Fault feature characterization method matched with extrusion domain mechanical signal

By calculating the matching extrusion transformation of mechanical fault signals under the synchronous extrusion transformation framework, the problem of traditional time-frequency analysis methods in emphasizing the divergence phenomenon in frequency signal characterization is solved, and more efficient fault characteristic characterization is achieved, which improves the mechanical signal Fault diagnosis accuracy.

CN118940149APending Publication Date: 2024-11-12CHENGDU UNIVERSITY OF TECHNOLOGY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410996420.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-24
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

Traditional time-frequency analysis methods have divergence when characterizing mechanical signals, especially when processing emphasize frequency signals, it is difficult to provide clear fault characteristics, limiting the fault diagnosis effect of mechanical signals.

Method used

Under the framework of synchronous extrusion transformation, by calculating the short-time Fourier transform and instantaneous frequency estimator of the mechanical fault signal, the matching instantaneous frequency estimator of the frequency band center, upper boundary and lower boundary is further calculated to obtain the matching extrusion transformation. This realizes fault characteristic characterization.

Benefits of technology

It significantly improves the energy focus of the time frequency distribution and improves the fault characteristic characterization accuracy of fault mechanical signals, especially when processing emphasize frequency signals, providing clearer time frequency characterization results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118940149A_ABST
    Figure CN118940149A_ABST
Patent Text Reader

Abstract

The invention discloses a fault feature characterization method matched with an extrusion domain mechanical signal, and the method comprises the following steps: S1, inputting a mechanical fault signal x (t) to be analyzed, t being time; s2, a window function # imgabs0 # is preset, sigma is a standard deviation, short-time Fourier transform Sx (t, omega) of the mechanical fault signal is calculated, and an instantaneous frequency estimator # imgabs1 # omega of the short-time Fourier transform Sx (t, omega) is calculated to represent frequency; s3, calculating a frequency band center # imgabs, a frequency band upper boundary # imgabs, a frequency band lower boundary # imgabs, a matching instantaneous frequency estimator # imgabs, a matching instantaneous frequency estimator # imgabs, a matching instantaneous frequency estimator # imgabs, a matching instantaneous frequency estimator # imgabs, a matching instantaneous frequency estimator # imgabs, and a matching instantaneous frequency estimator # imgabs, s4, calculating matched extrusion transformation Mx (t, omega) according to the relevant parameters calculated in the step S3; and S5, according to the matching extrusion transformation Mx (t, omega) in the step S4, performing fault feature representation on the mechanical fault signal.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of mechanical signal processing, and in particular relates to a fault feature characterization method for matching extrusion domain mechanical signals. Background Art

[0002] The time-frequency information characterization of mechanical equipment has always been one of the important links in fault diagnosis. Traditional time-frequency analysis methods mainly include short-time Fourier transform (STFT), wavelet transform (WT), S transform (ST) and generalized S transform (GST). They all use window functions to divide the signal, and then perform Fourier transform on it separately, thereby providing joint distribution information of time and frequency. However, limited by the Heisenberg uncertainty principle, the time-frequency distribution energy provided by traditional methods is very divergent, resulting in a fuzzy time-frequency characterization result. However, the time-frequency characterization with strong centralization can better characterize the fault characteristics of mechanical signals. For this reason, time-frequency post-processing methods have emerged.

[0003] Among them, the synchronized squeeze transform (SST) uses an instantaneous frequency estimator in the frequency direction to squeeze the scattered time-frequency coefficients onto the instantaneous frequency trajectory, which can greatly improve the focus of the time-frequency representation and support signal reconstruction. However, this method is only applicable to relatively stable signals. When characterizing strong frequency signals, the time-frequency representation results obtained by this method also have divergence. This is not conducive to the fault characterization of mechanical signals. Therefore, it is urgent to study a time-frequency analysis method for characterizing strong frequency signals. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a method for characterizing fault characteristics of mechanical signals in the matching extrusion domain. The present invention studies the characteristics of the signal frequency band under the framework of synchronous extrusion transform (SST) and constructs a new transform. The invention provides a new time-frequency transform method for characterizing the stressed frequency signal.

[0005] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is: a fault feature characterization method matching the mechanical signal of the extrusion domain, comprising the following steps:

[0006] S1. Input the mechanical fault signal x(t) to be analyzed, where t is time;

[0007] S2, preset window function Where σ is the standard deviation, and the short-time Fourier transform S of the mechanical fault signal is calculated. x (t,ω), and calculate its instantaneous frequency estimator

[0008]

[0009] in, is the imaginary unit of the complex number, u is the time variable, ω represents the frequency, Indicates S x The partial derivative of (t,ω) with respect to t, It means taking the real part of a complex number;

[0010] S3. Calculate the center of the frequency band Upper band edge Band lower edge Matching instantaneous frequency estimator Where k = 1, 2…K, K is the number of signal components;

[0011] S4, according to the relevant parameters calculated in step S3, calculate the matching extrusion transformation M x (t,ω);

[0012] S5, according to the matching extrusion transformation M in step S4 x (t,ω), characterize the fault characteristics of the mechanical fault signal.

[0013] Preferably, in step S3 and is represented as:

[0014]

[0015] in, dω is the frequency discrete interval;

[0016]

[0017] in, is the threshold value. It is recommended to take min(·) means taking the minimum value, represents the partial derivative with respect to ω;

[0018]

[0019] in, is the threshold value. It is recommended to take max(·) means taking the maximum value, represents the partial derivative with respect to ω;

[0020]

[0021] Preferably, the matching extrusion transformation M in step S4 x (t,ω) is defined as:

[0022]

[0023] where η is the frequency variable.

[0024] The idea of ​​the present invention is:

[0025] First, input the mechanical fault signal x(t) to be analyzed, where t is time;

[0026] Second, preset window function Where σ is the standard deviation, and the short-time Fourier transform S of the mechanical fault signal is calculated. x (t,ω), and calculate its instantaneous frequency estimator Where ω represents the frequency;

[0027] Third, calculate the center of the frequency band Upper band edge Band lower edge Matching instantaneous frequency estimator Where k = 1, 2…L, L is the number of signal components;

[0028] Fourth, according to the calculated relevant parameters, calculate the matching extrusion transformation M x (t,ω);

[0029] Fifth, according to the matching extrusion transformation M x (t,ω), characterize the fault characteristics of the mechanical fault signal.

[0030] The working principle of the present invention is as follows: input the mechanical fault signal x(t) to be analyzed, where t is time; preset the window function Where σ is the standard deviation, and the short-time Fourier transform S of the mechanical fault signal is calculated. x (t,ω), and calculate its instantaneous frequency estimator Where ω represents the frequency; calculate the center of the frequency band Upper band edge Band lower edge Matching instantaneous frequency estimator Where k = 1, 2 ... K, K is the number of signal components; according to the calculated relevant parameters, the matching squeeze transformation M is calculated. x (t,ω); according to the matching extrusion transformation M x (t, ω), and characterize the fault characteristics of the mechanical fault signal. The present invention can provide a new time-frequency transformation method for characterizing the strong frequency signal to characterize the characteristics of the faulty mechanical signal.

[0031] In view of the difficulties encountered by synchronized squeeze transform (SST) in characterizing stressed frequency signals, the present invention proposes a fault feature characterization method for matching squeeze domain mechanical signals. First, the faulty mechanical signal to be analyzed is input and its short-time Fourier transform result and instantaneous frequency estimator are calculated; then its frequency band center, frequency band upper boundary, frequency band lower boundary are calculated, and the instantaneous frequency estimator is matched; finally, a matching squeeze transform is obtained; the present invention shows good results in time-frequency characterization capability. When characterizing stressed frequency signals, it can significantly improve the energy focusing of the time-frequency distribution and improve the characterization accuracy of the fault characteristics of the faulty mechanical signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a flow chart of the present invention;

[0033] Figure 2 is the input signal of faulty mechanical equipment;

[0034] Figure 3 Spectral time spectrum diagrams processed by (a) synchronous extrusion transformation and (b) the method of the present invention respectively. DETAILED DESCRIPTION

[0035] The present invention will be further described below in conjunction with the accompanying drawings.

[0036] Example 1: See Figure 1 , a fault feature characterization method matching an extrusion domain mechanical signal, comprising the following steps:

[0037] S1. Input the mechanical fault signal x(t) to be analyzed, where t is time;

[0038] S2, preset window function Where σ is the standard deviation, and the short-time Fourier transform S of the mechanical fault signal is calculated. x (t,ω), and calculate its instantaneous frequency estimator

[0039]

[0040] in, is the imaginary unit of the complex number, u is the time variable, ω represents the frequency, Indicates S x The partial derivative of (t,ω) with respect to t, It means taking the real part of a complex number;

[0041] S3. Calculate the center of the frequency band Upper band edge Band lower edge Matching instantaneous frequency estimator Where k = 1, 2…K, K is the number of signal components;

[0042] S4, according to the relevant parameters calculated in step S3, calculate the matching extrusion transformation M x (t,ω);

[0043] S5, according to the matching extrusion transformation M in step S4 x (t,ω), characterize the fault characteristics of the mechanical fault signal.

[0044] Preferably, in step S3 and is represented as:

[0045]

[0046] in, dω is the frequency discrete interval;

[0047]

[0048] in, is the threshold value. It is recommended to take min(·) means taking the minimum value, represents the partial derivative with respect to ω;

[0049]

[0050] in, is the threshold value. It is recommended to take max(·) means taking the maximum value, represents the partial derivative with respect to ω;

[0051]

[0052] Preferably, the matching extrusion transformation M in step S4 x (y,ω) is defined as:

[0053]

[0054] where η is the frequency variable.

[0055] See also Figures 1 to 3 , let's take a faulty mechanical equipment signal as an example, Figure 2 The signal of the faulty mechanical equipment is displayed, with the horizontal axis representing time in seconds and the vertical axis representing amplitude. Figure 3The time-frequency spectrum is respectively processed by (a) synchronous extrusion transformation and (b) the method of the present invention, the horizontal axis represents time in seconds, and the vertical axis represents frequency in Hertz. The embodiment proves that the result obtained after the method of the present invention has better focusing, can better characterize the fault signal and other strong frequency signals, and thus achieve better fault feature characterization effect.

[0056] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A fault feature characterization method matching an extrusion domain mechanical signal, characterized in that: The following steps are involved: S1. Input the mechanical fault signal x(t) to be analyzed, where t is time; S2, preset window function Where σ is the standard deviation, and the short-time Fourier transform S of the mechanical fault signal is calculated. x (t,ω), and calculate its instantaneous frequency estimator in, is the imaginary unit of the complex number, u is the time variable, ω represents the frequency, Indicates S x The partial derivative of (t,ω) with respect to t, It means taking the real part of a complex number; S3. Calculate the center of the frequency band Upper band edge Band lower edge Matching instantaneous frequency estimator Where k = 1, 2…K, K is the number of signal components; S4, according to the relevant parameters calculated in step S3, calculate the matching extrusion transformation M x (t,ω); S5, according to the matching extrusion transformation M in step S4 x (t,ω), characterize the fault characteristics of the mechanical fault signal.

2. A method for characterizing fault characteristics of matching extrusion domain mechanical signals according to claim 1, characterized in that: In step S3 and is represented as: in, dω is the frequency discrete interval; in, is the threshold value. It is recommended to take min(·) means taking the minimum value, represents the partial derivative with respect to ω; in, is the threshold value. It is recommended to take max(·) means taking the maximum value, represents the partial derivative with respect to ω; 3. The method for characterizing fault characteristics of matching extrusion domain mechanical signals according to claim 1 is characterized in that: The matching extrusion transformation M in step S4 x (t,ω) is defined as: where η is the frequency variable.