Wind power main bearing fault diagnosis method and device and electronic equipment

Abstract

The invention relates to the technical field of wind power equipment fault monitoring, in particular to a wind power main bearing fault diagnosis method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the envelope spectrum analysis of vibration data, and obtaining fault impact feature data; obtaining an SDP image difference maximum parameter based on the fault impact feature data; constructing a training data set based on the image difference maximum parameter; and constructing a wind power main bearing fault diagnosis model based on the convolutional neural network model, and diagnosing a to-be-diagnosed bearing by using the trained wind power main bearing fault diagnosis model to obtain a diagnosis result. The frequency domain information is extracted from the original signals of the multiple sensors of the wind power main bearing and fused into the time frequency SDP image, full fusion of multi-source and multi-dimensional information is achieved, accurate judgment of faults is facilitated, the over-fitting risk of a deep learning model in the training process is fully restrained, and the fault monitoring efficiency and precision of wind power equipment are improved.

Description

Technical Field

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

[0002] As a key component of wind turbine transmission systems, the operating condition of wind turbine main bearings directly affects the performance and lifespan of wind turbine equipment. Therefore, research on component-level condition monitoring and fault diagnosis of wind turbine main bearings is of great significance. Due to the complex on-site environment, their vibration signals often exhibit strong non-stationary characteristics. Traditional signal processing methods either use a single data source or lack comprehensive feature extraction, leading to inaccurate final diagnosis.

[0003] Traditional bearing fault diagnosis methods based on one-dimensional vibration signals mainly include time-domain, frequency-domain, and time-frequency-domain methods. Since relying solely on time-frequency domain signals cannot directly reflect the fault characteristics of the unit, many researchers have attempted to transform one-dimensional vibration signals into images from an image processing perspective, then extract and classify fault features through manual judgment or with the help of intelligent models such as neural networks. However, traditional signal imagery + intelligent model approaches often suffer from drawbacks such as limited sensor placement and data utilization, insufficient comprehensiveness of image features, single-dimensional image features, and large, computationally intensive intelligent model algorithms that heavily rely on hardware computing power.

[0004] In summary, designing an efficient and accurate method for diagnosing wind turbine main bearing faults is a problem that urgently needs to be solved. Summary of the Invention

[0005] This application aims to at least partially address one of the technical problems in the related art.

[0006] Therefore, the first objective of this application is to propose a fault diagnosis method for wind turbine main bearings to address the limitations of existing technologies, the reliance on single data sources, and the lack of comprehensive graphical features.

[0007] The second objective of this application is to provide an apparatus.

[0008] The third objective of this application is to propose an electronic device.

[0009] The fourth objective of this application is to provide a computer-readable storage medium.

[0010] To achieve the above objectives, the first aspect of this application proposes a method for diagnosing faults in wind turbine main bearings, comprising:

[0011] Obtain vibration data of wind turbine main bearings;

[0012] Envelope spectrum analysis was performed on the vibration data to obtain fault impact characteristic data;

[0013] The parameter with the maximum difference in SDP images is obtained based on the fault impact feature data.

[0014] A training dataset is constructed based on the maximum image difference parameter;

[0015] A wind turbine main bearing fault diagnosis model is constructed based on a convolutional neural network model. The model is trained using the training dataset to obtain the completed wind turbine main bearing fault diagnosis model.

[0016] The wind turbine main bearing fault diagnosis model, which has been trained, is used to diagnose the bearing to be diagnosed, and the diagnosis results are obtained.

[0017] Preferably, acquiring the vibration data of the wind turbine main bearing includes:

[0018] When acquiring multi-sensor vibration data of the main bearing under different fault types, vibration data of the main bearing in the X and Y directions are acquired.

[0019] Preferably, the step of performing envelope spectrum analysis on the vibration data to obtain fault impact characteristic data includes:

[0020] The original vibration data collected in the X and Y directions of the main bearing are resampled using a sliding window to generate a time-domain signal with a sample length of 1024. Envelope spectrum analysis is then performed to obtain the frequency domain data of the fault impact characteristics.

[0021] Preferably, the envelope spectrum analysis includes:

[0022] For a given time-domain signal, the analytic signal is obtained by using the Hilbert transform.

[0023] Performing a Fourier transform on the analytic signal yields the amplitude-modulated signal, which is the envelope signal.

[0024] Preferably, the step of obtaining the maximum difference parameter of the SDP image based on the fault impact feature data includes:

[0025] Based on the fault impact characteristic data, one data point is selected for each fault type, and the time domain signal, frequency domain signal, time domain signal, and frequency domain signal in the X direction, respectively, are used as outputs to calculate the CW-SSIM index under different parameters. Based on the CW-SSIM index, the parameter with the maximum difference in the SDP image is obtained.

[0026] Preferably, the SDP method calculation formula is as follows:

[0027]

[0028]

[0029]

[0030] Where, x max and x min Let r(i) be the maximum and minimum amplitude values ​​of the original signal waveform, r(i) be the polar coordinate radius of the point, θ(i) and φ(i) be the angles of rotation of the point about the mirror symmetry plane in the counterclockwise and clockwise directions, respectively, and θ be the rotation angle of the specified mirror symmetry plane. This is the gain coefficient.

[0031] Preferably, the method for constructing a wind turbine main bearing fault diagnosis model based on a convolutional neural network model includes:

[0032] Construct a CNN model with a backbone consisting of three convolutional layers, three pooling layers, and a global average pooling layer. Add a random deactivation layer before the global average pooling layer. The output layer uses Softmax, and the number of output units is set according to different classification tasks. The activation function is ReLU.

[0033] To achieve the above objectives, a second aspect of this application provides a wind turbine main bearing fault diagnosis device, comprising:

[0034] The data acquisition module acquires vibration data of the wind turbine main bearing.

[0035] The envelope spectrum analysis module performs envelope spectrum analysis on the vibration data to obtain fault impact characteristic data;

[0036] The optimal parameter acquisition module obtains the parameter with the maximum difference in the SDP image based on the fault impact feature data;

[0037] The training dataset acquisition module constructs a training dataset based on the maximum image difference parameter.

[0038] The training module constructs a wind turbine main bearing fault diagnosis model based on a convolutional neural network model, and trains it using the training dataset to obtain the trained wind turbine main bearing fault diagnosis model.

[0039] The diagnostic module uses the trained wind turbine main bearing fault diagnosis model to diagnose the bearing to be diagnosed and obtain the diagnostic results.

[0040] To achieve the above objectives, a third aspect of this application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0041] The memory stores computer-executed instructions;

[0042] The processor executes computer execution instructions stored in the memory to implement the method described in any of the preceding descriptions.

[0043] To achieve the above objectives, a fourth aspect of this application provides a computer-readable storage medium, comprising computer-executable instructions stored therein, which, when executed by a processor, are used to implement the method described in any of the above embodiments.

[0044] This application provides a method for fault diagnosis of wind turbine main bearings. By extracting frequency domain information from the raw signals of multiple sensors on the wind turbine main bearing and fusing them into a time-frequency SDP image, it achieves full fusion of multi-source, multi-dimensional information. This allows a single SDP image to simultaneously contain raw vibration information and frequency domain information from multiple sensors, reflecting rich characteristics of the equipment's operating status and highlighting subtle differences between different operational faults. This provides more complete image features for CNN intelligent models, aiding in accurate fault identification and effectively suppressing the risk of overfitting during deep learning model training. The data originates from both X- and Y-axis measurement points on the wind turbine main bearing and includes time-domain information (raw signal) and frequency-domain information (envelope spectrum analysis), resulting in more comprehensive and specific features.

[0045] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0046] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0047] Figure 1 A flowchart of a first specific embodiment of a wind turbine main bearing fault diagnosis method provided by the present invention;

[0048] Figure 2 Flowchart for fault diagnosis of wind turbine main bearing;

[0049] Figure 3 This is a schematic diagram of the sensor location;

[0050] Figure 4 The original signal envelope spectrum in the X direction;

[0051] Figure 5 The original signal envelope spectrum in the Y direction;

[0052] Figure 6 This is a schematic diagram of the SDP principle in symmetric point mode.

[0053] Figure 7 This is an SDP time-frequency fusion diagram;

[0054] Figure 8 This is a structural block diagram of a wind turbine main bearing fault diagnosis device provided in an embodiment of the present invention. Detailed Implementation

[0055] The core of this invention is to provide a method, device, electronic device and storage medium for diagnosing wind turbine main bearing faults. By extracting frequency domain information from the raw signals of multiple sensors of the wind turbine main bearing and fusing them into a time-frequency SDP image, the invention achieves full fusion of multi-source and multi-dimensional information, thereby improving the efficiency and accuracy of bearing fault diagnosis.

[0056] To enable those skilled in the art to better understand the present invention, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] Please refer to Figure 1 , Figure 1 The flowchart illustrates a first specific embodiment of a wind turbine main bearing fault diagnosis method provided by the present invention; the specific operation steps are as follows:

[0058] Step S101: Obtain vibration data of the wind turbine main bearing;

[0059] When acquiring multi-sensor vibration data of the main bearing under different fault types, vibration data of the main bearing in the X and Y directions are acquired.

[0060] Step S102: Perform envelope spectrum analysis on the vibration data to obtain fault impact characteristic data;

[0061] The original vibration data collected in the X and Y directions of the main bearing are resampled using a sliding window to generate a time-domain signal with a sample length of 1024. Envelope spectrum analysis is then performed to obtain the frequency domain data of the fault impact characteristics.

[0062] The envelope spectrum analysis includes:

[0063] For a given time-domain signal, the analytic signal is obtained by using the Hilbert transform.

[0064] Performing a Fourier transform on the analytic signal yields the amplitude-modulated signal, which is the envelope signal.

[0065] Step S103: Obtain the parameter with the maximum difference in SDP images based on the fault impact feature data;

[0066] Based on the fault impact characteristic data, one data point is selected for each fault type, and the time domain signal, frequency domain signal, time domain signal, and frequency domain signal in the X direction, respectively, are used as outputs to calculate the CW-SSIM index under different parameters. Based on the CW-SSIM index, the parameter with the maximum difference in the SDP image is obtained.

[0067] The calculation formula for the SDP method is as follows:

[0068]

[0069]

[0070]

[0071] Where, x max and x min Let r(i) be the maximum and minimum amplitude values ​​of the original signal waveform, r(i) be the polar coordinate radius of the point, θ(i) and φ(i) be the angles of rotation of the point about the mirror symmetry plane in the counterclockwise and clockwise directions, respectively, and θ be the rotation angle of the specified mirror symmetry plane. This is the gain coefficient.

[0072] Step S104: Construct a training dataset based on the maximum image difference parameter;

[0073] Step S105: Construct a wind turbine main bearing fault diagnosis model based on a convolutional neural network model, and train it using the training dataset to obtain the trained wind turbine main bearing fault diagnosis model.

[0074] The wind turbine main bearing fault diagnosis model constructed based on the convolutional neural network model includes:

[0075] Construct a CNN model with a backbone consisting of three convolutional layers, three pooling layers, and a global average pooling layer. Add a random deactivation layer before the global average pooling layer. The output layer uses Softmax, and the number of output units is set according to different classification tasks. The activation function is ReLU.

[0076] Step S106: Use the trained wind turbine main bearing fault diagnosis model to diagnose the bearing to be diagnosed and obtain the diagnosis results.

[0077] This embodiment provides a method for diagnosing wind turbine main bearing faults. It uses the CW-SSIM index as a standard to optimize the parameters of the time-frequency fused SDP image. Under optimal parameters, the differences between time-frequency fused SDP images of different faults are maximized, which is more conducive to CNN recognition. By extracting frequency domain information from the raw signals of multiple sensors of the wind turbine main bearing and fusing them into a time-frequency SDP image, the method achieves full fusion of multi-source and multi-dimensional information. This allows a single SDP image to simultaneously possess the raw vibration information and frequency domain information from multiple sensors, reflecting rich equipment operating status characteristics and highlighting subtle differences between different operating faults. This provides more complete image features for the CNN intelligent model, aiding in accurate fault identification and effectively suppressing the risk of overfitting during the training process of the deep learning model.

[0078] Based on the above embodiments, this embodiment describes a method for diagnosing wind turbine main bearing faults, such as... Figure 2 As shown, the details are as follows:

[0079] Acquire multi-sensor vibration data under different fault types of wind turbine main bearings, such as Figure 3 As shown, when acquiring multi-sensor vibration data under different fault types of wind turbine main bearings, the sensor test points and installation methods are as follows: main bearing X direction, main bearing Y direction, low-frequency accelerometer sensor, fixed installation with openings.

[0080] Envelope spectrum analysis was performed simultaneously on the raw vibration data from the X and Y sensors.

[0081] The raw vibration data collected by the X and Y sensors are resampled using a sliding window to generate a time-domain signal with a sample length of 1024. Then, envelope spectrum analysis is performed to obtain frequency-domain data containing fault impact characteristics. At this step, the time-domain and frequency-domain data of the X and Y sensors have been obtained simultaneously.

[0082] For a given time-domain signal x(t), its Hilbert transform is:

[0083]

[0084] After the signal x(t) undergoes the Hilbert transform, the analytic signal of the signal is defined as follows:

[0085]

[0086] Performing a Fourier transform on the analytic signal yields the amplitude modulation signal:

[0087]

[0088] The analytic signal obtained by the Hilbert transform is a complex number. Its real part is the real signal itself, and its imaginary part is the Hilbert transform of the signal. The amplitude of the analytic signal is the envelope of the real signal, such as... Figure 4 , Figure 5 As shown, after envelope analysis, the original vibration signal yields an envelope signal, which can be used to extract signal features using time-domain and frequency-domain methods.

[0089] For each fault type, a data point (containing time and frequency domain data from both X and Y sensors) is randomly selected. Using the X-direction time-domain signal, X-direction frequency-domain signal, Y-direction time-domain signal, and Y-direction frequency-domain signal as inputs, a pair of petals is formed in each of the 1st, 2nd, 3rd, and 4th quadrants of the polar coordinate system. Parameters l = 1-10 with a step size of 1 and ξ = 10-45 with a step size of 5 are set, and the CW-SSIM index is calculated under different parameters. This process is repeated 5 times, and the CW-SSIM values ​​from the 5 calculations are summed. The parameters l and ξ that maximize the difference in SDP images between different faults when the CW-SSIM index is minimized are considered the optimal parameters. Figure 6 As shown, the principle of the SDP method is as follows:

[0090]

[0091]

[0092]

[0093] Where, x max and x min Let r(i) be the maximum and minimum amplitude values ​​of the original signal waveform, r(i) be the polar coordinate radius of the point, θ(i) and φ(i) be the angles of rotation of the point about the mirror symmetry plane in the counterclockwise and clockwise directions, respectively, θ be the rotation angle of the specified mirror symmetry plane (θ = 360m / n, m = 1, 2, ..., n, n is the number of mirror symmetry planes), and ζ be the gain coefficient.

[0094] In the complex wavelet domain, let c x ={c x,i |i=1,…,N} and c y ={c y,i ∣i=1,…,N} are two sets of coefficients extracted from the same spatial location in the same wavelet subband of the two compared images, where N is the number of pixels in the image. The CW-SSIM exponent can be obtained and rearranged as follows:

[0095]

[0096] Among them, C *Let C be the complex conjugate, and K be a very small positive constant that can be used to ensure computational robustness when the signal-to-noise ratio is low. The above equation can be analogized to SSIM, viewed as a product of two parts, the first of which depends on the modulus of the complex wavelet decomposition coefficients, C. x,i With C y,i When they are completely equal, the value is 1. The other part is related to the phase difference of the decomposition coefficients, C. x,i With C y,i When the phase difference is exactly the same, the value is 1. When the image x and image y change, the greater the difference between the images, the smaller the CW-SSIM value, reaching a peak of 1 when the two images are exactly the same.

[0097] Obtain SDP images and construct datasets that fuse time-frequency information from multiple sensors.

[0098] like Figure 7 As shown, under optimal SDP parameters, the X-direction time-domain signal, X-direction frequency-domain signal, Y-direction time-domain signal, and Y-direction frequency-domain signal are used as inputs, forming a pair of petals in each of the polar coordinate quadrants 1, 2, 3, and 4. The time and frequency information from the two sensors are fused to obtain an SDP image with fused multi-sensor time and frequency information. The dataset contains 300 data points each for normal operation, inner race fault, outer race fault, and rolling element fault, and is randomly divided into training and test sets in a 7:3 ratio.

[0099] To reduce training costs, a small CNN model was constructed: It takes a 256x256 image as input, and the backbone consists of three convolutional layers, three pooling layers, and a global average pooling layer (GAP). A dropout layer is added before the GAP layer to suppress overfitting. The output layer uses Softmax, and the number of output units is adjusted according to different classification tasks. The activation function is ReLU. The specific structural parameters of the CNN are shown in Table 1.

[0100] Table 1. Network Structure of CNN

[0101]

[0102] For model training and testing, the four states of normal, inner ring fault, outer ring fault, and rolling element fault are sequentially encoded as 0, 1, 2, and 3 before inputting into the network. In the TensorFlow environment, the network learning rate is set to 0.0001, the batch size is 32, the iteration is 50 rounds, and the image size is resized to 256x256. After training, the classification results are obtained, and the h5 weight file is saved, thus completing the model training and testing process.

[0103] For model fault diagnosis, in the TensorFlow environment, a pre-trained h5 file is loaded, and newly collected data during the process is input to obtain an SDP image that integrates time and frequency information. This image is then used as the model input to obtain the fault diagnosis results.

[0104] This invention provides a method for diagnosing wind turbine main bearing faults. By extracting frequency domain information from the raw signals of multiple sensors on the wind turbine main bearing and fusing them into a time-frequency SDP image, it achieves full fusion of multi-source and multi-dimensional information. This allows a single SDP image to simultaneously contain the raw vibration information and frequency domain information from multiple sensors, reflecting rich characteristics of the equipment's operating status and highlighting subtle differences in features between different operating faults. This provides more complete image features for CNN intelligent models, aiding in accurate fault identification and effectively suppressing the risk of overfitting during the training process of deep learning models. The CW-SSIM index is used as a standard to optimize the parameters of the SDP. Under the optimal parameters, the differences between the time-frequency fused SDP images of different faults are maximized, further facilitating CNN recognition.

[0105] Please refer to Figure 8 , Figure 8 This invention provides a structural block diagram of a wind turbine main bearing fault diagnosis device; the specific device may include:

[0106] Data acquisition module 100 acquires vibration data of the wind turbine main bearing;

[0107] The envelope spectrum analysis module 200 performs envelope spectrum analysis on the vibration data to obtain fault impact characteristic data;

[0108] The optimal parameter acquisition module 300 acquires the parameter with the maximum difference in the SDP image based on the fault impact feature data.

[0109] The training dataset acquisition module 400 constructs a training dataset based on the maximum image difference parameter.

[0110] Training module 500 constructs a wind turbine main bearing fault diagnosis model based on a convolutional neural network model, and trains it using the training dataset to obtain the trained wind turbine main bearing fault diagnosis model.

[0111] The diagnostic module 600 uses the trained wind turbine main bearing fault diagnosis model to diagnose the bearing to be diagnosed and obtain the diagnostic results.

[0112] This embodiment of a wind turbine main bearing fault diagnosis device is used to implement the aforementioned wind turbine main bearing fault diagnosis method. Therefore, the specific implementation of the wind turbine main bearing fault diagnosis device can be found in the embodiment section of the wind turbine main bearing fault diagnosis method above. For example, the data acquisition module 100, envelope spectrum analysis module 200, optimal parameter acquisition module 300, training dataset acquisition module 400, training module 500, and diagnosis module 600 are respectively used to implement steps S101, S102, S103, S104, S105, and S106 in the aforementioned wind turbine main bearing fault diagnosis method. Therefore, its specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0113] To implement the above embodiments, this application also proposes an electronic device, including: a processor and a memory communicatively connected to the processor; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to implement the method provided in the foregoing embodiments.

[0114] To implement the above embodiments, this application also proposes a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the methods provided in the foregoing embodiments.

[0115] To implement the above embodiments, this application also proposes a computer program product, including a computer program that, when executed by a processor, implements the methods provided in the foregoing embodiments.

[0116] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in this application comply with relevant laws and regulations and do not violate public order and good morals.

[0117] It should be noted that personal information collected from users should be used for legitimate and reasonable purposes and should not be shared or sold outside of these legitimate uses. Furthermore, such collection / sharing should only be conducted after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign an agreement / authorization that includes authorization of relevant user information before the user uses the function. In addition, any necessary steps must be taken to protect and safeguard access to such personal information data and ensure that others with access to personal information data comply with their privacy policies and procedures.

[0118] This application is intended to provide an implementation scheme for users to selectively prevent the use or access to their personal information data. Specifically, this disclosure is intended to provide hardware and / or software to prevent or block access to such personal information data. Once personal information data is no longer needed, risks can be minimized by restricting data collection and deleting data. Furthermore, where applicable, such personal information is de-identified to protect user privacy.

[0119] In the foregoing descriptions of the embodiments, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0120] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0121] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0122] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0123] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0124] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0125] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0126] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A method for diagnosing faults in wind turbine main bearings, characterized in that, include: Obtain vibration data of wind turbine main bearings; Envelope spectrum analysis was performed on the vibration data to obtain fault impact characteristic data; The parameter with the maximum difference in SDP images is obtained based on the fault impact feature data. A training dataset is constructed based on the maximum image difference parameter; A wind turbine main bearing fault diagnosis model is constructed based on a convolutional neural network model. The model is trained using the training dataset to obtain the completed wind turbine main bearing fault diagnosis model. The wind turbine main bearing fault diagnosis model, which has been trained, is used to diagnose the bearing to be diagnosed, and the diagnosis results are obtained.

2. The wind turbine main bearing fault diagnosis method according to claim 1, characterized in that, The acquisition of wind turbine main bearing vibration data includes: When acquiring multi-sensor vibration data of the main bearing under different fault types, vibration data of the main bearing in the X and Y directions are acquired.

3. The wind turbine main bearing fault diagnosis method according to claim 2, characterized in that, The process of performing envelope spectrum analysis on the vibration data to obtain fault impact characteristic data includes: The original vibration data collected in the X and Y directions of the main bearing are resampled using a sliding window to generate a time-domain signal with a sample length of 1024. Envelope spectrum analysis is then performed to obtain the frequency domain data of the fault impact characteristics.

4. The wind turbine main bearing fault diagnosis method according to claim 3, characterized in that, The envelope spectrum analysis includes: For a given time-domain signal, the analytic signal is obtained by using the Hilbert transform. Performing a Fourier transform on the analytic signal yields the amplitude-modulated signal, which is the envelope signal.

5. The wind turbine main bearing fault diagnosis method according to claim 1, characterized in that, The parameter for obtaining the maximum difference in SDP images based on the fault impact feature data includes: Based on the fault impact characteristic data, one data point is selected for each fault type, and the time domain signal, frequency domain signal, time domain signal, and frequency domain signal in the X direction, respectively, are used as outputs to calculate the CW-SSIM index under different parameters. Based on the CW-SSIM index, the parameter with the maximum difference in the SDP image is obtained.

6. The wind turbine main bearing fault diagnosis method according to claim 5, characterized in that, The calculation formula for the SDP method is as follows: Where, x max and x min Let r(i) be the maximum and minimum amplitude values ​​of the original signal waveform, r(i) be the polar coordinate radius of the point, θ(i) and φ(i) be the angles of rotation of the point about the mirror symmetry plane in the counterclockwise and clockwise directions, respectively, θ be the rotation angle of the specified mirror symmetry plane, and ζ be the gain coefficient.

7. The wind turbine main bearing fault diagnosis method according to claim 1, characterized in that, The wind turbine main bearing fault diagnosis model constructed based on the convolutional neural network model includes: Construct a CNN model with a backbone consisting of three convolutional layers, three pooling layers, and a global average pooling layer. Add a random deactivation layer before the global average pooling layer. The output layer uses Softmax, and the number of output units is set according to different classification tasks. The activation function is ReLU.

8. A fault diagnosis device for wind turbine main bearings, characterized in that, include: The data acquisition module acquires vibration data of the wind turbine main bearing. The envelope spectrum analysis module performs envelope spectrum analysis on the vibration data to obtain fault impact characteristic data; The optimal parameter acquisition module obtains the parameter with the maximum difference in the SDP image based on the fault impact feature data; The training dataset acquisition module constructs a training dataset based on the maximum image difference parameter. The training module constructs a wind turbine main bearing fault diagnosis model based on a convolutional neural network model, and trains it using the training dataset to obtain the trained wind turbine main bearing fault diagnosis model. The diagnostic module uses the trained wind turbine main bearing fault diagnosis model to diagnose the bearing to be diagnosed and obtain the diagnostic results.

9. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-7.

Images