Wind turbine generator fault identification early warning method and system

Abstract

The invention belongs to the technical field of fault identification and early warning, and discloses a wind turbine generator fault identification and early warning method and system. The method comprises the following steps: constructing an artificial intelligence model based on a cloud data center; deploying the artificial intelligence model to all wind turbine generator monitoring devices based on a cloud data center; acquiring real-time monitoring data of the wind turbine generator based on a wind turbine generator monitoring device; based on a wind turbine generator monitoring device, performing fault identification according to real-time monitoring data after data processing; based on a wind turbine generator monitoring device, performing fault early warning according to a real-time fault identification result; uploading the data to a cloud data center based on a wind turbine generator monitoring device; and generating a report based on the cloud data center according to the real-time monitoring data after data processing, the real-time fault identification result and the real-time fault early warning strategy. According to the invention, the problems of insufficient real-time performance, low accuracy and low intelligent degree in the prior art are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of fault identification and early warning, and in particular relates to a method and system for wind turbine fault identification and early warning. Background Art

[0002] With the development and popularization of clean energy, more and more wind power plants are put into use. Wind turbines are equipment that uses wind energy to generate electricity. They can effectively convert wind energy into electrical energy for users or transmit it to the power grid. The normal operation of wind turbines is the premise of the stability and reliability of wind power plants. During the operation of wind turbines, failures may occur due to various reasons, affecting the normal operation and life of wind turbines. Therefore, it is necessary to monitor wind turbines online to detect faults in wind turbines in a timely manner and send out early warning signals to remind staff to carry out maintenance.

[0003] The existing wind turbine fault identification and early warning technology has the following defects:

[0004] 1) Lack of real-time performance: Traditional fault identification methods often rely on periodic manual inspections or regular data analysis, which cannot achieve real-time monitoring and early warning, resulting in delayed fault discovery and processing, which may cause greater losses;

[0005] 2) Low accuracy: Existing technologies mostly use simple threshold judgment or rule-based methods to identify faults. These methods are difficult to deal with complex and changeable fault modes and are prone to false positives or false negatives.

[0006] 3) Low level of intelligence: The existing technology only realizes fault identification and early warning, but lacks the function of intelligent generation of fault reports. It requires staff to manually organize fault reports, which cannot meet the needs. Summary of the invention

[0007] In order to solve the problems of insufficient real-time performance, low accuracy and low intelligence in the prior art, the present invention aims to provide a wind turbine fault identification and warning method and system.

[0008] The technical solution adopted by the present invention is:

[0009] A wind turbine fault identification and early warning method comprises the following steps:

[0010] Based on the cloud data center, using artificial intelligence algorithms, we build wind turbine fault identification models, wind turbine fault warning models, and wind turbine fault report generation models;

[0011] Based on the cloud data center, the wind turbine fault identification model and the wind turbine fault warning model are deployed to all wind turbine monitoring devices connected to the cloud data center;

[0012] Based on the wind turbine monitoring device, real-time monitoring data of the wind turbine is collected, and the real-time monitoring data is processed to obtain the real-time monitoring data after data processing;

[0013] Based on the wind turbine monitoring device, according to the real-time monitoring data after data processing, the wind turbine fault identification model is used to perform fault identification and obtain real-time fault identification results;

[0014] Based on the wind turbine monitoring device, according to the real-time fault identification results, the wind turbine fault warning model is used to perform fault warning, obtain and respond to the real-time fault warning strategy;

[0015] Based on the wind turbine monitoring device, the real-time monitoring data, real-time fault identification results and real-time fault warning strategies are uploaded to the cloud data center after data processing;

[0016] Based on the cloud data center, according to the real-time monitoring data after data processing, the real-time fault identification results and the real-time fault warning strategy, the wind turbine fault report generation model is used to generate the report and obtain the real-time wind turbine fault report.

[0017] Furthermore, based on the cloud data center, using artificial intelligence algorithms, a wind turbine fault identification model, a wind turbine fault warning model, and a wind turbine fault report generation model are constructed, including the following steps:

[0018] Based on the cloud data center, historical monitoring data of several wind turbines are collected, and data processing is performed on the several historical monitoring data to obtain several processed historical monitoring data;

[0019] Based on some processed historical monitoring data, a wind turbine fault identification model is constructed using deep learning and image recognition algorithms, and some historical fault identification results are generated;

[0020] Based on several historical fault identification results, a wind turbine fault warning model is constructed using a reinforcement learning algorithm, and several historical fault warning strategies are generated;

[0021] Based on several processed historical monitoring data, several historical fault identification results and several historical fault warning strategies, a wind turbine fault report generation model is constructed using a deep learning algorithm.

[0022] Furthermore, the historical monitoring data includes historical sequence monitoring data and historical image monitoring data; the historical sequence monitoring data includes historical wind deviation monitoring data and historical unit operation monitoring data;

[0023] Real-time monitoring data includes real-time sequence monitoring data and real-time image monitoring data; real-time sequence monitoring data includes real-time wind deviation monitoring data and real-time unit operation monitoring data.

[0024] Furthermore, the wind turbine fault recognition model is constructed based on the FPN-LSTM-Attention-Elman algorithm, and the wind turbine fault recognition model includes an image feature extraction module constructed based on the FPN algorithm, a sequence feature extraction module constructed based on the LSTM algorithm, an attention weight module constructed based on the Attention mechanism, and a generator fault recognition module constructed based on the Elman algorithm;

[0025] The wind turbine fault warning model is built based on the MOPPO algorithm, and the wind turbine fault warning model is equipped with an objective function set, an experience replay pool, an Actor network, a Critic network, and an intelligent agent;

[0026] The wind turbine fault report generation model is constructed based on the cGAN-MLP algorithm, and the wind turbine fault report generation model includes a generator and a discriminator both constructed based on the RNN algorithm, and a conditional embedding module and a conditional processing module both constructed based on the MLP algorithm. The generator is connected to the conditional embedding module and the discriminator respectively, and the discriminator is connected to the conditional processing module.

[0027] Furthermore, based on the historical monitoring data after some data processing, a wind turbine fault recognition model is constructed using deep learning and image recognition algorithms, and several historical fault recognition results are generated, including the following steps:

[0028] The FPN-LSTM-Attention-Elman algorithm is used to build an initial wind turbine fault identification model, and the historical monitoring data after some data processing is divided into a model training set and a model test set;

[0029] According to the model training set, the initial wind turbine fault identification model is optimized and trained to obtain an optimized wind turbine fault identification model, and several historical fault identification results and corresponding historical weighted fusion features are generated;

[0030] According to the model test set, the optimized wind turbine fault identification model is tested to obtain the model test accuracy. If the model test accuracy is greater than the accuracy threshold, the final wind turbine fault identification model is output, otherwise, the optimization training continues.

[0031] Furthermore, based on several historical fault identification results, a reinforcement learning algorithm is used to construct a wind turbine fault warning model and generate several historical fault warning strategies, including the following steps:

[0032] Taking the fault warning strategy generation problem as the simulation environment, the MOPPO algorithm is used to build the initial wind turbine fault warning model, and the objective function set, experience replay pool, Actor network, Critic network and intelligent agent are set for the initial wind turbine fault warning model.

[0033] Analyze the historical fault recognition results to obtain several historical fault recognition states, define the state space of the intelligent agent based on the several historical fault recognition states, and define the action space of the intelligent agent based on several preset fault warning decision actions;

[0034] Based on any objective function in the objective function set and according to a number of historical fault identification results, the initial wind turbine fault warning model is pre-trained to obtain a pre-trained wind turbine fault warning model, and a number of historical fault warning strategies and corresponding historical fault warning experiences are generated;

[0035] Use the critic network of the pre-trained wind turbine fault warning model to obtain the rewards of several historical wind turbine fault warning strategies and obtain the optimized critic network;

[0036] According to several rewards, the Actor network of the pre-trained wind turbine fault warning model is optimized to obtain an optimized Actor network;

[0037] Traverse all objective functions in the objective function set and repeat the above steps to obtain the final Actor network and the final Critic network;

[0038] According to the final Actor network and the final Critic network, the final wind turbine fault warning model is obtained, and several historical fault warning experiences are stored in the experience replay pool.

[0039] Furthermore, according to a number of historical monitoring data after data processing, a number of historical fault identification results and a number of historical fault warning strategies, a deep learning algorithm is used to construct a wind turbine fault report generation model, including the following steps:

[0040] Using the cGAN-MLP algorithm, an initial wind turbine fault report generation model is constructed; the initial wind turbine fault report generation model includes an initial generator, an initial discriminator, a conditional embedding module, and a conditional processing module;

[0041] Set up a real wind turbine fault report for each processed historical monitoring data, the corresponding historical fault identification result and the corresponding historical fault warning strategy;

[0042] Combining a first loss function of an initial generator and a second loss function of an initial discriminator of an initial wind turbine fault report generation model to obtain a comprehensive loss function;

[0043] Extracting the historical weighted fusion features of the historical monitoring data after each data processing, the historical fault identification result features of the corresponding historical fault identification results, and the historical fault warning strategy features of the corresponding historical fault warning strategy;

[0044] Using the conditional information embedder of the initial wind turbine fault report generation model, conditionally embed each historical weighted fusion feature, historical fault identification result feature, historical fault warning strategy feature, and random noise to obtain several historical conditional information embedding features;

[0045] According to several historical condition information embedding features, an initial generator of an initial wind turbine fault report generation model is trained to obtain an optimized generator, and several wind turbine fault reports are generated;

[0046] Using a condition information processor, condition information processing is performed on the historical condition information embedding feature, the corresponding real wind turbine fault report, and the corresponding generated wind turbine fault report to obtain the historical condition information;

[0047] According to the real wind turbine fault report, the corresponding generated wind turbine fault report and the corresponding historical condition information, the initial discriminator is trained to obtain the optimized discriminator, and several historical data discrimination results are generated;

[0048] According to each generated wind turbine fault report and the corresponding historical data judgment result, a comprehensive loss function is used to obtain the historical comprehensive loss value in the training process;

[0049] If the historical loss value is lower than the loss value threshold, the final generator and the final discriminator are output, otherwise, the optimization training continues;

[0050] The final generator and the final discriminator are integrated to obtain the final wind turbine fault report generation model.

[0051] Further, based on the wind turbine monitoring device, according to the real-time monitoring data after data processing, the wind turbine fault identification model is used to perform fault identification to obtain a real-time fault identification result, including the following steps:

[0052] Based on the wind turbine monitoring device, the real-time monitoring data after data processing is analyzed to obtain the real-time sequence monitoring data after data processing and the real-time image monitoring data after data processing;

[0053] Use the image feature extraction module of the wind turbine fault recognition model to extract the real-time image features of the real-time image monitoring data after data processing;

[0054] Use the sequence feature extraction module of the wind turbine fault identification model to extract the real-time sequence features of the real-time sequence monitoring data after data processing;

[0055] According to the preset attention weight, the attention weight module of the wind turbine fault recognition model is used to perform weighted fusion on the real-time image data features and the real-time sequence data features to obtain the real-time weighted fusion features;

[0056] According to the real-time weighted fusion features, the generator fault identification module of the wind turbine fault identification model is used to perform fault identification and obtain real-time fault identification results.

[0057] Furthermore, based on the wind turbine monitoring device, according to the real-time fault identification result, the wind turbine fault warning model is used to perform fault warning, obtain and respond to the real-time fault warning strategy, including the following steps:

[0058] According to the real-time fault identification result, the most appropriate objective function is selected from the objective function set of the wind turbine fault warning model;

[0059] The real-time fault identification results are parsed to obtain a number of real-time fault identification states, and the state space of the intelligent agent is updated according to the number of real-time fault identification states to obtain an updated state space;

[0060] Extracting a number of historical fault warning experiences from the experience replay pool, and updating the action space of the agent according to a number of preset fault warning decision actions of the historical fault warning experiences, thereby obtaining an updated action space;

[0061] Based on the most appropriate objective function, the intelligent agent of the wind turbine fault warning model is used to control the Actor network, and the probability distribution of all possible fault warning decision actions in the updated action space corresponding to each real-time fault recognition state in the updated state space is generated in the updated action space;

[0062] The possible fault warning decision action with the highest probability distribution in the updated action space is used as the execution fault warning decision action of the real-time fault identification state, and the execution fault warning decision actions of all real-time fault identification states are integrated to obtain the real-time fault warning strategy;

[0063] According to the real-time fault warning strategy, a real-time fault warning signal, a first real-time fault control instruction of the wind turbine generator set and a second real-time fault control instruction of the auxiliary control device are generated;

[0064] In response to the real-time fault warning signal, a first real-time fault control instruction is sent to the corresponding wind turbine set, and a second real-time fault control instruction is sent to the corresponding auxiliary control device.

[0065] A wind turbine fault identification and early warning system is used to implement a wind turbine fault identification and early warning method. The system includes a cloud data center and a plurality of wind turbine monitoring devices. The cloud data center is communicatively connected to the plurality of wind turbine monitoring devices respectively. Each wind turbine monitoring device is arranged at a wind turbine in a power plant, and each wind turbine monitoring device is communicatively connected to a wind turbine and a corresponding auxiliary control device respectively.

[0066] The beneficial effects of the present invention are:

[0067] The present invention discloses a wind turbine fault identification and early warning method and system. Through real-time data collection, processing and model analysis, real-time monitoring and early warning of wind turbines are realized, the response time of fault discovery and processing is greatly shortened, and greater losses caused by delayed fault discovery are effectively avoided. Advanced artificial intelligence algorithms are used to construct wind turbine fault identification and wind turbine early warning models, which can more accurately identify complex and changeable fault modes, and formulate corresponding fault early warning strategies according to the fault conditions, reduce false alarms and missed alarms, improve the accuracy of fault diagnosis, and realize automatic fault early warning control. Not only automatic fault identification and early warning are realized, but also an intelligent fault report generation function is integrated, which automatically organizes and generates personalized fault reports, greatly reduces the burden on staff and improves overall work efficiency. The ability to quickly generate and execute early warning strategies is realized, so that a rapid response can be made when a fault occurs, and timely measures can be taken to prevent the fault from expanding.

[0068] Other beneficial effects of the present invention will be further described in the specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a flow chart of the wind turbine generator set fault identification and early warning method of the present invention.

[0070] Figure 2 It is a structural block diagram of the wind turbine fault identification and early warning system in the present invention. DETAILED DESCRIPTION

[0071] The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments.

[0072] Embodiment 1:

[0073] like Figure 1 As shown, this embodiment provides a wind turbine fault identification and early warning method, comprising the following steps:

[0074] S1: Based on the cloud data center, using artificial intelligence algorithms, construct a wind turbine fault identification model, a wind turbine fault warning model, and a wind turbine fault report generation model, including the following steps:

[0075] S1-1: Based on the cloud data center, historical monitoring data of several wind turbines are collected, and data processing is performed on the several historical monitoring data to obtain several processed historical monitoring data;

[0076] Historical monitoring data include historical sequence monitoring data and historical image monitoring data; historical sequence monitoring data include historical wind deviation monitoring data and historical unit operation monitoring data;

[0077] The wind deviation monitoring data include wind deviation parameters, wind deviation mean, wind deviation standard deviation and distribution model fitting parameters;

[0078] The wind deviation monitoring data is obtained through the wind deviation calculation program preset in the wind turbine monitoring device, including the following steps:

[0079] A-1: The wind turbine monitoring device collects the original monitoring parameters of the wind turbine, such as wind direction, wind speed, power and wind angle, and performs data quality processing such as data time alignment, missing value interpolation, and outlier filtering on the original parameters to obtain the monitoring parameters after quality processing;

[0080] A-2: Calculate the wind deviation of each wind turbine as follows:

[0081] 1) Calculation of wind deviation parameters:

[0082] Wind deviation parameter = |wind direction - wind angle|;

[0083] 2) Statistical calculation:

[0084] Mean wind deviation (sliding window) = sum (wind deviation value) / window size;

[0085] Standard deviation of wind deviation (sliding window) = sqrt(sum((wind deviation value - mean)^2) / (window size - 1));

[0086] 3) Distribution model fitting parameter calculation:

[0087] Normal distribution model:

[0088] Probability density function: f(x|μ,σ)=1 / sqrt(2×π×σ^2)×exp(-(x-μ)^2 / (2×σ^2));

[0089] In the formula, μ is the mean of wind deviation, σ is the standard deviation of wind deviation, and the mean and standard deviation of wind deviation are obtained through robust estimation;

[0090] t distribution model:

[0091] Probability density function: f(x|μ',σ',v)=Gamma((v+1) / 2) / (sqrt(v×π)×Gamma(v / 2)×σ'×(1+(x-μ')^2 / (v×σ'^2))^((v+1) / 2));

[0092] In the formula, μ' is the location parameter, σ' is the scale parameter, and v is the degree of freedom. The location parameter, scale parameter, and degree of freedom are estimated by methods such as maximum likelihood estimation;

[0093] For all units in the same scenario, calculate the overall mean and standard deviation of wind deviation, use robust estimation methods to reduce the impact of outliers, build probability models such as normal distribution or t distribution, and fit the overall deviation distribution;

[0094] A-3: Integrate the wind deviation parameters, wind deviation mean, wind deviation standard deviation and distribution model fitting parameters to obtain wind deviation monitoring data;

[0095] The unit operation monitoring data includes unsaturated operation monitoring data and design defect monitoring data;

[0096] Unsaturated operation monitoring data include wind turbine unit operation monitoring parameters such as wind speed, power, time, status and power;

[0097] Design defect monitoring data include wind turbine gearbox oil temperature, power, wind speed, temperature slope, power slope, kurtosis, zero-crossing rate and other unit operation monitoring parameters;

[0098] In addition to the above-mentioned data cleaning, normalization and other preprocessing, the principal component analysis algorithm is also used to reduce the data dimension, thereby improving the data quality and reducing the data dimension;

[0099] S1-2: Based on some historical monitoring data after data processing, use deep learning and image recognition algorithms to build a wind turbine fault recognition model and generate some historical fault recognition results;

[0100] The wind turbine fault recognition model is built based on Feature Pyramid Networks (FPN)-Long Short-Term Memory (LSTM)-Attention-Elman algorithm, and the wind turbine fault recognition model includes an image feature extraction module built based on the FPN algorithm, a sequence feature extraction module built based on the LSTM algorithm, an attention weight module built based on the Attention mechanism, and a generator fault recognition module built based on the Elman algorithm;

[0101] The convolutional neural network (CNN) algorithm is used to construct the basic network architecture of the graph feature extraction module. The basic network architecture includes alternately connected convolutional layers and pooling layers. Through the alternating connection of convolutional layers and pooling layers, multi-scale feature maps can be effectively extracted from the image, local features of the image can be captured, and the spatial dimension of the features can be reduced through the pooling layer, while maintaining important feature information. The outputs of all convolutional layers are horizontally connected in a top-down order to obtain a feature pyramid, which is used to extract feature maps of image data at different scales and perform multi-scale fusion to obtain high-level image features. The LSTM network realizes the memory function through the gate mechanism and can extract deep features of sequence data. The attention weight module is used to perform weighted fusion of image features and sequence features according to the preset attention weight value, strengthen the representation of key features that affect the analysis results, and improve the analysis accuracy. The motor group fault recognition module learns the deep information of the weighted fusion features and performs label prediction based on the weighted fusion features.

[0102] Based on some historical monitoring data after data processing, a wind turbine fault recognition model is constructed using deep learning and image recognition algorithms, and some historical fault recognition results are generated, including the following steps:

[0103] S1-2-1: Use the FPN-LSTM-Attention-Elman algorithm to build an initial wind turbine fault identification model, and divide the historical monitoring data after some data processing into a model training set and a model test set;

[0104] S1-2-2: According to the model training set, the initial wind turbine fault identification model is optimized and trained to obtain an optimized wind turbine fault identification model, and a number of historical fault identification results and corresponding historical weighted fusion features are generated;

[0105] S1-2-3: According to the model test set, the optimized wind turbine fault identification model is tested to obtain the model test accuracy. If the model test accuracy is greater than the accuracy threshold, the final wind turbine fault identification model is output, otherwise, the optimization training is continued;

[0106] S1-3: Based on several historical fault identification results, a reinforcement learning algorithm is used to build a wind turbine fault warning model and generate several historical fault warning strategies;

[0107] The wind turbine fault warning model is constructed based on the Multi-Objective Proximal Policy Optimization (MOPPO) algorithm, and the wind turbine fault warning model is equipped with an objective function set, an experience replay pool, an Actor network, a Critic network, and an intelligent agent.

[0108] Based on several historical fault identification results, a wind turbine fault warning model is constructed using a reinforcement learning algorithm, and several historical fault warning strategies are generated, including the following steps:

[0109] S1-3-1: Taking the fault warning strategy generation problem as the simulation environment, using the MOPPO algorithm, constructing the initial wind turbine fault warning model, and setting the objective function set, experience replay pool, Actor network, Critic network and intelligent agent for the initial wind turbine fault warning model;

[0110] The Actor network is responsible for outputting the probability distribution of the actions that should be taken in a given state. The goal is to learn an optimal strategy, that is, to maximize the long-term cumulative reward. In the continuous action space, the Actor network usually outputs a mean and an optional variance parameter to describe the probability distribution of the action. The Critic network is responsible for evaluating the value of a given state, that is, predicting the expected return that can be obtained by starting from this state and following the current strategy. It usually outputs a scalar value that represents the value of the state or the state-action value. The experience replay pool is used to store historical experience for reuse during training. The objective function set includes functions of multiple wind turbine fault warning objectives, including minimizing the fault warning cost objective, maximizing the wind turbine fault warning efficiency objective, minimizing the fault response time objective, minimizing the fault response loss objective, and maximizing the wind turbine operation reliability objective.

[0111] S1-3-2: Analyze the historical fault identification results to obtain several historical fault identification states, define the state space of the intelligent agent based on the several historical fault identification states, and define the action space of the intelligent agent based on several preset fault warning decision actions;

[0112] S1-3-3: Based on any objective function in the objective function set and according to a number of historical fault identification results, the initial wind turbine fault warning model is pre-trained to obtain a pre-trained wind turbine fault warning model, and a number of historical fault warning strategies and corresponding historical fault warning experiences are generated;

[0113] S1-3-4: Use the critic network of the pre-trained wind turbine fault warning model to obtain rewards for several historical wind turbine fault warning strategies and obtain an optimized critic network;

[0114] S1-3-5: According to a number of rewards, the Actor network of the pre-trained wind turbine fault warning model is optimized to obtain an optimized Actor network;

[0115] S1-3-6: Traverse all objective functions in the objective function set and repeat the above steps to obtain the final Actor network and the final Critic network;

[0116] S1-3-7: Based on the final Actor network and the final Critic network, the final wind turbine fault warning model is obtained, and several historical fault warning experiences are stored in the experience playback pool;

[0117] S1-4: Based on some historical monitoring data after data processing, some historical fault identification results and some historical fault warning strategies, a deep learning algorithm is used to build a wind turbine fault report generation model;

[0118] The wind turbine fault report generation model is constructed based on the conditional generative adversarial network (cGAN)-multilayer perceptron (MLP) algorithm, and the wind turbine fault report generation model includes a generator and a discriminator both constructed based on the recurrent neural network (RNN) algorithm, and a conditional embedding module and a conditional processing module both constructed based on the MLP algorithm. The generator is connected to the conditional embedding module and the discriminator respectively, and the discriminator is connected to the conditional processing module.

[0119] The conditional embedding module receives weighted fusion features, fault identification result features, fault warning strategy features and random noise, and uses a multi-layer perceptron (MLP) to convert the features into an embedding vector that the generator can understand. The embedding vector serves as conditional information to guide the generation process of the generator. The MLP algorithm can effectively handle nonlinear relationships and convert features into useful embedding representations. The conditional embedding module ensures that the generator can take into account the specific decision context when generating outputs, thereby improving the relevance and accuracy of the generation process. The generator receives the embedding vector from the conditional embedding module and other relevant information, and uses a recurrent neural network (RNN) algorithm to generate output that matches the input conditions, i.e., generates a report. The generator attempts to generate data that is real enough to deceive the discriminator. The generator can generate customized outputs based on decision information, thereby improving flexibility and accuracy. Adaptability; The discriminator receives the output from the generator and the output of the conditional processing module, and uses the RNN algorithm to determine whether the data generated by the generator is realistic enough, that is, whether it conforms to the actual situation of the real report. The discriminator guides the training process of the generator through feedback signals. The discriminator improves the quality and authenticity of the data generated by the generator through adversarial training. The adversarial process helps to improve the overall performance of the execution model and the consistency of the output; The conditional processing module receives part of the output from the generator and the output of the conditional embedding module, and uses the MLP algorithm to process this information to provide additional conditional information for the discriminator to help the discriminator better understand the context of the generator output. The conditional processing module enhances the model's ability to evaluate the generator output and improves the reliability of the entire model. By combining conditional information, the matching degree between the generator output and the actual application scenario is improved;

[0120] Based on some historical monitoring data after data processing, some historical fault identification results and some historical fault warning strategies, a wind turbine fault report generation model is constructed using a deep learning algorithm, including the following steps:

[0121] S1-4-1: Use the cGAN-MLP algorithm to build an initial wind turbine fault report generation model; the initial wind turbine fault report generation model includes an initial generator, an initial discriminator, a conditional embedding module, and a conditional processing module;

[0122] S1-4-2: Setting a real wind turbine fault report for each processed historical monitoring data, the corresponding historical fault identification result and the corresponding historical fault warning strategy;

[0123] S1-4-3: combining the first loss function of the initial generator and the second loss function of the initial discriminator of the initial wind turbine fault report generation model to obtain a comprehensive loss function;

[0124] S1-4-4: extracting the historical weighted fusion features of the historical monitoring data after each data processing, the historical fault identification result features of the corresponding historical fault identification results, and the historical fault warning strategy features of the corresponding historical fault warning strategy;

[0125] S1-4-5: Using the conditional information embedder of the initial wind turbine fault report generation model, conditionally embed each historical weighted fusion feature, historical fault identification result feature, historical fault warning strategy feature and random noise to obtain a number of historical conditional information embedding features;

[0126] S1-4-6: according to a number of historical condition information embedding features, an initial generator of an initial wind turbine fault report generation model is trained to obtain an optimized generator, and a number of generated wind turbine fault reports are generated;

[0127] S1-4-7: using a condition information processor, performing condition information processing on the historical condition information embedding feature, the corresponding real wind turbine fault report, and the corresponding generated wind turbine fault report to obtain the historical condition information;

[0128] S1-4-8: According to the real wind turbine fault report, the corresponding generated wind turbine fault report and the corresponding historical condition information, the initial discriminator is trained to obtain an optimized discriminator, and a number of historical data discrimination results are generated;

[0129] S1-4-9: Based on each generated wind turbine fault report and the corresponding historical data judgment result, a comprehensive loss function is used to obtain the historical comprehensive loss value during the training process;

[0130] S1-4-10: If the historical loss value is lower than the loss value threshold, the final generator and the final discriminator are output, otherwise, the optimization training continues;

[0131] S1-4-11: Integrate the final generator and the final discriminator to obtain the final wind turbine fault report generation model;

[0132] S2: Based on the cloud data center, the wind turbine fault identification model and the wind turbine fault warning model are deployed to all wind turbine monitoring devices connected to the cloud data center;

[0133] S3: Based on the wind turbine monitoring device, real-time monitoring data of the wind turbine is collected, and the real-time monitoring data is processed to obtain real-time monitoring data after data processing;

[0134] Real-time monitoring data includes real-time sequence monitoring data and real-time image monitoring data; real-time sequence monitoring data includes real-time wind deviation monitoring data and real-time unit operation monitoring data;

[0135] S4: Based on the wind turbine monitoring device, according to the real-time monitoring data after data processing, using the wind turbine fault identification model, fault identification is performed to obtain a real-time fault identification result, including the following steps:

[0136] S4-1: Based on the wind turbine monitoring device, the real-time monitoring data after data processing is analyzed to obtain the real-time sequence monitoring data after data processing and the real-time image monitoring data after data processing;

[0137] S4-2: Using the image feature extraction module of the wind turbine fault recognition model, extract the real-time image features of the real-time image monitoring data after data processing;

[0138] S4-3: Use the sequence feature extraction module of the wind turbine fault identification model to extract the real-time sequence features of the real-time sequence monitoring data after data processing;

[0139] S4-4: according to the preset attention weight, using the attention weight module of the wind turbine fault recognition model, weighted fusion is performed on the real-time image data features and the real-time sequence data features to obtain the real-time weighted fusion features;

[0140] S4-5: According to the real-time weighted fusion features, the generator fault identification module of the wind turbine fault identification model is used to perform fault identification to obtain a real-time fault identification result;

[0141] Fault identification results include wind deviation fault identification results, unsaturated operation fault identification results, and design defect fault identification results;

[0142] S5: Based on the wind turbine monitoring device, according to the real-time fault identification result, using the wind turbine fault warning model, fault warning is performed, and a real-time fault warning strategy is obtained and responded to, including the following steps:

[0143] S5-1: According to the real-time fault identification result, the most appropriate objective function is selected from the objective function set of the wind turbine fault warning model;

[0144] S5-2: parsing the real-time fault identification result to obtain a number of real-time fault identification states, and updating the state space of the agent according to the number of real-time fault identification states to obtain an updated state space;

[0145] S5-3: extracting a number of historical fault warning experiences from the experience replay pool, and updating the action space of the agent according to a number of preset fault warning decision actions of the historical fault warning experiences to obtain an updated action space;

[0146] S5-4: Based on the most appropriate objective function, the agent of the wind turbine fault warning model is used to control the Actor network, and in the updated action space, the probability distribution of all possible fault warning decision actions in the updated action space corresponding to each real-time fault identification state in the updated state space is generated;

[0147] S5-5: taking the possible fault warning decision action with the highest probability distribution in the updated action space as the execution fault warning decision action of the real-time fault identification state, integrating the execution fault warning decision actions of all real-time fault identification states, and obtaining the real-time fault warning strategy;

[0148] S5-6: generating a real-time fault warning signal, a first real-time fault control instruction for the wind turbine generator set, and a second real-time fault control instruction for the auxiliary control device according to the real-time fault warning strategy;

[0149] Fault warning signals include wind deviation fault warning signals, unsaturated operation fault warning signals, and design defect fault warning signals;

[0150] S5-7: In response to the real-time fault warning signal, a first real-time fault control instruction is sent to the corresponding wind turbine set, and a second real-time fault control instruction is sent to the corresponding auxiliary control device;

[0151] S6: Based on the wind turbine monitoring device, the real-time monitoring data, real-time fault identification results and real-time fault warning strategies are uploaded to the cloud data center after data processing;

[0152] S7: Based on the cloud data center, according to the real-time monitoring data after data processing, the real-time fault identification results and the real-time fault warning strategy, the wind turbine fault report generation model is used to generate a report to obtain a real-time wind turbine fault report, including the following steps:

[0153] S7-1: Based on the cloud data center, extract the real-time weighted fusion features of the real-time monitoring data after data processing, the real-time fault identification result features of the corresponding real-time fault identification results, and the real-time fault warning strategy features of the corresponding real-time fault warning strategy;

[0154] S7-2: Using the conditional information embedder of the wind turbine fault report generation model, conditionally embed the real-time weighted fusion feature, the real-time fault identification result feature, and the real-time fault warning strategy feature to obtain the real-time conditional information embedding feature;

[0155] S7-3: Use the generator of the wind turbine fault report generation model to generate a report based on the real-time condition information embedded features to obtain a real-time wind turbine fault report.

[0156] Embodiment 2:

[0157] like Figure 2 As shown, this embodiment provides a wind turbine fault identification and early warning system for implementing a wind turbine fault identification and early warning method. The system includes a cloud data center and a plurality of wind turbine monitoring devices. The cloud data center is respectively connected to the plurality of wind turbine monitoring devices for communication. Each wind turbine monitoring device is arranged at a wind turbine in a power plant, and each wind turbine monitoring device is respectively connected to a wind turbine and a corresponding auxiliary control device for communication.

[0158] A cloud data center is used to use artificial intelligence algorithms to build a wind turbine fault identification model, a wind turbine fault warning model, and a wind turbine fault report generation model; and deploy the wind turbine fault identification model and the wind turbine fault warning model to all wind turbine monitoring devices connected to the cloud data center;

[0159] The wind turbine monitoring device is used to collect real-time monitoring data of the wind turbine, and process the real-time monitoring data to obtain real-time monitoring data after data processing; based on the real-time monitoring data after data processing, use the wind turbine fault identification model to perform fault identification to obtain real-time fault identification results; based on the real-time fault identification results, use the wind turbine fault warning model to perform fault warning, obtain and respond to the real-time fault warning strategy; upload the real-time monitoring data after data processing, the real-time fault identification results and the real-time fault warning strategy to the cloud data center; based on the real-time monitoring data after data processing, the real-time fault identification results and the real-time fault warning strategy, use the wind turbine fault report generation model to generate a report to obtain a real-time wind turbine fault report.

[0160] The present invention discloses a wind turbine fault identification and early warning method and system. Through real-time data collection, processing and model analysis, real-time monitoring and early warning of wind turbines are realized, the response time of fault discovery and processing is greatly shortened, and greater losses caused by delayed fault discovery are effectively avoided. Advanced artificial intelligence algorithms are used to construct wind turbine fault identification and wind turbine early warning models, which can more accurately identify complex and changeable fault modes, and formulate corresponding fault early warning strategies according to the fault conditions, reduce false alarms and missed alarms, improve the accuracy of fault diagnosis, and realize automatic fault early warning control. Not only automatic fault identification and early warning are realized, but also an intelligent fault report generation function is integrated, which automatically organizes and generates personalized fault reports, greatly reduces the burden on staff and improves overall work efficiency. The ability to quickly generate and execute early warning strategies is realized, so that a rapid response can be made when a fault occurs, and timely measures can be taken to prevent the fault from expanding.

[0161] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A wind turbine fault identification and early warning method, characterized in that: The steps include: Based on the cloud data center, using artificial intelligence algorithms, we build wind turbine fault identification models, wind turbine fault warning models, and wind turbine fault report generation models; Based on the cloud data center, the wind turbine fault identification model and the wind turbine fault warning model are deployed to all wind turbine monitoring devices connected to the cloud data center; Based on the wind turbine monitoring device, real-time monitoring data of the wind turbine is collected, and the real-time monitoring data is processed to obtain the real-time monitoring data after data processing; Based on the wind turbine monitoring device, according to the real-time monitoring data after data processing, the wind turbine fault identification model is used to perform fault identification and obtain real-time fault identification results; Based on the wind turbine monitoring device, according to the real-time fault identification results, the wind turbine fault warning model is used to perform fault warning, obtain and respond to the real-time fault warning strategy; Based on the wind turbine monitoring device, the real-time monitoring data, real-time fault identification results and real-time fault warning strategies are uploaded to the cloud data center after data processing; Based on the cloud data center, according to the real-time monitoring data after data processing, the real-time fault identification results and the real-time fault warning strategy, the wind turbine fault report generation model is used to generate the report and obtain the real-time wind turbine fault report.

2. A wind turbine fault identification and early warning method according to claim 1, characterized in that: Based on the cloud data center, using artificial intelligence algorithms, a wind turbine fault identification model, a wind turbine fault warning model, and a wind turbine fault report generation model are constructed, including the following steps: Based on the cloud data center, historical monitoring data of several wind turbines are collected, and data processing is performed on the several historical monitoring data to obtain several processed historical monitoring data; Based on some processed historical monitoring data, a wind turbine fault identification model is constructed using deep learning and image recognition algorithms, and some historical fault identification results are generated; Based on several historical fault identification results, a wind turbine fault warning model is constructed using a reinforcement learning algorithm, and several historical fault warning strategies are generated; Based on several processed historical monitoring data, several historical fault identification results and several historical fault warning strategies, a wind turbine fault report generation model is constructed using a deep learning algorithm.

3. A wind turbine fault identification and early warning method according to claim 2, characterized in that: The historical monitoring data includes historical sequence monitoring data and historical image monitoring data; the historical sequence monitoring data includes historical wind deviation monitoring data and historical unit operation monitoring data; The real-time monitoring data includes real-time sequence monitoring data and real-time image monitoring data; the real-time sequence monitoring data includes real-time wind deviation monitoring data and real-time unit operation monitoring data.

4. A wind turbine fault identification and early warning method according to claim 3, characterized in that: The wind turbine fault recognition model is constructed based on the FPN-LSTM-Attention-Elman algorithm, and the wind turbine fault recognition model includes an image feature extraction module constructed based on the FPN algorithm, a sequence feature extraction module constructed based on the LSTM algorithm, an attention weight module constructed based on the Attention mechanism, and a generator fault recognition module constructed based on the Elman algorithm; The wind turbine fault warning model is constructed based on the MOPPO algorithm, and the wind turbine fault warning model is provided with an objective function set, an experience replay pool, an Actor network, a Critic network and an intelligent agent; The wind turbine fault report generation model is constructed based on the cGAN-MLP algorithm, and the wind turbine fault report generation model includes a generator and a discriminator both constructed based on the RNN algorithm, and a conditional embedding module and a conditional processing module both constructed based on the MLP algorithm. The generator is connected to the conditional embedding module and the discriminator respectively, and the discriminator is connected to the conditional processing module.

5. A wind turbine fault identification and early warning method according to claim 4, characterized in that: Based on some historical monitoring data after data processing, a wind turbine fault recognition model is constructed using deep learning and image recognition algorithms, and some historical fault recognition results are generated, including the following steps: The FPN-LSTM-Attention-Elman algorithm is used to build an initial wind turbine fault identification model, and the historical monitoring data after some data processing is divided into a model training set and a model test set; According to the model training set, the initial wind turbine fault identification model is optimized and trained to obtain an optimized wind turbine fault identification model, and several historical fault identification results and corresponding historical weighted fusion features are generated; According to the model test set, the optimized wind turbine fault identification model is tested to obtain the model test accuracy. If the model test accuracy is greater than the accuracy threshold, the final wind turbine fault identification model is output, otherwise, the optimization training continues.

6. A wind turbine fault identification and early warning method according to claim 5, characterized in that: Based on several historical fault identification results, a wind turbine fault warning model is constructed using a reinforcement learning algorithm, and several historical fault warning strategies are generated, including the following steps: Taking the fault warning strategy generation problem as the simulation environment, the MOPPO algorithm is used to build the initial wind turbine fault warning model, and the objective function set, experience replay pool, Actor network, Critic network and intelligent agent are set for the initial wind turbine fault warning model. Analyze the historical fault recognition results to obtain several historical fault recognition states, define the state space of the intelligent agent based on the several historical fault recognition states, and define the action space of the intelligent agent based on several preset fault warning decision actions; Based on any objective function in the objective function set and according to a number of historical fault identification results, the initial wind turbine fault warning model is pre-trained to obtain a pre-trained wind turbine fault warning model, and a number of historical fault warning strategies and corresponding historical fault warning experiences are generated; Use the critic network of the pre-trained wind turbine fault warning model to obtain the rewards of several historical wind turbine fault warning strategies and obtain the optimized critic network; According to several rewards, the Actor network of the pre-trained wind turbine fault warning model is optimized to obtain an optimized Actor network; Traverse all objective functions in the objective function set and repeat the above steps to obtain the final Actor network and the final Critic network; According to the final Actor network and the final Critic network, the final wind turbine fault warning model is obtained, and several historical fault warning experiences are stored in the experience replay pool.

7. A wind turbine fault identification and early warning method according to claim 6, characterized in that: Based on some historical monitoring data after data processing, some historical fault identification results and some historical fault warning strategies, a wind turbine fault report generation model is constructed using a deep learning algorithm, including the following steps: Using the cGAN-MLP algorithm, an initial wind turbine fault report generation model is constructed; the initial wind turbine fault report generation model includes an initial generator, an initial discriminator, a condition embedding module and a condition processing module; Set up a real wind turbine fault report for each processed historical monitoring data, the corresponding historical fault identification result and the corresponding historical fault warning strategy; Combining a first loss function of an initial generator and a second loss function of an initial discriminator of an initial wind turbine fault report generation model to obtain a comprehensive loss function; Extracting the historical weighted fusion features of the historical monitoring data after each data processing, the historical fault identification result features of the corresponding historical fault identification results, and the historical fault warning strategy features of the corresponding historical fault warning strategy; Using the conditional information embedder of the initial wind turbine fault report generation model, conditionally embed each historical weighted fusion feature, historical fault identification result feature, historical fault warning strategy feature, and random noise to obtain several historical conditional information embedding features; According to several historical condition information embedding features, an initial generator of an initial wind turbine fault report generation model is trained to obtain an optimized generator, and several wind turbine fault reports are generated; Using a condition information processor, condition information processing is performed on the historical condition information embedding feature, the corresponding real wind turbine fault report, and the corresponding generated wind turbine fault report to obtain the historical condition information; According to the real wind turbine fault report, the corresponding generated wind turbine fault report and the corresponding historical condition information, the initial discriminator is trained to obtain the optimized discriminator, and several historical data discrimination results are generated; According to each generated wind turbine fault report and the corresponding historical data judgment result, a comprehensive loss function is used to obtain the historical comprehensive loss value in the training process; If the historical loss value is lower than the loss value threshold, the final generator and the final discriminator are output, otherwise, the optimization training continues; The final generator and the final discriminator are integrated to obtain the final wind turbine fault report generation model.

8. A wind turbine fault identification and early warning method according to claim 7, characterized in that: Based on the wind turbine monitoring device, according to the real-time monitoring data after data processing, the wind turbine fault identification model is used to perform fault identification to obtain real-time fault identification results, including the following steps: Based on the wind turbine monitoring device, the real-time monitoring data after data processing is analyzed to obtain the real-time sequence monitoring data after data processing and the real-time image monitoring data after data processing; Use the image feature extraction module of the wind turbine fault recognition model to extract the real-time image features of the real-time image monitoring data after data processing; Use the sequence feature extraction module of the wind turbine fault identification model to extract the real-time sequence features of the real-time sequence monitoring data after data processing; According to the preset attention weight, the attention weight module of the wind turbine fault recognition model is used to perform weighted fusion on the real-time image data features and the real-time sequence data features to obtain the real-time weighted fusion features; According to the real-time weighted fusion features, the generator fault identification module of the wind turbine fault identification model is used to perform fault identification and obtain real-time fault identification results.

9. A wind turbine fault identification and early warning method according to claim 8, characterized in that: Based on the wind turbine monitoring device, according to the real-time fault identification result, the wind turbine fault warning model is used to perform fault warning, obtain and respond to the real-time fault warning strategy, including the following steps: According to the real-time fault identification result, the most appropriate objective function is selected from the objective function set of the wind turbine fault warning model; The real-time fault identification result is parsed to obtain a number of real-time fault identification states, and the state space of the intelligent agent is updated according to the number of real-time fault identification states to obtain an updated state space; Extracting a number of historical fault warning experiences from the experience replay pool, and updating the action space of the agent according to a number of preset fault warning decision actions of the historical fault warning experiences, thereby obtaining an updated action space; Based on the most appropriate objective function, the intelligent agent of the wind turbine fault warning model is used to control the Actor network, and the probability distribution of all possible fault warning decision actions in the updated action space corresponding to each real-time fault recognition state in the updated state space is generated in the updated action space; The possible fault warning decision action with the highest probability distribution in the updated action space is used as the execution fault warning decision action of the real-time fault identification state, and the execution fault warning decision actions of all real-time fault identification states are integrated to obtain the real-time fault warning strategy; According to the real-time fault warning strategy, a real-time fault warning signal, a first real-time fault control instruction of the wind turbine generator set and a second real-time fault control instruction of the auxiliary control device are generated; In response to the real-time fault warning signal, a first real-time fault control instruction is sent to the corresponding wind turbine set, and a second real-time fault control instruction is sent to the corresponding auxiliary control device.

10. A wind turbine fault identification and early warning system, used to implement the wind turbine fault identification and early warning method according to any one of claims 1 to 9, characterized in that: The system includes a cloud data center and several wind turbine monitoring devices. The cloud data center is communicatively connected to the several wind turbine monitoring devices respectively. Each of the wind turbine monitoring devices is arranged at a wind turbine in a power plant, and each wind turbine monitoring device is communicatively connected to a wind turbine and a corresponding auxiliary control device respectively.

Images