Gear weak fault extraction method and device, electronic equipment and storage medium

Abstract

The invention provides a gear weak fault extraction method and device, electronic equipment and a storage medium, and belongs to the field of fault detection, and the method comprises the steps: carrying out the denoising of a target gear signal based on an improved wavelet function, and obtaining a first gear signal; decomposing the first gear signal to obtain a plurality of gear intrinsic mode functions corresponding to the first gear signal; and determining a fault type of the gear based on the plurality of gear eigenmode functions corresponding to the first gear signal. According to the gear weak fault extraction method and device, the electronic equipment and the storage medium provided by the invention, the accuracy and reliability of gear weak fault extraction can be improved.

Description

Technical Field

[0001] The present disclosure belongs to the field of fault detection, and more specifically, relates to a gear weak fault extraction method and device, electronic equipment, and storage medium. Background Art

[0002] In many industrial equipment, such as machine tools, automobile transmissions, wind turbines, etc., gears are key components in mechanical equipment, and their operating status directly affects the overall performance and reliability of the equipment. Through gear fault detection, potential problems such as early wear and fatigue cracks can be discovered in time. For wear problems, timely detection can take measures when the tooth surface wear is light. If the gear fault is not discovered in time, it may gradually develop into serious damage, such as tooth root fracture, which will not only cause the gear itself to be scrapped, but also may cause damage to other connected components, greatly shortening the service life of the entire equipment. In some large equipment or high-load equipment, gear failure may cause serious safety accidents.

[0003] The current technology has the problem that the characteristics of weak fault vibration signals of gears are difficult to extract under strong noise background.

[0004] It can be seen that an accurate and reliable gear weak fault extraction method is urgently needed. Summary of the invention

[0005] The purpose of the present disclosure is to provide a gear weak fault extraction method and device, electronic equipment, and storage medium to improve the accuracy and reliability of gear weak fault extraction.

[0006] A first aspect of an embodiment of the present disclosure provides a method for extracting weak faults of gears, comprising: De-noising the target gear signal based on the improved wavelet function to obtain a first gear signal; Decomposing the first gear signal to obtain a plurality of gear eigenmode functions corresponding to the first gear signal; The fault type of the gear is determined based on the plurality of gear eigenmode functions corresponding to the first gear signal.

[0007] A second aspect of the embodiments of the present disclosure provides a gear weak fault extraction device, comprising: A denoising module, used for denoising the target gear signal based on an improved wavelet function to obtain a first gear signal; a decomposition module, configured to decompose the first gear signal to obtain a plurality of gear eigenmode functions corresponding to the first gear signal; A fault diagnosis module is used to determine a gear fault type based on a plurality of gear eigenmode functions corresponding to the first gear signal.

[0008] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the above-mentioned method for extracting weak faults of gears when executing the computer program.

[0009] In a fourth aspect of an embodiment of the present disclosure, a computer storage medium is provided, wherein the computer storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned gear weak fault extraction method are implemented.

[0010] The beneficial effects of a gear weak fault extraction method and device, electronic device, and storage medium provided by the embodiments of the present disclosure are: The present disclosure performs denoising processing on the gear vibration signal through an improved wavelet function, which can effectively remove the noise components in the signal, improve the signal-to-noise ratio of the signal, and provide a higher quality data basis for subsequent signal analysis. The present disclosure decomposes the first gear signal into multiple gear intrinsic mode functions through collective empirical mode decomposition. Each gear intrinsic mode function component represents different frequency and time characteristics, which can reflect the local characteristics of the signal more finely and help to discover weak fault signals hidden in complex signals. The present disclosure calculates the correlation coefficient between the gear intrinsic mode function components and the original signal, which can eliminate those false components with weak correlation with the original signal, retain the effective gear intrinsic mode function components, reduce the interference of subsequent analysis, and improve the accuracy and reliability of the extraction of gear weak fault features. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A schematic flow chart of a gear weak fault extraction method provided by an embodiment of the present disclosure; Figure 2 A structural block diagram of a gear weak fault extraction device provided by an embodiment of the present disclosure; Figure 3 A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0013] In the following description, specific details such as specific system structures and technologies are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present disclosure. However, it should be clear to those skilled in the art that the present disclosure may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obstructing the description of the present disclosure with unnecessary details.

[0014] In order to make the purpose, technical solutions and advantages of the present disclosure more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0015] Please refer to Figure 1 , Figure 1 A schematic flow chart of a method for extracting a weak gear fault according to an embodiment of the present disclosure is provided, and the method comprises: S101: De-noising the target gear signal based on the improved wavelet function to obtain a first gear signal.

[0016] In this embodiment, the improved wavelet function may be:

[0017] in, is the processed wavelet coefficient, representing the correlation between the wavelet and the gear vibration signal; is the wavelet coefficient before processing; is the regulating factor, ,Pick =0.618, is the first threshold, is the adjustment coefficient, is an integer.

[0018] The target gear signal is a plurality of initial wavelet coefficients obtained after the gear vibration signal is subjected to wavelet transform. The wavelet transform may be one of Haar wavelet, Daubechies wavelet and Symlets wavelet. The gear vibration signal may be obtained based on a vibration sensor.

[0019] The vibration signal of the first gear is the wavelet coefficient after denoising. There are multiple wavelet coefficients, which reflect the characteristics at different scales. The global and local characteristics of the vibration signal can be understood by analyzing the coefficients at different scales.

[0020] S102: Decompose the first gear signal to obtain a plurality of gear eigenmode functions corresponding to the first gear signal.

[0021] In this embodiment, the first gear signal can be decomposed based on ensemble empirical mode decomposition (EEMD) to obtain multiple gear intrinsic mode functions corresponding to the first gear signal. The gear intrinsic mode function is essentially an intrinsic mode function (IMF).

[0022] EEMD is a signal processing method that can decompose complex signals into multiple IMF components, each of which represents different frequency and time characteristics. The above IMF components can reflect the local characteristics of the signal more finely and help to extract the characteristics of weak fault signals.

[0023] S103: Determine the fault type of the gear based on a plurality of gear eigenmode functions corresponding to the first gear signal.

[0024] In this embodiment, correlation coefficients between multiple gear eigenmode functions and original signals may be calculated, and false gear eigenfunction components may be eliminated based on the correlation coefficients to obtain valid gear eigenmode functions.

[0025] By calculating the Teager energy operator of the above-mentioned effective gear eigenmode function and reconstructing it into a Teager energy spectrum, time-frequency analysis is performed on the above-mentioned reconstructed Teager energy spectrum to obtain the weak fault characteristics of the gear, and the weak fault characteristics of the gear are matched with the stored gear weak fault characteristics and their corresponding fault types to obtain the gear fault type.

[0026] From the above, it can be concluded that the present disclosure performs denoising processing on the gear vibration signal through the improved wavelet function, which can effectively remove the noise components in the signal, improve the signal-to-noise ratio of the signal, and provide a higher quality data basis for subsequent signal analysis. The present disclosure decomposes the first gear signal into multiple gear intrinsic mode functions through collective empirical mode decomposition. Each gear intrinsic mode function component represents different frequency and time characteristics, which can reflect the local characteristics of the signal more finely and help to discover weak fault signals hidden in complex signals. By calculating the correlation coefficient between the gear intrinsic mode function components and the original signal, the present disclosure can eliminate those false components with weak correlation with the original signal, retain the effective gear intrinsic mode function components, reduce the interference of subsequent analysis, and improve the accuracy and reliability of the extraction of gear weak fault features.

[0027] In one embodiment of the present disclosure, before denoising the target gear signal based on the improved wavelet function, the method further includes: Collecting raw data based on the first device, where the raw data is a gear vibration signal; The original data is processed based on wavelet transform to obtain a target gear signal, which is a plurality of wavelet coefficient data.

[0028] In this embodiment, the first device may be a vibration sensor, the original data is a vibration signal of the gear, and the wavelet transform may be one of Haar wavelet, Daubechies wavelet and Symlets wavelet. For example, a vibration sensor is used to detect the gear vibration signal, and the gear vibration signal is subjected to wavelet transform using Haar wavelet to obtain a target gear signal, which is a plurality of wavelet coefficient data.

[0029] From the above, it can be concluded that the present invention processes the original vibration signal through wavelet transform, and can decompose the signal into multiple wavelet coefficient data, namely the target gear signal, and performs multi-scale analysis on the signal, and extracts the characteristics of the signal at different frequencies and time scales, thereby providing reliable data support for subsequent denoising and fault analysis; the present invention decomposes the signal into different frequency levels, which helps the subsequent denoising step to more accurately identify and remove noise components, thereby improving the accuracy and reliability of gear weak fault detection.

[0030] In one embodiment of the present disclosure, the improved wavelet function is:

[0031] in, is the processed wavelet coefficient, representing the correlation between the wavelet and the gear vibration signal; is the wavelet coefficient before processing; is the regulating factor, ,Pick =0.618, is the first threshold, is the adjustment coefficient, is an integer.

[0032] In this embodiment, the wavelet function is improved, not only the classification of the threshold in the wavelet function is refined, but also the calculation formula is improved. The wavelet function contains a hard threshold function and a soft threshold function.

[0033] Among them, in the hard threshold function, when the absolute value of the wavelet coefficient is greater than the given threshold, it remains unchanged; when the absolute value of the wavelet coefficient is less than or equal to the threshold, it is set to zero. The characteristics are simple calculation and the ability to better retain larger wavelet coefficients, but it is discontinuous at the threshold, which may cause pseudo-Gibbs phenomenon in the reconstructed signal.

[0034] In the soft threshold function, when the absolute value of the wavelet coefficient is greater than the given threshold, it shrinks toward zero; when the absolute value of the wavelet coefficient is less than or equal to the threshold, it is set to zero and continuous at the threshold. The reconstructed signal is relatively smooth, but it will cause a certain degree of shrinkage of the larger wavelet coefficients, and some useful signal components may be lost.

[0035] Both the hard threshold function and the soft threshold function have their advantages and disadvantages. Therefore, an improved wavelet function is proposed here. This improvement is based on the soft threshold function in the wavelet function. Although the traditional soft threshold function is continuous at the threshold and the reconstructed signal is relatively smooth, it will shrink the larger wavelet coefficients to a certain extent and may lose some useful signal components. This may cause the denoised signal to be distorted to a certain extent compared with the original signal.

[0036] The improved soft threshold function may adopt a more flexible contraction rule, dynamically adjusting the contraction degree according to the size and position of the wavelet coefficients. In this way, the impact on the useful signal components can be minimized while removing the noise. The improved soft threshold function can remove the noise more effectively, while better retaining the useful components of the signal, thereby improving the signal quality after denoising. The present disclosure can effectively avoid the attenuation of the wavelet with a larger amplitude in the soft threshold method, and eliminate the constant deviation of the soft threshold function as much as possible.

[0037] From the above, it can be concluded that the present disclosure introduces adjustment factors and adjustment coefficients, so that the improved wavelet function can flexibly adjust the denoising strategy according to different wavelet coefficient characteristics, making the denoising process more accurate. This can not only remove noise more effectively, but also retain more signal details. The present disclosure can perform different degrees of processing according to the size and position of different wavelet coefficients, thereby reducing the loss of useful signal components and improving signal fidelity. The improved wavelet function maintains the continuity of the soft threshold function at the threshold, further avoiding the pseudo-Gibbs phenomenon in the reconstructed signal, making the signal smoother. Improve the accuracy and reliability of gear weak fault extraction.

[0038] In one embodiment of the present disclosure, determining the fault type of the gear based on a plurality of gear eigenmode functions corresponding to the first gear signal includes: Based on the first formula group and the plurality of gear eigenmode functions, correlation coefficients between the plurality of gear eigenmode functions and the original data are obtained; Based on the correlation coefficient, the effective gear eigenmode function is obtained; The fault type of the gear is determined based on the effective gear eigenmode function.

[0039] In this embodiment, based on the first formula group and the multiple gear eigenmode functions, the correlation coefficients between the multiple gear eigenmode functions and the original data are obtained, including: Performing mean calculation on the original data based on the first sub-formula to obtain first sub-data; Calculate the mean of the plurality of third data based on the second sub-formula to obtain the second sub-data; Obtain third sub-data based on the third sub-formula from the first sub-data and the second sub-data; Calculate the variance of the first sub-data based on the fourth sub-formula to obtain the fourth sub-data; Based on the fifth sub-formula, the variance of the second sub-data is calculated to obtain the fifth sub-data; The third sub-data, the fourth sub-data and the fifth sub-data are calculated based on the sixth sub-formula to obtain a correlation coefficient.

[0040] This embodiment can determine whether the gear intrinsic mode function corresponding to the correlation coefficient is valid based on the correlation coefficient and the first threshold value, eliminate false gear intrinsic mode functions, and retain valid gear intrinsic mode functions.

[0041] This embodiment can calculate the Teager energy operator of the above-mentioned effective gear eigenmode function and reconstruct it into a Teager energy spectrum, perform time-frequency analysis on the above-mentioned reconstructed Teager energy spectrum to obtain the weak fault characteristics of the gear, and match the weak fault characteristics of the gear with the stored gear weak fault characteristics and their corresponding fault types to obtain the gear fault type.

[0042] The first subformula can be:

[0043] in, is the mean of the original signal, is the number of original signals, For the The value of the original signal, That is the first sub-data.

[0044] The second sub-formula can be:

[0045] in, is the mean value of the gear eigenmode function component, is the number of gear eigenmode function components, For the The value of the gear eigenmode function, That is the second sub-data.

[0046] The third sub-formula can be:

[0047] in, is the covariance between the mean of the original signal and the mean of the gear eigenmode function component, is the number of original signal and gear eigenmode function components, This is the third sub-data.

[0048] The fourth sub-formula can be:

[0049] in, is the variance of the original signal, is the number of original signals, For the The value of the original signal, is the mean of the original signal, This is the fourth sub-data.

[0050] The fifth sub-formula can be:

[0051] in, is the variance of the gear eigenmode function component, is the mean value of the gear eigenmode function component, is the number of gear eigenmode function components, For the The value of the gear eigenmode function, This is the fifth sub-data.

[0052] The sixth sub-formula can be:

[0053] in, is the correlation coefficient between multiple gear eigenmode functions and original data, is the covariance between the mean of the original signal and the mean of the gear eigenmode function component, is the variance of the original signal, is the variance of the gear eigenmode function component.

[0054] The value of the correlation coefficient is between -1 and 1. If the correlation coefficient is close to 1, it means that the IMF component is highly positively correlated with the original signal; if it is close to -1, it means that it is highly negatively correlated; if it is close to 0, it means that the correlation between the two is weak.

[0055] The absolute value of the calculated correlation coefficient can be compared with the first threshold. The absolute value greater than or equal to the first threshold can be identified as a valid gear intrinsic mode function component, and the absolute value less than the first threshold can be identified as a false or invalid gear intrinsic mode function. The first threshold can be set artificially based on experience.

[0056] In this embodiment, the Teager energy operator of the effective gear eigenmode function component can be calculated by the second formula, and the second formula is:

[0057] in, For the Teager energy operator for the parameters of the effective gear eigenmodes, Indicates The effective gear eigenmode parameter components, Indicates The effective gear eigenmode parameter components, Indicates The effective gear eigenmode parameter components.

[0058] The Teager energy operator of each effective gear eigenmode function is obtained and combined into a first energy spectrum, which is essentially a Teager energy spectrum.

[0059] Time-frequency analysis can be performed using short-time Fourier transform, wavelet transform, Hilbert-Huang transform, etc. to obtain time-frequency features. The above time-frequency features can be compared with the corresponding time-frequency features when various gear faults occur to obtain the gear fault type.

[0060] From the above, it can be concluded that the present disclosure can effectively eliminate false gear intrinsic mode function components with weak correlation with the original signal by calculating the correlation coefficients between multiple gear intrinsic mode functions and original data, and screen out effective gear intrinsic mode function components based on these correlation coefficients, thereby reducing the impact of noise and interference on fault identification and improving the accuracy of fault identification. From calculating the statistical characteristics (such as mean, variance, covariance) of the original signal and the gear intrinsic mode function components, to calculating the correlation coefficient, to screening effective IMF components, and finally through the time-frequency analysis of the Teager energy spectrum, the entire process includes multiple calculation steps, which enhances the robustness of fault identification and improves the accuracy and reliability of gear weak fault extraction.

[0061] In this embodiment, the fault type of the gear is determined based on the effective gear eigenmode function, including: Obtaining a first energy spectrum based on the effective gear eigenmode function; Performing time-frequency analysis on the first energy spectrum based on at least two time-frequency algorithms to obtain a weak fault feature of the gear; The gear fault type is determined based on the gear weak fault characteristics.

[0062] In this embodiment, the Teager energy operator of the effective gear eigenmode function component can be calculated by the second formula:

[0063] in, For the Teager energy operator for the parameters of the effective gear eigenmodes, Indicates The effective gear eigenmode parameter components, Indicates The effective gear eigenmode parameter components, Indicates The effective gear eigenmode parameter components.

[0064] The above calculations are performed on the effective gear eigenmode functions in sequence to obtain the Teager energy operator of the effective gear eigenmode function and form a first energy spectrum. The first energy spectrum may be presented in the form of a sequence.

[0065] For example, here is a sequence of several effective gear eigenmode function components:

[0066] Then the first energy spectrum can be calculated as follows: ; ; … .

[0067] in, For the Teager energy operator for the parameters of the effective gear eigenmodes, For the The effective gear eigenmode functions, and Special processing can be performed as needed, such as taking adjacent values ​​or pre-set specific values.

[0068] By calculating the Teager energy operator of each effective gear eigenmode function, the transient changes and nonlinear characteristics of the first gear signal can be highlighted, which is helpful to extract the characteristics of weak fault signals.

[0069] Time-frequency analysis can be performed using short-time Fourier transform, wavelet transform, and Hilbert-Huang transform to obtain time-frequency features.

[0070] For example, short-time Fourier transform is used to perform time-frequency calculation on the first energy spectrum to obtain the time-frequency characteristics of the first energy spectrum. The time-frequency characteristics can be presented in the form of an image (time-frequency diagram). The horizontal axis of the time-frequency diagram represents time, and the vertical axis represents frequency. The color or grayscale value (in the case of a grayscale diagram) of each pixel point (or each small area) in the time-frequency diagram represents the energy intensity of the signal at that time and frequency.

[0071] In the time-frequency diagram obtained by short-time Fourier transform, the vibration signal energy of a normal gear is usually concentrated in a specific frequency region and is relatively stable in time.

[0072] From the overall morphology, when the gear is worn, the energy distribution in the time-frequency diagram becomes relatively dispersed. Wear causes the contact state of the tooth surface to change. In addition to the energy at the meshing frequency and its harmonics, some new, relatively weak high-frequency components will be generated. These high-frequency components appear in the time-frequency diagram as some additional, irregular energy spots or blurred areas above the original stripes, and these high-frequency energy areas will gradually expand as the degree of wear increases.

[0073] Tooth root cracks can cause periodic changes in gear stiffness, which are manifested in the time-frequency diagram as sidebands around the meshing frequency and its harmonics. The intervals of these sidebands are usually related to the rotation frequency of the gear, and appear as periodic small stripes on both sides of the original horizontal stripes. The appearance of sidebands is due to the modulation of the vibration characteristics of the gear during rotation caused by the cracks, and the energy size and interval changes can reflect the severity of the cracks.

[0074] A broken tooth is a serious fault that will show very obvious characteristics on the time-frequency graph. A broken tooth will produce a strong impact at the moment of breaking, causing a high-energy peak or a high-energy area to appear on the time-frequency graph. The frequency range of this area is relatively wide, including components from low frequency to high frequency. Afterwards, due to the loss of balance and normal meshing state of the gear, the energy distribution in the time-frequency graph will become extremely chaotic, with a large number of irregular high-energy areas and frequency components.

[0075] From the perspective of changing rules, the frequency components of normal gears are mainly the meshing frequency and its harmonics, and these frequencies remain basically stable during the entire operation process.

[0076] Wear causes changes in the energy of the meshing frequency and its harmonics. Generally speaking, as wear progresses, the energy at the meshing frequency will gradually decrease, while the harmonic components will become more complex. At the same time, the frequency and energy of the newly emerged high-frequency components will change with the degree of wear, for example, the frequency of the high-frequency components will gradually increase, and the energy will also gradually increase.

[0077] In addition to the meshing frequency and its harmonics, the focus is on the frequency of the sidebands. The frequency interval of the sidebands is related to the rotation speed of the gear. By analyzing the frequency interval of the sidebands, the rotation frequency of the gear can be determined, and then the fault characteristics related to rotation can be inferred. The energy change of the sidebands is also critical. As the crack expands, the energy of the sidebands usually increases gradually.

[0078] The frequency components of the gear vibration signal after tooth breakage become chaotic. In addition to the wide-band high-energy region generated by the initial impact, many frequency components related to non-meshing frequencies will appear during subsequent operation. The energy of these new frequency components will fluctuate at a high level without obvious regularity, which is significantly different from the frequency components of normal gears.

[0079] Alternatively, for example, the Hilbert-Huang transform is used to perform time-frequency calculation on the first energy spectrum to obtain a marginal spectrum of the first energy spectrum.

[0080] When the gear is running normally, the marginal spectrum shows some relatively stable frequency components. These frequency components are mainly related to the meshing frequency of the gear and its low-order harmonics. For example, for a gear with a fixed number of teeth and speed, its meshing frequency is fixed, which is manifested as an obvious peak on the marginal spectrum. In addition, since the energy distribution of normal vibration is relatively concentrated, the energy of other frequency components in the marginal spectrum is relatively low.

[0081] When gears wear, the meshing frequency and the amplitude of its harmonics in the marginal spectrum will change. Wear will cause the tooth shape to change, which will change the vibration energy during the meshing process. Because the contact state of the tooth surface changes after wear, more high-frequency vibration components are generated. Wear will reduce the amplitude of the meshing frequency, and the harmonic components will increase or become more complex.

[0082] Cracks in the tooth root will cause the stiffness of the gear to change. In the marginal spectrum, in addition to the meshing frequency and its harmonics, there will be some low-frequency modulation components. These low-frequency components are due to the existence of cracks, which causes the gear to produce periodic stiffness changes during rotation, thereby modulating the meshing frequency. For example, there will be sidebands below the meshing frequency, and their interval frequency is related to the rotation frequency of the gear.

[0083] Tooth breakage is a relatively serious fault, and the marginal spectrum will change significantly. At this time, the amplitude of the meshing frequency and its harmonics will increase significantly, and new and chaotic high-frequency components will appear. This is because tooth breakage will cause strong impact vibration and generate a large amount of high-frequency energy. At the same time, since tooth breakage destroys the normal meshing relationship of the gears, the vibration signal becomes more complex and unstable.

[0084] The above-mentioned time-frequency diagram or marginal spectrum can be identified through manual observation and analysis or through a neural network model to obtain the corresponding gear fault. The above-mentioned neural network model has been trained based on a data set formed by a large number of gear fault types and their corresponding time-frequency diagrams and marginal spectra.

[0085] From the above, it can be concluded that the present disclosure can highlight the transient changes and nonlinear characteristics of the gear signal by calculating the Teager energy operator of the effective gear intrinsic mode function component. These features may be difficult to capture in traditional signal processing methods, but they are crucial for detecting weak fault signals. The Teager energy operator is very sensitive to the local energy changes of the signal, which helps to extract more refined fault features. The present disclosure uses at least two time-frequency algorithms (such as short-time Fourier transform, wavelet transform, Hilbert-Huang transform, etc.) to perform time-frequency analysis on the first energy spectrum, which can reveal the characteristics of the signal from multiple angles. Different algorithms have their own advantages and disadvantages, and they can complement each other when used in combination to improve the comprehensiveness and accuracy of fault identification. Even if a certain algorithm performs poorly in a specific situation, other algorithms can provide useful information, thereby enhancing the robustness of the entire disclosure and improving the accuracy and reliability of gear weak fault extraction.

[0086] In one embodiment of the present disclosure, a gear weak fault extraction method further includes: Based on the first neural network model, multiple gear eigenmode functions are processed to determine the fault type of the gear.

[0087] In one embodiment of the present disclosure, before processing a plurality of gear eigenmode functions based on the first neural network model, the method includes: Decompose the original data based on wavelet transform to obtain multiple wavelet coefficients; Adjusting the second number of wavelet coefficients based on the first number of first coefficients to obtain a first data set; Acquire the gear eigenmode function when the gear fails to obtain a second data set; The second neural network model is trained based on the first data set and the second data set to obtain the first neural network model.

[0088] In this embodiment, the first neural network model can be a convolutional neural network, a long short-term memory network or a deep belief network, etc., and a suitable neural network model can be selected according to actual needs and data characteristics. The first data set is the fault wavelet coefficient of the gear and the gear eigenmode function when the gear fails, and the fault wavelet coefficient and the corresponding gear eigenmode function of the gear under different fault states are collected.

[0089] The fault wavelet coefficients can be obtained by performing wavelet transform on the gear vibration signal, and the gear intrinsic mode function can be obtained by performing ensemble empirical mode decomposition on the denoised signal.

[0090] The first data set should contain various types of gear fault types and their corresponding wavelet coefficients; the second data set should contain various types of gear fault types and their corresponding gear eigenmode functions.

[0091] This embodiment takes into account that the data sets of wavelet coefficients corresponding to different types of gear faults are relatively scarce, so the first coefficient of the first quantity is used to adjust the second number of wavelet coefficients. The above wavelet coefficients can be the wavelet coefficients when the gear works normally. This embodiment can multiply or add it with the first coefficient to obtain an adjusted wavelet coefficient. The wavelet coefficient corresponds to the wavelet coefficient generated when the gear fails. The fault type corresponding to the wavelet coefficient can be determined based on experience and historical data. The first number and the second number can be set manually, but both are positive integers.

[0092] In this embodiment, the obtained adjusted wavelet coefficients and their corresponding fault types can be sorted to form a first data set, and the second neural network model can be trained according to the first data set to obtain a third neural network model.

[0093] The second data set is the gear eigenmode function and its corresponding fault type when the gear fails. The third neural network model is trained by the second data set to obtain the first neural network model.

[0094] The wavelet coefficients obtained after wavelet transformation or the gear eigenmode function obtained after set empirical mode decomposition are input into the first neural network model to obtain the fault type of the gear.

[0095] From the above, it can be concluded that the present disclosure takes into account that the wavelet coefficient data set when gear failure occurs is relatively scarce. This embodiment adopts a strategy of adjusting the second number of wavelet coefficients by the first number of first coefficients, which alleviates the problem of data scarcity to a certain extent, so that an effective model can be trained even with limited data, reduces dependence on data volume, and improves the accuracy and reliability of gear weak fault extraction.

[0096] Corresponding to a gear weak fault extraction method in the above embodiment, Figure 2 This is a structural block diagram of a gear weak fault extraction device provided by an embodiment of the present disclosure. For ease of description, only the parts related to the embodiment of the present disclosure are shown. Figure 2 The gear weak fault extraction device 20 includes: a denoising module 21, a decomposition module 22 and a fault diagnosis module 23.

[0097] The denoising module 21 is used to denoise the target gear signal based on the improved wavelet function to obtain a first gear signal.

[0098] A decomposition module 22, used for decomposing the first gear signal to obtain a plurality of gear eigenmode functions corresponding to the first gear signal; The fault diagnosis module 23 is used to determine the fault type of the gear based on a plurality of gear eigenmode functions corresponding to the first gear signal.

[0099] In one embodiment of the present disclosure, a gear weak fault extraction device 20 further includes: a signal acquisition module; A signal acquisition module, used for collecting raw data based on the first device, where the raw data is a gear vibration signal; The original data is processed based on wavelet transform to obtain a target gear signal, which is a plurality of wavelet coefficient data.

[0100] In one embodiment of the present disclosure, the denoising module 21 is specifically used to improve the wavelet function:

[0101] in, is the processed wavelet coefficient, representing the correlation between the wavelet and the gear vibration signal; is the wavelet coefficient before processing; is the regulating factor, ,Pick =0.618, is the first threshold, is the adjustment coefficient, is an integer.

[0102] In one embodiment of the present disclosure, the fault diagnosis module 23 is specifically used to obtain correlation coefficients between the plurality of gear intrinsic mode functions and the original data based on the first formula group and the plurality of gear intrinsic mode functions; Based on the correlation coefficient, the effective gear eigenmode function is obtained; The fault type of the gear is determined based on the effective gear eigenmode function.

[0103] In one embodiment of the present disclosure, the fault diagnosis module 23 is further configured to obtain a first energy spectrum based on an effective gear eigenmode function; Performing time-frequency analysis on the first energy spectrum based on at least two time-frequency algorithms to obtain a weak fault feature of the gear; The gear fault type is determined based on the gear weak fault characteristics.

[0104] In one embodiment of the present disclosure, a gear weak fault extraction device 20 further includes: a model diagnosis module; The model diagnosis module is used to process multiple gear eigenmode functions based on the first neural network model to determine the fault type of the gear.

[0105] In one embodiment of the present disclosure, a gear weak fault extraction device 20 further includes: a model training module; A model training module is used to decompose the original data based on wavelet transform to obtain multiple wavelet coefficients; Adjusting the second number of wavelet coefficients based on the first number of first coefficients to obtain a first data set; Acquire the gear eigenmode function when the gear fails to obtain a second data set; The second neural network model is trained based on the first data set and the second data set to obtain the first neural network model.

[0106] See also Figure 3 , Figure 3 A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. Figure 3 The electronic device 300 in the embodiment shown may include: one or more processors 301, one or more input devices 302, one or more output devices 303 and one or more memories 304. The processors 301, input devices 302, output devices 303 and memories 304 communicate with each other via a communication bus 305. The memory 304 is used to store computer programs, which include program instructions. The processor 301 is used to execute the program instructions stored in the memory 304. The processor 301 is configured to call the program instructions to execute the functions of each module / unit in the above-mentioned device embodiments, such as Figure 2 The functions of modules 21 to 23 are shown.

[0107] It should be understood that in the embodiment of the present disclosure, the processor 301 may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0108] The input device 302 may include a touch panel, a fingerprint collection sensor (for collecting the user's fingerprint information and fingerprint direction information), a microphone, etc., and the output device 303 may include a display (LCD, etc.), a speaker, etc.

[0109] The memory 304 may include a read-only memory and a random access memory, and provide instructions and data to the processor 301. A portion of the memory 304 may also include a non-volatile random access memory. For example, the memory 304 may also store information about the device type.

[0110] In a specific implementation, the processor 301, input device 302, and output device 303 described in the embodiments of the present disclosure can execute the implementation methods described in the first and second embodiments of a gear weak fault extraction method provided in the embodiments of the present disclosure, and can also execute the implementation methods of the electronic device described in the embodiments of the present disclosure, which will not be repeated here.

[0111] In another embodiment of the present disclosure, a computer storage medium is provided, the computer storage medium stores a computer program, the computer program includes program instructions, the program instructions are executed by the processor to implement all or part of the process in the above embodiment method, and can also be completed by instructing related hardware through the computer program. The computer program can be stored in a computer storage medium, and the computer program can implement the steps of each of the above method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. Computer readable media can include: any entity or device that can carry computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0112] The computer storage medium may be an internal storage unit of the electronic device of any of the aforementioned embodiments, such as a hard disk or memory of the electronic device. The computer storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Furthermore, the computer storage medium may also include both an internal storage unit of the electronic device and an external storage device. The computer storage medium is used to store computer programs and other programs and data required by the electronic device. The computer storage medium may also be used to temporarily store data that has been output or is to be output.

[0113] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this disclosure.

[0114] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the electronic devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0115] In the several embodiments provided in the present application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces or units, or it can be an electrical, mechanical or other form of connection.

[0116] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present disclosure.

[0117] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0118] The above are only specific embodiments of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present disclosure, and these modifications or replacements should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. A gear weak fault extraction method, characterized in that: include: De-noising the target gear signal based on the improved wavelet function to obtain a first gear signal; Decomposing the first gear signal to obtain a plurality of gear eigenmode functions corresponding to the first gear signal; The fault type of the gear is determined based on the plurality of gear eigenmode functions corresponding to the first gear signal.

2. A gear weak fault extraction method as claimed in claim 1, characterized in that: Before the target gear signal is denoised based on the improved wavelet function, the method further includes: Collecting raw data based on the first device, where the raw data is a gear vibration signal; The original data is processed based on wavelet transform to obtain the target gear signal, and the target gear signal is a plurality of wavelet coefficient data.

3. A gear weak fault extraction method as claimed in claim 1, characterized in that: The improved wavelet function is: in, is the processed wavelet coefficient, representing the correlation between the wavelet and the gear vibration signal; is the wavelet coefficient before processing; is the regulating factor, , is the first threshold, is the adjustment coefficient, is an integer.

4. A gear weak fault extraction method as claimed in claim 2, characterized in that: The determining the fault type of the gear based on a plurality of gear eigenmode functions corresponding to the first gear signal comprises: Based on the first formula group and the plurality of gear eigenmode functions, obtaining correlation coefficients between the plurality of gear eigenmode functions and the original data; Obtaining a valid gear eigenmode function based on the correlation coefficient; A fault type of the gear is determined based on the effective eigenmode function of the gear.

5. A gear weak fault extraction method as claimed in claim 4, characterized in that: The determining the fault type of the gear based on the effective gear eigenmode function comprises: Obtaining a first energy spectrum based on the effective gear eigenmode function; Performing time-frequency analysis on the first energy spectrum based on at least two time-frequency algorithms to obtain a weak fault feature of the gear; A fault type of the gear is determined based on the gear weak fault feature.

6. A gear weak fault extraction method as claimed in claim 2, characterized in that: Also includes: The plurality of gear eigenmode functions are processed based on the first neural network model to determine the fault type of the gear.

7. A gear weak fault extraction method as claimed in claim 6, characterized in that: Before processing the plurality of gear eigenmode functions based on the first neural network model, the method comprises: Decomposing the original data based on wavelet transform to obtain multiple wavelet coefficients; Adjusting the second number of wavelet coefficients based on the first number of first coefficients to obtain a first data set; Acquire the gear eigenmode function when the gear fails to obtain a second data set; A second neural network model is trained based on the first data set and the second data set to obtain the first neural network model.

8. A gear weak fault extraction device, characterized in that: include: A denoising module, used for denoising the target gear signal based on an improved wavelet function to obtain a first gear signal; a decomposition module, configured to decompose the first gear signal to obtain a plurality of gear eigenmode functions corresponding to the first gear signal; A fault diagnosis module is used to determine a gear fault type based on a plurality of gear eigenmode functions corresponding to the first gear signal.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images