Fault judgment method and device for wind turbine generator and storage medium

By building a digital twin model of wind turbine units and training fault judgment model, and using real-time data to make fault judgments, the problem of low efficiency of traditional wind turbine units is solved, and efficient fault warning and the effect of reducing maintenance costs is achieved.

CN118885751APending Publication Date: 2024-11-01STATE POWER INVESTMENT CORPORATION RESEARCH INSTITUTE
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410972769.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-19
Publication Date
2024-11-01

AI Technical Summary

Technical Problem

The fault diagnosis method of traditional wind turbines has low fault judgment efficiency and low fault prediction accuracy, which cannot meet the development needs of modern wind farms, resulting in potential downtime, economic losses and unstable operation.

Method used

By building a digital twin model of the wind turbine, obtaining historical data for processing, training the target fault judgment model, and using real-time data for fault judgment, real-time monitoring and fault warning of the wind turbine status are achieved.

Benefits of technology

It improves the efficiency of fault judgment and prediction accuracy, avoids the expansion of faults or causes greater losses, improves the operating reliability and stability of wind turbines, and reduces maintenance costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118885751A_ABST
    Figure CN118885751A_ABST
Patent Text Reader

Abstract

The invention discloses a fault judgment method and device for a wind turbine generator and a storage medium, and the method comprises the steps: carrying out the three-dimensional modeling of the wind turbine generator, and constructing a digital twinborn model of the wind turbine generator; acquiring historical data of the wind turbine generator, and performing data processing on the historical data to obtain a first data set; obtaining a second data set based on the first data set and the digital twin model; training the initial fault judgment model based on the second data set to obtain a target fault judgment model; and collecting real-time data of the wind turbine generator, and performing fault judgment on the real-time data through the target fault judgment model to obtain a fault judgment result of the wind turbine generator. Therefore, fault judgment is performed on the real-time data of the wind turbine generator through the target fault judgment model, manual regular inspection can be replaced, real-time monitoring of the state of the wind turbine generator is realized, the fault judgment efficiency and the fault prediction accuracy are improved, fault expansion or greater loss is avoided, and the fault prediction accuracy is improved. And the operation reliability and stability of the wind turbine generator are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of fault judgment, and in particular to a fault judgment method, device and storage medium for a wind turbine generator set. Background Art

[0002] With the transformation of the global energy structure and the development of green energy, wind power generation, as a clean and renewable energy form, has received widespread attention and application. At present, the domestic wind power industry is developing rapidly, and the number of wind turbines has increased dramatically. As an important equipment for wind power generation, whether wind turbines work normally is closely related to daily life. Among them, timely fault identification of large rotating equipment on wind turbines plays a vital role in ensuring the normal operation of wind turbines. If a wind turbine fails and cannot be identified and repaired in time, it may cause downtime, which will affect the power generation, not only causing the loss of economic benefits, but also affecting people's normal life. Based on this, it is necessary to diagnose the fault of wind turbines in order to monitor and repair them in real time, ensure the safe and stable operation of wind turbines, and avoid failures. The traditional fault diagnosis method of wind turbines has the problems of low fault judgment efficiency and low fault prediction accuracy, which cannot meet the development needs of modern wind farms. Summary of the invention

[0003] The present invention aims to solve one of the technical problems in the related art at least to a certain extent.

[0004] To this end, the present invention proposes a fault judgment method for a wind turbine set, which performs fault judgment on the real-time data of the wind turbine set through a target fault judgment model, can replace manual periodic inspections, realize real-time monitoring of the status of the wind turbine set, and promptly issue an early warning when a fault occurs, thereby improving the fault judgment efficiency and the fault prediction accuracy, avoiding the expansion of faults or causing greater losses, improving the operating reliability and stability of the wind turbine set, and reducing maintenance costs.

[0005] Another object of the present invention is to provide a fault diagnosis device for a wind turbine generator set.

[0006] To achieve the above object, the present invention provides a fault diagnosis method for a wind turbine generator set, the method comprising:

[0007] Performing three-dimensional modeling on the wind turbine generator set and constructing a digital twin model of the wind turbine generator set;

[0008] Acquiring historical data of the wind turbine generator set, and performing data processing on the historical data to obtain a first data set;

[0009] Based on the first data set and the digital twin model, obtain a second data set;

[0010] Training an initial fault judgment model based on the second data set to obtain a target fault judgment model;

[0011] Real-time data of the wind turbine generator set is collected, and fault judgment is performed on the real-time data using the target fault judgment model to obtain a fault judgment result of the wind turbine generator set.

[0012] The fault judgment method of the wind turbine generator set according to the embodiment of the present invention may also have the following additional technical features:

[0013] In one embodiment of the present invention, the three-dimensional modeling of the wind turbine generator set to construct a digital twin model of the wind turbine generator set includes:

[0014] Obtaining wind turbine data and operation data of the wind turbine modeling;

[0015] Performing geometric modeling on the wind turbine generator set to obtain a geometric model of the wind turbine generator set;

[0016] Based on the wind turbine data, operation data and geometric model, a digital twin model of the wind turbine group is constructed.

[0017] In one embodiment of the present invention, the historical data of the wind turbine generator set includes historical vibration data and historical temperature data, and the acquiring the historical data of the wind turbine generator set and performing data processing on the historical data to obtain a first data set includes:

[0018] Processing the historical vibration data to obtain a third data set;

[0019] Processing the historical temperature data to obtain a fourth data set;

[0020] The first data set is obtained based on the third data set and the fourth data set.

[0021] In one embodiment of the present invention, obtaining a second data set based on the first data set and the digital twin model includes:

[0022] Mapping the first data set to the digital twin model, and determining a fault judgment result corresponding to each data in the first data set;

[0023] Extracting data features of the data in the first data set;

[0024] Based on the data features and the corresponding fault judgment results, a second data set is obtained.

[0025] In one embodiment of the present invention, the training of the initial fault judgment model based on the second data set to obtain the target fault judgment model includes:

[0026] Based on a preset ratio, dividing the second data set into a training data set and a test data set;

[0027] Using the training data set to train the initial fault judgment model to obtain a trained fault judgment model;

[0028] The trained fault judgment model is evaluated using the test data set, and a target risk detection model is determined based on the obtained evaluation result.

[0029] In one embodiment of the present invention, the initial fault judgment model is trained using the training data set to obtain a trained fault judgment model, including:

[0030] Inputting the training data set into the initial fault judgment model to obtain a predicted fault judgment result;

[0031] Based on the predicted fault judgment result and the actual fault result, a corresponding loss value is obtained through a loss function;

[0032] The network parameters in the initial fault judgment model are updated based on the loss value until the initial fault judgment model converges, or when the number of iterations of the network parameters reaches a preset number, a trained fault judgment model is obtained.

[0033] In one embodiment of the present invention, the method further comprises:

[0034] If the fault judgment result of the wind turbine generator set is a fault, outputting the fault type in the fault judgment result and triggering a fault alarm mechanism;

[0035] The location where the fault occurs is located based on the real-time data of the wind turbine generator set, and the location result is output.

[0036] To achieve the above object, the present invention provides a fault diagnosis device for a wind turbine generator set, the device comprising:

[0037] A modeling module, used to perform three-dimensional modeling of a wind turbine and construct a digital twin model of the wind turbine;

[0038] A first data processing module, used for acquiring historical data of the wind turbine generator set, and performing data processing on the historical data to obtain a first data set;

[0039] A second data processing module, configured to obtain a second data set based on the first data set and the digital twin model;

[0040] A training module, used for training the initial fault judgment model based on the second data set to obtain a target fault judgment model;

[0041] The fault judgment module is used to collect real-time data of the wind turbine generator set, perform fault judgment on the real-time data through the target fault judgment model, and obtain a fault judgment result of the wind turbine generator set.

[0042] The fault judgment method, device and storage medium of the wind turbine generator set of the embodiment of the present invention perform three-dimensional modeling on the wind turbine generator set to construct a digital twin model of the wind turbine generator set; obtain historical data of the wind turbine generator set, and perform data processing on the historical data to obtain a first data set; based on the first data set and the digital twin model, obtain a second data set; train the initial fault judgment model based on the second data set to obtain a target fault judgment model; collect real-time data of the wind turbine generator set, perform fault judgment on the real-time data through the target fault judgment model, and obtain a fault judgment result of the wind turbine generator set. Therefore, the present invention performs fault judgment on the real-time data of the wind turbine generator set through the target fault judgment model, which can replace manual regular inspections, realize real-time monitoring of the status of the wind turbine generator set, and issue an early warning in time when a fault occurs, thereby improving the fault judgment efficiency and fault prediction accuracy, avoiding the expansion of faults or causing greater losses, improving the operational reliability and stability of the wind turbine generator set, and reducing maintenance costs.

[0043] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0045] Figure 1 is a flow chart of a fault determination method for a wind turbine generator system according to an embodiment of the present invention;

[0046] Figure 2 4 is a structural diagram of a fault judgment device for a wind turbine generator set according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0048] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0049] The following describes a fault determination method and device for a wind turbine generator set according to an embodiment of the present invention with reference to the accompanying drawings.

[0050] Figure 1 It is a flow chart of a fault judgment method for a wind turbine generator set according to an embodiment of the present invention.

[0051] like Figure 1 As shown, the method includes:

[0052] S1, perform three-dimensional modeling of the wind turbine and build a digital twin model of the wind turbine.

[0053] In one embodiment of the present invention, the method of performing three-dimensional modeling on a wind turbine and constructing a digital twin model of the wind turbine may include the following steps:

[0054] S11, obtaining wind turbine data and operation data for wind turbine modeling.

[0055] S12, performing geometric modeling on the wind turbine generator set to obtain a geometric model of the wind turbine generator set.

[0056] S13, build a digital twin model of the wind turbine based on wind turbine data, operation data and geometric model.

[0057] In one embodiment of the present invention, the above wind turbine data may include geometric dimensions, material properties and operating conditions. In one embodiment of the present invention, the above wind turbine data may be obtained through measurement and / or CAD drawings and / or technical documents.

[0058] In one embodiment of the present invention, the operating data may include temperature, pressure and vibration, and the operating data may be operating data of the wind turbine under different working conditions. Also, in one embodiment of the present invention, the operating data of the wind turbine may be obtained by a sensor provided on the wind turbine.

[0059] Furthermore, in one embodiment of the present invention, the geometric model of the wind turbine generator set can be obtained by geometrically modeling the wind turbine generator set using 3D modeling software, wherein, in one embodiment of the present invention, the 3D modeling software includes but is not limited to SolidWorks, AutoCAD, and Blender.

[0060] And, in a real-time example of the present invention, the components and structures in the geometric model obtained above are adjusted and constructed based on the wind turbine data and operation data to obtain a digital twin model of the wind turbine. In one embodiment of the present invention, the digital twin model constructed above can simulate the behavior of the wind turbine under various operating conditions and has strong anti-interference ability.

[0061] S2, acquiring historical data of the wind turbine generator set, and processing the historical data to obtain a first data set.

[0062] In one embodiment of the present invention, the historical data may include historical temperature data and historical vibration data at each moment. Also, in a real-time example of the present invention, the historical data of the wind turbine in the past preset time can be obtained through the historical operation data of the wind turbine. The preset time can be set as needed, for example, the past 3 months.

[0063] In one embodiment of the present invention, the above-mentioned historical data of key parts of the wind turbine generator set can be obtained. In one embodiment of the present invention, the above-mentioned key parts include but are not limited to bearings, blades and transmission systems.

[0064] Furthermore, in one embodiment of the present invention, after obtaining the historical data of the wind turbine generator set through the above steps, data processing may be performed on the historical temperature data and the historical vibration data to obtain a first data set.

[0065] Specifically, in one embodiment of the present invention, the method for processing historical data to obtain a first data set may include the following steps:

[0066] S21, performing data processing on the historical vibration data to obtain a third data set.

[0067] S22, processing the historical temperature data to obtain a fourth data set.

[0068] S23, obtaining a first data set based on the third data set and the fourth data set.

[0069] In one embodiment of the present invention, the method for processing the historical vibration data to obtain the third data set may include the following steps:

[0070] Step 1: Perform wavelet transform on the historical vibration data to decompose the signal into sub-bands of different frequencies, and reconstruct the signal after removing the high-frequency noise to obtain a first vibration signal;

[0071] Step 2: Perform median filtering on the first vibration signal to remove impulse noise and abnormal values ​​to obtain a second vibration signal;

[0072] Step 3: Perform average processing on the continuous second vibration signals, smooth the second vibration signals and remove high-frequency noise to obtain a third data set.

[0073] Further, in one embodiment of the present invention, the method for processing the historical temperature data to obtain the fourth data set may include the following steps:

[0074] Step a: Perform moving average processing on the historical temperature data to smooth the historical temperature data and remove high-frequency noise to obtain first temperature data.

[0075] Step b: smoothing the first temperature data by Gaussian filtering to remove Gaussian noise and obtain second temperature data.

[0076] Step c: removing abnormal values ​​in the second temperature data to obtain a fourth data set.

[0077] In one embodiment of the present invention, outliers in the second temperature data may be detected and removed by a statistical method or a local outlier factor detection method (LOF) to obtain a fourth data set.

[0078] Furthermore, in one embodiment of the present invention, after obtaining the third data set and the fourth data set through the above steps, the historical temperature data in the third data set and the historical vibration data in the fourth data set at each moment can be determined as the first data set. If the historical temperature data in the third data set and the historical vibration data in the fourth data set cannot correspond, the corresponding data at that moment is removed to remove the abnormal data in advance and ensure the accuracy of the subsequent fault judgment results. For example, if there is corresponding historical temperature data in the third data set at a certain moment, and there is no corresponding historical vibration data in the fourth data set, the historical temperature data at that moment is removed.

[0079] S3, based on the first data set and the digital twin model, obtain a second data set.

[0080] In one embodiment of the present invention, after obtaining the first data set and the digital twin model through the above steps, a second data set can be obtained based on the first data set and the digital twin model, so that the initial fault judgment model can be subsequently trained through the second data set.

[0081] Specifically, in one embodiment of the present invention, the method for obtaining the second data set based on the first data set and the digital twin model may include the following steps:

[0082] S31, mapping the first data set to the digital twin model, and determining the fault judgment result corresponding to each data in the first data set.

[0083] S32, extracting data features of the data in the first data set.

[0084] S33, obtaining a second data set based on the data features and the corresponding fault judgment results.

[0085] In one embodiment of the present invention, the temperature data and vibration data corresponding to each moment of the first data set can be mapped to the digital twin model, so as to determine the fault judgment result corresponding to each data in the first data set through artificial experience, so as to mark the data at each moment in the first data set. In one embodiment of the present invention, the fault judgment result can include whether a fault occurs, and if it is determined that a fault occurs, the fault judgment result can also include the fault type.

[0086] And, in one embodiment of the present invention, the first data feature of the vibration data and the second data feature of the temperature data in the first data set can be extracted. Specifically, in one embodiment of the present invention, the amplitude, frequency, phase value, mean, variance, and peak value of the vibration data can be extracted as the first data feature through Fourier transform or wavelet transform. And, in one embodiment of the present invention, the temperature change rate and abnormal temperature points extracted from the temperature data can be used as the second data feature by analyzing the temperature change trend and temperature distribution.

[0087] Furthermore, in one embodiment of the present invention, after obtaining the first data feature and the second data feature through the above steps, the first data feature, the second data feature and the corresponding fault judgment result at each moment can be determined as a second data set.

[0088] S4, training the initial fault judgment model based on the second data set to obtain a target fault judgment model.

[0089] In one embodiment of the present invention, after the second data set is obtained through the above steps, the initial fault judgment model can be trained based on the second data set to obtain the target fault judgment model.

[0090] Specifically, in one embodiment of the present invention, the method of training the initial fault judgment model based on the second data set to obtain the target fault judgment model may include the following steps:

[0091] S41, dividing the second data set into a training data set and a test data set based on a preset ratio.

[0092] S42, using the training data set to train the initial fault judgment model to obtain a trained fault judgment model.

[0093] S43, using the test data set to evaluate the trained fault judgment model, and determining the target risk detection model based on the obtained evaluation results.

[0094] In one embodiment of the present invention, the second data set can be divided into a training data set and a test data set based on a preset ratio. For example, assuming that the preset ratio is 8:2, that is, 80% of the data in the first data set is randomly divided into a training data set and 20% of the data is divided into a test data set.

[0095] Further, in one embodiment of the present invention, the method of training the initial fault judgment model using the training data set to obtain the trained fault judgment model may include the following steps:

[0096] Step 1: Input the training data set into the initial fault judgment model to obtain the predicted fault judgment result;

[0097] Step 2: Based on the predicted fault judgment result and the actual fault result, the corresponding loss value is obtained through the loss function;

[0098] Step 3: Update the network parameters in the initial fault judgment model based on the loss value until the initial fault judgment model converges, or when the number of iterations of the network parameters reaches a preset number, a trained fault judgment model is obtained.

[0099] In one embodiment of the present invention, the initial fault judgment model may be an initial neural network model.

[0100] And, in one embodiment of the present invention, after the trained fault judgment model is obtained through the above steps, the test data in the test data set can be input into the trained fault judgment model to obtain the test result, and the evaluation result of the trained fault judgment model is determined based on the test result and the real result. In one embodiment of the present invention, the above evaluation result can be any one of the accuracy, recall rate, and F1 value.

[0101] Furthermore, in one embodiment of the present invention, if the above-mentioned evaluation result exceeds a preset threshold, the above-mentioned trained fault judgment model is determined as the target fault judgment model; otherwise, the network parameters of the model are continuously adjusted with the goal of minimizing the prediction error until the target fault judgment model is obtained.

[0102] In one embodiment of the present invention, by training the initial fault judgment model with the second data set, it is possible to perform fault judgment and testing on the wind turbine without interrupting the operation of the wind turbine, thereby predicting its performance and potential fault points more timely and accurately, and more accurately identifying wind turbine faults, reducing the possibility of misjudgment and missed judgment.

[0103] S5, collecting real-time data of the wind turbine generator set, performing fault judgment on the real-time data through a target fault judgment model, and obtaining a fault judgment result of the wind turbine generator set.

[0104] In one embodiment of the present invention, the real-time data may include vibration data and temperature data of the key components. Also, in one embodiment of the present invention, after collecting the real-time data of the wind turbine generator set, data features in the real-time data may be extracted, and the extracted data features may be input into the target fault judgment model to obtain the fault judgment result of the wind turbine generator set.

[0105] Furthermore, in one embodiment of the present invention, if the fault judgment result of the above-mentioned wind turbine is a fault, the fault type in the fault judgment result is output and the fault alarm mechanism is triggered, and the location of the fault is located based on the real-time data of the wind turbine, and the positioning result is output.

[0106] Furthermore, in one embodiment of the present invention, a method for locating the location of a fault based on real-time data of a wind turbine generator system and outputting the positioning result may include: determining the specific location of the fault through manual experience by comparing vibration data and temperature data collected at different locations, and determining the cause and severity of the fault through manual experience in combination with the changing trend of the temperature data, and outputting the specific location, cause and severity of the fault.

[0107] In one embodiment of the present invention, by inputting real-time data into the target fault judgment model, it is possible to replace manual periodic inspections, realize real-time monitoring of the status of wind turbines, and issue early warnings in time when faults occur, thereby avoiding the expansion of faults or causing greater losses, improving its operational reliability and stability, and reducing maintenance costs.

[0108] The embodiment of the present invention proposes a fault judgment method for a wind turbine, performs three-dimensional modeling on the wind turbine, and constructs a digital twin model of the wind turbine; obtains historical data of the wind turbine, and processes the historical data to obtain a first data set; obtains a second data set based on the first data set and the digital twin model; trains an initial fault judgment model based on the second data set to obtain a target fault judgment model; collects real-time data of the wind turbine, and performs fault judgment on the real-time data through the target fault judgment model to obtain a fault judgment result of the wind turbine. Therefore, the present invention performs fault judgment on the real-time data of the wind turbine through the target fault judgment model, which can replace manual regular inspections, realize real-time monitoring of the status of the wind turbine, and issue an early warning in time when a fault occurs, thereby improving the efficiency of fault judgment and the accuracy of fault prediction, avoiding the expansion of faults or causing greater losses, improving the operating reliability and stability of the wind turbine, and reducing maintenance costs.

[0109] In order to implement the above embodiment, Figure 2As shown, this embodiment also provides a fault judgment device 10 for a wind turbine generator set, which includes a modeling module 100, a first data processing module 200, a second data processing module 300, a training module 400, and a fault judgment module 500;

[0110] A modeling module 100 is used to perform three-dimensional modeling on the wind turbine and construct a digital twin model of the wind turbine;

[0111] A first data processing module 200 is used to obtain historical data of the wind turbine generator set and process the historical data to obtain a first data set;

[0112] A second data processing module 300, configured to obtain a second data set based on the first data set and the digital twin model;

[0113] A training module 400 is used to train the initial fault judgment model based on the second data set to obtain a target fault judgment model;

[0114] The fault judgment module 500 is used to collect real-time data of the wind turbine generator set, perform fault judgment on the real-time data through a target fault judgment model, and obtain a fault judgment result of the wind turbine generator set.

[0115] In one embodiment of the present invention, the modeling module 100 is specifically used for:

[0116] Obtain wind turbine data and operation data for wind turbine modeling;

[0117] Perform geometric modeling on the wind turbine generator set to obtain a geometric model of the wind turbine generator set;

[0118] Build a digital twin model of the wind turbine based on wind turbine data, operation data and geometric model.

[0119] Further, in one embodiment of the present invention, the historical data of the wind turbine generator set includes historical vibration data and historical temperature data; the first data processing module 200 is specifically used for:

[0120] Processing the historical vibration data to obtain a third data set;

[0121] Processing the historical temperature data to obtain a fourth data set;

[0122] Based on the third data set and the fourth data set, a first data set is obtained.

[0123] Furthermore, in one embodiment of the present invention, the second data processing module 300 is specifically used to:

[0124] Mapping the first data set to the digital twin model to determine a fault judgment result corresponding to each data in the first data set;

[0125] Extracting data features of data in the first data set;

[0126] Based on the data features and the corresponding fault judgment results, a second data set is obtained.

[0127] Furthermore, in one embodiment of the present invention, the training module 400 is specifically:

[0128] Based on a preset ratio, the second data set is divided into a training data set and a test data set;

[0129] Using the training data set to train the initial fault judgment model, to obtain a trained fault judgment model;

[0130] The trained fault judgment model is evaluated using the test data set, and the target risk detection model is determined based on the obtained evaluation results.

[0131] In one embodiment of the present invention, the training module 400 is further used for:

[0132] Input the training data set into the initial fault judgment model to obtain the predicted fault judgment result;

[0133] Based on the predicted fault judgment results and the actual fault results, the corresponding loss value is obtained through the loss function;

[0134] The network parameters in the initial fault judgment model are updated based on the loss value until the initial fault judgment model converges, or when the number of iterations of the network parameters reaches a preset number, a trained fault judgment model is obtained.

[0135] In one embodiment of the present invention, the above device is further used for:

[0136] If the fault judgment result of the wind turbine is a fault, the fault type in the fault judgment result is output and the fault alarm mechanism is triggered;

[0137] The fault location is located based on the real-time data of the wind turbine and the location result is output.

[0138] According to the fault judgment device of the wind turbine set of the embodiment of the present invention, the wind turbine set is three-dimensionally modeled to construct a digital twin model of the wind turbine set; the historical data of the wind turbine set is obtained, and the historical data is processed to obtain a first data set; based on the first data set and the digital twin model, a second data set is obtained; based on the second data set, the initial fault judgment model is trained to obtain a target fault judgment model; the real-time data of the wind turbine set is collected, and the real-time data is judged by the target fault judgment model to obtain the fault judgment result of the wind turbine set. Therefore, the present invention judges the real-time data of the wind turbine set by the target fault judgment model, which can replace manual regular inspections, realize real-time monitoring of the status of the wind turbine set, and issue an early warning in time when a fault occurs, thereby improving the efficiency of fault judgment and the accuracy of fault prediction, avoiding the expansion of faults or causing greater losses, improving the operational reliability and stability of the wind turbine set, and reducing maintenance costs.

[0139] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0140] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

Claims

1. A method for determining a fault of a wind turbine generator system, characterized in that: The method comprises: Performing three-dimensional modeling on the wind turbine generator set and constructing a digital twin model of the wind turbine generator set; Acquiring historical data of the wind turbine generator set, and performing data processing on the historical data to obtain a first data set; Based on the first data set and the digital twin model, obtain a second data set; Training an initial fault judgment model based on the second data set to obtain a target fault judgment model; Real-time data of the wind turbine generator set is collected, and fault judgment is performed on the real-time data using the target fault judgment model to obtain a fault judgment result of the wind turbine generator set.

2. The method according to claim 1, characterized in that The three-dimensional modeling of the wind turbine generator set and the construction of the digital twin model of the wind turbine generator set include: Obtaining wind turbine data and operation data of the wind turbine modeling; Performing geometric modeling on the wind turbine generator set to obtain a geometric model of the wind turbine generator set; Based on the wind turbine data, operation data and geometric model, a digital twin model of the wind turbine group is constructed.

3. The method according to claim 1, characterized in that The historical data of the wind turbine generator set includes historical vibration data and historical temperature data. The acquiring the historical data of the wind turbine generator set and performing data processing on the historical data to obtain a first data set includes: Processing the historical vibration data to obtain a third data set; Processing the historical temperature data to obtain a fourth data set; The first data set is obtained based on the third data set and the fourth data set.

4. The method according to claim 1, characterized in that: The obtaining of a second data set based on the first data set and the digital twin model comprises: Mapping the first data set to the digital twin model, and determining a fault judgment result corresponding to each data in the first data set; Extracting data features of the data in the first data set; Based on the data features and the corresponding fault judgment results, a second data set is obtained.

5. The method according to claim 1, characterized in that The training of the initial fault judgment model based on the second data set to obtain a target fault judgment model includes: Based on a preset ratio, dividing the second data set into a training data set and a test data set; Using the training data set to train the initial fault judgment model to obtain a trained fault judgment model; The trained fault judgment model is evaluated using the test data set, and a target risk detection model is determined based on the obtained evaluation result.

6. The method according to claim 5, characterized in that The method of training the initial fault judgment model using the training data set to obtain a trained fault judgment model includes: Inputting the training data set into the initial fault judgment model to obtain a predicted fault judgment result; Based on the predicted fault judgment result and the actual fault result, a corresponding loss value is obtained through a loss function; The network parameters in the initial fault judgment model are updated based on the loss value until the initial fault judgment model converges, or when the number of iterations of the network parameters reaches a preset number, a trained fault judgment model is obtained.

7. The method according to claim 1, characterized in that The method further comprises: If the fault judgment result of the wind turbine generator set is a fault, outputting the fault type in the fault judgment result and triggering a fault alarm mechanism; The location where the fault occurs is located based on the real-time data of the wind turbine generator set, and the location result is output.

8. A fault judgment device for a wind turbine generator set, characterized in that: The device comprises: A modeling module, used to perform three-dimensional modeling of a wind turbine and construct a digital twin model of the wind turbine; A first data processing module, used for acquiring historical data of the wind turbine generator set, and performing data processing on the historical data to obtain a first data set; A second data processing module, configured to obtain a second data set based on the first data set and the digital twin model; A training module, used for training the initial fault judgment model based on the second data set to obtain a target fault judgment model; The fault judgment module is used to collect real-time data of the wind turbine generator set, perform fault judgment on the real-time data through the target fault judgment model, and obtain a fault judgment result of the wind turbine generator set.

9. An electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

10. A computer storage medium, wherein: The computer storage medium stores computer executable instructions; after the computer executable instructions are executed by the processor, the method according to any one of claims 1 to 7 can be implemented.