibration signal noise reduction method of the present invention; Figure 2is a vibration velocity diagram of a measured signal S1(t) of an embodiment of the blasting vibration signal noise reduction method of the present invention; Figure 3 is a spectrum diagram of a measured signal S1(t) of an embodiment of the method for reducing noise of blasting vibration signals of the present invention; Figure 4 is a vibration velocity diagram of a measured signal S2(t) of an embodiment of the blasting vibration signal noise reduction method of the present invention; Figure 5 is a spectrum diagram of a measured signal S2(t) of an embodiment of the method for denoising a blasting vibration signal of the present invention; Figure 6 is a vibration velocity diagram of a measured signal S3(t) of an embodiment of the blasting vibration signal noise reduction method of the present invention; Figure 7 is a spectrum diagram of a measured signal S3(t) of an embodiment of the method for reducing noise of a blasting vibration signal of the present invention; Figure 8 is a fitness function variation diagram of the measured signal S1(t) of the embodiment of the blasting vibration signal noise reduction method of the present invention; Fig. 9 is a fitness function variation diagram of the measured signal S2(t) of the embodiment of the blasting vibration signal noise reduction method of the present invention; Fig.10 is a fitness function variation diagram of the measured signal S3(t) of the embodiment of the blasting vibration signal noise reduction method of the present invention; Fig.11 is a VMD decomposition result diagram of the measured signal S1(t) of the embodiment of the blasting vibration signal noise reduction method of the present invention; Fig.12 is a spectrum diagram of the measured signal S1(t) corresponding to the VMD decomposition result of the embodiment of the blasting vibration signal denoising method of the present invention; Fig.13 is a VMD decomposition result diagram of the measured signal S2(t) of the embodiment of the blasting vibration signal noise reduction method of the present invention; Fig.14 is a spectrum diagram of the measured signal S2(t) corresponding to the VMD decomposition result of the embodiment of the blasting vibration signal denoising method of the present invention; Fig.15 is a VMD decomposition result diagram of the measured signal S3(t) of the embodiment of the blasting vibration signal noise reduction method of the present invention; Fig.16 is a spectrum diagram of the measured signal S3(t) corresponding to the VMD decomposition result of the embodiment of the blasting vibration signal noise reduction method of the present invention; Fig.17 is a vibration velocity diagram of a reconstructed signal S1(t) of an embodiment of the method for denoising a blasting vibration signal of the present invention; Fig.18 is a spectrum diagram of a reconstructed signal S1(t) of an embodiment of the method for denoising a blasting vibration signal of the present invention; Fig.19 is a vibration velocity diagram of a reconstructed signal S2(t) of an embodiment of the method for denoising a blasting vibration signal of the present invention; Fig. 20 is a spectrum diagram of a reconstructed signal S2(t) of an embodiment of the method for denoising a blasting vibration signal of the present invention; Fig.21 is a vibration velocity diagram of a reconstructed signal S3(t) of an embodiment of the method for denoising a blasting vibration signal of the present invention; Fig. 22 is a spectrum diagram of a reconstructed signal S3(t) of an embodiment of the method for denoising a blasting vibration signal of the present invention; Fig.23 It is the spectrum diagram of the measured signal S1(t) after denoising the blasting vibration signal using different denoising methods; Fig.24 It is the spectrum diagram of the measured signal S2(t) after denoising the blasting vibration signal using different denoising methods; Fig.25 It is the spectrum diagram of the measured signal S3(t) after denoising the blasting vibration signal using different denoising methods. DETAILED DESCRIPTION

[0019] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0020] See also Figure 1 to Figure 25 As shown in the figure, a method for denoising blasting vibration signals is firstly to use the frost ice optimization algorithm (RIME) to modify the penalty factor of VMD. and the number of decomposition levels Adaptive optimization is performed, and then the original signal of the monitored tunnel blasting vibration is decomposed using the VMD with optimized parameters, the variance contribution rate of each modal component is calculated, the components with smaller variance contribution rate are eliminated, and the remaining components are processed using the wavelet threshold, and finally the processed components are reconstructed into the real signal of the blasting vibration after noise reduction. The specific steps include the following: S1, collecting the measured signal of tunnel blasting vibration containing noise; S2. Penalty factors for key parameters of VMD using the frost optimization algorithm and the number of decomposition levels Perform adaptive optimization to obtain the optimized VMD; S3, using the optimized VMD obtained in S2 to decompose the measured signal of tunnel blasting vibration, we get The intrinsic modal components; S4. In the measured signal of tunnel blasting vibration, the component with variance contribution rate less than 1% can be considered as high-frequency noise, and the remaining components can show most of the characteristics of the real signal of blasting vibration with less information loss; therefore, the calculation The variance contribution rate of the intrinsic modal components is calculated, and the components with variance contribution rate not exceeding 1% are removed, and the remaining intrinsic modal components are retained; S5, using the wavelet threshold to reduce the noise of the retained intrinsic mode components to obtain the denoised IMF components; S6. Reconstruct the denoised IMF components obtained in S5 to obtain the tunnel blasting vibration signal after denoising.

[0021] Figure 1 This is a flow chart of the present invention, and the specific process is: first determine the fitness function and input the tunnel blasting vibration noisy signal, then optimize the VMD parameters through the RIME algorithm, and then bring in the optimized parameters for VMD decomposition to determine whether the variance contribution rate is >1%. If not, remove it; if so, select the wavelet basis function and the number of decomposition layers according to the signal-to-noise ratio index, then use the wavelet threshold to reduce noise, and finally reconstruct the signal to end the process.

[0022] Preferably, S2 specifically includes the following steps: S201, first determine the number of individuals in the population in the RIME algorithm , maximum number of iterations And VMD parameters, then assign these two VMD parameters to each population individual and normalize them to the [0,1] interval, and then randomly generate the initial population. The VMD parameters include the penalty factor and the number of decomposition levels The value range of S202, performing VMD decomposition on the measured signal of tunnel blasting vibration containing noise according to the VMD parameters corresponding to each individual; S203, simulating the random movement of soft frost particles through the soft frost search strategy to explore the search space and avoid falling into the local optimal solution; then simulating the rapid growth and crossover behavior of hard frost particles in the same direction through the hard frost puncture mechanism to improve the convergence speed of VMD parameter optimization and the ability to jump out of the local optimal solution; S204, calculate and compare the fitness values ​​of the particles before and after the update. If the fitness value of the particle after the update is better than that before the update, the individual is replaced, otherwise, it remains unchanged; after each population update, the current optimal fitness value and the corresponding parameter combination are recorded; as the number of iterations increases, the population continues to evolve in a better direction, and finally outputs the penalty factor for the decomposition of the measured signal of the tunnel blasting vibration containing noise and the number of decomposition levels The optimal combination of is obtained to obtain the optimized VMD.

[0023] Preferably, after VMD decomposition of the measured signal of tunnel blasting vibration containing noise is performed according to the VMD parameters corresponding to each individual, the IMF component is obtained, and the minimum envelope entropy of the IMF component after VMD decomposition is selected as the fitness function to calculate the fitness value. The expression of envelope entropy is as follows: 1); 2); In the formula, is the envelope entropy; is the envelope value of the signal; is the normalized envelope value, i.e. The ratio of the amplitude of a frequency component to the total amplitude; is the total number of frequency components of the signal after Hirt transform.

[0024] Since the tunnel blasting vibration signal is a non-stationary random signal and is affected by the complex high-frequency noise of the surrounding environment during monitoring, the minimum envelope entropy of the decomposed IMF component is selected as the fitness function. The numerical change of the minimum envelope entropy is very sensitive to the high-frequency noise of the blasting vibration signal, and the minimum envelope entropy is calculated based on the envelope of the signal, which can adapt well to the changes of non-stationary signals. Most importantly, in the processing of tunnel blasting vibration signals, the minimum envelope entropy can ensure that the main characteristics of such signals, such as frequency distribution and energy distribution, are retained during the noise reduction process, ensuring the signal analysis effect of subsequent structural safety assessment.

[0025] Preferably, when simulating the random motion of the soft frost particles by the soft frost search strategy, the new position of the updated particle is calculated by the following formula: 3); Where: is the new position of the particle after the update; It is a frost population The best frost agent Particles; parameters is a random number in the range (-1,1), parameter is a random number in the range (0,1); It is an environmental factor; is the adhesion; It is The first of the frost The upper bound of the escape space of particles; It is The first of the frost The lower bound of the escape space of particles; is the adhesion coefficient; is the direction change factor during the iteration process, simulating the change of wind direction;

[0026] Preferably, when simulating the rapid growth and crossover behavior of frost particles in the same direction through the frost puncture mechanism, the new position of the updated particle is calculated by the following formula: 4); Where: It is The normalized value of the fitness value of the frost agent to indicate the A chance for Frost Agent to be selected. is a random number in the range (-1,1).

[0027] Preferably, S3 specifically includes the following steps: S301. Define optimization objectives: The goal of VMD is to decompose the tunnel blasting vibration signal into a finite number of IMFs (band-limited intrinsic mode functions), each of which should have a relatively narrow bandwidth, and reconstruct the original signal by the sum of all IMFs; The decomposed IMF components are expressed as: 5); Where: For the IMF components; For the The amplitude of the IMF component; For the The instantaneous frequency of the IMF components; For time; S302, constructing an optimization problem: constructing an optimization problem using a variational method, aiming to minimize the sum of squares of bandwidths of each IMF; S303, introduce penalty factor , and the alternating direction multiplier method (ADMM) is used to iteratively solve the optimization problem obtained by S302 to obtain the corresponding modal components and its center frequency ; S304, in each iteration, first fix the center frequency and optimize the mode of the IMF, then fix the mode of the IMF and optimize the center frequency; continue to iterate in this alternating manner, and the solution change between the current and subsequent iterations is less than , it is judged to have converged. Since the VMD decomposition of blasting vibration signals requires higher accuracy, it is usually selected As the convergence tolerance. After iteration, the modal component solution that achieves the optimization goal is the desired intrinsic modal component. Each intrinsic modal component represents a main frequency component in the measured signal of the tunnel blasting vibration containing noise, and is output The intrinsic mode components.

[0028] Preferably, the specific steps of S302 are as follows: Combining the classical Wiener filter, Hilbert-Huang transform and spectrum modulation, the expression of the optimization problem constructed by the variational method is obtained as follows: 6); Where: is the decomposed IMF component signal; is the center frequency of the corresponding signal; is the impulse function, and its expression is ; is the single side frequency corresponding to the IMF component; is the original signal before decomposition; is a mathematical expression that means the conditions must be met...; For time; For Find the partial score, that is, In Partial division is a mathematical expression in the process of solving a problem; It represents frequency shift, which is a mathematical expression in the solution process. Its function is to shift the center frequency from Move to baseband (0 frequency) to facilitate subsequent bandwidth calculation and signal separation; is an imaginary unit, that is, , which is similar to the common imaginary unit symbol in mathematics It's the same concept.

[0029] Then by introducing the quadratic penalty factor and Lagrange multipliers as the constraints for signal reconstruction, solve and constrain the optimization problem constructed by the above variational method, and change the expression of the optimization problem constructed by the variational method into: 7); Where: is the Lagrange multiplier; is the penalty factor; is the Lagrangian function; Represents the inner product operation.

[0030] Preferably, The modal components in the iteration and center frequency The expression is as follows: 8); 9); Where: for Frequency domain expression after Fourier transformation; For the In the iteration Frequency domain expressions of the natural mode components; For the The Lagrange multiplier in the iteration; For the In the iteration The center frequency of the modal components; Indicates the number of iterations; is the frequency; is a differential.

[0031] Formula 6) (minimizing the sum of squared bandwidths of each IMF) introduces the penalty factor and Lagrange multipliers After that, it becomes the expression of Lagrangian function, that is, Formula 7). By solving Formula 7), Formula 8) and Formula 9) are obtained. Therefore, Formula 8) and Formula 9) are the expressions of each modal component and its center frequency that satisfy "minimizing the sum of squares of the bandwidth of each IMF".

[0032] A further improvement of the blasting vibration signal noise reduction method of the present invention is that S5 specifically includes the following steps: S501, determine the wavelet basis function and the number of decomposition levels, and perform wavelet decomposition on the intrinsic mode components retained in S4 accordingly.

[0033] Tunnel blasting vibration signals usually have non-stationary and transient characteristics, while the dbN wavelet basis function has compact support and orthogonality, which can effectively capture the mutation and transient characteristics of the signal, thereby retaining the important characteristics of the blasting vibration signal in the noise reduction process; at the same time, tunnel blasting vibration signals are complex and changeable, and may have symmetry in some cases, and the sym wavelet basis function can better handle these symmetric parts; tunnel blasting vibration signals have a wide frequency domain range and a large frequency domain range of components, while the coif wavelet basis function has good time-frequency localization characteristics and can better handle different frequency components of the signal. These three wavelet basis functions have good decomposition effects on tunnel blasting vibration signal components, and the larger the number of decomposition layers, the richer the signal details that can be captured, but the amount of calculation is also greater. Therefore, three wavelet basis functions, dbN, sym, and coif, which have better decomposition effects on the components of tunnel blasting vibration signals, are selected, and the filtered modal components are decomposed by 1-10 layers respectively. The signal-to-noise ratio of the components after reconstruction under different combinations of wavelet basis functions and decomposition layers is calculated and compared. The wavelet basis function and decomposition layer number are determined, and the filtered IMF components are decomposed by wavelet to obtain the wavelet coefficients at different scales. (approximate coefficients) and (detail coefficient); S502, determining a threshold function and a threshold selection principle, specifically: determining a wavelet threshold by a fixed threshold criterion, and performing soft threshold processing on the wavelet coefficients according to the determined threshold; The size of the threshold determines the degree of separation between noise and signal. If the threshold is set too high, important information of the signal may be lost, and if it is set too low, the noise may not be effectively removed. The determination of the threshold size usually needs to be adjusted according to the characteristics of the signal and the noise level. There are four common criteria: unbiased risk estimation threshold, fixed threshold, heuristic threshold, and maximum and minimum threshold. The fixed threshold method is based on the standard deviation of the noise and the signal length. It can effectively remove the random noise existing in the tunnel blasting vibration signal. At the same time, it has a small amount of calculation, can quickly adapt to the non-stationarity of the blasting vibration signal, and provide a globally stable noise suppression level. Not only that, the measured signal of the noisy tunnel blasting vibration contains multiple frequency components. The fixed threshold can ensure that the noise can be effectively removed at different frequencies. Therefore, the present invention uses the fixed threshold criterion to determine the wavelet threshold; 10); Where: is the noise standard deviation, is the length of the signal; is the selected threshold.

[0034] Commonly used wavelet threshold functions include hard threshold function and soft threshold function. The hard threshold function sets the wavelet coefficients less than the threshold to 0, and the wavelet coefficients exceeding the threshold remain unchanged, while the soft threshold function shrinks the wavelet coefficients by a fixed amount. For the measured signal of tunnel blasting vibration containing noise, the soft threshold function can smooth the wavelet coefficients by linearly compressing the wavelet coefficients instead of directly setting them to zero, adapting to the non-stationarity of the signal, and better retaining the multi-frequency components in the signal, reducing the loss of frequency information caused by threshold processing. Therefore, the soft threshold function is used as the threshold function for wavelet threshold denoising, as follows: 11); Where: is the original wavelet coefficient, is the wavelet coefficient after processing.

[0035] S503 , using the wavelet coefficients after threshold processing, reconstruct the signal through inverse wavelet transform to obtain the denoised IMF component.

[0036] Preferably, the tunnel blasting vibration signal after noise reduction The expression is as follows: 12); Where: For the processed modal components; is the number of IMF components after screening.

[0037] This embodiment also discloses a computer device, including a memory and a processor; The memory is used to store a computer program executable on the processor; The processor is used to implement the steps of the above-mentioned blasting vibration signal noise reduction method when executing the computer program.

[0038] This embodiment further discloses a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the blasting vibration signal noise reduction method as described above are implemented.

[0039] In this embodiment, a mountain tunnel project is taken as an example to monitor the tunnel blasting vibration. The instrument used is L20-N blasting vibration meter with a sampling frequency of 2000 Hz. Three vibration velocity curves S1(t), S2(t) and S3(t) with different amplitudes at different times are collected. Figure 2~Figure 7 As shown in the figure, it is found that there is a lot of high-frequency noise in the vibration signal, and the monitored vibration signal needs to be denoised. Figure 2 , Figure 4 and Figure 6The horizontal axis is time, the unit is s, and the vertical axis is vibration speed, the unit is . Figure 3 , Figure 5 and Figure 7 The horizontal axis is the frequency in Hz, and the vertical axis is the amplitude.

[0040] The measured blasting vibration signal is subjected to RIME-optimized VMD decomposition. The number of individuals in the population is set to 10, the maximum number of iterations is set to 20, the penalty factor range is [100, 4000], and the decomposition layer range is [2, 10]. The parameters of VMD are optimized by RIME to obtain the optimal decomposition of the S1(t) signal. The value is 10, the penalty factor is 4000; the best decomposition of S2(t) signal The value is 7, the penalty factor is 287; the best decomposition of S3(t) signal The value is 7, the penalty factor is 1497. The change of the minimum envelope entropy value during the optimization process is shown in Figure 8~Figure 10 As shown, Figure 8~Figure 10 The horizontal axis is the number of iterations, and the vertical axis is the fitness value. The IMF components and their spectra obtained by decomposition are shown in Figure 11~Figure 16 shown. Fig.11 , Fig.13 and Fig.15 The horizontal axis is time, the unit is s, and the vertical axis is vibration speed, the unit is . Fig.12 , Fig.14 and Fig.16 The horizontal axis is the frequency in Hz, and the vertical axis is the amplitude. Fig.11 , Fig.13 and Fig.15 One to one correspondence, Fig.11 and Fig.12 From top to bottom, they are the S1(t)VMD decomposition results and spectrum diagrams of IMF1, IMF2, IMF3, IMF4, IMF5, IMF6, IMF7, IMF8, IMF9 and IMF10. Fig.13 and Fig.14 From top to bottom, they are the S2(t)VMD decomposition results and spectrum diagrams of IMF1, IMF2, IMF3, IMF4, IMF5, IMF6, and IMF7. Fig.15 and Fig.16 From top to bottom, they are the S3(t) VMD decomposition results and spectrum diagrams of IMF1, IMF2, IMF3, IMF4, IMF5, IMF6, and IMF7.

[0041] After the measured blasting vibration signal was decomposed by VMD, the obtained IMF component did not show the endpoint effect, and the modal aliasing phenomenon in the EMD decomposition was significantly suppressed. By observing its spectrum, it can be found that the boundaries of each frequency distribution are relatively clear, and they are arranged in order from low frequency to high frequency. After eliminating the IMF components with a variance contribution rate of less than 1%, the remaining IMF components are subjected to wavelet threshold denoising and reconstructed signals. The signal curve after denoising and the signal spectrum after denoising are shown in Figure 2. Figure 17 to Figure 22 As shown, Fig.17 , Fig.19 and Fig.21 The horizontal axis is time, the unit is s, and the vertical axis is vibration speed, the unit is , In the figure, the blue dotted line is the measured signal, and the red solid line is the reconstructed signal. Fig.18 , Fig. 20 and Fig. 22 The horizontal axis is the frequency in Hz, and the vertical axis is the amplitude.

[0042] Compare the signals before and after noise reduction (such as Figure 17 to Figure 22 As shown in Figure 2, it can be clearly seen that the reconstructed signal retains the basic characteristics of the original signal to a great extent. Fig.21 ) in the reconstruction process, the trend term IMF component in the decomposition was eliminated, effectively solving the problem of baseline drift. Further observation of the spectrum analysis results shows that the high-frequency noise above 400Hz is almost completely eliminated after noise reduction, and the response of the low-frequency band is also significantly reduced, which makes the frequency distribution of the reconstructed signal more consistent with the main frequency range of the blasting vibration signal. This shows that the blasting vibration signal noise reduction method of the present invention has a significant application effect in processing blasting vibration signal noise reduction.

[0043] In order to further evaluate the advantages of RIME-VMD combined with wavelet threshold method in blasting vibration signal denoising, the measured signal is denoised by using EMD, wavelet threshold and EMD+wavelet threshold as in the simulation signal, and the spectrum of the signal after denoising by each method is compared and analyzed. Figure 23~Figure 25 As shown, the X-axis coordinate is frequency in Hz, the Y-axis coordinate is amplitude, and the Z-axis coordinates are, from front to back, the signal without noise reduction, the signal processed by the present invention, the signal processed by EMD in the prior art, the signal processed by wavelet threshold in the prior art, and the signal processed by EMD and wavelet threshold in the prior art.

[0044] By observing the frequency distribution of the signal after denoising, it is found that in terms of the high frequency (greater than 400Hz) of the signal, the other three denoising methods have no obvious effect on removing the high-frequency noise of the blasting vibration signal. However, after the signal is denoised by the present invention, the high-frequency noise almost disappears, and the denoising effect is good. At the same time, in terms of low-frequency response, compared with the other three methods, the low-frequency response in the signal after denoising by the blasting vibration signal denoising method of the present invention is significantly reduced, and the main frequency distribution of the signal after denoising is also more reasonable.

[0045] The blasting vibration signal denoising method of the present invention adaptively optimizes VMD parameters through a frost optimization algorithm, and denoises the tunnel blasting vibration signal in combination with a wavelet threshold, thereby solving the technical problem in the prior art that the nonlinear characteristics of the tunnel blasting vibration signal and the complex factors of the construction environment interfere with the extraction of effective signal components from the actual signal containing noise, which has a great impact on the subsequent signal analysis; the denoising method can effectively suppress the modal aliasing phenomenon during signal decomposition, and solves the problem that the key parameter values ​​usually need to be determined based on experience in the variational mode decomposition algorithm.

[0046] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for reducing noise of blasting vibration signals, characterized in that: The steps include: S1, collecting the measured signal of tunnel blasting vibration containing noise; S2. Use the frost optimization algorithm to adaptively optimize the parameters of VMD to obtain the optimized VMD. The VMD parameters include the penalty factor and the number of decomposition levels ; S3, using the optimized VMD obtained in S2 to decompose the measured signal of tunnel blasting vibration, we get The intrinsic mode components; S4. Calculation The variance contribution rate of the intrinsic modal components is calculated, and the components with variance contribution rate not exceeding 1% are removed, and the remaining intrinsic modal components are retained; S5, using the wavelet threshold to reduce the noise of the retained intrinsic mode components to obtain the denoised IMF components; S6. Reconstruct the denoised IMF components obtained in S5 to obtain the tunnel blasting vibration signal after denoising.

2. The method for denoising blasting vibration signals according to claim 1, characterized in that: S2 specifically includes the following steps: S201, first determine the number of individuals in the population in the RIME algorithm , maximum number of iterations And VMD parameters, then assign VMD parameters to each population individual and normalize them to the [0,1] interval, and then randomly generate the initial population; S202, performing VMD decomposition on the measured signal of tunnel blasting vibration containing noise according to the VMD parameters corresponding to each individual; S203, simulating the random movement of soft frost particles through a soft frost search strategy; and simulating the rapid growth and crossover behavior of hard frost particles in the same direction through a hard frost puncture mechanism; S204, calculate and compare the fitness values ​​of the particles before and after the update. If the fitness value of the particle after the update is better than that before the update, the individual is replaced, otherwise, it remains unchanged; after each population update, the current optimal fitness value and the corresponding parameter combination are recorded; as the number of iterations increases, the population continues to evolve in a better direction, and finally outputs the penalty factor for the decomposition of the measured signal of the tunnel blasting vibration containing noise and the number of decomposition levels The optimal combination of is obtained to obtain the optimized VMD.

3. The method for denoising blasting vibration signals according to claim 2, characterized in that: After VMD decomposition of the measured signal of tunnel blasting vibration containing noise is performed according to the VMD parameters corresponding to each individual, the IMF component is obtained. The minimum envelope entropy of the IMF component after VMD decomposition is selected as the fitness function to calculate the fitness value. The expression of envelope entropy is as follows: 1)； 2)； In the formula, is the envelope entropy; is the envelope value of the signal; is the normalized envelope value, i.e. The ratio of the amplitude of a frequency component to the total amplitude; is the total number of frequency components of the signal after Hirt transform.

4. The method for denoising blasting vibration signals according to claim 2, characterized in that: When simulating the random motion of soft frost particles through the soft frost search strategy, the new position of the updated particles is calculated by the following formula: 3)； Where: is the new position of the particle after the update; It is a frost population The best frost agent Particles; parameters is a random number in the range (-1,1), parameter is a random number in the range (0,1); It is an environmental factor; is the adhesion; It is The first of the frost The upper bound of the escape space of particles; It is The first of the frost The lower bound of the escape space of particles; is the adhesion coefficient; is the direction change factor during the iteration process, simulating the change of wind direction; When simulating the rapid growth and crossover behavior of frost particles in the same direction through the frost penetration mechanism, the new position of the updated particle is calculated by the following formula: 4)； In the formula, It is The normalized value of the fitness value of each frost agent; is a random number in the range (-1,1).

5. The method for denoising blasting vibration signals according to claim 1, characterized in that: S3 specifically includes the following steps: S301, the decomposed IMF components are expressed as: 5)； Where: For the IMF components; For the The amplitude of the IMF component; For the The instantaneous frequency of the IMF components; For time; S302, constructing an optimization problem using a variational method; S303, introduce penalty factor , and the alternating direction multiplier method is used to iteratively solve the optimization problem obtained in S302 to obtain the corresponding modal components and its center frequency ; S304, in each iteration, first fix the center frequency and optimize the mode of the IMF, then fix the mode of the IMF and optimize the center frequency; continue to iterate in this alternating manner, and the solution change between the current and subsequent iterations is less than , it is judged to have converged; after iteration, the modal component solution that achieves the optimization goal is the desired inherent modal component. Each inherent modal component represents a main frequency component in the measured signal of the tunnel blasting vibration containing noise, and is output The intrinsic mode components.

6. The method for denoising blasting vibration signals according to claim 5, characterized in that: The specific steps of S302 are as follows: Combining the classical Wiener filter, Hilbert-Huang transform and spectrum modulation, the expression of the optimization problem constructed by the variational method is obtained as follows: 6)； Where: is the decomposed IMF component signal; is the center frequency of the corresponding signal; is the impulse function, and its expression is ; is the single side frequency corresponding to the IMF component; is the original signal before decomposition; For Find the partial score, that is, In To conduct partial division; is frequency shift; Then by introducing the penalty factor and Lagrange multipliers as the constraints for signal reconstruction, the expression of the optimization problem constructed by the variational method becomes: 7)； Where: is the Lagrange multiplier; is the penalty factor; is the Lagrangian function; represents the inner product operation; No. The modal components in the iteration and center frequency The expression is as follows: 8)； 9)； Where: for Frequency domain expression after Fourier transformation; For the In the iteration Frequency domain expressions of the natural mode components; For the The Lagrange multiplier in the iteration; For the In the iteration The center frequency of the modal components; Indicates the number of iterations; is the frequency.

7. The method for denoising blasting vibration signals according to claim 1, characterized in that: S5 specifically includes the following steps: S501, determining the wavelet basis function and the number of decomposition layers, and performing wavelet decomposition on the intrinsic mode components retained in S4 accordingly; S502, determining a threshold function and a threshold selection principle, specifically: determining a wavelet threshold by a fixed threshold criterion, and performing soft threshold processing on the wavelet coefficients according to the determined threshold; S503 , using the wavelet coefficients after threshold processing, reconstruct the signal through inverse wavelet transform to obtain the denoised IMF component.

8. The method for denoising blasting vibration signals according to claim 1, characterized in that: Tunnel blasting vibration signal after noise reduction The expression is as follows: 12)； In the formula, For the processed modal components; is the number of IMF components after screening.

9. A computer device, characterized in that: including memory and processor; The memory is used to store a computer program executable on the processor; The processor is used to implement the steps of the blasting vibration signal noise reduction method according to any one of claims 1 to 8 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the blasting vibration signal noise reduction method according to any one of claims 1 to 8 are implemented.

Images