A planetary gearbox fault detection method, device and storage medium

Abstract

The present application relates to a kind of planetary gearbox fault detection method, device and storage medium, the method comprises the following steps: obtaining the vibration signal of the planetary gearbox to be measured;The theoretical value of gear box fault characteristic frequency is calculated;Parameter combination optimization is carried out to the parameter combination of maximum correlation kurtosis deconvolution algorithm based on sparrow search algorithm, to obtain the best influence parameter combination;Based on the best influence parameter combination, parameter optimization is carried out to maximum correlation kurtosis deconvolution algorithm;The deconvolution signal of vibration signal is obtained using the maximum correlation kurtosis deconvolution algorithm after parameter optimization;Deconvolution signal is processed based on sparse coding shrinkage algorithm Denoising;Envelope demodulation operation is carried out to the noise reduction signal, and envelope spectrum is calculated;The theoretical value of gear fault characteristic frequency is compared with the spectral line in envelope spectrum, to distinguish the fault of gear box.Compared with prior art, the performance of maximum correlation kurtosis deconvolution algorithm is improved, with the advantage of high fault diagnosis precision.

Description

Technical Field

[0001] The present invention relates to the field of gearbox fault diagnosis methods, and in particular to a planetary gearbox fault detection method, device and storage medium. Background Art

[0002] A planetary gearbox typically consists of key components such as the sun gear, planet gears, ring gear, and planet carrier. The planet gears mesh with both the sun gear and the ring gear. If the ring gear is fixed and the sun gear rotates, the planet gears simultaneously rotate and orbit around the sun gear with the planet carrier. The vibration signal of a planetary gearbox consists of characteristic frequency components of each component, coupled vibration signals from the power input and load equipment, and background noise interference. During signal testing, the sensor is often mounted on the housing connected to the ring gear. In this case, the relative position between the meshing points of the sun gear-planet gear and planet gear-ring gear meshing pairs and the sensor changes with the rotation of the planet carrier, causing the vibration transmission path from the meshing point of the sun gear and planet gear teeth to the sensor to change. This time-varying vibration path can cause additional amplitude modulation in the shock sequence caused by localized damage to the sun gear and planet gears. Due to the complex motion and dynamic characteristics of a planetary gearbox, as well as the variable environmental excitation, the vibration test signal of the planetary gearbox is very complex, time-varying, and has significant modulation characteristics.

[0003] Considering the influence of strong noise and complex transmission paths, deconvolution methods are considered effective tools for extracting periodic fault pulses from planetary gearbox fault signatures embedded in background noise. The ability to adaptively design filters is crucial for accurately and completely extracting fault signature information. Two of the most commonly used deconvolution methods are minimum entropy deconvolution and maximum correlation kurtosis deconvolution. Minimum entropy deconvolution is susceptible to interference from random impulses in the signal and cannot distinguish between random and periodic impulses. Maximum correlation kurtosis deconvolution is sensitive to periodic fault impulses in the signal and can effectively highlight periodic impulses that are often buried in noise.

[0004] However, the two parameters, signal period T and deconvolution shift number M, have a crucial impact on the deconvolution performance. It is particularly important to select the optimal parameter combination [T, M] to improve the performance of the algorithm. At the same time, although the signal-to-noise ratio of the signal after maximum correlation kurtosis deconvolution is significantly improved, there is still a lot of noise interference, which makes the spectrum and envelope spectrum of the deconvolution signal blurred and difficult to identify. How to choose a noise reduction algorithm to further process the deconvolution signal to improve the accuracy of fault diagnosis has also become a problem that needs to be considered.

[0005] In summary, how to select the optimal parameter combination [T, M] to improve the performance of the maximum correlation kurtosis deconvolution algorithm and at the same time reduce the noise of the output deconvolution signal to obtain a reliable signal to identify the fault of the planetary gearbox is an urgent problem that needs to be solved. Summary of the Invention

[0006] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a planetary gearbox fault detection method, device and storage medium. The method optimizes the maximum correlation kurtosis deconvolution algorithm based on the sparrow search algorithm, and combines sparse coding contraction at the same time, which can effectively extract the planetary gearbox fault feature information hidden in the background noise, thereby realizing planetary gearbox fault detection.

[0007] The purpose of the present invention can be achieved by the following technical solutions:

[0008] A planetary gearbox fault detection method comprises the following steps:

[0009] Determine the planetary gearbox to be tested and obtain the gear structure parameters therein;

[0010] Obtain the vibration signal of the planetary gearbox to be tested;

[0011] Calculating the theoretical value of the gearbox fault characteristic frequency, the fault cycle, and the fault pulse interval based on the gear structural parameters;

[0012] Set the parameters of the sparrow search algorithm, and optimize the parameter combination [T, M] of the maximum correlation kurtosis deconvolution algorithm based on the sparrow search algorithm to obtain the best influencing parameter combination [T0, M0];

[0013] Based on the optimal influencing parameter combination, the maximum correlation kurtosis deconvolution algorithm is optimized, and the vibration signal is preprocessed using the maximum correlation kurtosis deconvolution algorithm with optimized parameters to obtain a deconvolution signal of the vibration signal;

[0014] Performing noise reduction processing on the deconvolution signal based on a sparse coding shrinkage algorithm to obtain a noise-reduced signal;

[0015] Performing envelope demodulation on the noise reduction signal and calculating an envelope spectrum;

[0016] The theoretical value of the gear fault characteristic frequency is compared with the spectrum line in the envelope spectrum to identify the gearbox fault.

[0017] Furthermore, the specific calculation formula for calculating the theoretical value of the gearbox fault characteristic frequency, the fault cycle and the fault pulse interval is:

[0018]

[0019]

[0020]

[0021]

[0022]

[0023] Among them, f s represents the characteristic frequency of sun gear fault, f p Indicates the characteristic frequency of planetary gear fault, f r Indicates the characteristic frequency of the inner ring gear fault, f m Indicates the meshing frequency, f c represents the sampling frequency, f represents the fault characteristic frequency, T1 represents the theoretical fault cycle, ΔT represents the fault pulse interval, Z s Indicates the number of sun gear teeth, Z p Indicates the number of planetary gear teeth, Z r Number of teeth of tabular internal gear ring, N P Indicates the number of planetary gears.

[0024] Furthermore, the parameters of the sparrow search algorithm are set as follows:

[0025] Performing filtering processing on the vibration signal using a maximum correlation kurtosis deconvolution algorithm to obtain an initial deconvolution signal;

[0026] The envelope spectrum entropy of the initial deconvolution signal is used as the fitness function of the sparrow search algorithm, and the minimum value of the envelope spectrum entropy is used as the basis for judging the optimization result of the sparrow search algorithm. The envelope spectrum entropy of the initial deconvolution signal is expressed as follows:

[0027]

[0028] Among them, p k Represents the probability distribution density of the envelope spectrum signal;

[0029] Furthermore, the optimization of the parameter combination of the maximum correlation kurtosis deconvolution algorithm based on the sparrow search algorithm to obtain the best influencing parameter combination includes the following steps:

[0030] The range of parameters T and M in the maximum correlation kurtosis deconvolution algorithm is defined as the population search space of the sparrow search algorithm, and X is defined as i =[T i ,M i ] is the position of sparrow i, that is, the i-th parameter combination to be optimized by the sparrow search algorithm; at this time X i The food concentration expression at is F=E(X i ), that is, the combination to be optimized [Ti ,M i ]The value of E corresponding to .

[0031] Based on the theoretical fault cycle T1, the optimization intervals of the sparrow search algorithm parameters and the maximum correlation kurtosis deconvolution algorithm parameters are defined, and the number of sparrows n, the maximum number of iterations, the proportion of discoverers, the proportion of scouts, and the safety threshold are set.

[0032] Define the maximum correlation kurtosis deconvolution filter with a length of L, a displacement of M, and a period of T∈[f c / f-10,f c / f+10], where f c is the sampling frequency, f is the fault characteristic frequency;

[0033] Initialize the position of the sparrows, randomly generate multiple parameter combinations [T, M] as the initial position of the sparrow search, and calculate the fitness value of each sparrow's position to obtain the optimal individual i in the initial population best and the position X of the optimal individual in the spatial domain best ;

[0034] Update the positions of discoverers, joiners, and scouts in the population, and compare E(X i ) size, update the optimal individual i in the population best and the position X of the optimal individual in the spatial domain best ;

[0035] If the maximum number of iterations is met, stop the iteration and output the best influencing parameter combination [T0, M0], otherwise continue to update and iterate.

[0036] Furthermore, based on the optimal influencing parameter combination [T0, M0], the maximum correlation kurtosis deconvolution algorithm is optimized, and the period parameter of the maximum correlation kurtosis deconvolution algorithm is set to T0, and the deconvolution shift parameter is set to M0;

[0037] The vibration signal is preprocessed using a maximum correlation kurtosis deconvolution algorithm with optimized parameters to obtain a deconvolution signal of the vibration signal, which is expressed as follows:

[0038]

[0039] y represents the deconvolution signal, x represents the vibration signal, F represents the maximum correlation kurtosis deconvolution filter coefficient with a length of L, and F=[f1 f2 … f L ] T , n represents the index value of the vibration signal sequence.

[0040] Furthermore, the calculation formula of the filter coefficient is:

[0041]

[0042] in,

[0043]

[0044]

[0045]

[0046] β=[y1y 1-T y2y 2-T … y N y N-T ] T

[0047] Wherein, N represents the length of the vibration signal sequence, r=M*T, M=M0, and T=T0.

[0048] Furthermore, the deconvolution signal is subjected to noise reduction processing based on the sparse coding shrinkage algorithm to obtain a noise reduction signal, which is expressed as:

[0049]

[0050] in, represents the denoised signal; y represents the deconvolution signal; α is the sparsity parameter that controls the probability density function, σ is the standard deviation of the white noise signal, d is the standard deviation of the deconvolution signal, and σ y is the standard deviation of the noise-reduced signal.

[0051] Furthermore, performing envelope demodulation operation on the noise reduction signal and calculating the envelope spectrum includes the following steps:

[0052] Perform Hilbert transform on the denoised signal s(t) to construct its analytical signal:

[0053] z(t)=s(t)+jy(t)

[0054] Where y(t) is the Hilbert transform of s(t), and its expression is:

[0055]

[0056] Where p is the Cauchy principal value;

[0057] The envelope signal is obtained by taking the modulus of the analytical signal:

[0058]

[0059] Performing Fourier transform on the envelope signal to obtain an envelope spectrum.

[0060] A planetary gearbox fault detection device comprises a memory and a processor, wherein the memory stores a computer program, and the processor calls the program instructions to execute the planetary gearbox fault detection method according to any one of claims 1 to 8.

[0061] A computer-readable storage medium comprises a computer program, wherein the computer program can be executed by a processor to implement the planetary gearbox fault detection method according to any one of claims 1 to 8.

[0062] Compared with the prior art, the present invention has the following beneficial effects:

[0063] 1. The present invention optimizes the maximum correlation kurtosis deconvolution algorithm by utilizing the sparrow search algorithm, and can adaptively search for the optimal parameter combination in the maximum correlation kurtosis deconvolution, thereby avoiding the influence of subjective parameter selection on the maximum correlation kurtosis deconvolution performance, effectively highlighting the periodic impact component in the vibration signal, and improving the performance of the maximum correlation kurtosis deconvolution algorithm;

[0064] 2. This invention utilizes a sparse coding contraction algorithm to process deconvoluted signals, effectively separating non-Gaussian fault impulse signals from Gaussian noise interference signals. Case analysis and comparative results demonstrate that this invention effectively addresses issues such as noise interference and severe transmission path attenuation in planetary gearbox fault signatures. It accurately extracts planetary gearbox fault signatures, enabling planetary gearbox fault detection and improving fault diagnosis accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 Flow chart of the implementation of the present invention;

[0066] Figure 2 It is the time domain waveform of the original vibration signal;

[0067] Figure 3 is the envelope spectrum of the original vibration signal;

[0068] Figure 4 It is a curve diagram showing the relationship between the envelope spectrum entropy and the number of iterations;

[0069] Figure 5 It is the time domain waveform of the deconvolution signal;

[0070] Figure 6 It is the time domain waveform of the noise reduction signal obtained after sparse coding shrinkage;

[0071] Figure 7 is the envelope spectrum of the denoised signal. DETAILED DESCRIPTION

[0072] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0073] like Figure 1 FIG. 1 is a flow chart of an implementation of the present invention, a planetary gearbox fault detection method, comprising the following steps:

[0074] S1. Determine the planetary gearbox to be tested and obtain the gear structure parameters therein;

[0075] S2. Obtaining the vibration signal of the planetary gearbox to be tested;

[0076] S3. Calculate the theoretical value of the gearbox fault characteristic frequency, fault cycle, and fault pulse interval based on the gear structural parameters;

[0077] S4. Setting various parameters of the sparrow search algorithm, optimizing the parameter combination [T, M] of the maximum correlation kurtosis deconvolution algorithm based on the sparrow search algorithm to obtain the best influencing parameter combination [T0, M0];

[0078] S5. Optimizing the parameters of the maximum correlation kurtosis deconvolution algorithm based on the optimal influencing parameter combination, and preprocessing the vibration signal using the maximum correlation kurtosis deconvolution algorithm with the optimized parameters to obtain a deconvolution signal of the vibration signal;

[0079] S6. performing noise reduction processing on the deconvolution signal based on a sparse coding shrinkage algorithm to obtain a noise-reduced signal;

[0080] S7, performing Hilbert envelope demodulation operation on the noise reduction signal and calculating the envelope spectrum;

[0081] S8. Compare the theoretical value of the gear fault characteristic frequency with the spectrum line in the envelope spectrum to identify the gearbox fault.

[0082] This example uses a wind turbine yaw planetary gearbox actually produced at a factory as the research object. This signal gearbox is a four-stage reduction transmission gearbox, and the structural parameters of the first-stage transmission component are shown in Table 1. A tiny dent was machined on the first-stage planetary gear tooth surface of the gearbox to simulate a minor tooth surface damage fault.

[0083] Table 1 Structural parameters of the first-stage transmission parts

[0084]

[0085] In this embodiment, the sensor model selected is PCB-352C33 acceleration sensor, the acquisition card is NI-9231, the chassis is cDAQ-9171, and the data acquisition and storage software is written using the Labview platform. The two sensors are vertically arranged at the first-stage transmission bearing support, and the sampling frequency is set to f. c The frequency is 12800 Hz, the sampling time t is 1 s, and the gearbox input motor speed is 940 r / min.

[0086] The time domain waveform of the planetary gear tooth surface collision is as follows: Figure 2 As shown in the figure, a small amount of impact can be vaguely seen in the time domain waveform, but the periodic characteristics of the impact caused by the fault are completely submerged by the background noise, and no regularity can be found. The envelope spectrum analysis of the measured signal is directly performed, and the results are as follows: Figure 3 As shown in the figure, it can be seen that the frequency corresponding to the amplitude peak is about 6Hz, which corresponds to the rotation frequency of the planetary gear of 5.22Hz. The frequencies of the other prominent amplitudes are seriously interfered by the surrounding frequency components, and the characteristic frequencies cannot be directly distinguished from the envelope spectrum.

[0087] In order to extract the fault characteristic frequency in the vibration signal, the present invention is used to process the original data. The maximum correlation kurtosis deconvolution algorithm is optimized based on the sparrow search algorithm, and sparse coding contraction is combined to perform planetary gearbox fault detection. Specifically, the following steps are included:

[0088] Step 1: Adsorb the acceleration sensor on the first-stage transmission bearing support of the planetary gearbox to be tested, and collect its vibration signal.

[0089] Step 2: Calculate the fault characteristic frequency, fault cycle, and fault pulse interval of each gearbox component based on the gear structure parameters. The specific calculation formula is:

[0090]

[0091]

[0092]

[0093]

[0094]

[0095] Among them, f s represents the characteristic frequency of sun gear fault, f p Indicates the characteristic frequency of planetary gear fault, f r Indicates the characteristic frequency of the inner ring gear fault, f m Indicates the meshing frequency, f crepresents the sampling frequency, f represents the fault characteristic frequency, T1 represents the theoretical fault cycle, ΔT represents the fault pulse interval, Z s Indicates the number of sun gear teeth, Z p Indicates the number of planetary gear teeth, Z r Number of teeth of tabular internal gear ring, N P Indicates the number of planetary gears.

[0096] Combined with the gearbox structural parameters, the theoretical values ​​of each characteristic frequency can be calculated as shown in Table 2

[0097] Table 2 Characteristic frequencies of various components of the gearbox

[0098]

[0099] Step 3: filtering the vibration signal using a maximum correlation kurtosis deconvolution algorithm to obtain an initial deconvolution signal;

[0100] The envelope spectrum entropy of the initial deconvolution signal is used as the fitness function of the sparrow search algorithm, and the minimum value of the envelope spectrum entropy is used as the basis for judging the optimization result of the sparrow search algorithm. The envelope spectrum entropy of the initial deconvolution signal is expressed as follows:

[0101]

[0102] Among them, p k Represents the probability distribution density of the envelope spectrum signal;

[0103] The range of parameters T and M in the maximum correlation kurtosis deconvolution algorithm is defined as the population search space of the sparrow search algorithm, and X is defined as i =[T i ,M i ] is the position of sparrow i, that is, the i-th parameter combination to be optimized by the sparrow search algorithm; at this time X i The food concentration expression at is F=E(X i ), that is, the combination to be optimized [T i ,M i ]The value of E corresponding to .

[0104] The adaptive selection process of parameters T and M in the maximum correlation kurtosis deconvolution algorithm is as follows:

[0105] Based on the theoretical fault cycle T1, the optimization intervals of the sparrow search algorithm parameters and the maximum correlation kurtosis deconvolution algorithm parameters are defined, and the number of sparrows n, the maximum number of iterations, the proportion of discoverers, the proportion of scouts, and the safety threshold are set.

[0106] Define the maximum correlation kurtosis deconvolution filter with a length of L, a displacement of M, and a period of T∈[f c / f-10,f c / f+10], where f c is the sampling frequency, f is the fault characteristic frequency;

[0107] Initialize the position of the sparrows, randomly generate multiple parameter combinations [T, M] as the initial position of the sparrow search, and calculate the fitness value of each sparrow's position to obtain the optimal individual i in the initial population best and the position X of the optimal individual in the spatial domain best ;

[0108] Update the positions of discoverers, joiners, and scouts in the population, and compare E(X i ) size, update the optimal individual i in the population best and the position X of the optimal individual in the spatial domain best ;

[0109] If the maximum number of iterations is met, stop the iteration and output the best influencing parameter combination [T0, M0], otherwise continue to update and iterate.

[0110] The relationship curve between the envelope spectrum entropy and the number of iterations is as follows: Figure 4 As shown in the figure, it can be seen that when the sparrow search algorithm iterates to the 14th generation, the envelope spectrum entropy value tends to be stable, and the corresponding optimal influencing parameter combination [T0, M0] is [1209, 6].

[0111] Step 4: Set the period parameter T0 of the maximum correlation kurtosis deconvolution algorithm to 1209 and the deconvolution displacement parameter M0 to 6. Use the maximum correlation kurtosis deconvolution algorithm with optimized parameters to preprocess the fault signal to obtain the deconvolution signal of the vibration signal. The processed signal is:

[0112]

[0113] y represents the deconvolution signal, x represents the vibration signal, F represents the maximum correlation kurtosis deconvolution filter coefficient with a length of L, and F=[f1 f2 … f L ] T , n represents the index value of the vibration signal sequence.

[0114] The filter coefficients are calculated as:

[0115]

[0116] in,

[0117]

[0118]

[0119]

[0120] β=[y1y 1-T y2y 2-T … y N y N-T ] T

[0121] Wherein, N represents the length of the vibration signal sequence, r=M*T, M=M0, and T=T0.

[0122] The obtained deconvolution signal waveform is as follows Figure 5 As shown in the figure, compared with the unprocessed vibration signal, it can be seen that some of the impact components that were originally masked by strong noise are effectively highlighted after deconvolution filtering.

[0123] Step 5: Use the sparse coding shrinkage algorithm to further reduce the noise of the deconvolution signal to obtain a noise-reduced signal. The signal after processing is:

[0124]

[0125] in, represents the denoised signal; y represents the deconvolution signal; α is the sparsity parameter that controls the probability density function. For non-Gaussian signals with impact characteristics, the parameter α can be set to 0.1; σ is the standard deviation of the white noise signal v, which can be estimated by the formula σ = MAD(y), where MAD(y) represents the median of the absolute deviations of the deconvolution signal y; d is the standard deviation of the deconvolution signal, and σ y is the standard deviation of the noise-reduced signal.

[0126] The deconvolution signal is subjected to denoising again by the sparse coding algorithm. The results are as follows: Figure 6 As shown in the figure, it can be seen that after the noise reduction process, a large amount of redundant interference noise is removed, while the periodic impact characteristics caused by the bump are well preserved.

[0127] Step 6: Perform Hilbert envelope demodulation on the noise reduction signal and calculate the envelope spectrum. The specific operations are as follows:

[0128] Perform Hilbert transform on the denoised signal s(t) to construct its analytical signal:

[0129] z(t)=s(t)+jy(t)

[0130] Where y(t) is the Hilbert transform of s(t), and its expression is:

[0131]

[0132] Where p is the Cauchy principal value;

[0133] Taking the modulus of the analytical signal, we get the envelope signal:

[0134]

[0135] Performing Fourier transform on the envelope signal to obtain an envelope spectrum.

[0136] The envelope demodulation of the signal after noise reduction is performed, and the result is as follows Figure 7 As shown in the envelope spectrum, it can be seen that the main frequency components are the characteristic frequency of the planetary gear tooth surface collision and its frequency multiple f p ~9f p .

[0137] Step 7: Compare the theoretical calculated value of the gear fault characteristic frequency with the spectrum line in the envelope spectrum, and it can be determined that the fault type of the planetary gearbox is the planetary gear tooth surface collision.

[0138] Another embodiment of the present invention provides a planetary gearbox fault detection device, including a memory and a processor. The memory stores a computer program, and the processor calls program instructions to execute the planetary gearbox fault detection method described in the above embodiment.

[0139] Another embodiment of the present invention further provides a computer-readable storage medium including a computer program, which can be executed by a processor to implement the planetary gearbox fault detection method as described in the above embodiment.

[0140] The above describes in detail the preferred embodiments of the present invention. It should be understood that those skilled in the art can make numerous modifications and variations based on the concepts of the present invention without inventive effort. Therefore, any technical solutions that can be derived by those skilled in the art through logical analysis, reasoning, or limited experimentation based on the concepts of the present invention and the prior art should be within the scope of protection defined by the claims.

Claims

1. A planetary gearbox fault detection method, characterized in that: The following steps are involved: Determine the planetary gearbox to be tested and obtain the gear structure parameters therein; Obtain the vibration signal of the planetary gearbox to be tested; Calculating the theoretical value of the gearbox fault characteristic frequency based on the gear structural parameters; Set the parameters of the sparrow search algorithm, and optimize the parameter combination of the maximum correlation kurtosis deconvolution algorithm based on the sparrow search algorithm to obtain the best influencing parameter combination; Based on the optimal influencing parameter combination, performing parameter optimization on the maximum correlation kurtosis deconvolution algorithm; Preprocessing the vibration signal using a maximum correlation kurtosis deconvolution algorithm with optimized parameters to obtain a deconvolution signal of the vibration signal; Performing noise reduction processing on the deconvolution signal based on a sparse coding shrinkage algorithm to obtain a noise-reduced signal; Performing envelope demodulation on the noise reduction signal and calculating an envelope spectrum; Comparing the theoretical value of the gearbox fault characteristic frequency with the spectrum line in the envelope spectrum to identify the gearbox fault; The parameters of the sparrow search algorithm are as follows: Performing filtering processing on the vibration signal using a maximum correlation kurtosis deconvolution algorithm to obtain an initial deconvolution signal; The envelope spectrum entropy of the initial deconvolution signal is used as the fitness function of the sparrow search algorithm, and the minimum value of the envelope spectrum entropy is used as the basis for judging the optimization result of the sparrow search algorithm. The envelope spectrum entropy of the initial deconvolution signal is expressed as follows: Among them, p k Represents the probability distribution density of the envelope spectrum signal; The parameter combination optimization of the maximum correlation kurtosis deconvolution algorithm is performed based on the sparrow search algorithm to obtain the best influencing parameter combination, including the following steps: The range of parameters T and M in the maximum correlation kurtosis deconvolution algorithm is defined as the population search space of the sparrow search algorithm; Define X i =[T i ,M i ] is the position of sparrow i, X i The food concentration expression at is F=E(X i ); Define the optimization interval of the sparrow search algorithm parameters and the maximum correlation kurtosis deconvolution algorithm parameters, set the number of sparrows n, the maximum number of iterations, the ratio of discoverers, the ratio of scouts and the safety threshold; Define the maximum correlation kurtosis deconvolution filter with a length of L, a displacement of M, and a period of T∈[f c / f-10,f c / f+10], where f c is the sampling frequency, f is the fault characteristic frequency; Initialize the position of the sparrows, randomly generate multiple parameter combinations [T, M] as the initial position of the sparrow search, and calculate the fitness value of each sparrow's position to obtain the optimal individual i in the initial population best and the position X of the optimal individual in the spatial domain best ; Update the positions of discoverers, joiners, and scouts in the population, and compare E(X i ) size, update the optimal individual i in the population best and the position X of the optimal individual in the spatial domain best ; If the maximum number of iterations is met, stop the iteration and output the best influencing parameter combination [T0, M0], otherwise continue to update and iterate.

2. A planetary gearbox fault detection method according to claim 1, characterized in that: The specific calculation formula for calculating the theoretical value of the gearbox fault characteristic frequency is: Among them, f s represents the characteristic frequency of sun gear fault, f p Indicates the characteristic frequency of planetary gear fault, f r Indicates the characteristic frequency of the internal gear ring fault, f m Indicates the meshing frequency, f c represents the sampling frequency, f represents the fault characteristic frequency, T1 represents the theoretical fault cycle, ΔT represents the fault pulse interval, Z s Indicates the number of sun gear teeth, Z p Indicates the number of planetary gear teeth, Z r Number of teeth of tabular internal gear ring, N P Indicates the number of planetary gears.

3. A planetary gearbox fault detection method according to claim 1, characterized in that: Based on the optimal influencing parameter combination [T0, M0], the maximum correlation kurtosis deconvolution algorithm is optimized, and the period parameter of the maximum correlation kurtosis deconvolution algorithm is set to T0 and the deconvolution shift parameter is set to M0; The vibration signal is preprocessed using a maximum correlation kurtosis deconvolution algorithm with optimized parameters to obtain a deconvolution signal of the vibration signal, which is expressed as follows: y represents the deconvolution signal, x represents the vibration signal, F represents the maximum correlation kurtosis deconvolution filter coefficient with a length of L, and F = [f1 f2…f L ] T , n represents the index value of the vibration signal sequence.

4. A planetary gearbox fault detection method according to claim 3, characterized in that: The calculation formula of the filter coefficient is: in, β=[y1y 1-T y2y 2-T …y N y N-T ] T Wherein, N represents the length of the vibration signal sequence, r=M*T, M=M0, and T=T0.

5. A planetary gearbox fault detection method according to claim 1, characterized in that: The deconvolution signal is subjected to noise reduction processing based on the sparse coding shrinkage algorithm to obtain a noise reduction signal, which is expressed as: Where s(t) represents the denoised signal; y represents the deconvolution signal; α is the sparsity parameter that controls the probability density function. σ is the standard deviation of the white noise signal, d is the standard deviation of the deconvolution signal, and σ y is the standard deviation of the noise-reduced signal.

6. A planetary gearbox fault detection method according to claim 1, characterized in that: Performing envelope demodulation on the noise reduction signal and calculating the envelope spectrum includes the following steps: Perform Hilbert transform on the denoised signal s(t) to construct its analytical signal: z(t)=s(t)+jy(t) Where y(t) is the Hilbert transform of s(t), and its expression is: Where p is the Cauchy principal value; The envelope signal is obtained by taking the modulus of the analytical signal: Performing Fourier transform on the envelope signal to obtain an envelope spectrum.

7. A planetary gearbox fault detection device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor calls the program to execute the planetary gearbox fault detection method according to any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that The method comprises a computer program, wherein the computer program can be executed by a processor to implement the planetary gearbox fault detection method according to any one of claims 1 to 6.

Images