Wind turbine generator gearbox bearing fault diagnosis method and device, electronic equipment and computer readable medium

Abstract

The invention discloses a wind turbine generator gearbox bearing fault diagnosis method and device, electronic equipment and a computer readable medium, and the method comprises the steps: carrying out the preprocessing of a detection signal of a wind turbine generator gearbox bearing; performing fault feature extraction on the preprocessed detection signal; and inputting the fault feature extraction into a fault diagnosis model to enable the fault diagnosis model to output a fault diagnosis result. The wind turbine generator gearbox bearing fault diagnosis method and device, the electronic equipment and the computer readable medium have the beneficial effects that the fault diagnosis precision can be improved based on signal preprocessing and the neural network model.

Description

Technical Field

[0001] The present invention relates to the field of wind power control technology, and in particular to a method, device, electronic equipment and computer-readable medium for diagnosing a bearing fault in a gearbox of a wind turbine generator set. Background Art

[0002] In actual engineering applications, wind turbine gearboxes have a high failure rate and long downtime. Once a failure occurs, it seriously affects the quality of power generation, increases maintenance costs and causes huge losses. As one of the key components of wind turbine gearboxes, the health of bearings is closely related to the safe and stable operation of the entire wind turbine gearbox.

[0003] At present, the fault diagnosis process of bearings mainly includes three steps: signal preprocessing, fault feature extraction, and fault classification and identification. However, due to the complexity of the application environment of wind turbine gearboxes and the variability of signal transmission paths, the existing fault diagnosis methods have many limitations, and there is currently no method that can accurately diagnose wind turbine gearbox bearing faults. Summary of the invention

[0004] The content of this application is used to introduce concepts in a brief form, which will be described in detail in the detailed implementation section below. The content of this application is not intended to identify the key features or essential features of the technical solution claimed for protection, nor is it intended to limit the scope of the technical solution claimed for protection.

[0005] Some embodiments of the present application propose a wind turbine gearbox bearing fault diagnosis method, device, electronic device and computer-readable medium to solve the technical problems mentioned in the above background technology section.

[0006] As a first aspect of the present application, some embodiments of the present application provide a method for diagnosing a wind turbine gearbox bearing fault, comprising: preprocessing a detection signal of a wind turbine gearbox bearing; extracting fault features from the preprocessed detection signal; and inputting the extracted fault features into a fault diagnosis model so that the fault diagnosis model outputs a fault diagnosis result.

[0007] Optionally, in some embodiments of the present application, the preprocessing of the detection signal of the wind turbine gearbox bearing includes:

[0008] A VMD algorithm enhanced by an IFOA algorithm is used to decompose multiple signal components from the detection signal.

[0009] Optionally, in some embodiments of the present application, extracting fault features from the preprocessed detection signal includes:

[0010] According to the correlation coefficient index, the signal components with greater correlation are selected to form a feature vector group.

[0011] Optionally, in some embodiments of the present application, the fault feature includes the feature vector group.

[0012] Optionally, in some embodiments of the present application, the fault diagnosis model is a recurrent neural network model.

[0013] Optionally, in some embodiments of the present application, the fault diagnosis model is a GRU model.

[0014] Optionally, in some embodiments of the present application, the wind turbine gearbox bearing fault diagnosis method further includes: training or providing the fault diagnosis model.

[0015] As a second aspect of the present application, some embodiments of the present application provide a wind turbine gearbox bearing fault diagnosis device, including: a signal module for preprocessing the detection signal of the wind turbine gearbox bearing; an extraction module for extracting fault features from the preprocessed detection signal; and a diagnosis module for inputting the fault feature extraction into a fault diagnosis model so that the fault diagnosis model outputs a fault diagnosis result.

[0016] As the third aspect of the present application, some embodiments of the present application provide an electronic device, comprising: one or more processors; a storage device on which one or more programs are stored, and when the one or more programs are executed by one or more processors, the one or more processors implement the method described in any implementation manner of the above-mentioned first aspect.

[0017] As a fourth aspect of the present application, some embodiments of the present application provide a computer-readable medium on which a computer program is stored, wherein when the program is executed by a processor, the method described in any implementation of the above-mentioned first aspect is implemented.

[0018] The beneficial effects of the present application are: providing a wind turbine gearbox bearing fault diagnosis method, device, electronic device and computer-readable medium that can improve fault diagnosis accuracy based on signal preprocessing and a neural network model.

[0019] More specifically, some embodiments of the present application may produce the following specific beneficial effects:

[0020] The IFOA algorithm is used to optimize the parameters of the VMD algorithm so as to effectively decompose the signal components from the detection signal;

[0021] Feature extraction based on signal components with good decomposition performance makes the trained neural network model have better application effect;

[0022] The GRU model is used to solve the problem of gradient disappearance or explosion in other neural network models. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The drawings constituting a part of this application are used to provide a further understanding of this application, so that other features, purposes and advantages of this application become more obvious. The illustrative embodiment drawings and their descriptions of this application are used to explain this application and do not constitute an improper limitation on this application.

[0024] In addition, throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that the components and elements are not necessarily drawn to scale.

[0025] In the attached picture:

[0026] Figure 1 It is a schematic diagram of the principle of a method for diagnosing a bearing fault of a wind turbine gearbox according to an embodiment of the present application;

[0027] Figure 2 It is a schematic diagram comparing the convergence capabilities of the IFOA algorithm and other optimization algorithms used in the wind turbine gearbox bearing fault diagnosis method according to an embodiment of the present application;

[0028] Figure 3 is a frequency domain diagram of a signal obtained by decomposition under working conditions according to an embodiment of the present application;

[0029] Figure 4 This is the comparison result between the E-VMD of this application and other EMD algorithms;

[0030] Figure 5 It is a schematic diagram of the process of expanding the data set of this application;

[0031] Figure 6 It is a loss curve and accuracy change diagram of the GRU model according to an embodiment of the present application;

[0032] Figure 7 is a schematic diagram of a confusion matrix output by a GRU model according to an embodiment of the present application;

[0033] Figure 8 This is a schematic diagram of the application of a wind turbine gearbox bearing fault diagnosis method in an embodiment according to an embodiment of the present application;

[0034] Fig. 9 It is a schematic diagram of the main steps of a method for diagnosing a bearing fault in a wind turbine gearbox according to an embodiment of the present application;

[0035] Fig.10 is a schematic diagram of the structure of a wind turbine gearbox bearing fault diagnosis device according to an embodiment of the present application;

[0036] Fig.11 It is a structural schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0037] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments set forth herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0038] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present disclosure can be combined with each other.

[0039] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0040] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0041] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0042] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0043] Reference Figure 1 As shown, in order to implement the wind turbine gearbox bearing fault diagnosis method of the present application, it is necessary to preferentially train or provide a fault diagnosis model.

[0044] Specifically, training or providing a fault diagnosis model includes:

[0045] Step 1: In the signal preprocessing stage, an E-VMD algorithm (Enhanced VMD, which refers to the VMD algorithm enhanced by IFOA) is proposed to improve (enhance) the VMD (Variational Mode Decomposition) algorithm through IFOA (Improved Fruit Fly Optimization Algorithm). That is, the IFOA algorithm (improved fruit fly algorithm) is used to optimize the parameters of the VMD algorithm to obtain the effective decomposition components in the signal.

[0046] The specific principle of step 1 is as follows:

[0047] First, this application briefly describes the basic theory and application steps of VMD. The mathematical principle of the VMD algorithm is shown in formula (1):

[0048]

[0049] In the formula, is the partial derivative with time, δ(t) is the impulse function, μ k (t) is the k-order intrinsic modal component (IMF) obtained after the original signal is decomposed by the VMD algorithm, and f k is the center frequency of the original data signal.

[0050] The constraint expressions it satisfies are as follows:

[0051]

[0052] Where s(t) is the input data signal and K is the total number of IMFs.

[0053] VMD abandons the recursive framework of EMD and transforms the signal decomposition problem into the solution of the variational problem. By introducing the penalty factor α and the Lagrangian multiplier λ, the constrained variational problem is further transformed into an unconstrained variational problem. The extended Lagrangian function obtained is:

[0054]

[0055] In the formula, the penalty factor α is introduced to represent the reconstruction accuracy of the signal, λ is the Lagrange multiplier, {μ k} is the kth order modal component obtained after VMD decomposition, {ω k} is {μ k} center frequency.

[0056] Based on the basic theory and practical application of VMD, this application briefly describes the application steps of VMD as follows:

[0057] 1. Load the nonlinear and discontinuous wind turbine gearbox bearing vibration signal data to be processed;

[0058] 2. μ k ,ω k , λ(ω) is initialized, and the number of iterations N=0;

[0059] 3. The number of iterations starts the loop in the form of N=N+1;

[0060] 4. As the number of iterations increases, μ is updated by the following formulas: k (ω),ω k ,λ(ω) three variables:

[0061]

[0062] 5. When the formula condition is met, the iteration loop stops. Otherwise, return to step 3 until the condition is met.

[0063]

[0064] Where s(ω) is the Fourier transform of the input data signal s(t), μ(ω) is the Fourier transform of the kth order component after VMD decomposition, and λ(ω) is the Lagrange multiplier μ k (t) is the Fourier transform of

[0065] Step 2: In the fault feature extraction process, the correlation coefficient index is introduced to screen the principal components with greater correlation to avoid the redundancy of the obtained components. At the same time, it is proposed to use multi-scale entropy (power spectrum entropy, singular spectrum entropy, energy entropy, approximate entropy, sample entropy, permutation entropy and fuzzy entropy) to construct the feature vector group to ensure that the extracted feature information is not missed to the greatest extent;

[0066] Step 3: In the fault classification and identification stage, the GRU model is introduced to implement fault diagnosis and solve the problem of gradient disappearance or explosion in traditional neural network recognition.

[0067] The specific principle of the GRU model in step 3 is as follows:

[0068] First, the constructed t groups of feature vectors are used as the input sequence x=(x1,x2,…,x t ), get two gating signals:

[0069] z t =σ(W (z) x t +U (z) h t-1 ) (8)

[0070] r t =σ(W(r) x t +U (r) h t-1 ) (9)

[0071] In formula (8) and formula (9), W (z) ,U (z) ,W (r) ,U (r) is a matrix with different weights.

[0072] The GRU gate state is obtained by the state information h of the previous moment. t-1 and the current input x t The update gate adds the two parts of information and obtains the update gate z through the sigmoid activation function. t , reset gate r t Obtained similarly.

[0073] When the reset gate is applied, the new memory content will be reset to store the past related information, which can be expressed as follows:

[0074] h' t =tanh(Wx t +r t Uh t-1 ) (10)

[0075] Where W, U are weight matrices, and tanh is the hyperbolic tangent function.

[0076] Finally, the final memory h needs to be calculated t , the current unit information will be retained and passed to the next unit. During this transmission process, the update gate determines the current memory content h' t With the previous moment h t-1 The state information that needs to be retained in . It can be expressed as follows:

[0077] h t =z t h t-1 +(1-z t )h' t (11)

[0078] y t =σ(W O h t ) (12)

[0079] Among them, W O is the weight matrix, z t is the activation result of the update gate, y t is the final output of GRU.

[0080] The improved variational mode decomposition-gated recurrent unit (E-VMD-GRU) method proposed in this application combines the E-VMD method with GRU and is applied to the fault diagnosis of wind turbine gearbox bearings. The E-VMD algorithm method can optimize the parameters of the traditional VMD algorithm without the problem of modal aliasing, and its fault classification and recognition accuracy is high.

[0081] Reference Figures 1 to 10 As shown, as a specific solution of this application, it is described as follows:

[0082] Step 1: In the signal preprocessing stage, an E-VMD algorithm combining IFOA and VMD algorithms is proposed. IFOA is used to optimize the parameters of the VMD algorithm to obtain the effective decomposition components in the signal.

[0083] In order to verify the effectiveness of the proposed E-VMD algorithm, this application selected the bearing fault data in the Case Western Reserve University Bearing Fault Public Dataset (CWRU) for verification. The basic situation of the bearing fault vibration signal obtained is: the load level is 3 horsepower, the bearing fault type is rolling element fault, and the fault diameter is 7 inches. For the convenience of subsequent description, this situation is briefly described as R7_3. In addition, in order to highlight the superiority of VMD, the EMD algorithm widely used in industry and the three improved algorithms EEMD, CEEMD, and CEEMDAN based on EMD are used for comparison with the E-VMD algorithm. Based on the fixed variable analysis strategy, the R7_3 (Table 1) data is also used for algorithm comparison.

[0084] Table 1 CWRU data label classification

[0085]

[0086] The original signal R7_3 is decomposed by the E-VMD algorithm: First, IFOA is used to optimize the parameters of the VMD algorithm. The optimization process is as follows: Figure 2 shown.

[0087] In the optimization process, this application applies different optimization algorithms to VMD for optimization comparison. The results show that the fitness value of the IFOA algorithm has converged to the global optimal solution of 7.354 in the 6th iteration, which is better than the optimization results of other algorithms, thus proving that the IFOA algorithm can efficiently implement the optimization process of VMD. The parameter K value after optimization is 17 and the penalty factor α is 160. In addition, for the convenience of analysis, the correlation coefficient is calculated to obtain 6 narrowband signal components decomposed by the E-VMD algorithm. The spectrum diagram of the obtained components is shown in the figure below: Figure 3As shown in the figure, it can be seen that within the same scale time, there is no overlap of the same characteristic frequency in the component frequency domain diagram, no mode aliasing phenomenon, and the components obtained by VMD decomposition retain the primary frequency, suppress the weak frequency components, and have excellent decomposition performance.

[0088] Based on the same process, the R7_3 data was processed using EMD, EEMD, CEEMD and CEEMDAN decomposition algorithms, the correlation coefficients of the obtained components were calculated, and the first 5 components were selected. The component spectrum diagram obtained is shown in the figure below: Figure 4 As shown. Comparing the results with the decomposition spectrum of the E-VMD algorithm, it can be found that within the same time scale, there are more characteristic frequencies overlapping, the EMD modal aliasing is more serious, and there is an endpoint effect. At the same time, the other three methods based on EMD improvement all have more or less modal aliasing problems and poor decomposition performance.

[0089] Through comparative analysis, it can be seen that the E-VMD algorithm method proposed in this application has better performance in the processing application of nonlinear and discontinuous stationary signals, and can well solve the modal aliasing problem existing in traditional algorithms.

[0090] Step 2: In the fault feature extraction process, the correlation coefficient index is used to screen the principal components with greater correlation to avoid the redundancy of the obtained components. At the same time, it is proposed to use multi-scale entropy (power spectrum entropy, singular spectrum entropy, energy entropy, approximate entropy, sample entropy, permutation entropy, fuzzy entropy) to construct a feature vector group to ensure that the extracted feature information is not missed to the greatest extent;

[0091] Step 3: In the fault classification and identification link, the GRU model is introduced to realize fault diagnosis and solve the problem of gradient disappearance or explosion in traditional neural network recognition.

[0092] Based on the effectiveness and superiority of the E-VMD algorithm, we verify whether the fusion of GRU and it can effectively implement the fault diagnosis strategy.

[0093] The data set selected in this application contains the gearbox bearing vibration acceleration data under 9 fault states and 1 normal state, where 9 can be calculated by combining 3 fault types and 3 fault diameters. Based on the selected fault data set, the bearing fault data set can be divided into 10 health states, as shown in Table 1.

[0094] In Table 1, I7_3 represents the inner ring fault data with a defect diameter of 0.007 inches under 3Hp-1730RPM conditions, R7_3 represents the rolling element fault data with a defect diameter of 0.007 inches under Hp-1730RPM conditions, and O7_3 represents the outer ring fault data with a defect diameter of 0.007 inches under Hp-1730RPM conditions. Similarly, the expressions of bearing fault types with defect diameters of 0.014 and 0.021 inches in the table have the same meaning and are not described. In order to simulate the complexity and diversity of vibration signals in the real world and avoid data overfitting, this application collects 12,000 discrete data points for each fault type and uses the overlapping sliding window method to expand the data set. The expansion process is as follows: Figure 5 shown.

[0095] The specific operation is as follows: the data is expanded using an overlapping sliding window of 4000 data points to complete the construction of a 1215×12000 data set.

[0096] After completing the construction of the data set, according to Figure 1 The overall strategy process implements the diagnostic strategy.

[0097] First, the E-VMD algorithm is used to complete the 1215×12000 data decomposition process. During the processing, due to the different number K of modal decompositions after optimization for each fault type, the dimensions of the feature vectors after decomposition are different, and the minimum K value of the normal state is 9. Therefore, for the convenience of analysis, we solve the correlation coefficients of the components corresponding to each group of data, and this application selects the first 5 components with larger correlation coefficients. At the same time, the features of the selected 5 components are extracted through the multi-scale entropy proposed in this application, thereby forming a 1215×35 feature vector group. In the process of forming the feature vector group, we can draw a conclusion that under the same bearing state, although the experimental samples are different, the feature vector difference value is small, and multiple features can be obtained. However, if different bearings are used, the difference value is larger. Therefore, it is also explained here that the use of multi-scale entropy to construct feature vectors can not only describe the characteristics of different operating states of bearings, but also be more conducive to fault classification and identification.

[0098] After E-VMD algorithm, correlation coefficient and multi-scale entropy processing, the constructed 1215×35 feature vector database is put into the GRU network as input, and the training set, validation set and test set are randomly divided into (0.6:0.2:0.2). During the training process, the single training iteration size is 10, the batch size is 300, and the Adam algorithm learning rate is 0.001. Figure 6 The loss accuracy curves of training samples and validation samples as they change with a single training iteration are given. Figure 6It can be seen that the loss curves of the training set and the validation set are very smooth, and the model is highly stable. After 20 experiments, the classification results are consistent, and the average accuracy has reached 99.5%, with no misclassification, and good robustness and applicability. The confusion matrix is ​​as follows Figure 7 shown.

[0099] Finally, in order to verify the superiority of the E-VMD algorithm-GRU, based on the same data and the divided labels in Table 1. This application combines the GRU model with four different algorithms: EMD, EEMD, CEEMD, and CEEMDAN. At the same time, the E-VMD algorithm is compared with three different models: SVM, BP, and LSTM. The results are shown in Table 2. Through comparative analysis, it can be seen that the method proposed in this application has high accuracy, excellent diagnostic effect, and strong robustness.

[0100] The present application is further described below based on a specific embodiment.

[0101] This application conducts actual engineering application research based on the high-speed bearing failure data of a certain offshore 2MW wind turbine gearbox.

[0102] In the analysis and research, in order to avoid overfitting, the data set was enhanced in the program in the same way to form a database of 1235 × 21000. The feature vector group was constructed by the E-VMD algorithm, correlation coefficient and multi-scale entropy.

[0103] Similarly, for the convenience of analysis, we still select the first 5 components with larger correlation coefficients after being processed by the E-VMD algorithm, and finally form a feature vector group (1235×35). Then, the feature vector group is input into the GRU network, and the training set, validation set and test set are randomly divided according to the ratio of 6:2:2. During the training process, the training parameters are consistent with those adopted in the simulation training. That is, the single training iteration size is 10, the data batch size is 300, and the Adam algorithm learning rate is 0.001. Through experiments, the loss accuracy of the training samples and the validation samples changes with the single training iteration. Figure 8 As shown. As can be seen from the figure, the loss curves of the training set and the validation set are very smooth, and the model stability is relatively high. According to the analysis of the changes in the loss curve, the loss curve decreases after the first 12 single training iterations, and the model begins to fit. After the 10th single training iteration, the loss curves overlap, and the absolute value of the slope eventually approaches 0, and the entire model converges. Finally, the changes in the accuracy of the training samples and the validation samples are observed. Also after the 10th single training iteration, the accuracy is close to 100%. After 20 experiments, the classification results are consistent, and the average accuracy reaches 99.5%. No misclassification occurs, and it has good robustness and applicability. The confusion matrix is ​​shown as follows Figure 8 shown.

[0104] In summary, the CRWU simulation experiment verifies the effectiveness and superiority of the combination of the E-VMD algorithm and the GRU model. The actual signal data further verifies the reliability of the strategy proposed in this application, providing a new method for bearing fault diagnosis of wind turbine transmission systems.

[0105] Based on the above technical solutions and embodiments, refer to Fig. 9 As shown, the present application provides a method for diagnosing a bearing fault in a wind turbine gearbox, including:

[0106] S101, pre-processing the detection signal of the wind turbine gearbox bearing.

[0107] S102: Extract fault features from the pre-processed detection signal.

[0108] S103, extracting the fault features and inputting them into a fault diagnosis model so that the fault diagnosis model outputs a fault diagnosis result.

[0109] Furthermore, the preprocessing of the detection signal of the wind turbine gearbox bearing includes:

[0110] The detection signal is decomposed into multiple signal components by combining the IFOA algorithm with the VMD algorithm.

[0111] Furthermore, the extraction of fault features from the preprocessed detection signal includes:

[0112] According to the correlation coefficient index, the signal components with greater correlation are selected to form a feature vector group.

[0113] Furthermore, the fault characteristics include the characteristic vector group.

[0114] Furthermore, the fault diagnosis model is a recurrent neural network model.

[0115] Furthermore, the fault diagnosis model is a GRU model.

[0116] Furthermore, the wind turbine gearbox bearing fault diagnosis method further includes: S104, training or providing the fault diagnosis model.

[0117] On the other hand, referring to Fig.10 As shown, some embodiments of the present application provide a wind turbine gearbox bearing fault diagnosis device including: a signal module, an extraction module and a diagnosis module.

[0118] The signal module is used to preprocess the detection signal of the wind turbine gearbox bearing; the extraction module is used to extract fault features from the preprocessed detection signal; and the diagnosis module is used to input the fault feature extraction into a fault diagnosis model so that the fault diagnosis model outputs a fault diagnosis result.

[0119] Reference Fig.11 As shown, the electronic device 800 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage device 808 into a random access memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the electronic device 800 are also stored. The processing device 801, the ROM 802, and the RAM 803 are connected to each other via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0120] Typically, the following devices may be connected to the I / O interface 805: input devices 806 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 807 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 808 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 809. The communication device 809 may allow the electronic device 800 to communicate with other devices wirelessly or by wire to exchange data. Although Fig.11 The electronic device 800 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead. Fig.11 Each block shown in the figure may represent one device, or may represent multiple devices as required.

[0121] In particular, according to some embodiments of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, some embodiments of the present disclosure include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In some such embodiments, the computer program can be downloaded and installed from the network through the communication device 809, or installed from the storage device 808, or installed from the ROM 802. When the computer program is executed by the processing device 801, the above-mentioned functions defined in the method of some embodiments of the present disclosure are executed.

[0122] It should be noted that the computer-readable medium described above in some embodiments of the present disclosure may be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0123] In some embodiments of the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device, or device. In some embodiments of the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in combination with an instruction execution system, device, or device. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0124] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperTextTransferProtocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an adhoc peer-to-peer network), as well as any currently known or future developed network.

[0125] The computer-readable medium may be included in the electronic device, or may exist independently without being installed in the electronic device.

[0126] Computer program code for performing the operations of some embodiments of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate 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 may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0127] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, which contains one or more executable instructions for implementing a specified logical function.

[0128] It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures.

[0129] For example, two boxes shown in succession may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of boxes in the block diagram and / or flow chart, can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0130] The units described in some embodiments of the present disclosure may be implemented by software or by hardware.

[0131] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0132] The above descriptions are only some preferred embodiments of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but should also cover other technical solutions formed by any combination of the above-mentioned technical features or their equivalent features without departing from the above-mentioned inventive concept. For example, the above-mentioned features are replaced with the technical features with similar functions disclosed in the embodiments of the present disclosure (but not limited to) and the technical solutions formed.

Claims

1. A method for diagnosing a bearing fault in a wind turbine gearbox, comprising: Preprocess the detection signal of the wind turbine gearbox bearing; Extract fault features from the preprocessed detection signal; The fault feature extraction is input into a fault diagnosis model so that the fault diagnosis model outputs a fault diagnosis result.

2. The wind turbine gearbox bearing fault diagnosis method according to claim 1, characterized in that: The preprocessing of the detection signal of the wind turbine gearbox bearing comprises: A VMD algorithm enhanced by an IFOA algorithm is used to decompose multiple signal components from the detection signal.

3. The wind turbine gearbox bearing fault diagnosis method according to claim 1, characterized in that: The fault feature extraction of the pre-processed detection signal includes: According to the correlation coefficient index, the signal components with greater correlation are selected to form a feature vector group.

4. The wind turbine gearbox bearing fault diagnosis method according to claim 3 is characterized in that: The fault characteristics include the characteristic vector group.

5. The wind turbine gearbox bearing fault diagnosis method according to claim 4, characterized in that: in, The fault diagnosis model is a recurrent neural network model.

6. The wind turbine gearbox bearing fault diagnosis method according to claim 5, characterized in that: in, The fault diagnosis model is a GRU model.

7. The wind turbine gearbox bearing fault diagnosis method according to any one of claims 1 to 6, characterized in that: Also includes: The fault diagnosis model is trained or provided.

8. A wind turbine gearbox bearing fault diagnosis device, comprising: A signal module, used for pre-processing the detection signal of the wind turbine gearbox bearing; An extraction module, used for extracting fault features from the detection signal after preprocessing; The diagnosis module is used to extract the fault characteristics and input them into a fault diagnosis model so that the fault diagnosis model outputs a fault diagnosis result.

9. An electronic device, comprising: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the processors implement the method according to any one of claims 1 to 7.

10. A computer readable medium having a computer program stored thereon, wherein: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images