SVMD combined wavelet threshold-based intelligent grinding system vibration signal denoising method and electronic equipment

Abstract

The invention discloses an intelligent grinding system vibration signal denoising method based on SVMD combined with a wavelet threshold and electronic equipment. The method comprises the following steps: acquiring vibration signals of the intelligent polishing detection equipment in all directions under different working conditions; decomposing the vibration signal into a plurality of mode components through continuous variational mode decomposition; calculating a correlation coefficient between the modal component and an original signal, and further distinguishing an effective component from a noise component; denoising the noise component by adopting a wavelet threshold to obtain a denoised component; and superposing the denoised component and the effective component to obtain a final denoised signal. Noise can be effectively separated under complex working conditions, noise interference is remarkably reduced, meanwhile, the original characteristics of signals are reserved to the maximum extent, the accuracy and stability of signal processing are improved, and the method is suitable for various complex signal environments.

Description

Technical Field

[0001] The present invention relates to the technical field of vibration signal processing of an intelligent polishing system, and more specifically, to a vibration signal denoising method and electronic equipment for an intelligent polishing system based on SVMD combined with wavelet threshold. Background Art

[0002] As an important maintenance tool to ensure the safe operation of railways, the intelligent rail grinding system must have efficient grinding capabilities and precise detection capabilities to achieve high-quality grinding effects in complex rail grinding working environments. The grinding process usually generates complex vibration signals, and these signals usually exhibit characteristics such as non-stationarity and strong noise. However, the vibration signals collected on-site contain multiple frequency components, and the noise will mask the true characteristics of the signal and affect subsequent analysis. Denoising can help improve the accuracy and reliability of the signal. Existing vibration signal denoising methods mainly include time domain methods, frequency domain methods, and time-frequency domain methods. Time domain methods include mean filtering and Kalman filtering, which usually rely on fixed filter parameters. These methods have poor adaptability to the non-stationarity and variability of signals and are difficult to effectively process complex vibration signals in intelligent polishing systems. Frequency domain methods include Fourier transform and wavelet transform, but they rely on the frequency domain characteristics of the signal. Although they can effectively remove noise in certain frequency bands, these methods usually assume that the noise has fixed frequency characteristics. For this reason, time-frequency analysis methods have gradually become the focus of research, and methods such as short-time Fourier transform and empirical mode decomposition have been introduced for signal decomposition and noise reduction, but these methods still have shortcomings in signal resolution and adaptability.

[0003] Therefore, it is necessary to develop a vibration signal denoising method and electronic equipment for an intelligent polishing system based on SVMD combined with wavelet threshold.

[0004] The information disclosed in the background technology section of the present invention is only intended to deepen the understanding of the general background technology of the present invention, and should not be regarded as acknowledging or suggesting in any form that the information constitutes the prior art already known to those skilled in the art. Summary of the invention

[0005] The present invention proposes a vibration signal denoising method and electronic equipment for an intelligent polishing system based on SVMD combined with wavelet threshold, which can effectively separate noise under complex working conditions and significantly reduce noise interference, while retaining the original characteristics of the signal to the greatest extent, improving the accuracy and stability of signal processing, and is suitable for a variety of complex signal environments.

[0006] In a first aspect, the embodiment of the present disclosure provides a vibration signal denoising method for an intelligent polishing system based on SVMD combined with wavelet threshold, comprising:

[0007] Obtain vibration signals of intelligent grinding and testing equipment in all directions under different working conditions;

[0008] Decomposing the vibration signal into multiple modal components through continuous variational mode decomposition;

[0009] Calculate the correlation coefficient between the modal component and the original signal to distinguish the effective component from the noise component;

[0010] The wavelet threshold is used to denoise the noise component to obtain the denoised component;

[0011] The denoised component is superimposed with the effective component to obtain the final denoised signal.

[0012] Preferably, decomposing the vibration signal into a plurality of modal components by continuous variational modal decomposition comprises:

[0013] Through continuous variational modal decomposition, the vibration signal is first decomposed into modal components and residual signals, and the residual signal is secondly decomposed into multiple modal components.

[0014] Preferably, the first constraint and the second constraint are determined for the first decomposition and the second decomposition respectively, the third constraint is determined based on the difference between the obtained modal component and the previous-order modal component, and the three constraints are minimized to obtain the modal component.

[0015] Preferably, the first constraint is:

[0016]

[0017] ω L is the center frequency of the Lth mode, and δ(t) is the Dirac distribution.

[0018] Preferably, the second constraint is:

[0019]

[0020] in,

[0021] Preferably, the third constraint is:

[0022]

[0023] in,

[0024] Preferably, the correlation coefficient is:

[0025]

[0026] Where x i represents the decomposed modal components, represents the mean value of the modal component, y i represents the original signal, Represents the mean of the original signal.

[0027] Preferably, distinguishing the effective component from the noise component comprises:

[0028] Determine the maximum correlation coefficient, and then determine the noise threshold. Define the component with a correlation coefficient greater than the noise threshold as the effective component, and the rest as noise components.

[0029] Preferably, denoising the noise component by using a wavelet threshold comprises:

[0030] The noise components are superimposed, and the db4 wavelet basis function and 5 wavelet decomposition layers are used to decompose the superimposed noise components to obtain 5 layers of high-frequency wavelet coefficients and 1 layer of low-frequency wavelet coefficients. The high-frequency wavelet coefficients of each layer are processed using a soft threshold, and the processed wavelet coefficients are reconstructed by wavelet to obtain the denoised components.

[0031] In a second aspect, an embodiment of the present disclosure further provides an electronic device, the electronic device comprising:

[0032] A memory storing executable instructions;

[0033] A processor runs the executable instructions in the memory to implement the vibration signal denoising method of the intelligent polishing system based on SVMD combined with wavelet threshold.

[0034] Its beneficial effects are:

[0035] (1) The present invention uses the SVMD method to process vibration signals, which can effectively decompose complex vibration signals into multiple frequency bands, each of which represents different signal characteristics. Through this decomposition method, the detailed features in the signal can be extracted more accurately, reducing the feature extraction error of traditional methods under high-frequency noise interference;

[0036] (2) When screening modal components, the present invention accurately classifies modal components through correlation coefficients to distinguish effective components from noise components. Only the noise components are denoised, while the effective components are directly used for signal reconstruction, avoiding the excessive weakening of the useful information of the original signal by traditional denoising methods, and can better retain the main features and information of the signal;

[0037] (3) Traditional denoising methods have limited adaptability to specific noise types, while the present invention can effectively adapt to mixed signals of different frequencies and characteristic noises by combining continuous variational mode decomposition with wavelet threshold denoising. Especially in complex backgrounds, the present invention can adaptively extract the noise component in the signal and perform denoising, making the denoising effect more stable and reliable in a variety of practical application scenarios.

[0038] The methods and apparatus of the present invention have other features and advantages that will be apparent from, or will be described in detail in, the accompanying drawings and subsequent detailed descriptions incorporated herein, which together serve to explain the specific principles of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The above and other objects, features and advantages of the present invention will become more apparent through a more detailed description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, wherein like reference numerals generally represent like components throughout the exemplary embodiments of the present invention.

[0040] Figure 1 A flowchart showing the steps of a vibration signal denoising method for an intelligent polishing system based on SVMD combined with wavelet threshold according to an embodiment of the present invention.

[0041] Figure 2 A schematic diagram of an original signal in the time domain according to an embodiment of the present invention is shown.

[0042] Figure 3 A schematic diagram of the original signal frequency domain according to an embodiment of the present invention is shown.

[0043] Figure 4 A schematic diagram of the time domain and frequency domain of each modal component according to an embodiment of the present invention is shown.

[0044] Figure 5 A schematic diagram of a time domain signal after denoising according to an embodiment of the present invention is shown.

[0045] Figure 6 A schematic diagram of a denoised signal frequency domain according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0046] The preferred embodiments of the present invention will be described in more detail below. Although the preferred embodiments of the present invention are described below, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0047] Figure 1A flowchart showing the steps of a vibration signal denoising method for an intelligent polishing system based on SVMD combined with wavelet threshold according to an embodiment of the present invention.

[0048] like Figure 1 As shown, the vibration signal denoising method of the intelligent polishing system based on SVMD combined with wavelet threshold includes:

[0049] Step 101, obtaining vibration signals of the intelligent grinding detection equipment in various directions under different working conditions;

[0050] Step 102, decomposing the vibration signal into multiple modal components by continuous variational modal decomposition;

[0051] Step 103, calculating the correlation coefficient between the modal component and the original signal, and then distinguishing the effective component from the noise component;

[0052] Step 104, denoising the noise component using a wavelet threshold to obtain a denoised component;

[0053] Step 105: superimpose the denoised component with the effective component to obtain the final denoised signal.

[0054] In one example, decomposing a vibration signal into a plurality of modal components by continuous variational modal decomposition includes:

[0055] Through continuous variational modal decomposition, the vibration signal is first decomposed into modal components and residual signals, and the residual signal is secondly decomposed into multiple modal components.

[0056] In one example, a first constraint and a second constraint are determined for the first decomposition and the second decomposition, respectively, a third constraint is determined based on the difference between the obtained modal component and the previous-order modal component, and the three constraints are minimized to obtain the modal component.

[0057] In one example, the first constraint is:

[0058]

[0059] ω L is the center frequency of the Lth mode, and δ(t) is the Dirac distribution.

[0060] In one example, the second constraint is:

[0061]

[0062] in,

[0063] In one example, the third constraint is:

[0064]

[0065] in,

[0066] In one example, the correlation coefficient is:

[0067]

[0068] Where x i represents the decomposed modal components, represents the mean value of the modal component, y i represents the original signal, Represents the mean of the original signal.

[0069] In one example, distinguishing valid components from noise components includes;

[0070] Determine the maximum correlation coefficient, and then determine the noise threshold. Define the component with a correlation coefficient greater than the noise threshold as the effective component, and the rest as noise components.

[0071] In one example, using a wavelet threshold to perform denoising on a noise component includes:

[0072] The noise components are superimposed, and the db4 wavelet basis function and 5 wavelet decomposition layers are used to decompose the superimposed noise components to obtain 5 layers of high-frequency wavelet coefficients and 1 layer of low-frequency wavelet coefficients. The high-frequency wavelet coefficients of each layer are processed using a soft threshold, and the processed wavelet coefficients are reconstructed by wavelet to obtain the denoised components.

[0073] Specifically, the vibration signals of the intelligent grinding and detection equipment in various directions under different working conditions are obtained. The noisy vibration signal f(t) is decomposed into intrinsic mode functions using continuous variational mode decomposition (SVMD). Each intrinsic mode function represents the components of the signal at different frequencies, namely the modal components. The vibration signal f(t) is decomposed into two signals:

[0074] f(t)=u L (t)+f r (t)

[0075] Among them, u L (t) is the Lth modal component, f r (t) is the residual signal, and the residual signal is further decomposed to obtain a series of modal components:

[0076]

[0077] Among them, u L (t) represents each modal component, f u (t) indicates a signal that has not been processed yet.

[0078] In order to decompose the modal components, the following criteria are established:

[0079] The decomposed modal components are compact near the center frequency, so the minimization constraint is achieved:

[0080]

[0081] ω L is the center frequency of the Lth mode, and δ(t) is the Dirac distribution.

[0082] in u L (t) at the frequency of the effective component, the residual signal f r The energy of (t) should be minimized, a constraint that is obtained by using an appropriate filter with a frequency response β L (ω) is realized, and then the second minimization criterion is obtained:

[0083]

[0084]

[0085] By minimizing the criterion J 1 and J 2 , the Lth modal component is obtained, but this modal component may be one of the previously obtained L-1th modal components. To avoid the above situation, u L (t) There should be less capacity near the center frequency of the acquired mode. This constraint is related to J 2 The constraints are the same and the frequency response of the filter used is:

[0086]

[0087] To ensure that the Lth modal component and the unprocessed part of the signal can be reconstructed into the original signal, the last constraint is imposed, and the formula is as follows:

[0088]

[0089] When L-1 constraints are known, the Lth modal component can be transformed into a constrained minimization problem:

[0090]

[0091] Where α is the equilibrium J 1 , J 2 , J 3 By introducing the quadratic penalty term and the Lagrange multiplier term, the optimization problem with constraints can be converted into an unconstrained optimization problem, and the alternating direction multiplier method is used to iteratively solve it, and finally several modal components are obtained.

[0092] In order to effectively distinguish the effective component and the noise component in the modal component, the Pearson correlation coefficient between each modal component and the original signal is calculated, denoted as ρ i , the Pearson correlation coefficient is used to indicate the degree of correlation between two sets of data. If the Pearson correlation coefficient is closer to 1, it means that the degree of correlation between the two is higher. If the coefficient is closer to 0, it means that the degree of correlation between the two groups is lower. The calculation formula is as follows:

[0093]

[0094] Where x i represents the decomposed modal components, represents the mean value of the modal component, y i represents the original signal, Represents the mean of the original signal.

[0095] Compare all correlation coefficients, and the largest correlation coefficient is recorded as ρ max , and ρ i ≤ρ max The component of / 10 is defined as the noise component, ρ i >ρ max The component of / 10 is defined as the effective component.

[0096] The noise components are superimposed, and the db4 wavelet basis function and 5 wavelet decomposition layers are used to perform wavelet decomposition on the superimposed noise components to obtain 5 layers of high-frequency wavelet coefficients and 1 layer of low-frequency wavelet coefficients. The high-frequency wavelet coefficients of each layer are processed by soft thresholding, and the processed wavelet coefficients are reconstructed by wavelet to obtain the denoised components. The denoised components and the effective components are superimposed to obtain the final denoised signal.

[0097] The present invention also provides an electronic device, which includes: a memory storing executable instructions; a processor, which runs the executable instructions in the memory to implement the above-mentioned intelligent polishing system vibration signal denoising method based on SVMD combined with wavelet threshold.

[0098] To facilitate understanding of the solutions and effects of the embodiments of the present invention, two specific application examples are given below. Those skilled in the art should understand that the examples are only for facilitating understanding of the present invention, and any specific details thereof are not intended to limit the present invention in any way.

[0099] Example 1

[0100] Figure 2 A schematic diagram of an original signal in the time domain according to an embodiment of the present invention is shown.

[0101] Figure 3A schematic diagram of the original signal frequency domain according to an embodiment of the present invention is shown.

[0102] Figure 4 A schematic diagram of the time domain and frequency domain of each modal component according to an embodiment of the present invention is shown.

[0103] Collect vibration signals of on-site intelligent grinding and testing equipment, Figure 2 , Figure 3 It represents the time domain and frequency domain distribution of the original intelligent grinding system vibration signal. The signal is mainly composed of several signals with main frequency characteristics. The original signal f(t) is decomposed to obtain a series of intrinsic mode components, and the signal is decomposed into modal components with different frequency characteristics, denoted as {IMF 1 (t),IMF 2 (t),...,IMF n (t)}, Figure 4 It means that after the original signal is decomposed by SVMD, 10 modal components are obtained, showing the time domain and frequency domain distribution of each modal component. After the original signal is decomposed by SVMD, each modal component represents the component of the signal in different frequency ranges.

[0104] Calculate the IMF for each modal component i The Pearson correlation coefficient of (t) and the original signal f(t) is used to indicate the degree of correlation between two sets of data. If the mutual correlation coefficient is closer to 1, it means that the degree of correlation between the two is higher, and if it is closer to 0, it means that the correlation between the two is lower. The original signal is decomposed by SVMD to obtain 10 modal components, and the correlation coefficient of each modal component is shown in Table 1.

[0105] Table 1 Correlation coefficients of each mode

[0106]

[0107] The correlation coefficient of the i-th modal component is denoted by ρ i , the maximum correlation coefficient is denoted as ρ max , and filter out ρ i ≤ρ max The component of / 10 is defined as the noise component, and ρ i >ρ max / 10 is recorded as the effective component, and the noise component is superimposed. From Table 1, we can see that the largest correlation coefficient is recorded as ρ max It is 0.9296. Through calculation, it can be known that IMF1, IMF2, IMF6, IMF8, and IMF10 are noise components, and they are superimposed, and the remaining components are effective components.

[0108] The noise components obtained by superposition are denoised by wavelet soft thresholding, and the db4 wavelet basis is used to decompose 5 layers to obtain the corresponding wavelet coefficients. The high-frequency wavelet coefficients are denoised by soft thresholding, and finally the denoised components are obtained by wavelet reconstruction.

[0109] Figure 5 A schematic diagram of a time domain signal after denoising according to an embodiment of the present invention is shown.

[0110] Figure 6 A schematic diagram of a denoised signal frequency domain according to an embodiment of the present invention is shown.

[0111] The effective component and the denoised component are superimposed to obtain the denoised signal. Figure 5 , Figure 6 The figure shows the time domain and frequency domain distribution of the vibration signal of the intelligent polishing system after denoising. It can be seen that the frequency components of the denoised signal are simpler, and the main characteristic components of the signal are retained. At the same time, the influence of noise is suppressed, making the signal spectrum more concentrated, and the main frequency components are highlighted, avoiding the interference of noise on signal analysis and improving the overall quality of the signal.

[0112] Example 2

[0113] The present disclosure provides an electronic device, which includes: a memory storing executable instructions; a processor, which runs the executable instructions in the memory to implement the above-mentioned intelligent polishing system vibration signal denoising method based on SVMD combined with wavelet threshold.

[0114] An electronic device according to an embodiment of the present disclosure includes a memory and a processor.

[0115] The memory is used to store non-temporary computer-readable instructions. Specifically, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc.

[0116] The processor may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions. In one embodiment of the present disclosure, the processor is used to run the computer-readable instructions stored in the memory.

[0117] Those skilled in the art should be able to understand that in order to solve the technical problem of how to obtain a good user experience, the present embodiment may also include well-known structures such as a communication bus and an interface, and these well-known structures should also be included in the protection scope of the present disclosure.

[0118] For detailed description of this embodiment, reference may be made to the corresponding descriptions in the aforementioned embodiments, which will not be repeated here.

[0119] Those skilled in the art should understand that the purpose of the above description of the embodiments of the present invention is only to exemplarily illustrate the beneficial effects of the embodiments of the present invention, and is not intended to limit the embodiments of the present invention to any given examples.

[0120] The embodiments of the present invention have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A vibration signal denoising method for an intelligent polishing system based on SVMD combined with wavelet threshold, characterized in that: include: Obtain vibration signals in all directions of intelligent grinding and testing equipment under different working conditions; Decomposing the vibration signal into multiple modal components through continuous variational mode decomposition; Calculate the correlation coefficient between the modal component and the original signal to distinguish the effective component from the noise component; The wavelet threshold is used to denoise the noise component to obtain the denoised component; The denoised component is superimposed with the effective component to obtain the final denoised signal.

2. According to claim 1, the vibration signal denoising method of the intelligent polishing system based on SVMD combined with wavelet threshold, wherein: Decomposing vibration signals into multiple modal components by continuous variational mode decomposition includes: Through continuous variational modal decomposition, the vibration signal is first decomposed into modal components and residual signals, and the residual signal is secondly decomposed into multiple modal components.

3. According to claim 2, the vibration signal denoising method of the intelligent polishing system based on SVMD combined with wavelet threshold, wherein: The first constraint and the second constraint are determined for the first decomposition and the second decomposition respectively, the third constraint is determined based on the difference between the obtained modal component and the previous-order modal component, and the three constraints are minimized to obtain the modal component.

4. According to claim 3, the vibration signal denoising method of the intelligent polishing system based on SVMD combined with wavelet threshold, wherein: The first constraint is: ω L is the center frequency of the Lth mode, and δ(t) is the Dirac distribution.

5. According to claim 3, the vibration signal denoising method of the intelligent polishing system based on SVMD combined with wavelet threshold, wherein: The second constraint is: in, 6. The vibration signal denoising method of the intelligent polishing system based on SVMD combined with wavelet threshold according to claim 3, wherein: The third constraint is: in, 7. The vibration signal denoising method of the intelligent polishing system based on SVMD combined with wavelet threshold according to claim 1, wherein: The correlation coefficient is: Where x i represents the decomposed modal components, represents the mean value of the modal component, y i represents the original signal, Represents the mean of the original signal.

8. The vibration signal denoising method of an intelligent polishing system based on SVMD combined with wavelet threshold according to claim 1, wherein: Distinguishing effective components from noise components includes; Determine the maximum correlation coefficient, and then determine the noise threshold. Define the component with a correlation coefficient greater than the noise threshold as the effective component, and the rest as noise components.

9. The vibration signal denoising method of an intelligent polishing system based on SVMD combined with wavelet threshold according to claim 1, wherein: De-noising using wavelet threshold for noise components includes: The noise components are superimposed, and the db4 wavelet basis function and 5 wavelet decomposition layers are used to decompose the superimposed noise components to obtain 5 layers of high-frequency wavelet coefficients and 1 layer of low-frequency wavelet coefficients. The high-frequency wavelet coefficients of each layer are processed using a soft threshold, and the processed wavelet coefficients are reconstructed by wavelet to obtain the denoised components.

10. An electronic device, characterized in that: The electronic device comprises: A memory storing executable instructions; A processor, wherein the processor runs the executable instructions in the memory to implement the vibration signal denoising method for an intelligent polishing system based on SVMD combined with wavelet threshold according to any one of claims 1 to 9.

Images