Rotating machinery impact fault feature extraction method and device and storage medium
By calculating the envelope spectrum kurtosis and squared envelope autocorrelation kurtosis of the vibration signal of rotating machinery impact response, the optimal demodulation frequency band is determined, which solves the problem of inaccurate demodulation frequency band selection in the prior art and realizes accurate extraction and diagnosis of rotating machinery impact fault characteristics.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- LIAOHE GASOLINEEUM EXPLORATION BUREAU CO LTD
- Filing Date
- 2024-11-27
- Publication Date
- 2026-05-29
AI Technical Summary
Existing technologies using envelope demodulation to extract features of rotating machinery impact faults suffer from inaccurate demodulation frequency band selection, leading to incorrect impact fault diagnosis results. This is especially problematic under strong non-Gaussian noise interference, where it is difficult to accurately extract fault features.
By calculating the envelope kurtosis of each sub-band signal, non-cyclic stationary frequency bands are marked step by step. The squared envelope autocorrelation kurtosis of the unmarked sub-band signal is calculated to determine the optimal demodulation frequency band. The impact response vibration signal within the optimal demodulation frequency band is then envelope demodulated to extract the impact fault characteristics of rotating machinery.
It enables accurate extraction and diagnosis of impact fault characteristics in rotating machinery under strong non-Gaussian noise interference, thus improving the accuracy of fault diagnosis.
Smart Images
Figure CN122108598A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of equipment fault diagnosis technology, specifically relating to a method for extracting features of rotating machinery impact faults, a device for extracting features of rotating machinery impact faults, a computer device, and a machine-readable storage medium. Background Technology
[0002] Due to the complex and variable operating environment, rolling bearings in rotating machinery are prone to deterioration, leading to equipment failure. Currently, numerous vibration signal fault diagnosis methods have been proposed and have achieved some success. However, under some complex operating conditions, the signal-to-noise ratio of vibration signals is low and there is non-Gaussian noise interference, which not only increases the difficulty of extracting impact fault features but also affects the accuracy of the diagnostic results.
[0003] To extract periodic impulse signals from noise, those skilled in the art have dedicated themselves to researching signal processing methods with superior performance, such as wavelet transform, blind deconvolution, adaptive signal decomposition, and demodulation analysis. Among these, envelope demodulation is a commonly used method for impulse fault feature extraction, offering wide applicability and high efficiency. Envelope demodulation primarily eliminates noise interference in the original signal by filtering the frequency band containing the fault. The core of this method lies in the selection of the demodulation frequency band; its correctness directly affects the accuracy of the demodulated impulse fault features. Most existing methods, such as kurtosis spectrum and autospectral analysis, rely on sparsity indices to design fault feature evaluation methods and select the optimal demodulation frequency band. However, when facing strong non-Gaussian noise interference, the interference frequency band is often selected instead of the fault band, leading to erroneous impulse fault diagnosis results.
[0004] Therefore, it is urgent to improve the method for selecting the optimal demodulation frequency band. Summary of the Invention
[0005] The purpose of this application is to provide a method for extracting features of rotating machinery impact faults, a device for extracting features of rotating machinery impact faults, a computer device, and a machine-readable storage medium, so as to overcome the technical problem of incorrect impact fault diagnosis caused by inaccurate selection of demodulation frequency band when extracting impact fault features using envelope demodulation method in the prior art.
[0006] To achieve the above objectives, the first aspect of this application provides a method for extracting features of rotating machinery impact faults, comprising: Acquire the impact response vibration signal caused by defects in rotating machinery; Calculate the envelope kurtosis of each sub-band signal, wherein the sub-band signal is a sub-band signal at different levels decomposed from the impact response vibration signal; The non-cyclic stationary frequency bands are marked stepwise according to the envelope spectrum kurtosis. Calculate the squared envelope autocorrelation kurtosis of the unlabeled subband signals and determine the maximum squared envelope autocorrelation kurtosis; The frequency band of the sub-band signal corresponding to the maximum squared envelope autocorrelation kurtosis is determined as the optimal demodulation frequency band; Envelope demodulation is performed on the impact response vibration signal within the optimal demodulation frequency band to obtain the impact fault characteristics of the rotating machinery.
[0007] In this embodiment of the application, calculating the envelope kurtosis of each sub-band signal includes: The impact response vibration signal is decomposed into sub-frequency band signals at different levels by wavelet transform, wherein the level division is performed using a binary tree structure. Perform Hibernate transform on each sub-band signal to obtain the envelope of each sub-band signal; Perform Fourier transform on each envelope to obtain the envelope spectrum of each sub-band signal. Calculate the kurtosis of each envelope spectrum, and obtain the envelope spectrum kurtosis of each sub-band signal one-to-one.
[0008] In this embodiment, the envelope kurtosis of the sub-band signal is calculated using the following formula: ; in, Indicates the first The first level The envelope kurtosis of the sub-band signal; Indicates the first The first level The first sub-band signal The envelope spectrum amplitude at each frequency point; Indicates the first The first level The average value of the envelope spectrum amplitude of all frequency points in the sub-band signal; Indicates the first The first level The total number of frequency points in each sub-band signal.
[0009] In this embodiment of the application, the non-cyclic stationary frequency bands are labeled stepwise according to the envelope spectrum kurtosis, including: Based on preset labeling rules, non-cyclic stationary frequency bands are labeled step by step according to the envelope spectrum kurtosis. The preset marking rules include at least the following: The number of sub-band signals retained in the hierarchy is greater than or equal to the number of sub-bands. The number of sub-band signals retained in the hierarchy, the first The depth of the level in the binary tree structure is greater than the first level. The depth of the hierarchy in the binary tree structure, for any level, the retained sub-band signal is the unlabeled sub-band signal remaining in the level after the non-cyclic stationary frequency bands of the level are labeled.
[0010] In this embodiment of the application, based on a preset marking rule, the non-cyclic stationary frequency band is marked stepwise according to the envelope spectrum kurtosis, including: Sort the envelope spectrum kurtosis in each level from largest to smallest; The frequency bands of all sub-band signals in the first and second levels from top to bottom are marked as non-cyclic stationary frequency bands; In the third level from top to bottom, the frequency bands of sub-band signals with envelope kurtosis in the last M positions are marked as non-cyclic stationary frequency bands, where the value of M is within a first preset range and M is a positive integer. In the fourth level from top to bottom, the frequency bands of sub-band signals with envelope kurtosis in the last M positions are marked as non-cyclic stationary frequency bands; In the fifth level from top to bottom, the frequency band of the sub-band signal with the envelope spectrum kurtosis in the last T position is marked as a non-cyclic stationary frequency band, where the value of T is within a second preset range and T is a positive integer.
[0011] In this embodiment of the application, the calculation of the squared envelope autocorrelation kurtosis of the unlabeled sub-band signal includes: Calculate the autocorrelation function value of the squared envelope of the unlabeled subband signal; A subsequence is selected from the sequence of autocorrelation function values of the sub-band signals, and the target sub-band signal is composed of the frequency point signals corresponding to the subsequences. Calculate the squared envelope autocorrelation kurtosis of the target subband signal.
[0012] In this embodiment of the application, a sub-sequence is selected from the sequence of autocorrelation function values of the sub-band signals, and the target sub-band signal is composed of the frequency point signals corresponding to the sub-sequences, including: Select the last N / 2 autocorrelation function values from the sequence of autocorrelation function values of the unlabeled sub-band signals, where N represents the total number of frequency points of the unlabeled sub-band signals; In the sequence of the last N / 2 selected autocorrelation function values, a subsequence is selected again, and the target sub-band signal is composed of the frequency signals corresponding to each frequency point of the subsequence.
[0013] In this embodiment, the squared envelope autocorrelation kurtosis of the target sub-band signal is calculated using the following formula: ; in, Indicates the first The first level The squared envelope autocorrelation kurtosis of the target sub-band signal; Indicates the first The first level The autocorrelation function value of the square envelope of the target sub-band signal; Indicates the first The first level The average value of the squared envelope autocorrelation function of all frequency points in the target sub-band signal; Indicates the frequency point number in the target sub-band signal; This indicates the total number of frequency points in the target sub-band signal.
[0014] In this embodiment of the application, the impact response vibration signal within the optimal demodulation frequency band is envelope demodulated to obtain the impact fault characteristics of the rotating machinery, including: Calculate the envelope spectrum of the impact response vibration signal within the optimal demodulation frequency band; The fault characteristic frequencies and / or amplitudes are identified from the calculated envelope spectrum, and the identified fault characteristic frequencies and / or amplitudes are used as the impact fault characteristics of rotating machinery.
[0015] A second aspect of this application provides a rotating machinery impact fault feature extraction device, comprising: The acquisition module is used to acquire the impact response vibration signal caused by defects in rotating machinery; The envelope spectrum kurtosis calculation module is used to calculate the envelope spectrum kurtosis of each sub-band signal, wherein the sub-band signal is a sub-band signal at different levels decomposed from the impact response vibration signal; The marking module is used to mark the non-cyclic stationary frequency bands stepwise according to the kurtosis of the envelope spectrum; The squared envelope autocorrelation kurtosis calculation module is used to calculate the squared envelope autocorrelation kurtosis of unlabeled sub-band signals and determine the maximum squared envelope autocorrelation kurtosis. The optimal demodulation frequency band determination module is used to determine the frequency band of the sub-band signal corresponding to the maximum squared envelope autocorrelation kurtosis as the optimal demodulation frequency band; The impact fault feature extraction module is used to perform envelope demodulation on the impact response vibration signal within the optimal demodulation frequency band to obtain the impact fault features of the rotating machinery.
[0016] In this embodiment of the application, calculating the envelope kurtosis of each sub-band signal includes: The impact response vibration signal is decomposed into sub-frequency band signals at different levels by wavelet transform, wherein the level division is performed using a binary tree structure. Perform Hibernate transform on each sub-band signal to obtain the envelope of each sub-band signal; Perform Fourier transform on each envelope to obtain the envelope spectrum of each sub-band signal. Calculate the kurtosis of each envelope spectrum, and obtain the envelope spectrum kurtosis of each sub-band signal one-to-one.
[0017] In this embodiment, the envelope kurtosis of the sub-band signal is calculated using the following formula: ; in, Indicates the first The first level The envelope kurtosis of the sub-band signal; Indicates the first The first level The first sub-band signal The envelope spectrum amplitude at each frequency point; Indicates the first The first level The average value of the envelope spectrum amplitude of all frequency points in the sub-band signal; Indicates the first The first level The total number of frequency points in each sub-band signal.
[0018] In this embodiment of the application, the non-cyclic stationary frequency bands are labeled stepwise according to the envelope spectrum kurtosis, including: Based on preset labeling rules, non-cyclic stationary frequency bands are labeled step by step according to the envelope spectrum kurtosis. The preset marking rules include at least the following: The number of sub-band signals retained in the hierarchy is greater than or equal to the number of sub-bands. The number of sub-band signals retained in the hierarchy, the first The depth of the level in the binary tree structure is greater than the first level. The depth of the hierarchy in the binary tree structure, for any level, the retained sub-band signal is the unlabeled sub-band signal remaining in the level after the non-cyclic stationary frequency bands of the level are labeled.
[0019] In this embodiment of the application, based on a preset marking rule, the non-cyclic stationary frequency band is marked stepwise according to the envelope spectrum kurtosis, including: Sort the envelope spectrum kurtosis in each level from largest to smallest; The frequency bands of all sub-band signals in the first and second levels from top to bottom are marked as non-cyclic stationary frequency bands; In the third level from top to bottom, the frequency bands of sub-band signals with envelope kurtosis in the last M positions are marked as non-cyclic stationary frequency bands, where the value of M is within a first preset range and M is a positive integer. In the fourth level from top to bottom, the frequency bands of sub-band signals with envelope kurtosis in the last M positions are marked as non-cyclic stationary frequency bands; In the fifth level from top to bottom, the frequency band of the sub-band signal with the envelope spectrum kurtosis in the last T position is marked as a non-cyclic stationary frequency band, where the value of T is within a second preset range and T is a positive integer.
[0020] In this embodiment of the application, the calculation of the squared envelope autocorrelation kurtosis of the unlabeled sub-band signal includes: Calculate the autocorrelation function value of the squared envelope of the unlabeled subband signal; A subsequence is selected from the sequence of autocorrelation function values of the sub-band signals, and the target sub-band signal is composed of the frequency point signals corresponding to the subsequences. Calculate the squared envelope autocorrelation kurtosis of the target subband signal.
[0021] In this embodiment of the application, a sub-sequence is selected from the sequence of autocorrelation function values of the sub-band signals, and the target sub-band signal is composed of the frequency point signals corresponding to the sub-sequences, including: Select the last N / 2 autocorrelation function values from the sequence of autocorrelation function values of the unlabeled sub-band signals, where N represents the total number of frequency points of the unlabeled sub-band signals; In the sequence of the last N / 2 selected autocorrelation function values, a subsequence is selected again, and the target sub-band signal is composed of the frequency signals corresponding to each frequency point of the subsequence.
[0022] In this embodiment, the squared envelope autocorrelation kurtosis of the target sub-band signal is calculated using the following formula: ; in, Indicates the first The first level The squared envelope autocorrelation kurtosis of the target sub-band signal; Indicates the first The first level The autocorrelation function value of the square envelope of the target sub-band signal; Indicates the first The first level The average value of the squared envelope autocorrelation function of all frequency points in the target sub-band signal; Indicates the frequency point number in the target sub-band signal; This indicates the total number of frequency points in the target sub-band signal.
[0023] In this embodiment of the application, the impact response vibration signal within the optimal demodulation frequency band is envelope demodulated to obtain the impact fault characteristics of the rotating machinery, including: Calculate the envelope spectrum of the impact response vibration signal within the optimal demodulation frequency band; The fault characteristic frequencies and / or amplitudes are identified from the calculated envelope spectrum, and the identified fault characteristic frequencies and / or amplitudes are used as the impact fault characteristics of rotating machinery.
[0024] A third aspect of this application provides a computer device, comprising: The memory is configured to store instructions; and The processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the rotating machinery impact fault feature extraction method according to the first aspect of this application.
[0025] A third aspect of this application provides a machine-readable storage medium storing instructions for causing a machine to perform the rotating machinery impact fault feature extraction method according to the first aspect of this application.
[0026] The above technical solution achieves accurate selection of the optimal demodulation frequency band through sequential evaluation, thereby improving the accuracy of rotating machinery impact fault feature extraction and impact fault diagnosis results. Specifically, the sequential evaluation refers to the sequential evaluation of each sub-frequency band in the impact response vibration signal using the envelope kurtosis index for evaluating cyclostationarity and the square envelope autocorrelation kurtosis index for evaluating impact.
[0027] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description
[0028] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings: Figure 1 A flowchart illustrating a method for extracting features of rotating machinery impact faults according to an embodiment of this application is shown schematically. Figure 2 The time-domain plot of the vibration signal to be analyzed is shown schematically. Figure 3 The spectrum of the vibration signal to be analyzed is schematically shown; Figure 4 This schematically illustrates the vibration signal to be analyzed. ESK Calculation results; Figure 5The optimal fault frequency band selection result of the vibration signal to be analyzed is schematically shown; Figure 6 The time-domain waveform of the selected optimal fault band filter signal is schematically shown; Figure 7 The envelope spectrum of the selected optimal fault band filtered signal is schematically shown; Figure 8 The diagram illustrates the optimal fault frequency band selection result of the vibration signal to be analyzed obtained using the Autogram method; Figure 9 The time-domain waveform of the optimal fault band filter signal selected by the Autogram method is illustrated schematically. Figure 10 The envelope spectrum of the optimal fault band filtered signal selected by the Autogram method is illustrated schematically. Figure 11 This schematic diagram illustrates the composition of a rotating machinery impact fault feature extraction device according to an embodiment of this application; Figure 12 A schematic block diagram of a computer device according to an embodiment of this application is shown. Detailed Implementation
[0029] The specific embodiments of this application will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the embodiments of this application.
[0030] If the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.
[0031] For the extraction of impact fault features using envelope demodulation, the accuracy of the demodulation frequency band is a key factor determining the accuracy of the impact fault diagnosis results. Accordingly, this application provides a method for extracting impact fault features in rotating machinery, which extracts impact fault features in the following manner: Acquire the impact response vibration signal caused by defects in rotating machinery; Calculate the envelope kurtosis of each sub-band signal, where the sub-band signals are sub-band signals at different levels decomposed from the impact response vibration signal; Non-cyclic stationary frequency bands are labeled step-by-step based on envelope kurtosis. Calculate the squared envelope autocorrelation kurtosis of the unlabeled subband signals and determine the maximum squared envelope autocorrelation kurtosis; The frequency band of the sub-band signal corresponding to the maximum squared envelope autocorrelation kurtosis is determined as the optimal demodulation frequency band; Envelope demodulation is performed on the impact response vibration signal within the optimal demodulation frequency band to obtain the impact fault characteristics of rotating machinery.
[0032] As is known, in rotating machinery, the impact response vibration signal caused by local defects such as rolling bearings has two main characteristics: impact and cyclic stationarity. This application, as described in the above embodiments, utilizes impact and cyclic stationarity to sequentially evaluate sub-bands. First, starting from the sparsity of the envelope spectrum of the cyclic stationary signal, the kurtosis of the envelope spectrum of the sub-band signal is used to evaluate cyclic stationarity, marking non-cyclic stationary frequency bands, screening out cyclic stationary sub-bands, and eliminating sub-bands that may interfere with subsequent impact evaluation, such as random pulse interference bands and electromagnetic noise interference bands. Then, the squared envelope autocorrelation kurtosis is used to evaluate the impact of the unmarked sub-band signal, and finally, the frequency band with the strongest impact is identified from the unmarked sub-band signal as the optimal fault frequency band, thus determining the optimal demodulation frequency band. By comprehensively and accurately evaluating impact and cyclic stationarity, the demodulation frequency band containing the richest fault information is selected, effectively improving the ability of the envelope demodulation method to resist strong non-Gaussian noise interference under no-priority conditions, and achieving accurate extraction of fault impact characteristics and accurate diagnosis of impact faults.
[0033] In a comparative embodiment, the Autogram method is used to select the optimal demodulation frequency band. The Autogram method primarily evaluates the cyclostationarity of each sub-band by using the time-domain envelope kurtosis of the impact response vibration signal as a cyclostationarity index, thereby finding the optimal fault frequency band. It is evident that the evaluation index of the envelope demodulation method based on the Autogram method is singular, easily leading to the selection of a frequency band subject to strong non-Gaussian noise interference as the optimal demodulation frequency band, thus causing fault diagnosis errors. Through the above comparison, it is clear that the embodiments of this application achieve accurate selection of the optimal demodulation frequency band, thereby achieving accurate diagnosis of impact faults.
[0034] Figure 1 A schematic flowchart of a method for extracting features of rotating machinery impact faults according to an embodiment of this application is shown. Figure 1As shown, the rotating machinery impact fault feature extraction method provided in this application includes steps 102 to 114. It is understood that the above-described rotating machinery impact fault feature extraction method may include all steps 102 to 114, or only some of them. As an optional embodiment of this application, a rotating machinery impact fault feature extraction method may include only steps 102 to 112.
[0035] Step 102: Obtain the impact response vibration signal caused by the defect in the rotating machinery. It is known that the impact response vibration signal is a periodic impact signal.
[0036] Step 104: Decompose the impact response vibration signal to obtain sub-band signals at different levels.
[0037] Specifically, in this application, the impact response vibration signal is decomposed into sub-frequency band signals at different levels by wavelet transform, wherein the hierarchical division is performed using a binary tree structure, that is, the binary tree is used as the basic structure.
[0038] In a preferred embodiment of this application, a binary tree is used as the basic structure, and the maximum overlap discrete wavelet transform is employed to decompose the impact response vibration signal.
[0039] Step 106: Calculate the envelope kurtosis of each sub-band signal.
[0040] As an optional embodiment of this application, the envelope kurtosis of each sub-band signal can be calculated in the following manner: Perform Hibernate transform on each sub-band signal to obtain the envelope of each sub-band signal; Perform Fourier transform on each envelope to obtain the envelope spectrum of each sub-band signal. Calculate the kurtosis of each envelope spectrum, and obtain the envelope spectrum kurtosis of each sub-band signal one-to-one.
[0041] For example, in a specific example: First, the envelope signal is obtained by performing a Hibernate transform on the sub-band signal using the following formula: (Formula 1); In Formula 1, Indicates the first The first level The envelope signal obtained after the Hibernate transform of the sub-band signal; Indicates the first The first level Individual frequency band signals; This represents the Hippel transform.
[0042] Then, the envelope signal is Fourier transformed using the following formula to obtain the envelope spectrum of the corresponding sub-band signal: (Formula 2); In Formula 2, Indicates the first The first level The envelope spectrum is obtained by performing a Fourier transform on the envelope signal corresponding to each sub-band signal. This represents the Fourier transform.
[0043] Finally, the envelope spectrum kurtosis is calculated using the following formula: (Formula 3); In Formula 3, Indicates the first The first level The envelope kurtosis of the sub-band signal; Indicates the first The first level The first sub-band signal The envelope spectrum amplitude at each frequency point; Indicates the first The first level The average value of the envelope spectrum amplitude of all frequency points in the sub-band signal; Indicates the first The first level The total number of frequency points in each sub-band signal.
[0044] Step 108: Mark the non-cyclic stationary frequency bands step by step according to the envelope spectrum kurtosis.
[0045] As a preferred embodiment of this application, non-cyclic stationary frequency bands are labeled progressively according to the envelope spectrum kurtosis based on a preset labeling rule. The preset labeling rule includes at least the following: The number of sub-band signals retained in the hierarchy is greater than or equal to the number of sub-bands. The number of sub-band signals retained in the hierarchy, the first The depth of the level in the binary tree structure is greater than the first level. The depth of the hierarchy in the binary tree structure, for any level, the retained sub-band signal is the unlabeled sub-band signal remaining in the level after the non-cyclic stationary frequency bands of the level are labeled.
[0046] The aforementioned preset marking rules are designed based on the cyclostationarity characteristics of sub-band signals at various levels. Because the envelope spectrum of cyclostationary signals is sparse, the number of cyclostationary frequency bands selected gradually increases as the level deepens. Based on this principle, cyclostationary frequency bands are retained level by level by marking non-cyclostationary frequency bands.
[0047] As an example, based on preset labeling rules, non-cyclic stationary frequency bands are labeled step by step according to the kurtosis of the envelope spectrum, specifically including the following steps: Sort the envelope spectrum kurtosis in each level from largest to smallest; The frequency bands of all sub-band signals in the first and second levels from top to bottom are marked as non-cyclic stationary frequency bands; In the third level from top to bottom, the frequency bands of sub-band signals with envelope kurtosis in the last M positions are marked as non-cyclic stationary frequency bands, where the value of M is within a first preset range and M is a positive integer. In the fourth level from top to bottom, the frequency bands of sub-band signals with envelope kurtosis in the last M positions are marked as non-cyclic stationary frequency bands; In the fifth level from top to bottom, the frequency band of the sub-band signal with the envelope spectrum kurtosis in the last T position is marked as a non-cyclic stationary frequency band, where the value of T is within a second preset range and T is a positive integer.
[0048] It should be understood that the first and second preset ranges need to be determined in conjunction with the specific application scenario. This application does not limit the specific values of the first and second preset ranges. Similarly, the values of M and T also need to be determined in conjunction with the specific application scenario.
[0049] For example, in a specific application: the value of M is... The value of T is , Indicates the level number, Indicates the first The total number of sub-bands at each level.
[0050] Step 110: Calculate the squared envelope autocorrelation kurtosis of the unlabeled sub-band signals and determine the maximum squared envelope autocorrelation kurtosis.
[0051] Specifically, in this application, the squared envelope autocorrelation kurtosis of the unlabeled sub-band signal is calculated in the following manner: Calculate the autocorrelation function value of the squared envelope of the unlabeled subband signal; Select a subsequence from the sequence of autocorrelation function values of the sub-band signals, and form the target sub-band signal from the signals of each frequency point corresponding to the subsequence; Calculate the squared envelope autocorrelation kurtosis of the target subband signal.
[0052] As an example, a subsequence is selected from the sequence of autocorrelation function values of the sub-band signals, and the target sub-band signal is composed of the frequency point signals corresponding to the subsequence, including: Select the last N / 2 autocorrelation function values from the sequence of autocorrelation function values of the unlabeled sub-band signals, where N represents the total number of frequency points of the unlabeled sub-band signals; In the sequence of the last N / 2 selected autocorrelation function values, a subsequence is selected again, and the target sub-band signal is composed of the frequency signals corresponding to each frequency point of the subsequence.
[0053] For example, in the above embodiment, the signals removed from the sequence of the last N / 2 autocorrelation function values are multiple frequency signals that oscillate, specifically: the frequency signal corresponding to zero time delay (maximum autocorrelation function value) and the frequency signal within a third preset range with zero time delay as the lower boundary.
[0054] As an example, the squared envelope autocorrelation kurtosis of the target sub-band signal is calculated using the following formula: (Formula 4); In Formula 4, Indicates the first The first level The squared envelope autocorrelation kurtosis of the target sub-band signal; Indicates the first The first level The autocorrelation function value of the square envelope of the target sub-band signal; Indicates the first The first level The average value of the squared envelope autocorrelation function of all frequency points in the target sub-band signal; Indicates the frequency point number in the target sub-band signal; This indicates the total number of frequency points in the target sub-band signal.
[0055] Step 112: Determine the frequency band of the sub-band signal corresponding to the maximum squared envelope autocorrelation kurtosis as the optimal demodulation frequency band.
[0056] Step 114: Envelope demodulation is performed on the impact response vibration signal within the optimal demodulation frequency band to obtain the impact fault characteristics of the rotating machinery.
[0057] As is known, the characteristics of impact failures in rotating machinery typically include the fault frequency and the amplitude of the fault impact signal obtained through envelope demodulation.
[0058] As an optional embodiment of this application, envelope demodulation is performed on the impact response vibration signal within the optimal demodulation frequency band to obtain the impact fault characteristics of the rotating machinery, specifically including the following steps: Calculate the envelope spectrum of the impact response vibration signal within the optimal demodulation frequency band; The fault characteristic frequencies and amplitudes are identified from the calculated envelope spectrum, and the identified fault characteristic frequencies and amplitudes are used as the impact fault characteristics of rotating machinery.
[0059] It should be understood that the above optional embodiments are not the only limitation on the envelope demodulation method of impact response vibration signal. It is known that after the optimal demodulation frequency band is known, impact fault features can be extracted based on other envelope demodulation methods. At the same time, the impact fault features are not limited to the fault feature frequency and amplitude, but need to be determined according to the local defective parts of the specific rotating machinery.
[0060] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.
[0061] In a specific application example, the above embodiment is applied to the vibration signal diagnosis of a rolling bearing in rotating machinery. This vibration signal consists of three parts: periodic impacts, noise, and random impacts. The periodic impact signal has a resonant frequency of 3200Hz, a damping ratio of 0.05, an impact repetition frequency of 100Hz, and a rotational frequency of 10Hz. Adding noise of appropriate energy reduces the signal-to-noise ratio of the periodic impact signal to -11dB. The random impact signal has a resonant frequency of 1900Hz, a damping ratio of 0.04, and 15 impacts. The vibration signal length is 60000, and the sampling rate is 12000Hz. Simulation of this vibration signal yields the following time-domain waveform: Figure 2 As shown in the figure, a large amount of noise and large-amplitude random impulses can be observed, but they cannot be removed to observe periodic impulse information. The spectrum of the simulated signal is as follows. Figure 3 As shown, the background noise is obvious, and resonance peaks corresponding to random and periodic impacts can be seen.
[0062] Combination Figures 2 to 10 As shown, in this application example, the process of extracting fault features from the above vibration signal specifically includes steps A1 to A8.
[0063] Step A1: Using a binary tree as the basic structure, the original vibration signal is decomposed into sub-band signals at different levels using the maximum overlap discrete wavelet transform. ,in, l For the number of levels, ; k This refers to the order of sub-band signals within the hierarchy. .
[0064] Step A2, use formula to pair all sub-band signals Perform a Hilbert transform to obtain the envelope signal. .
[0065] Step A3, apply Formula 2 to the envelope signal Perform a Fourier transform to obtain the envelope spectrum. .
[0066] Step A4: Calculate the kurtosis of the envelope spectrum of all sub-bands using Formula 3, where the kurtosis is... l Level 1 k The envelope kurtosis of the sub-band signal is denoted as... The result is as follows Figure 4 As shown, the periodic impact of the cyclic stationary component corresponds to the sub-bands within the frequency band range. The impact is significant, while random impacts correspond to sub-bands within the same frequency band. It is very small.
[0067] Step A5, according to Values are used to mark non-cyclic stationary frequency bands at each level. To preserve cyclic stationary frequency bands and filter out non-cyclic stationary frequency bands, the frequency bands at each level are... Sort from largest to smallest, then: in level 0, The frequency band with a value one position lower than the previous value in this level is marked as a non-cyclic stationary frequency band; in level 1, The frequency band with values in the last two positions of this level is marked as a non-cyclic stationary frequency band; in level 2, The frequency band with values in the last 3 digits of this level is marked as a non-cyclic stationary frequency band; in level 3, The frequency band with values in the last 7 bits of this level is marked as a non-cyclic stationary frequency band; in level 4, The frequency band with values in the last 14 bits of this level is marked as a non-cyclic stationary frequency band.
[0068] Step A6: Calculate the autocorrelation function of the squared envelope of all unlabeled sub-band signals. Select the latter half (i.e., half the autocorrelation function length) of the sequence formed by the autocorrelation function values of the sub-band signals. Then, select a starting point with an index of 0.1 from the selected sequence. N The cutoff point number is 0.45. N The subsequence is used as the target sub-band signal, and the corresponding sub-band signal is taken as the target sub-band signal. The squared envelope autocorrelation kurtosis is calculated using Formula 4. To evaluate the impulsive characteristics of each target sub-band signal. Furthermore, the sub-band signals marked in step A5... Set all values to 0.
[0069] The results are as follows Figure 5 As shown, only the sub-bands within the frequency range corresponding to the periodic impact were preserved, and the corresponding... Both are very large, while the sub-bands within the corresponding frequency range of random impacts are completely removed (corresponding to...). (0).
[0070] Step A7: Select the frequency band with the strongest impact as the optimal fault frequency band (optimal demodulation frequency band), that is: The largest frequency band is used as the optimal demodulation frequency band. In this application, it is the 3000~3750Hz frequency band, and the time-domain waveform of the filtered signal corresponding to this frequency band is as follows. Figure 6 As shown, its time-domain waveform no longer contains large-amplitude random impacts, and the signal-to-noise ratio is significantly higher than that of the original vibration signal, allowing for a clearer observation of the periodic impact characteristics.
[0071] Step A8: Calculate the envelope spectrum of the optimal fault frequency band signal to extract fault characteristic frequencies.
[0072] The results are as follows Figure 7 As shown, the fault characteristic frequency and its corresponding harmonics can be clearly observed in the envelope spectrum of the filtered signal. Furthermore, frequency conversion band characteristics related to frequency conversion modulation can also be observed, indicating that the above embodiments of this application achieve the extraction of fault characteristic frequencies. In addition, comparative analysis illustrates the beneficial effects of this application compared to the prior art, such as... Figure 8 The results shown are from the analysis using the Autogram method (Moshrefzadeh A and Fasana A. The Autogram: An effective approach for selecting the optimal demodulation band in rollingelement bearings diagnosis[J]. Mechanical Systems and Signal Processing, 2018, 105: 294-318). It can be observed that the Autogram method selects the 1500~2250Hz frequency band as the optimal fault frequency band. This frequency band falls within the random impulse frequency band, and its time-domain waveform is as follows... Figure 9 As shown, the time-domain waveform contains significant large-amplitude random impulses but lacks periodic impulse characteristics. Furthermore, its envelope spectrum is calculated, and the results are as follows... Figure 10 As shown, its envelope spectrum is dominated by noise, and fault characteristic frequency information cannot be observed. Clearly, the Autogram method is affected by random impact interference, incorrectly selecting the random impact frequency band as the optimal fault frequency band, thus failing to extract fault characteristic frequency information.
[0073] in, Figure 2 This is the time-domain plot of the vibration signal to be analyzed. Figure 3This is the spectrum of the vibration signal to be analyzed. Figure 4 The ESK calculation results are shown for the vibration signal to be analyzed. Figure 5 The optimal fault frequency band selection result is shown for the vibration signal to be analyzed. Figure 6 The time-domain waveform of the selected optimal fault band filter signal. Figure 7 This is the envelope spectrum of the selected optimal fault band filtered signal. Figure 8 The optimal fault frequency band selection result for the vibration signal to be analyzed is obtained using the Autogram method. Figure 9 The time-domain waveform of the optimal fault band filter signal selected for the Autogram method. Figure 10 The envelope spectrum of the optimal fault band filter signal selected by the Autogram method.
[0074] Corresponding to the rotating machinery impact fault feature extraction method in the above embodiments, Figure 11 The diagram illustrates the composition of a rotating machinery impact fault feature extraction device according to an embodiment of this application. For ease of explanation, only the parts related to the embodiments of this application are shown.
[0075] like Figure 11 As shown, the rotating machinery impact fault feature extraction device 400 includes: Acquisition module 410 is used to acquire the impact response vibration signal caused by defects in rotating machinery; Envelope spectrum kurtosis calculation module 420 is used to calculate the envelope spectrum kurtosis of each sub-band signal, wherein the sub-band signal is a sub-band signal at different levels decomposed from the impact response vibration signal; The labeling module 430 is used to label the non-cyclic stationary frequency band stepwise according to the envelope spectrum kurtosis; The squared envelope autocorrelation kurtosis calculation module 440 is used to calculate the squared envelope autocorrelation kurtosis of the unlabeled sub-band signal and determine the maximum squared envelope autocorrelation kurtosis. The optimal demodulation frequency band determination module 450 is used to determine the frequency band of the sub-frequency band signal corresponding to the maximum squared envelope autocorrelation kurtosis as the optimal demodulation frequency band; The impact fault feature extraction module 460 is used to perform envelope demodulation on the impact response vibration signal within the optimal demodulation frequency band to obtain the impact fault features of the rotating machinery.
[0076] As an embodiment of this application, the rotating machinery impact fault feature extraction device 400 can achieve the following: Figure 1 The embodiments shown and other related method embodiments.
[0077] The process by which each module of the rotating machinery impact fault feature extraction device 400 provided in this application implements its respective function can be found in the foregoing. Figure 1The descriptions of the embodiments shown and other related method embodiments are not repeated here.
[0078] It should be noted that the information interaction and execution process between the above modules are based on the same concept as the method embodiments of this application. Their specific functions and technical effects can be found in the method embodiments section, and will not be repeated here.
[0079] Figure 12 A schematic block diagram of a computer device according to an embodiment of the present application is shown. In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown below. Figure 12 As shown in the figure, the computer device includes a processor A01, a network interface A02, a display screen A04, an input device A05, and a memory (not shown) connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A06. The non-volatile storage medium A06 stores an operating system B01 and a computer program B02. The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 stored in the non-volatile storage medium A06. The network interface A02 is used for communication with external terminals via a network connection. When the computer program is executed by the processor A01, it implements a method for extracting features of rotating machinery impact faults. The display screen A04 can be a liquid crystal display (LCD) or an e-ink display. The input device A05 can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.
[0080] Those skilled in the art will understand that Figure 12 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0081] In one embodiment, the rotating machinery impact fault feature extraction device 400 provided in this application can be implemented as a computer program, and the computer program can be implemented in, for example... Figure 12 The computer device shown operates on the computer. The computer device's memory can store various program modules that constitute the rotating machinery impact fault feature extraction device 400. The computer program, composed of these program modules, causes the processor to execute the steps in the rotating machinery impact fault feature extraction methods of the various embodiments of this application described in this specification.
[0082] In one embodiment, this application also provides a machine-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the rotating machinery impact fault feature extraction method in the above embodiments.
[0083] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0084] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0085] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.
[0086] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.
Claims
1. A method for extracting features of rotating machinery impact faults, characterized in that, include: Acquire the impact response vibration signal caused by defects in rotating machinery; Calculate the envelope kurtosis of each sub-band signal, wherein the sub-band signal is a sub-band signal at different levels decomposed from the impact response vibration signal; The non-cyclic stationary frequency bands are marked stepwise according to the envelope spectrum kurtosis. Calculate the squared envelope autocorrelation kurtosis of the unlabeled subband signals and determine the maximum squared envelope autocorrelation kurtosis; The frequency band of the sub-band signal corresponding to the maximum squared envelope autocorrelation kurtosis is determined as the optimal demodulation frequency band; Envelope demodulation is performed on the impact response vibration signal within the optimal demodulation frequency band to obtain the impact fault characteristics of the rotating machinery.
2. The method for extracting features of rotating machinery impact faults according to claim 1, characterized in that, The calculation of the envelope kurtosis of each sub-band signal includes: The impact response vibration signal is decomposed into sub-frequency band signals at different levels by wavelet transform, wherein the level division is performed using a binary tree structure. Perform Hibernate transform on each sub-band signal to obtain the envelope of each sub-band signal; Perform Fourier transform on each envelope to obtain the envelope spectrum of each sub-band signal. Calculate the kurtosis of each envelope spectrum, and obtain the envelope spectrum kurtosis of each sub-band signal one-to-one.
3. The method for extracting features of rotating machinery impact faults according to claim 1 or 2, characterized in that, The envelope kurtosis of the sub-band signal is calculated using the following formula: ； in, Indicates the first The first level The envelope kurtosis of the sub-band signal; Indicates the first The first level The first sub-band signal The envelope spectrum amplitude at each frequency point; Indicates the first The first level The average value of the envelope spectrum amplitude of all frequency points in the sub-band signal; Indicates the first The first level The total number of frequency points in each sub-band signal.
4. The method for extracting features of rotating machinery impact faults according to claim 1, characterized in that, The non-cyclic stationary frequency bands are labeled stepwise according to the envelope spectrum kurtosis, including: Based on preset labeling rules, non-cyclic stationary frequency bands are labeled step by step according to the envelope spectrum kurtosis. The preset marking rules include: the first The number of sub-band signals retained in the hierarchy is greater than or equal to the number of sub-bands. The number of sub-band signals retained in the hierarchy, the first The depth of the level in the binary tree structure is greater than the first level. The depth of the hierarchy in the binary tree structure, for any level, the retained sub-band signal is the unlabeled sub-band signal remaining in the level after the non-cyclic stationary frequency bands of the level are labeled.
5. The method for extracting features of rotating machinery impact faults according to claim 4, characterized in that, Based on preset labeling rules, non-cyclic stationary frequency bands are labeled step-by-step according to the envelope spectrum kurtosis, including: Sort the envelope spectrum kurtosis in each level from largest to smallest; The frequency bands of all sub-band signals in the first and second levels from top to bottom are marked as non-cyclic stationary frequency bands; In the third level from top to bottom, the frequency bands of sub-band signals with envelope kurtosis in the last M positions are marked as non-cyclic stationary frequency bands, where M takes the value within a first preset range and M is a positive integer; In the fourth level from top to bottom, the frequency bands of sub-band signals with envelope kurtosis in the last M positions are marked as non-cyclic stationary frequency bands; In the fifth level from top to bottom, the frequency band of the sub-band signal with the envelope spectrum kurtosis in the last T position is marked as a non-cyclic stationary frequency band, where the value of T is within a second preset range and T is a positive integer.
6. The method for extracting features of rotating machinery impact faults according to claim 1, characterized in that, The calculation of the squared envelope autocorrelation kurtosis of the unlabeled sub-band signal includes: Calculate the autocorrelation function value of the squared envelope of the unlabeled subband signal; A subsequence is selected from the sequence of autocorrelation function values of the sub-band signals, and the target sub-band signal is composed of the frequency point signals corresponding to the subsequences. Calculate the squared envelope autocorrelation kurtosis of the target subband signal.
7. The method for extracting features of rotating machinery impact faults according to claim 6, characterized in that, A subsequence is selected from the sequence of autocorrelation function values of the sub-band signals, and the target sub-band signal is composed of the frequency point signals corresponding to the subsequences, including: Select the last N / 2 autocorrelation function values from the sequence of autocorrelation function values of the unlabeled sub-band signals, where N represents the total number of frequency points of the unlabeled sub-band signals; In the sequence of the last N / 2 selected autocorrelation function values, a subsequence is selected again, and the target sub-band signal is composed of the frequency signals corresponding to each frequency point of the subsequence.
8. The method for extracting features of rotating machinery impact faults according to claim 6, characterized in that, The squared envelope autocorrelation kurtosis of the target sub-band signal is calculated using the following formula: ； in, Indicates the first The first level The squared envelope autocorrelation kurtosis of the target sub-band signal; Indicates the first The first level The autocorrelation function value of the square envelope of the target sub-band signal; Indicates the first The first level The average value of the squared envelope autocorrelation function of all frequency points in the target sub-band signal; Indicates the frequency point number in the target sub-band signal; This indicates the total number of frequency points in the target sub-band signal.
9. The method for extracting features of rotating machinery impact faults according to claim 1, characterized in that, Envelope demodulation is performed on the impact response vibration signal within the optimal demodulation frequency band to obtain the impact fault characteristics of the rotating machinery, including: Calculate the envelope spectrum of the impact response vibration signal within the optimal demodulation frequency band; The fault characteristic frequencies and / or amplitudes are identified from the calculated envelope spectrum, and the identified fault characteristic frequencies and / or amplitudes are used as the impact fault characteristics of rotating machinery.
10. A device for extracting features of rotating machinery impact faults, characterized in that, include: The acquisition module is used to acquire the impact response vibration signal caused by defects in rotating machinery; The envelope spectrum kurtosis calculation module is used to calculate the envelope spectrum kurtosis of each sub-band signal, wherein the sub-band signal is a sub-band signal at different levels decomposed from the impact response vibration signal; The marking module is used to mark the non-cyclic stationary frequency bands stepwise according to the kurtosis of the envelope spectrum; The squared envelope autocorrelation kurtosis calculation module is used to calculate the squared envelope autocorrelation kurtosis of unlabeled sub-band signals and determine the maximum squared envelope autocorrelation kurtosis. The optimal demodulation frequency band determination module is used to determine the frequency band of the sub-band signal corresponding to the maximum squared envelope autocorrelation kurtosis as the optimal demodulation frequency band; The impact fault feature extraction module is used to perform envelope demodulation on the impact response vibration signal within the optimal demodulation frequency band to obtain the impact fault features of the rotating machinery.
11. The rotating machinery impact fault feature extraction device according to claim 10, characterized in that, The calculation of the envelope kurtosis of each sub-band signal includes: The impact response vibration signal is decomposed into sub-frequency band signals at different levels by wavelet transform, wherein the level division is performed using a binary tree structure. Perform Hibernate transform on each sub-band signal to obtain the envelope of each sub-band signal; Perform Fourier transform on each envelope to obtain the envelope spectrum of each sub-band signal. Calculate the kurtosis of each envelope spectrum, and obtain the envelope spectrum kurtosis of each sub-band signal one-to-one.
12. The rotating machinery impact fault feature extraction device according to claim 10 or 11, characterized in that, The envelope kurtosis of the sub-band signal is calculated using the following formula: ； in, Indicates the first The first level The envelope kurtosis of the sub-band signal; Indicates the first The first level The first sub-band signal The envelope spectrum amplitude at each frequency point; Indicates the first The first level The average value of the envelope spectrum amplitude of all frequency points in the sub-band signal; Indicates the first The first level The total number of frequency points in each sub-band signal.
13. The rotating machinery impact fault feature extraction device according to claim 10, characterized in that, The non-cyclic stationary frequency bands are labeled stepwise according to the envelope spectrum kurtosis, including: Based on preset labeling rules, non-cyclic stationary frequency bands are labeled step by step according to the envelope spectrum kurtosis. The preset marking rules include at least the following: The number of sub-band signals retained in the hierarchy is greater than or equal to the number of sub-bands. The number of sub-band signals retained in the hierarchy, the first The depth of the level in the binary tree structure is greater than the first level. The depth of the hierarchy in the binary tree structure, for any level, the retained sub-band signal is the unlabeled sub-band signal remaining in the level after the non-cyclic stationary frequency bands of the level are labeled.
14. The rotating machinery impact fault feature extraction device according to claim 13, characterized in that, Based on preset labeling rules, non-cyclic stationary frequency bands are labeled step-by-step according to the envelope spectrum kurtosis, including: Sort the envelope spectrum kurtosis in each level from largest to smallest; The frequency bands of all sub-band signals in the first and second levels from top to bottom are marked as non-cyclic stationary frequency bands; In the third level from top to bottom, the frequency bands of sub-band signals with envelope kurtosis in the last M positions are marked as non-cyclic stationary frequency bands, where M takes the value within a first preset range and M is a positive integer; In the fourth level from top to bottom, the frequency bands of sub-band signals with envelope kurtosis in the last M positions are marked as non-cyclic stationary frequency bands; In the fifth level from top to bottom, the frequency band of the sub-band signal with the envelope spectrum kurtosis in the last T position is marked as a non-cyclic stationary frequency band, where the value of T is within a second preset range and T is a positive integer.
15. The rotating machinery impact fault feature extraction device according to claim 10, characterized in that, The calculation of the squared envelope autocorrelation kurtosis of the unlabeled sub-band signal includes: Calculate the autocorrelation function value of the squared envelope of the unlabeled subband signal; A subsequence is selected from the sequence of autocorrelation function values of the sub-band signals, and the target sub-band signal is composed of the frequency point signals corresponding to the subsequences. Calculate the squared envelope autocorrelation kurtosis of the target subband signal.
16. The rotating machinery impact fault feature extraction device according to claim 15, characterized in that, A subsequence is selected from the sequence of autocorrelation function values of the sub-band signals, and the target sub-band signal is composed of the frequency point signals corresponding to the subsequences, including: Select the last N / 2 autocorrelation function values from the sequence of autocorrelation function values of the unlabeled sub-band signals, where N represents the total number of frequency points of the unlabeled sub-band signals; In the sequence of the last N / 2 selected autocorrelation function values, a subsequence is selected again, and the target sub-band signal is composed of the frequency signals corresponding to each frequency point of the subsequence.
17. The rotating machinery impact fault feature extraction device according to claim 15, characterized in that, The squared envelope autocorrelation kurtosis of the target sub-band signal is calculated using the following formula: ； in, Indicates the first The first level The squared envelope autocorrelation kurtosis of the target sub-band signal; Indicates the first The first level The autocorrelation function value of the square envelope of the target sub-band signal; Indicates the first The first level The average value of the squared envelope autocorrelation function of all frequency points in the target sub-band signal; Indicates the frequency point number in the target sub-band signal; This indicates the total number of frequency points in the target sub-band signal.
18. The rotating machinery impact fault feature extraction device according to claim 10, characterized in that, Envelope demodulation is performed on the impact response vibration signal within the optimal demodulation frequency band to obtain the impact fault characteristics of the rotating machinery, including: Calculate the envelope spectrum of the impact response vibration signal within the optimal demodulation frequency band; The fault characteristic frequencies and amplitudes are identified from the calculated envelope spectrum, and the identified fault characteristic frequencies and amplitudes are used as the impact fault characteristics of rotating machinery.
19. A computer device, characterized in that, include: The memory is configured to store instructions; as well as The processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the rotating machinery impact fault feature extraction method according to any one of claims 1 to 9.
20. A machine-readable storage medium, characterized in that, The machine-readable storage medium stores instructions for causing the machine to perform the rotating machinery impact fault feature extraction method according to any one of claims 1 to 9.