An early fault diagnosis method for rolling bearings based on adaptive variational mode decomposition

Through adaptive variational modal decomposition and sparrow algorithm optimization parameters, the problems of weak early fault characteristics and parameters need to be determined in advance are solved, and accurate fault diagnosis of rolling bearings and reduced noise interference are achieved, and diagnostic efficiency is improved.

CN115062665BActive Publication Date: 2024-11-19TAIYUAN UNIVERSITY OF TECHNOLOGY
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202210756586.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-06-30
Publication Date
2024-11-19
Estimated Expiration
2042-06-30

AI Technical Summary

Technical Problem

The early failure characteristics of rolling bearings are weak and the parameters in the variational modal decomposition need to be determined in advance, which affects the calculation efficiency and diagnostic accuracy.

Method used

Adaptive variational modal decomposition method is adopted to search the variational modal decomposition parameters through the sparrow algorithm, calculate the comprehensive fault index, filter the modal components and perform envelope demodulation analysis to determine the fault type.

Benefits of technology

It realizes accurate diagnosis of early rolling bearing failures, reduces noise interference, and improves the accuracy and diagnostic efficiency of fault feature extraction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115062665B_ABST
    Figure CN115062665B_ABST
Patent Text Reader

Abstract

The present invention belongs to the technical field of signal analysis and fault diagnosis, and discloses a rolling bearing early fault diagnosis method based on adaptive variational modal decomposition, comprising the following steps: S1, collecting the vibration signal of the rolling bearing; S2, taking the minimum comprehensive fault index as the objective function, and combining the sparrow algorithm to calculate the variational modal decomposition parameters (K, α) corresponding to the best sparrow individual position as the decomposition parameters, and performing variational modal decomposition on the original vibration signal; S3, removing the modal components less than the average power spectrum kurtosis to obtain a reconstructed signal; S4, performing envelope demodulation analysis on the reconstructed signal, so as to determine the fault type of the bearing. The present invention performs adaptive modal decomposition through the comprehensive fault index, can realize the separation of fault impact and background noise, and improve the accuracy of fault diagnosis.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of signal analysis and fault diagnosis, and in particular relates to an early fault diagnosis method of a rolling bearing based on adaptive variational mode decomposition. Background Art

[0002] As an important component of rotating machinery, the operating state of rolling bearings directly affects the working efficiency and overall performance of the mechanical system. During its operation, various faults are prone to occur. Early faults have the characteristics of weak characteristic information and low signal-to-noise ratio, which makes the generated periodic pulse signals often submerged in the background noise and difficult to identify and extract.

[0003] As a new method with a solid theoretical foundation, variational mode decomposition can separate signal components containing fault characteristics from the original vibration signal with low signal-to-noise ratio. Variational mode decomposition (VMD) has high computational efficiency and good noise robustness, and is widely used in many fields. However, when the VMD algorithm processes signals, it is necessary to predetermine parameters such as the number of modes K and the quadratic penalty factor α, which seriously affects the computational efficiency of VMD.

[0004] In order to improve the performance of VMD in the fault field, the parameters of VMD are optimized to achieve adaptive decomposition. Indicators such as set kurtosis, envelope spectrum kurtosis, and envelope entropy have been constructed in the prior art. These indicators help solve the problem of background noise interference and achieve accurate diagnosis of bearing faults. However, these indicators may lose some important information during fault diagnosis. Therefore, it is necessary to provide a new fault diagnosis method to optimize the parameters of VMD through more accurate fault indicators, extract weak fault features, and achieve accurate diagnosis of fault location and type. Summary of the invention

[0005] In view of the problems that the early fault characteristics of rolling bearings are weak and difficult to extract and the parameters in variational modal decomposition need to be determined in advance, the present invention proposes an adaptive variational modal decomposition method to achieve accurate diagnosis of early faults of rolling bearings.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: a rolling bearing early fault diagnosis method based on adaptive variational mode decomposition, comprising the following steps:

[0007] S1, collecting the vibration signal f(n) of the rolling bearing;

[0008] S2. Set the value range of the variational modal decomposition parameter and initialize the modal decomposition parameter; perform variational modal decomposition on the collected vibration signal f(n) according to the variational modal decomposition parameter, calculate the comprehensive fault index of each modal component, and select the minimum value as the comprehensive fault index under the current sparrow individual position; use the minimum comprehensive fault index as the objective function, use the sparrow algorithm to perform a global search on the variational modal decomposition parameter until the best sparrow individual position is obtained; use the variational modal decomposition parameter (K, α) corresponding to the best sparrow individual position as the decomposition parameter, perform variational modal decomposition on the original vibration signal, and obtain K modal components; wherein K represents the number of modes and α represents the penalty factor;

[0009] S3, calculating the power spectrum kurtosis of each modal component, comparing it with the average power spectrum kurtosis, discarding the modal components with a kurtosis smaller than the average power spectrum kurtosis, and adding the remaining modal components to obtain the reconstructed signal;

[0010] S4. Perform envelope demodulation analysis on the reconstructed signal to determine the fault type of the bearing.

[0011] In step S2, the calculation process of the comprehensive fault index is:

[0012] Perform Hilbert transform on each modal component obtained by variational modal decomposition to obtain the transformed signal g H (t);

[0013] According to the transformed signal g H (t) Calculate the power spectrum S(k) of each modal component;

[0014] Calculate the power spectrum kurtosis Q and power spectrum entropy H of each modal component according to the power spectrum S(k);

[0015] The comprehensive fault index CFI of each modal component is calculated according to the power spectrum kurtosis Q and power spectrum entropy H. The calculation formula is:

[0016] CFI=H·Q;.

[0017] In step S2, the power spectrum S(k) of each modal component is calculated as follows:

[0018]

[0019]

[0020] Where DFT represents discrete Fourier transform function.

[0021] In step S2, the calculation formulas for the power spectrum kurtosis Q and the power spectrum entropy H of each modal component are:

[0022]

[0023]

[0024] In the formula, m 4 Represents the fourth-order center distance, m 2 represents the second-order central moment, and P(S(k)) represents the probability that the power spectrum amplitude is S(k).

[0025] In step S3, the power spectrum kurtosis of each modal component is calculated as follows:

[0026] S301, calculate the power spectrum S(k) of each modal component, the calculation formula is:

[0027]

[0028] in, DFT represents discrete Fourier transform function; g(n) represents the modal component after the vibration signal is decomposed;

[0029] S302, calculate the power spectrum kurtosis KPS of each modal component i , the calculation formula is:

[0030]

[0031] Among them, m 4 Represents the fourth-order center distance, m 2 represents the second-order central moment, and P(S(k)) represents the probability that the power spectrum amplitude is S(k).

[0032] In step S3, the calculation formula of the average power spectrum kurtosis AKPS is:

[0033]

[0034] In step S2, the setting range of the variational mode decomposition parameters is: K∈[2, 10], α∈[100, 5000].

[0035] In step S2, when the sparrow algorithm is used to perform a global search on the variational modal decomposition parameters, if the set number of iteration cycles is reached, the sparrow individual position corresponding to the minimum value of the comprehensive fault index under the current individual sparrow positions is output as the best individual sparrow.

[0036] In the step S1, the vibration signal f(n) is collected by a vibration sensor and a data acquisition device.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. The present invention proposes a rolling bearing early fault diagnosis method based on adaptive variational modal decomposition, taking the minimum comprehensive fault index as the objective function, using the sparrow search algorithm to adaptively search for the optimal parameters of VMD, and performing variational modal decomposition. The power spectrum kurtosis of the modal components after decomposition is calculated, the components with larger power spectrum kurtosis are screened to reconstruct the signal, and the reconstructed signal is subjected to envelope demodulation analysis to determine the fault type of the bearing.

[0039] 2. The present invention proposes a new fault diagnosis index - comprehensive fault index, which can decompose complex signals into multiple components and separate fault impact and background noise.

[0040] 3. The present invention sets the average power spectrum kurtosis as a threshold, compares the power spectrum kurtosis of the signal with the threshold, removes the modal components smaller than the threshold, and reduces the interference of noise. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A schematic flow chart of a rolling bearing early fault diagnosis method based on adaptive variational mode decomposition provided by an embodiment of the present invention;

[0042] Figure 2 Schematic diagram of the process of the adaptive variational mode decomposition method in an embodiment of the present invention;

[0043] Figure 3 Schematic diagram of the experimental signal and its envelope spectrum used in the embodiment of the present invention;

[0044] Figure 4 A schematic diagram of various modal component signals and their envelope spectra obtained by performing adaptive variational modal decomposition processing on experimental signals according to an embodiment of the present invention;

[0045] Figure 5 Schematic diagram of the reconstructed signal and envelope spectrum in an embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0047] like Figure 1 As shown, an embodiment of the present invention provides a rolling bearing early fault diagnosis method based on adaptive variational mode decomposition, comprising the following steps:

[0048] S1. Collect the vibration signal f(n) of the rolling bearing.

[0049] Specifically, the vibration signal f(n) is collected by a vibration sensor and a data acquisition device. Specifically, the vibration signal f(n) can be collected by an acceleration sensor fixed on the bearing seat on the driving motor side and a related data acquisition device.

[0050] S2. Decompose the vibration signal f(n) by adaptive variational modal decomposition to obtain K modal components.

[0051] In this embodiment, firstly, the value range of the variational modal decomposition parameter (K, α) is set and initialized; the collected vibration signal f(n) is subjected to variational modal decomposition according to the initial parameters, and the comprehensive fault index is calculated; the minimum comprehensive fault index is used as the objective function, the Sparrow search algorithm (SSA) is used to perform a global search for the parameter (K, α), the parameter (K, α) corresponding to the best individual sparrow position is used as the decomposition parameter, the original vibration signal is subjected to variational modal decomposition, and K modal components are obtained.

[0052] Furthermore, the specific steps of step S2 are:

[0053] S201, initialize the sparrow search algorithm parameters, set the range of variational modal decomposition parameters K and α, in this embodiment, K∈[2, 10], α∈[100, 5000], determine the comprehensive fault index as the fitness function; the individual position of the sparrow corresponds to the variational modal decomposition parameters K and α.

[0054] S202, performing variational modal decomposition on the vibration signal f(n), calculating the comprehensive fault index of each modal component, and selecting the minimum value therein as the comprehensive fault index at the current individual position of the sparrow.

[0055] In this embodiment, the calculation method of the comprehensive fault index of each modal component is:

[0056] S2021. Perform Hilbert transform on each modal component after variational modal decomposition to obtain a transformed signal g H (n); its expression is:

[0057] g H (n) = Hilbert{g(n)}; (1)

[0058] Among them, g(n) represents a modal component after the vibration signal is decomposed.

[0059] S2022, according to the transformed signal g H (n) Calculate the power spectrum S(k) of the signal;

[0060] In this embodiment, the obtained transformation signal g H (n) performs N samplings and performs discrete Fourier transform to obtain the power spectrum S(k) of the signal. The calculation formula of the power spectrum S(k) of the signal is:

[0061]

[0062] Where DFT represents the discrete Fourier transform function, which is expressed as:

[0063]

[0064] Among them, n represents the length of the signal, k represents the length of the power spectrum, and N represents the number of sampling times.

[0065] S2023. Calculate the power spectrum kurtosis Q and the power spectrum entropy H according to the power spectrum S(k).

[0066] Among them, the calculation formulas of power spectrum kurtosis Q and power spectrum entropy H are:

[0067]

[0068]

[0069] In the formula, m 4 Represents the fourth-order center distance, m 2 represents the second-order central moment, and P(S(k)) represents the probability that the power spectrum amplitude is S(k).

[0070] S2024. Calculate the comprehensive fault index CFI (Comprehensive fault index) of each modal component according to the power spectrum kurtosis Q and the power spectrum entropy H. The calculation formula is:

[0071] CFI=H·Q; (6)

[0072] S2025. Compare the values ​​of the comprehensive fault index CFI of each modal component, and take the minimum value as the comprehensive fault index of the current sparrow individual position.

[0073] S203, using the comprehensive fault index as the fitness function for global search, continuously updating the position of the individual sparrow, repeating step S202, until the variational modal decomposition parameter (K, α) corresponding to the minimum value of the fitness function is found, or, when the set number of iteration cycles is reached, outputting the variational modal decomposition parameter (K, α) corresponding to the minimum value of the current fitness function, the variational modal decomposition parameter is the optimal individual sparrow position.

[0074] S204. Using the obtained variational modal decomposition parameters (K, α) as decomposition parameters, perform variational modal decomposition on the original vibration signal to obtain K modal components.

[0075] In this embodiment, the objective function objf used by the sparrow algorithm is the minimum comprehensive fault index (MCFI), which can be expressed as:

[0076]

[0077] S3. Calculate the power spectrum kurtosis of each modal component, compare it with the average power spectrum kurtosis, discard the modal components with a kurtosis smaller than the average power spectrum kurtosis, add the remaining modal components, and obtain the reconstructed signal.

[0078] In step S3, the power spectrum kurtosis calculation steps of each modal component are as follows:

[0079] S301, calculate the power spectrum S(k) of each modal component, the calculation formula is:

[0080]

[0081] in, DFT represents discrete Fourier transform function; g(n) represents a modal component after the vibration signal is decomposed.

[0082] S302, calculate the power spectrum kurtosis KPS of each modal component i , the calculation formula is:

[0083]

[0084] Among them, m 4 Represents the fourth-order center distance, m 2 represents the second-order central moment, P(S(k)) represents the probability that the power spectrum amplitude is S(k), KPS i represents the kurtosis of the power spectrum of the i-th modal component.

[0085] In step S3, the calculation formula of the average power spectrum kurtosis AKPS is:

[0086]

[0087] S4. Perform envelope demodulation analysis on the reconstructed signal to determine the fault type of the bearing.

[0088] In this embodiment, the experimental data comes from the Case Western Reserve University Bearing Data Center. A strong noise with a signal-to-noise ratio of 2 dB is added to the fault signal. The original signal and its envelope spectrum are as follows: Figure 3As shown, it can be found that the signal is subject to strong noise interference and the fault characteristics are completely submerged. Through the diagnostic method of the embodiment of the present invention, the minimum comprehensive fault index is used as the objective function, and the sparrow algorithm is used to adaptively search for the optimal parameters of VMD. The obtained parameters are: the number of modes K = 6, the penalty factor α = 1338, and the vibration signal is subjected to variational mode decomposition to obtain 6 IMF components, which are respectively recorded as IMF1~IMF6, as shown in Figure 4 As shown, it is a schematic diagram of each modal component signal IMF1~IMF6 and its envelope spectrum.

[0089] The power spectrum kurtosis of each modal component is calculated and compared with the average power spectrum kurtosis. The modal components with a kurtosis smaller than the average power spectrum kurtosis are discarded, that is, modal components 1, 2, and 6 are discarded to reduce the interference of noise. The remaining modal components 3, 4, and 5 are added to obtain the reconstructed signal. The threshold screening results are shown in Table 1.

[0090] Table 1 Threshold screening results

[0091] IMF 1 2 3 4 5 6 KPS 75.5597 42.6570 226.6138 226.3041 176.0750 77.1499 choose √ √ √

[0092] The reconstructed signal is Figure 5 As shown in (a), the reconstructed signal is subjected to envelope demodulation analysis, and the envelope spectrum obtained is as follows Figure 5 As shown in (b), the inner ring fault frequency and its multiple frequencies can be clearly seen from the envelope spectrum, thus confirming that the bearing has an inner ring fault.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A rolling bearing early fault diagnosis method based on adaptive variational mode decomposition, characterized in that: The following steps are involved: S1, collecting the vibration signal f (n) of the rolling bearing; S2. Set the value range of the variational modal decomposition parameters and initialize the modal decomposition parameters; perform variational modal decomposition on the collected vibration signal f (n) according to the variational modal decomposition parameters, calculate the comprehensive fault index of each modal component, and select the minimum value as the comprehensive fault index under the current sparrow individual position; use the minimum comprehensive fault index as the objective function, use the sparrow algorithm to perform a global search for the variational modal decomposition parameters until the best sparrow individual position is obtained; use the variational modal decomposition parameters (K, α) corresponding to the best sparrow individual position as decomposition parameters, perform variational modal decomposition on the original vibration signal, and obtain K modal components; where K represents the number of modes and α represents the penalty factor; S3, calculating the power spectrum kurtosis of each modal component, comparing it with the average power spectrum kurtosis, discarding the modal components with a kurtosis smaller than the average power spectrum kurtosis, and adding the remaining modal components to obtain the reconstructed signal; S4, performing envelope demodulation analysis on the reconstructed signal to determine the fault type of the bearing; in step S2, the calculation process of the comprehensive fault index is: Perform Hilbert transform on each modal component obtained by variational modal decomposition to obtain the transformed signal g H ( t ); According to the conversion signal g H ( t ) Calculate the power spectrum of each modal component S ( k ); According to the power spectrum S ( k ) Calculate the power spectrum kurtosis Q and power spectrum entropy H of each modal component; The comprehensive fault index of each modal component is calculated based on the power spectrum kurtosis Q and power spectrum entropy H CFI , the calculation formula is: CFI = H · Q 。 2. The rolling bearing early fault diagnosis method based on adaptive variational mode decomposition according to claim 1 is characterized in that: In step S2, the power spectrum of each modal component S ( k ) is calculated as: ; ; Where DFT represents discrete Fourier transform function.

3. The rolling bearing early fault diagnosis method based on adaptive variational mode decomposition according to claim 2 is characterized in that: In step S2, the calculation formulas for the power spectrum kurtosis Q and the power spectrum entropy H of each modal component are: ; ; In the formula, m 4 Represents the fourth-order center distance, m 2 represents the second-order central moment, and P(S(k)) represents the probability that the power spectrum amplitude is S(k).

4. The rolling bearing early fault diagnosis method based on adaptive variational mode decomposition according to claim 1 is characterized in that: In step S3, the power spectrum kurtosis of each modal component is calculated as follows: S301, calculate the power spectrum of each modal component S ( k ), the calculation formula is: ; in, ; DFT represents discrete Fourier transform function; g(n) represents the modal component after vibration signal decomposition; S302, calculate the power spectrum kurtosis of each modal component KPS i , The calculation formula is: ; Among them, m 4 Represents the fourth-order center distance, m 2 represents the second-order central moment, and P(S(k)) represents the probability that the power spectrum amplitude is S(k).

5. The rolling bearing early fault diagnosis method based on adaptive variational mode decomposition according to claim 4 is characterized in that: In step S3, the average power spectrum kurtosis AKPS The calculation formula is: 。 6. The rolling bearing early fault diagnosis method based on adaptive variational mode decomposition according to claim 1 is characterized in that: In step S2, the setting range of the variational mode decomposition parameters is: K∈[2, 10], α∈[100, 5000].

7. According to the method for early fault diagnosis of rolling bearings based on adaptive variational modal decomposition as described in claim 1, in step S2, when the sparrow algorithm is used to perform a global search on the variational modal decomposition parameters, if the set number of iteration cycles is reached, the sparrow individual position corresponding to the minimum value of the comprehensive fault index under the current sparrow individual positions is output as the best sparrow individual.

8. The rolling bearing early fault diagnosis method based on adaptive variational mode decomposition according to claim 1 is characterized in that: In step S1, the vibration signal f (n) Collected through vibration sensors and data acquisition equipment.

Citation Information

Cited By

  • Signal mode decomposition method and system for early fault diagnosis of rotating machinery

    CN119829991A

  • Signal Modal Decomposition Method and System for Early Fault Diagnosis of Rotating Machinery

    CN119829991B