Bridge dynamic strain signal noise reduction method, device, electronic device and storage medium
Through the combination of CEEMDAN and singular value decomposition, combined with the dual criterion of correlation coefficient and autocorrelation function, the problem of IMF component selection and noise removal in bridge dynamic strain signals is solved, and more efficient signal noise reduction is achieved and the bridge state is improved. Accuracy of the assessment.
Patent Information
- Application Number
- CN202210736662.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-06-27
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2042-06-27
AI Technical Summary
The prior art has modal aliasing and endpoint effects in bridge dynamic strain signal noise reduction, and it is difficult to accurately select effective Intrinsic Mode Function (IMF) components, affecting the noise reduction effect.
Using Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) combined with the dual criterion of singular value decomposition, correlation coefficient and autocorrelation function, the effective IMF component is selected and noise is removed, and pre-processed through an adaptive noise canceller to improve medium and low frequency noise filtering and removal. ability.
It effectively eliminates modal aliasing and endpoint effects, improves signal decomposition efficiency and accuracy, removes noise better, avoids loss of useful information, and improves the noise reduction effect of bridge dynamic strain signals.
Smart Images

Figure CN114997239B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of signal noise reduction, and in particular to a bridge dynamic strain signal noise reduction method, device, electronic equipment and storage medium. Background Art
[0002] Bridges are important infrastructure in my country and a key part of the land transportation network. However, many problems have been exposed in the process of using bridges. On the one hand, some bridges may have inherent deficiencies for some reason; on the other hand, with the development of social productivity, the traffic volume continues to increase, coupled with natural and man-made factors, the structural performance has been increasingly deteriorated, causing damage and destruction to the bridge structure; furthermore, some bridges are approaching the design service life and gradually have quality problems caused by aging. If they are not promptly inspected and reinforced, serious accidents may occur.
[0003] In order to ensure the safety and serviceability of the bridge structure throughout its service life, it is necessary to monitor the health of the bridge structure and evaluate the bridge status based on the monitoring results. Due to the noise interference in the bridge monitoring signal, it is difficult to accurately evaluate the bridge structure status directly based on the monitoring signal. Among them, dynamic strain is a physical quantity that reflects the dynamic stiffness of the bridge, which is extremely critical for evaluating the bridge status. Therefore, using a reasonable method to reduce the noise of the dynamic strain signal and minimize the impact of noise is very important for accurately evaluating the bridge status.
[0004] In the prior art, the commonly used denoising methods are mainly wavelet threshold denoising and EMD (Empirical Mode Decomposition) denoising. However, wavelet threshold denoising achieves denoising through a threshold, but there is no clear applicable scope for a variety of threshold criteria, so the denoising effect is greatly affected by subjectivity. EMD, on the other hand, has modal aliasing and endpoint effects due to the defects of the algorithm itself, and the denoising effect is not ideal. EEMD (Ensemble Empirical Mode Decomposition) and CEEMDAN (Complete EEMD with Adaptive Noise) are improvements to the EMD algorithm, which can improve the modal aliasing problem. However, after decomposing the signal into IMF (Intrinsic Mode Function) components, how to select effective IMF components and how to remove noise in the effective IMF components are still difficult. Summary of the invention
[0005] In view of this, it is necessary to provide a bridge dynamic strain signal denoising method, device, electronic device and storage medium, so as to select effective IMF components from IMF components and remove noise in the effective IMF components.
[0006] In order to achieve the above objectives, in a first aspect, the present invention provides a bridge dynamic strain signal noise reduction method, comprising:
[0007] Obtain the dynamic strain signal of the bridge to be de-noised;
[0008] Performing a preprocessing operation on the bridge dynamic strain signal to be denoised to obtain a preprocessed signal, and performing CEEMDAN decomposition on the preprocessed signal to obtain multiple IMF components and a residual component;
[0009] Determine the effective IMF component by using the correlation coefficient and the autocorrelation function, wherein the correlation coefficient is the correlation coefficient between each IMF component and the dynamic strain signal of the bridge to be de-noised, and the autocorrelation function is the normalized autocorrelation function of the IMF component whose correlation coefficient is greater than a preset threshold;
[0010] The effective IMF components are subjected to singular value decomposition, and effective singular values are selected using a singular value difference spectrum method. The effective singular values of the effective IMF components and the residual components are reconstructed to obtain a de-noised bridge dynamic strain signal.
[0011] Furthermore, the preprocessing operation is performed on the dynamic strain signal of the bridge to be de-noised to obtain a preprocessed signal, including:
[0012] Constructing an adaptive noise canceller, and using the adaptive noise canceller to preprocess the dynamic strain signal of the bridge to be de-noised, to obtain the preprocessed signal;
[0013] The constructing of the adaptive noise canceller specifically includes:
[0014] Performing wavelet decomposition on the dynamic strain signal of the bridge to be de-noised to obtain approximate signals and detail signals at different scales;
[0015] The adaptive noise canceller is constructed by using an adaptive filter and a low-frequency detail signal corresponding to a preset decomposition scale in wavelet decomposition.
[0016] Furthermore, the iterative algorithm of the adaptive filter includes:
[0017] y(n)=W T (n)X(n)
[0018] e(n)=d(n)-y(n)
[0019] W(n+1)=W(n)+2μe(n)x(n),
[0020] Among them, X(n) is the input signal, y(n) is the output signal, W(n) is the weight coefficient, d(n) is the desired signal, e(n) is the error signal, and μ is the step size.
[0021] Furthermore, the preprocessed signal is subjected to CEEMDAN decomposition to obtain a plurality of IMF components and a residual component, including:
[0022] Adding white noise to the preprocessed signal to obtain a first set signal, performing empirical mode decomposition on the first set signal, and solving a first IMF component;
[0023] Acquire a first residual component other than the first IMF component from the preprocessed signal, add the white noise to the first residual component to obtain a second set signal, perform empirical mode decomposition on the second set signal, and solve for a second IMF component;
[0024] Repeat the above decomposition steps until the residual component is a monotonic function and cannot be decomposed, and obtain multiple IMF components and a residual component.
[0025] Further, the determining of effective IMF components by using correlation coefficients and autocorrelation functions includes:
[0026] Calculating the correlation coefficient between each IMF component and the dynamic strain signal of the bridge to be de-noised, and determining the IMF component whose correlation coefficient is less than a preset correlation coefficient threshold as the first invalid IMF component;
[0027] Determine an IMF component whose correlation coefficient is greater than the preset correlation coefficient threshold, calculate a normalized autocorrelation function of the IMF component whose correlation coefficient is greater than the preset correlation coefficient threshold, and determine an IMF component that meets the characteristics of the noise autocorrelation function as a second invalid IMF component;
[0028] The remaining IMF components of the IMF components whose correlation coefficients are greater than the preset correlation coefficient threshold except the second invalid IMF component are determined as the valid IMF components.
[0029] Further, performing singular value decomposition on the effective IMF component includes:
[0030] Reconstructing the effective IMF components into a Hankel matrix using phase space;
[0031] Perform singular value decomposition on the Hankel matrix.
[0032] Furthermore, the effective order singular values of the effective IMF components and the residual components are reconstructed to obtain a de-noised bridge dynamic strain signal, including:
[0033] A reconstruction matrix is obtained according to an inverse process of singular value decomposition, and an inverse transformation of phase space reconstruction is performed based on the reconstruction matrix to obtain a reconstructed effective IMF component;
[0034] The reconstructed effective IMF component and the residual component are reconstructed to obtain the de-noised bridge dynamic strain signal.
[0035] In a second aspect, the present invention further provides a bridge dynamic strain signal noise reduction device, comprising:
[0036] An acquisition module, used for acquiring the dynamic strain signal of the bridge to be de-noised;
[0037] A decomposition module is used to perform a preprocessing operation on the dynamic strain signal of the bridge to be denoised to obtain a preprocessed signal, and perform CEEMDAN decomposition on the preprocessed signal to obtain multiple IMF components and a residual component;
[0038] A determination module is used to determine the effective IMF component using a correlation coefficient and an autocorrelation function, wherein the correlation coefficient is the correlation coefficient between each IMF component and the dynamic strain signal of the bridge to be denoised, and the autocorrelation function is a normalized autocorrelation function of the IMF component whose correlation coefficient is greater than a preset threshold.
[0039] The reconstruction module is used to perform singular value decomposition on the effective IMF component, select effective singular values using a singular value difference spectrum method, reconstruct the effective singular values of the effective IMF component and the residual component to obtain a de-noised bridge dynamic strain signal.
[0040] In a third aspect, the present invention also provides an electronic device comprising a memory and a processor, wherein the memory is used to store programs; and the processor is coupled to the memory and is used to execute the programs stored in the memory to implement the steps in the above-mentioned bridge dynamic strain signal denoising method.
[0041] In a fourth aspect, the present invention further provides a computer-readable storage medium for storing computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the above-mentioned bridge dynamic strain signal denoising method.
[0042] The beneficial effects of adopting the above-mentioned embodiment are as follows: the present invention can improve the filtering ability of medium and low frequency noise by preprocessing the dynamic strain signal of the bridge to be de-noised to obtain a preprocessed signal; by adopting CEEMDAN decomposition, modal aliasing and end effect phenomena are eliminated, while the amount of calculation is greatly reduced and the signal decomposition efficiency is improved; by selecting effective IMF components using the dual criteria of correlation coefficient and autocorrelation function, the effective IMF components can be selected more accurately to avoid the loss of useful information; by combining CEEMDAN decomposition with singular value decomposition, compared with other single denoising methods, noise can be better removed. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] Figure 1 A schematic diagram of a flow chart of an embodiment of a bridge dynamic strain signal denoising method provided by the present invention;
[0044] Figure 2 A schematic diagram of a process for performing CEEMDAN decomposition on a preprocessed signal to obtain multiple IMF components and one residual component provided by an embodiment of the present invention;
[0045] Figure 3 A schematic diagram of a process for determining effective IMF components using correlation coefficients and autocorrelation functions provided by an embodiment of the present invention;
[0046] Figure 4 A time history diagram of a real signal and a simulated signal provided by an embodiment of the present invention;
[0047] Figure 5 A comparison diagram of bridge simulation signals before and after preprocessing provided by an embodiment of the present invention;
[0048] Figure 6 A reference diagram of IMF components and residual components obtained by decomposing a preprocessed simulation signal through CEEMDAN according to an embodiment of the present invention;
[0049] Figure 7 A reference graph of normalized autocorrelation functions of IMF components greater than a correlation coefficient threshold provided by an embodiment of the present invention;
[0050] Figure 8 A comparison diagram of a real signal and a noise reduction signal provided by an embodiment of the present invention;
[0051] Fig. 9 An overall layout diagram of a bridge structure is provided for one embodiment of the present invention;
[0052] Fig.10 A time history diagram of a bridge dynamic strain signal is provided for one embodiment of the present invention, that is, a time history diagram of a dynamic strain signal before noise reduction;
[0053] Fig.11A reference diagram of IMF components and residual components obtained by decomposing a preprocessed signal through CEEMDAN according to an embodiment of the present invention;
[0054] Fig.12 Another reference graph of normalized autocorrelation functions of IMF components greater than a correlation coefficient threshold provided by an embodiment of the present invention;
[0055] Fig.13 A reference diagram of a dynamic strain signal after noise reduction provided by an embodiment of the present invention;
[0056] Fig.14 A comparison diagram of the power spectral density function of the dynamic strain signal before and after noise reduction provided by an embodiment of the present invention;
[0057] Fig.15 A schematic structural diagram of an embodiment of a bridge dynamic strain signal noise reduction device provided by the present invention;
[0058] Fig.16 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0059] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.
[0060] In the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically limited. Reference to "embodiment" in this article means that the specific features, structures or characteristics described in conjunction with the embodiment may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0061] The invention provides a bridge dynamic strain signal noise reduction method, device, electronic equipment and storage medium.
[0062] The following describes each one separately. Figure 1 , Figure 1 A flow chart of an embodiment of a bridge dynamic strain signal denoising method provided by the present invention is provided. A specific embodiment of the present invention discloses a bridge dynamic strain signal denoising method, comprising:
[0063] Step S101: obtaining a dynamic strain signal of a bridge to be de-noised;
[0064] Step S102: performing a preprocessing operation on the bridge dynamic strain signal to be de-noised to obtain a preprocessed signal, and performing CEEMDAN decomposition on the preprocessed signal to obtain multiple IMF components and a residual component;
[0065] Step S103: determining effective IMF components using correlation coefficients and autocorrelation functions, wherein the correlation coefficient is the correlation coefficient between each IMF component and the dynamic strain signal of the bridge to be de-noised, and the autocorrelation function is the normalized autocorrelation function of the IMF components whose correlation coefficients are greater than a preset threshold;
[0066] Step S104: singular value decomposition is performed on the effective IMF components, and effective singular values are selected using singular value difference spectrum method, and the effective singular values of the effective IMF components and the residual components are reconstructed to obtain the de-noised bridge dynamic strain signal.
[0067] Among them, the bridge dynamic strain signal to be de-noised can be a simulated signal generated by a computer, or it can be a real signal for testing the bridge, which is generally represented by y(t). The preprocessed signal can be represented by y1(t). It can be understood that by preprocessing the bridge dynamic strain signal to be de-noised, the bridge dynamic strain signal to be de-noised can be roughly filtered of noise. By adopting CEEMDAN decomposition, compared with EMD, EEMD, etc., the modal aliasing and end effect phenomena are eliminated, while the amount of calculation is greatly reduced and the signal decomposition efficiency is improved. By adopting the dual criteria of correlation coefficient and autocorrelation function to select effective IMF components, the effective components can be selected more accurately to avoid the loss of useful information. By performing singular value decomposition on the effective IMF components and selecting the effective singular values using the singular value difference spectrum method, the noise in the effective IMF components can be effectively removed. In addition, by combining CEEMDAN decomposition with singular value decomposition, the present invention can better remove noise compared with other single noise reduction methods.
[0068] The present invention preprocesses the dynamic strain signal of the bridge to be de-noised to obtain a preprocessed signal, thereby improving the filtering ability for medium and low frequency noise; by adopting CEEMDAN decomposition, modal aliasing and end effect phenomena are eliminated, while the amount of calculation is greatly reduced and the signal decomposition efficiency is improved; by selecting effective IMF components using the dual criteria of correlation coefficient and autocorrelation function, the effective IMF components can be selected more accurately to avoid the loss of useful information; by combining CEEMDAN decomposition with singular value decomposition, compared with other single denoising methods, the noise can be better removed.
[0069] In one embodiment of the present invention, a preprocessing operation is performed on the bridge dynamic strain signal to be de-noised to obtain a preprocessed signal, including:
[0070] An adaptive noise canceller is constructed, and the adaptive noise canceller is used to preprocess the dynamic strain signal of the bridge to be de-noised to obtain a preprocessed signal;
[0071] Construct an adaptive noise canceller, including:
[0072] The dynamic strain signal of the bridge to be de-noised is decomposed by wavelet to obtain approximate signals and detail signals at different scales;
[0073] The formula for wavelet decomposition is:
[0074]
[0075] ψ(t) is called the base wavelet or mother wavelet, a is the scale factor, b is the translation factor, a,b∈L 2 (R) and a≠0.
[0076] The decomposition algorithm of wavelet decomposition is:
[0077]
[0078]
[0079]
[0080] N1 is the number of sampling points, and the approximate signal frequency range is The detail signal frequency range is j=1,LJ, J is the number of decomposition levels.
[0081] An adaptive noise canceller is constructed by using an adaptive filter and a low-frequency detail signal corresponding to a preset decomposition scale in wavelet decomposition.
[0082] Among them, the iterative algorithm of the adaptive filter includes:
[0083] y(n)=W T (n)X(n)
[0084] e(n)=d(n)-y(n)
[0085] W(n+1)=W(n)+2μe(n)x(n),
[0086] X(n) is the input signal, y(n) is the output signal, W(n) is the weight coefficient, d(n) is the desired signal, e(n) is the error signal, and μ is the step size.
[0087] The detail signal corresponding to the maximum decomposition scale in wavelet decomposition is a low-frequency signal. By taking the detail component of the maximum decomposition scale as the reference signal during preprocessing, the ability to filter out medium and low frequency noise is improved.
[0088] In one embodiment of the present invention, see Figure 2 , Figure 2 A flow chart of performing CEEMDAN decomposition on a preprocessed signal to obtain multiple IMF components and one residual component is provided in accordance with an embodiment of the present invention, including:
[0089] Step S201: adding white noise to the preprocessed signal to obtain a first aggregate signal, performing empirical mode decomposition on the first aggregate signal, and solving a first IMF component;
[0090] Step S202: obtaining a first residual component except the first IMF component from the preprocessed signal, adding white noise to the first residual component to obtain a second set signal, performing empirical mode decomposition on the second set signal, and solving the second IMF component;
[0091] Step S203: Repeat the above decomposition steps until the residual component is a monotonic function and cannot be decomposed, thereby obtaining a plurality of IMF components and a residual component.
[0092] Specifically, white noise n can be added to the preprocessed signal y1(t) i (t) to obtain the first aggregate signal, and the first aggregate signal y1(t)+n i (t) Perform empirical mode decomposition to obtain the corresponding component IMF i1 , summing and averaging to get the first-order IMF component, that is, IMF1:
[0093]
[0094] Where N2 is the number of times white noise is added.
[0095] The residual component after the first-order decomposition, that is, the first residual component, is:
[0096] r1(t)=x(t)-IMF1,
[0097] Add white noise n to the first residual component i (t) to the second aggregate signal, for the second aggregate signal r1(t)+n i (t) Perform empirical mode decomposition to obtain the corresponding component IMF i2 , summing and averaging to get the second-order IMF component, that is, IMF2:
[0098]
[0099] Repeat the above steps until the residual component is a monotonic function that cannot be decomposed, and obtain multiple IMF components and a residual component. The decomposed signal can be expressed as:
[0100]
[0101] In one embodiment of the present invention, see Figure 3 , Figure 3 A flow chart of determining effective IMF components using correlation coefficients and autocorrelation functions provided by an embodiment of the present invention includes:
[0102] Step S301: calculating the correlation coefficient between each IMF component and the dynamic strain signal of the bridge to be de-noised, and determining the IMF component with a correlation coefficient less than a preset correlation coefficient threshold as the first invalid IMF component;
[0103] Step S302: determining the IMF components whose correlation coefficients are greater than a preset correlation coefficient threshold, calculating the normalized autocorrelation function of the IMF components whose correlation coefficients are greater than the preset correlation coefficient threshold, and determining the IMF components that meet the characteristics of the noise autocorrelation function as the second invalid IMF components;
[0104] Step S303: determining the remaining IMF components except the second invalid IMF component among the IMF components whose correlation coefficients are greater than a preset correlation coefficient threshold as valid IMF components.
[0105] Among them, the preset correlation coefficient threshold is: The IMF component with a correlation coefficient less than the threshold is the first invalid IMF component and can be eliminated; then, among the IMF components with a correlation coefficient greater than the threshold, the normalized autocorrelation function of the IMF components with a correlation coefficient greater than the threshold is calculated:
[0106]
[0107] Where T is the signal period, τ is the signal offset time, and then the second invalid IMF component that meets the characteristics of the noise autocorrelation function is eliminated, and the remaining IMF components are valid IMF components.
[0108] In one embodiment of the present invention, performing singular value decomposition on the effective IMF components includes:
[0109] The effective IMF components are reconstructed using phase space to generate the Hankel matrix;
[0110] Perform singular value decomposition on the Hankel matrix.
[0111] Specifically, the effective IMF components are reconstructed using phase space to generate an m×n Hankel matrix, which is:
[0112]
[0113] Where M = m + n - 1; the singular value decomposition of the Hankel matrix is calculated as follows:
[0114]
[0115] Where U is an m×m orthogonal matrix, V is an n×n orthogonal matrix, and S is a non-negative diagonal matrix whose main diagonal elements are σ1, L,σ r ,L 0,L,0, the first r elements are called Hankel singular values.
[0116] In one embodiment of the present invention, the effective singular values and residual components of the effective IMF components are reconstructed to obtain a de-noised bridge dynamic strain signal, including:
[0117] The reconstruction matrix is obtained according to the inverse process of singular value decomposition, and the inverse transformation of phase space reconstruction based on the reconstruction matrix is performed to obtain the reconstructed effective IMF components;
[0118] The reconstructed effective IMF components and residual components are reconstructed to obtain the denoised bridge dynamic strain signal.
[0119] It can be understood that the reconstruction matrix is obtained according to the inverse process of singular value decomposition, and then the inverse transformation of phase space reconstruction is performed to obtain the denoised IMF component, which is finally reconstructed to obtain the denoised signal, that is, the denoised bridge dynamic strain signal.
[0120] In order to better understand the present invention, the following is a detailed description of the present invention by combining the simulation signal and the real signal as the bridge dynamic strain signal to be de-noised. First, the simulation signal is taken as an example.
[0121] Step 1: The fundamental frequency of the bridge is generally below 10Hz. Design the simulation signal:
[0122] y(t)=cos(2π×t)+sin(2π×6t)+sin(2π×8t)+w,
[0123] Where w is random noise, sampling time is 100s, sampling frequency Fs = 128Hz, for easy signal observation, please refer to Figure 4 , Figure 4 A time chart of a real signal and a simulated signal provided in one embodiment of the present invention, wherein the real signal and the simulated signal are signals with a duration of 10 s intercepted from the sampled signal.
[0124] Step 2: Perform five-layer wavelet decomposition on y(t) to obtain five detail signals and one approximate signal.
[0125] Step 3: Combined with the adaptive noise canceller, the fifth layer detail signal is used as the reference signal of the adaptive noise canceller to preprocess y(t). For the comparison chart before and after preprocessing, please refer to Figure 5 , Figure 5A comparison diagram of bridge simulation signals before and after preprocessing provided by an embodiment of the present invention.
[0126] Step 4: The preprocessed signal is decomposed by CEEMDAN to obtain 15 IMF components and 1 residual component, see Figure 6 , Figure 6 A reference diagram of IMF components and residual components obtained by decomposing a preprocessed signal through CEEMDAN according to an embodiment of the present invention.
[0127] Add white noise n to the preprocessed signal y1(t) i (t), for the signal y1(t)+n i (t) performs empirical mode decomposition, sums and calculates the average value to obtain IMF1; the residual component after the first-order decomposition is r1(t) = x(t) - IMF1, and white noise n is added to the residual component i (t), for r1(t)+n i (t) Perform empirical mode decomposition to obtain IMF2; repeat the above steps until the residual component can no longer be decomposed.
[0128] Step 5: Calculate the correlation coefficient between each IMF component and y(t), with the correlation coefficient threshold being 0.0657. Figure 6 It can be seen that IMF1~IMF9 are greater than the threshold;
[0129] Step 6: Calculate the autocorrelation function of IMF1 to IMF9, see Figure 7 , Figure 7 A reference diagram of normalized autocorrelation functions of IMF components greater than the correlation coefficient threshold provided by an embodiment of the present invention. Among them, IMF1-IMF3, IMF7, and IMF9 meet the characteristics of noise autocorrelation functions, and they are removed, and the remaining components are used as effective IMF components.
[0130] Step 7: Perform singular value decomposition on the effective components, and then use the singular value difference spectrum method to select the effective singular values, perform the inverse process of singular value decomposition to obtain the reconstruction matrix, and then perform the inverse transformation of phase space reconstruction to obtain the denoised IMF components; finally, reconstruct them to obtain the denoised signal. Please refer to Figure 8 , Figure 8 The comparison diagram of the real signal and the noise reduction signal provided in one embodiment of the present invention is a comparison between a 10s duration signal intercepted from the noise reduction signal and the real signal.
[0131] In order to verify the superiority of the present invention, the same signal is denoised using wavelet threshold and CEEMDAN decomposition respectively, and the signal-to-noise ratio and root mean square error are used as evaluation indicators. The comparison results are shown in Table 1.
[0132] Table 1
[0133] CEEMDAN Wavelet Thresholding The present invention SNR / dB 7.65 11.90 18.75 RMSE 0.30 0.18 0.08
[0134] Through comparison, it is found that the present invention can indeed achieve a good noise reduction effect, the signal-to-noise ratio after noise reduction is greatly improved, and the root mean square error is small.
[0135] Then, take the real signal as an example. Specifically, take a five-span concrete continuous beam bridge in Wuhan as the engineering background. The cross section is a box section. In order to monitor the dynamic response of the continuous beam bridge in real time, the measuring points are arranged on the bottom plate of the middle section beam of the fourth span of the bridge to measure the dynamic strain. Fig. 9 , Fig. 9 An overall layout diagram of a bridge structure is provided for an embodiment of the present invention. The dynamic strain monitoring signal y(t) from 12:06:40 to 12:15:00 on July 20, 2019 is collected. The steps of this embodiment are similar to the steps of processing based on the simulation signal. To avoid redundancy, this section only describes the key parts.
[0136] See also Fig.10 , Fig.10 A time history diagram of a bridge dynamic strain signal is provided for one embodiment of the present invention, that is, a time history diagram of a dynamic strain signal before noise reduction. It can be seen from the diagram that the signal is interfered by noise.
[0137] After preprocessing the dynamic strain signal, it is decomposed by CEEMDAN to obtain 17 IMF components and 1 residual component. Fig.11 , Fig.11 Reference diagram of IMF components and residual components obtained by CEEMDAN decomposition of the preprocessed signal provided in one embodiment of the present invention. The correlation coefficient threshold is λ=0.0587, where the correlation coefficients of IMF2, IMF16 and y(t) are less than the threshold, and are invalid IMF components that can be removed.
[0138] See also Fig.12 , Fig.12 Another reference graph of normalized autocorrelation functions of IMF components greater than a correlation coefficient threshold provided by an embodiment of the present invention. It can be seen from the graph that IMF3 to IMF7 are invalid IMF components and can be eliminated.
[0139] Perform singular value decomposition denoising on the effective IMF components, see Fig.13 , Fig.13 A reference diagram of a dynamic strain signal after noise reduction provided by an embodiment of the present invention. By comparing the time history diagrams of the dynamic strain signal before and after noise reduction, it can be seen that the signal curve is smoother and the mutation points are significantly reduced.
[0140] See also Fig.14 , Fig.14A comparison diagram of the power spectral density function of the dynamic strain signal before and after noise reduction provided by an embodiment of the present invention; it can be seen that the power spectral density function of the signal after noise reduction is smoother and the structural fundamental frequency is more obvious, indicating that the present invention is suitable for noise reduction of dynamic strain signals of bridges.
[0141] In addition, a finite element model of the bridge superstructure is established to obtain the first two natural frequencies of the structure, and the ARMA method is used to identify the de-noised dynamic strain signal. The identification results are shown in Table 2.
[0142] Table 2
[0143] Finite element simulation results / Hz ARMA method recognition results / Hz Stage 1 2.32 2.51 Stage 2 4.63 4.29
[0144] It can be seen from Table 2 that the ARMA identification results are basically consistent with the finite element simulation results. Through the analysis of the above embodiments, it can be seen that the present invention can better reduce the noise of the bridge dynamic strain signal, more accurately identify the structural mode, and thus accurately evaluate the bridge state.
[0145] In order to better implement the bridge dynamic strain signal denoising method in the embodiment of the present invention, based on the bridge dynamic strain signal denoising method, correspondingly, please refer to Fig.15 , Fig.15 This is a schematic diagram of the structure of an embodiment of a bridge dynamic strain signal noise reduction device provided by the present invention. The embodiment of the present invention provides a bridge dynamic strain signal noise reduction device 1500, including:
[0146] An acquisition module 1501 is used to acquire a dynamic strain signal of a bridge to be de-noised;
[0147] Decomposition module 1502, used for performing preprocessing operation on the bridge dynamic strain signal to be de-noised to obtain a preprocessed signal, and performing CEEMDAN decomposition on the preprocessed signal to obtain multiple IMF components and a residual component;
[0148] The determination module 1503 is used to determine the effective IMF components by using the correlation coefficient and the autocorrelation function, wherein the correlation coefficient is the correlation coefficient between each IMF component and the dynamic strain signal of the bridge to be denoised, and the autocorrelation function is the normalized autocorrelation function of the IMF component whose correlation coefficient is greater than a preset threshold.
[0149] The reconstruction module 1504 is used to perform singular value decomposition on the effective IMF components, select effective singular values using the singular value difference spectrum method, reconstruct the effective order singular values and residual components of each effective IMF component, and obtain the de-noised bridge dynamic strain signal.
[0150] It should be noted here that: the device 1500 provided in the above embodiment can implement the technical solutions described in the above method embodiments. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above method embodiments, which will not be repeated here.
[0151] Based on the above-mentioned bridge dynamic strain signal denoising method, an embodiment of the present invention also provides an electronic device accordingly, including: a processor and a memory, and a computer program stored in the memory and executable on the processor; when the processor executes the computer program, the steps in the bridge dynamic strain signal denoising method of the above-mentioned embodiments are implemented.
[0152] Fig.16 1 shows a schematic diagram of the structure of an electronic device 1600 suitable for implementing the embodiment of the present invention. The electronic device in the embodiment of the present invention may include, but is not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle terminals (such as vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Fig.16 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.
[0153] The electronic device includes: a memory and a processor, wherein the processor here may be referred to as a processing device 1601 hereinafter, and the memory may include at least one of a read-only memory (ROM) 1602, a random access memory (RAM) 1603, and a storage device 1608 hereinafter, as specifically shown below:
[0154] like Fig.16 As shown, the electronic device 1600 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 1601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1602 or a program loaded from a storage device 1608 into a random access memory (RAM) 1603. In the RAM 1603, various programs and signals required for the operation of the electronic device 1600 are also stored. The processing device 1601, the ROM 1602, and the RAM 1603 are connected to each other via a bus 1604. An input / output (I / O) interface 1605 is also connected to the bus 1604.
[0155] Typically, the following devices may be connected to the I / O interface 1605: an input device 1606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 1607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 1608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1609. The communication device 1609 may allow the electronic device 1600 to communicate with other devices wirelessly or by wire to exchange signals. Although Fig.16The electronic device 1600 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.
[0156] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 1609, or installed from the storage device 1608, or installed from the ROM 1602. When the computer program is executed by the processing device 1601, the above-mentioned functions defined in the method of the embodiment of the present invention are executed.
[0157] Based on the above-mentioned bridge dynamic strain signal denoising method, an embodiment of the present invention also provides a computer-readable storage medium accordingly, which stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the bridge dynamic strain signal denoising method as described in the above-mentioned embodiments.
[0158] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.
[0159] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.
Claims
1. A bridge dynamic strain signal denoising method, characterized in that: include: Obtain the dynamic strain signal of the bridge to be de-noised; Performing a preprocessing operation on the bridge dynamic strain signal to be denoised to obtain a preprocessed signal, and performing CEEMDAN decomposition on the preprocessed signal to obtain multiple IMF components and a residual component; The preprocessing operation is performed on the dynamic strain signal of the bridge to be de-noised to obtain a preprocessed signal, including: Constructing an adaptive noise canceller, and using the adaptive noise canceller to preprocess the dynamic strain signal of the bridge to be de-noised, to obtain the preprocessed signal; The constructing of the adaptive noise canceller specifically includes: Performing wavelet decomposition on the dynamic strain signal of the bridge to be de-noised to obtain approximate signals and detail signals at different scales; The adaptive noise canceller is constructed by using an adaptive filter and a low-frequency detail signal corresponding to a preset decomposition scale in wavelet decomposition; The iterative algorithm of the adaptive filter includes: in, is the input signal, is the output signal, is the weight coefficient, is the expected signal, is the error signal, is the step length; Determine the effective IMF component by using the correlation coefficient and the autocorrelation function, wherein the correlation coefficient is the correlation coefficient between each IMF component and the dynamic strain signal of the bridge to be de-noised, and the autocorrelation function is the normalized autocorrelation function of the IMF component whose correlation coefficient is greater than a preset threshold; Determining the effective IMF component by using the correlation coefficient and the autocorrelation function includes: Calculating the correlation coefficient between each IMF component and the dynamic strain signal of the bridge to be de-noised, and determining the IMF component whose correlation coefficient is less than a preset correlation coefficient threshold as the first invalid IMF component; Determine an IMF component whose correlation coefficient is greater than the preset correlation coefficient threshold, calculate a normalized autocorrelation function of the IMF component whose correlation coefficient is greater than the preset correlation coefficient threshold, and determine an IMF component that meets the characteristics of the noise autocorrelation function as a second invalid IMF component; Determine the remaining IMF components of the IMF components whose correlation coefficients are greater than the preset correlation coefficient threshold except the second invalid IMF component as the valid IMF components; The effective IMF components are subjected to singular value decomposition, and effective singular values are selected using a singular value difference spectrum method. The effective singular values of the effective IMF components and the residual components are reconstructed to obtain a de-noised bridge dynamic strain signal.
2. The bridge dynamic strain signal denoising method according to claim 1, characterized in that: The preprocessed signal is subjected to CEEMDAN decomposition to obtain a plurality of IMF components and a residual component, including: Adding white noise to the preprocessed signal to obtain a first set signal, performing empirical mode decomposition on the first set signal, and solving a first IMF component; Acquire a first residual component other than the first IMF component from the preprocessed signal, add the white noise to the first residual component to obtain a second set signal, perform empirical mode decomposition on the second set signal, and solve for a second IMF component; Repeat the above decomposition steps until the residual component is a monotonic function and cannot be decomposed, and obtain multiple IMF components and a residual component.
3. The bridge dynamic strain signal denoising method according to claim 1, characterized in that: The performing singular value decomposition on the effective IMF component comprises: Reconstructing the effective IMF components into a Hankel matrix using phase space; Perform singular value decomposition on the Hankel matrix.
4. The bridge dynamic strain signal denoising method according to claim 3 is characterized in that: The reconstructing the effective singular value of the effective IMF component and the residual component to obtain the de-noised bridge dynamic strain signal includes: A reconstruction matrix is obtained according to an inverse process of singular value decomposition, and an inverse transformation of phase space reconstruction is performed based on the reconstruction matrix to obtain a reconstructed effective IMF component; The reconstructed effective IMF component and the residual component are reconstructed to obtain the de-noised bridge dynamic strain signal.
5. A bridge dynamic strain signal noise reduction device, characterized in that: include: An acquisition module, used for acquiring the dynamic strain signal of the bridge to be de-noised; A decomposition module is used to perform a preprocessing operation on the dynamic strain signal of the bridge to be denoised to obtain a preprocessed signal, and perform CEEMDAN decomposition on the preprocessed signal to obtain multiple IMF components and a residual component; The preprocessing operation is performed on the dynamic strain signal of the bridge to be de-noised to obtain a preprocessed signal, including: Constructing an adaptive noise canceller, and using the adaptive noise canceller to preprocess the dynamic strain signal of the bridge to be de-noised, to obtain the preprocessed signal; The constructing of the adaptive noise canceller specifically includes: Performing wavelet decomposition on the dynamic strain signal of the bridge to be de-noised to obtain approximate signals and detail signals at different scales; The adaptive noise canceller is constructed by using an adaptive filter and a low-frequency detail signal corresponding to a preset decomposition scale in wavelet decomposition; The iterative algorithm of the adaptive filter includes: in, is the input signal, is the output signal, is the weight coefficient, is the expected signal, is the error signal, is the step length; A determination module, used to determine effective IMF components using correlation coefficients and autocorrelation functions, wherein the correlation coefficient is the correlation coefficient between each IMF component and the dynamic strain signal of the bridge to be de-noised, and the autocorrelation function is a normalized autocorrelation function of the IMF component whose correlation coefficient is greater than a preset threshold; Determining the effective IMF component by using the correlation coefficient and the autocorrelation function includes: Calculating the correlation coefficient between each IMF component and the dynamic strain signal of the bridge to be de-noised, and determining the IMF component whose correlation coefficient is less than a preset correlation coefficient threshold as the first invalid IMF component; Determine an IMF component whose correlation coefficient is greater than the preset correlation coefficient threshold, calculate a normalized autocorrelation function of the IMF component whose correlation coefficient is greater than the preset correlation coefficient threshold, and determine an IMF component that meets the characteristics of the noise autocorrelation function as a second invalid IMF component; Determine the remaining IMF components of the IMF components whose correlation coefficients are greater than the preset correlation coefficient threshold except the second invalid IMF component as the valid IMF components; The reconstruction module is used to perform singular value decomposition on the effective IMF component, select effective singular values using a singular value difference spectrum method, reconstruct the effective singular values of the effective IMF component and the residual component to obtain a de-noised bridge dynamic strain signal.
6. An electronic device, characterized in that: It includes a memory and a processor, wherein the memory is used to store programs; the processor is coupled to the memory and is used to execute the programs stored in the memory to implement the steps in the bridge dynamic strain signal denoising method described in any one of claims 1 to 4.
7. A computer-readable storage medium, characterized in that: Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps of the bridge dynamic strain signal denoising method described in any one of claims 1 to 4 above.
