Method, system and device for detecting operating state of wind turbine and storage medium

Abstract

The application discloses a kind of detection methods, systems, equipment and storage medium of wind generator operating state.Therein, the method comprises: obtaining the vibration signal of wind generator in predetermined time period;Vibration signal is input to the feature extraction model trained in advance, and mutual information topology global vibration waveform feature matrix is obtained;Based on mutual information topology global vibration waveform feature matrix, optimization mutual information topology global vibration waveform feature matrix is obtained;Optimization mutual information topology global vibration waveform feature matrix is input to the classifier well trained in advance, and classification result is obtained;Based on classification result, the operating state of wind generator is detected.The technical scheme of the application, the operating state of wind generator is detected whether normal, to ensure the normal operation of wind generator.

Description

Technical Field

[0001] This application relates to the field of wind power generation technology, and in particular to a method, system, device and storage medium for detecting the operating status of a wind turbine. Background Technology

[0002] Because wind farms are typically located in remote areas, routine inspection and maintenance of wind turbines is often difficult and costly in terms of labor. Therefore, SCADA (Supervisory Control and Data Acquisition) systems have been introduced for wind turbine condition monitoring. SCADA systems collect a wide variety of data (e.g., vibration, speed, temperature, energy, and angle) in large quantities, and analyzing key data can enable condition monitoring of wind turbines. However, analyzing and interpreting the massive amounts of SCADA data from wind turbines is very difficult, and the reliability of the data analysis results is insufficient due to the environmental conditions and other operating factors affecting the wind turbines. Summary of the Invention

[0003] This application provides a method, system, device, and storage medium for detecting the operating status of a wind turbine. It can detect whether the operating status of a wind turbine is normal, thereby ensuring the normal operation of the wind turbine.

[0004] In a first aspect, embodiments of this application provide a method for detecting the operating state of a wind turbine, comprising: acquiring vibration signals of the wind turbine within a predetermined time period; inputting the vibration signals into a pre-trained feature extraction model to acquire a mutual information topology global vibration waveform feature matrix; wherein the feature extraction model has been pre-learned to acquire the ability to acquire a feature matrix based on feature information in the signal; acquiring an optimized mutual information topology global vibration waveform feature matrix based on the mutual information topology global vibration waveform feature matrix; inputting the optimized mutual information topology global vibration waveform feature matrix into a pre-trained classifier to obtain a classification result; wherein the classification result is used to indicate whether the operating state of the wind turbine is normal; and detecting the operating state of the wind turbine based on the classification result.

[0005] In this technical solution, the operating status of the wind turbine can be detected based on the acquired vibration signal of the wind turbine, so as to determine whether the operating status of the wind turbine is normal, thereby ensuring the normal operation of the wind turbine.

[0006] In one implementation, the feature extraction model includes a first convolutional neural network, a second convolutional neural network, and a graph neural network. The step of inputting the vibration signal into the pre-trained feature extraction model to obtain a mutual information topological global vibration waveform feature matrix includes: inputting the vibration signal into the first convolutional neural network to obtain multiple vibration waveform feature vectors; generating a mutual information input matrix based on the mutual information between the multiple vibration waveform feature vectors; inputting the mutual information input matrix into the second convolutional neural network to obtain a mutual information feature matrix; arranging the multiple vibration waveform feature vectors in a two-dimensional manner to obtain a global vibration waveform feature matrix; and inputting the mutual information feature matrix and the global vibration waveform feature matrix into the graph neural network to obtain the mutual information topological global vibration waveform feature matrix.

[0007] In one optional implementation, the step of inputting the vibration signal into the first convolutional neural network to obtain multiple vibration waveform feature vectors includes: performing convolution processing on the vibration signal to obtain a convolutional feature map; performing mean pooling processing on the convolutional feature map based on the feature matrix to obtain a pooling feature map; performing nonlinear activation on the pooling feature map to obtain an activation feature map; and obtaining the multiple vibration waveform feature vectors based on the activation feature map.

[0008] In one alternative implementation, the formula for obtaining the mutual information is:

[0009] in, and Each represents any two distinct feature vectors among the plurality of vibration waveforms. This represents the mutual information. Indicates the The first eigenvalue at each position in the middle, Indicates the The second eigenvalue at each position in the middle, For the and stated The joint probability density function, and This is the marginal density function.

[0010] In one implementation, the classifier includes multiple fully connected layers and a normalized index Softmax classification function. The step of inputting the optimized mutual information topological global vibration waveform feature matrix into the pre-trained classifier to obtain a classification result includes: expanding the optimized mutual information topological global vibration waveform feature matrix to obtain a classification feature vector; inputting the classification feature vector into the multiple fully connected layers to obtain an encoded classification feature vector; and inputting the encoded classification feature vector into the Softmax classification function to obtain the classification result.

[0011] Secondly, embodiments of this application provide a system for detecting the operating status of a wind turbine, comprising: an acquisition module for acquiring vibration signals of the wind turbine within a predetermined time period; a first processing module for inputting the vibration signals into a pre-trained feature extraction model to acquire a mutual information topology global vibration waveform feature matrix; wherein the feature extraction model has been pre-learned to acquire the ability to acquire a feature matrix based on feature information in the signal; a second processing module for acquiring an optimized mutual information topology global vibration waveform feature matrix based on the mutual information topology global vibration waveform feature matrix; a third processing module for inputting the optimized mutual information topology global vibration waveform feature matrix into a pre-trained classifier to obtain a classification result; wherein the classification result is used to indicate whether the operating status of the wind turbine is normal; and a detection module for detecting the operating status of the wind turbine based on the classification result.

[0012] In one implementation, the feature extraction model includes a first convolutional neural network, a second convolutional neural network, and a graph neural network. Specifically, the first processing module is used to: input the vibration signal into the first convolutional neural network to obtain multiple vibration waveform feature vectors; generate a mutual information input matrix based on the mutual information between the multiple vibration waveform feature vectors; input the mutual information input matrix into the second convolutional neural network to obtain a mutual information feature matrix; arrange the multiple vibration waveform feature vectors in a two-dimensional arrangement to obtain a global vibration waveform feature matrix; and input the mutual information feature matrix and the global vibration waveform feature matrix into the graph neural network to obtain the mutual information topology global vibration waveform feature matrix.

[0013] In one optional implementation, the first processing module is specifically used to: perform convolution processing on the vibration signal to obtain a convolution feature map; perform mean pooling processing on the convolution feature map based on the feature matrix to obtain a pooling feature map; perform nonlinear activation on the pooling feature map to obtain an activation feature map; and obtain the plurality of vibration waveform feature vectors based on the activation feature map.

[0014] In one alternative implementation, the formula for obtaining the mutual information is:

[0015] in, and Each represents any two distinct feature vectors among the plurality of vibration waveforms. This represents the mutual information. Indicates the The first eigenvalue at each position in the middle, Indicates the The second eigenvalue at each position in the middle, For the and stated The joint probability density function, and This is the marginal density function.

[0016] In one implementation, the classifier includes multiple fully connected layers and a Softmax classification function. Specifically, the third processing module is used to: expand the optimized mutual information topology global vibration waveform feature matrix to obtain a classification feature vector; input the classification feature vector into the multiple fully connected layers to obtain an encoded classification feature vector; and input the encoded classification feature vector into the Softmax classification function to obtain the classification result.

[0017] Thirdly, embodiments of this application provide an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the wind turbine operating state detection method as described in the first aspect.

[0018] Fourthly, embodiments of this application provide a computer-readable storage medium for storing instructions that, when executed, cause the method described in the first aspect to be implemented.

[0019] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the steps of the method for detecting the operating state of a wind turbine generator as described in the first aspect.

[0020] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description

[0021] The accompanying drawings are provided for a better understanding of this solution and do not constitute a limitation of this application. Wherein: Figure 1 This is a schematic diagram of a method for detecting the working status of a wind turbine provided in an embodiment of this application; Figure 2 This is a schematic diagram of another method for detecting the operating status of a wind turbine provided in an embodiment of this application; Figure 3 This is a schematic diagram of another method for detecting the working status of a wind turbine provided in an embodiment of this application; Figure 4 This is an application scenario diagram of a wind turbine operating status detection system provided in an embodiment of this application; Figure 5 This is a schematic diagram of the architecture of a method for detecting the operating status of a wind turbine provided in an embodiment of this application; Figure 6 This is a schematic diagram of a wind turbine operating status detection system provided in an embodiment of this application; Figure 7 This is a schematic block diagram of an example electronic device that can be used to implement embodiments of this application. Detailed Implementation

[0022] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of this application, including various details to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0023] In the description of this application, unless otherwise stated, " / " means "or", for example, A / B can mean A or B; "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. The various numerical designations such as "first", "second", etc., involved in this application are only for the convenience of description and are not used to limit the scope of the embodiments of this application, nor do they indicate the order of events.

[0024] Please see Figure 1 , Figure 1 This is a schematic diagram of a method for detecting the operating status of a wind turbine generator according to an embodiment of this application. Figure 1 As shown, the method may include, but is not limited to, the following steps: Step S101: Obtain the vibration signal of the wind turbine within a predetermined time period.

[0025] In the embodiments of this application, the wind turbine may be an offshore wind turbine; the predetermined time period is the time period before the current time point (e.g., one hour before the current time point).

[0026] In the embodiments of this application, the aforementioned vibration signal is at least one.

[0027] For example, by pre-deploying multiple vibration sensors at multiple different locations on the wind turbine, vibration signals at each time point within a predetermined time period are collected, thereby obtaining multiple vibration signals of the wind turbine within the predetermined time period.

[0028] Step S102: Input the vibration signal into the pre-trained feature extraction model to obtain the mutual information topology global vibration waveform feature matrix.

[0029] In the embodiments of this application, the feature extraction model has been pre-learned to obtain the mutual information topology global vibration waveform feature matrix based on the feature information in the vibration signal.

[0030] It is understandable that, due to the implicit regularity of the vibration signals generated by wind turbines during normal operation, the vibration mode characteristics will also change accordingly when the wind turbine's operating state is abnormal, and these vibration mode characteristics will have mutual correlations at different locations on the wind turbine. Based on these characteristics, this application uses a pre-trained feature extraction model to extract the temporal implicit features of the vibration signals at multiple locations on the wind turbine from the vibration signals.

[0031] Step S103: Obtain the optimized mutual information topology global vibration waveform feature matrix based on the mutual information topology global vibration waveform feature matrix.

[0032] For example, the distribution transfer optimization oriented towards the class-bounded closed domain can be performed on the mutual information topology global vibration waveform feature matrix to obtain the optimized mutual information topology global vibration waveform feature matrix.

[0033] Step S104: Input the optimized mutual information topology global vibration waveform feature matrix into the pre-trained classifier to obtain the classification result.

[0034] In the embodiments of this application, the above classification results are used to indicate whether the working status of the wind turbine is normal.

[0035] As an example, the optimized mutual information topology global vibration waveform feature matrix is ​​input into a pre-trained classifier to obtain the first classification result indicating that the wind turbine is in normal working condition.

[0036] As another example, the optimized mutual information topology global vibration waveform feature matrix is ​​input into a pre-trained classifier to obtain a second classification result indicating that the wind turbine is not in normal working condition.

[0037] Step S105: Based on the classification results, detect the operating status of the wind turbine.

[0038] As an example, in response to the classification result being the first classification result, the operating status of the wind turbine is determined to be normal.

[0039] As another example, in response to the classification result being the second classification result, the operating state of the wind turbine is determined to be abnormal.

[0040] By implementing the embodiments of this application, the operating status of the wind turbine can be detected based on the acquired vibration signal of the wind turbine, so as to determine whether the operating status of the wind turbine is normal, thereby ensuring the normal operation of the wind turbine.

[0041] In one implementation, the feature extraction model includes a first convolutional neural network, a second convolutional neural network, and a graph neural network. The vibration signal can then be processed sequentially through these multiple models to obtain a mutual information topological global vibration waveform feature matrix. For an example, please refer to [link to example]. Figure 2 , Figure 2 This is a schematic diagram of another method for detecting the operating status of a wind turbine provided in an embodiment of this application. Figure 2 As shown, the method may include, but is not limited to, the following steps: Step S201: Obtain the vibration signal of the wind turbine within a predetermined time period.

[0042] In the embodiments of this application, step S201 can be implemented in any of the various embodiments of this disclosure. This disclosure does not limit this and will not elaborate further.

[0043] Step S202: Input the vibration signal into the first convolutional neural network to obtain multiple vibration waveform feature vectors.

[0044] In the embodiments of this application, the vibration signal includes a waveform diagram of each vibration signal.

[0045] For example, the vibration signal is input into the first convolutional neural network to perform feature mining on the waveforms of each vibration signal in the vibration signal, extract the hidden feature distribution information of the waveforms of each vibration signal in the vibration signal, and thus obtain multiple vibration waveform feature vectors.

[0046] In one optional implementation, the vibration signal is input into a first convolutional neural network to obtain multiple vibration waveform feature vectors, including: performing convolution processing on the vibration signal to obtain a convolutional feature map; performing mean pooling processing on the convolutional feature map based on the feature matrix to obtain a pooling feature map; performing nonlinear activation on the pooling feature map to obtain an activation feature map; and obtaining multiple vibration waveform feature vectors based on the activation feature map.

[0047] Step S203: Generate a mutual information input matrix based on the mutual information between multiple vibration waveform feature vectors.

[0048] For example, the mutual information between every two vibration waveform feature vectors is calculated to obtain multiple mutual information, and each mutual information is used as an element in a matrix to obtain the mutual information input matrix.

[0049] In the embodiments of this application, the aforementioned mutual information is used to measure the influence of the mutuality between the vibration features at two locations of the wind turbine corresponding to the two vibration waveform feature vectors on the final working state classification judgment.

[0050] In one alternative implementation, the formula for obtaining the above mutual information is:

[0051] in, Indicates mutual information, and Each represents any two distinct eigenvectors from a plurality of vibration waveforms. express The first eigenvalue at each position in the middle, express The second eigenvalue at each position in the middle, for and The joint probability density function, and This is the marginal density function.

[0052] It is understandable that mutual information measures the relevance between two objects. Therefore, in classification problems, mutual information can be used to measure the impact of the relevance between two features on the final classification judgment, that is, to measure the impact of the relevance between vibration features at various locations of a wind turbine on the final classification judgment of anomalies in its operating state. If the contributions of two features to the confidence level of the final class probability are similar, then the mutual information between the two features is relatively large. Therefore, this application uses the contribution of each vibration waveform feature to the confidence level of the final class probability to further enhance the expression of implicit features in the vibration signal, thereby improving the accuracy of judging the anomalies in the operating state of the wind turbine.

[0053] Step S204: Input the mutual information input matrix into the second convolutional neural network to obtain the mutual information feature matrix.

[0054] For example, the mutual information input matrix is ​​input into the second convolutional neural network to extract the hidden correlation features at each position in the mutual information input matrix, that is, the mutual correlation features between the vibration hidden features at each position of the wind turbine. The above mutual correlation features are used as elements in the matrix to obtain the mutual information topological feature matrix.

[0055] Step S205: Arrange multiple vibration waveform feature vectors in a two-dimensional manner to obtain the global vibration waveform feature matrix.

[0056] Step S206: Input the mutual information feature matrix and the global vibration waveform feature matrix into the graph neural network to obtain the mutual information topology global vibration waveform feature matrix.

[0057] For example, multiple vibration waveform feature vectors are used as the feature representation of nodes, and the mutual information feature matrix is ​​used as the feature representation of the edges between nodes. The global vibration waveform feature matrix and the mutual information feature matrix obtained by arranging multiple vibration waveform feature vectors in two dimensions are input into the graph neural network to obtain the mutual information topology global vibration waveform feature matrix.

[0058] Step S207: Obtain the optimized mutual information topology global vibration waveform feature matrix based on the mutual information topology global vibration waveform feature matrix.

[0059] In the embodiments of this application, step S207 can be implemented in any of the various embodiments of this disclosure. This disclosure does not limit this and will not elaborate further.

[0060] Step S208: Input the optimized mutual information topology global vibration waveform feature matrix into the pre-trained classifier to obtain the classification result.

[0061] In the embodiments of this application, step S208 can be implemented in any of the various embodiments of this disclosure. This disclosure does not limit this and will not elaborate further.

[0062] Step S209: Based on the classification results, detect the operating status of the wind turbine.

[0063] In the embodiments of this application, step S209 can be implemented in any of the various embodiments of this disclosure. This disclosure does not limit this and will not elaborate further.

[0064] By implementing the embodiments of this application, the vibration signal of the wind turbine can be sequentially input into the first convolutional neural network, the second convolutional neural network, and the graph neural network to obtain the mutual information topology global vibration waveform feature matrix, and then perform subsequent processing, thereby realizing the detection of the working status of the wind turbine to determine whether the working status of the wind turbine is normal and ensuring the normal operation of the wind turbine.

[0065] By implementing the embodiments of this application, a mutual information topology global vibration waveform feature matrix can be obtained based on the vibration signal of the wind turbine within a predetermined time period. Based on the eigenvalues ​​of the mutual information topology global vibration waveform feature matrix, an optimized mutual information topology global vibration waveform feature matrix can be obtained. Thus, the working status of the wind turbine can be detected based on the optimized mutual information topology global vibration waveform feature matrix to determine whether the working status of the wind turbine is normal and ensure the normal operation of the wind turbine.

[0066] In one implementation of this application, the classifier includes multiple fully connected layers and a normalized exponential Softmax classification function. This allows the optimized mutual information topological global vibration waveform feature matrix to be sequentially input into the aforementioned fully connected layers and Softmax classification function to obtain the classification result. For an example, please refer to... Figure 3 , Figure 3 This is a schematic diagram of another method for detecting the operating status of a wind turbine provided in an embodiment of this application. Figure 3 As shown, the method may include, but is not limited to, the following steps: Step S301: Obtain the vibration signal of the wind turbine within a predetermined time period.

[0067] In the embodiments of this application, step S301 can be implemented in any of the various embodiments of this disclosure. This disclosure does not limit this and will not elaborate further.

[0068] Step S302: Input the vibration signal into the pre-trained feature extraction model to obtain the mutual information topology global vibration waveform feature matrix.

[0069] In the embodiments of this application, step S302 can be implemented in any of the various embodiments of this disclosure. This disclosure does not limit this and will not elaborate further.

[0070] Step S303: Based on the mutual information topology global vibration waveform feature matrix, obtain the optimized mutual information topology global vibration waveform feature matrix.

[0071] In the embodiments of this application, step S303 can be implemented in any of the various embodiments of this disclosure. This disclosure does not limit this and will not elaborate further.

[0072] Step S304: Expand the global vibration waveform feature matrix of the optimized mutual information topology to obtain the classification feature vector.

[0073] As an example, the global vibration waveform feature matrix of the optimized mutual information topology is expanded into row vectors to obtain the classification feature vector.

[0074] As another example, the global vibration waveform feature matrix of the optimized mutual information topology is expanded into column vectors to obtain the classification feature vector.

[0075] Step S305: Input the classification feature vector into multiple fully connected layers to obtain the encoded classification feature vector.

[0076] For example, the classification feature vector is input into multiple fully connected layers for fully connected encoding to obtain the encoded classification feature vector.

[0077] Step S306: Input the encoded classification feature vector into the Softmax classification function to obtain the classification result.

[0078] In the embodiments of this application, the formula for generating the classification result can be expressed as:

[0079] in, This means projecting the optimized mutual information topology global vibration waveform feature matrix into a vector. and This represents the weight matrix of each fully connected layer in the classifier. and This represents the bias matrix of each fully connected layer.

[0080] Step S307: Based on the classification results, detect the operating status of the wind turbine.

[0081] In the embodiments of this application, step S307 can be implemented in any of the various embodiments of this disclosure. This disclosure does not limit this and will not elaborate further.

[0082] By implementing the embodiments of this application, the obtained optimized mutual information topology global vibration waveform feature matrix can be input into a classifier including multiple fully connected layers and a Softmax classification function to obtain classification results to detect whether the working state of the wind turbine is normal, thus ensuring the normal operation of the wind turbine.

[0083] In some embodiments of this application, the method for detecting the operating status of a wind turbine may further include: issuing a fault warning message in response to an abnormal operating status of the wind turbine. This enables maintenance personnel of the wind turbine to promptly repair the faulty wind turbine based on the warning message, ensuring the normal operation of the wind turbine.

[0084] Please see Figure 4 , Figure 4 This is an application scenario diagram of a wind turbine operating status detection system provided in an embodiment of this application. For example... Figure 4 As shown, in this application scenario, the data first needs to be obtained from the wind turbine (e.g., as shown in the image). Figure 4 Vibration signals (e.g., as shown in F) collected by multiple vibration sensors over a predetermined time period. Figure 4 (as shown in C1, C2, C3); then the acquired vibration signal is input to a server equipped with a method for detecting the operating status of a wind turbine (e.g., Figure 4 As shown in S), the server can process the vibration signal using the wind turbine operating status detection method provided in any embodiment of this application to generate a classification result indicating whether the wind turbine operating status is normal.

[0085] Please see Figure 5 , Figure 5 This is a schematic diagram of the architecture of a method for detecting the operating status of a wind turbine provided in an embodiment of this application, as shown below. Figure 5 As shown, in the architecture of the wind turbine operating status detection method provided in this application embodiment, vibration signals collected by multiple vibration sensors deployed on the wind turbine over a predetermined time period are first acquired. Then, the waveforms of each vibration signal are passed through a first convolutional neural network model, acting as a filter, to obtain multiple vibration waveform feature vectors. Next, the mutual information between every two vibration waveform feature vectors is calculated to obtain a mutual information input matrix. Then, the mutual information input matrix is ​​passed through a second convolutional neural network model, acting as a feature extractor, to obtain a mutual information feature matrix. Next, the multiple vibration waveform feature vectors are arranged in a two-dimensional manner to obtain a global vibration waveform feature matrix. Then, the global vibration waveform feature matrix and the mutual information feature matrix are passed through a graph neural network model to obtain a mutual information topology global vibration waveform feature matrix. Next, the mutual information topology global vibration waveform feature matrix is ​​optimized using a feature manifold to obtain an optimized mutual information topology global vibration waveform feature matrix. Finally, the optimized mutual information topology global vibration waveform feature matrix is ​​passed through a classifier to obtain a classification result, which is used to indicate whether the wind turbine's operating status is normal.

[0086] Please see Figure 6 , Figure 6 This is a schematic diagram of a wind turbine operating status detection system provided in an embodiment of this application, as shown below. Figure 6 As shown, the system 600 includes: an acquisition module 601 for acquiring vibration signals of a wind turbine within a predetermined time period; a first processing module 602 for inputting the vibration signals into a pre-trained feature extraction model to acquire a mutual information topology global vibration waveform feature matrix; wherein the feature extraction model has been pre-learned to acquire the mutual information topology global vibration waveform feature matrix based on feature information in the vibration signals; a second processing module 603 for acquiring an optimized mutual information topology global vibration waveform feature matrix based on the mutual information topology global vibration waveform feature matrix; a third processing module 604 for inputting the optimized mutual information topology global vibration waveform feature matrix into a pre-trained classifier to obtain a classification result; wherein the classification result is used to indicate whether the wind turbine's operating state is normal; and a detection module 605 for detecting the wind turbine's operating state based on the classification result.

[0087] In one implementation, the feature extraction model includes a first convolutional neural network, a second convolutional neural network, and a graph neural network. Specifically, the first processing module 602 is used to: input a vibration signal into the first convolutional neural network to obtain multiple vibration waveform feature vectors; generate a mutual information input matrix based on the mutual information between the multiple vibration waveform feature vectors; input the mutual information input matrix into the second convolutional neural network to obtain a mutual information feature matrix; arrange the multiple vibration waveform feature vectors in a two-dimensional arrangement to obtain a global vibration waveform feature matrix; and input the mutual information feature matrix and the global vibration waveform feature matrix into the graph neural network to obtain a mutual information topological global vibration waveform feature matrix.

[0088] In one optional implementation, the first processing module 602 is specifically used to: perform convolution processing on the vibration signal to obtain a convolution feature map; perform mean pooling processing on the convolution feature map based on the feature matrix to obtain a pooling feature map; perform nonlinear activation on the pooling feature map to obtain an activation feature map; and obtain multiple vibration waveform feature vectors based on the activation feature map.

[0089] In one alternative implementation, the formula for obtaining mutual information is:

[0090] in, and Each represents any two distinct eigenvectors from a plurality of vibration waveforms. Indicates mutual information, express The first eigenvalue at each position in the middle, express The second eigenvalue at each position in the middle, for and The joint probability density function, and This is the marginal density function.

[0091] 60 In one implementation, the classifier includes multiple fully connected layers and a Softmax classification function. Specifically, the third processing module 604 is used to: expand the optimized mutual information topology global vibration waveform feature matrix to obtain a classification feature vector; input the classification feature vector into multiple fully connected layers to obtain an encoded classification feature vector; and input the encoded classification feature vector into the Softmax classification function to obtain a classification result.

[0092] The apparatus of this application embodiment can detect the operating status of a wind turbine based on the acquired vibration signal of the wind turbine, thereby determining whether the wind turbine is operating normally and ensuring the normal operation of the wind turbine.

[0093] Based on embodiments of this application, this application also provides a computer-readable storage medium, wherein computer instructions are used to cause a computer to execute a method for detecting the operating state of a wind turbine generator according to any of the foregoing embodiments provided in this application.

[0094] Please see Figure 7 ,like Figure 7 The diagram shown is a schematic block diagram of an example electronic device that can be used to implement embodiments of the present application. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present application described and / or claimed herein.

[0095] like Figure 7 As shown, device 700 includes a computing unit 701, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 702 or a computer program loaded from storage unit 708 into random access memory (RAM) 703. The RAM 703 may also store various programs and data required for the operation of device 700. The computing unit 701, ROM 702, and RAM 703 are interconnected via bus 704. An input / output (I / O) interface 705 is also connected to bus 704.

[0096] Multiple components in device 700 are connected to I / O interface 705, including: input unit 706, such as keyboard, mouse, etc.; output unit 707, such as various types of monitors, speakers, etc.; storage unit 708, such as disk, optical disk, etc.; and communication unit 709, such as network card, modem, wireless transceiver, etc. Communication unit 709 allows device 700 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0097] The computing unit 701 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 701 performs the various methods and processes described above, such as the method for detecting the operating state of a wind turbine. For example, in some embodiments, the method for detecting the operating state of a wind turbine can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 708. In some embodiments, part or all of the computer program can be loaded and / or installed on device 700 via ROM 702 and / or communication unit 709. When the computer program is loaded into RAM 703 and executed by the computing unit 701, one or more steps of the method for detecting the operating state of a wind turbine described above can be performed. Alternatively, in other embodiments, the computing unit 701 may be configured by any other suitable means (e.g., by means of firmware) to perform a method for detecting the operating state of the wind turbine.

[0098] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard parts (ASSPs), systems-on-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0099] The program code used to implement the methods of this application may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0100] In the context of this application, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on 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 fibers, compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0101] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a cathode ray tube (CRT) or LCD (Liquid Crystal Display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0102] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), the Internet, and blockchain networks.

[0103] Computer systems can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is established by computer programs running on the respective computers and having a client-server relationship with each other. A server can be a cloud server, also known as a cloud computing server or cloud host, a hosting product within the cloud computing service system that addresses the shortcomings of traditional physical hosts and VPS (Virtual Private Server) services, such as high management difficulty and weak business scalability. Servers can also be servers for distributed systems or servers integrated with blockchain technology.

[0104] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this application can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this application can be achieved, and this is not limited herein.

[0105] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for detecting the operating status of a wind turbine generator, characterized in that, include: Acquire vibration signals of wind turbines within a predetermined time period; The vibration signal is input into a pre-trained feature extraction model to obtain a mutual information topology global vibration waveform feature matrix; wherein, the feature extraction model has been pre-learned to obtain the feature matrix based on the feature information in the signal; The mutual information topology global vibration waveform feature matrix is ​​optimized by performing feature manifold optimization to obtain the optimized mutual information topology global vibration waveform feature matrix; The optimized mutual information topology global vibration waveform feature matrix is ​​input into a pre-trained classifier to obtain a classification result; wherein, the classification result is used to indicate whether the wind turbine is operating normally; Based on the classification results, the operating status of the wind turbine is detected; The feature extraction model includes a first convolutional neural network, a second convolutional neural network, and a graph neural network. The step of inputting the vibration signal into the pre-trained feature extraction model to obtain the mutual information topological global vibration waveform feature matrix includes: The vibration signal is input into the first convolutional neural network to obtain multiple vibration waveform feature vectors; Based on the mutual information among the multiple vibration waveform feature vectors, a mutual information input matrix is ​​generated; The mutual information input matrix is ​​input into the second convolutional neural network to obtain the mutual information feature matrix; The multiple vibration waveform feature vectors are arranged in a two-dimensional manner to obtain a global vibration waveform feature matrix. The mutual information feature matrix and the global vibration waveform feature matrix are input into the graph neural network to obtain the mutual information topology global vibration waveform feature matrix.

2. The method as described in claim 1, characterized in that, The step of inputting the vibration signal into the first convolutional neural network to obtain multiple vibration waveform feature vectors includes: The vibration signal is convolved to obtain a convolutional feature map. The convolutional feature map is subjected to mean pooling based on the feature matrix to obtain a pooled feature map; The pooled feature map is nonlinearly activated to obtain an activated feature map; Based on the activation feature map, the multiple vibration waveform feature vectors are obtained.

3. The method as described in claim 1, characterized in that, The formula for obtaining the mutual information is: in, This represents the mutual information. and Each represents any two distinct feature vectors among the plurality of vibration waveforms. Indicates the The first eigenvalue at each position in the middle, Indicates the The second eigenvalue at each position in the middle, For the and stated The joint probability density function, and This is the marginal density function.

4. The method as described in claim 1, characterized in that, The classifier includes multiple fully connected layers and a normalized exponential Softmax classification function. The step of inputting the optimized mutual information topological global vibration waveform feature matrix into the pre-trained classifier to obtain the classification result includes: Expand the optimized mutual information topology global vibration waveform feature matrix to obtain the classification feature vector; The classification feature vector is input into the multiple fully connected layers to obtain the encoded classification feature vector; The encoded classification feature vector is input into the Softmax classification function to obtain the classification result.

5. A system for detecting the operating status of a wind turbine generator, characterized in that, include: The acquisition module is used to acquire vibration signals of the wind turbine within a predetermined time period; The first processing module is used to input the vibration signal into a pre-trained feature extraction model to obtain a mutual information topology global vibration waveform feature matrix; wherein, the feature extraction model has been pre-learned to obtain the feature matrix based on the feature information in the signal; The second processing module is used to perform feature manifold optimization on the mutual information topology global vibration waveform feature matrix to obtain the optimized mutual information topology global vibration waveform feature matrix. The third processing module is used to input the optimized mutual information topology global vibration waveform feature matrix into a pre-trained classifier to obtain a classification result; wherein, the classification result is used to indicate whether the working state of the wind turbine is normal. The detection module is used to detect the operating status of the wind turbine based on the classification results; The feature extraction model includes a first convolutional neural network, a second convolutional neural network, and a graph neural network, wherein the first processing module is specifically used for: The vibration signal is input into the first convolutional neural network to obtain multiple vibration waveform feature vectors; Based on the mutual information among the multiple vibration waveform feature vectors, a mutual information input matrix is ​​generated; The mutual information input matrix is ​​input into the second convolutional neural network to obtain the mutual information feature matrix; The multiple vibration waveform feature vectors are arranged in a two-dimensional manner to obtain a global vibration waveform feature matrix. The mutual information feature matrix and the global vibration waveform feature matrix are input into the graph neural network to obtain the mutual information topology global vibration waveform feature matrix.

6. An electronic device, characterized in that, include: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores instructions that can be executed 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 method for detecting the operating status of a wind turbine generator as described in any one of claims 1 to 4.

7. A computer-readable storage medium for storing instructions, characterized in that, When the instruction is executed, the method as described in any one of claims 1 to 4 is implemented.

Images