Wind power generation equipment fault diagnosis method and device, electronic equipment and storage medium

Abstract

The application discloses a wind power generation equipment fault diagnosis method and device, electronic equipment and a storage medium, and relates to the technical field of wind power generators. The wind power generation equipment fault diagnosis method comprises the following steps: collecting a vibration signal of a wind power generation equipment; extracting a fault characteristic parameter from the vibration signal; classifying the fault characteristic parameter through a preset target support vector machine model and a target clustering analysis model, so as to obtain a first classification result and a second classification result; and determining a fault diagnosis result corresponding to the wind power generation equipment based on the first classification result and the second classification result. The application solves the technical problems of low accuracy and low efficiency of the current wind power generation equipment fault diagnosis method.

Description

Technical Field

[0001] This application relates to the field of wind turbine technology, and in particular to a method, apparatus, electronic device and storage medium for diagnosing faults in wind power generation equipment. Background Technology

[0002] The gearbox of a wind turbine is one of the core components of wind power generation equipment, undertaking the crucial task of transmitting and converting power during wind power generation. However, due to the long-term operation of wind turbines and the influence of harsh environmental conditions, gearboxes often face various faults, such as gear wear and bearing failure. Therefore, frequent equipment maintenance is necessary. Vibration signal analysis based on wind power equipment is currently the mainstream fault diagnosis method. However, due to the complex working environment and diverse fault types of gearboxes, traditional fault diagnosis methods may not be able to fully capture and accurately diagnose various faults. Therefore, current vibration analysis methods may have low accuracy in gearbox fault diagnosis. Moreover, vibration analysis-based fault diagnosis methods require offline analysis or manual operation, resulting in long detection cycles and low fault diagnosis efficiency. Summary of the Invention

[0003] The main objective of this application is to provide a method, device, electronic equipment, and storage medium for diagnosing faults in wind power generation equipment, aiming to solve the technical problems of low accuracy and low efficiency in current methods for diagnosing faults in wind power generation equipment.

[0004] To achieve the above objectives, this application provides a method for diagnosing faults in wind power generation equipment, the method comprising:

[0005] Collect vibration signals from wind power generation equipment;

[0006] Fault feature parameters are extracted from the vibration signal, and the fault feature parameters are classified by a preset target support vector machine model and a target clustering analysis model to obtain a first classification result and a second classification result.

[0007] Based on the first classification result and the second classification result, the fault diagnosis result corresponding to the wind power generation equipment is determined.

[0008] Optionally, the step of determining the fault diagnosis result corresponding to the wind power generation equipment based on the first classification result and the second classification result includes:

[0009] Extract the principal components corresponding to the first classification result and the second classification result, respectively;

[0010] Calculate the Mahalanobis distance between each principal component of the first classification result and each principal component of the second classification result;

[0011] Based on the Mahalanobis distance between the principal components, a target principal component is selected, and the classification evaluation index corresponding to the target principal component is set as the fault diagnosis result.

[0012] Optionally, after the step of collecting vibration signals from the wind power generation equipment, the method further includes:

[0013] The vibration signal is subjected to noise reduction, filtering, and data normalization processes in sequence to obtain an optimized vibration signal.

[0014] The vibration signal is updated by optimizing the vibration signal to obtain the updated vibration signal.

[0015] Optionally, the fault characteristic parameters include at least one of time-domain characteristic parameters, frequency-domain characteristic parameters, statistical characteristic parameters, and wavelet transform characteristic parameters, and the step of extracting fault characteristic parameters from the vibration signal includes:

[0016] The vibration signal is subjected to time-domain analysis to obtain time-domain characteristic parameters, wherein the time-domain characteristic parameters include mean, variance, kurtosis and kurtosis;

[0017] The vibration signal is subjected to frequency domain analysis to obtain frequency domain characteristic parameters, wherein the frequency domain characteristic parameters include spectral peaks and energy distribution;

[0018] The autocorrelation of the vibration signal and its correlation with other signals are calculated to obtain statistical characteristic parameters, wherein the statistical characteristic parameters include an autocorrelation function and a cross-correlation function.

[0019] Wavelet coefficients and wavelet coefficient energy are extracted from the vibration signal to obtain wavelet transform characteristic parameters.

[0020] Optionally, before the step of classifying the fault feature parameters using a preset target support vector machine model and a target clustering analysis model to obtain a first classification result and a second classification result, the method further includes:

[0021] Vibration signals were collected from wind power generation equipment under different operating conditions to obtain multiple sets of training samples, and the fault type label corresponding to each training sample was obtained.

[0022] The training samples labeled with fault types are input into the initial support vector machine model to obtain the corresponding third classification result;

[0023] Input the training samples without fault type labels into the initial clustering analysis model to obtain the corresponding fourth classification results;

[0024] Based on the third classification result and the fourth classification result, calculate the Mahalanobis distance between the principal components corresponding to the third classification result and the fourth classification result respectively;

[0025] If the Mahalanobis distance is greater than a preset threshold, the model parameters of the initial support vector machine model and the initial clustering analysis model are optimized and adjusted, and the execution step is returned: input the training samples with fault type labels into the initial support vector machine model to obtain the corresponding third classification result;

[0026] If the Mahalanobis distance is not greater than the preset threshold, then the current initial support vector machine model and the initial clustering analysis model are set to the target support vector machine model and the target clustering analysis model, respectively.

[0027] Optionally, after the step of determining the fault diagnosis result corresponding to the wind power generation equipment based on the first classification result and the second classification result, the method further includes:

[0028] Based on the fault diagnosis results, a fault warning signal is generated;

[0029] The fault warning signal is used to notify relevant personnel so that they can take the corresponding maintenance measures based on the fault diagnosis results.

[0030] Optionally, after the step of notifying relevant personnel via the fault warning signal, the method further includes:

[0031] Obtain the maintenance results of the wind power generation equipment, wherein the maintenance results include at least the fault type, severity, and vibration characteristics;

[0032] Based on the inspection results and the fault diagnosis results, the model parameters of the target support vector machine model and the target clustering analysis model are optimized.

[0033] This application also provides a wind power equipment fault diagnosis device, which is applied to a wind power equipment fault diagnosis device, and the wind power equipment fault diagnosis device includes:

[0034] The signal acquisition module is used to collect vibration signals from wind power generation equipment;

[0035] The fault classification module is used to extract fault feature parameters from the vibration signal, and classify the fault feature parameters by a preset target support vector machine model and a target clustering analysis model to obtain a first classification result and a second classification result.

[0036] The fault diagnosis module is used to determine the fault diagnosis result corresponding to the wind power generation equipment based on the first classification result and the second classification result.

[0037] This application also provides an electronic device, which is a physical device, comprising: a memory, a processor, and a program for the wind power equipment fault diagnosis method stored in the memory and executable on the processor. When the program for the wind power equipment fault diagnosis method is executed by the processor, it can implement the steps of the wind power equipment fault diagnosis method as described above.

[0038] This application also provides a computer-readable storage medium storing a program for implementing a fault diagnosis method for wind power generation equipment. When the program for the fault diagnosis method for wind power generation equipment is executed by a processor, it implements the steps of the fault diagnosis method for wind power generation equipment as described above.

[0039] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the wind power generation equipment fault diagnosis method described above.

[0040] This application provides a method, apparatus, electronic device, and computer-readable storage medium for fault diagnosis of wind power generation equipment. First, vibration signals from the wind power generation equipment are collected. Then, fault feature parameters are extracted from the vibration signals. These fault feature parameters are classified using a preset target support vector machine model and a target clustering analysis model, respectively, to obtain a first classification result and a second classification result. Based on the first and second classification results, the corresponding fault diagnosis result for the wind power generation equipment is determined. The technical solution of this application classifies fault feature parameters in the vibration signals of wind power generation equipment by combining a support vector machine model and a clustering model. The support vector machine model has a strong ability to characterize classification boundaries, while the clustering model can aggregate and classify fault feature parameters. The combined use of both can more accurately identify and distinguish different types of faults, improving the accuracy and reliability of fault diagnosis. Furthermore, the preset target support vector machine model and target clustering analysis model can automatically analyze and diagnose the collected vibration signals in real time, eliminating the need for manual operation and effectively improving the diagnostic efficiency of wind power generation equipment. Attached Figure Description

[0041] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0042] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart illustrating the first embodiment of the wind power equipment fault diagnosis method of this application;

[0044] Figure 2 This is a schematic diagram of the vibration signals collected by the sensors in the fault diagnosis method for wind power generation equipment of this application;

[0045] Figure 3 This is a schematic diagram of the vibration signal after preprocessing in the wind power equipment fault diagnosis method of this application.

[0046] Figure 4 This is a schematic diagram illustrating the extraction of vibration signal characteristic parameters at different locations in the wind power equipment fault diagnosis method of this application;

[0047] Figure 5 This is a flowchart illustrating the second embodiment of the wind power equipment fault diagnosis method of this application;

[0048] Figure 6 This is a schematic diagram illustrating the classification of two-dimensional sample points using a clustering algorithm incorporating support vector machines in the wind power equipment fault diagnosis method of this application.

[0049] Figure 7 This is a schematic diagram illustrating the classification of vibration signals in the fault diagnosis method for wind power generation equipment of this application;

[0050] Figure 8 This is a schematic diagram illustrating the overall process of model training and model application in the wind power equipment fault diagnosis method of this application;

[0051] Figure 9 This is a flowchart illustrating the first embodiment of the wind power equipment fault diagnosis method of this application;

[0052] Figure 10 This is a schematic diagram of the fault confirmation and model update process in the wind power equipment fault diagnosis method of this application;

[0053] Figure 11 This is a schematic diagram of the overall process of fault diagnosis and early warning for wind power generation equipment in the fault diagnosis method of this application;

[0054] Figure 12 This is a schematic diagram of the composition structure of the wind power equipment fault diagnosis device in the embodiments of this application;

[0055] Figure 13 This is a schematic diagram of the hardware operating environment involved in the wind power equipment fault diagnosis method in this application embodiment.

[0056] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0057] To make the above-mentioned objectives, features, and advantages of this application more apparent and understandable, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0058] Example 1

[0059] The gearbox in a wind turbine is one of the core components of a wind power generation system, playing a crucial role in power transmission and conversion. However, due to long-term operation and harsh environmental conditions, gearboxes often face various faults, such as gear wear and bearing failure. Early detection and resolution of these faults are essential for ensuring the normal operation of wind turbines. Existing vibration analysis-based methods for monitoring wind turbine gearbox faults suffer from the following problems: First, the accuracy of fault diagnosis is not high. Existing vibration analysis methods may lack accuracy in diagnosing gearbox faults. Due to the complex operating environment and diverse fault modes of gearboxes, traditional vibration analysis methods may not be able to fully capture and accurately diagnose various faults. Second, the reliability of fault diagnosis is insufficient. Some existing methods may lack reliability in gearbox fault diagnosis. Environmental noise, signal interference, or specific fault modes may reduce the reliability of diagnostic results. Third, the efficiency of fault monitoring is limited. Some existing methods may have limitations in fault monitoring efficiency. They often require offline analysis or manual operation, resulting in long monitoring cycles or the inability to obtain monitoring results.

[0060] The technical solution of this application aims to solve the technical problems of low accuracy and low diagnostic efficiency in the existing technology for wind turbine gearbox fault monitoring, so as to improve the accuracy, reliability and efficiency of fault diagnosis. To address these problems, this application provides a more accurate and reliable fault diagnosis method for wind power equipment by optimizing vibration signal acquisition, signal processing and fault diagnosis algorithm models, effectively improving the reliability and operating efficiency of wind turbines.

[0061] This application provides a method for diagnosing faults in wind power generation equipment. In the first embodiment of this method, refer to... Figure 1 The fault diagnosis method for wind power generation equipment includes:

[0062] Step S10: Collect vibration signals from the wind power generation equipment;

[0063] Step S20: Extract fault feature parameters from the vibration signal, and classify the fault feature parameters using a preset target support vector machine model and a target clustering analysis model to obtain a first classification result and a second classification result.

[0064] Step S30: Based on the first classification result and the second classification result, determine the fault diagnosis result corresponding to the wind power generation equipment.

[0065] In this embodiment, it should be noted that the wind power generation equipment is the gearbox in a wind turbine. The gearbox is one of the core components of a wind turbine, mainly responsible for transmission and power conversion. Therefore, it is necessary to monitor the working status of the gearbox in real time to prevent the wind turbine from malfunctioning due to gearbox failure. Specifically, the vibration signal of the wind power generation equipment can be collected by various sensors installed in the gearbox, and the corresponding vibration signal is transmitted back according to the data interface of each sensor. The sensors are installed at key locations in the wind turbine gearbox, such as gear shafts and bearings. The sensors can be acceleration sensors or vibration sensors, used to collect the vibration signal of the gearbox in real time. When selecting the installation location and sensor parameters, it is necessary to consider the specific wind turbine gearbox structure and operating conditions, as well as the monitoring target and expected fault type, and conduct experiments and verifications to ensure the accuracy and reliability of the vibration sensors. If sensors are already present at the key locations, only the sensor data needs to be connected. In one feasible embodiment, Figure 2 This is a schematic diagram of the vibration signal collected by the sensor, where the horizontal axis represents time and the vertical axis represents amplitude.

[0066] In another feasible embodiment, since traditional vibration sensors require wiring connections, while wireless sensor networks can provide a more convenient way to acquire signals. Therefore, in this embodiment, the use of a wireless sensor network can reduce wiring costs and complexity, and enable simultaneous monitoring and data acquisition from multiple monitoring points.

[0067] Furthermore, in this embodiment, the fault feature parameters are used to characterize the features of the wind power generation equipment in various dimensions. Based on these features, a diagnosis can be made as to whether the wind power generation equipment has malfunctioned and the type of malfunction that has occurred. Moreover, during the extraction of fault feature parameters, the collected vibration signals need to be converted into digital signals. The sampling frequency needs to be determined based on specific circumstances, and it should be as high as possible to reflect the frequency component changes caused by a fault in the wind power generation equipment. Further, this embodiment uses support vector machines (SVMs) and clustering algorithms as classification models for fault diagnosis to classify the extracted fault feature parameters. This fully utilizes the discriminative and generalization capabilities of SVMs, which can handle complex nonlinear problems and have clear classification boundaries. Clustering algorithms, on the other hand, can identify potential patterns and cluster structures in the data and have lower computational complexity for large-scale data processing. Combining the advantages of these two algorithms to classify the fault feature parameters of wind power generation equipment can further improve the accuracy and robustness of fault diagnosis for wind power generation equipment. After classifying the fault characteristics using the above model, the generated first and second classification results may include multiple fault types. Further screening and determination of the fault types in the first and second classification results are needed to obtain the fault type that best matches the current fault situation of the wind power generation equipment.

[0068] As an example, steps S10 to S30 include: acquiring vibration signals of the wind power generation equipment at various locations through data interfaces corresponding to sensors installed at each location of the wind power generation equipment, wherein the sensors include acceleration sensors and vibration sensors; converting the acquired vibration signals into digital signals and preprocessing them to filter out noise; extracting time-domain features, frequency-domain features, statistical features, and wavelet transform features from the vibration signals to obtain fault feature parameters; inputting the fault feature parameters into a preset target support vector machine model and a target clustering analysis model, respectively, and processing the fault feature parameters through the target support vector machine model and the target clustering analysis model to predict the fault type of the wind power generation equipment, and outputting a first classification result and a second classification result, wherein the first classification result and the second classification result each include a different number of fault types. It should be noted that fault types may also be excluded, which indicates that the wind power generation equipment currently does not have a fault; selecting the fault type that best matches the fault situation of the wind power generation equipment from the first classification result and the second classification result.

[0069] In another feasible embodiment, feature data of other dimensions of wind power generation equipment can also be collected. For example, corresponding feature data can be collected through acoustic analysis, thermal imaging monitoring and current monitoring, and fault feature parameters can be extracted. Similarly, the target support vector machine model and target clustering analysis model in the embodiments of this application can be used for classification and diagnosis, which further enriches the dimensions of feature data used for fault diagnosis, thereby improving the accuracy of fault diagnosis of wind power generation equipment.

[0070] In this application's embodiments, the key technical approach is a fault diagnosis algorithm combining support vector machines and clustering algorithms. This algorithm can improve the accuracy and efficiency of fault diagnosis while reducing the number of samples and computational complexity. By clustering and classifying vibration signals, different types of faults can be better identified and distinguished, thereby providing more accurate fault diagnosis results.

[0071] Further, the step of determining the fault diagnosis result corresponding to the wind power generation equipment based on the first classification result and the second classification result may include:

[0072] Step S31: Extract the principal components corresponding to the first classification result and the second classification result respectively;

[0073] Step S32: Calculate the Mahalanobis distance between each principal component of the first classification result and each principal component of the second classification result;

[0074] Step S33: Based on the Mahalanobis distance between the principal components, select the target principal component and set the classification evaluation index corresponding to the target principal component as the fault diagnosis result.

[0075] This application provides a method for determining the final fault diagnosis result based on the first classification result output by a target support vector machine model and the second classification result output by a target clustering analysis model. The method primarily utilizes Mahalanobis distance to filter the principal components in each classification result. The principal component analysis used in this application is a quantitative taxonomy method, mainly constructed using principal component analysis from multivariate statistics. It is suitable for situations where multiple trait indicators exist, and both classification units and traits are represented by multidimensional vectors. Therefore, by using principal component analysis, a multidimensional complex entity is reduced to less than three dimensions, and a graph showing the classification relationship is depicted in a less-than-three-dimensional space, improving the visualization of the data. In this application, the number of principal components can be determined according to actual needs and data characteristics, retaining the principal components that retain most of the data information. Each principal component has a separate classification evaluation index, i.e., the fault type.

[0076] In one feasible embodiment, the mathematical expression for calculating the Mahalanobis distance is as follows:

[0077] ;

[0078] in, For Mahalanobis distance, and and These are the coordinates of the first and second classification results in the two-dimensional coordinate system.

[0079] As an example, steps S31 to S33 include: obtaining classification requirements, and selecting multiple principal components from the first classification result and the second classification result respectively based on the classification requirements and the number of fault types in the first classification result and the second classification result; calculating the Mahalanobis distance between each principal component in the first classification result and each principal component in the second classification result according to the Mahalanobis distance calculation formula; selecting the two principal components with the closest Mahalanobis distance as target principal components; and using the classification evaluation index corresponding to the target principal components as the fault diagnosis result corresponding to the wind power generation equipment, wherein each principal component has a corresponding classification evaluation index to characterize the corresponding fault type and severity.

[0080] In addition, after the step of collecting vibration signals from the wind power generation equipment, the method further includes:

[0081] Step A10: The vibration signal is subjected to noise reduction, filtering and data normalization processes in sequence to obtain an optimized vibration signal;

[0082] Step A20: Update the vibration signal using the optimized vibration signal to obtain the updated vibration signal.

[0083] This application provides a method for preprocessing acquired vibration signals to filter out noise, improve signal quality and reliability, and facilitate subsequent feature parameter extraction and fault diagnosis. The denoising process employs moving average and median filtering methods; a low-pass filter is selected to remove high-frequency noise; and data normalization maps the signal amplitude to a fixed range.

[0084] As an example, steps A10 to A20 include: denoising the acquired vibration using a moving average method and a median filtering method. Specifically, the vibration signal is averaged using a sliding window to smooth the signal and reduce the influence of high-frequency noise; each data point in the vibration signal is replaced with the median of the data within the window using a sliding window to eliminate the influence of sudden noise and outliers; high-frequency noise in the vibration signal is then filtered out using a Butterworth low-pass filter, while retaining low-frequency components; and the amplitude of the vibration signal is then mapped to a fixed range (e.g., [0-1]) to ensure the consistency of amplitude between different signals, so that the difference is not too large.

[0085] The above preprocessing steps can reduce noise interference in vibration signals and extract the effective information needed for fault characteristics. Specific parameters such as filter type, window size, and normalization method can be selected and adjusted according to the actual situation and signal characteristics to achieve the best preprocessing effect. (Refer to...) Figure 3 As shown in the figure, a schematic diagram of the vibration signal after preprocessing is presented. The horizontal axis represents time, and the vertical axis represents amplitude. Figure 3 (a) is the original signal; (b) is the original signal containing noise; (c) is the signal after median filtering; (d) is the signal after wavelet denoising; and (e) is the signal after bilateral filtering.

[0086] Furthermore, the step of extracting fault feature parameters from the vibration signal includes:

[0087] Step S21: Perform time-domain analysis on the vibration signal to obtain time-domain characteristic parameters, wherein the time-domain characteristic parameters include mean, variance, kurtosis and kurtosis;

[0088] Step S22: Perform frequency domain analysis on the vibration signal to obtain frequency domain characteristic parameters, wherein the frequency domain characteristic parameters include spectral peaks and energy distribution;

[0089] Step S23: Calculate the autocorrelation of the vibration signal and its correlation with other signals to obtain statistical characteristic parameters, wherein the statistical characteristic parameters include an autocorrelation function and a cross-correlation function;

[0090] Step S24: Extract the wavelet coefficients and wavelet coefficient energy from the vibration signal to obtain wavelet transform characteristic parameters.

[0091] In this embodiment, fault feature parameters are formed by extracting feature parameters from various dimensions of the vibration signal. These fault feature parameters include at least time-domain feature parameters, frequency-domain feature parameters, statistical feature parameters, and wavelet transform feature parameters. These parameters can reflect the vibration characteristics and fault information of the wind power generation equipment, facilitating the analysis of corresponding fault diagnosis results. Specifically, refer to... Figure 4 The extracted feature signals include at least Fourier components ( Figure 4 a) and vibration signals on the pulse assembly ( Figure 4 (b) where the vertical axis represents amplitude and the horizontal axis represents time.

[0092] As an example, steps S21 to S24 include: calculating the average value of the vibration signal to obtain the mean, which reflects the overall level of the vibration signal; calculating the variance of the vibration signal to reflect the dispersion of the vibration signal; extracting the kurtosis of the vibration signal to describe the sharpness of the vibration signal and reflect the peak adjustment of the vibration signal; extracting the kurtosis of the vibration signal to describe the flatness of the vibration signal and reflect the low-peak characteristics of the vibration signal; converting the vibration signal to the frequency domain through Fourier transform to obtain the vibration spectrum; and reading the peak frequency from the spectrum. The frequency and amplitude are used to represent the active frequency components of the vibration signal; the energy distribution of the vibration signal in different frequency bands is calculated to describe the frequency domain characteristics of the vibration signal; the correlation between the vibration signal and itself is calculated to obtain the autocorrelation function, which reflects the periodicity and repetitiveness of the vibration signal; the correlation between the vibration signal and other signals is calculated to obtain the cross-correlation function, which reflects the correlation and mutual influence between the vibration signal and other signals; the vibration signal is decomposed into wavelet coefficients at different scales and frequencies to obtain wavelet coefficients at different scales and frequencies; the energy of the wavelet coefficients of the vibration signal at different scales and frequencies is calculated to describe the time-frequency characteristics of the vibration signal.

[0093] After feature extraction in the above steps, fault feature parameters related to the fault are extracted from the vibration signal. Specifically, the fault feature parameters are shown in Table 1 below.

[0094]

[0095] Table 1

[0096] This application provides a method for fault diagnosis of wind power generation equipment. First, vibration signals from the wind power generation equipment are collected. Then, fault feature parameters are extracted from the vibration signals. These fault feature parameters are classified using a preset target support vector machine model and a target clustering analysis model, respectively, to obtain a first classification result and a second classification result. Based on the first and second classification results, the fault diagnosis result corresponding to the wind power generation equipment is determined. The technical solution of this application combines a support vector machine model and a clustering model to classify fault feature parameters in the vibration signals of the wind power generation equipment. The support vector machine model has a strong ability to characterize classification boundaries, while the clustering model can aggregate and classify fault feature parameters. Using both together can more accurately identify and distinguish different types of faults, improving the accuracy and reliability of fault diagnosis. Furthermore, the preset target support vector machine model and target clustering analysis model can automatically analyze and diagnose the collected vibration signals in real time, eliminating the need for manual operation and effectively improving the diagnostic efficiency of wind power generation equipment.

[0097] Example 2

[0098] Furthermore, based on the first embodiment of this application, in another embodiment of this application, the same or similar content as in the first embodiment can be referred to the above description, and will not be repeated hereafter. Based on this, a method is provided for constructing and training the target support vector machine model and the target clustering analysis model before classifying the fault feature parameters using the target support vector machine model and the target clustering analysis model. Specifically, before the step S20 in which the fault feature parameters are classified using the preset target support vector machine model and the target clustering analysis model to obtain the first classification result and the second classification result, refer to... Figure 5 The method further includes:

[0099] Step B10: Collect vibration signals from wind power generation equipment under different operating conditions to obtain multiple sets of training samples, and obtain the fault type label corresponding to each training sample.

[0100] Step B20: Input the training samples with fault type labels into the initial support vector machine model to obtain the corresponding third classification result;

[0101] Step B30: Input the training samples without fault type labels into the initial clustering analysis model to obtain the corresponding fourth classification results;

[0102] Step B40: Based on the third classification result and the fourth classification result, calculate the Mahalanobis distance between the principal components corresponding to the third classification result and the fourth classification result respectively;

[0103] Step B50: If the Mahalanobis distance is greater than a preset threshold, the model parameters of the initial support vector machine model and the initial clustering analysis model are optimized and adjusted, and the process returns to the execution step: input the training samples with fault type labels into the initial support vector machine model to obtain the corresponding third classification result;

[0104] Step B60: If the Mahalanobis distance is not greater than the preset threshold, then the current initial support vector machine model and the initial clustering analysis model are set to the target support vector machine model and the target clustering analysis model, respectively.

[0105] This application provides a method for training a support vector machine (SVM) model and a clustering model based on collected training samples. The method primarily includes using vibration signals collected from wind power generation equipment under different operating conditions as training samples. These different operating conditions include wind power generation equipment in normal operating condition and wind power generation equipment operating under different fault types. Further, after collecting training samples under various fault types and normal conditions, each training sample is manually labeled with a fault type for subsequent supervised training of the SVM model. Regarding the clustering model, since clustering is an unsupervised learning algorithm, it primarily groups similar data points into the same category. Clustering algorithms can identify potential patterns and cluster structures in the data to complete classification. Therefore, unsupervised learning training can be performed using training samples without fault type labels. In one feasible embodiment, the clustering algorithm of the support vector machine is combined with the clustering algorithm for classifying two-dimensional sample points. Figure 6 As shown, (a), (b), (c), and (d) represent progressively higher levels of classification detail. Figure 6 In (d), the sample points are divided into four categories, where the horizontal and vertical coordinates are two-dimensional vector values ​​after dimensionality reduction of the multidimensional vector. As the number of cluster categories increases, the classification of sample points increases from 1 category to 4 categories. The clustering algorithm combined with the support vector machine can always ensure that some samples in each cluster are at the category boundary, which has a strong boundary characterization ability, proving the effectiveness of the technical solution of the embodiment of this application.

[0106] In addition, during the training of the initial support vector machine model and the initial clustering analysis model, the Mahalanobis distance between the principal components extracted from the classification results of the two different fault diagnosis models is mainly used to measure whether to continue optimization. Mahalanobis distance is an effective method for calculating the similarity between two unknown sample sets. Unlike Euclidean distance, it takes into account the relationship between various characteristics and is scale-independent, that is, independent of the measurement scale. It can be used to measure whether the performance of the two different initial support vector machine models and the initial clustering analysis model is stable, that is, whether the similarity of the corresponding classification results is close. When the similarity of the classification results of the fault diagnosis models using the two classification algorithms corresponding to the input training samples reaches a certain threshold (the Mahalanobis distance is less than the preset threshold), the model training can be considered complete and can be used for fault diagnosis of wind power generation equipment. If the Mahalanobis distance of the principal components corresponding to the classification results of the two models does not reach the preset threshold, the model parameters of the initial support vector machine model and the initial clustering analysis model need to be further optimized and adjusted. Specifically, the model parameters of the support vector machine include: a kernel function (in this embodiment, a Gaussian radial basis function can be selected to map the data to a high-dimensional feature space); a penalty parameter (C) to control the degree of penalty for misclassified samples, which can be adjusted according to the output results after model training; and a slack variable parameter (ε) to allow a certain degree of error in the classification process, which can be adjusted according to the output results after model training. The model parameters of the clustering analysis model include: clustering algorithm selection (in this embodiment, a hierarchical clustering method can be selected to cluster the fault vibration signal); and a distance metric (in this embodiment, Euclidean distance can be selected to measure the similarity or difference between samples). In addition, the model parameters of the initial support vector machine model and the initial clustering analysis model also include other model parameters disclosed in the prior art. Specific model parameter optimization methods can apply various mature model parameter adjustment methods, which will not be elaborated here.

[0107] In one feasible embodiment, the performance evaluation of each clustering algorithm is referenced. Figure 7The graph shows the number of vibration signals on the x-axis and the adjusted Rand index (ARI) for evaluating the clustering performance of different algorithms. Specifically, it includes: 1. SPARCWave Group: wavelet-based sparse K-means clustering; 2. SPARCWave: sparse clustering based on scattering transform; 3. K-means: K-means clustering; 4. Sparse K-means: sparse K-means clustering; 5. Giacofci: wavelet method for high-dimensional data clustering; 6. Antoniadis: wavelet clustering; 7. adjusted Rand index: clustering performance evaluation. This allows for the selection of the optimal clustering algorithm.

[0108] As an example, steps B10 to B60 include: collecting corresponding vibration signals from wind power generation equipment with different fault types, converting each vibration signal into a digital signal form, and extracting corresponding fault feature parameters. The fault feature parameters are described in steps S21 to S24, and will not be repeated here. Each fault feature parameter is used as multiple training samples. Fault type labels corresponding to each manually input training sample are obtained, including no fault and various fault types. The model parameters of the initial support vector machine model are initialized based on the support vector machine algorithm to obtain the initial support vector machine model. Each training sample and its corresponding fault type label are input into the initial support vector machine model, and classification prediction is performed using the initial support vector machine model to output the corresponding third classification result. The model parameters of the initial clustering analysis model are initialized based on the hierarchical clustering algorithm to obtain the initial clustering analysis model. Each training sample is input into the initial clustering analysis model, and classification prediction is performed using the initial clustering analysis model to output the corresponding fourth classification result. The third fault feature parameters are extracted. The principal components corresponding to the classification results and the fourth classification results are respectively identified; the Mahalanobis distance between the principal components corresponding to the third classification results and the fourth classification results is calculated, wherein the calculation steps of the Mahalanobis distance can refer to steps S31 to S33, and will not be repeated here; a preset number of target Mahalanobis distances are selected, wherein the preset number is the number of fault types to be classified; it is determined whether the target Mahalanobis distance is greater than a preset threshold; if the target Mahalanobis distance is greater than the preset threshold, the model parameters of the initial support vector machine model and the initial cluster analysis model are optimized and adjusted, and the process returns to steps B20 to B40; if the target Mahalanobis distance is not greater than the preset threshold, the iterative optimization of the model parameters is stopped, and the current initial support vector machine model and the initial cluster analysis model are set as the target support vector machine model and the target cluster analysis model, and the classification index of the principal component corresponding to each target Mahalanobis distance is used as the target classification index, with each classification index corresponding to one principal component, to complete the model training, wherein the target classification index is used as the fault type classification in the fault diagnosis results.

[0109] In one feasible embodiment, combining the embodiments of this application with the embodiments of the previous application, the process of model training and model application can be as follows: Figure 8As shown, firstly, time-domain features of the vibration signal are extracted to obtain training samples. Then, support vector machines (SVMs) are used to classify and predict training samples with fault labels, and cluster analysis is used to classify training samples without fault labels. SVMs are used to strengthen boundaries, and cluster analysis is used for automatic classification, resulting in classification result 1 and classification result 2. Principal components are then extracted, and the Mahalanobis distance between principal components is calculated to determine if it is less than a threshold. If not, the parameters of the corresponding models for SVM and cluster analysis are adjusted. If yes, the classification index of the principal components in the classification results is selected as the target classification index. Specifically, principal components need to be extracted for different fault types, resulting in a vibration signal fault diagnosis model composed of an SVM model and a cluster analysis model. During model application, the collected vibration signals are cleaned and input into the vibration signal fault diagnosis model composed of the SVM model and the cluster analysis model. The corresponding fault diagnosis results and warning signals are output for staff confirmation. The accuracy of the diagnosis is determined based on the confirmation results. If inaccurate, the model parameters of the SVM model and the cluster analysis model are further updated to complete model optimization.

[0110] In another feasible embodiment, a Support Vector Machine (SVM) is used to classify training samples already labeled with fault types. The trained SVM model can classify the training samples into different categories. Features are extracted and preprocessed for samples without fault type labels (including newly collected vibration signal data). The SVM model is then used to predict the classification of these training samples. These predictions are considered pseudo-labels, indicating that the training sample belongs to a certain category classified by the SVM. The predictions are then applied to a clustering algorithm along with the training samples without fault type labels. The clustering algorithm groups the training samples into different clusters. In the clustering results, the pseudo-labels of the training samples are matched with the actual clusters, associating the pseudo-labels with the clusters. The effectiveness of the clustering method is ensured by minimizing the distance between the pseudo-labels and the cluster centers (i.e., each cluster should contain one type of fault sample; and each cluster should be independent and non-overlapping).

[0111] This application combines a supervised support vector machine (SVM) algorithm with an unsupervised clustering algorithm, making full use of both labeled and unlabeled training samples to improve fault classification performance for wind turbine equipment. The SVM algorithm utilizes accurate label information for training, building an accurate model to support subsequent fault diagnosis. Meanwhile, the clustering algorithm discovers potential patterns and structures in the data, helping to identify new fault types or variations. By combining these two algorithms, this application fully leverages the advantages of supervised and unsupervised learning to improve fault classification accuracy and enhance fault diagnosis capabilities.

[0112] This application provides a model training method before applying a vibration signal diagnostic model. The method primarily uses vibration signals from wind power generation equipment under various operating conditions and manually input fault labels to form training samples. Initialized support vector machine (SVM) and clustering analysis models are then trained. Specifically, the Mahalanobis distance between principal components in the classification results is used to determine if training is complete. The models are then iteratively optimized to obtain more stable and high-performance SVM and clustering analysis models. This allows for subsequent fault vibration analysis of the vibration signals, automating the fault diagnosis of wind power generation equipment. This improves the accuracy and real-time performance of fault vibration analysis, providing users with fault information immediately for timely repair and maintenance by staff, reducing downtime, and further enhancing the reliability and operating efficiency of wind turbines.

[0113] Example 3

[0114] Furthermore, based on the first and second embodiments of this application, in another embodiment of this application, the content that is the same as or similar to that in Embodiment 1 can be referred to the above description, and will not be repeated hereafter. Based on this, after determining the fault diagnosis result corresponding to the wind power generation equipment in step S30 based on the first classification result and the second classification result, refer to... Figure 9 The method further includes:

[0115] Step C10: Generate a fault warning signal based on the fault diagnosis results;

[0116] Step C20: Notify relevant personnel through the fault warning signal so that they can take the corresponding maintenance measures based on the fault diagnosis results.

[0117] In this application embodiment, it should be noted that this application embodiment provides a fault early warning method for wind power generation equipment. The fault early warning signal is transmitted through an alarm system and remote communication to ensure timely handling of faults and reduce the fault risk and downtime of wind turbines.

[0118] As an example, steps C10 to C20: Based on the fault diagnosis results, a corresponding fault warning signal is generated, wherein the fault warning signal includes the fault type and the fault severity; based on the fault warning signal, relevant personnel are notified to take corresponding maintenance and repair measures.

[0119] Furthermore, after the step of notifying relevant personnel via the fault warning signal, the method may further include:

[0120] Step C30: Obtain the maintenance results of the wind power generation equipment, wherein the maintenance results include at least the fault type, severity, and vibration characteristics;

[0121] Step C40: Based on the inspection results and the fault diagnosis results, optimize the model parameters of the target support vector machine model and the target clustering analysis model.

[0122] In this embodiment, maintenance results regarding the fault type, severity, and corresponding vibration characteristics can be obtained based on on-site confirmation and verification by staff. Based on this feedback, the fault diagnosis model is further updated and optimized. By comparing the on-site confirmation results with the fault diagnosis results of the model, the accuracy of the fault diagnosis model is evaluated, and adjustments and improvements are made as needed. This includes optimizing the feature extraction algorithm, fault classification model (target support vector machine model and target), or threshold settings to improve the model's accuracy and robustness.

[0123] As an example, steps C30 to C40 include: when the system detects a fault warning signal, relevant personnel will receive a corresponding alarm or notification; the personnel will conduct an on-site inspection of the wind power equipment to confirm whether the fault actually exists, record the inspection results, and compare them with the fault warning signal; if the inspection results match the fault diagnosis results, the personnel will feed back the confirmation results to the system to confirm the existence of the fault; if the inspection results do not match the fault diagnosis results, the inspection results will be fed back to the system, and the inspection results will be used to adjust and optimize the model parameters, feature extraction algorithms, and preset thresholds of the target support vector machine model and the target clustering analysis model in the fault diagnosis model, so as to obtain a fault diagnosis model with higher accuracy and more stable performance.

[0124] In one feasible embodiment, the fault confirmation and model update process in this application is as follows: Figure 10 As shown, after outputting the fault diagnosis result through the vibration signal fault model, a corresponding warning signal is generated. For example, the warning signal includes: 1. Fault type: gearbox oil level too low, 2. Severity: Level II, 3. Suggested solution: stop the machine and add xx lubricating oil. Then, the staff confirms on-site whether it is the fault description in the warning signal. If not, the feature data in the current fault situation is extracted to update the model. If so, it is processed according to the operating procedures.

[0125] Furthermore, in one feasible embodiment, combining the technical solutions of the embodiments of this application and the first embodiment, the overall concept of a method for fault diagnosis and early warning of wind power generation equipment is as follows: Figure 11As shown, sensors are first installed in the wind power equipment, and the sensor locations are selected through experiments and tests. Vibration signals under different operating conditions are collected, and noise reduction and filtering are performed to clarify the vibration signals. Further, time and frequency domain features are extracted using methods such as FFT (Fast Fourier Transform Algorithm) and WT (wavelet transform). Fault diagnosis is then performed using a fault diagnosis model composed of support vector machines and hierarchical clustering models, outputting the fault diagnosis results. Based on the fault diagnosis results, a fault warning is issued, including the fault type, severity, and suggested solutions. On-site confirmation by staff is then conducted to determine if the actual fault situation matches the fault diagnosis results. Finally, the model is updated based on the confirmation results.

[0126] This application embodiment enables accurate monitoring and diagnosis of wind power equipment faults. Vibration analysis technology is used to acquire equipment vibration information, fault characteristic parameters are extracted through signal processing, and the fault type and severity are determined using a fault diagnosis model. This facilitates early warning and timely repair of gearbox faults, improving the reliability, safety, and operating efficiency of wind turbines. Furthermore, the technical solution of this application embodiment enables early fault detection and preventative maintenance, avoiding serious consequences from power equipment failures. This helps reduce maintenance costs, unnecessary equipment wear and tear, and maintenance cycles, thereby improving the overall economic benefits of wind turbines.

[0127] Example 4

[0128] This application also provides a wind power equipment fault diagnosis device, which is applied to wind power equipment fault diagnosis equipment, as described above. Figure 12 The wind power equipment fault diagnosis device includes:

[0129] Signal acquisition module 101 is used to acquire vibration signals from wind power generation equipment;

[0130] Fault classification module 102 is used to extract fault feature parameters from the vibration signal, and classify the fault feature parameters by a preset target support vector machine model and a target clustering analysis model to obtain a first classification result and a second classification result.

[0131] The fault diagnosis module 103 is used to determine the fault diagnosis result corresponding to the wind power generation equipment based on the first classification result and the second classification result.

[0132] Optionally, the fault diagnosis module 103 is further configured to:

[0133] Extract the principal components corresponding to the first classification result and the second classification result, respectively;

[0134] Calculate the Mahalanobis distance between each principal component of the first classification result and each principal component of the second classification result;

[0135] Based on the Mahalanobis distance between the principal components, a target principal component is selected, and the classification evaluation index corresponding to the target principal component is set as the fault diagnosis result.

[0136] Optionally, the wind power equipment fault diagnosis device further includes a signal preprocessing module, which is used for:

[0137] The vibration signal is subjected to noise reduction, filtering, and data normalization processes in sequence to obtain an optimized vibration signal.

[0138] The vibration signal is updated by optimizing the vibration signal to obtain the updated vibration signal.

[0139] Optionally, the fault characteristic parameters include at least one of time-domain characteristic parameters, frequency-domain characteristic parameters, statistical characteristic parameters, and wavelet transform characteristic parameters, and the fault classification module 102 is further used for:

[0140] The vibration signal is subjected to time-domain analysis to obtain time-domain characteristic parameters, wherein the time-domain characteristic parameters include mean, variance, kurtosis and kurtosis;

[0141] The vibration signal is subjected to frequency domain analysis to obtain frequency domain characteristic parameters, wherein the frequency domain characteristic parameters include spectral peaks and energy distribution;

[0142] The autocorrelation of the vibration signal and its correlation with other signals are calculated to obtain statistical characteristic parameters, wherein the statistical characteristic parameters include an autocorrelation function and a cross-correlation function.

[0143] Wavelet coefficients and wavelet coefficient energy are extracted from the vibration signal to obtain wavelet transform characteristic parameters.

[0144] Optionally, the wind power equipment fault diagnosis device further includes a model training module, which is used for:

[0145] Vibration signals were collected from wind power generation equipment under different operating conditions to obtain multiple sets of training samples, and the fault type label corresponding to each training sample was obtained.

[0146] The training samples labeled with fault types are input into the initial support vector machine model to obtain the corresponding third classification result;

[0147] Input the training samples without fault type labels into the initial clustering analysis model to obtain the corresponding fourth classification results;

[0148] Based on the third classification result and the fourth classification result, calculate the Mahalanobis distance between the principal components corresponding to the third classification result and the fourth classification result respectively;

[0149] If the Mahalanobis distance is greater than a preset threshold, the model parameters of the initial support vector machine model and the initial clustering analysis model are optimized and adjusted, and the execution step is returned: input the training samples with fault type labels into the initial support vector machine model to obtain the corresponding third classification result;

[0150] If the Mahalanobis distance is not greater than the preset threshold, then the current initial support vector machine model and the initial clustering analysis model are set to the target support vector machine model and the target clustering analysis model, respectively.

[0151] Optionally, the wind power equipment fault diagnosis device further includes a fault early warning module, which is used for:

[0152] Based on the fault diagnosis results, a fault warning signal is generated;

[0153] The fault warning signal is used to notify relevant personnel so that they can take the corresponding maintenance measures based on the fault diagnosis results.

[0154] Optionally, the wind power equipment fault diagnosis device further includes a model optimization module, which is used for:

[0155] Obtain the maintenance results of the wind power generation equipment, wherein the maintenance results include at least the fault type, severity, and vibration characteristics;

[0156] Based on the inspection results and the fault diagnosis results, the model parameters of the target support vector machine model and the target clustering analysis model are optimized.

[0157] The wind power equipment fault diagnosis device provided in this application adopts the wind power equipment fault diagnosis method in the above embodiments, solving the technical problems of low accuracy and low efficiency in current wind power equipment fault diagnosis methods. Compared with the prior art, the beneficial effects of the wind power equipment fault diagnosis device provided in this application are the same as those of the wind power equipment fault diagnosis method provided in the above embodiments, and other technical features in this wind power equipment fault diagnosis device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0158] Example 5

[0159] This application provides an electronic device, which includes: at least one processor; and a memory communicatively linked to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the wind power equipment fault diagnosis method in Embodiment 1 above.

[0160] The following is for reference. Figure 13 The diagram illustrates a structural schematic of an electronic device suitable for implementing embodiments of the present disclosure. The electronic devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable media players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 13 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0161] like Figure 13 As shown, the electronic device may include a processing unit 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 (ROM) or a program loaded from a storage device 1003 into a random access memory 1004 (RAM). The RAM 1004 also stores various programs and data required for the operation of the electronic device. The processing unit 1004, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also linked to the bus 1005.

[0162] Typically, the following systems can be linked to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the electronic device to communicate wirelessly or wiredly with other devices to exchange data. Although electronic devices with various systems are shown in the figures, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems may be implemented alternatively.

[0163] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by a processing device, it performs the functions defined above in the methods of embodiments of this disclosure.

[0164] The electronic device provided in this application employs the wind power equipment fault diagnosis method described in the above embodiments, solving the technical problems of low accuracy and low efficiency in current wind power equipment fault diagnosis methods. Compared with the prior art, the beneficial effects of the electronic device provided in this application are the same as those of the wind power equipment fault diagnosis method provided in Embodiment 1 above, and other technical features of this electronic device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0165] It should be understood that various parts of this disclosure can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0166] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0167] Example 6

[0168] This embodiment provides a computer-readable storage medium having computer-readable program instructions stored thereon, which are used to execute the method for diagnosing wind power generation equipment faults in Embodiment 1 above.

[0169] The computer-readable storage medium provided in this application embodiment may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical links having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM, or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.

[0170] The aforementioned computer-readable storage medium may be included in an electronic device or may exist independently without being assembled into an electronic device.

[0171] The aforementioned computer-readable storage medium carries one or more programs. When the aforementioned one or more programs are executed by an electronic device, the electronic device causes the electronic device to: collect vibration signals from the wind power generation equipment; extract fault feature parameters from the vibration signals; classify the fault feature parameters using a preset target support vector machine model and a target clustering analysis model, respectively, to obtain a first classification result and a second classification result; and determine the fault diagnosis result corresponding to the wind power generation equipment based on the first classification result and the second classification result.

[0172] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be linked to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be linked to an external computer (e.g., via the Internet using an Internet service provider).

[0173] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0174] The modules described in the embodiments of this disclosure can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0175] The computer-readable storage medium provided in this application stores computer-readable program instructions for executing the above-described wind power equipment fault diagnosis method, thus solving the technical problem of wind power equipment fault diagnosis. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the wind power equipment fault diagnosis method provided in the above-described embodiments, and will not be repeated here.

[0176] Example 7

[0177] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the wind power generation equipment fault diagnosis method described above.

[0178] The computer program product provided in this application solves the technical problems of low accuracy and low efficiency in current wind power equipment fault diagnosis methods. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as the beneficial effects of the wind power equipment fault diagnosis methods provided in the above embodiments, and will not be repeated here.

[0179] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent scope of this application.

Claims

1. A method for diagnosing faults in wind power generation equipment, characterized in that, The fault diagnosis method for wind power generation equipment includes: Collect vibration signals from wind power generation equipment; Fault feature parameters are extracted from the vibration signal, and the fault feature parameters are classified by a preset target support vector machine model and a target clustering analysis model to obtain a first classification result and a second classification result. Based on the first classification result and the second classification result, the fault diagnosis result corresponding to the wind power generation equipment is determined; The step of determining the fault diagnosis result corresponding to the wind power generation equipment based on the first classification result and the second classification result includes: Extract the principal components corresponding to the first classification result and the second classification result, respectively; Calculate the Mahalanobis distance between each principal component of the first classification result and each principal component of the second classification result; Based on the Mahalanobis distance between the principal components, a target principal component is selected, and the classification evaluation index corresponding to the target principal component is set as the fault diagnosis result. Prior to the step of classifying the fault feature parameters using a preset target support vector machine model and a target clustering analysis model to obtain a first classification result and a second classification result, the method further includes: Vibration signals were collected from wind power generation equipment under different operating conditions to obtain multiple sets of training samples, and the fault type label corresponding to each training sample was obtained. The training samples labeled with fault types are input into the initial support vector machine model to obtain the corresponding third classification result; Input the training samples without fault type labels into the initial clustering analysis model to obtain the corresponding fourth classification results; Based on the third classification result and the fourth classification result, calculate the Mahalanobis distance between the principal components corresponding to the third classification result and the fourth classification result respectively; If the Mahalanobis distance is greater than a preset threshold, the model parameters of the initial support vector machine model and the initial clustering analysis model are optimized and adjusted, and the execution step is returned: input the training samples with fault type labels into the initial support vector machine model to obtain the corresponding third classification result; If the Mahalanobis distance is not greater than the preset threshold, then the current initial support vector machine model and the initial clustering analysis model are set to the target support vector machine model and the target clustering analysis model, respectively.

2. The fault diagnosis method for wind power generation equipment as described in claim 1, characterized in that, After the step of collecting vibration signals from the wind power generation equipment, the method further includes: The vibration signal is subjected to noise reduction, filtering, and data normalization processes in sequence to obtain an optimized vibration signal. The vibration signal is updated by optimizing the vibration signal to obtain the updated vibration signal.

3. The method for fault diagnosis of wind power generation equipment as described in claim 1, characterized in that, The fault feature parameters include at least one of time-domain feature parameters, frequency-domain feature parameters, statistical feature parameters, and wavelet transform feature parameters. The step of extracting fault feature parameters from the vibration signal includes: The vibration signal is subjected to time-domain analysis to obtain time-domain characteristic parameters, wherein the time-domain characteristic parameters include mean, variance, kurtosis and kurtosis; The vibration signal is subjected to frequency domain analysis to obtain frequency domain characteristic parameters, wherein the frequency domain characteristic parameters include spectral peaks and energy distribution; The autocorrelation of the vibration signal and its correlation with other signals are calculated to obtain statistical characteristic parameters, wherein the statistical characteristic parameters include an autocorrelation function and a cross-correlation function. Wavelet coefficients and wavelet coefficient energy are extracted from the vibration signal to obtain wavelet transform characteristic parameters.

4. The method for fault diagnosis of wind power generation equipment as described in any one of claims 1-3, characterized in that, After the step of determining the fault diagnosis result corresponding to the wind power generation equipment based on the first classification result and the second classification result, the method further includes: Based on the fault diagnosis results, a fault warning signal is generated; The fault warning signal is used to notify relevant personnel so that they can take the corresponding maintenance measures based on the fault diagnosis results.

5. The fault diagnosis method for wind power generation equipment as described in claim 4, characterized in that, After the step of notifying relevant personnel via the fault warning signal, the method further includes: Obtain the maintenance results of the wind power generation equipment, wherein the maintenance results include at least the fault type, severity, and vibration characteristics; Based on the inspection results and the fault diagnosis results, the model parameters of the target support vector machine model and the target clustering analysis model are optimized.

6. A fault diagnosis device for wind power generation equipment, characterized in that, The wind power equipment fault diagnosis device includes: The signal acquisition module is used to collect vibration signals from wind power generation equipment; The fault classification module is used to extract fault feature parameters from the vibration signal, and classify the fault feature parameters by a preset target support vector machine model and a target clustering analysis model to obtain a first classification result and a second classification result. The fault diagnosis module is used to determine the fault diagnosis result corresponding to the wind power generation equipment based on the first classification result and the second classification result; The fault diagnosis module is further configured to: extract the principal components corresponding to the first classification result and the second classification result respectively; calculate the Mahalanobis distance between each principal component of the first classification result and each principal component of the second classification result; select a target principal component based on the Mahalanobis distance between each principal component, and set the classification evaluation index corresponding to the target principal component as the fault diagnosis result; The fault classification module is further configured to: collect vibration signals from wind power generation equipment under different operating conditions to obtain multiple sets of training samples, and obtain fault type labels corresponding to each training sample; input training samples with fault type labels into an initial support vector machine model to obtain corresponding third classification results; input training samples without fault type labels into an initial clustering analysis model to obtain corresponding fourth classification results; calculate the Mahalanobis distance between the principal components corresponding to the third and fourth classification results based on the third and fourth classification results; if the Mahalanobis distance is greater than a preset threshold, optimize and adjust the model parameters of the initial support vector machine model and the initial clustering analysis model, and return to the execution step: input training samples with fault type labels into the initial support vector machine model to obtain corresponding third classification results; if the Mahalanobis distance is not greater than the preset threshold, set the current initial support vector machine model and the initial clustering analysis model as the target support vector machine model and the target clustering analysis model, respectively.

7. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively linked to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the steps of the wind power equipment fault diagnosis method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program for implementing a method for diagnosing faults in wind power generation equipment, which is executed by a processor to implement the steps of the method for diagnosing faults in wind power generation equipment as described in any one of claims 1 to 5.

Images