Fan gearbox fault diagnosis method and system based on data fusion

Abstract

The invention belongs to the technical field of fan gear box fault diagnosis, and particularly relates to a fan gear box fault diagnosis method and system based on data fusion, the system comprises a data acquisition unit, a data processing unit and an output unit, features of different scales of fan gear box vibration data are extracted, feature fusion is realized by using a multi-scale convolution module, and a fault diagnosis result is obtained. The method can effectively alleviate the defect that a single scale feature cannot accurately express fault information, multiple scale features complement each other, the fault detection efficiency is improved, and the anti-interference capability is enhanced.

Description

Technical Field

[0001] This application belongs to the field of wind turbine gearbox fault diagnosis technology, specifically relating to a wind turbine gearbox fault diagnosis method and system based on data fusion. Background Technology

[0002] In wind turbines, the gearbox is a critical component of the transmission system, and maintaining its healthy condition is crucial for the entire wind turbine system. Therefore, an accurate and reliable wind turbine gearbox fault model is particularly important. Existing data-driven wind turbine gearbox fault diagnosis methods are mostly "black box" models with low interpretability. Therefore, it is necessary to design a reliable and accurate fault diagnosis method to accurately identify faults in their early stages and ensure the healthy operation of the wind turbine. Furthermore, wind turbine systems are complex and operate in harsh and variable environments. For gearbox fault diagnosis, simple time-domain features are sensitive to impact faults (such as broken gear teeth) but struggle to handle complex faults and signals with high background noise. Single frequency-domain features can effectively distinguish the correspondence between fault type and frequency, but their effectiveness is poor for non-stationary signals. Summary of the Invention

[0003] To address the aforementioned problems, this invention aims to provide a data-based method for wind turbine gearbox feature extraction, fusion, and fault diagnosis, overcoming the shortcomings of existing methods such as low interpretability and the inability of single-scale features to accurately represent data characteristics. This invention adopts the following technical solution: A method for fault diagnosis of wind turbine gearboxes based on data fusion includes the following steps: S1. Construct a dataset based on the collected vibration data of the wind turbine gearbox; S2. Process the constructed dataset and group the time series data according to rotational speed; S3. Extract the time domain, frequency domain, wavelet packet decomposition derived features, and spectral entropy features based on local feature scale decomposition for each data sample; S4. The extracted features are fused using a multi-scale convolution module to obtain the final data feature representation; S5. First, input the feature data into the data mapping layer to obtain the representation of the gearbox feature data in the hidden space; S6. Construct a multi-head attention model to discover global and local dependencies in feature data; S7. Construct a classifier to output the specific fault type.

[0004] Preferably, in step S2, linear interpolation is used to interpolate the missing values ​​in the dataset. x,y Fill in the blanks using the following formula: (1); Fill in the erroneous data with the mean; The vibration data for each revolution of the gearbox is divided into groups based on the gearbox rotation speed, and each group of data is used as a data sample.

[0005] Preferably, step S3 involves the following time-domain features: absolute mean, peak value, RMS value, root square amplitude, variance, peak-to-peak value, skewness index, kurtosis index, peak value index, waveform index, impulse index, margin index, coefficient of variation, minimum value, average value, rectified average value, 3rd-order center distance, 4th-order center distance, 3rd-order origin distance, and 4th-order origin distance, resulting in a time-domain feature vector. ; Frequency domain features: centroid frequency, mean square frequency, root mean square frequency, frequency variance, frequency standard deviation, spectral peak stability coefficient, relative energy of band 1, relative energy of band 2, relative energy of band 3, relative energy of band 4, and relative energy of band 5, yielding the frequency domain feature vector. .

[0006] Preferably, step S3, wavelet packet decomposition, can simultaneously perform multi-level decomposition of the low-frequency and high-frequency components of the signal, and its derived features can simultaneously reflect the time-domain and frequency-domain characteristics of the signal. The steps are as follows: First, wavelet packet decomposition is performed on each data sample, mathematically expressed as follows: (2); (3); in, It is the coefficient of the nth node in the j-th layer. and These are low-pass and high-pass filters, respectively; Secondly, after wavelet packet decomposition, relevant derived features are extracted: wavelet packet energy, wavelet packet energy ratio, wavelet packet node 1 energy, wavelet packet node 2 energy, wavelet packet node 3 energy, and wavelet packet node 4 energy, to obtain the time-frequency domain feature vector. .

[0007] Preferably, the steps for obtaining spectral entropy features based on local feature scale decomposition are as follows: (1) Extract the extreme points of the signal and find all local maxima and local minima of the signal; (2) Construct upper and lower envelopes, and interpolate the maximum points to form the upper envelope. Interpolation is performed on the minimum points to form the lower envelope. ; (3) Calculate the mean envelope. ; (4) Extract intrinsic scaling components by subtracting the mean envelope from the original signal to obtain candidate intrinsic scaling components. ; The condition for determining whether it is an effective intrinsic scale component is: The number of maxima and minima is equal and they alternate. The difference between the number of zero-crossing points and the number of extreme points does not exceed 1; if the above two requirements are met, then Considered as an intrinsic metric component, denoted as Otherwise, for Repeat the above decomposition process; (5) Recursive decomposition, remaining signal As a new input, continue to decompose until the remaining signal becomes a monotonic trend or noise; (6) After the local feature scale LCD decomposition is completed, the LCD energy spectral entropy, LCD singular value spectral entropy, LCD envelope spectral entropy, LCD component 1 energy, and LCD component 2 energy are extracted to obtain the LCD spectral entropy feature vector. .

[0008] Preferably, S4 performs feature fusion on the extracted features using a multi-scale convolution module, with the following steps: S41. The extracted feature vectors are directly concatenated to obtain... ,Right now , This serves as the input for the multi-scale convolution module; S411. Input After a convolution kernel The convolutional layer is obtained This convolutional layer can increase the number of channels, allowing data to interact with the next convolutional layer; S412. After three different operations, the result was obtained. , and The three operations are copying, convolution kernel... The convolutional layers and convolutional kernels are Convolutional layers; S413. , and Perform convolution kernel as The convolution operations respectively yielded , and ; S42. Will , and spliced ​​together , This is the final output of the proposed multi-scale feature extraction and fusion method.

[0009] Preferably, step S5 is for a given input sample , X From feature data The structure is achieved through the following linear transformation to represent the features of the input data in the hidden space, while also satisfying the dimensionality requirements of the multi-head attention module: (4); in, For the mapping weight matrix, For the bias vector, The mapped feature representation, To hide spatial dimensions; B The batch size of the input data. L For sequence length, D in For input dimensions.

[0010] Preferably, step S6 includes: S61. Single-head attention calculation ; S62. Multi-head attention is obtained by concatenating the outputs of all heads and performing a linear transformation. ; S63. After multi-head attention output, add residual connections and perform layer normalization to obtain... ; After S64.FFN output, residual connection and layer normalization are performed again to obtain: (13); Complete N The stacking process of self-attention layers is as follows: (14); The final output of the self-attention module is: (15).

[0011] Preferably, step S7 employs a three-layer KAN network structure. ,in Number of fault categories: The forward propagation process of the KAN network is as follows: (16); (17); (18); in For the first KAN transformation of the layer, This is the first layer output. This is the output of the second layer; Each transformation consists of a series of learnable B-spline basis functions: (19); in For the first A B-spline basis function, For the above , or , These are learnable weights; Final output This represents the predicted probability distribution for each fault category, which is the final fault diagnosis result for the wind turbine gearbox.

[0012] A wind turbine gearbox fault diagnosis system based on data fusion includes a data acquisition unit, a data processing unit, and an output unit; Data acquisition unit: Collects vibration data from the wind turbine gearbox and constructs a dataset; Data processing unit: Extracts time-domain, frequency-domain, wavelet packet decomposition-derived features and spectral entropy features based on local feature scale decomposition for each data sample; utilizes a multi-head attention model to discover global and local dependencies in the feature data; constructs a classifier to output the specific fault type; Output unit: Visualizes the results.

[0013] Compared with the prior art, the beneficial effects of this application are as follows: The design incorporates the above four types of features to create a fault extraction and fusion method. The multi-scale features enable complementary signal features, improve the detection rate of early faults, better adapt to complex faults, and enhance their anti-interference capabilities. Attached Figure Description

[0014] Figure 1 This is a flowchart of the fault diagnosis model. Figure 2 Here is a flowchart of the feature extraction and fusion process; Figure 3 This is a diagram showing the experimental results. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, not all of them. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] A data-based method for feature extraction, fusion, and fault diagnosis of wind turbine gearboxes includes the following steps: S1, Construct a dataset based on the collected vibration data of the wind turbine gearbox.

[0017] S2 processes the constructed dataset, fills in missing values, corrects erroneous data, and groups the time series data according to rotational speed.

[0018] Use linear interpolation to find missing values ​​in the dataset. x, y Fill in the blanks using the following formula: (1); For obviously erroneous data, fill in the mean. Divide the vibration data for each revolution of the gearbox into groups based on the gearbox speed, and use each group as a data sample.

[0019] S3 extracts the time domain, frequency domain, time-frequency domain, and spectral entropy features of each data sample based on local feature scale decomposition.

[0020] Time-domain features reflect the core characteristics of a signal, such as amplitude, volatility, periodicity, and impulsivity. The following time-domain features are extracted for each data sample: absolute mean, peak value, RMS value, root square amplitude, variance, peak-to-peak value, skewness index, kurtosis index, peak value index, waveform index, impulse index, margin index, coefficient of variation, minimum value, average value, rectified average value, 3rd-order center distance, 4th-order center distance, 3rd-order origin distance, and 4th-order origin distance, resulting in a time-domain feature vector. .

[0021] Frequency domain features can reveal the periodicity and frequency characteristics of a signal. For each data sample, the following frequency domain features are extracted: centroid frequency, mean square frequency, root mean square frequency, frequency variance, frequency standard deviation, spectral peak stability coefficient, relative energy of band 1, relative energy of band 2, relative energy of band 3, relative energy of band 4, and relative energy of band 5, resulting in a frequency domain feature vector. .

[0022] Wavelet packet decomposition can simultaneously perform multi-level decomposition of the low-frequency and high-frequency components of a signal, and its derived features can simultaneously reflect the time-domain and frequency-domain characteristics of the signal. First, wavelet packet decomposition is performed on each data sample, mathematically expressed as follows: (2); (3); in, It is the coefficient of the nth node in the j-th layer. and These are low-pass and high-pass filters, respectively. After wavelet packet decomposition, relevant derived features are extracted: wavelet packet energy, wavelet packet energy ratio, wavelet packet node 1 energy, wavelet packet node 2 energy, wavelet packet node 3 energy, and wavelet packet node 4 energy, resulting in a time-frequency domain feature vector. .

[0023] Spectral entropy features based on local feature scale decomposition have the advantage of high sensitivity to minor faults. Therefore, local feature scale decomposition is first performed on each data sample, as follows.

[0024] (1) Extract the extreme points of the signal and find all local maxima and local minima of the signal.

[0025] (2) Construct upper and lower envelopes, and interpolate the maximum points to form the upper envelope. Interpolation is performed on the minimum points to form the lower envelope. .

[0026] (3) Calculate the mean envelope. .

[0027] (4) Extract intrinsic scaling components by subtracting the mean envelope from the original signal to obtain candidate intrinsic scaling components. To determine whether it is an effective intrinsic scale component, the following conditions must be met: The number of maxima and minima is equal and they alternate. The difference between the number of zero-crossing points and the number of extreme points does not exceed 1. If both of the above requirements are met, then... Considered as an intrinsic metric component, denoted as Otherwise, for Repeat the above decomposition process.

[0028] (5) Recursive decomposition, remaining signal As a new input, the decomposition continues until the remaining signal becomes a monotonic trend or noise (intrinsic scale components cannot be extracted).

[0029] After the Local Feature Scale Decomposition (LCD) is completed, the LCD energy spectral entropy, LCD singular value spectral entropy, LCD envelope spectral entropy, LCD component 1 energy, and LCD component 2 energy are extracted to obtain the LCD spectral entropy eigenvector. .

[0030] S4 performs feature fusion on the extracted features using a multi-scale convolution module.

[0031] The extracted feature vectors are directly concatenated to obtain... ,Right now , This is the input for the multi-scale convolutional module. First, the input... After a convolution kernel The convolutional layer is obtained This convolutional layer can increase the number of channels, allowing data to interact with the next convolutional layer.

[0032] Secondly After three different operations, the result was obtained. , and The three operations are copying, convolution kernel... The convolutional layers and convolutional kernels are The convolutional layer.

[0033] Then, , and Perform convolution kernel as The convolution operations respectively yielded , and This convolutional layer not only increases nonlinearity but also reduces the number of parameters and computational cost without compromising the module's expressive power. Furthermore, the two convolutional kernels in this module are... The convolutional layers enable the fusion of data information between different channels.

[0034] Finally, , and spliced ​​together , This is the final output of the proposed multi-scale feature extraction and fusion method.

[0035] S5, first input the data into the data mapping layer to obtain the representation of the gearbox data in the hidden space.

[0036] For a given input sample (from feature data) The following linear transformation is used to represent the features of the input data in the hidden space and to meet the dimensionality requirements of the multi-head attention module.

[0037] (4); in, For the mapping weight matrix, For the bias vector, The mapped feature representation, To hide the spatial dimension.

[0038] S6 constructs a multi-head attention module, which includes multiple attention layers. Each attention layer consists of an attention mechanism, a residual normalization layer, and a feedforward network.

[0039] For the Layer input First, a query, key, and value matrix is ​​generated through linear transformation. For the... One point of attention: (5); (6); (7); in , , For the number of attention heads.

[0040] Single-head attention is calculated as follows: (8); scaling factor This is used to prevent the gradient from vanishing due to an excessively large inner product.

[0041] Multi-head attention works by concatenating the outputs of all heads and performing a linear transformation: (9); in This is for outputting the projection matrix.

[0042] After multi-head attention output, residual connections are added and layer normalization is performed: (10); The formula for calculating layer normalization is: (11); in: For layer normalization input, The characteristic mean, For characteristic variance, For learnable scaling and offset parameters, is the numerical stability constant.

[0043] The feedforward network consists of two linear transformations and a ReLU activation function: (12); in: , , , , This represents the intermediate dimension of the feedforward network.

[0044] After the FFN output, residual connections and layer normalization are performed again: (13); The complete stacking process of N self-attention layers is as follows: (14); The final output of the self-attention module is: (15); for Self-attention output.

[0045] S7. Construct the Kolmogorov-Arnold Network (KAN) classifier to output the specific fault type.

[0046] The Kolmogorov-Arnold Network, based on the Kolmogorov-Arnold representation theorem, achieves nonlinear mapping through learnable spline functions. This paper employs a three-layer KAN network structure. ,in This represents the number of fault categories.

[0047] The forward propagation process of the KAN network is as follows: (16); (17); (18); in For the first The KAN transform of each layer consists of a series of learnable B-spline basis functions: (19); in For the first A B-spline basis function, These are learnable weights.

[0048] Final output For each type of fault.

[0049] The above method yielded the following fault diagnosis results under four operating conditions on the MCC5-THU gearbox dataset: Figure 3 .from Figure 3As can be seen, in operating condition 1, the motor input shaft rotates at a constant speed of 1000 rpm, and the torque of the gearbox output shaft is between 0-10 Nm. In operating condition 2, the motor input shaft rotates at a constant speed of 2000 rpm, and the torque of the gearbox output shaft is between 0-20 Nm. In operating condition 3, the gear output torque is 10 Nm, and the motor input shaft rotates at a speed of 0-2000 rpm. In operating condition 4, the gear output torque is 20 Nm, and the motor input shaft rotates at a speed of 0-2000 rpm. Each operating condition includes data on the gearbox under five fault conditions: missing teeth, wear, pitting, fracture, and cracks, as well as data on its healthy operation.

[0050] A wind turbine gearbox fault diagnosis system based on data fusion includes a data acquisition unit, a data processing unit, and an output unit; Data acquisition unit: Collects vibration data from the wind turbine gearbox and constructs a dataset; Data processing unit: Extracts time-domain, frequency-domain, wavelet packet decomposition-derived features and spectral entropy features based on local feature scale decomposition for each data sample; utilizes a multi-head attention model to discover global and local dependencies in the feature data; constructs a classifier to output the specific fault type; Output unit: Visualizes the results.

[0051] The foregoing has shown and described the basic principles, main features, and advantages of this application. Those skilled in the art should understand that this application is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the application. Various changes and modifications can be made without departing from the spirit and scope of this application, and all such changes and modifications fall within the scope of the claims. The scope of protection of this application is defined by the appended claims and their equivalents.

Claims

1. A method for fault diagnosis of wind turbine gearboxes based on data fusion, characterized in that, Includes the following steps: S1. Construct a dataset based on the collected vibration data of the wind turbine gearbox; S2. Process the constructed dataset and group the time series data according to rotational speed; S3. Extract the time domain, frequency domain, wavelet packet decomposition derived features, and spectral entropy features based on local feature scale decomposition for each data sample; S4. The extracted features are fused using a multi-scale convolution module to obtain the final data feature representation; S5. First, input the feature data into the data mapping layer to obtain the representation of the gearbox feature data in the hidden space; S6. Construct a multi-head attention model to discover global and local dependencies in feature data; S7. Construct a classifier to output the specific fault type.

2. The wind turbine gearbox fault diagnosis method based on data fusion according to claim 1, characterized in that, In step S2, linear interpolation is used to correct the missing values ​​in the dataset. x,y Fill in the blanks using the following formula: （1）； Fill in the erroneous data with the mean; The vibration data for each revolution of the gearbox is divided into groups based on the gearbox rotation speed, and each group of data is used as a data sample.

3. The wind turbine gearbox fault diagnosis method based on data fusion according to claim 1, characterized in that, Step S3: Time-domain features: absolute mean, peak value, RMS value, root square amplitude, variance, peak-to-peak value, skewness index, kurtosis index, peak index, waveform index, impulse index, margin index, coefficient of variation, minimum value, average value, rectified average value, 3rd order center distance, 4th order center distance, 3rd order origin distance, 4th order origin distance, to obtain the time-domain feature vector. ; Frequency domain features: centroid frequency, mean square frequency, root mean square frequency, frequency variance, frequency standard deviation, spectral peak stability coefficient, relative energy of band 1, relative energy of band 2, relative energy of band 3, relative energy of band 4, and relative energy of band 5, yielding the frequency domain feature vector. .

4. The wind turbine gearbox fault diagnosis method based on data fusion according to claim 3, characterized in that, Step S3, wavelet packet decomposition, can simultaneously perform multi-level decomposition of the low-frequency and high-frequency components of a signal. Its derived features can simultaneously reflect the time-domain and frequency-domain characteristics of the signal. The steps are as follows: First, wavelet packet decomposition is performed on each data sample, mathematically expressed as follows: （2）； （3）； in, It is the coefficient of the nth node in the j-th layer. and These are low-pass and high-pass filters, respectively; Secondly, after wavelet packet decomposition, relevant derived features are extracted: wavelet packet energy, wavelet packet energy ratio, wavelet packet node 1 energy, wavelet packet node 2 energy, wavelet packet node 3 energy, and wavelet packet node 4 energy, to obtain the time-frequency domain feature vector. .

5. The wind turbine gearbox fault diagnosis method based on data fusion according to claim 1, characterized in that, The steps for obtaining spectral entropy features based on local feature scale decomposition are as follows: (1) Extract the extreme points of the signal and find all local maxima and local minima of the signal; (2) Construct upper and lower envelopes, and interpolate the maximum points to form the upper envelope. Interpolation is performed on the minimum points to form the lower envelope. ; (3) Calculate the mean envelope. ; (4) Extract intrinsic scaling components by subtracting the mean envelope from the original signal to obtain candidate intrinsic scaling components. ; The condition for determining whether it is an effective intrinsic scale component is: The number of maxima and minima is equal and they alternate. The difference between the number of zero-crossing points and the number of extreme points does not exceed 1; if the above two requirements are met, then Considered as an intrinsic metric component, denoted as Otherwise, for Repeat the above decomposition process; (5) Recursive decomposition, residual signal As a new input, continue to decompose until the remaining signal becomes a monotonic trend or noise; (6) After the local feature scale LCD decomposition is completed, the LCD energy spectral entropy, LCD singular value spectral entropy, LCD envelope spectral entropy, LCD component 1 energy, and LCD component 2 energy are extracted to obtain the LCD spectral entropy feature vector. .

6. The wind turbine gearbox fault diagnosis method based on data fusion according to claim 4, characterized in that, S4 performs feature fusion on the extracted features using a multi-scale convolution module, as follows: S41. The extracted feature vectors are directly concatenated to obtain... ,Right now , This serves as the input for the multi-scale convolution module; S411. Input After a convolution kernel The convolutional layer is obtained This convolutional layer can increase the number of channels, allowing data to interact with the next convolutional layer; S412. After three different operations, the result was obtained. , and The three operations are copying, convolution kernel... The convolutional layers and convolutional kernels are Convolutional layers; S413. , and Perform convolution kernel as The convolution operations respectively yielded , and ; S42. Will , and spliced ​​together , This is the final output of the proposed multi-scale feature extraction and fusion method.

7. The wind turbine gearbox fault diagnosis method based on data fusion according to claim 6, characterized in that, Step S5: For a given input sample , X From feature data The structure is achieved through the following linear transformation to represent the features of the input data in the hidden space, while also satisfying the dimensionality requirements of the multi-head attention module: （4）； in, For the mapping weight matrix, For the bias vector, The mapped feature representation, To hide spatial dimensions; B The batch size of the input data. L For sequence length, D in For input dimensions.

8. The wind turbine gearbox fault diagnosis method based on data fusion according to claim 7, characterized in that, Step S6 includes: S61. Single-head attention calculation ; S62. Multi-head attention is obtained by concatenating the outputs of all heads and performing a linear transformation. ; S63. After multi-head attention output, add residual connections and perform layer normalization to obtain... ; After S64.FFN output, residual connection and layer normalization are performed again to obtain: （13）； Complete N The stacking process of self-attention layers is as follows: （14）； The final output of the self-attention module is: （15）。 9. The wind turbine gearbox fault diagnosis method based on data fusion according to claim 7, characterized in that, Step S7 uses a three-layer KAN network structure. ,in Number of fault categories: The forward propagation process of the KAN network is as follows: （16）； （17）； （18）； in For the first KAN transformation of the layer, This is the first layer output. This is the output of the second layer; Each transformation consists of a series of learnable B-spline basis functions: （19）； in For the first A B-spline basis function, For the above , or , These are learnable weights; Final output This represents the predicted probability distribution for each fault category, which is the final fault diagnosis result for the wind turbine gearbox.

10. A wind turbine gearbox fault diagnosis system based on data fusion, employing the wind turbine gearbox fault diagnosis method based on data fusion as described in any one of claims 1-9, characterized in that, It includes a data acquisition unit, a data processing unit, and an output unit; Data acquisition unit: Collects vibration data from the wind turbine gearbox and constructs a dataset; Data processing unit: Extracts time-domain, frequency-domain, wavelet packet decomposition-derived features and spectral entropy features based on local feature scale decomposition for each data sample; utilizes a multi-head attention model to discover global and local dependencies in the feature data; constructs a classifier to output the specific fault type; Output unit: Visualizes the results.

Images