Safety guarantee method and system for emergency situation of wind turbine generator

Abstract

The invention discloses a safety guarantee method and system for an emergency situation of a wind turbine generator, and the method comprises the steps: collecting the operation data of the wind turbine generator in real time through a sensor, and carrying out the processing of the data; analyzing the processed data in real time, constructing a detection model, detecting the operation state of the wind turbine generator, and identifying and processing the fault of the wind turbine generator; and training the fault prediction model by using the historical fault data to predict the fault of the wind turbine generator. The operation state of the wind turbine generator is monitored in real time, potential faults are recognized and processed in time, fault prediction can be carried out through historical fault data, and the accuracy and timeliness of fault early warning are improved, so that the safety and stability of the wind turbine generator are remarkably improved, the shutdown loss and the maintenance cost are reduced, and the working efficiency is improved. And the continuous operation of the wind turbine generator in an emergency situation is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of fault management, and in particular to a safety guarantee method and system for sudden situations of wind turbines. Background Art

[0002] With the continued growth of global energy demand and increasing awareness of environmental protection, wind energy, as a clean, renewable energy source, has been widely used and rapidly developed. As a key device for wind energy conversion, the operational safety and stability of wind turbines are directly related to the efficiency and economic benefits of wind power generation. During wind turbine operation, early fault detection and resolution are key to ensuring their safe operation. Therefore, in recent years, research on sensor-based monitoring systems, data processing and analysis technologies, and fault detection and prediction algorithms has received widespread attention and in-depth development. Traditional wind turbine monitoring systems rely primarily on regular manual inspections and basic sensor data monitoring. This approach is not only inefficient but also fails to detect and resolve potential faults in a timely manner, easily leading to equipment damage and downtime losses.

[0003] While existing technologies can monitor and detect wind turbine faults to a certain extent, they still have many shortcomings. First, traditional monitoring systems have limited data acquisition and processing capabilities, making it difficult to fully and accurately reflect the operating status of wind turbines and effectively identify early-stage faults. Second, existing fault detection algorithms are mostly based on simple threshold judgments or single sensor data, lacking the ability to comprehensively analyze and process multi-source data, resulting in low accuracy and reliability in fault detection. Furthermore, existing fault prediction models often rely solely on historical data, ignoring the dynamic changes in real-time data, making it difficult to achieve early prediction and early warning of faults. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: the problem of untimely fault detection and inaccurate fault prediction during the operation of wind turbines due to insufficient data collection, limited processing and analysis capabilities, and imperfect automated fault handling mechanism.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a safety guarantee method for wind turbine generator system emergencies, comprising:

[0007] Collect wind turbine operating data in real time through sensors and process the data;

[0008] Analyze processed data in real time, build detection models, detect the operating status of wind turbines, identify wind turbine faults and handle them;

[0009] The fault prediction model is trained using historical fault data to predict wind turbine failures.

[0010] As a preferred solution of the safety assurance method for sudden situations of a wind turbine generator set described in the present invention, the operating data of the wind turbine generator set includes vibration data, temperature data, acoustic data, electrical data, environmental data, position and angle data, historical fault data and maintenance data.

[0011] As a preferred solution of the safety guarantee method for wind turbine generator system emergencies described in the present invention, the data processing includes pre-processing the collected data and fusing multi-source data;

[0012] Preprocessing includes data cleaning, data standardization, data noise reduction and feature extraction;

[0013] The formula of the state matrix of multi-source fusion data of wind turbines is expressed as:

[0014]

[0015] Among them, S(t) represents the state matrix of multi-source fusion data of wind turbine, N represents the number of sensors, T represents the upper limit of integration, λ represents the attenuation factor, W i represents the weight matrix of sensor i, x i (t) represents the data vector collected by sensor i at time t, t′ represents the integral variable, e -λ(T-t′) represents an exponential decay function;

[0016] All sensor data and weight matrices are normalized to [0, 1];

[0017] The value range of the state matrix of the multi-source fusion data of the wind turbine is also between [0, 1]. The larger the value, the better the state, and the smaller the value, the potential fault or abnormality.

[0018] As a preferred solution of the wind turbine emergency safety guarantee method described in the present invention, wherein: the real-time analysis of processed data, construction of a detection model, detection of the operating status of the wind turbine, identification of wind turbine faults and processing thereof include, through the value of the state matrix of the multi-source fusion data of the wind turbine, using the Bayesian network and random forest to construct a detection model, the formula is expressed as follows:

[0019]

[0020] Where D(t) represents the value of the detection model, B represents the number of decision trees in the random forest, and f b(x) represents the prediction function of the b-th decision tree, S(t) represents the state matrix of the wind turbine, N represents the number of sensors, P(S i (t)|Parents(S i (t))) represents the state conditional probability of sensor i in the Bayesian network, T represents the upper limit of integration, represents the total time, λ represents the attenuation factor, W i represents the weight matrix of sensor i, x i (t) represents the data vector collected by sensor i at time t, t′ represents the integral variable, e -λ(T-t′) represents an exponential decay function;

[0021] The value range of D(t) is between [0, 1], where 0 indicates the normal operation of the wind turbine and 1 indicates a serious fault or abnormality.

[0022] As a preferred solution of the wind turbine generator system emergency safety guarantee method of the present invention, wherein: the historical fault data training fault prediction model includes: an encoder extracting low-dimensional features through linear transformation and nonlinear activation function to remove data noise;

[0023] The decoder restores the low-dimensional features to the original data and verifies the effectiveness of feature extraction by comparing the reconstructed data with the original data.

[0024] The deep neural network uses the extracted low-dimensional features to build a fault prediction model through multiple layers of nonlinear transformations and outputs the state of the fault.

[0025] As a preferred solution of the wind turbine generator system emergency safety guarantee method of the present invention, wherein: the historical fault data training fault prediction model also includes, the encoder formula is expressed as:

[0026] z=σ(W e x+b e )

[0027] The decoder formula is expressed as:

[0028]

[0029] The deep neural network formula is expressed as:

[0030] h (l) =σ(W (l) h (l-1) +b (l) )(for l=1,2,…,L)

[0031] h (0) =z

[0032] y=σ(W (L)h (L-1) +b (L) )

[0033] Among them, z represents the feature vector encoded by the autoencoder, σ represents the activation function, and W e represents the weight matrix of the encoder, x represents the input data, b e represents the encoder bias term, Represents the reconstructed data after decoding, W d represents the weight matrix of the decoder, b d represents the decoder bias term, h (l) represents the activation output of the lth layer of the deep neural network, W (l) represents the weight matrix of the lth layer of the deep neural network, b (l) represents the bias term of the lth layer of the deep neural network, h (0) represents the initial input of the feature input after autoencoder encoding to the deep neural network, and y represents the final output of the deep neural network to predict the fault value;

[0034] The activation function σ selects the Sigmoid function;

[0035] Encoded feature z and decoded data The range of is [0, 1];

[0036] The value range of the deep neural network output y is [0, 1], which represents the probability of a fault occurring. The closer it is to 1, the higher the possibility of a fault occurring.

[0037] As a preferred solution of the wind turbine generator system emergency safety method of the present invention, the method of identifying and handling wind turbine generator system failures includes:

[0038] When D(t)<0.3, it is a minor fault, and the fault has little impact on the operation of the wind turbine;

[0039] The system will perform the following operations:

[0040] The system will continue to operate, but will start enhanced monitoring mode, recording data from all relevant sensors in real time to ensure that the fault characteristics are fully captured;

[0041] The system will issue an early warning notification to remind operation and maintenance personnel to pay attention, but will not automatically isolate the case.

[0042] When 0.3≤D(t)<0.7, it is a moderate fault, in which case the fault affects some functions of the wind turbine;

[0043] The system will perform the following operations:

[0044] The system will automatically isolate the fault area and stop the operation of related components to prevent the fault from spreading and affecting the normal operation of other components;

[0045] Start the backup system to take over the isolated part to ensure the normal operation of the wind turbine;

[0046] After receiving the notification, the operation and maintenance personnel will analyze the cause of the fault remotely or on-site, conduct a diagnosis based on the system's real-time monitoring data and historical data, and formulate a detailed maintenance plan;

[0047] When D(t)≥0.7, it is a serious fault, and the fault has a significant impact on the safe operation of the wind turbine;

[0048] The system will perform the following operations:

[0049] The system will immediately and comprehensively isolate the fault area and stop the operation of all related components to prevent the fault from spreading further;

[0050] Emergency start-up of the backup system to fully take over all functions of the isolated part to ensure the safe operation of the wind turbine as a whole;

[0051] The system will issue an emergency notification to inform the operation and maintenance personnel and management and activate the emergency response mechanism;

[0052] Operation and maintenance personnel must immediately rush to the site to conduct a comprehensive inspection and maintenance, and quickly locate and troubleshoot faults based on the real-time and historical data provided by the system.

[0053] A safety assurance method system for wind turbine generator system emergencies, wherein:

[0054] The data acquisition and processing module collects the operating data of the wind turbine in real time through sensors and processes the data;

[0055] Fault detection and processing module: analyzes processed data in real time, builds detection models, detects the operating status of wind turbines, identifies wind turbine faults and handles them;

[0056] The fault prediction module uses historical fault data to train a fault prediction model to predict wind turbine faults.

[0057] A computer device comprises: a memory and a processor; the memory stores a computer program, wherein the processor implements the steps of any one of the methods of the present invention when executing the computer program.

[0058] A computer-readable storage medium stores a computer program thereon, wherein the computer program, when executed by a processor, implements the steps of any one of the methods of the present invention.

[0059] The present invention provides a method for ensuring wind turbine safety during emergencies. This method uses multiple sensors to collect real-time operating data from wind turbines, and then comprehensively processes and analyzes this data to build an efficient fault detection and prediction model. This method not only monitors the operating status of wind turbines in real time, enabling the timely identification and resolution of potential faults, but also leverages historical fault data for fault prediction, improving the accuracy and timeliness of fault warnings. This significantly enhances the safety and stability of wind turbines, reduces downtime losses and maintenance costs, and ensures the continuous operation of wind turbines during emergencies. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0061] Figure 1 This is an overall flow chart of a method for ensuring safety of wind turbine generator system in case of emergencies provided by the first embodiment of the present invention. DETAILED DESCRIPTION

[0062] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0063] Example 1, with reference to Figure 1 , as one embodiment of the present invention, provides a safety assurance method for wind turbine generator system emergencies, comprising:

[0064] S1: Collect the operating data of wind turbines in real time through sensors and process the data.

[0065] The operating data of wind turbines include vibration data, temperature data, acoustic data, electrical data, environmental data, position and angle data, historical fault data and maintenance data.

[0066] Data processing includes preprocessing the collected data and multi-source data fusion.

[0067] Preprocessing includes data cleaning, data standardization, data noise reduction and feature extraction.

[0068] The formula of the state matrix of multi-source fusion data of wind turbines is expressed as:

[0069]

[0070] Among them, S(t) represents the state matrix of multi-source fusion data of wind turbine, N represents the number of sensors, T represents the upper limit of integration, λ represents the attenuation factor, W i represents the weight matrix of sensor i, x i (t) represents the data vector collected by sensor i at time t, t′ represents the integral variable, e -λ(T-t′) represents an exponential decay function.

[0071] All sensor data and weight matrices are normalized to [0, 1].

[0072] The value range of the state matrix of the multi-source fusion data of the wind turbine is also between [0, 1]. The larger the value, the better the state, and the smaller the value, the potential fault or abnormality.

[0073] Real-time wind turbine operating data is collected through a variety of sensors to ensure comprehensive and real-time monitoring. The collection of diverse data, including vibration, temperature, acoustics, electrical, environmental, position, and angle, comprehensively reflects the wind turbine's operating status. Preprocessing of the collected data, including data cleaning, standardization, noise reduction, and feature extraction, improves data quality and consistency. Multi-source data fusion generates a wind turbine status matrix, providing more comprehensive and accurate information on wind turbine operating status.

[0074] S2: Analyze the processed data in real time, build a detection model, detect the operating status of the wind turbine, identify wind turbine faults and handle them.

[0075] The real-time analysis of processed data, building a detection model, detecting the operating status of the wind turbine, identifying and processing wind turbine faults includes using the state matrix value of the multi-source fusion data of the wind turbine to build a detection model using the Bayesian network and random forest, and the formula is expressed as follows:

[0076]

[0077] Where D(t) represents the value of the detection model, B represents the number of decision trees in the random forest, and f b (x) represents the prediction function of the b-th decision tree, S(t) represents the state matrix of the wind turbine, N represents the number of sensors, P(S i (t)|Parents(S i (t))) represents the state conditional probability of sensor i in the Bayesian network, T represents the upper limit of integration, represents the total time, λ represents the attenuation factor, W irepresents the weight matrix of sensor i, x i (t) represents the data vector collected by sensor i at time t, t′ represents the integral variable, e -λ(T-t′) represents an exponential decay function.

[0078] The value range of D(t) is between [0, 1], where 0 indicates the normal operation of the wind turbine and 1 indicates a serious fault or abnormality.

[0079] By analyzing processed data in real time and building detection models, the system employs advanced algorithms such as Bayesian networks and random forests to efficiently and accurately monitor the operating status of wind turbines and identify potential faults. Based on the output of the detection model, faults of varying severity are classified and addressed, enabling automatic fault isolation and activation of backup systems, ensuring continuous and stable operation of the wind turbines.

[0080] S3: Use historical fault data to train a fault prediction model to predict wind turbine failures.

[0081] Training the fault prediction model with historical fault data includes: the encoder extracts low-dimensional features through linear transformation and nonlinear activation function to remove data noise.

[0082] The decoder restores the low-dimensional features to the original data and verifies the effectiveness of feature extraction by comparing the similarity between the reconstructed data and the original data.

[0083] The deep neural network uses the extracted low-dimensional features to build a fault prediction model through multiple layers of nonlinear transformations and outputs the state of the fault.

[0084] The encoder formula is expressed as:

[0085] z=σ(W e x+b e )

[0086] The decoder formula is expressed as:

[0087]

[0088] The deep neural network formula is expressed as:

[0089] h (l) =σ(W (l) h (l-1) +b (l) )(for l=1,2,…,L)

[0090] h (0) =z

[0091] y=σ(W (L) h (L-1) +b (L) )

[0092] Among them, z represents the feature vector encoded by the autoencoder, σ represents the activation function, and W e represents the weight matrix of the encoder, x represents the input data, b e represents the encoder bias term, Represents the reconstructed data after decoding, W d represents the weight matrix of the decoder, b d represents the decoder bias term, h (l) represents the activation output of the lth layer of the deep neural network, W (l) represents the weight matrix of the lth layer of the deep neural network, b (l) represents the bias term of the lth layer of the deep neural network, h (0) It represents the initial input of the feature input after autoencoder encoding to the deep neural network, and y represents the final output of the deep neural network to predict the fault value.

[0093] The activation function σ selects the Sigmoid function.

[0094] Encoded feature z and decoded data The value range is [0, 1].

[0095] The value range of the deep neural network output y is [0, 1], which represents the probability of a fault occurring. The closer it is to 1, the higher the possibility of a fault occurring.

[0096] By training a fault prediction model using historical fault data, wind turbine failures can be predicted in advance, providing early warning information and enabling maintenance personnel to take proactive measures to prevent further failures. A deep learning model using an encoder and decoder extracts low-dimensional features through linear transformations and nonlinear activation functions, removing data noise and improving the accuracy and robustness of the fault prediction model.

[0097] On the other hand, this embodiment also provides a wind turbine generator system for ensuring safety in emergency situations, which includes:

[0098] The data acquisition and processing module collects the operating data of the wind turbine in real time through sensors and processes the data.

[0099] The fault detection and processing module analyzes the processed data in real time, builds a detection model, detects the operating status of the wind turbine, identifies the wind turbine fault and handles it.

[0100] The fault prediction module uses historical fault data to train a fault prediction model to predict wind turbine faults.

[0101] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0102] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0103] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0104] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0105] In Example 2, a wind turbine was selected as the test subject and equipped with various sensors to collect real-time operating data. Experimental preparation included sensor installation, data acquisition system configuration, data processing and analysis platform development, and the collection and organization of historical fault data. Sensor types included vibration sensors, temperature sensors, acoustic sensors, electrical sensors, environmental sensors, and position and angle sensors. These sensors were installed at key locations on the wind turbine to ensure comprehensive coverage of its operating status.

[0106] First, the real-time data of the sensor is transmitted to the central data processing platform through the data acquisition system. The data acquisition system is equipped with an efficient communication protocol and data transmission mechanism to ensure timely data transmission and high-precision recording. The collected data includes the vibration data of the wind turbine (unit: m / s 2 ), temperature data (unit: °C), acoustic data (unit: dB), electrical data (unit: V), environmental data (unit: ppm), and position and angle data (unit: degrees). These data will serve as the basis for subsequent analysis.

[0107] Data processing includes data cleaning, data standardization, data noise reduction, and feature extraction. Data cleaning filters and corrects abnormal data. Data standardization converts data in different units to a common scale. Data noise reduction uses signal processing techniques to remove sensor noise and ensure data accuracy. Feature extraction uses statistical methods and machine learning algorithms to extract key features and form a multi-source fusion data state matrix for the wind turbine.

[0108] The constructed detection model uses a leaf-bayesian network and random forest algorithm to identify the operating status and potential faults of wind turbines through comprehensive analysis of real-time data. The input of the detection model is a multi-source fusion data state matrix, and the output is a fault detection value D(t), which ranges from 0 to 1. A larger D(t) value indicates a higher probability of fault.

[0109] The fault prediction model was trained using historical fault data. This data includes past fault records and operating parameters. An autoencoder was used for data dimensionality reduction and feature extraction, and a deep neural network (DNN) model was used for fault prediction. The prediction model was able to identify potential faults in advance and issue early warnings. The experimental data is shown in Table 1.

[0110] Table 1 Experimental data table

[0111]

[0112] The data in the above table shows that vibration, temperature, acoustics, electrical, environmental, and position and angle data vary significantly between wind turbines. In particular, the fault detection value D(t) effectively reflects the wind turbine's operating status and potential for failure. Based on the D(t) range, D(t) < 0.3 indicates a normal wind turbine, 0.3 ≤ D(t) < 0.7 indicates a moderate fault, and D(t) ≥ 0.7 indicates a severe fault.

[0113] From the data analysis results, the present invention collects and processes data in real time through multiple sensors, combines the Bayesian network and random forest algorithm to build a detection model, and can accurately identify the operating status and potential faults of wind turbines. For example, the D(t) values ​​of wind turbine 1 and wind turbine 2 are 0.10 and 0.20 respectively, both below 0.3, indicating that the wind turbines are in good operating condition and do not require special treatment. The D(t) values ​​of wind turbine 3 and wind turbine 4 are 0.35 and 0.50 respectively, between 0.3 and 0.7, indicating that there is a moderate fault, and it is necessary to start the enhanced monitoring mode and issue an early warning notification. The D(t) values ​​of wind turbine 5 and wind turbine 6 are 0.70 and 0.85 respectively, exceeding 0.7, indicating that there is a serious fault, and it is necessary to take isolation measures immediately and start the backup system to ensure the safe operation of the wind turbine.

[0114] Compared to existing technologies, the multi-sensor data acquisition and processing method of the present invention provides more comprehensive and accurate operating status information, improving the accuracy and timeliness of fault detection. A fault prediction model trained with historical fault data can proactively identify and warn of potential faults, reducing wind turbine downtime and maintenance costs, and improving system reliability and safety. This innovative and novel technical solution has significant application value and broad prospects for wind turbine safety.

[0115] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A safety guarantee method for sudden situations of wind turbines, characterized in that: include: Collect wind turbine operation data in real time through sensors and process the data; Analyze the processed data in real time, build detection models, detect the operating status of wind turbines, identify wind turbine faults and handle them; The fault prediction model is trained using historical fault data to predict wind turbine failures.

2. A wind turbine generator system emergency safety method as claimed in claim 1, characterized in that: The operation data of the wind turbine generator set includes vibration data, temperature data, acoustic data, electrical data, environmental data, position and angle data, historical fault data and maintenance data.

3. A wind turbine generator system emergency safety method as claimed in claim 2, characterized in that: Data processing includes preprocessing the collected data and fusing multi-source data; Preprocessing includes data cleaning, data standardization, data noise reduction and feature extraction; The formula of the state matrix of multi-source fusion data of wind turbines is expressed as: Among them, S(t) represents the state matrix of multi-source fusion data of wind turbines, N represents the number of sensors, T represents the upper limit of integration, λ represents the attenuation factor, and W i represents the weight matrix of sensor i, x i (t) represents the data vector collected by sensor i at time t, t′ represents the integral variable, e -λ(T-t′) represents an exponential decay function; All sensor data and weight matrices are normalized to [0, 1]; The value range of the state matrix of the multi-source fusion data of the wind turbine is also between [0, 1]. The larger the value, the better the state, and the smaller the value, the potential fault or abnormality.

4. A wind turbine generator system emergency safety guarantee method as claimed in claim 3, characterized in that: The real-time analysis of processed data, building a detection model, detecting the operating status of the wind turbine, identifying and processing the wind turbine fault includes building a detection model using the leaf-bayesian network and random forest through the value of the state matrix of the multi-source fusion data of the wind turbine, and the formula is expressed as follows: Where D(t) represents the value of the detection model, B represents the number of decision trees in the random forest, and f b (x) represents the prediction function of the bth decision tree, S(t) represents the state matrix of the wind turbine, N represents the number of sensors, P(S i (t)|Parents(S i (t))) represents the state conditional probability of sensor i in the Bayesian network, T represents the upper limit of integration, represents the total time, λ represents the attenuation factor, and W i represents the weight matrix of sensor i, x i (t) represents the data vector collected by sensor i at time t, t′ represents the integral variable, e -λ(T-t′) represents an exponential decay function; The value range of D(t) is between [0, 1], where 0 indicates the normal operation of the wind turbine and 1 indicates a serious fault or abnormality.

5. A wind turbine generator system emergency safety guarantee method as claimed in claim 4, characterized in that: The historical fault data training fault prediction model includes: the encoder extracts low-dimensional features through linear transformation and nonlinear activation function to remove data noise; The decoder restores the low-dimensional features to the original data and verifies the effectiveness of feature extraction by comparing the reconstructed data with the original data. The deep neural network uses the extracted low-dimensional features to build a fault prediction model through multiple layers of nonlinear transformations and outputs the state of the fault.

6. A wind turbine generator system emergency safety guarantee method as claimed in claim 5, characterized in that: The historical fault data training fault prediction model also includes an encoder formula expressed as: z=σ(W e x+b e ) The decoder formula is expressed as: The deep neural network formula is expressed as: h (l) =σ(W (l) h (l-1) +b (l) ) (for l=1,2,…,L) h (0) =z y=σ(W (L) h (L-1) +b (L) ) Among them, z represents the feature vector encoded by the autoencoder, σ represents the activation function, and W e represents the weight matrix of the encoder, x represents the input data, b e represents the encoder bias term, represents the reconstructed data after decoding, W d represents the weight matrix of the decoder, b d represents the decoder bias term, h (l) represents the activation output of the lth layer of the deep neural network, W (l) represents the weight matrix of the lth layer of the deep neural network, b (l) represents the bias term of the lth layer of the deep neural network, h (0) represents the initial input of the feature encoded by the autoencoder to the deep neural network, and y represents the final output of the deep neural network to predict the fault value; The activation function σ selects the Sigmoid function; Encode feature z and decode data The value range of is [0, 1]; The value range of the deep neural network output y is [0, 1], which represents the probability of a fault occurring. The closer it is to 1, the higher the possibility of a fault occurring.

7. A wind turbine generator system emergency safety guarantee method as claimed in claim 4, characterized in that: The identifying and processing of wind turbine faults includes: When D(t)<0.3, it is a minor fault, and the fault has little impact on the operation of the wind turbine; The system will perform the following operations: The system will continue to operate, but will start enhanced monitoring mode, recording data from all relevant sensors in real time to ensure that the fault characteristics are fully captured; The system will issue an early warning notification to remind operation and maintenance personnel to pay attention, but will not perform automatic isolation. When 0.3≤D(t)<0.7, it is a moderate fault, in which case the fault affects some functions of the wind turbine; The system will perform the following operations: The system will automatically isolate the fault area and stop the operation of related components to prevent the fault from spreading and affecting the normal operation of other components; Start the backup system to take over the isolated part to ensure the normal operation of the wind turbine as a whole; After receiving the notification, the operation and maintenance personnel will analyze the cause of the fault remotely or on-site, diagnose it based on the system's real-time monitoring data and historical data, and formulate a detailed maintenance plan; When D(t)≥0.7, it is a serious fault, and the fault has a significant impact on the safe operation of the wind turbine; The system will perform the following operations: The system will immediately and comprehensively isolate the fault area and stop the operation of all related components to prevent the fault from spreading further; Emergency start-up of the backup system to fully take over all functions of the isolated part to ensure the safe operation of the wind turbine as a whole; The system will issue an emergency notification to inform the operation and maintenance personnel and management and initiate the emergency response mechanism; Operation and maintenance personnel must immediately rush to the site to conduct a comprehensive inspection and maintenance, and quickly locate and troubleshoot faults based on the real-time and historical data provided by the system.

8. A wind turbine generator system safety assurance method system using any one of the methods described in claims 1 to 7, characterized in that: The data acquisition and processing module collects the operating data of the wind turbine in real time through sensors and processes the data; Fault detection and processing module, which analyzes the processed data in real time, builds a detection model, detects the operating status of wind turbines, identifies wind turbine faults and processes them; The fault prediction module uses historical fault data to train the fault prediction model and predict the fault of the wind turbine.

9. A computer device comprising: Memory and processor; The memory stores a computer program, characterized in that when the processor executes the computer program, steps of a safety guarantee method for sudden situations of a wind turbine are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, steps of a method for ensuring safety of sudden situations of a wind turbine generator set are implemented.

Images